Clinical Inquiries

Evidence summary

A double-blind RCT comparing chasteberry tree with placebo in 170 patients found a decrease in self-reported PMS symptom scores and an increase in response rate (defined as a 50% reduction in symptoms)—52% vs 24%—in the intervention group (number needed to treat [NNT]=3.5). Patients taking chasteberry tree had 1 occurrence of multiple abscesses and 1 of urticaria.¹

A prospective, open-label study of chasteberry tree for PMS symptoms in 43 patients found a 42% decrease in self-assessed PMS symptom scores, with the greatest improvement in pain, behavior changes, negative feelings, and fluid retention. No serious adverse events occurred.²

A third study comparing chasteberry tree with fluoxetine in 19 patients found a decrease in premenstrual symptom scores for both fluoxetine (13 of 19 patients) and chasteberry tree (11 of 19 patients). No statistically significant differences were noted between the 2 groups. Chasteberry tree was well tolerated; most adverse effects occurred in patients receiving fluoxetine. The most frequent adverse effects with chasteberry tree were nausea in 5 patients and headache in 4.³

Symptoms decrease significantly after 3 calcium treatment cycles
Two RCTs (33 and 466 patients) comparing 1000 and 1200 mg of calcium with placebo found a significant decrease in PMS symptoms after 3 treatment cycles.^4,5 Calcium improved negative affect, water retention, food cravings, and pain. In the first study, 73% of patients preferred taking calcium, compared with 15% who preferred placebo.

The second study found that, by the third treatment cycle, patients taking calcium had an overall 48% reduction in total symptom scores, compared with a 30% reduction in the control group. The most common adverse effects were headache, rhinitis, and nonspecific pain.

Watch out for neuropathy with high doses of pyridoxine
A meta-analysis of pyridoxine in doses from 50 to 600 mg per day for PMS included 9 RCTs. Relative to placebo, pyridoxine improved PMS symptom scores (odds ratio=2.32, 95% confidence interval, 1.95-2.54). The overall quality of studies was poor, however, with few subjects, widely varying doses, and different outcome measurements.⁶

Two subsequent RCTs of 40 and 96 patients that weren’t included in the meta-analysis failed to demonstrate reduced premenstrual symptoms.^7,8 Long-term use of pyridoxine in doses >200 mg/d can cause neuropathy.

Saffron shows promise in small study
A recent double-blind RCT evaluated the effect of 2 cycles of treatment with saffron (Crocus sativus L), 30 mg twice daily, on PMS symptoms in 50 patients. Nineteen patients in the saffron group showed a response, defined as 50% reduction in symptom severity, compared with 2 patients in the placebo group (NNT=2). The study found no statistically significant difference in frequency of adverse effects.⁹

Evidence for magnesium is sparse
Two RCTs comparing magnesium with placebo had low precision because of small numbers and short treatment duration.^10,11 The first (N=28) demonstrated reduced total Moos Menstrual Distress Questionnaire scores.¹⁰ The second study reported a decrease in fluid retention symptoms by 2 points on an 80-point scale (P<.009) at 2 months, but no difference in total score.¹¹

A further study, begun as an open trial of magnesium infusion for premenstrual dysphoric disorder (N=6), found a dramatic reduction in mood symptom scores. After converting to a randomized, blinded design (N=10), no difference was found compared with placebo.¹²

St. John’s wort, evening primrose oil don’t work
One randomized, double-blind controlled trial (N=125) of 600 mg St. John’s wort vs placebo over 2 cycles of treatment found no significant changes in symptom score from baseline.¹³ Two double-blind crossover studies of 27 and 38 patients found that evening primrose oil had no effect on PMS symptoms.^14,15

Recommendations

The Premenstrual Syndrome Guidelines of the American College of Obstetricians and Gynecologists (ACOG) state that calcium and magnesium have been shown to be effective in small trials and must be validated in larger trials before a strong evidence-based recommendation can be made. ACOG’s guidelines also report minimal effectiveness with vitamin B6 and vitamin E.¹⁶

Evidence summary

A double-blind RCT comparing chasteberry tree with placebo in 170 patients found a decrease in self-reported PMS symptom scores and an increase in response rate (defined as a 50% reduction in symptoms)—52% vs 24%—in the intervention group (number needed to treat [NNT]=3.5). Patients taking chasteberry tree had 1 occurrence of multiple abscesses and 1 of urticaria.¹

A prospective, open-label study of chasteberry tree for PMS symptoms in 43 patients found a 42% decrease in self-assessed PMS symptom scores, with the greatest improvement in pain, behavior changes, negative feelings, and fluid retention. No serious adverse events occurred.²

A third study comparing chasteberry tree with fluoxetine in 19 patients found a decrease in premenstrual symptom scores for both fluoxetine (13 of 19 patients) and chasteberry tree (11 of 19 patients). No statistically significant differences were noted between the 2 groups. Chasteberry tree was well tolerated; most adverse effects occurred in patients receiving fluoxetine. The most frequent adverse effects with chasteberry tree were nausea in 5 patients and headache in 4.³

Symptoms decrease significantly after 3 calcium treatment cycles
Two RCTs (33 and 466 patients) comparing 1000 and 1200 mg of calcium with placebo found a significant decrease in PMS symptoms after 3 treatment cycles.^4,5 Calcium improved negative affect, water retention, food cravings, and pain. In the first study, 73% of patients preferred taking calcium, compared with 15% who preferred placebo.

The second study found that, by the third treatment cycle, patients taking calcium had an overall 48% reduction in total symptom scores, compared with a 30% reduction in the control group. The most common adverse effects were headache, rhinitis, and nonspecific pain.

Watch out for neuropathy with high doses of pyridoxine
A meta-analysis of pyridoxine in doses from 50 to 600 mg per day for PMS included 9 RCTs. Relative to placebo, pyridoxine improved PMS symptom scores (odds ratio=2.32, 95% confidence interval, 1.95-2.54). The overall quality of studies was poor, however, with few subjects, widely varying doses, and different outcome measurements.⁶

Two subsequent RCTs of 40 and 96 patients that weren’t included in the meta-analysis failed to demonstrate reduced premenstrual symptoms.^7,8 Long-term use of pyridoxine in doses >200 mg/d can cause neuropathy.

Saffron shows promise in small study
A recent double-blind RCT evaluated the effect of 2 cycles of treatment with saffron (Crocus sativus L), 30 mg twice daily, on PMS symptoms in 50 patients. Nineteen patients in the saffron group showed a response, defined as 50% reduction in symptom severity, compared with 2 patients in the placebo group (NNT=2). The study found no statistically significant difference in frequency of adverse effects.⁹

Evidence for magnesium is sparse
Two RCTs comparing magnesium with placebo had low precision because of small numbers and short treatment duration.^10,11 The first (N=28) demonstrated reduced total Moos Menstrual Distress Questionnaire scores.¹⁰ The second study reported a decrease in fluid retention symptoms by 2 points on an 80-point scale (P<.009) at 2 months, but no difference in total score.¹¹

A further study, begun as an open trial of magnesium infusion for premenstrual dysphoric disorder (N=6), found a dramatic reduction in mood symptom scores. After converting to a randomized, blinded design (N=10), no difference was found compared with placebo.¹²

St. John’s wort, evening primrose oil don’t work
One randomized, double-blind controlled trial (N=125) of 600 mg St. John’s wort vs placebo over 2 cycles of treatment found no significant changes in symptom score from baseline.¹³ Two double-blind crossover studies of 27 and 38 patients found that evening primrose oil had no effect on PMS symptoms.^14,15

Recommendations

The Premenstrual Syndrome Guidelines of the American College of Obstetricians and Gynecologists (ACOG) state that calcium and magnesium have been shown to be effective in small trials and must be validated in larger trials before a strong evidence-based recommendation can be made. ACOG’s guidelines also report minimal effectiveness with vitamin B6 and vitamin E.¹⁶

Evidence summary

A double-blind RCT comparing chasteberry tree with placebo in 170 patients found a decrease in self-reported PMS symptom scores and an increase in response rate (defined as a 50% reduction in symptoms)—52% vs 24%—in the intervention group (number needed to treat [NNT]=3.5). Patients taking chasteberry tree had 1 occurrence of multiple abscesses and 1 of urticaria.¹

A prospective, open-label study of chasteberry tree for PMS symptoms in 43 patients found a 42% decrease in self-assessed PMS symptom scores, with the greatest improvement in pain, behavior changes, negative feelings, and fluid retention. No serious adverse events occurred.²

A third study comparing chasteberry tree with fluoxetine in 19 patients found a decrease in premenstrual symptom scores for both fluoxetine (13 of 19 patients) and chasteberry tree (11 of 19 patients). No statistically significant differences were noted between the 2 groups. Chasteberry tree was well tolerated; most adverse effects occurred in patients receiving fluoxetine. The most frequent adverse effects with chasteberry tree were nausea in 5 patients and headache in 4.³

Symptoms decrease significantly after 3 calcium treatment cycles
Two RCTs (33 and 466 patients) comparing 1000 and 1200 mg of calcium with placebo found a significant decrease in PMS symptoms after 3 treatment cycles.^4,5 Calcium improved negative affect, water retention, food cravings, and pain. In the first study, 73% of patients preferred taking calcium, compared with 15% who preferred placebo.

The second study found that, by the third treatment cycle, patients taking calcium had an overall 48% reduction in total symptom scores, compared with a 30% reduction in the control group. The most common adverse effects were headache, rhinitis, and nonspecific pain.

Watch out for neuropathy with high doses of pyridoxine
A meta-analysis of pyridoxine in doses from 50 to 600 mg per day for PMS included 9 RCTs. Relative to placebo, pyridoxine improved PMS symptom scores (odds ratio=2.32, 95% confidence interval, 1.95-2.54). The overall quality of studies was poor, however, with few subjects, widely varying doses, and different outcome measurements.⁶

Two subsequent RCTs of 40 and 96 patients that weren’t included in the meta-analysis failed to demonstrate reduced premenstrual symptoms.^7,8 Long-term use of pyridoxine in doses >200 mg/d can cause neuropathy.

Saffron shows promise in small study
A recent double-blind RCT evaluated the effect of 2 cycles of treatment with saffron (Crocus sativus L), 30 mg twice daily, on PMS symptoms in 50 patients. Nineteen patients in the saffron group showed a response, defined as 50% reduction in symptom severity, compared with 2 patients in the placebo group (NNT=2). The study found no statistically significant difference in frequency of adverse effects.⁹

Evidence for magnesium is sparse
Two RCTs comparing magnesium with placebo had low precision because of small numbers and short treatment duration.^10,11 The first (N=28) demonstrated reduced total Moos Menstrual Distress Questionnaire scores.¹⁰ The second study reported a decrease in fluid retention symptoms by 2 points on an 80-point scale (P<.009) at 2 months, but no difference in total score.¹¹

A further study, begun as an open trial of magnesium infusion for premenstrual dysphoric disorder (N=6), found a dramatic reduction in mood symptom scores. After converting to a randomized, blinded design (N=10), no difference was found compared with placebo.¹²

St. John’s wort, evening primrose oil don’t work
One randomized, double-blind controlled trial (N=125) of 600 mg St. John’s wort vs placebo over 2 cycles of treatment found no significant changes in symptom score from baseline.¹³ Two double-blind crossover studies of 27 and 38 patients found that evening primrose oil had no effect on PMS symptoms.^14,15

Recommendations

The Premenstrual Syndrome Guidelines of the American College of Obstetricians and Gynecologists (ACOG) state that calcium and magnesium have been shown to be effective in small trials and must be validated in larger trials before a strong evidence-based recommendation can be made. ACOG’s guidelines also report minimal effectiveness with vitamin B6 and vitamin E.¹⁶

Evidence summary

As many as 30% of patients with major depression fail to respond to treatment with a single antidepressant drug given in adequate dosage for an appropriate period.¹ Pharmacologic approaches such as switching antidepressant classes are often attempted first, followed by augmentation with another agent if needed.

The most widely studied medications used for augmentation are lithium and T₃.² An important limitation to their use is that most of the supporting evidence comes from studies of patients who didn’t respond initially to tricyclic antidepressants.

Lithium boosts response to antidepressants

A 2007 meta-analysis of 10 augmentation studies reported that adding lithium to various antidepressant agents increased the chances of clinical response 3-fold relative to placebo (odds ratio [OR]=3.11; 95% confidence interval [CI], 1.80-5.37), yielding a number needed to treat (NNT) of 5.³ The mean response rate was 41.2% in the lithium group and 14.4% in the placebo group (P<.001). The meta-analysis included only RCTs that enrolled subjects with unipolar or bipolar disorder (depressive phase) who were treated with any antidepressant plus lithium in any dose compared with placebo.

A previous meta-analysis published in 1999 concluded that a lithium dose sufficient to produce serum levels of at least 0.5 mEq/L and a minimum treatment duration of 2 weeks resulted in a pooled OR of response to lithium augmentation compared with placebo of 3.31 (95% CI, 1.46-7.53).⁴ Lithium augmentation is a reasonable alternative for depressed patients who don’t respond to conventional antidepressants.

