Clinical Inquiries

Evidence summary

Significant macrocytosis is usually defined as a mean corpuscular volume greater than 99 femtoliters (fL). The prevalence of macrocytosis (with or without anemia) ranges from 1.7% to 5.0%.^1-4 As many as 60% to 80% of primary care patients may not have anemia.^3,4

Because no study has looked specifically at evaluating macrocytosis without anemia, extrapolation from studies of all presentations of macrocytosis (with and without anemia) must help guide evaluation.^1,3,5-7 The causes of macrocytosis vary depending on the population studied (TABLE). In primary care, alcohol use and vitamin deficiency are common causes. Even after evaluation, approximately 10% of cases remain unexplained.³

TABLE
Causes of macrocytosis: What prospective studies show

Clues in the history, physical exam, and lab results

A history focusing specifically on alcohol use and medications—especially chemotherapeutics, antiretroviral drugs, and antiseizure medications—can provide important clues to the cause of macrocytosis. During the physical examination, look for signs consistent with chronic liver disease.

Laboratory studies can help identify vitamin deficiencies, liver disease, and thyroid disease. A normal serum B₁₂ level may not rule out a true B₁₂ deficiency, but normal levels of the metabolites methylmalonic acid and homocysteine do essentially rule it out.⁸ In this era of folic acid fortification, the utility of the serum folate level is uncertain. Several studies suggest empiric treatment with folic acid instead of testing for a deficiency when B₁₂ deficiency has been ruled out.^7,9

Liver disease—which may be confounded by alcohol abuse, medications, or cancer—is a common cause of macrocytosis.⁵ Hypothyroidism is rarely a cause, but may be more prevalent in the elderly.¹⁰

What these 2 tests may, or may not, tell you

Although several authorities recommend a peripheral smear and reticulocyte count to help evaluate macrocytic anemia, no specific recommendations exist for these tests in the absence of anemia. A peripheral smear can detect megaloblastic changes typical of B₁₂ and folate deficiency and other marrow disorders, especially myelodysplastic changes. Peripheral smear findings and reticulocytosis can also show evidence of hemolysis. However, megaloblastic changes and marrow-related changes on peripheral smear are typically seen with anemia.

In 2 prospective studies of primary care patients, 1 reported little diagnostic value for the peripheral smear,¹ and the other found that reticulocytosis rarely caused macrocytosis.⁵ A prospective study of 300 hospitalized patients with macrocytosis found that 100% of marrow disorders and hemolysis that caused macrocytosis also caused an associated anemia.⁷ A retrospective chart review of 113 cases of macrocytosis in outpatients found that general practitioners often didn’t order a peripheral smear and reticulocyte count to complete their diagnostic workups.⁴

Bone marrow biopsy may reveal dysplastic changes, but not a Dx

A prospective study of the utility of bone marrow biopsy in 124 elderly patients with macrocytosis found that as many as 60% were diagnosed by blood tests alone. All the remaining patients with unexplained macrocytosis underwent bone marrow biopsy, which showed early dysplastic changes in 39%, but did not provide a diagnosis in nearly 50%. Twelve percent were found to have myelodysplastic syndrome, but they had a mean hemoglobin of 8.5 g/dL.¹⁰

Recommendations

We were unable to find published guidelines for the evaluation of macrocytosis without anemia by the American Society of Hematology, the British Committee for Standards in Haematology, or in an authoritative hematology text.¹¹

Acknowledgments

The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Army, Department of Defense, or the US Government.

Evidence summary

Significant macrocytosis is usually defined as a mean corpuscular volume greater than 99 femtoliters (fL). The prevalence of macrocytosis (with or without anemia) ranges from 1.7% to 5.0%.^1-4 As many as 60% to 80% of primary care patients may not have anemia.^3,4

Because no study has looked specifically at evaluating macrocytosis without anemia, extrapolation from studies of all presentations of macrocytosis (with and without anemia) must help guide evaluation.^1,3,5-7 The causes of macrocytosis vary depending on the population studied (TABLE). In primary care, alcohol use and vitamin deficiency are common causes. Even after evaluation, approximately 10% of cases remain unexplained.³

TABLE
Causes of macrocytosis: What prospective studies show

Clues in the history, physical exam, and lab results

A history focusing specifically on alcohol use and medications—especially chemotherapeutics, antiretroviral drugs, and antiseizure medications—can provide important clues to the cause of macrocytosis. During the physical examination, look for signs consistent with chronic liver disease.

Laboratory studies can help identify vitamin deficiencies, liver disease, and thyroid disease. A normal serum B₁₂ level may not rule out a true B₁₂ deficiency, but normal levels of the metabolites methylmalonic acid and homocysteine do essentially rule it out.⁸ In this era of folic acid fortification, the utility of the serum folate level is uncertain. Several studies suggest empiric treatment with folic acid instead of testing for a deficiency when B₁₂ deficiency has been ruled out.^7,9

Liver disease—which may be confounded by alcohol abuse, medications, or cancer—is a common cause of macrocytosis.⁵ Hypothyroidism is rarely a cause, but may be more prevalent in the elderly.¹⁰

What these 2 tests may, or may not, tell you

Although several authorities recommend a peripheral smear and reticulocyte count to help evaluate macrocytic anemia, no specific recommendations exist for these tests in the absence of anemia. A peripheral smear can detect megaloblastic changes typical of B₁₂ and folate deficiency and other marrow disorders, especially myelodysplastic changes. Peripheral smear findings and reticulocytosis can also show evidence of hemolysis. However, megaloblastic changes and marrow-related changes on peripheral smear are typically seen with anemia.

In 2 prospective studies of primary care patients, 1 reported little diagnostic value for the peripheral smear,¹ and the other found that reticulocytosis rarely caused macrocytosis.⁵ A prospective study of 300 hospitalized patients with macrocytosis found that 100% of marrow disorders and hemolysis that caused macrocytosis also caused an associated anemia.⁷ A retrospective chart review of 113 cases of macrocytosis in outpatients found that general practitioners often didn’t order a peripheral smear and reticulocyte count to complete their diagnostic workups.⁴

Bone marrow biopsy may reveal dysplastic changes, but not a Dx

A prospective study of the utility of bone marrow biopsy in 124 elderly patients with macrocytosis found that as many as 60% were diagnosed by blood tests alone. All the remaining patients with unexplained macrocytosis underwent bone marrow biopsy, which showed early dysplastic changes in 39%, but did not provide a diagnosis in nearly 50%. Twelve percent were found to have myelodysplastic syndrome, but they had a mean hemoglobin of 8.5 g/dL.¹⁰

Recommendations

We were unable to find published guidelines for the evaluation of macrocytosis without anemia by the American Society of Hematology, the British Committee for Standards in Haematology, or in an authoritative hematology text.¹¹

Acknowledgments

The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Army, Department of Defense, or the US Government.

Evidence summary

Significant macrocytosis is usually defined as a mean corpuscular volume greater than 99 femtoliters (fL). The prevalence of macrocytosis (with or without anemia) ranges from 1.7% to 5.0%.^1-4 As many as 60% to 80% of primary care patients may not have anemia.^3,4

Because no study has looked specifically at evaluating macrocytosis without anemia, extrapolation from studies of all presentations of macrocytosis (with and without anemia) must help guide evaluation.^1,3,5-7 The causes of macrocytosis vary depending on the population studied (TABLE). In primary care, alcohol use and vitamin deficiency are common causes. Even after evaluation, approximately 10% of cases remain unexplained.³

TABLE
Causes of macrocytosis: What prospective studies show

Clues in the history, physical exam, and lab results

A history focusing specifically on alcohol use and medications—especially chemotherapeutics, antiretroviral drugs, and antiseizure medications—can provide important clues to the cause of macrocytosis. During the physical examination, look for signs consistent with chronic liver disease.

Laboratory studies can help identify vitamin deficiencies, liver disease, and thyroid disease. A normal serum B₁₂ level may not rule out a true B₁₂ deficiency, but normal levels of the metabolites methylmalonic acid and homocysteine do essentially rule it out.⁸ In this era of folic acid fortification, the utility of the serum folate level is uncertain. Several studies suggest empiric treatment with folic acid instead of testing for a deficiency when B₁₂ deficiency has been ruled out.^7,9

Liver disease—which may be confounded by alcohol abuse, medications, or cancer—is a common cause of macrocytosis.⁵ Hypothyroidism is rarely a cause, but may be more prevalent in the elderly.¹⁰

What these 2 tests may, or may not, tell you

Although several authorities recommend a peripheral smear and reticulocyte count to help evaluate macrocytic anemia, no specific recommendations exist for these tests in the absence of anemia. A peripheral smear can detect megaloblastic changes typical of B₁₂ and folate deficiency and other marrow disorders, especially myelodysplastic changes. Peripheral smear findings and reticulocytosis can also show evidence of hemolysis. However, megaloblastic changes and marrow-related changes on peripheral smear are typically seen with anemia.

In 2 prospective studies of primary care patients, 1 reported little diagnostic value for the peripheral smear,¹ and the other found that reticulocytosis rarely caused macrocytosis.⁵ A prospective study of 300 hospitalized patients with macrocytosis found that 100% of marrow disorders and hemolysis that caused macrocytosis also caused an associated anemia.⁷ A retrospective chart review of 113 cases of macrocytosis in outpatients found that general practitioners often didn’t order a peripheral smear and reticulocyte count to complete their diagnostic workups.⁴

Bone marrow biopsy may reveal dysplastic changes, but not a Dx

A prospective study of the utility of bone marrow biopsy in 124 elderly patients with macrocytosis found that as many as 60% were diagnosed by blood tests alone. All the remaining patients with unexplained macrocytosis underwent bone marrow biopsy, which showed early dysplastic changes in 39%, but did not provide a diagnosis in nearly 50%. Twelve percent were found to have myelodysplastic syndrome, but they had a mean hemoglobin of 8.5 g/dL.¹⁰

Recommendations

We were unable to find published guidelines for the evaluation of macrocytosis without anemia by the American Society of Hematology, the British Committee for Standards in Haematology, or in an authoritative hematology text.¹¹

Acknowledgments

The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Army, Department of Defense, or the US Government.

Evidence summary

Rhabdomyolysis is a well-described clinical syndrome resulting from injury to skeletal muscle and subsequent release of cellular contents into the extracellular fluid and circulation. It can lead to many complications, including renal failure, disseminated intravascular coagulation, and even death in 5% of cases.² The leading causes of rhabdomyolysis include trauma, soft tissue compression, alcohol, drugs, infections, seizures, and exercise.³

Only half of patients experience muscle pain.² Elevations occur in multiple serum markers, including CK, myoglobin, aldolase, lactate dehydrogenase, alanine aminotransferase, and aspartate aminotransferase, in either plasma or urine.^4,5

Variable elevations, ranging from mild to extreme, that are discovered incidentally after exercise may cause clinical uncertainty.

No clear consensus defines CK levels in rhabdomyolysis

CK is the primary serum marker for rhabdomyolysis. It’s highly sensitive, but not specific. No clear consensus exists on what threshold of CK elevation correlates with clinically relevant disease.⁶ A relationship between CK elevation and the severity of disease has been established (>6000 IU/L predicts renal failure), but patients can have significant morbidity with only moderately elevated CK levels.^7,8 Normal reference ranges for serum CK are 55 to 170 IU/L for males, and 30 to 135 IU/L for females.⁹

Recent definitions of rhabdomyolysis have been established to address muscle toxicity from lipid-lowering medications. The United States Food and Drug Administration specifies a CK level of more than 50 times the upper limit of normal (ULN)—or 10,000 IU/L—accompanied by organ damage, usually renal compromise.⁶ The National Lipid Association’s Muscle Safety Expert Panel has defined rhabdomyolysis as any evidence of muscle cell destruction regardless of the CK level and a causal relationship to a change in renal function. The panel further subdivides CK elevations into categories of mild (<10 times ULN), moderate (10-49 times ULN), and marked (≥50 times ULN).⁶

Exercise elevates CK level, but consider other factors, too

Although exercise is known to elevate CK, it produces a wide range of levels, based on a host of variables.^3,10 Increases in CK are more pronounced in males, blacks, and untrained people; age doesn’t seem to be a factor.^10,11 Higher-intensity, longer-duration, and weight-bearing exercise (eccentric muscular contractions and downhill running) cause the greatest rises in CK.¹⁰ Other influences include temperature, altitude, gravitational forces, noise, and vibration.

No studies firmly establish a normal range of CK elevation from moderate exercise; better data are available for extreme athletes, such as long-distance runners and triathletes. One study found that mean total CK elevations 24 hours after a marathon were 3322 IU/L (22.3 times baseline) for men and 946 IU/L (8.6 times baseline) for women.⁴ Another study showed that triathletes had a 12-fold mean increase in CK levels as long as 24 hours after the race.¹²

Eccentric exercises significantly raise CK

Exercise programs that include eccentric muscle contractions can result in significant serum CK elevations. One study followed 203 participants to evaluate the magnitude of CK elevation and the effect on renal function produced by exercise.³ After performing 50 maximal eccentric elbow flexor contractions, 55% of participants had CK elevations >2000 IU/L at 4 days after exercise; 25% had CK elevations >10,000 IU/L; 13% had levels >20,000 IU/L. None showed any evidence of renal compromise on clinical follow-up. Another study found significant increases in CK (approximate mean of 15,000 IU/L) after repetitive eccentric elbow flexor contractions in college-age males.¹³

Eccentric weight lifting and similar activities, like downhill running, may result in an increase in serum CK levels of 10 to 20 times normal, whereas other nonweight-bearing exercises and exercise involving no or minimal eccentric contractions, such as swimming and cycling, cause only nominal increases in serum CK.¹⁰

Recommendations

No formal guidelines from authoritative sources are available.

