Berkowitz RI, Wadden TA, Tershakovec AM, Cronquist JL. Behavior Therapy and Sibutramine for the Treatment of Adolescent ObesityA Randomized Controlled Trial. JAMA. 2003;289(14):1805-1812. doi:10.1001/jama.289.14.1805
Author Affiliations: Department of Psychiatry, Weight and Eating Disorders Program, University of Pennsylvania School of Medicine, Philadelphia (Drs Berkowitz and Wadden, and Ms Cronquist); and Departments of Psychiatry (Dr Berkowitz) and Pediatrics (Dr Tershakovec), The Children's Hospital of Philadelphia and University of Pennsylvania, Philadelphia.
Context Adolescent obesity is becoming a national public health problem. Weight-loss
medications including sibutramine facilitate weight control in adults and
could be used with obese adolescents in combination with behavior therapy
Objective To examine whether increased weight loss in obese adolescents is induced
when sibutramine is added to a family-based, behavioral weight control program.
Design, Setting, and Participants Randomized, double-blind, placebo-controlled trial consisting of 82
adolescents aged 13 to 17 years with a body mass index (BMI) of 32 to 44 conducted
from March 1999 to August 2002 at a university-based clinic for 6 months,
followed by open-label treatment during months 7 to 12.
Interventions For the first 6 months, participants received either BT and sibutramine
or BT and placebo. From months 7 to 12, all participants received sibutramine
in open-label treatment.
Main Outcome Measures Percentage change in BMI; systolic and diastolic blood pressure and
pulse; and hunger.
Results In intention-to-treat analysis at month 6, participants in the BT and
sibutramine group lost a mean (SD) of 7.8 kg (6.3 kg) and had an 8.5% (6.8%)
reduction in BMI, which was significantly more than weight loss of 3.2 kg
(6.1 kg) and reduction in BMI of 4.0% (5.4%) in the BT and placebo group.
Significantly greater reductions in hunger (P = .002)
also were reported by participants who received BT and sibutramine. From months
7 to 12, adolescents initially treated with sibutramine gained 0.8 kg (10.5
kg) with continued use of the medication, whereas those who switched from
placebo to sibutramine lost an additional 1.3 kg (5.4 kg). Medication dose
was reduced (n = 23) or discontinued (n = 10) to manage increases in blood
pressure, pulse rate, or other symptoms.
Conclusions The addition of sibutramine to a comprehensive behavioral program induced
significantly more weight loss than did BT and placebo. Until more extensive
safety and efficacy data are available, medications for weight loss should
be used only on an experimental basis in adolescents and children.
Adolescent obesity is rapidly becoming a national public health problem.
The prevalence of this disorder increased from 5% to 11% from the 1980s to
1994 and to 15.5% by 2000.1- 3 This
increase has been accompanied by a dramatic increase in type 2 diabetes mellitus
and related health complications.4- 11 There
has been relatively little controlled research on the treatment of adolescent
obesity. A comprehensive behavioral approach appears to be the most effective
treatment,12- 14 but
most studies have reported mean weight losses of only 1 to 4 kg. Participants
typically remained obese at the end of therapy.
Weight-loss medications, including sibutramine and orlistat, facilitate
weight control in adults15- 22 and
potentially could be used with obese adolescents. No weight-loss agents are
currently Food and Drug Administration approved for children younger than
16 years. We agree with Dietz8 that weight-loss
medications should be used only on an experimental basis with adolescents
and solely "as an adjunct to behavior modification, family therapy, increased
activity . . ." and that ". . . the initial studies of these agents should
be blinded and placebo controlled." The aim of the present study was to increase
weight loss in obese adolescents by combining a comprehensive behavioral program
with pharmacotherapy. This study is to our knowledge the first randomized,
placebo-controlled trial of sibutramine in the treatment of obese youth.