T₃ works well with tricyclics, but what about newer antidepressants?

A 1996 meta-analysis of 8 studies with a total of 292 patients found that patients who received T₃ augmentation of tricyclic antidepressant therapy were twice as likely to respond as controls (relative response=2.09; 95% CI, 1.31-3.32).⁵ The corresponding pooled absolute difference in response rate was 23.2% with a corresponding NNT of 4.3.

T₃ dosage in the studies ranged from 25 to 50 mcg/day and duration varied from 7 to 35 days. Analysis of data from RCTs alone produced a lower pooled relative response of 1.53 (95% CI, 0.70-3.35) and an NNT of 12.5. A major drawback of T₃ augmentation is that little information is available about its efficacy in combination with newer antidepressant agents.

T₃ may have a slight edge over lithium in side effects

As a part of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, lithium and T₃ augmentation were compared directly in patients who had failed 2 medication treatments for depression. A total of 142 adult outpatients with nonpsychotic major depressive disorder who had not achieved remission with (or who were intolerant to) citalopram, and who similarly had no luck with either a therapeutic switch or an augmentation trial, were randomly assigned to lithium or T₃ augmentation for as long as 14 weeks.

After a mean of 9.6 weeks of treatment, remission rates were 15.9% with lithium and 24.7% with T₃, although the difference between the 2 drugs wasn’t statistically significant. Because T₃ has fewer side effects and is easier to use, the researchers suggested that it may have a slight advantage over lithium.

Atypical antipsychotic agents are another alternative

A 2007 meta-analysis of 10 clinical trials involving 1500 outpatients studied the efficacy of augmenting standard antidepressants with atypical antipsychotic agents to treat resistant major depressive disorder.⁶ Across the trials, the pooled risk ratio (RR) for remission was 1.75 (95% CI, 1.36-2.24) and for response rates was 1.35 (95% CI, 1.13-1.63). Pooled rates for remission and response were 47.4% vs 22.3% (NNT=4) and 57.2% vs 35.4% (NNT=4.6), respectively.

Although the meta-analysis found no difference in overall discontinuation rates (P=.929), the rate of discontinuation because of adverse events was lower among placebo-treated patients (RR=3.38, P<.0001). These results suggest a role for atypical antipsychotic agents in augmenting standard antidepressants for treatment-resistant major depressive disorder.

Recommendations

Both the American Psychiatric Association⁷ and the Institute for Clinical Systems Improvement⁸ recommend considering a trial of lithium or thyroid augmentation for patients who respond only partially to initial antidepressant therapy. Many experienced clinicians consider lithium to be the most effective adjunct. The APA cautions, however, that its Major Depressive Disorder Practice Guideline, published in 2000, is no longer current.⁹ An update is expected by the end of the year.

Other options include maximizing the initial treatment, switching to another agent, or augmenting initial treatment with another antidepressant agent or psychotherapy. Before prescribing any additional treatments for patients who fail to respond to initial antidepressant therapy, however, primary care physicians need to be mindful of the fact that empirical data regarding the relative effectiveness of these strategies are limited—and that they should consider whether the patients they’re treating have bipolar—rather than unipolar—depression.

Evidence summary

As many as 30% of patients with major depression fail to respond to treatment with a single antidepressant drug given in adequate dosage for an appropriate period.¹ Pharmacologic approaches such as switching antidepressant classes are often attempted first, followed by augmentation with another agent if needed.

The most widely studied medications used for augmentation are lithium and T₃.² An important limitation to their use is that most of the supporting evidence comes from studies of patients who didn’t respond initially to tricyclic antidepressants.

Lithium boosts response to antidepressants

A 2007 meta-analysis of 10 augmentation studies reported that adding lithium to various antidepressant agents increased the chances of clinical response 3-fold relative to placebo (odds ratio [OR]=3.11; 95% confidence interval [CI], 1.80-5.37), yielding a number needed to treat (NNT) of 5.³ The mean response rate was 41.2% in the lithium group and 14.4% in the placebo group (P<.001). The meta-analysis included only RCTs that enrolled subjects with unipolar or bipolar disorder (depressive phase) who were treated with any antidepressant plus lithium in any dose compared with placebo.

A previous meta-analysis published in 1999 concluded that a lithium dose sufficient to produce serum levels of at least 0.5 mEq/L and a minimum treatment duration of 2 weeks resulted in a pooled OR of response to lithium augmentation compared with placebo of 3.31 (95% CI, 1.46-7.53).⁴ Lithium augmentation is a reasonable alternative for depressed patients who don’t respond to conventional antidepressants.

T₃ works well with tricyclics, but what about newer antidepressants?

A 1996 meta-analysis of 8 studies with a total of 292 patients found that patients who received T₃ augmentation of tricyclic antidepressant therapy were twice as likely to respond as controls (relative response=2.09; 95% CI, 1.31-3.32).⁵ The corresponding pooled absolute difference in response rate was 23.2% with a corresponding NNT of 4.3.

T₃ dosage in the studies ranged from 25 to 50 mcg/day and duration varied from 7 to 35 days. Analysis of data from RCTs alone produced a lower pooled relative response of 1.53 (95% CI, 0.70-3.35) and an NNT of 12.5. A major drawback of T₃ augmentation is that little information is available about its efficacy in combination with newer antidepressant agents.

T₃ may have a slight edge over lithium in side effects

As a part of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, lithium and T₃ augmentation were compared directly in patients who had failed 2 medication treatments for depression. A total of 142 adult outpatients with nonpsychotic major depressive disorder who had not achieved remission with (or who were intolerant to) citalopram, and who similarly had no luck with either a therapeutic switch or an augmentation trial, were randomly assigned to lithium or T₃ augmentation for as long as 14 weeks.

After a mean of 9.6 weeks of treatment, remission rates were 15.9% with lithium and 24.7% with T₃, although the difference between the 2 drugs wasn’t statistically significant. Because T₃ has fewer side effects and is easier to use, the researchers suggested that it may have a slight advantage over lithium.

Atypical antipsychotic agents are another alternative

A 2007 meta-analysis of 10 clinical trials involving 1500 outpatients studied the efficacy of augmenting standard antidepressants with atypical antipsychotic agents to treat resistant major depressive disorder.⁶ Across the trials, the pooled risk ratio (RR) for remission was 1.75 (95% CI, 1.36-2.24) and for response rates was 1.35 (95% CI, 1.13-1.63). Pooled rates for remission and response were 47.4% vs 22.3% (NNT=4) and 57.2% vs 35.4% (NNT=4.6), respectively.

Although the meta-analysis found no difference in overall discontinuation rates (P=.929), the rate of discontinuation because of adverse events was lower among placebo-treated patients (RR=3.38, P<.0001). These results suggest a role for atypical antipsychotic agents in augmenting standard antidepressants for treatment-resistant major depressive disorder.

Recommendations

Both the American Psychiatric Association⁷ and the Institute for Clinical Systems Improvement⁸ recommend considering a trial of lithium or thyroid augmentation for patients who respond only partially to initial antidepressant therapy. Many experienced clinicians consider lithium to be the most effective adjunct. The APA cautions, however, that its Major Depressive Disorder Practice Guideline, published in 2000, is no longer current.⁹ An update is expected by the end of the year.

Other options include maximizing the initial treatment, switching to another agent, or augmenting initial treatment with another antidepressant agent or psychotherapy. Before prescribing any additional treatments for patients who fail to respond to initial antidepressant therapy, however, primary care physicians need to be mindful of the fact that empirical data regarding the relative effectiveness of these strategies are limited—and that they should consider whether the patients they’re treating have bipolar—rather than unipolar—depression.

Evidence summary

As many as 30% of patients with major depression fail to respond to treatment with a single antidepressant drug given in adequate dosage for an appropriate period.¹ Pharmacologic approaches such as switching antidepressant classes are often attempted first, followed by augmentation with another agent if needed.

The most widely studied medications used for augmentation are lithium and T₃.² An important limitation to their use is that most of the supporting evidence comes from studies of patients who didn’t respond initially to tricyclic antidepressants.

Lithium boosts response to antidepressants

A 2007 meta-analysis of 10 augmentation studies reported that adding lithium to various antidepressant agents increased the chances of clinical response 3-fold relative to placebo (odds ratio [OR]=3.11; 95% confidence interval [CI], 1.80-5.37), yielding a number needed to treat (NNT) of 5.³ The mean response rate was 41.2% in the lithium group and 14.4% in the placebo group (P<.001). The meta-analysis included only RCTs that enrolled subjects with unipolar or bipolar disorder (depressive phase) who were treated with any antidepressant plus lithium in any dose compared with placebo.

A previous meta-analysis published in 1999 concluded that a lithium dose sufficient to produce serum levels of at least 0.5 mEq/L and a minimum treatment duration of 2 weeks resulted in a pooled OR of response to lithium augmentation compared with placebo of 3.31 (95% CI, 1.46-7.53).⁴ Lithium augmentation is a reasonable alternative for depressed patients who don’t respond to conventional antidepressants.

T₃ works well with tricyclics, but what about newer antidepressants?

A 1996 meta-analysis of 8 studies with a total of 292 patients found that patients who received T₃ augmentation of tricyclic antidepressant therapy were twice as likely to respond as controls (relative response=2.09; 95% CI, 1.31-3.32).⁵ The corresponding pooled absolute difference in response rate was 23.2% with a corresponding NNT of 4.3.

T₃ dosage in the studies ranged from 25 to 50 mcg/day and duration varied from 7 to 35 days. Analysis of data from RCTs alone produced a lower pooled relative response of 1.53 (95% CI, 0.70-3.35) and an NNT of 12.5. A major drawback of T₃ augmentation is that little information is available about its efficacy in combination with newer antidepressant agents.

T₃ may have a slight edge over lithium in side effects

As a part of the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, lithium and T₃ augmentation were compared directly in patients who had failed 2 medication treatments for depression. A total of 142 adult outpatients with nonpsychotic major depressive disorder who had not achieved remission with (or who were intolerant to) citalopram, and who similarly had no luck with either a therapeutic switch or an augmentation trial, were randomly assigned to lithium or T₃ augmentation for as long as 14 weeks.

After a mean of 9.6 weeks of treatment, remission rates were 15.9% with lithium and 24.7% with T₃, although the difference between the 2 drugs wasn’t statistically significant. Because T₃ has fewer side effects and is easier to use, the researchers suggested that it may have a slight advantage over lithium.

Atypical antipsychotic agents are another alternative

A 2007 meta-analysis of 10 clinical trials involving 1500 outpatients studied the efficacy of augmenting standard antidepressants with atypical antipsychotic agents to treat resistant major depressive disorder.⁶ Across the trials, the pooled risk ratio (RR) for remission was 1.75 (95% CI, 1.36-2.24) and for response rates was 1.35 (95% CI, 1.13-1.63). Pooled rates for remission and response were 47.4% vs 22.3% (NNT=4) and 57.2% vs 35.4% (NNT=4.6), respectively.

Although the meta-analysis found no difference in overall discontinuation rates (P=.929), the rate of discontinuation because of adverse events was lower among placebo-treated patients (RR=3.38, P<.0001). These results suggest a role for atypical antipsychotic agents in augmenting standard antidepressants for treatment-resistant major depressive disorder.

Recommendations

Both the American Psychiatric Association⁷ and the Institute for Clinical Systems Improvement⁸ recommend considering a trial of lithium or thyroid augmentation for patients who respond only partially to initial antidepressant therapy. Many experienced clinicians consider lithium to be the most effective adjunct. The APA cautions, however, that its Major Depressive Disorder Practice Guideline, published in 2000, is no longer current.⁹ An update is expected by the end of the year.

Other options include maximizing the initial treatment, switching to another agent, or augmenting initial treatment with another antidepressant agent or psychotherapy. Before prescribing any additional treatments for patients who fail to respond to initial antidepressant therapy, however, primary care physicians need to be mindful of the fact that empirical data regarding the relative effectiveness of these strategies are limited—and that they should consider whether the patients they’re treating have bipolar—rather than unipolar—depression.

TABLE
What’s fruit juice and what’s not

Evidence summary

One of every 6 American children is overweight or at risk of becoming overweight.¹ Overweight children are more likely than normal-weight children to be overweight as adults; they’re at significant risk for morbidity and mortality from hypertension, cardiovascular disease, and diabetes in adulthood. Establishing sound nutritional habits—including appropriate consumption of fruit juices, fruit drinks, and other sugar-sweetened beverages—early in life plays an important role in preventing overweight in later childhood and adulthood.²

Fruit juice/obesity link is controversial

During the transition to table foods between 4 and 11 months of age, the top 3 nonmilk sources of carbohydrate in an infant’s diet are infant cereal, 100% juice, and bananas.² One in 5 infants routinely drinks juice before 6 months of age.³ Consuming 100% juice and fruit-flavored drinks can contribute to excess energy intake and displace other nutrient-dense foods in the child’s diet.

