Sarah Safranek, MLIS

EVIDENCE SUMMARY

A post-hoc analysis of a prospective RCT (the Diabetes Prevention Program) comprising 3081 patients with impaired glucose metabolism who received metformin, a lifestyle modification program, or no intervention (placebo) found that metformin reduced the risk of developing diabetes only for patients in the highest risk quartile over 2.8 years. Lifestyle modification reduced diabetes risk in all patients.¹

Investigators stratified patients who met National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) criteria for metabolic syndrome into risk quartiles for progression to diabetes using a model they developed based on 7 parameters: fasting plasma glucose, hemoglobin A1c, history of high blood glucose, waist:hip ratio, waist circumference, triglycerides, and height (TABLE¹). The model reasonably fit outcomes—the percentage of patients in each quartile who developed diabetes—with an area under the receiver operator characteristic curve of 0.73 (a measure of diagnostic accuracy where 1 is a perfect predictor and 0.5 is random).

Lifestyle modification reduced risk in all quartiles with progressively greater effect as risk increased (lowest risk quartile: ARR=4.9%, 3-year NNT=20.4; highest risk quartile: ARR=28.3%; 3-year NNT=3.5).

There were 2 key weaknesses of the risk model: It wasn’t validated in a separate population and the true incidence of diabetes among patients taking placebo was higher than predicted. The investigators compared their risk prediction model results with the Framingham Risk Score (FRS) for diabetes and found that they correlated well, although the FRS results were consistently about 6% (absolute) higher when corrected for duration. (The FRS calculator is available online at www.framinghamheartstudy.org/risk-functions/diabetes/.)

Lifestyle change reduces diabetes risk more than metformin

The original Diabetes Prevention Program found that intensive lifestyle intervention and metformin reduced the number of diabetes cases over 2.8 years among 3234 patients at risk for developing diabetes.²

Compared with no intervention, fewer patients developed diabetes with either metformin or lifestyle improvement, although lifestyle change had the larger effect (no intervention: 11 cases per 100 person-years; metformin: 7.8 cases; 95% confidence interval [CI], 6.8-8.8; ARR=3.2% per year vs no intervention; lifestyle improvements: 4.8 cases; 95% CI, 4.1-5.7; ARR=6.2% per year vs no intervention).

A 10-year follow-up study to the Diabetes Prevention Program found that, compared with no intervention, both metformin and lifestyle interventions continued to be associated with a lower incidence of diabetes (no intervention: 7.8 cases per 100 person-years; 95% CI, 4.8-6.5; metformin: 6.4 cases; 95% CI, 4.2-5.7; ARR=1.4% per year; lifestyle interventions: 5.3 cases; 95% CI, 5.1-6.8; ARR=2.5% per year).³

Researchers originally randomized 3234 patients with body mass index ≥24 kg/m², fasting blood sugar 95 to 125 mg/dL, and 2-hour post 75-gm glucose value of 149 to 199 mg/dL to 3 groups: intensive lifestyle modification (weight loss goal of 7%, 150 minutes a week of exercise), metformin (850 mg twice daily), and no inter­vention. After the 2.8-year follow-up period, 2766 patients continued for another 5.7 years of follow-up. Investigators offered group lifestyle counseling to all patients and continued metformin at the same dose in the second group.

Earlier study shows an effect for metformin, but with a caveat

An earlier RCT found that metformin reduced the risk of developing diabetes in patients with metabolic syndrome.⁴ Investigators randomized 70 patients to metformin (250 mg 3 times daily) or placebo for a year. Fewer patients developed diabetes with metformin (3% vs 16.2%, P=.011; NNT=7.6) and more had a normal glucose tolerance test result (84.9% vs 51.4%, P=.011; NNT=3). However, by current American Diabetes Association criteria, half of the subjects had early diabetes at baseline.

Metformin lowers fasting blood sugar, but may not reverse metabolic syndrome

A post-hoc analysis of another RCT found that metformin reduced fasting plasma glucose (FPG) levels in patients with upper-body obesity and metabolic syndrome (by 1999 World Health Organization criteria but not NCEP ATP III criteria).⁵

Investigators randomized 457 patients to metformin 850 mg once daily or placebo and followed them for a year. FPG levels decreased with metformin but increased with placebo (reduction FPG 5.9 mg/dL vs increase FPG 12.3 mg/dL; P<.04). The investigators didn’t report whether any patients developed diabetes.

However, another RCT (155 patients) that compared metformin 850 mg twice daily with placebo in subjects with metabolic syndrome but without diabetes found greater normalization of FPG (5% vs 0%; P=.005), but no reversal of metabolic syndrome or change in Framingham 10-year risk score after 12 weeks.⁶

EVIDENCE SUMMARY

A post-hoc analysis of a prospective RCT (the Diabetes Prevention Program) comprising 3081 patients with impaired glucose metabolism who received metformin, a lifestyle modification program, or no intervention (placebo) found that metformin reduced the risk of developing diabetes only for patients in the highest risk quartile over 2.8 years. Lifestyle modification reduced diabetes risk in all patients.¹

Investigators stratified patients who met National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) criteria for metabolic syndrome into risk quartiles for progression to diabetes using a model they developed based on 7 parameters: fasting plasma glucose, hemoglobin A1c, history of high blood glucose, waist:hip ratio, waist circumference, triglycerides, and height (TABLE¹). The model reasonably fit outcomes—the percentage of patients in each quartile who developed diabetes—with an area under the receiver operator characteristic curve of 0.73 (a measure of diagnostic accuracy where 1 is a perfect predictor and 0.5 is random).

Lifestyle modification reduced risk in all quartiles with progressively greater effect as risk increased (lowest risk quartile: ARR=4.9%, 3-year NNT=20.4; highest risk quartile: ARR=28.3%; 3-year NNT=3.5).

There were 2 key weaknesses of the risk model: It wasn’t validated in a separate population and the true incidence of diabetes among patients taking placebo was higher than predicted. The investigators compared their risk prediction model results with the Framingham Risk Score (FRS) for diabetes and found that they correlated well, although the FRS results were consistently about 6% (absolute) higher when corrected for duration. (The FRS calculator is available online at www.framinghamheartstudy.org/risk-functions/diabetes/.)

Lifestyle change reduces diabetes risk more than metformin

The original Diabetes Prevention Program found that intensive lifestyle intervention and metformin reduced the number of diabetes cases over 2.8 years among 3234 patients at risk for developing diabetes.²

Compared with no intervention, fewer patients developed diabetes with either metformin or lifestyle improvement, although lifestyle change had the larger effect (no intervention: 11 cases per 100 person-years; metformin: 7.8 cases; 95% confidence interval [CI], 6.8-8.8; ARR=3.2% per year vs no intervention; lifestyle improvements: 4.8 cases; 95% CI, 4.1-5.7; ARR=6.2% per year vs no intervention).

A 10-year follow-up study to the Diabetes Prevention Program found that, compared with no intervention, both metformin and lifestyle interventions continued to be associated with a lower incidence of diabetes (no intervention: 7.8 cases per 100 person-years; 95% CI, 4.8-6.5; metformin: 6.4 cases; 95% CI, 4.2-5.7; ARR=1.4% per year; lifestyle interventions: 5.3 cases; 95% CI, 5.1-6.8; ARR=2.5% per year).³

Researchers originally randomized 3234 patients with body mass index ≥24 kg/m², fasting blood sugar 95 to 125 mg/dL, and 2-hour post 75-gm glucose value of 149 to 199 mg/dL to 3 groups: intensive lifestyle modification (weight loss goal of 7%, 150 minutes a week of exercise), metformin (850 mg twice daily), and no inter­vention. After the 2.8-year follow-up period, 2766 patients continued for another 5.7 years of follow-up. Investigators offered group lifestyle counseling to all patients and continued metformin at the same dose in the second group.

Earlier study shows an effect for metformin, but with a caveat

An earlier RCT found that metformin reduced the risk of developing diabetes in patients with metabolic syndrome.⁴ Investigators randomized 70 patients to metformin (250 mg 3 times daily) or placebo for a year. Fewer patients developed diabetes with metformin (3% vs 16.2%, P=.011; NNT=7.6) and more had a normal glucose tolerance test result (84.9% vs 51.4%, P=.011; NNT=3). However, by current American Diabetes Association criteria, half of the subjects had early diabetes at baseline.

Metformin lowers fasting blood sugar, but may not reverse metabolic syndrome

A post-hoc analysis of another RCT found that metformin reduced fasting plasma glucose (FPG) levels in patients with upper-body obesity and metabolic syndrome (by 1999 World Health Organization criteria but not NCEP ATP III criteria).⁵

Investigators randomized 457 patients to metformin 850 mg once daily or placebo and followed them for a year. FPG levels decreased with metformin but increased with placebo (reduction FPG 5.9 mg/dL vs increase FPG 12.3 mg/dL; P<.04). The investigators didn’t report whether any patients developed diabetes.

However, another RCT (155 patients) that compared metformin 850 mg twice daily with placebo in subjects with metabolic syndrome but without diabetes found greater normalization of FPG (5% vs 0%; P=.005), but no reversal of metabolic syndrome or change in Framingham 10-year risk score after 12 weeks.⁶

EVIDENCE SUMMARY

A post-hoc analysis of a prospective RCT (the Diabetes Prevention Program) comprising 3081 patients with impaired glucose metabolism who received metformin, a lifestyle modification program, or no intervention (placebo) found that metformin reduced the risk of developing diabetes only for patients in the highest risk quartile over 2.8 years. Lifestyle modification reduced diabetes risk in all patients.¹

Investigators stratified patients who met National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) criteria for metabolic syndrome into risk quartiles for progression to diabetes using a model they developed based on 7 parameters: fasting plasma glucose, hemoglobin A1c, history of high blood glucose, waist:hip ratio, waist circumference, triglycerides, and height (TABLE¹). The model reasonably fit outcomes—the percentage of patients in each quartile who developed diabetes—with an area under the receiver operator characteristic curve of 0.73 (a measure of diagnostic accuracy where 1 is a perfect predictor and 0.5 is random).

Lifestyle modification reduced risk in all quartiles with progressively greater effect as risk increased (lowest risk quartile: ARR=4.9%, 3-year NNT=20.4; highest risk quartile: ARR=28.3%; 3-year NNT=3.5).

There were 2 key weaknesses of the risk model: It wasn’t validated in a separate population and the true incidence of diabetes among patients taking placebo was higher than predicted. The investigators compared their risk prediction model results with the Framingham Risk Score (FRS) for diabetes and found that they correlated well, although the FRS results were consistently about 6% (absolute) higher when corrected for duration. (The FRS calculator is available online at www.framinghamheartstudy.org/risk-functions/diabetes/.)

Lifestyle change reduces diabetes risk more than metformin

The original Diabetes Prevention Program found that intensive lifestyle intervention and metformin reduced the number of diabetes cases over 2.8 years among 3234 patients at risk for developing diabetes.²

Compared with no intervention, fewer patients developed diabetes with either metformin or lifestyle improvement, although lifestyle change had the larger effect (no intervention: 11 cases per 100 person-years; metformin: 7.8 cases; 95% confidence interval [CI], 6.8-8.8; ARR=3.2% per year vs no intervention; lifestyle improvements: 4.8 cases; 95% CI, 4.1-5.7; ARR=6.2% per year vs no intervention).

A 10-year follow-up study to the Diabetes Prevention Program found that, compared with no intervention, both metformin and lifestyle interventions continued to be associated with a lower incidence of diabetes (no intervention: 7.8 cases per 100 person-years; 95% CI, 4.8-6.5; metformin: 6.4 cases; 95% CI, 4.2-5.7; ARR=1.4% per year; lifestyle interventions: 5.3 cases; 95% CI, 5.1-6.8; ARR=2.5% per year).³

Researchers originally randomized 3234 patients with body mass index ≥24 kg/m², fasting blood sugar 95 to 125 mg/dL, and 2-hour post 75-gm glucose value of 149 to 199 mg/dL to 3 groups: intensive lifestyle modification (weight loss goal of 7%, 150 minutes a week of exercise), metformin (850 mg twice daily), and no inter­vention. After the 2.8-year follow-up period, 2766 patients continued for another 5.7 years of follow-up. Investigators offered group lifestyle counseling to all patients and continued metformin at the same dose in the second group.

Earlier study shows an effect for metformin, but with a caveat

An earlier RCT found that metformin reduced the risk of developing diabetes in patients with metabolic syndrome.⁴ Investigators randomized 70 patients to metformin (250 mg 3 times daily) or placebo for a year. Fewer patients developed diabetes with metformin (3% vs 16.2%, P=.011; NNT=7.6) and more had a normal glucose tolerance test result (84.9% vs 51.4%, P=.011; NNT=3). However, by current American Diabetes Association criteria, half of the subjects had early diabetes at baseline.

Metformin lowers fasting blood sugar, but may not reverse metabolic syndrome

A post-hoc analysis of another RCT found that metformin reduced fasting plasma glucose (FPG) levels in patients with upper-body obesity and metabolic syndrome (by 1999 World Health Organization criteria but not NCEP ATP III criteria).⁵

Investigators randomized 457 patients to metformin 850 mg once daily or placebo and followed them for a year. FPG levels decreased with metformin but increased with placebo (reduction FPG 5.9 mg/dL vs increase FPG 12.3 mg/dL; P<.04). The investigators didn’t report whether any patients developed diabetes.

However, another RCT (155 patients) that compared metformin 850 mg twice daily with placebo in subjects with metabolic syndrome but without diabetes found greater normalization of FPG (5% vs 0%; P=.005), but no reversal of metabolic syndrome or change in Framingham 10-year risk score after 12 weeks.⁶

Evidence summary

A cross-sectional study found no correlation between knuckle popping and osteoarthritis (OA) of the hand.¹ Investigators recruited 300 consecutive patients (ages 45 years and older, mean age 63 years) and evaluated them for a history of habitual knuckle popping (74 of 300 patients, mean duration 35 years) and hand arthritis or dysfunction. Investigators excluded patients with neuromuscular, inflammatory, or malignant diseases.

Investigators found OA equally in both patients who did and didn’t pop their knuckles (12 of 74 vs 36 of 226, respectively; P nonsignificant); joint swelling was more common in participants with a history of knuckle popping (84% vs 6%; P<.01). Investigators didn’t describe how OA was diagnosed or specify which joints were affected.