Acknowledgments

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

Evidence summary

Rhabdomyolysis is a well-described clinical syndrome resulting from injury to skeletal muscle and subsequent release of cellular contents into the extracellular fluid and circulation. It can lead to many complications, including renal failure, disseminated intravascular coagulation, and even death in 5% of cases.² The leading causes of rhabdomyolysis include trauma, soft tissue compression, alcohol, drugs, infections, seizures, and exercise.³

Only half of patients experience muscle pain.² Elevations occur in multiple serum markers, including CK, myoglobin, aldolase, lactate dehydrogenase, alanine aminotransferase, and aspartate aminotransferase, in either plasma or urine.^4,5

Variable elevations, ranging from mild to extreme, that are discovered incidentally after exercise may cause clinical uncertainty.

No clear consensus defines CK levels in rhabdomyolysis

CK is the primary serum marker for rhabdomyolysis. It’s highly sensitive, but not specific. No clear consensus exists on what threshold of CK elevation correlates with clinically relevant disease.⁶ A relationship between CK elevation and the severity of disease has been established (>6000 IU/L predicts renal failure), but patients can have significant morbidity with only moderately elevated CK levels.^7,8 Normal reference ranges for serum CK are 55 to 170 IU/L for males, and 30 to 135 IU/L for females.⁹

Recent definitions of rhabdomyolysis have been established to address muscle toxicity from lipid-lowering medications. The United States Food and Drug Administration specifies a CK level of more than 50 times the upper limit of normal (ULN)—or 10,000 IU/L—accompanied by organ damage, usually renal compromise.⁶ The National Lipid Association’s Muscle Safety Expert Panel has defined rhabdomyolysis as any evidence of muscle cell destruction regardless of the CK level and a causal relationship to a change in renal function. The panel further subdivides CK elevations into categories of mild (<10 times ULN), moderate (10-49 times ULN), and marked (≥50 times ULN).⁶

Exercise elevates CK level, but consider other factors, too

Although exercise is known to elevate CK, it produces a wide range of levels, based on a host of variables.^3,10 Increases in CK are more pronounced in males, blacks, and untrained people; age doesn’t seem to be a factor.^10,11 Higher-intensity, longer-duration, and weight-bearing exercise (eccentric muscular contractions and downhill running) cause the greatest rises in CK.¹⁰ Other influences include temperature, altitude, gravitational forces, noise, and vibration.

No studies firmly establish a normal range of CK elevation from moderate exercise; better data are available for extreme athletes, such as long-distance runners and triathletes. One study found that mean total CK elevations 24 hours after a marathon were 3322 IU/L (22.3 times baseline) for men and 946 IU/L (8.6 times baseline) for women.⁴ Another study showed that triathletes had a 12-fold mean increase in CK levels as long as 24 hours after the race.¹²

Eccentric exercises significantly raise CK

Exercise programs that include eccentric muscle contractions can result in significant serum CK elevations. One study followed 203 participants to evaluate the magnitude of CK elevation and the effect on renal function produced by exercise.³ After performing 50 maximal eccentric elbow flexor contractions, 55% of participants had CK elevations >2000 IU/L at 4 days after exercise; 25% had CK elevations >10,000 IU/L; 13% had levels >20,000 IU/L. None showed any evidence of renal compromise on clinical follow-up. Another study found significant increases in CK (approximate mean of 15,000 IU/L) after repetitive eccentric elbow flexor contractions in college-age males.¹³

Eccentric weight lifting and similar activities, like downhill running, may result in an increase in serum CK levels of 10 to 20 times normal, whereas other nonweight-bearing exercises and exercise involving no or minimal eccentric contractions, such as swimming and cycling, cause only nominal increases in serum CK.¹⁰

Recommendations

No formal guidelines from authoritative sources are available.

Acknowledgments

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

Evidence summary

Rhabdomyolysis is a well-described clinical syndrome resulting from injury to skeletal muscle and subsequent release of cellular contents into the extracellular fluid and circulation. It can lead to many complications, including renal failure, disseminated intravascular coagulation, and even death in 5% of cases.² The leading causes of rhabdomyolysis include trauma, soft tissue compression, alcohol, drugs, infections, seizures, and exercise.³

Only half of patients experience muscle pain.² Elevations occur in multiple serum markers, including CK, myoglobin, aldolase, lactate dehydrogenase, alanine aminotransferase, and aspartate aminotransferase, in either plasma or urine.^4,5

Variable elevations, ranging from mild to extreme, that are discovered incidentally after exercise may cause clinical uncertainty.

No clear consensus defines CK levels in rhabdomyolysis

CK is the primary serum marker for rhabdomyolysis. It’s highly sensitive, but not specific. No clear consensus exists on what threshold of CK elevation correlates with clinically relevant disease.⁶ A relationship between CK elevation and the severity of disease has been established (>6000 IU/L predicts renal failure), but patients can have significant morbidity with only moderately elevated CK levels.^7,8 Normal reference ranges for serum CK are 55 to 170 IU/L for males, and 30 to 135 IU/L for females.⁹

Recent definitions of rhabdomyolysis have been established to address muscle toxicity from lipid-lowering medications. The United States Food and Drug Administration specifies a CK level of more than 50 times the upper limit of normal (ULN)—or 10,000 IU/L—accompanied by organ damage, usually renal compromise.⁶ The National Lipid Association’s Muscle Safety Expert Panel has defined rhabdomyolysis as any evidence of muscle cell destruction regardless of the CK level and a causal relationship to a change in renal function. The panel further subdivides CK elevations into categories of mild (<10 times ULN), moderate (10-49 times ULN), and marked (≥50 times ULN).⁶

Exercise elevates CK level, but consider other factors, too

Although exercise is known to elevate CK, it produces a wide range of levels, based on a host of variables.^3,10 Increases in CK are more pronounced in males, blacks, and untrained people; age doesn’t seem to be a factor.^10,11 Higher-intensity, longer-duration, and weight-bearing exercise (eccentric muscular contractions and downhill running) cause the greatest rises in CK.¹⁰ Other influences include temperature, altitude, gravitational forces, noise, and vibration.

No studies firmly establish a normal range of CK elevation from moderate exercise; better data are available for extreme athletes, such as long-distance runners and triathletes. One study found that mean total CK elevations 24 hours after a marathon were 3322 IU/L (22.3 times baseline) for men and 946 IU/L (8.6 times baseline) for women.⁴ Another study showed that triathletes had a 12-fold mean increase in CK levels as long as 24 hours after the race.¹²

Eccentric exercises significantly raise CK

Exercise programs that include eccentric muscle contractions can result in significant serum CK elevations. One study followed 203 participants to evaluate the magnitude of CK elevation and the effect on renal function produced by exercise.³ After performing 50 maximal eccentric elbow flexor contractions, 55% of participants had CK elevations >2000 IU/L at 4 days after exercise; 25% had CK elevations >10,000 IU/L; 13% had levels >20,000 IU/L. None showed any evidence of renal compromise on clinical follow-up. Another study found significant increases in CK (approximate mean of 15,000 IU/L) after repetitive eccentric elbow flexor contractions in college-age males.¹³

Eccentric weight lifting and similar activities, like downhill running, may result in an increase in serum CK levels of 10 to 20 times normal, whereas other nonweight-bearing exercises and exercise involving no or minimal eccentric contractions, such as swimming and cycling, cause only nominal increases in serum CK.¹⁰

Recommendations

No formal guidelines from authoritative sources are available.

Acknowledgments

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

Evidence summary

Breast stimulation works

A systematic review of 6 trials with a total of 719 participants compared unilateral breast stimulation with no intervention to induce labor in women at term.¹ In 2 trials, the 300 participants stimulated 1 breast for 1 hour a day for 3 days; in the other 4 trials, 419 women stimulated either breast alternately for 3 hours a day.

Breast stimulation significantly reduced the number of women who hadn’t gone into labor at 72 hours compared with nonintervention (62.7% vs 93.6%; relative risk [RR]=0.67; 95% confidence interval [CI], 0.60-0.74; number needed to treat [NNT]=3.2). This result remained significant when primiparous and multiparous women were analyzed separately, but not in the 1 trial (37 participants) that reported on women with an unfavorable cervix.¹ Breast stimulation also reduced postpartum hemorrhage (0.7% vs 6%; RR=0.16; 95% CI, 0.03-0.87; NNT=18.8).

None of the trials documented uterine hyperstimulation or meconium-stained fluid, and they didn’t find significant differences in the rates of cesarean section (9% vs 10%; RR=0.90; 95% CI, 0.38-2.12). One trial reported 4 perinatal deaths in a high-risk population—3 in the breast stimulation group and 1 in the control group—but this finding was not replicated in any other trials.¹

Acupuncture may also help

A Cochrane systematic review evaluating acupuncture for inducing labor identified 1 randomized trial with methodologic flaws (allowing no conclusion), 3 case series, and 2 nonrandomized trials. The first case series used electroacupuncture at 38 to 42 weeks to successfully induce labor in 21 of 31 women. The second series, using acupuncture with and without electrical stimulation, induced labor in 10 of 12 women at 19 to 43 weeks. The third study induced labor with electroacupuncture in 78% of 41 women—34 term and postterm patients and 7 with a fetal demise.

In the first nonrandomized trial, 20 of 27 women at term who received electroacupuncture gave birth between 39 weeks 3 days and 40 weeks, compared with 47 of 102 women in the control group. In the second study, 31 of 35 women who received electroacupuncture reported increased intensity of contractions compared with none of the 35 women in the control group.²

Sexual intercourse induces pregnancy, not labor

A Cochrane systematic review found 1 observational study of the effect of sexual intercourse on cervical ripening and induction of labor among 28 women at term. The study evaluated cervical Bishop scores after sexual intercourse with intravaginal semen deposition for 3 consecutive nights compared with no intercourse. The difference in Bishop score was not significant (1.0 with coitus vs 0.5 controls; P>.05) and no difference in the number of women who delivered within 3 days was noted in the 2 groups (RR=0.99; 95% CI, 0.45-2.20).³

Castor oil made women nauseous

A Cochrane systematic review included 1 study of women at term with intact membranes who received 1 dose of castor oil to induce labor (n=52) compared with no treatment (n=48). All women who received castor oil felt nauseous, but no other differences were noted between the castor oil group and controls in rates of cesarean section, meconium-stained fluid, or Apgar scores less than 7 at 5 minutes.⁴

Herbal preparations raise many questions—and some concerns

A prospective controlled trial that randomized 192 women to receive either red raspberry leaf tablets (1.2 g twice daily beginning at 32 weeks) or placebo found no significant differences in duration of the first stage of labor or birth outcomes. Symptom surveys of 108 women revealed no adverse effects.⁵

A retrospective cohort study compared pregnancy outcomes in 54 women who took evening primrose oil (500 mg 3 times a day beginning at 37 weeks) with a matched group of 54 women who did not take it. The study found no significant differences between the groups in maternal age, Apgar scores, or days of gestation. However, the women taking evening primrose oil tended to have protracted active phase of labor, prolonged rupture of membranes, arrest of descent, and oxytocin augmentation (none of which were statistically significant).⁶

A systematic review of the literature on black cohosh in pregnancy found no trials that evaluated its efficacy for inducing labor.⁷ A review of herbal preparations used by midwives reported a case of an infant with low Apgar scores after black cohosh use.⁸ Both articles described black cohosh as potentially unsafe because of the lack of trials demonstrating safety and possible deleterious estrogenic effects.

A systematic review of the literature on blue cohosh found only in vitro studies of efficacy (increased estradiol-induced transcription in estrogen-responsive cells and increased tone in excised guinea pig uteri) and 3 case reports of maternal adverse events after ingestion (perinatal stroke, congestive heart failure with shock, and multiorgan hypoxic injury).⁹ In vitro evidence presented in the review suggests that blue cohosh may have teratogenic, embryotoxic, and oxytocic effects.

Recommendations

We found no recommendations from professional organizations regarding folk methods to stimulate labor. The authors of a survey of herbal preparations used by midwives do not advocate their use in pregnancy because of unknown risks to the fetus, although “raspberry leaf tea does not seem to have any significant pharmacologic activity and is probably safe.”⁸

Evidence summary

Breast stimulation works

A systematic review of 6 trials with a total of 719 participants compared unilateral breast stimulation with no intervention to induce labor in women at term.¹ In 2 trials, the 300 participants stimulated 1 breast for 1 hour a day for 3 days; in the other 4 trials, 419 women stimulated either breast alternately for 3 hours a day.

