Joan Nashelsky, MLS

Evidence summary

Five randomized controlled trials have examined lung cancer mortality after screening chest x-rays. In the first trial—the only one that included former as well as current smokers and nonsmokers—subjects were randomized to undergo chest x-ray studies every 6 months, or at baseline and again at the end of the 3-year study. After 3 years, there was no statistically significant mortality difference with more frequent chest x-rays.^1,2

Another trial involved male smokers who were randomized to undergo chest x-ray and sputum cytology either every 6 months or after 3 years. After 3 years, both groups were screened annually with chest x-ray alone for an additional 3 years. There was no significant difference in lung cancer mortality at any point, including at a 15-year post-trial follow-up.³Both studies showed earlier detection and longer survivorship of lung cancer among screened vs nonscreened groups due to lead-time bias (because the cancer was detected earlier from screening vs clinical diagnosis, it falsely appears to prolong survival). Overall mortality was the same in both groups.

The National Cancer Institute (NCI) sponsored 3 randomized controlled trials on lung cancer screening for male smokers involving 3 major medical centers. The studies were designed to determine the incremental benefit of adding sputum cytology to chest x-ray screening. In 2 of the NCI studies, participants were randomly assigned to receive annual chest x-ray only or a dual screen with annual chest x-ray and sputum cytologies every 4 months. In both studies, there was no statistical difference in lung cancer mortality between the 2 groups.^4-6The third NCI study randomized participants to chest x-ray and sputum cytology either every 4 months or annually. Again, there was no significant difference in lung cancer mortality,⁴even after an extended follow-up of 20.5 years.⁷Adding sputum cytology to chest x-ray only improved lung cancer detection rates over chest x-ray alone.

A significant limitation of the 5 studies presented is that no true control or non-screening groups determined the real efficacy of screening chest x-rays vs no screening. The goal of a study of a screening program is to detect a disease early enough so that treatment can alter mortality. These uncontrolled studies of routine screening chest x-rays, no matter how frequently performed, do not meet this criteria for current and former smokers.

Recommendations from others

The US Preventive Services Task Force does not recommend for or against screening asymptomatic or high-risk persons for lung cancer with either low-dose computed tomography (CT), chest x-ray, sputum cytology, or a combination of these tests.⁸ The American Cancer Society and American Academy of Family Physicians recommend against the use of chest x-ray or sputum cytology in asymptomatic high-risk persons.^9,10The American College of Chest Physicians recommends against the use of serial chest x-rays for individuals without symptoms or without a history of cancer.¹¹ They do not comment about high-risk groups—that is, current or former smokers.

Evidence summary

Five randomized controlled trials have examined lung cancer mortality after screening chest x-rays. In the first trial—the only one that included former as well as current smokers and nonsmokers—subjects were randomized to undergo chest x-ray studies every 6 months, or at baseline and again at the end of the 3-year study. After 3 years, there was no statistically significant mortality difference with more frequent chest x-rays.^1,2

Another trial involved male smokers who were randomized to undergo chest x-ray and sputum cytology either every 6 months or after 3 years. After 3 years, both groups were screened annually with chest x-ray alone for an additional 3 years. There was no significant difference in lung cancer mortality at any point, including at a 15-year post-trial follow-up.³Both studies showed earlier detection and longer survivorship of lung cancer among screened vs nonscreened groups due to lead-time bias (because the cancer was detected earlier from screening vs clinical diagnosis, it falsely appears to prolong survival). Overall mortality was the same in both groups.

The National Cancer Institute (NCI) sponsored 3 randomized controlled trials on lung cancer screening for male smokers involving 3 major medical centers. The studies were designed to determine the incremental benefit of adding sputum cytology to chest x-ray screening. In 2 of the NCI studies, participants were randomly assigned to receive annual chest x-ray only or a dual screen with annual chest x-ray and sputum cytologies every 4 months. In both studies, there was no statistical difference in lung cancer mortality between the 2 groups.^4-6The third NCI study randomized participants to chest x-ray and sputum cytology either every 4 months or annually. Again, there was no significant difference in lung cancer mortality,⁴even after an extended follow-up of 20.5 years.⁷Adding sputum cytology to chest x-ray only improved lung cancer detection rates over chest x-ray alone.

A significant limitation of the 5 studies presented is that no true control or non-screening groups determined the real efficacy of screening chest x-rays vs no screening. The goal of a study of a screening program is to detect a disease early enough so that treatment can alter mortality. These uncontrolled studies of routine screening chest x-rays, no matter how frequently performed, do not meet this criteria for current and former smokers.

Recommendations from others

The US Preventive Services Task Force does not recommend for or against screening asymptomatic or high-risk persons for lung cancer with either low-dose computed tomography (CT), chest x-ray, sputum cytology, or a combination of these tests.⁸ The American Cancer Society and American Academy of Family Physicians recommend against the use of chest x-ray or sputum cytology in asymptomatic high-risk persons.^9,10The American College of Chest Physicians recommends against the use of serial chest x-rays for individuals without symptoms or without a history of cancer.¹¹ They do not comment about high-risk groups—that is, current or former smokers.

Evidence summary

Five randomized controlled trials have examined lung cancer mortality after screening chest x-rays. In the first trial—the only one that included former as well as current smokers and nonsmokers—subjects were randomized to undergo chest x-ray studies every 6 months, or at baseline and again at the end of the 3-year study. After 3 years, there was no statistically significant mortality difference with more frequent chest x-rays.^1,2

Another trial involved male smokers who were randomized to undergo chest x-ray and sputum cytology either every 6 months or after 3 years. After 3 years, both groups were screened annually with chest x-ray alone for an additional 3 years. There was no significant difference in lung cancer mortality at any point, including at a 15-year post-trial follow-up.³Both studies showed earlier detection and longer survivorship of lung cancer among screened vs nonscreened groups due to lead-time bias (because the cancer was detected earlier from screening vs clinical diagnosis, it falsely appears to prolong survival). Overall mortality was the same in both groups.

The National Cancer Institute (NCI) sponsored 3 randomized controlled trials on lung cancer screening for male smokers involving 3 major medical centers. The studies were designed to determine the incremental benefit of adding sputum cytology to chest x-ray screening. In 2 of the NCI studies, participants were randomly assigned to receive annual chest x-ray only or a dual screen with annual chest x-ray and sputum cytologies every 4 months. In both studies, there was no statistical difference in lung cancer mortality between the 2 groups.^4-6The third NCI study randomized participants to chest x-ray and sputum cytology either every 4 months or annually. Again, there was no significant difference in lung cancer mortality,⁴even after an extended follow-up of 20.5 years.⁷Adding sputum cytology to chest x-ray only improved lung cancer detection rates over chest x-ray alone.

A significant limitation of the 5 studies presented is that no true control or non-screening groups determined the real efficacy of screening chest x-rays vs no screening. The goal of a study of a screening program is to detect a disease early enough so that treatment can alter mortality. These uncontrolled studies of routine screening chest x-rays, no matter how frequently performed, do not meet this criteria for current and former smokers.

Recommendations from others

The US Preventive Services Task Force does not recommend for or against screening asymptomatic or high-risk persons for lung cancer with either low-dose computed tomography (CT), chest x-ray, sputum cytology, or a combination of these tests.⁸ The American Cancer Society and American Academy of Family Physicians recommend against the use of chest x-ray or sputum cytology in asymptomatic high-risk persons.^9,10The American College of Chest Physicians recommends against the use of serial chest x-rays for individuals without symptoms or without a history of cancer.¹¹ They do not comment about high-risk groups—that is, current or former smokers.

Evidence summary

Breast cancer is the second leading cause of cancer death among American women; 1 in 8 women will be diagnosed with breast cancer in her lifetime, and 1 in 30 will die of it.³ Breast cancer screening and mammography have become almost synonymous. But physical examinations by clinicians or women themselves remain important methods of screening to consider.

Breast self-examination is appealing as a patient-centered, inexpensive, noninvasive procedure that empowers women and is universally available. However, a recent Cochrane review found no evidence of benefit from self-screening.

Two large RCTs, conducted in St Petersburg, Russia (122,471 women) and Shanghai, China (266,064 women), were found. Both studies used cluster randomization (by worksite) and involved large numbers of women who were meticulously trained in proper breast self-examination technique and had numerousreinforcement sessions. Study compliance and follow-up were excellent. Outcomes assessment was explicitly blinded in the Shanghai study. Neither trial demonstrated a reduction in breast cancer mortality or improvement in the number or stage of cancers detected during 9 to 11 years of follow-up, but there is evidence for harm: a nearly 2-fold increase in false-positive results, physician visits, and biopsies for benign disease.⁴

No trials comparing screening clinical breast examinations alone to no screening have been reported, but good indirect evidence of efficacy comes from the results of the Canadian National Breast Screening Study-2 (CNBSS-2).⁵ A total of 39,405 women aged 50 to 59 years were randomized to screening with clinical exams plus mammography or clinical exams alone. Other large RCTs have shown a consistent benefit to mammography screening for women of this age (in-depth independent reviews of recent criticism of the trials have concluded that their flaws do not negate mammography’s efficacy in reducing breast cancer mortality).^3,6 The CNBSS-2 trial showed no mortality advantage when mammography was added to an annual, standardized 10- to 15-minute breast examination, implying that careful, detailed, annual clinical breast examinations may be as effective as a mammography screening program.³

Recommendations from others

The US Preventive Services Task Force found insufficient evidence to recommend for or against routine clinical exams alone to screen for breast cancer, or to recommend for or against teaching or performing routine breast self-examination.³ The Canadian Task Force on Preventive Health Services recommends against teaching self-examination to women aged 40 to 69 years due to “fair evidence of no benefit and good evidence of harm.”^7,8

The American Cancer Society continues to recommend periodic clinical exams,⁶ and women who choose to do self-examination should receive instruction and have their technique reviewed during periodic health examinations; it is acceptable for women to choose not to do self-examinations. The American Academy of Family Physicians concludes that the evidence is insufficient to recommend for or against breast self-examination.⁹ The American College of Obstetricians and Gynecologists recommends both.¹⁰

Evidence summary

Breast cancer is the second leading cause of cancer death among American women; 1 in 8 women will be diagnosed with breast cancer in her lifetime, and 1 in 30 will die of it.³ Breast cancer screening and mammography have become almost synonymous. But physical examinations by clinicians or women themselves remain important methods of screening to consider.

Breast self-examination is appealing as a patient-centered, inexpensive, noninvasive procedure that empowers women and is universally available. However, a recent Cochrane review found no evidence of benefit from self-screening.