The role of fruit juice consumption in childhood obesity is controversial. In 1 group of 168 children 2 to 5 years of age, 9% of children who drank >12 oz of fruit juice per day were overweight, compared with 3% of those who drank <12 oz daily.⁴

A recent review of 21 studies found 6 (3 longitudinal and 3 cross-sectional) that supported a relationship between juice intake and weight and 15 (9 longitudinal and 6 cross-sectional) that suggested no link between 100% fruit juice consumption and overweight in children or adolescents.⁵

Regardless of the relationship between fruit juice and obesity, it is important to emphasize that breast milk provides essential nutrients and immune protection for the growing infant. Breast milk should remain the recommended source of nutrition through the first year of life.^6,7

Sugar-sweetened drinks: Short on nutrition, long on risk

Sugar-sweetened beverages (labeled as fruit drinks) often replace whole fruits in childhood diets.⁸ As a result, children may fail to meet recommendations for intake of whole fruits and vegetables, which contain fiber and nutrients essential to growth.

Consumption of fruit drinks and other sugar-sweetened beverages by American children has increased 135% since 1977; such drinks account for roughly 9% of daily energy intake.⁹ Data from the Feeding Infants and Toddlers Survey (FITS) indicate that 100% fruit juice and sugar-sweetened beverages are now the second and third leading sources, respectively, of energy (and carbohydrates) for American children between 1 and 2 years of age.³

Data from the National Health and Nutrition Examination Survey (NHANES) suggest that overweight children and adolescents consume more sugar-sweetened beverages than those who are not overweight.¹⁰ Other cohort data show that children who regularly consume sugar-sweetened beverages are twice as likely to be overweight by 5 years of age as children who don’t.¹¹ However, 1 cohort of 521 children followed longitudinally from 5 to 9 years showed no association between sugar-sweetened beverage intake and body fat.¹²

A recent systematic review of 30 studies (15 cross-sectional, 10 prospective cohort, and 5 experimental trials) supports a link between consumption of sugar-sweetened beverages and childhood obesity.¹³

What about tooth decay?

Excess intake of both sugar-sweetened beverages¹⁴ and fruit juice¹⁵ has been associated with increased risk of dental caries. Excess intake is defined as more than 6 oz per day in children 1 to 6 years of age and more than 12 oz per day in children 7 to 18 years. To help reduce the risk for dental caries, children should drink juice from a cup.

Recommendations

The American Academy of Family Physicians,¹⁶ American Academy of Pediatrics,⁶ American Heart Association,¹⁷ and World Health Organization¹⁸ all recommend breast milk as the preferred source of infant nutrition for the first 6 (preferably 12) months of life. The US Preventive Services Task Force recently emphasized the need for primary care physicians to further promote breastfeeding efforts.¹⁹

Infants shouldn’t be given fruit juice before 6 months of age.¹⁷ If juice is offered, it should be 100% fruit juice in a cup, not a bottle. Children 1 to 6 years of age should drink no more than 4 to 6 oz of 100% fruit juice per day. Children 7 to 18 years of age should limit intake to 12 oz of 100% fruit juice per day.¹⁷ Infants, children, and adolescents shouldn’t drink unpasteurized juice.²⁰

The American Heart Association recommends that children 1 to 3 years of age consume the equivalent of 1 cup of whole fruit per day. Children from 4 to 13 years should consume 1.5 cups per day.¹⁷

TABLE
What’s fruit juice and what’s not

Evidence summary

One of every 6 American children is overweight or at risk of becoming overweight.¹ Overweight children are more likely than normal-weight children to be overweight as adults; they’re at significant risk for morbidity and mortality from hypertension, cardiovascular disease, and diabetes in adulthood. Establishing sound nutritional habits—including appropriate consumption of fruit juices, fruit drinks, and other sugar-sweetened beverages—early in life plays an important role in preventing overweight in later childhood and adulthood.²

Fruit juice/obesity link is controversial

During the transition to table foods between 4 and 11 months of age, the top 3 nonmilk sources of carbohydrate in an infant’s diet are infant cereal, 100% juice, and bananas.² One in 5 infants routinely drinks juice before 6 months of age.³ Consuming 100% juice and fruit-flavored drinks can contribute to excess energy intake and displace other nutrient-dense foods in the child’s diet.

The role of fruit juice consumption in childhood obesity is controversial. In 1 group of 168 children 2 to 5 years of age, 9% of children who drank >12 oz of fruit juice per day were overweight, compared with 3% of those who drank <12 oz daily.⁴

A recent review of 21 studies found 6 (3 longitudinal and 3 cross-sectional) that supported a relationship between juice intake and weight and 15 (9 longitudinal and 6 cross-sectional) that suggested no link between 100% fruit juice consumption and overweight in children or adolescents.⁵

Regardless of the relationship between fruit juice and obesity, it is important to emphasize that breast milk provides essential nutrients and immune protection for the growing infant. Breast milk should remain the recommended source of nutrition through the first year of life.^6,7

Sugar-sweetened drinks: Short on nutrition, long on risk

Sugar-sweetened beverages (labeled as fruit drinks) often replace whole fruits in childhood diets.⁸ As a result, children may fail to meet recommendations for intake of whole fruits and vegetables, which contain fiber and nutrients essential to growth.

Consumption of fruit drinks and other sugar-sweetened beverages by American children has increased 135% since 1977; such drinks account for roughly 9% of daily energy intake.⁹ Data from the Feeding Infants and Toddlers Survey (FITS) indicate that 100% fruit juice and sugar-sweetened beverages are now the second and third leading sources, respectively, of energy (and carbohydrates) for American children between 1 and 2 years of age.³

Data from the National Health and Nutrition Examination Survey (NHANES) suggest that overweight children and adolescents consume more sugar-sweetened beverages than those who are not overweight.¹⁰ Other cohort data show that children who regularly consume sugar-sweetened beverages are twice as likely to be overweight by 5 years of age as children who don’t.¹¹ However, 1 cohort of 521 children followed longitudinally from 5 to 9 years showed no association between sugar-sweetened beverage intake and body fat.¹²

A recent systematic review of 30 studies (15 cross-sectional, 10 prospective cohort, and 5 experimental trials) supports a link between consumption of sugar-sweetened beverages and childhood obesity.¹³

What about tooth decay?

Excess intake of both sugar-sweetened beverages¹⁴ and fruit juice¹⁵ has been associated with increased risk of dental caries. Excess intake is defined as more than 6 oz per day in children 1 to 6 years of age and more than 12 oz per day in children 7 to 18 years. To help reduce the risk for dental caries, children should drink juice from a cup.

Recommendations

The American Academy of Family Physicians,¹⁶ American Academy of Pediatrics,⁶ American Heart Association,¹⁷ and World Health Organization¹⁸ all recommend breast milk as the preferred source of infant nutrition for the first 6 (preferably 12) months of life. The US Preventive Services Task Force recently emphasized the need for primary care physicians to further promote breastfeeding efforts.¹⁹

Infants shouldn’t be given fruit juice before 6 months of age.¹⁷ If juice is offered, it should be 100% fruit juice in a cup, not a bottle. Children 1 to 6 years of age should drink no more than 4 to 6 oz of 100% fruit juice per day. Children 7 to 18 years of age should limit intake to 12 oz of 100% fruit juice per day.¹⁷ Infants, children, and adolescents shouldn’t drink unpasteurized juice.²⁰

The American Heart Association recommends that children 1 to 3 years of age consume the equivalent of 1 cup of whole fruit per day. Children from 4 to 13 years should consume 1.5 cups per day.¹⁷

TABLE
What’s fruit juice and what’s not

Evidence summary

One of every 6 American children is overweight or at risk of becoming overweight.¹ Overweight children are more likely than normal-weight children to be overweight as adults; they’re at significant risk for morbidity and mortality from hypertension, cardiovascular disease, and diabetes in adulthood. Establishing sound nutritional habits—including appropriate consumption of fruit juices, fruit drinks, and other sugar-sweetened beverages—early in life plays an important role in preventing overweight in later childhood and adulthood.²

Fruit juice/obesity link is controversial

During the transition to table foods between 4 and 11 months of age, the top 3 nonmilk sources of carbohydrate in an infant’s diet are infant cereal, 100% juice, and bananas.² One in 5 infants routinely drinks juice before 6 months of age.³ Consuming 100% juice and fruit-flavored drinks can contribute to excess energy intake and displace other nutrient-dense foods in the child’s diet.

The role of fruit juice consumption in childhood obesity is controversial. In 1 group of 168 children 2 to 5 years of age, 9% of children who drank >12 oz of fruit juice per day were overweight, compared with 3% of those who drank <12 oz daily.⁴

A recent review of 21 studies found 6 (3 longitudinal and 3 cross-sectional) that supported a relationship between juice intake and weight and 15 (9 longitudinal and 6 cross-sectional) that suggested no link between 100% fruit juice consumption and overweight in children or adolescents.⁵

Regardless of the relationship between fruit juice and obesity, it is important to emphasize that breast milk provides essential nutrients and immune protection for the growing infant. Breast milk should remain the recommended source of nutrition through the first year of life.^6,7

Sugar-sweetened drinks: Short on nutrition, long on risk

Sugar-sweetened beverages (labeled as fruit drinks) often replace whole fruits in childhood diets.⁸ As a result, children may fail to meet recommendations for intake of whole fruits and vegetables, which contain fiber and nutrients essential to growth.

Consumption of fruit drinks and other sugar-sweetened beverages by American children has increased 135% since 1977; such drinks account for roughly 9% of daily energy intake.⁹ Data from the Feeding Infants and Toddlers Survey (FITS) indicate that 100% fruit juice and sugar-sweetened beverages are now the second and third leading sources, respectively, of energy (and carbohydrates) for American children between 1 and 2 years of age.³

Data from the National Health and Nutrition Examination Survey (NHANES) suggest that overweight children and adolescents consume more sugar-sweetened beverages than those who are not overweight.¹⁰ Other cohort data show that children who regularly consume sugar-sweetened beverages are twice as likely to be overweight by 5 years of age as children who don’t.¹¹ However, 1 cohort of 521 children followed longitudinally from 5 to 9 years showed no association between sugar-sweetened beverage intake and body fat.¹²

A recent systematic review of 30 studies (15 cross-sectional, 10 prospective cohort, and 5 experimental trials) supports a link between consumption of sugar-sweetened beverages and childhood obesity.¹³

What about tooth decay?

Excess intake of both sugar-sweetened beverages¹⁴ and fruit juice¹⁵ has been associated with increased risk of dental caries. Excess intake is defined as more than 6 oz per day in children 1 to 6 years of age and more than 12 oz per day in children 7 to 18 years. To help reduce the risk for dental caries, children should drink juice from a cup.

Recommendations

The American Academy of Family Physicians,¹⁶ American Academy of Pediatrics,⁶ American Heart Association,¹⁷ and World Health Organization¹⁸ all recommend breast milk as the preferred source of infant nutrition for the first 6 (preferably 12) months of life. The US Preventive Services Task Force recently emphasized the need for primary care physicians to further promote breastfeeding efforts.¹⁹

Infants shouldn’t be given fruit juice before 6 months of age.¹⁷ If juice is offered, it should be 100% fruit juice in a cup, not a bottle. Children 1 to 6 years of age should drink no more than 4 to 6 oz of 100% fruit juice per day. Children 7 to 18 years of age should limit intake to 12 oz of 100% fruit juice per day.¹⁷ Infants, children, and adolescents shouldn’t drink unpasteurized juice.²⁰

The American Heart Association recommends that children 1 to 3 years of age consume the equivalent of 1 cup of whole fruit per day. Children from 4 to 13 years should consume 1.5 cups per day.¹⁷

Evidence summary

Trochanteric bursitis is characterized by chronic intermittent lateral hip pain caused by inflammation of the trochanteric bursae. The bursae can become inflamed at the gluteus medius tendon, iliotibial tract, or gluteus minimus during repetitive flexing of the hip. Several conditions are associated with trochanteric bursitis (TABLE).

Trochanteric bursitis peaks in the fourth to sixth decades of life. One retrospective cohort study found the prevalence to be 1.8 cases per 1000 patients per year in primary care; 79% of cases occurred in women.¹

TABLE
Conditions associated with trochanteric bursitis

No studies have compared conservative treatments

Most review articles refer to initial treatment with rest, physical therapy, stretching, and NSAIDs. These treatments were described in textbooks and articles from the 1940s and 1950s.

No studies comparing conservative treatments were found. Few reports discuss physical therapy for trochanteric bursitis.

Corticosteroid injection has the best evidential support

Corticosteroid injection for treating trochanteric bursitis is supported by the best evidence in the available literature. No controlled trials have compared injection with placebo, however.