Breast stimulation significantly reduced the number of women who hadn’t gone into labor at 72 hours compared with nonintervention (62.7% vs 93.6%; relative risk [RR]=0.67; 95% confidence interval [CI], 0.60-0.74; number needed to treat [NNT]=3.2). This result remained significant when primiparous and multiparous women were analyzed separately, but not in the 1 trial (37 participants) that reported on women with an unfavorable cervix.¹ Breast stimulation also reduced postpartum hemorrhage (0.7% vs 6%; RR=0.16; 95% CI, 0.03-0.87; NNT=18.8).

None of the trials documented uterine hyperstimulation or meconium-stained fluid, and they didn’t find significant differences in the rates of cesarean section (9% vs 10%; RR=0.90; 95% CI, 0.38-2.12). One trial reported 4 perinatal deaths in a high-risk population—3 in the breast stimulation group and 1 in the control group—but this finding was not replicated in any other trials.¹

Acupuncture may also help

A Cochrane systematic review evaluating acupuncture for inducing labor identified 1 randomized trial with methodologic flaws (allowing no conclusion), 3 case series, and 2 nonrandomized trials. The first case series used electroacupuncture at 38 to 42 weeks to successfully induce labor in 21 of 31 women. The second series, using acupuncture with and without electrical stimulation, induced labor in 10 of 12 women at 19 to 43 weeks. The third study induced labor with electroacupuncture in 78% of 41 women—34 term and postterm patients and 7 with a fetal demise.

In the first nonrandomized trial, 20 of 27 women at term who received electroacupuncture gave birth between 39 weeks 3 days and 40 weeks, compared with 47 of 102 women in the control group. In the second study, 31 of 35 women who received electroacupuncture reported increased intensity of contractions compared with none of the 35 women in the control group.²

Sexual intercourse induces pregnancy, not labor

A Cochrane systematic review found 1 observational study of the effect of sexual intercourse on cervical ripening and induction of labor among 28 women at term. The study evaluated cervical Bishop scores after sexual intercourse with intravaginal semen deposition for 3 consecutive nights compared with no intercourse. The difference in Bishop score was not significant (1.0 with coitus vs 0.5 controls; P>.05) and no difference in the number of women who delivered within 3 days was noted in the 2 groups (RR=0.99; 95% CI, 0.45-2.20).³

Castor oil made women nauseous

A Cochrane systematic review included 1 study of women at term with intact membranes who received 1 dose of castor oil to induce labor (n=52) compared with no treatment (n=48). All women who received castor oil felt nauseous, but no other differences were noted between the castor oil group and controls in rates of cesarean section, meconium-stained fluid, or Apgar scores less than 7 at 5 minutes.⁴

Herbal preparations raise many questions—and some concerns

A prospective controlled trial that randomized 192 women to receive either red raspberry leaf tablets (1.2 g twice daily beginning at 32 weeks) or placebo found no significant differences in duration of the first stage of labor or birth outcomes. Symptom surveys of 108 women revealed no adverse effects.⁵

A retrospective cohort study compared pregnancy outcomes in 54 women who took evening primrose oil (500 mg 3 times a day beginning at 37 weeks) with a matched group of 54 women who did not take it. The study found no significant differences between the groups in maternal age, Apgar scores, or days of gestation. However, the women taking evening primrose oil tended to have protracted active phase of labor, prolonged rupture of membranes, arrest of descent, and oxytocin augmentation (none of which were statistically significant).⁶

A systematic review of the literature on black cohosh in pregnancy found no trials that evaluated its efficacy for inducing labor.⁷ A review of herbal preparations used by midwives reported a case of an infant with low Apgar scores after black cohosh use.⁸ Both articles described black cohosh as potentially unsafe because of the lack of trials demonstrating safety and possible deleterious estrogenic effects.

A systematic review of the literature on blue cohosh found only in vitro studies of efficacy (increased estradiol-induced transcription in estrogen-responsive cells and increased tone in excised guinea pig uteri) and 3 case reports of maternal adverse events after ingestion (perinatal stroke, congestive heart failure with shock, and multiorgan hypoxic injury).⁹ In vitro evidence presented in the review suggests that blue cohosh may have teratogenic, embryotoxic, and oxytocic effects.

Recommendations

We found no recommendations from professional organizations regarding folk methods to stimulate labor. The authors of a survey of herbal preparations used by midwives do not advocate their use in pregnancy because of unknown risks to the fetus, although “raspberry leaf tea does not seem to have any significant pharmacologic activity and is probably safe.”⁸

Evidence summary

Breast stimulation works

A systematic review of 6 trials with a total of 719 participants compared unilateral breast stimulation with no intervention to induce labor in women at term.¹ In 2 trials, the 300 participants stimulated 1 breast for 1 hour a day for 3 days; in the other 4 trials, 419 women stimulated either breast alternately for 3 hours a day.

Breast stimulation significantly reduced the number of women who hadn’t gone into labor at 72 hours compared with nonintervention (62.7% vs 93.6%; relative risk [RR]=0.67; 95% confidence interval [CI], 0.60-0.74; number needed to treat [NNT]=3.2). This result remained significant when primiparous and multiparous women were analyzed separately, but not in the 1 trial (37 participants) that reported on women with an unfavorable cervix.¹ Breast stimulation also reduced postpartum hemorrhage (0.7% vs 6%; RR=0.16; 95% CI, 0.03-0.87; NNT=18.8).

None of the trials documented uterine hyperstimulation or meconium-stained fluid, and they didn’t find significant differences in the rates of cesarean section (9% vs 10%; RR=0.90; 95% CI, 0.38-2.12). One trial reported 4 perinatal deaths in a high-risk population—3 in the breast stimulation group and 1 in the control group—but this finding was not replicated in any other trials.¹

Acupuncture may also help

A Cochrane systematic review evaluating acupuncture for inducing labor identified 1 randomized trial with methodologic flaws (allowing no conclusion), 3 case series, and 2 nonrandomized trials. The first case series used electroacupuncture at 38 to 42 weeks to successfully induce labor in 21 of 31 women. The second series, using acupuncture with and without electrical stimulation, induced labor in 10 of 12 women at 19 to 43 weeks. The third study induced labor with electroacupuncture in 78% of 41 women—34 term and postterm patients and 7 with a fetal demise.

In the first nonrandomized trial, 20 of 27 women at term who received electroacupuncture gave birth between 39 weeks 3 days and 40 weeks, compared with 47 of 102 women in the control group. In the second study, 31 of 35 women who received electroacupuncture reported increased intensity of contractions compared with none of the 35 women in the control group.²

Sexual intercourse induces pregnancy, not labor

A Cochrane systematic review found 1 observational study of the effect of sexual intercourse on cervical ripening and induction of labor among 28 women at term. The study evaluated cervical Bishop scores after sexual intercourse with intravaginal semen deposition for 3 consecutive nights compared with no intercourse. The difference in Bishop score was not significant (1.0 with coitus vs 0.5 controls; P>.05) and no difference in the number of women who delivered within 3 days was noted in the 2 groups (RR=0.99; 95% CI, 0.45-2.20).³

Castor oil made women nauseous

A Cochrane systematic review included 1 study of women at term with intact membranes who received 1 dose of castor oil to induce labor (n=52) compared with no treatment (n=48). All women who received castor oil felt nauseous, but no other differences were noted between the castor oil group and controls in rates of cesarean section, meconium-stained fluid, or Apgar scores less than 7 at 5 minutes.⁴

Herbal preparations raise many questions—and some concerns

A prospective controlled trial that randomized 192 women to receive either red raspberry leaf tablets (1.2 g twice daily beginning at 32 weeks) or placebo found no significant differences in duration of the first stage of labor or birth outcomes. Symptom surveys of 108 women revealed no adverse effects.⁵

A retrospective cohort study compared pregnancy outcomes in 54 women who took evening primrose oil (500 mg 3 times a day beginning at 37 weeks) with a matched group of 54 women who did not take it. The study found no significant differences between the groups in maternal age, Apgar scores, or days of gestation. However, the women taking evening primrose oil tended to have protracted active phase of labor, prolonged rupture of membranes, arrest of descent, and oxytocin augmentation (none of which were statistically significant).⁶

A systematic review of the literature on black cohosh in pregnancy found no trials that evaluated its efficacy for inducing labor.⁷ A review of herbal preparations used by midwives reported a case of an infant with low Apgar scores after black cohosh use.⁸ Both articles described black cohosh as potentially unsafe because of the lack of trials demonstrating safety and possible deleterious estrogenic effects.

A systematic review of the literature on blue cohosh found only in vitro studies of efficacy (increased estradiol-induced transcription in estrogen-responsive cells and increased tone in excised guinea pig uteri) and 3 case reports of maternal adverse events after ingestion (perinatal stroke, congestive heart failure with shock, and multiorgan hypoxic injury).⁹ In vitro evidence presented in the review suggests that blue cohosh may have teratogenic, embryotoxic, and oxytocic effects.

Recommendations

We found no recommendations from professional organizations regarding folk methods to stimulate labor. The authors of a survey of herbal preparations used by midwives do not advocate their use in pregnancy because of unknown risks to the fetus, although “raspberry leaf tea does not seem to have any significant pharmacologic activity and is probably safe.”⁸

Evidence summary

Few short-term joint complaints, no long-term skeletal harm

A 1997 database review compiled reports of skeletally immature patients ranging in age from 4 days to 26 years who were exposed to quinolones.¹ Thirty-one reports met search term criteria, for a total of 7045 patients. No incidences of quinolone-associated arthralgia were documented in 30 reports (>5000 patients). The review didn’t report the incidence of tendonopathy. One report of 1795 pediatric patients documented a small incidence of arthralgias (~1.5%), which was considered to be reversible and no more than expected for a comparable quinolone-naïve population.

Follow-up data on safety and adverse findings, from as long as 12 years after treatment, were reported for 530 (28%) of the 7045 patients. Changes in skeletal growth were evaluated using various diagnostic techniques. Clinical observation was the most common method of assessment (N=357), however. The follow-up data revealed no arthropathy or abnormal skeletal growth (rate=0%; estimated 95% confidence interval [CI]=0%-0.04%).

A prospective study published in 2006 monitored joint toxicities (swelling, tenderness, or restricted movement) during acute treatment with ciprofloxacin as well as skeletal growth at follow-up based on physical examination.² Preterm neonates with septicemia were treated with either ciprofloxacin (n=48) or other antibiotics (n=66). Forty infants in the ciprofloxacin group completed an average of 28 months of follow-up. No complaints or physical findings of osteoarticular joint abnormalities or skeletal growth delay were noted in either group during acute treatment or at follow-up. The incidence of tendonopathy was not reported.

Arthralgias, myalgias are transient

A large multicenter, prospective, non-blinded cohort study evaluated adverse effects in children receiving fluoroquinolones versus other antibiotics.³ Duration of fluoroquinolone use was 1 to 23 days. Arthralgias or myalgias, which were only evaluated clinically, occurred more often in children receiving fluoroquinolones—10 of 276 children (3.6%) vs 1 of 249 (0.3%), respectively (odds ratio [OR]=9.3; 95% CI, 1.2-195; P=.02). All events occurred within the first 2 weeks of fluoroquinolone treatment and resolved within 20 days. No tendonopathies were reported.

Tendon rupture is rare, especially in children

A 1996 study reported the incidence of tendon disorders related to fluoroquinolones using drug surveillance data from the general population. The average age of the patients was 55 years.⁴

The author estimated the risk of tendon rupture associated with norfloxacin or ofloxacin to be 1 case per 23,130 days of treatment and only 1 case per 779,600 days of ciprofloxacin treatment. The estimated risk would likely be even lower in children, the author noted, because the risk of tendon rupture increases with age.

Recommendations

Ciprofloxacin is the only fluoroquinolone approved by the US Food and Drug Administration for pediatric indications. The FDA recently ordered the addition of a Boxed Warning to fluoroquinolones regarding the increased risk of tendonitis and tendon rupture. The FDA made no comments specifically about children or adolescents, and stated that the risks are increased in people older than 60.

The American Academy of Pediatrics recommends limiting fluoroquinolone use to children with infections caused by multidrug-resistant pathogens or children for whom parenteral therapy is not feasible and no other effective oral medication is available.⁵

The Agency for Healthcare Research and Quality (AHRQ) recommends fluoroquinolones as first-line treatment for children with uncomplicated gonorrhea who weigh more than 45 kg,⁶ and second-line therapy for children with bacterial meningitis,⁷ nongonococcal urethritis, chlamydia,⁶ or pelvic inflammatory disease.⁸

Acknowledgments

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the US Government.

Evidence summary

Few short-term joint complaints, no long-term skeletal harm

A 1997 database review compiled reports of skeletally immature patients ranging in age from 4 days to 26 years who were exposed to quinolones.¹ Thirty-one reports met search term criteria, for a total of 7045 patients. No incidences of quinolone-associated arthralgia were documented in 30 reports (>5000 patients). The review didn’t report the incidence of tendonopathy. One report of 1795 pediatric patients documented a small incidence of arthralgias (~1.5%), which was considered to be reversible and no more than expected for a comparable quinolone-naïve population.

Follow-up data on safety and adverse findings, from as long as 12 years after treatment, were reported for 530 (28%) of the 7045 patients. Changes in skeletal growth were evaluated using various diagnostic techniques. Clinical observation was the most common method of assessment (N=357), however. The follow-up data revealed no arthropathy or abnormal skeletal growth (rate=0%; estimated 95% confidence interval [CI]=0%-0.04%).

