Wednesday, September 06, 2006

Should Physicians Prescribe High-Dose Lipitor to Prevent Strokes?

The Boston Globe just reported about how Pfizer Inc, the world's biggest drug company, is "using the results of a study it sponsored to persuade doctors to switch patients to pricier doses of its blockbuster drug, Lipitor."

Pfizer has a simple strategy for maintaining market share: The company has sent thousands of sales representatives to doctors' offices to convince them that higher doses of Lipitor are more effective than generic statins in protecting patients' hearts. In particular, the sales reps point to a Pfizer-sponsored study that said Lipitor in high doses reduces the chance that stroke survivors will be stricken again.
'It's been part of our strategy that we ought to be encouraging physicians to move their patients to the higher doses,' Peter Brandt , Pfizer's newly promoted leader of US pharmaceuticals, told analysts in a recent call.

Pfizer also makes more money from higher doses of Lipitor. A 10 milligram Lipitor pill costs $2.44 . The 40 milligram and 80 milligram doses cost $3.33 apiece.
The Globe also noted that:

The Pfizer-sponsored study on Lipitor's impact on stroke was written by 11 authors who were either company employees or reported financial ties to it. The New England Journal of Medicine article linked daily , 80 milligram Lipitor doses with a 16 percent reduction in the risk of a repeat stroke.

An editorial, also published last month in the journal, could provide another driver for Lipitor revenue growth. The study adds to 'gathering momentum' that could convince insurers to count statin use as a measure of 'quality' after-stroke care, according to the editorial. Dr. David M. Kent , the editorial's author, reported receiving grant support from Pfizer.
But what does the study actually tell us? [The article was: The Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) Investigators. High-dose atorvastatin after stroke or transient ischemic attack. N Engl J Med 2006; 355: 549-559.] In the spirit of evidence-based medicine, let us examine how the study was done and what its actual results were.

In summary, this was a randomized controlled trial (RCT) which compared high-dose atorvastatin (80 mg/day) with placebo. Eligible patients were over 18, and had a stroke or transient ischemic attack (TIA) within one to six months. Patients with hemmorhagic stroke were "included if they were deemed by the investigator to be at risk for ischemic stroke or coronary heart disease." Patients had to be ambulatory, with at least somewhat preserved functional status (Rank score less than four), and had to have an LDL cholesterol between 100 md/dl and 190. Patients were excluded for a variety of reasons, including atrial fibrillation and sub-arachnoid hemmorhage, most of which were not specifically listed (and only found in a 2003 article which I could not find on-line).

Note that this study was not designed to determine whether lower dose atorvastatin, atorvastatin with dose-adjustment to achieve a particular cholesterol value, or treatment with another cholesterol lowering drug would affect the risk of stroke. Thus, one cannot tell whether any benefits conferred by atorvastatin at a fixed high dose could have been achieved with a lower dose, an adjusted dose, or another cholesterol lowering drug. Also note that the study's results may not apply to patients with conditions which would have excluded them from this study, but it is not immediately obvious who those patients might be.

The study's main results were as follows. The key result emphasized by Pfizer was that the rate of strokes in the patients treated with atorvastatin (265/2365, 11.2%) was lower than that in patients on placebo (311/2366, 13.1%) over an average of 4.9 years of observation. However, although the rate of ischemic strokes (those due to reduced blood flow) was lower in patients treated with atorvastatin compared with placebo (0.78 hazard ratio), the rate of hemmorhagic strokes (those due to bleeding) was higher in patients treated with atorvastatin (1.66 ratio). Importantly, the death rate of patients treated with atorvastatin (216/2365, 9.1%) during the study was higher than the death rate of patients given placebo (211, 8.9%), although the difference did not reach statistical significance, i.e., they could have been due to chance alone. Numerically more patients who received atorvastatin died from cancer, infection accidental or violent death, and "other causes" and "unclassified reasons" than did those who received placebo, although these differences also were too small to be statistically significant.

The article's discussion section noted the increased risk of hemmorhagic stroke, and suggested "the potential risk of recurrent hemmorghage should be considered when one is deciding to administer a statin to patients who have had a hemmorhagic stroke." I could not find data in the article, however, that indicated the risk of a hemmorhagic stroke after atorvastatin administration was only found in patients who had already had such a stroke. The discussion did not address the increase in total death rate in patients who received atorvastatin.

The accompanying editorial noted above [Kent DM. Stroke - an equal opportunity for the initiation of statin therapy. N Engl J Med 2006/ 355: 613-615] also attributed the increase in hemmorhagic strokes to the enrollment of patients with previous hemmorhagic strokes, and failed to mention the increase in total death rate in patients given atorvastatin.

These results are, in my humble opinion, troubling. The use of atorvastatin is being promoted as a means to reduce stroke, and other serious cardiovascular events. Such events can be fatal. If the only effect of the drug is to reduce these events, the death rate in patients treated with atorvastatin should be numerically lower, or at least no higher than that in those treated with placebo. Although it is possible that the increased death rate actually found was due to chance alone, it may be hinting at some adverse effect of the drug in this setting that may be offsetting its beneficial effects. Based on this single study, one cannot conclude that the overall benefits of high-dose atorvastatin for patients with previous strokes unequivocally outweigh the possible harms of the drug in this situation.

What are the lessons here? Per the Boston Globe, Dr John Abramson, author of "Overdosed America," suggested, "because this is a drug company-sponsored study, what we're seeing is just tunnel vision of the effect of high-dose Lipitor on stroke." I would agree that physicians ought to be very cautious about using this single study as a reason to prescribe high-dose Lipitor to prevent recurrent stroke.

This case, and numerous others discussed on Health Care Renewal suggest that
  • studies sponsored by organizations with vested interests in the results coming out in a particular way ought to be interpreted with caution.
  • physicians ought to be very skeptical about enthusiastic opinions about drugs expressed by authors who have financial arrangments with the drug's manufacturer
  • physicians ought to be very skeptical about how those with products or services to sell interpret their own sponsored clinical research when marketing their wares

1 comment:

Anonymous said...

I agree with you. But in most of the cases, if not all of them, the doctors are the ones to be informed properly of any lipitor dosage. The ones who should do it are pharmacists or pharmacist scientists in universities so there will be no suspicion regarding the reasons behind the result.