Case Reports

With the increasing utilization of telemedicine since the COVID-19 pandemic, it is critical that clinicians have an appropriate understanding of the application of virtual care resources, including teledermatology. We present a case series of 3 patients to demonstrate the clinical utility of teledermatology in reducing the time to diagnosis of various rare and/or aggressive cutaneous malignancies, including Merkel cell carcinoma, malignant melanoma, and atypical fibroxanthoma. Cases were obtained from one large Midwestern medical center during the month of July 2021. Each case presented includes a description of the initial teledermatology presentation and reviews the clinical timeline from initial consultation submission to in-person clinic visit with lesion biopsy. This case series demonstrates real-world examples of how teledermatology can be utilized to expedite the care of specific vulnerable patient populations.

Teledermatology is a rapidly growing digital resource with specific utility in triaging patients to determine those requiring in-person evaluation for early and accurate detection of skin malignancies. Approximately one-third of teledermatology consultations result in face-to-face clinical encounters, with malignant neoplasms being the leading cause for biopsy.^1,2 For specific populations, such as geriatric and immunocompromised patients, teledermatology may serve as a valuable tool, particularly in the wake of the COVID-19 pandemic. Furthermore, telemedicine may aid in addressing health disparities within the field of medicine and ultimately may improve access to care for vulnerable populations.³ Along with increasing access to specific subspecialty expertise, the use of teledermatology may reduce health care costs and improve the overall quality of care delivered to patients.^4,5

We describe the clinical utility of teledermatology in triaging and diagnosing skin malignancies through a series of 3 cases obtained from digital image review at one large Midwestern medical center during the month of July 2021. Three unique cases with a final diagnosis of a rare or aggressive skin cancer were selected as examples, including a 75-year-old man with Merkle cell carcinoma, a 55-year-old man with aggressive pT3b malignant melanoma, and a 72-year-old man with an atypical fibroxanthoma. A clinical timeline of each case is presented, including the time intervals from initial image submission to image review, image submission to face-to-face clinical encounter, and image submission to final diagnosis. In all cases, the primary care provider submitted an order for teledermatology, and the teledermatology team obtained the images.

Case Series

Patient 1—Images of the right hand of a 75-year-old man with a medical history of basal cell carcinoma were submitted for teledermatology consultation utilizing store-and-forward image-capturing technology (day 1). The patient history provided with image submission indicated that the lesion had been present for 6 months and there were no associated symptoms. Clinical imaging demonstrated a pink-red pearly papule located on the proximal fourth digit of the dorsal aspect of the right hand (Figure 1). One day following the teledermatology request (day 2), the patient’s case was reviewed and triaged for an in-person visit. The patient was brought to clinic on day 34, and a biopsy was performed. On day 36, dermatopathology results indicated a diagnosis of Merkel cell carcinoma. On day 37, the patient was referred to surgical oncology, and on day 44, the patient underwent an initial surgical oncology visit with a plan for wide local excision of the right fourth digit with right axillary sentinel lymph node biopsy.

Patient 2—Images of the left flank of a 55-year-old man were submitted for teledermatology consultation via store-and-forward technology (day 1). A patient history provided with the image indicated that the lesion had been present for months to years and there were no associated symptoms, but the lesion recently had changed in color and size. Teledermatology images were reviewed on day 3 and demonstrated a 2- to 3-cm brown plaque on the left flank with color variegation and a prominent red papule protruding centrally (Figure 2). The patient was scheduled for an urgent in-person visit with biopsy. On day 6, the patient presented to clinic and an excision biopsy was performed. Dermatopathology was ordered with a RUSH indication, with results on day 7 revealing a pT3b malignant melanoma. An urgent consultation to surgical oncology was placed on the same day, and the patient underwent an initial surgical oncology visit on day 24 with a plan for wide local excision with left axillary and inguinal sentinel lymph node biopsy.

Patient 3—Images of the left ear of a 72-year-old man were submitted for teledermatology consultation utilizing review via store-and-forward technology (day 1). A patient history indicated that the lesion had been present for 3 months with associated bleeding. Image review demonstrated a solitary pearly pink papule located on the crura of the antihelix (Figure 3). Initial teledermatology consultation was reviewed on day 2 with notification of the need for in-person evaluation. The patient presented to clinic on day 33 for a biopsy, with dermatopathology results on day 36 consistent with an atypical fibroxanthoma. The patient was scheduled for Mohs micrographic surgery on day 37 and underwent surgical treatment on day 64.

Comment

Teledermatology consultations from all patients demonstrated adequate image quality to be able to evaluate the lesion of concern and yielded a request for in-person evaluation with possible biopsy (Table). In this case series, the average time interval from teledermatology consultation placement to teledermatology image report was 2 days (range, 1–3 days). The average time from teledermatology consultation placement to face-to-face encounter with biopsy was 24.3 days for the 3 cases presented in this series (range, 6–34 days). The initial surgical oncology visits took place an average of 34 days after the initial teledermatology consultation was placed for the 2 patients requiring referral (44 days for patient 1; 24 days for patient 2). For patient 3, Mohs micrographic surgery was required for treatment, which was scheduled by day 37 and subsequently performed on day 64.

When specifically looking at the diagnosis of cutaneous malignancies, studies have found that the incidence of skin cancer detection is similar for teledermatology compared to in-person clinic visits.^6,7 Creighton-Smith et al⁶ performed a retrospective cohort study comparing prebiopsy and postbiopsy diagnostic accuracy and detection rates of skin cancer between store-and-forward technology and face-to-face consultation. When adjusting for possible compounding factors including personal and family history of skin cancer, there was no notable difference in detection rates of any skin cancer, including melanoma and nonmelanoma skin cancers. Furthermore, the 2 cohorts of patients were found to have similar prebiopsy and postbiopsy diagnostic concordance, with similar times from consultation being placed to requested biopsy and time from biopsy to final treatment.⁶

Clarke et al⁷ similarly analyzed the accuracy of store-and-forward teledermatology and found that there was overall concordance in diagnosis when comparing clinical dermatologists to teledermatologists. Moreover, when melanocytic lesions were excluded from the study, the decision to biopsy did not differ substantially.⁷

Areas of further study include determining what percentage of teledermatology lesions of concern for malignancy were proven to be skin cancer after in-person evaluation and biopsy, as well as investigating the effectiveness of teledermatology for melanocytic lesions, which frequently are removed from analysis in large-scale teledermatology studies.

Although teledermatology has substantial clinical utility and may serve as a great resource for specific populations, including geriatric patients and those who are immunocompromised, it is important to recognize notable limitations. Specifically, brief history and image review should not serve as replacements for a face-to-face visit with physical examination in cases where the diagnosis remains uncertain or when high-risk skin malignancies are suspected or included in the differential. Certain aggressive cutaneous malignancies such as Merkel cell carcinoma may appear as less aggressive via teledermatology due to restrictions of technology.

Conclusion

Teledermatology has had a major impact on the way health care is delivered to patients and may increase access to care, reducing unnecessary in-person visits and decreasing the number of in-person visit no-shows. With the appropriate use of a brief clinical history and image review, teledermatology can be effective to evaluate specific lesions of concern. We report 3 unique cases identified during a 1-month period at a large Midwestern medical center. These cases serve as important examples of the application of teledermatology in reducing the time to diagnosis of aggressive skin malignancies. Further research on the clinical utility of teledermatology is warranted.

Acknowledgments—The authors thank the additional providers from the University of Wisconsin and William S. Middleton Memorial Veterans Hospital (both in Madison, Wisconsin) involved in the medical care of the patients included in this case series.

With the increasing utilization of telemedicine since the COVID-19 pandemic, it is critical that clinicians have an appropriate understanding of the application of virtual care resources, including teledermatology. We present a case series of 3 patients to demonstrate the clinical utility of teledermatology in reducing the time to diagnosis of various rare and/or aggressive cutaneous malignancies, including Merkel cell carcinoma, malignant melanoma, and atypical fibroxanthoma. Cases were obtained from one large Midwestern medical center during the month of July 2021. Each case presented includes a description of the initial teledermatology presentation and reviews the clinical timeline from initial consultation submission to in-person clinic visit with lesion biopsy. This case series demonstrates real-world examples of how teledermatology can be utilized to expedite the care of specific vulnerable patient populations.

Teledermatology is a rapidly growing digital resource with specific utility in triaging patients to determine those requiring in-person evaluation for early and accurate detection of skin malignancies. Approximately one-third of teledermatology consultations result in face-to-face clinical encounters, with malignant neoplasms being the leading cause for biopsy.^1,2 For specific populations, such as geriatric and immunocompromised patients, teledermatology may serve as a valuable tool, particularly in the wake of the COVID-19 pandemic. Furthermore, telemedicine may aid in addressing health disparities within the field of medicine and ultimately may improve access to care for vulnerable populations.³ Along with increasing access to specific subspecialty expertise, the use of teledermatology may reduce health care costs and improve the overall quality of care delivered to patients.^4,5

We describe the clinical utility of teledermatology in triaging and diagnosing skin malignancies through a series of 3 cases obtained from digital image review at one large Midwestern medical center during the month of July 2021. Three unique cases with a final diagnosis of a rare or aggressive skin cancer were selected as examples, including a 75-year-old man with Merkle cell carcinoma, a 55-year-old man with aggressive pT3b malignant melanoma, and a 72-year-old man with an atypical fibroxanthoma. A clinical timeline of each case is presented, including the time intervals from initial image submission to image review, image submission to face-to-face clinical encounter, and image submission to final diagnosis. In all cases, the primary care provider submitted an order for teledermatology, and the teledermatology team obtained the images.

Case Series

Patient 1—Images of the right hand of a 75-year-old man with a medical history of basal cell carcinoma were submitted for teledermatology consultation utilizing store-and-forward image-capturing technology (day 1). The patient history provided with image submission indicated that the lesion had been present for 6 months and there were no associated symptoms. Clinical imaging demonstrated a pink-red pearly papule located on the proximal fourth digit of the dorsal aspect of the right hand (Figure 1). One day following the teledermatology request (day 2), the patient’s case was reviewed and triaged for an in-person visit. The patient was brought to clinic on day 34, and a biopsy was performed. On day 36, dermatopathology results indicated a diagnosis of Merkel cell carcinoma. On day 37, the patient was referred to surgical oncology, and on day 44, the patient underwent an initial surgical oncology visit with a plan for wide local excision of the right fourth digit with right axillary sentinel lymph node biopsy.

Patient 2—Images of the left flank of a 55-year-old man were submitted for teledermatology consultation via store-and-forward technology (day 1). A patient history provided with the image indicated that the lesion had been present for months to years and there were no associated symptoms, but the lesion recently had changed in color and size. Teledermatology images were reviewed on day 3 and demonstrated a 2- to 3-cm brown plaque on the left flank with color variegation and a prominent red papule protruding centrally (Figure 2). The patient was scheduled for an urgent in-person visit with biopsy. On day 6, the patient presented to clinic and an excision biopsy was performed. Dermatopathology was ordered with a RUSH indication, with results on day 7 revealing a pT3b malignant melanoma. An urgent consultation to surgical oncology was placed on the same day, and the patient underwent an initial surgical oncology visit on day 24 with a plan for wide local excision with left axillary and inguinal sentinel lymph node biopsy.

