Systematically Excluding Therapies from a Systematic Review

This is puzzling. When we saw a new systematic review of therapies for childhood obesity, we were intrigued. It popped up in JAMA Network Open on Friday, clearly focused on central (abdominal) obesity in youth. The authors used waist circumference (WC), waist-to-height ratio, waist-to-hip ratio, and WC z-score as measures of this.

But there is a gaping hole in this analysis.

Excluding Pharmacotherapies That Work

The puzzling thing is that, though the authors include pharmacotherapy in their review, they don’t include most of the pharmacotherapies that pediatric obesity medicine clinicians actually use. In fact, they include only three medications in their analysis: orlistat, metformin, and fluoxetine. At best, these drugs have a modest effect on obesity. Orlistat is indicated, but not frequently used. Metformin, albeit not indicated for obesity, is commonly used for this purpose. Fluoxetine (Prozac) is neither indicated nor commonly used. The evidence that it even works for this purpose is of “low certainty.”

Meanwhile, the paper makes no mention whatsoever of a number of really effective therapies with an indication for treating obesity in pediatric patients. These are phentermine, phentermine-topiramate, liraglutide, and semaglutide.

Sweeping Conclusions

So, because they exclude effective drugs from their systematic review of obesity therapies, the authors feel free to conclude that pharmacotherapy does not work. In making their conclusion, they do not bother to qualify it by specifying the marginally effective drugs they considered. They applied it to pharmacotherapy broadly.

In contrast, they conclude that behavioral therapy or combinations of dietary and physical activity interventions “significantly reduced central obesity.” But they do not mention the fact that the modest effects they documented, though statistically significant, may not be clinically significant.

No Look, No See

The lesson here is that when people don’t look for something, they will not see it. These authors did not look for effective pharmacotherapy, so they did not find it.

What’s more, they did not look for clinically meaningful effects on pediatric obesity. After reviewing this paper, one of the world’s leading pediatric obesity experts, Aaron Kelly, summed it up for us succinctly:

“Yet another study showing modest and clinically insignificant treatment effects with lifestyle and behavioral treatment approaches.”

This study reflects a mindset stuck on viewing obesity as a problem of bad behavior. Sadly, that remains common, even in an era of better medical insights into obesity.

Click here for the study in JAMA Network Open. For perspective on the benefits of truly effective obesity care in childhood, click here.

Gears, photograph by Thomas Claveirole, licensed under CC BY-SA 2.0

Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.

April 15, 2025