A randomized, double-blind, placebo-controlled trial was conducted for
6 months (phase 1), after which all participants received sibutramine in an
open-label extension for an additional 6 months (phase 2) (Figure 1). We hypothesized that participants who received behavior
therapy (BT) and sibutramine would lose significantly more weight during phase
1 than those who received BT and placebo. Participants, parents, and all study
personnel were blinded to treatment condition during phase 1. Only the research
pharmacist was aware of treatment status. This study was approved by the institutional
review boards of the University of Pennsylvania and The Children's Hospital
of Philadelphia and was conducted from March 1999 to August 2002 at the Weight
and Eating Disorders Program at the University of Pennsylvania School of Medicine,
Candidates were boys and postmenarcheal girls aged 13 to 17 years who
had a body mass index (BMI, calculated as weight in kilograms divided by the
square of height in meters) of 32 to 44. Contraindications to participation
included cardiovascular disease (including arrhythmias); type 1 or 2 diabetes
mellitus; major psychiatric disorders; pregnancy; use of a weight-loss medication
or a weight loss of 5 kg or more in the prior 6 months; use of medications
promoting weight gain (eg, oral steroids); use of medications contraindicated
with use of sibutramine23; or cigarette smoking.
Adolescents, accompanied by a parent or guardian, completed a behavioral assessment,
conducted by a staff psychologist or psychiatrist. Written informed consent
was obtained from the parent and assent from the adolescent. The adolescent's
primary care physician performed a history and physical examination to exclude
the noted contraindications.
Behavioral Protocol. Adolescents in both treatment
conditions received the same comprehensive family-based behavioral weight-loss
program delivered following detailed treatment manuals.24- 26 In
phase 1, participants attended 13 weekly group sessions followed up by 6 biweekly
group sessions. In phase 2, group sessions were held biweekly from months
7 to 9 and monthly from months 10 to 12. Parents met separately in group sessions
held on the same schedule as the adolescents' meetings. Groups were led by
dietitians, psychologists, or psychiatrists.
Adolescents in both treatment groups were instructed to consume a 1200
to 1500 kcal/d diet of conventional foods, with approximately 30% from fat,
15% from protein, and the remainder from carbohydrate. They were prescribed
an eventual goal of walking or engaging in similar aerobic activity for 120
minutes per week or more. Participants kept daily eating and activity logs
that they submitted at each session. The content of the parents' sessions
paralleled that of their children's sessions. Parents praised their adolescents
for adhering to the diet and activity program.
Medication Protocol. During phase 1, all participants
received placebo (single blind) the first week. At week 2, they received either
placebo or 5 mg/d of sibutramine. In medication-treated participants, sibutramine
was increased to 10 mg/d at week 3 and to 15 mg/d at week 7. Knoll Pharmaceutical
Co and Abbott Laboratories manufactured and provided both the placebo and
sibutramine for the study. Placebo capsules looked exactly like sibutramine
capsules and were dispersed in the same way.
Increases in blood pressure (BP) and pulse have been reported previously
in adults who received sibutramine.15,23 Thus,
in the present study, participants whose systolic or diastolic BP increased
from baseline by 10 mm Hg or more (or who had increases in pulse rate of ≥15%)
for 2 or more consecutive visits had their medication dose reduced in 5-mg
decrements until acceptable BP and pulse rate values were obtained. Sibutramine
was discontinued in participants in whom dose reductions did not reverse the
10-mm Hg or more increase or in whom systolic or diastolic BP increased 20
mm Hg or more at any single visit. During phase 2, all participants were treated
with sibutramine following the same dose titration schedule used in phase
Weight was measured at each treatment visit and at the major assessments
(ie, baseline and months 3, 6, 9, and 12). Height and waist circumference
were measured at the major assessments following standard techniques.27
Adherence to the lifestyle program was assessed by the number of days
of food records completed during phase 1. Medication adherence during this
same period was evaluated by pill counts. Hunger was evaluated at the major
assessments by the Eating Inventory (range, 0-14).28 Maternal
level of eduction was assessed on a 6-point scale ranging from some high school
to doctoral degree (some high school = 1, high school graduate = 2, some college
= 3, college graduate = 4, master's degree = 5, and doctoral degree = 6).