Patient 3—Images of the left ear of a 72-year-old man were submitted for teledermatology consultation utilizing review via store-and-forward technology (day 1). A patient history indicated that the lesion had been present for 3 months with associated bleeding. Image review demonstrated a solitary pearly pink papule located on the crura of the antihelix (Figure 3). Initial teledermatology consultation was reviewed on day 2 with notification of the need for in-person evaluation. The patient presented to clinic on day 33 for a biopsy, with dermatopathology results on day 36 consistent with an atypical fibroxanthoma. The patient was scheduled for Mohs micrographic surgery on day 37 and underwent surgical treatment on day 64.

Comment

Teledermatology consultations from all patients demonstrated adequate image quality to be able to evaluate the lesion of concern and yielded a request for in-person evaluation with possible biopsy (Table). In this case series, the average time interval from teledermatology consultation placement to teledermatology image report was 2 days (range, 1–3 days). The average time from teledermatology consultation placement to face-to-face encounter with biopsy was 24.3 days for the 3 cases presented in this series (range, 6–34 days). The initial surgical oncology visits took place an average of 34 days after the initial teledermatology consultation was placed for the 2 patients requiring referral (44 days for patient 1; 24 days for patient 2). For patient 3, Mohs micrographic surgery was required for treatment, which was scheduled by day 37 and subsequently performed on day 64.

When specifically looking at the diagnosis of cutaneous malignancies, studies have found that the incidence of skin cancer detection is similar for teledermatology compared to in-person clinic visits.^6,7 Creighton-Smith et al⁶ performed a retrospective cohort study comparing prebiopsy and postbiopsy diagnostic accuracy and detection rates of skin cancer between store-and-forward technology and face-to-face consultation. When adjusting for possible compounding factors including personal and family history of skin cancer, there was no notable difference in detection rates of any skin cancer, including melanoma and nonmelanoma skin cancers. Furthermore, the 2 cohorts of patients were found to have similar prebiopsy and postbiopsy diagnostic concordance, with similar times from consultation being placed to requested biopsy and time from biopsy to final treatment.⁶

Clarke et al⁷ similarly analyzed the accuracy of store-and-forward teledermatology and found that there was overall concordance in diagnosis when comparing clinical dermatologists to teledermatologists. Moreover, when melanocytic lesions were excluded from the study, the decision to biopsy did not differ substantially.⁷

Areas of further study include determining what percentage of teledermatology lesions of concern for malignancy were proven to be skin cancer after in-person evaluation and biopsy, as well as investigating the effectiveness of teledermatology for melanocytic lesions, which frequently are removed from analysis in large-scale teledermatology studies.

Although teledermatology has substantial clinical utility and may serve as a great resource for specific populations, including geriatric patients and those who are immunocompromised, it is important to recognize notable limitations. Specifically, brief history and image review should not serve as replacements for a face-to-face visit with physical examination in cases where the diagnosis remains uncertain or when high-risk skin malignancies are suspected or included in the differential. Certain aggressive cutaneous malignancies such as Merkel cell carcinoma may appear as less aggressive via teledermatology due to restrictions of technology.

Conclusion

Teledermatology has had a major impact on the way health care is delivered to patients and may increase access to care, reducing unnecessary in-person visits and decreasing the number of in-person visit no-shows. With the appropriate use of a brief clinical history and image review, teledermatology can be effective to evaluate specific lesions of concern. We report 3 unique cases identified during a 1-month period at a large Midwestern medical center. These cases serve as important examples of the application of teledermatology in reducing the time to diagnosis of aggressive skin malignancies. Further research on the clinical utility of teledermatology is warranted.

Acknowledgments—The authors thank the additional providers from the University of Wisconsin and William S. Middleton Memorial Veterans Hospital (both in Madison, Wisconsin) involved in the medical care of the patients included in this case series.

With the increasing utilization of telemedicine since the COVID-19 pandemic, it is critical that clinicians have an appropriate understanding of the application of virtual care resources, including teledermatology. We present a case series of 3 patients to demonstrate the clinical utility of teledermatology in reducing the time to diagnosis of various rare and/or aggressive cutaneous malignancies, including Merkel cell carcinoma, malignant melanoma, and atypical fibroxanthoma. Cases were obtained from one large Midwestern medical center during the month of July 2021. Each case presented includes a description of the initial teledermatology presentation and reviews the clinical timeline from initial consultation submission to in-person clinic visit with lesion biopsy. This case series demonstrates real-world examples of how teledermatology can be utilized to expedite the care of specific vulnerable patient populations.

Teledermatology is a rapidly growing digital resource with specific utility in triaging patients to determine those requiring in-person evaluation for early and accurate detection of skin malignancies. Approximately one-third of teledermatology consultations result in face-to-face clinical encounters, with malignant neoplasms being the leading cause for biopsy.^1,2 For specific populations, such as geriatric and immunocompromised patients, teledermatology may serve as a valuable tool, particularly in the wake of the COVID-19 pandemic. Furthermore, telemedicine may aid in addressing health disparities within the field of medicine and ultimately may improve access to care for vulnerable populations.³ Along with increasing access to specific subspecialty expertise, the use of teledermatology may reduce health care costs and improve the overall quality of care delivered to patients.^4,5

We describe the clinical utility of teledermatology in triaging and diagnosing skin malignancies through a series of 3 cases obtained from digital image review at one large Midwestern medical center during the month of July 2021. Three unique cases with a final diagnosis of a rare or aggressive skin cancer were selected as examples, including a 75-year-old man with Merkle cell carcinoma, a 55-year-old man with aggressive pT3b malignant melanoma, and a 72-year-old man with an atypical fibroxanthoma. A clinical timeline of each case is presented, including the time intervals from initial image submission to image review, image submission to face-to-face clinical encounter, and image submission to final diagnosis. In all cases, the primary care provider submitted an order for teledermatology, and the teledermatology team obtained the images.

Case Series

Patient 1—Images of the right hand of a 75-year-old man with a medical history of basal cell carcinoma were submitted for teledermatology consultation utilizing store-and-forward image-capturing technology (day 1). The patient history provided with image submission indicated that the lesion had been present for 6 months and there were no associated symptoms. Clinical imaging demonstrated a pink-red pearly papule located on the proximal fourth digit of the dorsal aspect of the right hand (Figure 1). One day following the teledermatology request (day 2), the patient’s case was reviewed and triaged for an in-person visit. The patient was brought to clinic on day 34, and a biopsy was performed. On day 36, dermatopathology results indicated a diagnosis of Merkel cell carcinoma. On day 37, the patient was referred to surgical oncology, and on day 44, the patient underwent an initial surgical oncology visit with a plan for wide local excision of the right fourth digit with right axillary sentinel lymph node biopsy.

Patient 2—Images of the left flank of a 55-year-old man were submitted for teledermatology consultation via store-and-forward technology (day 1). A patient history provided with the image indicated that the lesion had been present for months to years and there were no associated symptoms, but the lesion recently had changed in color and size. Teledermatology images were reviewed on day 3 and demonstrated a 2- to 3-cm brown plaque on the left flank with color variegation and a prominent red papule protruding centrally (Figure 2). The patient was scheduled for an urgent in-person visit with biopsy. On day 6, the patient presented to clinic and an excision biopsy was performed. Dermatopathology was ordered with a RUSH indication, with results on day 7 revealing a pT3b malignant melanoma. An urgent consultation to surgical oncology was placed on the same day, and the patient underwent an initial surgical oncology visit on day 24 with a plan for wide local excision with left axillary and inguinal sentinel lymph node biopsy.

Patient 3—Images of the left ear of a 72-year-old man were submitted for teledermatology consultation utilizing review via store-and-forward technology (day 1). A patient history indicated that the lesion had been present for 3 months with associated bleeding. Image review demonstrated a solitary pearly pink papule located on the crura of the antihelix (Figure 3). Initial teledermatology consultation was reviewed on day 2 with notification of the need for in-person evaluation. The patient presented to clinic on day 33 for a biopsy, with dermatopathology results on day 36 consistent with an atypical fibroxanthoma. The patient was scheduled for Mohs micrographic surgery on day 37 and underwent surgical treatment on day 64.

Comment

Teledermatology consultations from all patients demonstrated adequate image quality to be able to evaluate the lesion of concern and yielded a request for in-person evaluation with possible biopsy (Table). In this case series, the average time interval from teledermatology consultation placement to teledermatology image report was 2 days (range, 1–3 days). The average time from teledermatology consultation placement to face-to-face encounter with biopsy was 24.3 days for the 3 cases presented in this series (range, 6–34 days). The initial surgical oncology visits took place an average of 34 days after the initial teledermatology consultation was placed for the 2 patients requiring referral (44 days for patient 1; 24 days for patient 2). For patient 3, Mohs micrographic surgery was required for treatment, which was scheduled by day 37 and subsequently performed on day 64.

When specifically looking at the diagnosis of cutaneous malignancies, studies have found that the incidence of skin cancer detection is similar for teledermatology compared to in-person clinic visits.^6,7 Creighton-Smith et al⁶ performed a retrospective cohort study comparing prebiopsy and postbiopsy diagnostic accuracy and detection rates of skin cancer between store-and-forward technology and face-to-face consultation. When adjusting for possible compounding factors including personal and family history of skin cancer, there was no notable difference in detection rates of any skin cancer, including melanoma and nonmelanoma skin cancers. Furthermore, the 2 cohorts of patients were found to have similar prebiopsy and postbiopsy diagnostic concordance, with similar times from consultation being placed to requested biopsy and time from biopsy to final treatment.⁶

Clarke et al⁷ similarly analyzed the accuracy of store-and-forward teledermatology and found that there was overall concordance in diagnosis when comparing clinical dermatologists to teledermatologists. Moreover, when melanocytic lesions were excluded from the study, the decision to biopsy did not differ substantially.⁷

Areas of further study include determining what percentage of teledermatology lesions of concern for malignancy were proven to be skin cancer after in-person evaluation and biopsy, as well as investigating the effectiveness of teledermatology for melanocytic lesions, which frequently are removed from analysis in large-scale teledermatology studies.

Although teledermatology has substantial clinical utility and may serve as a great resource for specific populations, including geriatric patients and those who are immunocompromised, it is important to recognize notable limitations. Specifically, brief history and image review should not serve as replacements for a face-to-face visit with physical examination in cases where the diagnosis remains uncertain or when high-risk skin malignancies are suspected or included in the differential. Certain aggressive cutaneous malignancies such as Merkel cell carcinoma may appear as less aggressive via teledermatology due to restrictions of technology.

