Am Fam Physician. 2008;78(11):online
Background: Because lifestyle modifications are associated with only modest weight loss and high rates of recidivism, interest in pharmacologic treatments of obesity has been increasing. Rucker and colleagues conducted a systematic review and meta-analysis of weight-loss medications to assess effectiveness and the potential for adverse effects in adults.
The Study: The authors searched several electronic databases and reference lists from all relevant identified papers for randomized, placebo-controlled clinical trials of antiobesity drugs. To be eligible, the study had to be a double-blind randomized controlled trial lasting at least one year. Study outcomes needed to include changes in weight, cardiovascular risk factors, cardiovascular morbidity and mortality, and overall mortality.
Results: The search identified 30 trials that met inclusion criteria. Sixteen studies (10,631 participants) tracked the effects of orlistat (Xenical), 10 studies (2,623 participants) used sibutramine (Meridia), and four studies (6,635 participants) used rimonabant (Acomplia; not available in the United States). Participants had similar demographic profiles across all trials. Overall, about 67 to 75 percent of participants were women, and about 90 percent were white. The mean age was 45 to 50 years, and the mean body mass index was 35 to 36 kg per m2. Almost all studies excluded participants with obesity secondary to endocrine conditions, as well as those with uncontrolled hypertension, psychiatric conditions, medical illness, previous bariatric surgery, and those receiving drugs that affect body weight. Women who were pregnant or lactating were also excluded.
Attrition rates were high, averaging 30 percent for orlistat and 40 percent for sibutramine and rimonabant. The overall methodologic quality of the studies was poor. End points were inconsistently reported, descriptions of randomization were excluded, allocation concealment was not described, and the blinding of outcome assessors was not mentioned.
Studies reported that participants lost an average of 6 lb, 6 oz (2.9 kg) more using orlistat than placebo. The results for sibutramine and rimonabant were 9 lb, 5 oz (4.2 kg) and 10 lb, 6 oz (4.7 kg), respectively. Compared with placebo, the percentage of participants achieving 5 percent weight loss increased by 21 percent for orlistat, 32 percent for sibutramine, and 33 percent for rimonabant. The incidence of type 2 diabetes was reduced from 9.0 to 6.2 percent with orlistat. Measures of glycemic control in participants with diabetes were significantly improved in one study of rimonabant, but changes in glycemic variables were inconsistently reported in studies of sibutramine.
Orlistat was also associated with improvements in measures of total cholesterol, low-density lipoprotein cholesterol, and blood pressure levels. However, orlistat users also had lower concentrations of high-density lipoprotein (HDL) cholesterol and increased reports of gastrointestinal upsets.
Sibutramine and rimonabant were associated with improvements in HDL cholesterol and triglyceride levels, and rimonabant also improved blood pressure. However, sibutra-mine was associated with increased blood pressure, and rimonabant was associated with increased rates of mood disorders.
No data were found on the effect of these three drugs on cardiovascular morbidity or overall mortality.
Conclusion: The authors conclude that although all three drugs were associated with an average placebo-subtracted weight loss of less than 11 lb (5 kg), the drugs differed in secondary outcomes and adverse effect profiles.