Lipids, lipoproteins, serum glucose, and insulin levels were measured
at the major assessments following an overnight fast (Quest Diagnosis Incorporated,
Horsham, Pa). An electrocardiogram was also obtained at these times. Systolic
and diastolic BP, as well as pulse rate, were measured at the major assessments
(and biweekly for the first 8 months and then monthly) with a Dinamap monitor
(XL Model 9300, Johnson & Johnson, New Brunswick, NJ). Assessments were
obtained after the adolescent stayed still for at least 5 minutes. Three readings
were taken at 1-minute intervals; the second and third measures were averaged.
Adverse events were assessed and recorded at each medical visit.
The present study was powered to detect a 4.0% difference in baseline
BMI between the 2 treatment groups. The study was powered using the effect
size statistic d by Cohen,29 defined
as the difference between the 2 treatment groups divided by the pooled SD.
The pooled SD was estimated to be 5%. Given a 4.0% difference in percentage
change in BMI and an SD of 5%, this corresponds to a large effect size (Cohen d = .80).29 Assuming a 2-tail
analysis, α = .05 and an SD of 5%, 93% power was obtained to detect
a percentage BMI reduction as small as 4% between the 2 treatment groups.29
Changes in degree of obesity were evaluated by changes in weight, BMI,
and BMI z scores,30 as
well as by percentage change in initial BMI. This latter measure controls
for changes in height over time, as well as for differences among participants
in baseline BMI. Differences between treatment conditions in changes on these
measures from baseline to month 6 (phase 1) and from months 7 to 12 (phase
2) were examined using a 2 × 3 (condition × time) repeated measures
analysis of variance. An intention-to-treat analysis was used in which missing
data were replaced by the participant's baseline body weight (ie, a baseline-carried-forward
analysis). A completers analysis (with weight data for completers at 6 and
12 months) and a mixed-model analysis also were performed. The family-wise
error term for each analysis was set at .05, and all data were analyzed using
SAS version 8.0 (SAS Institute, Cary, NC). P≤.05
was considered statistically significant. Results of all analyses yielded
the same statistical conclusions. Only the results of the intention-to-treat
analysis for percentage change in BMI are presented.
Secondary outcome measures, including BP, pulse, and hunger, also were
analyzed using repeated measures analysis of variance. For each variable,
the family-wise error term was set at .05. All secondary analyses were conducted
using treatment completers. This was done, in part, to minimize potential
masking of changes in BP or pulse that might occur with an intention-to-treat
analysis, which would replace missing BP or pulse values with baseline values;
vital signs have been shown to increase while taking sibutramine.
Characteristics of the 82 patients who were randomized to treatment
are shown in Table 1 and Figure 1. There were no significant differences
between treatment groups on any of the baseline measures or in attrition from
At month 6 during phase 1 (placebo controlled), participants in the
BT and sibutramine group lost a mean (SD) of 7.8 kg (6.3 kg) equal to an 8.5%
(6.8%) reduction in initial BMI (Figure 2 and Table 2). In contrast,
adolescents treated with BT and placebo lost 3.2 kg (6.1 kg) equal to a significantly
smaller 4.0% (5.4%) reduction in BMI (effect size, 0.73; 95% confidence interval
[CI], 0.28-1.18; P = .001). More than twice as many
adolescents in the BT and sibutramine group reduced their initial BMI by 10%
(P = .02) and 15% (P = .02)
compared with those treated by BT and placebo (Table 3). In Table 2,
statistically significant differences between groups also were observed in
changes in waist circumference and BMI z score.