Conclusion

Teledermatology has had a major impact on the way health care is delivered to patients and may increase access to care, reducing unnecessary in-person visits and decreasing the number of in-person visit no-shows. With the appropriate use of a brief clinical history and image review, teledermatology can be effective to evaluate specific lesions of concern. We report 3 unique cases identified during a 1-month period at a large Midwestern medical center. These cases serve as important examples of the application of teledermatology in reducing the time to diagnosis of aggressive skin malignancies. Further research on the clinical utility of teledermatology is warranted.

Acknowledgments—The authors thank the additional providers from the University of Wisconsin and William S. Middleton Memorial Veterans Hospital (both in Madison, Wisconsin) involved in the medical care of the patients included in this case series.

To the Editor:

A 42-year man presented with hollowing of the upper eyelid and skin discoloration of the left periorbital area of 10 years’ duration. He was a professional mixed martial arts fighter with a history of 2 surgeries for retinal detachment of the left eye 13 years prior to the current presentation. The patient also has macular scarring in the left eye. He denied a history of facial fracture, reconstructive surgery, or other medical conditions. His visual acuity was unknown; however, he did not require corrective glasses. He used 3 prescription ophthalmic eye drops—dorzolamide hydrochloride plus timolol maleate, 10 mL; brimonidine tartrate ophthalmic solution 0.15%, 5 mL; and latanoprost ophthalmic solution 0.005%, 125 μg/2.5 mL—in the left eye to lower intraocular pressure, as therapy for glaucoma. If left untreated, glaucoma can lead to vision loss.

Physical examination revealed periorbital hyperpigmentation on the left side; hypertrichosis and eyelash trichomegaly compared to the right side; and a deep left upper orbital sulcus compared to the right side (Figure). The patient was alert and oriented to person, place, and time. Extraocular movement was intact bilaterally, and his pupillary reflex was symmetric. No tenderness was noted over the affected area on palpation; no subcutaneous masses or lesions were observed or palpated. There was no ocular discharge, the conjunctiva was pink, and the sclera was white bilaterally.

The differential diagnosis included professional trauma-induced orbital changes, nevus of Ota (oculomucodermal melanocytosis), prostaglandin-associated periorbitopathy (PAP), and melasma. Although the patient sustained an injury that caused retinal detachment, he never experienced an orbital bone fracture; additionally, a fracture would not explain the skin discoloration or longer eyelashes. Periorbital nevus of Ota most commonly manifests as a unilateral scleral and brown-bluish skin discoloration but does not cause hollowing of the orbital sulcus or affect the length and thickness of eyelashes. Melasma—bilateral skin hyperpigmentation that most commonly affects women—can be induced by oral contraceptives, antibiotics, heat, sun exposure, and pregnancy. It does not affect the color of the iris or the depth of the scleral sulcus, and it does not increase the length and thickness of eyelashes. Based on the clinical presentation and a review of the eye drops used, he was diagnosed with PAP due to prolonged use of latanoprost ophthalmic solution. The patient was referred to an ophthalmologist for consideration of a switch to a different class of medication.

Of the 3 eye drops used by this patient, latanoprost, a prostaglandin analog, decreases intraocular pressure and is known to cause PAP. This condition comprises a constellation of changes, including upper eyelid ptosis, deepening of the upper eyelid sulcus, involution of dermatochalasis, periorbital fat atrophy, mild enophthalmos (sunken eye), inferior scleral show, increased prominence of eyelid vessels, and tight eyelids.¹ Latanoprost most often produces these findings, but all prostaglandin ophthalmic medications can, including the dual-indication bimatoprost, which was approved by the US Food and Drug Administration to reduce elevated intraocular pressure in patients with open-angle glaucoma or ocular hypertension but also is used to grow darker, thicker, and longer eyelashes. Clinicians who prescribe bimatoprost for this cosmetic indication should be mindful of the potential for PAP and discuss it with patients.

The prescribing information (PI) for bimatoprost (Latisse; Allergan) does not list PAP as an adverse reaction observed in the 4-month multicenter, double-blind, randomized, vehicle-controlled study of bimatoprost (as Latisse) in 278 adults.² The PI does list “periorbital and lid changes associated with periorbital fat atrophy and skin tightness resulting in deepening of eyelid sulcus and eyelid ptosis” as an adverse reaction in postmarketing experience. However, according to the PI, the frequency of these adverse reactions cannot be established, as the reporting of such incidents was voluntary and the size of the treated population was uncertain.²

Prostaglandins can cause periorbitopathy by several mechanisms; one speculated cause is that this group of medications might provoke smooth muscle contraction. Prostaglandin medications also have an affinity for fat cells¹; atrophy of fat cells can lead to enophthalmos and deepening upper eyelid sulcus. In an observational study of 105 participants who were using a prostaglandin in 1 eye for longer than 1 month (the other eye was used as a control), the overall frequency of prostaglandin-associated periorbitopathy was 93.3% in the bimatoprost group, 41.4% in the latanoprost group, and 70% in the travoprost group, while the frequency of deepening of the upper eyelid sulcus was 80% in the bimatoprost group, 15.7% in the latanoprost group, and 45% in the travoprost group.³ These changes may not be as striking when a patient is using a prostaglandin ophthalmic medication in both eyes and may not be noticed even by the patient. It is prudent for the clinician to take a baseline photograph of the patient when these medications are prescribed to observe for early signs of periorbitopathy. These adverse effects may not be reversible when the medication is discontinued⁴ and have been observed as early as 4 to 6 weeks after the start of treatment.⁵

Our patient was counseled that his constellation of PAP findings potentially could be partially reversed over months or even a year or longer if the offending agent was discontinued. However, he was cautioned that cessation of latanoprost first needed to be discussed with his ophthalmologist, who would determine if there was a suitable alternative to a prostaglandin analog for him. The patient’s only concern was the aesthetic appearance of the left periorbital area. A hyaluronic acid filler or fat grafting can be considered for correction of orbital sulcus hollowing; however, we could not locate any long-term studies in which such corrective treatments were applied for PAP. Our patient continues to use latanoprost with no change in the frequency of use. There have been no further changes or progression in the physical appearance of the left eye or periorbital area. The patient has not undergone any corrective treatments.

To the Editor:

A 42-year man presented with hollowing of the upper eyelid and skin discoloration of the left periorbital area of 10 years’ duration. He was a professional mixed martial arts fighter with a history of 2 surgeries for retinal detachment of the left eye 13 years prior to the current presentation. The patient also has macular scarring in the left eye. He denied a history of facial fracture, reconstructive surgery, or other medical conditions. His visual acuity was unknown; however, he did not require corrective glasses. He used 3 prescription ophthalmic eye drops—dorzolamide hydrochloride plus timolol maleate, 10 mL; brimonidine tartrate ophthalmic solution 0.15%, 5 mL; and latanoprost ophthalmic solution 0.005%, 125 μg/2.5 mL—in the left eye to lower intraocular pressure, as therapy for glaucoma. If left untreated, glaucoma can lead to vision loss.

Physical examination revealed periorbital hyperpigmentation on the left side; hypertrichosis and eyelash trichomegaly compared to the right side; and a deep left upper orbital sulcus compared to the right side (Figure). The patient was alert and oriented to person, place, and time. Extraocular movement was intact bilaterally, and his pupillary reflex was symmetric. No tenderness was noted over the affected area on palpation; no subcutaneous masses or lesions were observed or palpated. There was no ocular discharge, the conjunctiva was pink, and the sclera was white bilaterally.

The differential diagnosis included professional trauma-induced orbital changes, nevus of Ota (oculomucodermal melanocytosis), prostaglandin-associated periorbitopathy (PAP), and melasma. Although the patient sustained an injury that caused retinal detachment, he never experienced an orbital bone fracture; additionally, a fracture would not explain the skin discoloration or longer eyelashes. Periorbital nevus of Ota most commonly manifests as a unilateral scleral and brown-bluish skin discoloration but does not cause hollowing of the orbital sulcus or affect the length and thickness of eyelashes. Melasma—bilateral skin hyperpigmentation that most commonly affects women—can be induced by oral contraceptives, antibiotics, heat, sun exposure, and pregnancy. It does not affect the color of the iris or the depth of the scleral sulcus, and it does not increase the length and thickness of eyelashes. Based on the clinical presentation and a review of the eye drops used, he was diagnosed with PAP due to prolonged use of latanoprost ophthalmic solution. The patient was referred to an ophthalmologist for consideration of a switch to a different class of medication.

Of the 3 eye drops used by this patient, latanoprost, a prostaglandin analog, decreases intraocular pressure and is known to cause PAP. This condition comprises a constellation of changes, including upper eyelid ptosis, deepening of the upper eyelid sulcus, involution of dermatochalasis, periorbital fat atrophy, mild enophthalmos (sunken eye), inferior scleral show, increased prominence of eyelid vessels, and tight eyelids.¹ Latanoprost most often produces these findings, but all prostaglandin ophthalmic medications can, including the dual-indication bimatoprost, which was approved by the US Food and Drug Administration to reduce elevated intraocular pressure in patients with open-angle glaucoma or ocular hypertension but also is used to grow darker, thicker, and longer eyelashes. Clinicians who prescribe bimatoprost for this cosmetic indication should be mindful of the potential for PAP and discuss it with patients.

The prescribing information (PI) for bimatoprost (Latisse; Allergan) does not list PAP as an adverse reaction observed in the 4-month multicenter, double-blind, randomized, vehicle-controlled study of bimatoprost (as Latisse) in 278 adults.² The PI does list “periorbital and lid changes associated with periorbital fat atrophy and skin tightness resulting in deepening of eyelid sulcus and eyelid ptosis” as an adverse reaction in postmarketing experience. However, according to the PI, the frequency of these adverse reactions cannot be established, as the reporting of such incidents was voluntary and the size of the treated population was uncertain.²

Prostaglandins can cause periorbitopathy by several mechanisms; one speculated cause is that this group of medications might provoke smooth muscle contraction. Prostaglandin medications also have an affinity for fat cells¹; atrophy of fat cells can lead to enophthalmos and deepening upper eyelid sulcus. In an observational study of 105 participants who were using a prostaglandin in 1 eye for longer than 1 month (the other eye was used as a control), the overall frequency of prostaglandin-associated periorbitopathy was 93.3% in the bimatoprost group, 41.4% in the latanoprost group, and 70% in the travoprost group, while the frequency of deepening of the upper eyelid sulcus was 80% in the bimatoprost group, 15.7% in the latanoprost group, and 45% in the travoprost group.³ These changes may not be as striking when a patient is using a prostaglandin ophthalmic medication in both eyes and may not be noticed even by the patient. It is prudent for the clinician to take a baseline photograph of the patient when these medications are prescribed to observe for early signs of periorbitopathy. These adverse effects may not be reversible when the medication is discontinued⁴ and have been observed as early as 4 to 6 weeks after the start of treatment.⁵

Our patient was counseled that his constellation of PAP findings potentially could be partially reversed over months or even a year or longer if the offending agent was discontinued. However, he was cautioned that cessation of latanoprost first needed to be discussed with his ophthalmologist, who would determine if there was a suitable alternative to a prostaglandin analog for him. The patient’s only concern was the aesthetic appearance of the left periorbital area. A hyaluronic acid filler or fat grafting can be considered for correction of orbital sulcus hollowing; however, we could not locate any long-term studies in which such corrective treatments were applied for PAP. Our patient continues to use latanoprost with no change in the frequency of use. There have been no further changes or progression in the physical appearance of the left eye or periorbital area. The patient has not undergone any corrective treatments.

