The agency that’s supposed to protect you from dangerous medicine sometimes approves it first. That’s not a conspiracy theory. That’s documented history, and it has happened more than once.

Introduction: The Trust We Place in a Three-Letter Agency

Most of us live by a simple assumption: if a drug is on pharmacy shelves, it’s safe. The FDA reviewed it. Scientists tested it. Regulators signed off on it. That means it’s fine, right?

Except that’s not always true. And this article is going to prove it to you.

The U.S. Food and Drug Administration is one of the most powerful regulatory bodies on the planet. Its approval is the golden ticket for pharmaceutical companies, the green light that transforms a lab molecule into a product sold to millions of people. We trust it with our lives, literally. Yet five times in relatively recent history, the FDA extended that golden ticket to drugs that turned out to be genuinely dangerous. Drugs that caused heart attacks. Drugs that triggered fatal liver failure. Drugs that destroyed lives before anyone sounded the alarm.

The recalls that followed were not quiet corrections. They were humiliations, lawsuits, congressional investigations, and in some cases, criminal charges. They raised uncomfortable questions about whether the system designed to protect American patients is, in some structural ways, set up to fail them.

This is not a hit piece on the FDA. The agency does enormous good and protects people from countless harmful products every year. But understanding where it has failed, and why, is essential if we are ever going to demand a better system. So let’s pull back the curtain.

These are five dangerous drugs the FDA approved and later recalled, and the shocking inside story of how it happened each time.

1. Vioxx: The FDA Approved a Painkiller That May Have Caused 60,000 Deaths

Vioxx, the brand name for rofecoxib, was Merck’s blockbuster answer to arthritis pain. When the FDA approved it in 1999, it was celebrated as a breakthrough. Here was a painkiller that worked as well as older drugs like ibuprofen but, crucially, was easier on the stomach. For people with chronic pain and sensitive digestive systems, it felt like a genuine gift.

At its peak, Vioxx was prescribed to more than 20 million Americans. Merck promoted it aggressively, and doctors prescribed it enthusiastically. The FDA had approved it. What could go wrong?

As it turned out, quite a lot.

What the data was saying before the approval:

Internal Merck studies, later revealed during litigation, showed signs of cardiovascular risk years before the recall. A landmark clinical trial called VIGOR, completed in 2000, showed that patients on Vioxx were five times more likely to suffer a heart attack compared to those taking naproxen. Merck’s explanation at the time was that naproxen must be cardioprotective, not that Vioxx was dangerous. The FDA accepted this reasoning, at least partially, and simply required an updated label warning.

That was not enough.

By 2004, a large long-term safety study called APPROVe confirmed what the warning signs had been telegraphing for years. Vioxx significantly increased the risk of serious cardiovascular events, including heart attacks and strokes, in patients who took it for eighteen months or more. Merck voluntarily withdrew the drug from the market in September 2004.

The aftermath was staggering. FDA scientist Dr. David Graham, in a Senate testimony that sent shockwaves through Washington, estimated that Vioxx had caused between 88,000 and 139,000 heart attacks, of which 30 to 40 percent were likely fatal. That placed the death toll somewhere in the range of 26,000 to 55,000 Americans. Other estimates pushed the figure higher.

Why did the FDA approve it in the first place?

This is where it gets uncomfortable. The FDA’s approval process relied heavily on data submitted by Merck itself. That is not unusual. Pharmaceutical companies fund the clinical trials that form the basis of every drug approval. The FDA reviews the data, but it rarely conducts independent trials. Critics have long argued that this creates a structural conflict of interest.

David Graham’s Senate testimony went further. He argued that the FDA had developed a culture that prioritized approving drugs over protecting patients, partly because its budget had become partially dependent on “user fees” paid by pharmaceutical companies seeking approval. When the regulator is partly funded by the industry it regulates, objectivity becomes structurally complicated.

Merck eventually settled lawsuits totaling approximately $4.85 billion. It was one of the largest pharmaceutical settlements in history.