Another cross-sectional study also found no correlation between habitual knuckle popping of the metacarpal phalangeal joint and the prevalence of OA in that joint.² Investigators recruited 28 patients (mean age 78.5 years; 23 women and 5 men) from a Jewish home for the aged and asked them whether they had habitually cracked their knuckles during their lifetime. They then performed clinical and radiographic hand examinations (excluding patients with a history of ­traumatic injury, rheumatoid arthritis, gout, chondrocalcinosis, and hemochromatosis).

Knuckle popping didn’t correlate with OA of the metacarpal phalanges (1 of 15 knuckle popping patients vs 5 of 13 patients who didn’t pop their knuckles; P=.06). All 6 patients with radiographic evidence of OA showed involvement at the metacarpal phalangeal and distal interphalangeal joints, whether or not they popped their knuckles.

A case control study found no correlation between OA in the hands and habitual knuckle popping.³ Investigators recruited 215 patients 50 to 89 years old who had received a radiograph of their right hand during the previous 5 years and divided them into cases with OA (135 patients), and controls without OA (80 patients). Patients completed questionnaires assessing the prevalence (20%), frequency (1 to 20 times per day), and duration (26 to 36 years) of knuckle popping.

Patients most commonly popped proximal interphalangeal joints (15.9%) followed by metacarpal phalangeal joints (13.5%), distal interphalangeal joints (6.1%), and first carpal metacarpal joints (2.3%). OA most often affected the distal interphalangeal joint (68.4%), followed by the first carpal metacarpal (57.1%), proximal interphalangeal (54.1%), and metacarpal phalangeal joints (28.6%). Investigators found no difference in the prevalence of knuckle popping between cases and controls (18% in cases vs 23.2% in controls; P=.361).

When investigators evaluated total knuckle popping exposure in “crack years” (number of times per day multiplied by years) in the distal interphalangeal or metacarpal phalangeal joints, they found no significant association between crack years and OA (distal interphalangeal joint, mean 108 crack years; metacarpal phalangeal joint, mean 75 crack years).

50 years of knuckle popping without ill effects

An n-of-1 case control study found similar results.⁴ The researcher, a physician, popped only the knuckles of his left hand, twice a day, for 50 years. He compared his hands at the end of the trial and found no arthritis in either hand and no visible differences.

But knuckle popping does have a downside

A paper described 2 case reports of acute injuries sustained during attempted knuckle popping—a partial tear of the ulnar collateral ligament of the thumb and subluxation of the extensor tendon of the fifth digit.⁵ Both injuries were associated with forceful manipulation of the digits, and both resolved with conservative management within 4 weeks.

Evidence summary

A cross-sectional study found no correlation between knuckle popping and osteoarthritis (OA) of the hand.¹ Investigators recruited 300 consecutive patients (ages 45 years and older, mean age 63 years) and evaluated them for a history of habitual knuckle popping (74 of 300 patients, mean duration 35 years) and hand arthritis or dysfunction. Investigators excluded patients with neuromuscular, inflammatory, or malignant diseases.

Investigators found OA equally in both patients who did and didn’t pop their knuckles (12 of 74 vs 36 of 226, respectively; P nonsignificant); joint swelling was more common in participants with a history of knuckle popping (84% vs 6%; P<.01). Investigators didn’t describe how OA was diagnosed or specify which joints were affected.

Another cross-sectional study also found no correlation between habitual knuckle popping of the metacarpal phalangeal joint and the prevalence of OA in that joint.² Investigators recruited 28 patients (mean age 78.5 years; 23 women and 5 men) from a Jewish home for the aged and asked them whether they had habitually cracked their knuckles during their lifetime. They then performed clinical and radiographic hand examinations (excluding patients with a history of ­traumatic injury, rheumatoid arthritis, gout, chondrocalcinosis, and hemochromatosis).

Knuckle popping didn’t correlate with OA of the metacarpal phalanges (1 of 15 knuckle popping patients vs 5 of 13 patients who didn’t pop their knuckles; P=.06). All 6 patients with radiographic evidence of OA showed involvement at the metacarpal phalangeal and distal interphalangeal joints, whether or not they popped their knuckles.

A case control study found no correlation between OA in the hands and habitual knuckle popping.³ Investigators recruited 215 patients 50 to 89 years old who had received a radiograph of their right hand during the previous 5 years and divided them into cases with OA (135 patients), and controls without OA (80 patients). Patients completed questionnaires assessing the prevalence (20%), frequency (1 to 20 times per day), and duration (26 to 36 years) of knuckle popping.

Patients most commonly popped proximal interphalangeal joints (15.9%) followed by metacarpal phalangeal joints (13.5%), distal interphalangeal joints (6.1%), and first carpal metacarpal joints (2.3%). OA most often affected the distal interphalangeal joint (68.4%), followed by the first carpal metacarpal (57.1%), proximal interphalangeal (54.1%), and metacarpal phalangeal joints (28.6%). Investigators found no difference in the prevalence of knuckle popping between cases and controls (18% in cases vs 23.2% in controls; P=.361).

When investigators evaluated total knuckle popping exposure in “crack years” (number of times per day multiplied by years) in the distal interphalangeal or metacarpal phalangeal joints, they found no significant association between crack years and OA (distal interphalangeal joint, mean 108 crack years; metacarpal phalangeal joint, mean 75 crack years).

50 years of knuckle popping without ill effects

An n-of-1 case control study found similar results.⁴ The researcher, a physician, popped only the knuckles of his left hand, twice a day, for 50 years. He compared his hands at the end of the trial and found no arthritis in either hand and no visible differences.

But knuckle popping does have a downside

A paper described 2 case reports of acute injuries sustained during attempted knuckle popping—a partial tear of the ulnar collateral ligament of the thumb and subluxation of the extensor tendon of the fifth digit.⁵ Both injuries were associated with forceful manipulation of the digits, and both resolved with conservative management within 4 weeks.

Evidence summary

A cross-sectional study found no correlation between knuckle popping and osteoarthritis (OA) of the hand.¹ Investigators recruited 300 consecutive patients (ages 45 years and older, mean age 63 years) and evaluated them for a history of habitual knuckle popping (74 of 300 patients, mean duration 35 years) and hand arthritis or dysfunction. Investigators excluded patients with neuromuscular, inflammatory, or malignant diseases.

Investigators found OA equally in both patients who did and didn’t pop their knuckles (12 of 74 vs 36 of 226, respectively; P nonsignificant); joint swelling was more common in participants with a history of knuckle popping (84% vs 6%; P<.01). Investigators didn’t describe how OA was diagnosed or specify which joints were affected.

Another cross-sectional study also found no correlation between habitual knuckle popping of the metacarpal phalangeal joint and the prevalence of OA in that joint.² Investigators recruited 28 patients (mean age 78.5 years; 23 women and 5 men) from a Jewish home for the aged and asked them whether they had habitually cracked their knuckles during their lifetime. They then performed clinical and radiographic hand examinations (excluding patients with a history of ­traumatic injury, rheumatoid arthritis, gout, chondrocalcinosis, and hemochromatosis).

Knuckle popping didn’t correlate with OA of the metacarpal phalanges (1 of 15 knuckle popping patients vs 5 of 13 patients who didn’t pop their knuckles; P=.06). All 6 patients with radiographic evidence of OA showed involvement at the metacarpal phalangeal and distal interphalangeal joints, whether or not they popped their knuckles.

A case control study found no correlation between OA in the hands and habitual knuckle popping.³ Investigators recruited 215 patients 50 to 89 years old who had received a radiograph of their right hand during the previous 5 years and divided them into cases with OA (135 patients), and controls without OA (80 patients). Patients completed questionnaires assessing the prevalence (20%), frequency (1 to 20 times per day), and duration (26 to 36 years) of knuckle popping.

Patients most commonly popped proximal interphalangeal joints (15.9%) followed by metacarpal phalangeal joints (13.5%), distal interphalangeal joints (6.1%), and first carpal metacarpal joints (2.3%). OA most often affected the distal interphalangeal joint (68.4%), followed by the first carpal metacarpal (57.1%), proximal interphalangeal (54.1%), and metacarpal phalangeal joints (28.6%). Investigators found no difference in the prevalence of knuckle popping between cases and controls (18% in cases vs 23.2% in controls; P=.361).

When investigators evaluated total knuckle popping exposure in “crack years” (number of times per day multiplied by years) in the distal interphalangeal or metacarpal phalangeal joints, they found no significant association between crack years and OA (distal interphalangeal joint, mean 108 crack years; metacarpal phalangeal joint, mean 75 crack years).

50 years of knuckle popping without ill effects

An n-of-1 case control study found similar results.⁴ The researcher, a physician, popped only the knuckles of his left hand, twice a day, for 50 years. He compared his hands at the end of the trial and found no arthritis in either hand and no visible differences.

But knuckle popping does have a downside

A paper described 2 case reports of acute injuries sustained during attempted knuckle popping—a partial tear of the ulnar collateral ligament of the thumb and subluxation of the extensor tendon of the fifth digit.⁵ Both injuries were associated with forceful manipulation of the digits, and both resolved with conservative management within 4 weeks.

EVIDENCE SUMMARY

A Cochrane systematic review (19 RCTs; 3335 patients, total) of newer-generation antidepressants for treating depression in adolescents found that, overall, they produced both a small decrease in symptom severity scores and an increased risk of suicide-related outcomes.¹

Three SSRIs slightly lower one symptom severity score

Investigators performed a meta-analysis of all trials (14 RCTs; 2490 patients, total) that used the same standardized symptom severity score (the Children’s Depression Rating Scale—Revised [CDRS-R], range 17 to 113 points) to evaluate the following medications: fluoxetine, sertraline, escitalopram, citalopram, paroxetine, venlafaxine, and mirtazapine.¹

All participants were outpatients who met criteria for a primary diagnosis of major depression, excluding comorbid conditions. The CDRS-R scores were evaluated by clinicians; the mean baseline score was 57 (40 is considered a threshold score for diagnosis, and above 60 indicates severe symptoms). Only 5 trials reported patients’ self-rated depression symptom severity (in patients taking fluoxetine and paroxetine) and none reported improvement. Treatment courses ranged from 8 to 12 weeks.

As a group, the newer antidepressants slightly reduced CDRS-R scores in adolescents (by 4.21 points, 95% confidence interval [CI], 0.41-5.95) but increased suicide-related outcomes (relative risk [RR]=1.47; 95% CI, 0.99-2.19). The individual antidepressants fluoxetine, sertraline, and escitalopram each produced statistically significant but clinically small reductions in CDRS-R scores of 5% to 10% without significantly increasing suicide-related outcomes (TABLE¹). The other medications evaluated individually didn’t improve CDRS-R scores, and only venlafaxine increased suicide-related outcomes.

Other symptom severity scores show no improvement with SSRIs

Five additional RCTs not included in the meta-analysis that used standardized symptom severity scores other than the CDRS-R (Schedule for Affective Disorders and Schizophrenia for School-Aged Children [K-SADS], Montgomery-Asberg Depression Rating Scale [MADR], and Hamilton Depression Rating Scale [HAM-D]) found no improvement with fluoxetine (2 RCTs; 63 patients, total), citalopram (one RCT, 233 patients), or paroxetine (2 RCTs; 466 patients, total).

Certain drugs cause significantly more adverse events than placebo

Ten RCTs evaluated adverse events in adolescents treated with fluoxetine, escitalopram, citalopram, and paroxetine and reported a small increase over placebo when all medications were combined as a group (RR=1.11; 95% CI, 1.05-1.17). Investigators reported that the individual antidepressants fluoxetine, escitalopram, venlafaxine, and mirtazapine produced significantly more adverse events than placebo (P values not given). No studies compared antidepressant medications against each other for either efficacy or potential harms.

RECOMMENDATIONS

A newly revised expert guideline recommends treating mildly depressed adolescents with a specific psychological therapy—individual cognitive behavioral therapy, interpersonal therapy, family therapy, or psychodynamic psychotherapy—for at least 3 months.²

For adolescents with moderate to severe depression, the guideline advocates psychotherapy with the option of adding fluoxetine, although using antidepressants in adolescents who haven’t at least tried psychotherapy is outside of the drug’s indications.

The guideline also recommends careful monitoring for adverse effects and close review of mental state—weekly for the first 4 weeks of treatment, for example. If fluoxetine doesn’t help, sertraline and citalopram are recommended as alternatives.

EVIDENCE SUMMARY

A Cochrane systematic review (19 RCTs; 3335 patients, total) of newer-generation antidepressants for treating depression in adolescents found that, overall, they produced both a small decrease in symptom severity scores and an increased risk of suicide-related outcomes.¹

Three SSRIs slightly lower one symptom severity score

Investigators performed a meta-analysis of all trials (14 RCTs; 2490 patients, total) that used the same standardized symptom severity score (the Children’s Depression Rating Scale—Revised [CDRS-R], range 17 to 113 points) to evaluate the following medications: fluoxetine, sertraline, escitalopram, citalopram, paroxetine, venlafaxine, and mirtazapine.¹

All participants were outpatients who met criteria for a primary diagnosis of major depression, excluding comorbid conditions. The CDRS-R scores were evaluated by clinicians; the mean baseline score was 57 (40 is considered a threshold score for diagnosis, and above 60 indicates severe symptoms). Only 5 trials reported patients’ self-rated depression symptom severity (in patients taking fluoxetine and paroxetine) and none reported improvement. Treatment courses ranged from 8 to 12 weeks.

As a group, the newer antidepressants slightly reduced CDRS-R scores in adolescents (by 4.21 points, 95% confidence interval [CI], 0.41-5.95) but increased suicide-related outcomes (relative risk [RR]=1.47; 95% CI, 0.99-2.19). The individual antidepressants fluoxetine, sertraline, and escitalopram each produced statistically significant but clinically small reductions in CDRS-R scores of 5% to 10% without significantly increasing suicide-related outcomes (TABLE¹). The other medications evaluated individually didn’t improve CDRS-R scores, and only venlafaxine increased suicide-related outcomes.

Other symptom severity scores show no improvement with SSRIs

Five additional RCTs not included in the meta-analysis that used standardized symptom severity scores other than the CDRS-R (Schedule for Affective Disorders and Schizophrenia for School-Aged Children [K-SADS], Montgomery-Asberg Depression Rating Scale [MADR], and Hamilton Depression Rating Scale [HAM-D]) found no improvement with fluoxetine (2 RCTs; 63 patients, total), citalopram (one RCT, 233 patients), or paroxetine (2 RCTs; 466 patients, total).