In phase 2 (open-label sibutramine treatment), participants who were
originally treated with placebo and were switched to sibutramine from months
7 to 12 lost an additional 1.3 kg (5.4 kg) during this period, reducing their
baseline BMI by an additional 2.4% (5.0%). In contrast, participants originally
treated with sibutramine who continued medication gained 0.8 kg (10.5 kg)
during months 7 to 12 equal to a 0.2% (5.4%) reduction in initial BMI (BMI
decreased even with weight gain because of increases in height). The difference
in percentage reduction in BMI between groups (months 7-12) approached significance
(effect size, –0.22; 95% CI, –0.66 to 0.21; P = .057).
From baseline to month 12, participants treated throughout the study
with BT and sibutramine lost a total of 7.0 kg (9.3 kg) equal to an 8.6% (9.9%)
reduction in initial BMI. Those participants in the BT and placebo group who
were switched at month 7 to sibutramine lost a total of 4.5 kg (8.8 kg) equal
to a 6.4% (8.3%) reduction in initial BMI. The difference between groups at
month 12 was not significant (Figure 2).
Weight loss was not related at any time to participants' baseline characteristics,
including maternal BMI or education.
During the first 6 months, participants in the BT and sibutramine group
monitored their food intake for a mean (SD) of 58.9 days (44.9 days) compared
with 61.3 days (38.2 days) for those in the BT and placebo group. The effects
of treatment condition (placebo vs sibutramine) and self-monitoring on percentage
reduction in BMI at month 6 were examined using stepwise linear regression.
Number of food records completed during this period accounted for 17.7% of
the variance (R2 = 0.177) in the percentage
change in BMI. Treatment group accounted for an additional 12.9% (R2 = 0.129), bringing the total variance explained to 30.6%
(R2 = 0.306, P<.001). Figure 3 presents weight change data according
to treatment condition and the frequency of participants' self-monitoring.
Participants treated with BT and sibutramine used (determined by pill
count) 79.1% (21.7%) of the total number of pills they were prescribed during
the first 6 months compared with 78.3% (15.9%) for adolescents in the BT and
placebo group. In the BT and sibutramine group, the total dose of medication
taken correlated significantly with percentage reduction in initial BMI at
month 6 (r = 0.44, P = .003).
Participants in this group compared with those treated with placebo reported
significantly greater reductions in hunger at month 3 (effect size, 0.39;
95% CI, –0.07 to 0.86; P<.001) and month
6 (effect size, 0.43; 95% CI, –0.06 to 0.86; P =
.002). Mean (SD) changes were –2.1 (3.3) vs –0.3 (3.2) and –2.1
(3.3) vs –0.71 (3.7) at the 2 periods, respectively. No significant
differences in hunger were observed at month 12.
There were no differences between groups during phases 1 or 2 in changes
in lipids, triglycerides, serum insulin, serum glucose, or homeostasis model
of insulin sensitivity (HOMA).31 Collapsing
across conditions, however, at month 12 there was a significant increase in
high-density lipoprotein cholesterol (P = .001) and
significant reductions in serum insulin (P<.001)
and HOMA (P<.001) (Table 4). At this time, there were significant correlations between
percentage change in BMI and percentage reduction in total cholesterol (r = 0.30, P = .02), low-density
lipoprotein cholesterol (r = 0.33, P = .01), insulin (r = 0.30, P = .02), glucose (r = 0.30, P = .02), triglycerides (r = 0.36, P = .005), and HOMA (r = 0.34, P = .01). Percentage reduction in BMI also correlated with percentage
increase in high-density lipoprotein cholesterol (r =
–0.32, P = .01).