A prospective study published in 2006 monitored joint toxicities (swelling, tenderness, or restricted movement) during acute treatment with ciprofloxacin as well as skeletal growth at follow-up based on physical examination.² Preterm neonates with septicemia were treated with either ciprofloxacin (n=48) or other antibiotics (n=66). Forty infants in the ciprofloxacin group completed an average of 28 months of follow-up. No complaints or physical findings of osteoarticular joint abnormalities or skeletal growth delay were noted in either group during acute treatment or at follow-up. The incidence of tendonopathy was not reported.

Arthralgias, myalgias are transient

A large multicenter, prospective, non-blinded cohort study evaluated adverse effects in children receiving fluoroquinolones versus other antibiotics.³ Duration of fluoroquinolone use was 1 to 23 days. Arthralgias or myalgias, which were only evaluated clinically, occurred more often in children receiving fluoroquinolones—10 of 276 children (3.6%) vs 1 of 249 (0.3%), respectively (odds ratio [OR]=9.3; 95% CI, 1.2-195; P=.02). All events occurred within the first 2 weeks of fluoroquinolone treatment and resolved within 20 days. No tendonopathies were reported.

Tendon rupture is rare, especially in children

A 1996 study reported the incidence of tendon disorders related to fluoroquinolones using drug surveillance data from the general population. The average age of the patients was 55 years.⁴

The author estimated the risk of tendon rupture associated with norfloxacin or ofloxacin to be 1 case per 23,130 days of treatment and only 1 case per 779,600 days of ciprofloxacin treatment. The estimated risk would likely be even lower in children, the author noted, because the risk of tendon rupture increases with age.

Recommendations

Ciprofloxacin is the only fluoroquinolone approved by the US Food and Drug Administration for pediatric indications. The FDA recently ordered the addition of a Boxed Warning to fluoroquinolones regarding the increased risk of tendonitis and tendon rupture. The FDA made no comments specifically about children or adolescents, and stated that the risks are increased in people older than 60.

The American Academy of Pediatrics recommends limiting fluoroquinolone use to children with infections caused by multidrug-resistant pathogens or children for whom parenteral therapy is not feasible and no other effective oral medication is available.⁵

The Agency for Healthcare Research and Quality (AHRQ) recommends fluoroquinolones as first-line treatment for children with uncomplicated gonorrhea who weigh more than 45 kg,⁶ and second-line therapy for children with bacterial meningitis,⁷ nongonococcal urethritis, chlamydia,⁶ or pelvic inflammatory disease.⁸

Acknowledgments

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the US Government.

Evidence summary

Few short-term joint complaints, no long-term skeletal harm

A 1997 database review compiled reports of skeletally immature patients ranging in age from 4 days to 26 years who were exposed to quinolones.¹ Thirty-one reports met search term criteria, for a total of 7045 patients. No incidences of quinolone-associated arthralgia were documented in 30 reports (>5000 patients). The review didn’t report the incidence of tendonopathy. One report of 1795 pediatric patients documented a small incidence of arthralgias (~1.5%), which was considered to be reversible and no more than expected for a comparable quinolone-naïve population.

Follow-up data on safety and adverse findings, from as long as 12 years after treatment, were reported for 530 (28%) of the 7045 patients. Changes in skeletal growth were evaluated using various diagnostic techniques. Clinical observation was the most common method of assessment (N=357), however. The follow-up data revealed no arthropathy or abnormal skeletal growth (rate=0%; estimated 95% confidence interval [CI]=0%-0.04%).

A prospective study published in 2006 monitored joint toxicities (swelling, tenderness, or restricted movement) during acute treatment with ciprofloxacin as well as skeletal growth at follow-up based on physical examination.² Preterm neonates with septicemia were treated with either ciprofloxacin (n=48) or other antibiotics (n=66). Forty infants in the ciprofloxacin group completed an average of 28 months of follow-up. No complaints or physical findings of osteoarticular joint abnormalities or skeletal growth delay were noted in either group during acute treatment or at follow-up. The incidence of tendonopathy was not reported.

Arthralgias, myalgias are transient

A large multicenter, prospective, non-blinded cohort study evaluated adverse effects in children receiving fluoroquinolones versus other antibiotics.³ Duration of fluoroquinolone use was 1 to 23 days. Arthralgias or myalgias, which were only evaluated clinically, occurred more often in children receiving fluoroquinolones—10 of 276 children (3.6%) vs 1 of 249 (0.3%), respectively (odds ratio [OR]=9.3; 95% CI, 1.2-195; P=.02). All events occurred within the first 2 weeks of fluoroquinolone treatment and resolved within 20 days. No tendonopathies were reported.

Tendon rupture is rare, especially in children

A 1996 study reported the incidence of tendon disorders related to fluoroquinolones using drug surveillance data from the general population. The average age of the patients was 55 years.⁴

The author estimated the risk of tendon rupture associated with norfloxacin or ofloxacin to be 1 case per 23,130 days of treatment and only 1 case per 779,600 days of ciprofloxacin treatment. The estimated risk would likely be even lower in children, the author noted, because the risk of tendon rupture increases with age.

Recommendations

Ciprofloxacin is the only fluoroquinolone approved by the US Food and Drug Administration for pediatric indications. The FDA recently ordered the addition of a Boxed Warning to fluoroquinolones regarding the increased risk of tendonitis and tendon rupture. The FDA made no comments specifically about children or adolescents, and stated that the risks are increased in people older than 60.

The American Academy of Pediatrics recommends limiting fluoroquinolone use to children with infections caused by multidrug-resistant pathogens or children for whom parenteral therapy is not feasible and no other effective oral medication is available.⁵

The Agency for Healthcare Research and Quality (AHRQ) recommends fluoroquinolones as first-line treatment for children with uncomplicated gonorrhea who weigh more than 45 kg,⁶ and second-line therapy for children with bacterial meningitis,⁷ nongonococcal urethritis, chlamydia,⁶ or pelvic inflammatory disease.⁸

Acknowledgments

The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the US Government.

Evidence summary

Exercise therapy for osteoarthritis aims at reducing pain and disability by improving muscle strength, range of motion, joint stability, and fitness.^1,2 A Cochrane systematic review of hip and knee osteoarthritis³ defined exercise therapy as activities performed actively, passively, or against resistance. The outcomes were reported as effect sizes (ES): Typically, an ES of 0.2 is considered small, 0.5 medium, and 0.8 large.

Exercise, regardless of intensity, may ease pain of mild disease

Land-based therapeutic exercise demonstrated a combined treatment effect of 0.39 (95% confidence interval [CI], 0.30-0.47) for reducing pain. Individual treatments (ES=0.52; 95% CI, 0.32-0.72) or exercise classes (ES=0.47; 95% CI, 0.34-0.60) had a larger effect than home-based programs (ES=0.28; 95% CI, 0.16-0.40). However, these studies only included patients with early or mildly symptomatic osteoarthritis, which may limit their applicability to more severe disease.

Another systematic review examined the effect of therapeutic exercise on hip and knee osteoarthritis.⁴ The standardized mean difference in ES for pain reduction was moderate, 0.46 (95% CI, 0.35-0.57). For self-reported physical functioning, the standardized mean difference was small, 0.33 (95% CI, 0.23-0.43). Study limitations included enrollment of patients with only early or mild osteoarthritis, low numbers of intent-to-treat studies, short-term nature of the studies, and inadequate sample sizes.

Advanced osteoarthritis

Both aerobic and strengthening exercises seem to help

A 2005 systematic review examined aerobic walking and quadriceps strengthening for knee osteoarthritis using pain as the primary outcome and self-reported disability as the secondary outcome.⁶ In the exercise group, the weighted, pooled ES for pain reduction and self-reported disability was moderate at 0.52 (95% CI, 0.34-0.70) and 0.46 (95% CI, 0.25-0.67) respectively. The quadriceps strengthening group showed a small weighted, pooled ES for pain reduction of 0.32 (95% CI, 0.23-0.42) and disability of 0.32 (95% CI, 0.23-0.41); pain and disability measures were variable. None of the included clinical trials were blinded.

Two other reviews examined strengthening and aerobic exercises for osteoarthritis, including studies using isokinetic, concentric, and eccentric strengthening; water therapy; walking; yoga; and tai chi.^7,8 The exercises appeared to have beneficial effects on pain. However, results from the studies of each type of exercise weren’t combined and ES wasn’t calculated.

No reviews assessed the effect of exercise on reducing stiffness.

Recommendations

The American College of Rheumatology recommends that nonpharmacologic therapy for osteoarthritis include education, aerobic exercise programs, physical therapy with range of motion, and muscle strengthening.⁹ The American Geriatric Society and the Centers for Disease Control and Prevention also recommend regular exercise.^10,11

Evidence summary

Exercise therapy for osteoarthritis aims at reducing pain and disability by improving muscle strength, range of motion, joint stability, and fitness.^1,2 A Cochrane systematic review of hip and knee osteoarthritis³ defined exercise therapy as activities performed actively, passively, or against resistance. The outcomes were reported as effect sizes (ES): Typically, an ES of 0.2 is considered small, 0.5 medium, and 0.8 large.

Exercise, regardless of intensity, may ease pain of mild disease

Land-based therapeutic exercise demonstrated a combined treatment effect of 0.39 (95% confidence interval [CI], 0.30-0.47) for reducing pain. Individual treatments (ES=0.52; 95% CI, 0.32-0.72) or exercise classes (ES=0.47; 95% CI, 0.34-0.60) had a larger effect than home-based programs (ES=0.28; 95% CI, 0.16-0.40). However, these studies only included patients with early or mildly symptomatic osteoarthritis, which may limit their applicability to more severe disease.

Another systematic review examined the effect of therapeutic exercise on hip and knee osteoarthritis.⁴ The standardized mean difference in ES for pain reduction was moderate, 0.46 (95% CI, 0.35-0.57). For self-reported physical functioning, the standardized mean difference was small, 0.33 (95% CI, 0.23-0.43). Study limitations included enrollment of patients with only early or mild osteoarthritis, low numbers of intent-to-treat studies, short-term nature of the studies, and inadequate sample sizes.

Advanced osteoarthritis

Both aerobic and strengthening exercises seem to help

A 2005 systematic review examined aerobic walking and quadriceps strengthening for knee osteoarthritis using pain as the primary outcome and self-reported disability as the secondary outcome.⁶ In the exercise group, the weighted, pooled ES for pain reduction and self-reported disability was moderate at 0.52 (95% CI, 0.34-0.70) and 0.46 (95% CI, 0.25-0.67) respectively. The quadriceps strengthening group showed a small weighted, pooled ES for pain reduction of 0.32 (95% CI, 0.23-0.42) and disability of 0.32 (95% CI, 0.23-0.41); pain and disability measures were variable. None of the included clinical trials were blinded.

Two other reviews examined strengthening and aerobic exercises for osteoarthritis, including studies using isokinetic, concentric, and eccentric strengthening; water therapy; walking; yoga; and tai chi.^7,8 The exercises appeared to have beneficial effects on pain. However, results from the studies of each type of exercise weren’t combined and ES wasn’t calculated.

No reviews assessed the effect of exercise on reducing stiffness.

Recommendations

The American College of Rheumatology recommends that nonpharmacologic therapy for osteoarthritis include education, aerobic exercise programs, physical therapy with range of motion, and muscle strengthening.⁹ The American Geriatric Society and the Centers for Disease Control and Prevention also recommend regular exercise.^10,11

Evidence summary

Exercise therapy for osteoarthritis aims at reducing pain and disability by improving muscle strength, range of motion, joint stability, and fitness.^1,2 A Cochrane systematic review of hip and knee osteoarthritis³ defined exercise therapy as activities performed actively, passively, or against resistance. The outcomes were reported as effect sizes (ES): Typically, an ES of 0.2 is considered small, 0.5 medium, and 0.8 large.

Exercise, regardless of intensity, may ease pain of mild disease

Land-based therapeutic exercise demonstrated a combined treatment effect of 0.39 (95% confidence interval [CI], 0.30-0.47) for reducing pain. Individual treatments (ES=0.52; 95% CI, 0.32-0.72) or exercise classes (ES=0.47; 95% CI, 0.34-0.60) had a larger effect than home-based programs (ES=0.28; 95% CI, 0.16-0.40). However, these studies only included patients with early or mildly symptomatic osteoarthritis, which may limit their applicability to more severe disease.

Another systematic review examined the effect of therapeutic exercise on hip and knee osteoarthritis.⁴ The standardized mean difference in ES for pain reduction was moderate, 0.46 (95% CI, 0.35-0.57). For self-reported physical functioning, the standardized mean difference was small, 0.33 (95% CI, 0.23-0.43). Study limitations included enrollment of patients with only early or mild osteoarthritis, low numbers of intent-to-treat studies, short-term nature of the studies, and inadequate sample sizes.

Advanced osteoarthritis

Both aerobic and strengthening exercises seem to help

A 2005 systematic review examined aerobic walking and quadriceps strengthening for knee osteoarthritis using pain as the primary outcome and self-reported disability as the secondary outcome.⁶ In the exercise group, the weighted, pooled ES for pain reduction and self-reported disability was moderate at 0.52 (95% CI, 0.34-0.70) and 0.46 (95% CI, 0.25-0.67) respectively. The quadriceps strengthening group showed a small weighted, pooled ES for pain reduction of 0.32 (95% CI, 0.23-0.42) and disability of 0.32 (95% CI, 0.23-0.41); pain and disability measures were variable. None of the included clinical trials were blinded.