Certain drugs cause significantly more adverse events than placebo

Ten RCTs evaluated adverse events in adolescents treated with fluoxetine, escitalopram, citalopram, and paroxetine and reported a small increase over placebo when all medications were combined as a group (RR=1.11; 95% CI, 1.05-1.17). Investigators reported that the individual antidepressants fluoxetine, escitalopram, venlafaxine, and mirtazapine produced significantly more adverse events than placebo (P values not given). No studies compared antidepressant medications against each other for either efficacy or potential harms.

RECOMMENDATIONS

A newly revised expert guideline recommends treating mildly depressed adolescents with a specific psychological therapy—individual cognitive behavioral therapy, interpersonal therapy, family therapy, or psychodynamic psychotherapy—for at least 3 months.²

For adolescents with moderate to severe depression, the guideline advocates psychotherapy with the option of adding fluoxetine, although using antidepressants in adolescents who haven’t at least tried psychotherapy is outside of the drug’s indications.

The guideline also recommends careful monitoring for adverse effects and close review of mental state—weekly for the first 4 weeks of treatment, for example. If fluoxetine doesn’t help, sertraline and citalopram are recommended as alternatives.

EVIDENCE SUMMARY

A Cochrane systematic review (19 RCTs; 3335 patients, total) of newer-generation antidepressants for treating depression in adolescents found that, overall, they produced both a small decrease in symptom severity scores and an increased risk of suicide-related outcomes.¹

Three SSRIs slightly lower one symptom severity score

Investigators performed a meta-analysis of all trials (14 RCTs; 2490 patients, total) that used the same standardized symptom severity score (the Children’s Depression Rating Scale—Revised [CDRS-R], range 17 to 113 points) to evaluate the following medications: fluoxetine, sertraline, escitalopram, citalopram, paroxetine, venlafaxine, and mirtazapine.¹

All participants were outpatients who met criteria for a primary diagnosis of major depression, excluding comorbid conditions. The CDRS-R scores were evaluated by clinicians; the mean baseline score was 57 (40 is considered a threshold score for diagnosis, and above 60 indicates severe symptoms). Only 5 trials reported patients’ self-rated depression symptom severity (in patients taking fluoxetine and paroxetine) and none reported improvement. Treatment courses ranged from 8 to 12 weeks.

As a group, the newer antidepressants slightly reduced CDRS-R scores in adolescents (by 4.21 points, 95% confidence interval [CI], 0.41-5.95) but increased suicide-related outcomes (relative risk [RR]=1.47; 95% CI, 0.99-2.19). The individual antidepressants fluoxetine, sertraline, and escitalopram each produced statistically significant but clinically small reductions in CDRS-R scores of 5% to 10% without significantly increasing suicide-related outcomes (TABLE¹). The other medications evaluated individually didn’t improve CDRS-R scores, and only venlafaxine increased suicide-related outcomes.

Other symptom severity scores show no improvement with SSRIs

Five additional RCTs not included in the meta-analysis that used standardized symptom severity scores other than the CDRS-R (Schedule for Affective Disorders and Schizophrenia for School-Aged Children [K-SADS], Montgomery-Asberg Depression Rating Scale [MADR], and Hamilton Depression Rating Scale [HAM-D]) found no improvement with fluoxetine (2 RCTs; 63 patients, total), citalopram (one RCT, 233 patients), or paroxetine (2 RCTs; 466 patients, total).

Certain drugs cause significantly more adverse events than placebo

Ten RCTs evaluated adverse events in adolescents treated with fluoxetine, escitalopram, citalopram, and paroxetine and reported a small increase over placebo when all medications were combined as a group (RR=1.11; 95% CI, 1.05-1.17). Investigators reported that the individual antidepressants fluoxetine, escitalopram, venlafaxine, and mirtazapine produced significantly more adverse events than placebo (P values not given). No studies compared antidepressant medications against each other for either efficacy or potential harms.

RECOMMENDATIONS

A newly revised expert guideline recommends treating mildly depressed adolescents with a specific psychological therapy—individual cognitive behavioral therapy, interpersonal therapy, family therapy, or psychodynamic psychotherapy—for at least 3 months.²

For adolescents with moderate to severe depression, the guideline advocates psychotherapy with the option of adding fluoxetine, although using antidepressants in adolescents who haven’t at least tried psychotherapy is outside of the drug’s indications.

The guideline also recommends careful monitoring for adverse effects and close review of mental state—weekly for the first 4 weeks of treatment, for example. If fluoxetine doesn’t help, sertraline and citalopram are recommended as alternatives.

EVIDENCE SUMMARY

A systematic review of 6 RCTs that evaluated SSRIs and SNRIs (fluoxetine, sertraline, paroxetine, venlafaxine) found them all to be effective for reducing hot flash frequency and symptom scores in women with previous breast cancer¹ (TABLE^1,2).

A 2006 meta-analysis combined the results of 7 RCTs (each evaluating a single SSRI [fluoxetine, paroxetine] or SNRI [venlafaxine]) and found that as a group, they reduced mean hot flash frequency (–1.13 hot flashes/d; 95% confidence interval [CI], –1.70 to –0.57) in women with and without breast cancer.² No trial compared medications head to head, and the populations differed among studies, so that investigators couldn’t determine a single best agent.

Clonidine and gabapentin decrease hot flash frequency

The 2006 meta-analysis also included 10 RCTs (743 patients) that studied clonidine in women with and without a history of breast cancer, and 2 RCTs (479 patients) that evaluated gabapentin in women with breast cancer.² Both drugs reduced mean hot flash frequency (clonidine: –0.95 hot flashes/d, 95% CI, –1.44 to –0.47 at 4 weeks and –1.63 hot flashes/d, 95% CI, –2.76 to –0.05 at 8 weeks; gabapentin: –2.05 hot flashes/d; 95% CI, −2.80 to –1.30).

Phytoestrogens: The jury is still out

A meta-analysis of 43 RCTs (4364 patients) evaluated phytoestrogens that included dietary soy, soy extracts, red clover extracts, genistein extracts, and other types of phytoestrogens.³ The data from the only 5 RCTs (300 patients) that could be combined showed no effect from red clover extract on hot flash frequency. However, another 4 individual trials that couldn’t be combined each found that extracts with high levels of the phytoestrogen genistein (>30 mg/d) did reduce frequency. Investigators reported that many of the trials were small and had a high risk of bias.

A meta-analysis of 16 RCTs (2027 patients) that assessed black cohosh found that it didn’t reduce hot flash frequency (3 RCTs, 393 patients) or symptom severity scores (4 RCTs, 357 patients).⁴ Investigators reported high heterogeneity and recommended further research.

Nonpharmacologic therapies and vitamin E don’t help

Systematic reviews found that relaxation therapy (4 RCTs, 281 patients), exercise (3 RCTs, 454 patients), and acupuncture (8 RCTs, 414 patients) didn’t reduce hot flashes.^5-7 In another review, vitamin E (1 RCT, 105 patients), homeopathy (2 RCTs, 124 patients), and magnetic devices (1 RCT, 11 patients) also produced no benefit.¹

EVIDENCE SUMMARY

A systematic review of 6 RCTs that evaluated SSRIs and SNRIs (fluoxetine, sertraline, paroxetine, venlafaxine) found them all to be effective for reducing hot flash frequency and symptom scores in women with previous breast cancer¹ (TABLE^1,2).

A 2006 meta-analysis combined the results of 7 RCTs (each evaluating a single SSRI [fluoxetine, paroxetine] or SNRI [venlafaxine]) and found that as a group, they reduced mean hot flash frequency (–1.13 hot flashes/d; 95% confidence interval [CI], –1.70 to –0.57) in women with and without breast cancer.² No trial compared medications head to head, and the populations differed among studies, so that investigators couldn’t determine a single best agent.

Clonidine and gabapentin decrease hot flash frequency

The 2006 meta-analysis also included 10 RCTs (743 patients) that studied clonidine in women with and without a history of breast cancer, and 2 RCTs (479 patients) that evaluated gabapentin in women with breast cancer.² Both drugs reduced mean hot flash frequency (clonidine: –0.95 hot flashes/d, 95% CI, –1.44 to –0.47 at 4 weeks and –1.63 hot flashes/d, 95% CI, –2.76 to –0.05 at 8 weeks; gabapentin: –2.05 hot flashes/d; 95% CI, −2.80 to –1.30).

Phytoestrogens: The jury is still out

A meta-analysis of 43 RCTs (4364 patients) evaluated phytoestrogens that included dietary soy, soy extracts, red clover extracts, genistein extracts, and other types of phytoestrogens.³ The data from the only 5 RCTs (300 patients) that could be combined showed no effect from red clover extract on hot flash frequency. However, another 4 individual trials that couldn’t be combined each found that extracts with high levels of the phytoestrogen genistein (>30 mg/d) did reduce frequency. Investigators reported that many of the trials were small and had a high risk of bias.

A meta-analysis of 16 RCTs (2027 patients) that assessed black cohosh found that it didn’t reduce hot flash frequency (3 RCTs, 393 patients) or symptom severity scores (4 RCTs, 357 patients).⁴ Investigators reported high heterogeneity and recommended further research.

Nonpharmacologic therapies and vitamin E don’t help

Systematic reviews found that relaxation therapy (4 RCTs, 281 patients), exercise (3 RCTs, 454 patients), and acupuncture (8 RCTs, 414 patients) didn’t reduce hot flashes.^5-7 In another review, vitamin E (1 RCT, 105 patients), homeopathy (2 RCTs, 124 patients), and magnetic devices (1 RCT, 11 patients) also produced no benefit.¹

EVIDENCE SUMMARY

A systematic review of 6 RCTs that evaluated SSRIs and SNRIs (fluoxetine, sertraline, paroxetine, venlafaxine) found them all to be effective for reducing hot flash frequency and symptom scores in women with previous breast cancer¹ (TABLE^1,2).

A 2006 meta-analysis combined the results of 7 RCTs (each evaluating a single SSRI [fluoxetine, paroxetine] or SNRI [venlafaxine]) and found that as a group, they reduced mean hot flash frequency (–1.13 hot flashes/d; 95% confidence interval [CI], –1.70 to –0.57) in women with and without breast cancer.² No trial compared medications head to head, and the populations differed among studies, so that investigators couldn’t determine a single best agent.

Clonidine and gabapentin decrease hot flash frequency

The 2006 meta-analysis also included 10 RCTs (743 patients) that studied clonidine in women with and without a history of breast cancer, and 2 RCTs (479 patients) that evaluated gabapentin in women with breast cancer.² Both drugs reduced mean hot flash frequency (clonidine: –0.95 hot flashes/d, 95% CI, –1.44 to –0.47 at 4 weeks and –1.63 hot flashes/d, 95% CI, –2.76 to –0.05 at 8 weeks; gabapentin: –2.05 hot flashes/d; 95% CI, −2.80 to –1.30).

Phytoestrogens: The jury is still out

A meta-analysis of 43 RCTs (4364 patients) evaluated phytoestrogens that included dietary soy, soy extracts, red clover extracts, genistein extracts, and other types of phytoestrogens.³ The data from the only 5 RCTs (300 patients) that could be combined showed no effect from red clover extract on hot flash frequency. However, another 4 individual trials that couldn’t be combined each found that extracts with high levels of the phytoestrogen genistein (>30 mg/d) did reduce frequency. Investigators reported that many of the trials were small and had a high risk of bias.

A meta-analysis of 16 RCTs (2027 patients) that assessed black cohosh found that it didn’t reduce hot flash frequency (3 RCTs, 393 patients) or symptom severity scores (4 RCTs, 357 patients).⁴ Investigators reported high heterogeneity and recommended further research.

Nonpharmacologic therapies and vitamin E don’t help

Systematic reviews found that relaxation therapy (4 RCTs, 281 patients), exercise (3 RCTs, 454 patients), and acupuncture (8 RCTs, 414 patients) didn’t reduce hot flashes.^5-7 In another review, vitamin E (1 RCT, 105 patients), homeopathy (2 RCTs, 124 patients), and magnetic devices (1 RCT, 11 patients) also produced no benefit.¹

EVIDENCE SUMMARY

A Cochrane systematic review of the effects of caffeine on pregnancy identified 2 studies, only one of which addressed the question of maternal caffeine intake and infant birth weight.¹ The double-blind RCT evaluating caffeine intake during pregnancy found no significant differences in birth weight or length of gestation between women who drank regular coffee and women who drank decaffeinated coffee.²

At 16 weeks’ gestation, investigators randomized 1207 pregnant women who reported daily intake of at least 3 cups of regular coffee to drink unlabeled instant coffee (which was either regular or decaffeinated) for the rest of their pregnancy. The women were allowed to request as much of their assigned instant coffee as they wanted.

Subjects were recruited from among all women with uncomplicated, singleton pregnancies who were expected to deliver at a Danish university hospital during the study period. Investigators interviewed the women at 20, 25, and 34 weeks to determine coffee consumption (including both coffee provided by the investigators and other coffee), consumption of other caffeinated beverages, and smoking status.

The difference in caffeine intake between the groups didn’t correspond to significant differences in birth weight (16 g lighter with caffeinated coffee; 95% confidence interval [CI], −40 g to 73 g; P=.48) or birth length (0.03 cm longer with caffeinated coffee; 95% CI, −0.29 to 0.22) among infants born to the 1150 women who completed the study.

Limitations of the study include randomizing women after 16 weeks’ gestation and the observation that many women correctly guessed which type of coffee they received (35% of women drinking caffeinated coffee and 49% of women drinking decaf).

A caffeine effect, but with study limitations

The Cochrane systematic review (described above) and a meta-analysis of 9 prospective cohort studies with a total of 90,000 patients that evaluated maternal caffeine intake found that it was associated with increased low birth weight, intrauterine growth restriction (IUGR), or small for gestational age (SGA) infants.³

Researchers assessed caffeine consumption from coffee or other sources either by questionnaire (5 studies) or interview (4 studies) at various times during pregnancy, mostly in the first or second trimester, and assigned subjects to 4 intake categories: none, low (50-149 mg/d), moderate (150-349 mg/d), and high (>350 mg/d).

Compared with no caffeine, all levels of caffeine intake were associated with increased rates of low birth weight, IUGR, or SGA (low intake: relative risk [RR]=1.13; 95% CI, 1.06-1.21; moderate intake: RR=1.38; 95% CI, 1.18-1.62; high intake: RR=1.60; 95% CI, 1.24-2.08).

