Comments on FDA Obesity Trial Guidelines
April 12, 2007
Docket No. 2007D-0040
There is now evidence, at least for some drugs, of long-term benefits on morbidity and mortality for pharmacologic agents used to treat chronic conditions such as hypertension, or abnormally high lipid levels. Unfortunately, such crucial studies are lacking for obesity drugs. The current FDA-approved labeling of the two anti-obesity drugs approved for long-term use, orlistat (Xenical) and sibutramine (Meridia) even states this in the clinical trials sections saying, ”The long-term effects of [drug] on morbidity or mortality associated with obesity have not been established”. We hope that the revised guidelines will include a requirement that will finally provide this information.
The clinical significance of weight loss
Drug-induced weight loss in overweight and obese individuals has been shown to improve markers of cardiovascular risk factors (elevated blood pressure, LDL, and glucose), but that has not yet translated into a reduction in morbidity or mortality. The question remains as to whether the general overweight population would actually benefit from taking weight loss drugs long-term as this data does not yet exist. A meta-analysis of Douketis et al., cited as a reference in the guidelines, concludes that, “weight loss of ≥5% is not consistently associated with improvements in cardiovascular risk factors, and when improvements occur they are mainly in high-risk groups and appear to be intervention specific.”
The need for long-term clinical trials
Since obesity is a chronic condition and since patients tend to regain weight when taken off drug, treatment would be expected to be chronic (many years). Given these facts, clinical trials of one year’s duration, as recommended here, do not provide an adequate foundation for an evaluation of long-term safety and efficacy. Therefore, we think randomized, placebo-controlled trials should extend to two or three years, enabling one to answer some basic questions: What happens to patients over time if they continue on drug; is efficacy maintained or lost? Does safety appear to improve or to worsen? Do new serious adverse effects appear later? The time factor is especially important in children as the long-term consequences of drug intake are unique.
The trial size is being set such that it will pick up adverse events that occur at a frequency of ≥3% in the placebo group vs. 4.5% in treated patients. Thus, rare or unusual adverse events will be missed, and these may be the most important as they be the most serious. Significantly lengthening the trials would provide added sensitivity.
The guidelines state that the primary endpoint should be either a statistically significant mean weight loss of 5% vs. placebo or that at least 35% of patients lose ≥5% of baseline BW (and that the loss be statistically significant and double that of the placebo group). This endpoint seems fairly arbitrary based on the limited data in the meta analysis of Douketis et al.: in the 20 studies they analyzed for drug-induced weight loss, a high percentage of patients were lost to follow-up (usually 30%-50%), only 5 out of 20 studies were longer than one year, and only small numbers of patients were studied (most had about 200/group). The median weight loss should certainly be presented as well as the mean.
Lifestyle modification (diet and exercise) is to be the first step for patients, but it is not clear that these steps would continue and with how much guidance while on drug. There needs to be clarification as to how long this initial period of lifestyle modification would last, what duration would constitute a “sufficient” length of trial, what would be the criteria for failure, and whether this would continue while on drug. It is hard to see what the incentive would be for a manufacturer to succeed with lifestyle changes if it meant that the patient was not eligible for the trial. It is important that this aspect be defined, monitored, and encouraged for the duration of the trial, and beyond, in order for it to become a lifetime habit for patients. Taking a pill is not in itself a satisfactory approach to weight loss.
Standard of care and concomitant medications
A presumed benefit of drug treatment is to improve blood pressure, lipids, and glycemia, yet patients are receiving concurrent medications for these conditions. Furthermore, these medications are to be adjusted during the study; as a result, there needs to be clarification as to how this will be done without either breaking the blind or sorting out what interacting factors are at work.
General safety assessment of weight-management products
For drugs taken chronically, the roles of genotoxicity and carcinogenicity become increasingly important and need to be studied, especially when given to a young population. Genotoxicity and carcinogenicity were not mentioned in the safety assessment.
Weight-management products used in combination
This section focuses only on efficacy, but for long-term use, safety is at least as important. For this reason, treatment groups for the single agents alone might be more important than placebo in Phase 3 trials.
FDA should require carcinogenicity studies for the combination even if studies have been done with the individual drugs alone. It is not possible to predict what the effects of taking two (or more) drugs together might be as opposed to any one drug alone. There is no advantage in treating obesity if it ends up increasing the probability of tumors.
Weight management products for patients with medication-induced weight gain
In order to counter a weight gain due to a drug, it is necessary to have some knowledge of the mechanism of that weight gain. For the glitazones (rosiglitazone and pioglitazone), for example, the weight gain can be due to edema and/or fat deposition (for fat deposition, PPARgamma receptor activation induces the activity of genes that promote formation of and fat accumulation in adipocytes, part of the mechanism for how the drugs work).
Furthermore, since it is harder to lose weight than to prevent its gain, it may not make sense to wait until there is a 5% weight gain, as suggested in the guidelines. It would be helpful if there were good animal models for the two drugs taken together in order to get some ideas as to whether adding a second drug was effective and what the long-term consequences might be. Are there obese animal models and would they be required for these studies? If not, is it known whether normal weight animals respond the same as obese ones?
Weight management in pediatric patients
The comments pertaining to trials for adults apply here as well; certainly, one year is an even less adequate check on safety and efficacy. The drugs should be tested thoroughly first in young animals to see how they affect growth and development.
If treatment is to actually prevent development of type 2 diabetes, the time frame needs to be defined.
Elizabeth Barbehenn, Ph.D.
Sidney M. Wolfe, M.D.
Public Citizen’s Health Research Group