A randomized, prospective, open comparison trial at a rheumatology clinic assigned patients with trochanteric bursitis to 6-, 12-, or 24-mg doses of betamethasone mixed with 1% lidocaine.² Seventy-seven percent of patients had improved at 1 week, 69% at 6 weeks, and 61% at 26 weeks. Notably, a significant difference was found at 26 weeks in the number of patients with sustained pain improvement who had received 24 mg of steroid (P<.0123) compared with patients who received the lower doses. The authors didn’t report side effects or complications.

A prospective, noncomparative cohort study investigated 72 patients in a rheumatology clinic who hadn’t improved after at least 2 weeks of treatment with NSAIDs, analgesics, or ointments.³ Of the 59 patients who consented to steroid injections, 42 improved after 1 injection of 40 mg methylprednisolone with 2 mL of 2% lidocaine, 13 improved after a second injection 3 weeks later, and the remaining 4 improved after a third injection. Improvement was defined as disappearance of pain and disability. Six patients (8%) experienced a recurrence of bursitis during a 2-year follow-up period. No local or systemic complications were associated with the corticosteroids or anesthetic solution.

Two retrospective studies also documented the efficacy of corticosteroid injection. One investigated treatment of 36 patients in a rheumatology practice.⁴ All received methylprednisolone (40-80 mg) or triamcinolone (20-40 mg), and all improved. Two thirds of the patients were symptom free after 1 or 2 injections. Symptoms usually resolved within 2 days to several months (typically 1 or 2 weeks) postinjection. About 25% of the patients relapsed within 2 years.

Another retrospective cohort study of 164 British patients found that those who received a corticosteroid injection were 2.7 times more likely to have recovered at 5 years than patients who had not received an injection (odds ratio=0.4; 95% confidence interval, 0.1-1.0).¹

When to consider surgery

Surgical treatment may be necessary for patients with refractory trochanteric bursitis. Several case studies^5-7 demonstrate successful outcomes with a variety of surgical techniques, including trochanteric reduction osteotomy and iliotibial band release. Newer techniques involve arthroscopic bursectomy.

Recommendations

UpToDate⁸ recommends conservative treatment initially. For persistent cases, a corticosteroid injection should be given and repeated in 6 weeks if pain persists. Surgery may be considered if these measures don’t relieve symptoms and pain lasts longer than 1 year.

The American Academy of Orthopaedic Surgeons similarly recommends NSAIDs and activity modification followed by corticosteroid injection.⁹ Surgery is rarely indicated.

Evidence summary

Trochanteric bursitis is characterized by chronic intermittent lateral hip pain caused by inflammation of the trochanteric bursae. The bursae can become inflamed at the gluteus medius tendon, iliotibial tract, or gluteus minimus during repetitive flexing of the hip. Several conditions are associated with trochanteric bursitis (TABLE).

Trochanteric bursitis peaks in the fourth to sixth decades of life. One retrospective cohort study found the prevalence to be 1.8 cases per 1000 patients per year in primary care; 79% of cases occurred in women.¹

TABLE
Conditions associated with trochanteric bursitis

No studies have compared conservative treatments

Most review articles refer to initial treatment with rest, physical therapy, stretching, and NSAIDs. These treatments were described in textbooks and articles from the 1940s and 1950s.

No studies comparing conservative treatments were found. Few reports discuss physical therapy for trochanteric bursitis.

Corticosteroid injection has the best evidential support

Corticosteroid injection for treating trochanteric bursitis is supported by the best evidence in the available literature. No controlled trials have compared injection with placebo, however.

A randomized, prospective, open comparison trial at a rheumatology clinic assigned patients with trochanteric bursitis to 6-, 12-, or 24-mg doses of betamethasone mixed with 1% lidocaine.² Seventy-seven percent of patients had improved at 1 week, 69% at 6 weeks, and 61% at 26 weeks. Notably, a significant difference was found at 26 weeks in the number of patients with sustained pain improvement who had received 24 mg of steroid (P<.0123) compared with patients who received the lower doses. The authors didn’t report side effects or complications.

A prospective, noncomparative cohort study investigated 72 patients in a rheumatology clinic who hadn’t improved after at least 2 weeks of treatment with NSAIDs, analgesics, or ointments.³ Of the 59 patients who consented to steroid injections, 42 improved after 1 injection of 40 mg methylprednisolone with 2 mL of 2% lidocaine, 13 improved after a second injection 3 weeks later, and the remaining 4 improved after a third injection. Improvement was defined as disappearance of pain and disability. Six patients (8%) experienced a recurrence of bursitis during a 2-year follow-up period. No local or systemic complications were associated with the corticosteroids or anesthetic solution.

Two retrospective studies also documented the efficacy of corticosteroid injection. One investigated treatment of 36 patients in a rheumatology practice.⁴ All received methylprednisolone (40-80 mg) or triamcinolone (20-40 mg), and all improved. Two thirds of the patients were symptom free after 1 or 2 injections. Symptoms usually resolved within 2 days to several months (typically 1 or 2 weeks) postinjection. About 25% of the patients relapsed within 2 years.

Another retrospective cohort study of 164 British patients found that those who received a corticosteroid injection were 2.7 times more likely to have recovered at 5 years than patients who had not received an injection (odds ratio=0.4; 95% confidence interval, 0.1-1.0).¹

When to consider surgery

Surgical treatment may be necessary for patients with refractory trochanteric bursitis. Several case studies^5-7 demonstrate successful outcomes with a variety of surgical techniques, including trochanteric reduction osteotomy and iliotibial band release. Newer techniques involve arthroscopic bursectomy.

Recommendations

UpToDate⁸ recommends conservative treatment initially. For persistent cases, a corticosteroid injection should be given and repeated in 6 weeks if pain persists. Surgery may be considered if these measures don’t relieve symptoms and pain lasts longer than 1 year.

The American Academy of Orthopaedic Surgeons similarly recommends NSAIDs and activity modification followed by corticosteroid injection.⁹ Surgery is rarely indicated.

Evidence summary

Trochanteric bursitis is characterized by chronic intermittent lateral hip pain caused by inflammation of the trochanteric bursae. The bursae can become inflamed at the gluteus medius tendon, iliotibial tract, or gluteus minimus during repetitive flexing of the hip. Several conditions are associated with trochanteric bursitis (TABLE).

Trochanteric bursitis peaks in the fourth to sixth decades of life. One retrospective cohort study found the prevalence to be 1.8 cases per 1000 patients per year in primary care; 79% of cases occurred in women.¹

TABLE
Conditions associated with trochanteric bursitis

No studies have compared conservative treatments

Most review articles refer to initial treatment with rest, physical therapy, stretching, and NSAIDs. These treatments were described in textbooks and articles from the 1940s and 1950s.

No studies comparing conservative treatments were found. Few reports discuss physical therapy for trochanteric bursitis.

Corticosteroid injection has the best evidential support

Corticosteroid injection for treating trochanteric bursitis is supported by the best evidence in the available literature. No controlled trials have compared injection with placebo, however.

A randomized, prospective, open comparison trial at a rheumatology clinic assigned patients with trochanteric bursitis to 6-, 12-, or 24-mg doses of betamethasone mixed with 1% lidocaine.² Seventy-seven percent of patients had improved at 1 week, 69% at 6 weeks, and 61% at 26 weeks. Notably, a significant difference was found at 26 weeks in the number of patients with sustained pain improvement who had received 24 mg of steroid (P<.0123) compared with patients who received the lower doses. The authors didn’t report side effects or complications.

A prospective, noncomparative cohort study investigated 72 patients in a rheumatology clinic who hadn’t improved after at least 2 weeks of treatment with NSAIDs, analgesics, or ointments.³ Of the 59 patients who consented to steroid injections, 42 improved after 1 injection of 40 mg methylprednisolone with 2 mL of 2% lidocaine, 13 improved after a second injection 3 weeks later, and the remaining 4 improved after a third injection. Improvement was defined as disappearance of pain and disability. Six patients (8%) experienced a recurrence of bursitis during a 2-year follow-up period. No local or systemic complications were associated with the corticosteroids or anesthetic solution.

Two retrospective studies also documented the efficacy of corticosteroid injection. One investigated treatment of 36 patients in a rheumatology practice.⁴ All received methylprednisolone (40-80 mg) or triamcinolone (20-40 mg), and all improved. Two thirds of the patients were symptom free after 1 or 2 injections. Symptoms usually resolved within 2 days to several months (typically 1 or 2 weeks) postinjection. About 25% of the patients relapsed within 2 years.

Another retrospective cohort study of 164 British patients found that those who received a corticosteroid injection were 2.7 times more likely to have recovered at 5 years than patients who had not received an injection (odds ratio=0.4; 95% confidence interval, 0.1-1.0).¹

When to consider surgery

Surgical treatment may be necessary for patients with refractory trochanteric bursitis. Several case studies^5-7 demonstrate successful outcomes with a variety of surgical techniques, including trochanteric reduction osteotomy and iliotibial band release. Newer techniques involve arthroscopic bursectomy.

Recommendations

UpToDate⁸ recommends conservative treatment initially. For persistent cases, a corticosteroid injection should be given and repeated in 6 weeks if pain persists. Surgery may be considered if these measures don’t relieve symptoms and pain lasts longer than 1 year.

The American Academy of Orthopaedic Surgeons similarly recommends NSAIDs and activity modification followed by corticosteroid injection.⁹ Surgery is rarely indicated.

Evidence summary

External hemorrhoids originate below the dentate line and become acutely painful with thrombosis. They can cause perianal pruritus and excoriation because of interference with perianal hygiene. Internal hemorrhoids become symptomatic when they bleed or prolapse (TABLE).

TABLE
Classification of symptomatic internal hemorrhoids

For thrombosed external hemorrhoids, surgery works best

Few studies have evaluated the best treatment for thrombosed external hemorrhoids. A retrospective study of 231 patients treated conservatively or surgically found that the 48.5% of patients treated surgically had a lower recurrence rate than the conservative group (number needed to treat [NNT]=2 for recurrence at mean follow-up of 7.6 months) and earlier resolution of symptoms (average 3.9 days compared with 24 days for conservative treatment).¹

Another retrospective analysis of 340 patients who underwent outpatient excision of thrombosed external hemorrhoids under local anesthesia reported a low recurrence rate of 6.5% at a mean follow-up of 17.3 months.²

A prospective, randomized controlled trial (RCT) of 98 patients treated nonsurgically found improved pain relief with a combination of topical nifedipine 0.3% and lidocaine 1.5% compared with lidocaine alone. The NNT for complete pain relief at 7 days was 3.³

Conventional hemorrhoidectomy beats stapling

Many studies have evaluated the best treatment for prolapsed hemorrhoids. A Cochrane systematic review of 12 RCTs that compared conventional hemorrhoidectomy with stapled hemorrhoidectomy in patients with grades I to III hemorrhoids found a lower rate of recurrence (follow-up ranged from 6 to 39 months) in patients who had conventional hemorrhoidectomy (NNT=14).⁴ Conventional hemorrhoidectomy showed a nonsignificant trend in decreased bleeding and decreased incontinence.

A second systematic review of 25 studies, including some that were of lower quality, showed a higher recurrence rate at 1 year with stapled hemorrhoidectomy than with conventional surgery.⁵

Nonoperative techniques? Consider rubber band ligation

A systematic review of 3 poor-quality trials comparing rubber band ligation with excisional hemorrhoidectomy in patients with grade III hemorrhoids found that excisional hemorrhoidectomy produced better long-term symptom control but more immediate postoperative complications of anal stenosis and hemorrhage.⁶ Rubber band ligation had the lowest recurrence rate at 12 months compared with the other nonoperative techniques of sclerotherapy and infrared coagulation.⁷

Fiber supplements help relieve symptoms

A Cochrane systematic review of 7 RCTs enrolling a total of 378 patients with grade I to III hemorrhoids evaluated the effect of fiber supplements on pain, itching, and bleeding. Persistent hemorrhoid symptoms decreased by 53% in the group receiving fiber.⁸

When surgical hemorrhoidectomy is recommended

The American Society of Colon and Rectal Surgeons recommends adequate fluid and fiber intake for all patients with symptomatic hemorrhoids. For grade I to III hemorrhoids, the society states that banding is usually most effective. When office treatments fail, the society recommends surgical hemorrhoidectomy (SOR: B).

The society recommends excision of thrombosed hemorrhoids less than 72 hours old and expectant treatment with analgesia and sitz baths for thrombosed hemorrhoids present for longer than 72 hours (SOR: B).⁹

The American Gastroenterological Association recommends excision of symptomatic thrombosed external hemorrhoids that present early. Surgical hemorrhoidectomy should be reserved for when conservative treatment fails and for patients with symptomatic grade III and IV hemorrhoids.¹⁰

Evidence summary

External hemorrhoids originate below the dentate line and become acutely painful with thrombosis. They can cause perianal pruritus and excoriation because of interference with perianal hygiene. Internal hemorrhoids become symptomatic when they bleed or prolapse (TABLE).

