Inside GSK's $1.1B Kafka-esque Diabetes Nightmare

Although GlaxosmithKline (GSK) paid $520 million to settle 10,700 lawsuits filed over its diabetes drug Avandia in 2010, the company potentially remains on the hook for an estimated $580 million more in liabilities from the patients who have refused to settle. The potential scope of those liabilities became a little more clear a couple of days ago when the federal judge handling all the remaining Avandia cases (an unknown number) issued an easy-to-read ruling on which scientific experts will be allowed into the trials and which will not. Unfortunately for GSK, the plaintiffs won the right to present three of their anti-Avandia boffins to a jury.

The cases stem from several studies and an FDA decision that pointed to an increased number of heart attacks among diabetics taking Avandia. The ruling is a triple headache for GSK's management because its logic suggests that the evidence GSK needs to prove that Avandia is innocent does not exist, and that it is now impossible for GSK to get it. Worse, the very evidence GSK was asked to gather to prove the case either way is now being used against GSK even though it's not the gold-standard evidence that is actually needed to settle the question. It's pharma's version of a Kafka-esque nightmare.

Judge Cynthia Rufe wrote that the type of study that should have been done would have been to test Avandia in diabetics versus a placebo, and then to measure which group suffered more heart problems. Unfortunately, creating such a test would require turning a blind eye while the diabetics in the placebo group went untreated and the folks in the Avandia group potentially died of myocardial infarctions. That would be unethical and so it wasn't done. Instead, Avandia was tested against Actos, a similar drug made by Takeda, in a number of trials. As diabetics have a higher risk of heart problems anyway, adverse events showed up in both groups. The Avandia people seemed to have a higher rate of heart attacks, but that wasn't statistically significant, Rufe wrote.

The Cleveland Clinic's Dr. Steve Nissen did an infamous meta-study of all the available Avandia data from 42 different trials and found there was a statistically significant higher rate of heart problems with Avandia -- a 43 percent higher risk.

GSK argued that meta-studies are far from the gold-standard of evidence (they mash together apples and oranges, after all) and thus solid conclusions can't be drawn. Indeed, one study the FDA asked GSK to do on the issue didn't show a statistically significant risk increase. The other was called off when the FDA -- convinced in large part by the Nissen info-drop -- decided that waiting to see how many Avandia patients dropped dead was not kosher.

Thus GSK has been deprived of the very evidence it most needs to counter its critics.

The first of the upcoming trials, initially scheduled to start Feb. 22, features a plaintiff who died. Many of those who settled in the first two rounds took only small sums, less than $70,000. In death cases, verdicts and settlements can go much higher. Which is why UBS analyst Gbola Amusa's estimate that GSK's total liabilities on Avandia after the previous settlements still stand at around $1.1 billion.

Related:

More from CBS News