Two other reviews examined strengthening and aerobic exercises for osteoarthritis, including studies using isokinetic, concentric, and eccentric strengthening; water therapy; walking; yoga; and tai chi.^7,8 The exercises appeared to have beneficial effects on pain. However, results from the studies of each type of exercise weren’t combined and ES wasn’t calculated.

No reviews assessed the effect of exercise on reducing stiffness.

Recommendations

The American College of Rheumatology recommends that nonpharmacologic therapy for osteoarthritis include education, aerobic exercise programs, physical therapy with range of motion, and muscle strengthening.⁹ The American Geriatric Society and the Centers for Disease Control and Prevention also recommend regular exercise.^10,11

Evidence summary

Few studies have compared treatment options for oropharyngeal candidiasis in immunocompetent infants. In a survey of 312 health care providers, approximately 75% of the respondents reported treating thrush with oral nystatin, citing fewer side effects and lower cost.¹ However, nystatin has proved less effective than either miconazole gel or oral fluconazole.

Nystatin is safe and available, but other options work better

Miconazole vs nystatin. An unblinded RCT assigned 83 immunocompetent infants with culture-positive oral thrush to receive either 25 mg miconazole oral gel (not commercially available in the United States) or nystatin suspension (1 mL of 100,000 IU/mL) qid after meals. The clinical cure rate, defined as absence of plaques by day 12, was significantly higher in the miconazole group (99% for miconazole, 54% for nystatin; P<.0001, number needed to treat [NNT]=2). The eradication rate, confirmed by cultures collected in a blinded manner on the day of clinical cure, was also higher in the miconazole group (55.7% for miconazole, 15.2% for nystatin; P<.0001, NNT=3). In successfully treated patients, infection recurred with similar frequency in both treatment groups within 4 weeks (miconazole, 12.4%; nystatin, 13.0%). Side effects—mostly vomiting and, infrequently, diarrhea—were rare in both groups (miconazole, 4.5%; nystatin 3.5%).²

An earlier, unblinded RCT of 95 infants compared miconazole gel to 2 nystatin oral gels (gel A: 250,000 IU/g with 250,000 IU administered as single dose; gel B: 100,000 IU/g with 50,000 IU administered as single dose). Each medication was given qid over the course of 8 to 14 days. The study confirmed higher clinical cure rates with miconazole gel (85.1% for miconazole vs 42.8% for nystatin gel A [P<.0007, NNT=2] and 48.5% for nystatin gel B [P<.004, NNT=3]).³

Fluconazole vs nystatin. In the only prospective RCT (unblinded) to compare oral suspensions of fluconazole and nystatin, 34 infants were randomized to receive either nystatin (1 mL of 100,000 IU/mL) qid for 10 days or fluconazole (3 mg/kg) once a day for 7 days. Mothers of breastfed infants applied nystatin cream to their nipples twice a day for the duration of the infant’s treatment. The clinical cure rate—defined as absence of oral plaques at the end of therapy (day 10 for the nystatin group, day 7 for the fluconazole group)—was significantly higher in the group treated with fluconazole (100% for fluconazole, 32% for nystatin; P<.0001, NNT=2). The eradication rate was also higher with fluconazole (73.3% for fluconazole, 5.6% for nystatin; P<.0001, NNT=2). The patients treated with fluconazole experienced no side effects.⁴ Fluconazole has been shown to be effective, safe, and easy to use to treat thrush in immunocompromised children,⁵ but has not been approved by the FDA for use in healthy infants.

Gentian violet is effective, but messy and irritating

A retrospective cohort study that reviewed 69 cases of oral thrush showed that gentian violet achieved a 75% cure rate in an average of 11 days (compared to 55% in 10 days for nystatin). Both treatments shortened the duration of illness compared with the average of 34 days for untreated children.⁶ However, gentian violet can stain skin and clothes, and case studies have shown an association with ulceration of the buccal mucosa.⁷

Recommendations

A thorough literature search through the Cochrane Database Systematic Reviews, Agency for Healthcare Research and Quality, National Guideline Clearinghouse, and Medline did not yield any guidelines or consensus statements from other organizations or specialty groups on treating oropharyngeal candidiasis in infants. Neither the American Academy of Pediatrics nor the Infectious Diseases Society of America has issued applicable practice guidelines.

Evidence summary

Few studies have compared treatment options for oropharyngeal candidiasis in immunocompetent infants. In a survey of 312 health care providers, approximately 75% of the respondents reported treating thrush with oral nystatin, citing fewer side effects and lower cost.¹ However, nystatin has proved less effective than either miconazole gel or oral fluconazole.

Nystatin is safe and available, but other options work better

Miconazole vs nystatin. An unblinded RCT assigned 83 immunocompetent infants with culture-positive oral thrush to receive either 25 mg miconazole oral gel (not commercially available in the United States) or nystatin suspension (1 mL of 100,000 IU/mL) qid after meals. The clinical cure rate, defined as absence of plaques by day 12, was significantly higher in the miconazole group (99% for miconazole, 54% for nystatin; P<.0001, number needed to treat [NNT]=2). The eradication rate, confirmed by cultures collected in a blinded manner on the day of clinical cure, was also higher in the miconazole group (55.7% for miconazole, 15.2% for nystatin; P<.0001, NNT=3). In successfully treated patients, infection recurred with similar frequency in both treatment groups within 4 weeks (miconazole, 12.4%; nystatin, 13.0%). Side effects—mostly vomiting and, infrequently, diarrhea—were rare in both groups (miconazole, 4.5%; nystatin 3.5%).²

An earlier, unblinded RCT of 95 infants compared miconazole gel to 2 nystatin oral gels (gel A: 250,000 IU/g with 250,000 IU administered as single dose; gel B: 100,000 IU/g with 50,000 IU administered as single dose). Each medication was given qid over the course of 8 to 14 days. The study confirmed higher clinical cure rates with miconazole gel (85.1% for miconazole vs 42.8% for nystatin gel A [P<.0007, NNT=2] and 48.5% for nystatin gel B [P<.004, NNT=3]).³

Fluconazole vs nystatin. In the only prospective RCT (unblinded) to compare oral suspensions of fluconazole and nystatin, 34 infants were randomized to receive either nystatin (1 mL of 100,000 IU/mL) qid for 10 days or fluconazole (3 mg/kg) once a day for 7 days. Mothers of breastfed infants applied nystatin cream to their nipples twice a day for the duration of the infant’s treatment. The clinical cure rate—defined as absence of oral plaques at the end of therapy (day 10 for the nystatin group, day 7 for the fluconazole group)—was significantly higher in the group treated with fluconazole (100% for fluconazole, 32% for nystatin; P<.0001, NNT=2). The eradication rate was also higher with fluconazole (73.3% for fluconazole, 5.6% for nystatin; P<.0001, NNT=2). The patients treated with fluconazole experienced no side effects.⁴ Fluconazole has been shown to be effective, safe, and easy to use to treat thrush in immunocompromised children,⁵ but has not been approved by the FDA for use in healthy infants.

Gentian violet is effective, but messy and irritating

A retrospective cohort study that reviewed 69 cases of oral thrush showed that gentian violet achieved a 75% cure rate in an average of 11 days (compared to 55% in 10 days for nystatin). Both treatments shortened the duration of illness compared with the average of 34 days for untreated children.⁶ However, gentian violet can stain skin and clothes, and case studies have shown an association with ulceration of the buccal mucosa.⁷

Recommendations

A thorough literature search through the Cochrane Database Systematic Reviews, Agency for Healthcare Research and Quality, National Guideline Clearinghouse, and Medline did not yield any guidelines or consensus statements from other organizations or specialty groups on treating oropharyngeal candidiasis in infants. Neither the American Academy of Pediatrics nor the Infectious Diseases Society of America has issued applicable practice guidelines.

Evidence summary

Few studies have compared treatment options for oropharyngeal candidiasis in immunocompetent infants. In a survey of 312 health care providers, approximately 75% of the respondents reported treating thrush with oral nystatin, citing fewer side effects and lower cost.¹ However, nystatin has proved less effective than either miconazole gel or oral fluconazole.

Nystatin is safe and available, but other options work better

Miconazole vs nystatin. An unblinded RCT assigned 83 immunocompetent infants with culture-positive oral thrush to receive either 25 mg miconazole oral gel (not commercially available in the United States) or nystatin suspension (1 mL of 100,000 IU/mL) qid after meals. The clinical cure rate, defined as absence of plaques by day 12, was significantly higher in the miconazole group (99% for miconazole, 54% for nystatin; P<.0001, number needed to treat [NNT]=2). The eradication rate, confirmed by cultures collected in a blinded manner on the day of clinical cure, was also higher in the miconazole group (55.7% for miconazole, 15.2% for nystatin; P<.0001, NNT=3). In successfully treated patients, infection recurred with similar frequency in both treatment groups within 4 weeks (miconazole, 12.4%; nystatin, 13.0%). Side effects—mostly vomiting and, infrequently, diarrhea—were rare in both groups (miconazole, 4.5%; nystatin 3.5%).²

An earlier, unblinded RCT of 95 infants compared miconazole gel to 2 nystatin oral gels (gel A: 250,000 IU/g with 250,000 IU administered as single dose; gel B: 100,000 IU/g with 50,000 IU administered as single dose). Each medication was given qid over the course of 8 to 14 days. The study confirmed higher clinical cure rates with miconazole gel (85.1% for miconazole vs 42.8% for nystatin gel A [P<.0007, NNT=2] and 48.5% for nystatin gel B [P<.004, NNT=3]).³

Fluconazole vs nystatin. In the only prospective RCT (unblinded) to compare oral suspensions of fluconazole and nystatin, 34 infants were randomized to receive either nystatin (1 mL of 100,000 IU/mL) qid for 10 days or fluconazole (3 mg/kg) once a day for 7 days. Mothers of breastfed infants applied nystatin cream to their nipples twice a day for the duration of the infant’s treatment. The clinical cure rate—defined as absence of oral plaques at the end of therapy (day 10 for the nystatin group, day 7 for the fluconazole group)—was significantly higher in the group treated with fluconazole (100% for fluconazole, 32% for nystatin; P<.0001, NNT=2). The eradication rate was also higher with fluconazole (73.3% for fluconazole, 5.6% for nystatin; P<.0001, NNT=2). The patients treated with fluconazole experienced no side effects.⁴ Fluconazole has been shown to be effective, safe, and easy to use to treat thrush in immunocompromised children,⁵ but has not been approved by the FDA for use in healthy infants.

Gentian violet is effective, but messy and irritating

A retrospective cohort study that reviewed 69 cases of oral thrush showed that gentian violet achieved a 75% cure rate in an average of 11 days (compared to 55% in 10 days for nystatin). Both treatments shortened the duration of illness compared with the average of 34 days for untreated children.⁶ However, gentian violet can stain skin and clothes, and case studies have shown an association with ulceration of the buccal mucosa.⁷

Recommendations

A thorough literature search through the Cochrane Database Systematic Reviews, Agency for Healthcare Research and Quality, National Guideline Clearinghouse, and Medline did not yield any guidelines or consensus statements from other organizations or specialty groups on treating oropharyngeal candidiasis in infants. Neither the American Academy of Pediatrics nor the Infectious Diseases Society of America has issued applicable practice guidelines.

Evidence summary

Breast augmentation is one of the most popular plastic surgeries in the United States; an estimated 291,350 such procedures were performed in 2005.¹ Breast cancer occurs in 1 of every 8 women; a projected 32,000 women who received breast implants in 2003 will develop cancer.¹ Available research has focused on retrospective and prospective designs because of the ethical limitations of experimental designs. No US studies directly compare mammography with alternate screening methods, such as sonography or magnetic resonance imaging.

With implants: Lower screening sensitivity but similar prognosis

Studies show that augmentation decreases the sensitivity of screening mammography but doesn’t affect breast cancer prognosis.² A 2004 prospective cohort study of 986,270 women found that, among asymptomatic women diagnosed with breast cancer (40 augmented, 238 nonaugmented), the sensitivity of screening mammograms was lower in women with breast implants (45%; 95% confidence interval [CI], 29.3%–61.5%) than those without (66.8%; 95% CI, 60.4%–72.8%); P=.008).² Similarly, in symptomatic women diagnosed with breast cancer (41 augmented, 145 nonaugmented), screening sensitivity was lower in the augmented women (73.2%) than the nonaugmented women (81.4%)—although the results weren’t significant (P=.25).

Despite lower screening sensitivity, breast tumors in asymptomatic women, whether augmented or not, had similar characteristics, except for larger tumor size (3 mm) at diagnosis in augmented women. Symptomatic women with breast implants had cancers that were smaller, lower-grade, and more likely to be estrogen receptor dependent and invasive (P=.052) compared with nonaugmented women. The authors concluded that augmentation doesn’t influence the prognostic characteristics of tumors, and they recommended screening mammography at appropriate intervals.