A major limitation of the meta-analysis was that 8 of the included studies were identified by the reviewers as having quality problems. The reviewers also identified additional cohort studies, not included in the meta-analysis, which failed to show any association between caffeine intake and poor pregnancy outcomes.

Results of best-quality study prove clinically trivial

The best-quality prospective cohort study in the review described above was also the largest, comprising two-thirds of the total patients. It found a statistically significant, but clinically trivial, association between caffeine intake and birth weight.⁴

Investigators from Norway’s Institute of Public Health mailed surveys to 106,707 pregnant Norwegian women and recruited 59,123 with uncomplicated singleton pregnancies. The survey assessed diet and lifestyle at several stages of pregnancy and correlated caffeine intake with birth weight, gestational length, and SGA deliveries. Investigators calculated caffeine intake from coffee and other dietary sources (tea and chocolate).

Higher caffeine intake was associated with a small reduction in birth weight (8 g/100 mg/d of additional caffeine intake; 95% CI, −10 to −6 g/100 mg/d). Higher intake was also associated with increasing likelihood of SGA birth, a finding of borderline significance (odds ratio [OR]=1.18; 95% CI, 1.00-1.38, comparing intake <50 mg/d with 51-200 mg/d; OR=1.62; 95% CI, 1.26-2.29, comparing <50 mg/d with 201-300 mg/d; and OR=1.62; 95% CI, 1.15-2.29, comparing <50 mg/d with >300 mg/d).

EVIDENCE SUMMARY

A Cochrane systematic review of the effects of caffeine on pregnancy identified 2 studies, only one of which addressed the question of maternal caffeine intake and infant birth weight.¹ The double-blind RCT evaluating caffeine intake during pregnancy found no significant differences in birth weight or length of gestation between women who drank regular coffee and women who drank decaffeinated coffee.²

At 16 weeks’ gestation, investigators randomized 1207 pregnant women who reported daily intake of at least 3 cups of regular coffee to drink unlabeled instant coffee (which was either regular or decaffeinated) for the rest of their pregnancy. The women were allowed to request as much of their assigned instant coffee as they wanted.

Subjects were recruited from among all women with uncomplicated, singleton pregnancies who were expected to deliver at a Danish university hospital during the study period. Investigators interviewed the women at 20, 25, and 34 weeks to determine coffee consumption (including both coffee provided by the investigators and other coffee), consumption of other caffeinated beverages, and smoking status.

The difference in caffeine intake between the groups didn’t correspond to significant differences in birth weight (16 g lighter with caffeinated coffee; 95% confidence interval [CI], −40 g to 73 g; P=.48) or birth length (0.03 cm longer with caffeinated coffee; 95% CI, −0.29 to 0.22) among infants born to the 1150 women who completed the study.

Limitations of the study include randomizing women after 16 weeks’ gestation and the observation that many women correctly guessed which type of coffee they received (35% of women drinking caffeinated coffee and 49% of women drinking decaf).

A caffeine effect, but with study limitations

The Cochrane systematic review (described above) and a meta-analysis of 9 prospective cohort studies with a total of 90,000 patients that evaluated maternal caffeine intake found that it was associated with increased low birth weight, intrauterine growth restriction (IUGR), or small for gestational age (SGA) infants.³

Researchers assessed caffeine consumption from coffee or other sources either by questionnaire (5 studies) or interview (4 studies) at various times during pregnancy, mostly in the first or second trimester, and assigned subjects to 4 intake categories: none, low (50-149 mg/d), moderate (150-349 mg/d), and high (>350 mg/d).

Compared with no caffeine, all levels of caffeine intake were associated with increased rates of low birth weight, IUGR, or SGA (low intake: relative risk [RR]=1.13; 95% CI, 1.06-1.21; moderate intake: RR=1.38; 95% CI, 1.18-1.62; high intake: RR=1.60; 95% CI, 1.24-2.08).

A major limitation of the meta-analysis was that 8 of the included studies were identified by the reviewers as having quality problems. The reviewers also identified additional cohort studies, not included in the meta-analysis, which failed to show any association between caffeine intake and poor pregnancy outcomes.

Results of best-quality study prove clinically trivial

The best-quality prospective cohort study in the review described above was also the largest, comprising two-thirds of the total patients. It found a statistically significant, but clinically trivial, association between caffeine intake and birth weight.⁴

Investigators from Norway’s Institute of Public Health mailed surveys to 106,707 pregnant Norwegian women and recruited 59,123 with uncomplicated singleton pregnancies. The survey assessed diet and lifestyle at several stages of pregnancy and correlated caffeine intake with birth weight, gestational length, and SGA deliveries. Investigators calculated caffeine intake from coffee and other dietary sources (tea and chocolate).

Higher caffeine intake was associated with a small reduction in birth weight (8 g/100 mg/d of additional caffeine intake; 95% CI, −10 to −6 g/100 mg/d). Higher intake was also associated with increasing likelihood of SGA birth, a finding of borderline significance (odds ratio [OR]=1.18; 95% CI, 1.00-1.38, comparing intake <50 mg/d with 51-200 mg/d; OR=1.62; 95% CI, 1.26-2.29, comparing <50 mg/d with 201-300 mg/d; and OR=1.62; 95% CI, 1.15-2.29, comparing <50 mg/d with >300 mg/d).

EVIDENCE SUMMARY

A Cochrane systematic review of the effects of caffeine on pregnancy identified 2 studies, only one of which addressed the question of maternal caffeine intake and infant birth weight.¹ The double-blind RCT evaluating caffeine intake during pregnancy found no significant differences in birth weight or length of gestation between women who drank regular coffee and women who drank decaffeinated coffee.²

At 16 weeks’ gestation, investigators randomized 1207 pregnant women who reported daily intake of at least 3 cups of regular coffee to drink unlabeled instant coffee (which was either regular or decaffeinated) for the rest of their pregnancy. The women were allowed to request as much of their assigned instant coffee as they wanted.

Subjects were recruited from among all women with uncomplicated, singleton pregnancies who were expected to deliver at a Danish university hospital during the study period. Investigators interviewed the women at 20, 25, and 34 weeks to determine coffee consumption (including both coffee provided by the investigators and other coffee), consumption of other caffeinated beverages, and smoking status.

The difference in caffeine intake between the groups didn’t correspond to significant differences in birth weight (16 g lighter with caffeinated coffee; 95% confidence interval [CI], −40 g to 73 g; P=.48) or birth length (0.03 cm longer with caffeinated coffee; 95% CI, −0.29 to 0.22) among infants born to the 1150 women who completed the study.

Limitations of the study include randomizing women after 16 weeks’ gestation and the observation that many women correctly guessed which type of coffee they received (35% of women drinking caffeinated coffee and 49% of women drinking decaf).

A caffeine effect, but with study limitations

The Cochrane systematic review (described above) and a meta-analysis of 9 prospective cohort studies with a total of 90,000 patients that evaluated maternal caffeine intake found that it was associated with increased low birth weight, intrauterine growth restriction (IUGR), or small for gestational age (SGA) infants.³

Researchers assessed caffeine consumption from coffee or other sources either by questionnaire (5 studies) or interview (4 studies) at various times during pregnancy, mostly in the first or second trimester, and assigned subjects to 4 intake categories: none, low (50-149 mg/d), moderate (150-349 mg/d), and high (>350 mg/d).

Compared with no caffeine, all levels of caffeine intake were associated with increased rates of low birth weight, IUGR, or SGA (low intake: relative risk [RR]=1.13; 95% CI, 1.06-1.21; moderate intake: RR=1.38; 95% CI, 1.18-1.62; high intake: RR=1.60; 95% CI, 1.24-2.08).

A major limitation of the meta-analysis was that 8 of the included studies were identified by the reviewers as having quality problems. The reviewers also identified additional cohort studies, not included in the meta-analysis, which failed to show any association between caffeine intake and poor pregnancy outcomes.

Results of best-quality study prove clinically trivial

The best-quality prospective cohort study in the review described above was also the largest, comprising two-thirds of the total patients. It found a statistically significant, but clinically trivial, association between caffeine intake and birth weight.⁴

Investigators from Norway’s Institute of Public Health mailed surveys to 106,707 pregnant Norwegian women and recruited 59,123 with uncomplicated singleton pregnancies. The survey assessed diet and lifestyle at several stages of pregnancy and correlated caffeine intake with birth weight, gestational length, and SGA deliveries. Investigators calculated caffeine intake from coffee and other dietary sources (tea and chocolate).

Higher caffeine intake was associated with a small reduction in birth weight (8 g/100 mg/d of additional caffeine intake; 95% CI, −10 to −6 g/100 mg/d). Higher intake was also associated with increasing likelihood of SGA birth, a finding of borderline significance (odds ratio [OR]=1.18; 95% CI, 1.00-1.38, comparing intake <50 mg/d with 51-200 mg/d; OR=1.62; 95% CI, 1.26-2.29, comparing <50 mg/d with 201-300 mg/d; and OR=1.62; 95% CI, 1.15-2.29, comparing <50 mg/d with >300 mg/d).

EVIDENCE SUMMARY

A Cochrane systematic review of 5 RCTs with a total of 7314 patients evaluated cardiovascular outcomes after 4.7 years follow-up in patients with diabetes who were treated for hypertension to either “lower” or “standard” target blood pressures.¹ One trial in the review (ACCORD, 4734 patients) compared outcomes from significantly lower and standard systolic blood pressures (119/64 mm Hg vs 134/71 mm Hg; P<.0001) in patients with diabetes and either cardiovascular disease or 2 risk factors for cardiovascular disease. The authors evaluated outcomes based on achieved systolic blood pressures rather than intention to treat.

They found a reduced incidence of stroke (risk ratio [RR]=0.58; 95% confidence interval [CI], 0.39-0.88; P=.009; number needed to treat [NNT]=91) but no change in mortality (RR=1.05; 95% CI, 0.84-1.30) at lower blood pressures. Achieving the lower systolic blood pressure increased the number of serious adverse effects, however (RR=2.58; 95% CI, 1.70-3.91; P<.0001; absolute risk increase=2%; number needed to harm=50).

Four RCTs (2580 patients) in the systematic review compared clinical outcomes produced by achieving significantly lower or standard diastolic blood pressure targets (128/76 mm Hg vs 135/83 mm Hg; P<.0001). The trials found no significant difference in total mortality (RR=0.73; 95% CI, 0.53-1.01), stroke (RR=0.67; 95% CI, 0.42-1.05), myocardial infarction (RR=0.95; 95% CI, 0.64-1.40), or congestive heart failure (RR=1.06; 95% CI, 0.58-1.92). Sensitivity analysis of trials comparing diastolic blood pressure targets below 80 mm Hg and below 90 mm Hg showed similar results.

The 4 RCTs didn’t report end-stage renal failure or total serious adverse events. The authors stated that there was a high risk of selection bias in favor of lower blood pressure targets.

Patients with CKD

A systematic review and meta-analysis of 11 RCTs (9287 patients) compared outcomes of achieving lower blood pressure targets or standard targets in patients with CKD. Intensive blood pressure treatment reduced the risk of kidney failure only in patients with proteinuria at baseline (hazard ratio [HR]=0.73; 95% CI, 0.62-0.86; 5 trials, 1703 patients).² Investigators didn’t report the degree of proteinuria for all the trials, but in one trial, patients had proteinuria of 1 to 3 g/d.

Achieved blood pressures in the intensive therapy group averaged 7.7/4.9 mm Hg lower, with pressures typically ranging from 75 to 80 mm Hg diastolic and 125 to 135 mm Hg systolic. Intensive blood pressure lowering didn’t reduce kidney failure in patients without baseline proteinuria (HR=1.12; 95% CI, 0.67-1.87; 3 trials, 1218 patients). Nor did it reduce death (RR=0.94; 95% CI, 0.84-1.05; 10 trials, 6788 patients) or major cardiovascular outcomes (RR=1.09; 95% CI, 0.83-1.42; 5 trials, 5308 patients).

Patients with coronary artery disease

A meta-analysis of 15 RCTs (66,504 patients) that evaluated tight control of hypertension (≤130/80 mm Hg) compared with standard control (<140/90 mm Hg) in patients with coronary artery disease found reduced rates of heart failure (RR=0.73; 95% CI, 0.64-0.84; 10 trials, 37,990 patients) and stroke (RR=0.82; 95% CI, 0.69-0.98; 9 trials, 8344 patients) but increased rates of hypotension (RR=2.19; 95% CI, 1.80-2.66; 6 trials, 17,836 patients).³

Achieving lower blood pressure targets didn’t reduce all-cause mortality (RR=0.96; 95% CI, 0.89-1.04; 13 trials, 39,262 patients), cardiovascular mortality (RR=0.96; 95% CI, 0.86-1.07; 11 trials, 38,452 patients), myocardial infarction (RR=0.92; 95% CI, 0.85-1.00; 14 trials, 39,696 patients), or angina (RR=0.92; 95% CI, 0.84-1.0; 11 trials, 28,007 patients).  But it slightly increased the need for revascularization (RR=1.06; 95% CI, 1.01-1.12; 11 trials, 38,450 patients).

The SPRINT trial: Promising results for intensive treatment of some patients

The Systolic Blood Pressure Intervention Trial (SPRINT), a large RCT, found that targeting systolic blood pressures below 120 mm Hg (compared with a target below 140 mm Hg) in middle-aged and older patients with increased cardiovascular risk reduced a composite outcome that included cardiovascular death by 25%.⁴

Researchers recruited 9361 patients older than 50 years (mean age 68 years; >28% older than 75 years) with systolic blood pressure between 130 and 180 mm Hg and increased cardiovascular risk defined by one or more of the following: preexisting cardiovascular disease, CKD with estimated glomerular filtration rate between 20 and 60 mL/min/1.73 m², age >75 years, and Framingham 10-year risk of 15% or more. They excluded patients with diabetes or previous stroke.

Patients were randomized to intensive treatment (target systolic BP <120; mean achieved 121.4) or standard treatment (target systolic BP <140; mean achieved 136.2). Treatment typically comprised 3 (intensive) or 2 (standard) agents. The primary outcome was a composite of myocardial infarction, acute coronary syndrome, stroke, congestive heart failure, or cardiovascular death.

The study, which was originally intended to run for 5 years, was stopped at 3.26 years based on positive results. Intensive treatment improved the primary composite outcome overall (1.65% vs 2.19%; HR=0.75; 95% CI, 0.64-0.89; P<.001; NNT=61 over 3.26 years), all-cause mortality (HR=0.73; 95% CI, 0.60-0.90; P=.003; NNT=90), and cardiovascular death (HR=0.57; 95% CI, 0.38-0.85; P=.005; NNT=172).

