PURLs

ILLUSTRATIVE CASE

The parents of an otherwise healthy 10-week-old breastfed baby girl bring her in for the second time in 2 weeks because of her in-consolable crying. Physical examination and work-up remain normal, and you again diagnose colic. What can you suggest to help decrease the baby’s crying, other than the usual dietary advice?

Colic affects up to 28% of infants, causing considerable stress for parents and for their health care providers.² Indeed, in the first 3 months of a baby’s life, crying is the No. 1 reason for pediatric visits.³ Parents often perceive—incorrectly—that the inconsolable crying is either a sign of serious illness or a result of poor parenting skills.⁴

A distressing problem, with few remedies
Despite the ubiquity and frustration that accompany colic, its exact etiology remains unclear and effective treatments remain elusive. With very little quality evidence to support interventions for colicky infants, we often have nothing more than grandmotherly advice to offer parents of babies with this vexing condition.

Current guidelines recommend only one strategy for breastfeeding mothers: a lowallergen diet.⁵ However, recent studies suggest that low counts of intestinal lactobacilli may play a role in colic and have documented improved symptoms after treatment with lactobacilli compared with treatment with simethicone.^6-8 Infant formulas that contain probiotics are now available, as a result.

Although the results of the recent studies have been promising, they were not double-blinded or placebo-controlled. The study detailed here is the first to provide compelling evidence for a safe intervention for colicky breastfed infants.

STUDY SUMMARY: Lactobacilli cut crying time

In a randomized, double-blinded, placebo-controlled trial, Savino et al randomly assigned 50 exclusively breastfed colicky infants ages 2 to 16 weeks to receive either L reuteri DSM 17 938 (10⁸ colony-forming units) or placebo daily for 21 days. Diagnosis of colic was made according to the modified Wessel’s criteria—fussy crying for ≥3 hours per day for ≥3 days per week in the week before enrollment. Their mothers were told to avoid ingesting cow’s milk during the course of the study.

Term infants adequate for gestational age were eligible for inclusion in the study. Exclusion criteria included evidence of chronic illness or gastrointestinal disorders, any intake of probiotics and/or antibiotics in the week preceding recruitment, and any formula feeding.

Parents and providers were blinded during the study, and they reported daily crying time, stool characteristics, adverse events, and growth patterns. An identical looking and tasting triglyceride oil without live bacteria was used in the placebo group. Each infant received 5 drops of L reuteri or placebo each morning 30 minutes before the morning feeding.

The primary outcome was a reduction of average crying time to <3 hours a day by Day 21. A secondary outcome was the number of infants in each group who experienced a 50% decrease in the daily average crying time from baseline on Days 7, 14, and 21 of the study.

Initially, the babies were divided equally between the control and intervention groups, but 4 participants in the control group were later excluded from analysis (1 had fever, 1 had reflux, and the parents of 2 infants did not complete the analysis).

At the time of enrollment, no significant differences were noted between the intervention and control groups regarding type of delivery, sex, age, family history of gastrointestinal disorders, growth parameters, and median daily crying time (370 minutes for the probiotics group vs 300 minutes for the placebo group).

By Day 21, the number of infants with crying times >3 hours was significantly lower in the treatment group compared with the placebo group (4 vs 12, respectively; P=.009). At all stages in the study, crying time for those in the treatment group was less than in the placebo group; median crying times for the intervention group were 95, 60, and 35 minutes per day, vs 185, 150, and 90 minutes for the controls, at 7, 14, and 21 days, respectively.

The number of infants with a 50% reduction in crying time was significantly greater in the treatment group than in the placebo group on Days 7, 14, and 21 (TABLE), although both groups saw an increase in the number of children whose average crying time had dropped by 50% as time went by. The number needed to treat to reduce crying time by 50% on Day 21 was 4.

There were no differences between the groups in growth, weight gain, frequency of stools, or incidence of regurgitation or constipation. No adverse events related to the treatment were reported.

TABLE
Babies respond* to probiotics

WHAT’S NEW: We have an evidence-based remedy that’s safe and effective

This study represents the first randomized, double-blinded, placebo-controlled investigation of probiotics to reduce infant colic. The researchers’ focus on patient-oriented outcomes and their solid study design move the notion of probiotics’ efficacy from conjecture to evidence. Furthermore, the study documents the safety of the intervention in the treatment group. This study increases our evidence-based armamentarium for treating colic, and family physicians should consider prescribing probiotics for healthy breastfed infants with colic.

CAVEATS: Will it work for bottle-fed infants?

This study was conducted in exclusively breast-fed, healthy infants whose mothers avoided dietary cow’s milk, which limits its applicability to a more general infant population. The study was funded by the makers of the probiotic, but the rigorous study design with random allocation, double-blind design, and intention-to-treat analysis makes bias unlikely. Although no adverse effects were reported during this study, there is little available evidence about the long-term effects of probiotics in infants. As L reuteri are naturally occurring gut bacteria, however, it seems unlikely that it would be harmful in the long term.

CHALLENGES TO IMPLEMENTATION: Parents will need to purchase the probiotics

As with any non–FDA-regulated product, it will be important to guide patients toward reputable manufacturers to ensure homogeneity of dosing. A 29-day supply of BioGaia probiotic drops (100 million units once a day), which costs $37 according to the manufacturer’s Web site, http://www.biogaia.com/consumer/biogaia-probiotic-products/probiotic-drops, should be affordable for most parents. Otherwise, little stands in the way of using this therapy to reduce the crying and subsequent stress associated with infant colic.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

The parents of an otherwise healthy 10-week-old breastfed baby girl bring her in for the second time in 2 weeks because of her in-consolable crying. Physical examination and work-up remain normal, and you again diagnose colic. What can you suggest to help decrease the baby’s crying, other than the usual dietary advice?

Colic affects up to 28% of infants, causing considerable stress for parents and for their health care providers.² Indeed, in the first 3 months of a baby’s life, crying is the No. 1 reason for pediatric visits.³ Parents often perceive—incorrectly—that the inconsolable crying is either a sign of serious illness or a result of poor parenting skills.⁴

A distressing problem, with few remedies
Despite the ubiquity and frustration that accompany colic, its exact etiology remains unclear and effective treatments remain elusive. With very little quality evidence to support interventions for colicky infants, we often have nothing more than grandmotherly advice to offer parents of babies with this vexing condition.

Current guidelines recommend only one strategy for breastfeeding mothers: a lowallergen diet.⁵ However, recent studies suggest that low counts of intestinal lactobacilli may play a role in colic and have documented improved symptoms after treatment with lactobacilli compared with treatment with simethicone.^6-8 Infant formulas that contain probiotics are now available, as a result.

Although the results of the recent studies have been promising, they were not double-blinded or placebo-controlled. The study detailed here is the first to provide compelling evidence for a safe intervention for colicky breastfed infants.

STUDY SUMMARY: Lactobacilli cut crying time

In a randomized, double-blinded, placebo-controlled trial, Savino et al randomly assigned 50 exclusively breastfed colicky infants ages 2 to 16 weeks to receive either L reuteri DSM 17 938 (10⁸ colony-forming units) or placebo daily for 21 days. Diagnosis of colic was made according to the modified Wessel’s criteria—fussy crying for ≥3 hours per day for ≥3 days per week in the week before enrollment. Their mothers were told to avoid ingesting cow’s milk during the course of the study.

Term infants adequate for gestational age were eligible for inclusion in the study. Exclusion criteria included evidence of chronic illness or gastrointestinal disorders, any intake of probiotics and/or antibiotics in the week preceding recruitment, and any formula feeding.

Parents and providers were blinded during the study, and they reported daily crying time, stool characteristics, adverse events, and growth patterns. An identical looking and tasting triglyceride oil without live bacteria was used in the placebo group. Each infant received 5 drops of L reuteri or placebo each morning 30 minutes before the morning feeding.

The primary outcome was a reduction of average crying time to <3 hours a day by Day 21. A secondary outcome was the number of infants in each group who experienced a 50% decrease in the daily average crying time from baseline on Days 7, 14, and 21 of the study.

Initially, the babies were divided equally between the control and intervention groups, but 4 participants in the control group were later excluded from analysis (1 had fever, 1 had reflux, and the parents of 2 infants did not complete the analysis).

At the time of enrollment, no significant differences were noted between the intervention and control groups regarding type of delivery, sex, age, family history of gastrointestinal disorders, growth parameters, and median daily crying time (370 minutes for the probiotics group vs 300 minutes for the placebo group).

By Day 21, the number of infants with crying times >3 hours was significantly lower in the treatment group compared with the placebo group (4 vs 12, respectively; P=.009). At all stages in the study, crying time for those in the treatment group was less than in the placebo group; median crying times for the intervention group were 95, 60, and 35 minutes per day, vs 185, 150, and 90 minutes for the controls, at 7, 14, and 21 days, respectively.

The number of infants with a 50% reduction in crying time was significantly greater in the treatment group than in the placebo group on Days 7, 14, and 21 (TABLE), although both groups saw an increase in the number of children whose average crying time had dropped by 50% as time went by. The number needed to treat to reduce crying time by 50% on Day 21 was 4.

There were no differences between the groups in growth, weight gain, frequency of stools, or incidence of regurgitation or constipation. No adverse events related to the treatment were reported.

TABLE
Babies respond* to probiotics

WHAT’S NEW: We have an evidence-based remedy that’s safe and effective

This study represents the first randomized, double-blinded, placebo-controlled investigation of probiotics to reduce infant colic. The researchers’ focus on patient-oriented outcomes and their solid study design move the notion of probiotics’ efficacy from conjecture to evidence. Furthermore, the study documents the safety of the intervention in the treatment group. This study increases our evidence-based armamentarium for treating colic, and family physicians should consider prescribing probiotics for healthy breastfed infants with colic.

CAVEATS: Will it work for bottle-fed infants?

This study was conducted in exclusively breast-fed, healthy infants whose mothers avoided dietary cow’s milk, which limits its applicability to a more general infant population. The study was funded by the makers of the probiotic, but the rigorous study design with random allocation, double-blind design, and intention-to-treat analysis makes bias unlikely. Although no adverse effects were reported during this study, there is little available evidence about the long-term effects of probiotics in infants. As L reuteri are naturally occurring gut bacteria, however, it seems unlikely that it would be harmful in the long term.

CHALLENGES TO IMPLEMENTATION: Parents will need to purchase the probiotics

As with any non–FDA-regulated product, it will be important to guide patients toward reputable manufacturers to ensure homogeneity of dosing. A 29-day supply of BioGaia probiotic drops (100 million units once a day), which costs $37 according to the manufacturer’s Web site, http://www.biogaia.com/consumer/biogaia-probiotic-products/probiotic-drops, should be affordable for most parents. Otherwise, little stands in the way of using this therapy to reduce the crying and subsequent stress associated with infant colic.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

The parents of an otherwise healthy 10-week-old breastfed baby girl bring her in for the second time in 2 weeks because of her in-consolable crying. Physical examination and work-up remain normal, and you again diagnose colic. What can you suggest to help decrease the baby’s crying, other than the usual dietary advice?

Colic affects up to 28% of infants, causing considerable stress for parents and for their health care providers.² Indeed, in the first 3 months of a baby’s life, crying is the No. 1 reason for pediatric visits.³ Parents often perceive—incorrectly—that the inconsolable crying is either a sign of serious illness or a result of poor parenting skills.⁴

A distressing problem, with few remedies
Despite the ubiquity and frustration that accompany colic, its exact etiology remains unclear and effective treatments remain elusive. With very little quality evidence to support interventions for colicky infants, we often have nothing more than grandmotherly advice to offer parents of babies with this vexing condition.

Current guidelines recommend only one strategy for breastfeeding mothers: a lowallergen diet.⁵ However, recent studies suggest that low counts of intestinal lactobacilli may play a role in colic and have documented improved symptoms after treatment with lactobacilli compared with treatment with simethicone.^6-8 Infant formulas that contain probiotics are now available, as a result.

Although the results of the recent studies have been promising, they were not double-blinded or placebo-controlled. The study detailed here is the first to provide compelling evidence for a safe intervention for colicky breastfed infants.

STUDY SUMMARY: Lactobacilli cut crying time

In a randomized, double-blinded, placebo-controlled trial, Savino et al randomly assigned 50 exclusively breastfed colicky infants ages 2 to 16 weeks to receive either L reuteri DSM 17 938 (10⁸ colony-forming units) or placebo daily for 21 days. Diagnosis of colic was made according to the modified Wessel’s criteria—fussy crying for ≥3 hours per day for ≥3 days per week in the week before enrollment. Their mothers were told to avoid ingesting cow’s milk during the course of the study.

Term infants adequate for gestational age were eligible for inclusion in the study. Exclusion criteria included evidence of chronic illness or gastrointestinal disorders, any intake of probiotics and/or antibiotics in the week preceding recruitment, and any formula feeding.

Parents and providers were blinded during the study, and they reported daily crying time, stool characteristics, adverse events, and growth patterns. An identical looking and tasting triglyceride oil without live bacteria was used in the placebo group. Each infant received 5 drops of L reuteri or placebo each morning 30 minutes before the morning feeding.

The primary outcome was a reduction of average crying time to <3 hours a day by Day 21. A secondary outcome was the number of infants in each group who experienced a 50% decrease in the daily average crying time from baseline on Days 7, 14, and 21 of the study.

Initially, the babies were divided equally between the control and intervention groups, but 4 participants in the control group were later excluded from analysis (1 had fever, 1 had reflux, and the parents of 2 infants did not complete the analysis).

At the time of enrollment, no significant differences were noted between the intervention and control groups regarding type of delivery, sex, age, family history of gastrointestinal disorders, growth parameters, and median daily crying time (370 minutes for the probiotics group vs 300 minutes for the placebo group).

By Day 21, the number of infants with crying times >3 hours was significantly lower in the treatment group compared with the placebo group (4 vs 12, respectively; P=.009). At all stages in the study, crying time for those in the treatment group was less than in the placebo group; median crying times for the intervention group were 95, 60, and 35 minutes per day, vs 185, 150, and 90 minutes for the controls, at 7, 14, and 21 days, respectively.

The number of infants with a 50% reduction in crying time was significantly greater in the treatment group than in the placebo group on Days 7, 14, and 21 (TABLE), although both groups saw an increase in the number of children whose average crying time had dropped by 50% as time went by. The number needed to treat to reduce crying time by 50% on Day 21 was 4.

There were no differences between the groups in growth, weight gain, frequency of stools, or incidence of regurgitation or constipation. No adverse events related to the treatment were reported.

TABLE
Babies respond* to probiotics

WHAT’S NEW: We have an evidence-based remedy that’s safe and effective

This study represents the first randomized, double-blinded, placebo-controlled investigation of probiotics to reduce infant colic. The researchers’ focus on patient-oriented outcomes and their solid study design move the notion of probiotics’ efficacy from conjecture to evidence. Furthermore, the study documents the safety of the intervention in the treatment group. This study increases our evidence-based armamentarium for treating colic, and family physicians should consider prescribing probiotics for healthy breastfed infants with colic.

CAVEATS: Will it work for bottle-fed infants?

This study was conducted in exclusively breast-fed, healthy infants whose mothers avoided dietary cow’s milk, which limits its applicability to a more general infant population. The study was funded by the makers of the probiotic, but the rigorous study design with random allocation, double-blind design, and intention-to-treat analysis makes bias unlikely. Although no adverse effects were reported during this study, there is little available evidence about the long-term effects of probiotics in infants. As L reuteri are naturally occurring gut bacteria, however, it seems unlikely that it would be harmful in the long term.

CHALLENGES TO IMPLEMENTATION: Parents will need to purchase the probiotics

As with any non–FDA-regulated product, it will be important to guide patients toward reputable manufacturers to ensure homogeneity of dosing. A 29-day supply of BioGaia probiotic drops (100 million units once a day), which costs $37 according to the manufacturer’s Web site, http://www.biogaia.com/consumer/biogaia-probiotic-products/probiotic-drops, should be affordable for most parents. Otherwise, little stands in the way of using this therapy to reduce the crying and subsequent stress associated with infant colic.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

A mother brings in her 14-year-old daughter for a routine check-up. The girl has no chronic medical problems and an unremarkable physical exam. When you’re alone with your patient, you inquire about substance abuse and sexual activity. She denies both. What questions would you ask to screen for depression?

Estimates of the prevalence of adolescent depression range from 3% to 9%,^2-4 and nearly 20% of teens will experience a depressive disorder before the age of 20.² But less than half of depressed adolescents are diagnosed or treated.²

Depressed teens face multiple risks
Teens with depressive disorders are at elevated risk, not only for poor family and social relationships and difficulties at school, but also for early pregnancy, substance abuse, hospitalization, recurrent episodes of depression, and suicide.^2,4 Thirteen percent of adolescents have seriously contemplated suicide, and 6.3% have made a suicide attempt in the previous 12 months.⁵

The US Preventive Services Task Force (USPSTF) recommends screening all adolescents for depression—provided that effective treatments and counseling are available for those who need it.⁶ Nearly all primary care clinicians agree that it is important to screen for adolescent depression, yet many feel hindered by both a lack of training and time constraints.² The study by Richardson et al¹ shows that targeted screening can be effective, even when time is tight.

STUDY SUMMARY: 2-question screen is fairly accurate

The Patient Health Questionnaire (PHQ)-9 is a simple and reasonably accurate test for depression in adults.⁷ A much shorter version, using only the first 2 of the PHQ-9’s questions, is an effective screening tool for adults.⁸ Richardson et al evaluated this brief screen—the PHQ-2—for adolescent depression.¹

The researchers invited 4000 teens (ages 13-17) who had seen a clinician within the previous 12 months to participate in a mailed survey, with parental or guardian approval. The survey included questions about age, gender, height, weight, sedentary and functional behaviors, and overall health, as well as depressive symptoms identified with the PHQ-2. This simple screen asks patients to rate how often in the past 2 weeks they have had:

1) a depressed mood, and/or

2) a lack of pleasure in usual activities.

Each question is scored from 0 to 3, with 0=not at all, and 3=nearly every day.

Next, the authors randomly selected 271 respondents with scores of ≥3 and 228 respondents with scores <3, matched for age and gender. Of those, 89% (n=444) participated in a longer telephone interview, which included the PHQ-9 and the Diagnostic Interview Schedule for Children (DISC-IV). Participants were predominantly female (60%), Caucasian (71%), and from urban areas (83%), with a mean household income of $57,442.