Two other prospective cohort studies produced similar findings. A 2006 study of 4082 breast cancer patients concluded that mammography yielded a false-negative rate of 41.4% in augmented patients compared with 8.8% in nonaugmented patients (P<.0001).³ However, both augmented (n=129) and nonaugmented (n=3953) women had a comparable prognosis at diagnosis. The authors of the studies suggested diagnostic mammography for augmented patients and correlation with physical exam findings.

An earlier study of 2956 cancer patients found that mammography detected an abnormal breast mass in 66.3% of augmented women compared with 94.6% of nonaugmented women (P=.001).⁴ No significant differences were noted in cancer characteristics at diagnosis or survival rates (P=.78). The authors of this study concluded that mammography should be used for augmented women until a more effective screening tool is found.

Sonography vs mammography: The jury is still out

Although studies comparing screening methods have not been performed in the United States, a small Taiwanese study directly compared ultrasound to mammography in 105 women without breast implants. This retrospective cohort study found sonography to be a more useful diagnostic tool than mammography in Taiwanese women.⁵ Sonography had the highest sensitivity (87.5%) compared to physical examination (50.0%) and mammography (25.5%).

Sonography was recommended as the imaging tool for Asian women with smaller, denser breasts. However, it is unclear whether this result applies to US women or women who have undergone breast augmentation surgery.

Training in implant imaging is needed

Mammography appears to be the most effective screening method for women with breast implants. Despite the small differences in cancer characteristics at diagnosis between augmented and nonaugmented women, overall prognosis and survival rates are no different.^1-3,6 This is true whether a screening mammogram or diagnostic mammogram is used. In any case, all available findings suggest that clinicians who perform mammography should be trained in imaging the augmented breast.^6-8

Recommendations

The National Cancer Institute indicates that the best screening method for augmented women is mammography performed at a facility with employees trained in implant imaging.⁷ The American College of Radiology’s practice guidelines affirm that mammography is the best imaging tool available.⁸ The American College of Obstetrics and Gynecology and the US Preventive Services Task Force don’t comment on screening augmented women.

Acknowledgments

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

Evidence summary

Breast augmentation is one of the most popular plastic surgeries in the United States; an estimated 291,350 such procedures were performed in 2005.¹ Breast cancer occurs in 1 of every 8 women; a projected 32,000 women who received breast implants in 2003 will develop cancer.¹ Available research has focused on retrospective and prospective designs because of the ethical limitations of experimental designs. No US studies directly compare mammography with alternate screening methods, such as sonography or magnetic resonance imaging.

With implants: Lower screening sensitivity but similar prognosis

Studies show that augmentation decreases the sensitivity of screening mammography but doesn’t affect breast cancer prognosis.² A 2004 prospective cohort study of 986,270 women found that, among asymptomatic women diagnosed with breast cancer (40 augmented, 238 nonaugmented), the sensitivity of screening mammograms was lower in women with breast implants (45%; 95% confidence interval [CI], 29.3%–61.5%) than those without (66.8%; 95% CI, 60.4%–72.8%); P=.008).² Similarly, in symptomatic women diagnosed with breast cancer (41 augmented, 145 nonaugmented), screening sensitivity was lower in the augmented women (73.2%) than the nonaugmented women (81.4%)—although the results weren’t significant (P=.25).

Despite lower screening sensitivity, breast tumors in asymptomatic women, whether augmented or not, had similar characteristics, except for larger tumor size (3 mm) at diagnosis in augmented women. Symptomatic women with breast implants had cancers that were smaller, lower-grade, and more likely to be estrogen receptor dependent and invasive (P=.052) compared with nonaugmented women. The authors concluded that augmentation doesn’t influence the prognostic characteristics of tumors, and they recommended screening mammography at appropriate intervals.

Two other prospective cohort studies produced similar findings. A 2006 study of 4082 breast cancer patients concluded that mammography yielded a false-negative rate of 41.4% in augmented patients compared with 8.8% in nonaugmented patients (P<.0001).³ However, both augmented (n=129) and nonaugmented (n=3953) women had a comparable prognosis at diagnosis. The authors of the studies suggested diagnostic mammography for augmented patients and correlation with physical exam findings.

An earlier study of 2956 cancer patients found that mammography detected an abnormal breast mass in 66.3% of augmented women compared with 94.6% of nonaugmented women (P=.001).⁴ No significant differences were noted in cancer characteristics at diagnosis or survival rates (P=.78). The authors of this study concluded that mammography should be used for augmented women until a more effective screening tool is found.

Sonography vs mammography: The jury is still out

Although studies comparing screening methods have not been performed in the United States, a small Taiwanese study directly compared ultrasound to mammography in 105 women without breast implants. This retrospective cohort study found sonography to be a more useful diagnostic tool than mammography in Taiwanese women.⁵ Sonography had the highest sensitivity (87.5%) compared to physical examination (50.0%) and mammography (25.5%).

Sonography was recommended as the imaging tool for Asian women with smaller, denser breasts. However, it is unclear whether this result applies to US women or women who have undergone breast augmentation surgery.

Training in implant imaging is needed

Mammography appears to be the most effective screening method for women with breast implants. Despite the small differences in cancer characteristics at diagnosis between augmented and nonaugmented women, overall prognosis and survival rates are no different.^1-3,6 This is true whether a screening mammogram or diagnostic mammogram is used. In any case, all available findings suggest that clinicians who perform mammography should be trained in imaging the augmented breast.^6-8

Recommendations

The National Cancer Institute indicates that the best screening method for augmented women is mammography performed at a facility with employees trained in implant imaging.⁷ The American College of Radiology’s practice guidelines affirm that mammography is the best imaging tool available.⁸ The American College of Obstetrics and Gynecology and the US Preventive Services Task Force don’t comment on screening augmented women.

Acknowledgments

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

Evidence summary

Breast augmentation is one of the most popular plastic surgeries in the United States; an estimated 291,350 such procedures were performed in 2005.¹ Breast cancer occurs in 1 of every 8 women; a projected 32,000 women who received breast implants in 2003 will develop cancer.¹ Available research has focused on retrospective and prospective designs because of the ethical limitations of experimental designs. No US studies directly compare mammography with alternate screening methods, such as sonography or magnetic resonance imaging.

With implants: Lower screening sensitivity but similar prognosis

Studies show that augmentation decreases the sensitivity of screening mammography but doesn’t affect breast cancer prognosis.² A 2004 prospective cohort study of 986,270 women found that, among asymptomatic women diagnosed with breast cancer (40 augmented, 238 nonaugmented), the sensitivity of screening mammograms was lower in women with breast implants (45%; 95% confidence interval [CI], 29.3%–61.5%) than those without (66.8%; 95% CI, 60.4%–72.8%); P=.008).² Similarly, in symptomatic women diagnosed with breast cancer (41 augmented, 145 nonaugmented), screening sensitivity was lower in the augmented women (73.2%) than the nonaugmented women (81.4%)—although the results weren’t significant (P=.25).

Despite lower screening sensitivity, breast tumors in asymptomatic women, whether augmented or not, had similar characteristics, except for larger tumor size (3 mm) at diagnosis in augmented women. Symptomatic women with breast implants had cancers that were smaller, lower-grade, and more likely to be estrogen receptor dependent and invasive (P=.052) compared with nonaugmented women. The authors concluded that augmentation doesn’t influence the prognostic characteristics of tumors, and they recommended screening mammography at appropriate intervals.

Two other prospective cohort studies produced similar findings. A 2006 study of 4082 breast cancer patients concluded that mammography yielded a false-negative rate of 41.4% in augmented patients compared with 8.8% in nonaugmented patients (P<.0001).³ However, both augmented (n=129) and nonaugmented (n=3953) women had a comparable prognosis at diagnosis. The authors of the studies suggested diagnostic mammography for augmented patients and correlation with physical exam findings.

An earlier study of 2956 cancer patients found that mammography detected an abnormal breast mass in 66.3% of augmented women compared with 94.6% of nonaugmented women (P=.001).⁴ No significant differences were noted in cancer characteristics at diagnosis or survival rates (P=.78). The authors of this study concluded that mammography should be used for augmented women until a more effective screening tool is found.

Sonography vs mammography: The jury is still out

Although studies comparing screening methods have not been performed in the United States, a small Taiwanese study directly compared ultrasound to mammography in 105 women without breast implants. This retrospective cohort study found sonography to be a more useful diagnostic tool than mammography in Taiwanese women.⁵ Sonography had the highest sensitivity (87.5%) compared to physical examination (50.0%) and mammography (25.5%).

Sonography was recommended as the imaging tool for Asian women with smaller, denser breasts. However, it is unclear whether this result applies to US women or women who have undergone breast augmentation surgery.

Training in implant imaging is needed

Mammography appears to be the most effective screening method for women with breast implants. Despite the small differences in cancer characteristics at diagnosis between augmented and nonaugmented women, overall prognosis and survival rates are no different.^1-3,6 This is true whether a screening mammogram or diagnostic mammogram is used. In any case, all available findings suggest that clinicians who perform mammography should be trained in imaging the augmented breast.^6-8

Recommendations

The National Cancer Institute indicates that the best screening method for augmented women is mammography performed at a facility with employees trained in implant imaging.⁷ The American College of Radiology’s practice guidelines affirm that mammography is the best imaging tool available.⁸ The American College of Obstetrics and Gynecology and the US Preventive Services Task Force don’t comment on screening augmented women.

Acknowledgments

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

Evidence summary

A systematic review evaluating whether antepartum perineal massage reduced perineal trauma included 3 RCTs with a total of 1941 primagravidas and 493 multigravidas.¹ Women were randomized to receive either instruction in perineal massage or no instruction.

Beginning at 34 weeks, women or their partners performed perineal massage for 4 minutes, 3 to 4 times a week,² or once a day for 10 minutes.^3,4 Massage was performed by inserting 1 or 2 fingers 3 to 5 cm into the vagina and sweeping downward and from side to side, using almond oil for lubrication.

Birth attendants were blinded to patients’ assignments. All studies evaluated immediate postpartum outcomes; 1 study included a 3-month follow-up questionnaire.³

Massage reduces trauma, but less is more

Massage reduced perineal trauma requiring suturing by 10% among primigravid patients, compared with controls (relative risk [RR]=0.90; 95% confidence interval [CI], 0.84-0.96; number needed to treat [NNT]=14). Subgroup analysis revealed an inverse relationship between reduced trauma and frequency of massage: Primagravidas who massaged fewer than 1.5 times a week showed a 17% reduction (RR=0.83; 95% CI, 0.75-0.92; NNT=9), compared with an 8% reduction for women who massaged 1.5 to 3.4 times a week (RR=0.92; 95% CI, 0.85-1.00; NNT=22) and a statistically insignificant 7% reduction for the group that massaged more than 3.5 times a week (RR=0.93; 95% CI, 0.86-1.02).

Perineal massage reduced the incidence of episiotomy by 15% among primigravidas compared with controls (RR=0.85; 95% CI, 0.74-0.97; NNT=20); the largest reduction occurred in primagravidas who massaged as often as 1.5 times per week (RR=0.72; 95% CI, 0.57-0.91). This effect was not seen in primagravidas who massaged more often. Multigravid patients didn’t experience a statistically significant reduction in episiotomy.

Perineal massage didn’t affect the overall incidence of first- and second-degree perineal lacerations (first-degree laceration: RR=0.95; 95% CI, 0.78-1.16; second-degree laceration: RR=0.98; 95% CI, 0.84-1.15), nor the incidence of third- or fourth-degree lacerations (RR=0.81; 95% CI, 0.56-1.15). No difference was noted in the incidence of instrument delivery (RR=0.94; 95% CI, 0.81-1.08).

Massage means less pain for multigravidas

Massage reduced postpartum perineal pain in multigravidas, according to a questionnaire administered at 3 months in 1 study, to which 376 of 493 women (76%) responded (RR=0.45; 95% CI, 0.02-0.87; NNT=13).³ A subgroup of women who massaged more often than 3.5 times a week had a larger reduction in pain (RR=0.51; 95% CI, 0.33-0.79; NNT=11). At 3 months, massage produced no difference in rates of dyspareunia, sexual satisfaction, or incontinence of urine or feces when compared with standard care.

Recommendations

We found no expert or advocacy group guidelines on this topic.

Evidence summary

A systematic review evaluating whether antepartum perineal massage reduced perineal trauma included 3 RCTs with a total of 1941 primagravidas and 493 multigravidas.¹ Women were randomized to receive either instruction in perineal massage or no instruction.

Beginning at 34 weeks, women or their partners performed perineal massage for 4 minutes, 3 to 4 times a week,² or once a day for 10 minutes.^3,4 Massage was performed by inserting 1 or 2 fingers 3 to 5 cm into the vagina and sweeping downward and from side to side, using almond oil for lubrication.

Birth attendants were blinded to patients’ assignments. All studies evaluated immediate postpartum outcomes; 1 study included a 3-month follow-up questionnaire.³

Massage reduces trauma, but less is more

Massage reduced perineal trauma requiring suturing by 10% among primigravid patients, compared with controls (relative risk [RR]=0.90; 95% confidence interval [CI], 0.84-0.96; number needed to treat [NNT]=14). Subgroup analysis revealed an inverse relationship between reduced trauma and frequency of massage: Primagravidas who massaged fewer than 1.5 times a week showed a 17% reduction (RR=0.83; 95% CI, 0.75-0.92; NNT=9), compared with an 8% reduction for women who massaged 1.5 to 3.4 times a week (RR=0.92; 95% CI, 0.85-1.00; NNT=22) and a statistically insignificant 7% reduction for the group that massaged more than 3.5 times a week (RR=0.93; 95% CI, 0.86-1.02).