However, intensive treatment didn’t significantly improve the primary composite outcome in these subgroups:

• female patients (HR=0.84; 95% CI, 0.62-1.14)
• black patients (HR=0.77; 95% CI, 0.55-1.06)
• patients with preexisting CKD (HR=0.82; 95% CI, 0.63-1.07) or cardiovascular disease (HR=0.83; 95% CI, 0.62-1.09)
• patients younger than 75 years (HR=0.80; 95% CI, 0.64-1.00)
• patients with systolic blood pressures higher than 132 mm Hg (BP >132 to <145 mm Hg, HR=0.77; 95% CI, 0.57-1.03; BP ≥145 mm Hg, HR=0.83; 95% CI, 0.63-1.09).

While the SPRINT study found that targeting systolic BP below 120 mm Hg reduced a composite outcome that included cardiovascular death, it didn’t improve this outcome in certain patient subgroups.

Intensive treatment also produced more net serious adverse events (HR=1.88; 4.7% vs 2.5%; P<.001), including: ≥30% decrease of glomerular filtration rates to values below 60 mL/min/1.73 m² (HR=3.49; 95% CI, 2.44-5.10; P<.001), syncope (HR=1.44; 3.5% vs 2.4%; P=.003), hypotension (HR=1.70; 3.4% vs 2.0%; P<.001), and electrolyte abnormalities (HR=1.38; 3.8% vs 2.8%; P=.006). It didn’t cause injurious falls (HR=1.00; P=.97) or orthostatic hypotension in clinic (HR=0.88; 16.6% vs 18.3%; P=.01).

Guidelines for patients with peripheral artery disease, previous stroke

A national guideline by an expert panel recommended treating patients with hypertension who have peripheral artery disease or previous stroke to standard values for the general population: <140/90 mm Hg if ages 60 years or younger, <150/90 mm Hg if older than 60 years.⁵ 

EVIDENCE SUMMARY

A Cochrane systematic review of 5 RCTs with a total of 7314 patients evaluated cardiovascular outcomes after 4.7 years follow-up in patients with diabetes who were treated for hypertension to either “lower” or “standard” target blood pressures.¹ One trial in the review (ACCORD, 4734 patients) compared outcomes from significantly lower and standard systolic blood pressures (119/64 mm Hg vs 134/71 mm Hg; P<.0001) in patients with diabetes and either cardiovascular disease or 2 risk factors for cardiovascular disease. The authors evaluated outcomes based on achieved systolic blood pressures rather than intention to treat.

They found a reduced incidence of stroke (risk ratio [RR]=0.58; 95% confidence interval [CI], 0.39-0.88; P=.009; number needed to treat [NNT]=91) but no change in mortality (RR=1.05; 95% CI, 0.84-1.30) at lower blood pressures. Achieving the lower systolic blood pressure increased the number of serious adverse effects, however (RR=2.58; 95% CI, 1.70-3.91; P<.0001; absolute risk increase=2%; number needed to harm=50).

Four RCTs (2580 patients) in the systematic review compared clinical outcomes produced by achieving significantly lower or standard diastolic blood pressure targets (128/76 mm Hg vs 135/83 mm Hg; P<.0001). The trials found no significant difference in total mortality (RR=0.73; 95% CI, 0.53-1.01), stroke (RR=0.67; 95% CI, 0.42-1.05), myocardial infarction (RR=0.95; 95% CI, 0.64-1.40), or congestive heart failure (RR=1.06; 95% CI, 0.58-1.92). Sensitivity analysis of trials comparing diastolic blood pressure targets below 80 mm Hg and below 90 mm Hg showed similar results.

The 4 RCTs didn’t report end-stage renal failure or total serious adverse events. The authors stated that there was a high risk of selection bias in favor of lower blood pressure targets.

Patients with CKD

A systematic review and meta-analysis of 11 RCTs (9287 patients) compared outcomes of achieving lower blood pressure targets or standard targets in patients with CKD. Intensive blood pressure treatment reduced the risk of kidney failure only in patients with proteinuria at baseline (hazard ratio [HR]=0.73; 95% CI, 0.62-0.86; 5 trials, 1703 patients).² Investigators didn’t report the degree of proteinuria for all the trials, but in one trial, patients had proteinuria of 1 to 3 g/d.

Achieved blood pressures in the intensive therapy group averaged 7.7/4.9 mm Hg lower, with pressures typically ranging from 75 to 80 mm Hg diastolic and 125 to 135 mm Hg systolic. Intensive blood pressure lowering didn’t reduce kidney failure in patients without baseline proteinuria (HR=1.12; 95% CI, 0.67-1.87; 3 trials, 1218 patients). Nor did it reduce death (RR=0.94; 95% CI, 0.84-1.05; 10 trials, 6788 patients) or major cardiovascular outcomes (RR=1.09; 95% CI, 0.83-1.42; 5 trials, 5308 patients).

Patients with coronary artery disease

A meta-analysis of 15 RCTs (66,504 patients) that evaluated tight control of hypertension (≤130/80 mm Hg) compared with standard control (<140/90 mm Hg) in patients with coronary artery disease found reduced rates of heart failure (RR=0.73; 95% CI, 0.64-0.84; 10 trials, 37,990 patients) and stroke (RR=0.82; 95% CI, 0.69-0.98; 9 trials, 8344 patients) but increased rates of hypotension (RR=2.19; 95% CI, 1.80-2.66; 6 trials, 17,836 patients).³

Achieving lower blood pressure targets didn’t reduce all-cause mortality (RR=0.96; 95% CI, 0.89-1.04; 13 trials, 39,262 patients), cardiovascular mortality (RR=0.96; 95% CI, 0.86-1.07; 11 trials, 38,452 patients), myocardial infarction (RR=0.92; 95% CI, 0.85-1.00; 14 trials, 39,696 patients), or angina (RR=0.92; 95% CI, 0.84-1.0; 11 trials, 28,007 patients).  But it slightly increased the need for revascularization (RR=1.06; 95% CI, 1.01-1.12; 11 trials, 38,450 patients).

The SPRINT trial: Promising results for intensive treatment of some patients

The Systolic Blood Pressure Intervention Trial (SPRINT), a large RCT, found that targeting systolic blood pressures below 120 mm Hg (compared with a target below 140 mm Hg) in middle-aged and older patients with increased cardiovascular risk reduced a composite outcome that included cardiovascular death by 25%.⁴

Researchers recruited 9361 patients older than 50 years (mean age 68 years; >28% older than 75 years) with systolic blood pressure between 130 and 180 mm Hg and increased cardiovascular risk defined by one or more of the following: preexisting cardiovascular disease, CKD with estimated glomerular filtration rate between 20 and 60 mL/min/1.73 m², age >75 years, and Framingham 10-year risk of 15% or more. They excluded patients with diabetes or previous stroke.

Patients were randomized to intensive treatment (target systolic BP <120; mean achieved 121.4) or standard treatment (target systolic BP <140; mean achieved 136.2). Treatment typically comprised 3 (intensive) or 2 (standard) agents. The primary outcome was a composite of myocardial infarction, acute coronary syndrome, stroke, congestive heart failure, or cardiovascular death.

The study, which was originally intended to run for 5 years, was stopped at 3.26 years based on positive results. Intensive treatment improved the primary composite outcome overall (1.65% vs 2.19%; HR=0.75; 95% CI, 0.64-0.89; P<.001; NNT=61 over 3.26 years), all-cause mortality (HR=0.73; 95% CI, 0.60-0.90; P=.003; NNT=90), and cardiovascular death (HR=0.57; 95% CI, 0.38-0.85; P=.005; NNT=172).

However, intensive treatment didn’t significantly improve the primary composite outcome in these subgroups:

• female patients (HR=0.84; 95% CI, 0.62-1.14)
• black patients (HR=0.77; 95% CI, 0.55-1.06)
• patients with preexisting CKD (HR=0.82; 95% CI, 0.63-1.07) or cardiovascular disease (HR=0.83; 95% CI, 0.62-1.09)
• patients younger than 75 years (HR=0.80; 95% CI, 0.64-1.00)
• patients with systolic blood pressures higher than 132 mm Hg (BP >132 to <145 mm Hg, HR=0.77; 95% CI, 0.57-1.03; BP ≥145 mm Hg, HR=0.83; 95% CI, 0.63-1.09).

While the SPRINT study found that targeting systolic BP below 120 mm Hg reduced a composite outcome that included cardiovascular death, it didn’t improve this outcome in certain patient subgroups.

Intensive treatment also produced more net serious adverse events (HR=1.88; 4.7% vs 2.5%; P<.001), including: ≥30% decrease of glomerular filtration rates to values below 60 mL/min/1.73 m² (HR=3.49; 95% CI, 2.44-5.10; P<.001), syncope (HR=1.44; 3.5% vs 2.4%; P=.003), hypotension (HR=1.70; 3.4% vs 2.0%; P<.001), and electrolyte abnormalities (HR=1.38; 3.8% vs 2.8%; P=.006). It didn’t cause injurious falls (HR=1.00; P=.97) or orthostatic hypotension in clinic (HR=0.88; 16.6% vs 18.3%; P=.01).

Guidelines for patients with peripheral artery disease, previous stroke

A national guideline by an expert panel recommended treating patients with hypertension who have peripheral artery disease or previous stroke to standard values for the general population: <140/90 mm Hg if ages 60 years or younger, <150/90 mm Hg if older than 60 years.⁵ 

EVIDENCE SUMMARY

A Cochrane systematic review of 5 RCTs with a total of 7314 patients evaluated cardiovascular outcomes after 4.7 years follow-up in patients with diabetes who were treated for hypertension to either “lower” or “standard” target blood pressures.¹ One trial in the review (ACCORD, 4734 patients) compared outcomes from significantly lower and standard systolic blood pressures (119/64 mm Hg vs 134/71 mm Hg; P<.0001) in patients with diabetes and either cardiovascular disease or 2 risk factors for cardiovascular disease. The authors evaluated outcomes based on achieved systolic blood pressures rather than intention to treat.

They found a reduced incidence of stroke (risk ratio [RR]=0.58; 95% confidence interval [CI], 0.39-0.88; P=.009; number needed to treat [NNT]=91) but no change in mortality (RR=1.05; 95% CI, 0.84-1.30) at lower blood pressures. Achieving the lower systolic blood pressure increased the number of serious adverse effects, however (RR=2.58; 95% CI, 1.70-3.91; P<.0001; absolute risk increase=2%; number needed to harm=50).

Four RCTs (2580 patients) in the systematic review compared clinical outcomes produced by achieving significantly lower or standard diastolic blood pressure targets (128/76 mm Hg vs 135/83 mm Hg; P<.0001). The trials found no significant difference in total mortality (RR=0.73; 95% CI, 0.53-1.01), stroke (RR=0.67; 95% CI, 0.42-1.05), myocardial infarction (RR=0.95; 95% CI, 0.64-1.40), or congestive heart failure (RR=1.06; 95% CI, 0.58-1.92). Sensitivity analysis of trials comparing diastolic blood pressure targets below 80 mm Hg and below 90 mm Hg showed similar results.

The 4 RCTs didn’t report end-stage renal failure or total serious adverse events. The authors stated that there was a high risk of selection bias in favor of lower blood pressure targets.

Patients with CKD

A systematic review and meta-analysis of 11 RCTs (9287 patients) compared outcomes of achieving lower blood pressure targets or standard targets in patients with CKD. Intensive blood pressure treatment reduced the risk of kidney failure only in patients with proteinuria at baseline (hazard ratio [HR]=0.73; 95% CI, 0.62-0.86; 5 trials, 1703 patients).² Investigators didn’t report the degree of proteinuria for all the trials, but in one trial, patients had proteinuria of 1 to 3 g/d.

Achieved blood pressures in the intensive therapy group averaged 7.7/4.9 mm Hg lower, with pressures typically ranging from 75 to 80 mm Hg diastolic and 125 to 135 mm Hg systolic. Intensive blood pressure lowering didn’t reduce kidney failure in patients without baseline proteinuria (HR=1.12; 95% CI, 0.67-1.87; 3 trials, 1218 patients). Nor did it reduce death (RR=0.94; 95% CI, 0.84-1.05; 10 trials, 6788 patients) or major cardiovascular outcomes (RR=1.09; 95% CI, 0.83-1.42; 5 trials, 5308 patients).

Patients with coronary artery disease

A meta-analysis of 15 RCTs (66,504 patients) that evaluated tight control of hypertension (≤130/80 mm Hg) compared with standard control (<140/90 mm Hg) in patients with coronary artery disease found reduced rates of heart failure (RR=0.73; 95% CI, 0.64-0.84; 10 trials, 37,990 patients) and stroke (RR=0.82; 95% CI, 0.69-0.98; 9 trials, 8344 patients) but increased rates of hypotension (RR=2.19; 95% CI, 1.80-2.66; 6 trials, 17,836 patients).³

Achieving lower blood pressure targets didn’t reduce all-cause mortality (RR=0.96; 95% CI, 0.89-1.04; 13 trials, 39,262 patients), cardiovascular mortality (RR=0.96; 95% CI, 0.86-1.07; 11 trials, 38,452 patients), myocardial infarction (RR=0.92; 95% CI, 0.85-1.00; 14 trials, 39,696 patients), or angina (RR=0.92; 95% CI, 0.84-1.0; 11 trials, 28,007 patients).  But it slightly increased the need for revascularization (RR=1.06; 95% CI, 1.01-1.12; 11 trials, 38,450 patients).

The SPRINT trial: Promising results for intensive treatment of some patients

The Systolic Blood Pressure Intervention Trial (SPRINT), a large RCT, found that targeting systolic blood pressures below 120 mm Hg (compared with a target below 140 mm Hg) in middle-aged and older patients with increased cardiovascular risk reduced a composite outcome that included cardiovascular death by 25%.⁴

Researchers recruited 9361 patients older than 50 years (mean age 68 years; >28% older than 75 years) with systolic blood pressure between 130 and 180 mm Hg and increased cardiovascular risk defined by one or more of the following: preexisting cardiovascular disease, CKD with estimated glomerular filtration rate between 20 and 60 mL/min/1.73 m², age >75 years, and Framingham 10-year risk of 15% or more. They excluded patients with diabetes or previous stroke.