Compared with the DISC-IV—which the researchers considered the gold standard—the PHQ-2 had a sensitivity of 74% and a specificity of 75% at a cut point score ≥3; the sensitivity and specificity were 96% and 82%, respectively, for detecting young people who met the criteria for major depression on the PHQ-9. The area under the receiver operating curve was 0.84 (95% confidence interval, 0.75-0.92), meaning that the PHQ-2 correctly classified 84% of the participants as depressed or not depressed.

PHQ-2 helps identify related symptoms
Most of those with false-positive screens had other mental health problems. These included depressive symptoms that did not meet the criteria for major depression, an episode of major depression within the past year (but not in the last month), significant psychosocial impairment, and clinically significant anxiety symptoms.

WHAT’S NEW: Screening can be quick

Prior to this study, most validated tools for depression screening of adolescents were relatively time-consuming, and not likely to be performed during routine visits. The PHQ-2 is a reasonably accurate screen that requires minimal time (and minimal training).

CAVEATS: Study assessed a homogenous group

This study included mostly white girls from urban areas, relatively few of whom had public insurance. Whether the results are applicable to teens from different backgrounds is unclear. While the accuracy of the PHQ-2 was not perfect, almost 95% of those with a positive screen had some psychological problems.

CHALLENGES TO IMPLEMENTATION: Physicians may lack psych resources

Routinely using a 2-question screen for adolescent depression is unlikely to interfere with workflow in most practices. However, the USPSTF recommends screening teens only when there are systems in place to ensure accurate diagnosis, psychotherapy, and follow-up. Unfortunately, not all clinicians are adequately trained to diagnose or treat depressed teens, and some may lack access to appropriate psychotherapy referrals or consultation.

Despite the benefit of medications such as selective serotonin reuptake inhibitors (SSRIs) for teens with major depression, the antidepressants carry some risk. The black box warning for suicidality among adolescents treated with SSRIs⁹ necessitates accurate diagnosis, informed consent, and appropriate follow-up with clinicians who are comfortable treating adolescents.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

A mother brings in her 14-year-old daughter for a routine check-up. The girl has no chronic medical problems and an unremarkable physical exam. When you’re alone with your patient, you inquire about substance abuse and sexual activity. She denies both. What questions would you ask to screen for depression?

Estimates of the prevalence of adolescent depression range from 3% to 9%,^2-4 and nearly 20% of teens will experience a depressive disorder before the age of 20.² But less than half of depressed adolescents are diagnosed or treated.²

Depressed teens face multiple risks
Teens with depressive disorders are at elevated risk, not only for poor family and social relationships and difficulties at school, but also for early pregnancy, substance abuse, hospitalization, recurrent episodes of depression, and suicide.^2,4 Thirteen percent of adolescents have seriously contemplated suicide, and 6.3% have made a suicide attempt in the previous 12 months.⁵

The US Preventive Services Task Force (USPSTF) recommends screening all adolescents for depression—provided that effective treatments and counseling are available for those who need it.⁶ Nearly all primary care clinicians agree that it is important to screen for adolescent depression, yet many feel hindered by both a lack of training and time constraints.² The study by Richardson et al¹ shows that targeted screening can be effective, even when time is tight.

STUDY SUMMARY: 2-question screen is fairly accurate

The Patient Health Questionnaire (PHQ)-9 is a simple and reasonably accurate test for depression in adults.⁷ A much shorter version, using only the first 2 of the PHQ-9’s questions, is an effective screening tool for adults.⁸ Richardson et al evaluated this brief screen—the PHQ-2—for adolescent depression.¹

The researchers invited 4000 teens (ages 13-17) who had seen a clinician within the previous 12 months to participate in a mailed survey, with parental or guardian approval. The survey included questions about age, gender, height, weight, sedentary and functional behaviors, and overall health, as well as depressive symptoms identified with the PHQ-2. This simple screen asks patients to rate how often in the past 2 weeks they have had:

1) a depressed mood, and/or

2) a lack of pleasure in usual activities.

Each question is scored from 0 to 3, with 0=not at all, and 3=nearly every day.

Next, the authors randomly selected 271 respondents with scores of ≥3 and 228 respondents with scores <3, matched for age and gender. Of those, 89% (n=444) participated in a longer telephone interview, which included the PHQ-9 and the Diagnostic Interview Schedule for Children (DISC-IV). Participants were predominantly female (60%), Caucasian (71%), and from urban areas (83%), with a mean household income of $57,442.

Compared with the DISC-IV—which the researchers considered the gold standard—the PHQ-2 had a sensitivity of 74% and a specificity of 75% at a cut point score ≥3; the sensitivity and specificity were 96% and 82%, respectively, for detecting young people who met the criteria for major depression on the PHQ-9. The area under the receiver operating curve was 0.84 (95% confidence interval, 0.75-0.92), meaning that the PHQ-2 correctly classified 84% of the participants as depressed or not depressed.

PHQ-2 helps identify related symptoms
Most of those with false-positive screens had other mental health problems. These included depressive symptoms that did not meet the criteria for major depression, an episode of major depression within the past year (but not in the last month), significant psychosocial impairment, and clinically significant anxiety symptoms.

WHAT’S NEW: Screening can be quick

Prior to this study, most validated tools for depression screening of adolescents were relatively time-consuming, and not likely to be performed during routine visits. The PHQ-2 is a reasonably accurate screen that requires minimal time (and minimal training).

CAVEATS: Study assessed a homogenous group

This study included mostly white girls from urban areas, relatively few of whom had public insurance. Whether the results are applicable to teens from different backgrounds is unclear. While the accuracy of the PHQ-2 was not perfect, almost 95% of those with a positive screen had some psychological problems.

CHALLENGES TO IMPLEMENTATION: Physicians may lack psych resources

Routinely using a 2-question screen for adolescent depression is unlikely to interfere with workflow in most practices. However, the USPSTF recommends screening teens only when there are systems in place to ensure accurate diagnosis, psychotherapy, and follow-up. Unfortunately, not all clinicians are adequately trained to diagnose or treat depressed teens, and some may lack access to appropriate psychotherapy referrals or consultation.

Despite the benefit of medications such as selective serotonin reuptake inhibitors (SSRIs) for teens with major depression, the antidepressants carry some risk. The black box warning for suicidality among adolescents treated with SSRIs⁹ necessitates accurate diagnosis, informed consent, and appropriate follow-up with clinicians who are comfortable treating adolescents.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

A mother brings in her 14-year-old daughter for a routine check-up. The girl has no chronic medical problems and an unremarkable physical exam. When you’re alone with your patient, you inquire about substance abuse and sexual activity. She denies both. What questions would you ask to screen for depression?

Estimates of the prevalence of adolescent depression range from 3% to 9%,^2-4 and nearly 20% of teens will experience a depressive disorder before the age of 20.² But less than half of depressed adolescents are diagnosed or treated.²

Depressed teens face multiple risks
Teens with depressive disorders are at elevated risk, not only for poor family and social relationships and difficulties at school, but also for early pregnancy, substance abuse, hospitalization, recurrent episodes of depression, and suicide.^2,4 Thirteen percent of adolescents have seriously contemplated suicide, and 6.3% have made a suicide attempt in the previous 12 months.⁵

The US Preventive Services Task Force (USPSTF) recommends screening all adolescents for depression—provided that effective treatments and counseling are available for those who need it.⁶ Nearly all primary care clinicians agree that it is important to screen for adolescent depression, yet many feel hindered by both a lack of training and time constraints.² The study by Richardson et al¹ shows that targeted screening can be effective, even when time is tight.

STUDY SUMMARY: 2-question screen is fairly accurate

The Patient Health Questionnaire (PHQ)-9 is a simple and reasonably accurate test for depression in adults.⁷ A much shorter version, using only the first 2 of the PHQ-9’s questions, is an effective screening tool for adults.⁸ Richardson et al evaluated this brief screen—the PHQ-2—for adolescent depression.¹

The researchers invited 4000 teens (ages 13-17) who had seen a clinician within the previous 12 months to participate in a mailed survey, with parental or guardian approval. The survey included questions about age, gender, height, weight, sedentary and functional behaviors, and overall health, as well as depressive symptoms identified with the PHQ-2. This simple screen asks patients to rate how often in the past 2 weeks they have had:

1) a depressed mood, and/or

2) a lack of pleasure in usual activities.

Each question is scored from 0 to 3, with 0=not at all, and 3=nearly every day.

Next, the authors randomly selected 271 respondents with scores of ≥3 and 228 respondents with scores <3, matched for age and gender. Of those, 89% (n=444) participated in a longer telephone interview, which included the PHQ-9 and the Diagnostic Interview Schedule for Children (DISC-IV). Participants were predominantly female (60%), Caucasian (71%), and from urban areas (83%), with a mean household income of $57,442.

Compared with the DISC-IV—which the researchers considered the gold standard—the PHQ-2 had a sensitivity of 74% and a specificity of 75% at a cut point score ≥3; the sensitivity and specificity were 96% and 82%, respectively, for detecting young people who met the criteria for major depression on the PHQ-9. The area under the receiver operating curve was 0.84 (95% confidence interval, 0.75-0.92), meaning that the PHQ-2 correctly classified 84% of the participants as depressed or not depressed.

PHQ-2 helps identify related symptoms
Most of those with false-positive screens had other mental health problems. These included depressive symptoms that did not meet the criteria for major depression, an episode of major depression within the past year (but not in the last month), significant psychosocial impairment, and clinically significant anxiety symptoms.

WHAT’S NEW: Screening can be quick

Prior to this study, most validated tools for depression screening of adolescents were relatively time-consuming, and not likely to be performed during routine visits. The PHQ-2 is a reasonably accurate screen that requires minimal time (and minimal training).

CAVEATS: Study assessed a homogenous group

This study included mostly white girls from urban areas, relatively few of whom had public insurance. Whether the results are applicable to teens from different backgrounds is unclear. While the accuracy of the PHQ-2 was not perfect, almost 95% of those with a positive screen had some psychological problems.

CHALLENGES TO IMPLEMENTATION: Physicians may lack psych resources

Routinely using a 2-question screen for adolescent depression is unlikely to interfere with workflow in most practices. However, the USPSTF recommends screening teens only when there are systems in place to ensure accurate diagnosis, psychotherapy, and follow-up. Unfortunately, not all clinicians are adequately trained to diagnose or treat depressed teens, and some may lack access to appropriate psychotherapy referrals or consultation.

Despite the benefit of medications such as selective serotonin reuptake inhibitors (SSRIs) for teens with major depression, the antidepressants carry some risk. The black box warning for suicidality among adolescents treated with SSRIs⁹ necessitates accurate diagnosis, informed consent, and appropriate follow-up with clinicians who are comfortable treating adolescents.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

A 60-year-old man with cardiovascular disease and diabetes comes to your clinic for a routine check of his dyslipidemia, for which he is on statin therapy. His fasting lipid panel shows a low-density lipoprotein (LDL) of 70 mg/dL, triglycerides of 200 mg/dL, and a high-density lipoprotein (HDL) of 30 mg/dL. Should you recommend adding fenofibrate?

Patients with type 2 diabetes are at increased risk for cardiovascular events. National Cholesterol Education Program Adult Treatment Panel (NCEP ATP) III guidelines recommend treatment of dyslipidemia for all patients with high risk of cardiovascular events to an LDL goal of <100 mg/dL (optional <70 mg/dL), HDL goal of >40 mg/dL, and triglyceride goal of <150 mg/dL. These recommendations include the use of combination therapy with a statin and fenofibrate for patients who have elevated triglycerides and low HDL cholesterol despite being on statin therapy alone.²

Survival benefit of combo therapy remains unproven
We know that fibrate therapy alone in patients with type 2 diabetes reduces major cardiovascular events.^3,4 We also know that adding a fibrate to statin therapy can help patients reach their HDL and triglyceride targets. However, the survival benefit of the fibrate-statin combination over that of a statin alone has not been proven. In addition, there have been concerns about the increased risk of adverse effects with the combination. In fact, the overall benefits (and risks) of combining fibrates and statins for patients with diabetes and dyslipidemia were not addressed in a large randomized trial until the study we report on here.

STUDY SUMMARY: Statin + fibrate = minimal benefit for most patients

The Action to Control Cardiovascular Risk in Diabetes (ACCORD) study is among the largest trials conducted in adults with type 2 diabetes at high risk of cardiovascular events.⁵ The study examined 3 approaches to lowering the risk of major cardiovascular events: intensive lowering of blood sugar levels compared with standard blood sugar treatment; intensive lowering of blood pressure (BP) compared with standard BP treatment; and treatment of lipids with 2 drugs—a fibrate plus a statin—compared with a statin alone. This summary focuses on the lipid arm of the ACCORD study.¹

All patients in the study had type 2 diabetes and a hemoglobin A1c ≥7.5%. The study included patients ages 40 to 79 years with clinical evidence of cardiovascular disease and patients ages 55 to 79 years with either subclinical cardiovascular disease or ≥2 cardiovascular risk factors in addition to diabetes.

The lipid arm enrolled patients who had an LDL cholesterol of 60 to 180 mg/dL, an HDL cholesterol <55 mg/dL for women and blacks and <50 mg/dL for all other groups, and a triglyceride level <750 mg/dL for those not receiving lipid therapy and <400 mg/dL for those on lipid therapy. Enrollees (N=5518) were started on open-label simvastatin 20 mg, titrated up as needed to reach the LDL goal, then randomized to receive either fenofibrate or placebo 1 month later. The mean duration of follow-up was 4.7 years.

The primary outcome was the first occurrence of a major cardiovascular event— nonfatal myocardial infarction (MI), nonfatal stroke, or death from a cardiovascular cause. The annual rate of the primary outcome was 2.2% (n=291) in the fenofibrate-statin group and 2.4% (n=310) in the placebo group, a nonsignificant difference (P=.32).

The results were reported by sex. The primary outcome rate for men during the 4.7-year follow-up was 11.2% in the fenofibrate-statin group vs 13.3% in the placebo group; for women, the outcome rates were 9.1% in the treatment group and 6.6% in the placebo group (P =.01). These rates suggest a small benefit for men, and harm for women.

Subgroup analysis showed additional benefit from fenofibrate in patients with a combination of a high baseline triglyceride level (≥204 mg/dL) and very low baseline HDL cholesterol (≤34 mg/dL), representing about 16% of the study participants. The primary outcome rate for patients in this subgroup was 12.4% in the fenofibrate-statin group and 17.3% in the placebo group (P=.057); number needed to treat (NNT)=20 patients for 4.7 years to prevent 1 major cardiovascular event.

Harm was similar in both groups. A small number of patients had elevations of alanine aminotransferase of >3 times the upper limit of normal (1.9% in the fenofibrate-statin group and 1.5% in the statin group). The study drug was discontinued in 66 patients (2.4%) in the fenofibrate-statin group (and the placebo was discontinued in 30 patients [1.1%] in the statin group). The fenofibrate or statin dose was reduced in 440 patients (15.9%) in the fenofibrate-statin group and in 194 patients (7%) in the statin group due to a decrease in estimated glomerular filtration rate. There was no significant difference in the incidence of hemodialysis and end-stage renal disease (75 patients in the fenofibrate-statin group vs 77 patients in the statin group).

WHAT’S NEW: We have evidence that combo therapy doesn’t further reduce risk

This study examined a previously unaddressed question, the role of combination fibrate-statin therapy in high-risk patients with type 2 diabetes. The findings do not support the routine use of combination therapy compared with a statin alone for most patients with diabetes. Overall, combination therapy with simvastatin and fenofibrate did not lower the risk of MI, stroke, or death from cardiovascular disease more than simvastatin alone.

This trial showed that women with diabetes and hyperlipidemia should not be treated with both a statin and a fibrate. Men appeared to have a very small benefit from combination therapy (NNT=50). Patients with a baseline HDL ≤34 mg/dL and baseline triglyceride ≥204 mg/dL appeared to benefit from the combination, but this group constituted only 16% of the patients in this trial and the difference had borderline statistical significance. Nonetheless, it may be reasonable to treat such patients with combination therapy until a definitive study is done.

CAVEATS: Statin dose did not match standard practice

This study used a low dose of statin. The average daily simvastatin dose was 22.3 mg in the fenofibrate-statin group and 22.4 mg in the placebo group. This constitutes low-dose therapy compared with doses routinely used in practice (ie, 40 or 80 mg). A higher dose of simvastatin may have negated any outcome differences.

CHALLENGES TO IMPLEMENTATION: This “practice changer” conflicts with NCEP guidelines

The current NCEP ATP III guidelines recommend combination fibrate-statin therapy for all patients when statin therapy alone is not adequate to achieve lipid goals. This is a major challenge to our recommendation against using this combination for most patients with diabetes. Some physicians may choose to follow the ATP III guidelines rather than the new evidence because they feel more confident adhering to national guidelines.

Clinical inertia is another challenge, as clinicians may be hesitant to stop therapy in patients already on a fibrate-statin combination. Finally, specialists may continue to use fibrate-statin combinations in all patients with diabetes who do not achieve lipid goals on a statin, and family physicians may hesitate to contradict their recommendations.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

A 60-year-old man with cardiovascular disease and diabetes comes to your clinic for a routine check of his dyslipidemia, for which he is on statin therapy. His fasting lipid panel shows a low-density lipoprotein (LDL) of 70 mg/dL, triglycerides of 200 mg/dL, and a high-density lipoprotein (HDL) of 30 mg/dL. Should you recommend adding fenofibrate?

Patients with type 2 diabetes are at increased risk for cardiovascular events. National Cholesterol Education Program Adult Treatment Panel (NCEP ATP) III guidelines recommend treatment of dyslipidemia for all patients with high risk of cardiovascular events to an LDL goal of <100 mg/dL (optional <70 mg/dL), HDL goal of >40 mg/dL, and triglyceride goal of <150 mg/dL. These recommendations include the use of combination therapy with a statin and fenofibrate for patients who have elevated triglycerides and low HDL cholesterol despite being on statin therapy alone.²

Survival benefit of combo therapy remains unproven
We know that fibrate therapy alone in patients with type 2 diabetes reduces major cardiovascular events.^3,4 We also know that adding a fibrate to statin therapy can help patients reach their HDL and triglyceride targets. However, the survival benefit of the fibrate-statin combination over that of a statin alone has not been proven. In addition, there have been concerns about the increased risk of adverse effects with the combination. In fact, the overall benefits (and risks) of combining fibrates and statins for patients with diabetes and dyslipidemia were not addressed in a large randomized trial until the study we report on here.

