Article

Deconditioning among hospitalized older adults contributes to significant decline in posthospitalization functional ability, physical performance, and physical activity.^1-10 Previous hospital-to-home interventions have targeted improving function and physical activity, including recent programs leveraging home telehealth as a feasible and potentially effective mode of delivering in-home exercise and rehabilitation.^11-14 However, pilot interventions have shown mixed effectiveness.^11,12,14 This study expands on a previously published intervention describing a pilot home telehealth program for veterans posthospital discharge that demonstrated significant 6-month improvement in physical activity as well as trends in physical function improvement, including among those with cognitive impairment.¹⁵ Factors that contribute to improved outcomes are the focus of the present study.

Key factors underlying the complexity of hospital-to-home transitions include hospitalization elements (ie, reason for admission and length of stay), associated posthospital syndromes (ie, postdischarge falls, medication changes, cognitive impairment, and pain), and postdischarge health care application (ie, physical therapy and hospital readmission).^16-18 These factors may be associated with postdischarge functional ability, physical performance, and physical activity, but their direct influence on intervention outcomes is unclear (Figure 1).^5,7,9,16-20 The objective of this study was to examine the influence of hospitalization, posthospital syndrome, and postdischarge health care application factors on outcomes of a US Department of Veterans Affairs (VA) Video Connect (VVC) intervention to enhance function and physical activity in older adults posthospital discharge.

Methods

The previous analysis reported on patient characteristics, program feasibility, and preliminary outcomes.^13,15 The current study reports on relationships between hospitalization, posthospital syndrome, and postdischarge health care application factors and change in key outcomes, namely postdischarge self-reported functional ability, physical performance, and physical activity from baseline to endpoint.

Participants provided written informed consent. The protocol and consent forms were approved by the VA Ann Arbor Healthcare System (VAAAHS) Research and Development Committee, and the project was registered on clinicaltrials.gov (NCT04045054).

Intervention

The pilot program targeted older adults following recent hospital discharge from VAAAHS. Participants were eligible if they were aged ≥ 50 years, had been discharged following an inpatient stay in the past 1 to 2 weeks, evaluated by physical therapy during hospitalization with stated rehabilitation goals on discharge, and followed by a VAAAHS primary care physician. Participants were either recruited during hospital admission or shortly after discharge.¹³

An experienced physical activity trainer (PAT) supported the progression of participants’ rehabilitation goals via a home exercise program and coached the patient and caregiver to optimize functional ability, physical performance, and physical activity. The PAT was a nonlicensed research assistant with extensive experience in applying standard physical activity enhancement protocols (eg, increased walking) to older adults with comorbidities. Participation in the program lasted about 6 months. Initiation of the PAT program was delayed if the patient was already receiving postdischarge home-based or outpatient physical therapy. The PAT contacted the patient weekly via VVC for the first 6 weeks, then monthly for a total of 6 months. Each contact included information on optimal walking form, injury prevention, program progression, and ways to incorporate sit-to-stand transitions, nonsitting behavior, and walking into daily routines. The initial VVC contact lasted about 60 minutes and subsequent sessions lasted about 30 minutes.¹³

Demographic characteristics were self-reported by participants and included age, sex, race, years of education, and marital status. Clinical characteristics were obtained from each participant’s electronic health record (EHR), including copay status, index hospitalization length of stay, admission diagnosis, and postsurgery status (postsurgery vs nonpostsurgery). Intervention adherence was tracked as the number of PAT sessions attended.

Posthospital Syndrome Factors

Participant falls (categorized as those who reported a fall vs those who did not) and medication changes (number of changes reported, including new medication, discontinued medication, dose changes, medication changes, or changes in medication schedule) were reported by participants or caregivers during each VVC contact. Participants completed the Montreal Cognitive Assessment (MoCA) at baseline, and were dichotomized into 2 groups: no cognitive impairment (MoCA score ≥ 26) and mild to moderate cognitive impairment (MoCA score 10-25).^13,21

Participants rated how much pain interfered with their normal daily activities since the previous VVC session on a 5-point Likert scale (1, not at all; to 5, extremely).²² Similar to prior research, participants were placed into 2 groups based on their mean pain interference score (individuals with scores from 1.0 to 2.0 in 1 group, and individuals with > 2.0 in another).^23-25 Participants were separated into a no or mild pain interference group and a moderate to severe pain interference group. Hospital readmissions (VA and non-VA) and postdischarge physical therapy outcomes were obtained from the participant’s EHR, including primary care visits.

Outcomes

Outcomes were collected at baseline (posthospital discharge) and 6 months postenrollment.

Self-Reported Functional Ability. This measure is provided by participants or caregivers and measured by the Katz Index of Independence in Activities of Daily Living (ADL), Lawton and Brody Instrumental ADL Scale (IADL), Nagi Disability Model, and Rosow-Breslau Scale. The Katz ADL assesses the ability to complete 6 self-care activities and awards 1 point for independence and 0 if the individual is dependent (total score range, 0-6).²⁶ The Lawton and Brody IADL measures an individual’s independence in 8 instrumental ADLs; it awards 1 point for independence and 0 if the individual is dependent (total score range, 0-8).²⁷ The Nagi Disability Model evaluates an individual’s difficulty performing 5 tasks (total score range, 0-5) and tallies the number of items with a response other than “no difficulty at all” (higher total score indicates greater difficulty). ²⁸ The Rosow-Breslau Scale is a 3-item measure of mobility disability; individual responses are 0 (no help) and 1 (requires help or unable); higher total score (range, 0-3) indicates greater disability.²⁹

Physical Performance. Measured using the Short Physical Performance Battery (SPPB), which evaluates standing balance, sit to stand, and walking performance. Scores range from 0 to 4 on the balance, gait speed, and chair stand tests, for a total composite score between 0 and 12 (higher score indicates better performance).³⁰

Physical Activity. Measured using actigraphy, namely a physical activity monitor adherent to the thigh (activ-PAL3TM, PAL Technologies Ltd., Glasgow, UK).³¹ Participants were instructed to wear the activPal for ≥ 1 week. Participants with a minimum of 5 days of wear were included in this analysis.

Data Analyses

Analyses were performed using SPSS software version 29.0.³² Continuous variables were summarized using mean (SD) or median and IQR using the weighted average method; categorical variables were summarized using frequencies and percentages. Baseline scores on outcome variables were compared by categorical hospitalization, posthospital syndrome, and postdischarge health care application factor variables using Mann-Whitney U tests. The differences between outcome variables from baseline to endpoint were then calculated to produce change scores. Relationships between the number of PAT sessions attended and baseline outcomes and outcome change scores were estimated using Spearman correlations. Relationships between categorical factors (hospitalization, posthospital syndrome, and postdischarge health care application) and outcome variable change scores (which were normally distributed) were examined using Mann-Whitney U tests. Relationships with continuous hospitalization (length of stay) and posthospital syndrome factors (medication changes) were estimated using Spearman correlations. Effect sizes (ES) were estimated with Cohen d; small (d = 0.2), medium (d = 0.5), or large (d ≥ 0.8). Missing data were handled using pairwise deletion.³³ Therefore, sample sizes were reported for each analysis. For all statistical tests, P < .05 was considered significant.

Results

Twenty-four individuals completed the pilot intervention.¹⁵ Mean (SD) age was 73.6 (8.1) years (range, 64-93 years) and participants were predominantly White males (Table 1). Eight participants had a high school education only and 13 had more than a high school education. Diagnoses at admission included 9 patients with orthopedic/musculoskeletal conditions (6 were for joint replacement), 6 patients with vascular/pulmonary conditions, and 4 with gastrointestinal/renal/urological conditions. Of the 11 postsurgery participants, 7 were orthopedic, 4 were gastrointestinal, and 1 was peripheral vascular.

Baseline outcome scores did not differ significantly between groups, except individuals with moderate to severe pain interference reported a significantly lower IADL score (median [IQR] 4 [2-7]) than individuals with mild or moderate pain interference (median [IQR] 8 [7-8]; P = .02) (Table 2). The mean (SD) number of PAT sessions attended was 9.3 (3.7) (range, 3-19). There were no significant relationships between number of sessions attended and any baseline outcome variables or outcome change scores.

Hospitalization Factors

Participants who were postsurgery tended to have greater improvement than individuals who were nonpostsurgery in ADLs (median [IQR] 0 [0-1.5]; ES, 0.6; P = .10) and SPPB (median [IQR] 2 [1.5-9]; ES, 0.9; P = .07), but the improvements were not statistically significant (Table 3). Mean (SD) length of stay of the index hospitalization was 6.7 (6.1) days. Longer length of stay was significantly correlated with an increase in Nagi score (ρ, 0.45; 95% CI, 0.01-0.75). There were no other significant or trending relationships between length of stay and outcome variables.

Posthospital Syndrome Factors

The 16 participants with mild to moderate cognitive impairment had less improvement in ADLs (median [IQR] 0 [0-1]) than the 8 participants with no impairment (median [IQR] 0 [-0.75 to 0]; ES, -1.1; P = .04). Change in outcome variables from baseline to endpoint did not significantly differ between the 8 patients who reported a fall compared with the 13 who did not, nor were any trends observed. Change in outcome variables from baseline to endpoint also did not significantly differ between the 8 participants who reported no or mild pain interference compared with the 10 patients with moderate to severe pain interference, nor were any trends observed. Mean (SD) number of medication changes was 2.5 (1.6). Higher number of medication changes was significantly correlated with a decrease in Rosow-Breslau score (ρ, -0.47; 95% CI, -0.76 to -0.02). There were no other significant or trending relationships between number of medication changes and outcome variables.

Postdischarge Health Care Application Factors

The 16 participants who attended posthospital physical therapy trended towards less improvement in IADLs (median [IQR] 0 [-0.5 to 1.5]; ES, -0.7; P = .11) and SPPB (median [IQR] 2 [-3.0 to 4.5]; ES, -0.5; P = .15) than the 8 patients with no postdischarge physical therapy. Eleven participants were readmitted, while 13 had no readmissions in their medical records between baseline and endpoint. Participants with ≥ 1 readmission experienced a greater increase in Rosow-Breslau score (median [IQR] 0 [-0.5 to 1.0]) than those not readmitted (median [IQR] 0 [-1.25 to 0.25]; ES, 1.0; P = .03). Borderline greater improvement in number of steps was found in those not readmitted (median [IQR] 3365.6 [274.4-7710.9]) compared with those readmitted (median [IQR] 319.9 [-136.1 to 774.5]; ES, -1.3; P = .05). Patients who were readmitted also tended to have lower and not statistically significant improvements in SPPB (median [IQR] 1 [-4.0 to 5.3]) compared with those not readmitted (median [IQR] 2 [0.3-3.8]; ES, -0.5; P = .17) (Table 3).

Discussion

This study examined the association between hospitalization, posthospital syndrome, and postdischarge health care use in patients undergoing a VVC-based intervention following hospital discharge. Participants who had no or mild cognitive impairment, no readmissions, higher medication changes, and a shorter hospital length of stay tended to experience lower disability, including in mobility and ADLs. This suggests individuals who are less clinically complex may be more likely to benefit from this type of virtual rehabilitation program. These findings are consistent with clinical experiences; home-based programs to improve physical activity posthospital discharge can be challenging for those who were medically ill (and did not undergo a specific surgical procedure), cognitively impaired, and become acutely ill and trigger hospital readmission. ¹⁵ For example, the sample in this study had higher rates of falls, pain, and readmissions compared to previous research.^2,3,34-39

The importance of posthospital syndrome in the context of recovery of function and health at home following hospitalization is well documented.^16-18 The potential impact of posthospital syndrome on physical activity-focused interventions is less understood. In our analysis, participants with mild or moderate cognitive impairment tended to become more dependent in their ADLs, while those with no cognitive impairment tended to become more independent in their ADLs. This functional decline over time is perhaps expected in persons with cognitive impairment, but the significant difference with a large ES warrants further consideration on how to tailor interventions to better promote functional recovery in these individuals.^40,41 While some cognitive decline may not be preventable, this finding supports the need to promote healthy cognitive aging, identify declines in cognition, and work to mitigate additional decline. Programs specifically designed to promote function and physical activity in older adults with cognitive impairment are needed, especially during care transitions.^41-43

While participants reported that falls and pain interference did not have a significant impact on change in outcomes between baseline and endpoint, these areas need further investigation. Falls and pain have been associated with function and physical activity in older adults.^42-46 Pain is common, yet underappreciated during older adult hospital-to-home transitions.^11,12,45,46 There is a need for more comprehensive assessment of pain (including pain intensity) and qualitative research.

Hospitalization and postdischarge health care application factors may have a significant impact on home-telehealth physical activity intervention success. Individuals who were postsurgery tended to have greater improvements in ADLs and physical performance. Most postsurgery participants had joint replacement surgery. Postsurgery status may not be modifiable, but it is important to note expected differences in recovery between medical and surgical admissions and the need to tailor care based on admission diagnosis. Those with a longer length of hospital stay may be considered at higher risk of suboptimal outcomes postdischarge, which indicates an opportunity for targeting resources and support, in addition to efforts of reducing length of stay where possible.⁴⁷

Readmissions were significantly related to a change in Rosow-Breslau mobility disability score. This may indicate the detrimental impact a readmission can have on increasing mobility and physical activity postdischarge, or the potential of this pilot program to impact readmissions by increasing mobility and physical activity, contrary to prior physical exercise interventions.^5,7,9,48 With 5% to 79% of readmissions considered preventable, continued efforts and program dissemination and implementation to address preventable readmissions are warranted.⁴⁹ Individuals with postdischarge physical therapy (prior to beginning the pilot program) tended to demonstrate less improvement in disability and physical performance. This relationship needs further investigation; the 2 groups did not appear to have significant differences at baseline, albeit with a small sample size. It is possible they experienced initial improvements with postdischarge physical therapy and plateaued or had little further reserve to improve upon entering the VVC program.

Strengths and Limitations

This pilot program provided evaluative data on the use of VVC to enhance function and physical activity in older adults posthospital discharge. It included individual (eg, fall, pain, cognitive impairment) and health service (eg, readmission, physical therapy) level factors as predictors of function and physical activity posthospitalization.^5,7,9,15-19

The results of this pilot project stem from a small sample lacking diversity in terms of race, ethnicity, and sex. There was some variation in baseline and endpoints between participants, and when hospitalization, posthospital syndrome, and postdischarge health care application factors were collected. The majority of participants were recruited within a month postdischarge, and the program lasted about 6 months. Data collection was attempted at regular PAT contacts, but there was some variation in when visits occurred based on participant availability and preference. Some participants had missing data, which was handled using pairwise deletion.³³ Larger studies are needed to confirm the findings of this study, particularly the trends that did not reach statistical significance. Home health services other than physical therapy (eg, nursing, occupational therapy) were not fully accounted for and should be considered in future research.

Conclusions

In patients undergoing a 6-month pilot VVC-based physical activity intervention posthospital discharge, improvements in mobility and disability were most likely in those who had no cognitive impairment and were not readmitted. Larger sample and qualitative investigations are necessary to optimize outcomes for patients who meet these clinical profiles.

Deconditioning among hospitalized older adults contributes to significant decline in posthospitalization functional ability, physical performance, and physical activity.^1-10 Previous hospital-to-home interventions have targeted improving function and physical activity, including recent programs leveraging home telehealth as a feasible and potentially effective mode of delivering in-home exercise and rehabilitation.^11-14 However, pilot interventions have shown mixed effectiveness.^11,12,14 This study expands on a previously published intervention describing a pilot home telehealth program for veterans posthospital discharge that demonstrated significant 6-month improvement in physical activity as well as trends in physical function improvement, including among those with cognitive impairment.¹⁵ Factors that contribute to improved outcomes are the focus of the present study.

Key factors underlying the complexity of hospital-to-home transitions include hospitalization elements (ie, reason for admission and length of stay), associated posthospital syndromes (ie, postdischarge falls, medication changes, cognitive impairment, and pain), and postdischarge health care application (ie, physical therapy and hospital readmission).^16-18 These factors may be associated with postdischarge functional ability, physical performance, and physical activity, but their direct influence on intervention outcomes is unclear (Figure 1).^5,7,9,16-20 The objective of this study was to examine the influence of hospitalization, posthospital syndrome, and postdischarge health care application factors on outcomes of a US Department of Veterans Affairs (VA) Video Connect (VVC) intervention to enhance function and physical activity in older adults posthospital discharge.

Methods

The previous analysis reported on patient characteristics, program feasibility, and preliminary outcomes.^13,15 The current study reports on relationships between hospitalization, posthospital syndrome, and postdischarge health care application factors and change in key outcomes, namely postdischarge self-reported functional ability, physical performance, and physical activity from baseline to endpoint.

Participants provided written informed consent. The protocol and consent forms were approved by the VA Ann Arbor Healthcare System (VAAAHS) Research and Development Committee, and the project was registered on clinicaltrials.gov (NCT04045054).

Intervention

The pilot program targeted older adults following recent hospital discharge from VAAAHS. Participants were eligible if they were aged ≥ 50 years, had been discharged following an inpatient stay in the past 1 to 2 weeks, evaluated by physical therapy during hospitalization with stated rehabilitation goals on discharge, and followed by a VAAAHS primary care physician. Participants were either recruited during hospital admission or shortly after discharge.¹³

An experienced physical activity trainer (PAT) supported the progression of participants’ rehabilitation goals via a home exercise program and coached the patient and caregiver to optimize functional ability, physical performance, and physical activity. The PAT was a nonlicensed research assistant with extensive experience in applying standard physical activity enhancement protocols (eg, increased walking) to older adults with comorbidities. Participation in the program lasted about 6 months. Initiation of the PAT program was delayed if the patient was already receiving postdischarge home-based or outpatient physical therapy. The PAT contacted the patient weekly via VVC for the first 6 weeks, then monthly for a total of 6 months. Each contact included information on optimal walking form, injury prevention, program progression, and ways to incorporate sit-to-stand transitions, nonsitting behavior, and walking into daily routines. The initial VVC contact lasted about 60 minutes and subsequent sessions lasted about 30 minutes.¹³

Demographic characteristics were self-reported by participants and included age, sex, race, years of education, and marital status. Clinical characteristics were obtained from each participant’s electronic health record (EHR), including copay status, index hospitalization length of stay, admission diagnosis, and postsurgery status (postsurgery vs nonpostsurgery). Intervention adherence was tracked as the number of PAT sessions attended.

Posthospital Syndrome Factors

Participant falls (categorized as those who reported a fall vs those who did not) and medication changes (number of changes reported, including new medication, discontinued medication, dose changes, medication changes, or changes in medication schedule) were reported by participants or caregivers during each VVC contact. Participants completed the Montreal Cognitive Assessment (MoCA) at baseline, and were dichotomized into 2 groups: no cognitive impairment (MoCA score ≥ 26) and mild to moderate cognitive impairment (MoCA score 10-25).^13,21

Participants rated how much pain interfered with their normal daily activities since the previous VVC session on a 5-point Likert scale (1, not at all; to 5, extremely).²² Similar to prior research, participants were placed into 2 groups based on their mean pain interference score (individuals with scores from 1.0 to 2.0 in 1 group, and individuals with > 2.0 in another).^23-25 Participants were separated into a no or mild pain interference group and a moderate to severe pain interference group. Hospital readmissions (VA and non-VA) and postdischarge physical therapy outcomes were obtained from the participant’s EHR, including primary care visits.

Outcomes

Outcomes were collected at baseline (posthospital discharge) and 6 months postenrollment.

Self-Reported Functional Ability. This measure is provided by participants or caregivers and measured by the Katz Index of Independence in Activities of Daily Living (ADL), Lawton and Brody Instrumental ADL Scale (IADL), Nagi Disability Model, and Rosow-Breslau Scale. The Katz ADL assesses the ability to complete 6 self-care activities and awards 1 point for independence and 0 if the individual is dependent (total score range, 0-6).²⁶ The Lawton and Brody IADL measures an individual’s independence in 8 instrumental ADLs; it awards 1 point for independence and 0 if the individual is dependent (total score range, 0-8).²⁷ The Nagi Disability Model evaluates an individual’s difficulty performing 5 tasks (total score range, 0-5) and tallies the number of items with a response other than “no difficulty at all” (higher total score indicates greater difficulty). ²⁸ The Rosow-Breslau Scale is a 3-item measure of mobility disability; individual responses are 0 (no help) and 1 (requires help or unable); higher total score (range, 0-3) indicates greater disability.²⁹

Physical Performance. Measured using the Short Physical Performance Battery (SPPB), which evaluates standing balance, sit to stand, and walking performance. Scores range from 0 to 4 on the balance, gait speed, and chair stand tests, for a total composite score between 0 and 12 (higher score indicates better performance).³⁰

Physical Activity. Measured using actigraphy, namely a physical activity monitor adherent to the thigh (activ-PAL3TM, PAL Technologies Ltd., Glasgow, UK).³¹ Participants were instructed to wear the activPal for ≥ 1 week. Participants with a minimum of 5 days of wear were included in this analysis.

Data Analyses

Analyses were performed using SPSS software version 29.0.³² Continuous variables were summarized using mean (SD) or median and IQR using the weighted average method; categorical variables were summarized using frequencies and percentages. Baseline scores on outcome variables were compared by categorical hospitalization, posthospital syndrome, and postdischarge health care application factor variables using Mann-Whitney U tests. The differences between outcome variables from baseline to endpoint were then calculated to produce change scores. Relationships between the number of PAT sessions attended and baseline outcomes and outcome change scores were estimated using Spearman correlations. Relationships between categorical factors (hospitalization, posthospital syndrome, and postdischarge health care application) and outcome variable change scores (which were normally distributed) were examined using Mann-Whitney U tests. Relationships with continuous hospitalization (length of stay) and posthospital syndrome factors (medication changes) were estimated using Spearman correlations. Effect sizes (ES) were estimated with Cohen d; small (d = 0.2), medium (d = 0.5), or large (d ≥ 0.8). Missing data were handled using pairwise deletion.³³ Therefore, sample sizes were reported for each analysis. For all statistical tests, P < .05 was considered significant.

Results

Twenty-four individuals completed the pilot intervention.¹⁵ Mean (SD) age was 73.6 (8.1) years (range, 64-93 years) and participants were predominantly White males (Table 1). Eight participants had a high school education only and 13 had more than a high school education. Diagnoses at admission included 9 patients with orthopedic/musculoskeletal conditions (6 were for joint replacement), 6 patients with vascular/pulmonary conditions, and 4 with gastrointestinal/renal/urological conditions. Of the 11 postsurgery participants, 7 were orthopedic, 4 were gastrointestinal, and 1 was peripheral vascular.

Baseline outcome scores did not differ significantly between groups, except individuals with moderate to severe pain interference reported a significantly lower IADL score (median [IQR] 4 [2-7]) than individuals with mild or moderate pain interference (median [IQR] 8 [7-8]; P = .02) (Table 2). The mean (SD) number of PAT sessions attended was 9.3 (3.7) (range, 3-19). There were no significant relationships between number of sessions attended and any baseline outcome variables or outcome change scores.

Hospitalization Factors

Participants who were postsurgery tended to have greater improvement than individuals who were nonpostsurgery in ADLs (median [IQR] 0 [0-1.5]; ES, 0.6; P = .10) and SPPB (median [IQR] 2 [1.5-9]; ES, 0.9; P = .07), but the improvements were not statistically significant (Table 3). Mean (SD) length of stay of the index hospitalization was 6.7 (6.1) days. Longer length of stay was significantly correlated with an increase in Nagi score (ρ, 0.45; 95% CI, 0.01-0.75). There were no other significant or trending relationships between length of stay and outcome variables.

Posthospital Syndrome Factors

The 16 participants with mild to moderate cognitive impairment had less improvement in ADLs (median [IQR] 0 [0-1]) than the 8 participants with no impairment (median [IQR] 0 [-0.75 to 0]; ES, -1.1; P = .04). Change in outcome variables from baseline to endpoint did not significantly differ between the 8 patients who reported a fall compared with the 13 who did not, nor were any trends observed. Change in outcome variables from baseline to endpoint also did not significantly differ between the 8 participants who reported no or mild pain interference compared with the 10 patients with moderate to severe pain interference, nor were any trends observed. Mean (SD) number of medication changes was 2.5 (1.6). Higher number of medication changes was significantly correlated with a decrease in Rosow-Breslau score (ρ, -0.47; 95% CI, -0.76 to -0.02). There were no other significant or trending relationships between number of medication changes and outcome variables.

Postdischarge Health Care Application Factors

The 16 participants who attended posthospital physical therapy trended towards less improvement in IADLs (median [IQR] 0 [-0.5 to 1.5]; ES, -0.7; P = .11) and SPPB (median [IQR] 2 [-3.0 to 4.5]; ES, -0.5; P = .15) than the 8 patients with no postdischarge physical therapy. Eleven participants were readmitted, while 13 had no readmissions in their medical records between baseline and endpoint. Participants with ≥ 1 readmission experienced a greater increase in Rosow-Breslau score (median [IQR] 0 [-0.5 to 1.0]) than those not readmitted (median [IQR] 0 [-1.25 to 0.25]; ES, 1.0; P = .03). Borderline greater improvement in number of steps was found in those not readmitted (median [IQR] 3365.6 [274.4-7710.9]) compared with those readmitted (median [IQR] 319.9 [-136.1 to 774.5]; ES, -1.3; P = .05). Patients who were readmitted also tended to have lower and not statistically significant improvements in SPPB (median [IQR] 1 [-4.0 to 5.3]) compared with those not readmitted (median [IQR] 2 [0.3-3.8]; ES, -0.5; P = .17) (Table 3).

Discussion

This study examined the association between hospitalization, posthospital syndrome, and postdischarge health care use in patients undergoing a VVC-based intervention following hospital discharge. Participants who had no or mild cognitive impairment, no readmissions, higher medication changes, and a shorter hospital length of stay tended to experience lower disability, including in mobility and ADLs. This suggests individuals who are less clinically complex may be more likely to benefit from this type of virtual rehabilitation program. These findings are consistent with clinical experiences; home-based programs to improve physical activity posthospital discharge can be challenging for those who were medically ill (and did not undergo a specific surgical procedure), cognitively impaired, and become acutely ill and trigger hospital readmission. ¹⁵ For example, the sample in this study had higher rates of falls, pain, and readmissions compared to previous research.^2,3,34-39

The importance of posthospital syndrome in the context of recovery of function and health at home following hospitalization is well documented.^16-18 The potential impact of posthospital syndrome on physical activity-focused interventions is less understood. In our analysis, participants with mild or moderate cognitive impairment tended to become more dependent in their ADLs, while those with no cognitive impairment tended to become more independent in their ADLs. This functional decline over time is perhaps expected in persons with cognitive impairment, but the significant difference with a large ES warrants further consideration on how to tailor interventions to better promote functional recovery in these individuals.^40,41 While some cognitive decline may not be preventable, this finding supports the need to promote healthy cognitive aging, identify declines in cognition, and work to mitigate additional decline. Programs specifically designed to promote function and physical activity in older adults with cognitive impairment are needed, especially during care transitions.^41-43

While participants reported that falls and pain interference did not have a significant impact on change in outcomes between baseline and endpoint, these areas need further investigation. Falls and pain have been associated with function and physical activity in older adults.^42-46 Pain is common, yet underappreciated during older adult hospital-to-home transitions.^11,12,45,46 There is a need for more comprehensive assessment of pain (including pain intensity) and qualitative research.

Hospitalization and postdischarge health care application factors may have a significant impact on home-telehealth physical activity intervention success. Individuals who were postsurgery tended to have greater improvements in ADLs and physical performance. Most postsurgery participants had joint replacement surgery. Postsurgery status may not be modifiable, but it is important to note expected differences in recovery between medical and surgical admissions and the need to tailor care based on admission diagnosis. Those with a longer length of hospital stay may be considered at higher risk of suboptimal outcomes postdischarge, which indicates an opportunity for targeting resources and support, in addition to efforts of reducing length of stay where possible.⁴⁷

Readmissions were significantly related to a change in Rosow-Breslau mobility disability score. This may indicate the detrimental impact a readmission can have on increasing mobility and physical activity postdischarge, or the potential of this pilot program to impact readmissions by increasing mobility and physical activity, contrary to prior physical exercise interventions.^5,7,9,48 With 5% to 79% of readmissions considered preventable, continued efforts and program dissemination and implementation to address preventable readmissions are warranted.⁴⁹ Individuals with postdischarge physical therapy (prior to beginning the pilot program) tended to demonstrate less improvement in disability and physical performance. This relationship needs further investigation; the 2 groups did not appear to have significant differences at baseline, albeit with a small sample size. It is possible they experienced initial improvements with postdischarge physical therapy and plateaued or had little further reserve to improve upon entering the VVC program.

Strengths and Limitations

This pilot program provided evaluative data on the use of VVC to enhance function and physical activity in older adults posthospital discharge. It included individual (eg, fall, pain, cognitive impairment) and health service (eg, readmission, physical therapy) level factors as predictors of function and physical activity posthospitalization.^5,7,9,15-19

The results of this pilot project stem from a small sample lacking diversity in terms of race, ethnicity, and sex. There was some variation in baseline and endpoints between participants, and when hospitalization, posthospital syndrome, and postdischarge health care application factors were collected. The majority of participants were recruited within a month postdischarge, and the program lasted about 6 months. Data collection was attempted at regular PAT contacts, but there was some variation in when visits occurred based on participant availability and preference. Some participants had missing data, which was handled using pairwise deletion.³³ Larger studies are needed to confirm the findings of this study, particularly the trends that did not reach statistical significance. Home health services other than physical therapy (eg, nursing, occupational therapy) were not fully accounted for and should be considered in future research.

Conclusions

In patients undergoing a 6-month pilot VVC-based physical activity intervention posthospital discharge, improvements in mobility and disability were most likely in those who had no cognitive impairment and were not readmitted. Larger sample and qualitative investigations are necessary to optimize outcomes for patients who meet these clinical profiles.

Deconditioning among hospitalized older adults contributes to significant decline in posthospitalization functional ability, physical performance, and physical activity.^1-10 Previous hospital-to-home interventions have targeted improving function and physical activity, including recent programs leveraging home telehealth as a feasible and potentially effective mode of delivering in-home exercise and rehabilitation.^11-14 However, pilot interventions have shown mixed effectiveness.^11,12,14 This study expands on a previously published intervention describing a pilot home telehealth program for veterans posthospital discharge that demonstrated significant 6-month improvement in physical activity as well as trends in physical function improvement, including among those with cognitive impairment.¹⁵ Factors that contribute to improved outcomes are the focus of the present study.

Key factors underlying the complexity of hospital-to-home transitions include hospitalization elements (ie, reason for admission and length of stay), associated posthospital syndromes (ie, postdischarge falls, medication changes, cognitive impairment, and pain), and postdischarge health care application (ie, physical therapy and hospital readmission).^16-18 These factors may be associated with postdischarge functional ability, physical performance, and physical activity, but their direct influence on intervention outcomes is unclear (Figure 1).^5,7,9,16-20 The objective of this study was to examine the influence of hospitalization, posthospital syndrome, and postdischarge health care application factors on outcomes of a US Department of Veterans Affairs (VA) Video Connect (VVC) intervention to enhance function and physical activity in older adults posthospital discharge.

Methods

The previous analysis reported on patient characteristics, program feasibility, and preliminary outcomes.^13,15 The current study reports on relationships between hospitalization, posthospital syndrome, and postdischarge health care application factors and change in key outcomes, namely postdischarge self-reported functional ability, physical performance, and physical activity from baseline to endpoint.

Participants provided written informed consent. The protocol and consent forms were approved by the VA Ann Arbor Healthcare System (VAAAHS) Research and Development Committee, and the project was registered on clinicaltrials.gov (NCT04045054).

Intervention

The pilot program targeted older adults following recent hospital discharge from VAAAHS. Participants were eligible if they were aged ≥ 50 years, had been discharged following an inpatient stay in the past 1 to 2 weeks, evaluated by physical therapy during hospitalization with stated rehabilitation goals on discharge, and followed by a VAAAHS primary care physician. Participants were either recruited during hospital admission or shortly after discharge.¹³

An experienced physical activity trainer (PAT) supported the progression of participants’ rehabilitation goals via a home exercise program and coached the patient and caregiver to optimize functional ability, physical performance, and physical activity. The PAT was a nonlicensed research assistant with extensive experience in applying standard physical activity enhancement protocols (eg, increased walking) to older adults with comorbidities. Participation in the program lasted about 6 months. Initiation of the PAT program was delayed if the patient was already receiving postdischarge home-based or outpatient physical therapy. The PAT contacted the patient weekly via VVC for the first 6 weeks, then monthly for a total of 6 months. Each contact included information on optimal walking form, injury prevention, program progression, and ways to incorporate sit-to-stand transitions, nonsitting behavior, and walking into daily routines. The initial VVC contact lasted about 60 minutes and subsequent sessions lasted about 30 minutes.¹³

Demographic characteristics were self-reported by participants and included age, sex, race, years of education, and marital status. Clinical characteristics were obtained from each participant’s electronic health record (EHR), including copay status, index hospitalization length of stay, admission diagnosis, and postsurgery status (postsurgery vs nonpostsurgery). Intervention adherence was tracked as the number of PAT sessions attended.