To the Editor:

A 42-year man presented with hollowing of the upper eyelid and skin discoloration of the left periorbital area of 10 years’ duration. He was a professional mixed martial arts fighter with a history of 2 surgeries for retinal detachment of the left eye 13 years prior to the current presentation. The patient also has macular scarring in the left eye. He denied a history of facial fracture, reconstructive surgery, or other medical conditions. His visual acuity was unknown; however, he did not require corrective glasses. He used 3 prescription ophthalmic eye drops—dorzolamide hydrochloride plus timolol maleate, 10 mL; brimonidine tartrate ophthalmic solution 0.15%, 5 mL; and latanoprost ophthalmic solution 0.005%, 125 μg/2.5 mL—in the left eye to lower intraocular pressure, as therapy for glaucoma. If left untreated, glaucoma can lead to vision loss.

Physical examination revealed periorbital hyperpigmentation on the left side; hypertrichosis and eyelash trichomegaly compared to the right side; and a deep left upper orbital sulcus compared to the right side (Figure). The patient was alert and oriented to person, place, and time. Extraocular movement was intact bilaterally, and his pupillary reflex was symmetric. No tenderness was noted over the affected area on palpation; no subcutaneous masses or lesions were observed or palpated. There was no ocular discharge, the conjunctiva was pink, and the sclera was white bilaterally.

The differential diagnosis included professional trauma-induced orbital changes, nevus of Ota (oculomucodermal melanocytosis), prostaglandin-associated periorbitopathy (PAP), and melasma. Although the patient sustained an injury that caused retinal detachment, he never experienced an orbital bone fracture; additionally, a fracture would not explain the skin discoloration or longer eyelashes. Periorbital nevus of Ota most commonly manifests as a unilateral scleral and brown-bluish skin discoloration but does not cause hollowing of the orbital sulcus or affect the length and thickness of eyelashes. Melasma—bilateral skin hyperpigmentation that most commonly affects women—can be induced by oral contraceptives, antibiotics, heat, sun exposure, and pregnancy. It does not affect the color of the iris or the depth of the scleral sulcus, and it does not increase the length and thickness of eyelashes. Based on the clinical presentation and a review of the eye drops used, he was diagnosed with PAP due to prolonged use of latanoprost ophthalmic solution. The patient was referred to an ophthalmologist for consideration of a switch to a different class of medication.

Of the 3 eye drops used by this patient, latanoprost, a prostaglandin analog, decreases intraocular pressure and is known to cause PAP. This condition comprises a constellation of changes, including upper eyelid ptosis, deepening of the upper eyelid sulcus, involution of dermatochalasis, periorbital fat atrophy, mild enophthalmos (sunken eye), inferior scleral show, increased prominence of eyelid vessels, and tight eyelids.¹ Latanoprost most often produces these findings, but all prostaglandin ophthalmic medications can, including the dual-indication bimatoprost, which was approved by the US Food and Drug Administration to reduce elevated intraocular pressure in patients with open-angle glaucoma or ocular hypertension but also is used to grow darker, thicker, and longer eyelashes. Clinicians who prescribe bimatoprost for this cosmetic indication should be mindful of the potential for PAP and discuss it with patients.

The prescribing information (PI) for bimatoprost (Latisse; Allergan) does not list PAP as an adverse reaction observed in the 4-month multicenter, double-blind, randomized, vehicle-controlled study of bimatoprost (as Latisse) in 278 adults.² The PI does list “periorbital and lid changes associated with periorbital fat atrophy and skin tightness resulting in deepening of eyelid sulcus and eyelid ptosis” as an adverse reaction in postmarketing experience. However, according to the PI, the frequency of these adverse reactions cannot be established, as the reporting of such incidents was voluntary and the size of the treated population was uncertain.²

Prostaglandins can cause periorbitopathy by several mechanisms; one speculated cause is that this group of medications might provoke smooth muscle contraction. Prostaglandin medications also have an affinity for fat cells¹; atrophy of fat cells can lead to enophthalmos and deepening upper eyelid sulcus. In an observational study of 105 participants who were using a prostaglandin in 1 eye for longer than 1 month (the other eye was used as a control), the overall frequency of prostaglandin-associated periorbitopathy was 93.3% in the bimatoprost group, 41.4% in the latanoprost group, and 70% in the travoprost group, while the frequency of deepening of the upper eyelid sulcus was 80% in the bimatoprost group, 15.7% in the latanoprost group, and 45% in the travoprost group.³ These changes may not be as striking when a patient is using a prostaglandin ophthalmic medication in both eyes and may not be noticed even by the patient. It is prudent for the clinician to take a baseline photograph of the patient when these medications are prescribed to observe for early signs of periorbitopathy. These adverse effects may not be reversible when the medication is discontinued⁴ and have been observed as early as 4 to 6 weeks after the start of treatment.⁵

Our patient was counseled that his constellation of PAP findings potentially could be partially reversed over months or even a year or longer if the offending agent was discontinued. However, he was cautioned that cessation of latanoprost first needed to be discussed with his ophthalmologist, who would determine if there was a suitable alternative to a prostaglandin analog for him. The patient’s only concern was the aesthetic appearance of the left periorbital area. A hyaluronic acid filler or fat grafting can be considered for correction of orbital sulcus hollowing; however, we could not locate any long-term studies in which such corrective treatments were applied for PAP. Our patient continues to use latanoprost with no change in the frequency of use. There have been no further changes or progression in the physical appearance of the left eye or periorbital area. The patient has not undergone any corrective treatments.

Cutaneous reactions associated with the Pfizer-BioNTech COVID-19 vaccine have been reported worldwide since December 2020. Local injection site reactions (<1%) such as erythema, swelling, delayed local reactions (1%–10%), morbilliform rash, urticarial reactions, pityriasis rosea, Rowell syndrome, and lichen planus have been reported following the Pfizer-BioNTech COVID-19 vaccine.¹ Cutaneous reactions reported in association with the Sinovac-Coronavac COVID-19 vaccine include swelling, redness, itching, discoloration, induration (1%–10%), urticaria, petechial rash, and exacerbation of psoriasis at the local injection site (<1%).²

We describe 7 patients from Turkey who presented with various dermatologic problems 5 to 28 days after COVID-19 vaccination, highlighting the possibility of early and late cutaneous reactions related to the vaccine (Table).

Case Reports

Patient 1—A 44-year-old woman was admitted to the dermatology clinic with painful lesions on the trunk of 3 days’ duration. Dermatologic examination revealed grouped erythematous vesicles showing dermatomal spread in the right thoracolumbar (dermatome T10) region. The patient reported that she had received 2 doses of the Sinovac-Coronavac vaccine (doses 1 and 2) and 2 doses of the BioNTech COVID-19 vaccine (doses 3 and 4); the rash had developed 28 days after she received the 4th dose. Her medical history was unremarkable. The lesions regressed after 1 week of treatment with oral valacyclovir 1000 mg 3 times daily, but she developed postherpetic neuralgia 1 week after starting treatment, which resolved after 8 weeks.

Patient 2—A 68-year-old woman presented to the dermatology clinic for evaluation of painful sores on the upper lip of 1 day’s duration. She had a history of rheumatoid arthritis, hypertension, and atopy and was currently taking prednisone and etanercept. Dermatologic examination revealed grouped vesicles on an erythematous base on the upper lip. A diagnosis of herpes labialis was made. The patient reported that she had received a third dose of the Sinovac-Coronavac vaccine 10 days prior to the appearance of the lesions. Her symptoms resolved completely within 2 weeks of treatment with topical acyclovir.

Patient 3—A 64-year-old woman was admitted to the hospital with pain, redness, and watery sores on and around the left eyelid of 2 days’ duration. Dermatologic evaluation revealed the erythematous surface of the left eyelid and periorbital area showed partial crusts, clustered vesicles, erythema, and edema. Additionally, the conjunctiva was purulent and erythematous. The patient’s medical history was notable for allergic asthma, hypertension, anxiety, and depression. For this reason, the patient was prescribed an angiotensin receptor blocker and a selective serotonin reuptake inhibitor. She noted that a similar rash had developed around the left eye 6 years prior that was diagnosed as herpes zoster (HZ). She also reported that she had received 2 doses of the Sinovac-Coronavac COVID-19 vaccine followed by 1 dose of the BioNTech COVID-19 vaccine, which she had received 2 weeks before the rash developed. The patient was treated at the eye clinic and was found to have ocular involvement. Ophthalmology was consulted and a diagnosis of herpes zoster ophthalmicus (HZO) was made. Systemic valacyclovir treatment was initiated, resulting in clinical improvement within 3 weeks.

Patient 4—A 75-year-old man was admitted to the hospital with chest and back pain and widespread muscle pain of several days’ duration. His medical history was remarkable for diabetes mellitus, hypertension, depression, and coronary artery bypass surgery. A medication history revealed treatment with a β-blocker, acetylsalicylic acid, a calcium channel blocker, a dipeptidyl peptidase 4 inhibitor, and a selective serotonin reuptake inhibitor. Dermatologic examination revealed grouped vesicles on an erythematous background in dermatome T5 on the right chest and back. A diagnosis of HZ was made. The patient reported that he had received 2 doses of the Sinovac-Coronavac vaccine followed by 1 dose of the Pfizer-BioNTech vaccine 2 weeks prior to the current presentation. He was treated with valacyclovir for 1 week, and his symptoms resolved entirely within 3 weeks.

Patient 5—A 50-year-old woman presented to the hospital for evaluation of painful sores on the back, chest, groin, and abdomen of 10 days’ duration. The lesions initially had developed 7 days after receiving the BioNTech COVID-19 vaccine; she previously had received 2 doses of the Sinovac-Coronavac vaccine. The patient had a history of untreated psoriasis. Dermatologic examination revealed grouped vesicles on an erythematous background in the T2–L2 dermatomes on the left side of the trunk. A diagnosis of HZ was made. The lesions resolved after 1 week of treatment with systemic valacyclovir; however, she subsequently developed postherpetic neuralgia, hypoesthesia, and postinflammatory hyperpigmentation in the affected regions.