STUDY SUMMARY: Statin + fibrate = minimal benefit for most patients

The Action to Control Cardiovascular Risk in Diabetes (ACCORD) study is among the largest trials conducted in adults with type 2 diabetes at high risk of cardiovascular events.⁵ The study examined 3 approaches to lowering the risk of major cardiovascular events: intensive lowering of blood sugar levels compared with standard blood sugar treatment; intensive lowering of blood pressure (BP) compared with standard BP treatment; and treatment of lipids with 2 drugs—a fibrate plus a statin—compared with a statin alone. This summary focuses on the lipid arm of the ACCORD study.¹

All patients in the study had type 2 diabetes and a hemoglobin A1c ≥7.5%. The study included patients ages 40 to 79 years with clinical evidence of cardiovascular disease and patients ages 55 to 79 years with either subclinical cardiovascular disease or ≥2 cardiovascular risk factors in addition to diabetes.

The lipid arm enrolled patients who had an LDL cholesterol of 60 to 180 mg/dL, an HDL cholesterol <55 mg/dL for women and blacks and <50 mg/dL for all other groups, and a triglyceride level <750 mg/dL for those not receiving lipid therapy and <400 mg/dL for those on lipid therapy. Enrollees (N=5518) were started on open-label simvastatin 20 mg, titrated up as needed to reach the LDL goal, then randomized to receive either fenofibrate or placebo 1 month later. The mean duration of follow-up was 4.7 years.

The primary outcome was the first occurrence of a major cardiovascular event— nonfatal myocardial infarction (MI), nonfatal stroke, or death from a cardiovascular cause. The annual rate of the primary outcome was 2.2% (n=291) in the fenofibrate-statin group and 2.4% (n=310) in the placebo group, a nonsignificant difference (P=.32).

The results were reported by sex. The primary outcome rate for men during the 4.7-year follow-up was 11.2% in the fenofibrate-statin group vs 13.3% in the placebo group; for women, the outcome rates were 9.1% in the treatment group and 6.6% in the placebo group (P =.01). These rates suggest a small benefit for men, and harm for women.

Subgroup analysis showed additional benefit from fenofibrate in patients with a combination of a high baseline triglyceride level (≥204 mg/dL) and very low baseline HDL cholesterol (≤34 mg/dL), representing about 16% of the study participants. The primary outcome rate for patients in this subgroup was 12.4% in the fenofibrate-statin group and 17.3% in the placebo group (P=.057); number needed to treat (NNT)=20 patients for 4.7 years to prevent 1 major cardiovascular event.

Harm was similar in both groups. A small number of patients had elevations of alanine aminotransferase of >3 times the upper limit of normal (1.9% in the fenofibrate-statin group and 1.5% in the statin group). The study drug was discontinued in 66 patients (2.4%) in the fenofibrate-statin group (and the placebo was discontinued in 30 patients [1.1%] in the statin group). The fenofibrate or statin dose was reduced in 440 patients (15.9%) in the fenofibrate-statin group and in 194 patients (7%) in the statin group due to a decrease in estimated glomerular filtration rate. There was no significant difference in the incidence of hemodialysis and end-stage renal disease (75 patients in the fenofibrate-statin group vs 77 patients in the statin group).

WHAT’S NEW: We have evidence that combo therapy doesn’t further reduce risk

This study examined a previously unaddressed question, the role of combination fibrate-statin therapy in high-risk patients with type 2 diabetes. The findings do not support the routine use of combination therapy compared with a statin alone for most patients with diabetes. Overall, combination therapy with simvastatin and fenofibrate did not lower the risk of MI, stroke, or death from cardiovascular disease more than simvastatin alone.

This trial showed that women with diabetes and hyperlipidemia should not be treated with both a statin and a fibrate. Men appeared to have a very small benefit from combination therapy (NNT=50). Patients with a baseline HDL ≤34 mg/dL and baseline triglyceride ≥204 mg/dL appeared to benefit from the combination, but this group constituted only 16% of the patients in this trial and the difference had borderline statistical significance. Nonetheless, it may be reasonable to treat such patients with combination therapy until a definitive study is done.

CAVEATS: Statin dose did not match standard practice

This study used a low dose of statin. The average daily simvastatin dose was 22.3 mg in the fenofibrate-statin group and 22.4 mg in the placebo group. This constitutes low-dose therapy compared with doses routinely used in practice (ie, 40 or 80 mg). A higher dose of simvastatin may have negated any outcome differences.

CHALLENGES TO IMPLEMENTATION: This “practice changer” conflicts with NCEP guidelines

The current NCEP ATP III guidelines recommend combination fibrate-statin therapy for all patients when statin therapy alone is not adequate to achieve lipid goals. This is a major challenge to our recommendation against using this combination for most patients with diabetes. Some physicians may choose to follow the ATP III guidelines rather than the new evidence because they feel more confident adhering to national guidelines.

Clinical inertia is another challenge, as clinicians may be hesitant to stop therapy in patients already on a fibrate-statin combination. Finally, specialists may continue to use fibrate-statin combinations in all patients with diabetes who do not achieve lipid goals on a statin, and family physicians may hesitate to contradict their recommendations.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

A 60-year-old man with cardiovascular disease and diabetes comes to your clinic for a routine check of his dyslipidemia, for which he is on statin therapy. His fasting lipid panel shows a low-density lipoprotein (LDL) of 70 mg/dL, triglycerides of 200 mg/dL, and a high-density lipoprotein (HDL) of 30 mg/dL. Should you recommend adding fenofibrate?

Patients with type 2 diabetes are at increased risk for cardiovascular events. National Cholesterol Education Program Adult Treatment Panel (NCEP ATP) III guidelines recommend treatment of dyslipidemia for all patients with high risk of cardiovascular events to an LDL goal of <100 mg/dL (optional <70 mg/dL), HDL goal of >40 mg/dL, and triglyceride goal of <150 mg/dL. These recommendations include the use of combination therapy with a statin and fenofibrate for patients who have elevated triglycerides and low HDL cholesterol despite being on statin therapy alone.²

Survival benefit of combo therapy remains unproven
We know that fibrate therapy alone in patients with type 2 diabetes reduces major cardiovascular events.^3,4 We also know that adding a fibrate to statin therapy can help patients reach their HDL and triglyceride targets. However, the survival benefit of the fibrate-statin combination over that of a statin alone has not been proven. In addition, there have been concerns about the increased risk of adverse effects with the combination. In fact, the overall benefits (and risks) of combining fibrates and statins for patients with diabetes and dyslipidemia were not addressed in a large randomized trial until the study we report on here.

STUDY SUMMARY: Statin + fibrate = minimal benefit for most patients

The Action to Control Cardiovascular Risk in Diabetes (ACCORD) study is among the largest trials conducted in adults with type 2 diabetes at high risk of cardiovascular events.⁵ The study examined 3 approaches to lowering the risk of major cardiovascular events: intensive lowering of blood sugar levels compared with standard blood sugar treatment; intensive lowering of blood pressure (BP) compared with standard BP treatment; and treatment of lipids with 2 drugs—a fibrate plus a statin—compared with a statin alone. This summary focuses on the lipid arm of the ACCORD study.¹

All patients in the study had type 2 diabetes and a hemoglobin A1c ≥7.5%. The study included patients ages 40 to 79 years with clinical evidence of cardiovascular disease and patients ages 55 to 79 years with either subclinical cardiovascular disease or ≥2 cardiovascular risk factors in addition to diabetes.

The lipid arm enrolled patients who had an LDL cholesterol of 60 to 180 mg/dL, an HDL cholesterol <55 mg/dL for women and blacks and <50 mg/dL for all other groups, and a triglyceride level <750 mg/dL for those not receiving lipid therapy and <400 mg/dL for those on lipid therapy. Enrollees (N=5518) were started on open-label simvastatin 20 mg, titrated up as needed to reach the LDL goal, then randomized to receive either fenofibrate or placebo 1 month later. The mean duration of follow-up was 4.7 years.

The primary outcome was the first occurrence of a major cardiovascular event— nonfatal myocardial infarction (MI), nonfatal stroke, or death from a cardiovascular cause. The annual rate of the primary outcome was 2.2% (n=291) in the fenofibrate-statin group and 2.4% (n=310) in the placebo group, a nonsignificant difference (P=.32).

The results were reported by sex. The primary outcome rate for men during the 4.7-year follow-up was 11.2% in the fenofibrate-statin group vs 13.3% in the placebo group; for women, the outcome rates were 9.1% in the treatment group and 6.6% in the placebo group (P =.01). These rates suggest a small benefit for men, and harm for women.

Subgroup analysis showed additional benefit from fenofibrate in patients with a combination of a high baseline triglyceride level (≥204 mg/dL) and very low baseline HDL cholesterol (≤34 mg/dL), representing about 16% of the study participants. The primary outcome rate for patients in this subgroup was 12.4% in the fenofibrate-statin group and 17.3% in the placebo group (P=.057); number needed to treat (NNT)=20 patients for 4.7 years to prevent 1 major cardiovascular event.

Harm was similar in both groups. A small number of patients had elevations of alanine aminotransferase of >3 times the upper limit of normal (1.9% in the fenofibrate-statin group and 1.5% in the statin group). The study drug was discontinued in 66 patients (2.4%) in the fenofibrate-statin group (and the placebo was discontinued in 30 patients [1.1%] in the statin group). The fenofibrate or statin dose was reduced in 440 patients (15.9%) in the fenofibrate-statin group and in 194 patients (7%) in the statin group due to a decrease in estimated glomerular filtration rate. There was no significant difference in the incidence of hemodialysis and end-stage renal disease (75 patients in the fenofibrate-statin group vs 77 patients in the statin group).

WHAT’S NEW: We have evidence that combo therapy doesn’t further reduce risk

This study examined a previously unaddressed question, the role of combination fibrate-statin therapy in high-risk patients with type 2 diabetes. The findings do not support the routine use of combination therapy compared with a statin alone for most patients with diabetes. Overall, combination therapy with simvastatin and fenofibrate did not lower the risk of MI, stroke, or death from cardiovascular disease more than simvastatin alone.

This trial showed that women with diabetes and hyperlipidemia should not be treated with both a statin and a fibrate. Men appeared to have a very small benefit from combination therapy (NNT=50). Patients with a baseline HDL ≤34 mg/dL and baseline triglyceride ≥204 mg/dL appeared to benefit from the combination, but this group constituted only 16% of the patients in this trial and the difference had borderline statistical significance. Nonetheless, it may be reasonable to treat such patients with combination therapy until a definitive study is done.

CAVEATS: Statin dose did not match standard practice

This study used a low dose of statin. The average daily simvastatin dose was 22.3 mg in the fenofibrate-statin group and 22.4 mg in the placebo group. This constitutes low-dose therapy compared with doses routinely used in practice (ie, 40 or 80 mg). A higher dose of simvastatin may have negated any outcome differences.

CHALLENGES TO IMPLEMENTATION: This “practice changer” conflicts with NCEP guidelines

The current NCEP ATP III guidelines recommend combination fibrate-statin therapy for all patients when statin therapy alone is not adequate to achieve lipid goals. This is a major challenge to our recommendation against using this combination for most patients with diabetes. Some physicians may choose to follow the ATP III guidelines rather than the new evidence because they feel more confident adhering to national guidelines.

Clinical inertia is another challenge, as clinicians may be hesitant to stop therapy in patients already on a fibrate-statin combination. Finally, specialists may continue to use fibrate-statin combinations in all patients with diabetes who do not achieve lipid goals on a statin, and family physicians may hesitate to contradict their recommendations.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A 67-year-old man comes in for a follow-up visit after being hospitalized for atrial fibrillation with a rapid ventricular rate. Before being discharged, he was put on warfarin and metoprolol, and his heart rate today is 96 bpm. You consider increasing the dose of his beta-blocker. What should his target heart rate be?

Guidelines cite stringent targets
The American College of Cardiology/American Heart Association Task Force/European Society of Cardiology (ACC/AHA/ESC) guidelines make no definite recommendations about heart rate targets. The guidelines do indicate, however, that rate control criteria vary based on age, “but usually involve achieving ventricular rates between 60 and 80 [bpm] at rest and between 90 and 115 [bpm] during moderate exercise.”⁴

This guidance is based on data from epidemiologic studies suggesting that faster heart rates in sinus rhythm may increase mortality from cardiovascular causes.⁵ However, strict control often requires higher doses of rate-controlling medications, which can lead to adverse events such as symptomatic bradycardia, dizziness, and syncope, as well as pacemaker implantation.

Pooled data suggest a more relaxed rate is better
A retrospective analysis of pooled data from the rate-control arms of the AFFIRM and RACE trials found no difference in all-cause mortality between the more stringent rate-control group in AFFIRM and the more lenient control in RACE.⁶ This finding suggested that more lenient heart rate targets may be preferred to avoid the adverse effects often associated with the higher doses of rate-controlling drugs needed to achieve strict control. The Rate Control Efficacy in Permanent Atrial Fibrillation: a Comparison between Lenient versus Strict Rate Control II (RACE II) study we report on here provides strong evidence in favor of lenient rate control.

STUDY SUMMARY: Lenient control is as effective, easier to achieve

RACE II was the first RCT to directly compare lenient rate control (resting heart rate <110 bpm) with strict rate control (resting heart rate <80 bpm, and <110 bpm during moderate exercise). This prospective, multi-center study in Holland randomized patients with permanent atrial fibrillation (N=614) to either a lenient or strict rate-control group. Eligibility criteria were (1) permanent atrial fibrillation for up to 12 months; (2) ≤80 years of age (3) mean resting heart rate >80 bpm; and (4) current use of oral anticoagulation therapy (or aspirin, in the absence of risk factors for thromboembolic complications).

Patients received various doses of beta-blockers, nondihydropyridine calcium-channel blockers, or digoxin, singly or in combination as needed to reach the target heart rate. In both groups, the resting heart rate was determined by 12-lead electrocardiogram after the patient remained in a supine position for 2 to 3 minutes. In the strict-control group, heart rate was also measured during moderate exercise on a stationary bicycle after the resting rate goal had been achieved. In addition, patients in the strict-control group wore a Holter monitor for 24 hours to check for bradycardia.

Participants in both groups were seen every 2 weeks until their heart rate goals were achieved, with follow-up at 1, 2, and 3 years. The primary composite outcome included death from cardiovascular causes; hospitalization for heart failure, stroke, systemic embolism, major bleeding, or life-threatening adverse effects of rate-control drugs; arrhythmic events, including sustained ventricular tachycardia, syncope, or cardiac arrest; and implantation of a pacemaker or cardioverter-defibrillator.

At the end of 3 years, the estimated cumulative incidence of the primary outcome was 12.9% in the lenient-control group vs 14.9% in the strict-control group. The absolute difference was -2.0 (90% confidence interval [CI], -7.6 to 3.5); a 90% CI was acceptable because the study only tested whether lenient control was worse than strict control. The frequency of reported symptoms and adverse events was similar between the 2 groups, but the lenient-control group had fewer visits for rate control (75 vs 684; P<.001), required fewer medications, and took lower doses of some medications.

Heart rate targets were met in 97.7% of patients in the lenient-control group, compared with 67% in the strict-control group (P<.001). Of those not meeting the strict control targets, 25% were due to an adverse medication event. There were no differences between the 2 groups in symptoms or in New York Heart Association functional class status.

WHAT'S NEW: Now we know: It doesn’t pay to go too low

A heart rate <80 at rest and <110 during exercise is difficult to maintain. This more stringent target often requires high dosages of drugs and/or multiple medications, which may lead to adverse effects. This RCT—the first to compare outcomes in patients with lenient vs strict heart rate control—found that morbidity and mortality were similar between the 2 groups. This means that, in many cases, patients will need less medication—leading to a reduction in risk of side effects and interactions.

CAVEATS: Unblinded study excluded very old, high risk

This was not a blinded study, so both patients and providers knew the target heart rates. However, the major outcomes were determined with relative objectivity and were not different between the 2 groups, so it is unlikely that this knowledge would have a major effect on the results. Nonetheless, this is a single study, and the findings are not yet supported by other large, prospective studies.

The researchers did not enroll patients >80 years, who have a higher incidence of atrial fibrillation and are less likely than younger patients to tolerate higher doses of rate-controlling medications. Also excluded were sedentary patients and those with a history of stroke, which resulted in a lower-risk study population. However, 40% of the subjects had a CHADS score ≥2 (an indication of high risk of stroke in patients with atrial fibrillation), and subgroup analysis found that the results applied to higher-risk groups.

Finally, it is possible that it may take longer than 3 years (the duration of study follow-up) for higher ventricular rates to result in adverse cardiovascular outcomes and that there could be a benefit of strict rate control over a longer period of time.

CHALLENGES TO IMPLEMENTATION: Guidelines do not reflect these findings

These findings are not yet incorporated into the ACC/AHA/ESC guidelines or those issued by other organizations. Clinical inertia may stop some physicians from reducing medications for patients with atrial fibrillation, but in general, both doctors and patients should welcome an easing of the drug burden.

Click here to view PURL METHODOLOGY

Illustrative case

A 67-year-old man comes in for a follow-up visit after being hospitalized for atrial fibrillation with a rapid ventricular rate. Before being discharged, he was put on warfarin and metoprolol, and his heart rate today is 96 bpm. You consider increasing the dose of his beta-blocker. What should his target heart rate be?

Guidelines cite stringent targets
The American College of Cardiology/American Heart Association Task Force/European Society of Cardiology (ACC/AHA/ESC) guidelines make no definite recommendations about heart rate targets. The guidelines do indicate, however, that rate control criteria vary based on age, “but usually involve achieving ventricular rates between 60 and 80 [bpm] at rest and between 90 and 115 [bpm] during moderate exercise.”⁴

This guidance is based on data from epidemiologic studies suggesting that faster heart rates in sinus rhythm may increase mortality from cardiovascular causes.⁵ However, strict control often requires higher doses of rate-controlling medications, which can lead to adverse events such as symptomatic bradycardia, dizziness, and syncope, as well as pacemaker implantation.

Pooled data suggest a more relaxed rate is better
A retrospective analysis of pooled data from the rate-control arms of the AFFIRM and RACE trials found no difference in all-cause mortality between the more stringent rate-control group in AFFIRM and the more lenient control in RACE.⁶ This finding suggested that more lenient heart rate targets may be preferred to avoid the adverse effects often associated with the higher doses of rate-controlling drugs needed to achieve strict control. The Rate Control Efficacy in Permanent Atrial Fibrillation: a Comparison between Lenient versus Strict Rate Control II (RACE II) study we report on here provides strong evidence in favor of lenient rate control.