The legacy: Vioxx became the defining case study for post-market surveillance failure. It changed the conversation about how the FDA monitors drugs after they reach the market, and it remains the most cited example of why “approved” does not always mean “safe.”

2. Fen-Phen: The FDA Approved a Diet Drug Combo That Damaged Hearts

Few pharmaceutical stories capture the desperation of America’s relationship with weight loss quite like fen-phen. The nickname combined two drugs: fenfluramine and phentermine. When used together, they produced remarkable weight loss results, and throughout the early 1990s, the combination became one of the most popular off-label prescriptions in America.

Here’s the thing: the FDA had approved each drug individually, but it had never approved them as a combination. Doctors were prescribing them together based on a 1992 clinical study that showed impressive results. The FDA was aware the combo was being used, but it had not formally blessed the pairing.

Then, in 1996, the FDA approved a new version of the fenfluramine component called dexfenfluramine, sold under the brand name Redux, which was the most popular formulation of the pair. By that point, an estimated 6 million Americans had used fen-phen or Redux.

The cardiac connection nobody wanted to find:

In July 1997, a team of researchers from the Mayo Clinic published an alarming report in the New England Journal of Medicine. They had identified 24 women, most of whom had no prior heart disease risk factors, who developed unusual heart valve abnormalities after taking fen-phen. The damage was specific and distinctive. The valves looked like those seen in patients with carcinoid tumors, a completely unrelated condition.

The FDA moved quickly after that report, commissioning an emergency survey. Results came back showing that roughly 30 percent of patients who had been on the drug combination for an average of twelve months had abnormal echocardiograms, even if they had no symptoms. The FDA did not wait for more data.

Both fenfluramine and dexfenfluramine were pulled from the market in September 1997. Phentermine, which had not been implicated in the cardiac findings, remains available today.

Why did this happen?

The fen-phen case highlights a different kind of regulatory failure: the problem of combination use. The FDA approved these drugs individually, but the real-world clinical environment combined them. By the time the drug combination was ubiquitous, the FDA had an enormous political problem on its hands. Pulling approval retroactively is not straightforward. The public loved fen-phen. Doctors were prescribing it widely. There was no formal mechanism to withdraw an unapproved combination.

The manufacturer of Redux, Wyeth (now part of Pfizer), paid out more than $21 billion in settlements and judgments over the following years. It was one of the costliest pharmaceutical liability disasters in history.

The legacy: Fen-phen became a cautionary tale about off-label use, combination therapies, and what happens when commercial popularity outpaces safety vigilance. It also demonstrated how quickly the FDA can act when the evidence is unambiguous, which raises the question of why the same speed is not always applied earlier in the process.

3. Rezulin: The FDA Kept a Liver-Destroying Diabetes Drug on the Market While Europe Pulled It

Rezulin, the brand name for troglitazone, was approved by the FDA in January 1997 for the treatment of Type 2 diabetes. It was a novel drug in a new class called thiazolidinediones, drugs that work by improving the body’s sensitivity to insulin. For a country where diabetes was already reaching epidemic proportions, a new mechanism of action was genuinely exciting.

The FDA approval was, in hindsight, rushed.

The agency approved Rezulin in just 72 days under an expedited review program designed to fast-track drugs that addressed serious conditions with limited treatment options. The urgency was understandable. But speed and thoroughness are not always compatible.

The liver failure problem:

Before the FDA’s approval was even finalized, there were signs of trouble. During clinical trials, a small number of patients developed elevated liver enzymes, a classic warning signal of liver stress or damage. The FDA required Merck-Parke Davis (the manufacturer) to include liver monitoring warnings on the label, but it approved the drug.

Over the following two years, reports of serious liver damage began accumulating. Patients were developing acute liver failure, a condition that can be rapidly fatal. By 1997, Britain’s equivalent of the FDA, the Medicines Control Agency, had seen enough. It withdrew Rezulin from the UK market.

The FDA did not follow. Instead, it required more label updates. Additional liver monitoring requirements were added. The warnings got stronger. The drug stayed on shelves.

Japan pulled it in 1999. The FDA still did not act.

