It may rank as one of the costliest sentences to run in a medical journal. On May 21, The New England Journal of Medicine broke its usual embargo and posted online a paper by scientists who had been studying data on a diabetes drug, rosiglitazone.

Dr Steve Nissen and Kathy Wolski’s summary of 42 patient trials into rosiglitazone, sold under the name Avandia by its makers GlaxoSmith-Kline (GSK), had a devastating pay-off: “Rosiglitazone was associated with a significant increase in the risk of myocardial infarction and with an increase in the risk of death from cardiovascular causes that had border-line significance.”

In other words, diabetics who took the drug ran a 43 per cent higher risk of heart attack (a 1 per cent risk rose to 1.43 per cent) and, possibly, a 64 per cent higher risk of death from cardiovascular disease (a 1 per cent risk rose to 1.64 per cent) than patients who took a placebo or an alternative medicine, according to Dr Nissen, a cardiologist, and Wolski, a statistician. Within days of publication, £9 billion had been wiped off the value of GSK.

The company fought back with a letter to The Lancet, deriding Nissen’s meta-analysis methodology and claiming that the side-effects of Avandia, approved for use in the US and Europe in 1999, were comparable to those of rival diabetes medications. GSK’s letter, written by Dr Ronald Krall, the company’s chief medical officer, revealed that its own studies were beginning to point the same way, findings that the company had passed to regulators. GSK’s Avandia-induced headache was exacerbated last week by a small study published in the journal Diabetes Care, linking the drug to a loss of bone density in men. A possible similar effect in women had already been reported.

The Food and Drug Administration (FDA) in America will meet at the end of this month to review safety data on Avandia; it will be expected to focus on the risk of heart attack rather than bone density. Withdrawal is possible but unlikely; the odds are that GSK will be asked to include a prominent “black box” warning on Avandia’s label, highlighting the risk of heart attack (the label already cautions about heart failure as a side-effect, and it is not recommended for patients with congestive heart disease). In Europe, the European Agency for the Evaluation of Medicinal Products, which approved rosiglitazone in 2000, had already demanded postlicensing surveillance to monitor the risk of heart attack. Its label had been updated several times to include further safety warnings.

Krall suggests, rightly, that conclusive answers on Avandia’s safety will come only from two long-term GSK trials of diabetes patients, one of which passed an interim safety review. “We should stay the course and allow ongoing trials to provide their definitive answers,” Krall writes. The trials will report by 2009.

But the ripples from the controversy continue to spread. GSK now faces a class-action lawsuit from investors who say that, had GSK disclosed its own data casting Avandia in a riskier light, they would not have bought GSK shares and would not have lost money when its share price fell to a two-year low. The company was also accused of trying to intimidate Dr John Buse, an eminent American diabetician and a longstanding critic of Avandia (Buse accepted an apology).

Whether fairly or not, Avandia risks joining Vioxx – a painkiller voluntarily withdrawn by Merck after being linked to heart attack and stroke – as a totem of the darker practices in the world of Big Pharma: the dogged pursuit of remedies for the ailments of an indulgent West (obesity, diabetes, heart disease); the relentless research and marketing campaigns devoted to me-too medications that offer no substantial benefit over existing therapies; the steep prices charged for branded drugs; the perceived financial cosiness between companies, regulatory agencies and researchers; the selective disclosure of information that shows off experimental therapies in the best possible light; the heavy-handed tactics allegedly employed to silence critics.

In GSK’s defence, it is one of the more transparent companies, registering all its clinical trials at their inception; indeed, Nissen put his study together from information freely available on GSK’s own website. GSK also rejects that Avandia is not an innovative medicine. Dr Alastair Benbow, GSK’s European medical director, says that Avandia’s risks have been overblown: “Science, not hype, will be the king here. Avandia will be seen to have a place in the overall management of type 2 diabetes.” Two weeks ago, at a meeting of the American Diabetes Association in Chicago, Nissen defended his study to a packed, sometimes hostile group of diabetes specialists, some of whom believe that he sparked an unwarranted scare. Even The Lancet, whose editor Dr Richard Horton once described industry-funded academics as having “bargained with the Devil”, has sounded a note of sympathy for GSK.