STUDY SUMMARY: Lenient control is as effective, easier to achieve

RACE II was the first RCT to directly compare lenient rate control (resting heart rate <110 bpm) with strict rate control (resting heart rate <80 bpm, and <110 bpm during moderate exercise). This prospective, multi-center study in Holland randomized patients with permanent atrial fibrillation (N=614) to either a lenient or strict rate-control group. Eligibility criteria were (1) permanent atrial fibrillation for up to 12 months; (2) ≤80 years of age (3) mean resting heart rate >80 bpm; and (4) current use of oral anticoagulation therapy (or aspirin, in the absence of risk factors for thromboembolic complications).

Patients received various doses of beta-blockers, nondihydropyridine calcium-channel blockers, or digoxin, singly or in combination as needed to reach the target heart rate. In both groups, the resting heart rate was determined by 12-lead electrocardiogram after the patient remained in a supine position for 2 to 3 minutes. In the strict-control group, heart rate was also measured during moderate exercise on a stationary bicycle after the resting rate goal had been achieved. In addition, patients in the strict-control group wore a Holter monitor for 24 hours to check for bradycardia.

Participants in both groups were seen every 2 weeks until their heart rate goals were achieved, with follow-up at 1, 2, and 3 years. The primary composite outcome included death from cardiovascular causes; hospitalization for heart failure, stroke, systemic embolism, major bleeding, or life-threatening adverse effects of rate-control drugs; arrhythmic events, including sustained ventricular tachycardia, syncope, or cardiac arrest; and implantation of a pacemaker or cardioverter-defibrillator.

At the end of 3 years, the estimated cumulative incidence of the primary outcome was 12.9% in the lenient-control group vs 14.9% in the strict-control group. The absolute difference was -2.0 (90% confidence interval [CI], -7.6 to 3.5); a 90% CI was acceptable because the study only tested whether lenient control was worse than strict control. The frequency of reported symptoms and adverse events was similar between the 2 groups, but the lenient-control group had fewer visits for rate control (75 vs 684; P<.001), required fewer medications, and took lower doses of some medications.

Heart rate targets were met in 97.7% of patients in the lenient-control group, compared with 67% in the strict-control group (P<.001). Of those not meeting the strict control targets, 25% were due to an adverse medication event. There were no differences between the 2 groups in symptoms or in New York Heart Association functional class status.

WHAT'S NEW: Now we know: It doesn’t pay to go too low

A heart rate <80 at rest and <110 during exercise is difficult to maintain. This more stringent target often requires high dosages of drugs and/or multiple medications, which may lead to adverse effects. This RCT—the first to compare outcomes in patients with lenient vs strict heart rate control—found that morbidity and mortality were similar between the 2 groups. This means that, in many cases, patients will need less medication—leading to a reduction in risk of side effects and interactions.

CAVEATS: Unblinded study excluded very old, high risk

This was not a blinded study, so both patients and providers knew the target heart rates. However, the major outcomes were determined with relative objectivity and were not different between the 2 groups, so it is unlikely that this knowledge would have a major effect on the results. Nonetheless, this is a single study, and the findings are not yet supported by other large, prospective studies.

The researchers did not enroll patients >80 years, who have a higher incidence of atrial fibrillation and are less likely than younger patients to tolerate higher doses of rate-controlling medications. Also excluded were sedentary patients and those with a history of stroke, which resulted in a lower-risk study population. However, 40% of the subjects had a CHADS score ≥2 (an indication of high risk of stroke in patients with atrial fibrillation), and subgroup analysis found that the results applied to higher-risk groups.

Finally, it is possible that it may take longer than 3 years (the duration of study follow-up) for higher ventricular rates to result in adverse cardiovascular outcomes and that there could be a benefit of strict rate control over a longer period of time.

CHALLENGES TO IMPLEMENTATION: Guidelines do not reflect these findings

These findings are not yet incorporated into the ACC/AHA/ESC guidelines or those issued by other organizations. Clinical inertia may stop some physicians from reducing medications for patients with atrial fibrillation, but in general, both doctors and patients should welcome an easing of the drug burden.

Click here to view PURL METHODOLOGY

Illustrative case

A 67-year-old man comes in for a follow-up visit after being hospitalized for atrial fibrillation with a rapid ventricular rate. Before being discharged, he was put on warfarin and metoprolol, and his heart rate today is 96 bpm. You consider increasing the dose of his beta-blocker. What should his target heart rate be?

Guidelines cite stringent targets
The American College of Cardiology/American Heart Association Task Force/European Society of Cardiology (ACC/AHA/ESC) guidelines make no definite recommendations about heart rate targets. The guidelines do indicate, however, that rate control criteria vary based on age, “but usually involve achieving ventricular rates between 60 and 80 [bpm] at rest and between 90 and 115 [bpm] during moderate exercise.”⁴

This guidance is based on data from epidemiologic studies suggesting that faster heart rates in sinus rhythm may increase mortality from cardiovascular causes.⁵ However, strict control often requires higher doses of rate-controlling medications, which can lead to adverse events such as symptomatic bradycardia, dizziness, and syncope, as well as pacemaker implantation.

Pooled data suggest a more relaxed rate is better
A retrospective analysis of pooled data from the rate-control arms of the AFFIRM and RACE trials found no difference in all-cause mortality between the more stringent rate-control group in AFFIRM and the more lenient control in RACE.⁶ This finding suggested that more lenient heart rate targets may be preferred to avoid the adverse effects often associated with the higher doses of rate-controlling drugs needed to achieve strict control. The Rate Control Efficacy in Permanent Atrial Fibrillation: a Comparison between Lenient versus Strict Rate Control II (RACE II) study we report on here provides strong evidence in favor of lenient rate control.

STUDY SUMMARY: Lenient control is as effective, easier to achieve

RACE II was the first RCT to directly compare lenient rate control (resting heart rate <110 bpm) with strict rate control (resting heart rate <80 bpm, and <110 bpm during moderate exercise). This prospective, multi-center study in Holland randomized patients with permanent atrial fibrillation (N=614) to either a lenient or strict rate-control group. Eligibility criteria were (1) permanent atrial fibrillation for up to 12 months; (2) ≤80 years of age (3) mean resting heart rate >80 bpm; and (4) current use of oral anticoagulation therapy (or aspirin, in the absence of risk factors for thromboembolic complications).

Patients received various doses of beta-blockers, nondihydropyridine calcium-channel blockers, or digoxin, singly or in combination as needed to reach the target heart rate. In both groups, the resting heart rate was determined by 12-lead electrocardiogram after the patient remained in a supine position for 2 to 3 minutes. In the strict-control group, heart rate was also measured during moderate exercise on a stationary bicycle after the resting rate goal had been achieved. In addition, patients in the strict-control group wore a Holter monitor for 24 hours to check for bradycardia.

Participants in both groups were seen every 2 weeks until their heart rate goals were achieved, with follow-up at 1, 2, and 3 years. The primary composite outcome included death from cardiovascular causes; hospitalization for heart failure, stroke, systemic embolism, major bleeding, or life-threatening adverse effects of rate-control drugs; arrhythmic events, including sustained ventricular tachycardia, syncope, or cardiac arrest; and implantation of a pacemaker or cardioverter-defibrillator.

At the end of 3 years, the estimated cumulative incidence of the primary outcome was 12.9% in the lenient-control group vs 14.9% in the strict-control group. The absolute difference was -2.0 (90% confidence interval [CI], -7.6 to 3.5); a 90% CI was acceptable because the study only tested whether lenient control was worse than strict control. The frequency of reported symptoms and adverse events was similar between the 2 groups, but the lenient-control group had fewer visits for rate control (75 vs 684; P<.001), required fewer medications, and took lower doses of some medications.

Heart rate targets were met in 97.7% of patients in the lenient-control group, compared with 67% in the strict-control group (P<.001). Of those not meeting the strict control targets, 25% were due to an adverse medication event. There were no differences between the 2 groups in symptoms or in New York Heart Association functional class status.

WHAT'S NEW: Now we know: It doesn’t pay to go too low

A heart rate <80 at rest and <110 during exercise is difficult to maintain. This more stringent target often requires high dosages of drugs and/or multiple medications, which may lead to adverse effects. This RCT—the first to compare outcomes in patients with lenient vs strict heart rate control—found that morbidity and mortality were similar between the 2 groups. This means that, in many cases, patients will need less medication—leading to a reduction in risk of side effects and interactions.

CAVEATS: Unblinded study excluded very old, high risk

This was not a blinded study, so both patients and providers knew the target heart rates. However, the major outcomes were determined with relative objectivity and were not different between the 2 groups, so it is unlikely that this knowledge would have a major effect on the results. Nonetheless, this is a single study, and the findings are not yet supported by other large, prospective studies.

The researchers did not enroll patients >80 years, who have a higher incidence of atrial fibrillation and are less likely than younger patients to tolerate higher doses of rate-controlling medications. Also excluded were sedentary patients and those with a history of stroke, which resulted in a lower-risk study population. However, 40% of the subjects had a CHADS score ≥2 (an indication of high risk of stroke in patients with atrial fibrillation), and subgroup analysis found that the results applied to higher-risk groups.

Finally, it is possible that it may take longer than 3 years (the duration of study follow-up) for higher ventricular rates to result in adverse cardiovascular outcomes and that there could be a benefit of strict rate control over a longer period of time.

CHALLENGES TO IMPLEMENTATION: Guidelines do not reflect these findings

These findings are not yet incorporated into the ACC/AHA/ESC guidelines or those issued by other organizations. Clinical inertia may stop some physicians from reducing medications for patients with atrial fibrillation, but in general, both doctors and patients should welcome an easing of the drug burden.

Click here to view PURL METHODOLOGY

Illustrative case

A 10-year-old boy comes in with his mother for a well-child check-up. His BMI is 40 kg/m²—above the 99th percentile for his age and up from 37 a year ago. His blood pressure is 120/84 mm Hg. What treatment, if any, should you offer for his obesity?

Childhood obesity is a global epidemic. In the United States, 19.6% of children ages 6 through 11 and 18.1% of 12- to 19-year-olds are obese, a 3-fold increase in the last 30 years.³ Without intervention, most obese adolescents will become obese adults, threatening to reverse the progress in slowing cardiovascular morbidity and mortality that has occurred over the past few decades.³

Obese kids get adult diseases
Obesity is a risk factor for a variety of chronic conditions, including cardiovascular disease, cerebrovascular disease, and arthritis. Severe obesity is also associated with higher mortality rates.⁴ Unfortunately, these comorbidities are not limited to adulthood.

“Adult” diseases, such as obstructive sleep apnea, dyslipidemia, and type 2 diabetes, are increasingly seen in children and adolescents.¹ Nutritional deficits such as vitamin D and iron deficiency are often seen in obese children, as well.⁵ There are also psychological ramifications of childhood obesity, including social isolation and depression.⁶

The USPSTF recently upgraded its recommendation regarding obesity screening in children ages 6 and older from I (insufficient evidence) to B (a positive grade based on high or moderate certainty of the benefit of the intervention), citing new evidence in favor of screening and treating or referring children when appropriate.² The systematic review we report on here, which formed the basis for the USPSTF’s upgrade, focused on management options for children identified as overweight or obese.

STUDY SUMMARY: Intense, comprehensive efforts pay off

This systematic review¹ included studies of children ages 4 to 18 years who were overweight (defined as a body mass index [BMI] in the 85th to 94th percentile for age and sex) or obese (either a BMI at or above the 95th percentile for age and sex or a BMI >30 kg/m²). The researchers found 25 trials—15 of behavioral interventions alone and 10 that combined behavioral and pharmacologic interventions—that met their criteria: The studies focused on weight loss and/or maintenance, reported outcomes ≥6 months from baseline, and were conducted (or feasible) in a primary care setting.

Behavioral interventions were categorized by treatment intensity (as measured by hours of contact, which ranged from <10 hours to >75) and comprehensiveness (including nutritional counseling, physical activity counseling or participation, and counseling on behavioral management techniques). Weight outcomes were categorized as short-term (6-12 months since treatment initiation) or maintenance (≥12 months after the end of active treatment).

The 15 behavioral intervention trials included 1258 children ages 4 to 18 years, most of whom were obese. Most trials were small and reported high retention rates. All had beneficial effects on weight in the intervention group compared with the controls, but not all changes were statistically significant. Higher intensity and more comprehensive programs had better outcomes.

The largest effects were in 3 moderate- to high-intensity, comprehensive weight management programs with ≥26 hours of contact. These 3 trials demonstrated a difference in BMI of 1.9 to 3.3 in the intervention groups at 12 months compared with the controls. (A 3.3 difference in BMI is equal to approximately 13 lb in an 8-year-old and 17 lb in a 12-year-old.)

Four behavioral intervention studies reported outcomes ≥12 months after completing the intervention (range 15-48 months). Three of the 4 reported continued beneficial effects on weight after the active treatment period, but the effects were markedly attenuated.

The only adverse effect reported in the trials of behavioral interventions was the injury rate among children in an exercise program, but it was minimal: One fracture was reported, vs no injuries for the controls. No differences were reported in height, eating disorders, or depression. However, fewer than half of the behavioral intervention trials reported on adverse effects.

Weight loss drugs have modest effects
Ten trials combining pharmacologic and behavioral interventions involved a total of 1294 obese adolescents ages 12 to 19. All evaluated short-term weight loss effects of either sibutramine (10-15 mg/d) or orlistat (120 mg tid). Trials ranged from 3 to 12 months. Participants in both the control and intervention groups received behavioral counseling.

The trials all favored the treatment groups, although not all of the results were statistically significant. Trials of longer duration (12 months) had more favorable results than those lasting 6 months.

The largest sibutramine trial (n=498) reported a mean BMI reduction of 2.9 in the treatment group, compared with a reduction of 0.3 in the control group (P<.001). This corresponds to an average weight loss of 14 lb in the intervention group, vs 4.2 lb in the control group, after 12 months.

The largest orlistat trial (n=539) reported a mean BMI reduction in the treatment group of 0.6, vs 0.3 in the control group (P<.001)—an average weight loss of 4.2 lb in the intervention group, compared with 2.1 lb among the controls after 12 months. None of the trials evaluated weight change after cessation of the study drug, and none compared orlistat with sibutramine.

Adverse effects in the sibutramine-treated patients were primarily cardiovascular and gastrointestinal. Cardiovascular effects included tachycardia and increases in systolic and diastolic blood pressure. The differences between the intervention and control groups were small, and no differences were observed in discontinuation rates caused by adverse events. Nor were differences reported in growth and maturation between the intervention and control groups.

Adverse effects in the orlistat-treated patients were also low and similar in the intervention and control groups. Gastrointestinal effects were common. The number needed to harm (NNH) for fatty or oily stools was 2,⁴ and the NNH for fecal incontinence was 12.⁵

WHAT'S NEW: Clinicians treating obese kids have cause for optimism

Although the trials included in this review were heterogeneous and many were small, this systematic review provides evidence that intensive, comprehensive behavioral weight loss interventions for obese children can be effective up to 12 months after the conclusion of the program. Family physicians should consider referring obese children and adolescents to such programs—or finding ways to provide supportive strategies themselves.

Sibutramine and orlistat may be helpful in the context of comprehensive, intensive behavioral interventions, although there is no follow-up data to demonstrate long-term safety and weight maintenance after the medication is stopped.

CAVEATS: Little is known about long-term safety of the drugs

There have been few randomized trials of pharmacologic interventions in adolescents and none evaluating weight maintenance after 12 months (or discontinuation of treatment), or assessing long-term safety of the medication.

Sibutramine is not approved by the US Food and Drug Administration (FDA) for use in children or adolescents.⁷ Orlistat is currently approved only for individuals over the age of 12.⁸

In January 2010, an additional contraindication was added to the sibutramine drug label, stating that it is not to be used in patients with a history of cardiovascular disease.⁹ And the FDA is currently investigating a rare association between orlistat and liver injury, although no conclusions have been released.¹⁰ Children and adolescents are particularly vulnerable to long-term side effects, given their relatively young age at the time of drug initiation, so we urge caution with the use of these drugs in this patient population.

CHALLENGES TO IMPLEMENTATION: Intensive approach may be hard to reproduce

Implementation of high-intensity comprehensive interventions for obese children faces a number of roadblocks, including limited availability of programs, cost, and reimbursement. Most of the intensive interventions in these trials took place in specialty centers rather than in primary care offices. Replicating them could require a referral—or significant resources within the primary care setting itself. Yet many, if not most, insurance policies still do not cover such extensive lifestyle interventions. (For information on weight loss interventions for adults, see “Weight loss strategies that really work”).

None of these trials reported on cost or cost effectiveness. Despite the considerable cost of a comprehensive obesity management program, however, a successful weight-maintenance model could be a worthwhile investment in long-term health.

Lastly, the results of this trial should not negate the importance of obesity prevention efforts by parents, who are in the best position to reverse the childhood obesity epidemic.¹¹

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999; awarded by the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A 10-year-old boy comes in with his mother for a well-child check-up. His BMI is 40 kg/m²—above the 99th percentile for his age and up from 37 a year ago. His blood pressure is 120/84 mm Hg. What treatment, if any, should you offer for his obesity?

Childhood obesity is a global epidemic. In the United States, 19.6% of children ages 6 through 11 and 18.1% of 12- to 19-year-olds are obese, a 3-fold increase in the last 30 years.³ Without intervention, most obese adolescents will become obese adults, threatening to reverse the progress in slowing cardiovascular morbidity and mortality that has occurred over the past few decades.³

Obese kids get adult diseases
Obesity is a risk factor for a variety of chronic conditions, including cardiovascular disease, cerebrovascular disease, and arthritis. Severe obesity is also associated with higher mortality rates.⁴ Unfortunately, these comorbidities are not limited to adulthood.

“Adult” diseases, such as obstructive sleep apnea, dyslipidemia, and type 2 diabetes, are increasingly seen in children and adolescents.¹ Nutritional deficits such as vitamin D and iron deficiency are often seen in obese children, as well.⁵ There are also psychological ramifications of childhood obesity, including social isolation and depression.⁶

The USPSTF recently upgraded its recommendation regarding obesity screening in children ages 6 and older from I (insufficient evidence) to B (a positive grade based on high or moderate certainty of the benefit of the intervention), citing new evidence in favor of screening and treating or referring children when appropriate.² The systematic review we report on here, which formed the basis for the USPSTF’s upgrade, focused on management options for children identified as overweight or obese.