Posthospital Syndrome Factors

Participant falls (categorized as those who reported a fall vs those who did not) and medication changes (number of changes reported, including new medication, discontinued medication, dose changes, medication changes, or changes in medication schedule) were reported by participants or caregivers during each VVC contact. Participants completed the Montreal Cognitive Assessment (MoCA) at baseline, and were dichotomized into 2 groups: no cognitive impairment (MoCA score ≥ 26) and mild to moderate cognitive impairment (MoCA score 10-25).^13,21

Participants rated how much pain interfered with their normal daily activities since the previous VVC session on a 5-point Likert scale (1, not at all; to 5, extremely).²² Similar to prior research, participants were placed into 2 groups based on their mean pain interference score (individuals with scores from 1.0 to 2.0 in 1 group, and individuals with > 2.0 in another).^23-25 Participants were separated into a no or mild pain interference group and a moderate to severe pain interference group. Hospital readmissions (VA and non-VA) and postdischarge physical therapy outcomes were obtained from the participant’s EHR, including primary care visits.

Outcomes

Outcomes were collected at baseline (posthospital discharge) and 6 months postenrollment.

Self-Reported Functional Ability. This measure is provided by participants or caregivers and measured by the Katz Index of Independence in Activities of Daily Living (ADL), Lawton and Brody Instrumental ADL Scale (IADL), Nagi Disability Model, and Rosow-Breslau Scale. The Katz ADL assesses the ability to complete 6 self-care activities and awards 1 point for independence and 0 if the individual is dependent (total score range, 0-6).²⁶ The Lawton and Brody IADL measures an individual’s independence in 8 instrumental ADLs; it awards 1 point for independence and 0 if the individual is dependent (total score range, 0-8).²⁷ The Nagi Disability Model evaluates an individual’s difficulty performing 5 tasks (total score range, 0-5) and tallies the number of items with a response other than “no difficulty at all” (higher total score indicates greater difficulty). ²⁸ The Rosow-Breslau Scale is a 3-item measure of mobility disability; individual responses are 0 (no help) and 1 (requires help or unable); higher total score (range, 0-3) indicates greater disability.²⁹

Physical Performance. Measured using the Short Physical Performance Battery (SPPB), which evaluates standing balance, sit to stand, and walking performance. Scores range from 0 to 4 on the balance, gait speed, and chair stand tests, for a total composite score between 0 and 12 (higher score indicates better performance).³⁰

Physical Activity. Measured using actigraphy, namely a physical activity monitor adherent to the thigh (activ-PAL3TM, PAL Technologies Ltd., Glasgow, UK).³¹ Participants were instructed to wear the activPal for ≥ 1 week. Participants with a minimum of 5 days of wear were included in this analysis.

Data Analyses

Analyses were performed using SPSS software version 29.0.³² Continuous variables were summarized using mean (SD) or median and IQR using the weighted average method; categorical variables were summarized using frequencies and percentages. Baseline scores on outcome variables were compared by categorical hospitalization, posthospital syndrome, and postdischarge health care application factor variables using Mann-Whitney U tests. The differences between outcome variables from baseline to endpoint were then calculated to produce change scores. Relationships between the number of PAT sessions attended and baseline outcomes and outcome change scores were estimated using Spearman correlations. Relationships between categorical factors (hospitalization, posthospital syndrome, and postdischarge health care application) and outcome variable change scores (which were normally distributed) were examined using Mann-Whitney U tests. Relationships with continuous hospitalization (length of stay) and posthospital syndrome factors (medication changes) were estimated using Spearman correlations. Effect sizes (ES) were estimated with Cohen d; small (d = 0.2), medium (d = 0.5), or large (d ≥ 0.8). Missing data were handled using pairwise deletion.³³ Therefore, sample sizes were reported for each analysis. For all statistical tests, P < .05 was considered significant.

Results

Twenty-four individuals completed the pilot intervention.¹⁵ Mean (SD) age was 73.6 (8.1) years (range, 64-93 years) and participants were predominantly White males (Table 1). Eight participants had a high school education only and 13 had more than a high school education. Diagnoses at admission included 9 patients with orthopedic/musculoskeletal conditions (6 were for joint replacement), 6 patients with vascular/pulmonary conditions, and 4 with gastrointestinal/renal/urological conditions. Of the 11 postsurgery participants, 7 were orthopedic, 4 were gastrointestinal, and 1 was peripheral vascular.

Baseline outcome scores did not differ significantly between groups, except individuals with moderate to severe pain interference reported a significantly lower IADL score (median [IQR] 4 [2-7]) than individuals with mild or moderate pain interference (median [IQR] 8 [7-8]; P = .02) (Table 2). The mean (SD) number of PAT sessions attended was 9.3 (3.7) (range, 3-19). There were no significant relationships between number of sessions attended and any baseline outcome variables or outcome change scores.

Hospitalization Factors

Participants who were postsurgery tended to have greater improvement than individuals who were nonpostsurgery in ADLs (median [IQR] 0 [0-1.5]; ES, 0.6; P = .10) and SPPB (median [IQR] 2 [1.5-9]; ES, 0.9; P = .07), but the improvements were not statistically significant (Table 3). Mean (SD) length of stay of the index hospitalization was 6.7 (6.1) days. Longer length of stay was significantly correlated with an increase in Nagi score (ρ, 0.45; 95% CI, 0.01-0.75). There were no other significant or trending relationships between length of stay and outcome variables.

Posthospital Syndrome Factors

The 16 participants with mild to moderate cognitive impairment had less improvement in ADLs (median [IQR] 0 [0-1]) than the 8 participants with no impairment (median [IQR] 0 [-0.75 to 0]; ES, -1.1; P = .04). Change in outcome variables from baseline to endpoint did not significantly differ between the 8 patients who reported a fall compared with the 13 who did not, nor were any trends observed. Change in outcome variables from baseline to endpoint also did not significantly differ between the 8 participants who reported no or mild pain interference compared with the 10 patients with moderate to severe pain interference, nor were any trends observed. Mean (SD) number of medication changes was 2.5 (1.6). Higher number of medication changes was significantly correlated with a decrease in Rosow-Breslau score (ρ, -0.47; 95% CI, -0.76 to -0.02). There were no other significant or trending relationships between number of medication changes and outcome variables.

Postdischarge Health Care Application Factors

The 16 participants who attended posthospital physical therapy trended towards less improvement in IADLs (median [IQR] 0 [-0.5 to 1.5]; ES, -0.7; P = .11) and SPPB (median [IQR] 2 [-3.0 to 4.5]; ES, -0.5; P = .15) than the 8 patients with no postdischarge physical therapy. Eleven participants were readmitted, while 13 had no readmissions in their medical records between baseline and endpoint. Participants with ≥ 1 readmission experienced a greater increase in Rosow-Breslau score (median [IQR] 0 [-0.5 to 1.0]) than those not readmitted (median [IQR] 0 [-1.25 to 0.25]; ES, 1.0; P = .03). Borderline greater improvement in number of steps was found in those not readmitted (median [IQR] 3365.6 [274.4-7710.9]) compared with those readmitted (median [IQR] 319.9 [-136.1 to 774.5]; ES, -1.3; P = .05). Patients who were readmitted also tended to have lower and not statistically significant improvements in SPPB (median [IQR] 1 [-4.0 to 5.3]) compared with those not readmitted (median [IQR] 2 [0.3-3.8]; ES, -0.5; P = .17) (Table 3).

Discussion

This study examined the association between hospitalization, posthospital syndrome, and postdischarge health care use in patients undergoing a VVC-based intervention following hospital discharge. Participants who had no or mild cognitive impairment, no readmissions, higher medication changes, and a shorter hospital length of stay tended to experience lower disability, including in mobility and ADLs. This suggests individuals who are less clinically complex may be more likely to benefit from this type of virtual rehabilitation program. These findings are consistent with clinical experiences; home-based programs to improve physical activity posthospital discharge can be challenging for those who were medically ill (and did not undergo a specific surgical procedure), cognitively impaired, and become acutely ill and trigger hospital readmission. ¹⁵ For example, the sample in this study had higher rates of falls, pain, and readmissions compared to previous research.^2,3,34-39

The importance of posthospital syndrome in the context of recovery of function and health at home following hospitalization is well documented.^16-18 The potential impact of posthospital syndrome on physical activity-focused interventions is less understood. In our analysis, participants with mild or moderate cognitive impairment tended to become more dependent in their ADLs, while those with no cognitive impairment tended to become more independent in their ADLs. This functional decline over time is perhaps expected in persons with cognitive impairment, but the significant difference with a large ES warrants further consideration on how to tailor interventions to better promote functional recovery in these individuals.^40,41 While some cognitive decline may not be preventable, this finding supports the need to promote healthy cognitive aging, identify declines in cognition, and work to mitigate additional decline. Programs specifically designed to promote function and physical activity in older adults with cognitive impairment are needed, especially during care transitions.^41-43

While participants reported that falls and pain interference did not have a significant impact on change in outcomes between baseline and endpoint, these areas need further investigation. Falls and pain have been associated with function and physical activity in older adults.^42-46 Pain is common, yet underappreciated during older adult hospital-to-home transitions.^11,12,45,46 There is a need for more comprehensive assessment of pain (including pain intensity) and qualitative research.

Hospitalization and postdischarge health care application factors may have a significant impact on home-telehealth physical activity intervention success. Individuals who were postsurgery tended to have greater improvements in ADLs and physical performance. Most postsurgery participants had joint replacement surgery. Postsurgery status may not be modifiable, but it is important to note expected differences in recovery between medical and surgical admissions and the need to tailor care based on admission diagnosis. Those with a longer length of hospital stay may be considered at higher risk of suboptimal outcomes postdischarge, which indicates an opportunity for targeting resources and support, in addition to efforts of reducing length of stay where possible.⁴⁷

Readmissions were significantly related to a change in Rosow-Breslau mobility disability score. This may indicate the detrimental impact a readmission can have on increasing mobility and physical activity postdischarge, or the potential of this pilot program to impact readmissions by increasing mobility and physical activity, contrary to prior physical exercise interventions.^5,7,9,48 With 5% to 79% of readmissions considered preventable, continued efforts and program dissemination and implementation to address preventable readmissions are warranted.⁴⁹ Individuals with postdischarge physical therapy (prior to beginning the pilot program) tended to demonstrate less improvement in disability and physical performance. This relationship needs further investigation; the 2 groups did not appear to have significant differences at baseline, albeit with a small sample size. It is possible they experienced initial improvements with postdischarge physical therapy and plateaued or had little further reserve to improve upon entering the VVC program.

Strengths and Limitations

This pilot program provided evaluative data on the use of VVC to enhance function and physical activity in older adults posthospital discharge. It included individual (eg, fall, pain, cognitive impairment) and health service (eg, readmission, physical therapy) level factors as predictors of function and physical activity posthospitalization.^5,7,9,15-19

The results of this pilot project stem from a small sample lacking diversity in terms of race, ethnicity, and sex. There was some variation in baseline and endpoints between participants, and when hospitalization, posthospital syndrome, and postdischarge health care application factors were collected. The majority of participants were recruited within a month postdischarge, and the program lasted about 6 months. Data collection was attempted at regular PAT contacts, but there was some variation in when visits occurred based on participant availability and preference. Some participants had missing data, which was handled using pairwise deletion.³³ Larger studies are needed to confirm the findings of this study, particularly the trends that did not reach statistical significance. Home health services other than physical therapy (eg, nursing, occupational therapy) were not fully accounted for and should be considered in future research.

Conclusions

In patients undergoing a 6-month pilot VVC-based physical activity intervention posthospital discharge, improvements in mobility and disability were most likely in those who had no cognitive impairment and were not readmitted. Larger sample and qualitative investigations are necessary to optimize outcomes for patients who meet these clinical profiles.

More than 38 million people in the United States (12%) have diabetes mellitus (DM), though 1 in 5 are unaware they have DM.¹ The prevalence among veterans is even more substantial, impacting nearly 25% of those who received care from the US Department of Veterans Affairs (VA).² DM can lead to increased health care costs in addition to various complications (eg, cardiovascular, renal), especially if left uncontrolled.^1,3 similar impact is found in the perioperative period (defined as at or around the time of an operation), as multiple studies have found that uncontrolled preoperative DM can result in worsened surgical outcomes, including longer hospital stays, more infectious complications, and higher perioperative mortality.^4-6

In contrast, adequate glycemic control assessed with blood glucose levels has been shown to decrease the incidence of postoperative infections.⁷ Optimizing glycemic control during hospital stays, especially postsurgery, has become the standard of care, with most health systems establishing specific protocols. In current literature, most studies examining DM management in the perioperative period are focused on postoperative care, with little attention to the preoperative period.^4,6,7

One study found that patients with poor presurgery glycemic control assessed by hemoglobin A_1c (HbA_1c) levels were more likely to remain hyperglycemic during and after surgery. 8 Blood glucose levels < 200 mg/dL can lead to an increased risk of infection and impaired wound healing, meaning a well-controlled HbA_1c before a procedure serves as a potential factor for success.⁹ The 2025 American Diabetes Association (ADA) Standards of Care (SOC) recommendation is to target HbA_1c < 8% whenever possible, and some health systems require lower levels (eg, < 7% or 7.5%).¹⁰ With that goal in mind and knowing that preoperative hyperglycemia has been shown to be a contributing factor in the delay or cancellation of surgical cases, an argument can be made that attention to preoperative DM management also should be a focus for health care systems performing surgeries.^8,9,11

Attention to glucose control during preoperative care offers an opportunity to screen for DM in patients who may not have been screened otherwise and to standardize perioperative DM management. Since DM disproportionately impacts veterans, this is a pertinent issue to the VA. Veterans can be more susceptible to complications if DM is left uncontrolled prior to surgery. To determine readiness for surgery and control of comorbid conditions such as DM before a planned surgery, facilities often perform a preoperative clinic assessment, often in a multidisciplinary clinic.

At Veteran Health Indiana (VHI), a presurgery clinic visit involving the primary surgery service (physician, nurse practitioner, and/or a physician assistant) is conducted 1 to 2 months prior to the planned procedure to determine whether a patient is ready for surgery. During this visit, patients receive a packet with instructions for various tasks and medications, such as applying topical antibiotic prophylaxis on the anticipated surgical site. This is documented in the form of a note in the VHI Computerized Patient Record System (CPRS). The medication instructions are provided according to the preferences of the surgical team. These may be templated notes that contain general directions on the timing and dosing of specific medications, in addition to instructions for holding or reducing doses when appropriate. The instructions can be tailored by the team conducting the preoperative visit (eg, “Take 20 units of insulin glargine the day before surgery” vs “Take half of your long-acting insulin the night before surgery”). Specific to DM, VHI has a nurse-driven day of surgery glucose assessment where point-of-care blood glucose is collected during preoperative holding for most patients.

There is limited research assessing the level of preoperative glycemic control and the incidence of complications in a veteran population. The objective of this study was to gain a baseline understanding of what, if any, standardization exists for preoperative instructions for DM medications and to assess the level of preoperative glycemic control and postoperative complications in patients with DM undergoing major elective surgical procedures.

Methods

This retrospective, single-center chart review was conducted at VHI. The Indiana University and VHI institutional review boards determined that this quality improvement project was exempt from review.

The primary outcome was the number of patients with surgical procedures delayed or canceled due to hyperglycemia or hypoglycemia. Hyperglycemia was defined as blood glucose > 180 mg/dL and hypoglycemia was defined as < 70 mg/dL, slight variations from the current ADA SOC preoperative specific recommendation of a blood glucose reading of 100 to 180 mg/dL within 4 hours of surgery.¹⁰ The standard outpatient hypoglycemia definition of blood glucose < 70 mg/dL was chosen because the current goal (< 100 mg/dL) was not the standard in previous ADA SOCs that were in place during the study period. Specifically, the 2018 ADA SOC did not provide preoperative recommendations and the 2019-2021 ADA SOC recommended 80 to 180 mg/dL.^10,12-18 For patients who had multiple preoperative blood glucose measurements, the first recorded glucose on the day of the procedure was used.

The secondary outcomes of this study were focused on the preoperative process/care at VHI and postoperative glycemic control. The preoperative process included examining whether medication instructions were given and their quality. Additionally, the number of interventions for hyperglycemia and hypoglycemia were required immediately prior to surgery and the average preoperative HbA_1c (measured within 3 months prior to surgery) were collected and analyzed. For postoperative glycemic control, average blood glucose measurements and number of hypoglycemic (< 70 mg/dL) and hyperglycemic (> 180 mg/dL) events were measured in addition to the frequency of changes made at discharge to patients’ DM medication regimens.

The safety outcome of this study assessed commonly observed postoperative complications and was examined up to 30 days postsurgery. These included acute kidney injury (defined using Kidney Disease: Improving Global Outcomes 2012, the standard during the study period), nonfatal myocardial infarction, nonfatal stroke, and surgical site infections, which were identified from the discharge summary written by the primary surgery service.¹⁹ All-cause mortality also was collected.

Patients were included if they were admitted for major elective surgeries and had a diagnosis of either type 1 or type 2 DM on their problem list, determined by International Classification of Diseases, Tenth Revision codes. Major elective surgery was defined as a procedure that would likely result in a hospital admission of > 24 hours. Of note, patients may have been included in this study more than once if they had > 1 procedure at least 30 days apart and met inclusion criteria within the time frame. Patients were excluded if they were taking no DM medications or chronic steroids (at any dose), residing in a long-term care facility, being managed by a non-VA clinician prior to surgery, or missing a preoperative blood glucose measurement.

All data were collected from the CPRS. A list of surgical cases involving patients with DM who were scheduled to undergo major elective surgeries from January 1, 2018, to December 31, 2021, at VHI was generated. The list was randomized to a smaller number (N = 394) for data collection due to the time and resource constraints for a pharmacy residency project. All data were deidentified and stored in a secured VA server to protect patient confidentiality. Descriptive statistics were used for all results.

Results

Initially, 2362 surgeries were identified. A randomized sample of 394 charts were reviewed and 131 cases met inclusion criteria. Each case involved a unique patient (Figure). The most common reasons for exclusion were 143 patients with diet-controlled DM and 78 nonelective surgeries. The mean (SD) age of patients was 68 (8) years, and the most were male (98.5%) and White (76.3%) (Table 1). 

At baseline, 45 of 131 patients (34.4%) had coronary artery disease and 29 (22.1%) each had autonomic neuropathy and chronic kidney disease. Most surgeries were conducted by orthopedic (32.1%) and peripheral vascular (21.4%) specialties. The mean (SD) length of surgery was 4.6 (2.6) hours and of hospital length of stay was 4 (4) days. No patients stayed longer than the 30-day safety outcome follow-up period. All patients had type 2 DM and took a mean 2 DM medications. The 63 patients taking insulin had a mean (SD) total daily dose of 99 (77) U (Table 2). A preoperative HbA1c was collected in 116 patients within 3 months of surgery, with a mean HbA_1c of 7.0% (range, 5.3-10.7).

No patients had surgeries delayed or canceled because of uncontrolled DM on the day of surgery. The mean preoperative blood glucose level was 146 mg/dL (range, 73-365) (Table 3). No patients had a preoperative blood glucose level of < 70 mg/dL and 19 (14.5%) had a blood glucose level > 180 mg/dL. Among patients with hyperglycemia immediately prior to surgery, 6 (31.6%) had documentation of insulin being provided.

For this sample of patients, the preoperative clinic visit was conducted a mean 22 days prior to the planned surgery date. Among the 131 included patients, 122 (93.1%) had documentation of receiving instructions for DM medications. Among patients who had documented receipt of instructions, only 30 (24.6%) had instructions specifically tailored to their regimen rather than a generic templated form. The mean (SD) preoperative blood glucose was similar for those who received specific perioperative DM instructions at 146 (50) mg/dL when compared with those who did not at 147 (45) mg/dL. The mean (SD) preoperative blood glucose reading for those who had no documentation of receipt of perioperative instructions was 126 (54) mg/dL compared with 147 (46) mg/dL for those who did.

The mean number of postoperative blood glucose events per day was negligible for hypoglycemia and more frequent for hyperglycemia with a mean of 2 events per day. The mean postoperative blood glucose range was 121 to 247 mg/dL with most readings < 180 mg/dL. Upon discharge, most patients continued their home DM regimen with 5 patients (3.8%) having changes made to their regimen upon discharge.

Very few postoperative complications were identified from chart review. The most frequently observed postoperative complications were acute kidney injury, surgical site infections, and nonfatal stroke. There were no documented nonfatal myocardial infarctions. Two patients (1.5%) died within 30 days of the surgery; neither death was deemed to have been related to poor perioperative glycemic control.

Discussion

To our knowledge, this retrospective chart review was the first study to assess preoperative DM management and postoperative complications in a veteran population. VHI is a large, tertiary, level 1a, academic medical center that serves approximately 62,000 veterans annually and performs about 5000 to 6000 surgeries annually, a total that is increasing following the COVID-19 pandemic.²⁰ This study found that the current process of a presurgery clinic visit and day of surgery glucose assessment has prevented surgical delays or cancellations.

Most patients included in this study were well controlled at baseline in accordance with the 2025 ADA SOC HbA_1c recommendation of a preoperative HbA1c of < 8%, which may have contributed to no surgical delays or cancellations.¹⁰ However, not all patients had HbA1c collected within 3 months of surgery or even had one collected at all. Despite the ADA SOC providing no explicit recommendation for universal HbA_1c screening prior to elective procedures, its importance cannot be understated given the body of evidence demonstrating poor outcomes with uncontrolled preoperative DM.^8,10 The glycemic control at baseline may have contributed to the very few postsurgical complications observed in this study.

Although the current process at VHI prevented surgical delays and cancellations in this sample, there are still identified areas for improvement. One area is the instructions the patients received. Patients with DM are often prescribed ≥ 1 medication or a combination of insulins, noninsulin injectables, and oral DM medications, and this study population was no different. Because these medications may influence the anesthesia and perioperative periods, the ADA has specific guidance for altering administration schedules in the days leading up to surgery.¹⁰

Inappropriate administration of DM medications could lead to perioperative hypoglycemia or hyperglycemia, possibly causing surgical delays, case cancellations, and/or postoperative complications.21 Although these data reveal the specificity and documented receipt that the preoperative DM instructions did not impact the first recorded preoperative blood glucose, future studies should examine patient confidence in how to properly administer their DM medications prior to surgery. It is vital that patients receive clear instructions in accordance with the ADA SOC on whether to continue, hold, or adjust the dose of their medications to prevent fluctuations in blood glucose levels in the perioperative period, ensure safety with anesthesia, and prevent postoperative complications such as acute kidney injury. Of note, compliance with guideline recommendations for medication instructions was not examined because the data collection time frame expanded over multiple years and the recommendations have evolved each year as new data emerge.

Preoperative DM Management

The first key takeaway from this study is to ensure patients are ready for surgery with a formal assessment (typically in the form of a clinic visit) prior to the surgery. One private sector health system published their approach to this by administering an automatic preoperative HbA_1c screening for those with a DM diagnosis and all patients with a random plasma glucose ≥ 200 mg/dL.²² Additionally, if the patient's HbA_1c level was not at goal prior to surgery (≥ 8% for those with known DM and ≥ 6.5% with no known DM), patients were referred to endocrinology for further management. Increasing attention to the preoperative visit and extending HbA_1c testing to all patients regardless of DM status also provides an opportunity to identify individuals living with undiagnosed DM.¹

Even though there was no difference in the mean preoperative blood glucose level based on receipt or specificity of preoperative DM instructions, a second takeaway from this study is the importance of ensuring patients receive clear instructions on their DM medication schedule in the perioperative period. A practical first step may be updating the templates used by the primary surgery teams and providing education to the clinicians in the clinic on how to personalize the visits. Because the current preoperative DM process at VHI is managed by the primary surgical team in a clinic visit, there is an opportunity to shift this responsibility to other health care professionals, such as pharmacists—a change shown to reduce unintended omission of home medications following surgery during hospitalization and reduce costs.^23,24

Limitations

This study relied on data included in the patient chart. These data include medication interventions made immediately prior to surgery, which can sometimes be inaccurately charted or difficult to find as they are not documented in the typical medication administration record. Also, the safety outcomes were collected from a discharge summary written by different clinicians, which may lead to information bias. Special attention was taken to ensure these data points were collected as accurately as possible, but it is possible some data may be inaccurate from unintentional human error. Additionally, the safety outcome was limited to a 30-day follow-up, but encompassed the entire length of postoperative stay for all included patients. Finally, given this study was retrospective with no comparison group and the intent was to improve processes at VHI, only hypotheses and potential interventions can be generated from this study. Future prospective studies with larger sample sizes and comparator groups are needed to draw further conclusions.

Conclusions

This study found that the current presurgery process at VHI appears to be successful in preventing surgical delays or cancellations due to hyperglycemia or hypoglycemia. Optimizing DM management can improve surgical outcomes by decreasing rates of postoperative complications, and this study added additional evidence in support of that in a unique population: veterans. Insight on the awareness of preoperative blood glucose management should be gleaned from this study, and based on this sample and site, the preadmission screening process and instructions provided to patients can serve as 2 starting points for optimizing elective surgery.

More than 38 million people in the United States (12%) have diabetes mellitus (DM), though 1 in 5 are unaware they have DM.¹ The prevalence among veterans is even more substantial, impacting nearly 25% of those who received care from the US Department of Veterans Affairs (VA).² DM can lead to increased health care costs in addition to various complications (eg, cardiovascular, renal), especially if left uncontrolled.^1,3 similar impact is found in the perioperative period (defined as at or around the time of an operation), as multiple studies have found that uncontrolled preoperative DM can result in worsened surgical outcomes, including longer hospital stays, more infectious complications, and higher perioperative mortality.^4-6

In contrast, adequate glycemic control assessed with blood glucose levels has been shown to decrease the incidence of postoperative infections.⁷ Optimizing glycemic control during hospital stays, especially postsurgery, has become the standard of care, with most health systems establishing specific protocols. In current literature, most studies examining DM management in the perioperative period are focused on postoperative care, with little attention to the preoperative period.^4,6,7

One study found that patients with poor presurgery glycemic control assessed by hemoglobin A_1c (HbA_1c) levels were more likely to remain hyperglycemic during and after surgery. 8 Blood glucose levels < 200 mg/dL can lead to an increased risk of infection and impaired wound healing, meaning a well-controlled HbA_1c before a procedure serves as a potential factor for success.⁹ The 2025 American Diabetes Association (ADA) Standards of Care (SOC) recommendation is to target HbA_1c < 8% whenever possible, and some health systems require lower levels (eg, < 7% or 7.5%).¹⁰ With that goal in mind and knowing that preoperative hyperglycemia has been shown to be a contributing factor in the delay or cancellation of surgical cases, an argument can be made that attention to preoperative DM management also should be a focus for health care systems performing surgeries.^8,9,11

Attention to glucose control during preoperative care offers an opportunity to screen for DM in patients who may not have been screened otherwise and to standardize perioperative DM management. Since DM disproportionately impacts veterans, this is a pertinent issue to the VA. Veterans can be more susceptible to complications if DM is left uncontrolled prior to surgery. To determine readiness for surgery and control of comorbid conditions such as DM before a planned surgery, facilities often perform a preoperative clinic assessment, often in a multidisciplinary clinic.

At Veteran Health Indiana (VHI), a presurgery clinic visit involving the primary surgery service (physician, nurse practitioner, and/or a physician assistant) is conducted 1 to 2 months prior to the planned procedure to determine whether a patient is ready for surgery. During this visit, patients receive a packet with instructions for various tasks and medications, such as applying topical antibiotic prophylaxis on the anticipated surgical site. This is documented in the form of a note in the VHI Computerized Patient Record System (CPRS). The medication instructions are provided according to the preferences of the surgical team. These may be templated notes that contain general directions on the timing and dosing of specific medications, in addition to instructions for holding or reducing doses when appropriate. The instructions can be tailored by the team conducting the preoperative visit (eg, “Take 20 units of insulin glargine the day before surgery” vs “Take half of your long-acting insulin the night before surgery”). Specific to DM, VHI has a nurse-driven day of surgery glucose assessment where point-of-care blood glucose is collected during preoperative holding for most patients.

There is limited research assessing the level of preoperative glycemic control and the incidence of complications in a veteran population. The objective of this study was to gain a baseline understanding of what, if any, standardization exists for preoperative instructions for DM medications and to assess the level of preoperative glycemic control and postoperative complications in patients with DM undergoing major elective surgical procedures.

Methods

This retrospective, single-center chart review was conducted at VHI. The Indiana University and VHI institutional review boards determined that this quality improvement project was exempt from review.

The primary outcome was the number of patients with surgical procedures delayed or canceled due to hyperglycemia or hypoglycemia. Hyperglycemia was defined as blood glucose > 180 mg/dL and hypoglycemia was defined as < 70 mg/dL, slight variations from the current ADA SOC preoperative specific recommendation of a blood glucose reading of 100 to 180 mg/dL within 4 hours of surgery.¹⁰ The standard outpatient hypoglycemia definition of blood glucose < 70 mg/dL was chosen because the current goal (< 100 mg/dL) was not the standard in previous ADA SOCs that were in place during the study period. Specifically, the 2018 ADA SOC did not provide preoperative recommendations and the 2019-2021 ADA SOC recommended 80 to 180 mg/dL.^10,12-18 For patients who had multiple preoperative blood glucose measurements, the first recorded glucose on the day of the procedure was used.

The secondary outcomes of this study were focused on the preoperative process/care at VHI and postoperative glycemic control. The preoperative process included examining whether medication instructions were given and their quality. Additionally, the number of interventions for hyperglycemia and hypoglycemia were required immediately prior to surgery and the average preoperative HbA_1c (measured within 3 months prior to surgery) were collected and analyzed. For postoperative glycemic control, average blood glucose measurements and number of hypoglycemic (< 70 mg/dL) and hyperglycemic (> 180 mg/dL) events were measured in addition to the frequency of changes made at discharge to patients’ DM medication regimens.

The safety outcome of this study assessed commonly observed postoperative complications and was examined up to 30 days postsurgery. These included acute kidney injury (defined using Kidney Disease: Improving Global Outcomes 2012, the standard during the study period), nonfatal myocardial infarction, nonfatal stroke, and surgical site infections, which were identified from the discharge summary written by the primary surgery service.¹⁹ All-cause mortality also was collected.

Patients were included if they were admitted for major elective surgeries and had a diagnosis of either type 1 or type 2 DM on their problem list, determined by International Classification of Diseases, Tenth Revision codes. Major elective surgery was defined as a procedure that would likely result in a hospital admission of > 24 hours. Of note, patients may have been included in this study more than once if they had > 1 procedure at least 30 days apart and met inclusion criteria within the time frame. Patients were excluded if they were taking no DM medications or chronic steroids (at any dose), residing in a long-term care facility, being managed by a non-VA clinician prior to surgery, or missing a preoperative blood glucose measurement.

All data were collected from the CPRS. A list of surgical cases involving patients with DM who were scheduled to undergo major elective surgeries from January 1, 2018, to December 31, 2021, at VHI was generated. The list was randomized to a smaller number (N = 394) for data collection due to the time and resource constraints for a pharmacy residency project. All data were deidentified and stored in a secured VA server to protect patient confidentiality. Descriptive statistics were used for all results.

Results

Initially, 2362 surgeries were identified. A randomized sample of 394 charts were reviewed and 131 cases met inclusion criteria. Each case involved a unique patient (Figure). The most common reasons for exclusion were 143 patients with diet-controlled DM and 78 nonelective surgeries. The mean (SD) age of patients was 68 (8) years, and the most were male (98.5%) and White (76.3%) (Table 1). 

At baseline, 45 of 131 patients (34.4%) had coronary artery disease and 29 (22.1%) each had autonomic neuropathy and chronic kidney disease. Most surgeries were conducted by orthopedic (32.1%) and peripheral vascular (21.4%) specialties. The mean (SD) length of surgery was 4.6 (2.6) hours and of hospital length of stay was 4 (4) days. No patients stayed longer than the 30-day safety outcome follow-up period. All patients had type 2 DM and took a mean 2 DM medications. The 63 patients taking insulin had a mean (SD) total daily dose of 99 (77) U (Table 2). A preoperative HbA1c was collected in 116 patients within 3 months of surgery, with a mean HbA_1c of 7.0% (range, 5.3-10.7).