Patients were randomized to intensive treatment (target systolic BP <120; mean achieved 121.4) or standard treatment (target systolic BP <140; mean achieved 136.2). Treatment typically comprised 3 (intensive) or 2 (standard) agents. The primary outcome was a composite of myocardial infarction, acute coronary syndrome, stroke, congestive heart failure, or cardiovascular death.

The study, which was originally intended to run for 5 years, was stopped at 3.26 years based on positive results. Intensive treatment improved the primary composite outcome overall (1.65% vs 2.19%; HR=0.75; 95% CI, 0.64-0.89; P<.001; NNT=61 over 3.26 years), all-cause mortality (HR=0.73; 95% CI, 0.60-0.90; P=.003; NNT=90), and cardiovascular death (HR=0.57; 95% CI, 0.38-0.85; P=.005; NNT=172).

However, intensive treatment didn’t significantly improve the primary composite outcome in these subgroups:

• female patients (HR=0.84; 95% CI, 0.62-1.14)
• black patients (HR=0.77; 95% CI, 0.55-1.06)
• patients with preexisting CKD (HR=0.82; 95% CI, 0.63-1.07) or cardiovascular disease (HR=0.83; 95% CI, 0.62-1.09)
• patients younger than 75 years (HR=0.80; 95% CI, 0.64-1.00)
• patients with systolic blood pressures higher than 132 mm Hg (BP >132 to <145 mm Hg, HR=0.77; 95% CI, 0.57-1.03; BP ≥145 mm Hg, HR=0.83; 95% CI, 0.63-1.09).

While the SPRINT study found that targeting systolic BP below 120 mm Hg reduced a composite outcome that included cardiovascular death, it didn’t improve this outcome in certain patient subgroups.

Intensive treatment also produced more net serious adverse events (HR=1.88; 4.7% vs 2.5%; P<.001), including: ≥30% decrease of glomerular filtration rates to values below 60 mL/min/1.73 m² (HR=3.49; 95% CI, 2.44-5.10; P<.001), syncope (HR=1.44; 3.5% vs 2.4%; P=.003), hypotension (HR=1.70; 3.4% vs 2.0%; P<.001), and electrolyte abnormalities (HR=1.38; 3.8% vs 2.8%; P=.006). It didn’t cause injurious falls (HR=1.00; P=.97) or orthostatic hypotension in clinic (HR=0.88; 16.6% vs 18.3%; P=.01).

Guidelines for patients with peripheral artery disease, previous stroke

A national guideline by an expert panel recommended treating patients with hypertension who have peripheral artery disease or previous stroke to standard values for the general population: <140/90 mm Hg if ages 60 years or younger, <150/90 mm Hg if older than 60 years.⁵ 

EVIDENCE SUMMARY

Six cohort studies found that patients who underwent ASD for subacromial impingement had improved pain and function scores at 4.5 to 12 years after surgery (TABLE^1-7). Weaknesses of the overall data set include use of heterogeneous outcome measures across studies, lack of sham surgical controls, and lack of blinding.

ASD improves pain and function slightly more than other treatments

One prospective and one retrospective cohort trial compared ASD with another intervention. In the prospective trial, ASD was associated with a 10% better combined pain and function score than open acromioplasty at 12 years.¹ In the retrospective trial, ASD was also associated with a 10% better combined pain and function score than prolonged physical therapy in patients older than 57 years (the median age of study participants) but not patients younger than 57 years.²

Two other studies found improvements in pain and function

Two other prospective cohort studies didn’t use a comparison group but followed changes in standardized shoulder pain and function scores for 5 to 6 years after ASD. In one study, pain decreased 6 points on a 10-point visual analog scale by 6 months postop (P<.001).³ In both studies, a 9% to 10% improvement in function was seen between 6 months and 5 to 6 years after surgery.^3,4

A third cohort study that asked patients about overall pain and satisfaction 8 to 11 years after ASD found that most were “very” or “quite” satisfied and half were pain-free.^5,6

Rotator cuff tears found less likely with ASD

An anatomic study obtained ultrasounds of patients 13 to 17 years after ASD and compared the findings to rotator cuff ultrasounds of the general population.⁷ Patients who had ASD were 22% less likely to demonstrate rotator cuff tears at the end of the study (no statistics were reported to measure significance).

RECOMMENDATIONS

Guidelines from the Washington State Department of Labor and Industry state that patients who should undergo isolated subacromial decompression (with or without acromioplasty) need to have documented subacromial impingement syndrome with magnetic resonance imaging evidence of rotator cuff tendonopathy or tear, have undergone 12 weeks of conservative therapy (including at least active assisted range of motion and home-based exercises), and have had a subacromial injection with a local anesthetic that has provided documented relief of pain.⁸

No current guidelines are available from national or international orthopedic or sports medicine organizations.

EVIDENCE SUMMARY

Six cohort studies found that patients who underwent ASD for subacromial impingement had improved pain and function scores at 4.5 to 12 years after surgery (TABLE^1-7). Weaknesses of the overall data set include use of heterogeneous outcome measures across studies, lack of sham surgical controls, and lack of blinding.

ASD improves pain and function slightly more than other treatments

One prospective and one retrospective cohort trial compared ASD with another intervention. In the prospective trial, ASD was associated with a 10% better combined pain and function score than open acromioplasty at 12 years.¹ In the retrospective trial, ASD was also associated with a 10% better combined pain and function score than prolonged physical therapy in patients older than 57 years (the median age of study participants) but not patients younger than 57 years.²

Two other studies found improvements in pain and function

Two other prospective cohort studies didn’t use a comparison group but followed changes in standardized shoulder pain and function scores for 5 to 6 years after ASD. In one study, pain decreased 6 points on a 10-point visual analog scale by 6 months postop (P<.001).³ In both studies, a 9% to 10% improvement in function was seen between 6 months and 5 to 6 years after surgery.^3,4

A third cohort study that asked patients about overall pain and satisfaction 8 to 11 years after ASD found that most were “very” or “quite” satisfied and half were pain-free.^5,6

Rotator cuff tears found less likely with ASD

An anatomic study obtained ultrasounds of patients 13 to 17 years after ASD and compared the findings to rotator cuff ultrasounds of the general population.⁷ Patients who had ASD were 22% less likely to demonstrate rotator cuff tears at the end of the study (no statistics were reported to measure significance).

RECOMMENDATIONS

Guidelines from the Washington State Department of Labor and Industry state that patients who should undergo isolated subacromial decompression (with or without acromioplasty) need to have documented subacromial impingement syndrome with magnetic resonance imaging evidence of rotator cuff tendonopathy or tear, have undergone 12 weeks of conservative therapy (including at least active assisted range of motion and home-based exercises), and have had a subacromial injection with a local anesthetic that has provided documented relief of pain.⁸

No current guidelines are available from national or international orthopedic or sports medicine organizations.

EVIDENCE SUMMARY

Six cohort studies found that patients who underwent ASD for subacromial impingement had improved pain and function scores at 4.5 to 12 years after surgery (TABLE^1-7). Weaknesses of the overall data set include use of heterogeneous outcome measures across studies, lack of sham surgical controls, and lack of blinding.

ASD improves pain and function slightly more than other treatments

One prospective and one retrospective cohort trial compared ASD with another intervention. In the prospective trial, ASD was associated with a 10% better combined pain and function score than open acromioplasty at 12 years.¹ In the retrospective trial, ASD was also associated with a 10% better combined pain and function score than prolonged physical therapy in patients older than 57 years (the median age of study participants) but not patients younger than 57 years.²

Two other studies found improvements in pain and function

Two other prospective cohort studies didn’t use a comparison group but followed changes in standardized shoulder pain and function scores for 5 to 6 years after ASD. In one study, pain decreased 6 points on a 10-point visual analog scale by 6 months postop (P<.001).³ In both studies, a 9% to 10% improvement in function was seen between 6 months and 5 to 6 years after surgery.^3,4

A third cohort study that asked patients about overall pain and satisfaction 8 to 11 years after ASD found that most were “very” or “quite” satisfied and half were pain-free.^5,6

Rotator cuff tears found less likely with ASD

An anatomic study obtained ultrasounds of patients 13 to 17 years after ASD and compared the findings to rotator cuff ultrasounds of the general population.⁷ Patients who had ASD were 22% less likely to demonstrate rotator cuff tears at the end of the study (no statistics were reported to measure significance).

RECOMMENDATIONS

Guidelines from the Washington State Department of Labor and Industry state that patients who should undergo isolated subacromial decompression (with or without acromioplasty) need to have documented subacromial impingement syndrome with magnetic resonance imaging evidence of rotator cuff tendonopathy or tear, have undergone 12 weeks of conservative therapy (including at least active assisted range of motion and home-based exercises), and have had a subacromial injection with a local anesthetic that has provided documented relief of pain.⁸

No current guidelines are available from national or international orthopedic or sports medicine organizations.

EVIDENCE SUMMARY

A large systematic review evaluated the relationship between dietary soy intake and risk of a primary breast cancer diagnosis. It included 7 prospective cohort studies, which comprised the best quality evidence available (numerous other reviewed studies were of lower quality). The review found no significant association between dietary soy intake and primary breast cancer (TABLE^1-6).

Investigators either surveyed women for intake of soy isoflavones or soy foods or products (tofu, soybeans, lentils, miso) or measured urinary or plasma levels of soy isoflavones. They adjusted for age, alcohol use, smoking status, body mass index, caloric intake, and hormone replacement therapy, then followed subjects for 7 to 23 years, comparing the risk of breast cancer for the lowest and highest levels of soy intake. 

Six of the prospective cohort studies found no association between soy intake and breast cancer risk; one study, comprising 4% of the total population, found a lower risk with higher soy intake (effect size=0.44; 95% confidence interval [CI], 0.26-0.73; an effect size of 0.2 is considered small, 0.6 medium, and 1.2 large). The authors didn’t do a meta-analysis of the prospective cohort studies.

Other cohort studies yield similar findings

Four other large systematic reviews evaluating soy intake and breast cancer risk incorporated a total of 6 individual prospective cohort studies that weren’t included in the previously described review (again, these studies comprised the best quality evidence within the reviews). The 6 studies found no association between soy intake and breast cancer risk.

In 2 of the studies, investigators surveyed postmenopausal women and followed them for 4 to 8 years.² Investigators in another study adjusted for age, family and gynecologic history, hormone and medication use, exercise, and other factors.³ In 2 other studies, investigators evaluated population subsets that consumed the most vs the fewest servings per week or kilograms per year of soy foods.⁴ The sixth study compared low with high intake of soy foods and miso.⁵

Soy intake after breast cancer diagnosis reduces recurrence risk in most studies

Most prospective cohort studies evaluating the association between dietary soy intake after breast cancer diagnosis found an overall 21% decrease in recurrence with high soy intake and a 15% reduction in mortality (TABLE^1-6).

Investigators in a meta-analysis of 5 studies that followed women for 4 to 7 years after first breast cancer diagnosis found that higher soy intake was associated with lower mortality but not less recurrence in women who were estrogen receptor positive. Both recurrence and mortality were decreased in estrogen receptor negative women.⁶

The study also found lower recurrence and mortality in premenopausal women with higher soy intake (recurrence hazard ratio [HR]=0.91; 95% CI, 0.72-1.14; mortality HR=0.78; 95% CI, 0.69-0.88). In postmenopausal women, higher intake was likewise associated with improvement of both outcomes (recurrence HR=0.67; 95% CI, 0.56-0.80; mortality HR=0.81; 95% CI, 0.73-0.91).

An earlier meta-analysis of 4 prospective cohort studies, 2 of which were not included above, also found reduced risk of breast cancer recurrence in groups with high vs low soy isoflavone intake (HR=0.84; 95% CI, 0.70-0.99).⁷ Women taking tamoxifen showed no difference in mortality or recurrence risk associated with soy intake.

An additional small prospective cohort study (n=256) found similar reductions in recurrence and mortality associated with higher consumption of soy protein.⁸

EVIDENCE SUMMARY

A large systematic review evaluated the relationship between dietary soy intake and risk of a primary breast cancer diagnosis. It included 7 prospective cohort studies, which comprised the best quality evidence available (numerous other reviewed studies were of lower quality). The review found no significant association between dietary soy intake and primary breast cancer (TABLE^1-6).

Investigators either surveyed women for intake of soy isoflavones or soy foods or products (tofu, soybeans, lentils, miso) or measured urinary or plasma levels of soy isoflavones. They adjusted for age, alcohol use, smoking status, body mass index, caloric intake, and hormone replacement therapy, then followed subjects for 7 to 23 years, comparing the risk of breast cancer for the lowest and highest levels of soy intake. 

Six of the prospective cohort studies found no association between soy intake and breast cancer risk; one study, comprising 4% of the total population, found a lower risk with higher soy intake (effect size=0.44; 95% confidence interval [CI], 0.26-0.73; an effect size of 0.2 is considered small, 0.6 medium, and 1.2 large). The authors didn’t do a meta-analysis of the prospective cohort studies.

Other cohort studies yield similar findings

Four other large systematic reviews evaluating soy intake and breast cancer risk incorporated a total of 6 individual prospective cohort studies that weren’t included in the previously described review (again, these studies comprised the best quality evidence within the reviews). The 6 studies found no association between soy intake and breast cancer risk.

In 2 of the studies, investigators surveyed postmenopausal women and followed them for 4 to 8 years.² Investigators in another study adjusted for age, family and gynecologic history, hormone and medication use, exercise, and other factors.³ In 2 other studies, investigators evaluated population subsets that consumed the most vs the fewest servings per week or kilograms per year of soy foods.⁴ The sixth study compared low with high intake of soy foods and miso.⁵

Soy intake after breast cancer diagnosis reduces recurrence risk in most studies

Most prospective cohort studies evaluating the association between dietary soy intake after breast cancer diagnosis found an overall 21% decrease in recurrence with high soy intake and a 15% reduction in mortality (TABLE^1-6).

Investigators in a meta-analysis of 5 studies that followed women for 4 to 7 years after first breast cancer diagnosis found that higher soy intake was associated with lower mortality but not less recurrence in women who were estrogen receptor positive. Both recurrence and mortality were decreased in estrogen receptor negative women.⁶

The study also found lower recurrence and mortality in premenopausal women with higher soy intake (recurrence hazard ratio [HR]=0.91; 95% CI, 0.72-1.14; mortality HR=0.78; 95% CI, 0.69-0.88). In postmenopausal women, higher intake was likewise associated with improvement of both outcomes (recurrence HR=0.67; 95% CI, 0.56-0.80; mortality HR=0.81; 95% CI, 0.73-0.91).