Two large RCTs, conducted in St Petersburg, Russia (122,471 women) and Shanghai, China (266,064 women), were found. Both studies used cluster randomization (by worksite) and involved large numbers of women who were meticulously trained in proper breast self-examination technique and had numerousreinforcement sessions. Study compliance and follow-up were excellent. Outcomes assessment was explicitly blinded in the Shanghai study. Neither trial demonstrated a reduction in breast cancer mortality or improvement in the number or stage of cancers detected during 9 to 11 years of follow-up, but there is evidence for harm: a nearly 2-fold increase in false-positive results, physician visits, and biopsies for benign disease.⁴

No trials comparing screening clinical breast examinations alone to no screening have been reported, but good indirect evidence of efficacy comes from the results of the Canadian National Breast Screening Study-2 (CNBSS-2).⁵ A total of 39,405 women aged 50 to 59 years were randomized to screening with clinical exams plus mammography or clinical exams alone. Other large RCTs have shown a consistent benefit to mammography screening for women of this age (in-depth independent reviews of recent criticism of the trials have concluded that their flaws do not negate mammography’s efficacy in reducing breast cancer mortality).^3,6 The CNBSS-2 trial showed no mortality advantage when mammography was added to an annual, standardized 10- to 15-minute breast examination, implying that careful, detailed, annual clinical breast examinations may be as effective as a mammography screening program.³

Recommendations from others

The US Preventive Services Task Force found insufficient evidence to recommend for or against routine clinical exams alone to screen for breast cancer, or to recommend for or against teaching or performing routine breast self-examination.³ The Canadian Task Force on Preventive Health Services recommends against teaching self-examination to women aged 40 to 69 years due to “fair evidence of no benefit and good evidence of harm.”^7,8

The American Cancer Society continues to recommend periodic clinical exams,⁶ and women who choose to do self-examination should receive instruction and have their technique reviewed during periodic health examinations; it is acceptable for women to choose not to do self-examinations. The American Academy of Family Physicians concludes that the evidence is insufficient to recommend for or against breast self-examination.⁹ The American College of Obstetricians and Gynecologists recommends both.¹⁰

Evidence summary

Breast cancer is the second leading cause of cancer death among American women; 1 in 8 women will be diagnosed with breast cancer in her lifetime, and 1 in 30 will die of it.³ Breast cancer screening and mammography have become almost synonymous. But physical examinations by clinicians or women themselves remain important methods of screening to consider.

Breast self-examination is appealing as a patient-centered, inexpensive, noninvasive procedure that empowers women and is universally available. However, a recent Cochrane review found no evidence of benefit from self-screening.

Two large RCTs, conducted in St Petersburg, Russia (122,471 women) and Shanghai, China (266,064 women), were found. Both studies used cluster randomization (by worksite) and involved large numbers of women who were meticulously trained in proper breast self-examination technique and had numerousreinforcement sessions. Study compliance and follow-up were excellent. Outcomes assessment was explicitly blinded in the Shanghai study. Neither trial demonstrated a reduction in breast cancer mortality or improvement in the number or stage of cancers detected during 9 to 11 years of follow-up, but there is evidence for harm: a nearly 2-fold increase in false-positive results, physician visits, and biopsies for benign disease.⁴

No trials comparing screening clinical breast examinations alone to no screening have been reported, but good indirect evidence of efficacy comes from the results of the Canadian National Breast Screening Study-2 (CNBSS-2).⁵ A total of 39,405 women aged 50 to 59 years were randomized to screening with clinical exams plus mammography or clinical exams alone. Other large RCTs have shown a consistent benefit to mammography screening for women of this age (in-depth independent reviews of recent criticism of the trials have concluded that their flaws do not negate mammography’s efficacy in reducing breast cancer mortality).^3,6 The CNBSS-2 trial showed no mortality advantage when mammography was added to an annual, standardized 10- to 15-minute breast examination, implying that careful, detailed, annual clinical breast examinations may be as effective as a mammography screening program.³

Recommendations from others

The US Preventive Services Task Force found insufficient evidence to recommend for or against routine clinical exams alone to screen for breast cancer, or to recommend for or against teaching or performing routine breast self-examination.³ The Canadian Task Force on Preventive Health Services recommends against teaching self-examination to women aged 40 to 69 years due to “fair evidence of no benefit and good evidence of harm.”^7,8

The American Cancer Society continues to recommend periodic clinical exams,⁶ and women who choose to do self-examination should receive instruction and have their technique reviewed during periodic health examinations; it is acceptable for women to choose not to do self-examinations. The American Academy of Family Physicians concludes that the evidence is insufficient to recommend for or against breast self-examination.⁹ The American College of Obstetricians and Gynecologists recommends both.¹⁰

Evidence summary

In the United States, approximately 5000 infants with moderate-to-profound hearing loss are born annually.¹ Affected children graduate high school averaging 4th-grade academic performance skills.² Efforts to reduce the impact on these children have focused on early diagnosis and treatment.

A systematic review gathered studies comparing universal hearing screening with selective screening.¹ Most included studies used a 2-stage universal screening protocol. Infants who failed initial testing were retested within 12 weeks. Testing methods included otoacoustic emissions (OAE) and auditory brainstem response (ABR). Infants who failed the second test were referred for audiological evaluation. Using these data, a hypothetical model was created, which found that 1441 newborns would need to be screened to diagnose 1 additional case of moderate-to-profound permanent hearing loss before 10 months of age (at cost of 200 extra referrals for false-positives). Sensitivity and specificity of the hypothetical model’s 2-stage screening was 85% and 97%, respectively. The estimated positive predictive value was 6.7%.^1,3

Individually, OAE and ABR accurately diagnose neonatal hearing loss. One multicenter cohort of 2995 infants measured test performance of OAE and ABR against the gold standard (visual reinforcement audiometry performed at 8–12 months).⁴ The authors used a receiver operating characteristics (ROC) curve to plot speech awareness thresholds for both tests. When middle-ear pathology and progressive hearing loss were excluded, the area under the ROC curves for ABR and OAE were 0.91 and 0.94, respectively, indication that both tests had excellent test accuracy (a perfect test would have an area under the curve of 1.0).

Strategies based on selective screening of high-risk infants fails to identify permanent hearing loss in many affected infants. In a cohort study of more than 10,000 infants, only 43% of infants with permanent hearing loss were identified with selective versus universal screening. Most affected infants would have been missed using risk-based criteria.⁵

Limited evidence suggests that early identification of infants with permanent hearing loss improves language skills. In a retrospective cohort study of 150 infants examining language outcomes, participants were grouped according to age at identification of hearing loss.⁶ All participants received comprehensive in-home language intervention services plus amplification devices.

Of the 85 children with normal cognitive ability, the mean receptive and expressive language quotients at 13 to 36 months were higher in the early-identified group vs the late-identified group (receptive language quotients, 79.6 vs 64.6, P<.001; expressive language quotients, 78.3 vs 63.1, P<.001). Total language quotient was also higher in the early group (language quotients, 79 vs 64; P<.001).

The conclusions were limited by multiple factors: retrospective study design, cohort selection drawn from different hospitals during different time periods, unblinded participant selection, and unblended outcome assessments. Other published studies have inconclusive outcome data. The Cochrane Collaboration published a systematic review in which no studies were found that fulfilled the inclusion criteria to evaluate the effectiveness of universal hearing screening.⁷

Recommendations from others

The Joint Committee on Infant Hearing recommended universal neonatal hearing screening during hospital birth admission in their Year 2000 Position Statement.⁸ For infants whose hearing is impaired on re-screening, the committee recommends audiology referral and medical evaluation to rule out associated conditions before age 3 months. They further recommend interventional services begin before age 6 months for infants with confirmed hearing loss.

The US Preventive Services Task Force does not recommend for or against universal hearing screening, citing insufficient outcomes data.⁹

Evidence summary

In the United States, approximately 5000 infants with moderate-to-profound hearing loss are born annually.¹ Affected children graduate high school averaging 4th-grade academic performance skills.² Efforts to reduce the impact on these children have focused on early diagnosis and treatment.

A systematic review gathered studies comparing universal hearing screening with selective screening.¹ Most included studies used a 2-stage universal screening protocol. Infants who failed initial testing were retested within 12 weeks. Testing methods included otoacoustic emissions (OAE) and auditory brainstem response (ABR). Infants who failed the second test were referred for audiological evaluation. Using these data, a hypothetical model was created, which found that 1441 newborns would need to be screened to diagnose 1 additional case of moderate-to-profound permanent hearing loss before 10 months of age (at cost of 200 extra referrals for false-positives). Sensitivity and specificity of the hypothetical model’s 2-stage screening was 85% and 97%, respectively. The estimated positive predictive value was 6.7%.^1,3

Individually, OAE and ABR accurately diagnose neonatal hearing loss. One multicenter cohort of 2995 infants measured test performance of OAE and ABR against the gold standard (visual reinforcement audiometry performed at 8–12 months).⁴ The authors used a receiver operating characteristics (ROC) curve to plot speech awareness thresholds for both tests. When middle-ear pathology and progressive hearing loss were excluded, the area under the ROC curves for ABR and OAE were 0.91 and 0.94, respectively, indication that both tests had excellent test accuracy (a perfect test would have an area under the curve of 1.0).

Strategies based on selective screening of high-risk infants fails to identify permanent hearing loss in many affected infants. In a cohort study of more than 10,000 infants, only 43% of infants with permanent hearing loss were identified with selective versus universal screening. Most affected infants would have been missed using risk-based criteria.⁵

Limited evidence suggests that early identification of infants with permanent hearing loss improves language skills. In a retrospective cohort study of 150 infants examining language outcomes, participants were grouped according to age at identification of hearing loss.⁶ All participants received comprehensive in-home language intervention services plus amplification devices.

Of the 85 children with normal cognitive ability, the mean receptive and expressive language quotients at 13 to 36 months were higher in the early-identified group vs the late-identified group (receptive language quotients, 79.6 vs 64.6, P<.001; expressive language quotients, 78.3 vs 63.1, P<.001). Total language quotient was also higher in the early group (language quotients, 79 vs 64; P<.001).

The conclusions were limited by multiple factors: retrospective study design, cohort selection drawn from different hospitals during different time periods, unblinded participant selection, and unblended outcome assessments. Other published studies have inconclusive outcome data. The Cochrane Collaboration published a systematic review in which no studies were found that fulfilled the inclusion criteria to evaluate the effectiveness of universal hearing screening.⁷

Recommendations from others

The Joint Committee on Infant Hearing recommended universal neonatal hearing screening during hospital birth admission in their Year 2000 Position Statement.⁸ For infants whose hearing is impaired on re-screening, the committee recommends audiology referral and medical evaluation to rule out associated conditions before age 3 months. They further recommend interventional services begin before age 6 months for infants with confirmed hearing loss.

The US Preventive Services Task Force does not recommend for or against universal hearing screening, citing insufficient outcomes data.⁹

Evidence summary

In the United States, approximately 5000 infants with moderate-to-profound hearing loss are born annually.¹ Affected children graduate high school averaging 4th-grade academic performance skills.² Efforts to reduce the impact on these children have focused on early diagnosis and treatment.

A systematic review gathered studies comparing universal hearing screening with selective screening.¹ Most included studies used a 2-stage universal screening protocol. Infants who failed initial testing were retested within 12 weeks. Testing methods included otoacoustic emissions (OAE) and auditory brainstem response (ABR). Infants who failed the second test were referred for audiological evaluation. Using these data, a hypothetical model was created, which found that 1441 newborns would need to be screened to diagnose 1 additional case of moderate-to-profound permanent hearing loss before 10 months of age (at cost of 200 extra referrals for false-positives). Sensitivity and specificity of the hypothetical model’s 2-stage screening was 85% and 97%, respectively. The estimated positive predictive value was 6.7%.^1,3

Individually, OAE and ABR accurately diagnose neonatal hearing loss. One multicenter cohort of 2995 infants measured test performance of OAE and ABR against the gold standard (visual reinforcement audiometry performed at 8–12 months).⁴ The authors used a receiver operating characteristics (ROC) curve to plot speech awareness thresholds for both tests. When middle-ear pathology and progressive hearing loss were excluded, the area under the ROC curves for ABR and OAE were 0.91 and 0.94, respectively, indication that both tests had excellent test accuracy (a perfect test would have an area under the curve of 1.0).

Strategies based on selective screening of high-risk infants fails to identify permanent hearing loss in many affected infants. In a cohort study of more than 10,000 infants, only 43% of infants with permanent hearing loss were identified with selective versus universal screening. Most affected infants would have been missed using risk-based criteria.⁵

Limited evidence suggests that early identification of infants with permanent hearing loss improves language skills. In a retrospective cohort study of 150 infants examining language outcomes, participants were grouped according to age at identification of hearing loss.⁶ All participants received comprehensive in-home language intervention services plus amplification devices.

