The history of obesity drugs — from amphetamines to Ozempic

That demand has driven a century of pharmaceutical ambition, repeated failure, and occasional genuine breakthrough. Understanding the history of obesity pharmacology is not just a history lesson. It is the context within which today's GLP-1 drugs acquire their real significance — and the context that makes the pharmaceutical industry's cautious enthusiasm for them more comprehensible than the cultural conversation sometimes suggests.

The amphetamine era began in the 1930s and ran, in various forms, for four decades. Smith, Kline and French launched Benzedrine in 1937 — initially as an over-the-counter inhaler for nasal congestion, quickly repurposed when users discovered its appetite-suppressing and energy-enhancing effects. Amphetamines work through norepinephrine and dopamine release, which suppresses appetite reliably but comes packaged with cardiovascular strain, dependence potential, and the crash that follows the drug's duration of action. Prescriptions for amphetamine-based weight loss drugs ran in the millions through the 1950s and 1960s. The harm was gradual — dependency built slowly, cardiovascular events didn't cluster obviously enough to trigger immediate regulatory action — and the cultural framing was of a modern convenience, a pill for the problem of too much appetite. The reckoning came eventually. By the early 1970s, the DEA had scheduled amphetamines, their use contracted, and the era of unambiguous stimulant prescription for obesity was largely over. Phentermine, a related but less potent and less addictive stimulant, remained — it's still FDA-approved for short-term weight management today, which gives it one of the longer FDA histories in the weight loss category, even if it's rarely discussed as the legacy amphetamine drug it structurally resembles.

Fenfluramine arrived in the 1970s as a different mechanism: it released serotonin rather than dopamine and norepinephrine, producing appetite suppression through a different pathway and with a different side effect profile than amphetamines. Modest efficacy, modest use. Then, in 1992, a Rochester endocrinologist named Michael Weintraub published a clinical trial showing that combining fenfluramine with phentermine — fen-phen — produced weight loss substantially greater than either drug alone, with an apparently tolerable side effect profile over the trial period. The study landed in a medical environment primed to receive it. By 1996, roughly 18 million Americans had been prescribed fen-phen. Prescriptions were written not by weight management specialists but by general practitioners responding to patient demand, with a duration and a patient population that extended far beyond what the original trial had evaluated.

In 1997, cardiac echocardiograms in a Mayo Clinic case series identified valvular heart disease — specifically aortic and mitral regurgitation — in women who had been taking fen-phen. Twenty-four cases initially. The FDA issued a public health advisory. More cases were identified. Within weeks, the manufacturers withdrew fenfluramine and dexfenfluramine voluntarily. Phentermine, not associated with the valvulopathy, remained. The mechanism was serotonin's effect on cardiac valve tissue — the same pathway that had made the drug an appetite suppressant at the brain level was toxic to valve leaflets at the periphery. The FDA later received reports linking fen-phen to pulmonary arterial hypertension as well. The litigation extended for years. The episode taught the FDA and the industry that weight loss drugs prescribed at population scale, to patients who might not have been served by specialists or monitored for cardiovascular endpoints, could produce harms that only became visible when millions of people had been exposed.

The shadow of fen-phen fell on every obesity drug development program that followed it. It fell heavily. Sibutramine (Meridia) reached the US market in 1997, the same year fen-phen was withdrawn. It worked by blocking the reuptake of norepinephrine and serotonin — mechanistically more like an SNRI antidepressant than an amphetamine, and with a side effect profile that included modest blood pressure increases but seemed initially manageable. Abbott Laboratories ran it as a mainstream obesity treatment through the late 1990s and 2000s. The SCOUT trial, a cardiovascular outcomes study in high-risk patients, found that sibutramine increased the risk of nonfatal myocardial infarction and stroke. The FDA requested voluntary withdrawal in 2010. Gone.

Rimonabant wrote a different cautionary chapter. The endocannabinoid system's role in appetite had been recognized from the observation that marijuana reliably produced the munchies — so the question was what would happen if you blocked the receptor. Sanofi-Aventis developed rimonabant as a CB1 receptor antagonist and it worked as predicted on body weight, reducing it by amounts that were clinically meaningful and accompanied by favorable metabolic markers. It was approved in Europe as Acomplia in 2006. It was never approved in the United States — the FDA declined, and the reason was psychiatric. Blocking CB1 receptors in the brain doesn't just suppress appetite; it produces anxiety, depression, and in a subset of users, suicidal ideation. The European Medicines Agency withdrew its approval in 2008. Several other CB1 antagonists in development were quietly terminated. Rimonabant had been a genuine pharmacological insight — the mechanism was valid, the weight loss was real — but the brain is not a modular system where you can block an appetite pathway without reaching adjacent ones.

