Wednesday, September 05, 2007

Why Might Switching Statins Be Bad for You?

At the risk of being thought redundant, I am going to take on a topic already covered today in the health care blogsphere, mainly because I think there is even more that can be added to the discussion.

There has been some media attention to this poster describing an observational study sponsored by Pfizer Inc., and authored mainly by Pfizer employees, presented at the European Society of Cardiology Congress 2007 [Phillips B, Roberts C, Rudolph AE, Morant S, Aziz F, O'Regan CP. Switching statins: the impact on patient outcomes. Br J Cardiol, in press.] AP coverage is here, and Pfizer's press release is here.

The poster described an observational study that used data from a primary-care data-base. The study compared a cohort of patients who were switched from atorvastatin (Lipitor) to simvastatin to a cohort who remained on atorvastatin. Patients who were switched were more likely to die or sustain major cardiovascular events, even controlling for certain patient characteristics.

The study's media coverage reiterated the authors' conclusion that it is bad to switch patients from Lipitor, a more expensive, brand-name statin, to generic simvastatin. For instance, a Reuters article stated, "switching was associated with a 30 percent increase in the relative risk of major cardiovascular events, including heart attacks, strokes and death."

But the study's observational design allowed for other interpretations.

Dr Peter Rost, writing on the BrandweekNRX blog, fumed, "Lies, damned lies, observational trials. And the morons who write about them." Rost noted that people may switch from one statin to another because the first drug failed to lower their cholesterol. Such people may be resistant to statin treatment, and hence more likely to have a bad outcome for that reason, not because of the switch itself.

A Wall Street Journal Health blog post interviewed Dr Mark Fendrick, who also suggested that people may resent switching drugs for economic reasons, and hence be less likely to comply with their new prescriptions.

I can think of two more plausible reasons that people who switch may have worse results than those who don't, regardless of the particular drugs involved. First, people switched because of side-effects of the first drug may be more prone to have side-effects on the second drug, and thus fail to take it, or a sufficient dose. Second, people switched because they were non-compliant with the first drug may not be complaint with the second drug too.

Thus, there are many reasons that the data from this study could simply indicate that people switched from one drug to another are more likely to have bad outcomes than people who stay on the first drug they are given.

To really test the hypothesis that switching from Lipitor to simvastatin produces worse results than staying on Lipitor would require a randomized, controlled trial.

But isn't it odd that the investigators didn't compare the results of people who switched from Lipitor to simvastatin to people who switched from say, simvastatin to Lipitor, or from Lipitor to a third drug? One wonders whether their failure to consider alternative explanations of their data had something to do with the economic interests of their employer?

It's also too bad that the media coverage of this study did not initially consider that there may be interpretations of its results other than those provided by its commercially sponsored study investigators.

2 comments:

Peter Rost said...

You are very right about your assumptions. Actually they are mentioned in the press release, but with a foolish explanation:

A secondary analysis of the same data showed that patients who were switched from Lipitor to simvastatin were more than twice as likely to discontinue their treatment compared to those who remained on Lipitor therapy (20.5 percent versus 7.62 percent, p<0.001). The reasons for discontinuation were not available from the database, though disruption in treatment has been associated with poor adherence in previous studies of statins and other medications.

So, part of the explanation is exactly what you suggested! But the regular journalists didn't pick up on that . . .

Anonymous said...

Where is Merck's rebuttal?!