University of Pennsylvania School of Medicine, 3600 Market Street, Suite 738, Philadelphia, PA 19104. E-mail: email@example.com
Objective: This study assessed whether adding orlistat to sibutramine would induce further weight loss in patients who previously had lost weight while taking sibutramine alone.
Research Methods and Procedures: Patients were 34 women with a mean age of 44.1 ± 10.4 years, weight of 89.4 ± 13.8 kg, and body mass index (BMI) of 33.9 ± 4.9 kg/m2 who had lost an average of 11.6 ± 9.2% of initial weight during the prior 1 year of treatment by sibutramine combined with lifestyle modification. Patients were randomly assigned, in double-blind fashion, to sibutramine plus orlistat or sibutramine plus placebo. In addition to medication, participants were provided five brief lifestyle modification visits during the 16-week continuation trial.
Results: Mean body weight did not change significantly in either treatment condition during the 16 weeks. The addition of orlistat to sibutramine did not induce further weight loss as compared with treatment by sibutramine alone (mean changes = +0.1 ± 4.1 kg vs. +0.5 ± 2.1 kg, respectively).
Discussion: These results must be interpreted with caution because of the study's small sample size. The findings, however, suggest that the combination of sibutramine and orlistat is unlikely to have additive effects that will yield mean losses ≥15% of initial weight, as desired by many obese individuals.
Two medications, sibutramine (Meridia; Knoll Pharmaceutical Co., Mt. Olive, NJ) (1) and orlistat (Xenical; Roche Laboratories, Nutley, NJ) (2), are currently approved by the Food and Drug Administration for weight loss and the maintenance of weight loss. Sibutramine is a combined norepinephrine-serotonin re-uptake inhibitor, whereas orlistat is a gastric and pancreatic lipase inhibitor. In controlled trials, sibutramine (15 mg once a day) was associated at 1 year with a 7% reduction in initial weight (1,3) and orlistat (120 mg, three times a day [TID]) with a 10% reduction (2,4,5,6). In both cases, the difference in weight loss between the medication and placebo conditions (i.e., placebo-subtracted weight loss) was approximately 4% to 5%.
Obese individuals want to lose two to three times more weight than is typically possible with current medications (7). Several investigators have suggested that larger weight losses might be achieved by combining weight loss agents (8,9,10). The present pilot study explored the benefits of adding orlistat to sibutramine in obese women who had lost an average of 11.6 ± 9.2% of their initial weight during 1 year of treatment by sibutramine alone. All women in the pilot study continued to receive sibutramine for 16 weeks; in addition, half of them were randomly assigned to orlistat and the other half to placebo. These two medications would appear to be excellent candidates for combined therapy because of their different mechanisms of action.
Research Methods and Procedures
Patients were 34 volunteers from a group of 43 women who had completed a 1-year treatment program that combined sibutramine (10 to 15 mg/d) with different amounts of lifestyle modification. As described in a separate report (11), the 43 participants lost an average of 12.0 ± 9.6 kg at 1 year, but there were marked differences among patients based on the program of lifestyle modification they received.
The 34 volunteers in the continuation study were told that all participants would receive sibutramine for an additional 16 weeks and that half of them also would be assigned at random (in double-blind fashion) to orlistat and the other half to placebo. The stated goal of the study was to determine whether the addition of orlistat would be associated with greater weight loss (or better maintenance of weight loss) than would continued treatment by sibutramine alone. Patients gave their informed consent to participate in the continuation study, which was approved by the University of Pennsylvania's Committee on Studies Involving Human Beings. Patients’ characteristics, before randomization, are shown in Table 1. ANOVA showed that patients in the two conditions did not differ significantly on any of the baseline measures, including weight loss during the prior 1-year program.
Table 1. Patients’ characteristics before randomization to sibutramine or sibutramine plus orlistat for the 16-week continuation study
Sibutramine plus placebo (N= 17)
Sibutramine plus orlistat (N= 17)
There were no significant differences among groups on any of the above variables.
44.3 ± 10.4
43.9 ± 10.7
90.1 ± 14.4
88.7 ± 13.5
163.6 ± 5.4
161.3 ± 10.1
33.6 ± 4.8
34.2 ± 5.1
Age of onset of obesity (year)
17.6 ± 10.5
14.1 ± 9.3
First year weight loss (kg) on sibutramine alone
9.8 ± 8.5
13.4 ± 9.7
At baseline (i.e., week 52), all patients met with a physician (R. I. B.) who examined their health and told them to continue to take sibutramine (10 to 15 mg once a day) in the morning. In addition, they were instructed to take one capsule of the investigational medication within ± 1 hour of lunch, dinner, and an evening snack. We decided not to prescribe orlistat in the morning because 14 of 34 (41%) patients indicated that they ate breakfast infrequently (i.e., 0 to 3 times a week). Of the 20 remaining participants, 12 (i.e., 35% of the total sample) reported that they usually ate a breakfast that was determined to contain ≤10 g of fat. For most women, evening snacking appeared to present a greater risk for overeating than did breakfast.
Patients were instructed to limit their fat intake to a maximum of 20 g per meal (or snack), and 60 g per day, to minimize possible gastrointestinal events, including oily stools, oily spotting, fecal urgency, and related side effects (2,4,5,6). They were warned that they would not be able to predict the temporal occurrence of such events. Participants were instructed to take the medication three times a day, at the designated times, even if they missed a meal or snack. This was done to facilitate their taking the medication as regularly as possible. Patients were also instructed to take a multivitamin supplement every morning to prevent possible decreases in levels of fat-soluble vitamins.
At week 53, patients returned to see the physician who assessed their response to both sibutramine and the experimental medication. Follow-up medical visits were scheduled at weeks 56 and 68 (or more frequently, as needed).
All patients met with a registered dietitian or doctoral-level psychologist for 30 minutes at weeks 52, 56, 60, 64, and 68. At the first visit, patients’ energy requirements were calculated, and they were instructed to consume a diet of 1200 to 1600 kcal/d, representing a deficit of approximately 600 to 850 kcal/d. Patients were told to consume a balanced diet (of their choosing) with approximately 20% of calories from protein, 50% from carbohydrate, and ≤30% from fat. They were provided handouts on topics that included the Food Guide Pyramid, food labels, low-fat cooking, and meal planning. Each month the practitioner reviewed patients’ food diaries and medication compliance. Participants also set monthly activity goals with an eventual objective of exercising five times a week for 30 to 40 minutes per bout.
Weight was measured at each visit with patients dressed in light clothing and without shoes. At week 68 (i.e., end of study), participants indicated whether they believed they had been assigned to orlistat or placebo. In addition, they completed a symptom inventory that assessed, for the prior week, the number of days that they had experienced various gastrointestinal events.
Attrition and Statistical Analyses
Three patients treated by sibutramine plus orlistat (i.e., combined therapy) and five treated by sibutramine alone discontinued treatment prematurely. Table 2 summarizes the reasons for attrition and patients’ weight loss at the time. A chi square test revealed no significant differences in dropout between conditions. Differences in weight loss between conditions during the 16-week trial were compared using analysis of covariance, with weight loss at the end of the first year of treatment (by sibutramine alone) taken as the covariate. Data were analyzed using both an end-point analysis (which included only treatment completers) and a last-observation-carried-forward analysis. The two sets of analyses reached the same statistical conclusions.
Figure 1 shows that body weight was essentially unchanged in both conditions during the 16-week continuation trial. ANOVA revealed neither an effect of time nor treatment condition. Thus, contrary to our hypothesis, the addition of orlistat to sibutramine did not significantly increase weight loss (or improve the maintenance of weight loss) as compared with the continued use of sibutramine alone (see Table 3).
Table 3. Change in weight (kg) for patients in two conditions
Sibutramine plus placebo
Sibutramine plus orlistat
EPA = end-point-analysis; N= 16, 10, 12, and 12 at weeks 56, 60, 64, and 68, respectively, for sibutramine plus placebo; N= 14, 15, 15, and 14 at weeks 56, 60, 64, and 68, respectively, for sibutramine plus orlistat.
LOCF = last-observation-carried-forward analysis; N= 17 for both treatment conditions at all times.