Benbow insisted that GSK did not take issue with Nissen’s analysis per se but with further media reporting, which he said may have led patients to abandon Avandia and thus put themselves at risk. Surely Benbow can’t blame the media for following it up? He says: “Patients don’t necessarily understand the science behind these figures. You can blame the way it [the 43 per cent excess risk of heart attack] is presented as a big-ticket number when the actual numbers [the increase in risk] were very small.”

Nissen, a heart specialist at the Cleveland Clinic in Ohio, is a seasoned whistle-blower: he wrote a key paper that led to the withdrawal of Vioxx, and has previously criticised the FDA for what he regards as a lax attitude towards drug safety. GSK has speculated that the timing of Nissen’s paper is suspicious (it coincided with US government meetings on the FDA’s future), and may have more to do with politics than patient safety.

Why would GSK question Nissen’s motives? Benbow insists it is patient welfare. But it would be naive not to mention the bottom line – profits.

Until 2003, pharmaceuticals was the most profitable industry in the world (it now comes third, behind crude oil production and commercial banking). Pfizer is at the top of the food chain with its cholesterol-buster Lipitor. It is by far and away the world’s bestselling drug, reeling in $12 billion (£6 billion) in sales a year.

Every pharmaceutical company dreams of enjoying a comparably large slice of the healthcare pie, estimated to be worth $660 billion (£328 billion) by 2020, which is why companies like developing “blockbusters” – defined as drugs expected to reap at least $1 billion in sales a year. These tend to be for common complaints, such as heart disease, asthma, diabetes, arthritis or depression. GSK’s biggest hitter is Advair, an antiasthma medicine, which brings in $3.2 billion (£1.6 billion) annually. Avandia was second with $1.3 billion (£0.65 million).

On average, blockbusters take about 15 years to travel from laboratory to market, a journey that costs around $1 billion. Around four in ten medicines fail at the final hurdle, the Phase 3 clinical trial (when tested in patients against placebo or another drug). Because of this, companies spread their bets by having several therapies in development at the same time. This costly approach is used by Big Pharma to justify high prices on the market.

The blockbuster approach, however, is under attack. According to a report by the accountants PricewaterhouseCoopers published this month, the pharmaceutical industry business model is “economically unsustainable”. Drug companies, the report noted, spend twice as much on research and development than ten years ago but produce half as many drugs.

Share prices have suffered on the back of leaky drug-development pipelines, soaring sales and marketing costs, the prospects of lawsuits (Merck faces more than 11,000 patient lawsuits over Vioxx) and the appearance of the National Institute for Health and Clinical Excellence (NICE), which vets drugs to ensure that they deliver value for money in a cash-strapped NHS. In addition, the finding that not all patients respond similarly to drugs is driving research towards personalised medicines and away from a one-blockbuster-treats-all philosophy. These challenges are, slowly, leading Big Pharma to expand its horizons beyond blockbusters. This month, GSK opened a £50 million imaging centre, a collaboration with Imperial College and the Medical Research Council, which is designed to speed up (and cut the cost of) drug development. It is a tacit admission that companies are finding it harder to go it alone in the hunt for innovative new medicines.

Another option for restocking a lacklustre drugs pipeline is to snap up small biotech companies with promising new products. These products have already survived part of the risk-laden journey to regulatory approval. Innovation can apply to the selling, as well as the development, of drugs. To increase the chances of winning NICE approval, some companies, including GSK, are offering to reimburse the NHS if their drugs do not benefit patients. Many potential blockbusters, such as the anticancer drugs Herceptin and Iressa, benefit only a quarter or a third of patients who take them. NICE is said to be considering favourably an offer along these lines from the makers of Velcade, a treatment for multiple myeloma that it previously rejected.

For now, it remains unclear how the Avandia saga will play out. Nissen has uncovered evidence of a small excess risk, but his meta-analysis is fragile; GSK is entitled to suggest that the risk has been overplayed. But its reaction has not enhanced the industry’s reputation – the company has hit out at Nissen, the NEJM, and the journalists who have followed up the story. All the while, GSK’s analysis of Avandia’s safety profile provides scant reassurance. Benbow’s view that patients don’t understand the numbers is somewhat patronising.

In any case, doctors understand the science, don’t they? Let us turn for clarity to an editorial in the British Medical Journal, which concludes: “Doctors will need to revisit the indication for (rosiglitazone) on a case-by-case basis, bearing in mind that several alternatives are cheaper, supported by robust evidence and now perhaps safer.”