STUDY SUMMARY: Intense, comprehensive efforts pay off

This systematic review¹ included studies of children ages 4 to 18 years who were overweight (defined as a body mass index [BMI] in the 85th to 94th percentile for age and sex) or obese (either a BMI at or above the 95th percentile for age and sex or a BMI >30 kg/m²). The researchers found 25 trials—15 of behavioral interventions alone and 10 that combined behavioral and pharmacologic interventions—that met their criteria: The studies focused on weight loss and/or maintenance, reported outcomes ≥6 months from baseline, and were conducted (or feasible) in a primary care setting.

Behavioral interventions were categorized by treatment intensity (as measured by hours of contact, which ranged from <10 hours to >75) and comprehensiveness (including nutritional counseling, physical activity counseling or participation, and counseling on behavioral management techniques). Weight outcomes were categorized as short-term (6-12 months since treatment initiation) or maintenance (≥12 months after the end of active treatment).

The 15 behavioral intervention trials included 1258 children ages 4 to 18 years, most of whom were obese. Most trials were small and reported high retention rates. All had beneficial effects on weight in the intervention group compared with the controls, but not all changes were statistically significant. Higher intensity and more comprehensive programs had better outcomes.

The largest effects were in 3 moderate- to high-intensity, comprehensive weight management programs with ≥26 hours of contact. These 3 trials demonstrated a difference in BMI of 1.9 to 3.3 in the intervention groups at 12 months compared with the controls. (A 3.3 difference in BMI is equal to approximately 13 lb in an 8-year-old and 17 lb in a 12-year-old.)

Four behavioral intervention studies reported outcomes ≥12 months after completing the intervention (range 15-48 months). Three of the 4 reported continued beneficial effects on weight after the active treatment period, but the effects were markedly attenuated.

The only adverse effect reported in the trials of behavioral interventions was the injury rate among children in an exercise program, but it was minimal: One fracture was reported, vs no injuries for the controls. No differences were reported in height, eating disorders, or depression. However, fewer than half of the behavioral intervention trials reported on adverse effects.

Weight loss drugs have modest effects
Ten trials combining pharmacologic and behavioral interventions involved a total of 1294 obese adolescents ages 12 to 19. All evaluated short-term weight loss effects of either sibutramine (10-15 mg/d) or orlistat (120 mg tid). Trials ranged from 3 to 12 months. Participants in both the control and intervention groups received behavioral counseling.

The trials all favored the treatment groups, although not all of the results were statistically significant. Trials of longer duration (12 months) had more favorable results than those lasting 6 months.

The largest sibutramine trial (n=498) reported a mean BMI reduction of 2.9 in the treatment group, compared with a reduction of 0.3 in the control group (P<.001). This corresponds to an average weight loss of 14 lb in the intervention group, vs 4.2 lb in the control group, after 12 months.

The largest orlistat trial (n=539) reported a mean BMI reduction in the treatment group of 0.6, vs 0.3 in the control group (P<.001)—an average weight loss of 4.2 lb in the intervention group, compared with 2.1 lb among the controls after 12 months. None of the trials evaluated weight change after cessation of the study drug, and none compared orlistat with sibutramine.

Adverse effects in the sibutramine-treated patients were primarily cardiovascular and gastrointestinal. Cardiovascular effects included tachycardia and increases in systolic and diastolic blood pressure. The differences between the intervention and control groups were small, and no differences were observed in discontinuation rates caused by adverse events. Nor were differences reported in growth and maturation between the intervention and control groups.

Adverse effects in the orlistat-treated patients were also low and similar in the intervention and control groups. Gastrointestinal effects were common. The number needed to harm (NNH) for fatty or oily stools was 2,⁴ and the NNH for fecal incontinence was 12.⁵

WHAT'S NEW: Clinicians treating obese kids have cause for optimism

Although the trials included in this review were heterogeneous and many were small, this systematic review provides evidence that intensive, comprehensive behavioral weight loss interventions for obese children can be effective up to 12 months after the conclusion of the program. Family physicians should consider referring obese children and adolescents to such programs—or finding ways to provide supportive strategies themselves.

Sibutramine and orlistat may be helpful in the context of comprehensive, intensive behavioral interventions, although there is no follow-up data to demonstrate long-term safety and weight maintenance after the medication is stopped.

CAVEATS: Little is known about long-term safety of the drugs

There have been few randomized trials of pharmacologic interventions in adolescents and none evaluating weight maintenance after 12 months (or discontinuation of treatment), or assessing long-term safety of the medication.

Sibutramine is not approved by the US Food and Drug Administration (FDA) for use in children or adolescents.⁷ Orlistat is currently approved only for individuals over the age of 12.⁸

In January 2010, an additional contraindication was added to the sibutramine drug label, stating that it is not to be used in patients with a history of cardiovascular disease.⁹ And the FDA is currently investigating a rare association between orlistat and liver injury, although no conclusions have been released.¹⁰ Children and adolescents are particularly vulnerable to long-term side effects, given their relatively young age at the time of drug initiation, so we urge caution with the use of these drugs in this patient population.

CHALLENGES TO IMPLEMENTATION: Intensive approach may be hard to reproduce

Implementation of high-intensity comprehensive interventions for obese children faces a number of roadblocks, including limited availability of programs, cost, and reimbursement. Most of the intensive interventions in these trials took place in specialty centers rather than in primary care offices. Replicating them could require a referral—or significant resources within the primary care setting itself. Yet many, if not most, insurance policies still do not cover such extensive lifestyle interventions. (For information on weight loss interventions for adults, see “Weight loss strategies that really work”).

None of these trials reported on cost or cost effectiveness. Despite the considerable cost of a comprehensive obesity management program, however, a successful weight-maintenance model could be a worthwhile investment in long-term health.

Lastly, the results of this trial should not negate the importance of obesity prevention efforts by parents, who are in the best position to reverse the childhood obesity epidemic.¹¹

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999; awarded by the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A 10-year-old boy comes in with his mother for a well-child check-up. His BMI is 40 kg/m²—above the 99th percentile for his age and up from 37 a year ago. His blood pressure is 120/84 mm Hg. What treatment, if any, should you offer for his obesity?

Childhood obesity is a global epidemic. In the United States, 19.6% of children ages 6 through 11 and 18.1% of 12- to 19-year-olds are obese, a 3-fold increase in the last 30 years.³ Without intervention, most obese adolescents will become obese adults, threatening to reverse the progress in slowing cardiovascular morbidity and mortality that has occurred over the past few decades.³

Obese kids get adult diseases
Obesity is a risk factor for a variety of chronic conditions, including cardiovascular disease, cerebrovascular disease, and arthritis. Severe obesity is also associated with higher mortality rates.⁴ Unfortunately, these comorbidities are not limited to adulthood.

“Adult” diseases, such as obstructive sleep apnea, dyslipidemia, and type 2 diabetes, are increasingly seen in children and adolescents.¹ Nutritional deficits such as vitamin D and iron deficiency are often seen in obese children, as well.⁵ There are also psychological ramifications of childhood obesity, including social isolation and depression.⁶

The USPSTF recently upgraded its recommendation regarding obesity screening in children ages 6 and older from I (insufficient evidence) to B (a positive grade based on high or moderate certainty of the benefit of the intervention), citing new evidence in favor of screening and treating or referring children when appropriate.² The systematic review we report on here, which formed the basis for the USPSTF’s upgrade, focused on management options for children identified as overweight or obese.

STUDY SUMMARY: Intense, comprehensive efforts pay off

This systematic review¹ included studies of children ages 4 to 18 years who were overweight (defined as a body mass index [BMI] in the 85th to 94th percentile for age and sex) or obese (either a BMI at or above the 95th percentile for age and sex or a BMI >30 kg/m²). The researchers found 25 trials—15 of behavioral interventions alone and 10 that combined behavioral and pharmacologic interventions—that met their criteria: The studies focused on weight loss and/or maintenance, reported outcomes ≥6 months from baseline, and were conducted (or feasible) in a primary care setting.

Behavioral interventions were categorized by treatment intensity (as measured by hours of contact, which ranged from <10 hours to >75) and comprehensiveness (including nutritional counseling, physical activity counseling or participation, and counseling on behavioral management techniques). Weight outcomes were categorized as short-term (6-12 months since treatment initiation) or maintenance (≥12 months after the end of active treatment).

The 15 behavioral intervention trials included 1258 children ages 4 to 18 years, most of whom were obese. Most trials were small and reported high retention rates. All had beneficial effects on weight in the intervention group compared with the controls, but not all changes were statistically significant. Higher intensity and more comprehensive programs had better outcomes.

The largest effects were in 3 moderate- to high-intensity, comprehensive weight management programs with ≥26 hours of contact. These 3 trials demonstrated a difference in BMI of 1.9 to 3.3 in the intervention groups at 12 months compared with the controls. (A 3.3 difference in BMI is equal to approximately 13 lb in an 8-year-old and 17 lb in a 12-year-old.)

Four behavioral intervention studies reported outcomes ≥12 months after completing the intervention (range 15-48 months). Three of the 4 reported continued beneficial effects on weight after the active treatment period, but the effects were markedly attenuated.

The only adverse effect reported in the trials of behavioral interventions was the injury rate among children in an exercise program, but it was minimal: One fracture was reported, vs no injuries for the controls. No differences were reported in height, eating disorders, or depression. However, fewer than half of the behavioral intervention trials reported on adverse effects.

Weight loss drugs have modest effects
Ten trials combining pharmacologic and behavioral interventions involved a total of 1294 obese adolescents ages 12 to 19. All evaluated short-term weight loss effects of either sibutramine (10-15 mg/d) or orlistat (120 mg tid). Trials ranged from 3 to 12 months. Participants in both the control and intervention groups received behavioral counseling.

The trials all favored the treatment groups, although not all of the results were statistically significant. Trials of longer duration (12 months) had more favorable results than those lasting 6 months.

The largest sibutramine trial (n=498) reported a mean BMI reduction of 2.9 in the treatment group, compared with a reduction of 0.3 in the control group (P<.001). This corresponds to an average weight loss of 14 lb in the intervention group, vs 4.2 lb in the control group, after 12 months.

The largest orlistat trial (n=539) reported a mean BMI reduction in the treatment group of 0.6, vs 0.3 in the control group (P<.001)—an average weight loss of 4.2 lb in the intervention group, compared with 2.1 lb among the controls after 12 months. None of the trials evaluated weight change after cessation of the study drug, and none compared orlistat with sibutramine.

Adverse effects in the sibutramine-treated patients were primarily cardiovascular and gastrointestinal. Cardiovascular effects included tachycardia and increases in systolic and diastolic blood pressure. The differences between the intervention and control groups were small, and no differences were observed in discontinuation rates caused by adverse events. Nor were differences reported in growth and maturation between the intervention and control groups.

Adverse effects in the orlistat-treated patients were also low and similar in the intervention and control groups. Gastrointestinal effects were common. The number needed to harm (NNH) for fatty or oily stools was 2,⁴ and the NNH for fecal incontinence was 12.⁵

WHAT'S NEW: Clinicians treating obese kids have cause for optimism

Although the trials included in this review were heterogeneous and many were small, this systematic review provides evidence that intensive, comprehensive behavioral weight loss interventions for obese children can be effective up to 12 months after the conclusion of the program. Family physicians should consider referring obese children and adolescents to such programs—or finding ways to provide supportive strategies themselves.

Sibutramine and orlistat may be helpful in the context of comprehensive, intensive behavioral interventions, although there is no follow-up data to demonstrate long-term safety and weight maintenance after the medication is stopped.

CAVEATS: Little is known about long-term safety of the drugs

There have been few randomized trials of pharmacologic interventions in adolescents and none evaluating weight maintenance after 12 months (or discontinuation of treatment), or assessing long-term safety of the medication.

Sibutramine is not approved by the US Food and Drug Administration (FDA) for use in children or adolescents.⁷ Orlistat is currently approved only for individuals over the age of 12.⁸

In January 2010, an additional contraindication was added to the sibutramine drug label, stating that it is not to be used in patients with a history of cardiovascular disease.⁹ And the FDA is currently investigating a rare association between orlistat and liver injury, although no conclusions have been released.¹⁰ Children and adolescents are particularly vulnerable to long-term side effects, given their relatively young age at the time of drug initiation, so we urge caution with the use of these drugs in this patient population.

CHALLENGES TO IMPLEMENTATION: Intensive approach may be hard to reproduce

Implementation of high-intensity comprehensive interventions for obese children faces a number of roadblocks, including limited availability of programs, cost, and reimbursement. Most of the intensive interventions in these trials took place in specialty centers rather than in primary care offices. Replicating them could require a referral—or significant resources within the primary care setting itself. Yet many, if not most, insurance policies still do not cover such extensive lifestyle interventions. (For information on weight loss interventions for adults, see “Weight loss strategies that really work”).

None of these trials reported on cost or cost effectiveness. Despite the considerable cost of a comprehensive obesity management program, however, a successful weight-maintenance model could be a worthwhile investment in long-term health.

Lastly, the results of this trial should not negate the importance of obesity prevention efforts by parents, who are in the best position to reverse the childhood obesity epidemic.¹¹

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999; awarded by the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A mother brings her 6-year-old son to the emergency department (ED) for treatment of forearm pain after a bicycle accident. clinical examination reveals a swollen and tender wrist. A radiograph confirms a diagnosis of a nondisplaced distal radial fracture. After proper stabilization, the little boy is discharged home, with a visit to his primary care physician scheduled for the following week. if he were your patient, what would you prescribe for outpatient analgesia?

Musculoskeletal trauma is a common pediatric presentation, in both emergency and office settings. In fact, it is estimated that by age 15, one-half to two-thirds of children will have fractured a bone.⁴ Physicians commonly prescribe nonsteroidal anti-inflammatory drugs (NSAIDs) and opioids—especially acetaminophen with codeine—as analgesia for children with fractures,⁵ but few studies have directly compared these medications in pediatric patients.

No consensus on analgesia for musculoskeletal pain in kids
Pain associated with an acute fracture is substantial, and most children who incur fractures are managed at home, and thus require effective and well-tolerated oral analgesia. However, prescribing practices vary widely, and there is no consensus regarding the first-line medication for kids with fracture.

A Cochrane review of adult postoperative pain concluded that NSAIDs are effective, and they are commonly prescribed to adult patients for various types of pain.⁶ Fewer studies of pain control in children exist. Before the 2009 study reported on here, there were just 2 RCTs that addressed pediatric musculoskeletal pain in patients presenting to the ED.

In single-dose studies, ibuprofen comes out ahead
The smaller of the 2 trials (N=66) compared ibuprofen alone vs ibuprofen plus oxycodone for suspected orthopedic injury. The researchers found that pain relief was equivalent, but the oxycodone group had more adverse effects.² The larger trial (N=336) compared ibuprofen, acetaminophen, and codeine for acute pediatric musculoskeletal injuries. An hour after receiving their study drug, children in the ibuprofen group had significantly greater reduction in pain than those in either the acetaminophen group or the codeine group. They were also more likely to report adequate analgesia.³ Neither study followed patients after discharge from the ED.

STUDY SUMMARY: New RCT evaluates pain relief once patients go home

The Drendel study was a randomized, controlled, double-blind trial of outpatient analgesia for pediatric fractures.¹ The investigators randomized 336 children ages 4 to 18 years with radiographically confirmed arm fractures to a suspension of either ibuprofen (10 mg/kg) or acetaminophen with codeine (1 mg/kg codeine component per dose), which are recommended dosages. They enrolled a convenience sample of children with nondisplaced fractures that did not require reduction in the ED.

Children were excluded if they weighed more than 60 kg, preferred tablets to liquid medication, sought care more than 12 hours after injury, or had developmental delays or contraindications to any study medication. Also excluded were children—or their parents—who did not speak English and those who were inaccessible by telephone for follow-up.

Study groups had similar baseline demographic and fracture characteristics, and similar pain scores. Patients and their parents were blinded to the assigned drug; all received the same discharge instructions and 2 doses of a rescue medication (the alternate study drug). The primary outcome was use of rescue medication due to failure of the assigned study drug. Secondary outcomes included decrease in pain score, functional outcomes (play, school, eating, sleeping), and satisfaction with the medication.

During the 72 hours after discharge from the ED, patients and parents filled out a standard diary recording pain and medication use. The diaries were returned by mail. Follow-up was good, with about 75% of diaries returned.

Ibuprofen users had fewer problems
Analysis of 244 diaries revealed that less rescue medication was used in the ibuprofen group, although the difference was not statistically significant (20.3% vs 31% [absolute risk reduction, 10.7%], 95% confidence interval, -0.2% to 21.6%). Decrease in mean pain score was the same in both groups. Functional status the day after the injury was better in the ibuprofen group compared with the acetaminophen/codeine group. In addition, 50.9% of patients in the acetaminophen/codeine group reported adverse events, vs 29.5% of those in the ibuprofen group (number needed to harm=4.7).

At the study’s end, children were more satisfied with ibuprofen. Only 10% of patients who took ibuprofen said they would not use it for future fractures; in comparison, 27.5% of patients in the acetaminophen/codeine group said they would not choose to use codeine again.

The authors followed participants for 1 to 4 years through orthopedic clinic records and telephone calls for any long-term adverse orthopedic outcomes. Four cases of refracture at the same site occurred (1.6%), 3 of which were in the acetaminophen/codeine group. There were no cases of nonunion.

WHAT'S NEW: Ibuprofen emerges as first-line agent for kids

Both ibuprofen and acetaminophen with codeine are commonly prescribed for outpatient pediatric analgesia, but this is the first study to compare them head-to-head for outpatient management of postfracture pain. Ibuprofen worked at least as well as acetaminophen with codeine for fracture pain control, and had fewer adverse effects. Children given ibuprofen were better able to eat and play than those given acetaminophen with codeine—an important patient-oriented functional outcome. Patients and their parents were also more satisfied when ED physicians prescribed ibuprofen. This study is consistent with short-term (single-dose) studies and confirms that ibuprofen should be the first-line agent for outpatient analgesia in this group.