An earlier meta-analysis of 4 prospective cohort studies, 2 of which were not included above, also found reduced risk of breast cancer recurrence in groups with high vs low soy isoflavone intake (HR=0.84; 95% CI, 0.70-0.99).⁷ Women taking tamoxifen showed no difference in mortality or recurrence risk associated with soy intake.

An additional small prospective cohort study (n=256) found similar reductions in recurrence and mortality associated with higher consumption of soy protein.⁸

EVIDENCE SUMMARY

A large systematic review evaluated the relationship between dietary soy intake and risk of a primary breast cancer diagnosis. It included 7 prospective cohort studies, which comprised the best quality evidence available (numerous other reviewed studies were of lower quality). The review found no significant association between dietary soy intake and primary breast cancer (TABLE^1-6).

Investigators either surveyed women for intake of soy isoflavones or soy foods or products (tofu, soybeans, lentils, miso) or measured urinary or plasma levels of soy isoflavones. They adjusted for age, alcohol use, smoking status, body mass index, caloric intake, and hormone replacement therapy, then followed subjects for 7 to 23 years, comparing the risk of breast cancer for the lowest and highest levels of soy intake. 

Six of the prospective cohort studies found no association between soy intake and breast cancer risk; one study, comprising 4% of the total population, found a lower risk with higher soy intake (effect size=0.44; 95% confidence interval [CI], 0.26-0.73; an effect size of 0.2 is considered small, 0.6 medium, and 1.2 large). The authors didn’t do a meta-analysis of the prospective cohort studies.

Other cohort studies yield similar findings

Four other large systematic reviews evaluating soy intake and breast cancer risk incorporated a total of 6 individual prospective cohort studies that weren’t included in the previously described review (again, these studies comprised the best quality evidence within the reviews). The 6 studies found no association between soy intake and breast cancer risk.

In 2 of the studies, investigators surveyed postmenopausal women and followed them for 4 to 8 years.² Investigators in another study adjusted for age, family and gynecologic history, hormone and medication use, exercise, and other factors.³ In 2 other studies, investigators evaluated population subsets that consumed the most vs the fewest servings per week or kilograms per year of soy foods.⁴ The sixth study compared low with high intake of soy foods and miso.⁵

Soy intake after breast cancer diagnosis reduces recurrence risk in most studies

Most prospective cohort studies evaluating the association between dietary soy intake after breast cancer diagnosis found an overall 21% decrease in recurrence with high soy intake and a 15% reduction in mortality (TABLE^1-6).

Investigators in a meta-analysis of 5 studies that followed women for 4 to 7 years after first breast cancer diagnosis found that higher soy intake was associated with lower mortality but not less recurrence in women who were estrogen receptor positive. Both recurrence and mortality were decreased in estrogen receptor negative women.⁶

The study also found lower recurrence and mortality in premenopausal women with higher soy intake (recurrence hazard ratio [HR]=0.91; 95% CI, 0.72-1.14; mortality HR=0.78; 95% CI, 0.69-0.88). In postmenopausal women, higher intake was likewise associated with improvement of both outcomes (recurrence HR=0.67; 95% CI, 0.56-0.80; mortality HR=0.81; 95% CI, 0.73-0.91).

An earlier meta-analysis of 4 prospective cohort studies, 2 of which were not included above, also found reduced risk of breast cancer recurrence in groups with high vs low soy isoflavone intake (HR=0.84; 95% CI, 0.70-0.99).⁷ Women taking tamoxifen showed no difference in mortality or recurrence risk associated with soy intake.

An additional small prospective cohort study (n=256) found similar reductions in recurrence and mortality associated with higher consumption of soy protein.⁸

EVIDENCE SUMMARY

Two RCTs that evaluated qHPV in young men for preventing outcomes associated with the 4 HPV subtypes in the vaccine (6, 11, 16, and 18) found that it reduced them by 50% to 66% using an intention-to-treat protocol (TABLE^1-4).

Vaccination reduces AIN and persistent infection in MSM

The first RCT evaluated a subset of 602 MSM from the second, larger RCT for preventing AIN and persistent HPV infection.¹ The intention-to-treat population included men with 5 or fewer lifetime sexual partners who had engaged in insertive or receptive anal intercourse or oral sex within the last year, were not necessarily HPV-negative at enrollment, and received at least one dose of vaccine (or placebo).

The vaccine reduced AIN associated with the 4 HPV types (6.3 vs 12.6 events per 100 person-years; relative risk reduction [RRR]=50.3%; 95% confidence interval [CI], 25.7-67.2; number needed to treat [NNT]=16 to prevent one AIN case per year) and with HPV of any type (13 vs 17.5 events per 100 person-years; RRR=25.7%; 95% CI, -1.1 to 45.6). It also reduced the rate of persistent HPV infection with the 4 HPV vaccine subtypes (8.8 vs 21.6 events per 100 person-years; RRR=59.4%; 95% CI, 43%-71%; NNT=8 to prevent one persistent HPV infection per year).

Investigators in the study also evaluated vaccine efficacy in a smaller subset (194 men) using per-protocol analysis and found higher prevention rates (78% for AIN due to HPV types 6, 11, 16, and 18). Investigators followed these subjects every 6 months for 36 months with polymerase chain reaction testing for HPV DNA, high-resolution anoscopy with anal cytology, and anal biopsy and histology if there were atypia.

The vaccine decreases persistent HPV infection and external genital lesions

The second RCT, including both MSM and heterosexual men, found that qHPV vaccine reduced rates of persistent HPV infection by 48%, and external genital lesions (condylomata or intraepithelial neoplasia involving the penis, perineum, or perianal area) by 66% associated with HPV types 6, 11, 16, and 18 using the intention-to-treat protocol.²

Investigators used the same protocols used in the first RCT, and the per-protocol population again had higher prevention rates (84% for any HPV type, 90% against the 4 vaccine types). The only adverse effect of the vaccine was injection site pain (57% vs 51% with placebo; P<.001).

The vaccine also helps older MSM

A nonconcurrent cohort study that evaluated qHPV vaccination among older MSM with previously treated high-grade AIN found a 50% decrease in recurrence rates in the 2 years after vaccination.³ Investigators recruited HIV-negative men, some of whom chose vaccination (not randomized), and followed them for 2 years. Study limitations included using medical records for data collection and the predominance of white, nonsmoking men with private insurance.

A post-hoc analysis of older men without previous anal condylomata (210 men) or with treated condylomata and no recurrence in the year before vaccination (103 men) found that qHPV vaccination was associated with 55% lower rates of anal condylomata.⁴

RECOMMENDATIONS

The Centers for Disease Control and Prevention’s Advisory Committee on Immunization Practices recommends routine use of qHPV vaccine in males ages 11 through 21 years, and optional use in unvaccinated men as old as 26 years.⁵

EVIDENCE SUMMARY

Two RCTs that evaluated qHPV in young men for preventing outcomes associated with the 4 HPV subtypes in the vaccine (6, 11, 16, and 18) found that it reduced them by 50% to 66% using an intention-to-treat protocol (TABLE^1-4).

Vaccination reduces AIN and persistent infection in MSM

The first RCT evaluated a subset of 602 MSM from the second, larger RCT for preventing AIN and persistent HPV infection.¹ The intention-to-treat population included men with 5 or fewer lifetime sexual partners who had engaged in insertive or receptive anal intercourse or oral sex within the last year, were not necessarily HPV-negative at enrollment, and received at least one dose of vaccine (or placebo).

The vaccine reduced AIN associated with the 4 HPV types (6.3 vs 12.6 events per 100 person-years; relative risk reduction [RRR]=50.3%; 95% confidence interval [CI], 25.7-67.2; number needed to treat [NNT]=16 to prevent one AIN case per year) and with HPV of any type (13 vs 17.5 events per 100 person-years; RRR=25.7%; 95% CI, -1.1 to 45.6). It also reduced the rate of persistent HPV infection with the 4 HPV vaccine subtypes (8.8 vs 21.6 events per 100 person-years; RRR=59.4%; 95% CI, 43%-71%; NNT=8 to prevent one persistent HPV infection per year).

Investigators in the study also evaluated vaccine efficacy in a smaller subset (194 men) using per-protocol analysis and found higher prevention rates (78% for AIN due to HPV types 6, 11, 16, and 18). Investigators followed these subjects every 6 months for 36 months with polymerase chain reaction testing for HPV DNA, high-resolution anoscopy with anal cytology, and anal biopsy and histology if there were atypia.

The vaccine decreases persistent HPV infection and external genital lesions

The second RCT, including both MSM and heterosexual men, found that qHPV vaccine reduced rates of persistent HPV infection by 48%, and external genital lesions (condylomata or intraepithelial neoplasia involving the penis, perineum, or perianal area) by 66% associated with HPV types 6, 11, 16, and 18 using the intention-to-treat protocol.²

Investigators used the same protocols used in the first RCT, and the per-protocol population again had higher prevention rates (84% for any HPV type, 90% against the 4 vaccine types). The only adverse effect of the vaccine was injection site pain (57% vs 51% with placebo; P<.001).

The vaccine also helps older MSM

A nonconcurrent cohort study that evaluated qHPV vaccination among older MSM with previously treated high-grade AIN found a 50% decrease in recurrence rates in the 2 years after vaccination.³ Investigators recruited HIV-negative men, some of whom chose vaccination (not randomized), and followed them for 2 years. Study limitations included using medical records for data collection and the predominance of white, nonsmoking men with private insurance.

A post-hoc analysis of older men without previous anal condylomata (210 men) or with treated condylomata and no recurrence in the year before vaccination (103 men) found that qHPV vaccination was associated with 55% lower rates of anal condylomata.⁴

RECOMMENDATIONS

The Centers for Disease Control and Prevention’s Advisory Committee on Immunization Practices recommends routine use of qHPV vaccine in males ages 11 through 21 years, and optional use in unvaccinated men as old as 26 years.⁵

EVIDENCE SUMMARY

Two RCTs that evaluated qHPV in young men for preventing outcomes associated with the 4 HPV subtypes in the vaccine (6, 11, 16, and 18) found that it reduced them by 50% to 66% using an intention-to-treat protocol (TABLE^1-4).

Vaccination reduces AIN and persistent infection in MSM

The first RCT evaluated a subset of 602 MSM from the second, larger RCT for preventing AIN and persistent HPV infection.¹ The intention-to-treat population included men with 5 or fewer lifetime sexual partners who had engaged in insertive or receptive anal intercourse or oral sex within the last year, were not necessarily HPV-negative at enrollment, and received at least one dose of vaccine (or placebo).

The vaccine reduced AIN associated with the 4 HPV types (6.3 vs 12.6 events per 100 person-years; relative risk reduction [RRR]=50.3%; 95% confidence interval [CI], 25.7-67.2; number needed to treat [NNT]=16 to prevent one AIN case per year) and with HPV of any type (13 vs 17.5 events per 100 person-years; RRR=25.7%; 95% CI, -1.1 to 45.6). It also reduced the rate of persistent HPV infection with the 4 HPV vaccine subtypes (8.8 vs 21.6 events per 100 person-years; RRR=59.4%; 95% CI, 43%-71%; NNT=8 to prevent one persistent HPV infection per year).

Investigators in the study also evaluated vaccine efficacy in a smaller subset (194 men) using per-protocol analysis and found higher prevention rates (78% for AIN due to HPV types 6, 11, 16, and 18). Investigators followed these subjects every 6 months for 36 months with polymerase chain reaction testing for HPV DNA, high-resolution anoscopy with anal cytology, and anal biopsy and histology if there were atypia.

The vaccine decreases persistent HPV infection and external genital lesions

The second RCT, including both MSM and heterosexual men, found that qHPV vaccine reduced rates of persistent HPV infection by 48%, and external genital lesions (condylomata or intraepithelial neoplasia involving the penis, perineum, or perianal area) by 66% associated with HPV types 6, 11, 16, and 18 using the intention-to-treat protocol.²

Investigators used the same protocols used in the first RCT, and the per-protocol population again had higher prevention rates (84% for any HPV type, 90% against the 4 vaccine types). The only adverse effect of the vaccine was injection site pain (57% vs 51% with placebo; P<.001).

The vaccine also helps older MSM

A nonconcurrent cohort study that evaluated qHPV vaccination among older MSM with previously treated high-grade AIN found a 50% decrease in recurrence rates in the 2 years after vaccination.³ Investigators recruited HIV-negative men, some of whom chose vaccination (not randomized), and followed them for 2 years. Study limitations included using medical records for data collection and the predominance of white, nonsmoking men with private insurance.

A post-hoc analysis of older men without previous anal condylomata (210 men) or with treated condylomata and no recurrence in the year before vaccination (103 men) found that qHPV vaccination was associated with 55% lower rates of anal condylomata.⁴

RECOMMENDATIONS

The Centers for Disease Control and Prevention’s Advisory Committee on Immunization Practices recommends routine use of qHPV vaccine in males ages 11 through 21 years, and optional use in unvaccinated men as old as 26 years.⁵

EVIDENCE SUMMARY

Women with PCOS who are not seeking pregnancy commonly have symptoms such as excessive weight, hirsutism, and menstrual irregularities. This review focuses on interventions to manage those symptoms. The TABLE summarizes the results of the interventions.

Lifestyle modification improves symptoms; no benefit to adding metformin

A Cochrane meta-analysis of 6 RCTs with 164 patients compared LSM (with diet and exercise) and no or minimal intervention. LSM reduced weight more than minimal intervention (mean difference [MD]=-3.5 kg; 95% confidence interval [CI], -4.5 to -2.0).¹ It also improved hirsutism, assessed with the 36-point FG score, where a lower score corresponds to less hirsutism (MD=-1.2 points, 95% CI, -2.4 to -0.1). No data were available on menstrual regularity.

A double-blind RCT comparing LSM alone with LSM plus metformin in 114 patients with PCOS found no difference in mean weight reduction (-2- to -3 kg, data from graph), ovulation rate, or androgen levels at 6 months.² Six patients dropped out of the LSM-with-metformin group, whereas no patients dropped out of the LSM-alone group.