No patients had surgeries delayed or canceled because of uncontrolled DM on the day of surgery. The mean preoperative blood glucose level was 146 mg/dL (range, 73-365) (Table 3). No patients had a preoperative blood glucose level of < 70 mg/dL and 19 (14.5%) had a blood glucose level > 180 mg/dL. Among patients with hyperglycemia immediately prior to surgery, 6 (31.6%) had documentation of insulin being provided.

For this sample of patients, the preoperative clinic visit was conducted a mean 22 days prior to the planned surgery date. Among the 131 included patients, 122 (93.1%) had documentation of receiving instructions for DM medications. Among patients who had documented receipt of instructions, only 30 (24.6%) had instructions specifically tailored to their regimen rather than a generic templated form. The mean (SD) preoperative blood glucose was similar for those who received specific perioperative DM instructions at 146 (50) mg/dL when compared with those who did not at 147 (45) mg/dL. The mean (SD) preoperative blood glucose reading for those who had no documentation of receipt of perioperative instructions was 126 (54) mg/dL compared with 147 (46) mg/dL for those who did.

The mean number of postoperative blood glucose events per day was negligible for hypoglycemia and more frequent for hyperglycemia with a mean of 2 events per day. The mean postoperative blood glucose range was 121 to 247 mg/dL with most readings < 180 mg/dL. Upon discharge, most patients continued their home DM regimen with 5 patients (3.8%) having changes made to their regimen upon discharge.

Very few postoperative complications were identified from chart review. The most frequently observed postoperative complications were acute kidney injury, surgical site infections, and nonfatal stroke. There were no documented nonfatal myocardial infarctions. Two patients (1.5%) died within 30 days of the surgery; neither death was deemed to have been related to poor perioperative glycemic control.

Discussion

To our knowledge, this retrospective chart review was the first study to assess preoperative DM management and postoperative complications in a veteran population. VHI is a large, tertiary, level 1a, academic medical center that serves approximately 62,000 veterans annually and performs about 5000 to 6000 surgeries annually, a total that is increasing following the COVID-19 pandemic.²⁰ This study found that the current process of a presurgery clinic visit and day of surgery glucose assessment has prevented surgical delays or cancellations.

Most patients included in this study were well controlled at baseline in accordance with the 2025 ADA SOC HbA_1c recommendation of a preoperative HbA1c of < 8%, which may have contributed to no surgical delays or cancellations.¹⁰ However, not all patients had HbA1c collected within 3 months of surgery or even had one collected at all. Despite the ADA SOC providing no explicit recommendation for universal HbA_1c screening prior to elective procedures, its importance cannot be understated given the body of evidence demonstrating poor outcomes with uncontrolled preoperative DM.^8,10 The glycemic control at baseline may have contributed to the very few postsurgical complications observed in this study.

Although the current process at VHI prevented surgical delays and cancellations in this sample, there are still identified areas for improvement. One area is the instructions the patients received. Patients with DM are often prescribed ≥ 1 medication or a combination of insulins, noninsulin injectables, and oral DM medications, and this study population was no different. Because these medications may influence the anesthesia and perioperative periods, the ADA has specific guidance for altering administration schedules in the days leading up to surgery.¹⁰

Inappropriate administration of DM medications could lead to perioperative hypoglycemia or hyperglycemia, possibly causing surgical delays, case cancellations, and/or postoperative complications.21 Although these data reveal the specificity and documented receipt that the preoperative DM instructions did not impact the first recorded preoperative blood glucose, future studies should examine patient confidence in how to properly administer their DM medications prior to surgery. It is vital that patients receive clear instructions in accordance with the ADA SOC on whether to continue, hold, or adjust the dose of their medications to prevent fluctuations in blood glucose levels in the perioperative period, ensure safety with anesthesia, and prevent postoperative complications such as acute kidney injury. Of note, compliance with guideline recommendations for medication instructions was not examined because the data collection time frame expanded over multiple years and the recommendations have evolved each year as new data emerge.

Preoperative DM Management

The first key takeaway from this study is to ensure patients are ready for surgery with a formal assessment (typically in the form of a clinic visit) prior to the surgery. One private sector health system published their approach to this by administering an automatic preoperative HbA_1c screening for those with a DM diagnosis and all patients with a random plasma glucose ≥ 200 mg/dL.²² Additionally, if the patient's HbA_1c level was not at goal prior to surgery (≥ 8% for those with known DM and ≥ 6.5% with no known DM), patients were referred to endocrinology for further management. Increasing attention to the preoperative visit and extending HbA_1c testing to all patients regardless of DM status also provides an opportunity to identify individuals living with undiagnosed DM.¹

Even though there was no difference in the mean preoperative blood glucose level based on receipt or specificity of preoperative DM instructions, a second takeaway from this study is the importance of ensuring patients receive clear instructions on their DM medication schedule in the perioperative period. A practical first step may be updating the templates used by the primary surgery teams and providing education to the clinicians in the clinic on how to personalize the visits. Because the current preoperative DM process at VHI is managed by the primary surgical team in a clinic visit, there is an opportunity to shift this responsibility to other health care professionals, such as pharmacists—a change shown to reduce unintended omission of home medications following surgery during hospitalization and reduce costs.^23,24

Limitations

This study relied on data included in the patient chart. These data include medication interventions made immediately prior to surgery, which can sometimes be inaccurately charted or difficult to find as they are not documented in the typical medication administration record. Also, the safety outcomes were collected from a discharge summary written by different clinicians, which may lead to information bias. Special attention was taken to ensure these data points were collected as accurately as possible, but it is possible some data may be inaccurate from unintentional human error. Additionally, the safety outcome was limited to a 30-day follow-up, but encompassed the entire length of postoperative stay for all included patients. Finally, given this study was retrospective with no comparison group and the intent was to improve processes at VHI, only hypotheses and potential interventions can be generated from this study. Future prospective studies with larger sample sizes and comparator groups are needed to draw further conclusions.

Conclusions

This study found that the current presurgery process at VHI appears to be successful in preventing surgical delays or cancellations due to hyperglycemia or hypoglycemia. Optimizing DM management can improve surgical outcomes by decreasing rates of postoperative complications, and this study added additional evidence in support of that in a unique population: veterans. Insight on the awareness of preoperative blood glucose management should be gleaned from this study, and based on this sample and site, the preadmission screening process and instructions provided to patients can serve as 2 starting points for optimizing elective surgery.

More than 38 million people in the United States (12%) have diabetes mellitus (DM), though 1 in 5 are unaware they have DM.¹ The prevalence among veterans is even more substantial, impacting nearly 25% of those who received care from the US Department of Veterans Affairs (VA).² DM can lead to increased health care costs in addition to various complications (eg, cardiovascular, renal), especially if left uncontrolled.^1,3 similar impact is found in the perioperative period (defined as at or around the time of an operation), as multiple studies have found that uncontrolled preoperative DM can result in worsened surgical outcomes, including longer hospital stays, more infectious complications, and higher perioperative mortality.^4-6

In contrast, adequate glycemic control assessed with blood glucose levels has been shown to decrease the incidence of postoperative infections.⁷ Optimizing glycemic control during hospital stays, especially postsurgery, has become the standard of care, with most health systems establishing specific protocols. In current literature, most studies examining DM management in the perioperative period are focused on postoperative care, with little attention to the preoperative period.^4,6,7

One study found that patients with poor presurgery glycemic control assessed by hemoglobin A_1c (HbA_1c) levels were more likely to remain hyperglycemic during and after surgery. 8 Blood glucose levels < 200 mg/dL can lead to an increased risk of infection and impaired wound healing, meaning a well-controlled HbA_1c before a procedure serves as a potential factor for success.⁹ The 2025 American Diabetes Association (ADA) Standards of Care (SOC) recommendation is to target HbA_1c < 8% whenever possible, and some health systems require lower levels (eg, < 7% or 7.5%).¹⁰ With that goal in mind and knowing that preoperative hyperglycemia has been shown to be a contributing factor in the delay or cancellation of surgical cases, an argument can be made that attention to preoperative DM management also should be a focus for health care systems performing surgeries.^8,9,11

Attention to glucose control during preoperative care offers an opportunity to screen for DM in patients who may not have been screened otherwise and to standardize perioperative DM management. Since DM disproportionately impacts veterans, this is a pertinent issue to the VA. Veterans can be more susceptible to complications if DM is left uncontrolled prior to surgery. To determine readiness for surgery and control of comorbid conditions such as DM before a planned surgery, facilities often perform a preoperative clinic assessment, often in a multidisciplinary clinic.

At Veteran Health Indiana (VHI), a presurgery clinic visit involving the primary surgery service (physician, nurse practitioner, and/or a physician assistant) is conducted 1 to 2 months prior to the planned procedure to determine whether a patient is ready for surgery. During this visit, patients receive a packet with instructions for various tasks and medications, such as applying topical antibiotic prophylaxis on the anticipated surgical site. This is documented in the form of a note in the VHI Computerized Patient Record System (CPRS). The medication instructions are provided according to the preferences of the surgical team. These may be templated notes that contain general directions on the timing and dosing of specific medications, in addition to instructions for holding or reducing doses when appropriate. The instructions can be tailored by the team conducting the preoperative visit (eg, “Take 20 units of insulin glargine the day before surgery” vs “Take half of your long-acting insulin the night before surgery”). Specific to DM, VHI has a nurse-driven day of surgery glucose assessment where point-of-care blood glucose is collected during preoperative holding for most patients.

There is limited research assessing the level of preoperative glycemic control and the incidence of complications in a veteran population. The objective of this study was to gain a baseline understanding of what, if any, standardization exists for preoperative instructions for DM medications and to assess the level of preoperative glycemic control and postoperative complications in patients with DM undergoing major elective surgical procedures.

Methods

This retrospective, single-center chart review was conducted at VHI. The Indiana University and VHI institutional review boards determined that this quality improvement project was exempt from review.

The primary outcome was the number of patients with surgical procedures delayed or canceled due to hyperglycemia or hypoglycemia. Hyperglycemia was defined as blood glucose > 180 mg/dL and hypoglycemia was defined as < 70 mg/dL, slight variations from the current ADA SOC preoperative specific recommendation of a blood glucose reading of 100 to 180 mg/dL within 4 hours of surgery.¹⁰ The standard outpatient hypoglycemia definition of blood glucose < 70 mg/dL was chosen because the current goal (< 100 mg/dL) was not the standard in previous ADA SOCs that were in place during the study period. Specifically, the 2018 ADA SOC did not provide preoperative recommendations and the 2019-2021 ADA SOC recommended 80 to 180 mg/dL.^10,12-18 For patients who had multiple preoperative blood glucose measurements, the first recorded glucose on the day of the procedure was used.

The secondary outcomes of this study were focused on the preoperative process/care at VHI and postoperative glycemic control. The preoperative process included examining whether medication instructions were given and their quality. Additionally, the number of interventions for hyperglycemia and hypoglycemia were required immediately prior to surgery and the average preoperative HbA_1c (measured within 3 months prior to surgery) were collected and analyzed. For postoperative glycemic control, average blood glucose measurements and number of hypoglycemic (< 70 mg/dL) and hyperglycemic (> 180 mg/dL) events were measured in addition to the frequency of changes made at discharge to patients’ DM medication regimens.

The safety outcome of this study assessed commonly observed postoperative complications and was examined up to 30 days postsurgery. These included acute kidney injury (defined using Kidney Disease: Improving Global Outcomes 2012, the standard during the study period), nonfatal myocardial infarction, nonfatal stroke, and surgical site infections, which were identified from the discharge summary written by the primary surgery service.¹⁹ All-cause mortality also was collected.

Patients were included if they were admitted for major elective surgeries and had a diagnosis of either type 1 or type 2 DM on their problem list, determined by International Classification of Diseases, Tenth Revision codes. Major elective surgery was defined as a procedure that would likely result in a hospital admission of > 24 hours. Of note, patients may have been included in this study more than once if they had > 1 procedure at least 30 days apart and met inclusion criteria within the time frame. Patients were excluded if they were taking no DM medications or chronic steroids (at any dose), residing in a long-term care facility, being managed by a non-VA clinician prior to surgery, or missing a preoperative blood glucose measurement.

All data were collected from the CPRS. A list of surgical cases involving patients with DM who were scheduled to undergo major elective surgeries from January 1, 2018, to December 31, 2021, at VHI was generated. The list was randomized to a smaller number (N = 394) for data collection due to the time and resource constraints for a pharmacy residency project. All data were deidentified and stored in a secured VA server to protect patient confidentiality. Descriptive statistics were used for all results.

Results

Initially, 2362 surgeries were identified. A randomized sample of 394 charts were reviewed and 131 cases met inclusion criteria. Each case involved a unique patient (Figure). The most common reasons for exclusion were 143 patients with diet-controlled DM and 78 nonelective surgeries. The mean (SD) age of patients was 68 (8) years, and the most were male (98.5%) and White (76.3%) (Table 1). 

At baseline, 45 of 131 patients (34.4%) had coronary artery disease and 29 (22.1%) each had autonomic neuropathy and chronic kidney disease. Most surgeries were conducted by orthopedic (32.1%) and peripheral vascular (21.4%) specialties. The mean (SD) length of surgery was 4.6 (2.6) hours and of hospital length of stay was 4 (4) days. No patients stayed longer than the 30-day safety outcome follow-up period. All patients had type 2 DM and took a mean 2 DM medications. The 63 patients taking insulin had a mean (SD) total daily dose of 99 (77) U (Table 2). A preoperative HbA1c was collected in 116 patients within 3 months of surgery, with a mean HbA_1c of 7.0% (range, 5.3-10.7).

No patients had surgeries delayed or canceled because of uncontrolled DM on the day of surgery. The mean preoperative blood glucose level was 146 mg/dL (range, 73-365) (Table 3). No patients had a preoperative blood glucose level of < 70 mg/dL and 19 (14.5%) had a blood glucose level > 180 mg/dL. Among patients with hyperglycemia immediately prior to surgery, 6 (31.6%) had documentation of insulin being provided.

For this sample of patients, the preoperative clinic visit was conducted a mean 22 days prior to the planned surgery date. Among the 131 included patients, 122 (93.1%) had documentation of receiving instructions for DM medications. Among patients who had documented receipt of instructions, only 30 (24.6%) had instructions specifically tailored to their regimen rather than a generic templated form. The mean (SD) preoperative blood glucose was similar for those who received specific perioperative DM instructions at 146 (50) mg/dL when compared with those who did not at 147 (45) mg/dL. The mean (SD) preoperative blood glucose reading for those who had no documentation of receipt of perioperative instructions was 126 (54) mg/dL compared with 147 (46) mg/dL for those who did.

The mean number of postoperative blood glucose events per day was negligible for hypoglycemia and more frequent for hyperglycemia with a mean of 2 events per day. The mean postoperative blood glucose range was 121 to 247 mg/dL with most readings < 180 mg/dL. Upon discharge, most patients continued their home DM regimen with 5 patients (3.8%) having changes made to their regimen upon discharge.

Very few postoperative complications were identified from chart review. The most frequently observed postoperative complications were acute kidney injury, surgical site infections, and nonfatal stroke. There were no documented nonfatal myocardial infarctions. Two patients (1.5%) died within 30 days of the surgery; neither death was deemed to have been related to poor perioperative glycemic control.

Discussion

To our knowledge, this retrospective chart review was the first study to assess preoperative DM management and postoperative complications in a veteran population. VHI is a large, tertiary, level 1a, academic medical center that serves approximately 62,000 veterans annually and performs about 5000 to 6000 surgeries annually, a total that is increasing following the COVID-19 pandemic.²⁰ This study found that the current process of a presurgery clinic visit and day of surgery glucose assessment has prevented surgical delays or cancellations.

Most patients included in this study were well controlled at baseline in accordance with the 2025 ADA SOC HbA_1c recommendation of a preoperative HbA1c of < 8%, which may have contributed to no surgical delays or cancellations.¹⁰ However, not all patients had HbA1c collected within 3 months of surgery or even had one collected at all. Despite the ADA SOC providing no explicit recommendation for universal HbA_1c screening prior to elective procedures, its importance cannot be understated given the body of evidence demonstrating poor outcomes with uncontrolled preoperative DM.^8,10 The glycemic control at baseline may have contributed to the very few postsurgical complications observed in this study.

Although the current process at VHI prevented surgical delays and cancellations in this sample, there are still identified areas for improvement. One area is the instructions the patients received. Patients with DM are often prescribed ≥ 1 medication or a combination of insulins, noninsulin injectables, and oral DM medications, and this study population was no different. Because these medications may influence the anesthesia and perioperative periods, the ADA has specific guidance for altering administration schedules in the days leading up to surgery.¹⁰

Inappropriate administration of DM medications could lead to perioperative hypoglycemia or hyperglycemia, possibly causing surgical delays, case cancellations, and/or postoperative complications.21 Although these data reveal the specificity and documented receipt that the preoperative DM instructions did not impact the first recorded preoperative blood glucose, future studies should examine patient confidence in how to properly administer their DM medications prior to surgery. It is vital that patients receive clear instructions in accordance with the ADA SOC on whether to continue, hold, or adjust the dose of their medications to prevent fluctuations in blood glucose levels in the perioperative period, ensure safety with anesthesia, and prevent postoperative complications such as acute kidney injury. Of note, compliance with guideline recommendations for medication instructions was not examined because the data collection time frame expanded over multiple years and the recommendations have evolved each year as new data emerge.

Preoperative DM Management

The first key takeaway from this study is to ensure patients are ready for surgery with a formal assessment (typically in the form of a clinic visit) prior to the surgery. One private sector health system published their approach to this by administering an automatic preoperative HbA_1c screening for those with a DM diagnosis and all patients with a random plasma glucose ≥ 200 mg/dL.²² Additionally, if the patient's HbA_1c level was not at goal prior to surgery (≥ 8% for those with known DM and ≥ 6.5% with no known DM), patients were referred to endocrinology for further management. Increasing attention to the preoperative visit and extending HbA_1c testing to all patients regardless of DM status also provides an opportunity to identify individuals living with undiagnosed DM.¹

Even though there was no difference in the mean preoperative blood glucose level based on receipt or specificity of preoperative DM instructions, a second takeaway from this study is the importance of ensuring patients receive clear instructions on their DM medication schedule in the perioperative period. A practical first step may be updating the templates used by the primary surgery teams and providing education to the clinicians in the clinic on how to personalize the visits. Because the current preoperative DM process at VHI is managed by the primary surgical team in a clinic visit, there is an opportunity to shift this responsibility to other health care professionals, such as pharmacists—a change shown to reduce unintended omission of home medications following surgery during hospitalization and reduce costs.^23,24

Limitations

This study relied on data included in the patient chart. These data include medication interventions made immediately prior to surgery, which can sometimes be inaccurately charted or difficult to find as they are not documented in the typical medication administration record. Also, the safety outcomes were collected from a discharge summary written by different clinicians, which may lead to information bias. Special attention was taken to ensure these data points were collected as accurately as possible, but it is possible some data may be inaccurate from unintentional human error. Additionally, the safety outcome was limited to a 30-day follow-up, but encompassed the entire length of postoperative stay for all included patients. Finally, given this study was retrospective with no comparison group and the intent was to improve processes at VHI, only hypotheses and potential interventions can be generated from this study. Future prospective studies with larger sample sizes and comparator groups are needed to draw further conclusions.

Conclusions

This study found that the current presurgery process at VHI appears to be successful in preventing surgical delays or cancellations due to hyperglycemia or hypoglycemia. Optimizing DM management can improve surgical outcomes by decreasing rates of postoperative complications, and this study added additional evidence in support of that in a unique population: veterans. Insight on the awareness of preoperative blood glucose management should be gleaned from this study, and based on this sample and site, the preadmission screening process and instructions provided to patients can serve as 2 starting points for optimizing elective surgery.

Type 2 diabetes mellitus (T2DM) is a chronic disease becoming more prevalent each year and is the seventh-leading cause of death in the United States.¹ The most common reason for hospitalization for patients with T2DM is uncontrolled glycemic levels.² Nearly 25% of the US Department of Veterans Affairs (VA) patient population has T2DM.³ T2DM is the leading cause of blindness, end-stage renal disease, and amputation for VA patients.⁴

According to the 2023 American Diabetes Association (ADA) guidelines, treatment goals of T2DM include eliminating symptoms, preventing or delaying complications, and attaining glycemic goals. A typical hemoglobin A_1c (HbA_1c) goal range is < 7%, but individual goals can vary up to < 9% due to a multitude of factors, including patient comorbidities and clinical status.⁵

Initial treatment recommendations are nonpharmacologic and include comprehensive lifestyle interventions such as optimizing nutrition, physical activity, and behavioral therapy. When pharmacologic therapy is required, metformin is the preferred first-line treatment for the majority of newly diagnosed patients with T2DM and should be added to continued lifestyle management.⁵ If HbA_1c levels remains above goal, the 2023 ADA guidelines recommend adding a second medication, including but not limited to insulin, a glucagonlike peptide-1 receptor agonist (GLP-1RA), or a sodium-glucose cotransporter 2 inhibitor. Medication choice is largely based on the patient’s concomitant conditions (eg, atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease). The 2023 ADA guidelines suggest initiating insulin therapy when a patient's blood glucose ≥ 300 mg/dL, HbA_1c > 10%, or if the patient has symptoms of hyperglycemia, even at initial diagnosis. Initiating medications to minimize or avoid hypoglycemia is a priority, especially in high-risk individuals.⁵

Clinical evidence shows that GLP-1RAs may provide similar glycemic control to insulin with lower risk of hypoglycemia.⁶ Other reported benefits of GLP-1RAs include weight loss, blood pressure reduction, and improved lipid levels. The most common adverse events (AEs) with GLP-1RAs are gastrointestinal. Including GLP-1RAs in T2DM pharmacotherapy may lower the risk of hypoglycemia, especially in patients at high risk of hypoglycemia.

The 2023 ADA guidelines indicate that it is appropriate to initiate GLP-]1RAs in patients on insulin.⁵ However, while GLP-1RAs do not increase the risk of hypoglycemia independently, combination treatment with GLP-1RAs and insulin can still result in hypoglycemia.⁶ Insulin is the key suspect of this hypoglycemic risk.⁷ Thus, if insulin dosage can be reduced or discontinued, this might reduce the risk of hypoglycemia.

The literature is limited on how the addition of a GLP-1RA to insulin treatment will affect the patient's daily insulin doses, particularly for the veteran population. The goal of this study is to examine this gap in current research by examining semaglutide, which is the current formulary preferred GLP-1RA at the VA.

Semaglutide is subcutaneously initiated at a dose of 0.25 mg once weekly for 4 weeks to reduce gastrointestinal symptoms, then increased to 0.5 mg weekly. Additional increases to a maintenance dose of 1 mg or 2 mg weekly can occur to achieve glycemic goals. The SUSTAIN-FORTE randomized controlled trial sought to determine whether there was a difference in HbA_1c level reduction and significant weight loss with the 2-mg vs 1-mg dose.⁸ Patients in the trial were taking metformin but needed additional medication to control their HbA_1c. They were not using insulin and may or may not have been taking sulfonylureas prior to semaglutide initiation. Semaglutide 2 mg was found to significantly improve HbA_1c control and promote weight loss compared with semaglutide 1 mg, while maintaining a similar safety profile.

Because this study involved patients who required additional HbA_1c control, although semaglutide reduced HbA_1c, not all patients were able to reduce their other diabetes medications, which depended on the baseline HbA_1c level and the level upon completion of semaglutide titration. Dose reductions for the patients’ other T2DM medications were not reported at trial end. SUSTAIN-FORTE established titration up to semaglutide 2 mg as effective for HbA_1c reduction, although it did not study patients also on insulin.⁸

Insulin is associated with hypoglycemic risk, weight gain, and other AEs.^7,8 This study analyzed whether increasing semaglutide could reduce insulin doses and therefore reduce risk of AEs in patients with T2DM.

Methods

A retrospective, single-center, chart review was conducted at VA Sioux Falls Health Care System (VASFHCS). Data were collected through manual review of VASFHCS electronic medical records. Patients aged ≥ 18 years with active prescriptions for at least once-daily insulin who were initiated on 2-mg weekly dose of semaglutide at the VASFHCS clinical pharmacy practitioner medication management clinic between January 1, 2021, and September 1, 2023, were included. VASFHCS clinical pharmacy practitioners have a scope of practice that allows them to initiate, modify, or discontinue medication therapy within medication management clinics.

The most frequently used prandial insulin at VASFHCS is insulin aspart, and the most frequently used basal insulin is insulin glargine. Patients were retrospectively monitored as they progressed from baseline (the point in time where semaglutide 0.5 mg was initiated) to ≥ 3 months on semaglutide 2-mg therapy. Patients were excluded if they previously used a GLP-1RA or if they were on sliding scale insulin without an exact daily dosage.

The primary endpoint was the percent change in total daily insulin dose from baseline to each dose increase after receiving semaglutide 2 mg for ≥ 3 months. Secondary endpoints included changes in daily prandial insulin dose, daily basal insulin dose, HbA_1c, and number of hypoglycemic events reported. Data collected included age, race, weight, body mass index, total daily prandial insulin dose, total daily basal insulin dose, HbA_1c, and hypoglycemic events reported at the visit when semaglutide was initiated.

Statistical Analysis

The sample size was calculated prior to data collection, and it was determined that for α = .05, 47 patients were needed to achieve 95% power. The primary endpoint was assessed using a paired t test, as were each secondary endpoint. Results with P < .05 were considered statistically significant.

Results

Sixty-two patients were included. The mean HbA_1c level at baseline was 7.7%, the baseline mean prandial and insulin daily doses were 41.5 units and 85.1 units, respectively (Table 1) From baseline to initiation of a semaglutide 1-mg dose, the daily insulin dose changed –5.6% (95% CI, 2.2-14.0; P = .008). From baseline to 2-mg dose initiation daily insulin changed -22.2% (95% CI, 22.0-35.1; P < .001) and for patients receiving semaglutide 2 mg for ≥ 3 months it changed -36.9% (95% CI, 37.4-56.5; P < .001) (Figure).

After receiving the 2-mg dose for ≥ 3 months, the mean daily dose of prandial insulin decreased from 41.5 units to 24.6 units (95% CI, 12.6-21.2; P < .001); mean daily dose of basal insulin decreased from 85.1 units to 52.1 units (95% CI, 23.9-42.0; P < .001); and mean HbA_1c level decreased from 7.7% to 7.1% (95% CI, 0.3-0.8; P < .001). Mean number of hypoglycemic events reported was not statistically significant, changing from 3.6 to 3.2 (95% CI, –0.6 to 0.1; P = .21) (Table 2).

Discussion

This study investigated the effect of subcutaneous semaglutide dose escalation on total daily insulin dose for patients with T2DM. There was a statistically significant decrease in total daily insulin dose from baseline to 1 mg initiation; this decrease continued with further insulin dose reduction seen at the 2-mg dose initiation and additional insulin dose reduction at ≥ 3 months at this dose. It was hypothesized there would be a significant total daily insulin dose reduction at some point, especially when transitioning from the semaglutide 1-mg to the 2-mg dose, based on previous research. ^9,10 The additional reduction in daily insulin dose when continuing on semaglutide 2 mg for ≥ 3 months was an unanticipated but added benefit, showing that if tolerated, maintaining the 2-mg dose will help patients reduce their insulin doses.

In terms of secondary endpoints, there was a statistically significant decrease in mean total daily dose individually for prandial and basal insulin from baseline to ≥ 3 months after semaglutide 2 mg initiation. The change in HbA_1c level was also statistically significant and decreased from baseline, even as insulin doses were reduced. This change in HbA_1c level was expected; previous literature has shown a significant link between improving HbA_1c control when semaglutide doses are increased to 2 mg weekly.¹⁰ Due to having been shown in previous trials, it was expected that HbA_1c levels would decrease even when the insulin doses were being reduced.¹⁰ Insulin dose reduction can potentially be added to the growing evidence of semaglutide benefits. The change in the number of hypoglycemic events was not statistically significant, which was unexpected since previous research show a trend in patients taking GLP-1RAs having fewer hypoglycemic events than those taking insulin.⁶ Further investigation with a larger sample size and prospective trial could determine whether this result is an outlier. In this study, there was no increase in HbA_1c or hypoglycemic events reported with increasing semaglutide doses, which provides further evidence of the safety of semaglutide even at higher doses.

These data suggest that for a patient with T2DM who is already taking insulin, the recommended titration of semaglutide is to start with 0.5 mg and titrate up to a 2-mg subcutaneous weekly dose and to then continue at that dose. As long as the 2-mg dose is tolerated, it will provide patients with the most HbA_1c control and lead to a reduction of their total daily insulin doses according to these results.

Strengths and Limitations

This study compared patient data at different points. This method did not require a second distinct control group, which would potentially introduce confounding factors, such as different baseline characteristics. Another strength is that documentation was available for all patients throughout the study so no one was lost to follow-up. This allowed comprehensive data collection and provided a stronger conclusion given the completeness of the data from baseline to follow-up.

Limitations include the retrospective design and small sample size. In addition, the study design did not allow for randomization. There is no documentation of adherence to medication regimen, which was difficult to determine due to the retrospective nature. Other changes to the patients’ medication regimen were not collected in aggregate and thus, it is possible the total daily insulin dose was impacted by other medication changes. There is also potential for inconsistent documentation of the patients’ true total daily insulin dose in the medical record, thus leading to inaccuracy of recorded data.

Conclusions

A small sample of veterans with T2DM had statistically significant reductions in total daily insulin dose when subcutaneous semaglutide was initiated, as well as after each dose increase. There was also a statistically significant reduction in HbA_1c levels from baseline even as patient insulin doses were reduced. These results support the current practice of using semaglutide to treat T2DM, suggesting it may be safe and effective at reducing HbA_1c levels as the dose is titrated up to 2 mg. There was no statistically significant change in the number of hypoglycemic events reported as semaglutide was titrated up. Thus, when semaglutide is increased to the maximum recommended dose of 2 mg for T2DM, patients may experience a reduction of their total daily dose of insulin and HbA_1c levels. These benefits may reduce the risk of insulin-related AEs while maintaining appropriate glycemic control.

Type 2 diabetes mellitus (T2DM) is a chronic disease becoming more prevalent each year and is the seventh-leading cause of death in the United States.¹ The most common reason for hospitalization for patients with T2DM is uncontrolled glycemic levels.² Nearly 25% of the US Department of Veterans Affairs (VA) patient population has T2DM.³ T2DM is the leading cause of blindness, end-stage renal disease, and amputation for VA patients.⁴

According to the 2023 American Diabetes Association (ADA) guidelines, treatment goals of T2DM include eliminating symptoms, preventing or delaying complications, and attaining glycemic goals. A typical hemoglobin A_1c (HbA_1c) goal range is < 7%, but individual goals can vary up to < 9% due to a multitude of factors, including patient comorbidities and clinical status.⁵

Initial treatment recommendations are nonpharmacologic and include comprehensive lifestyle interventions such as optimizing nutrition, physical activity, and behavioral therapy. When pharmacologic therapy is required, metformin is the preferred first-line treatment for the majority of newly diagnosed patients with T2DM and should be added to continued lifestyle management.⁵ If HbA_1c levels remains above goal, the 2023 ADA guidelines recommend adding a second medication, including but not limited to insulin, a glucagonlike peptide-1 receptor agonist (GLP-1RA), or a sodium-glucose cotransporter 2 inhibitor. Medication choice is largely based on the patient’s concomitant conditions (eg, atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease). The 2023 ADA guidelines suggest initiating insulin therapy when a patient's blood glucose ≥ 300 mg/dL, HbA_1c > 10%, or if the patient has symptoms of hyperglycemia, even at initial diagnosis. Initiating medications to minimize or avoid hypoglycemia is a priority, especially in high-risk individuals.⁵

Clinical evidence shows that GLP-1RAs may provide similar glycemic control to insulin with lower risk of hypoglycemia.⁶ Other reported benefits of GLP-1RAs include weight loss, blood pressure reduction, and improved lipid levels. The most common adverse events (AEs) with GLP-1RAs are gastrointestinal. Including GLP-1RAs in T2DM pharmacotherapy may lower the risk of hypoglycemia, especially in patients at high risk of hypoglycemia.

The 2023 ADA guidelines indicate that it is appropriate to initiate GLP-]1RAs in patients on insulin.⁵ However, while GLP-1RAs do not increase the risk of hypoglycemia independently, combination treatment with GLP-1RAs and insulin can still result in hypoglycemia.⁶ Insulin is the key suspect of this hypoglycemic risk.⁷ Thus, if insulin dosage can be reduced or discontinued, this might reduce the risk of hypoglycemia.

The literature is limited on how the addition of a GLP-1RA to insulin treatment will affect the patient's daily insulin doses, particularly for the veteran population. The goal of this study is to examine this gap in current research by examining semaglutide, which is the current formulary preferred GLP-1RA at the VA.