Patient 6—A 37-year-old woman presented to the hospital with redness, swelling, and itching all over the body of 3 days’ duration. The patient noted that the rash would subside and reappear throughout the day. Her medical history was unremarkable, except for COVID-19 infection 6 months prior. She had received a second dose of the BioNTech vaccine 20 days prior to development of symptoms. Dermatologic examination revealed widespread erythematous urticarial plaques. A diagnosis of acute urticaria was made. The patient recovered completely after 1 week of treatment with a systemic steroid and 3 weeks of antihistamine treatment.

Patient 7—A 63-year-old woman presented to the hospital with widespread itching and rash that appeared 5 days after the first dose of the BioNTech COVID-19 vaccine. The patient reported that the rash resolved spontaneously within a few hours but then reappeared. Her medical history revealed that she was taking tamoxifen for breast cancer and that she previously had received 2 doses of the Sinovac-Coronavac vaccine. Dermatologic examination revealed erythematous urticarial plaques on the trunk and arms. A diagnosis of urticaria was made, and her symptoms resolved after 6 weeks of antihistamine treatment.

Comment

Skin lesions associated with COVID-19 infection have been reported worldwide^3,4 as well as dermatologic reactions following COVID-19 vaccination. In one case from Turkey, HZ infection was reported in a 68-year-old man 5 days after he received a second dose of the COVID-19 vaccine.⁵ In another case, HZ infection developed in a 78-year-old man 5 days after COVID-19 vaccination.⁶ Numerous cases of HZ infection developing within 1 to 26 days of COVID-19 vaccination have been reported worldwide.^7-9

In a study conducted in the United States, 40 skin reactions associated with the COVID-19 vaccine were investigated; of these cases, 87.5% (35/40) were reported as varicella-zoster virus, and 12.5% (5/40) were reported as herpes simplex reactivation; 54% (19/35) and 80% (4/5) of these cases, respectively, were associated with the Pfizer-BioNTech vaccine.¹⁰ The average age of patients who developed a skin reaction was 46 years, and 70% (28/40) were women. The time to onset of the reaction was 2 to 13 days after vaccination, and symptoms were reported to improve within 7 days on average.¹⁰

Another study from Spain examined 405 vaccine-related skin reactions, 40.2% of which were related to the Pfizer-BioNTech vaccine. Among them, 80.2% occurred in women; 13.8% of cases were diagnosed as varicella-zoster virus or HZ virus reactivation, and 14.6% were urticaria. Eighty reactions (21%) were classified as severe/very severe and 81% required treatment.¹¹ One study reported 414 skin reactions from the COVID-19 vaccine from December 2020 to February 2021; of these cases, 83% occurred after the Moderna vaccine, which is not available in Turkey, and 17% occurred after the Pfizer-BioNTech vaccine.¹²A systematic review of 91 patients who developed HZ infection after COVID-19 vaccination reported that 10% (9/91) of cases were receiving immunosuppressive therapy and 13% (12/91) had an autoimmune disease.⁷ In our case series, it is known that at least 2 of the patients (patients 2 and 5), including 1 patient with rheumatoid arthritis (patient 2) who was on immunosuppressive treatment, had autoimmune disorders. However, reports in the literature indicate that most patients with autoimmune inflammatory rheumatic diseases remain stable after vaccination.¹³

Herpes zoster ophthalmicus is a rare form of HZ caused by involvement of the ophthalmic branch of the trigeminal nerve that manifests as vesicular lesions and retinitis, uveitis, keratitis, conjunctivitis, and pain on an erythematous background. Two cases of women who developed HZO infection after Pfizer-BioNTech vaccination were reported in the literature.¹⁴ Although patient 3 in our case series had a history of HZO 6 years prior, the possibility of the Pfizer-BioNTech vaccine triggering HZO should be taken into consideration.

Although cutaneous reactions after the Sinovac-Coronavac vaccine were observed in only 1 of 7 patients in our case series, skin reactions after Sinovac-Coronavac (an inactivated viral vaccine) have been reported in the literature. In one study, after a total of 35,229 injections, the incidence of cutaneous adverse events due to Sinovac-Coronavac was reported to be 0.94% and 0.70% after the first and second doses, respectively.¹⁵ Therefore, further study results are needed to directly attribute the reactions to COVID-19 vaccination.

Conclusion

Our case series highlights that clinicians should be vigilant in diagnosing cutaneous reactions following COVID-19 vaccination early to prevent potential complications. Early recognition of reactions is crucial, and the prognosis can be improved with appropriate treatment. Despite the potential dermatologic adverse effects of the COVID-19 vaccine, the most effective way to protect against serious COVID-19 infection is to continue to be vaccinated.

Cutaneous reactions associated with the Pfizer-BioNTech COVID-19 vaccine have been reported worldwide since December 2020. Local injection site reactions (<1%) such as erythema, swelling, delayed local reactions (1%–10%), morbilliform rash, urticarial reactions, pityriasis rosea, Rowell syndrome, and lichen planus have been reported following the Pfizer-BioNTech COVID-19 vaccine.¹ Cutaneous reactions reported in association with the Sinovac-Coronavac COVID-19 vaccine include swelling, redness, itching, discoloration, induration (1%–10%), urticaria, petechial rash, and exacerbation of psoriasis at the local injection site (<1%).²

We describe 7 patients from Turkey who presented with various dermatologic problems 5 to 28 days after COVID-19 vaccination, highlighting the possibility of early and late cutaneous reactions related to the vaccine (Table).

Case Reports

Patient 1—A 44-year-old woman was admitted to the dermatology clinic with painful lesions on the trunk of 3 days’ duration. Dermatologic examination revealed grouped erythematous vesicles showing dermatomal spread in the right thoracolumbar (dermatome T10) region. The patient reported that she had received 2 doses of the Sinovac-Coronavac vaccine (doses 1 and 2) and 2 doses of the BioNTech COVID-19 vaccine (doses 3 and 4); the rash had developed 28 days after she received the 4th dose. Her medical history was unremarkable. The lesions regressed after 1 week of treatment with oral valacyclovir 1000 mg 3 times daily, but she developed postherpetic neuralgia 1 week after starting treatment, which resolved after 8 weeks.

Patient 2—A 68-year-old woman presented to the dermatology clinic for evaluation of painful sores on the upper lip of 1 day’s duration. She had a history of rheumatoid arthritis, hypertension, and atopy and was currently taking prednisone and etanercept. Dermatologic examination revealed grouped vesicles on an erythematous base on the upper lip. A diagnosis of herpes labialis was made. The patient reported that she had received a third dose of the Sinovac-Coronavac vaccine 10 days prior to the appearance of the lesions. Her symptoms resolved completely within 2 weeks of treatment with topical acyclovir.

Patient 3—A 64-year-old woman was admitted to the hospital with pain, redness, and watery sores on and around the left eyelid of 2 days’ duration. Dermatologic evaluation revealed the erythematous surface of the left eyelid and periorbital area showed partial crusts, clustered vesicles, erythema, and edema. Additionally, the conjunctiva was purulent and erythematous. The patient’s medical history was notable for allergic asthma, hypertension, anxiety, and depression. For this reason, the patient was prescribed an angiotensin receptor blocker and a selective serotonin reuptake inhibitor. She noted that a similar rash had developed around the left eye 6 years prior that was diagnosed as herpes zoster (HZ). She also reported that she had received 2 doses of the Sinovac-Coronavac COVID-19 vaccine followed by 1 dose of the BioNTech COVID-19 vaccine, which she had received 2 weeks before the rash developed. The patient was treated at the eye clinic and was found to have ocular involvement. Ophthalmology was consulted and a diagnosis of herpes zoster ophthalmicus (HZO) was made. Systemic valacyclovir treatment was initiated, resulting in clinical improvement within 3 weeks.

Patient 4—A 75-year-old man was admitted to the hospital with chest and back pain and widespread muscle pain of several days’ duration. His medical history was remarkable for diabetes mellitus, hypertension, depression, and coronary artery bypass surgery. A medication history revealed treatment with a β-blocker, acetylsalicylic acid, a calcium channel blocker, a dipeptidyl peptidase 4 inhibitor, and a selective serotonin reuptake inhibitor. Dermatologic examination revealed grouped vesicles on an erythematous background in dermatome T5 on the right chest and back. A diagnosis of HZ was made. The patient reported that he had received 2 doses of the Sinovac-Coronavac vaccine followed by 1 dose of the Pfizer-BioNTech vaccine 2 weeks prior to the current presentation. He was treated with valacyclovir for 1 week, and his symptoms resolved entirely within 3 weeks.

Patient 5—A 50-year-old woman presented to the hospital for evaluation of painful sores on the back, chest, groin, and abdomen of 10 days’ duration. The lesions initially had developed 7 days after receiving the BioNTech COVID-19 vaccine; she previously had received 2 doses of the Sinovac-Coronavac vaccine. The patient had a history of untreated psoriasis. Dermatologic examination revealed grouped vesicles on an erythematous background in the T2–L2 dermatomes on the left side of the trunk. A diagnosis of HZ was made. The lesions resolved after 1 week of treatment with systemic valacyclovir; however, she subsequently developed postherpetic neuralgia, hypoesthesia, and postinflammatory hyperpigmentation in the affected regions.

Patient 6—A 37-year-old woman presented to the hospital with redness, swelling, and itching all over the body of 3 days’ duration. The patient noted that the rash would subside and reappear throughout the day. Her medical history was unremarkable, except for COVID-19 infection 6 months prior. She had received a second dose of the BioNTech vaccine 20 days prior to development of symptoms. Dermatologic examination revealed widespread erythematous urticarial plaques. A diagnosis of acute urticaria was made. The patient recovered completely after 1 week of treatment with a systemic steroid and 3 weeks of antihistamine treatment.

Patient 7—A 63-year-old woman presented to the hospital with widespread itching and rash that appeared 5 days after the first dose of the BioNTech COVID-19 vaccine. The patient reported that the rash resolved spontaneously within a few hours but then reappeared. Her medical history revealed that she was taking tamoxifen for breast cancer and that she previously had received 2 doses of the Sinovac-Coronavac vaccine. Dermatologic examination revealed erythematous urticarial plaques on the trunk and arms. A diagnosis of urticaria was made, and her symptoms resolved after 6 weeks of antihistamine treatment.