Metformin decreases BMI more than thiazolidinediones

In a meta-analysis of 10 RCTs (459 patients) comparing the effects of metformin and thiazolidinediones (TZDs), metformin reduced body mass index (BMI) more than TZDs at 3 months (weighted mean difference [WMD]=-2.5 kg/m²; 95% CI, -3.3 to -.6) and 6 months (WMD=-0.70 kg/m²; 95% CI, -0.76 to -0.65).³

In a prospective cohort dose-comparison study, 201 women with PCOS received either metformin 1000 mg or 1500 to 1700 mg daily for 6 months. Patients were asked not to modify their diet or exercise routines. In both dosage groups, patients lost weight from baseline (-3 kg; P<.01), and the number of menstrual cycles increased (0.7 per 6 months; P<.001).⁴ No clear dose-response relationship was observed.

Spironolactone can significantly reduce hirsutism

A systematic review identified 4 studies (132 patients) of antiandrogen therapy for hirsutism in PCOS. The 3 studies that used the FG score as an outcome all showed significant reductions in hirsutism after 6 to 12 months of treatment with spironolactone.⁵

Lifestyle modification and metformin both reduce body weight, but lifestyle modification may be better tolerated.

A 6-month RCT of 198 patients with PCOS compared outcomes for spironolactone (50 mg/d), metformin (1000 mg/d), or both. Combined therapy was marginally better than either agent alone for reducing the FG score (end score for combined therapy 9.1 vs 9.6 for spironolactone and 9.7 for metformin, an absolute difference for combined therapy vs spironolactone of -0.5 FG points or -1.4%; P<.05).⁶

OCPs normalize menstrual cycles and reduce hirsutism

A Cochrane review evaluating the effects of OCPs on patients with PCOS included 4 RCTs (104 patients) that compared OCPs with metformin (1500-2000 mg/d) and 2 RCTs (70 patients) that compared the combination of an OCP and metformin with the OCP alone. Use of an OCP was much more likely to normalize menstrual cycling than metformin alone (2 trials, N=35; odds ratio [OR]=12; 95% CI, 2.2-100). Combining an OCP with metformin resulted in slightly better FG scores than an OCP alone (1 trial, N=40; WMD=-2.8 points; 95% CI, -5.4 to -0.17).⁷ There was no difference in the final BMI between patients taking an OCP alone, metformin alone, or both.

An RCT of 35 patients compared the effect on insulin levels of an OCP with rosiglitazone 4 mg/d and also looked at menstrual cycling as a secondary outcome. The study found no difference in effect on insulin levels in the 2 groups. All patients taking the OCP reported regular menstrual cycles at the end of the study compared with 75% of the patients taking rosiglitazone (P=.7).⁸ The study was underpowered to find a difference, however.

Statins alone don’t affect hirsutism, menstruation, or BMI

A Cochrane review identified 4 RCTs (244 women, ages 18-39 years) that compared a statin alone with placebo, another agent, or another agent plus a statin.⁹ One RCT of 48 patients found that a statin combined with an OCP improved hirsutism compared with a statin alone. Two RCTs (85 patients) found that statins didn’t lead to resumption of regular menstrual cycles. Statins also didn’t alter BMI in 3 studies of 105 patients.

Trials report no adverse effects, but VTE may be a concern with OCPs

A meta-analysis evaluated the safety of metformin, OCPs, and antiandrogens in 22 clinical trials with 1335 patients, primarily PCOS patients. The trials reported no cases of lactic acidosis with metformin, no drug-induced liver injury with antiandrogens, and no venous thromboembolism (VTE) with OCPs. The meta-analysis authors noted, however, that in a cohort trial of 1.6 million Danish women followed for 15 years, OCPs were associated with a 2- to 3-fold increase in risk of VTE, with higher risks linked to higher ethinyl estradiol content.¹⁰

RECOMMENDATIONS

A 2009 practice bulletin from The American College of Obestetrics and Gynecology (ACOG) recommends OCPs, progestin, metformin, and TZDs for anovulation and amenorrhea in patients with PCOS. OCPs, antiandrogens, metformin, eflornithine, and mechanical hair removal are recommended for hirsutism. ACOG advocates LSM, insulin-sensitizing agents (such as metformin), and statins to prevent cardiovascular disease and diabetes.¹¹

EVIDENCE SUMMARY

Women with PCOS who are not seeking pregnancy commonly have symptoms such as excessive weight, hirsutism, and menstrual irregularities. This review focuses on interventions to manage those symptoms. The TABLE summarizes the results of the interventions.

Lifestyle modification improves symptoms; no benefit to adding metformin

A Cochrane meta-analysis of 6 RCTs with 164 patients compared LSM (with diet and exercise) and no or minimal intervention. LSM reduced weight more than minimal intervention (mean difference [MD]=-3.5 kg; 95% confidence interval [CI], -4.5 to -2.0).¹ It also improved hirsutism, assessed with the 36-point FG score, where a lower score corresponds to less hirsutism (MD=-1.2 points, 95% CI, -2.4 to -0.1). No data were available on menstrual regularity.

A double-blind RCT comparing LSM alone with LSM plus metformin in 114 patients with PCOS found no difference in mean weight reduction (-2- to -3 kg, data from graph), ovulation rate, or androgen levels at 6 months.² Six patients dropped out of the LSM-with-metformin group, whereas no patients dropped out of the LSM-alone group.

Metformin decreases BMI more than thiazolidinediones

In a meta-analysis of 10 RCTs (459 patients) comparing the effects of metformin and thiazolidinediones (TZDs), metformin reduced body mass index (BMI) more than TZDs at 3 months (weighted mean difference [WMD]=-2.5 kg/m²; 95% CI, -3.3 to -.6) and 6 months (WMD=-0.70 kg/m²; 95% CI, -0.76 to -0.65).³

In a prospective cohort dose-comparison study, 201 women with PCOS received either metformin 1000 mg or 1500 to 1700 mg daily for 6 months. Patients were asked not to modify their diet or exercise routines. In both dosage groups, patients lost weight from baseline (-3 kg; P<.01), and the number of menstrual cycles increased (0.7 per 6 months; P<.001).⁴ No clear dose-response relationship was observed.

Spironolactone can significantly reduce hirsutism

A systematic review identified 4 studies (132 patients) of antiandrogen therapy for hirsutism in PCOS. The 3 studies that used the FG score as an outcome all showed significant reductions in hirsutism after 6 to 12 months of treatment with spironolactone.⁵

Lifestyle modification and metformin both reduce body weight, but lifestyle modification may be better tolerated.

A 6-month RCT of 198 patients with PCOS compared outcomes for spironolactone (50 mg/d), metformin (1000 mg/d), or both. Combined therapy was marginally better than either agent alone for reducing the FG score (end score for combined therapy 9.1 vs 9.6 for spironolactone and 9.7 for metformin, an absolute difference for combined therapy vs spironolactone of -0.5 FG points or -1.4%; P<.05).⁶

OCPs normalize menstrual cycles and reduce hirsutism

A Cochrane review evaluating the effects of OCPs on patients with PCOS included 4 RCTs (104 patients) that compared OCPs with metformin (1500-2000 mg/d) and 2 RCTs (70 patients) that compared the combination of an OCP and metformin with the OCP alone. Use of an OCP was much more likely to normalize menstrual cycling than metformin alone (2 trials, N=35; odds ratio [OR]=12; 95% CI, 2.2-100). Combining an OCP with metformin resulted in slightly better FG scores than an OCP alone (1 trial, N=40; WMD=-2.8 points; 95% CI, -5.4 to -0.17).⁷ There was no difference in the final BMI between patients taking an OCP alone, metformin alone, or both.

An RCT of 35 patients compared the effect on insulin levels of an OCP with rosiglitazone 4 mg/d and also looked at menstrual cycling as a secondary outcome. The study found no difference in effect on insulin levels in the 2 groups. All patients taking the OCP reported regular menstrual cycles at the end of the study compared with 75% of the patients taking rosiglitazone (P=.7).⁸ The study was underpowered to find a difference, however.

Statins alone don’t affect hirsutism, menstruation, or BMI

A Cochrane review identified 4 RCTs (244 women, ages 18-39 years) that compared a statin alone with placebo, another agent, or another agent plus a statin.⁹ One RCT of 48 patients found that a statin combined with an OCP improved hirsutism compared with a statin alone. Two RCTs (85 patients) found that statins didn’t lead to resumption of regular menstrual cycles. Statins also didn’t alter BMI in 3 studies of 105 patients.

Trials report no adverse effects, but VTE may be a concern with OCPs

A meta-analysis evaluated the safety of metformin, OCPs, and antiandrogens in 22 clinical trials with 1335 patients, primarily PCOS patients. The trials reported no cases of lactic acidosis with metformin, no drug-induced liver injury with antiandrogens, and no venous thromboembolism (VTE) with OCPs. The meta-analysis authors noted, however, that in a cohort trial of 1.6 million Danish women followed for 15 years, OCPs were associated with a 2- to 3-fold increase in risk of VTE, with higher risks linked to higher ethinyl estradiol content.¹⁰

RECOMMENDATIONS

A 2009 practice bulletin from The American College of Obestetrics and Gynecology (ACOG) recommends OCPs, progestin, metformin, and TZDs for anovulation and amenorrhea in patients with PCOS. OCPs, antiandrogens, metformin, eflornithine, and mechanical hair removal are recommended for hirsutism. ACOG advocates LSM, insulin-sensitizing agents (such as metformin), and statins to prevent cardiovascular disease and diabetes.¹¹

EVIDENCE SUMMARY

Women with PCOS who are not seeking pregnancy commonly have symptoms such as excessive weight, hirsutism, and menstrual irregularities. This review focuses on interventions to manage those symptoms. The TABLE summarizes the results of the interventions.

Lifestyle modification improves symptoms; no benefit to adding metformin

A Cochrane meta-analysis of 6 RCTs with 164 patients compared LSM (with diet and exercise) and no or minimal intervention. LSM reduced weight more than minimal intervention (mean difference [MD]=-3.5 kg; 95% confidence interval [CI], -4.5 to -2.0).¹ It also improved hirsutism, assessed with the 36-point FG score, where a lower score corresponds to less hirsutism (MD=-1.2 points, 95% CI, -2.4 to -0.1). No data were available on menstrual regularity.

A double-blind RCT comparing LSM alone with LSM plus metformin in 114 patients with PCOS found no difference in mean weight reduction (-2- to -3 kg, data from graph), ovulation rate, or androgen levels at 6 months.² Six patients dropped out of the LSM-with-metformin group, whereas no patients dropped out of the LSM-alone group.

Metformin decreases BMI more than thiazolidinediones

In a meta-analysis of 10 RCTs (459 patients) comparing the effects of metformin and thiazolidinediones (TZDs), metformin reduced body mass index (BMI) more than TZDs at 3 months (weighted mean difference [WMD]=-2.5 kg/m²; 95% CI, -3.3 to -.6) and 6 months (WMD=-0.70 kg/m²; 95% CI, -0.76 to -0.65).³

In a prospective cohort dose-comparison study, 201 women with PCOS received either metformin 1000 mg or 1500 to 1700 mg daily for 6 months. Patients were asked not to modify their diet or exercise routines. In both dosage groups, patients lost weight from baseline (-3 kg; P<.01), and the number of menstrual cycles increased (0.7 per 6 months; P<.001).⁴ No clear dose-response relationship was observed.

Spironolactone can significantly reduce hirsutism

A systematic review identified 4 studies (132 patients) of antiandrogen therapy for hirsutism in PCOS. The 3 studies that used the FG score as an outcome all showed significant reductions in hirsutism after 6 to 12 months of treatment with spironolactone.⁵

Lifestyle modification and metformin both reduce body weight, but lifestyle modification may be better tolerated.

A 6-month RCT of 198 patients with PCOS compared outcomes for spironolactone (50 mg/d), metformin (1000 mg/d), or both. Combined therapy was marginally better than either agent alone for reducing the FG score (end score for combined therapy 9.1 vs 9.6 for spironolactone and 9.7 for metformin, an absolute difference for combined therapy vs spironolactone of -0.5 FG points or -1.4%; P<.05).⁶

OCPs normalize menstrual cycles and reduce hirsutism

A Cochrane review evaluating the effects of OCPs on patients with PCOS included 4 RCTs (104 patients) that compared OCPs with metformin (1500-2000 mg/d) and 2 RCTs (70 patients) that compared the combination of an OCP and metformin with the OCP alone. Use of an OCP was much more likely to normalize menstrual cycling than metformin alone (2 trials, N=35; odds ratio [OR]=12; 95% CI, 2.2-100). Combining an OCP with metformin resulted in slightly better FG scores than an OCP alone (1 trial, N=40; WMD=-2.8 points; 95% CI, -5.4 to -0.17).⁷ There was no difference in the final BMI between patients taking an OCP alone, metformin alone, or both.

An RCT of 35 patients compared the effect on insulin levels of an OCP with rosiglitazone 4 mg/d and also looked at menstrual cycling as a secondary outcome. The study found no difference in effect on insulin levels in the 2 groups. All patients taking the OCP reported regular menstrual cycles at the end of the study compared with 75% of the patients taking rosiglitazone (P=.7).⁸ The study was underpowered to find a difference, however.

Statins alone don’t affect hirsutism, menstruation, or BMI

A Cochrane review identified 4 RCTs (244 women, ages 18-39 years) that compared a statin alone with placebo, another agent, or another agent plus a statin.⁹ One RCT of 48 patients found that a statin combined with an OCP improved hirsutism compared with a statin alone. Two RCTs (85 patients) found that statins didn’t lead to resumption of regular menstrual cycles. Statins also didn’t alter BMI in 3 studies of 105 patients.

Trials report no adverse effects, but VTE may be a concern with OCPs

A meta-analysis evaluated the safety of metformin, OCPs, and antiandrogens in 22 clinical trials with 1335 patients, primarily PCOS patients. The trials reported no cases of lactic acidosis with metformin, no drug-induced liver injury with antiandrogens, and no venous thromboembolism (VTE) with OCPs. The meta-analysis authors noted, however, that in a cohort trial of 1.6 million Danish women followed for 15 years, OCPs were associated with a 2- to 3-fold increase in risk of VTE, with higher risks linked to higher ethinyl estradiol content.¹⁰

RECOMMENDATIONS

A 2009 practice bulletin from The American College of Obestetrics and Gynecology (ACOG) recommends OCPs, progestin, metformin, and TZDs for anovulation and amenorrhea in patients with PCOS. OCPs, antiandrogens, metformin, eflornithine, and mechanical hair removal are recommended for hirsutism. ACOG advocates LSM, insulin-sensitizing agents (such as metformin), and statins to prevent cardiovascular disease and diabetes.¹¹