CAVEATS: Study did not address NSAIDs’ effect on bone healing

In theory, ibuprofen—like other NSAIDs—can diminish the proinflammatory milieu required for bone turnover and fracture healing. Chart reviews of up to 4 years after the incident fracture found no evidence that ibuprofen delayed healing or increased rates of refracture. However, this study was neither designed nor powered to examine this outcome. Previous studies have found no conclusive evidence that short-term use of NSAIDs impairs fracture healing.^7,8

Results apply only to simple fractures. Patients in this study did not require manipulation or reduction of their fracture, limiting the scope of the authors’ recommendation to simple arm fractures. More severe injury may require narcotic analgesia. One can assume, based on this and other supporting literature, that the findings extrapolate to other similarly painful pediatric musculoskeletal injuries.²

Twenty-five percent of subjects were lost to follow-up. Follow-up diaries were available from about 75% of the participants. It is possible that a clearer beneficial outcome would have been found with 1 of the analgesics studied if the response rate had been higher. Because this study is consistent with the previous ED-only studies comparing ibuprofen with acetaminophen plus codeine, however, it is unlikely that a higher response rate would find ibuprofen inferior to acetaminophen plus codeine.

CHALLENGES TO IMPLEMENTATION: Parents—or patients—may expect an Rx

Prescribing an effective, common, inexpensive, and well-tolerated oral medication should have no barriers to implementation. Still, use of an over-the-counter medication, however effective, may face resistance from patients or parents who expect “something more” for fracture pain.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A mother brings her 6-year-old son to the emergency department (ED) for treatment of forearm pain after a bicycle accident. clinical examination reveals a swollen and tender wrist. A radiograph confirms a diagnosis of a nondisplaced distal radial fracture. After proper stabilization, the little boy is discharged home, with a visit to his primary care physician scheduled for the following week. if he were your patient, what would you prescribe for outpatient analgesia?

Musculoskeletal trauma is a common pediatric presentation, in both emergency and office settings. In fact, it is estimated that by age 15, one-half to two-thirds of children will have fractured a bone.⁴ Physicians commonly prescribe nonsteroidal anti-inflammatory drugs (NSAIDs) and opioids—especially acetaminophen with codeine—as analgesia for children with fractures,⁵ but few studies have directly compared these medications in pediatric patients.

No consensus on analgesia for musculoskeletal pain in kids
Pain associated with an acute fracture is substantial, and most children who incur fractures are managed at home, and thus require effective and well-tolerated oral analgesia. However, prescribing practices vary widely, and there is no consensus regarding the first-line medication for kids with fracture.

A Cochrane review of adult postoperative pain concluded that NSAIDs are effective, and they are commonly prescribed to adult patients for various types of pain.⁶ Fewer studies of pain control in children exist. Before the 2009 study reported on here, there were just 2 RCTs that addressed pediatric musculoskeletal pain in patients presenting to the ED.

In single-dose studies, ibuprofen comes out ahead
The smaller of the 2 trials (N=66) compared ibuprofen alone vs ibuprofen plus oxycodone for suspected orthopedic injury. The researchers found that pain relief was equivalent, but the oxycodone group had more adverse effects.² The larger trial (N=336) compared ibuprofen, acetaminophen, and codeine for acute pediatric musculoskeletal injuries. An hour after receiving their study drug, children in the ibuprofen group had significantly greater reduction in pain than those in either the acetaminophen group or the codeine group. They were also more likely to report adequate analgesia.³ Neither study followed patients after discharge from the ED.

STUDY SUMMARY: New RCT evaluates pain relief once patients go home

The Drendel study was a randomized, controlled, double-blind trial of outpatient analgesia for pediatric fractures.¹ The investigators randomized 336 children ages 4 to 18 years with radiographically confirmed arm fractures to a suspension of either ibuprofen (10 mg/kg) or acetaminophen with codeine (1 mg/kg codeine component per dose), which are recommended dosages. They enrolled a convenience sample of children with nondisplaced fractures that did not require reduction in the ED.

Children were excluded if they weighed more than 60 kg, preferred tablets to liquid medication, sought care more than 12 hours after injury, or had developmental delays or contraindications to any study medication. Also excluded were children—or their parents—who did not speak English and those who were inaccessible by telephone for follow-up.

Study groups had similar baseline demographic and fracture characteristics, and similar pain scores. Patients and their parents were blinded to the assigned drug; all received the same discharge instructions and 2 doses of a rescue medication (the alternate study drug). The primary outcome was use of rescue medication due to failure of the assigned study drug. Secondary outcomes included decrease in pain score, functional outcomes (play, school, eating, sleeping), and satisfaction with the medication.

During the 72 hours after discharge from the ED, patients and parents filled out a standard diary recording pain and medication use. The diaries were returned by mail. Follow-up was good, with about 75% of diaries returned.

Ibuprofen users had fewer problems
Analysis of 244 diaries revealed that less rescue medication was used in the ibuprofen group, although the difference was not statistically significant (20.3% vs 31% [absolute risk reduction, 10.7%], 95% confidence interval, -0.2% to 21.6%). Decrease in mean pain score was the same in both groups. Functional status the day after the injury was better in the ibuprofen group compared with the acetaminophen/codeine group. In addition, 50.9% of patients in the acetaminophen/codeine group reported adverse events, vs 29.5% of those in the ibuprofen group (number needed to harm=4.7).

At the study’s end, children were more satisfied with ibuprofen. Only 10% of patients who took ibuprofen said they would not use it for future fractures; in comparison, 27.5% of patients in the acetaminophen/codeine group said they would not choose to use codeine again.

The authors followed participants for 1 to 4 years through orthopedic clinic records and telephone calls for any long-term adverse orthopedic outcomes. Four cases of refracture at the same site occurred (1.6%), 3 of which were in the acetaminophen/codeine group. There were no cases of nonunion.

WHAT'S NEW: Ibuprofen emerges as first-line agent for kids

Both ibuprofen and acetaminophen with codeine are commonly prescribed for outpatient pediatric analgesia, but this is the first study to compare them head-to-head for outpatient management of postfracture pain. Ibuprofen worked at least as well as acetaminophen with codeine for fracture pain control, and had fewer adverse effects. Children given ibuprofen were better able to eat and play than those given acetaminophen with codeine—an important patient-oriented functional outcome. Patients and their parents were also more satisfied when ED physicians prescribed ibuprofen. This study is consistent with short-term (single-dose) studies and confirms that ibuprofen should be the first-line agent for outpatient analgesia in this group.

CAVEATS: Study did not address NSAIDs’ effect on bone healing

In theory, ibuprofen—like other NSAIDs—can diminish the proinflammatory milieu required for bone turnover and fracture healing. Chart reviews of up to 4 years after the incident fracture found no evidence that ibuprofen delayed healing or increased rates of refracture. However, this study was neither designed nor powered to examine this outcome. Previous studies have found no conclusive evidence that short-term use of NSAIDs impairs fracture healing.^7,8

Results apply only to simple fractures. Patients in this study did not require manipulation or reduction of their fracture, limiting the scope of the authors’ recommendation to simple arm fractures. More severe injury may require narcotic analgesia. One can assume, based on this and other supporting literature, that the findings extrapolate to other similarly painful pediatric musculoskeletal injuries.²

Twenty-five percent of subjects were lost to follow-up. Follow-up diaries were available from about 75% of the participants. It is possible that a clearer beneficial outcome would have been found with 1 of the analgesics studied if the response rate had been higher. Because this study is consistent with the previous ED-only studies comparing ibuprofen with acetaminophen plus codeine, however, it is unlikely that a higher response rate would find ibuprofen inferior to acetaminophen plus codeine.

CHALLENGES TO IMPLEMENTATION: Parents—or patients—may expect an Rx

Prescribing an effective, common, inexpensive, and well-tolerated oral medication should have no barriers to implementation. Still, use of an over-the-counter medication, however effective, may face resistance from patients or parents who expect “something more” for fracture pain.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A mother brings her 6-year-old son to the emergency department (ED) for treatment of forearm pain after a bicycle accident. clinical examination reveals a swollen and tender wrist. A radiograph confirms a diagnosis of a nondisplaced distal radial fracture. After proper stabilization, the little boy is discharged home, with a visit to his primary care physician scheduled for the following week. if he were your patient, what would you prescribe for outpatient analgesia?

Musculoskeletal trauma is a common pediatric presentation, in both emergency and office settings. In fact, it is estimated that by age 15, one-half to two-thirds of children will have fractured a bone.⁴ Physicians commonly prescribe nonsteroidal anti-inflammatory drugs (NSAIDs) and opioids—especially acetaminophen with codeine—as analgesia for children with fractures,⁵ but few studies have directly compared these medications in pediatric patients.

No consensus on analgesia for musculoskeletal pain in kids
Pain associated with an acute fracture is substantial, and most children who incur fractures are managed at home, and thus require effective and well-tolerated oral analgesia. However, prescribing practices vary widely, and there is no consensus regarding the first-line medication for kids with fracture.

A Cochrane review of adult postoperative pain concluded that NSAIDs are effective, and they are commonly prescribed to adult patients for various types of pain.⁶ Fewer studies of pain control in children exist. Before the 2009 study reported on here, there were just 2 RCTs that addressed pediatric musculoskeletal pain in patients presenting to the ED.

In single-dose studies, ibuprofen comes out ahead
The smaller of the 2 trials (N=66) compared ibuprofen alone vs ibuprofen plus oxycodone for suspected orthopedic injury. The researchers found that pain relief was equivalent, but the oxycodone group had more adverse effects.² The larger trial (N=336) compared ibuprofen, acetaminophen, and codeine for acute pediatric musculoskeletal injuries. An hour after receiving their study drug, children in the ibuprofen group had significantly greater reduction in pain than those in either the acetaminophen group or the codeine group. They were also more likely to report adequate analgesia.³ Neither study followed patients after discharge from the ED.

STUDY SUMMARY: New RCT evaluates pain relief once patients go home

The Drendel study was a randomized, controlled, double-blind trial of outpatient analgesia for pediatric fractures.¹ The investigators randomized 336 children ages 4 to 18 years with radiographically confirmed arm fractures to a suspension of either ibuprofen (10 mg/kg) or acetaminophen with codeine (1 mg/kg codeine component per dose), which are recommended dosages. They enrolled a convenience sample of children with nondisplaced fractures that did not require reduction in the ED.

Children were excluded if they weighed more than 60 kg, preferred tablets to liquid medication, sought care more than 12 hours after injury, or had developmental delays or contraindications to any study medication. Also excluded were children—or their parents—who did not speak English and those who were inaccessible by telephone for follow-up.

Study groups had similar baseline demographic and fracture characteristics, and similar pain scores. Patients and their parents were blinded to the assigned drug; all received the same discharge instructions and 2 doses of a rescue medication (the alternate study drug). The primary outcome was use of rescue medication due to failure of the assigned study drug. Secondary outcomes included decrease in pain score, functional outcomes (play, school, eating, sleeping), and satisfaction with the medication.

During the 72 hours after discharge from the ED, patients and parents filled out a standard diary recording pain and medication use. The diaries were returned by mail. Follow-up was good, with about 75% of diaries returned.

Ibuprofen users had fewer problems
Analysis of 244 diaries revealed that less rescue medication was used in the ibuprofen group, although the difference was not statistically significant (20.3% vs 31% [absolute risk reduction, 10.7%], 95% confidence interval, -0.2% to 21.6%). Decrease in mean pain score was the same in both groups. Functional status the day after the injury was better in the ibuprofen group compared with the acetaminophen/codeine group. In addition, 50.9% of patients in the acetaminophen/codeine group reported adverse events, vs 29.5% of those in the ibuprofen group (number needed to harm=4.7).

At the study’s end, children were more satisfied with ibuprofen. Only 10% of patients who took ibuprofen said they would not use it for future fractures; in comparison, 27.5% of patients in the acetaminophen/codeine group said they would not choose to use codeine again.

The authors followed participants for 1 to 4 years through orthopedic clinic records and telephone calls for any long-term adverse orthopedic outcomes. Four cases of refracture at the same site occurred (1.6%), 3 of which were in the acetaminophen/codeine group. There were no cases of nonunion.

WHAT'S NEW: Ibuprofen emerges as first-line agent for kids

Both ibuprofen and acetaminophen with codeine are commonly prescribed for outpatient pediatric analgesia, but this is the first study to compare them head-to-head for outpatient management of postfracture pain. Ibuprofen worked at least as well as acetaminophen with codeine for fracture pain control, and had fewer adverse effects. Children given ibuprofen were better able to eat and play than those given acetaminophen with codeine—an important patient-oriented functional outcome. Patients and their parents were also more satisfied when ED physicians prescribed ibuprofen. This study is consistent with short-term (single-dose) studies and confirms that ibuprofen should be the first-line agent for outpatient analgesia in this group.

CAVEATS: Study did not address NSAIDs’ effect on bone healing

In theory, ibuprofen—like other NSAIDs—can diminish the proinflammatory milieu required for bone turnover and fracture healing. Chart reviews of up to 4 years after the incident fracture found no evidence that ibuprofen delayed healing or increased rates of refracture. However, this study was neither designed nor powered to examine this outcome. Previous studies have found no conclusive evidence that short-term use of NSAIDs impairs fracture healing.^7,8

Results apply only to simple fractures. Patients in this study did not require manipulation or reduction of their fracture, limiting the scope of the authors’ recommendation to simple arm fractures. More severe injury may require narcotic analgesia. One can assume, based on this and other supporting literature, that the findings extrapolate to other similarly painful pediatric musculoskeletal injuries.²

Twenty-five percent of subjects were lost to follow-up. Follow-up diaries were available from about 75% of the participants. It is possible that a clearer beneficial outcome would have been found with 1 of the analgesics studied if the response rate had been higher. Because this study is consistent with the previous ED-only studies comparing ibuprofen with acetaminophen plus codeine, however, it is unlikely that a higher response rate would find ibuprofen inferior to acetaminophen plus codeine.

CHALLENGES TO IMPLEMENTATION: Parents—or patients—may expect an Rx

Prescribing an effective, common, inexpensive, and well-tolerated oral medication should have no barriers to implementation. Still, use of an over-the-counter medication, however effective, may face resistance from patients or parents who expect “something more” for fracture pain.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant is a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

James M, a 43-year-old self-employed mechanic, came to see you 2 weeks ago because of neck pain radiating to his right shoulder, arm, forearm, and dorsum of his hand. You diagnosed acute right-sided cervical radiculopathy and prescribed a nonsteroidal anti-inflammatory drug.

Today he’s back in your office, reporting that he has experienced only minimal transient relief. You reassure him that the pain will subside within a few months, but james wants to know if you can give him something to speed up his recovery and enable him to return to work.

Each year in the United States, approximately 85 out of every 100,000 adults develop cervical radiculopathy²—a neurologic condition characterized by dysfunction of a cervical spinal nerve, the roots of the nerve, or both. In addition to pain in the neck and the arm on the affected side, patients often develop sensory loss, loss of motor function, and/or reflex changes in the affected nerve-root distribution.

Most patients respond to conservative measures
A nonsurgical approach is the preferred first-line treatment strategy for cervical radiculopathy.³ The Bone and Joint Decade 2000-2010 Task Force on Neck Pain and Its Associated Disorders—an international network of experts in a number of specialties—found no evidence that surgery provides better long-term outcomes than more conservative treatment.³ Approximately 80% to 90% of patients respond to a conservative approach, with improvements in pain, function, and mood in 3 to 6 months.^4,5

There are numerous conservative therapies for cervical radiculopathy, including oral analgesics, rest, cervical traction, short-term immobilization with a cervical collar, PT, a short course of oral corticosteroids, and perineural steroid injections.^4-6 These therapies may be used singly or in combination. Until now, however, no high-quality RCTs compared the efficacy of various nonsurgical treatment modalities for acute cervical radiculopathy—and their effectiveness is still subject to debate.

STUDY SUMMARY: Initially, both Tx modes beat wait-and-see

The study by Kuijper et al¹ is the first RCT to compare the effectiveness of PT, cervical collars, and a wait-and-see strategy in alleviating symptoms of cervical radiculopathy. Enrollees (N=205) were men and women ages 18 to 75 years who were referred by general practitioners in 3 Dutch hospitals. All the participants had a diagnosis of cervical radiculopathy confirmed by a neurologist. In addition, all the cases were of recent onset, with symptoms of <1 month’s duration at the time of enrollment. Patients with clinical signs of cord compression and those who had previously been treated with either PT or a cervical collar for this episode were excluded.

The researchers randomized the participants into 3 groups: PT, cervical collar, or control. All the groups were comparable at baseline.

Those in the PT group received twice weekly therapy for 6 weeks, with a focus on mobilizing and stabilizing the cervical spine. They were also taught to perform home exercises and advised to do the exercises daily.

Patients in the cervical collar group were given a semi-hard, snugly fitted collar and instructed to wear it during the day for 3 weeks—and to rest as much as possible. They were weaned from the collar over the course of another 3 weeks.

Participants in the control group were simply told to follow their normal daily routine as much as possible. All 3 groups were permitted to take oral pain medication as needed.

The primary outcome measures were changes over time in neck and arm pain scores, using 2 validated measurement tools: a 100-mm visual analog scale (VAS) and a 100-point neck disability index (NDI). Both tools were used at 3 weeks, 6 weeks, and 6 months. Secondary outcomes were treatment satisfaction (as measured on a 5-point scale), use of opiates, and working status.

By 6 months, differences virtually disappeared
Both the active and passive interventions reduced arm and neck pain faster than the wait-and-see strategy. At 6 weeks, participants in both the PT and cervical collar groups reported a 31-mm reduction in arm pain (P=.007 and .006, respectively), compared with a 19-mm reduction for those in the control group (P=.006). This is a clinically meaningful difference.

The rate of reduction in neck pain over the first 6 weeks was: PT group, 2.4 mm/week, P=.002; cervical collar group, 2.8 mm/week, P=<.001; and control group, 0.9 mm/week. The rate of reduction in the NDI was 1.4 points per week for the control group vs 2.3 points per week for the cervical collar group (P= .024). The PT group fared no better on the NDI measure than the control group. This may reflect the fact that the index predominantly measured disability caused by neck pain, whereas arm pain scores,—which were highest initially—showed the greatest improvement, the authors note.

At 6 months, pain and disability had almost resolved for all the patients, regardless of their treatment group, and secondary outcomes—treatment satisfaction, analgesic use, and working status—were similar for all 3 groups.

WHAT'S NEW: High-quality RCT supports PT and cervical collar

Some investigators have advocated the short-term use of immobilization with either a cervical collar or a cervical pillow during sleep. Until now, however, there was no conclusive evidence about the benefits of this approach.