Of the 85 children with normal cognitive ability, the mean receptive and expressive language quotients at 13 to 36 months were higher in the early-identified group vs the late-identified group (receptive language quotients, 79.6 vs 64.6, P<.001; expressive language quotients, 78.3 vs 63.1, P<.001). Total language quotient was also higher in the early group (language quotients, 79 vs 64; P<.001).

The conclusions were limited by multiple factors: retrospective study design, cohort selection drawn from different hospitals during different time periods, unblinded participant selection, and unblended outcome assessments. Other published studies have inconclusive outcome data. The Cochrane Collaboration published a systematic review in which no studies were found that fulfilled the inclusion criteria to evaluate the effectiveness of universal hearing screening.⁷

Recommendations from others

The Joint Committee on Infant Hearing recommended universal neonatal hearing screening during hospital birth admission in their Year 2000 Position Statement.⁸ For infants whose hearing is impaired on re-screening, the committee recommends audiology referral and medical evaluation to rule out associated conditions before age 3 months. They further recommend interventional services begin before age 6 months for infants with confirmed hearing loss.

The US Preventive Services Task Force does not recommend for or against universal hearing screening, citing insufficient outcomes data.⁹

Evidence summary

Because of the nature of major surgery, there are practical and ethical barriers to true randomized controlled trials (RCTs) comparing bariatric surgery with placebo or to no intervention. However, multiple RCTs have compared the weight-reducing effects of different bariatric surgical techniques against each other.¹ All studies included patients who had a BMI >40 kg/m², or a BMI >35 kg/m² with at least 1 comorbidity, such as cardiovascular disease, sleep apnea, uncontrolled type 2 diabetes, or weight-induced physical problems interfering with performance of daily life activities. It is these study inclusion criteria that, by default, have become widely accepted indications for bariatric surgery. Weight loss in all RCTs was substantial, ranging from 50 to 100 kg over 6 months to 1 year. Comorbid factors associated with obesity showed either resolution or improvement after surgery in 91% of patients.

Patients with a BMI >40 have substantially more serious health consequences and a reduced life expectancy. Obesity significantly impairs quality of life, and the risk of morbidity and mortality increases with the degree of obesity.² Those who are extremely obese often do not have sustained benefit from more conservative treatment. The benefits of nonsurgical treatment are significantly limited by the failure to maintain reduced body weight.

A large literature of controlled and uncontrolled cohort studies show that surgery has produced the longest period of sustained weight loss.³ A recent meta-analysis proved bariatric surgery not only efficacious for weight loss, but showed that a substantial majority of patients with diabetes, hyperlipidemia, hypertension, and obstructive sleep apnea experienced complete resolution or significant improvement of their comorbid condition after surgery.⁴

The possibility of significant adverse effects remains. The postoperative mortality rate for bariatric surgery is approximately 0.2%. Reoperation is required for up to 25% of patients within 5 years. Other complications are wound infection, staple failure, vitamin deficiency, diarrhea, and hemorrhage.³ The long-term health effects of bariatric surgery are not well known.

Recommendations from others

The NIH statement “Gastrointestinal Surgery for Severe Obesity” concluded that the benefits outweigh the risks and that surgical treatment is reasonable in those who strongly desire substantial weight loss and have life-threatening comorbid conditions.²

Clinical guidelines developed by the National Heart, Lung, and Blood Institute Expert Panel on the identification, evaluation, and treatment of obesity for adults recommend that bariatric surgery be an option for carefully selected patients with clinically severe obesity (BMI >40 or >35 with comorbid conditions) when less invasive methods of weight loss have failed and the patient is at high risk for obesity-associated morbidity and mortality.¹

The American Gastroenterological Association (AGA) medical position statement on obesity finds surgical therapy to be the most effective approach for achieving long-term weight loss. The AGA recommends surgery for patients with a BMI >40, or those with BMI >35 and 1 or more severe obesity-related medical complication (eg, hypertension, heart failure, or sleep apnea) if they have been unable to achieve or maintain weight loss with conventional therapy, have acceptable operative risks, and are able to comply with long-term treatment and follow-up.⁵

The American College of Preventive Medicine, in its policy statement on weight management counseling, recommends limiting surgical therapy for obesity to severely obese patients, defined as BMI >40.⁶

Evidence summary

Because of the nature of major surgery, there are practical and ethical barriers to true randomized controlled trials (RCTs) comparing bariatric surgery with placebo or to no intervention. However, multiple RCTs have compared the weight-reducing effects of different bariatric surgical techniques against each other.¹ All studies included patients who had a BMI >40 kg/m², or a BMI >35 kg/m² with at least 1 comorbidity, such as cardiovascular disease, sleep apnea, uncontrolled type 2 diabetes, or weight-induced physical problems interfering with performance of daily life activities. It is these study inclusion criteria that, by default, have become widely accepted indications for bariatric surgery. Weight loss in all RCTs was substantial, ranging from 50 to 100 kg over 6 months to 1 year. Comorbid factors associated with obesity showed either resolution or improvement after surgery in 91% of patients.

Patients with a BMI >40 have substantially more serious health consequences and a reduced life expectancy. Obesity significantly impairs quality of life, and the risk of morbidity and mortality increases with the degree of obesity.² Those who are extremely obese often do not have sustained benefit from more conservative treatment. The benefits of nonsurgical treatment are significantly limited by the failure to maintain reduced body weight.

A large literature of controlled and uncontrolled cohort studies show that surgery has produced the longest period of sustained weight loss.³ A recent meta-analysis proved bariatric surgery not only efficacious for weight loss, but showed that a substantial majority of patients with diabetes, hyperlipidemia, hypertension, and obstructive sleep apnea experienced complete resolution or significant improvement of their comorbid condition after surgery.⁴

The possibility of significant adverse effects remains. The postoperative mortality rate for bariatric surgery is approximately 0.2%. Reoperation is required for up to 25% of patients within 5 years. Other complications are wound infection, staple failure, vitamin deficiency, diarrhea, and hemorrhage.³ The long-term health effects of bariatric surgery are not well known.

Recommendations from others

The NIH statement “Gastrointestinal Surgery for Severe Obesity” concluded that the benefits outweigh the risks and that surgical treatment is reasonable in those who strongly desire substantial weight loss and have life-threatening comorbid conditions.²

Clinical guidelines developed by the National Heart, Lung, and Blood Institute Expert Panel on the identification, evaluation, and treatment of obesity for adults recommend that bariatric surgery be an option for carefully selected patients with clinically severe obesity (BMI >40 or >35 with comorbid conditions) when less invasive methods of weight loss have failed and the patient is at high risk for obesity-associated morbidity and mortality.¹

The American Gastroenterological Association (AGA) medical position statement on obesity finds surgical therapy to be the most effective approach for achieving long-term weight loss. The AGA recommends surgery for patients with a BMI >40, or those with BMI >35 and 1 or more severe obesity-related medical complication (eg, hypertension, heart failure, or sleep apnea) if they have been unable to achieve or maintain weight loss with conventional therapy, have acceptable operative risks, and are able to comply with long-term treatment and follow-up.⁵

The American College of Preventive Medicine, in its policy statement on weight management counseling, recommends limiting surgical therapy for obesity to severely obese patients, defined as BMI >40.⁶

Evidence summary

Because of the nature of major surgery, there are practical and ethical barriers to true randomized controlled trials (RCTs) comparing bariatric surgery with placebo or to no intervention. However, multiple RCTs have compared the weight-reducing effects of different bariatric surgical techniques against each other.¹ All studies included patients who had a BMI >40 kg/m², or a BMI >35 kg/m² with at least 1 comorbidity, such as cardiovascular disease, sleep apnea, uncontrolled type 2 diabetes, or weight-induced physical problems interfering with performance of daily life activities. It is these study inclusion criteria that, by default, have become widely accepted indications for bariatric surgery. Weight loss in all RCTs was substantial, ranging from 50 to 100 kg over 6 months to 1 year. Comorbid factors associated with obesity showed either resolution or improvement after surgery in 91% of patients.

Patients with a BMI >40 have substantially more serious health consequences and a reduced life expectancy. Obesity significantly impairs quality of life, and the risk of morbidity and mortality increases with the degree of obesity.² Those who are extremely obese often do not have sustained benefit from more conservative treatment. The benefits of nonsurgical treatment are significantly limited by the failure to maintain reduced body weight.

A large literature of controlled and uncontrolled cohort studies show that surgery has produced the longest period of sustained weight loss.³ A recent meta-analysis proved bariatric surgery not only efficacious for weight loss, but showed that a substantial majority of patients with diabetes, hyperlipidemia, hypertension, and obstructive sleep apnea experienced complete resolution or significant improvement of their comorbid condition after surgery.⁴

The possibility of significant adverse effects remains. The postoperative mortality rate for bariatric surgery is approximately 0.2%. Reoperation is required for up to 25% of patients within 5 years. Other complications are wound infection, staple failure, vitamin deficiency, diarrhea, and hemorrhage.³ The long-term health effects of bariatric surgery are not well known.

Recommendations from others

The NIH statement “Gastrointestinal Surgery for Severe Obesity” concluded that the benefits outweigh the risks and that surgical treatment is reasonable in those who strongly desire substantial weight loss and have life-threatening comorbid conditions.²

Clinical guidelines developed by the National Heart, Lung, and Blood Institute Expert Panel on the identification, evaluation, and treatment of obesity for adults recommend that bariatric surgery be an option for carefully selected patients with clinically severe obesity (BMI >40 or >35 with comorbid conditions) when less invasive methods of weight loss have failed and the patient is at high risk for obesity-associated morbidity and mortality.¹

The American Gastroenterological Association (AGA) medical position statement on obesity finds surgical therapy to be the most effective approach for achieving long-term weight loss. The AGA recommends surgery for patients with a BMI >40, or those with BMI >35 and 1 or more severe obesity-related medical complication (eg, hypertension, heart failure, or sleep apnea) if they have been unable to achieve or maintain weight loss with conventional therapy, have acceptable operative risks, and are able to comply with long-term treatment and follow-up.⁵

The American College of Preventive Medicine, in its policy statement on weight management counseling, recommends limiting surgical therapy for obesity to severely obese patients, defined as BMI >40.⁶

Evidence summary

Onset of diabetes in childhood with ketoacidosis and insulin dependency has traditionally been sufficient to diagnose type 1 diabetes, while onset in older, obese patients with primary insulin resistance suggested type 2 diabetes. Unfortunately, features of type 1 and type 2 diabetes may be present in the same patient, making differentiation difficult. No diagnostic studies in the literature were identified that definitively demonstrate how to separate type 1 from type 2 diabetes.