Semaglutide is subcutaneously initiated at a dose of 0.25 mg once weekly for 4 weeks to reduce gastrointestinal symptoms, then increased to 0.5 mg weekly. Additional increases to a maintenance dose of 1 mg or 2 mg weekly can occur to achieve glycemic goals. The SUSTAIN-FORTE randomized controlled trial sought to determine whether there was a difference in HbA_1c level reduction and significant weight loss with the 2-mg vs 1-mg dose.⁸ Patients in the trial were taking metformin but needed additional medication to control their HbA_1c. They were not using insulin and may or may not have been taking sulfonylureas prior to semaglutide initiation. Semaglutide 2 mg was found to significantly improve HbA_1c control and promote weight loss compared with semaglutide 1 mg, while maintaining a similar safety profile.

Because this study involved patients who required additional HbA_1c control, although semaglutide reduced HbA_1c, not all patients were able to reduce their other diabetes medications, which depended on the baseline HbA_1c level and the level upon completion of semaglutide titration. Dose reductions for the patients’ other T2DM medications were not reported at trial end. SUSTAIN-FORTE established titration up to semaglutide 2 mg as effective for HbA_1c reduction, although it did not study patients also on insulin.⁸

Insulin is associated with hypoglycemic risk, weight gain, and other AEs.^7,8 This study analyzed whether increasing semaglutide could reduce insulin doses and therefore reduce risk of AEs in patients with T2DM.

Methods

A retrospective, single-center, chart review was conducted at VA Sioux Falls Health Care System (VASFHCS). Data were collected through manual review of VASFHCS electronic medical records. Patients aged ≥ 18 years with active prescriptions for at least once-daily insulin who were initiated on 2-mg weekly dose of semaglutide at the VASFHCS clinical pharmacy practitioner medication management clinic between January 1, 2021, and September 1, 2023, were included. VASFHCS clinical pharmacy practitioners have a scope of practice that allows them to initiate, modify, or discontinue medication therapy within medication management clinics.

The most frequently used prandial insulin at VASFHCS is insulin aspart, and the most frequently used basal insulin is insulin glargine. Patients were retrospectively monitored as they progressed from baseline (the point in time where semaglutide 0.5 mg was initiated) to ≥ 3 months on semaglutide 2-mg therapy. Patients were excluded if they previously used a GLP-1RA or if they were on sliding scale insulin without an exact daily dosage.

The primary endpoint was the percent change in total daily insulin dose from baseline to each dose increase after receiving semaglutide 2 mg for ≥ 3 months. Secondary endpoints included changes in daily prandial insulin dose, daily basal insulin dose, HbA_1c, and number of hypoglycemic events reported. Data collected included age, race, weight, body mass index, total daily prandial insulin dose, total daily basal insulin dose, HbA_1c, and hypoglycemic events reported at the visit when semaglutide was initiated.

Statistical Analysis

The sample size was calculated prior to data collection, and it was determined that for α = .05, 47 patients were needed to achieve 95% power. The primary endpoint was assessed using a paired t test, as were each secondary endpoint. Results with P < .05 were considered statistically significant.

Results

Sixty-two patients were included. The mean HbA_1c level at baseline was 7.7%, the baseline mean prandial and insulin daily doses were 41.5 units and 85.1 units, respectively (Table 1) From baseline to initiation of a semaglutide 1-mg dose, the daily insulin dose changed –5.6% (95% CI, 2.2-14.0; P = .008). From baseline to 2-mg dose initiation daily insulin changed -22.2% (95% CI, 22.0-35.1; P < .001) and for patients receiving semaglutide 2 mg for ≥ 3 months it changed -36.9% (95% CI, 37.4-56.5; P < .001) (Figure).

After receiving the 2-mg dose for ≥ 3 months, the mean daily dose of prandial insulin decreased from 41.5 units to 24.6 units (95% CI, 12.6-21.2; P < .001); mean daily dose of basal insulin decreased from 85.1 units to 52.1 units (95% CI, 23.9-42.0; P < .001); and mean HbA_1c level decreased from 7.7% to 7.1% (95% CI, 0.3-0.8; P < .001). Mean number of hypoglycemic events reported was not statistically significant, changing from 3.6 to 3.2 (95% CI, –0.6 to 0.1; P = .21) (Table 2).

Discussion

This study investigated the effect of subcutaneous semaglutide dose escalation on total daily insulin dose for patients with T2DM. There was a statistically significant decrease in total daily insulin dose from baseline to 1 mg initiation; this decrease continued with further insulin dose reduction seen at the 2-mg dose initiation and additional insulin dose reduction at ≥ 3 months at this dose. It was hypothesized there would be a significant total daily insulin dose reduction at some point, especially when transitioning from the semaglutide 1-mg to the 2-mg dose, based on previous research. ^9,10 The additional reduction in daily insulin dose when continuing on semaglutide 2 mg for ≥ 3 months was an unanticipated but added benefit, showing that if tolerated, maintaining the 2-mg dose will help patients reduce their insulin doses.

In terms of secondary endpoints, there was a statistically significant decrease in mean total daily dose individually for prandial and basal insulin from baseline to ≥ 3 months after semaglutide 2 mg initiation. The change in HbA_1c level was also statistically significant and decreased from baseline, even as insulin doses were reduced. This change in HbA_1c level was expected; previous literature has shown a significant link between improving HbA_1c control when semaglutide doses are increased to 2 mg weekly.¹⁰ Due to having been shown in previous trials, it was expected that HbA_1c levels would decrease even when the insulin doses were being reduced.¹⁰ Insulin dose reduction can potentially be added to the growing evidence of semaglutide benefits. The change in the number of hypoglycemic events was not statistically significant, which was unexpected since previous research show a trend in patients taking GLP-1RAs having fewer hypoglycemic events than those taking insulin.⁶ Further investigation with a larger sample size and prospective trial could determine whether this result is an outlier. In this study, there was no increase in HbA_1c or hypoglycemic events reported with increasing semaglutide doses, which provides further evidence of the safety of semaglutide even at higher doses.

These data suggest that for a patient with T2DM who is already taking insulin, the recommended titration of semaglutide is to start with 0.5 mg and titrate up to a 2-mg subcutaneous weekly dose and to then continue at that dose. As long as the 2-mg dose is tolerated, it will provide patients with the most HbA_1c control and lead to a reduction of their total daily insulin doses according to these results.

Strengths and Limitations

This study compared patient data at different points. This method did not require a second distinct control group, which would potentially introduce confounding factors, such as different baseline characteristics. Another strength is that documentation was available for all patients throughout the study so no one was lost to follow-up. This allowed comprehensive data collection and provided a stronger conclusion given the completeness of the data from baseline to follow-up.

Limitations include the retrospective design and small sample size. In addition, the study design did not allow for randomization. There is no documentation of adherence to medication regimen, which was difficult to determine due to the retrospective nature. Other changes to the patients’ medication regimen were not collected in aggregate and thus, it is possible the total daily insulin dose was impacted by other medication changes. There is also potential for inconsistent documentation of the patients’ true total daily insulin dose in the medical record, thus leading to inaccuracy of recorded data.

Conclusions

A small sample of veterans with T2DM had statistically significant reductions in total daily insulin dose when subcutaneous semaglutide was initiated, as well as after each dose increase. There was also a statistically significant reduction in HbA_1c levels from baseline even as patient insulin doses were reduced. These results support the current practice of using semaglutide to treat T2DM, suggesting it may be safe and effective at reducing HbA_1c levels as the dose is titrated up to 2 mg. There was no statistically significant change in the number of hypoglycemic events reported as semaglutide was titrated up. Thus, when semaglutide is increased to the maximum recommended dose of 2 mg for T2DM, patients may experience a reduction of their total daily dose of insulin and HbA_1c levels. These benefits may reduce the risk of insulin-related AEs while maintaining appropriate glycemic control.

Type 2 diabetes mellitus (T2DM) is a chronic disease becoming more prevalent each year and is the seventh-leading cause of death in the United States.¹ The most common reason for hospitalization for patients with T2DM is uncontrolled glycemic levels.² Nearly 25% of the US Department of Veterans Affairs (VA) patient population has T2DM.³ T2DM is the leading cause of blindness, end-stage renal disease, and amputation for VA patients.⁴

According to the 2023 American Diabetes Association (ADA) guidelines, treatment goals of T2DM include eliminating symptoms, preventing or delaying complications, and attaining glycemic goals. A typical hemoglobin A_1c (HbA_1c) goal range is < 7%, but individual goals can vary up to < 9% due to a multitude of factors, including patient comorbidities and clinical status.⁵

Initial treatment recommendations are nonpharmacologic and include comprehensive lifestyle interventions such as optimizing nutrition, physical activity, and behavioral therapy. When pharmacologic therapy is required, metformin is the preferred first-line treatment for the majority of newly diagnosed patients with T2DM and should be added to continued lifestyle management.⁵ If HbA_1c levels remains above goal, the 2023 ADA guidelines recommend adding a second medication, including but not limited to insulin, a glucagonlike peptide-1 receptor agonist (GLP-1RA), or a sodium-glucose cotransporter 2 inhibitor. Medication choice is largely based on the patient’s concomitant conditions (eg, atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease). The 2023 ADA guidelines suggest initiating insulin therapy when a patient's blood glucose ≥ 300 mg/dL, HbA_1c > 10%, or if the patient has symptoms of hyperglycemia, even at initial diagnosis. Initiating medications to minimize or avoid hypoglycemia is a priority, especially in high-risk individuals.⁵

Clinical evidence shows that GLP-1RAs may provide similar glycemic control to insulin with lower risk of hypoglycemia.⁶ Other reported benefits of GLP-1RAs include weight loss, blood pressure reduction, and improved lipid levels. The most common adverse events (AEs) with GLP-1RAs are gastrointestinal. Including GLP-1RAs in T2DM pharmacotherapy may lower the risk of hypoglycemia, especially in patients at high risk of hypoglycemia.

The 2023 ADA guidelines indicate that it is appropriate to initiate GLP-]1RAs in patients on insulin.⁵ However, while GLP-1RAs do not increase the risk of hypoglycemia independently, combination treatment with GLP-1RAs and insulin can still result in hypoglycemia.⁶ Insulin is the key suspect of this hypoglycemic risk.⁷ Thus, if insulin dosage can be reduced or discontinued, this might reduce the risk of hypoglycemia.

The literature is limited on how the addition of a GLP-1RA to insulin treatment will affect the patient's daily insulin doses, particularly for the veteran population. The goal of this study is to examine this gap in current research by examining semaglutide, which is the current formulary preferred GLP-1RA at the VA.

Semaglutide is subcutaneously initiated at a dose of 0.25 mg once weekly for 4 weeks to reduce gastrointestinal symptoms, then increased to 0.5 mg weekly. Additional increases to a maintenance dose of 1 mg or 2 mg weekly can occur to achieve glycemic goals. The SUSTAIN-FORTE randomized controlled trial sought to determine whether there was a difference in HbA_1c level reduction and significant weight loss with the 2-mg vs 1-mg dose.⁸ Patients in the trial were taking metformin but needed additional medication to control their HbA_1c. They were not using insulin and may or may not have been taking sulfonylureas prior to semaglutide initiation. Semaglutide 2 mg was found to significantly improve HbA_1c control and promote weight loss compared with semaglutide 1 mg, while maintaining a similar safety profile.

Because this study involved patients who required additional HbA_1c control, although semaglutide reduced HbA_1c, not all patients were able to reduce their other diabetes medications, which depended on the baseline HbA_1c level and the level upon completion of semaglutide titration. Dose reductions for the patients’ other T2DM medications were not reported at trial end. SUSTAIN-FORTE established titration up to semaglutide 2 mg as effective for HbA_1c reduction, although it did not study patients also on insulin.⁸

Insulin is associated with hypoglycemic risk, weight gain, and other AEs.^7,8 This study analyzed whether increasing semaglutide could reduce insulin doses and therefore reduce risk of AEs in patients with T2DM.

Methods

A retrospective, single-center, chart review was conducted at VA Sioux Falls Health Care System (VASFHCS). Data were collected through manual review of VASFHCS electronic medical records. Patients aged ≥ 18 years with active prescriptions for at least once-daily insulin who were initiated on 2-mg weekly dose of semaglutide at the VASFHCS clinical pharmacy practitioner medication management clinic between January 1, 2021, and September 1, 2023, were included. VASFHCS clinical pharmacy practitioners have a scope of practice that allows them to initiate, modify, or discontinue medication therapy within medication management clinics.

The most frequently used prandial insulin at VASFHCS is insulin aspart, and the most frequently used basal insulin is insulin glargine. Patients were retrospectively monitored as they progressed from baseline (the point in time where semaglutide 0.5 mg was initiated) to ≥ 3 months on semaglutide 2-mg therapy. Patients were excluded if they previously used a GLP-1RA or if they were on sliding scale insulin without an exact daily dosage.

The primary endpoint was the percent change in total daily insulin dose from baseline to each dose increase after receiving semaglutide 2 mg for ≥ 3 months. Secondary endpoints included changes in daily prandial insulin dose, daily basal insulin dose, HbA_1c, and number of hypoglycemic events reported. Data collected included age, race, weight, body mass index, total daily prandial insulin dose, total daily basal insulin dose, HbA_1c, and hypoglycemic events reported at the visit when semaglutide was initiated.

Statistical Analysis

The sample size was calculated prior to data collection, and it was determined that for α = .05, 47 patients were needed to achieve 95% power. The primary endpoint was assessed using a paired t test, as were each secondary endpoint. Results with P < .05 were considered statistically significant.

Results

Sixty-two patients were included. The mean HbA_1c level at baseline was 7.7%, the baseline mean prandial and insulin daily doses were 41.5 units and 85.1 units, respectively (Table 1) From baseline to initiation of a semaglutide 1-mg dose, the daily insulin dose changed –5.6% (95% CI, 2.2-14.0; P = .008). From baseline to 2-mg dose initiation daily insulin changed -22.2% (95% CI, 22.0-35.1; P < .001) and for patients receiving semaglutide 2 mg for ≥ 3 months it changed -36.9% (95% CI, 37.4-56.5; P < .001) (Figure).

After receiving the 2-mg dose for ≥ 3 months, the mean daily dose of prandial insulin decreased from 41.5 units to 24.6 units (95% CI, 12.6-21.2; P < .001); mean daily dose of basal insulin decreased from 85.1 units to 52.1 units (95% CI, 23.9-42.0; P < .001); and mean HbA_1c level decreased from 7.7% to 7.1% (95% CI, 0.3-0.8; P < .001). Mean number of hypoglycemic events reported was not statistically significant, changing from 3.6 to 3.2 (95% CI, –0.6 to 0.1; P = .21) (Table 2).

Discussion

This study investigated the effect of subcutaneous semaglutide dose escalation on total daily insulin dose for patients with T2DM. There was a statistically significant decrease in total daily insulin dose from baseline to 1 mg initiation; this decrease continued with further insulin dose reduction seen at the 2-mg dose initiation and additional insulin dose reduction at ≥ 3 months at this dose. It was hypothesized there would be a significant total daily insulin dose reduction at some point, especially when transitioning from the semaglutide 1-mg to the 2-mg dose, based on previous research. ^9,10 The additional reduction in daily insulin dose when continuing on semaglutide 2 mg for ≥ 3 months was an unanticipated but added benefit, showing that if tolerated, maintaining the 2-mg dose will help patients reduce their insulin doses.

In terms of secondary endpoints, there was a statistically significant decrease in mean total daily dose individually for prandial and basal insulin from baseline to ≥ 3 months after semaglutide 2 mg initiation. The change in HbA_1c level was also statistically significant and decreased from baseline, even as insulin doses were reduced. This change in HbA_1c level was expected; previous literature has shown a significant link between improving HbA_1c control when semaglutide doses are increased to 2 mg weekly.¹⁰ Due to having been shown in previous trials, it was expected that HbA_1c levels would decrease even when the insulin doses were being reduced.¹⁰ Insulin dose reduction can potentially be added to the growing evidence of semaglutide benefits. The change in the number of hypoglycemic events was not statistically significant, which was unexpected since previous research show a trend in patients taking GLP-1RAs having fewer hypoglycemic events than those taking insulin.⁶ Further investigation with a larger sample size and prospective trial could determine whether this result is an outlier. In this study, there was no increase in HbA_1c or hypoglycemic events reported with increasing semaglutide doses, which provides further evidence of the safety of semaglutide even at higher doses.

These data suggest that for a patient with T2DM who is already taking insulin, the recommended titration of semaglutide is to start with 0.5 mg and titrate up to a 2-mg subcutaneous weekly dose and to then continue at that dose. As long as the 2-mg dose is tolerated, it will provide patients with the most HbA_1c control and lead to a reduction of their total daily insulin doses according to these results.

Strengths and Limitations

This study compared patient data at different points. This method did not require a second distinct control group, which would potentially introduce confounding factors, such as different baseline characteristics. Another strength is that documentation was available for all patients throughout the study so no one was lost to follow-up. This allowed comprehensive data collection and provided a stronger conclusion given the completeness of the data from baseline to follow-up.

Limitations include the retrospective design and small sample size. In addition, the study design did not allow for randomization. There is no documentation of adherence to medication regimen, which was difficult to determine due to the retrospective nature. Other changes to the patients’ medication regimen were not collected in aggregate and thus, it is possible the total daily insulin dose was impacted by other medication changes. There is also potential for inconsistent documentation of the patients’ true total daily insulin dose in the medical record, thus leading to inaccuracy of recorded data.

Conclusions

A small sample of veterans with T2DM had statistically significant reductions in total daily insulin dose when subcutaneous semaglutide was initiated, as well as after each dose increase. There was also a statistically significant reduction in HbA_1c levels from baseline even as patient insulin doses were reduced. These results support the current practice of using semaglutide to treat T2DM, suggesting it may be safe and effective at reducing HbA_1c levels as the dose is titrated up to 2 mg. There was no statistically significant change in the number of hypoglycemic events reported as semaglutide was titrated up. Thus, when semaglutide is increased to the maximum recommended dose of 2 mg for T2DM, patients may experience a reduction of their total daily dose of insulin and HbA_1c levels. These benefits may reduce the risk of insulin-related AEs while maintaining appropriate glycemic control.

Diabetes mellitus (DM) is a national health crisis affecting > 38 million people (11.6%) in the United States.¹ American Indian and Alaska Native (AI/AN) adults are disproportionately affected, with a prevalence of 14.5%—the highest among all racial and ethnic groups.¹ Type 2 DM (T2DM) accounts for 90% to 95% of all DM cases and is a leading cause of morbidity and mortality due to its association with cardiovascular disease, kidney failure, and other complications.²

Maintaining glycemic control is important for managing T2DM and preventing microvascular and macrovascular complications.³ The cornerstone of diabetes self-management has been patient self-monitored blood glucose (SMBG) using finger-stick glucometers.⁴ However, SMBG provides measurements from a single point in time and requires frequent, painful, and inconvenient finger pricks, leading to decreased adherence.^5,6 These limitations negatively affect patient engagement and overall glycemic control.⁷

Continuous glucose monitors (CGMs) offer real-time, continuous glucose readings and trends.⁸ CGMs improve glycemic control and reduce hypoglycemic episodes in patients who are insulin-dependent.^9,10 Flash glucose monitors, a type of CGM that requires scanning to obtain glucose readings, provide similar benefits.¹¹ Despite these demonstrated advantages, research has primarily focused on insulin-dependent populations, leaving a significant gap in understanding the effect of CGMs on patients with T2DM who are not insulin-dependent.¹²

Given the high prevalence of T2DM among AI/AN populations and the potential benefits of CGMs, this study sought to evaluate the effect of CGM use on glycemic control and other health metrics in patients with non–insulin-dependent T2DM in an AI/AN population. This focus addresses a critical knowledge gap and may inform clinical practices and policies to improve diabetes management in this high-risk group.

Methods

A retrospective observational study was conducted using deidentified electronic health records (EHRs) from 2019 to 2024 at a federally operated outpatient Indian Health Service (IHS) clinic serving an AI/AN population in the IHS Portland Area (Oregon, Washington, Idaho). The study protocol was reviewed and deemed exempt by institutional review boards at Washington State University and the Portland Area IHS.

Study Population

This study included patients diagnosed with non–insulin-dependent T2DM, had used a CGM for ≥ 1 year, and had hemoglobin A_1c (HbA_1c) measurements within 4 months prior to CGM initiation (baseline) and within ± 4 months after 1 year of CGM use. For other health metrics, including blood pressure (BP), weight, low-density lipoprotein cholesterol (LDL-C), and estimated glomerular filtration rate (eGFR), this study required measurements within 6 months before CGM initiation and within 6 months after 1 year of CGM use. The baseline HbA_1c in the dataset ranged from 5.3% to > 14%.

Patients were excluded if they used insulin during the study period, had incomplete laboratory or clinical data for the required time frame, or had < 1 year of CGM use. The dataset did not include detailed information on oral DM medications; thus, we could not report or account for the type or number of oral hypoglycemic agents used by the patients. The IHS clinical applications coordinator compiled the dataset from the EHR, identifying patients who were prescribed and received a CGM at the clinic. All patients used the Abbott Freestyle Libre CGM, the only formulary CGM available at the clinic during the study period.

A 1-year follow-up endpoint was selected for several reasons: (1) to capture potential seasonal variations in diet and activity; (2) to align with the clinic’s standard practice of annual comprehensive diabetes evaluations; and (3) to allow sufficient time for patients to adapt to CGM use and reflect any meaningful changes in glycemic control.

All patients received standard DM care according to clinic protocols, which included DM self-management education and training. Patients met with the diabetes educator at least once, during which the educator emphasized making informed decisions using CGM data, such as adjusting dietary choices and physical activity levels to manage blood glucose concentrations effectively.

A total of 302 patients were initially identified. After applying exclusion criteria, 132 were excluded due to insulin use, and 77 were excluded due to incomplete HbA_1c data within the specified time frames (Figure 1). The final sample included 93 patients.

Measures

The primary outcome was the change in HbA_1c levels from baseline to 1 year after CGM initiation. Secondary outcomes included changes in weight, systolic and diastolic BP, LDL-C concentrations, and eGFR. For the primary outcome, HbA_1c values were collected within a grace period of ± 4 months from the baseline and 1-year time points. The laboratory’s upper reporting limit for HbA_1c was 14%; values reported as “> 14%” were recorded as 14.1% for data analysis, although the actual values could have been higher.

For secondary outcomes, data were included if measurements were obtained within ± 6 months of the baseline and 1-year time points. Patients who did not have measurements within these time frames for specific metrics were excluded from secondary outcome analysis but remained in the overall study if they met the criteria for HbA_1c and CGM use.

Statistical Analysis

Statistical analysis was performed using R statistical software version 4.4.2. Paired t tests were conducted to compare baseline and 1-year follow- up measurements for variables with parametric distributions. Wilcoxon signed-rank test was used for nonparametric data. A linear regression analysis was conducted to examine the relationship between baseline HbA_1c levels and the change in HbA_1c after 1 year of CGM use. Differences were considered significant at P < .05 set a priori. To guide future research, a posthoc power analysis was performed using Cohen’s d to estimate the required sample sizes for detecting significant effects, assuming a similar population.

Results

The study included 93 patients, with a mean (SD) age of 55 (13) years (range, 29-83 years). Of the participants, 56 were female (60%) and 37 were male (40%). All participants were identified as AI/AN and had non–insulin-dependent T2DM.

Primary Outcomes

A significant reduction in HbA_1c levels was observed after 1 year of CGM use. The mean (SD) baseline HbA1c was 9.5% (2.4%), which decreased to 7.6% (2.2%) at 1-year follow-up (Table 1). This difference represents a mean change of -1.86% (2.4%) (95% CI, -2.35 to -1.37; P < .001 [paired t test, -7.53]).

A linear regression model evaluated the relationship between baseline HbA_1c (predictor) and the change in HbA_1c after 1 year (outcome). The change in HbA_1c was calculated as the difference between 1-year follow-up and baseline values. The regression model revealed a significant negative association between baseline HbA_1c and the change in HbA_1c (Β = -0.576; P < .001), indicating that higher baseline HbA_1c values were associated with greater reductions in HbA_1c over the year. The regression equation was: Change in HbA_1c = 3.587 – 0.576 × Baseline HbA_1c

The regression coefficient for baseline HbA_1c was -0.576 (standard error, 0.083; t = -6.931; P < .001), indicating that for each 1% increase in baseline HbA_1c, the reduction of HbA_1c after 1 year increased by approximately 0.576% (Figure 2). The model explained 34.6% of the variance in HbA_1c change (R² = .345; adjusted R² = .338).

Secondary Outcomes

Systolic BP decreased by a mean (SD) -4.9 (17) mm Hg; 95% CI, -8.6 to -1.11; P = .01, paired t test). However, no significant change was observed for diastolic BP (P = .77, paired t test). Similarly, no significant changes were observed in weight, LDL-C concentrations, or eGFR after 1 year of CGM use. A posthoc power analysis indicated that the study was underpowered to detect smaller effect sizes in secondary outcomes. For example, sample size estimates indicated that detecting significant changes in weight and LDL-C concentrations would require sample sizes of 152 and 220 patients, respectively (Table 2).

Discussion

This study found a clinically significant reduction in HbA_1c levels after 1 year among AI/AN patients with non–insulin-dependent T2DM who used CGMs. The mean HbA_1c decreased 1.9%, from 9.5% at baseline to 7.6% after 1 year. This reduction is not only statistically significant (P < .001), it is clinically meaningful—even a 1% decrease in HbA_1c is associated with substantial reductions in the risk of microvascular complications.³ The magnitude of the HbA_1c reduction observed suggests CGM use may be associated with improved glycemic control in this high-risk population. By achieving lower HbA_1c levels, patients may experience improved long-term health outcomes and a reduced burden of DM-related complications.

Changes in oral DM medications during the study period may have contributed to the observed improvements in HbA_1c levels. While the dataset lacked detailed information on types or dosages of oral hypoglycemic agents used, adjustments in medication regimens are common in DM management and could significantly affect glycemic control. The inability to account for these changes results in an inability to attribute the improvements in HbA_1c solely to CGM use. Future studies should collect comprehensive medication data to better isolate the effects of CGM use from other treatment modifications.

Another factor that may have contributed to the improved glycemic control is the DM self-management education and training patients received as part of standard care. Patients met with diabetes educators at least once and learned how to use the CGM device and interpret the data for self-management decisions. This education may have enhanced patient engagement and empowerment, enabling them to make informed choices about diet, physical activity, and medication adherence. Studies have shown that DM self-management education can significantly improve glycemic control and patient outcomes.¹³ By combining the CGM technology with targeted education, patients may have been better equipped to manage their condition, contributing to the observed reduction in HbA_1c levels. Future studies should consider synergistic effects of CGM use and DM education when evaluating interventions for glycemic control.

The significant reduction in HbA_1c indicates CGM use is associated with improved glycemic control in non–insulin-dependent T2DM. The linear regression analysis suggests patients with poorer glycemic control at baseline experienced greater reductions in HbA_1c over the course of 1 year. This finding aligns with previous studies that have shown greater HbA_1c reductions in patients with higher initial levels when using CGMs. Yaron et al reported similar findings: higher baseline HbA_1c levels predicted more substantial improvements with CGM use in patients with T2DM on insulin therapy.¹⁴

This study contributes to existing research by examining the association between CGM use and glycemic control in patients with non– insulin-dependent T2DM within an AI/AN population, a group that has been underreported in previous studies. Most prior research has focused on insulin-dependent patients or populations with different ethnic backgrounds.12 By focusing on patients with non–insulin-dependent T2DM, this study highlights the broader applicability of CGMs beyond traditional use, showcasing their potential association with benefits in earlier stages of DM management. Targeting the AI/AN population addresses a critical knowledge gap, given the disproportionately high prevalence of T2DM and associated complications in this group. The findings of this study suggest integrating CGM technology into the standard care of AI/AN patients with non–insulin-dependent T2DM may be associated with improved glycemic control and may help reduce health disparities.

The modest decrease in systolic BP observed in this study may indicate potential cardiovascular benefits associated with CGM use, possibly due to improved glycemic control and increased patient engagement in self-management. However, given the limited sample size and exclusion criteria, the study lacked sufficient power to detect significant associations between CGM use and other secondary outcomes such as BP, weight, LDL-C, and eGFR. Therefore, the significant finding with systolic BP should be interpreted with caution.

The lack of significant changes in secondary outcomes may be attributed to the study’s limited sample size and the relatively short duration for observing changes in these parameters. Larger studies are needed to assess the full impact of CGM on these variables. The required sample sizes for achieving adequate power in future studies were calculated, highlighting the utility of our study as a pilot, providing critical data for the design of larger, adequately powered studies.

Limitations

The retrospective design of this study limits causal inferences. Moreover, potential confounding variables were not controlled, such as changes in medication regimens (other than insulin use), dietary counseling, or physical activity. Additionally, we could not account for the type or number of oral DM medications prescribed to patients. The dataset included only information on insulin use, without detailed records of other antidiabetic medications. This limitation may have influenced the observed change in glycemic control, as variations in medication regimens could affect HbA_1c levels.

Because this study lacked a comparator group, the effect of CGM use cannot be definitively isolated from other factors (eg, medication changes, dietary modifications, or physical activity). Moreover, CGM devices can be costly and are not universally covered by all insurance or IHS programs, potentially limiting widespread implementation. Policy-level restrictions and patient-specific barriers may also hinder feasibility in other settings.

The small sample size may limit the generalizability of the findings. Of the initial 302 patients, about 69% were excluded due to insulin use or incomplete laboratory data. A ± 4-month window was selected to balance data quality with real-world practices. Extending this window further (eg, ± 6 months) might have included more participants but risked diluting the 1-year endpoint consistency. The lack of statistical significance in secondary metrics may be due to insufficient power rather than the absence of an effect.

Exclusion of patients due to incomplete data may have introduced selection bias. However, patients were included in the overall analysis if they met the criteria for HbA_1c and CGM use, even if they lacked data for secondary outcomes. Additionally, the laboratory’s upper reporting limit for HbA_1c was 14%, with values above this reported as “> 14%.” For analysis, these were recorded as 14.1%, which may underestimate the true baseline HbA_1c levels and impact of the assessment of change. This occurred for 4 of the 93 patients included.

All patients used the Freestyle Libre CGM, which may limit the generalizability of the findings to other CGM brands or models. Differences in device features, accuracy, scanning frequency, and user experience may influence outcomes, and results might differ with other CGM technologies. The dataset did not include patients’ scanning frequency because this metric was not consistently included in the EHRs.

Conclusions

This study found that CGM use was significantly associated with improved glycemic control in patients with non–insulin-dependent T2DM within an AI/AN population, particularly among patients with higher baseline HbA_1c levels. The findings suggest that CGMs may be a valuable tool for managing T2DM beyond insulin-dependent populations.

Additional research with larger sample sizes, control groups, and extended follow-up periods is recommended to explore long-term benefits and impacts on other health metrics. The sample size estimates derived from this study serve as a valuable resource for researchers designing future studies aimed at addressing these gaps. Future research that expands on our findings by including larger, more diverse cohorts, accounting for medication use, and exploring different CGM technologies will enhance understanding and contribute to more effective diabetes management strategies for varied populations.

Diabetes mellitus (DM) is a national health crisis affecting > 38 million people (11.6%) in the United States.¹ American Indian and Alaska Native (AI/AN) adults are disproportionately affected, with a prevalence of 14.5%—the highest among all racial and ethnic groups.¹ Type 2 DM (T2DM) accounts for 90% to 95% of all DM cases and is a leading cause of morbidity and mortality due to its association with cardiovascular disease, kidney failure, and other complications.²

Maintaining glycemic control is important for managing T2DM and preventing microvascular and macrovascular complications.³ The cornerstone of diabetes self-management has been patient self-monitored blood glucose (SMBG) using finger-stick glucometers.⁴ However, SMBG provides measurements from a single point in time and requires frequent, painful, and inconvenient finger pricks, leading to decreased adherence.^5,6 These limitations negatively affect patient engagement and overall glycemic control.⁷

Continuous glucose monitors (CGMs) offer real-time, continuous glucose readings and trends.⁸ CGMs improve glycemic control and reduce hypoglycemic episodes in patients who are insulin-dependent.^9,10 Flash glucose monitors, a type of CGM that requires scanning to obtain glucose readings, provide similar benefits.¹¹ Despite these demonstrated advantages, research has primarily focused on insulin-dependent populations, leaving a significant gap in understanding the effect of CGMs on patients with T2DM who are not insulin-dependent.¹²

Given the high prevalence of T2DM among AI/AN populations and the potential benefits of CGMs, this study sought to evaluate the effect of CGM use on glycemic control and other health metrics in patients with non–insulin-dependent T2DM in an AI/AN population. This focus addresses a critical knowledge gap and may inform clinical practices and policies to improve diabetes management in this high-risk group.