In Table 5, mean (SD) systolic
BP decreased 3.6 mm Hg (8.6 mm Hg) at month 3 in the BT and placebo participants
compared with a significant increase of 1.8 mm Hg (10.7 mm Hg) in adolescents
treated with sibutramine (effect size, 0.55; 95% CI, 0.10-1.00; P = .02). Values for these 2 groups at month 6 were –4.0 mm Hg
(8.9 mm Hg) and 0.4 mm Hg (9.0 mm Hg), respectively. The difference between
groups approached significance (effect size, 0.45; 95% CI, –0.02 to
0.92; P = .06). There were no other significant differences
between groups at any period in changes in systolic or diastolic BP. Table 5 also shows that pulse rate increased
significantly (5-6/min) at months 3 and 6 in participants who received BT
and sibutramine compared with BT and placebo. Similarly, during months 7 to
12, pulse rate increased by 7.6/min (12.7/min) in participants who were switched
from placebo to sibutramine.
During phase 1, medication dose was reduced or discontinued in 19 of
43 participants in the BT and sibutramine condition in response to elevations
in BP, pulse rate, or both. Thus, the values reported in Table 5 reflect deliberate efforts to minimize increases. To estimate
the effects of sibutramine with BP and pulse, in the absence of such intervention,
baseline values for the subset of 19 participants were compared with their
values at the time dose reduction was triggered. In these 19 adolescents,
systolic BP increased from a baseline of 112.3 mm Hg (13.8 mm Hg) to 122.3
mm Hg (14.4 mm Hg) and diastolic BP rose from 55.6 mm Hg (5.1 mm Hg) to 64.2
mm Hg (8.3 mm Hg) (P<.001 for both). Pulse rate
increased from 77.5/min (11.5/min) to 91.8/min (12.7/min) (P<.001).
When adverse events (Table 6)
characterized as elevations in BP and pulse rate, BP alone, or pulse rate
alone are combined, differences approach significance (P = .06). One participant had ventricular premature complexes (VPCs)
after 6 months of treatment with sibutramine. The participant was asymptomatic
and had no other significant changes on electrocardiographic examination.
The participant continued to have VPCs in the 6 months following discontinuation
of sibutramine, following a course similar to that of benign VPCs that may
be observed during adolescence. A second adolescent was observed to have VPCs
at month 9, 3 weeks following discontinuation of medication because of elevations
of BP and pulse rate; by month 12, there was no evidence of VPCs.
Ventricular premature complexes and extrasystoles have been reported
in sibutramine postmarketing surveillance.23 During
the full 12-month study, sibutramine was reduced to 10 mg in 16 participants,
to 5 mg in 7 additional adolescents, and discontinued in 10 participants (6
because of increased BP and/or pulse rate, 2 for ecchymoses, 1 for VPCs, and
1 because of rash of unclear etiology). Reductions were made primarily to
manage elevations in BP and pulse rate.
The addition of sibutramine with a comprehensive behavioral program
induced significantly more weight loss in obese adolescents than traditional
behavioral treatment (with placebo). Participants who received both BT and
pharmacotherapy during the first 6 months decreased their initial BMI more
than those who received BT and placebo, and reported significantly greater
reductions in hunger. Results of this controlled trial of sibutramine in adolescents
suggest there may be benefits of combining behavioral and pharmacological
The potential benefits of weight-loss medication were further suggested
in the open-label phase. The BMI decreased an additional 2.4% (during months
7-12) in participants who were switched from placebo to sibutramine, and adolescents
originally treated with medication maintained their weight loss at month 12
with continued use of sibutramine. Additional placebo-controlled trials are
needed to determine if sibutramine and orlistat facilitate the maintenance
of weight loss in obese adolescents, as has been shown in adults.20,34 Because data are not available beyond
month 12 in the present study, changes in weight (eg, weight regain) following
discontinuation of medication are unknown.