A patient’s age may suggest, but does not reliably distinguish, diabetes types. A study of 569 new-onset type 1 and type 2 diabetic children and adolescents showed that older age was only weakly associated with type 2 diagnosis (odds ratio [OR]= 1.4 for each 1-year increment in age; 95% confidence interval [CI], 1.3–1.6).² In fact, newly diagnosed 12-year-old children have an equal incidence of type 1 as type 2 diabetes. Likewise, adults with type 2 phenotype (no initial insulin requirement) can present with positive autoantibodies typically found in younger type 1 patients. Older patients who fit this profile have been classified as type 1.5 diabetes or latent autoimmune disease in adults (LADA).³

A history of diabetic ketoacidosis (DKA) also does not reliably distinguish between types 1 and 2. A retrospective chart review gathered data on adults over 18 years of age who were admitted for DKA in a urban US hospital. Many patients with DKA were subsequently diagnosed with type 2 diabetes. Rates of type 2 diabetes in patients with DKA varied by race: 47% of Hispanics, 44% of African Americans, and 17% of Caucasians had type 2 diabetes.⁴

The overlapping presence of autoantibodies in both types of diabetes limits their use (TABLE). Autoantibodies do predict an earlier need for insulin. One prevalence study of 101 type 2 adult patients found 20% were positive for glutamic acid decarboxylase autoantibody (GADAb), which was positively associated with insulin dependence at 4 years postdiagnosis (OR=5.8; 95% CI, 1.8–18.9).⁵ Eighty percent of patients with autoantibodies required insulin compared with 41% of patients without autoantibodies. Another study in young adults with type 2 or unclassified diabetes from Sweden found 93% of patients who were GADAb+ required insulin at 3 years, compared with 51% who were GADAb–(OR=18.8; 95% CI, 1.8–191).⁶

One motivation to study autoantibody testing is a potential benefit in preserving pancreatic function. Kobayashi proposed treating those with autoantibody-positive diabetes (presumed type 1 or type 1.5) with insulin immediately, while initiating oral medications in those who test negative (presumed type 2 diabetes). This approach lacks significant patient-oriented outcome data, but his small RCT of 55 patients was encouraging. With a 3-year follow-up rate of 89%, early insulin use in GADAb+ patients preserved C-peptide levels and possibly prolonged pancreatic beta cell survival.⁷ Insulin dependency, defined as needing insulin for survival, occurred in 47% of controls (who received oral sulfonylureas) and only 13% of patients receiving insulin (number needed to treat [NNT]= 4; P=.043).⁷ The theoretical benefit is that if beta cell exhaustion can be delayed, endogenous insulin production could be maintained to assist prevention of damaging postprandial glucose spikes.

Although daily variation in serum insulin levels limits its use, C-peptide levels show more promise. Random C-peptide levels were superior to fasting or glucagon stimulated levels in 1093 patients, who were followed for 3 years to confirm insulin requirements. Using a receiver operating characteristic (ROC) curve, the area under the curve for random C-peptide levels to distinguish diabetes types was 0.98 (95% CI, 0.97–0.99).⁸ For patients under the optimal cutoff of 0.5 nmol/L, the positive predictive value was 96% for diagnosing type 1 and the likelihood ratio was 22.5.

Finally, the ratio of adiponectin to leptin hormone may show diagnostic merit. Adipocytes secrete adiponectin which acts as an insulin sensitizer, antiatherogenic and anti-inflammatory agent. Obesity and type 2 phenotype correlate with lower levels of adiponectin, but are associated with higher levels of leptin hormone, another molecule secreted by adipocytes. A recent case-control study of children aged 6 to 21 years analyzed adiponectin and leptin hormone levels in patients with classical type 1 and 2 diabetes, as determined by 2 pediatric endocrinologists; interestingly, 29% of the type 1 patients were autoantibody negative.⁹ After plotting a ROC curve, they found the area under the curve was 0.97 (95% CI, 0.93–1.00). At an adiponectin-to-leptin ratio cutoff less than 0.7, they found the sensitivity to diagnose type 2 was 88% (95% CI, 64–99%), the specificity was 90% (95% CI, 77–97), and the likelihood ratio for a positive test was 8.8.⁹

TABLE 1
Antibody markers and diabetes type

Recommendations from others

The National Academy of Clinical Biochemistry and the American Association of Clinical Endocrinologists recommend against routine testing of insulin, C-peptide, autoantibodies and genetic markers.^1,10 Guidelines from the American Diabetes Association admit that many diabetic individuals do not easily fit into a distinct diagnostic category; however, they only provide criteria for the general diagnosis of diabetes, not specific criteria to distinguish type 1 from type 2.¹¹

Acknowledgments

The opinions and assertions contained herein are the private views of the author and are not to be construed as official, or as reflecting the views of the US Air Force medical department or the US Air Force at large.

Evidence summary

Onset of diabetes in childhood with ketoacidosis and insulin dependency has traditionally been sufficient to diagnose type 1 diabetes, while onset in older, obese patients with primary insulin resistance suggested type 2 diabetes. Unfortunately, features of type 1 and type 2 diabetes may be present in the same patient, making differentiation difficult. No diagnostic studies in the literature were identified that definitively demonstrate how to separate type 1 from type 2 diabetes.

A patient’s age may suggest, but does not reliably distinguish, diabetes types. A study of 569 new-onset type 1 and type 2 diabetic children and adolescents showed that older age was only weakly associated with type 2 diagnosis (odds ratio [OR]= 1.4 for each 1-year increment in age; 95% confidence interval [CI], 1.3–1.6).² In fact, newly diagnosed 12-year-old children have an equal incidence of type 1 as type 2 diabetes. Likewise, adults with type 2 phenotype (no initial insulin requirement) can present with positive autoantibodies typically found in younger type 1 patients. Older patients who fit this profile have been classified as type 1.5 diabetes or latent autoimmune disease in adults (LADA).³

A history of diabetic ketoacidosis (DKA) also does not reliably distinguish between types 1 and 2. A retrospective chart review gathered data on adults over 18 years of age who were admitted for DKA in a urban US hospital. Many patients with DKA were subsequently diagnosed with type 2 diabetes. Rates of type 2 diabetes in patients with DKA varied by race: 47% of Hispanics, 44% of African Americans, and 17% of Caucasians had type 2 diabetes.⁴

The overlapping presence of autoantibodies in both types of diabetes limits their use (TABLE). Autoantibodies do predict an earlier need for insulin. One prevalence study of 101 type 2 adult patients found 20% were positive for glutamic acid decarboxylase autoantibody (GADAb), which was positively associated with insulin dependence at 4 years postdiagnosis (OR=5.8; 95% CI, 1.8–18.9).⁵ Eighty percent of patients with autoantibodies required insulin compared with 41% of patients without autoantibodies. Another study in young adults with type 2 or unclassified diabetes from Sweden found 93% of patients who were GADAb+ required insulin at 3 years, compared with 51% who were GADAb–(OR=18.8; 95% CI, 1.8–191).⁶

One motivation to study autoantibody testing is a potential benefit in preserving pancreatic function. Kobayashi proposed treating those with autoantibody-positive diabetes (presumed type 1 or type 1.5) with insulin immediately, while initiating oral medications in those who test negative (presumed type 2 diabetes). This approach lacks significant patient-oriented outcome data, but his small RCT of 55 patients was encouraging. With a 3-year follow-up rate of 89%, early insulin use in GADAb+ patients preserved C-peptide levels and possibly prolonged pancreatic beta cell survival.⁷ Insulin dependency, defined as needing insulin for survival, occurred in 47% of controls (who received oral sulfonylureas) and only 13% of patients receiving insulin (number needed to treat [NNT]= 4; P=.043).⁷ The theoretical benefit is that if beta cell exhaustion can be delayed, endogenous insulin production could be maintained to assist prevention of damaging postprandial glucose spikes.

Although daily variation in serum insulin levels limits its use, C-peptide levels show more promise. Random C-peptide levels were superior to fasting or glucagon stimulated levels in 1093 patients, who were followed for 3 years to confirm insulin requirements. Using a receiver operating characteristic (ROC) curve, the area under the curve for random C-peptide levels to distinguish diabetes types was 0.98 (95% CI, 0.97–0.99).⁸ For patients under the optimal cutoff of 0.5 nmol/L, the positive predictive value was 96% for diagnosing type 1 and the likelihood ratio was 22.5.

Finally, the ratio of adiponectin to leptin hormone may show diagnostic merit. Adipocytes secrete adiponectin which acts as an insulin sensitizer, antiatherogenic and anti-inflammatory agent. Obesity and type 2 phenotype correlate with lower levels of adiponectin, but are associated with higher levels of leptin hormone, another molecule secreted by adipocytes. A recent case-control study of children aged 6 to 21 years analyzed adiponectin and leptin hormone levels in patients with classical type 1 and 2 diabetes, as determined by 2 pediatric endocrinologists; interestingly, 29% of the type 1 patients were autoantibody negative.⁹ After plotting a ROC curve, they found the area under the curve was 0.97 (95% CI, 0.93–1.00). At an adiponectin-to-leptin ratio cutoff less than 0.7, they found the sensitivity to diagnose type 2 was 88% (95% CI, 64–99%), the specificity was 90% (95% CI, 77–97), and the likelihood ratio for a positive test was 8.8.⁹

TABLE 1
Antibody markers and diabetes type

Recommendations from others

The National Academy of Clinical Biochemistry and the American Association of Clinical Endocrinologists recommend against routine testing of insulin, C-peptide, autoantibodies and genetic markers.^1,10 Guidelines from the American Diabetes Association admit that many diabetic individuals do not easily fit into a distinct diagnostic category; however, they only provide criteria for the general diagnosis of diabetes, not specific criteria to distinguish type 1 from type 2.¹¹

Acknowledgments

The opinions and assertions contained herein are the private views of the author and are not to be construed as official, or as reflecting the views of the US Air Force medical department or the US Air Force at large.

Evidence summary

Onset of diabetes in childhood with ketoacidosis and insulin dependency has traditionally been sufficient to diagnose type 1 diabetes, while onset in older, obese patients with primary insulin resistance suggested type 2 diabetes. Unfortunately, features of type 1 and type 2 diabetes may be present in the same patient, making differentiation difficult. No diagnostic studies in the literature were identified that definitively demonstrate how to separate type 1 from type 2 diabetes.

A patient’s age may suggest, but does not reliably distinguish, diabetes types. A study of 569 new-onset type 1 and type 2 diabetic children and adolescents showed that older age was only weakly associated with type 2 diagnosis (odds ratio [OR]= 1.4 for each 1-year increment in age; 95% confidence interval [CI], 1.3–1.6).² In fact, newly diagnosed 12-year-old children have an equal incidence of type 1 as type 2 diabetes. Likewise, adults with type 2 phenotype (no initial insulin requirement) can present with positive autoantibodies typically found in younger type 1 patients. Older patients who fit this profile have been classified as type 1.5 diabetes or latent autoimmune disease in adults (LADA).³

A history of diabetic ketoacidosis (DKA) also does not reliably distinguish between types 1 and 2. A retrospective chart review gathered data on adults over 18 years of age who were admitted for DKA in a urban US hospital. Many patients with DKA were subsequently diagnosed with type 2 diabetes. Rates of type 2 diabetes in patients with DKA varied by race: 47% of Hispanics, 44% of African Americans, and 17% of Caucasians had type 2 diabetes.⁴

The overlapping presence of autoantibodies in both types of diabetes limits their use (TABLE). Autoantibodies do predict an earlier need for insulin. One prevalence study of 101 type 2 adult patients found 20% were positive for glutamic acid decarboxylase autoantibody (GADAb), which was positively associated with insulin dependence at 4 years postdiagnosis (OR=5.8; 95% CI, 1.8–18.9).⁵ Eighty percent of patients with autoantibodies required insulin compared with 41% of patients without autoantibodies. Another study in young adults with type 2 or unclassified diabetes from Sweden found 93% of patients who were GADAb+ required insulin at 3 years, compared with 51% who were GADAb–(OR=18.8; 95% CI, 1.8–191).⁶

One motivation to study autoantibody testing is a potential benefit in preserving pancreatic function. Kobayashi proposed treating those with autoantibody-positive diabetes (presumed type 1 or type 1.5) with insulin immediately, while initiating oral medications in those who test negative (presumed type 2 diabetes). This approach lacks significant patient-oriented outcome data, but his small RCT of 55 patients was encouraging. With a 3-year follow-up rate of 89%, early insulin use in GADAb+ patients preserved C-peptide levels and possibly prolonged pancreatic beta cell survival.⁷ Insulin dependency, defined as needing insulin for survival, occurred in 47% of controls (who received oral sulfonylureas) and only 13% of patients receiving insulin (number needed to treat [NNT]= 4; P=.043).⁷ The theoretical benefit is that if beta cell exhaustion can be delayed, endogenous insulin production could be maintained to assist prevention of damaging postprandial glucose spikes.

Although daily variation in serum insulin levels limits its use, C-peptide levels show more promise. Random C-peptide levels were superior to fasting or glucagon stimulated levels in 1093 patients, who were followed for 3 years to confirm insulin requirements. Using a receiver operating characteristic (ROC) curve, the area under the curve for random C-peptide levels to distinguish diabetes types was 0.98 (95% CI, 0.97–0.99).⁸ For patients under the optimal cutoff of 0.5 nmol/L, the positive predictive value was 96% for diagnosing type 1 and the likelihood ratio was 22.5.

Finally, the ratio of adiponectin to leptin hormone may show diagnostic merit. Adipocytes secrete adiponectin which acts as an insulin sensitizer, antiatherogenic and anti-inflammatory agent. Obesity and type 2 phenotype correlate with lower levels of adiponectin, but are associated with higher levels of leptin hormone, another molecule secreted by adipocytes. A recent case-control study of children aged 6 to 21 years analyzed adiponectin and leptin hormone levels in patients with classical type 1 and 2 diabetes, as determined by 2 pediatric endocrinologists; interestingly, 29% of the type 1 patients were autoantibody negative.⁹ After plotting a ROC curve, they found the area under the curve was 0.97 (95% CI, 0.93–1.00). At an adiponectin-to-leptin ratio cutoff less than 0.7, they found the sensitivity to diagnose type 2 was 88% (95% CI, 64–99%), the specificity was 90% (95% CI, 77–97), and the likelihood ratio for a positive test was 8.8.⁹

TABLE 1
Antibody markers and diabetes type

Recommendations from others

The National Academy of Clinical Biochemistry and the American Association of Clinical Endocrinologists recommend against routine testing of insulin, C-peptide, autoantibodies and genetic markers.^1,10 Guidelines from the American Diabetes Association admit that many diabetic individuals do not easily fit into a distinct diagnostic category; however, they only provide criteria for the general diagnosis of diabetes, not specific criteria to distinguish type 1 from type 2.¹¹

Acknowledgments

The opinions and assertions contained herein are the private views of the author and are not to be construed as official, or as reflecting the views of the US Air Force medical department or the US Air Force at large.

Evidence summary

A recent Cochrane systematic review summarizes the clinical effects of digoxin when used for patients with heart failure in normal sinus rhythm. Thirteen studies including 7896 participants, most of whom had systolic dysfunction, met the criteria for inclusion. Ninety-four percent of all study participants came from a single large randomized placebo-controlled trial.³ Because the studies did not all measure the same outcomes, subgroup analyses were performed.

Four studies with 1096 participants contributed to the findings on clinical status, 12 studies with 7262 participants contributed to the findings of hospitalization and 8 studies including 7755 patients contributed to the data on mortality. Patients receiving digoxin experienced reduced rates of hospitalization due to worsening heart failure (odds ratio [OR]=0.68; 95% confidence interval [CI], 0.61–0.75; number needed to treat [NNT]=13–17) and less clinical deterioration (OR=0.31; 95% CI, 0.21–0.43; NNT=3–61). The wide range in NNT for the reduction in clinical deterioration reflects varying baseline rates of worsening clinical status found among the 12 studies for patients receiving placebo. The narrow CI associated with the odds ratio for reduced rates of clinical deterioration reflects the fact that the majority of patients whose clinical status was evaluated as an outcome came from a single large study, the DIG trial.³ This trial followed 6800 patients with NYHA classifications I to III. Ninety-four percent of patients in this trial were additionally on angiotensin-converting enzyme (ACE) inhibitors and 82% were taking diuretics. Patients were followed for a mean of 37 months.

A subgroup analysis of 988 patients with diastolic dysfunction (ejection fraction >45%) in this study³ suggested no clear benefits or harms when digoxin was used in combination with other therapies vs placebo; however, it did show a positive trend towards the combined outcome of reduced hospitalizations and less clinical deterioration (relative risk [RR]=0.82; 95% CI, 0.63–1.07). Increased rates of supraventricular dysrhythmias (RR=2.08; 95% CI, 1.44–2.99; number needed to harm [NNH]=77) and second- and third-degree heart block were demonstrated for patients receiving digoxin (RR=2.93; 95% CI, 1.61–5.34; NNH=125). There was no difference in mortality between patients receiving digoxin or those receiving placebo (OR=0.98; 95% CI, 0.89–1.09).¹

A post-hoc subgroup analysis focusing only on sex-based differences in the DIG trial suggested women benefit less than men from reduced hospitalizations: –4.2% (95% CI, –8.9 to 0.5) vs –8.9% (95% CI, –11.4 to –6.5) (P=.053).² When a multivariable analysis was performed, digoxin use for women was associated with a higher risk of mortality (adjusted hazard ratio vs placebo=1.23; 95% CI, 1.02–1.47).²

Two randomized controlled withdrawal studies, in which patients who were being treated with digoxin had it discontinued, were also included in the systematic review. These patients’ clinical outcomes were then compared with persons who had continued to receive digoxin for the duration of the trial. Six parallel design studies, in which patients taking digoxin underwent a washout period before being randomized to either digoxin or placebo, were also included in the evaluation of digoxin’s effect on clinical status. Because these patients had already demonstrated the ability to tolerate digoxin, these studies may have been biased in favor of digoxin.^4,5

Recommendations from others

The American College of Cardiology/ American Heart Association⁶ and Heart Failure Society of America⁷ guidelines both recommend that digoxin be used in NYHA class II–III patients in sinus rhythm who remain symptomatic on standard therapy (described as ACE inhibitors, diuretics, and beta-blockers). Guidelines from the Scottish Intercollegiate Society,⁸ the European Society of Cardiology,⁹ and the American Medical Directors Association¹⁰ all offer similar recommendations.

Evidence summary

A recent Cochrane systematic review summarizes the clinical effects of digoxin when used for patients with heart failure in normal sinus rhythm. Thirteen studies including 7896 participants, most of whom had systolic dysfunction, met the criteria for inclusion. Ninety-four percent of all study participants came from a single large randomized placebo-controlled trial.³ Because the studies did not all measure the same outcomes, subgroup analyses were performed.

Four studies with 1096 participants contributed to the findings on clinical status, 12 studies with 7262 participants contributed to the findings of hospitalization and 8 studies including 7755 patients contributed to the data on mortality. Patients receiving digoxin experienced reduced rates of hospitalization due to worsening heart failure (odds ratio [OR]=0.68; 95% confidence interval [CI], 0.61–0.75; number needed to treat [NNT]=13–17) and less clinical deterioration (OR=0.31; 95% CI, 0.21–0.43; NNT=3–61). The wide range in NNT for the reduction in clinical deterioration reflects varying baseline rates of worsening clinical status found among the 12 studies for patients receiving placebo. The narrow CI associated with the odds ratio for reduced rates of clinical deterioration reflects the fact that the majority of patients whose clinical status was evaluated as an outcome came from a single large study, the DIG trial.³ This trial followed 6800 patients with NYHA classifications I to III. Ninety-four percent of patients in this trial were additionally on angiotensin-converting enzyme (ACE) inhibitors and 82% were taking diuretics. Patients were followed for a mean of 37 months.

A subgroup analysis of 988 patients with diastolic dysfunction (ejection fraction >45%) in this study³ suggested no clear benefits or harms when digoxin was used in combination with other therapies vs placebo; however, it did show a positive trend towards the combined outcome of reduced hospitalizations and less clinical deterioration (relative risk [RR]=0.82; 95% CI, 0.63–1.07). Increased rates of supraventricular dysrhythmias (RR=2.08; 95% CI, 1.44–2.99; number needed to harm [NNH]=77) and second- and third-degree heart block were demonstrated for patients receiving digoxin (RR=2.93; 95% CI, 1.61–5.34; NNH=125). There was no difference in mortality between patients receiving digoxin or those receiving placebo (OR=0.98; 95% CI, 0.89–1.09).¹

A post-hoc subgroup analysis focusing only on sex-based differences in the DIG trial suggested women benefit less than men from reduced hospitalizations: –4.2% (95% CI, –8.9 to 0.5) vs –8.9% (95% CI, –11.4 to –6.5) (P=.053).² When a multivariable analysis was performed, digoxin use for women was associated with a higher risk of mortality (adjusted hazard ratio vs placebo=1.23; 95% CI, 1.02–1.47).²

Two randomized controlled withdrawal studies, in which patients who were being treated with digoxin had it discontinued, were also included in the systematic review. These patients’ clinical outcomes were then compared with persons who had continued to receive digoxin for the duration of the trial. Six parallel design studies, in which patients taking digoxin underwent a washout period before being randomized to either digoxin or placebo, were also included in the evaluation of digoxin’s effect on clinical status. Because these patients had already demonstrated the ability to tolerate digoxin, these studies may have been biased in favor of digoxin.^4,5

Recommendations from others

The American College of Cardiology/ American Heart Association⁶ and Heart Failure Society of America⁷ guidelines both recommend that digoxin be used in NYHA class II–III patients in sinus rhythm who remain symptomatic on standard therapy (described as ACE inhibitors, diuretics, and beta-blockers). Guidelines from the Scottish Intercollegiate Society,⁸ the European Society of Cardiology,⁹ and the American Medical Directors Association¹⁰ all offer similar recommendations.

Evidence summary

A recent Cochrane systematic review summarizes the clinical effects of digoxin when used for patients with heart failure in normal sinus rhythm. Thirteen studies including 7896 participants, most of whom had systolic dysfunction, met the criteria for inclusion. Ninety-four percent of all study participants came from a single large randomized placebo-controlled trial.³ Because the studies did not all measure the same outcomes, subgroup analyses were performed.

Four studies with 1096 participants contributed to the findings on clinical status, 12 studies with 7262 participants contributed to the findings of hospitalization and 8 studies including 7755 patients contributed to the data on mortality. Patients receiving digoxin experienced reduced rates of hospitalization due to worsening heart failure (odds ratio [OR]=0.68; 95% confidence interval [CI], 0.61–0.75; number needed to treat [NNT]=13–17) and less clinical deterioration (OR=0.31; 95% CI, 0.21–0.43; NNT=3–61). The wide range in NNT for the reduction in clinical deterioration reflects varying baseline rates of worsening clinical status found among the 12 studies for patients receiving placebo. The narrow CI associated with the odds ratio for reduced rates of clinical deterioration reflects the fact that the majority of patients whose clinical status was evaluated as an outcome came from a single large study, the DIG trial.³ This trial followed 6800 patients with NYHA classifications I to III. Ninety-four percent of patients in this trial were additionally on angiotensin-converting enzyme (ACE) inhibitors and 82% were taking diuretics. Patients were followed for a mean of 37 months.

A subgroup analysis of 988 patients with diastolic dysfunction (ejection fraction >45%) in this study³ suggested no clear benefits or harms when digoxin was used in combination with other therapies vs placebo; however, it did show a positive trend towards the combined outcome of reduced hospitalizations and less clinical deterioration (relative risk [RR]=0.82; 95% CI, 0.63–1.07). Increased rates of supraventricular dysrhythmias (RR=2.08; 95% CI, 1.44–2.99; number needed to harm [NNH]=77) and second- and third-degree heart block were demonstrated for patients receiving digoxin (RR=2.93; 95% CI, 1.61–5.34; NNH=125). There was no difference in mortality between patients receiving digoxin or those receiving placebo (OR=0.98; 95% CI, 0.89–1.09).¹

A post-hoc subgroup analysis focusing only on sex-based differences in the DIG trial suggested women benefit less than men from reduced hospitalizations: –4.2% (95% CI, –8.9 to 0.5) vs –8.9% (95% CI, –11.4 to –6.5) (P=.053).² When a multivariable analysis was performed, digoxin use for women was associated with a higher risk of mortality (adjusted hazard ratio vs placebo=1.23; 95% CI, 1.02–1.47).²

Two randomized controlled withdrawal studies, in which patients who were being treated with digoxin had it discontinued, were also included in the systematic review. These patients’ clinical outcomes were then compared with persons who had continued to receive digoxin for the duration of the trial. Six parallel design studies, in which patients taking digoxin underwent a washout period before being randomized to either digoxin or placebo, were also included in the evaluation of digoxin’s effect on clinical status. Because these patients had already demonstrated the ability to tolerate digoxin, these studies may have been biased in favor of digoxin.^4,5

Recommendations from others

The American College of Cardiology/ American Heart Association⁶ and Heart Failure Society of America⁷ guidelines both recommend that digoxin be used in NYHA class II–III patients in sinus rhythm who remain symptomatic on standard therapy (described as ACE inhibitors, diuretics, and beta-blockers). Guidelines from the Scottish Intercollegiate Society,⁸ the European Society of Cardiology,⁹ and the American Medical Directors Association¹⁰ all offer similar recommendations.

Evidence summary

Multiple studies and reviews have compared tissue adhesives with sutures or adhesive strips for wound closure. A Cochrane review found 10 studies, which included 970 patients in the emergency-room setting. Review of these articles found no significant difference in cosmetic appearance between tissue adhesive closure and standard suture closure with a 3-month follow-up period in acute, linear wounds under low tension. Wound erythema (number needed to treat [NNT]=10) and dehiscence rates (NNT=25) were lower for tissue adhesives.¹ In the 6 studies that reported time data, treatment with tissue adhesive took 4.7 fewer minutes. In all 6 studies that reported patients’ perception of pain, pain was significantly less with tissue adhesive (weighted mean difference=13.7 mm [on 100-mm scale]; 95% confidence interval [CI], –20.0 to –6.9).

A multicenter, randomized trial studied 924 wounds (383 traumatic, 541 surgical) and reported no difference in cosmetic appearance upon grading by both a clinician and the patients themselves.² This study was not included in the Cochrane review because of the inclusion of surgical wounds. In a clinical trial reported after the Cochrane review, Holger and colleagues³ studied tissue adhesives against standard wound closure using either nylon or absorbable gut sutures. The study included 145 patients, 84 of whom had at least a 9-month follow-up. No significant difference was noted in a visual analog grading scale, with a 10- to 15-mm difference (out of 100 mm) considered significant.³ Tissue adhesive closure was, on average, 5.7 minutes faster than standard wound closure with sutures for superficial lacerations. Pain outcomes in the studies showed that closure with tissue adhesive was less painful due to the lack of a need for anesthesia.⁴

Recommendations from others

No major guidelines were found regarding the use of skin adhesives for wound closure.

Evidence summary

Multiple studies and reviews have compared tissue adhesives with sutures or adhesive strips for wound closure. A Cochrane review found 10 studies, which included 970 patients in the emergency-room setting. Review of these articles found no significant difference in cosmetic appearance between tissue adhesive closure and standard suture closure with a 3-month follow-up period in acute, linear wounds under low tension. Wound erythema (number needed to treat [NNT]=10) and dehiscence rates (NNT=25) were lower for tissue adhesives.¹ In the 6 studies that reported time data, treatment with tissue adhesive took 4.7 fewer minutes. In all 6 studies that reported patients’ perception of pain, pain was significantly less with tissue adhesive (weighted mean difference=13.7 mm [on 100-mm scale]; 95% confidence interval [CI], –20.0 to –6.9).

A multicenter, randomized trial studied 924 wounds (383 traumatic, 541 surgical) and reported no difference in cosmetic appearance upon grading by both a clinician and the patients themselves.² This study was not included in the Cochrane review because of the inclusion of surgical wounds. In a clinical trial reported after the Cochrane review, Holger and colleagues³ studied tissue adhesives against standard wound closure using either nylon or absorbable gut sutures. The study included 145 patients, 84 of whom had at least a 9-month follow-up. No significant difference was noted in a visual analog grading scale, with a 10- to 15-mm difference (out of 100 mm) considered significant.³ Tissue adhesive closure was, on average, 5.7 minutes faster than standard wound closure with sutures for superficial lacerations. Pain outcomes in the studies showed that closure with tissue adhesive was less painful due to the lack of a need for anesthesia.⁴

Recommendations from others

No major guidelines were found regarding the use of skin adhesives for wound closure.

Evidence summary

Multiple studies and reviews have compared tissue adhesives with sutures or adhesive strips for wound closure. A Cochrane review found 10 studies, which included 970 patients in the emergency-room setting. Review of these articles found no significant difference in cosmetic appearance between tissue adhesive closure and standard suture closure with a 3-month follow-up period in acute, linear wounds under low tension. Wound erythema (number needed to treat [NNT]=10) and dehiscence rates (NNT=25) were lower for tissue adhesives.¹ In the 6 studies that reported time data, treatment with tissue adhesive took 4.7 fewer minutes. In all 6 studies that reported patients’ perception of pain, pain was significantly less with tissue adhesive (weighted mean difference=13.7 mm [on 100-mm scale]; 95% confidence interval [CI], –20.0 to –6.9).

A multicenter, randomized trial studied 924 wounds (383 traumatic, 541 surgical) and reported no difference in cosmetic appearance upon grading by both a clinician and the patients themselves.² This study was not included in the Cochrane review because of the inclusion of surgical wounds. In a clinical trial reported after the Cochrane review, Holger and colleagues³ studied tissue adhesives against standard wound closure using either nylon or absorbable gut sutures. The study included 145 patients, 84 of whom had at least a 9-month follow-up. No significant difference was noted in a visual analog grading scale, with a 10- to 15-mm difference (out of 100 mm) considered significant.³ Tissue adhesive closure was, on average, 5.7 minutes faster than standard wound closure with sutures for superficial lacerations. Pain outcomes in the studies showed that closure with tissue adhesive was less painful due to the lack of a need for anesthesia.⁴

Recommendations from others

No major guidelines were found regarding the use of skin adhesives for wound closure.

Evidence summary

A 1998 meta-analysis of 35 exploratory cohort studies published between 1966 and 1996 included a total of 5892 women with postmenopausal bleeding.¹ TVUS evaluations were followed by endometrial tissue sampling and results were compared. Using endometrial thickness of ≤5 mm as the threshold, ultrasound was very accurate at ruling out patients with endometrial cancer but only fair at diagnosing cancer (likelihood ratio for a positive test [LR+]=2.5; LR for a negative test [LR–]=0.06). In addition, the 5-mm threshold was accurate at ruling out any endometrial abnormality (cancer, polyp, atypical hyperplasia: LR– = 0.01). The authors suggested that TVUS can reliably rule out significant endometrial disease among postmenopausal women with vaginal bleeding.

A 2002 meta-analysis of 57 cohort studies, without consistently applied reference standards, published between 1966 and 2000 included a total of 9031 women with postmenopausal bleeding.² Because many of the studies were felt to use inadequately stringent criteria for diagnosis, the authors limited their final analysis to only 4 studies. They concluded that a negative result using a 5-mm threshold rules out endometrial pathology with fair certainty (LR– = 0.21).

Recommendations from others

A Consensus Conference Statement from the Society of Radiologists in Ultrasound recommended that either TVUS or endometrial biopsy could be used in the initial evaluation of patients with postmenopausal bleeding.³ Using a threshold of >5 mm as abnormal, they concluded that the sensitivities of TVUS and endometrial biopsy are comparable when “sufficient tissue” is obtained with endometrial biopsy. They felt that data was currently insufficient to clearly state which technique is more effective.

TABLE
Differential diagnosis of postmenopausal bleeding

Evidence summary

A 1998 meta-analysis of 35 exploratory cohort studies published between 1966 and 1996 included a total of 5892 women with postmenopausal bleeding.¹ TVUS evaluations were followed by endometrial tissue sampling and results were compared. Using endometrial thickness of ≤5 mm as the threshold, ultrasound was very accurate at ruling out patients with endometrial cancer but only fair at diagnosing cancer (likelihood ratio for a positive test [LR+]=2.5; LR for a negative test [LR–]=0.06). In addition, the 5-mm threshold was accurate at ruling out any endometrial abnormality (cancer, polyp, atypical hyperplasia: LR– = 0.01). The authors suggested that TVUS can reliably rule out significant endometrial disease among postmenopausal women with vaginal bleeding.

A 2002 meta-analysis of 57 cohort studies, without consistently applied reference standards, published between 1966 and 2000 included a total of 9031 women with postmenopausal bleeding.² Because many of the studies were felt to use inadequately stringent criteria for diagnosis, the authors limited their final analysis to only 4 studies. They concluded that a negative result using a 5-mm threshold rules out endometrial pathology with fair certainty (LR– = 0.21).

Recommendations from others

A Consensus Conference Statement from the Society of Radiologists in Ultrasound recommended that either TVUS or endometrial biopsy could be used in the initial evaluation of patients with postmenopausal bleeding.³ Using a threshold of >5 mm as abnormal, they concluded that the sensitivities of TVUS and endometrial biopsy are comparable when “sufficient tissue” is obtained with endometrial biopsy. They felt that data was currently insufficient to clearly state which technique is more effective.

TABLE
Differential diagnosis of postmenopausal bleeding

Evidence summary

A 1998 meta-analysis of 35 exploratory cohort studies published between 1966 and 1996 included a total of 5892 women with postmenopausal bleeding.¹ TVUS evaluations were followed by endometrial tissue sampling and results were compared. Using endometrial thickness of ≤5 mm as the threshold, ultrasound was very accurate at ruling out patients with endometrial cancer but only fair at diagnosing cancer (likelihood ratio for a positive test [LR+]=2.5; LR for a negative test [LR–]=0.06). In addition, the 5-mm threshold was accurate at ruling out any endometrial abnormality (cancer, polyp, atypical hyperplasia: LR– = 0.01). The authors suggested that TVUS can reliably rule out significant endometrial disease among postmenopausal women with vaginal bleeding.

A 2002 meta-analysis of 57 cohort studies, without consistently applied reference standards, published between 1966 and 2000 included a total of 9031 women with postmenopausal bleeding.² Because many of the studies were felt to use inadequately stringent criteria for diagnosis, the authors limited their final analysis to only 4 studies. They concluded that a negative result using a 5-mm threshold rules out endometrial pathology with fair certainty (LR– = 0.21).

Recommendations from others

A Consensus Conference Statement from the Society of Radiologists in Ultrasound recommended that either TVUS or endometrial biopsy could be used in the initial evaluation of patients with postmenopausal bleeding.³ Using a threshold of >5 mm as abnormal, they concluded that the sensitivities of TVUS and endometrial biopsy are comparable when “sufficient tissue” is obtained with endometrial biopsy. They felt that data was currently insufficient to clearly state which technique is more effective.

TABLE
Differential diagnosis of postmenopausal bleeding

Evidence summary

A systematic review of antibiotic prophylaxis for recurrent AOM examined 9 RCTs with a total of 958 children. Recurrent AOM was defined as 3 or more episodes per 6 to 18 months. The studies were low to moderate in quality (mean methodologic quality score of 11.8 out of 29 possible points). The most commonly used antibiotics were amoxicillin, cotrimoxazole, and sulfamethoxazole, given for 3 to 24 months (dosing not reported).

