Eli Lilly/Boehringer Ingelheim published a study claiming that duloxetine (Cymbalta) was an effective treatment for pain in depressed patients. Nothing new – they’ve run several such studies. The results were published in the Journal of Clinical Psychiatry in November 2007. Three wise readers (Jay Griffith, Joseph Hasley, and Daniel Severn) noted serious issues with the study and submitted a letter to the editor, which was then published in the June 2008 issue. It was noted that the patients were unclearly described: What kind of pain were they experiencing? The study also noted that patients weren’t taking pain-relieving medications for six months prior to the start of the study, at least not on a “regular basis,” a term that was not defined in the paper. Don't most patients who experience serious pain take analgesic medication at least somewhat regularly? Finally, and most importantly, the difference between Cymbalta and placebo was “statistically significant,” but more importantly (and ignored by the study authors), the difference was small. On an 11-point rating scale, the average difference in pain ratings favored Cymbalta by less than a point. Griffith and colleagues note, accurately, that the small advantage for Cymbalta “is not robust from the standpoint of clinical practice.” I’m always glad to see that there are a few readers of medical journals who are willing to take the time to pen a good letter to the editor in which the massive inadequacies of a study are noted. If we’re going to have evidence based medicine, we might want to make sure the evidence is of somewhat palatable quality, eh?
The kicker is that the lead author of the Cymbalta study, Stephan Brecht of Boehringer Ingelheim opted not to offer a response to Griffith et al.’s letter. So I suppose Brecht is conceding that the patient population was poorly defined and that Cymbalta’s advantage over placebo was meager. So this kinda runs counter to the conclusions of the study, which claimed in part that duloxetine is an effective painkiller. Not a big surprise, given that a prior analysis also cried foul about Cymbalta’s claim to successfully treat pain in depression.
Yet Cymbalta continues to fly off pharmacy shelves. Are physicians really this poorly trained at understanding scientific literature? “Gee, the ads say that Depression Hurts and the rep handed me these journal reprints that prove it's a painkiller, so now I’m writing Cymbalta scripts like there’s no tomorrow!”For failing to note that Cymbalta's effects over placebo were small, and for refusing to reply to the concerns about their study, I hereby nominate the authors of the November 2007 Cymbalta study (especially the lead author) for a Golden Goblet Award. Your dedication to obfuscation is notable -- keep up the bad work.