Orlistat was the exception to the pattern of cardiovascular and psychiatric failures, and it remains the only obesity drug to survive the entire subsequent period without being withdrawn. It works by inhibiting pancreatic lipase in the gut, blocking absorption of approximately thirty percent of ingested fat. The mechanism is local rather than central — it doesn't act on the brain at all — which is why it doesn't produce the cardiovascular or psychiatric effects that killed other drugs. What it does produce is celebrated in an unflattering way by anyone who has read an orlistat side-effect disclosure: the unabsorbed fat has to go somewhere, and it goes through the GI tract with considerable urgency and limited warning. The clinical term is fecal incontinence. The informal terms are less clinical. Orlistat received prescription approval as Xenical in 1999 and over-the-counter approval as alli in 2007. Its efficacy is modest — average weight loss of roughly three to four kilograms over placebo in trials — and adherence is poor, for reasons that do not require elaboration. It persists on the market not because it's excellent but because it's safe, and in this therapeutic category, safety without cardiovascular or psychiatric liability is a genuine distinguishing feature.

The 2010s produced a cluster of approvals that came after a long drought. Phentermine-topiramate extended-release (Qsymia, FDA-approved 2012) paired the old stimulant with an anticonvulsant that had been observed to cause weight loss as a side effect. The combination produced greater efficacy than either agent alone, and the cardiovascular profile appeared acceptable. The FDA approved it with a Risk Evaluation and Mitigation Strategy due to topiramate's teratogenicity. Naltrexone-bupropion extended-release (Contrave, 2014) combined an opioid antagonist with an atypical antidepressant; the combination's effect on hypothalamic and reward circuits produced modest average weight loss with a side effect profile dominated by GI symptoms and a blood pressure concern that required a cardiovascular outcomes trial. Liraglutide at 3 mg (Saxenda, 2014) was the first GLP-1 receptor agonist approved specifically for chronic weight management — at a higher dose than its diabetes indication required and with daily injection rather than once-weekly administration. The weight loss data were meaningful but not transformative by later standards — average of around eight percent body weight versus placebo. The field had options it had not had before, but nothing that changed the fundamental clinical math.

Then semaglutide at 2.4 mg weekly (Wegovy, FDA-approved June 2021). The STEP 1 trial published in the New England Journal of Medicine showed average body weight reduction of 14.9 percent over sixty-eight weeks versus 2.4 percent for placebo. This was a different number. Not in the range of existing pharmacotherapy. In the range of bariatric surgery — the intervention that, for decades, had been the only thing that reliably produced and maintained that magnitude of weight loss. The SELECT cardiovascular outcomes trial, published in 2023, showed that semaglutide at the obesity dose reduced major adverse cardiovascular events in adults with obesity and established cardiovascular disease by twenty percent. The drug was not just producing weight loss. It was producing cardiovascular benefit beyond what weight loss alone would predict, suggesting direct vascular effects of GLP-1 receptor agonism. The clinical calculus shifted in ways that the regulatory and insurance industries are still adjusting to.

Tirzepatide, a dual GIP and GLP-1 receptor agonist (Zepbound, FDA-approved for obesity November 2023), produced SURMOUNT trial data showing average body weight reductions of twenty to twenty-two percent — exceeding semaglutide and approaching outcomes seen with Roux-en-Y gastric bypass. The mechanism of the additional efficacy is still being characterized: GIP receptor agonism appears to complement GLP-1 receptor agonism through effects on adipose tissue and potentially through enhanced neurological satiety signaling. Whatever the mechanism, the numbers were different from anything the obesity pharmacology field had generated before, from any molecule since fen-phen or DNP — and those were removed because they killed people.

The pipeline behind these drugs reflects an industry that now believes obesity pharmacology can work. Retatrutide, a triple agonist at GIP, GLP-1, and glucagon receptors, produced Phase 2 data with average weight reductions exceeding twenty-four percent. Cagrilintide, an amylin analog being developed in combination with semaglutide as cagri-sema, has produced encouraging early data. Survodutide and mazdutide are in various trial stages. Oral formulations of semaglutide and other GLP-1 agonists are in development, which would remove the injection requirement that creates adherence barriers for some patients.

What the obesity drug history teaches is not comfortable. It teaches that the pharmaceutical industry has repeatedly put drugs into millions of bodies with insufficient evidence of long-term safety, that regulators have sometimes caught the problem late, and that the demand for effective weight management pharmacology has been strong enough to outrun caution in ways that caused real harm. DNP. Amphetamines. Fen-phen. Sibutramine. Rimonabant. The graveyard of withdrawn obesity drugs is not evidence that obesity pharmacology is inherently doomed — it's evidence that the specific mechanisms those drugs targeted were either toxic at scale or incompatible with the brain's deeply integrated appetite, reward, and affective systems.

GLP-1 receptor agonism appears to have found a mechanism that avoids those failure modes. The physiological pathway it engages is one the body uses naturally — these drugs amplify a signal that already exists, rather than hijacking a parallel system or antagonizing a receptor with broad effects. The safety data accumulating from post-marketing surveillance of semaglutide across many millions of patient-years is not revealing the kind of cardiovascular or psychiatric signals that ended earlier drugs. That is not a guarantee. It is a meaningful observation about a class that has more clinical time behind it than any obesity drug since orlistat. The history of this field is a history of enthusiasm followed by harm followed by withdrawal. The current generation of drugs has not followed that pattern. Whether they eventually will, or whether they represent the genuine pharmaceutical reckoning with obesity that the field has been trying to produce since the 1930s, is a question that cannot yet be answered with certainty — but is being asked with better evidence than any previous generation of obesity pharmacology could offer.