Children taking antibiotics had 0.11 (95% confidence interval [CI], 0.03–0.19) fewer episodes of recurrent AOM per patient-month than those taking placebo. The rate in the control group was 0.19 (95% CI, 0.13–0.26). Nine children would have to be treated per month to prevent 1 ear infection (NNT=9; 95% CI, 5–33). Only 2 of the 9 studies had statistically significant results; both used sulfisoxazole for 10 to 12 weeks and were of similar methodologic quality (12.5 out 29 points).

A trend towards a better outcome in studies that used sulfisoxazole did not reach significance compared with those using other medications (ie, ampicillin, amoxicillin, cotrimoxazole). Shorter treatment intervals (<6 months) trended toward being more effective than longer intervals, but this also did not reach significance. Children with more frequent episodes of AOM did no better than those with less frequent episodes.¹

Since that review was published, another study of prophylaxis for ear infections had been published. This randomized, double blind, placebo-controlled study enrolled 194 children aged 3 months to 6 years with at least 3 documented AOM episodes in the preceding 6 months. The children were given amoxicillin (20 mg/kg/d) either once daily (n=55) or divided twice daily (n=44) or placebo (n=59). Excluding 36 noncompliant subjects, the percentages without a recurrent episode were 63% for the placebo group, 64% for the once-daily amoxicillin group, and 61% for the twice-daily amoxicillin group. There was no significant difference in the incidence of new AOM episodes among the children in the 3 groups.²

A review article states: “Many children with acute otitis media do not benefit from antimicrobial therapy because the cause of their illness is not bacterial or the infection is cleared by the immune system without use of a drug. At present, we do not have clinical criteria for distinguishing which children are in need of antibiotic therapy for AOM.”³ The lack of criteria for determining which children need antibiotic therapy for AOM makes it more difficult to select children for antibiotic prophylaxis against recurrent AOM.

Recommendations from others

The American Academy of Pediatrics and the American Academy of Family Physicians do not address antibiotic prophylaxis for recurrent episodes of otitis media in their guidelines. Both groups recommend modification of risk factors to decrease recurrent AOM, including promoting breastfeeding during the first 6 months, avoiding bottle-propping, reducing or eliminating pacifier use in the second 6 months of life, and eliminating exposure to secondhand smoke.

They also recommend pneumococcal conjugate vaccine to reduce vaccine-serotype pneumococcal otitis and live-attenuated influenza vaccine during respiratory virus season for children aged >2 years.

Evidence summary

A systematic review of antibiotic prophylaxis for recurrent AOM examined 9 RCTs with a total of 958 children. Recurrent AOM was defined as 3 or more episodes per 6 to 18 months. The studies were low to moderate in quality (mean methodologic quality score of 11.8 out of 29 possible points). The most commonly used antibiotics were amoxicillin, cotrimoxazole, and sulfamethoxazole, given for 3 to 24 months (dosing not reported).

Children taking antibiotics had 0.11 (95% confidence interval [CI], 0.03–0.19) fewer episodes of recurrent AOM per patient-month than those taking placebo. The rate in the control group was 0.19 (95% CI, 0.13–0.26). Nine children would have to be treated per month to prevent 1 ear infection (NNT=9; 95% CI, 5–33). Only 2 of the 9 studies had statistically significant results; both used sulfisoxazole for 10 to 12 weeks and were of similar methodologic quality (12.5 out 29 points).

A trend towards a better outcome in studies that used sulfisoxazole did not reach significance compared with those using other medications (ie, ampicillin, amoxicillin, cotrimoxazole). Shorter treatment intervals (<6 months) trended toward being more effective than longer intervals, but this also did not reach significance. Children with more frequent episodes of AOM did no better than those with less frequent episodes.¹

Since that review was published, another study of prophylaxis for ear infections had been published. This randomized, double blind, placebo-controlled study enrolled 194 children aged 3 months to 6 years with at least 3 documented AOM episodes in the preceding 6 months. The children were given amoxicillin (20 mg/kg/d) either once daily (n=55) or divided twice daily (n=44) or placebo (n=59). Excluding 36 noncompliant subjects, the percentages without a recurrent episode were 63% for the placebo group, 64% for the once-daily amoxicillin group, and 61% for the twice-daily amoxicillin group. There was no significant difference in the incidence of new AOM episodes among the children in the 3 groups.²

A review article states: “Many children with acute otitis media do not benefit from antimicrobial therapy because the cause of their illness is not bacterial or the infection is cleared by the immune system without use of a drug. At present, we do not have clinical criteria for distinguishing which children are in need of antibiotic therapy for AOM.”³ The lack of criteria for determining which children need antibiotic therapy for AOM makes it more difficult to select children for antibiotic prophylaxis against recurrent AOM.

Recommendations from others

The American Academy of Pediatrics and the American Academy of Family Physicians do not address antibiotic prophylaxis for recurrent episodes of otitis media in their guidelines. Both groups recommend modification of risk factors to decrease recurrent AOM, including promoting breastfeeding during the first 6 months, avoiding bottle-propping, reducing or eliminating pacifier use in the second 6 months of life, and eliminating exposure to secondhand smoke.

They also recommend pneumococcal conjugate vaccine to reduce vaccine-serotype pneumococcal otitis and live-attenuated influenza vaccine during respiratory virus season for children aged >2 years.

Evidence summary

A systematic review of antibiotic prophylaxis for recurrent AOM examined 9 RCTs with a total of 958 children. Recurrent AOM was defined as 3 or more episodes per 6 to 18 months. The studies were low to moderate in quality (mean methodologic quality score of 11.8 out of 29 possible points). The most commonly used antibiotics were amoxicillin, cotrimoxazole, and sulfamethoxazole, given for 3 to 24 months (dosing not reported).

Children taking antibiotics had 0.11 (95% confidence interval [CI], 0.03–0.19) fewer episodes of recurrent AOM per patient-month than those taking placebo. The rate in the control group was 0.19 (95% CI, 0.13–0.26). Nine children would have to be treated per month to prevent 1 ear infection (NNT=9; 95% CI, 5–33). Only 2 of the 9 studies had statistically significant results; both used sulfisoxazole for 10 to 12 weeks and were of similar methodologic quality (12.5 out 29 points).

A trend towards a better outcome in studies that used sulfisoxazole did not reach significance compared with those using other medications (ie, ampicillin, amoxicillin, cotrimoxazole). Shorter treatment intervals (<6 months) trended toward being more effective than longer intervals, but this also did not reach significance. Children with more frequent episodes of AOM did no better than those with less frequent episodes.¹

Since that review was published, another study of prophylaxis for ear infections had been published. This randomized, double blind, placebo-controlled study enrolled 194 children aged 3 months to 6 years with at least 3 documented AOM episodes in the preceding 6 months. The children were given amoxicillin (20 mg/kg/d) either once daily (n=55) or divided twice daily (n=44) or placebo (n=59). Excluding 36 noncompliant subjects, the percentages without a recurrent episode were 63% for the placebo group, 64% for the once-daily amoxicillin group, and 61% for the twice-daily amoxicillin group. There was no significant difference in the incidence of new AOM episodes among the children in the 3 groups.²

A review article states: “Many children with acute otitis media do not benefit from antimicrobial therapy because the cause of their illness is not bacterial or the infection is cleared by the immune system without use of a drug. At present, we do not have clinical criteria for distinguishing which children are in need of antibiotic therapy for AOM.”³ The lack of criteria for determining which children need antibiotic therapy for AOM makes it more difficult to select children for antibiotic prophylaxis against recurrent AOM.

Recommendations from others

The American Academy of Pediatrics and the American Academy of Family Physicians do not address antibiotic prophylaxis for recurrent episodes of otitis media in their guidelines. Both groups recommend modification of risk factors to decrease recurrent AOM, including promoting breastfeeding during the first 6 months, avoiding bottle-propping, reducing or eliminating pacifier use in the second 6 months of life, and eliminating exposure to secondhand smoke.

They also recommend pneumococcal conjugate vaccine to reduce vaccine-serotype pneumococcal otitis and live-attenuated influenza vaccine during respiratory virus season for children aged >2 years.

FIGURE 1
Bariatric surgical techniques for weight loss

Evidence summary

A systematic review comparing bariatric surgery with conventional medical therapy for obesity included 1 randomized controlled trial and the Swedish Obesity Study, a large cohort study with matched controls. Surgery produced 23 to 28 kg more weight loss at 2 years.¹ The study demonstrated 33% ± 10% weight loss for gastric bypass and 0% for medical therapy (not described) at 2 years,² and 25% ± 6% loss vs 0.9% gain at 8 years.³ Among bariatric surgical techniques, patients undergoing gastric bypass lost more weight than those with gastroplasty (using staples to partition the stomach, either horizontally or vertically ( Figure 1 ) (P=.057, not significant) or gastric banding (placing a constricting ring around the stomach) (P<.05) at 8 years.³

The same systematic review assessed multiple randomized controlled trials comparing gastric bypass with gastroplasty and found greater weight loss, fewer revisions, and more side effects from gastric bypass ( Figure 2 ).¹ Five trials comparing gastric bypass with horizontal gastroplasty demonstrated significantly greater weight loss from gastric bypass. Five other trials comparing weight loss from gastric bypass with vertical gastroplasty produced mixed results, with 3 trials favoring gastric bypass and 2 showing no difference.¹ Fewer patients required revision after gastric bypass (0%–4%) compared with vertical gastroplasty (9%) or horizontal gastroplasty (19%–40%). One included trial found that postoperative dumping syndrome (28% vs 0%, P<0.05) and heartburn (59% vs 32%, P<.05) were more common with gastric bypass than with gastroplasty.¹

Bariatric surgery, including gastric bypass, improves a variety of obesity-related comorbid conditions. Diabetes prevalence decreased among gastric bypass patients at 2 years (0.0% vs 4.7%, P<0.005) and 8 years (3.6% vs 18.5%, P<.0005) compared with those receiving medical therapy.^2,3 In a case series involving 154 diabetic gastric bypass patients, diabetes resolved for 83% by 1 year, and for 86% at 5 to 7 years.⁴ In several case series, most patients became euglycemic and discontinued insulin or oral agents.

In the Swedish Obesity Study, hypertriglyceridemia decreased postoperatively but hypercholesterolemia did not.⁵ In a case series, bariatric surgery reduced triglycerides (50%) as well as total cholesterol (15%) (P<.05 for both) at 6 months and significantly increased high-density lipoprotein cholesterol levels at 1 and 5 years.⁶

Bariatric surgery significantly lowered the incidence of hypertension at 2 years (3.2%) compared with conventional treatment (9.9%), but after 8 years this difference disappeared.^2,3,5 However, in multiple large case series with morbidly obese patients, hypertension resolved or improved. The largest study showed resolution of hypertension for 69% at 1 to 2 years (91% follow-up), 66% at 5 to 7 years (50% follow-up), and 51% at 10 to 12 years (37% follow-up).⁴

Bariatric surgery improved obstructive sleep apnea and obesity hypoventilation syndrome in 2 case series. In one, Epworth Sleepiness Scale scores, minimum O₂ saturation, and other measures improved significantly (P<.001) by 3 to 21 months after surgery.⁷

In another case series, menstrual irregularities decreased from 40.4% to 4.6% following surgery (P<.001) among women who lost 50% of their excess weight.⁸ The incidence of urinary stress incontinence also decreased significantly (61.2% to 11.6%, P<.001 in this study⁸ ). The Swedish Obesity Study found significant improvements in Health-Related Quality of Life scores at 2 years with surgery vs conventional treatment.⁹

Bariatric surgery, including gastric bypass, has significant postoperative morbidity and mortality. Thirteen percent of patients in the Swedish Obesity Study experienced peri-operative complications, including pulmonary symptoms (6.2%), abdominal infection (2.1%), wound complications (1.8%), bleeding (0.9%), thromboembolic events (0.8%), and other miscellaneous complications (4.8%). Postoperative complications required reoperation for 2.2% of surgical patients, and there were 4 postoperative deaths (0.2% of the operative patients; 3 due to leakage, and 1 due to a technical laparoscopic error).²

Nutritional and vitamin deficiencies are common following gastric bypass, including deficiencies of vitamin B12, iron, folate, and calcium. Lifelong nutritional supplementation is generally necessary following this procedure.¹⁰

FIGURE 2
Long-term weight loss with bariatric surgery

Recommendations from others

A 1991 National Institutes of Health consensus conference suggested consideration of obesity surgery for patients with a body-mass index ≥40, or ≥35 plus severe obesity-related medical comorbidities (such as severe sleep apnea, obesity hypoventilation syndrome, obesity-related cardiomyopathy, or severe diabetes) who have not been successfully treated with non-surgical attempts at weight reduction.