Comment

Skin lesions associated with COVID-19 infection have been reported worldwide^3,4 as well as dermatologic reactions following COVID-19 vaccination. In one case from Turkey, HZ infection was reported in a 68-year-old man 5 days after he received a second dose of the COVID-19 vaccine.⁵ In another case, HZ infection developed in a 78-year-old man 5 days after COVID-19 vaccination.⁶ Numerous cases of HZ infection developing within 1 to 26 days of COVID-19 vaccination have been reported worldwide.^7-9

In a study conducted in the United States, 40 skin reactions associated with the COVID-19 vaccine were investigated; of these cases, 87.5% (35/40) were reported as varicella-zoster virus, and 12.5% (5/40) were reported as herpes simplex reactivation; 54% (19/35) and 80% (4/5) of these cases, respectively, were associated with the Pfizer-BioNTech vaccine.¹⁰ The average age of patients who developed a skin reaction was 46 years, and 70% (28/40) were women. The time to onset of the reaction was 2 to 13 days after vaccination, and symptoms were reported to improve within 7 days on average.¹⁰

Another study from Spain examined 405 vaccine-related skin reactions, 40.2% of which were related to the Pfizer-BioNTech vaccine. Among them, 80.2% occurred in women; 13.8% of cases were diagnosed as varicella-zoster virus or HZ virus reactivation, and 14.6% were urticaria. Eighty reactions (21%) were classified as severe/very severe and 81% required treatment.¹¹ One study reported 414 skin reactions from the COVID-19 vaccine from December 2020 to February 2021; of these cases, 83% occurred after the Moderna vaccine, which is not available in Turkey, and 17% occurred after the Pfizer-BioNTech vaccine.¹²A systematic review of 91 patients who developed HZ infection after COVID-19 vaccination reported that 10% (9/91) of cases were receiving immunosuppressive therapy and 13% (12/91) had an autoimmune disease.⁷ In our case series, it is known that at least 2 of the patients (patients 2 and 5), including 1 patient with rheumatoid arthritis (patient 2) who was on immunosuppressive treatment, had autoimmune disorders. However, reports in the literature indicate that most patients with autoimmune inflammatory rheumatic diseases remain stable after vaccination.¹³

Herpes zoster ophthalmicus is a rare form of HZ caused by involvement of the ophthalmic branch of the trigeminal nerve that manifests as vesicular lesions and retinitis, uveitis, keratitis, conjunctivitis, and pain on an erythematous background. Two cases of women who developed HZO infection after Pfizer-BioNTech vaccination were reported in the literature.¹⁴ Although patient 3 in our case series had a history of HZO 6 years prior, the possibility of the Pfizer-BioNTech vaccine triggering HZO should be taken into consideration.

Although cutaneous reactions after the Sinovac-Coronavac vaccine were observed in only 1 of 7 patients in our case series, skin reactions after Sinovac-Coronavac (an inactivated viral vaccine) have been reported in the literature. In one study, after a total of 35,229 injections, the incidence of cutaneous adverse events due to Sinovac-Coronavac was reported to be 0.94% and 0.70% after the first and second doses, respectively.¹⁵ Therefore, further study results are needed to directly attribute the reactions to COVID-19 vaccination.

Conclusion

Our case series highlights that clinicians should be vigilant in diagnosing cutaneous reactions following COVID-19 vaccination early to prevent potential complications. Early recognition of reactions is crucial, and the prognosis can be improved with appropriate treatment. Despite the potential dermatologic adverse effects of the COVID-19 vaccine, the most effective way to protect against serious COVID-19 infection is to continue to be vaccinated.

Cutaneous reactions associated with the Pfizer-BioNTech COVID-19 vaccine have been reported worldwide since December 2020. Local injection site reactions (<1%) such as erythema, swelling, delayed local reactions (1%–10%), morbilliform rash, urticarial reactions, pityriasis rosea, Rowell syndrome, and lichen planus have been reported following the Pfizer-BioNTech COVID-19 vaccine.¹ Cutaneous reactions reported in association with the Sinovac-Coronavac COVID-19 vaccine include swelling, redness, itching, discoloration, induration (1%–10%), urticaria, petechial rash, and exacerbation of psoriasis at the local injection site (<1%).²

We describe 7 patients from Turkey who presented with various dermatologic problems 5 to 28 days after COVID-19 vaccination, highlighting the possibility of early and late cutaneous reactions related to the vaccine (Table).

Case Reports

Patient 1—A 44-year-old woman was admitted to the dermatology clinic with painful lesions on the trunk of 3 days’ duration. Dermatologic examination revealed grouped erythematous vesicles showing dermatomal spread in the right thoracolumbar (dermatome T10) region. The patient reported that she had received 2 doses of the Sinovac-Coronavac vaccine (doses 1 and 2) and 2 doses of the BioNTech COVID-19 vaccine (doses 3 and 4); the rash had developed 28 days after she received the 4th dose. Her medical history was unremarkable. The lesions regressed after 1 week of treatment with oral valacyclovir 1000 mg 3 times daily, but she developed postherpetic neuralgia 1 week after starting treatment, which resolved after 8 weeks.

Patient 2—A 68-year-old woman presented to the dermatology clinic for evaluation of painful sores on the upper lip of 1 day’s duration. She had a history of rheumatoid arthritis, hypertension, and atopy and was currently taking prednisone and etanercept. Dermatologic examination revealed grouped vesicles on an erythematous base on the upper lip. A diagnosis of herpes labialis was made. The patient reported that she had received a third dose of the Sinovac-Coronavac vaccine 10 days prior to the appearance of the lesions. Her symptoms resolved completely within 2 weeks of treatment with topical acyclovir.

Patient 3—A 64-year-old woman was admitted to the hospital with pain, redness, and watery sores on and around the left eyelid of 2 days’ duration. Dermatologic evaluation revealed the erythematous surface of the left eyelid and periorbital area showed partial crusts, clustered vesicles, erythema, and edema. Additionally, the conjunctiva was purulent and erythematous. The patient’s medical history was notable for allergic asthma, hypertension, anxiety, and depression. For this reason, the patient was prescribed an angiotensin receptor blocker and a selective serotonin reuptake inhibitor. She noted that a similar rash had developed around the left eye 6 years prior that was diagnosed as herpes zoster (HZ). She also reported that she had received 2 doses of the Sinovac-Coronavac COVID-19 vaccine followed by 1 dose of the BioNTech COVID-19 vaccine, which she had received 2 weeks before the rash developed. The patient was treated at the eye clinic and was found to have ocular involvement. Ophthalmology was consulted and a diagnosis of herpes zoster ophthalmicus (HZO) was made. Systemic valacyclovir treatment was initiated, resulting in clinical improvement within 3 weeks.

Patient 4—A 75-year-old man was admitted to the hospital with chest and back pain and widespread muscle pain of several days’ duration. His medical history was remarkable for diabetes mellitus, hypertension, depression, and coronary artery bypass surgery. A medication history revealed treatment with a β-blocker, acetylsalicylic acid, a calcium channel blocker, a dipeptidyl peptidase 4 inhibitor, and a selective serotonin reuptake inhibitor. Dermatologic examination revealed grouped vesicles on an erythematous background in dermatome T5 on the right chest and back. A diagnosis of HZ was made. The patient reported that he had received 2 doses of the Sinovac-Coronavac vaccine followed by 1 dose of the Pfizer-BioNTech vaccine 2 weeks prior to the current presentation. He was treated with valacyclovir for 1 week, and his symptoms resolved entirely within 3 weeks.

Patient 5—A 50-year-old woman presented to the hospital for evaluation of painful sores on the back, chest, groin, and abdomen of 10 days’ duration. The lesions initially had developed 7 days after receiving the BioNTech COVID-19 vaccine; she previously had received 2 doses of the Sinovac-Coronavac vaccine. The patient had a history of untreated psoriasis. Dermatologic examination revealed grouped vesicles on an erythematous background in the T2–L2 dermatomes on the left side of the trunk. A diagnosis of HZ was made. The lesions resolved after 1 week of treatment with systemic valacyclovir; however, she subsequently developed postherpetic neuralgia, hypoesthesia, and postinflammatory hyperpigmentation in the affected regions.

Patient 6—A 37-year-old woman presented to the hospital with redness, swelling, and itching all over the body of 3 days’ duration. The patient noted that the rash would subside and reappear throughout the day. Her medical history was unremarkable, except for COVID-19 infection 6 months prior. She had received a second dose of the BioNTech vaccine 20 days prior to development of symptoms. Dermatologic examination revealed widespread erythematous urticarial plaques. A diagnosis of acute urticaria was made. The patient recovered completely after 1 week of treatment with a systemic steroid and 3 weeks of antihistamine treatment.

Patient 7—A 63-year-old woman presented to the hospital with widespread itching and rash that appeared 5 days after the first dose of the BioNTech COVID-19 vaccine. The patient reported that the rash resolved spontaneously within a few hours but then reappeared. Her medical history revealed that she was taking tamoxifen for breast cancer and that she previously had received 2 doses of the Sinovac-Coronavac vaccine. Dermatologic examination revealed erythematous urticarial plaques on the trunk and arms. A diagnosis of urticaria was made, and her symptoms resolved after 6 weeks of antihistamine treatment.

Comment

Skin lesions associated with COVID-19 infection have been reported worldwide^3,4 as well as dermatologic reactions following COVID-19 vaccination. In one case from Turkey, HZ infection was reported in a 68-year-old man 5 days after he received a second dose of the COVID-19 vaccine.⁵ In another case, HZ infection developed in a 78-year-old man 5 days after COVID-19 vaccination.⁶ Numerous cases of HZ infection developing within 1 to 26 days of COVID-19 vaccination have been reported worldwide.^7-9

In a study conducted in the United States, 40 skin reactions associated with the COVID-19 vaccine were investigated; of these cases, 87.5% (35/40) were reported as varicella-zoster virus, and 12.5% (5/40) were reported as herpes simplex reactivation; 54% (19/35) and 80% (4/5) of these cases, respectively, were associated with the Pfizer-BioNTech vaccine.¹⁰ The average age of patients who developed a skin reaction was 46 years, and 70% (28/40) were women. The time to onset of the reaction was 2 to 13 days after vaccination, and symptoms were reported to improve within 7 days on average.¹⁰

Another study from Spain examined 405 vaccine-related skin reactions, 40.2% of which were related to the Pfizer-BioNTech vaccine. Among them, 80.2% occurred in women; 13.8% of cases were diagnosed as varicella-zoster virus or HZ virus reactivation, and 14.6% were urticaria. Eighty reactions (21%) were classified as severe/very severe and 81% required treatment.¹¹ One study reported 414 skin reactions from the COVID-19 vaccine from December 2020 to February 2021; of these cases, 83% occurred after the Moderna vaccine, which is not available in Turkey, and 17% occurred after the Pfizer-BioNTech vaccine.¹²A systematic review of 91 patients who developed HZ infection after COVID-19 vaccination reported that 10% (9/91) of cases were receiving immunosuppressive therapy and 13% (12/91) had an autoimmune disease.⁷ In our case series, it is known that at least 2 of the patients (patients 2 and 5), including 1 patient with rheumatoid arthritis (patient 2) who was on immunosuppressive treatment, had autoimmune disorders. However, reports in the literature indicate that most patients with autoimmune inflammatory rheumatic diseases remain stable after vaccination.¹³

Herpes zoster ophthalmicus is a rare form of HZ caused by involvement of the ophthalmic branch of the trigeminal nerve that manifests as vesicular lesions and retinitis, uveitis, keratitis, conjunctivitis, and pain on an erythematous background. Two cases of women who developed HZO infection after Pfizer-BioNTech vaccination were reported in the literature.¹⁴ Although patient 3 in our case series had a history of HZO 6 years prior, the possibility of the Pfizer-BioNTech vaccine triggering HZO should be taken into consideration.