One earlier RCT (N=493) compared 5 treatment modalities—traction, positioning, collar, placebo tablets, and heat treatment—and found no significant difference in pain and ability to work.⁷ That trial was done nearly 15 years ago, however, and the investigators did not use validated outcome scales. Therefore, the trial would not meet current RCT standards.

The study we report on here leaves little doubt that the 2 treatments reviewed—PT and cervical collar—provide more rapid relief than a wait-and-see approach.

CAVEATS: Pain meds still needed, unanswered questions remain

Although the cervical collar and PT groups had less pain at 3 and 6 weeks compared with the controls—and all 3 groups showed equal improvement at study’s end—the researchers found little difference in use of analgesics. Data on adherence to treatment was recorded by patients, so treatment adherence may not be completely accurate.⁸

Patients without severe arm pain or signs of muscle weakness were not included in this study, so we don’t know whether individuals with less severe cervical radiculopathy would benefit from these treatments. What’s more, this study focused only on new cases of acute cervical radiculopathy, and the findings may not apply to patients with chronic, recurrent, or persistent symptoms.

The apparent contradiction in the finding that both immobilization and PT are beneficial does not have a clear scientific explanation. The researchers hypothesize that immobilizing the neck with a collar reduces foraminal root compression and inflammation; this could explain the larger reduction in arm pain compared with neck pain and neck disability found in this study. The mechanism of pain reduction with PT is unclear, although it is probably related to the restoration of the neck musculature’s strength and range of motion.

Cost is another issue. A cervical collar and rest is at least as effective as PT for recent onset cervical radiculopathy, but the collar costs only about $20—far less than the cost of 12 sessions of therapy.

One final caveat: Any patient with persistent or worsening symptoms should undergo additional evaluation, including imaging.

CHALLENGES TO IMPLEMENTATION: Rest is contrary to usual approach

Some physicians may not agree with the recommendation to encourage rest. Indeed, rest and immobilization are contrary to the usual recommendation for musculoskeletal injuries—to resume activity as soon as possible.

Patients might not like wearing a collar for a variety of personal reasons, such as cosmetic appearance or limitations of motion. On the other hand, some patients may feel that their pain is too severe to be able to participate in PT—which may also be too expensive for, or not readily available to, some patients.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant number UL1RR024999 from the National Center for Research Resources; the grant was a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

James M, a 43-year-old self-employed mechanic, came to see you 2 weeks ago because of neck pain radiating to his right shoulder, arm, forearm, and dorsum of his hand. You diagnosed acute right-sided cervical radiculopathy and prescribed a nonsteroidal anti-inflammatory drug.

Today he’s back in your office, reporting that he has experienced only minimal transient relief. You reassure him that the pain will subside within a few months, but james wants to know if you can give him something to speed up his recovery and enable him to return to work.

Each year in the United States, approximately 85 out of every 100,000 adults develop cervical radiculopathy²—a neurologic condition characterized by dysfunction of a cervical spinal nerve, the roots of the nerve, or both. In addition to pain in the neck and the arm on the affected side, patients often develop sensory loss, loss of motor function, and/or reflex changes in the affected nerve-root distribution.

Most patients respond to conservative measures
A nonsurgical approach is the preferred first-line treatment strategy for cervical radiculopathy.³ The Bone and Joint Decade 2000-2010 Task Force on Neck Pain and Its Associated Disorders—an international network of experts in a number of specialties—found no evidence that surgery provides better long-term outcomes than more conservative treatment.³ Approximately 80% to 90% of patients respond to a conservative approach, with improvements in pain, function, and mood in 3 to 6 months.^4,5

There are numerous conservative therapies for cervical radiculopathy, including oral analgesics, rest, cervical traction, short-term immobilization with a cervical collar, PT, a short course of oral corticosteroids, and perineural steroid injections.^4-6 These therapies may be used singly or in combination. Until now, however, no high-quality RCTs compared the efficacy of various nonsurgical treatment modalities for acute cervical radiculopathy—and their effectiveness is still subject to debate.

STUDY SUMMARY: Initially, both Tx modes beat wait-and-see

The study by Kuijper et al¹ is the first RCT to compare the effectiveness of PT, cervical collars, and a wait-and-see strategy in alleviating symptoms of cervical radiculopathy. Enrollees (N=205) were men and women ages 18 to 75 years who were referred by general practitioners in 3 Dutch hospitals. All the participants had a diagnosis of cervical radiculopathy confirmed by a neurologist. In addition, all the cases were of recent onset, with symptoms of <1 month’s duration at the time of enrollment. Patients with clinical signs of cord compression and those who had previously been treated with either PT or a cervical collar for this episode were excluded.

The researchers randomized the participants into 3 groups: PT, cervical collar, or control. All the groups were comparable at baseline.

Those in the PT group received twice weekly therapy for 6 weeks, with a focus on mobilizing and stabilizing the cervical spine. They were also taught to perform home exercises and advised to do the exercises daily.

Patients in the cervical collar group were given a semi-hard, snugly fitted collar and instructed to wear it during the day for 3 weeks—and to rest as much as possible. They were weaned from the collar over the course of another 3 weeks.

Participants in the control group were simply told to follow their normal daily routine as much as possible. All 3 groups were permitted to take oral pain medication as needed.

The primary outcome measures were changes over time in neck and arm pain scores, using 2 validated measurement tools: a 100-mm visual analog scale (VAS) and a 100-point neck disability index (NDI). Both tools were used at 3 weeks, 6 weeks, and 6 months. Secondary outcomes were treatment satisfaction (as measured on a 5-point scale), use of opiates, and working status.

By 6 months, differences virtually disappeared
Both the active and passive interventions reduced arm and neck pain faster than the wait-and-see strategy. At 6 weeks, participants in both the PT and cervical collar groups reported a 31-mm reduction in arm pain (P=.007 and .006, respectively), compared with a 19-mm reduction for those in the control group (P=.006). This is a clinically meaningful difference.

The rate of reduction in neck pain over the first 6 weeks was: PT group, 2.4 mm/week, P=.002; cervical collar group, 2.8 mm/week, P=<.001; and control group, 0.9 mm/week. The rate of reduction in the NDI was 1.4 points per week for the control group vs 2.3 points per week for the cervical collar group (P= .024). The PT group fared no better on the NDI measure than the control group. This may reflect the fact that the index predominantly measured disability caused by neck pain, whereas arm pain scores,—which were highest initially—showed the greatest improvement, the authors note.

At 6 months, pain and disability had almost resolved for all the patients, regardless of their treatment group, and secondary outcomes—treatment satisfaction, analgesic use, and working status—were similar for all 3 groups.

WHAT'S NEW: High-quality RCT supports PT and cervical collar

Some investigators have advocated the short-term use of immobilization with either a cervical collar or a cervical pillow during sleep. Until now, however, there was no conclusive evidence about the benefits of this approach.

One earlier RCT (N=493) compared 5 treatment modalities—traction, positioning, collar, placebo tablets, and heat treatment—and found no significant difference in pain and ability to work.⁷ That trial was done nearly 15 years ago, however, and the investigators did not use validated outcome scales. Therefore, the trial would not meet current RCT standards.

The study we report on here leaves little doubt that the 2 treatments reviewed—PT and cervical collar—provide more rapid relief than a wait-and-see approach.

CAVEATS: Pain meds still needed, unanswered questions remain

Although the cervical collar and PT groups had less pain at 3 and 6 weeks compared with the controls—and all 3 groups showed equal improvement at study’s end—the researchers found little difference in use of analgesics. Data on adherence to treatment was recorded by patients, so treatment adherence may not be completely accurate.⁸

Patients without severe arm pain or signs of muscle weakness were not included in this study, so we don’t know whether individuals with less severe cervical radiculopathy would benefit from these treatments. What’s more, this study focused only on new cases of acute cervical radiculopathy, and the findings may not apply to patients with chronic, recurrent, or persistent symptoms.

The apparent contradiction in the finding that both immobilization and PT are beneficial does not have a clear scientific explanation. The researchers hypothesize that immobilizing the neck with a collar reduces foraminal root compression and inflammation; this could explain the larger reduction in arm pain compared with neck pain and neck disability found in this study. The mechanism of pain reduction with PT is unclear, although it is probably related to the restoration of the neck musculature’s strength and range of motion.

Cost is another issue. A cervical collar and rest is at least as effective as PT for recent onset cervical radiculopathy, but the collar costs only about $20—far less than the cost of 12 sessions of therapy.

One final caveat: Any patient with persistent or worsening symptoms should undergo additional evaluation, including imaging.

CHALLENGES TO IMPLEMENTATION: Rest is contrary to usual approach

Some physicians may not agree with the recommendation to encourage rest. Indeed, rest and immobilization are contrary to the usual recommendation for musculoskeletal injuries—to resume activity as soon as possible.

Patients might not like wearing a collar for a variety of personal reasons, such as cosmetic appearance or limitations of motion. On the other hand, some patients may feel that their pain is too severe to be able to participate in PT—which may also be too expensive for, or not readily available to, some patients.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant number UL1RR024999 from the National Center for Research Resources; the grant was a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

James M, a 43-year-old self-employed mechanic, came to see you 2 weeks ago because of neck pain radiating to his right shoulder, arm, forearm, and dorsum of his hand. You diagnosed acute right-sided cervical radiculopathy and prescribed a nonsteroidal anti-inflammatory drug.

Today he’s back in your office, reporting that he has experienced only minimal transient relief. You reassure him that the pain will subside within a few months, but james wants to know if you can give him something to speed up his recovery and enable him to return to work.

Each year in the United States, approximately 85 out of every 100,000 adults develop cervical radiculopathy²—a neurologic condition characterized by dysfunction of a cervical spinal nerve, the roots of the nerve, or both. In addition to pain in the neck and the arm on the affected side, patients often develop sensory loss, loss of motor function, and/or reflex changes in the affected nerve-root distribution.

Most patients respond to conservative measures
A nonsurgical approach is the preferred first-line treatment strategy for cervical radiculopathy.³ The Bone and Joint Decade 2000-2010 Task Force on Neck Pain and Its Associated Disorders—an international network of experts in a number of specialties—found no evidence that surgery provides better long-term outcomes than more conservative treatment.³ Approximately 80% to 90% of patients respond to a conservative approach, with improvements in pain, function, and mood in 3 to 6 months.^4,5

There are numerous conservative therapies for cervical radiculopathy, including oral analgesics, rest, cervical traction, short-term immobilization with a cervical collar, PT, a short course of oral corticosteroids, and perineural steroid injections.^4-6 These therapies may be used singly or in combination. Until now, however, no high-quality RCTs compared the efficacy of various nonsurgical treatment modalities for acute cervical radiculopathy—and their effectiveness is still subject to debate.

STUDY SUMMARY: Initially, both Tx modes beat wait-and-see

The study by Kuijper et al¹ is the first RCT to compare the effectiveness of PT, cervical collars, and a wait-and-see strategy in alleviating symptoms of cervical radiculopathy. Enrollees (N=205) were men and women ages 18 to 75 years who were referred by general practitioners in 3 Dutch hospitals. All the participants had a diagnosis of cervical radiculopathy confirmed by a neurologist. In addition, all the cases were of recent onset, with symptoms of <1 month’s duration at the time of enrollment. Patients with clinical signs of cord compression and those who had previously been treated with either PT or a cervical collar for this episode were excluded.

The researchers randomized the participants into 3 groups: PT, cervical collar, or control. All the groups were comparable at baseline.

Those in the PT group received twice weekly therapy for 6 weeks, with a focus on mobilizing and stabilizing the cervical spine. They were also taught to perform home exercises and advised to do the exercises daily.

Patients in the cervical collar group were given a semi-hard, snugly fitted collar and instructed to wear it during the day for 3 weeks—and to rest as much as possible. They were weaned from the collar over the course of another 3 weeks.

Participants in the control group were simply told to follow their normal daily routine as much as possible. All 3 groups were permitted to take oral pain medication as needed.

The primary outcome measures were changes over time in neck and arm pain scores, using 2 validated measurement tools: a 100-mm visual analog scale (VAS) and a 100-point neck disability index (NDI). Both tools were used at 3 weeks, 6 weeks, and 6 months. Secondary outcomes were treatment satisfaction (as measured on a 5-point scale), use of opiates, and working status.

By 6 months, differences virtually disappeared
Both the active and passive interventions reduced arm and neck pain faster than the wait-and-see strategy. At 6 weeks, participants in both the PT and cervical collar groups reported a 31-mm reduction in arm pain (P=.007 and .006, respectively), compared with a 19-mm reduction for those in the control group (P=.006). This is a clinically meaningful difference.

The rate of reduction in neck pain over the first 6 weeks was: PT group, 2.4 mm/week, P=.002; cervical collar group, 2.8 mm/week, P=<.001; and control group, 0.9 mm/week. The rate of reduction in the NDI was 1.4 points per week for the control group vs 2.3 points per week for the cervical collar group (P= .024). The PT group fared no better on the NDI measure than the control group. This may reflect the fact that the index predominantly measured disability caused by neck pain, whereas arm pain scores,—which were highest initially—showed the greatest improvement, the authors note.

At 6 months, pain and disability had almost resolved for all the patients, regardless of their treatment group, and secondary outcomes—treatment satisfaction, analgesic use, and working status—were similar for all 3 groups.

WHAT'S NEW: High-quality RCT supports PT and cervical collar

Some investigators have advocated the short-term use of immobilization with either a cervical collar or a cervical pillow during sleep. Until now, however, there was no conclusive evidence about the benefits of this approach.

One earlier RCT (N=493) compared 5 treatment modalities—traction, positioning, collar, placebo tablets, and heat treatment—and found no significant difference in pain and ability to work.⁷ That trial was done nearly 15 years ago, however, and the investigators did not use validated outcome scales. Therefore, the trial would not meet current RCT standards.

The study we report on here leaves little doubt that the 2 treatments reviewed—PT and cervical collar—provide more rapid relief than a wait-and-see approach.

CAVEATS: Pain meds still needed, unanswered questions remain

Although the cervical collar and PT groups had less pain at 3 and 6 weeks compared with the controls—and all 3 groups showed equal improvement at study’s end—the researchers found little difference in use of analgesics. Data on adherence to treatment was recorded by patients, so treatment adherence may not be completely accurate.⁸

Patients without severe arm pain or signs of muscle weakness were not included in this study, so we don’t know whether individuals with less severe cervical radiculopathy would benefit from these treatments. What’s more, this study focused only on new cases of acute cervical radiculopathy, and the findings may not apply to patients with chronic, recurrent, or persistent symptoms.

The apparent contradiction in the finding that both immobilization and PT are beneficial does not have a clear scientific explanation. The researchers hypothesize that immobilizing the neck with a collar reduces foraminal root compression and inflammation; this could explain the larger reduction in arm pain compared with neck pain and neck disability found in this study. The mechanism of pain reduction with PT is unclear, although it is probably related to the restoration of the neck musculature’s strength and range of motion.

Cost is another issue. A cervical collar and rest is at least as effective as PT for recent onset cervical radiculopathy, but the collar costs only about $20—far less than the cost of 12 sessions of therapy.

One final caveat: Any patient with persistent or worsening symptoms should undergo additional evaluation, including imaging.

CHALLENGES TO IMPLEMENTATION: Rest is contrary to usual approach

Some physicians may not agree with the recommendation to encourage rest. Indeed, rest and immobilization are contrary to the usual recommendation for musculoskeletal injuries—to resume activity as soon as possible.

Patients might not like wearing a collar for a variety of personal reasons, such as cosmetic appearance or limitations of motion. On the other hand, some patients may feel that their pain is too severe to be able to participate in PT—which may also be too expensive for, or not readily available to, some patients.

Acknowledgement
The PURLs Surveillance System is supported in part by Grant number UL1RR024999 from the National Center for Research Resources; the grant was a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A 24-year-old woman comes to your clinic because of frequent urination for the past 2 to 3 days. She is not taking any medication, but does take a daily prenatal vitamin because she and her husband are trying to conceive. After your examination, you order a urinalysis and perform a urine pregnancy test. The urinalysis shows bacteriuria ≥100,000 cfu/ml, and the pregnancy test is positive.

What will you prescribe to treat her urinary tract infection?

Antibacterial agents are among the most commonly used medications during pregnancy because treatment of infections is critical to both maternal and fetal well-being.¹ Urinary tract infections (UTIs) are the most common medical complication during pregnancy, with Escherichia coli contributing to roughly 90% of the infections.² Screening for and treating asymptomatic bacteriuria is also recommended during pregnancy to prevent pyelonephritis and increased maternal and fetal morbidity.³ In addition, UTIs are common in reproductive-age women who may not know they are pregnant or who become pregnant during treatment with antibiotics. And nitrofurantoin and sulfonamides are commonly prescribed antibiotics for the treatment of UTIs, both in pregnant women and women of reproductive age.

Prior warnings only address near-term pregnancy
Prior to the study detailed in this PURL, no clinical trials had reported a teratogenic risk associated with either nitrofurantoin (current pregnancy category B) or sulfonamide (current pregnancy category C).⁴ It is recommended, however, that both of these antibacterials be avoided in pregnant women who are near term because of the risk of hemolytic disease in patients with glucose-6-phosphate dehydrogenase deficiency associated with nitrofurantoin and the risk of kernicterus in neonates exposed to sulfamethoxazole.⁵

But a rise in E coli resistance to penicillins (resistance to amoxicillin, for example, can be as high as 30-40%⁶) has led to greater use of nitrofurantoin. The drug has been viewed as a safe and effective alternative treatment for UTIs associated with E coli.⁷ Indeed, nitrofurantoin has been considered to be the preferred antibiotic for bacteriuria suppression, as both ampicillin and cephalosporins can interfere with the normal gastrointestinal flora. Thus, nitrofurantoin is used extensively in pregnant women. Sulfonamides are also prescribed for pregnant women, although not as frequently.^7,8

STUDY SUMMARY: First trimester use linked to many defects

The study by Crider et al¹ was based on the National Birth Defects Prevention Study, an ongoing, population-based case control study of an estimated annual birth population of roughly 482,000, including cases identified by birth defects surveillance registries in 10 states.⁹ The researchers identified pregnancies affected by any of 30 types of birth defects from 1997 to 2003 (n=13,155). The controls (n=4941) were randomly selected from similar geographic locations, and matched for race/ethnicity, age, and prepregnancy body mass index. Exposure to antibacterials from 1 month prepregnancy through the end of the first trimester was recorded.