Selected patients should be well-informed and motivated, with acceptable operative risk. A multidisciplinary team with medical, surgical, psychiatric, and nutritional expertise should evaluate patients who are candidates for surgery. An experienced surgeon, working in a clinical setting with adequate support for all aspects of management and assessment, should perform the surgery.

Lifelong medical surveillance is necessary after surgery, and patients should be selected who are likely to comply with this.¹¹

FIGURE 1
Bariatric surgical techniques for weight loss

Evidence summary

A systematic review comparing bariatric surgery with conventional medical therapy for obesity included 1 randomized controlled trial and the Swedish Obesity Study, a large cohort study with matched controls. Surgery produced 23 to 28 kg more weight loss at 2 years.¹ The study demonstrated 33% ± 10% weight loss for gastric bypass and 0% for medical therapy (not described) at 2 years,² and 25% ± 6% loss vs 0.9% gain at 8 years.³ Among bariatric surgical techniques, patients undergoing gastric bypass lost more weight than those with gastroplasty (using staples to partition the stomach, either horizontally or vertically ( Figure 1 ) (P=.057, not significant) or gastric banding (placing a constricting ring around the stomach) (P<.05) at 8 years.³

The same systematic review assessed multiple randomized controlled trials comparing gastric bypass with gastroplasty and found greater weight loss, fewer revisions, and more side effects from gastric bypass ( Figure 2 ).¹ Five trials comparing gastric bypass with horizontal gastroplasty demonstrated significantly greater weight loss from gastric bypass. Five other trials comparing weight loss from gastric bypass with vertical gastroplasty produced mixed results, with 3 trials favoring gastric bypass and 2 showing no difference.¹ Fewer patients required revision after gastric bypass (0%–4%) compared with vertical gastroplasty (9%) or horizontal gastroplasty (19%–40%). One included trial found that postoperative dumping syndrome (28% vs 0%, P<0.05) and heartburn (59% vs 32%, P<.05) were more common with gastric bypass than with gastroplasty.¹

Bariatric surgery, including gastric bypass, improves a variety of obesity-related comorbid conditions. Diabetes prevalence decreased among gastric bypass patients at 2 years (0.0% vs 4.7%, P<0.005) and 8 years (3.6% vs 18.5%, P<.0005) compared with those receiving medical therapy.^2,3 In a case series involving 154 diabetic gastric bypass patients, diabetes resolved for 83% by 1 year, and for 86% at 5 to 7 years.⁴ In several case series, most patients became euglycemic and discontinued insulin or oral agents.

In the Swedish Obesity Study, hypertriglyceridemia decreased postoperatively but hypercholesterolemia did not.⁵ In a case series, bariatric surgery reduced triglycerides (50%) as well as total cholesterol (15%) (P<.05 for both) at 6 months and significantly increased high-density lipoprotein cholesterol levels at 1 and 5 years.⁶

Bariatric surgery significantly lowered the incidence of hypertension at 2 years (3.2%) compared with conventional treatment (9.9%), but after 8 years this difference disappeared.^2,3,5 However, in multiple large case series with morbidly obese patients, hypertension resolved or improved. The largest study showed resolution of hypertension for 69% at 1 to 2 years (91% follow-up), 66% at 5 to 7 years (50% follow-up), and 51% at 10 to 12 years (37% follow-up).⁴

Bariatric surgery improved obstructive sleep apnea and obesity hypoventilation syndrome in 2 case series. In one, Epworth Sleepiness Scale scores, minimum O₂ saturation, and other measures improved significantly (P<.001) by 3 to 21 months after surgery.⁷

In another case series, menstrual irregularities decreased from 40.4% to 4.6% following surgery (P<.001) among women who lost 50% of their excess weight.⁸ The incidence of urinary stress incontinence also decreased significantly (61.2% to 11.6%, P<.001 in this study⁸ ). The Swedish Obesity Study found significant improvements in Health-Related Quality of Life scores at 2 years with surgery vs conventional treatment.⁹

Bariatric surgery, including gastric bypass, has significant postoperative morbidity and mortality. Thirteen percent of patients in the Swedish Obesity Study experienced peri-operative complications, including pulmonary symptoms (6.2%), abdominal infection (2.1%), wound complications (1.8%), bleeding (0.9%), thromboembolic events (0.8%), and other miscellaneous complications (4.8%). Postoperative complications required reoperation for 2.2% of surgical patients, and there were 4 postoperative deaths (0.2% of the operative patients; 3 due to leakage, and 1 due to a technical laparoscopic error).²

Nutritional and vitamin deficiencies are common following gastric bypass, including deficiencies of vitamin B12, iron, folate, and calcium. Lifelong nutritional supplementation is generally necessary following this procedure.¹⁰

FIGURE 2
Long-term weight loss with bariatric surgery

Recommendations from others

A 1991 National Institutes of Health consensus conference suggested consideration of obesity surgery for patients with a body-mass index ≥40, or ≥35 plus severe obesity-related medical comorbidities (such as severe sleep apnea, obesity hypoventilation syndrome, obesity-related cardiomyopathy, or severe diabetes) who have not been successfully treated with non-surgical attempts at weight reduction.

Selected patients should be well-informed and motivated, with acceptable operative risk. A multidisciplinary team with medical, surgical, psychiatric, and nutritional expertise should evaluate patients who are candidates for surgery. An experienced surgeon, working in a clinical setting with adequate support for all aspects of management and assessment, should perform the surgery.

Lifelong medical surveillance is necessary after surgery, and patients should be selected who are likely to comply with this.¹¹

FIGURE 1
Bariatric surgical techniques for weight loss

Evidence summary

A systematic review comparing bariatric surgery with conventional medical therapy for obesity included 1 randomized controlled trial and the Swedish Obesity Study, a large cohort study with matched controls. Surgery produced 23 to 28 kg more weight loss at 2 years.¹ The study demonstrated 33% ± 10% weight loss for gastric bypass and 0% for medical therapy (not described) at 2 years,² and 25% ± 6% loss vs 0.9% gain at 8 years.³ Among bariatric surgical techniques, patients undergoing gastric bypass lost more weight than those with gastroplasty (using staples to partition the stomach, either horizontally or vertically ( Figure 1 ) (P=.057, not significant) or gastric banding (placing a constricting ring around the stomach) (P<.05) at 8 years.³

The same systematic review assessed multiple randomized controlled trials comparing gastric bypass with gastroplasty and found greater weight loss, fewer revisions, and more side effects from gastric bypass ( Figure 2 ).¹ Five trials comparing gastric bypass with horizontal gastroplasty demonstrated significantly greater weight loss from gastric bypass. Five other trials comparing weight loss from gastric bypass with vertical gastroplasty produced mixed results, with 3 trials favoring gastric bypass and 2 showing no difference.¹ Fewer patients required revision after gastric bypass (0%–4%) compared with vertical gastroplasty (9%) or horizontal gastroplasty (19%–40%). One included trial found that postoperative dumping syndrome (28% vs 0%, P<0.05) and heartburn (59% vs 32%, P<.05) were more common with gastric bypass than with gastroplasty.¹

Bariatric surgery, including gastric bypass, improves a variety of obesity-related comorbid conditions. Diabetes prevalence decreased among gastric bypass patients at 2 years (0.0% vs 4.7%, P<0.005) and 8 years (3.6% vs 18.5%, P<.0005) compared with those receiving medical therapy.^2,3 In a case series involving 154 diabetic gastric bypass patients, diabetes resolved for 83% by 1 year, and for 86% at 5 to 7 years.⁴ In several case series, most patients became euglycemic and discontinued insulin or oral agents.

In the Swedish Obesity Study, hypertriglyceridemia decreased postoperatively but hypercholesterolemia did not.⁵ In a case series, bariatric surgery reduced triglycerides (50%) as well as total cholesterol (15%) (P<.05 for both) at 6 months and significantly increased high-density lipoprotein cholesterol levels at 1 and 5 years.⁶

Bariatric surgery significantly lowered the incidence of hypertension at 2 years (3.2%) compared with conventional treatment (9.9%), but after 8 years this difference disappeared.^2,3,5 However, in multiple large case series with morbidly obese patients, hypertension resolved or improved. The largest study showed resolution of hypertension for 69% at 1 to 2 years (91% follow-up), 66% at 5 to 7 years (50% follow-up), and 51% at 10 to 12 years (37% follow-up).⁴

Bariatric surgery improved obstructive sleep apnea and obesity hypoventilation syndrome in 2 case series. In one, Epworth Sleepiness Scale scores, minimum O₂ saturation, and other measures improved significantly (P<.001) by 3 to 21 months after surgery.⁷

In another case series, menstrual irregularities decreased from 40.4% to 4.6% following surgery (P<.001) among women who lost 50% of their excess weight.⁸ The incidence of urinary stress incontinence also decreased significantly (61.2% to 11.6%, P<.001 in this study⁸ ). The Swedish Obesity Study found significant improvements in Health-Related Quality of Life scores at 2 years with surgery vs conventional treatment.⁹

Bariatric surgery, including gastric bypass, has significant postoperative morbidity and mortality. Thirteen percent of patients in the Swedish Obesity Study experienced peri-operative complications, including pulmonary symptoms (6.2%), abdominal infection (2.1%), wound complications (1.8%), bleeding (0.9%), thromboembolic events (0.8%), and other miscellaneous complications (4.8%). Postoperative complications required reoperation for 2.2% of surgical patients, and there were 4 postoperative deaths (0.2% of the operative patients; 3 due to leakage, and 1 due to a technical laparoscopic error).²

Nutritional and vitamin deficiencies are common following gastric bypass, including deficiencies of vitamin B12, iron, folate, and calcium. Lifelong nutritional supplementation is generally necessary following this procedure.¹⁰

FIGURE 2
Long-term weight loss with bariatric surgery

Recommendations from others

A 1991 National Institutes of Health consensus conference suggested consideration of obesity surgery for patients with a body-mass index ≥40, or ≥35 plus severe obesity-related medical comorbidities (such as severe sleep apnea, obesity hypoventilation syndrome, obesity-related cardiomyopathy, or severe diabetes) who have not been successfully treated with non-surgical attempts at weight reduction.

Selected patients should be well-informed and motivated, with acceptable operative risk. A multidisciplinary team with medical, surgical, psychiatric, and nutritional expertise should evaluate patients who are candidates for surgery. An experienced surgeon, working in a clinical setting with adequate support for all aspects of management and assessment, should perform the surgery.

Lifelong medical surveillance is necessary after surgery, and patients should be selected who are likely to comply with this.¹¹