I was broad-sided: Rolling meta-analyses to detect drug side effects

Anthony De Carlo's story:
In an auto accident on 6/15/01 in New Jersey, I was broadsided by a woman on a phone running a red light. I was in the hospital three days, banged up. I hurt my knee and back when I was hit.

I took Bextra for two years, 
and Celebrex. I have had one heart attack and two smaller heart attacks. An EKG revealed I have had three silent strokes, one in 2003 while I was taking Bextra and also Celebrex. I later had the EKG confirmed at a teaching hospital in New York City. I was also diagnosed with a deteriorating heart valve, which will probably require surgery.

Due to complications, I am disabled and can't work.


Though studies had been performed about the side effects of Bextra, the results had not been put together. If they had been, the dangers of Bextra would have been known years sooner. Instead, many people like Anthony suffered from major side effects. Eventually, Pfizer took Bextra off the market in 2005, acknowledging its dangerous increase in heart attacks and strokes.

Drug makers have removed several other drugs from the market after learning of their health problems. In the most recent issue of the Archives of Internal Medicine, Dr. Joseph Ross and his colleagues describe a method of statistically pooling data on drug effects in a rolling meta-analysis that could have informed Merck of the cardiovascular effects of Vioxx years before it withdrew the highly profitable drug from the market. Dr. Michael Steinman mentioned Bextra as a case in point, and added several other examples: Bayer removed the cholesterol-lowering drug Baycol from the market in 2001 after reports of serious muscle problems. Novartis withdrew Zelnorm, a drug for irritable bowel syndrome, from the market after learning of its increased risk of heart problems.

Advice: Read the story of a whistleblower on the BBC website.

Read another story about Vioxx. Thanks to Natasha Singer for the source story in today’s New York Times.