Methods

A retrospective observational study was conducted using deidentified electronic health records (EHRs) from 2019 to 2024 at a federally operated outpatient Indian Health Service (IHS) clinic serving an AI/AN population in the IHS Portland Area (Oregon, Washington, Idaho). The study protocol was reviewed and deemed exempt by institutional review boards at Washington State University and the Portland Area IHS.

Study Population

This study included patients diagnosed with non–insulin-dependent T2DM, had used a CGM for ≥ 1 year, and had hemoglobin A_1c (HbA_1c) measurements within 4 months prior to CGM initiation (baseline) and within ± 4 months after 1 year of CGM use. For other health metrics, including blood pressure (BP), weight, low-density lipoprotein cholesterol (LDL-C), and estimated glomerular filtration rate (eGFR), this study required measurements within 6 months before CGM initiation and within 6 months after 1 year of CGM use. The baseline HbA_1c in the dataset ranged from 5.3% to > 14%.

Patients were excluded if they used insulin during the study period, had incomplete laboratory or clinical data for the required time frame, or had < 1 year of CGM use. The dataset did not include detailed information on oral DM medications; thus, we could not report or account for the type or number of oral hypoglycemic agents used by the patients. The IHS clinical applications coordinator compiled the dataset from the EHR, identifying patients who were prescribed and received a CGM at the clinic. All patients used the Abbott Freestyle Libre CGM, the only formulary CGM available at the clinic during the study period.

A 1-year follow-up endpoint was selected for several reasons: (1) to capture potential seasonal variations in diet and activity; (2) to align with the clinic’s standard practice of annual comprehensive diabetes evaluations; and (3) to allow sufficient time for patients to adapt to CGM use and reflect any meaningful changes in glycemic control.

All patients received standard DM care according to clinic protocols, which included DM self-management education and training. Patients met with the diabetes educator at least once, during which the educator emphasized making informed decisions using CGM data, such as adjusting dietary choices and physical activity levels to manage blood glucose concentrations effectively.

A total of 302 patients were initially identified. After applying exclusion criteria, 132 were excluded due to insulin use, and 77 were excluded due to incomplete HbA_1c data within the specified time frames (Figure 1). The final sample included 93 patients.

Measures

The primary outcome was the change in HbA_1c levels from baseline to 1 year after CGM initiation. Secondary outcomes included changes in weight, systolic and diastolic BP, LDL-C concentrations, and eGFR. For the primary outcome, HbA_1c values were collected within a grace period of ± 4 months from the baseline and 1-year time points. The laboratory’s upper reporting limit for HbA_1c was 14%; values reported as “> 14%” were recorded as 14.1% for data analysis, although the actual values could have been higher.

For secondary outcomes, data were included if measurements were obtained within ± 6 months of the baseline and 1-year time points. Patients who did not have measurements within these time frames for specific metrics were excluded from secondary outcome analysis but remained in the overall study if they met the criteria for HbA_1c and CGM use.

Statistical Analysis

Statistical analysis was performed using R statistical software version 4.4.2. Paired t tests were conducted to compare baseline and 1-year follow- up measurements for variables with parametric distributions. Wilcoxon signed-rank test was used for nonparametric data. A linear regression analysis was conducted to examine the relationship between baseline HbA_1c levels and the change in HbA_1c after 1 year of CGM use. Differences were considered significant at P < .05 set a priori. To guide future research, a posthoc power analysis was performed using Cohen’s d to estimate the required sample sizes for detecting significant effects, assuming a similar population.

Results

The study included 93 patients, with a mean (SD) age of 55 (13) years (range, 29-83 years). Of the participants, 56 were female (60%) and 37 were male (40%). All participants were identified as AI/AN and had non–insulin-dependent T2DM.

Primary Outcomes

A significant reduction in HbA_1c levels was observed after 1 year of CGM use. The mean (SD) baseline HbA1c was 9.5% (2.4%), which decreased to 7.6% (2.2%) at 1-year follow-up (Table 1). This difference represents a mean change of -1.86% (2.4%) (95% CI, -2.35 to -1.37; P < .001 [paired t test, -7.53]).

A linear regression model evaluated the relationship between baseline HbA_1c (predictor) and the change in HbA_1c after 1 year (outcome). The change in HbA_1c was calculated as the difference between 1-year follow-up and baseline values. The regression model revealed a significant negative association between baseline HbA_1c and the change in HbA_1c (Β = -0.576; P < .001), indicating that higher baseline HbA_1c values were associated with greater reductions in HbA_1c over the year. The regression equation was: Change in HbA_1c = 3.587 – 0.576 × Baseline HbA_1c

The regression coefficient for baseline HbA_1c was -0.576 (standard error, 0.083; t = -6.931; P < .001), indicating that for each 1% increase in baseline HbA_1c, the reduction of HbA_1c after 1 year increased by approximately 0.576% (Figure 2). The model explained 34.6% of the variance in HbA_1c change (R² = .345; adjusted R² = .338).

Secondary Outcomes

Systolic BP decreased by a mean (SD) -4.9 (17) mm Hg; 95% CI, -8.6 to -1.11; P = .01, paired t test). However, no significant change was observed for diastolic BP (P = .77, paired t test). Similarly, no significant changes were observed in weight, LDL-C concentrations, or eGFR after 1 year of CGM use. A posthoc power analysis indicated that the study was underpowered to detect smaller effect sizes in secondary outcomes. For example, sample size estimates indicated that detecting significant changes in weight and LDL-C concentrations would require sample sizes of 152 and 220 patients, respectively (Table 2).

Discussion

This study found a clinically significant reduction in HbA_1c levels after 1 year among AI/AN patients with non–insulin-dependent T2DM who used CGMs. The mean HbA_1c decreased 1.9%, from 9.5% at baseline to 7.6% after 1 year. This reduction is not only statistically significant (P < .001), it is clinically meaningful—even a 1% decrease in HbA_1c is associated with substantial reductions in the risk of microvascular complications.³ The magnitude of the HbA_1c reduction observed suggests CGM use may be associated with improved glycemic control in this high-risk population. By achieving lower HbA_1c levels, patients may experience improved long-term health outcomes and a reduced burden of DM-related complications.

Changes in oral DM medications during the study period may have contributed to the observed improvements in HbA_1c levels. While the dataset lacked detailed information on types or dosages of oral hypoglycemic agents used, adjustments in medication regimens are common in DM management and could significantly affect glycemic control. The inability to account for these changes results in an inability to attribute the improvements in HbA_1c solely to CGM use. Future studies should collect comprehensive medication data to better isolate the effects of CGM use from other treatment modifications.

Another factor that may have contributed to the improved glycemic control is the DM self-management education and training patients received as part of standard care. Patients met with diabetes educators at least once and learned how to use the CGM device and interpret the data for self-management decisions. This education may have enhanced patient engagement and empowerment, enabling them to make informed choices about diet, physical activity, and medication adherence. Studies have shown that DM self-management education can significantly improve glycemic control and patient outcomes.¹³ By combining the CGM technology with targeted education, patients may have been better equipped to manage their condition, contributing to the observed reduction in HbA_1c levels. Future studies should consider synergistic effects of CGM use and DM education when evaluating interventions for glycemic control.

The significant reduction in HbA_1c indicates CGM use is associated with improved glycemic control in non–insulin-dependent T2DM. The linear regression analysis suggests patients with poorer glycemic control at baseline experienced greater reductions in HbA_1c over the course of 1 year. This finding aligns with previous studies that have shown greater HbA_1c reductions in patients with higher initial levels when using CGMs. Yaron et al reported similar findings: higher baseline HbA_1c levels predicted more substantial improvements with CGM use in patients with T2DM on insulin therapy.¹⁴

This study contributes to existing research by examining the association between CGM use and glycemic control in patients with non– insulin-dependent T2DM within an AI/AN population, a group that has been underreported in previous studies. Most prior research has focused on insulin-dependent patients or populations with different ethnic backgrounds.12 By focusing on patients with non–insulin-dependent T2DM, this study highlights the broader applicability of CGMs beyond traditional use, showcasing their potential association with benefits in earlier stages of DM management. Targeting the AI/AN population addresses a critical knowledge gap, given the disproportionately high prevalence of T2DM and associated complications in this group. The findings of this study suggest integrating CGM technology into the standard care of AI/AN patients with non–insulin-dependent T2DM may be associated with improved glycemic control and may help reduce health disparities.

The modest decrease in systolic BP observed in this study may indicate potential cardiovascular benefits associated with CGM use, possibly due to improved glycemic control and increased patient engagement in self-management. However, given the limited sample size and exclusion criteria, the study lacked sufficient power to detect significant associations between CGM use and other secondary outcomes such as BP, weight, LDL-C, and eGFR. Therefore, the significant finding with systolic BP should be interpreted with caution.

The lack of significant changes in secondary outcomes may be attributed to the study’s limited sample size and the relatively short duration for observing changes in these parameters. Larger studies are needed to assess the full impact of CGM on these variables. The required sample sizes for achieving adequate power in future studies were calculated, highlighting the utility of our study as a pilot, providing critical data for the design of larger, adequately powered studies.

Limitations

The retrospective design of this study limits causal inferences. Moreover, potential confounding variables were not controlled, such as changes in medication regimens (other than insulin use), dietary counseling, or physical activity. Additionally, we could not account for the type or number of oral DM medications prescribed to patients. The dataset included only information on insulin use, without detailed records of other antidiabetic medications. This limitation may have influenced the observed change in glycemic control, as variations in medication regimens could affect HbA_1c levels.

Because this study lacked a comparator group, the effect of CGM use cannot be definitively isolated from other factors (eg, medication changes, dietary modifications, or physical activity). Moreover, CGM devices can be costly and are not universally covered by all insurance or IHS programs, potentially limiting widespread implementation. Policy-level restrictions and patient-specific barriers may also hinder feasibility in other settings.

The small sample size may limit the generalizability of the findings. Of the initial 302 patients, about 69% were excluded due to insulin use or incomplete laboratory data. A ± 4-month window was selected to balance data quality with real-world practices. Extending this window further (eg, ± 6 months) might have included more participants but risked diluting the 1-year endpoint consistency. The lack of statistical significance in secondary metrics may be due to insufficient power rather than the absence of an effect.

Exclusion of patients due to incomplete data may have introduced selection bias. However, patients were included in the overall analysis if they met the criteria for HbA_1c and CGM use, even if they lacked data for secondary outcomes. Additionally, the laboratory’s upper reporting limit for HbA_1c was 14%, with values above this reported as “> 14%.” For analysis, these were recorded as 14.1%, which may underestimate the true baseline HbA_1c levels and impact of the assessment of change. This occurred for 4 of the 93 patients included.

All patients used the Freestyle Libre CGM, which may limit the generalizability of the findings to other CGM brands or models. Differences in device features, accuracy, scanning frequency, and user experience may influence outcomes, and results might differ with other CGM technologies. The dataset did not include patients’ scanning frequency because this metric was not consistently included in the EHRs.

Conclusions

This study found that CGM use was significantly associated with improved glycemic control in patients with non–insulin-dependent T2DM within an AI/AN population, particularly among patients with higher baseline HbA_1c levels. The findings suggest that CGMs may be a valuable tool for managing T2DM beyond insulin-dependent populations.

Additional research with larger sample sizes, control groups, and extended follow-up periods is recommended to explore long-term benefits and impacts on other health metrics. The sample size estimates derived from this study serve as a valuable resource for researchers designing future studies aimed at addressing these gaps. Future research that expands on our findings by including larger, more diverse cohorts, accounting for medication use, and exploring different CGM technologies will enhance understanding and contribute to more effective diabetes management strategies for varied populations.

Diabetes mellitus (DM) is a national health crisis affecting > 38 million people (11.6%) in the United States.¹ American Indian and Alaska Native (AI/AN) adults are disproportionately affected, with a prevalence of 14.5%—the highest among all racial and ethnic groups.¹ Type 2 DM (T2DM) accounts for 90% to 95% of all DM cases and is a leading cause of morbidity and mortality due to its association with cardiovascular disease, kidney failure, and other complications.²

Maintaining glycemic control is important for managing T2DM and preventing microvascular and macrovascular complications.³ The cornerstone of diabetes self-management has been patient self-monitored blood glucose (SMBG) using finger-stick glucometers.⁴ However, SMBG provides measurements from a single point in time and requires frequent, painful, and inconvenient finger pricks, leading to decreased adherence.^5,6 These limitations negatively affect patient engagement and overall glycemic control.⁷

Continuous glucose monitors (CGMs) offer real-time, continuous glucose readings and trends.⁸ CGMs improve glycemic control and reduce hypoglycemic episodes in patients who are insulin-dependent.^9,10 Flash glucose monitors, a type of CGM that requires scanning to obtain glucose readings, provide similar benefits.¹¹ Despite these demonstrated advantages, research has primarily focused on insulin-dependent populations, leaving a significant gap in understanding the effect of CGMs on patients with T2DM who are not insulin-dependent.¹²

Given the high prevalence of T2DM among AI/AN populations and the potential benefits of CGMs, this study sought to evaluate the effect of CGM use on glycemic control and other health metrics in patients with non–insulin-dependent T2DM in an AI/AN population. This focus addresses a critical knowledge gap and may inform clinical practices and policies to improve diabetes management in this high-risk group.

Methods

A retrospective observational study was conducted using deidentified electronic health records (EHRs) from 2019 to 2024 at a federally operated outpatient Indian Health Service (IHS) clinic serving an AI/AN population in the IHS Portland Area (Oregon, Washington, Idaho). The study protocol was reviewed and deemed exempt by institutional review boards at Washington State University and the Portland Area IHS.

Study Population

This study included patients diagnosed with non–insulin-dependent T2DM, had used a CGM for ≥ 1 year, and had hemoglobin A_1c (HbA_1c) measurements within 4 months prior to CGM initiation (baseline) and within ± 4 months after 1 year of CGM use. For other health metrics, including blood pressure (BP), weight, low-density lipoprotein cholesterol (LDL-C), and estimated glomerular filtration rate (eGFR), this study required measurements within 6 months before CGM initiation and within 6 months after 1 year of CGM use. The baseline HbA_1c in the dataset ranged from 5.3% to > 14%.

Patients were excluded if they used insulin during the study period, had incomplete laboratory or clinical data for the required time frame, or had < 1 year of CGM use. The dataset did not include detailed information on oral DM medications; thus, we could not report or account for the type or number of oral hypoglycemic agents used by the patients. The IHS clinical applications coordinator compiled the dataset from the EHR, identifying patients who were prescribed and received a CGM at the clinic. All patients used the Abbott Freestyle Libre CGM, the only formulary CGM available at the clinic during the study period.

A 1-year follow-up endpoint was selected for several reasons: (1) to capture potential seasonal variations in diet and activity; (2) to align with the clinic’s standard practice of annual comprehensive diabetes evaluations; and (3) to allow sufficient time for patients to adapt to CGM use and reflect any meaningful changes in glycemic control.

All patients received standard DM care according to clinic protocols, which included DM self-management education and training. Patients met with the diabetes educator at least once, during which the educator emphasized making informed decisions using CGM data, such as adjusting dietary choices and physical activity levels to manage blood glucose concentrations effectively.

A total of 302 patients were initially identified. After applying exclusion criteria, 132 were excluded due to insulin use, and 77 were excluded due to incomplete HbA_1c data within the specified time frames (Figure 1). The final sample included 93 patients.

Measures

The primary outcome was the change in HbA_1c levels from baseline to 1 year after CGM initiation. Secondary outcomes included changes in weight, systolic and diastolic BP, LDL-C concentrations, and eGFR. For the primary outcome, HbA_1c values were collected within a grace period of ± 4 months from the baseline and 1-year time points. The laboratory’s upper reporting limit for HbA_1c was 14%; values reported as “> 14%” were recorded as 14.1% for data analysis, although the actual values could have been higher.

For secondary outcomes, data were included if measurements were obtained within ± 6 months of the baseline and 1-year time points. Patients who did not have measurements within these time frames for specific metrics were excluded from secondary outcome analysis but remained in the overall study if they met the criteria for HbA_1c and CGM use.

Statistical Analysis

Statistical analysis was performed using R statistical software version 4.4.2. Paired t tests were conducted to compare baseline and 1-year follow- up measurements for variables with parametric distributions. Wilcoxon signed-rank test was used for nonparametric data. A linear regression analysis was conducted to examine the relationship between baseline HbA_1c levels and the change in HbA_1c after 1 year of CGM use. Differences were considered significant at P < .05 set a priori. To guide future research, a posthoc power analysis was performed using Cohen’s d to estimate the required sample sizes for detecting significant effects, assuming a similar population.

Results

The study included 93 patients, with a mean (SD) age of 55 (13) years (range, 29-83 years). Of the participants, 56 were female (60%) and 37 were male (40%). All participants were identified as AI/AN and had non–insulin-dependent T2DM.

Primary Outcomes

A significant reduction in HbA_1c levels was observed after 1 year of CGM use. The mean (SD) baseline HbA1c was 9.5% (2.4%), which decreased to 7.6% (2.2%) at 1-year follow-up (Table 1). This difference represents a mean change of -1.86% (2.4%) (95% CI, -2.35 to -1.37; P < .001 [paired t test, -7.53]).

A linear regression model evaluated the relationship between baseline HbA_1c (predictor) and the change in HbA_1c after 1 year (outcome). The change in HbA_1c was calculated as the difference between 1-year follow-up and baseline values. The regression model revealed a significant negative association between baseline HbA_1c and the change in HbA_1c (Β = -0.576; P < .001), indicating that higher baseline HbA_1c values were associated with greater reductions in HbA_1c over the year. The regression equation was: Change in HbA_1c = 3.587 – 0.576 × Baseline HbA_1c

The regression coefficient for baseline HbA_1c was -0.576 (standard error, 0.083; t = -6.931; P < .001), indicating that for each 1% increase in baseline HbA_1c, the reduction of HbA_1c after 1 year increased by approximately 0.576% (Figure 2). The model explained 34.6% of the variance in HbA_1c change (R² = .345; adjusted R² = .338).

Secondary Outcomes

Systolic BP decreased by a mean (SD) -4.9 (17) mm Hg; 95% CI, -8.6 to -1.11; P = .01, paired t test). However, no significant change was observed for diastolic BP (P = .77, paired t test). Similarly, no significant changes were observed in weight, LDL-C concentrations, or eGFR after 1 year of CGM use. A posthoc power analysis indicated that the study was underpowered to detect smaller effect sizes in secondary outcomes. For example, sample size estimates indicated that detecting significant changes in weight and LDL-C concentrations would require sample sizes of 152 and 220 patients, respectively (Table 2).

Discussion

This study found a clinically significant reduction in HbA_1c levels after 1 year among AI/AN patients with non–insulin-dependent T2DM who used CGMs. The mean HbA_1c decreased 1.9%, from 9.5% at baseline to 7.6% after 1 year. This reduction is not only statistically significant (P < .001), it is clinically meaningful—even a 1% decrease in HbA_1c is associated with substantial reductions in the risk of microvascular complications.³ The magnitude of the HbA_1c reduction observed suggests CGM use may be associated with improved glycemic control in this high-risk population. By achieving lower HbA_1c levels, patients may experience improved long-term health outcomes and a reduced burden of DM-related complications.

Changes in oral DM medications during the study period may have contributed to the observed improvements in HbA_1c levels. While the dataset lacked detailed information on types or dosages of oral hypoglycemic agents used, adjustments in medication regimens are common in DM management and could significantly affect glycemic control. The inability to account for these changes results in an inability to attribute the improvements in HbA_1c solely to CGM use. Future studies should collect comprehensive medication data to better isolate the effects of CGM use from other treatment modifications.

Another factor that may have contributed to the improved glycemic control is the DM self-management education and training patients received as part of standard care. Patients met with diabetes educators at least once and learned how to use the CGM device and interpret the data for self-management decisions. This education may have enhanced patient engagement and empowerment, enabling them to make informed choices about diet, physical activity, and medication adherence. Studies have shown that DM self-management education can significantly improve glycemic control and patient outcomes.¹³ By combining the CGM technology with targeted education, patients may have been better equipped to manage their condition, contributing to the observed reduction in HbA_1c levels. Future studies should consider synergistic effects of CGM use and DM education when evaluating interventions for glycemic control.

The significant reduction in HbA_1c indicates CGM use is associated with improved glycemic control in non–insulin-dependent T2DM. The linear regression analysis suggests patients with poorer glycemic control at baseline experienced greater reductions in HbA_1c over the course of 1 year. This finding aligns with previous studies that have shown greater HbA_1c reductions in patients with higher initial levels when using CGMs. Yaron et al reported similar findings: higher baseline HbA_1c levels predicted more substantial improvements with CGM use in patients with T2DM on insulin therapy.¹⁴

This study contributes to existing research by examining the association between CGM use and glycemic control in patients with non– insulin-dependent T2DM within an AI/AN population, a group that has been underreported in previous studies. Most prior research has focused on insulin-dependent patients or populations with different ethnic backgrounds.12 By focusing on patients with non–insulin-dependent T2DM, this study highlights the broader applicability of CGMs beyond traditional use, showcasing their potential association with benefits in earlier stages of DM management. Targeting the AI/AN population addresses a critical knowledge gap, given the disproportionately high prevalence of T2DM and associated complications in this group. The findings of this study suggest integrating CGM technology into the standard care of AI/AN patients with non–insulin-dependent T2DM may be associated with improved glycemic control and may help reduce health disparities.

The modest decrease in systolic BP observed in this study may indicate potential cardiovascular benefits associated with CGM use, possibly due to improved glycemic control and increased patient engagement in self-management. However, given the limited sample size and exclusion criteria, the study lacked sufficient power to detect significant associations between CGM use and other secondary outcomes such as BP, weight, LDL-C, and eGFR. Therefore, the significant finding with systolic BP should be interpreted with caution.

The lack of significant changes in secondary outcomes may be attributed to the study’s limited sample size and the relatively short duration for observing changes in these parameters. Larger studies are needed to assess the full impact of CGM on these variables. The required sample sizes for achieving adequate power in future studies were calculated, highlighting the utility of our study as a pilot, providing critical data for the design of larger, adequately powered studies.

Limitations

The retrospective design of this study limits causal inferences. Moreover, potential confounding variables were not controlled, such as changes in medication regimens (other than insulin use), dietary counseling, or physical activity. Additionally, we could not account for the type or number of oral DM medications prescribed to patients. The dataset included only information on insulin use, without detailed records of other antidiabetic medications. This limitation may have influenced the observed change in glycemic control, as variations in medication regimens could affect HbA_1c levels.

Because this study lacked a comparator group, the effect of CGM use cannot be definitively isolated from other factors (eg, medication changes, dietary modifications, or physical activity). Moreover, CGM devices can be costly and are not universally covered by all insurance or IHS programs, potentially limiting widespread implementation. Policy-level restrictions and patient-specific barriers may also hinder feasibility in other settings.

The small sample size may limit the generalizability of the findings. Of the initial 302 patients, about 69% were excluded due to insulin use or incomplete laboratory data. A ± 4-month window was selected to balance data quality with real-world practices. Extending this window further (eg, ± 6 months) might have included more participants but risked diluting the 1-year endpoint consistency. The lack of statistical significance in secondary metrics may be due to insufficient power rather than the absence of an effect.

Exclusion of patients due to incomplete data may have introduced selection bias. However, patients were included in the overall analysis if they met the criteria for HbA_1c and CGM use, even if they lacked data for secondary outcomes. Additionally, the laboratory’s upper reporting limit for HbA_1c was 14%, with values above this reported as “> 14%.” For analysis, these were recorded as 14.1%, which may underestimate the true baseline HbA_1c levels and impact of the assessment of change. This occurred for 4 of the 93 patients included.

All patients used the Freestyle Libre CGM, which may limit the generalizability of the findings to other CGM brands or models. Differences in device features, accuracy, scanning frequency, and user experience may influence outcomes, and results might differ with other CGM technologies. The dataset did not include patients’ scanning frequency because this metric was not consistently included in the EHRs.

Conclusions

This study found that CGM use was significantly associated with improved glycemic control in patients with non–insulin-dependent T2DM within an AI/AN population, particularly among patients with higher baseline HbA_1c levels. The findings suggest that CGMs may be a valuable tool for managing T2DM beyond insulin-dependent populations.

Additional research with larger sample sizes, control groups, and extended follow-up periods is recommended to explore long-term benefits and impacts on other health metrics. The sample size estimates derived from this study serve as a valuable resource for researchers designing future studies aimed at addressing these gaps. Future research that expands on our findings by including larger, more diverse cohorts, accounting for medication use, and exploring different CGM technologies will enhance understanding and contribute to more effective diabetes management strategies for varied populations.

Cardiovascular disease (CVD) is the leading cause of death among women in the United States.¹ Most CVD is due to the buildup of plaque (ie, cholesterol, proteins, calcium, and inflammatory cells) in artery walls.² The plaque may lead to atherosclerotic cardiovascular disease (ASCVD), which includes coronary heart disease, cerebrovascular disease, peripheral artery disease, and aortic atherosclerotic disease.^2,3 Control and reduction of ASCVD risk factors, including high cholesterol levels, elevated blood pressure, insulin resistance, smoking, and a sedentary lifestyle, can contribute to a reduction in ASCVD morbidity and mortality.² People with type 2 diabetes mellitus (T2DM) have an increased prevalence of lipid abnormalities, contributing to their high risk of ASCVD.^4,5

The prescribing of statins (3-hydroxy-3-methyl-glutaryl-coenzmye A reductase inhibitors) is the cornerstone of lipid-lowering therapy and cardiovascular risk reduction for primary and secondary prevention of ASCVD.⁶ The American Diabetes Association (ADA) and American College of Cardiology/American Heart Association (ACC/AHA) recommend moderate- to high-intensity statins for primary prevention in patients with T2DM and high-intensity statins for secondary prevention in those with or without diabetes when not contraindicated.^4,5,7 Despite eligibility according to guideline recommendations, research predominantly shows that women are less likely to receive statin therapy; however, this trend is improving. ^[6,8-11] To explain the sex differences in statin use, Nanna et al found that there is a combination of women being offered statin therapy less frequently, declining therapy more frequently, and discontinuing treatment more frequently.¹¹ One possibility for discontinuing treatment could be statin-associated muscle symptoms (SAMS), which occur in about 10% of patients.¹² The incidence of adverse effects (AEs) may be related to the way statins are metabolized.

Pharmacogenomic testing is free for veterans through the US Department of Veterans Affairs (VA) PHASER program, which offers information and recommendations for a panel of 11 gene variants. The panel includes genes related to common medication classes such as anticoagulants, antiplatelets, proton pump inhibitors, nonsteroidal anti-inflammatory drugs, opioids, antidepressants, and statins. The VA PHASER panel includes the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene, which is predominantly expressed in the liver and facilitates the hepatic uptake of most statins.^13,14 A reduced function of SLCO1B1 can lead to higher statin levels, resulting in increased concentrations that may potentially cause SAMS.^13,14 Some alleles associated with reduced function include SLCO1B1*5, *15, *23, *31, and *46 to *49, whereas others are associated with increased function, such as SLCO1B1 *14 and *20 (Appendix).¹⁵ Supporting evidence shows the SLCO1B1*5 nucleotide polymorphism increases plasma levels of simvastatin and atorvastatin, affecting effectiveness or toxicity. ¹³ Females tend to have a lower body weight and higher percentage of body fat compared with males, which might lead to higher concentrations of lipophilic drugs, including atorvastatin and simvastatin, which may be exacerbated by decreased function of SLCO1B1*5.¹⁵ With pharmacogenomic testing, therapeutic recommendations can be made to improve the overall safety and efficacy of statins, thus improving adherence using a patient-specific approach.^14,15

Methods

Carl Vinson VA Medical Center (CVVAMC) serves about 42,000 veterans in Central and South Georgia, of which about 15% are female. Of the female veterans enrolled in care, 63% identify as Black, 27% White, and 1.5% as Asian, American Indian/Alaska Native, or Native Hawaiian/Other Pacific Islander. The 2020 Veterans Chartbook report showed that female veterans and minority racial and ethnic groups had worse access to health care and higher mortality rates than their male and non-Hispanic White counterparts.¹⁶

The Primary Care Equity Dashboard (PCED) was developed to engage the VA health care workforce in the process of identifying and addressing inequities in local patient populations.¹⁷ Using electronic quality measure data, the PCED provides Veterans Integrated Service Network-level and facility-level performance on several metrics.¹⁸ The PCED had not been previously used at the CVVAMC, and few publications or quality improvement projects regarding its use have been reported by the VA Office of Health Equity. PCED helped identify disparities when comparing female to male patients in the prescribing of statin therapy for patients with CVD and statin therapy for patients with T2DM.

VA PHASER pharmacogenomic analyses provided an opportunity to expand this quality improvement project. Sanford Health and the VA collaborated on the PHASER program to offer free genetic testing for veterans. The program launched in 2019 and expanded to various VA sites, including CVVAMC in March 2023. This program has been extended to December 31, 2025.

The primary objective of this quality improvement project was to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk. Secondary outcomes included increased pharmacogenomic testing and the assessment of pharmacogenomic results related to statin therapy. This project was approved by the CVVAMC Pharmacy and Therapeutics Committee. The PCED was used to identify female veterans with T2DM and/or CVD without an active prescription for a statin between July and October 2023. A review of Computerized Patient Record System patient charts was completed to screen for prespecified inclusion and exclusion criteria. Veterans were included if they were assigned female at birth, were enrolled in care at CVVAMC, and had a diagnosis of T2DM or CVD (history of myocardial infarction, coronary bypass graft, percutaneous coronary intervention, or other revascularization in any setting).

Veterans were excluded if they were currently pregnant, trying to conceive, breastfeeding, had a T1DM diagnosis, had previously documented hypersensitivity to a statin, active liver failure or decompensated cirrhosis, previously documented statin-associated rhabdomyolysis or autoimmune myopathy, an active prescription for a proprotein convertase subtilisin/kexin type 9 inhibitor, or previously documented statin intolerance (defined as the inability to tolerate ≥ 3 statins, with ≥ 1 prescribed at low intensity or alternate-day dosing). The female veterans were compared to 2 comparators: the facility's male veterans and the VA national average, identified via the PCED.

Once a veteran was screened, they were telephoned between October 2023 and February 2024 and provided education on statin use and pharmacogenomic testing using a standardized note template. An order was placed for participants who provided verbal consent for pharmacogenomic testing. Those who agreed to statin initiation were referred to a clinical pharmacist practitioner (CPP) who contacted them at a later date to prescribe a statin following the recommendations of the 2019 ACC/AHA and 2023 ADA guidelines and pharmacogenomic testing, if applicable.^4,5,7 Appropriate monitoring and follow-up occurred at the discretion of each CPP. Data collection included: age, race, diagnoses (T2DM, CVD, or both), baseline lipid panel (total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein), hepatic function, name and dose of statin, reasons for declining statin therapy, and pharmacogenomic testing results related to SLCO1B1.

Results

At baseline in July 2023, 77.8% of female veterans with T2DM were prescribed a statin, which exceeded the national VA average (77.0%), but was below the rate for male veterans (78.7%) in the facility comparator group.17 Additionally, 82.2% of females with CVD were prescribed a statin, which was below the national VA average of 86.0% and the 84.9% of male veterans in the facility comparator group.17 The PCED identified 189 female veterans from July 2023 to October 2023 who may benefit from statin therapy. Thirty-three females met the exclusion criteria. Of the 156 included veterans, 129 (82.7%) were successfully contacted and 27 (17.3%) could not be reached by telephone after 3 attempts (Figure 1). The 129 female veterans contacted had a mean age of 59 years and the majority were Black (82.9%) (Table 1).

Primary Outcomes

Of the 129 contacted veterans, 31 (24.0%) had a non-VA statin prescription, 13 (10.1%) had an active VA statin prescription, and 85 (65.9%) did not have a statin prescription, despite being eligible. Statin adherence was confirmed with participants, and the medication list was updated accordingly.

Of the 85 veterans with no active statin therapy, 37 (43.5%) accepted a new statin prescription and 48 (56.5%) declined. There were various reasons provided for declining statin therapy: 17 participants (35.4%) declined due to concern for AEs (Table 2).

From July 2023 to March 2024, the percentage of female veterans with active statin therapy with T2DM increased from 77.8% to 79.0%. For those with active statin therapy with CVD, usage increased from 82.2% to 90.2%, which exceeded the national VA average and facility male comparator group (Figures 2 and 3).¹⁷

Secondary Outcomes

Seventy-one of 129 veterans (55.0%) gave verbal consent, and 47 (66.2%) completed the pharmacogenomic testing; 58 (45.0%) declined. Five veterans (10.6%) had a known SLCO1B1 allele variant present. One veteran required a change in statin therapy based on the results (eAppendix).