TABLE
Classification of symptomatic internal hemorrhoids

For thrombosed external hemorrhoids, surgery works best

Few studies have evaluated the best treatment for thrombosed external hemorrhoids. A retrospective study of 231 patients treated conservatively or surgically found that the 48.5% of patients treated surgically had a lower recurrence rate than the conservative group (number needed to treat [NNT]=2 for recurrence at mean follow-up of 7.6 months) and earlier resolution of symptoms (average 3.9 days compared with 24 days for conservative treatment).¹

Another retrospective analysis of 340 patients who underwent outpatient excision of thrombosed external hemorrhoids under local anesthesia reported a low recurrence rate of 6.5% at a mean follow-up of 17.3 months.²

A prospective, randomized controlled trial (RCT) of 98 patients treated nonsurgically found improved pain relief with a combination of topical nifedipine 0.3% and lidocaine 1.5% compared with lidocaine alone. The NNT for complete pain relief at 7 days was 3.³

Conventional hemorrhoidectomy beats stapling

Many studies have evaluated the best treatment for prolapsed hemorrhoids. A Cochrane systematic review of 12 RCTs that compared conventional hemorrhoidectomy with stapled hemorrhoidectomy in patients with grades I to III hemorrhoids found a lower rate of recurrence (follow-up ranged from 6 to 39 months) in patients who had conventional hemorrhoidectomy (NNT=14).⁴ Conventional hemorrhoidectomy showed a nonsignificant trend in decreased bleeding and decreased incontinence.

A second systematic review of 25 studies, including some that were of lower quality, showed a higher recurrence rate at 1 year with stapled hemorrhoidectomy than with conventional surgery.⁵

Nonoperative techniques? Consider rubber band ligation

A systematic review of 3 poor-quality trials comparing rubber band ligation with excisional hemorrhoidectomy in patients with grade III hemorrhoids found that excisional hemorrhoidectomy produced better long-term symptom control but more immediate postoperative complications of anal stenosis and hemorrhage.⁶ Rubber band ligation had the lowest recurrence rate at 12 months compared with the other nonoperative techniques of sclerotherapy and infrared coagulation.⁷

Fiber supplements help relieve symptoms

A Cochrane systematic review of 7 RCTs enrolling a total of 378 patients with grade I to III hemorrhoids evaluated the effect of fiber supplements on pain, itching, and bleeding. Persistent hemorrhoid symptoms decreased by 53% in the group receiving fiber.⁸

When surgical hemorrhoidectomy is recommended

The American Society of Colon and Rectal Surgeons recommends adequate fluid and fiber intake for all patients with symptomatic hemorrhoids. For grade I to III hemorrhoids, the society states that banding is usually most effective. When office treatments fail, the society recommends surgical hemorrhoidectomy (SOR: B).

The society recommends excision of thrombosed hemorrhoids less than 72 hours old and expectant treatment with analgesia and sitz baths for thrombosed hemorrhoids present for longer than 72 hours (SOR: B).⁹

The American Gastroenterological Association recommends excision of symptomatic thrombosed external hemorrhoids that present early. Surgical hemorrhoidectomy should be reserved for when conservative treatment fails and for patients with symptomatic grade III and IV hemorrhoids.¹⁰

Evidence summary

External hemorrhoids originate below the dentate line and become acutely painful with thrombosis. They can cause perianal pruritus and excoriation because of interference with perianal hygiene. Internal hemorrhoids become symptomatic when they bleed or prolapse (TABLE).

TABLE
Classification of symptomatic internal hemorrhoids

For thrombosed external hemorrhoids, surgery works best

Few studies have evaluated the best treatment for thrombosed external hemorrhoids. A retrospective study of 231 patients treated conservatively or surgically found that the 48.5% of patients treated surgically had a lower recurrence rate than the conservative group (number needed to treat [NNT]=2 for recurrence at mean follow-up of 7.6 months) and earlier resolution of symptoms (average 3.9 days compared with 24 days for conservative treatment).¹

Another retrospective analysis of 340 patients who underwent outpatient excision of thrombosed external hemorrhoids under local anesthesia reported a low recurrence rate of 6.5% at a mean follow-up of 17.3 months.²

A prospective, randomized controlled trial (RCT) of 98 patients treated nonsurgically found improved pain relief with a combination of topical nifedipine 0.3% and lidocaine 1.5% compared with lidocaine alone. The NNT for complete pain relief at 7 days was 3.³

Conventional hemorrhoidectomy beats stapling

Many studies have evaluated the best treatment for prolapsed hemorrhoids. A Cochrane systematic review of 12 RCTs that compared conventional hemorrhoidectomy with stapled hemorrhoidectomy in patients with grades I to III hemorrhoids found a lower rate of recurrence (follow-up ranged from 6 to 39 months) in patients who had conventional hemorrhoidectomy (NNT=14).⁴ Conventional hemorrhoidectomy showed a nonsignificant trend in decreased bleeding and decreased incontinence.

A second systematic review of 25 studies, including some that were of lower quality, showed a higher recurrence rate at 1 year with stapled hemorrhoidectomy than with conventional surgery.⁵

Nonoperative techniques? Consider rubber band ligation

A systematic review of 3 poor-quality trials comparing rubber band ligation with excisional hemorrhoidectomy in patients with grade III hemorrhoids found that excisional hemorrhoidectomy produced better long-term symptom control but more immediate postoperative complications of anal stenosis and hemorrhage.⁶ Rubber band ligation had the lowest recurrence rate at 12 months compared with the other nonoperative techniques of sclerotherapy and infrared coagulation.⁷

Fiber supplements help relieve symptoms

A Cochrane systematic review of 7 RCTs enrolling a total of 378 patients with grade I to III hemorrhoids evaluated the effect of fiber supplements on pain, itching, and bleeding. Persistent hemorrhoid symptoms decreased by 53% in the group receiving fiber.⁸

When surgical hemorrhoidectomy is recommended

The American Society of Colon and Rectal Surgeons recommends adequate fluid and fiber intake for all patients with symptomatic hemorrhoids. For grade I to III hemorrhoids, the society states that banding is usually most effective. When office treatments fail, the society recommends surgical hemorrhoidectomy (SOR: B).

The society recommends excision of thrombosed hemorrhoids less than 72 hours old and expectant treatment with analgesia and sitz baths for thrombosed hemorrhoids present for longer than 72 hours (SOR: B).⁹

The American Gastroenterological Association recommends excision of symptomatic thrombosed external hemorrhoids that present early. Surgical hemorrhoidectomy should be reserved for when conservative treatment fails and for patients with symptomatic grade III and IV hemorrhoids.¹⁰

Evidence summary

Several studies have contrasted physical examination findings associated with testicular torsion with findings related to epididymitis and torsion appendix testis. Neither the presence nor absence of any particular physical examination sign excludes the diagnosis of testicular torsion.

A consecutive case series evaluated 245 boys, newborn to 18 years of age, with acute scrotal swelling. None of the 125 subjects who had an intact cremasteric reflex had ipsilateral testicular torsion. The cremasteric reflex was absent in all 56 subjects with testicular torsion.¹ An absent cremasteric reflex in boys with acute scrotal swelling had a sensitivity of 100% (95% confidence interval [CI], 91%-100%), a specificity of 66% (95% CI, 59%-72%), and a likelihood ratio of a negative test (presence of a cremasteric reflex) of 0.01 (95% CI, 0.001-0.21).

A retrospective study reviewed the records of 90 hospitalized patients, 18 years or younger, who were discharged with a diagnosis of testicular torsion, epididymitis, or torsion appendix testis. The cremasteric reflex was absent, and testicular tenderness present, in all 13 patients with testicular torsion. The presence or absence of other physical exam findings—such as abnormal testicular lie, tender epididymis, and scrotal erythema or edema—didn’t exclude testicular torsion ( TABLE ).²

TABLE
The patients all had testicular torsion, but how helpful were the exam findings?

But cremasteric reflex and testicular torsion can coexist

Isolated case reports have demonstrated the presence of the cremasteric reflex with an eventual diagnosis of testicular torsion.^3,4 All of these studies were limited by a small number of patients and their retrospective nature.

Recommendations

When evaluating patients suspected to have testicular torsion, the European Society for Pediatric Urology (ESPU) recommends looking for absence of a cremasteric reflex and abnormal testicular position.⁵ The ESPU notes that “in many cases it is not easy to determine the cause of acute scrotum based on history and physical examination alone.”⁵ The society recommends using Doppler ultrasound as an adjunct to the history and physical.

UpToDate notes that “the diagnosis of testicular torsion can be made clinically,” but states that “radiologic evaluation (a color Doppler ultrasound or nuclear scan of the scrotum) should be undertaken if the certainty of the diagnosis is in question and the performance of imaging studies will not significantly delay treatment.”⁶

The American College of Radiology recommends color Doppler ultrasound (CDU) or radionuclide scrotal imaging (RNSI) to evaluate testicular perfusion. The group notes that “although some authors still suggest immediate surgical exploration in patients with a strong clinical impression of testicular ischemia, if either CDU or RNSI is readily available and can be performed within 30 to 60 minutes of the request to simultaneously prepare an operating room, there is ample evidence that fewer patients with infection will be operated on.”⁷

Evidence summary

Several studies have contrasted physical examination findings associated with testicular torsion with findings related to epididymitis and torsion appendix testis. Neither the presence nor absence of any particular physical examination sign excludes the diagnosis of testicular torsion.

A consecutive case series evaluated 245 boys, newborn to 18 years of age, with acute scrotal swelling. None of the 125 subjects who had an intact cremasteric reflex had ipsilateral testicular torsion. The cremasteric reflex was absent in all 56 subjects with testicular torsion.¹ An absent cremasteric reflex in boys with acute scrotal swelling had a sensitivity of 100% (95% confidence interval [CI], 91%-100%), a specificity of 66% (95% CI, 59%-72%), and a likelihood ratio of a negative test (presence of a cremasteric reflex) of 0.01 (95% CI, 0.001-0.21).

A retrospective study reviewed the records of 90 hospitalized patients, 18 years or younger, who were discharged with a diagnosis of testicular torsion, epididymitis, or torsion appendix testis. The cremasteric reflex was absent, and testicular tenderness present, in all 13 patients with testicular torsion. The presence or absence of other physical exam findings—such as abnormal testicular lie, tender epididymis, and scrotal erythema or edema—didn’t exclude testicular torsion ( TABLE ).²

TABLE
The patients all had testicular torsion, but how helpful were the exam findings?

But cremasteric reflex and testicular torsion can coexist

Isolated case reports have demonstrated the presence of the cremasteric reflex with an eventual diagnosis of testicular torsion.^3,4 All of these studies were limited by a small number of patients and their retrospective nature.

Recommendations

When evaluating patients suspected to have testicular torsion, the European Society for Pediatric Urology (ESPU) recommends looking for absence of a cremasteric reflex and abnormal testicular position.⁵ The ESPU notes that “in many cases it is not easy to determine the cause of acute scrotum based on history and physical examination alone.”⁵ The society recommends using Doppler ultrasound as an adjunct to the history and physical.

UpToDate notes that “the diagnosis of testicular torsion can be made clinically,” but states that “radiologic evaluation (a color Doppler ultrasound or nuclear scan of the scrotum) should be undertaken if the certainty of the diagnosis is in question and the performance of imaging studies will not significantly delay treatment.”⁶

The American College of Radiology recommends color Doppler ultrasound (CDU) or radionuclide scrotal imaging (RNSI) to evaluate testicular perfusion. The group notes that “although some authors still suggest immediate surgical exploration in patients with a strong clinical impression of testicular ischemia, if either CDU or RNSI is readily available and can be performed within 30 to 60 minutes of the request to simultaneously prepare an operating room, there is ample evidence that fewer patients with infection will be operated on.”⁷

Evidence summary

Several studies have contrasted physical examination findings associated with testicular torsion with findings related to epididymitis and torsion appendix testis. Neither the presence nor absence of any particular physical examination sign excludes the diagnosis of testicular torsion.

A consecutive case series evaluated 245 boys, newborn to 18 years of age, with acute scrotal swelling. None of the 125 subjects who had an intact cremasteric reflex had ipsilateral testicular torsion. The cremasteric reflex was absent in all 56 subjects with testicular torsion.¹ An absent cremasteric reflex in boys with acute scrotal swelling had a sensitivity of 100% (95% confidence interval [CI], 91%-100%), a specificity of 66% (95% CI, 59%-72%), and a likelihood ratio of a negative test (presence of a cremasteric reflex) of 0.01 (95% CI, 0.001-0.21).

A retrospective study reviewed the records of 90 hospitalized patients, 18 years or younger, who were discharged with a diagnosis of testicular torsion, epididymitis, or torsion appendix testis. The cremasteric reflex was absent, and testicular tenderness present, in all 13 patients with testicular torsion. The presence or absence of other physical exam findings—such as abnormal testicular lie, tender epididymis, and scrotal erythema or edema—didn’t exclude testicular torsion ( TABLE ).²

TABLE
The patients all had testicular torsion, but how helpful were the exam findings?

