Am Fam Physician. 2009;80(10):1042-1045
Author disclosure: Nothing to disclose.
Purpose
Each month, three presenters review an interesting journal article in a conversational manner. These articles involve “hot topics” that affect family physicians or “bust” commonly held medical myths. The presenters give their opinions about the clinical value of the individual study discussed. The opinions reflect the views of the presenters, not those of AFP or the AAFP.
This Month's Article
Yusuf S, Diener HC, Sacco RL, et al. Telmisartan to prevent recurrent stroke and cardiovascular events. N Engl J Med. 2008;359(12):1225–1237.
Does blood pressure reduction with telmisartan (Micardis) prevent recurrent stroke and adverse cardiovascular events when used after a stroke?
Mark: We thought we knew the beta blocker well—it lowers blood pressure, but unfortunately, likely doesn't change the outcomes we worry about in persons with hypertension who don't have coronary artery disease (CAD), such as stroke or myocardial infarction (MI).1 In fact, although beta blockers are clearly indicated in persons with CAD, they may be harmful when used to treat hypertension in those who don't have CAD.2 Can the same be said of angiotensin receptor blockers (ARBs)?
What does this article say?
Mark: This is a randomized study of 20,332 patients; 10,146 were randomized to receive telmisartan, and the rest to receive placebo. Patients were eligible if they were 55 years or older, had an ischemic stroke within 90 days of randomization, and were “stable.” After the first 6,000 patients, the protocol was changed to allow those 50 to 54 years of age to participate if they had a stroke within 120 days and if they also had two additional risk factors. This study was part of the PRoFESS study, which compared clopidogrel (Plavix) with aspirin/dipyridamole (Aggrenox) to prevent recurrent stroke.3
The end points were things we care about; the primary end point was recurrent stroke, and the secondary end point was major cardiovascular events, including death from cardiovascular causes, MI, and new or worsening congestive heart failure (CHF). New onset of diabetes was also a secondary end point.
Despite lowering blood pressure, telmisartan did nothing to prevent stroke, MI, CHF, cardiovascular deaths, or new-onset diabetes. And, telmisartan was associated with more hypotension (P < .001), syncope (P < .001), diarrhea (P = .02), nausea (P = .01), and atrial fibrillation (P = .006), among other adverse effects. So, not only did it do nothing good, but it also caused harm. There was an NNT of infinity (no benefit) and NNH of 100 for syncope, and an NNH of 330 for atrial fibrillation—no benefit, but important adverse effects.
Bob: And this was despite a systolic blood pressure reduction of 8.3 mm Hg at one month with telmisartan, and an overall systolic blood pressure reduction that favored telmisartan by 3.8 mm Hg. Overall diastolic blood pressure also favored telmisartan by 2 mm Hg.
Should we believe this study?
Andrea: You're probably wondering to yourself, “How come I haven't heard about this?” When a statin shows a minimal benefit in a flawed study, with an NNT of 250 in patients with an elevated C-reactive protein level (the JUPITER study),6 it is given advance publication in the New England Journal of Medicine, the American Heart Association calls a press conference, and everyone in the media reports the new miracle breakthrough. When a standard therapy that everyone has bought into fails to provide any benefit, it is hurriedly buried in the graveyard of obscure news stories.
Bob: Andrea gives an excellent example of publication bias, in which positive results are handled differently than negative results. This bias is present on many levels: researchers want to publish positive results because they are more likely to be accepted into more famous journals; editors prefer positive results because they get more publicity for the journal; and the mass media more readily report positive results as “groundbreaking” or “lifesaving” to increase viewership. Although studies with negative results may be important to physicians and our patients, these studies often get far less publicity. Could you imagine an evening newscaster reporting on a study of a drug that had no effect? Plus, nobody likes to be told that the drug they have been taking faithfully for 10 years hasn't been helping.
Andrea: The other thing about this study is that it shows, once again, that you can't rely on surrogate outcomes. Yes, telmisartan reduced blood pressure, but it did not reduce adverse outcomes.
Mark: This isn't the only negative study about ARBs to come out recently.7 Irbesartan (Avapro) has been shown to have no benefit in patients with CHF and preserved left ventricular ejection fraction.
Bob: Why this difference between angiotensin-converting enzyme (ACE) inhibitors and ARBs? ACE inhibitors release kinins that cause vasodilatation via nitric oxide, something that ARBs don't do. Also, ACE inhibitors prevent the stimulation of angiotensin receptor types 1 and 2; ARBs only block angiotensin receptor type 1.
What should the family physician do?
Andrea: ARBs should not be used as a substitute for ACE inhibitors. Stick with ACE inhibitors first line. Patients who do not tolerate ACE inhibitors (e.g., because of cough, angioedema) can be considered for an ARB. But realize that, although we have a wealth of surrogate outcomes, we have very little good outcome data on ARBs.
Mark: Verapamil and diltiazem (Cardizem) can be used to treat microalbuminuria in patients who do not tolerate ACE inhibitors and ARBs.
Bob: I agree. And be wary of publication bias—we will keep doing our best to give you the straight story.
Main Points
Stroke prevention after a stroke: ACE inhibitors = good; ARBs = same as placebo.8
Hypertension without CAD: ACE inhibitors = good; beta blockers = bad (unless the patient has a history of CAD)1,2; ARBs = likely bad.4,5,8
Diabetes with proteinuria and cardiovascular outcomes: ACE inhibitors = good; ARBs = same as placebo.4,9
EBM Points
Surrogate markers, such as blood pressure control, may not reflect patient outcomes. It is possible to lower blood pressure, but not help the patient.
Negative studies are less likely to get published than positive studies. This is called publication bias, and it is responsible for the overwhelmingly positive nature of the literature. Even when negative studies are published, they are less likely to receive the attention by the media and medical establishment that positive studies do.