It was not until March 2000, three years after the British withdrawal, that the FDA finally removed Rezulin from the U.S. market. By that point, the drug had been linked to at least 90 deaths from liver failure in the United States, with many researchers putting the true figure considerably higher. An estimated 1.9 million Americans had taken it.

Why did the FDA wait so long after Europe acted?

A Los Angeles Times investigative series published in 2000 examined this question in exhaustive detail. Their reporting suggested that FDA medical officer Dr. John Gueriguian had recommended against approving Rezulin back in 1996, citing the liver enzyme data. He was removed from the review team. His replacement approved the drug.

The reporting also suggested that after the British withdrawal, pressure from both the manufacturer and physicians who believed the drug was effective for certain patients contributed to the FDA’s reluctance to act decisively. The agency’s post-market safety review process was, critics argued, too deferential to the drug’s sponsor and too slow to respond to international signals.

The legacy: Rezulin’s story is particularly damning because it shows not just an approval failure, but a failure of ongoing surveillance and response. The FDA had multiple opportunities to act. International peers acted. The FDA waited, and people died. The case directly influenced later reforms to the FDA’s post-market safety review process.

4. Baycol: A Cholesterol Drug That Destroyed Muscles and Killed Patients

By the late 1990s, statins had become among the most prescribed drugs in the world. The evidence for their ability to lower LDL cholesterol and reduce cardiovascular risk was overwhelming. So when Bayer AG introduced cerivastatin, sold under the brand name Baycol in the U.S. and Lipobay in Europe, in 1997, the FDA approved it as part of a well-established, trusted class of medications.

The thinking was logical. Statins as a class had a good track record. Cerivastatin was a new statin. It should be similarly safe.

That logic proved to be fatally flawed.

Rhabdomyolysis: the muscle-destroying side effect:

All statins carry a small risk of a condition called rhabdomyolysis, a breakdown of muscle tissue that releases proteins into the bloodstream and can cause severe kidney damage and death. For most statins, this risk is rare and manageable. Baycol, it turned out, carried a dramatically higher risk than its competitors, particularly at higher doses and when combined with a fibrate drug called gemfibrozil, another cholesterol-lowering medication that many patients were taking simultaneously.

The FDA began receiving reports of rhabdomyolysis cases in Baycol patients. In August 2001, Bayer voluntarily withdrew Baycol from the market worldwide. By that point, the drug had been linked to at least 31 deaths in the United States alone, and more than 100 deaths globally. Thousands of patients had suffered serious muscle damage.

Why did the FDA not catch this earlier?

The Baycol case underscores the limits of pre-approval clinical trials. Those trials were not large enough, long enough, or structured in a way that would have reliably detected a relatively rare but serious interaction effect. The combination of cerivastatin with gemfibrozil, common in clinical practice, was not adequately studied before approval.

This speaks to a fundamental challenge in drug regulation. Clinical trials are designed to demonstrate efficacy and identify common side effects. They are not well-designed to detect rare adverse events, particularly those that emerge from drug-drug interactions that are common in the real world but not always replicated in controlled trial settings.

Bayer paid more than $1.2 billion to settle thousands of lawsuits related to Baycol injuries and deaths. The company maintained that it acted responsibly by withdrawing the drug voluntarily once the risk profile became clear.

The legacy: Baycol intensified the push for better pharmacovigilance systems and larger post-market safety studies. It also raised important questions about whether a new drug in an established class should face a higher evidentiary bar for safety, rather than benefiting from an assumption that it shares the safety profile of its predecessors.

5. Propulsid: A Heartburn Drug That Caused Fatal Heart Arrhythmias in Children

Cisapride, sold under the brand name Propulsid, was approved by the FDA in 1993 for the treatment of gastroesophageal reflux disease, better known as GERD or acid reflux. It worked by increasing the movement of food through the digestive tract. For patients with severe reflux, it was genuinely helpful.

The drug was also widely prescribed off-label for infants and children with feeding problems and reflux, a population for whom the original clinical trials had not been conducted.

The heart rhythm problem:

Cisapride is what pharmacologists call a QT-prolonging drug. QT prolongation refers to an abnormality in the electrical cycle of the heart that can, in some patients, trigger a type of dangerous arrhythmia called torsades de pointes, which can lead to sudden cardiac death. The risk is particularly elevated in people who have other QT-prolonging conditions or who are taking other drugs that also prolong the QT interval.

The FDA knew about this risk before approval. The original label included warnings about cardiac risks. But the warnings were not strong enough, and they were not effectively communicated to the prescribers who were using the drug in children.

By the late 1990s, the FDA had received reports of more than 270 cases of serious cardiac arrhythmias and 70 deaths associated with cisapride. A disproportionate number involved infants and children. In 2000, Johnson and Johnson voluntarily withdrew Propulsid from the general market, though it continued to be available through a restricted access program for adult patients with severe conditions who had no other options.

Why were children particularly at risk?

Children metabolize drugs differently than adults. Their hearts are more sensitive to QT-prolonging effects. The off-label use of Propulsid in infants was based on limited data from small studies, and pediatric dosing guidelines were inconsistent. Parents and pediatricians were administering a drug with known cardiac risks to vulnerable patients based on extrapolated adult data and anecdotal evidence.

The FDA’s role here is complicated. It cannot legally prohibit physicians from prescribing approved drugs off-label. But it can require stronger warnings, restrict distribution, and communicate risks more forcefully to prescribers. Critics argued that the FDA was too slow and too gentle in all three areas.

The legacy: The Propulsid case became a landmark in the debate about pediatric drug testing. It directly contributed to the Best Pharmaceuticals for Children Act of 2002 and the Pediatric Research Equity Act of 2003, which together created new incentives and requirements for pharmaceutical companies to study their drugs specifically in pediatric populations. It was a concrete policy change born from a documented failure.

Why Does the FDA Keep Approving Dangerous Drugs? The Structural Answers

Reading these five stories in sequence, a reasonable person might ask: is this just bad luck, or is there something broken in the system? The honest answer is that there are several structural features of the FDA approval process that make these failures not just possible but, arguably, predictable.

The user fee problem:

Since the Prescription Drug User Fee Act of 1992, the FDA has collected fees from pharmaceutical companies for each new drug application it reviews. This funding now accounts for a significant portion of the FDA’s drug review budget. The arrangement has clear benefits: it funds a larger, faster review process. But it also creates a dynamic where the FDA’s operational funding depends on approving drugs. An agency that rarely approves drugs collects fewer fees. This is not a conspiracy; it is a structural incentive, and incentives shape behavior in ways that are often invisible to the people inside the system.

The pre-market vs. post-market imbalance:

The FDA devotes enormous resources to pre-market review, the process of evaluating clinical trial data before a drug is approved. It devotes comparatively fewer resources to post-market surveillance, the ongoing monitoring of drugs after they enter widespread use. Yet as the Vioxx, Rezulin, and Baycol cases all demonstrate, the most important safety data often emerges only after millions of real-world patients with diverse health profiles and drug combinations have taken a medication for years.

The manufacturer-funded trial problem:

Every drug that goes through the FDA approval process is supported by clinical trial data funded by the company that wants to sell the drug. This is standard practice globally, not an American peculiarity. But it means that the FDA is, by design, evaluating data produced by parties with a financial interest in a particular outcome. Independent replication of drug trials before approval is rare.

The revolving door:

A 2018 study published in the British Medical Journal examined the career trajectories of FDA medical officers and found that a substantial proportion left the agency to work for the pharmaceutical industry. This is legal and common. It is also a dynamic that critics argue may subtly influence decision-making within the FDA, as regulators may be reluctant to develop adversarial relationships with companies they may later wish to work for.

The Human Cost: A Data Snapshot

The following table summarizes the five drugs discussed, their approval dates, recall dates, estimated affected patients, and known death tolls. These figures come from peer-reviewed literature, congressional testimony, and investigative journalism. Where exact figures are disputed, ranges are provided.