But cremasteric reflex and testicular torsion can coexist

Isolated case reports have demonstrated the presence of the cremasteric reflex with an eventual diagnosis of testicular torsion.^3,4 All of these studies were limited by a small number of patients and their retrospective nature.

Recommendations

When evaluating patients suspected to have testicular torsion, the European Society for Pediatric Urology (ESPU) recommends looking for absence of a cremasteric reflex and abnormal testicular position.⁵ The ESPU notes that “in many cases it is not easy to determine the cause of acute scrotum based on history and physical examination alone.”⁵ The society recommends using Doppler ultrasound as an adjunct to the history and physical.

UpToDate notes that “the diagnosis of testicular torsion can be made clinically,” but states that “radiologic evaluation (a color Doppler ultrasound or nuclear scan of the scrotum) should be undertaken if the certainty of the diagnosis is in question and the performance of imaging studies will not significantly delay treatment.”⁶

The American College of Radiology recommends color Doppler ultrasound (CDU) or radionuclide scrotal imaging (RNSI) to evaluate testicular perfusion. The group notes that “although some authors still suggest immediate surgical exploration in patients with a strong clinical impression of testicular ischemia, if either CDU or RNSI is readily available and can be performed within 30 to 60 minutes of the request to simultaneously prepare an operating room, there is ample evidence that fewer patients with infection will be operated on.”⁷

Evidence summary

The Centers for Disease Control and Prevention (CDC) doesn’t use the term obesity to describe weight in children. Instead, the CDC defines overweight as body mass index (BMI) or weight-for-length for age and sex greater than the 95th percentile. Children above the 85th percentile are called “at risk for overweight.” A national expert panel recently recommended changing “at risk for overweight” to “overweight” and “overweight” to “obese”—the terminology used in this Clinical Inquiry.³

Studies find a clear connection

A 2008 systematic review found 25 prospective or retrospective longitudinal studies that examined the risk of overweight in adulthood based on overweight in childhood or adolescence. Studies had to include at least 1 anthropomorphic measurement before age 18 and at least 1 after age 18. The informativeness and validity of the studies were assessed using a standard evaluation tool. Because the review sought to provide results that could be generalized to large populations, it didn’t include studies of specific populations, such as former premature infants.

All of the 13 studies judged to be high quality found an elevated RR or odds ratio (OR) for adult obesity among participants who had been overweight as children. The authors didn’t calculate a composite measure of effect, but RRs in the individual studies ranged from 1.9 to 10.1.⁴

Older overweight children are at heightened risk

The systematic review considered children (≤12 years) and adolescents (>12 years) separately. Four high-quality studies assessed how many overweight children became overweight adults; 2 high-quality studies examined how many overweight children became obese adults. Overweight children had RRs between 1.9 and 3.6 for being overweight in adulthood compared with average-weight children; 1 study reported an OR of 7.0.

One study showed older overweight children to be at greater risk than younger children for overweight in adulthood: children who were overweight at 2 years of age had an RR of 2.7, whereas children who were overweight at 11 years had an RR of 3.6.

Obese children had similar results. A study of 4 age cohorts showed that children who were obese at 1 or 2 years of age had an OR of 1.3 for obesity in adulthood compared with average-weight peers. Obese children in the 3- to 5-year-old cohort had an OR of 4.7; obese children in the 6- to 9-year-old cohort had an OR of 8.8; and obese 10- to 14-year-olds had an OR of 22.3.

Risk increases with age

As with children, overweight adolescents had a higher risk of being overweight in adulthood. And the association between older age and higher ORs persisted into adolescence. One study found an OR of 17.5 for adult overweight among youngsters who were overweight at 10 to 14 years of age and an OR of 22.3 for adolescents who were overweight at 15 to 17 years.

Boys are at greater risk than girls

The systematic review also revealed sex differences. Two studies showed that overweight or obese boys were not only more likely to be overweight in adulthood than their average-weight peers (OR=15.0 in 1 study; RR=9.8 in the other), but also more likely to be overweight later in life than overweight or obese girls. The girls had an OR of 12.0 for adult overweight in 1 study and an RR of 6.8 in the other.

Recommendations

The American Academy of Family Physicians emphasizes that weight management in childhood is an important goal, but notes a lack of evidence regarding the effectiveness of screening and treating overweight in children.⁵ The American Academy of Pediatrics recommends calculating and plotting BMI yearly to identify excessive weight gain.⁶

The United States Preventive Services Task Force, citing lack of evidence for treatment benefit, finds insufficient evidence for or against screening for overweight in children.⁷

Evidence summary

The Centers for Disease Control and Prevention (CDC) doesn’t use the term obesity to describe weight in children. Instead, the CDC defines overweight as body mass index (BMI) or weight-for-length for age and sex greater than the 95th percentile. Children above the 85th percentile are called “at risk for overweight.” A national expert panel recently recommended changing “at risk for overweight” to “overweight” and “overweight” to “obese”—the terminology used in this Clinical Inquiry.³

Studies find a clear connection

A 2008 systematic review found 25 prospective or retrospective longitudinal studies that examined the risk of overweight in adulthood based on overweight in childhood or adolescence. Studies had to include at least 1 anthropomorphic measurement before age 18 and at least 1 after age 18. The informativeness and validity of the studies were assessed using a standard evaluation tool. Because the review sought to provide results that could be generalized to large populations, it didn’t include studies of specific populations, such as former premature infants.

All of the 13 studies judged to be high quality found an elevated RR or odds ratio (OR) for adult obesity among participants who had been overweight as children. The authors didn’t calculate a composite measure of effect, but RRs in the individual studies ranged from 1.9 to 10.1.⁴

Older overweight children are at heightened risk

The systematic review considered children (≤12 years) and adolescents (>12 years) separately. Four high-quality studies assessed how many overweight children became overweight adults; 2 high-quality studies examined how many overweight children became obese adults. Overweight children had RRs between 1.9 and 3.6 for being overweight in adulthood compared with average-weight children; 1 study reported an OR of 7.0.

One study showed older overweight children to be at greater risk than younger children for overweight in adulthood: children who were overweight at 2 years of age had an RR of 2.7, whereas children who were overweight at 11 years had an RR of 3.6.

Obese children had similar results. A study of 4 age cohorts showed that children who were obese at 1 or 2 years of age had an OR of 1.3 for obesity in adulthood compared with average-weight peers. Obese children in the 3- to 5-year-old cohort had an OR of 4.7; obese children in the 6- to 9-year-old cohort had an OR of 8.8; and obese 10- to 14-year-olds had an OR of 22.3.

Risk increases with age

As with children, overweight adolescents had a higher risk of being overweight in adulthood. And the association between older age and higher ORs persisted into adolescence. One study found an OR of 17.5 for adult overweight among youngsters who were overweight at 10 to 14 years of age and an OR of 22.3 for adolescents who were overweight at 15 to 17 years.

Boys are at greater risk than girls

The systematic review also revealed sex differences. Two studies showed that overweight or obese boys were not only more likely to be overweight in adulthood than their average-weight peers (OR=15.0 in 1 study; RR=9.8 in the other), but also more likely to be overweight later in life than overweight or obese girls. The girls had an OR of 12.0 for adult overweight in 1 study and an RR of 6.8 in the other.

Recommendations

The American Academy of Family Physicians emphasizes that weight management in childhood is an important goal, but notes a lack of evidence regarding the effectiveness of screening and treating overweight in children.⁵ The American Academy of Pediatrics recommends calculating and plotting BMI yearly to identify excessive weight gain.⁶

The United States Preventive Services Task Force, citing lack of evidence for treatment benefit, finds insufficient evidence for or against screening for overweight in children.⁷

Evidence summary

The Centers for Disease Control and Prevention (CDC) doesn’t use the term obesity to describe weight in children. Instead, the CDC defines overweight as body mass index (BMI) or weight-for-length for age and sex greater than the 95th percentile. Children above the 85th percentile are called “at risk for overweight.” A national expert panel recently recommended changing “at risk for overweight” to “overweight” and “overweight” to “obese”—the terminology used in this Clinical Inquiry.³

Studies find a clear connection

A 2008 systematic review found 25 prospective or retrospective longitudinal studies that examined the risk of overweight in adulthood based on overweight in childhood or adolescence. Studies had to include at least 1 anthropomorphic measurement before age 18 and at least 1 after age 18. The informativeness and validity of the studies were assessed using a standard evaluation tool. Because the review sought to provide results that could be generalized to large populations, it didn’t include studies of specific populations, such as former premature infants.

All of the 13 studies judged to be high quality found an elevated RR or odds ratio (OR) for adult obesity among participants who had been overweight as children. The authors didn’t calculate a composite measure of effect, but RRs in the individual studies ranged from 1.9 to 10.1.⁴

Older overweight children are at heightened risk

The systematic review considered children (≤12 years) and adolescents (>12 years) separately. Four high-quality studies assessed how many overweight children became overweight adults; 2 high-quality studies examined how many overweight children became obese adults. Overweight children had RRs between 1.9 and 3.6 for being overweight in adulthood compared with average-weight children; 1 study reported an OR of 7.0.

One study showed older overweight children to be at greater risk than younger children for overweight in adulthood: children who were overweight at 2 years of age had an RR of 2.7, whereas children who were overweight at 11 years had an RR of 3.6.

Obese children had similar results. A study of 4 age cohorts showed that children who were obese at 1 or 2 years of age had an OR of 1.3 for obesity in adulthood compared with average-weight peers. Obese children in the 3- to 5-year-old cohort had an OR of 4.7; obese children in the 6- to 9-year-old cohort had an OR of 8.8; and obese 10- to 14-year-olds had an OR of 22.3.

Risk increases with age

As with children, overweight adolescents had a higher risk of being overweight in adulthood. And the association between older age and higher ORs persisted into adolescence. One study found an OR of 17.5 for adult overweight among youngsters who were overweight at 10 to 14 years of age and an OR of 22.3 for adolescents who were overweight at 15 to 17 years.

Boys are at greater risk than girls

The systematic review also revealed sex differences. Two studies showed that overweight or obese boys were not only more likely to be overweight in adulthood than their average-weight peers (OR=15.0 in 1 study; RR=9.8 in the other), but also more likely to be overweight later in life than overweight or obese girls. The girls had an OR of 12.0 for adult overweight in 1 study and an RR of 6.8 in the other.

Recommendations

The American Academy of Family Physicians emphasizes that weight management in childhood is an important goal, but notes a lack of evidence regarding the effectiveness of screening and treating overweight in children.⁵ The American Academy of Pediatrics recommends calculating and plotting BMI yearly to identify excessive weight gain.⁶

The United States Preventive Services Task Force, citing lack of evidence for treatment benefit, finds insufficient evidence for or against screening for overweight in children.⁷

Evidence summary

Postherpetic neuralgia (PHN) is defined as pain lasting 1 to 3 months after resolution of acute herpes zoster (shingles) rash. It occurs in approximately 10% to 15% of patients and can cause significant morbidity.

Tricyclic antidepressants provide effective pain relief

Five systematic reviews have concluded that tricyclic antidepressants (TCAs) are effective treatments for PHN.^1-5 Amitriptyline, the best studied TCA, provides at least moderate pain relief in two-thirds of patients with a pooled number needed to treat (NNT) for TCAs of 2.64 (95% confidence interval [CI], 2.1-3.54)⁵ (TABLE).

Selective serotonin reuptake inhibitors—including fluoxetine, paroxetine, citalopram, and sertraline—have been studied in a variety of neuropathic pain syndromes, but not for treating PHN.¹

TABLE
What’s the NNT for drugs used to treat postherpetic neuralgia?

Anticonvulsants help, too

Five systematic reviews found gabapentin to be effective, with a range of NNT from 2.8 to 5.3 for as much as 50% pain reduction based on the visual analog score (VAS).^2-6 Pregabalin is also effective, with an NNT of 4.93 (95% CI, 3.34-6.07) for up to 50% pain reduction.^7,8 Limited data are available concerning the effectiveness of valproate.⁵

A look at the role of narcotics

Four systematic reviews found that controlled-release oxycodone reduced pain by 50%, based on the VAS.^2-5 Another systematic review reported only limited evidence of effectiveness.⁶ In pooled results from systematic reviews, opioids decreased pain by 50% on the VAS (NNT=2.67; 95% CI, 2.10-3.77).⁶

An RCT of 76 patients demonstrated that morphine, with methadone as backup, both reduced the intensity of pain and relieved pain more than placebo.⁹

Tramadol, a selective opioid agonist, showed moderate effectiveness in a small RCT (N=125), with an NNT of 4.76 (95% CI, 2.61-26.97).^3,5,6 The mean pain intensity, degree of pain relief, and amount of rescue medication required were all better in the tramadol group than the placebo group.

