Back grey_arrow_rt.gif
Rimonabant: obituary for a wonder drug - Editorial
  The Lancet Aug 14 2010
S Matthijs Boekholdt aEmail Address, Ron JG Peters a a Department of Cardiology, Academic Medical Center, 1100DD Amsterdam, Netherlands
Abdominal obesity leads to a range of cardiovascular risk factors, including insulin resistance or overt diabetes, hypertension, hypertriglyceridaemia, and low HDL cholesterol. In theory, interventions to reverse this process should be simple: regular physical activity and a healthier diet. In practice, however, these targets are rarely achieved. A drug that could help to reduce weight and reverse these metabolic risk factors is eagerly awaited.
Rimonabant is a selective blocker of the cannabinoid-1 receptor. Stimulation of this receptor, for example by cannabinoids, can lead to increased appetite, feelings of wellbeing, and relaxation; anxiety and paranoia might also occur. In preliminary studies, rimonabant facilitated weight loss and had beneficial metabolic effects, even independently of weight reduction. Four trials in obese patients with additional cardiovascular risk factors showed that, in combination with a hypocaloric diet, rimonabant produced modest but sustained reductions in weight, fasting insulin, glycosylated haemoglobin, and triglyceride concentrations, and an increase in HDL cholesterol concentrations.14 As such, reviews hailed the arrival of a wonder drug.
In 2006, the European Medicines Agency approved the marketing of rimonabant in Europe. In 2007, however, new contraindications were formulated by the Agency. Also in 2007, the US Food and Drug Administration denied approval in the USA. Reasons for concern were neuropsychiatric side-effects, including depression and suicide attempts.
The clinical benefits of rimonabant were explored in the STRADIVARIUS and CRESCENDO trials.5, 6 STRADIVARIUS used intracoronary ultrasound to assess the effects of rimonabant on atheroma volume. Disappointingly, no effect on the primary outcome was observed, and the rate of neuropsychiatric side-effects was higher in the rimonabant group than in the control group (43·4% vs 28·4%). CRESCENDO was a large, randomised, placebo-controlled trial designed to test the efficacy of rimonabant in reducing the rates of cardiovascular events. In 2008, the European Medicines Agency halted the marketing authorisation for rimonabant in Europe because postmarketing registries and ongoing clinical trials suggested an increased risk of suicide.7 This led the sponsor to terminate CRESCENDO and all other research involving rimonabant.
In The Lancet today, the CRESCENDO investigators report their truncated trial.6 The primary efficacy outcome showed a hazard ratio of 0·97 (95% CI 0·841·12, p=0·68). Mean exposure to rimonabant was 13·8 months (at least 33 months and up to 50 months of study drug treatment were planned). In cardiovascular preventive drug-therapy, there tends to be a lag phase between the improvement in risk profile and the onset of protection from cardiovascular events. Therefore, in addition to reduced statistical power, early termination of the trial might have precluded a difference in the efficacy outcome. Serious psychiatric side-effects were more frequent with rimonabant than with placebo (2·5% vs 1·3%). Suicide attempts (nine patients on drug vs five on placebo), and completed suicides (four vs one, respectively) also occurred more frequently in the drug group. The investigators do not report whether the participants who committed suicide had depression at baseline. In view of the known neuropsychiatric side-effects of rimonabant, it is unclear why depression was not an exclusion criterion.
If CRESCENDO had been completed, a large improvement in cardiovascular outcome could have outweighed a small risk of serious adverse events, in a similar way to, for example, the large cardiovascular benefits of statins outweighing a small risk of rhabdomyolysis. However, any mortality associated with cardiovascular preventive therapy is generally viewed as unacceptable. The preventive approach is fundamentally different from curative therapy for a potentially lethal illness. Therefore the investigators' comparison between suicide risk during rimonabant use and intracranial haemorrhage during thrombolysis for myocardial infarction does not seem valid. Even if a net clinical benefit of rimonabant had been recorded in CRESCENDO, its external validity would have been doubtful. Although the investigators stressed that rimonabant should not be seen as a miracle drug for weight loss, the public would probably see it as such. Medical supervision of rimonabant use, including periodic screening for neuropsychiatric side-effects, is unlikely to be strict in daily practice. Also, if used for cosmetic weight-reduction, the drug might be used episodically rather than long term, as in CRESCENDO. On the basis of these arguments, the European Medicines Agency's decision seems appropriate. Similar arguments were made in the Agency's decision to suspend sibutramine, another weight-loss drug with a slightly increased risk of cardiovascular side-effects.8
The story of rimonabant might be tarnished by low public and regulatory acceptance for a drug designed to counteract the consequences of our abundant lifestyle. Focus should now return to motivating patients to control their caloric intake and increase physical activity. Although cumbersome, this approach is causal and safe.9 New strategies to more effectively achieve these lifestyle changes are needed.
We declare that we have no conflicts of interest.
  icon paper stack View Older Articles   Back to Top