Perineal massage reduced the incidence of episiotomy by 15% among primigravidas compared with controls (RR=0.85; 95% CI, 0.74-0.97; NNT=20); the largest reduction occurred in primagravidas who massaged as often as 1.5 times per week (RR=0.72; 95% CI, 0.57-0.91). This effect was not seen in primagravidas who massaged more often. Multigravid patients didn’t experience a statistically significant reduction in episiotomy.

Perineal massage didn’t affect the overall incidence of first- and second-degree perineal lacerations (first-degree laceration: RR=0.95; 95% CI, 0.78-1.16; second-degree laceration: RR=0.98; 95% CI, 0.84-1.15), nor the incidence of third- or fourth-degree lacerations (RR=0.81; 95% CI, 0.56-1.15). No difference was noted in the incidence of instrument delivery (RR=0.94; 95% CI, 0.81-1.08).

Massage means less pain for multigravidas

Massage reduced postpartum perineal pain in multigravidas, according to a questionnaire administered at 3 months in 1 study, to which 376 of 493 women (76%) responded (RR=0.45; 95% CI, 0.02-0.87; NNT=13).³ A subgroup of women who massaged more often than 3.5 times a week had a larger reduction in pain (RR=0.51; 95% CI, 0.33-0.79; NNT=11). At 3 months, massage produced no difference in rates of dyspareunia, sexual satisfaction, or incontinence of urine or feces when compared with standard care.

Recommendations

We found no expert or advocacy group guidelines on this topic.

Evidence summary

A systematic review evaluating whether antepartum perineal massage reduced perineal trauma included 3 RCTs with a total of 1941 primagravidas and 493 multigravidas.¹ Women were randomized to receive either instruction in perineal massage or no instruction.

Beginning at 34 weeks, women or their partners performed perineal massage for 4 minutes, 3 to 4 times a week,² or once a day for 10 minutes.^3,4 Massage was performed by inserting 1 or 2 fingers 3 to 5 cm into the vagina and sweeping downward and from side to side, using almond oil for lubrication.

Birth attendants were blinded to patients’ assignments. All studies evaluated immediate postpartum outcomes; 1 study included a 3-month follow-up questionnaire.³

Massage reduces trauma, but less is more

Massage reduced perineal trauma requiring suturing by 10% among primigravid patients, compared with controls (relative risk [RR]=0.90; 95% confidence interval [CI], 0.84-0.96; number needed to treat [NNT]=14). Subgroup analysis revealed an inverse relationship between reduced trauma and frequency of massage: Primagravidas who massaged fewer than 1.5 times a week showed a 17% reduction (RR=0.83; 95% CI, 0.75-0.92; NNT=9), compared with an 8% reduction for women who massaged 1.5 to 3.4 times a week (RR=0.92; 95% CI, 0.85-1.00; NNT=22) and a statistically insignificant 7% reduction for the group that massaged more than 3.5 times a week (RR=0.93; 95% CI, 0.86-1.02).

Perineal massage reduced the incidence of episiotomy by 15% among primigravidas compared with controls (RR=0.85; 95% CI, 0.74-0.97; NNT=20); the largest reduction occurred in primagravidas who massaged as often as 1.5 times per week (RR=0.72; 95% CI, 0.57-0.91). This effect was not seen in primagravidas who massaged more often. Multigravid patients didn’t experience a statistically significant reduction in episiotomy.

Perineal massage didn’t affect the overall incidence of first- and second-degree perineal lacerations (first-degree laceration: RR=0.95; 95% CI, 0.78-1.16; second-degree laceration: RR=0.98; 95% CI, 0.84-1.15), nor the incidence of third- or fourth-degree lacerations (RR=0.81; 95% CI, 0.56-1.15). No difference was noted in the incidence of instrument delivery (RR=0.94; 95% CI, 0.81-1.08).

Massage means less pain for multigravidas

Massage reduced postpartum perineal pain in multigravidas, according to a questionnaire administered at 3 months in 1 study, to which 376 of 493 women (76%) responded (RR=0.45; 95% CI, 0.02-0.87; NNT=13).³ A subgroup of women who massaged more often than 3.5 times a week had a larger reduction in pain (RR=0.51; 95% CI, 0.33-0.79; NNT=11). At 3 months, massage produced no difference in rates of dyspareunia, sexual satisfaction, or incontinence of urine or feces when compared with standard care.

Recommendations

We found no expert or advocacy group guidelines on this topic.

Evidence summary

Brief counseling works

Good evidence suggests that physician-administered smoking cessation counseling lasting less than 3 minutes improves quit rates.¹ A Cochrane analysis of pooled data from 17 randomized trials that compared brief advice to no advice or usual care showed a small but significant increase in the odds of smoking cessation (odds ratio [OR]=1.74; 95% confidence interval [CI], 1.48-2.05).² The absolute difference in cessation rate was about 2.5% (number needed to treat [NNT]=40).

Another systematic review of 188 RCTs concluded that an estimated 2% (95% CI, 1%-3%; P<.001) of all smokers stopped smoking and did not relapse for as long as a year after receiving advice and encouragement to quit smoking from their physician in a single routine consultation.³

NRT is effective and safe for heart patients

NRT reduces withdrawal symptoms associated with stopping smoking by partially replacing nicotine in the blood. Abstinence rates are superior to placebo based on a Cochrane review (OR=1.77; 95% CI, 1.66-1.88; NNT=20; 95% CI, 17-23).⁴ The Cochrane review also concluded that all commercially available forms of NRT are effective for smoking cessation. Also, recent studies have established no association between NRT and further cardiac events.¹

Antidepressants are good treatment options

Bupropion acts by increasing brain levels of dopamine and norepinephrine and is a nicotine antagonist. A large double-blind, placebo controlled trial compared the relative efficacy of sustained-release bupropion (n=244), nicotine patch (n=244), bupropion plus nicotine patch (n=245), and placebo (n=160).⁵ At 1 year, the bupropion groups had higher self-reported point-prevalence abstinence rates (abstinence during the previous 7 days) than the placebo and nicotine-patch-alone groups (bupropion 30%, placebo 16%, nicotine-patch-alone 16%; absolute risk reduction [ARR]=0.14, NNT=7, P<.001).

Continuous abstinence (abstinence from quit date) was also higher for the bupropion groups compared with placebo (bupropion 18%, placebo 6%; ARR=0.12; NNT=8; P<.001). Adding nicotine replacement to bupropion therapy increased 1-year smoking cessation rates by 5% over bupropion alone but was not statistically significant.

A Cochrane review assessing the efficacy of antidepressants for smoking cessation showed that, when used as monotherapy, bupropion (31 trials; OR=1.94; 95% CI, 1.72-2.19) and nortriptyline (4 trials; OR=2.34; 95% CI, 1.61-3.41) both doubled the odds of smoking cessation.⁶

Another option: Varenicline

Varenicline, a partial agonist at the α4β2 nicotinic acetylcholine receptor, aids smoking cessation by relieving nicotine withdrawal symptoms. A Cochrane meta-analysis concluded that varenicline resulted in significantly greater continuous abstinence at 12 months than placebo (OR=3.22; 95% CI, 2.43-4.27; NNT=8; 95% CI, 5-11).⁷

Recommendations

The US Preventive Service Task Force (USPSTF) strongly recommends that clinicians screen all adults for tobacco use and provide tobacco cessation interventions as needed.⁸ The USPSTF’s Clinical Practice Guideline for treating tobacco dependence recommends following a 5-step (5 A’s) intervention for smoking cessation in patients willing to quit.¹

1. Ask the patient about smoking status at every visit.
2. Advise the patient to stop smoking.
3. Assess the patient’s willingness to quit.
4. Assist the patient by setting a date to quit smoking, providing self-help materials, and recommending the use of pharmacologic agents.
5. Arrange for follow-up visits.

Evidence summary

Brief counseling works

Good evidence suggests that physician-administered smoking cessation counseling lasting less than 3 minutes improves quit rates.¹ A Cochrane analysis of pooled data from 17 randomized trials that compared brief advice to no advice or usual care showed a small but significant increase in the odds of smoking cessation (odds ratio [OR]=1.74; 95% confidence interval [CI], 1.48-2.05).² The absolute difference in cessation rate was about 2.5% (number needed to treat [NNT]=40).

Another systematic review of 188 RCTs concluded that an estimated 2% (95% CI, 1%-3%; P<.001) of all smokers stopped smoking and did not relapse for as long as a year after receiving advice and encouragement to quit smoking from their physician in a single routine consultation.³

NRT is effective and safe for heart patients

NRT reduces withdrawal symptoms associated with stopping smoking by partially replacing nicotine in the blood. Abstinence rates are superior to placebo based on a Cochrane review (OR=1.77; 95% CI, 1.66-1.88; NNT=20; 95% CI, 17-23).⁴ The Cochrane review also concluded that all commercially available forms of NRT are effective for smoking cessation. Also, recent studies have established no association between NRT and further cardiac events.¹

Antidepressants are good treatment options

Bupropion acts by increasing brain levels of dopamine and norepinephrine and is a nicotine antagonist. A large double-blind, placebo controlled trial compared the relative efficacy of sustained-release bupropion (n=244), nicotine patch (n=244), bupropion plus nicotine patch (n=245), and placebo (n=160).⁵ At 1 year, the bupropion groups had higher self-reported point-prevalence abstinence rates (abstinence during the previous 7 days) than the placebo and nicotine-patch-alone groups (bupropion 30%, placebo 16%, nicotine-patch-alone 16%; absolute risk reduction [ARR]=0.14, NNT=7, P<.001).

Continuous abstinence (abstinence from quit date) was also higher for the bupropion groups compared with placebo (bupropion 18%, placebo 6%; ARR=0.12; NNT=8; P<.001). Adding nicotine replacement to bupropion therapy increased 1-year smoking cessation rates by 5% over bupropion alone but was not statistically significant.

A Cochrane review assessing the efficacy of antidepressants for smoking cessation showed that, when used as monotherapy, bupropion (31 trials; OR=1.94; 95% CI, 1.72-2.19) and nortriptyline (4 trials; OR=2.34; 95% CI, 1.61-3.41) both doubled the odds of smoking cessation.⁶

Another option: Varenicline

Varenicline, a partial agonist at the α4β2 nicotinic acetylcholine receptor, aids smoking cessation by relieving nicotine withdrawal symptoms. A Cochrane meta-analysis concluded that varenicline resulted in significantly greater continuous abstinence at 12 months than placebo (OR=3.22; 95% CI, 2.43-4.27; NNT=8; 95% CI, 5-11).⁷

Recommendations

The US Preventive Service Task Force (USPSTF) strongly recommends that clinicians screen all adults for tobacco use and provide tobacco cessation interventions as needed.⁸ The USPSTF’s Clinical Practice Guideline for treating tobacco dependence recommends following a 5-step (5 A’s) intervention for smoking cessation in patients willing to quit.¹

1. Ask the patient about smoking status at every visit.
2. Advise the patient to stop smoking.
3. Assess the patient’s willingness to quit.
4. Assist the patient by setting a date to quit smoking, providing self-help materials, and recommending the use of pharmacologic agents.
5. Arrange for follow-up visits.

Evidence summary

Brief counseling works

Good evidence suggests that physician-administered smoking cessation counseling lasting less than 3 minutes improves quit rates.¹ A Cochrane analysis of pooled data from 17 randomized trials that compared brief advice to no advice or usual care showed a small but significant increase in the odds of smoking cessation (odds ratio [OR]=1.74; 95% confidence interval [CI], 1.48-2.05).² The absolute difference in cessation rate was about 2.5% (number needed to treat [NNT]=40).

Another systematic review of 188 RCTs concluded that an estimated 2% (95% CI, 1%-3%; P<.001) of all smokers stopped smoking and did not relapse for as long as a year after receiving advice and encouragement to quit smoking from their physician in a single routine consultation.³

NRT is effective and safe for heart patients

NRT reduces withdrawal symptoms associated with stopping smoking by partially replacing nicotine in the blood. Abstinence rates are superior to placebo based on a Cochrane review (OR=1.77; 95% CI, 1.66-1.88; NNT=20; 95% CI, 17-23).⁴ The Cochrane review also concluded that all commercially available forms of NRT are effective for smoking cessation. Also, recent studies have established no association between NRT and further cardiac events.¹

Antidepressants are good treatment options

Bupropion acts by increasing brain levels of dopamine and norepinephrine and is a nicotine antagonist. A large double-blind, placebo controlled trial compared the relative efficacy of sustained-release bupropion (n=244), nicotine patch (n=244), bupropion plus nicotine patch (n=245), and placebo (n=160).⁵ At 1 year, the bupropion groups had higher self-reported point-prevalence abstinence rates (abstinence during the previous 7 days) than the placebo and nicotine-patch-alone groups (bupropion 30%, placebo 16%, nicotine-patch-alone 16%; absolute risk reduction [ARR]=0.14, NNT=7, P<.001).