Discussion

This project aimed to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk and increase pharmacogenomic testing using the PCED and care managed by CPPs. The results of this quality improvement project illustrated that both metrics have improved at CVVAMC as a result of the intervention. The results in both metrics now exceed the PCED national VA average, and the CVD metric also exceeds that of the facility male comparator group. While there was only a 1.2% increase from July 2023 to March 2024 for patients with T2DM, there was an 8.0% increase for patients with CVD. Despite standardized education on statin use, more veterans declined therapy than accepted it, mostly due to concern for AEs. Recording the reasons for declining statin therapy offered valuable insight that can be used in additional discussions with veterans and clinicians.

Pharmacogenomics gives clinicians the unique opportunity to take a proactive approach to better predict drug responses, potentially allowing for less trial and error with medications, fewer AEs, greater trust in the clinician, and improved medication adherence. The CPPs incorporated pharmacogenomic testing into their practice, which led to identifying 5 SLCO1B1 gene abnormalities. The PCED served as a powerful tool for advancing equity-focused quality improvement initiatives on a local level and was crucial in prioritizing the detection of veterans potentially receiving suboptimal care.

Limitations

The nature of “cold calls” made it challenging to establish contact for inclusion in this study. An alternative to increase engagement could have been scheduled phone or face-to-face visits. While the use of the PCED was crucial, data did not account for statins listed in the non-VA medication list. All 31 patients with statins prescribed outside the VA had a start date added to provide the most accurate representation of the data moving forward.

Another limitation in this project was its small sample size and population. CVVAMC serves about 6200 female veterans, with roughly 63% identifying as Black. The preponderance of Black individuals (83%) in this project is typical for the female patient population at CVVAMC but may not reflect the demographics of other populations. Other limitations to this project consisted of scheduling conflicts. Appointments for laboratory draws at community-based outpatient clinics were subject to availability, which resulted in some delay in completion of pharmacogenomic testing.

Conclusions

CPPs can help reduce inequity in health care delivery. Increased incorporation of the PCED into regular practice within the VA is recommended to continue addressing sex disparities in statin use, diabetes control, blood pressure management, cancer screenings, and vaccination needs. CVVAMC plans to expand its use through another quality improvement project focused on reducing sex disparities in blood pressure management. Improving educational resources made available to veterans on the importance of statin therapy and potential to mitigate AEs through use of the VA PHASER program also would be helpful. This project successfully improved CVVAMC metrics for female veterans appropriately prescribed statin therapy and increased access to pharmacogenomic testing. Most importantly, it helped close the sex-based gap in CVD risk reduction care.

Cardiovascular disease (CVD) is the leading cause of death among women in the United States.¹ Most CVD is due to the buildup of plaque (ie, cholesterol, proteins, calcium, and inflammatory cells) in artery walls.² The plaque may lead to atherosclerotic cardiovascular disease (ASCVD), which includes coronary heart disease, cerebrovascular disease, peripheral artery disease, and aortic atherosclerotic disease.^2,3 Control and reduction of ASCVD risk factors, including high cholesterol levels, elevated blood pressure, insulin resistance, smoking, and a sedentary lifestyle, can contribute to a reduction in ASCVD morbidity and mortality.² People with type 2 diabetes mellitus (T2DM) have an increased prevalence of lipid abnormalities, contributing to their high risk of ASCVD.^4,5

The prescribing of statins (3-hydroxy-3-methyl-glutaryl-coenzmye A reductase inhibitors) is the cornerstone of lipid-lowering therapy and cardiovascular risk reduction for primary and secondary prevention of ASCVD.⁶ The American Diabetes Association (ADA) and American College of Cardiology/American Heart Association (ACC/AHA) recommend moderate- to high-intensity statins for primary prevention in patients with T2DM and high-intensity statins for secondary prevention in those with or without diabetes when not contraindicated.^4,5,7 Despite eligibility according to guideline recommendations, research predominantly shows that women are less likely to receive statin therapy; however, this trend is improving. ^[6,8-11] To explain the sex differences in statin use, Nanna et al found that there is a combination of women being offered statin therapy less frequently, declining therapy more frequently, and discontinuing treatment more frequently.¹¹ One possibility for discontinuing treatment could be statin-associated muscle symptoms (SAMS), which occur in about 10% of patients.¹² The incidence of adverse effects (AEs) may be related to the way statins are metabolized.

Pharmacogenomic testing is free for veterans through the US Department of Veterans Affairs (VA) PHASER program, which offers information and recommendations for a panel of 11 gene variants. The panel includes genes related to common medication classes such as anticoagulants, antiplatelets, proton pump inhibitors, nonsteroidal anti-inflammatory drugs, opioids, antidepressants, and statins. The VA PHASER panel includes the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene, which is predominantly expressed in the liver and facilitates the hepatic uptake of most statins.^13,14 A reduced function of SLCO1B1 can lead to higher statin levels, resulting in increased concentrations that may potentially cause SAMS.^13,14 Some alleles associated with reduced function include SLCO1B1*5, *15, *23, *31, and *46 to *49, whereas others are associated with increased function, such as SLCO1B1 *14 and *20 (Appendix).¹⁵ Supporting evidence shows the SLCO1B1*5 nucleotide polymorphism increases plasma levels of simvastatin and atorvastatin, affecting effectiveness or toxicity. ¹³ Females tend to have a lower body weight and higher percentage of body fat compared with males, which might lead to higher concentrations of lipophilic drugs, including atorvastatin and simvastatin, which may be exacerbated by decreased function of SLCO1B1*5.¹⁵ With pharmacogenomic testing, therapeutic recommendations can be made to improve the overall safety and efficacy of statins, thus improving adherence using a patient-specific approach.^14,15

Methods

Carl Vinson VA Medical Center (CVVAMC) serves about 42,000 veterans in Central and South Georgia, of which about 15% are female. Of the female veterans enrolled in care, 63% identify as Black, 27% White, and 1.5% as Asian, American Indian/Alaska Native, or Native Hawaiian/Other Pacific Islander. The 2020 Veterans Chartbook report showed that female veterans and minority racial and ethnic groups had worse access to health care and higher mortality rates than their male and non-Hispanic White counterparts.¹⁶

The Primary Care Equity Dashboard (PCED) was developed to engage the VA health care workforce in the process of identifying and addressing inequities in local patient populations.¹⁷ Using electronic quality measure data, the PCED provides Veterans Integrated Service Network-level and facility-level performance on several metrics.¹⁸ The PCED had not been previously used at the CVVAMC, and few publications or quality improvement projects regarding its use have been reported by the VA Office of Health Equity. PCED helped identify disparities when comparing female to male patients in the prescribing of statin therapy for patients with CVD and statin therapy for patients with T2DM.

VA PHASER pharmacogenomic analyses provided an opportunity to expand this quality improvement project. Sanford Health and the VA collaborated on the PHASER program to offer free genetic testing for veterans. The program launched in 2019 and expanded to various VA sites, including CVVAMC in March 2023. This program has been extended to December 31, 2025.

The primary objective of this quality improvement project was to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk. Secondary outcomes included increased pharmacogenomic testing and the assessment of pharmacogenomic results related to statin therapy. This project was approved by the CVVAMC Pharmacy and Therapeutics Committee. The PCED was used to identify female veterans with T2DM and/or CVD without an active prescription for a statin between July and October 2023. A review of Computerized Patient Record System patient charts was completed to screen for prespecified inclusion and exclusion criteria. Veterans were included if they were assigned female at birth, were enrolled in care at CVVAMC, and had a diagnosis of T2DM or CVD (history of myocardial infarction, coronary bypass graft, percutaneous coronary intervention, or other revascularization in any setting).

Veterans were excluded if they were currently pregnant, trying to conceive, breastfeeding, had a T1DM diagnosis, had previously documented hypersensitivity to a statin, active liver failure or decompensated cirrhosis, previously documented statin-associated rhabdomyolysis or autoimmune myopathy, an active prescription for a proprotein convertase subtilisin/kexin type 9 inhibitor, or previously documented statin intolerance (defined as the inability to tolerate ≥ 3 statins, with ≥ 1 prescribed at low intensity or alternate-day dosing). The female veterans were compared to 2 comparators: the facility's male veterans and the VA national average, identified via the PCED.

Once a veteran was screened, they were telephoned between October 2023 and February 2024 and provided education on statin use and pharmacogenomic testing using a standardized note template. An order was placed for participants who provided verbal consent for pharmacogenomic testing. Those who agreed to statin initiation were referred to a clinical pharmacist practitioner (CPP) who contacted them at a later date to prescribe a statin following the recommendations of the 2019 ACC/AHA and 2023 ADA guidelines and pharmacogenomic testing, if applicable.^4,5,7 Appropriate monitoring and follow-up occurred at the discretion of each CPP. Data collection included: age, race, diagnoses (T2DM, CVD, or both), baseline lipid panel (total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein), hepatic function, name and dose of statin, reasons for declining statin therapy, and pharmacogenomic testing results related to SLCO1B1.

Results

At baseline in July 2023, 77.8% of female veterans with T2DM were prescribed a statin, which exceeded the national VA average (77.0%), but was below the rate for male veterans (78.7%) in the facility comparator group.17 Additionally, 82.2% of females with CVD were prescribed a statin, which was below the national VA average of 86.0% and the 84.9% of male veterans in the facility comparator group.17 The PCED identified 189 female veterans from July 2023 to October 2023 who may benefit from statin therapy. Thirty-three females met the exclusion criteria. Of the 156 included veterans, 129 (82.7%) were successfully contacted and 27 (17.3%) could not be reached by telephone after 3 attempts (Figure 1). The 129 female veterans contacted had a mean age of 59 years and the majority were Black (82.9%) (Table 1).

Primary Outcomes

Of the 129 contacted veterans, 31 (24.0%) had a non-VA statin prescription, 13 (10.1%) had an active VA statin prescription, and 85 (65.9%) did not have a statin prescription, despite being eligible. Statin adherence was confirmed with participants, and the medication list was updated accordingly.

Of the 85 veterans with no active statin therapy, 37 (43.5%) accepted a new statin prescription and 48 (56.5%) declined. There were various reasons provided for declining statin therapy: 17 participants (35.4%) declined due to concern for AEs (Table 2).

From July 2023 to March 2024, the percentage of female veterans with active statin therapy with T2DM increased from 77.8% to 79.0%. For those with active statin therapy with CVD, usage increased from 82.2% to 90.2%, which exceeded the national VA average and facility male comparator group (Figures 2 and 3).¹⁷

Secondary Outcomes

Seventy-one of 129 veterans (55.0%) gave verbal consent, and 47 (66.2%) completed the pharmacogenomic testing; 58 (45.0%) declined. Five veterans (10.6%) had a known SLCO1B1 allele variant present. One veteran required a change in statin therapy based on the results (eAppendix).

Discussion

This project aimed to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk and increase pharmacogenomic testing using the PCED and care managed by CPPs. The results of this quality improvement project illustrated that both metrics have improved at CVVAMC as a result of the intervention. The results in both metrics now exceed the PCED national VA average, and the CVD metric also exceeds that of the facility male comparator group. While there was only a 1.2% increase from July 2023 to March 2024 for patients with T2DM, there was an 8.0% increase for patients with CVD. Despite standardized education on statin use, more veterans declined therapy than accepted it, mostly due to concern for AEs. Recording the reasons for declining statin therapy offered valuable insight that can be used in additional discussions with veterans and clinicians.

Pharmacogenomics gives clinicians the unique opportunity to take a proactive approach to better predict drug responses, potentially allowing for less trial and error with medications, fewer AEs, greater trust in the clinician, and improved medication adherence. The CPPs incorporated pharmacogenomic testing into their practice, which led to identifying 5 SLCO1B1 gene abnormalities. The PCED served as a powerful tool for advancing equity-focused quality improvement initiatives on a local level and was crucial in prioritizing the detection of veterans potentially receiving suboptimal care.

Limitations

The nature of “cold calls” made it challenging to establish contact for inclusion in this study. An alternative to increase engagement could have been scheduled phone or face-to-face visits. While the use of the PCED was crucial, data did not account for statins listed in the non-VA medication list. All 31 patients with statins prescribed outside the VA had a start date added to provide the most accurate representation of the data moving forward.

Another limitation in this project was its small sample size and population. CVVAMC serves about 6200 female veterans, with roughly 63% identifying as Black. The preponderance of Black individuals (83%) in this project is typical for the female patient population at CVVAMC but may not reflect the demographics of other populations. Other limitations to this project consisted of scheduling conflicts. Appointments for laboratory draws at community-based outpatient clinics were subject to availability, which resulted in some delay in completion of pharmacogenomic testing.

Conclusions

CPPs can help reduce inequity in health care delivery. Increased incorporation of the PCED into regular practice within the VA is recommended to continue addressing sex disparities in statin use, diabetes control, blood pressure management, cancer screenings, and vaccination needs. CVVAMC plans to expand its use through another quality improvement project focused on reducing sex disparities in blood pressure management. Improving educational resources made available to veterans on the importance of statin therapy and potential to mitigate AEs through use of the VA PHASER program also would be helpful. This project successfully improved CVVAMC metrics for female veterans appropriately prescribed statin therapy and increased access to pharmacogenomic testing. Most importantly, it helped close the sex-based gap in CVD risk reduction care.

Cardiovascular disease (CVD) is the leading cause of death among women in the United States.¹ Most CVD is due to the buildup of plaque (ie, cholesterol, proteins, calcium, and inflammatory cells) in artery walls.² The plaque may lead to atherosclerotic cardiovascular disease (ASCVD), which includes coronary heart disease, cerebrovascular disease, peripheral artery disease, and aortic atherosclerotic disease.^2,3 Control and reduction of ASCVD risk factors, including high cholesterol levels, elevated blood pressure, insulin resistance, smoking, and a sedentary lifestyle, can contribute to a reduction in ASCVD morbidity and mortality.² People with type 2 diabetes mellitus (T2DM) have an increased prevalence of lipid abnormalities, contributing to their high risk of ASCVD.^4,5

The prescribing of statins (3-hydroxy-3-methyl-glutaryl-coenzmye A reductase inhibitors) is the cornerstone of lipid-lowering therapy and cardiovascular risk reduction for primary and secondary prevention of ASCVD.⁶ The American Diabetes Association (ADA) and American College of Cardiology/American Heart Association (ACC/AHA) recommend moderate- to high-intensity statins for primary prevention in patients with T2DM and high-intensity statins for secondary prevention in those with or without diabetes when not contraindicated.^4,5,7 Despite eligibility according to guideline recommendations, research predominantly shows that women are less likely to receive statin therapy; however, this trend is improving. ^[6,8-11] To explain the sex differences in statin use, Nanna et al found that there is a combination of women being offered statin therapy less frequently, declining therapy more frequently, and discontinuing treatment more frequently.¹¹ One possibility for discontinuing treatment could be statin-associated muscle symptoms (SAMS), which occur in about 10% of patients.¹² The incidence of adverse effects (AEs) may be related to the way statins are metabolized.

Pharmacogenomic testing is free for veterans through the US Department of Veterans Affairs (VA) PHASER program, which offers information and recommendations for a panel of 11 gene variants. The panel includes genes related to common medication classes such as anticoagulants, antiplatelets, proton pump inhibitors, nonsteroidal anti-inflammatory drugs, opioids, antidepressants, and statins. The VA PHASER panel includes the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene, which is predominantly expressed in the liver and facilitates the hepatic uptake of most statins.^13,14 A reduced function of SLCO1B1 can lead to higher statin levels, resulting in increased concentrations that may potentially cause SAMS.^13,14 Some alleles associated with reduced function include SLCO1B1*5, *15, *23, *31, and *46 to *49, whereas others are associated with increased function, such as SLCO1B1 *14 and *20 (Appendix).¹⁵ Supporting evidence shows the SLCO1B1*5 nucleotide polymorphism increases plasma levels of simvastatin and atorvastatin, affecting effectiveness or toxicity. ¹³ Females tend to have a lower body weight and higher percentage of body fat compared with males, which might lead to higher concentrations of lipophilic drugs, including atorvastatin and simvastatin, which may be exacerbated by decreased function of SLCO1B1*5.¹⁵ With pharmacogenomic testing, therapeutic recommendations can be made to improve the overall safety and efficacy of statins, thus improving adherence using a patient-specific approach.^14,15

Methods

Carl Vinson VA Medical Center (CVVAMC) serves about 42,000 veterans in Central and South Georgia, of which about 15% are female. Of the female veterans enrolled in care, 63% identify as Black, 27% White, and 1.5% as Asian, American Indian/Alaska Native, or Native Hawaiian/Other Pacific Islander. The 2020 Veterans Chartbook report showed that female veterans and minority racial and ethnic groups had worse access to health care and higher mortality rates than their male and non-Hispanic White counterparts.¹⁶

The Primary Care Equity Dashboard (PCED) was developed to engage the VA health care workforce in the process of identifying and addressing inequities in local patient populations.¹⁷ Using electronic quality measure data, the PCED provides Veterans Integrated Service Network-level and facility-level performance on several metrics.¹⁸ The PCED had not been previously used at the CVVAMC, and few publications or quality improvement projects regarding its use have been reported by the VA Office of Health Equity. PCED helped identify disparities when comparing female to male patients in the prescribing of statin therapy for patients with CVD and statin therapy for patients with T2DM.

VA PHASER pharmacogenomic analyses provided an opportunity to expand this quality improvement project. Sanford Health and the VA collaborated on the PHASER program to offer free genetic testing for veterans. The program launched in 2019 and expanded to various VA sites, including CVVAMC in March 2023. This program has been extended to December 31, 2025.

The primary objective of this quality improvement project was to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk. Secondary outcomes included increased pharmacogenomic testing and the assessment of pharmacogenomic results related to statin therapy. This project was approved by the CVVAMC Pharmacy and Therapeutics Committee. The PCED was used to identify female veterans with T2DM and/or CVD without an active prescription for a statin between July and October 2023. A review of Computerized Patient Record System patient charts was completed to screen for prespecified inclusion and exclusion criteria. Veterans were included if they were assigned female at birth, were enrolled in care at CVVAMC, and had a diagnosis of T2DM or CVD (history of myocardial infarction, coronary bypass graft, percutaneous coronary intervention, or other revascularization in any setting).

Veterans were excluded if they were currently pregnant, trying to conceive, breastfeeding, had a T1DM diagnosis, had previously documented hypersensitivity to a statin, active liver failure or decompensated cirrhosis, previously documented statin-associated rhabdomyolysis or autoimmune myopathy, an active prescription for a proprotein convertase subtilisin/kexin type 9 inhibitor, or previously documented statin intolerance (defined as the inability to tolerate ≥ 3 statins, with ≥ 1 prescribed at low intensity or alternate-day dosing). The female veterans were compared to 2 comparators: the facility's male veterans and the VA national average, identified via the PCED.

Once a veteran was screened, they were telephoned between October 2023 and February 2024 and provided education on statin use and pharmacogenomic testing using a standardized note template. An order was placed for participants who provided verbal consent for pharmacogenomic testing. Those who agreed to statin initiation were referred to a clinical pharmacist practitioner (CPP) who contacted them at a later date to prescribe a statin following the recommendations of the 2019 ACC/AHA and 2023 ADA guidelines and pharmacogenomic testing, if applicable.^4,5,7 Appropriate monitoring and follow-up occurred at the discretion of each CPP. Data collection included: age, race, diagnoses (T2DM, CVD, or both), baseline lipid panel (total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein), hepatic function, name and dose of statin, reasons for declining statin therapy, and pharmacogenomic testing results related to SLCO1B1.

Results

At baseline in July 2023, 77.8% of female veterans with T2DM were prescribed a statin, which exceeded the national VA average (77.0%), but was below the rate for male veterans (78.7%) in the facility comparator group.17 Additionally, 82.2% of females with CVD were prescribed a statin, which was below the national VA average of 86.0% and the 84.9% of male veterans in the facility comparator group.17 The PCED identified 189 female veterans from July 2023 to October 2023 who may benefit from statin therapy. Thirty-three females met the exclusion criteria. Of the 156 included veterans, 129 (82.7%) were successfully contacted and 27 (17.3%) could not be reached by telephone after 3 attempts (Figure 1). The 129 female veterans contacted had a mean age of 59 years and the majority were Black (82.9%) (Table 1).

Primary Outcomes

Of the 129 contacted veterans, 31 (24.0%) had a non-VA statin prescription, 13 (10.1%) had an active VA statin prescription, and 85 (65.9%) did not have a statin prescription, despite being eligible. Statin adherence was confirmed with participants, and the medication list was updated accordingly.

Of the 85 veterans with no active statin therapy, 37 (43.5%) accepted a new statin prescription and 48 (56.5%) declined. There were various reasons provided for declining statin therapy: 17 participants (35.4%) declined due to concern for AEs (Table 2).

From July 2023 to March 2024, the percentage of female veterans with active statin therapy with T2DM increased from 77.8% to 79.0%. For those with active statin therapy with CVD, usage increased from 82.2% to 90.2%, which exceeded the national VA average and facility male comparator group (Figures 2 and 3).¹⁷

Secondary Outcomes

Seventy-one of 129 veterans (55.0%) gave verbal consent, and 47 (66.2%) completed the pharmacogenomic testing; 58 (45.0%) declined. Five veterans (10.6%) had a known SLCO1B1 allele variant present. One veteran required a change in statin therapy based on the results (eAppendix).

Discussion

This project aimed to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk and increase pharmacogenomic testing using the PCED and care managed by CPPs. The results of this quality improvement project illustrated that both metrics have improved at CVVAMC as a result of the intervention. The results in both metrics now exceed the PCED national VA average, and the CVD metric also exceeds that of the facility male comparator group. While there was only a 1.2% increase from July 2023 to March 2024 for patients with T2DM, there was an 8.0% increase for patients with CVD. Despite standardized education on statin use, more veterans declined therapy than accepted it, mostly due to concern for AEs. Recording the reasons for declining statin therapy offered valuable insight that can be used in additional discussions with veterans and clinicians.

Pharmacogenomics gives clinicians the unique opportunity to take a proactive approach to better predict drug responses, potentially allowing for less trial and error with medications, fewer AEs, greater trust in the clinician, and improved medication adherence. The CPPs incorporated pharmacogenomic testing into their practice, which led to identifying 5 SLCO1B1 gene abnormalities. The PCED served as a powerful tool for advancing equity-focused quality improvement initiatives on a local level and was crucial in prioritizing the detection of veterans potentially receiving suboptimal care.

Limitations

The nature of “cold calls” made it challenging to establish contact for inclusion in this study. An alternative to increase engagement could have been scheduled phone or face-to-face visits. While the use of the PCED was crucial, data did not account for statins listed in the non-VA medication list. All 31 patients with statins prescribed outside the VA had a start date added to provide the most accurate representation of the data moving forward.

Another limitation in this project was its small sample size and population. CVVAMC serves about 6200 female veterans, with roughly 63% identifying as Black. The preponderance of Black individuals (83%) in this project is typical for the female patient population at CVVAMC but may not reflect the demographics of other populations. Other limitations to this project consisted of scheduling conflicts. Appointments for laboratory draws at community-based outpatient clinics were subject to availability, which resulted in some delay in completion of pharmacogenomic testing.

Conclusions

CPPs can help reduce inequity in health care delivery. Increased incorporation of the PCED into regular practice within the VA is recommended to continue addressing sex disparities in statin use, diabetes control, blood pressure management, cancer screenings, and vaccination needs. CVVAMC plans to expand its use through another quality improvement project focused on reducing sex disparities in blood pressure management. Improving educational resources made available to veterans on the importance of statin therapy and potential to mitigate AEs through use of the VA PHASER program also would be helpful. This project successfully improved CVVAMC metrics for female veterans appropriately prescribed statin therapy and increased access to pharmacogenomic testing. Most importantly, it helped close the sex-based gap in CVD risk reduction care.

THE DIAGNOSIS: Lepromatous Leprosy

Histopathology showed collections of epithelioid to sarcoidal granulomas throughout the dermis and clustered around nerve bundles with a grenz zone at the dermoepidermal junction. Fite stain was positive for acid-fast bacteria, which were confirmed to be Mycobacterium leprae by by the National Hansen’s Disease program. Based on these findings, a diagnosis of lepromatous leprosy (LL) was made. The patient was treated by the infectious disease department with multidrug therapy that included monthly rifampin, moxifloxacin, and minocycline; weekly methotrexate with daily folic acid; and an extended prednisone taper with prophylactic cholecalciferol.

Lepromatous leprosy is characterized by high antibody titers to the acid-fast, gram-positive bacillus Mycobacterium leprae as well as a high bacillary load.¹ Patients typically present with muscle weakness, anesthetic skin patches, and claw hands. Patients also may present with foot drop, ulcerations of the hands and feet, autonomic dysfunction with anhidrosis or impaired sweating, and localized alopecia.² Over months to years, LL may progress to extensive sensory loss and indurated lesions that infiltrate the skin and cause thickening, especially on the face (known as leonine facies). Furthermore, LL is characterized by extensive bilaterally symmetric cutaneous lesions with poorly defined borders and raised indurated centers.³

Lepromatous leprosy transmission is not fully understood but is thought to occur via airborne droplets from coughing/sneezing and nasal secretions.² Histopathology generally shows a dense and diffuse granulomatous infiltrate that involves the dermis but is separated from the epidermis by a zone of collagen (grenz zone).³ Histology is characterized by the presence of lymphocytes and numerous foamy macrophages (lepra or Virchow cells) containing M leprae organisms. In persistent lesions, the high density of uncleared bacilli forms spherical cytoplasmic clumps known as globi within enlarged foamy histiocytes (Figure 1).⁴ The macrophages form granulomatous lesions in the skin and around nerve bundles, resulting in tissue damage and decreased sensation. The current standard of care for LL is a multidrug combination of dapsone, rifampin, and clofazimine. Early diagnosis and complete treatment of LL is crucial, as this approach typically leads to complete cure of the disease.

The differential diagnosis for LL includes granuloma annulare (GA), mycosis fungoides (MF), sarcoidosis, and subacute cutaneous lupus erythematosus (SCLE). Granuloma annulare is a noninfectious inflammatory granulomatous skin disease that manifests in a localized, generalized, or subcutaneous pattern. Localized GA is the most common form and manifests as self-resolving, flesh-colored or erythematous papules or plaques limited to the extremities.^5,6 Generalized GA is defined by more than 10 widespread annular plaques involving the trunk and extremities and can persist for decades.⁶ This form can be associated with hyperlipidemia, diabetes, autoimmune disease and immunodeficiency (eg, HIV), and rarely with lymphoma or solid tumors. On histology, GA shows necrobiosis surrounded by palisading histiocytes and mucin (palisading GA) or patchy interstitial histiocytes and lymphocytes (interstitial GA)(Figure 2).⁶ This palisading pattern differs from the histiocytes in LL, which contain numerous acid-fast bacilli and bacterial clumps. Topical and intralesional corticosteroids are first-line therapies for GA.

Mycosis fungoides is a cutaneous T-cell lymphoma characterized by proliferation of CD4+ T cells.⁷ In the early stages of MF, patients may present with multiple erythematous and scaly patches, plaques, or nodules that most commonly develop on unexposed areas of the skin, but specific variants frequently may cause lesions on the face or scalp.⁸ Tumors may be solitary, localized, or generalized and may be observed alongside patches and plaques or in the absence of cutaneous lesions.⁷ The pathologic features of MF include fibrosis of the papillary dermis, individual haloed atypical lymphocytes in the epidermis, and atypical lymphoid cells with cerebriform nuclei (Figure 3).⁹ Granulomatous MF is characterized by diffuse nodular and perivascular infiltrates of histiocytes with small lymphocytes without atypia, eosinophils, and plasma cells. Small lymphocytes with cerebriform nuclei and larger lymphocytes with hyperconvoluted nuclei also may be seen, in addition to multinucleated histiocytic giant cells. Although MF commonly manifests with epidermotropism, it typically is absent in granulomatous MF (GMF).¹⁰ Granulomatous MF may manifest similarly to LL. Noduloulcerative lesions and infiltration of atypical lymphocytes into the epidermis (epidermotropism) are much more common in GMF than in LL; however, although ulcerative nodules are not a common feature in patients with leprosy (except during reactional states [ie, Lucio phenomenon]) or secondary to neuropathies, they also can occur in LL.¹¹ In GMF, the infiltrate does not follow a specific pattern, whereas LL infiltrates tend to follow a nerve distribution. Treatment for MF is determined by disease severity.¹² First-line therapy includes local corticosteroids and phototherapy with UVB irradiation.

Sarcoidosis is a multisystem disease that demonstrates nonspecific clinical manifestations affecting the lungs, eyes, liver, and skin.¹³ Environmental exposures to silica and inorganic matter have been linked to an increased risk for sarcoidosis, with patients presenting with fatigue, fever, and arthralgia.¹³ Skin manifestations include subcutaneous nodules, polymorphous plaques, and erythema nodosum—nodosum—the most common cutaneous presentation of sarcoidosis. Erythema nodosum manifests as symmetrically distributed, nonulcerative, painful red nodules on the skin, especially the lower legs. The histopathology of sarcoidosis shows noncaseating granulomas with activated T-lymphocytes, epithelioid cells, and multinucleated giant cells (Figure 4). Although granulomas occur in both LL and sarcoidosis, those in sarcoidosis typically consist of epithelioid cells surrounded by a rim of lymphocytes, whereas LL granulomas contain foamy histiocytes and multinucleated giant cells. Treatment of sarcoidosis depends on disease progression and generally involves oral corticosteroids, followed by corticosteroid-sparing regimens.

Subacute cutaneous lupus erythematosus is a chronic autoimmune disease that predominantly affects younger women. Common findings in SCLE include red scaly plaques and ring-shaped lesions on sun-exposed areas of the skin.¹⁴ Subacute cutaneous lupus erythematosus primarily is characterized by a photosensitive rash, often with arthralgia, myalgia, and/or oral ulcers; less commonly, a small percentage of patients can experience central nervous system involvement, vasculitis, or nephritis. The histologic findings of SCLE include hydropic degeneration of the basal cell layer and periadnexal infiltrates (Figure 5). The incidence of SCLE often is associated with anti-Ro (SSA) and anti-La (SSB) antibodies.¹⁵ Treatment of SCLE focuses on managing skin symptoms with corticosteroids, antimalarials, and sun protection.

THE DIAGNOSIS: Lepromatous Leprosy

Histopathology showed collections of epithelioid to sarcoidal granulomas throughout the dermis and clustered around nerve bundles with a grenz zone at the dermoepidermal junction. Fite stain was positive for acid-fast bacteria, which were confirmed to be Mycobacterium leprae by by the National Hansen’s Disease program. Based on these findings, a diagnosis of lepromatous leprosy (LL) was made. The patient was treated by the infectious disease department with multidrug therapy that included monthly rifampin, moxifloxacin, and minocycline; weekly methotrexate with daily folic acid; and an extended prednisone taper with prophylactic cholecalciferol.

Lepromatous leprosy is characterized by high antibody titers to the acid-fast, gram-positive bacillus Mycobacterium leprae as well as a high bacillary load.¹ Patients typically present with muscle weakness, anesthetic skin patches, and claw hands. Patients also may present with foot drop, ulcerations of the hands and feet, autonomic dysfunction with anhidrosis or impaired sweating, and localized alopecia.² Over months to years, LL may progress to extensive sensory loss and indurated lesions that infiltrate the skin and cause thickening, especially on the face (known as leonine facies). Furthermore, LL is characterized by extensive bilaterally symmetric cutaneous lesions with poorly defined borders and raised indurated centers.³

Lepromatous leprosy transmission is not fully understood but is thought to occur via airborne droplets from coughing/sneezing and nasal secretions.² Histopathology generally shows a dense and diffuse granulomatous infiltrate that involves the dermis but is separated from the epidermis by a zone of collagen (grenz zone).³ Histology is characterized by the presence of lymphocytes and numerous foamy macrophages (lepra or Virchow cells) containing M leprae organisms. In persistent lesions, the high density of uncleared bacilli forms spherical cytoplasmic clumps known as globi within enlarged foamy histiocytes (Figure 1).⁴ The macrophages form granulomatous lesions in the skin and around nerve bundles, resulting in tissue damage and decreased sensation. The current standard of care for LL is a multidrug combination of dapsone, rifampin, and clofazimine. Early diagnosis and complete treatment of LL is crucial, as this approach typically leads to complete cure of the disease.

The differential diagnosis for LL includes granuloma annulare (GA), mycosis fungoides (MF), sarcoidosis, and subacute cutaneous lupus erythematosus (SCLE). Granuloma annulare is a noninfectious inflammatory granulomatous skin disease that manifests in a localized, generalized, or subcutaneous pattern. Localized GA is the most common form and manifests as self-resolving, flesh-colored or erythematous papules or plaques limited to the extremities.^5,6 Generalized GA is defined by more than 10 widespread annular plaques involving the trunk and extremities and can persist for decades.⁶ This form can be associated with hyperlipidemia, diabetes, autoimmune disease and immunodeficiency (eg, HIV), and rarely with lymphoma or solid tumors. On histology, GA shows necrobiosis surrounded by palisading histiocytes and mucin (palisading GA) or patchy interstitial histiocytes and lymphocytes (interstitial GA)(Figure 2).⁶ This palisading pattern differs from the histiocytes in LL, which contain numerous acid-fast bacilli and bacterial clumps. Topical and intralesional corticosteroids are first-line therapies for GA.