Weight loss continued through month 6 in the BT and sibutramine group
and then plateaued. Weight loss tends to plateau in obese adults after 6 months
of treatment with BT or pharmacotherapy.35 We
can only speculate that counterregulatory biological mechanisms, reduced total
energy expenditure, or behavioral fatigue may limit weight reduction following
about a 10% weight loss.36,37
There was a 6.4% decrease in initial BMI at month 12 for participants
who started taking placebo and then switched to sibutramine at month 6. This
compares with an 8.5% reduction for those participants who started taking
sibutramine at month 6. The smaller weight loss in the former group when switched
to sibutramine may have been attributable to the reduced number of behavior
modification sessions provided during the second 6 months. Further research
is needed to evaluate the optimal timing and duration of pharmacological treatment
when combined with BT.
Similar to studies of adults,15,23 we
observed sibutramine-related increases in pulse rate of 5 to 8/min. Moreover,
during the first 6 months of treatment, 19 of 43 adolescents experienced significant
elevations in BP and pulse rate that required reductions in the dose of sibutramine.
During the entire study, 5 participants had marked and sustained increases
in BP (≥10 mm Hg) that required discontinuation of the medication. These
cases underscore the need to closely monitor vital signs in patients prescribed
sibutramine. Our dose-reduction strategy prevented mean systolic and diastolic
BP from rising significantly during the year. However, larger and longer studies
are needed of BP and pulse rate in adolescents treated with sibutramine.
The Committee for Proprietary Medicinal Products38 in
Europe recently investigated concerns about the cardiovascular safety of sibutramine.
The committee noted increases in BP and pulse rate in some patients and also
concluded that current recommendations for identifying contraindications to
treatment and for close monitoring of BP and pulse rate adequately addressed
safety concerns. With these precautions, the committee determined that the
benefit/risk balance of sibutramine was favorable for obese adults and recommended
reduction or discontinuation of dose in response to significant elevations
in BP and pulse rate.
The rapid increase in the prevalence of type 2 diabetes mellitus in
children and adolescents is likely related to the increase in the prevalence
of obesity.39 The high mean baseline insulin
and HOMA levels for the participants in this study are consistent with significant
insulin resistance, suggesting that the adolescents in our study are at increased
risk of developing type 2 diabetes mellitus. However, with mean decreases
in BMI of less than 10%, participants' insulin and HOMA levels decreased 20%
and 23% at month 12, respectively. Decreases in these values suggest the positive
effects of weight loss in adolescents at risk for type 2 diabetes mellitus.
Moreover, the 8% increase in high-density lipoprotein cholesterol suggests
the benefits of weight loss and increased physical activity in decreasing
the risk of cardiovascular disease.40 Further
investigations, however, are needed to confirm these findings that were observed
during the open-label phase of our study.
This study revealed an apparent interplay between medication and behavior.
The medication was associated with significant reductions in hunger that likely
facilitated adherence to a low-calorie diet (thus, producing more weight loss).18 On the other hand, as reported in adults,41 adherence to self-monitoring optimized response to
treatment with medication. Adolescents who adhered closely to the behavioral
program, measured by the number of days they completed their food records,
and who were treated with sibutramine achieved the greatest decrease in BMI
during the first 6 months. The behavioral and pharmacological treatments appeared
to have additive effects that maximized weight loss.32,33
Data were not collected concerning whether participants could guess
if they were taking sibutramine or placebo. It is possible that those who
thought they received the drug (ie, who experienced adverse effects) worked
harder in the behavioral program to achieve weight reduction. The adherence
measure (ie, number of completed food records), however, did not reveal differences
between participants who received placebo or sibutramine.
Results of this placebo-controlled trial showed that the addition of
sibutramine with BT significantly increased weight loss in obese adolescents
compared with traditional BT alone (with placebo). These findings suggest
that weight loss medication may be of benefit to adolescents. Sibutramine,
however, must be carefully monitored in adolescents, as in adults, to control
increases in BP and pulse rate. Moreover, larger and longer studies are needed
to assess the benefits and costs of pharmacological treatment in obese adolescents.
Until more extensive safety and efficacy data are available, we agree with
Dietz8 that weight-loss medications should
be used only on an experimental basis for adolescents.