Continuous abstinence (abstinence from quit date) was also higher for the bupropion groups compared with placebo (bupropion 18%, placebo 6%; ARR=0.12; NNT=8; P<.001). Adding nicotine replacement to bupropion therapy increased 1-year smoking cessation rates by 5% over bupropion alone but was not statistically significant.

A Cochrane review assessing the efficacy of antidepressants for smoking cessation showed that, when used as monotherapy, bupropion (31 trials; OR=1.94; 95% CI, 1.72-2.19) and nortriptyline (4 trials; OR=2.34; 95% CI, 1.61-3.41) both doubled the odds of smoking cessation.⁶

Another option: Varenicline

Varenicline, a partial agonist at the α4β2 nicotinic acetylcholine receptor, aids smoking cessation by relieving nicotine withdrawal symptoms. A Cochrane meta-analysis concluded that varenicline resulted in significantly greater continuous abstinence at 12 months than placebo (OR=3.22; 95% CI, 2.43-4.27; NNT=8; 95% CI, 5-11).⁷

Recommendations

The US Preventive Service Task Force (USPSTF) strongly recommends that clinicians screen all adults for tobacco use and provide tobacco cessation interventions as needed.⁸ The USPSTF’s Clinical Practice Guideline for treating tobacco dependence recommends following a 5-step (5 A’s) intervention for smoking cessation in patients willing to quit.¹

1. Ask the patient about smoking status at every visit.
2. Advise the patient to stop smoking.
3. Assess the patient’s willingness to quit.
4. Assist the patient by setting a date to quit smoking, providing self-help materials, and recommending the use of pharmacologic agents.
5. Arrange for follow-up visits.

Evidence summary

The number of children 2 years and older who are overweight has tripled in the past 2 decades; the current prevalence of over-weight children and adolescents in the US is 15%.¹ By contrast with adults—in whom overweight and obesity are defined separately as a body mass index (BMI) above 25 kg/m² and 30 kg/m², respectively—overweight and obesity are synonymous in children and are defined as a BMI above the 95th percentile for age and sex.² Children and adolescents with a BMI between the 85th and 95th percentiles are considered at risk for overweight.²

Overweight children are vulnerable to adverse health outcomes, including insulin resistance, hyperlipidemia, hypertension, depression, sleep apnea, asthma, steatohepatitis, genu varum, and slipped capital femoral epiphysis.² The many variables that have been suggested to influence childhood obesity include birth weight, gestational age, parental obesity, socioeconomic status, single parent household, and birth order.^3-5

Birth weight and later BMI: Consistently linked

A systematic review of 19 longitudinal, observational studies comparing birth weight with later BMI indicates that the association between the 2 is positive and consistent in multiple cohorts.³ Eleven studies focused on outcomes in childhood; another 8 measured outcomes into adulthood.

Fifteen of the 19 studies (79%), ranging in size from 1028 to 92,940 subjects, found a positive association between birth weight and later BMI. However, the data were too heterogeneous to combine into a single summary measure. One representative study quantified the relative risk (RR) for severe obesity (>95th percentile BMI) at 5 years of age as 1.7 (95% confidence interval [CI], 1.2-2.9) for birth weights between the 85th and 94th percentiles and 1.8 (95% CI, 1.1-2.9) for birth weights greater than the 95th percentile.³ Studies that didn’t find such an association had smaller sample sizes (137 to 432 subjects) and, therefore, may have lacked the power to detect an association.

Gestational diabetes. A subsequent retrospective cohort survey of 14,881 children born to mothers with gestational diabetes—and controlled for age, sex, and Tanner stage—found that the odds ratio (OR) for adolescent overweight was 1.4 (95% CI, 1.2-1.6) for each 1-kg increment in birth weight.⁴ The correlation persisted (OR=1.3; 95% CI, 1.1-1.5) when other covariates were controlled (television viewing, physical activity, energy intake, breastfeeding duration, birth order, household income, mother’s smoking, dietary restraint, and mother’s current BMI).

Large for gestational age. A US national cohort study of 3192 children adjusted for gestational age, found that large-for-gestational-age (LGA) infants with birth weights above the 90th percentile remained longer and heavier through 83 months of life.⁵ The triceps and subscapular skinfold measurements at 3 years of age for children born LGA were virtually identical to those of children born appropriate for gestational age, but by 6 years of age, skinfold measurements had diverged considerably, to more than 0.60 standard deviations. The researchers concluded that intrauterine growth is associated with obesity in early childhood.

Finally, a large Chinese population-based, case-control study (N=1322), found birth weight above 4.0 kg to be a risk factor for obesity in preschool-age children (OR=3.77; 95% CI, 2.06-6.29).⁶ The absolute rate of overweight increased from 8% to 26% among LGA infants.

In adolescence, parental weight may be a factor

A prospective cohort study of 1993 white LGA infants found a greater propensity to become obese in adolescence, but only if their mothers or fathers were also obese (RR=5.7).⁷ Children with lean parents did not have an increased risk of being over-weight in adolescence.

Recommendations

Although major organizations don’t focus on infant birth weight as a predictor of overweight, they do address childhood obesity. The American Academy of Pediatrics states that genetic, environmental, or combinations of risk factors predisposing children to obesity can and should be identified.² The US Preventive Services Task Force concludes that the evidence is insufficient to recommend for or against routine screening for overweight in children and adolescents as a means of preventing adverse health outcomes (Grade I recommendation).¹

Acknowledgements

The opinions and assertions contained herein are the private view of the author and not to be construed as official or as reflecting the view of the US Air Force Medical Service or the US Air Force at large.

Evidence summary

The number of children 2 years and older who are overweight has tripled in the past 2 decades; the current prevalence of over-weight children and adolescents in the US is 15%.¹ By contrast with adults—in whom overweight and obesity are defined separately as a body mass index (BMI) above 25 kg/m² and 30 kg/m², respectively—overweight and obesity are synonymous in children and are defined as a BMI above the 95th percentile for age and sex.² Children and adolescents with a BMI between the 85th and 95th percentiles are considered at risk for overweight.²

Overweight children are vulnerable to adverse health outcomes, including insulin resistance, hyperlipidemia, hypertension, depression, sleep apnea, asthma, steatohepatitis, genu varum, and slipped capital femoral epiphysis.² The many variables that have been suggested to influence childhood obesity include birth weight, gestational age, parental obesity, socioeconomic status, single parent household, and birth order.^3-5

Birth weight and later BMI: Consistently linked

A systematic review of 19 longitudinal, observational studies comparing birth weight with later BMI indicates that the association between the 2 is positive and consistent in multiple cohorts.³ Eleven studies focused on outcomes in childhood; another 8 measured outcomes into adulthood.

Fifteen of the 19 studies (79%), ranging in size from 1028 to 92,940 subjects, found a positive association between birth weight and later BMI. However, the data were too heterogeneous to combine into a single summary measure. One representative study quantified the relative risk (RR) for severe obesity (>95th percentile BMI) at 5 years of age as 1.7 (95% confidence interval [CI], 1.2-2.9) for birth weights between the 85th and 94th percentiles and 1.8 (95% CI, 1.1-2.9) for birth weights greater than the 95th percentile.³ Studies that didn’t find such an association had smaller sample sizes (137 to 432 subjects) and, therefore, may have lacked the power to detect an association.

Gestational diabetes. A subsequent retrospective cohort survey of 14,881 children born to mothers with gestational diabetes—and controlled for age, sex, and Tanner stage—found that the odds ratio (OR) for adolescent overweight was 1.4 (95% CI, 1.2-1.6) for each 1-kg increment in birth weight.⁴ The correlation persisted (OR=1.3; 95% CI, 1.1-1.5) when other covariates were controlled (television viewing, physical activity, energy intake, breastfeeding duration, birth order, household income, mother’s smoking, dietary restraint, and mother’s current BMI).

Large for gestational age. A US national cohort study of 3192 children adjusted for gestational age, found that large-for-gestational-age (LGA) infants with birth weights above the 90th percentile remained longer and heavier through 83 months of life.⁵ The triceps and subscapular skinfold measurements at 3 years of age for children born LGA were virtually identical to those of children born appropriate for gestational age, but by 6 years of age, skinfold measurements had diverged considerably, to more than 0.60 standard deviations. The researchers concluded that intrauterine growth is associated with obesity in early childhood.

Finally, a large Chinese population-based, case-control study (N=1322), found birth weight above 4.0 kg to be a risk factor for obesity in preschool-age children (OR=3.77; 95% CI, 2.06-6.29).⁶ The absolute rate of overweight increased from 8% to 26% among LGA infants.

In adolescence, parental weight may be a factor

A prospective cohort study of 1993 white LGA infants found a greater propensity to become obese in adolescence, but only if their mothers or fathers were also obese (RR=5.7).⁷ Children with lean parents did not have an increased risk of being over-weight in adolescence.

Recommendations

Although major organizations don’t focus on infant birth weight as a predictor of overweight, they do address childhood obesity. The American Academy of Pediatrics states that genetic, environmental, or combinations of risk factors predisposing children to obesity can and should be identified.² The US Preventive Services Task Force concludes that the evidence is insufficient to recommend for or against routine screening for overweight in children and adolescents as a means of preventing adverse health outcomes (Grade I recommendation).¹

Acknowledgements

The opinions and assertions contained herein are the private view of the author and not to be construed as official or as reflecting the view of the US Air Force Medical Service or the US Air Force at large.

Evidence summary

The number of children 2 years and older who are overweight has tripled in the past 2 decades; the current prevalence of over-weight children and adolescents in the US is 15%.¹ By contrast with adults—in whom overweight and obesity are defined separately as a body mass index (BMI) above 25 kg/m² and 30 kg/m², respectively—overweight and obesity are synonymous in children and are defined as a BMI above the 95th percentile for age and sex.² Children and adolescents with a BMI between the 85th and 95th percentiles are considered at risk for overweight.²

Overweight children are vulnerable to adverse health outcomes, including insulin resistance, hyperlipidemia, hypertension, depression, sleep apnea, asthma, steatohepatitis, genu varum, and slipped capital femoral epiphysis.² The many variables that have been suggested to influence childhood obesity include birth weight, gestational age, parental obesity, socioeconomic status, single parent household, and birth order.^3-5

Birth weight and later BMI: Consistently linked

A systematic review of 19 longitudinal, observational studies comparing birth weight with later BMI indicates that the association between the 2 is positive and consistent in multiple cohorts.³ Eleven studies focused on outcomes in childhood; another 8 measured outcomes into adulthood.

Fifteen of the 19 studies (79%), ranging in size from 1028 to 92,940 subjects, found a positive association between birth weight and later BMI. However, the data were too heterogeneous to combine into a single summary measure. One representative study quantified the relative risk (RR) for severe obesity (>95th percentile BMI) at 5 years of age as 1.7 (95% confidence interval [CI], 1.2-2.9) for birth weights between the 85th and 94th percentiles and 1.8 (95% CI, 1.1-2.9) for birth weights greater than the 95th percentile.³ Studies that didn’t find such an association had smaller sample sizes (137 to 432 subjects) and, therefore, may have lacked the power to detect an association.

Gestational diabetes. A subsequent retrospective cohort survey of 14,881 children born to mothers with gestational diabetes—and controlled for age, sex, and Tanner stage—found that the odds ratio (OR) for adolescent overweight was 1.4 (95% CI, 1.2-1.6) for each 1-kg increment in birth weight.⁴ The correlation persisted (OR=1.3; 95% CI, 1.1-1.5) when other covariates were controlled (television viewing, physical activity, energy intake, breastfeeding duration, birth order, household income, mother’s smoking, dietary restraint, and mother’s current BMI).

Large for gestational age. A US national cohort study of 3192 children adjusted for gestational age, found that large-for-gestational-age (LGA) infants with birth weights above the 90th percentile remained longer and heavier through 83 months of life.⁵ The triceps and subscapular skinfold measurements at 3 years of age for children born LGA were virtually identical to those of children born appropriate for gestational age, but by 6 years of age, skinfold measurements had diverged considerably, to more than 0.60 standard deviations. The researchers concluded that intrauterine growth is associated with obesity in early childhood.

Finally, a large Chinese population-based, case-control study (N=1322), found birth weight above 4.0 kg to be a risk factor for obesity in preschool-age children (OR=3.77; 95% CI, 2.06-6.29).⁶ The absolute rate of overweight increased from 8% to 26% among LGA infants.

In adolescence, parental weight may be a factor

A prospective cohort study of 1993 white LGA infants found a greater propensity to become obese in adolescence, but only if their mothers or fathers were also obese (RR=5.7).⁷ Children with lean parents did not have an increased risk of being over-weight in adolescence.

Recommendations

Although major organizations don’t focus on infant birth weight as a predictor of overweight, they do address childhood obesity. The American Academy of Pediatrics states that genetic, environmental, or combinations of risk factors predisposing children to obesity can and should be identified.² The US Preventive Services Task Force concludes that the evidence is insufficient to recommend for or against routine screening for overweight in children and adolescents as a means of preventing adverse health outcomes (Grade I recommendation).¹

Acknowledgements

The opinions and assertions contained herein are the private view of the author and not to be construed as official or as reflecting the view of the US Air Force Medical Service or the US Air Force at large.