Mycosis fungoides is a cutaneous T-cell lymphoma characterized by proliferation of CD4+ T cells.⁷ In the early stages of MF, patients may present with multiple erythematous and scaly patches, plaques, or nodules that most commonly develop on unexposed areas of the skin, but specific variants frequently may cause lesions on the face or scalp.⁸ Tumors may be solitary, localized, or generalized and may be observed alongside patches and plaques or in the absence of cutaneous lesions.⁷ The pathologic features of MF include fibrosis of the papillary dermis, individual haloed atypical lymphocytes in the epidermis, and atypical lymphoid cells with cerebriform nuclei (Figure 3).⁹ Granulomatous MF is characterized by diffuse nodular and perivascular infiltrates of histiocytes with small lymphocytes without atypia, eosinophils, and plasma cells. Small lymphocytes with cerebriform nuclei and larger lymphocytes with hyperconvoluted nuclei also may be seen, in addition to multinucleated histiocytic giant cells. Although MF commonly manifests with epidermotropism, it typically is absent in granulomatous MF (GMF).¹⁰ Granulomatous MF may manifest similarly to LL. Noduloulcerative lesions and infiltration of atypical lymphocytes into the epidermis (epidermotropism) are much more common in GMF than in LL; however, although ulcerative nodules are not a common feature in patients with leprosy (except during reactional states [ie, Lucio phenomenon]) or secondary to neuropathies, they also can occur in LL.¹¹ In GMF, the infiltrate does not follow a specific pattern, whereas LL infiltrates tend to follow a nerve distribution. Treatment for MF is determined by disease severity.¹² First-line therapy includes local corticosteroids and phototherapy with UVB irradiation.

Sarcoidosis is a multisystem disease that demonstrates nonspecific clinical manifestations affecting the lungs, eyes, liver, and skin.¹³ Environmental exposures to silica and inorganic matter have been linked to an increased risk for sarcoidosis, with patients presenting with fatigue, fever, and arthralgia.¹³ Skin manifestations include subcutaneous nodules, polymorphous plaques, and erythema nodosum—nodosum—the most common cutaneous presentation of sarcoidosis. Erythema nodosum manifests as symmetrically distributed, nonulcerative, painful red nodules on the skin, especially the lower legs. The histopathology of sarcoidosis shows noncaseating granulomas with activated T-lymphocytes, epithelioid cells, and multinucleated giant cells (Figure 4). Although granulomas occur in both LL and sarcoidosis, those in sarcoidosis typically consist of epithelioid cells surrounded by a rim of lymphocytes, whereas LL granulomas contain foamy histiocytes and multinucleated giant cells. Treatment of sarcoidosis depends on disease progression and generally involves oral corticosteroids, followed by corticosteroid-sparing regimens.

Subacute cutaneous lupus erythematosus is a chronic autoimmune disease that predominantly affects younger women. Common findings in SCLE include red scaly plaques and ring-shaped lesions on sun-exposed areas of the skin.¹⁴ Subacute cutaneous lupus erythematosus primarily is characterized by a photosensitive rash, often with arthralgia, myalgia, and/or oral ulcers; less commonly, a small percentage of patients can experience central nervous system involvement, vasculitis, or nephritis. The histologic findings of SCLE include hydropic degeneration of the basal cell layer and periadnexal infiltrates (Figure 5). The incidence of SCLE often is associated with anti-Ro (SSA) and anti-La (SSB) antibodies.¹⁵ Treatment of SCLE focuses on managing skin symptoms with corticosteroids, antimalarials, and sun protection.

THE DIAGNOSIS: Lepromatous Leprosy

Histopathology showed collections of epithelioid to sarcoidal granulomas throughout the dermis and clustered around nerve bundles with a grenz zone at the dermoepidermal junction. Fite stain was positive for acid-fast bacteria, which were confirmed to be Mycobacterium leprae by by the National Hansen’s Disease program. Based on these findings, a diagnosis of lepromatous leprosy (LL) was made. The patient was treated by the infectious disease department with multidrug therapy that included monthly rifampin, moxifloxacin, and minocycline; weekly methotrexate with daily folic acid; and an extended prednisone taper with prophylactic cholecalciferol.

Lepromatous leprosy is characterized by high antibody titers to the acid-fast, gram-positive bacillus Mycobacterium leprae as well as a high bacillary load.¹ Patients typically present with muscle weakness, anesthetic skin patches, and claw hands. Patients also may present with foot drop, ulcerations of the hands and feet, autonomic dysfunction with anhidrosis or impaired sweating, and localized alopecia.² Over months to years, LL may progress to extensive sensory loss and indurated lesions that infiltrate the skin and cause thickening, especially on the face (known as leonine facies). Furthermore, LL is characterized by extensive bilaterally symmetric cutaneous lesions with poorly defined borders and raised indurated centers.³

Lepromatous leprosy transmission is not fully understood but is thought to occur via airborne droplets from coughing/sneezing and nasal secretions.² Histopathology generally shows a dense and diffuse granulomatous infiltrate that involves the dermis but is separated from the epidermis by a zone of collagen (grenz zone).³ Histology is characterized by the presence of lymphocytes and numerous foamy macrophages (lepra or Virchow cells) containing M leprae organisms. In persistent lesions, the high density of uncleared bacilli forms spherical cytoplasmic clumps known as globi within enlarged foamy histiocytes (Figure 1).⁴ The macrophages form granulomatous lesions in the skin and around nerve bundles, resulting in tissue damage and decreased sensation. The current standard of care for LL is a multidrug combination of dapsone, rifampin, and clofazimine. Early diagnosis and complete treatment of LL is crucial, as this approach typically leads to complete cure of the disease.

The differential diagnosis for LL includes granuloma annulare (GA), mycosis fungoides (MF), sarcoidosis, and subacute cutaneous lupus erythematosus (SCLE). Granuloma annulare is a noninfectious inflammatory granulomatous skin disease that manifests in a localized, generalized, or subcutaneous pattern. Localized GA is the most common form and manifests as self-resolving, flesh-colored or erythematous papules or plaques limited to the extremities.^5,6 Generalized GA is defined by more than 10 widespread annular plaques involving the trunk and extremities and can persist for decades.⁶ This form can be associated with hyperlipidemia, diabetes, autoimmune disease and immunodeficiency (eg, HIV), and rarely with lymphoma or solid tumors. On histology, GA shows necrobiosis surrounded by palisading histiocytes and mucin (palisading GA) or patchy interstitial histiocytes and lymphocytes (interstitial GA)(Figure 2).⁶ This palisading pattern differs from the histiocytes in LL, which contain numerous acid-fast bacilli and bacterial clumps. Topical and intralesional corticosteroids are first-line therapies for GA.

Mycosis fungoides is a cutaneous T-cell lymphoma characterized by proliferation of CD4+ T cells.⁷ In the early stages of MF, patients may present with multiple erythematous and scaly patches, plaques, or nodules that most commonly develop on unexposed areas of the skin, but specific variants frequently may cause lesions on the face or scalp.⁸ Tumors may be solitary, localized, or generalized and may be observed alongside patches and plaques or in the absence of cutaneous lesions.⁷ The pathologic features of MF include fibrosis of the papillary dermis, individual haloed atypical lymphocytes in the epidermis, and atypical lymphoid cells with cerebriform nuclei (Figure 3).⁹ Granulomatous MF is characterized by diffuse nodular and perivascular infiltrates of histiocytes with small lymphocytes without atypia, eosinophils, and plasma cells. Small lymphocytes with cerebriform nuclei and larger lymphocytes with hyperconvoluted nuclei also may be seen, in addition to multinucleated histiocytic giant cells. Although MF commonly manifests with epidermotropism, it typically is absent in granulomatous MF (GMF).¹⁰ Granulomatous MF may manifest similarly to LL. Noduloulcerative lesions and infiltration of atypical lymphocytes into the epidermis (epidermotropism) are much more common in GMF than in LL; however, although ulcerative nodules are not a common feature in patients with leprosy (except during reactional states [ie, Lucio phenomenon]) or secondary to neuropathies, they also can occur in LL.¹¹ In GMF, the infiltrate does not follow a specific pattern, whereas LL infiltrates tend to follow a nerve distribution. Treatment for MF is determined by disease severity.¹² First-line therapy includes local corticosteroids and phototherapy with UVB irradiation.

Sarcoidosis is a multisystem disease that demonstrates nonspecific clinical manifestations affecting the lungs, eyes, liver, and skin.¹³ Environmental exposures to silica and inorganic matter have been linked to an increased risk for sarcoidosis, with patients presenting with fatigue, fever, and arthralgia.¹³ Skin manifestations include subcutaneous nodules, polymorphous plaques, and erythema nodosum—nodosum—the most common cutaneous presentation of sarcoidosis. Erythema nodosum manifests as symmetrically distributed, nonulcerative, painful red nodules on the skin, especially the lower legs. The histopathology of sarcoidosis shows noncaseating granulomas with activated T-lymphocytes, epithelioid cells, and multinucleated giant cells (Figure 4). Although granulomas occur in both LL and sarcoidosis, those in sarcoidosis typically consist of epithelioid cells surrounded by a rim of lymphocytes, whereas LL granulomas contain foamy histiocytes and multinucleated giant cells. Treatment of sarcoidosis depends on disease progression and generally involves oral corticosteroids, followed by corticosteroid-sparing regimens.

Subacute cutaneous lupus erythematosus is a chronic autoimmune disease that predominantly affects younger women. Common findings in SCLE include red scaly plaques and ring-shaped lesions on sun-exposed areas of the skin.¹⁴ Subacute cutaneous lupus erythematosus primarily is characterized by a photosensitive rash, often with arthralgia, myalgia, and/or oral ulcers; less commonly, a small percentage of patients can experience central nervous system involvement, vasculitis, or nephritis. The histologic findings of SCLE include hydropic degeneration of the basal cell layer and periadnexal infiltrates (Figure 5). The incidence of SCLE often is associated with anti-Ro (SSA) and anti-La (SSB) antibodies.¹⁵ Treatment of SCLE focuses on managing skin symptoms with corticosteroids, antimalarials, and sun protection.

To the Editor:  

For many years, topical treatment of plaque psoriasis was limited to steroids, calcineurin inhibitors, vitamin D analogs, retinoids, coal tar products, and anthralin. In recent years, 2 new nonsteroidal treatment options with alternative mechanisms of action, roflumilast 0.3% and tapinarof 1%, have been approved by the US Food and Drug Administration.¹ Roflumilast 0.3%, a topical phosphodiesterase 4 inhibitor, was shown in phase 3 clinical trials to reach an Investigator Global Assessment response of 37.5% to 42.2% in 8 weeks using once-daily application with minimal cutaneous adverse effects.¹ Furthermore, it has demonstrated efficacy in treating psoriasis in intertriginous areas in subset analyses.¹ Tapinarof is an aryl hydrocarbon receptor agonist that suppresses Th17 cell differentiation by downregulating IL-17, IL-22, and IL-23.¹ In phase 3 clinical trials, 35% to 40% of patients who used tapinarof cream 1% once daily demonstrated improvement in psoriasis compared with 6% who used the vehicle alone.² In these studies, 18% to 24% of patients who used tapinarof cream 1% experienced folliculitis.²

Acute generalized exanthematous pustulosis (AGEP) is a nonfollicular pustular drug reaction with systemic symptoms that typically occurs within 2 weeks of exposure to an inciting medication. Systemic antibiotics are the most commonly reported cause of AGEP.³ There are few reports in the literature of AGEP induced by topical agents.^4,5 We report a case of AGEP in a young man following the use of tapinarof cream 1%.

A 23-year-old man with a history of psoriasis presented to the emergency department with fever and a pustular rash. One week prior to presentation, he developed a pustular eruption around plaques of psoriasis on the arms and legs. The patient had been prescribed tapinarof cream 1% by an outside dermatologist and was applying the medication to the affected areas once daily for 1 month prior to onset of symptoms. He discontinued tapinarof a few days prior to the eruption starting, but the rash progressed centrifugally and was associated with fevers and fatigue despite treatment with a brief course of empiric cephalexin prescribed by his primary care provider.

At presentation to our institution, the patient had widespread erythematous patches studded with pustules located on the arms, legs, and flexural areas as well as plaques of psoriasis involving approximately 20% of the body surface area (Figure 1). Furthermore, the patient was noted to have large noninflammatory bullae along the legs. The new eruption occurred on areas that were both treated and spared from the tapinarof cream 1%. Laboratory evaluation showed neutrophil-­predominant leukocytosis (white blood cell count, 15.9×10³/µL ­[reference range, 4.0-11.0×10³/µL]; absolute neutrophil count, 10.3×10³/µL [reference range, 1.5-8.0×10³/µL]), absolute eosinophilia (1930/µL [reference range, 0-0.5×10³/µL]), hypocalcemia (8.4 mg/dL ­[reference range, 8.5-10.5 mg/dL]), and a mild transaminitis ­(aspartate aminotransferase, 37 IU/L [reference range, 10-40 IU/L]; alanine aminotransferase, 53 IU/L ­[reference range, 7-56 U/L]). Histopathology demonstrated spongiosis with subcorneal and intraepidermal pustules and mixed dermal inflammation containing eosinophils (Figure 2). Direct immunofluorescence revealed mild granular staining of C3 at the basement membrane zone.

The patient was started on 1 mg/kg/d of prednisone tapered over 20 days, and he rapidly improved. Alanine aminotransferase levels peaked at 120 IU/L 2 weeks later. At that time, he had complete resolution of the original eruption and was transitioned to topical steroids for continued management of the psoriasis (Figure 3).

The differential diagnosis for our patient included AGEP, generalized pustular psoriasis (GPP), miliaria pustulosa, generalized cutaneous candidiasis, exuberant allergic contact dermatitis (ACD), and linear IgA bullous dermatosis (LABD). Based on the clinical manifestations, laboratory results, and histopathologic evaluation, we made the diagnosis of AGEP secondary to tapinarof with systemic absorption. Acute generalized exanthematous pustulosis has been reported with topical use of morphine and diphenhydramine, among other agents.^4,5 To our knowledge, AGEP due to tapinarof cream 1% has not been reported. In the original clinical trials of tapinarof, folliculitis was contained to sites of application.² Our patient developed pustules at sites distant to areas of application, as well as systemic symptoms and laboratory abnormalities, indicating a systemic reaction. It can be difficult to distinguish AGEP clinically and histologically from GPP. Both conditions can manifest with fever, hypocalcemia, and sterile pustules on a background of erythema that favors intertriginous areas.⁶ Infection, rapid oral steroid withdrawal, pregnancy, and rarely oral medications have been reported causes of GPP.⁶ Our patient did not have any of these exposures. There is overlap in the histology of AGEP and GPP. One retrospective series compared histologic samples to help distinguish these 2 entities. Reliable markers that favored AGEP over GPP included eosinophilic spongiosis, interface dermatitis, and dermal eosinophilia (>2/mm²).⁷ In contrast, the presence of CD161 positivity in the dermis with at least 10 cells favored a diagnosis of GPP.⁷ In our case, the presence of spongiosis with eosinophils in the dermis favored a diagnosis of AGEP over GPP. 

Miliaria pustulosa is a benign condition caused by the occlusion of the epidermal portion of eccrine glands related to either high fever or hot and humid environmental conditions. While it can be present in intertriginous areas like AGEP, miliaria pustulosa can be seen extensively on the back, most commonly in immobile hospitalized patients.⁸ Generalized cutaneous candidiasis usually is caused by the yeast Candida albicans and can take on multiple morphologies, including folliculitis.⁹ The eruption may be disseminated but often is accentuated in intertriginous areas and the anogenital folds. Predisposing factors include immunosuppression, endocrinopathies, recent use of systemic antibiotics or steroids, chemotherapy, and indwelling catheters.⁹ Outside of recent antibiotic use, our patient did not have any risk factors for miliaria pustulosa, making this diagnosis unlikely.

Given the presence of overlapping bullae along the lower extremities, an exuberant ACD and LABD were considered. Bullae formation can occur in ACD secondary to robust inflammation and edema leading to acantholysis.¹⁰ While a delayed hypersensitivity reaction to topical tapinarof cream 1% was considered given that the patient used the medication for approximately 1 month prior to the onset of symptoms, it would be unlikely for ACD to present with a concomitant pustular eruption. Linear IgA bullous dermatosis is an autoimmune blistering disease in which antibodies target bullous pemphigoid antigen 2, and there is characteristically linear deposition of IgA at the dermal-epidermal junction that leads to subepidermal blistering.¹¹ This often manifests clinically as widespread tense vesicles in an annular or string-of-pearls appearance. However, morphologies can vary, and large bullae may be seen. In adults, LABD typically is associated with inflammatory bowel disease, malignancy, or medications, notably vancomycin.^11,12 Our patient did not have any of these predisposing factors, and his biopsy for direct immunofluorescence did not reveal the classic pattern described above.

Interestingly, there have been reports in the literature of bullous AGEP in the setting of oral anti-infectives. One report described a 62-year-old woman who developed widespread nonfollicular pustules with multiple tense serous blisters 24 hours after taking oral terbinafine.¹³ Another case described an 80-year-old woman with a similar presentation following a course of ciprofloxacin (although the timeline of medication administration was not described).¹⁴ In this case, patch testing to the culprit medication reproduced the response.¹⁴ In both cases, a biopsy revealed subcorneal and intraepidermal pustules with marked dermal edema.^13,14 As previously described, spongiosis is a common feature of AGEP. We hypothesize that, similar to these reports, our patient had a robust inflammatory response leading to spongiosis, acantholysis, and blister formation secondary to AGEP.

Dermatologists should be aware of this case of AGEP secondary to tapinarof cream 1%, as reports in the literature are rare and it is a reminder that topical medications can cause serious systemic reactions.

To the Editor:  

For many years, topical treatment of plaque psoriasis was limited to steroids, calcineurin inhibitors, vitamin D analogs, retinoids, coal tar products, and anthralin. In recent years, 2 new nonsteroidal treatment options with alternative mechanisms of action, roflumilast 0.3% and tapinarof 1%, have been approved by the US Food and Drug Administration.¹ Roflumilast 0.3%, a topical phosphodiesterase 4 inhibitor, was shown in phase 3 clinical trials to reach an Investigator Global Assessment response of 37.5% to 42.2% in 8 weeks using once-daily application with minimal cutaneous adverse effects.¹ Furthermore, it has demonstrated efficacy in treating psoriasis in intertriginous areas in subset analyses.¹ Tapinarof is an aryl hydrocarbon receptor agonist that suppresses Th17 cell differentiation by downregulating IL-17, IL-22, and IL-23.¹ In phase 3 clinical trials, 35% to 40% of patients who used tapinarof cream 1% once daily demonstrated improvement in psoriasis compared with 6% who used the vehicle alone.² In these studies, 18% to 24% of patients who used tapinarof cream 1% experienced folliculitis.²

Acute generalized exanthematous pustulosis (AGEP) is a nonfollicular pustular drug reaction with systemic symptoms that typically occurs within 2 weeks of exposure to an inciting medication. Systemic antibiotics are the most commonly reported cause of AGEP.³ There are few reports in the literature of AGEP induced by topical agents.^4,5 We report a case of AGEP in a young man following the use of tapinarof cream 1%.

A 23-year-old man with a history of psoriasis presented to the emergency department with fever and a pustular rash. One week prior to presentation, he developed a pustular eruption around plaques of psoriasis on the arms and legs. The patient had been prescribed tapinarof cream 1% by an outside dermatologist and was applying the medication to the affected areas once daily for 1 month prior to onset of symptoms. He discontinued tapinarof a few days prior to the eruption starting, but the rash progressed centrifugally and was associated with fevers and fatigue despite treatment with a brief course of empiric cephalexin prescribed by his primary care provider.

At presentation to our institution, the patient had widespread erythematous patches studded with pustules located on the arms, legs, and flexural areas as well as plaques of psoriasis involving approximately 20% of the body surface area (Figure 1). Furthermore, the patient was noted to have large noninflammatory bullae along the legs. The new eruption occurred on areas that were both treated and spared from the tapinarof cream 1%. Laboratory evaluation showed neutrophil-­predominant leukocytosis (white blood cell count, 15.9×10³/µL ­[reference range, 4.0-11.0×10³/µL]; absolute neutrophil count, 10.3×10³/µL [reference range, 1.5-8.0×10³/µL]), absolute eosinophilia (1930/µL [reference range, 0-0.5×10³/µL]), hypocalcemia (8.4 mg/dL ­[reference range, 8.5-10.5 mg/dL]), and a mild transaminitis ­(aspartate aminotransferase, 37 IU/L [reference range, 10-40 IU/L]; alanine aminotransferase, 53 IU/L ­[reference range, 7-56 U/L]). Histopathology demonstrated spongiosis with subcorneal and intraepidermal pustules and mixed dermal inflammation containing eosinophils (Figure 2). Direct immunofluorescence revealed mild granular staining of C3 at the basement membrane zone.

The patient was started on 1 mg/kg/d of prednisone tapered over 20 days, and he rapidly improved. Alanine aminotransferase levels peaked at 120 IU/L 2 weeks later. At that time, he had complete resolution of the original eruption and was transitioned to topical steroids for continued management of the psoriasis (Figure 3).

The differential diagnosis for our patient included AGEP, generalized pustular psoriasis (GPP), miliaria pustulosa, generalized cutaneous candidiasis, exuberant allergic contact dermatitis (ACD), and linear IgA bullous dermatosis (LABD). Based on the clinical manifestations, laboratory results, and histopathologic evaluation, we made the diagnosis of AGEP secondary to tapinarof with systemic absorption. Acute generalized exanthematous pustulosis has been reported with topical use of morphine and diphenhydramine, among other agents.^4,5 To our knowledge, AGEP due to tapinarof cream 1% has not been reported. In the original clinical trials of tapinarof, folliculitis was contained to sites of application.² Our patient developed pustules at sites distant to areas of application, as well as systemic symptoms and laboratory abnormalities, indicating a systemic reaction. It can be difficult to distinguish AGEP clinically and histologically from GPP. Both conditions can manifest with fever, hypocalcemia, and sterile pustules on a background of erythema that favors intertriginous areas.⁶ Infection, rapid oral steroid withdrawal, pregnancy, and rarely oral medications have been reported causes of GPP.⁶ Our patient did not have any of these exposures. There is overlap in the histology of AGEP and GPP. One retrospective series compared histologic samples to help distinguish these 2 entities. Reliable markers that favored AGEP over GPP included eosinophilic spongiosis, interface dermatitis, and dermal eosinophilia (>2/mm²).⁷ In contrast, the presence of CD161 positivity in the dermis with at least 10 cells favored a diagnosis of GPP.⁷ In our case, the presence of spongiosis with eosinophils in the dermis favored a diagnosis of AGEP over GPP. 

Miliaria pustulosa is a benign condition caused by the occlusion of the epidermal portion of eccrine glands related to either high fever or hot and humid environmental conditions. While it can be present in intertriginous areas like AGEP, miliaria pustulosa can be seen extensively on the back, most commonly in immobile hospitalized patients.⁸ Generalized cutaneous candidiasis usually is caused by the yeast Candida albicans and can take on multiple morphologies, including folliculitis.⁹ The eruption may be disseminated but often is accentuated in intertriginous areas and the anogenital folds. Predisposing factors include immunosuppression, endocrinopathies, recent use of systemic antibiotics or steroids, chemotherapy, and indwelling catheters.⁹ Outside of recent antibiotic use, our patient did not have any risk factors for miliaria pustulosa, making this diagnosis unlikely.

Given the presence of overlapping bullae along the lower extremities, an exuberant ACD and LABD were considered. Bullae formation can occur in ACD secondary to robust inflammation and edema leading to acantholysis.¹⁰ While a delayed hypersensitivity reaction to topical tapinarof cream 1% was considered given that the patient used the medication for approximately 1 month prior to the onset of symptoms, it would be unlikely for ACD to present with a concomitant pustular eruption. Linear IgA bullous dermatosis is an autoimmune blistering disease in which antibodies target bullous pemphigoid antigen 2, and there is characteristically linear deposition of IgA at the dermal-epidermal junction that leads to subepidermal blistering.¹¹ This often manifests clinically as widespread tense vesicles in an annular or string-of-pearls appearance. However, morphologies can vary, and large bullae may be seen. In adults, LABD typically is associated with inflammatory bowel disease, malignancy, or medications, notably vancomycin.^11,12 Our patient did not have any of these predisposing factors, and his biopsy for direct immunofluorescence did not reveal the classic pattern described above.

Interestingly, there have been reports in the literature of bullous AGEP in the setting of oral anti-infectives. One report described a 62-year-old woman who developed widespread nonfollicular pustules with multiple tense serous blisters 24 hours after taking oral terbinafine.¹³ Another case described an 80-year-old woman with a similar presentation following a course of ciprofloxacin (although the timeline of medication administration was not described).¹⁴ In this case, patch testing to the culprit medication reproduced the response.¹⁴ In both cases, a biopsy revealed subcorneal and intraepidermal pustules with marked dermal edema.^13,14 As previously described, spongiosis is a common feature of AGEP. We hypothesize that, similar to these reports, our patient had a robust inflammatory response leading to spongiosis, acantholysis, and blister formation secondary to AGEP.

Dermatologists should be aware of this case of AGEP secondary to tapinarof cream 1%, as reports in the literature are rare and it is a reminder that topical medications can cause serious systemic reactions.

To the Editor:  

For many years, topical treatment of plaque psoriasis was limited to steroids, calcineurin inhibitors, vitamin D analogs, retinoids, coal tar products, and anthralin. In recent years, 2 new nonsteroidal treatment options with alternative mechanisms of action, roflumilast 0.3% and tapinarof 1%, have been approved by the US Food and Drug Administration.¹ Roflumilast 0.3%, a topical phosphodiesterase 4 inhibitor, was shown in phase 3 clinical trials to reach an Investigator Global Assessment response of 37.5% to 42.2% in 8 weeks using once-daily application with minimal cutaneous adverse effects.¹ Furthermore, it has demonstrated efficacy in treating psoriasis in intertriginous areas in subset analyses.¹ Tapinarof is an aryl hydrocarbon receptor agonist that suppresses Th17 cell differentiation by downregulating IL-17, IL-22, and IL-23.¹ In phase 3 clinical trials, 35% to 40% of patients who used tapinarof cream 1% once daily demonstrated improvement in psoriasis compared with 6% who used the vehicle alone.² In these studies, 18% to 24% of patients who used tapinarof cream 1% experienced folliculitis.²

Acute generalized exanthematous pustulosis (AGEP) is a nonfollicular pustular drug reaction with systemic symptoms that typically occurs within 2 weeks of exposure to an inciting medication. Systemic antibiotics are the most commonly reported cause of AGEP.³ There are few reports in the literature of AGEP induced by topical agents.^4,5 We report a case of AGEP in a young man following the use of tapinarof cream 1%.

A 23-year-old man with a history of psoriasis presented to the emergency department with fever and a pustular rash. One week prior to presentation, he developed a pustular eruption around plaques of psoriasis on the arms and legs. The patient had been prescribed tapinarof cream 1% by an outside dermatologist and was applying the medication to the affected areas once daily for 1 month prior to onset of symptoms. He discontinued tapinarof a few days prior to the eruption starting, but the rash progressed centrifugally and was associated with fevers and fatigue despite treatment with a brief course of empiric cephalexin prescribed by his primary care provider.

At presentation to our institution, the patient had widespread erythematous patches studded with pustules located on the arms, legs, and flexural areas as well as plaques of psoriasis involving approximately 20% of the body surface area (Figure 1). Furthermore, the patient was noted to have large noninflammatory bullae along the legs. The new eruption occurred on areas that were both treated and spared from the tapinarof cream 1%. Laboratory evaluation showed neutrophil-­predominant leukocytosis (white blood cell count, 15.9×10³/µL ­[reference range, 4.0-11.0×10³/µL]; absolute neutrophil count, 10.3×10³/µL [reference range, 1.5-8.0×10³/µL]), absolute eosinophilia (1930/µL [reference range, 0-0.5×10³/µL]), hypocalcemia (8.4 mg/dL ­[reference range, 8.5-10.5 mg/dL]), and a mild transaminitis ­(aspartate aminotransferase, 37 IU/L [reference range, 10-40 IU/L]; alanine aminotransferase, 53 IU/L ­[reference range, 7-56 U/L]). Histopathology demonstrated spongiosis with subcorneal and intraepidermal pustules and mixed dermal inflammation containing eosinophils (Figure 2). Direct immunofluorescence revealed mild granular staining of C3 at the basement membrane zone.

The patient was started on 1 mg/kg/d of prednisone tapered over 20 days, and he rapidly improved. Alanine aminotransferase levels peaked at 120 IU/L 2 weeks later. At that time, he had complete resolution of the original eruption and was transitioned to topical steroids for continued management of the psoriasis (Figure 3).

The differential diagnosis for our patient included AGEP, generalized pustular psoriasis (GPP), miliaria pustulosa, generalized cutaneous candidiasis, exuberant allergic contact dermatitis (ACD), and linear IgA bullous dermatosis (LABD). Based on the clinical manifestations, laboratory results, and histopathologic evaluation, we made the diagnosis of AGEP secondary to tapinarof with systemic absorption. Acute generalized exanthematous pustulosis has been reported with topical use of morphine and diphenhydramine, among other agents.^4,5 To our knowledge, AGEP due to tapinarof cream 1% has not been reported. In the original clinical trials of tapinarof, folliculitis was contained to sites of application.² Our patient developed pustules at sites distant to areas of application, as well as systemic symptoms and laboratory abnormalities, indicating a systemic reaction. It can be difficult to distinguish AGEP clinically and histologically from GPP. Both conditions can manifest with fever, hypocalcemia, and sterile pustules on a background of erythema that favors intertriginous areas.⁶ Infection, rapid oral steroid withdrawal, pregnancy, and rarely oral medications have been reported causes of GPP.⁶ Our patient did not have any of these exposures. There is overlap in the histology of AGEP and GPP. One retrospective series compared histologic samples to help distinguish these 2 entities. Reliable markers that favored AGEP over GPP included eosinophilic spongiosis, interface dermatitis, and dermal eosinophilia (>2/mm²).⁷ In contrast, the presence of CD161 positivity in the dermis with at least 10 cells favored a diagnosis of GPP.⁷ In our case, the presence of spongiosis with eosinophils in the dermis favored a diagnosis of AGEP over GPP. 

Miliaria pustulosa is a benign condition caused by the occlusion of the epidermal portion of eccrine glands related to either high fever or hot and humid environmental conditions. While it can be present in intertriginous areas like AGEP, miliaria pustulosa can be seen extensively on the back, most commonly in immobile hospitalized patients.⁸ Generalized cutaneous candidiasis usually is caused by the yeast Candida albicans and can take on multiple morphologies, including folliculitis.⁹ The eruption may be disseminated but often is accentuated in intertriginous areas and the anogenital folds. Predisposing factors include immunosuppression, endocrinopathies, recent use of systemic antibiotics or steroids, chemotherapy, and indwelling catheters.⁹ Outside of recent antibiotic use, our patient did not have any risk factors for miliaria pustulosa, making this diagnosis unlikely.

Given the presence of overlapping bullae along the lower extremities, an exuberant ACD and LABD were considered. Bullae formation can occur in ACD secondary to robust inflammation and edema leading to acantholysis.¹⁰ While a delayed hypersensitivity reaction to topical tapinarof cream 1% was considered given that the patient used the medication for approximately 1 month prior to the onset of symptoms, it would be unlikely for ACD to present with a concomitant pustular eruption. Linear IgA bullous dermatosis is an autoimmune blistering disease in which antibodies target bullous pemphigoid antigen 2, and there is characteristically linear deposition of IgA at the dermal-epidermal junction that leads to subepidermal blistering.¹¹ This often manifests clinically as widespread tense vesicles in an annular or string-of-pearls appearance. However, morphologies can vary, and large bullae may be seen. In adults, LABD typically is associated with inflammatory bowel disease, malignancy, or medications, notably vancomycin.^11,12 Our patient did not have any of these predisposing factors, and his biopsy for direct immunofluorescence did not reveal the classic pattern described above.

Interestingly, there have been reports in the literature of bullous AGEP in the setting of oral anti-infectives. One report described a 62-year-old woman who developed widespread nonfollicular pustules with multiple tense serous blisters 24 hours after taking oral terbinafine.¹³ Another case described an 80-year-old woman with a similar presentation following a course of ciprofloxacin (although the timeline of medication administration was not described).¹⁴ In this case, patch testing to the culprit medication reproduced the response.¹⁴ In both cases, a biopsy revealed subcorneal and intraepidermal pustules with marked dermal edema.^13,14 As previously described, spongiosis is a common feature of AGEP. We hypothesize that, similar to these reports, our patient had a robust inflammatory response leading to spongiosis, acantholysis, and blister formation secondary to AGEP.