Crider et al interviewed all the participants up to 24 months after delivery to obtain their exposure history to penicillins, erythromycins, nitrofurantoin, sulfonamides, cephalosporins, quinolones, tetracyclines, other miscellaneous beta-lactams, aminoglycosides, antimycobacterial agents, and other antibiotics. (Exposure to antivirals, antifungals, and antiparasitic agents was not addressed.) Women who didn’t know whether they had been exposed to these agents or could not remember the timing of exposure were excluded.

Overall, antibacterial use ranged from 2% to 5.8%, and peaked in the third month of pregnancy. Penicillins were the most commonly used antibiotics. Odds ratios obtained for birth defects were adjusted for confounders such as maternal age, race, education level, prepregnancy body mass index, time from estimated date of delivery to the interview, use of folic acid or multivitamins from 1 month prior to pregnancy through the first month, and periconceptional smoking and/or alcohol use.

Nitrofurantoin was associated with anophthalmia or microphthalmos (adjusted odds ratio [AOR]=3.7; 95% confidence interval [CI], 1.1-12.2), hypoplastic left heart syndrome (AOR=4.2; 95% CI, 1.9-9.1), atrial septal defects (AOR=1.9; 95% CI, 1.1-3.4), and cleft lip with cleft palate (AOR=2.1; 95% CI, 1.2-3.9).

Sulfonamides were associated with anencephaly (AOR=3.4; 95% CI, 1.3-8.8), hypoplastic left heart syndrome (AOR=3.2; 95% CI, 1.3-7.6), coarctation of the aorta (AOR=2.7; 95% CI, 1.3-5.6), choanal atresia (AOR=8.0; 95% CI, 2.7-23.5), transverse limb deficiency (AOR=2.5; 95% CI, 1.0-5.9), and diaphragmatic hernia (AOR=2.4; 95% CI, 1.1-5.4).

Some links between other antibiotics and birth defects were also found. For example, erythromycins were associated with anencephaly and transverse limb deficiency, penicillins with intercalary limb deficiency, and cephalosporins with atrial septal defects. The authors noted, however, that these agents, which are commonly prescribed for pregnant women, were not associated with many birth defects—and that because of limited sample sizes for these drug classes, the associations may be spurious.

WHAT'S NEW: A large-scale study provides evidence of risk

Previous case studies and meta-analysis have shown no link between the use of nitrofurantoin and congenital abnormalities.⁸ Similarly, sulfonamides have not appeared to pose significant teratogenic risk. This is the first large-scale study evaluating the risk of birth defects associated with antibiotic use during pregnancy, and therefore provides evidence of risk not previously available.

CAVEATS: Study design raises questions of recall bias

The retrospective case-control methodology used in this study leaves open the possibility of recall bias, misclassification bias, and confounding bias. The length of time from actual exposure to data collection could affect the accuracy of participants’ memories. The data gathered were not cross-verified against medical records, and other issues, such as the possible effect of medications for other infections (eg, antivirals and antifungals), could not be measured. However, women who did not know or were unsure of their medication exposure history were excluded from the analysis, which should reduce the risk of this potential bias.

The investigators also controlled for several important sources of potential confounding bias, and the reporting rates were similar among participants in both the case and control groups. These measures provide some assurance that the outcomes are valid.

It would be unethical (and extraordinarily expensive) to conduct a prospective randomized controlled trial to confirm these findings. Case-control methodology is the most practical way to assess for the risk of birth defects, and our literature review suggests that this is the most rigorous study to date. In our view, the potential harm from continuing to use these antibiotics for pregnant women and women who may become pregnant far outweighs the risk that these findings may be erroneous.

That said, a final caveat is the fact that even a several-fold increase in the risk of a rare major birth defect such as those reported in this study is still a rare risk. There may be clinical situations in which the benefits of using nitrofurantoin or sulfonamides in women who are or may become pregnant outweigh the potential risks.

CHALLENGES TO IMPLEMENTATION: Finding an alternative treatment

The main challenge to implementing this new recommendation lies in choosing alternative antibiotics with which to treat UTIs in reproductive-age women and bacteriuria in pregnancy. Obtaining a pregnancy test in sexually active patients of reproductive age who are not using a reliable form of contraception seems like a prudent first step.

If the pregnancy test is positive, cephalexin should be a good initial choice until the results of culture and sensitivities are available. In the event of Enterococcus infection (for which cephalosporins are not active) or other organisms resistant to cephalosporins, the sensitivity results should provide guidance.³

Acknowledgment
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant was a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A 24-year-old woman comes to your clinic because of frequent urination for the past 2 to 3 days. She is not taking any medication, but does take a daily prenatal vitamin because she and her husband are trying to conceive. After your examination, you order a urinalysis and perform a urine pregnancy test. The urinalysis shows bacteriuria ≥100,000 cfu/ml, and the pregnancy test is positive.

What will you prescribe to treat her urinary tract infection?

Antibacterial agents are among the most commonly used medications during pregnancy because treatment of infections is critical to both maternal and fetal well-being.¹ Urinary tract infections (UTIs) are the most common medical complication during pregnancy, with Escherichia coli contributing to roughly 90% of the infections.² Screening for and treating asymptomatic bacteriuria is also recommended during pregnancy to prevent pyelonephritis and increased maternal and fetal morbidity.³ In addition, UTIs are common in reproductive-age women who may not know they are pregnant or who become pregnant during treatment with antibiotics. And nitrofurantoin and sulfonamides are commonly prescribed antibiotics for the treatment of UTIs, both in pregnant women and women of reproductive age.

Prior warnings only address near-term pregnancy
Prior to the study detailed in this PURL, no clinical trials had reported a teratogenic risk associated with either nitrofurantoin (current pregnancy category B) or sulfonamide (current pregnancy category C).⁴ It is recommended, however, that both of these antibacterials be avoided in pregnant women who are near term because of the risk of hemolytic disease in patients with glucose-6-phosphate dehydrogenase deficiency associated with nitrofurantoin and the risk of kernicterus in neonates exposed to sulfamethoxazole.⁵

But a rise in E coli resistance to penicillins (resistance to amoxicillin, for example, can be as high as 30-40%⁶) has led to greater use of nitrofurantoin. The drug has been viewed as a safe and effective alternative treatment for UTIs associated with E coli.⁷ Indeed, nitrofurantoin has been considered to be the preferred antibiotic for bacteriuria suppression, as both ampicillin and cephalosporins can interfere with the normal gastrointestinal flora. Thus, nitrofurantoin is used extensively in pregnant women. Sulfonamides are also prescribed for pregnant women, although not as frequently.^7,8

STUDY SUMMARY: First trimester use linked to many defects

The study by Crider et al¹ was based on the National Birth Defects Prevention Study, an ongoing, population-based case control study of an estimated annual birth population of roughly 482,000, including cases identified by birth defects surveillance registries in 10 states.⁹ The researchers identified pregnancies affected by any of 30 types of birth defects from 1997 to 2003 (n=13,155). The controls (n=4941) were randomly selected from similar geographic locations, and matched for race/ethnicity, age, and prepregnancy body mass index. Exposure to antibacterials from 1 month prepregnancy through the end of the first trimester was recorded.

Crider et al interviewed all the participants up to 24 months after delivery to obtain their exposure history to penicillins, erythromycins, nitrofurantoin, sulfonamides, cephalosporins, quinolones, tetracyclines, other miscellaneous beta-lactams, aminoglycosides, antimycobacterial agents, and other antibiotics. (Exposure to antivirals, antifungals, and antiparasitic agents was not addressed.) Women who didn’t know whether they had been exposed to these agents or could not remember the timing of exposure were excluded.

Overall, antibacterial use ranged from 2% to 5.8%, and peaked in the third month of pregnancy. Penicillins were the most commonly used antibiotics. Odds ratios obtained for birth defects were adjusted for confounders such as maternal age, race, education level, prepregnancy body mass index, time from estimated date of delivery to the interview, use of folic acid or multivitamins from 1 month prior to pregnancy through the first month, and periconceptional smoking and/or alcohol use.

Nitrofurantoin was associated with anophthalmia or microphthalmos (adjusted odds ratio [AOR]=3.7; 95% confidence interval [CI], 1.1-12.2), hypoplastic left heart syndrome (AOR=4.2; 95% CI, 1.9-9.1), atrial septal defects (AOR=1.9; 95% CI, 1.1-3.4), and cleft lip with cleft palate (AOR=2.1; 95% CI, 1.2-3.9).

Sulfonamides were associated with anencephaly (AOR=3.4; 95% CI, 1.3-8.8), hypoplastic left heart syndrome (AOR=3.2; 95% CI, 1.3-7.6), coarctation of the aorta (AOR=2.7; 95% CI, 1.3-5.6), choanal atresia (AOR=8.0; 95% CI, 2.7-23.5), transverse limb deficiency (AOR=2.5; 95% CI, 1.0-5.9), and diaphragmatic hernia (AOR=2.4; 95% CI, 1.1-5.4).

Some links between other antibiotics and birth defects were also found. For example, erythromycins were associated with anencephaly and transverse limb deficiency, penicillins with intercalary limb deficiency, and cephalosporins with atrial septal defects. The authors noted, however, that these agents, which are commonly prescribed for pregnant women, were not associated with many birth defects—and that because of limited sample sizes for these drug classes, the associations may be spurious.

WHAT'S NEW: A large-scale study provides evidence of risk

Previous case studies and meta-analysis have shown no link between the use of nitrofurantoin and congenital abnormalities.⁸ Similarly, sulfonamides have not appeared to pose significant teratogenic risk. This is the first large-scale study evaluating the risk of birth defects associated with antibiotic use during pregnancy, and therefore provides evidence of risk not previously available.

CAVEATS: Study design raises questions of recall bias

The retrospective case-control methodology used in this study leaves open the possibility of recall bias, misclassification bias, and confounding bias. The length of time from actual exposure to data collection could affect the accuracy of participants’ memories. The data gathered were not cross-verified against medical records, and other issues, such as the possible effect of medications for other infections (eg, antivirals and antifungals), could not be measured. However, women who did not know or were unsure of their medication exposure history were excluded from the analysis, which should reduce the risk of this potential bias.

The investigators also controlled for several important sources of potential confounding bias, and the reporting rates were similar among participants in both the case and control groups. These measures provide some assurance that the outcomes are valid.

It would be unethical (and extraordinarily expensive) to conduct a prospective randomized controlled trial to confirm these findings. Case-control methodology is the most practical way to assess for the risk of birth defects, and our literature review suggests that this is the most rigorous study to date. In our view, the potential harm from continuing to use these antibiotics for pregnant women and women who may become pregnant far outweighs the risk that these findings may be erroneous.

That said, a final caveat is the fact that even a several-fold increase in the risk of a rare major birth defect such as those reported in this study is still a rare risk. There may be clinical situations in which the benefits of using nitrofurantoin or sulfonamides in women who are or may become pregnant outweigh the potential risks.

CHALLENGES TO IMPLEMENTATION: Finding an alternative treatment

The main challenge to implementing this new recommendation lies in choosing alternative antibiotics with which to treat UTIs in reproductive-age women and bacteriuria in pregnancy. Obtaining a pregnancy test in sexually active patients of reproductive age who are not using a reliable form of contraception seems like a prudent first step.

If the pregnancy test is positive, cephalexin should be a good initial choice until the results of culture and sensitivities are available. In the event of Enterococcus infection (for which cephalosporins are not active) or other organisms resistant to cephalosporins, the sensitivity results should provide guidance.³

Acknowledgment
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant was a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

Illustrative case

A 24-year-old woman comes to your clinic because of frequent urination for the past 2 to 3 days. She is not taking any medication, but does take a daily prenatal vitamin because she and her husband are trying to conceive. After your examination, you order a urinalysis and perform a urine pregnancy test. The urinalysis shows bacteriuria ≥100,000 cfu/ml, and the pregnancy test is positive.

What will you prescribe to treat her urinary tract infection?

Antibacterial agents are among the most commonly used medications during pregnancy because treatment of infections is critical to both maternal and fetal well-being.¹ Urinary tract infections (UTIs) are the most common medical complication during pregnancy, with Escherichia coli contributing to roughly 90% of the infections.² Screening for and treating asymptomatic bacteriuria is also recommended during pregnancy to prevent pyelonephritis and increased maternal and fetal morbidity.³ In addition, UTIs are common in reproductive-age women who may not know they are pregnant or who become pregnant during treatment with antibiotics. And nitrofurantoin and sulfonamides are commonly prescribed antibiotics for the treatment of UTIs, both in pregnant women and women of reproductive age.

Prior warnings only address near-term pregnancy
Prior to the study detailed in this PURL, no clinical trials had reported a teratogenic risk associated with either nitrofurantoin (current pregnancy category B) or sulfonamide (current pregnancy category C).⁴ It is recommended, however, that both of these antibacterials be avoided in pregnant women who are near term because of the risk of hemolytic disease in patients with glucose-6-phosphate dehydrogenase deficiency associated with nitrofurantoin and the risk of kernicterus in neonates exposed to sulfamethoxazole.⁵

But a rise in E coli resistance to penicillins (resistance to amoxicillin, for example, can be as high as 30-40%⁶) has led to greater use of nitrofurantoin. The drug has been viewed as a safe and effective alternative treatment for UTIs associated with E coli.⁷ Indeed, nitrofurantoin has been considered to be the preferred antibiotic for bacteriuria suppression, as both ampicillin and cephalosporins can interfere with the normal gastrointestinal flora. Thus, nitrofurantoin is used extensively in pregnant women. Sulfonamides are also prescribed for pregnant women, although not as frequently.^7,8

STUDY SUMMARY: First trimester use linked to many defects

The study by Crider et al¹ was based on the National Birth Defects Prevention Study, an ongoing, population-based case control study of an estimated annual birth population of roughly 482,000, including cases identified by birth defects surveillance registries in 10 states.⁹ The researchers identified pregnancies affected by any of 30 types of birth defects from 1997 to 2003 (n=13,155). The controls (n=4941) were randomly selected from similar geographic locations, and matched for race/ethnicity, age, and prepregnancy body mass index. Exposure to antibacterials from 1 month prepregnancy through the end of the first trimester was recorded.

Crider et al interviewed all the participants up to 24 months after delivery to obtain their exposure history to penicillins, erythromycins, nitrofurantoin, sulfonamides, cephalosporins, quinolones, tetracyclines, other miscellaneous beta-lactams, aminoglycosides, antimycobacterial agents, and other antibiotics. (Exposure to antivirals, antifungals, and antiparasitic agents was not addressed.) Women who didn’t know whether they had been exposed to these agents or could not remember the timing of exposure were excluded.

Overall, antibacterial use ranged from 2% to 5.8%, and peaked in the third month of pregnancy. Penicillins were the most commonly used antibiotics. Odds ratios obtained for birth defects were adjusted for confounders such as maternal age, race, education level, prepregnancy body mass index, time from estimated date of delivery to the interview, use of folic acid or multivitamins from 1 month prior to pregnancy through the first month, and periconceptional smoking and/or alcohol use.

Nitrofurantoin was associated with anophthalmia or microphthalmos (adjusted odds ratio [AOR]=3.7; 95% confidence interval [CI], 1.1-12.2), hypoplastic left heart syndrome (AOR=4.2; 95% CI, 1.9-9.1), atrial septal defects (AOR=1.9; 95% CI, 1.1-3.4), and cleft lip with cleft palate (AOR=2.1; 95% CI, 1.2-3.9).

Sulfonamides were associated with anencephaly (AOR=3.4; 95% CI, 1.3-8.8), hypoplastic left heart syndrome (AOR=3.2; 95% CI, 1.3-7.6), coarctation of the aorta (AOR=2.7; 95% CI, 1.3-5.6), choanal atresia (AOR=8.0; 95% CI, 2.7-23.5), transverse limb deficiency (AOR=2.5; 95% CI, 1.0-5.9), and diaphragmatic hernia (AOR=2.4; 95% CI, 1.1-5.4).

Some links between other antibiotics and birth defects were also found. For example, erythromycins were associated with anencephaly and transverse limb deficiency, penicillins with intercalary limb deficiency, and cephalosporins with atrial septal defects. The authors noted, however, that these agents, which are commonly prescribed for pregnant women, were not associated with many birth defects—and that because of limited sample sizes for these drug classes, the associations may be spurious.

WHAT'S NEW: A large-scale study provides evidence of risk

Previous case studies and meta-analysis have shown no link between the use of nitrofurantoin and congenital abnormalities.⁸ Similarly, sulfonamides have not appeared to pose significant teratogenic risk. This is the first large-scale study evaluating the risk of birth defects associated with antibiotic use during pregnancy, and therefore provides evidence of risk not previously available.

CAVEATS: Study design raises questions of recall bias

The retrospective case-control methodology used in this study leaves open the possibility of recall bias, misclassification bias, and confounding bias. The length of time from actual exposure to data collection could affect the accuracy of participants’ memories. The data gathered were not cross-verified against medical records, and other issues, such as the possible effect of medications for other infections (eg, antivirals and antifungals), could not be measured. However, women who did not know or were unsure of their medication exposure history were excluded from the analysis, which should reduce the risk of this potential bias.

The investigators also controlled for several important sources of potential confounding bias, and the reporting rates were similar among participants in both the case and control groups. These measures provide some assurance that the outcomes are valid.

It would be unethical (and extraordinarily expensive) to conduct a prospective randomized controlled trial to confirm these findings. Case-control methodology is the most practical way to assess for the risk of birth defects, and our literature review suggests that this is the most rigorous study to date. In our view, the potential harm from continuing to use these antibiotics for pregnant women and women who may become pregnant far outweighs the risk that these findings may be erroneous.

That said, a final caveat is the fact that even a several-fold increase in the risk of a rare major birth defect such as those reported in this study is still a rare risk. There may be clinical situations in which the benefits of using nitrofurantoin or sulfonamides in women who are or may become pregnant outweigh the potential risks.

CHALLENGES TO IMPLEMENTATION: Finding an alternative treatment

The main challenge to implementing this new recommendation lies in choosing alternative antibiotics with which to treat UTIs in reproductive-age women and bacteriuria in pregnancy. Obtaining a pregnancy test in sexually active patients of reproductive age who are not using a reliable form of contraception seems like a prudent first step.

If the pregnancy test is positive, cephalexin should be a good initial choice until the results of culture and sensitivities are available. In the event of Enterococcus infection (for which cephalosporins are not active) or other organisms resistant to cephalosporins, the sensitivity results should provide guidance.³

Acknowledgment
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources; the grant was a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY