Pharmaceuticals: Perils for pill pushers
By Andrew Jack
Published: September 21 2010 21:32 | Last updated: September 21 2010 21:32
When the US Food and Drug Administration presented an award last week to honour the 96-year-old Frances Kelsey, a former employee who helped save America from thalidomide, the mood in the agency’s corridors was not purely one of celebration.
Back in the 1960s, her actions blocked US approval of the morning sickness pill that caused thousands of children in Europe to be born with birth defects. Yet 50 years later, in the agency in Washington and at its counterpart in London, the European Medicines Agency, officials still fret over such life-and-death decisions.
In the coming days, they must choose whether to withdraw from the market GlaxoSmithKline’s top-selling Avandia diabetes drug, just one of a number of recent “blockbuster” medicines that have been subject to intensifying – and fiercely divergent – opinions over safety.
Critics such as Steven Nissen, a cardiologist at the Cleveland Clinic in the US who three years ago helped stoke public and political concern – and a sharp drop in prescriptions – by suggesting that the medicine triggered heart attacks, says: “Avandia is the worst drug safety catastrophe in our lifetime. It’s really a mess.”
To others, the regulators have done an increasingly effective job over the years in protecting patients, but are coming under growing pressure to gather and release ever greater amounts of data. Hans-Georg Eichler, senior medical officer at the EMA, says: “We are scrutinising drugs far more thoroughly than 10 or 20 years ago. I would like to think we have become much better at approving safer and more effective drugs.”
Squeezed in the middle is the pharmaceutical industry, the engine of drug development, which is seeing its business model put under severe strain by costly new requirements, pressures on pricing and the limitations of scientific progress.
Increasingly working with their foreign counterparts and in ever earlier consultation with pharmaceutical companies, regulators set standards for safety and efficacy for new drugs, authorise clinical trials to test them against and decide whether they meet the thresholds for approval.
They continue to process data on side effects, and weigh the risks and benefits, withdrawing drugs or changing authorised uses. Doctors may prescribe beyond these uses.
Thalidomide marked a turning point in the history of drug regulation, leading authorities around the world to impose higher approval standards to ensure drugs were tested for safety as well as efficacy.
A more recent landmark came in 2004, when Merck of the US withdrew its blockbuster painkiller Vioxx after concerns that it was causing heart attacks in patients. That heralded a fresh period of political criticism of – and introspection among – regulators. The result has been a more cautious “safety first” approach that has further raised the bar for approval.
Vioxx was highly effective in treating pain in some patients and avoided the stomach bleeding associated with the previous generation of drugs. Critics claim data on the drug’s side effects were played down, while aggressive marketing led to its inappropriate use in many for whom the risks outweighed the benefits.
GlaxoSmithKline’s handling of Avandia – also known by the generic name rosiglitazone – has many of the same elements. Launched in 2000 as a drug designed to help control diabetes, by 2006 was generating more than $3bn in annual sales. But data began emerging suggesting that it could also be contributing to cardiovascular problems including heart attacks.
A US Senate inquiry concluded in January: “GSK executives attempted to intimidate independent physicians, focused on strategies to minimise or misrepresent findings that Avandia may increase cardiovascular risk, and sought ways to downplay findings that a competing drug might reduce cardiovascular risk.”
A series of articles in the British Medical Journal this month reviewing the history of Avandia led its editor to suggest that the drug “should never have been licensed”. The BMJ called for a substantial weakening of the pharmaceuticals industry’s control over clinical trials. Some experts questioned both the short duration of the trials undertaken by GSK and the use of “surrogate end-points” based on blood tests to evaluate its effect on diabetes, rather than “hard end-points” such as whether it extended life or caused coronaries.
The drug was approved, although the EMA also required GSK to begin a longer-term trial measuring its cardiovascular effects. The study, called Record, was lambasted at a hearing this summer by Thomas Marciniak, an FDA official, who listed weaknesses in its design and implementation, and claimed his agency would never have authorised it.
GSK maintains it has always fully co-operated with the regulators on the growing number of studies it ran; that there are flaws in the studies that criticised Avandia, some conducted by researchers working with Japan’s Takeda Pharmaceutical on its rival Actos treatment; and that the results of the best-constructed and most relevant clinical trials still show no greater risk of serious cardiovascular problems than among diabetics on other therapies.
In balancing the evidence, one difficulty is untangling the different “confounding” factors that may have caused the cardiovascular problems identified in patients. “Diabetes itself increases coronary heart disease,” says one regulator, arguing that the case against Avandia is not easy to make. As with Vioxx, many people have multiple medical conditions and take a range of drugs.
A second problem is the uncertainty of rapidly evolving scientific understanding. Today, with more knowledge of the risk that Avandia and similar drugs may cause cardiovascular problems, there is greater – though still far from unanimous – agreement on what clinical studies should measure. A decade ago, that consensus was less well established. Even if longer and more detailed studies could have provided clearer answers then, they would also have required more time.
Subsequent advances in understanding can also make new studies difficult. Critics argue that the risk of cardiovascular problems with Avandia make it unethical even to compare the drug against Actos in patients. As a result, the FDA has suspended a trial designed to do that.
‘When practised in the public eye, medicine can be messy’
If Avandia, GlaxoSmithKline’s controversial diabetes drug, is largely the product of the regulatory system of a decade ago, a cancer medicine called Avastin is a symbol of more recent fine-tuning by authorities.
Avastin, developed by Genentech, Roche’s now wholly owned US subsidiary, is a lucrative blockbuster. Its possible withdrawal from the market as a treatment for advanced breast cancer (though not for other cancers), mooted this week but deferred until December, would cause sales to drop. Its problems have depressed the Swiss pharmaceutical group’s share price.
There, however, the similarities end. Avandia is a chemical-based drug that has racked up 14m patient years in use over the past decade, designed to slow the progression of diabetes. Avastin is a much more expensive biological drug given to far fewer patients, aimed at pushing back what would otherwise be imminent death.
Where Avandia was granted approval with a relatively wide range of uses in diabetes following large-scale clinical trials – however imperfect – Avastin was granted “accelerated approval” in spite of scientific dissent. That clearance carried the explicit condition that that decision would be reviewed once its effect in additional patients had been analysed in further studies.
These further data – the Ribbon-1 and Avado trials – have shown that Avastin in metastatic breast cancer does not extend patients’ survival, probably because the drug’s ability to prevent the tumour growing was offset by its toxicity, causing bleeding and heart failure. Many observers expect it to be withdrawn as a result.
“Avastin in breast cancer is rapidly becoming a classic case study of how messy science-based medicine can be when practised in the public eye and debated among pharmaceutical companies, the government and patient advocacy groups,” David Gorski, an oncologist and blogger, wrote recently.
Some now suggest that Avastin – like Herceptin, Roche’s other breast cancer drug – should be studied to see if it would work only if given to a subgroup of patients with a suitable genetic make-up. The problem is that such research is costly and uncertain: few drugs yet have such clear genetic markers in place.
More rapid approvals where the benefits of a new drug are strong – coupled with tough follow-up testing and a curb on excessive marketing – can meanwhile help. One danger is a reluctance to withdraw the drug subsequently even if data are weak.
Another is that this course defies the traditional mass-market approach to drug pricing. Offering it to far fewer people would cut revenues, unless its efficacy in a smaller subpopulation is sufficient to persuade healthcare systems to pay that much more for it.
Ray Hill, president of the British Pharmacological Society and a former Merck executive, says running more and longer studies comes at the expense of slower access by patients to more effective new treatments. “I still have nightmares that we may have made a mistake in withdrawing Vioxx,” he says, citing examples of patients with arthritis and other painful conditions for whom the drug was the most effective treatment but who now have to seek inferior alternatives.
He also cautions that the much higher cost of longer and larger trials risks reducing pharmaceutical research and stunting innovation. For example, the entire class of Cox 2 painkillers, of which Vioxx was only one, was in effect killed by the withdrawal, as the FDA began to demand much bigger pre-approval trials.
Statistics back up the argument: while the proportion of new drugs submitted that are approved has been relatively constant in recent years, the absolute numbers put forward for scrutiny are well below their peak of 15 years ago. Terry Maguire, an adviser to the UK’s medicine safety committee, argues that in spite of the “mind-boggling complexity” of the science and the fact that “data come from companies, which have vested interests”, regulation works well
Dr Nissen disagrees. He says the FDA lacks the resources to scrutinise trial results and reports of side effects in detail. Others suggest the difficulty is still greater at the EMA, which spends less time sifting through patient data, instead accepting summaries provided by the companies. He also wants to see criminal penalties for companies that fail to submit full details of studies they conduct.
Ironically, his own 2007 “meta-analysis” of Avandia, which examined different studies of the drug, was possible precisely because GSK had not only submitted its studies to the regulators but also made some of them public. That was partly the result of a deal with regulators earlier that decade when the company was fined for tardy reporting of suicidal feelings in children taking another of its drugs, the antidepressant Seroxat.
In fact, since the initial approval of Avandia, much has changed. Harder end-points have been agreed, systems to identify side effects have improved and regulators much more frequently demand – and companies implement – “post-marketing studies” to identify problems. Prof Hill says many of the necessary reforms are now in place. “If you look back to the mid-1990s, the industry was getting what it wanted and regulators were being bullied,” he adds. “Now it’s the other way round.”
He also cautions that it would be impractical to wrest control of trials from drug companies entirely: while governments, clinicians and academics could contribute more, they have neither the money nor the expertise to do as much as industry.
Yet further evolution is still thought necessary in the shifting relationship among regulators, patients, healthcare systems and drug companies. The first aspect is redoubled research to identify clearer genetic markers and other signals to understand which patients respond best to new drugs and which would suffer and should avoid them. Combined with new diagnostics, that would help to identify safer, more effectively targeted treatments.
In response, companies will have to shift further from the intensive selling of drugs to as many patients as possible, towards more “evidence-driven marketing” justified by data showing greater efficacy in smaller sub-groups of patients. The problem is that such markers are still difficult to identify and costly to implement.
Also badly needed is the creation, in combination with governments and health services, of more detailed electronic systems to supplement the information from clinical trials, which study small numbers of “ideal” patients. Monitoring risks and benefits of medicines once they are widely prescribed shows how they behave in ordinary patients, often with multiple diseases and taking a range of drugs.
“There is very substantial under-reporting of adverse drug reactions,” says John Thompson, a pharmacologist at Cardiff University. More data would help in drug development but would also impose extra requirements on regulators and doctors to restrict the use of drugs as problems emerge.
Finally, greater transparency will be needed, if only to counter suspicions of excessive industry influence. That involves removing inconsistencies in the extent and speed of information provided by companies on their trials; and circulating information as side effects are identified. The EMA board next month will consider just those sorts of proposals to explain its decision-making in more detail, as well as to declare any potential conflicts of interest among its advisers.
But the EMA’s Prof Eichler stresses that greater information will also need to be interpreted carefully by doctors, the public and politicians alike. “There is still a lot of uncertainty about the risks and benefits of drugs, which will not go away any time soon,” he says. “If people want certainty, very soon we will not have any new drugs.”