Dermatologists should be aware of this case of AGEP secondary to tapinarof cream 1%, as reports in the literature are rare and it is a reminder that topical medications can cause serious systemic reactions.

The Diagnosis: Glomangiomyoma

The nail unit excision specimen showed collections of cuboidal cells and spindled cells within the corium that were consistent with a diagnosis of a glomangiomyoma, a rare glomus tumor variant (Figure). Glomus tumors are benign neoplasms comprising glomus bodies, which are arteriovenous anastomoses involved in thermoregulation.¹ They develop in areas densely populated by glomus bodies, including the fingers, toes, and subungual areas. Glomus tumors most commonly develop in middle-aged women.² Clinically, they manifest with a characteristic triad of intense pain, point tenderness, and cold sensitivity and may appear as reddish-pink or blue macules under the nail plate and/or longitudinal erythronychia.^2-6 The presence of multiple glomus tumors is associated with neurofibromatosis type 1.⁷ 

Advanced imaging including ultrasonography and magnetic resonance imaging (MRI) may help confirm the diagnosis but may not be cost effective, as excision with histopathology is needed to relieve symptoms and render a definitive diagnosis. Radiography is highly insensitive in identifying bone erosions associated with glomus tumors.⁸ With ultrasonography, glomus tumors appear hypoechoic; with Doppler ultrasonography, they appear hypervascular. With MRI, glomus tumors appear as well-defined nodular lesions with hypointense signal intensity on T1-weighted sequence and hyperintense signal intensity on T2-weighted sequence, with strong enhancement using gadolinium-based contrast.^9,10 On histopathology, a glomus tumor appears as a nodular tumor with sheets of oval-nucleated cells arranged in multicellular layers surrounding blood vessels and are immunoreactive for α-smooth muscle actin, muscle-specific actin, and type IV collagen.^11,12 

There are several glomus tumor variants. The most common is a solid glomus tumor, which predominantly is composed of glomus cells, followed by glomangioma, which mainly is composed of blood vessels. Glomangiomyoma, which mostly is composed of smooth muscle cells, is the rarest variant.¹³ 

While glomus tumors are common in the subungual areas, it is an uncommon location for glomangiomyomas, which have been reported in the nail unit in only 7 prior case reports identified through searches of PubMed and Google Scholar using the terms glomangiomyoma, glomangiomyoma nail, and subungual glomangiomyoma (Table).^13-19 Glomangiomyomas more commonly are described in solid organs, including the stomach, kidney, pancreas, and bladder.¹⁶ The mean age of patients with subungual glomangiomyomas, including our patient, was 40.4 years (range, 3-61 years), with the majority being female (75.0% [6/8]). Most patients presented with fingernail involvement (75.0% [6/8]), nail dystrophy (eg, nail plate thinning, longitudinal grooves, splinter hemorrhages, longitudinal erythronychia)(62.5% [5/8]), and intermittent pain and/or point tenderness in the affected nail (75.0% [6/8]).^13-19 Notably, only our patient had longitudinal erythronychia as a clinical feature, and only one other case described MRI findings, which included a lobulated mass with intense contrast and distal phalanx destruction.¹⁸ One patient was a 3-year-old girl with a family history of generalized multiple glomangiomyomas. Although subungual glomangiomyoma was not confirmed on histopathology, the diagnosis in this patient was presumed based on her family history.¹³ On histopathology, glomangiomyomas are composed of oval-nucleated cells surrounding blood vessels. These oval-nucleated cells then gradually transition to smooth muscle cells.²⁰ 

A myxoid cyst is composed of a pseudocyst, which lacks a cyst lining, and is a result of synovial fluid from the distal interphalangeal joint entering the pseudocyst space.² It typically manifests with a longitudinal groove in the nail plate. A flesh-colored nodule may be appreciated between the cuticle and the distal interphalangeal joint.² The depth of the longitudinal groove may vary depending on the volume of synovial fluid within the myxoid cyst.²¹ In a series of 35 cases of subungual myxoid cysts, none manifested with longitudinal erythronychia. Due to their composition, myxoid cysts can be distinguished easily from solid tumors of the nail unit via transillumination.²² Pain is a much less common with myxoid cysts vs glomus tumors, as the filling of the pseudocyst space with synovial fluid typically is gradual, allowing the surrounding tissue to accommodate and adapt over time.²¹ In equivocal cases, MRI or high-resolution ultrasonography may be used to distinguish myxoid cysts and glomus tumors.⁸ Histopathology shows accumulation of mucin in the dermis with surrounding fibrous stroma.²³

Subungual neuromas are painful benign tumors that develop due to disorganized neural proliferation following disruption to peripheral nerves secondary to trauma or surgery. In 3 case reports, subungual neuromas manifested as painful subungual nodules, with proximal nail plate ridging, or onycholysis.^24-26 Since neuromas have only rarely been described in the subungual region, reports of MRI and ultrasonography findings are unknown. Histopathology is needed to distinguish neuromas from glomus tumors. Histopathology shows an acapsular structure consisting of disorganized spindle-cell proliferation and nerve fibers arranged in a tangle of fascicles within fibrotic tissue.²⁵ On immunochemistry, spindle cells typically are positive for cellular antigen protein S100.²⁶ 

Leiomyomas are benign neoplasms derived from smooth muscle, typically localized to the uterus or gastrointestinal tract, and have been described rarely in the nail unit.^27,28 It is hypothesized that subungual leiomyomas originate from the vascular smooth muscle in the subcutaneous layer of the nail unit.²⁸ Like glomus tumors, leiomyomas of the subungual region often manifest with pain and longitudinal erythronychia.^27-30 Subungual leiomyomas may be distinguished from glomus tumors via advanced imaging techniques, including ultrasonography and MRI. Cutaneous leiomyomas have been described with mild to moderate internal low flow vascularity on Doppler ultrasonography, while glomus tumors typically reveal high internal vascularity.²⁸ Biopsy with histopathology is needed for definitive diagnosis. On histopathology, leiomyomas demonstrate bland-appearing spindle-shaped cells with elongated nuclei arranged in fascicles.²⁷ They typically are positive for α-smooth muscle actin and caldesmon on immunostaining. 

Eccrine spiradenomas are benign adnexal tumors likely of apocrine origin with limited case reports in the literature.^31,32 Clinically, eccrine spiradenomas involving the nail unit may manifest with longitudinal nail splitting of the nail or as a papule on the proximal nail fold, with associated tenderness.^31,32 In a report of a 50-year-old woman with a histopathologically confirmed eccrine spiradenoma manifesting with longitudinal splitting of the nail and pain in the proximal nail fold, the mass appeared hypoechoic on ultrasonography with increased intramass vascularity on Doppler, while MRI showed an intensely enhancing lesion.³¹ These imaging features, combined with a classically manifesting feature of pain, make eccrine spiradenomas difficult to distinguish from glomus tumors; therefore, histopathologic examination can provide a definitive diagnosis, and surgical excision is used for treatment.³¹ On histopathology, these tumors are well circumscribed and composed of both small dark basaloid cells with peripheral compact nuclei and larger cells with central pale nuclei, which may be arranged in tubules.^31,32

The Diagnosis: Glomangiomyoma

The nail unit excision specimen showed collections of cuboidal cells and spindled cells within the corium that were consistent with a diagnosis of a glomangiomyoma, a rare glomus tumor variant (Figure). Glomus tumors are benign neoplasms comprising glomus bodies, which are arteriovenous anastomoses involved in thermoregulation.¹ They develop in areas densely populated by glomus bodies, including the fingers, toes, and subungual areas. Glomus tumors most commonly develop in middle-aged women.² Clinically, they manifest with a characteristic triad of intense pain, point tenderness, and cold sensitivity and may appear as reddish-pink or blue macules under the nail plate and/or longitudinal erythronychia.^2-6 The presence of multiple glomus tumors is associated with neurofibromatosis type 1.⁷ 

Advanced imaging including ultrasonography and magnetic resonance imaging (MRI) may help confirm the diagnosis but may not be cost effective, as excision with histopathology is needed to relieve symptoms and render a definitive diagnosis. Radiography is highly insensitive in identifying bone erosions associated with glomus tumors.⁸ With ultrasonography, glomus tumors appear hypoechoic; with Doppler ultrasonography, they appear hypervascular. With MRI, glomus tumors appear as well-defined nodular lesions with hypointense signal intensity on T1-weighted sequence and hyperintense signal intensity on T2-weighted sequence, with strong enhancement using gadolinium-based contrast.^9,10 On histopathology, a glomus tumor appears as a nodular tumor with sheets of oval-nucleated cells arranged in multicellular layers surrounding blood vessels and are immunoreactive for α-smooth muscle actin, muscle-specific actin, and type IV collagen.^11,12 

There are several glomus tumor variants. The most common is a solid glomus tumor, which predominantly is composed of glomus cells, followed by glomangioma, which mainly is composed of blood vessels. Glomangiomyoma, which mostly is composed of smooth muscle cells, is the rarest variant.¹³ 

While glomus tumors are common in the subungual areas, it is an uncommon location for glomangiomyomas, which have been reported in the nail unit in only 7 prior case reports identified through searches of PubMed and Google Scholar using the terms glomangiomyoma, glomangiomyoma nail, and subungual glomangiomyoma (Table).^13-19 Glomangiomyomas more commonly are described in solid organs, including the stomach, kidney, pancreas, and bladder.¹⁶ The mean age of patients with subungual glomangiomyomas, including our patient, was 40.4 years (range, 3-61 years), with the majority being female (75.0% [6/8]). Most patients presented with fingernail involvement (75.0% [6/8]), nail dystrophy (eg, nail plate thinning, longitudinal grooves, splinter hemorrhages, longitudinal erythronychia)(62.5% [5/8]), and intermittent pain and/or point tenderness in the affected nail (75.0% [6/8]).^13-19 Notably, only our patient had longitudinal erythronychia as a clinical feature, and only one other case described MRI findings, which included a lobulated mass with intense contrast and distal phalanx destruction.¹⁸ One patient was a 3-year-old girl with a family history of generalized multiple glomangiomyomas. Although subungual glomangiomyoma was not confirmed on histopathology, the diagnosis in this patient was presumed based on her family history.¹³ On histopathology, glomangiomyomas are composed of oval-nucleated cells surrounding blood vessels. These oval-nucleated cells then gradually transition to smooth muscle cells.²⁰ 

A myxoid cyst is composed of a pseudocyst, which lacks a cyst lining, and is a result of synovial fluid from the distal interphalangeal joint entering the pseudocyst space.² It typically manifests with a longitudinal groove in the nail plate. A flesh-colored nodule may be appreciated between the cuticle and the distal interphalangeal joint.² The depth of the longitudinal groove may vary depending on the volume of synovial fluid within the myxoid cyst.²¹ In a series of 35 cases of subungual myxoid cysts, none manifested with longitudinal erythronychia. Due to their composition, myxoid cysts can be distinguished easily from solid tumors of the nail unit via transillumination.²² Pain is a much less common with myxoid cysts vs glomus tumors, as the filling of the pseudocyst space with synovial fluid typically is gradual, allowing the surrounding tissue to accommodate and adapt over time.²¹ In equivocal cases, MRI or high-resolution ultrasonography may be used to distinguish myxoid cysts and glomus tumors.⁸ Histopathology shows accumulation of mucin in the dermis with surrounding fibrous stroma.²³

Subungual neuromas are painful benign tumors that develop due to disorganized neural proliferation following disruption to peripheral nerves secondary to trauma or surgery. In 3 case reports, subungual neuromas manifested as painful subungual nodules, with proximal nail plate ridging, or onycholysis.^24-26 Since neuromas have only rarely been described in the subungual region, reports of MRI and ultrasonography findings are unknown. Histopathology is needed to distinguish neuromas from glomus tumors. Histopathology shows an acapsular structure consisting of disorganized spindle-cell proliferation and nerve fibers arranged in a tangle of fascicles within fibrotic tissue.²⁵ On immunochemistry, spindle cells typically are positive for cellular antigen protein S100.²⁶ 

Leiomyomas are benign neoplasms derived from smooth muscle, typically localized to the uterus or gastrointestinal tract, and have been described rarely in the nail unit.^27,28 It is hypothesized that subungual leiomyomas originate from the vascular smooth muscle in the subcutaneous layer of the nail unit.²⁸ Like glomus tumors, leiomyomas of the subungual region often manifest with pain and longitudinal erythronychia.^27-30 Subungual leiomyomas may be distinguished from glomus tumors via advanced imaging techniques, including ultrasonography and MRI. Cutaneous leiomyomas have been described with mild to moderate internal low flow vascularity on Doppler ultrasonography, while glomus tumors typically reveal high internal vascularity.²⁸ Biopsy with histopathology is needed for definitive diagnosis. On histopathology, leiomyomas demonstrate bland-appearing spindle-shaped cells with elongated nuclei arranged in fascicles.²⁷ They typically are positive for α-smooth muscle actin and caldesmon on immunostaining. 

Eccrine spiradenomas are benign adnexal tumors likely of apocrine origin with limited case reports in the literature.^31,32 Clinically, eccrine spiradenomas involving the nail unit may manifest with longitudinal nail splitting of the nail or as a papule on the proximal nail fold, with associated tenderness.^31,32 In a report of a 50-year-old woman with a histopathologically confirmed eccrine spiradenoma manifesting with longitudinal splitting of the nail and pain in the proximal nail fold, the mass appeared hypoechoic on ultrasonography with increased intramass vascularity on Doppler, while MRI showed an intensely enhancing lesion.³¹ These imaging features, combined with a classically manifesting feature of pain, make eccrine spiradenomas difficult to distinguish from glomus tumors; therefore, histopathologic examination can provide a definitive diagnosis, and surgical excision is used for treatment.³¹ On histopathology, these tumors are well circumscribed and composed of both small dark basaloid cells with peripheral compact nuclei and larger cells with central pale nuclei, which may be arranged in tubules.^31,32

The Diagnosis: Glomangiomyoma

The nail unit excision specimen showed collections of cuboidal cells and spindled cells within the corium that were consistent with a diagnosis of a glomangiomyoma, a rare glomus tumor variant (Figure). Glomus tumors are benign neoplasms comprising glomus bodies, which are arteriovenous anastomoses involved in thermoregulation.¹ They develop in areas densely populated by glomus bodies, including the fingers, toes, and subungual areas. Glomus tumors most commonly develop in middle-aged women.² Clinically, they manifest with a characteristic triad of intense pain, point tenderness, and cold sensitivity and may appear as reddish-pink or blue macules under the nail plate and/or longitudinal erythronychia.^2-6 The presence of multiple glomus tumors is associated with neurofibromatosis type 1.⁷ 

Advanced imaging including ultrasonography and magnetic resonance imaging (MRI) may help confirm the diagnosis but may not be cost effective, as excision with histopathology is needed to relieve symptoms and render a definitive diagnosis. Radiography is highly insensitive in identifying bone erosions associated with glomus tumors.⁸ With ultrasonography, glomus tumors appear hypoechoic; with Doppler ultrasonography, they appear hypervascular. With MRI, glomus tumors appear as well-defined nodular lesions with hypointense signal intensity on T1-weighted sequence and hyperintense signal intensity on T2-weighted sequence, with strong enhancement using gadolinium-based contrast.^9,10 On histopathology, a glomus tumor appears as a nodular tumor with sheets of oval-nucleated cells arranged in multicellular layers surrounding blood vessels and are immunoreactive for α-smooth muscle actin, muscle-specific actin, and type IV collagen.^11,12 

There are several glomus tumor variants. The most common is a solid glomus tumor, which predominantly is composed of glomus cells, followed by glomangioma, which mainly is composed of blood vessels. Glomangiomyoma, which mostly is composed of smooth muscle cells, is the rarest variant.¹³ 

While glomus tumors are common in the subungual areas, it is an uncommon location for glomangiomyomas, which have been reported in the nail unit in only 7 prior case reports identified through searches of PubMed and Google Scholar using the terms glomangiomyoma, glomangiomyoma nail, and subungual glomangiomyoma (Table).^13-19 Glomangiomyomas more commonly are described in solid organs, including the stomach, kidney, pancreas, and bladder.¹⁶ The mean age of patients with subungual glomangiomyomas, including our patient, was 40.4 years (range, 3-61 years), with the majority being female (75.0% [6/8]). Most patients presented with fingernail involvement (75.0% [6/8]), nail dystrophy (eg, nail plate thinning, longitudinal grooves, splinter hemorrhages, longitudinal erythronychia)(62.5% [5/8]), and intermittent pain and/or point tenderness in the affected nail (75.0% [6/8]).^13-19 Notably, only our patient had longitudinal erythronychia as a clinical feature, and only one other case described MRI findings, which included a lobulated mass with intense contrast and distal phalanx destruction.¹⁸ One patient was a 3-year-old girl with a family history of generalized multiple glomangiomyomas. Although subungual glomangiomyoma was not confirmed on histopathology, the diagnosis in this patient was presumed based on her family history.¹³ On histopathology, glomangiomyomas are composed of oval-nucleated cells surrounding blood vessels. These oval-nucleated cells then gradually transition to smooth muscle cells.²⁰ 

A myxoid cyst is composed of a pseudocyst, which lacks a cyst lining, and is a result of synovial fluid from the distal interphalangeal joint entering the pseudocyst space.² It typically manifests with a longitudinal groove in the nail plate. A flesh-colored nodule may be appreciated between the cuticle and the distal interphalangeal joint.² The depth of the longitudinal groove may vary depending on the volume of synovial fluid within the myxoid cyst.²¹ In a series of 35 cases of subungual myxoid cysts, none manifested with longitudinal erythronychia. Due to their composition, myxoid cysts can be distinguished easily from solid tumors of the nail unit via transillumination.²² Pain is a much less common with myxoid cysts vs glomus tumors, as the filling of the pseudocyst space with synovial fluid typically is gradual, allowing the surrounding tissue to accommodate and adapt over time.²¹ In equivocal cases, MRI or high-resolution ultrasonography may be used to distinguish myxoid cysts and glomus tumors.⁸ Histopathology shows accumulation of mucin in the dermis with surrounding fibrous stroma.²³

Subungual neuromas are painful benign tumors that develop due to disorganized neural proliferation following disruption to peripheral nerves secondary to trauma or surgery. In 3 case reports, subungual neuromas manifested as painful subungual nodules, with proximal nail plate ridging, or onycholysis.^24-26 Since neuromas have only rarely been described in the subungual region, reports of MRI and ultrasonography findings are unknown. Histopathology is needed to distinguish neuromas from glomus tumors. Histopathology shows an acapsular structure consisting of disorganized spindle-cell proliferation and nerve fibers arranged in a tangle of fascicles within fibrotic tissue.²⁵ On immunochemistry, spindle cells typically are positive for cellular antigen protein S100.²⁶ 

Leiomyomas are benign neoplasms derived from smooth muscle, typically localized to the uterus or gastrointestinal tract, and have been described rarely in the nail unit.^27,28 It is hypothesized that subungual leiomyomas originate from the vascular smooth muscle in the subcutaneous layer of the nail unit.²⁸ Like glomus tumors, leiomyomas of the subungual region often manifest with pain and longitudinal erythronychia.^27-30 Subungual leiomyomas may be distinguished from glomus tumors via advanced imaging techniques, including ultrasonography and MRI. Cutaneous leiomyomas have been described with mild to moderate internal low flow vascularity on Doppler ultrasonography, while glomus tumors typically reveal high internal vascularity.²⁸ Biopsy with histopathology is needed for definitive diagnosis. On histopathology, leiomyomas demonstrate bland-appearing spindle-shaped cells with elongated nuclei arranged in fascicles.²⁷ They typically are positive for α-smooth muscle actin and caldesmon on immunostaining. 

Eccrine spiradenomas are benign adnexal tumors likely of apocrine origin with limited case reports in the literature.^31,32 Clinically, eccrine spiradenomas involving the nail unit may manifest with longitudinal nail splitting of the nail or as a papule on the proximal nail fold, with associated tenderness.^31,32 In a report of a 50-year-old woman with a histopathologically confirmed eccrine spiradenoma manifesting with longitudinal splitting of the nail and pain in the proximal nail fold, the mass appeared hypoechoic on ultrasonography with increased intramass vascularity on Doppler, while MRI showed an intensely enhancing lesion.³¹ These imaging features, combined with a classically manifesting feature of pain, make eccrine spiradenomas difficult to distinguish from glomus tumors; therefore, histopathologic examination can provide a definitive diagnosis, and surgical excision is used for treatment.³¹ On histopathology, these tumors are well circumscribed and composed of both small dark basaloid cells with peripheral compact nuclei and larger cells with central pale nuclei, which may be arranged in tubules.^31,32

The requisites of government are that there be sufficiency of food, sufficiency of military equipment, and the confidence of the people in their ruler.

Analects by Confucius¹

From ancient festivals to modern holidays, autumn has long been associated with the gathering of the harvest. Friends and families come together around tables laden with delicious food to enjoy the pleasures of peace and plenty. During these celebrations, we must never forget that without the strength of the nation’s military and the service of its veterans, this freedom and abundance would not be possible. Our debt of gratitude to the current and former members of the armed services makes the fact that a substantial minority experiences food insecurity not only a human tragedy, but a travesty of the nation’s promise to support those who wear or have worn the uniform.

The National Defense Authorization Act for Fiscal Year 2020 charged the Secretary of Defense to investigate food insecurity among active-duty service members and their dependents.² The RAND Corporation conducted the assessment and, based on the results of its analysis, made recommendations to reduce hunger among armed forces members and their families.³

The RAND study found that 10% of active-duty military met US Department of Agriculture (USDA) criteria for very low food security; another 15% were classified as having low food security. The USDA defines food insecurity with hunger as “reports of multiple indications of disrupted eating patterns and reduced food intake.” USDA defines low food security as “reports of reduced quality, variety, or desirability of diet. Little or no indication of reduced food intake.”⁴

As someone who grew up on an Army base with the commissary a short trip from military housing, I was unpleasantly surprised that food insecurity was more common among in-service members living on post. I was even more dismayed to read that a variety of factors constrained 14% of active-duty military experiencing food insecurity to seek public assistance to feed themselves and their families. As with so many health care and social services, (eg, mental health care), those wearing the uniform were concerned that participating in a food assistance program would damage their career or stigmatize them. Others did not seek help, perhaps because they believed they were not eligible, and in many cases were correct: they did not qualify for food banks or food stamps due to receiving other benefits. A variety of factors contribute to periods of food insecurity among military families, including remote or rural bases that lack access to grocery stores or jobs for partners or other family members, and low base military pay.⁵

Food insecurity is an even more serious concern among veterans who are frequently older and have more comorbidities, often leading to unemployment and homelessness. Feeding America, the nation’s largest organization of community food banks, estimates that 1 in 9 working-age veterans are food insecure.⁵ US Department of Veterans Affairs (VA) statistics indicate that veterans are 7% more likely to experience food insecurity than other sectors of the population.⁶ The Veterans Health Administration has recognized that food insecurity is directly related to medical problems already common among veterans, including diabetes, obesity, and depression. Women and minority veterans are the most at risk of food insecurity.⁷

Recognizing that many veterans are at risk of food insecurity, the US Department of Defense and VA have taken steps to try and reduce hunger among those who serve. In response to the shocking statistic that food insecurity was found in 27% of Iraq and Afghanistan veterans, the VA and Rockefeller Foundation are partnering on the Food as Medicine initiative to improve veteran nutrition as a means of improving nutrition-related health consequences of food insecurity.⁸

Like many federal practitioners, I was unaware of the food insecurity assistance available to active-duty service members or veterans, or how to help individuals access it. In addition to the resources outlined in the Table, there are many community-based options open to anyone, including veterans and service members. 

I have written columns on many difficult issues in my years as the Editor-in-Chief of Federal Practitioner, but personally this is one of the most distressing editorials I have ever published. That individuals dedicated to defending our rights and protecting our safety should be compelled to go hungry or not know if they have enough money at the end of the month to buy food is manifestly unjust. It is challenging when faced with such a large-scale injustice to think we cannot make a difference, but that resignation or abdication only magnifies this inequity. I have a friend who kept giving back even after they retired from federal service: they volunteered at a community garden and brought produce to the local food bank and helped distribute it. That may seem too much for those still working yet almost anyone can pick up a few items on their weekly shopping trip and donate them to a food drive. 

As we approach Veterans Day, let’s not just express our gratitude to our military and veterans in words but in deeds like feeding the hungry and urging elected representatives to fulfill their commitment to ensure that service members and veterans and their families do not experience food insecurity. Confucian wisdom written in a very distant time and vastly dissimilar context still rings true: there are direct and critical links between food and trust and between hunger and the military.¹

The requisites of government are that there be sufficiency of food, sufficiency of military equipment, and the confidence of the people in their ruler.

Analects by Confucius¹

From ancient festivals to modern holidays, autumn has long been associated with the gathering of the harvest. Friends and families come together around tables laden with delicious food to enjoy the pleasures of peace and plenty. During these celebrations, we must never forget that without the strength of the nation’s military and the service of its veterans, this freedom and abundance would not be possible. Our debt of gratitude to the current and former members of the armed services makes the fact that a substantial minority experiences food insecurity not only a human tragedy, but a travesty of the nation’s promise to support those who wear or have worn the uniform.

The National Defense Authorization Act for Fiscal Year 2020 charged the Secretary of Defense to investigate food insecurity among active-duty service members and their dependents.² The RAND Corporation conducted the assessment and, based on the results of its analysis, made recommendations to reduce hunger among armed forces members and their families.³

The RAND study found that 10% of active-duty military met US Department of Agriculture (USDA) criteria for very low food security; another 15% were classified as having low food security. The USDA defines food insecurity with hunger as “reports of multiple indications of disrupted eating patterns and reduced food intake.” USDA defines low food security as “reports of reduced quality, variety, or desirability of diet. Little or no indication of reduced food intake.”⁴

As someone who grew up on an Army base with the commissary a short trip from military housing, I was unpleasantly surprised that food insecurity was more common among in-service members living on post. I was even more dismayed to read that a variety of factors constrained 14% of active-duty military experiencing food insecurity to seek public assistance to feed themselves and their families. As with so many health care and social services, (eg, mental health care), those wearing the uniform were concerned that participating in a food assistance program would damage their career or stigmatize them. Others did not seek help, perhaps because they believed they were not eligible, and in many cases were correct: they did not qualify for food banks or food stamps due to receiving other benefits. A variety of factors contribute to periods of food insecurity among military families, including remote or rural bases that lack access to grocery stores or jobs for partners or other family members, and low base military pay.⁵

Food insecurity is an even more serious concern among veterans who are frequently older and have more comorbidities, often leading to unemployment and homelessness. Feeding America, the nation’s largest organization of community food banks, estimates that 1 in 9 working-age veterans are food insecure.⁵ US Department of Veterans Affairs (VA) statistics indicate that veterans are 7% more likely to experience food insecurity than other sectors of the population.⁶ The Veterans Health Administration has recognized that food insecurity is directly related to medical problems already common among veterans, including diabetes, obesity, and depression. Women and minority veterans are the most at risk of food insecurity.⁷

Recognizing that many veterans are at risk of food insecurity, the US Department of Defense and VA have taken steps to try and reduce hunger among those who serve. In response to the shocking statistic that food insecurity was found in 27% of Iraq and Afghanistan veterans, the VA and Rockefeller Foundation are partnering on the Food as Medicine initiative to improve veteran nutrition as a means of improving nutrition-related health consequences of food insecurity.⁸

Like many federal practitioners, I was unaware of the food insecurity assistance available to active-duty service members or veterans, or how to help individuals access it. In addition to the resources outlined in the Table, there are many community-based options open to anyone, including veterans and service members. 

I have written columns on many difficult issues in my years as the Editor-in-Chief of Federal Practitioner, but personally this is one of the most distressing editorials I have ever published. That individuals dedicated to defending our rights and protecting our safety should be compelled to go hungry or not know if they have enough money at the end of the month to buy food is manifestly unjust. It is challenging when faced with such a large-scale injustice to think we cannot make a difference, but that resignation or abdication only magnifies this inequity. I have a friend who kept giving back even after they retired from federal service: they volunteered at a community garden and brought produce to the local food bank and helped distribute it. That may seem too much for those still working yet almost anyone can pick up a few items on their weekly shopping trip and donate them to a food drive. 

As we approach Veterans Day, let’s not just express our gratitude to our military and veterans in words but in deeds like feeding the hungry and urging elected representatives to fulfill their commitment to ensure that service members and veterans and their families do not experience food insecurity. Confucian wisdom written in a very distant time and vastly dissimilar context still rings true: there are direct and critical links between food and trust and between hunger and the military.¹

The requisites of government are that there be sufficiency of food, sufficiency of military equipment, and the confidence of the people in their ruler.

Analects by Confucius¹

From ancient festivals to modern holidays, autumn has long been associated with the gathering of the harvest. Friends and families come together around tables laden with delicious food to enjoy the pleasures of peace and plenty. During these celebrations, we must never forget that without the strength of the nation’s military and the service of its veterans, this freedom and abundance would not be possible. Our debt of gratitude to the current and former members of the armed services makes the fact that a substantial minority experiences food insecurity not only a human tragedy, but a travesty of the nation’s promise to support those who wear or have worn the uniform.

The National Defense Authorization Act for Fiscal Year 2020 charged the Secretary of Defense to investigate food insecurity among active-duty service members and their dependents.² The RAND Corporation conducted the assessment and, based on the results of its analysis, made recommendations to reduce hunger among armed forces members and their families.³

The RAND study found that 10% of active-duty military met US Department of Agriculture (USDA) criteria for very low food security; another 15% were classified as having low food security. The USDA defines food insecurity with hunger as “reports of multiple indications of disrupted eating patterns and reduced food intake.” USDA defines low food security as “reports of reduced quality, variety, or desirability of diet. Little or no indication of reduced food intake.”⁴

As someone who grew up on an Army base with the commissary a short trip from military housing, I was unpleasantly surprised that food insecurity was more common among in-service members living on post. I was even more dismayed to read that a variety of factors constrained 14% of active-duty military experiencing food insecurity to seek public assistance to feed themselves and their families. As with so many health care and social services, (eg, mental health care), those wearing the uniform were concerned that participating in a food assistance program would damage their career or stigmatize them. Others did not seek help, perhaps because they believed they were not eligible, and in many cases were correct: they did not qualify for food banks or food stamps due to receiving other benefits. A variety of factors contribute to periods of food insecurity among military families, including remote or rural bases that lack access to grocery stores or jobs for partners or other family members, and low base military pay.⁵

Food insecurity is an even more serious concern among veterans who are frequently older and have more comorbidities, often leading to unemployment and homelessness. Feeding America, the nation’s largest organization of community food banks, estimates that 1 in 9 working-age veterans are food insecure.⁵ US Department of Veterans Affairs (VA) statistics indicate that veterans are 7% more likely to experience food insecurity than other sectors of the population.⁶ The Veterans Health Administration has recognized that food insecurity is directly related to medical problems already common among veterans, including diabetes, obesity, and depression. Women and minority veterans are the most at risk of food insecurity.⁷

Recognizing that many veterans are at risk of food insecurity, the US Department of Defense and VA have taken steps to try and reduce hunger among those who serve. In response to the shocking statistic that food insecurity was found in 27% of Iraq and Afghanistan veterans, the VA and Rockefeller Foundation are partnering on the Food as Medicine initiative to improve veteran nutrition as a means of improving nutrition-related health consequences of food insecurity.⁸

Like many federal practitioners, I was unaware of the food insecurity assistance available to active-duty service members or veterans, or how to help individuals access it. In addition to the resources outlined in the Table, there are many community-based options open to anyone, including veterans and service members. 

I have written columns on many difficult issues in my years as the Editor-in-Chief of Federal Practitioner, but personally this is one of the most distressing editorials I have ever published. That individuals dedicated to defending our rights and protecting our safety should be compelled to go hungry or not know if they have enough money at the end of the month to buy food is manifestly unjust. It is challenging when faced with such a large-scale injustice to think we cannot make a difference, but that resignation or abdication only magnifies this inequity. I have a friend who kept giving back even after they retired from federal service: they volunteered at a community garden and brought produce to the local food bank and helped distribute it. That may seem too much for those still working yet almost anyone can pick up a few items on their weekly shopping trip and donate them to a food drive. 

As we approach Veterans Day, let’s not just express our gratitude to our military and veterans in words but in deeds like feeding the hungry and urging elected representatives to fulfill their commitment to ensure that service members and veterans and their families do not experience food insecurity. Confucian wisdom written in a very distant time and vastly dissimilar context still rings true: there are direct and critical links between food and trust and between hunger and the military.¹