These numbers are not abstractions. Each represents a person who trusted the system and was failed by it. A parent who gave their infant a heartburn medication based on a doctor’s recommendation. A diabetic patient who took a pill their physician prescribed in good faith. A chronic pain sufferer who wanted to live normally.

The combined human cost of these five approvals is measured in tens of thousands of deaths and hundreds of thousands of injuries. The combined financial cost in litigation and settlements exceeds $28 billion.

What Has Changed Since These Disasters?

It would be unfair to stop the story at the failures without acknowledging the genuine reforms that followed. The FDA today is, in important ways, a different institution than it was in the 1990s when several of these approvals were made.

The FDA Amendments Act of 2007 gave the agency new authority to require post-market safety studies, mandate risk evaluation and mitigation strategies (REMS), and issue safety communications more rapidly. It was a direct legislative response to Vioxx.

The Sentinel System, launched in 2008, created an active surveillance network that allows the FDA to monitor drug safety signals across large electronic health databases covering hundreds of millions of patients. It has enabled the detection of safety signals far faster than the old passive reporting system.

Pediatric testing requirements were significantly strengthened in the early 2000s, as noted in the Propulsid discussion. More drugs are now formally tested in children before being marketed to them.

Transparency requirements have improved, with clinical trial registration and results reporting now mandatory under ClinicalTrials.gov. This makes it harder for unflattering trial data to disappear quietly into a filing cabinet.

These are meaningful improvements. But critics, including many within the FDA itself, argue that the fundamental tension between speed and safety remains unresolved. The approval process still relies heavily on manufacturer-funded data. Post-market surveillance, while better, remains under-resourced relative to pre-market review. And the political and commercial pressure to approve drugs quickly has not diminished.

How to Protect Yourself: What Patients Can Actually Do

The uncomfortable truth is that no individual patient can fully compensate for systemic failures in drug regulation. You cannot personally replicate clinical trials. But you can make more informed decisions.

Ask about the drug’s approval timeline. Drugs approved under expedited review programs like Breakthrough Therapy Designation, Accelerated Approval, or Priority Review have been evaluated with less complete data than standard approvals. That does not make them bad. It means the safety profile is less certain. Ask your doctor whether the drug has had post-market safety issues.

Check the FDA’s drug database. The FDA maintains a publicly accessible database of drug safety communications and recalls at fda.gov. If you are taking a medication you are uncertain about, you can search for recent communications about it.

Ask about alternatives. As several of these cases show, dangerous drugs often had safer alternatives available. Vioxx was competing with older NSAIDs that, while imperfect, did not carry the same cardiac risk. Asking your doctor whether there is a longer-established alternative with a more complete safety record is always reasonable.

Report side effects. The FDA’s MedWatch system allows patients and healthcare providers to report suspected adverse drug reactions. The post-market surveillance systems that eventually flagged Vioxx, Baycol, and others relied heavily on these voluntary reports. Your report could contribute to detecting a safety signal that protects future patients.

According to a study analyzing FDA drug safety communications, post-approval safety issues are identified in roughly one-third of all new molecular entities within ten years of their approval. That figure should not terrify you into avoiding necessary medications. But it should permanently retire the notion that “FDA approved” is a guarantee of safety rather than a statement of probability.

The Pharmaceutical Industry’s Role: Villain, Partner, or Both?

It would be convenient to cast pharmaceutical companies as straightforward villains in these stories. The reality is messier.

Drug development is extraordinarily expensive. A successful drug costs, by various estimates, between $1 billion and $2.5 billion to develop from molecule to market. Companies that invest that money have a legitimate interest in recouping it. Patents expire. Competition is fierce. The pressure to bring drugs to market quickly and to present data in the most favorable light is not a moral failing unique to bad actors. It is a rational response to the financial architecture of the industry.

That said, the Vioxx case in particular revealed behavior that went beyond commercial enthusiasm. Internal documents shown during litigation suggested that Merck executives were aware of the cardiovascular signal before it was publicly disclosed, and that the company worked actively to minimize or reframe that signal. Several scientists who raised concerns internally described being sidelined.

