« ADVANCE ACCORD | Main | Social Reform »
April 26, 2008
TORCH Analysis
American Family Physician (paid subscribers only) has a great analysis of the TORCH study. TORCH (Toward a Revolution in COPD Health) was a large, randomized, double-blind trial studying a treatment for COPD (emphysema). The particular treatment they were studying has great theory behind it, but came under question a few years ago for increasing deaths in kids treated for asthma with the combination.
What Mark Graber points out in this Journal Club format is that not only is the mortality not affected by these drugs alone or in combination, but the benefits touted for the drug are not very meaningful. The NNT (Number Needed to Treat) to prevent an exacerbation of COPD is 4, which is good. But the NNT for one year to prevent a hospitalization is 33. That means that 33 patients have to be treated with this medicine to even keep one out of the hospital. That is of arguable meaningfulness.
Where the study is deceptive is in downplaying the increase in cases of pneumonia. The NNH (number need to harm) for one year for pneumonia is 41. If 41 people are treated for one year, there will be one additional case of pneumonia in the combination treatment group. These did not translate to deaths, apparently. But certainly the benefit is offset by the harm.
The Journal Club also points out that there were a lot of dropouts from the study, and all the drugs were made by the drug company sponsoring the study, with no comparison with other helpful treatments in COPD.
It is refreshing to see an increase in such analyses.
Clinical Iatrogenesis | By Robert Maddox | 09:24 PM