Evidence for topical therapy is limited

The anesthetic lidocaine patch 5% has shown efficacy in PHN with allodynia based on 3 RCTs of lower quality (short duration, recruitment of patients who had improved on lidocaine previously, no report of baseline levels of pain); the NNT was 2 (95% CI, 1.4-3.3).¹⁰ A systematic review of these 3 RCTs concluded that evidence is insufficient to recommend the lidocaine patch as treatment for PHN.¹⁰

Capsaicin, a topical counterirritant, reduced pain in fewer than 20% of patients in 2 RCTs reported in systematic reviews, with an NNT of 3.26 (95% CI, 2.26-5.85).^2-6 Blinding was limited in these studies because of the stinging associated with treatment.

Recommendations

A 2004 practice parameter of the American Academy of Neurology recommends TCAs (amitriptyline, nortriptyline, desipramine, and maprotiline), gabapentin, pregabalin, opioids, topical lidocaine, and capsaicin to treat PHN (level of evidence: A), but notes that amitriptyline has significant cardiac effects in the elderly compared with nortriptyline and desipramine.³

In 2006, the European Federation of Neurological Societies determined that TCAs, gabapentin, pregabalin, and opioids had established efficacy (level of evidence: A), but recommended opioids as second-line therapy because of potential adverse events with long-term use.⁴ The federation’s guidelines designate capsaicin, tramadol, topical lidocaine, and valproate as drugs with lower efficacy or limited strength of evidence (level of evidence: B). Nevertheless, they recommend considering topical lidocaine for elderly patients with allodynia and small areas of pain.⁴

Evidence summary

Postherpetic neuralgia (PHN) is defined as pain lasting 1 to 3 months after resolution of acute herpes zoster (shingles) rash. It occurs in approximately 10% to 15% of patients and can cause significant morbidity.

Tricyclic antidepressants provide effective pain relief

Five systematic reviews have concluded that tricyclic antidepressants (TCAs) are effective treatments for PHN.^1-5 Amitriptyline, the best studied TCA, provides at least moderate pain relief in two-thirds of patients with a pooled number needed to treat (NNT) for TCAs of 2.64 (95% confidence interval [CI], 2.1-3.54)⁵ (TABLE).

Selective serotonin reuptake inhibitors—including fluoxetine, paroxetine, citalopram, and sertraline—have been studied in a variety of neuropathic pain syndromes, but not for treating PHN.¹

TABLE
What’s the NNT for drugs used to treat postherpetic neuralgia?

Anticonvulsants help, too

Five systematic reviews found gabapentin to be effective, with a range of NNT from 2.8 to 5.3 for as much as 50% pain reduction based on the visual analog score (VAS).^2-6 Pregabalin is also effective, with an NNT of 4.93 (95% CI, 3.34-6.07) for up to 50% pain reduction.^7,8 Limited data are available concerning the effectiveness of valproate.⁵

A look at the role of narcotics

Four systematic reviews found that controlled-release oxycodone reduced pain by 50%, based on the VAS.^2-5 Another systematic review reported only limited evidence of effectiveness.⁶ In pooled results from systematic reviews, opioids decreased pain by 50% on the VAS (NNT=2.67; 95% CI, 2.10-3.77).⁶

An RCT of 76 patients demonstrated that morphine, with methadone as backup, both reduced the intensity of pain and relieved pain more than placebo.⁹

Tramadol, a selective opioid agonist, showed moderate effectiveness in a small RCT (N=125), with an NNT of 4.76 (95% CI, 2.61-26.97).^3,5,6 The mean pain intensity, degree of pain relief, and amount of rescue medication required were all better in the tramadol group than the placebo group.

Evidence for topical therapy is limited

The anesthetic lidocaine patch 5% has shown efficacy in PHN with allodynia based on 3 RCTs of lower quality (short duration, recruitment of patients who had improved on lidocaine previously, no report of baseline levels of pain); the NNT was 2 (95% CI, 1.4-3.3).¹⁰ A systematic review of these 3 RCTs concluded that evidence is insufficient to recommend the lidocaine patch as treatment for PHN.¹⁰

Capsaicin, a topical counterirritant, reduced pain in fewer than 20% of patients in 2 RCTs reported in systematic reviews, with an NNT of 3.26 (95% CI, 2.26-5.85).^2-6 Blinding was limited in these studies because of the stinging associated with treatment.

Recommendations

A 2004 practice parameter of the American Academy of Neurology recommends TCAs (amitriptyline, nortriptyline, desipramine, and maprotiline), gabapentin, pregabalin, opioids, topical lidocaine, and capsaicin to treat PHN (level of evidence: A), but notes that amitriptyline has significant cardiac effects in the elderly compared with nortriptyline and desipramine.³

In 2006, the European Federation of Neurological Societies determined that TCAs, gabapentin, pregabalin, and opioids had established efficacy (level of evidence: A), but recommended opioids as second-line therapy because of potential adverse events with long-term use.⁴ The federation’s guidelines designate capsaicin, tramadol, topical lidocaine, and valproate as drugs with lower efficacy or limited strength of evidence (level of evidence: B). Nevertheless, they recommend considering topical lidocaine for elderly patients with allodynia and small areas of pain.⁴

Evidence summary

Postherpetic neuralgia (PHN) is defined as pain lasting 1 to 3 months after resolution of acute herpes zoster (shingles) rash. It occurs in approximately 10% to 15% of patients and can cause significant morbidity.

Tricyclic antidepressants provide effective pain relief

Five systematic reviews have concluded that tricyclic antidepressants (TCAs) are effective treatments for PHN.^1-5 Amitriptyline, the best studied TCA, provides at least moderate pain relief in two-thirds of patients with a pooled number needed to treat (NNT) for TCAs of 2.64 (95% confidence interval [CI], 2.1-3.54)⁵ (TABLE).

Selective serotonin reuptake inhibitors—including fluoxetine, paroxetine, citalopram, and sertraline—have been studied in a variety of neuropathic pain syndromes, but not for treating PHN.¹

TABLE
What’s the NNT for drugs used to treat postherpetic neuralgia?

Anticonvulsants help, too

Five systematic reviews found gabapentin to be effective, with a range of NNT from 2.8 to 5.3 for as much as 50% pain reduction based on the visual analog score (VAS).^2-6 Pregabalin is also effective, with an NNT of 4.93 (95% CI, 3.34-6.07) for up to 50% pain reduction.^7,8 Limited data are available concerning the effectiveness of valproate.⁵

A look at the role of narcotics

Four systematic reviews found that controlled-release oxycodone reduced pain by 50%, based on the VAS.^2-5 Another systematic review reported only limited evidence of effectiveness.⁶ In pooled results from systematic reviews, opioids decreased pain by 50% on the VAS (NNT=2.67; 95% CI, 2.10-3.77).⁶

An RCT of 76 patients demonstrated that morphine, with methadone as backup, both reduced the intensity of pain and relieved pain more than placebo.⁹

Tramadol, a selective opioid agonist, showed moderate effectiveness in a small RCT (N=125), with an NNT of 4.76 (95% CI, 2.61-26.97).^3,5,6 The mean pain intensity, degree of pain relief, and amount of rescue medication required were all better in the tramadol group than the placebo group.

Evidence for topical therapy is limited

The anesthetic lidocaine patch 5% has shown efficacy in PHN with allodynia based on 3 RCTs of lower quality (short duration, recruitment of patients who had improved on lidocaine previously, no report of baseline levels of pain); the NNT was 2 (95% CI, 1.4-3.3).¹⁰ A systematic review of these 3 RCTs concluded that evidence is insufficient to recommend the lidocaine patch as treatment for PHN.¹⁰

Capsaicin, a topical counterirritant, reduced pain in fewer than 20% of patients in 2 RCTs reported in systematic reviews, with an NNT of 3.26 (95% CI, 2.26-5.85).^2-6 Blinding was limited in these studies because of the stinging associated with treatment.

Recommendations

A 2004 practice parameter of the American Academy of Neurology recommends TCAs (amitriptyline, nortriptyline, desipramine, and maprotiline), gabapentin, pregabalin, opioids, topical lidocaine, and capsaicin to treat PHN (level of evidence: A), but notes that amitriptyline has significant cardiac effects in the elderly compared with nortriptyline and desipramine.³

In 2006, the European Federation of Neurological Societies determined that TCAs, gabapentin, pregabalin, and opioids had established efficacy (level of evidence: A), but recommended opioids as second-line therapy because of potential adverse events with long-term use.⁴ The federation’s guidelines designate capsaicin, tramadol, topical lidocaine, and valproate as drugs with lower efficacy or limited strength of evidence (level of evidence: B). Nevertheless, they recommend considering topical lidocaine for elderly patients with allodynia and small areas of pain.⁴

Evidence summary

The evidence supporting improved health outcomes resulting from group prenatal care is limited. We found 1 RCT,¹ 1 matched-cohort study,² and several pilot studies with descriptive analysis.^3-5 All data sets used a trademarked group prenatal care model, CenteringPregnancy. The TABLE summarizes the outcomes of group and individual prenatal care reported in the studies.

Fewer preterm births

One large, unblinded RCT investigated the effect of group prenatal care on a cohort of young, mostly minority women of low economic status. Women who received group prenatal care had fewer preterm births than those who received traditional care (number needed to treat [NNT]=25; P=.045).¹

A single cohort study compared pregnant teenagers enrolled in the CenteringPregnancy program with 2 clinic convenience samples. The group care recipients had significantly lower preterm delivery rates (NNT=7; P<.02).³ The study design, and therefore the detected relationship of group care to pregnancy-associated outcomes, may be particularly subject to selection bias.

Birth weight data are inconsistent

The matched cohort study recorded higher birth weights among infants born to mothers in group prenatal care.² Subset analysis of preterm infants born to mothers in group care showed average birth weights approximately 400 g higher than those in individual care (P<.05).² The RCT, however, found no clinically or statistically significant differences in birth weights between intervention and control groups.¹

TABLE
Pregnancy outcomes: Group vs individual prenatal care

Group care boosts breastfeeding, knowledge, and satisfaction

The RCT and the cohort study showed increased rates of breastfeeding initiation (NNT=8 and 6, respectively).^1,3 The RCT demonstrated that patients in group care more often had adequate prenatal care (NNT=16).¹ One cohort trial found that women enrolled in group prenatal care used the emergency department less during the third trimester (NNT=2, P=.001).⁴

Several studies have reported improved pregnancy knowledge and high levels of satisfaction with group prenatal care. The RCT showed increased knowledge and readiness for labor, and higher satisfaction compared with individual care (P<.001 for all outcomes).¹ Lower-quality studies of group care support these findings.^3-5

An innovative model that requires further study

Group prenatal care is a relatively new, innovative model of care, and limited data are available for review. The evidence from 1 RCT and 1 cohort study supports the protective effect of group prenatal care against preterm delivery for women at higher risk of adverse pregnancy outcomes.^1,2 Trends toward improved health outcomes were found in lower-quality studies; the trends were large enough to have potential clinical significance. These preliminary findings should be evaluated as primary health outcomes in future research to define the optimal population for group care.

Recommendations

No published guidelines or textbook recommendations exist for group-based prenatal care. In other areas of medical care, including diabetes and low back pain, specialty societies such as the American Diabetes Association and systematic reviews have supported practice changes, including group visits, to improve care.^6,7

Evidence summary

The evidence supporting improved health outcomes resulting from group prenatal care is limited. We found 1 RCT,¹ 1 matched-cohort study,² and several pilot studies with descriptive analysis.^3-5 All data sets used a trademarked group prenatal care model, CenteringPregnancy. The TABLE summarizes the outcomes of group and individual prenatal care reported in the studies.

Fewer preterm births

One large, unblinded RCT investigated the effect of group prenatal care on a cohort of young, mostly minority women of low economic status. Women who received group prenatal care had fewer preterm births than those who received traditional care (number needed to treat [NNT]=25; P=.045).¹

A single cohort study compared pregnant teenagers enrolled in the CenteringPregnancy program with 2 clinic convenience samples. The group care recipients had significantly lower preterm delivery rates (NNT=7; P<.02).³ The study design, and therefore the detected relationship of group care to pregnancy-associated outcomes, may be particularly subject to selection bias.

Birth weight data are inconsistent

The matched cohort study recorded higher birth weights among infants born to mothers in group prenatal care.² Subset analysis of preterm infants born to mothers in group care showed average birth weights approximately 400 g higher than those in individual care (P<.05).² The RCT, however, found no clinically or statistically significant differences in birth weights between intervention and control groups.¹

TABLE
Pregnancy outcomes: Group vs individual prenatal care

Group care boosts breastfeeding, knowledge, and satisfaction

The RCT and the cohort study showed increased rates of breastfeeding initiation (NNT=8 and 6, respectively).^1,3 The RCT demonstrated that patients in group care more often had adequate prenatal care (NNT=16).¹ One cohort trial found that women enrolled in group prenatal care used the emergency department less during the third trimester (NNT=2, P=.001).⁴

Several studies have reported improved pregnancy knowledge and high levels of satisfaction with group prenatal care. The RCT showed increased knowledge and readiness for labor, and higher satisfaction compared with individual care (P<.001 for all outcomes).¹ Lower-quality studies of group care support these findings.^3-5

An innovative model that requires further study

Group prenatal care is a relatively new, innovative model of care, and limited data are available for review. The evidence from 1 RCT and 1 cohort study supports the protective effect of group prenatal care against preterm delivery for women at higher risk of adverse pregnancy outcomes.^1,2 Trends toward improved health outcomes were found in lower-quality studies; the trends were large enough to have potential clinical significance. These preliminary findings should be evaluated as primary health outcomes in future research to define the optimal population for group care.