That is not a systemic problem. It is a cultural one. And it points to the need for regulatory systems that do not rely on the good faith of commercial entities with enormous financial stakes in specific outcomes.

The World Health Organization’s essential medicines framework and similar international frameworks attempt to create reference points for drug safety and efficacy that are independent of commercial interests. Advocates argue that the U.S. should move further in this direction, including through greater public funding of independent clinical trials.

The Global Comparison: How Other Countries Handle Drug Safety

The United States is not the only country that has approved drugs later found to be dangerous. But it has sometimes been slower than its international peers to act on emerging safety signals, as the Rezulin case illustrates vividly.

The European Medicines Agency takes a slightly different approach in several respects. It operates with a more committee-driven review process, incorporating input from member states’ national agencies. It has, in several high-profile cases, been faster to withdraw or restrict problematic drugs than the FDA.

Canada’s Health Canada agency operates with greater transparency in some areas, publishing more detailed reasons for both approvals and refusals.

This is not to suggest that foreign regulators are uniformly superior. Each system has its own failure modes, its own commercial pressures, and its own track record of both successes and mistakes. But the comparative perspective is useful because it demonstrates that the specific structural choices embedded in the FDA’s system are choices, not inevitabilities. Different designs produce different outcomes.

The Question Nobody Wants to Answer

Here is the question that hovers over all five of these case studies: if we know the structural features that contribute to these failures, why have they not been more comprehensively addressed?

The answer is political. Pharmaceutical companies are among the most powerful lobbying forces in Washington. The industry spends hundreds of millions of dollars annually on federal lobbying. Drug approval reform that would slow the pipeline or increase the cost of development is, predictably, vigorously opposed.

Patient advocacy groups, which might be expected to align with safety-first regulatory approaches, are complicated. Many receive pharmaceutical industry funding. Some have argued for faster approval processes for drugs targeting serious diseases, prioritizing access over certainty. These are not unreasonable positions, but they complicate the reform landscape.

Academic researchers who study drug safety often depend on grant funding from sources that include pharmaceutical companies, creating conflicts of interest that, while not determinative, add complexity to the policy conversation.

The result is a system that has improved at the margins, that has genuinely incorporated important lessons from disasters like Vioxx and Rezulin, but that has not fundamentally restructured the relationship between the FDA and the industries it regulates.

Conclusion: Approved Is Not the Same as Safe

The five stories in this article share a common thread. They are not stories about rogue scientists or corrupt officials, though elements of both appear in some cases. They are stories about a system that was designed under certain assumptions: that companies would share all relevant data, that clinical trials would capture important safety signals, that post-market monitoring would catch what pre-market review missed.

Those assumptions have repeatedly proven optimistic.

The FDA is not a broken institution. It protects Americans from thousands of harmful products every year. The drugs that cause serious harm represent a small fraction of the medications it approves, and the conditions those drugs were designed to treat were real and sometimes severe.

But “better than nothing” is not the same as “good enough.” When 20 million people take a painkiller that increases their heart attack risk, and the FDA knew about early warning signs and accepted a manufacturer’s reassurance rather than demanding more data, that is a failure worth examining seriously.

The lesson is not that you should refuse medications your doctor prescribes. Modern medicine is extraordinarily powerful, and the suffering that goes untreated because of excessive medication fear is real and quantifiable.

The lesson is that you are entitled to ask questions. That “the FDA approved it” is the beginning of a conversation, not the end of one. That the history of drug regulation is, among other things, a history of human beings making consequential decisions under uncertainty, with commercial pressures tugging in one direction and public health interests pulling in another.

The people who died from Vioxx, from Rezulin, from Propulsid, were not naive or careless. They were patients who trusted a system. The least we can do is understand how that system works, where it has failed, and what we can demand from it going forward.

CTA: Your Next Step

If this article opened your eyes to the complexity behind drug approvals, share it with someone you know who takes prescription medications. This is information every patient deserves to have.

Drop a comment below: Which of these five cases surprised you most, and what do you think the most important reform to the FDA approval process would be?

This article is intended for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare provider before making decisions about your medications.