Although cutaneous reactions after the Sinovac-Coronavac vaccine were observed in only 1 of 7 patients in our case series, skin reactions after Sinovac-Coronavac (an inactivated viral vaccine) have been reported in the literature. In one study, after a total of 35,229 injections, the incidence of cutaneous adverse events due to Sinovac-Coronavac was reported to be 0.94% and 0.70% after the first and second doses, respectively.¹⁵ Therefore, further study results are needed to directly attribute the reactions to COVID-19 vaccination.

Conclusion

Our case series highlights that clinicians should be vigilant in diagnosing cutaneous reactions following COVID-19 vaccination early to prevent potential complications. Early recognition of reactions is crucial, and the prognosis can be improved with appropriate treatment. Despite the potential dermatologic adverse effects of the COVID-19 vaccine, the most effective way to protect against serious COVID-19 infection is to continue to be vaccinated.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

Clinicians usually measure renal function by using surrogate markers because directly measuring glomerular filtration rate (GFR) is not routinely feasible in a clinical setting.^1,2 Creatinine (Cr) and cystatin C (CysC) are the 2 main surrogate molecules used to estimate GFR.³

Creatine is a molecule nonenzymatically converted into Cr, weighing only 113 Da in skeletal muscles.⁴ It is then filtered at the glomeruli and secreted at the proximal tubules of the kidneys. However, serum Cr (sCr) levels are affected by several factors, including age, biological sex, liver function, diet, and muscle mass.⁵ Historically, sCr levels also are affected by race.⁵ In an early study of factors affecting accurate GFR, researchers reported that self-identified African American patients had a 16% higher GFR than those who did not when using Cr.⁶ Despite this, the inclusion of Cr on a basic metabolic panel has allowed automatic reporting of an estimated GFR using sCr (eGFR_Cr) to be readily available.⁷

In comparison to Cr, CysC is an endogenous protein weighing 13 kDa produced by all nucleated cells.^8,9 CysC is filtered by the kidney at the glomeruli and completely reabsorbed and catabolized by epithelial cells at the proximal tubule.⁹ Since production is not dependent on skeletal muscle, there are fewer physiological impacts on serum concentration of CysC. Levels of CysC may be elevated by factors shown in the Table.

Estimating Glomerular Filtration Rates

Multiple equations were developed to mitigate the impact of extraneous factors on the accuracy of an eGFR_Cr. The first widely used equation that included a variable adjustment for race was the Modification of Diet in Renal Disease study, presented in 2006.¹⁰ The equation increased the accuracy of eGFR_Cr further by adjusting for sex and age. It was followed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation in 2009, which was more accurate at higher GFR levels.¹¹

CysC was simultaneously studied as an alternative to Cr with multiple equation iterations shown to be viable in various populations as early as 2003.^12-15 However, it was not until 2012 that an equation for the use of CysC was offered for widespread use as an alternative to Cr alongside further refinement of the CKD-EPI equation for Cr.¹⁶ A new formula was presented in 2021 to use both sCr and serum CysC levels to obtain a more accurate estimation of GFR.¹⁷ Research continues its effort to accurately estimate GFR for diagnosing kidney disease and assessing comorbidities relating to decreased kidney function.³

All historical equations attempted to mitigate the potential impact of race on sCr level when calculating eGFR_Crby assigning a separate variable for African American patients. As an unintended adverse effect, these equations may have led to discrimination by having a different equation for African American patients.¹⁸ Moreover, these Cr-based equations remain less accurate in patients with varied muscle mass, such as older patients, bodybuilders, athletes, and individuals with varied extremes of daily protein intake.^1,8,9,19Several medications can also directly affect Cr clearance, reducing its ability to act as a surrogate for kidney function.¹In this case report, we discuss an African American patient with high muscle mass and protein intake who was initially diagnosed with kidney disease based on an elevated Cr and found to be misdiagnosed based on the use of CysC for a more accurate GFR estimation.

A 35-year-old African American man serving in the military and recently diagnosed with HIV was referred to a nephrology clinic for further evaluation of an acute elevation in sCr. Before treatment for HIV, a brief record review showed a baseline Cr of about 1.3 mg/dL, with an eGFR_Cr of 75 mL/min/1.73 m².²⁰ In the same month, the patient was prescribed bictegravir/emtricitabine/tenofovir alafenamide, an HIV drug with nephrotoxic potential.²¹ The patient's total viral load remained low, and CD4 count remained > 500 after initiation of the HIV treatment. He was in his normal state of health and had no known contributory history before his HIV diagnosis. Cr readings peaked at 1.83 mg/dL after starting the HIV treatment and remained elevated to 1.73 mg/dL over the next few months, corresponding to CKD stage 3A. Because bictegravir/emtricitabine/tenofovir alafenamide is cleared by the kidneys and has a nephrotoxic profile, the clinical care team considered dosage adjustment or a medication switch given his observed elevated eGFR_Cr based on the CKD-EPI 2021 equation for Cr alone. It was also noted that the patient had a similar Cr spike to 1.83 mg/dL in 2018 without any identifiable renal insult or symptoms (Figure).

Diagnostic Evaluation

The primary care team ordered a renal ultrasound and referred the patient to the nephrology clinic. The nephrologist ordered the following laboratory studies: urine microalbumin to Cr ratio, basic metabolic panel (BMP), comprehensive metabolic panel (CMP), urinalysis, urine protein, urine Cr, parathyroid hormone level, hemoglobin A_1c, complement component 3/4 panels, antinuclear and antineutrophil cytoplasmic antibodies titers, glomerular basement membrane antibody titer, urine light chains, serum protein electrophoresis, κ/λ ratio, viral hepatitis panel, and rapid plasma reagin testing. Much of this laboratory evaluation served to rule out any secondary causes of kidney disease, including autoimmune disease, monoclonal or polyclonal gammopathies, diabetic nephropathy or glomerulosclerosis, and nephrotic or nephritic syndromes.

All laboratory studies returned within normal limits; no proteinuria was discovered on urinalysis, and no abnormalities were visualized on renal ultrasound. Bictegravir/emtricitabine/tenofovir alafenamide nephrotoxicity was highest among the differential diagnoses due to the timing of Cr elevation coinciding with the initiation of the medications. The patient's CysC level was 0.85 mg/dL with a calculated eGFR_Cys of 125 mL/min/1.73 m². The calculated sCR and serum cystatin C (eGFR_Cr-Cys) using the new 2021 equation and when adjusting for body surface area placed his eGFR at 92 mL/min/1.73 m².²⁰

The patient’s eGFR_Cysreassured the care team that the patient’s renal function was not acutely or chronically impacted by bictegravir/emtricitabine/tenofovir alafenamide, resulting in avoidance of unnecessary dosage adjustment or discontinuation of the HIV treatment. The patient reported a chronic habit of protein and creatine supplementation and bodybuilding, which likely further compounded the discrepancy between eGFR_Cr and eGFR_Cys and explained his previous elevation in Cr in 2018.

Follow-up

The patient underwent serial monitoring that revealed a stable Cr and unremarkable eGFR, ruling out CKD. There has been no evidence of worsening kidney disease to date, and the patient remained on his initial HIV regimen.

This case shows the importance of using CysC as an alternative or confirmatory marker compared with sCr to estimate GFR in patients with high muscle mass and/or high creatine intake, such as many in the US Department of Defense (DoD) and US Department of Veterans Affairs (VA) patient populations. In the presented case, recorded Cr levels climbed from baseline Cr with the initiation of bictegravir/emtricitabine/tenofovir alafenamide. This raised the concern that HIV treatment was leading to the development of kidney damage.²²

Diagnosis of kidney disease as opposed to the normal decline of eGFR with age in individuals without intrinsic CKD requires GFR ≥ 60 mL/min/1.73 m² with kidney damage (proteinuria or radiological abnormalities, etc) or GFR < 135 to 140 mL/min/1.73 m²minus the patient’s age in years.²³ The patient’s Cr peak at 1.83 mg/dL in 2018 led to an inappropriate diagnosis of kidney disease stage 3a based on an eGFR_Cr (2021 equation) of 52 mL/min/1.73 m² when not corrected for body surface area.²⁰ However, using the new 2021 equation using both Cr and CysC, the patient’s eGFR_Cr-Cyswas 92 mL/min/1.73 m² after a correction for body surface area.

The 2009 CKD-EPI recommended the calculation of eGFR based on SCr concentration using age, sex, and race while the 2021 CKD-EPI recommended the exclusion of race.³ Both equations are less accurate in African American patients, individuals taking medications that interfere with Cr secretion and assay, and patients taking creatine supplements, high daily protein intake, or with high muscle mass.⁷ These settings result in a decreased eGFR_Cr without corresponding eGFR_Cys changes. Using SCr and CysC together, the eGFR_Cr-Cys yields improved concordance to measured GFR across race groups compared to GFR estimation based on Cr alone, which can avoid unnecessary expensive diagnostic workup, inappropriate kidney disease diagnosis, incorrect dosing of drugs, and accurately represent the military readiness of patients. Interestingly, in African American patients with recently diagnosed HIV, CKD-EPI using both Cr and CysC without race inclusion led to only a 2.9% overestimation of GFR and was the only equation with no statistically significant bias compared with measured GFR.²⁴

A March 2023 case involving an otherwise healthy 26-year-old male active-duty US Navy member with a history of excessive protein supplement intake and intense exercise < 24 hours before laboratory work was diagnosed with CKD after a measured Cr of 16 mg/dL and an eGFR_Cr of 4 mL/min/1.73 m² without any other evidence of kidney disease. His CysC remained within normal limits, resulting in a normal eGFR_Cys of 121 mL/min/1.73 m², indicating no CKD. His Cr and eGFR recovered 10 days after his clinic visit and cessation of his supplement intake. These findings may not be uncommon given that 65% of active-duty military use protein supplements and 38% use other performance-enhancing supplements, such as creatine, according to a study.²⁵

Unfortunately, the BMP/CMP traditionally used at VA centers use the eGFR_Cr equation, and it is unknown how many primary care practitioners recognize the limitations of these metabolic panels on accurate estimation of kidney function. However, in 2022 an expert panel including VA physicians recommended the immediate use of eGFR_Cr-Cys or eGFR_Cys for confirmatory testing and potentially screening of CKD.²⁶ A small number of VAs have since adopted this recommendation, which should lead to fewer misdiagnoses among US military members as clinicians should now have access to more accurate measurements of GFR.