Recommendations

No published guidelines or textbook recommendations exist for group-based prenatal care. In other areas of medical care, including diabetes and low back pain, specialty societies such as the American Diabetes Association and systematic reviews have supported practice changes, including group visits, to improve care.^6,7

Evidence summary

The evidence supporting improved health outcomes resulting from group prenatal care is limited. We found 1 RCT,¹ 1 matched-cohort study,² and several pilot studies with descriptive analysis.^3-5 All data sets used a trademarked group prenatal care model, CenteringPregnancy. The TABLE summarizes the outcomes of group and individual prenatal care reported in the studies.

Fewer preterm births

One large, unblinded RCT investigated the effect of group prenatal care on a cohort of young, mostly minority women of low economic status. Women who received group prenatal care had fewer preterm births than those who received traditional care (number needed to treat [NNT]=25; P=.045).¹

A single cohort study compared pregnant teenagers enrolled in the CenteringPregnancy program with 2 clinic convenience samples. The group care recipients had significantly lower preterm delivery rates (NNT=7; P<.02).³ The study design, and therefore the detected relationship of group care to pregnancy-associated outcomes, may be particularly subject to selection bias.

Birth weight data are inconsistent

The matched cohort study recorded higher birth weights among infants born to mothers in group prenatal care.² Subset analysis of preterm infants born to mothers in group care showed average birth weights approximately 400 g higher than those in individual care (P<.05).² The RCT, however, found no clinically or statistically significant differences in birth weights between intervention and control groups.¹

TABLE
Pregnancy outcomes: Group vs individual prenatal care

Group care boosts breastfeeding, knowledge, and satisfaction

The RCT and the cohort study showed increased rates of breastfeeding initiation (NNT=8 and 6, respectively).^1,3 The RCT demonstrated that patients in group care more often had adequate prenatal care (NNT=16).¹ One cohort trial found that women enrolled in group prenatal care used the emergency department less during the third trimester (NNT=2, P=.001).⁴

Several studies have reported improved pregnancy knowledge and high levels of satisfaction with group prenatal care. The RCT showed increased knowledge and readiness for labor, and higher satisfaction compared with individual care (P<.001 for all outcomes).¹ Lower-quality studies of group care support these findings.^3-5

An innovative model that requires further study

Group prenatal care is a relatively new, innovative model of care, and limited data are available for review. The evidence from 1 RCT and 1 cohort study supports the protective effect of group prenatal care against preterm delivery for women at higher risk of adverse pregnancy outcomes.^1,2 Trends toward improved health outcomes were found in lower-quality studies; the trends were large enough to have potential clinical significance. These preliminary findings should be evaluated as primary health outcomes in future research to define the optimal population for group care.

Recommendations

No published guidelines or textbook recommendations exist for group-based prenatal care. In other areas of medical care, including diabetes and low back pain, specialty societies such as the American Diabetes Association and systematic reviews have supported practice changes, including group visits, to improve care.^6,7

Evidence summary

The Centers for Disease Control and Prevention defines overweight in children as weight-for-length greater than the 95th percentile for sex in children younger than 24 months and body mass index (BMI) greater than the 95th percentile for age and sex in children >24 months.

Breastfeeding is protective

Breastfed infants are less likely to be over-weight later in life than infants fed formula. A meta-analysis of 9 studies found that 7 showed a significantly lower risk of overweight among children who were breastfed (odds ratio [OR]=0.78; 95% confidence interval [CI], 0.71-0.85).¹

Four of the studies demonstrated that longer duration of breastfeeding offered greater protection than shorter duration. Two of the 4 studies defined longer duration as more than 6 months, 1 defined it as more than 3 months, and 1 examined breastfeeding for periods of less than 1 week, 1 week to 1 month, 2 to 3 months, 4 to 6 months, 7 to 9 months, and longer than 9 months, showing a duration-dependent decrease in risk. The other studies in the meta-analysis evaluated never-breastfed vs ever-breastfed infants.¹

Higher birth weight increases risk

Several meta-analyses report that birth weight is an early risk factor for later overweight. One found a positive association between birth weight and over-weight in childhood in 9 of 11 studies.² Another meta-analysis found a positive association in 25 of 28 studies that examined birth weight and BMI in childhood.³ These descriptive meta-analyses didn’t calculate pooled odds ratios (ORs) because of heterogeneity of the ages included and methods used to measure obesity.

A high rate of weight gain in infancy is also a risk factor for later overweight. One descriptive meta-analysis reported that 13 of 15 studies found a positive association between weight gain in the first year of life and overweight later in childhood, although overall OR and relative risk weren’t reported.⁴ A large cohort study found that each 100 g per month increase in weight gain above the mean (820 g per month) during the first 4 months of life increased the odds of overweight at 7 years of age by 38% (OR=1.38; 95% CI, 1.32-1.44).⁵

Socioeconomic status is a factor

Low socioeconomic status in infancy or early childhood increases the risk of childhood overweight, perhaps because of less breastfeeding and more smoking, among other factors.^6,7 Socioeconomic status was determined using the International Standard Classification of Occupations; children whose parents worked at unskilled manual labor jobs or were unemployed were considered in the lowest socioeconomic group.^6,7

A Brazilian study found that children born in the lowest socioeconomic group had BMI measurements at 18 years of age that were an average of 1.21 kg/m² higher than children in the highest socioeconomic group (P<.05). The study controlled for birth weight, maternal smoking, gestational age, and level of schooling eventually achieved by the child.⁸

Maternal overweight or obesity during the child’s infancy also increases the risk of childhood overweight.^9,10 Infants of obese parents were more likely to be overweight at 7 years, compared with children whose mothers were normal weight (OR=10.44; 95% CI, 5.11-21.23).⁹

Recommendations

The American Academy of Pediatrics (AAP) cites prevention of overweight as a potential benefit of breastfeeding.¹¹ The American Academy of Family Physicians notes that obese mothers should be especially encouraged to breastfeed.¹² The American Medical Association-AAP Expert Panel recommends breastfeeding; safe, free movement; and no television for infants to decrease the risk of later over-weight.¹³

Evidence summary

The Centers for Disease Control and Prevention defines overweight in children as weight-for-length greater than the 95th percentile for sex in children younger than 24 months and body mass index (BMI) greater than the 95th percentile for age and sex in children >24 months.

Breastfeeding is protective

Breastfed infants are less likely to be over-weight later in life than infants fed formula. A meta-analysis of 9 studies found that 7 showed a significantly lower risk of overweight among children who were breastfed (odds ratio [OR]=0.78; 95% confidence interval [CI], 0.71-0.85).¹

Four of the studies demonstrated that longer duration of breastfeeding offered greater protection than shorter duration. Two of the 4 studies defined longer duration as more than 6 months, 1 defined it as more than 3 months, and 1 examined breastfeeding for periods of less than 1 week, 1 week to 1 month, 2 to 3 months, 4 to 6 months, 7 to 9 months, and longer than 9 months, showing a duration-dependent decrease in risk. The other studies in the meta-analysis evaluated never-breastfed vs ever-breastfed infants.¹

Higher birth weight increases risk

Several meta-analyses report that birth weight is an early risk factor for later overweight. One found a positive association between birth weight and over-weight in childhood in 9 of 11 studies.² Another meta-analysis found a positive association in 25 of 28 studies that examined birth weight and BMI in childhood.³ These descriptive meta-analyses didn’t calculate pooled odds ratios (ORs) because of heterogeneity of the ages included and methods used to measure obesity.

A high rate of weight gain in infancy is also a risk factor for later overweight. One descriptive meta-analysis reported that 13 of 15 studies found a positive association between weight gain in the first year of life and overweight later in childhood, although overall OR and relative risk weren’t reported.⁴ A large cohort study found that each 100 g per month increase in weight gain above the mean (820 g per month) during the first 4 months of life increased the odds of overweight at 7 years of age by 38% (OR=1.38; 95% CI, 1.32-1.44).⁵

Socioeconomic status is a factor

Low socioeconomic status in infancy or early childhood increases the risk of childhood overweight, perhaps because of less breastfeeding and more smoking, among other factors.^6,7 Socioeconomic status was determined using the International Standard Classification of Occupations; children whose parents worked at unskilled manual labor jobs or were unemployed were considered in the lowest socioeconomic group.^6,7

A Brazilian study found that children born in the lowest socioeconomic group had BMI measurements at 18 years of age that were an average of 1.21 kg/m² higher than children in the highest socioeconomic group (P<.05). The study controlled for birth weight, maternal smoking, gestational age, and level of schooling eventually achieved by the child.⁸

Maternal overweight or obesity during the child’s infancy also increases the risk of childhood overweight.^9,10 Infants of obese parents were more likely to be overweight at 7 years, compared with children whose mothers were normal weight (OR=10.44; 95% CI, 5.11-21.23).⁹

Recommendations

The American Academy of Pediatrics (AAP) cites prevention of overweight as a potential benefit of breastfeeding.¹¹ The American Academy of Family Physicians notes that obese mothers should be especially encouraged to breastfeed.¹² The American Medical Association-AAP Expert Panel recommends breastfeeding; safe, free movement; and no television for infants to decrease the risk of later over-weight.¹³

Evidence summary

The Centers for Disease Control and Prevention defines overweight in children as weight-for-length greater than the 95th percentile for sex in children younger than 24 months and body mass index (BMI) greater than the 95th percentile for age and sex in children >24 months.

Breastfeeding is protective

Breastfed infants are less likely to be over-weight later in life than infants fed formula. A meta-analysis of 9 studies found that 7 showed a significantly lower risk of overweight among children who were breastfed (odds ratio [OR]=0.78; 95% confidence interval [CI], 0.71-0.85).¹

Four of the studies demonstrated that longer duration of breastfeeding offered greater protection than shorter duration. Two of the 4 studies defined longer duration as more than 6 months, 1 defined it as more than 3 months, and 1 examined breastfeeding for periods of less than 1 week, 1 week to 1 month, 2 to 3 months, 4 to 6 months, 7 to 9 months, and longer than 9 months, showing a duration-dependent decrease in risk. The other studies in the meta-analysis evaluated never-breastfed vs ever-breastfed infants.¹

Higher birth weight increases risk

Several meta-analyses report that birth weight is an early risk factor for later overweight. One found a positive association between birth weight and over-weight in childhood in 9 of 11 studies.² Another meta-analysis found a positive association in 25 of 28 studies that examined birth weight and BMI in childhood.³ These descriptive meta-analyses didn’t calculate pooled odds ratios (ORs) because of heterogeneity of the ages included and methods used to measure obesity.

A high rate of weight gain in infancy is also a risk factor for later overweight. One descriptive meta-analysis reported that 13 of 15 studies found a positive association between weight gain in the first year of life and overweight later in childhood, although overall OR and relative risk weren’t reported.⁴ A large cohort study found that each 100 g per month increase in weight gain above the mean (820 g per month) during the first 4 months of life increased the odds of overweight at 7 years of age by 38% (OR=1.38; 95% CI, 1.32-1.44).⁵

Socioeconomic status is a factor

Low socioeconomic status in infancy or early childhood increases the risk of childhood overweight, perhaps because of less breastfeeding and more smoking, among other factors.^6,7 Socioeconomic status was determined using the International Standard Classification of Occupations; children whose parents worked at unskilled manual labor jobs or were unemployed were considered in the lowest socioeconomic group.^6,7

A Brazilian study found that children born in the lowest socioeconomic group had BMI measurements at 18 years of age that were an average of 1.21 kg/m² higher than children in the highest socioeconomic group (P<.05). The study controlled for birth weight, maternal smoking, gestational age, and level of schooling eventually achieved by the child.⁸

Maternal overweight or obesity during the child’s infancy also increases the risk of childhood overweight.^9,10 Infants of obese parents were more likely to be overweight at 7 years, compared with children whose mothers were normal weight (OR=10.44; 95% CI, 5.11-21.23).⁹

Recommendations

The American Academy of Pediatrics (AAP) cites prevention of overweight as a potential benefit of breastfeeding.¹¹ The American Academy of Family Physicians notes that obese mothers should be especially encouraged to breastfeed.¹² The American Medical Association-AAP Expert Panel recommends breastfeeding; safe, free movement; and no television for infants to decrease the risk of later over-weight.¹³