Note: 95th percentile confidence intervals = 2.1 to −0.5
A second ANOVA examined the effect of prior 1-year weight loss and treatment. Patients were divided into two groups based on whether they had lost <10% of their initial weight in the prior 1-year study or ≥10% (resulting in a 2 × 2 ANOVA). (The mean loss for the 16 patients in the first group was 3.3 ± 3.2%, whereas that for the 18 participants in the second group was 18.9 ± 5.8%.) Patients who had reduced <10% during the earlier trial lost 1.2 ± 3.2 kg during the 16-week continuation study, independent of which medications they received. By contrast, those who had lost >10% of weight in the prior trial gained 1.7 ± 2.6 kg during the 16-week study, yielding a significant difference between groups (p < 0.01). Figure 2 shows that women who had lost <10% of weight in the prior 1-year trial tended to lose more weight in the continuation study if assigned to orlistat plus sibutramine rather than to sibutramine alone (−2.6 ± 4.9 kg vs. −0.4 ± 1.2 kg, respectively). The difference, however, between conditions was not statistically significant.
A final subanalysis examined weight change in eight women who were assigned to the combination of orlistat plus sibutramine and had lost 5% to 14% of initial weight in the prior 1-year study. These women were selected, because all had responded to sibutramine (i.e., achieved a 5% weight loss) but had not lost so much weight (i.e., ≥15%) as to make further weight reduction unlikely with orlistat. These participants lost an average of 8.4 ± 4.4% of initial weight in the 1-year trial. In the 16-week continuation study, their mean weight increased by 0.2 ± 5.1 kg. Thus, even in highly selected patients, who were thought to be the most likely to benefit from combination therapy, adding orlistat to sibutramine did not increase weight loss.
Of the 17 patients assigned to sibutramine plus placebo, 7 took 10 mg/d of sibutramine and 10 received 15 mg/d. In the sibutramine plus orlistat group, 6 took 10 mg/d of sibutramine while the other 11 participants received 15 mg/d. All patients had been prescribed the 15 mg/d dose in the original 1-year study but, by the end of the year, it had been reduced to 10 mg/d in 13 of 34 women to control side-effects that included insomnia and increased blood pressure and pulse. These reductions occurred before patients began the 16-week continuation trial. There were no significant differences in weight change during the 16-week trial between patients who received the 10 mg/d vs. 15 mg/d dose.
Determination of Treatment Condition
Of the 14 patients assigned to sibutramine plus orlistat, 12 correctly identified their treatment condition at the end of the trial, as did 10 of 12 assigned to sibutramine plus placebo. A chi square test revealed that the percentage of correct identifications (84.6%) was significantly (p < 0.05) greater than that expected by chance. Thus, patients appeared to know whether they had received orlistat.
Table 4 presents patients’ reports of gastrointestinal symptoms during the last week of the trial. Fifty percent of patients treated by combined therapy reported experiencing soft stool and increased frequency of bowel movements at least 1 day of the week, as compared with only 9.1% of patients treated by sibutramine alone. Similarly, 42.9% of combined-therapy patients reported oily evacuation and fecal urgency at least 1 day of the week as compared with 0% and 9.1%, respectively, of patients treated by sibutramine alone. Although three of these four differences were statistically significant at the 0.05 level, none was significant at the 0.004 level, the level required if Bonferroni's correction for multiple tests were used.
Table 4. Patients’ report of side effects at week 68
During physician visits, patients did not report any unusual symptoms that could not be attributed to either sibutramine or orlistat alone. Thus, combining the two medications did not appear to result in any unexpected side-effects.
This study's principal finding was that adding orlistat to sibutramine did not significantly increase weight loss in obese women who had previously lost 11.6% of initial weight during 1 year of treatment by sibutramine alone. The two medications did not appear to have additive effects, a finding that disappointed several patients who had hoped, as we had, that they could lose approximately 10% of weight with the first medication and then an additional 10% with the second. The data revealed a trend for patients, who in the prior 1-year trial had lost less than 10% of their initial weight with sibutramine, to lose additional weight by also taking orlistat. However, their loss of only 2.6 kg at the end of 16 weeks was modest and did not differ significantly from that of patients who received sibutramine plus placebo. In addition, it is possible that these patients would have lost 2.6 kg if treated by orlistat alone (without combining it with sibutramine).