The VA spends about $18 billion (excluding dialysis) for care for 1.1 to 2.5 million VA patients with CKD.²⁷ The majority of these diagnoses were undoubtedly made using the eGFR_Cr equation, raising the question of how many may be misdiagnosed. Assessment with CysC is currently relatively expensive, but it will likely become more affordable as the use of CysC as a confirmatory test increases.⁵ The cost of a sCr test is about $2.50, while CysC costs about $10.60, with variation from laboratory to laboratory.²⁸ By comparison, a renal ultrasound costs $99 to $140 for uninsured patients.²⁹ Furthermore, the cost of CysC testing is likely to trend downward as more facilities adopt the use of CysC measurements, which can be run on the same analytical equipment currently used for Cr measurements. Currently, most laboratories do not have established assays to use in-house and thus require CysC to be sent out to a laboratory, which increases result time and makes Cr a more attractive option. As more laboratories adopt assays for CysC, the cost of reagents will further decrease.

Given such considerations, confirmation testing of kidney function with CysC in specific patient populations with decreased eGFR_Cr without other features of CKD can offer great medical and financial benefits. A 2023 KDIGO report noted that many individuals may be mistakenly diagnosed with CKD when using eGFR_Cr.³ KDIGO noted that a 2013 meta-analysis of 90,000 individuals found that with a Cr-based eGFR of 45 to 59 mL/min/1.73 m² (42%) had a CysC-based eGFR of ≥ 60 mL/min/1.73 m². An eGFR_Cr of 45 to 59 represents 54% of all patients with CKD, amounting to millions of people (including current and former military personnel).^3,29-31 Correcting a misdiagnosis of CKD would bring significant relief to patients and save millions in health care spending.

Conclusions

In patients who meet CKD criteria using eGFR_Cr but without other features of CKD, we recommend using confirmatory CysC levels and the eGFR_Cr-Cys equation. This will align care with the KDIGO guidelines and could be a cost-effective step toward improving military patient care. Further work in this area should focus on determining the knowledge gaps in primary care practitioners’ understanding of the limits of eGFR_Cr, the potential mitigation of concomitant CysC testing in equivocal CKD cases, and the cost-effectiveness and increased utilization of CysC.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.

Hypertension is a major risk factor for heart disease, stroke, and kidney disease.The prevalence of hypertension increases with age, primarily due to age-related changes in arterial physiology.¹ For older adults, current guidelines regarding blood pressure (BP) treatment goals vary. The American Heart Association/American College of Cardiology 2017 clinical practice guidelines recommend a systolic BP (SBP) treatment goal of < 130 mm Hg for community-dwelling, ambulatory, noninstitutionalized adults aged ≥ 65 years; whereas the American College of Physicians/American Academy of Family Physicians recommend a goal of < 150 mm Hg for those aged ≥ 60 years without comorbidities and < 140 mm Hg for those with increased cardiovascular risk.^1-3 Regardless of the specific threshold, agreement that some degree of BP control even in those with advanced age improves outcomes.²

First-line therapy for uncomplicated hypertension includes thiazide diuretics, long-acting calcium channel blockers, and renin-angiotensin system inhibitors. When choosing between these options, it is recommended to engage in shared decision making and to consider the patient’s comorbidities. Among patients who are likely to require a second agent (eg, if initial BP is > 20/10 mm Hg above goal), it is recommended to begin both drugs at the same time, preferably benazepril plus amlodipine due to the reduction in cardiovascular events reported in the ACCOMPLISH trial.⁴ If BP remains elevated despite 2 agents at moderate to maximum doses, it is important to investigate for secondary hypertension causes and to explore medication adherence as possible etiologies of treatment failure. Older adults are often at higher risk of adverse drug events due to age-related changes in pharmacodynamics. Despite this, there are no guidelines for choosing between different classes of antihypertensives in this population. We present a case of thiazide-induced hyponatremia in an older adult and review the risks of thiazide use in this population.

Case Presentation

A man aged > 90 years was admitted to the hospital after a syncopal episode. His history was significant for hypertension, hyperlipidemia, and vitamin D deficiency. At the time, his home medications were amlodipine 5 mg daily, atorvastatin 40 mg daily, ergocalciferol 50,000 IU weekly, and polyethylene glycol 17 g daily as needed. His syncope workup was unremarkable and included negative orthostatic vital signs, normal serial troponins, an electrocardiogram without ischemic changes, normal serum creatinine, sodium, and glucose, and a head computed tomography without any acute abnormality. Throughout the patient’s hospital stay, he had multiple elevated SBP readings, including many > 200 mm Hg. On discharge, in addition to continuing his home medications, he was started on valsartan 20 mg daily and enrolled in a remote BP monitoring program.

Three weeks later, the patient was seen by their primary care practitioner for follow-up. He reported adherence to his antihypertensive regimen. However, his remote BP monitoring revealed persistently elevated BPs, with an average of 179/79 mm Hg, a high of 205/85 mm Hg, and a low of 150/67 mm Hg over the previous 7 days. Laboratory tests obtained at the visit were notable for serum sodium of 138 mmol/L and potassium of 4.1 mmol/L. His weight was 87 kg. Given persistently elevated BP readings, in addition to continuing his amlodipine 5 mg daily and valsartan 20 mg daily, he was started on hydrochlorothiazide 25 mg daily, with plans to repeat a basic metabolic panel in 2 weeks.

Two weeks later, he fell after getting out of his bed. On examination, he was noted to have dry mucous membranes, and although no formal delirium screening was performed, he was able to repeat the months of the year backward. Vital signs were notable for positive postural hypertension, and his laboratory tests revealed a normal serum creatinine, serum sodium of 117 mmol/L (reference range, 135-145 mmol/L), serum potassium of 3.2 mmol/L (reference range, 3.0-5.0 mmol/L), a low serum osmolality, and urine sodium of 35 mmol/L most consistent with hypovolemic hypoosmotic hyponatremia secondary to thiazide initiation. The patient’s hydrochlorothiazide was discontinued, and he was admitted to the hospital for close monitoring. His sodium levels gradually normalized over the next 2 weeks without any other intervention.

Discussion

Although thiazide diuretics are recommended as first-line therapy for uncomplicated hypertension, they are known to cause electrolyte abnormalities, including hypomagnesemia, hypokalemia, and hyponatremia.⁴ These metabolic derangements are more likely to occur in older adults. One study of adults aged ≥ 65 years found that at 9 months of follow-up, 14.3% of new thiazide users had developed a thiazide-related metabolic adverse event (hyponatremia < 135 mmol/L, hypokalemia < 3.5 mmol/L, and decrease in estimated glomerular filtration rate by > 25%) compared with 6.0% of nonusers (P < .001; number needed to harm [NNH] = 12).⁵ In addition, 3.8% of new thiazide users had an emergency department visit or were hospitalized for complications related to thiazides compared with only 2.0% of nonusers (P = .02; NNH = 56).⁵ Independent risk factors for thiazide-induced hyponatremia include high-comorbidity burden, low body weight, low-normal or unmeasured serum sodium, low potassium, and aged > 70 years.^5-7 Each 10-year increment in age is associated with a 2-fold increase in risk, suggesting that older adults are at a much higher risk for hyponatremia than their younger peers.⁶

Despite their designation as a first-line option for uncomplicated hypertension, thiazide diuretics may cause more harm than good in some older adults, especially those with additional risk factors for thiazide-induced hyponatremia. In this population, these adverse effects should be discussed before starting thiazides for the treatment of hypertension. If thiazides are initiated, they should be started at the lowest possible dose, and plans made to monitor bloodwork within 1 to 2 weeks of initiation or dose change and periodically thereafter while the patient remains on the therapy.

Medication Management in Older Adults

Due to the risks of medication use in older adults, the phrase “start low, go slow” is commonly used in geriatric medicine to describe the optimal method for initiation and up-titration of new medication with the hope of mitigating adverse drug events. In our case, we started valsartan at 20 mg daily—one-fourth the recommended initial dose. Although this strategy is reasonable to “start low,” we were not surprised to find that the patient’s BP did not markedly improve on such a low dose. The team could have increased the valsartan dose to a therapeutically efficacious dose before choosing to add another hypertensive agent. In alignment with geriatric prescribing principles, starting at the lowest possible dose of hydrochlorothiazide is recommended.⁵ However, the clinician started hydrochlorothiazide at 25 mg daily, potentially increasing this patient’s risk of electrolyte abnormalities and eventual fall.

Managing hypertension also invites a discussion of polypharmacy and medication adherence. Older adults are at risk of polypharmacy, defined as the prescription of 5 or more medications.⁸ Polypharmacy is associated with increased hospitalizations, higher costs of care for individuals and health care systems, increased risks of adverse drug events, medication nonadherence, and lower quality of life for patients.⁹ In some situations, the risks of polypharmacy may outweigh the benefits of using multiple antihypertensives with different mechanisms of action if patients can reach their BP goal on the maximum dose of a single agent. For patients taking multiple antihypertensives, it is important to routinely monitor BP and assess whether deprescribing is indicated. Cognitive impairment and decreased social support may affect medication adherence for older adults.⁶ Clinicians should be aware of strategies, such as medication reminders and pillboxes, to increase antihypertensive medication adherence. Polypills that contain 2 antihypertensives can be another tool used to manage older adults to increase adherence and decrease health care costs.¹⁰

A current strategy that encompasses discussing many, if not all, of these noted elements is the Institute for Healthcare Improvement’s Age-Friendly Health System. This framework uses evidence-based tools to provide care for older adults across all clinical settings and highlights the 4Ms: what matters, medication, mentation, and mobility.¹¹ Medication considers whether a medication is necessary, whether its use has benefits that outweigh the risks, and how it interacts with what matters, mentation, and mobility. In particular, what matters plays an important role in hypertension management in older adults given the recommended target BP differs, depending on which specialty organization guideline is followed. By better understanding what matters to patients, including their goals and priorities, clinicians can engage patients in shared decision making and provide individualized recommendations based on geriatric principles (eg, start low, go slow, principles of medication adherence) and patient comorbidities (eg, medical history and risk factors for hyponatremia) to help patients make a more informed choice about their antihypertensive treatment regimen (Figure).

Conclusions

This case illustrates the need for a specialized approach to hypertension management in older adults and the risks of thiazide diuretics in this population. Clinicians should consider BP goals, patient-specific factors, and principles of medication management in older adults. If initiating thiazide therapy, discuss the risks associated with use, start at the lowest possible dose, and monitor bloodwork within 1 to 2 weeks of initiation/dose change and periodically thereafter while the patient remains on the therapy to decrease the risk of adverse events. Finally, the Institute for Healthcare Improvement’s Age-Friendly Health System framework can be a useful when considering the addition of a new medication in an older adult’s treatment plan.

Acknowledgments

This material is the result of work supported with resources and the use of facilities at the New England Geriatrics Research, Education, and Clinical Center, Veterans Affairs Boston Healthcare System, and the Cincinnati VeteransAffairs Medical Center.