Patients who had lost ≥10% of initial weight in the prior 1-year trial appeared to receive little benefit from combined therapy; they gained 1.7 kg during the 16-week continuation study, as did patients treated by sibutramine alone. This finding suggests that there may be limits to the amount of weight that most obese individuals can lose (and maintain) with currently approved medications (8), as well as with behavioral interventions (12). This limit appears to be 10% to 15% of initial weight. Efforts to push beyond this limit may be thwarted by a toxic environment (13) that discourages physical activity while encouraging consumption of a high-fat diet, as well as by compensatory biological responses (14,15) that decrease energy expenditure. Whether alone or together, these factors appear to return weight toward the 10% mark, if not toward baseline (16,17,18). Andersen et al.(19), for example, used a very low calorie diet to induce an average loss of approximately 15% of initial weight but found that patients maintained a loss of only 10% at the end of 1 year, despite their receiving 30 mg/d dexfenfluramine throughout the trial. Hill et al. (20) similarly found during a 1-year follow-up that patients regained about one quarter of their 11% reduction in initial weight, despite receiving orlistat (120 mg TID) for the full follow-up period. From this perspective, it is not surprising that our most successful patients, who had lost an average of 18.9% in the prior 1-year trial, tended to gain weight in the 16-week continuation study, whether they received sibutramine alone or sibutramine plus orlistat. Even when a subanalysis was conducted on eight women who had lost a mean of only 8.6% of initial weight in the prior trial, they were found to gain 0.2 kg during the 16-week continuation study while receiving sibutramine plus orlistat.
Future medications, or combinations of medications, may well be capable of inducing and sustaining larger weight losses (8). This was the promise of the fenfluramine-phentermine combination until the fenfluramines were removed from the market in 1997 because of their association with valvular heart disease (21).
Results of the present study must be interpreted with caution because of our small sample size. Clearly, further studies are needed that have adequate power to detect clinically significant differences. In designing our investigation, we estimated that patients treated by orlistat plus sibutramine would lose 3.0 ± 3.0 kg during the 16-week trial, whereas those who received sibutramine alone would have a mean weight change of 0.0 ± 3.0 kg. With a sample size of 34, the power to detect this difference was 0.81 (α = 0.05, two-tailed test). We thought that, even with this small sample, we would be able to detect at least a trend toward significant differences between the two conditions.
In addition to increasing the sample size (and including men), investigators may wish to use alternative study designs such as comparing sibutramine (plus placebo) to orlistat (plus placebo) to the two medications combined (i.e., sibutramine plus orlistat). There are also a variety of options for sequencing the medications that include prescribing both from the outset or introducing the second medication only after the patient has met a weight-related criterion such as a 5% loss, a 2-month weight loss plateau, or significant weight regain. In addition, longer trials (≥1 year) will be needed to determine whether combined therapy improves the maintenance of weight loss. The present study was limited to 16 weeks, because we were interested primarily in whether adding orlistat to sibutramine would induce further weight loss. Maximum weight loss with medication typically occurs in the first 16 to 26 weeks (8).
We intentionally used a modest behavioral intervention in the continuation study to reveal most clearly the effects of the medications. Use of a more intensive lifestyle intervention may well have increased the size of the weight losses produced by combined therapy (as well as by sibutramine alone), as shown in a previous study (11).
The present findings raise questions about whether it is possible to conduct truly blinded evaluations of orlistat. At the end of the 16-week trial, all but 4 of 26 patients correctly identified their treatment condition. We suspect that the gastrointestinal side-effects associated with orlistat enabled patients to discern their treatment assignment. The problem, however, of patients becoming “unblinded” is not unique to orlistat. Nearly two decades ago Brownell and Stunkard(22) showed that 71% of patients correctly identified whether they had been assigned to fenfluramine or placebo.
In summary, results of this pilot study revealed little benefit of adding orlistat to sibutramine in patients who had previously lost 11.6% of initial weight on sibutramine. Additional studies, however, that include larger sample sizes, as well as different experimental designs, are needed to reach definitive conclusions about the possible benefits of combining these two medications.
The authors thank Dr. Richard Prus-Wisniewski for his contributions to this study, as well as Knoll Pharmaceutical Co. for providing the sibutramine used in the trial. Orlistat was purchased by the authors. This study was supported, in part, by a National Institute of Mental Health Research Scientist Development Award and by grants from Knoll Pharmaceutical Company and Novartis Nutrition Co. Drs. Berkowitz, Sarwer, and Wadden have all received speaking honoraria from Knoll Pharmaceutical Co. and Roche Laboratories, which manufacture sibutramine and orlistat, respectively. Drs. Berkowitz and Wadden also serve (or have served) as consultants to both companies.