Lancet Comment questions benefit of statins in primary prevention

January 25, 2007 Shelley Wood

London, UK - Authors of a Comment in the January 20, 2007 issue of the Lancet are calling for a revision of the 2001 US National Cholesterol Education Program guidelines, saying that statin therapy for primary prevention in women or in people over age 65 is not supported by the bulk of the evidence [1]. But others counter that while the magnitude of the benefit may not be as large in primary-prevention patients as in secondary prevention, there is clear benefit, particularly in higher-risk groups.

"In support of statin therapy for the primary prevention of this disease in women and people aged over 65 years, the guidelines cite seven and nine randomized trials, respectively. Yet not one of the studies provides such evidence," Dr John Abramson (Harvard Medical School, Boston, MA) and Dr Jim Wright (University of British Columbia, Vancouver) write in the Lancet.

The authors of the Comment drew their conclusions after looking at eight randomized clinical trials that compared statins with placebo and included primary-prevention populations at increased risk of developing CAD. Overall benefit was estimated on the basis of total mortality and total serious adverse events. They report that total mortality was not reduced by statins. In the two trials that reported total serious adverse events, these events were not reduced or increased by statins.

When only cardiovascular events were considered, Abramson and Wright report that statins reduced this outcome to a statistically significant degree; however, the absolute risk reduction of 1.5% was small. Moreover, this would mean that 67 people would have to be treated for five years to prevent one event. Furthermore, they argue, this benefit might be confined to high-risk men between the ages of 30 and 69. Abramson and Wright contend that statins did not reduce coronary heart disease events in the almost 11 000 women in pooled trials, nor in men and women older than 69 (n=3239).

Relative risk by event and group


Outcome
RR
95% CI

Mortality, all
0.95
0.89-1.01

Serious adverse events, all*
1.01
0.97-1.05

CVD events, all
0.82
0.77-0.87

CVD events, women
0.98
0.85-1.12

CVD events, men and women over age 69
0.94
0.77-1.15




*Reported in two trials only

To download table as a slide, click on slide logo below

In an interview with heartwire, Wright explained that their analysis was conducted in response to the 2005 meta-analysis reported by the Cholesterol Treatment Trialists (CTT) collaborators, which found a benefit of statins in a combined cohort of primary- and secondary-prevention populations and in a cohort of patients with no coronary heart disease at baseline, many of whom had preexisting cerebrovascular or peripheral vascular disease [2].

"Because they have individual participants' data, the CTT collaborators have the unique opportunity to analyze the data for the 41 354 people in the true primary-prevention group," Abramson and Wright contend.

To heartwire, Wright elaborated: "Our information, which is the best that's available in the published literature, is not as good as what the CTT collaborators have. They basically are in a better position to publish this information specifically in the primary-prevention group and have the world be able to look at it and have doctors be able to see that the evidence isn't there, for some groups of people.

"It appears to me that they are deliberately not providing that information. . . . If you look at their paper, they're academics, but they do have relationships with a number of the companies that market these drugs. Maybe they honestly believe that [public access to data] could mislead, or that they don't have all the answers yet, or all those things, but personally I find it shocking that they wouldn't just provide access. And if they don't want to analyze it and publish it themselves, then make it available for other people to do so. We'd be more than happy to do that."

In the meantime, Wright thinks physicians should be honest with their patients about the lack of evidence in low-risk patients and the low absolute difference in relative risk, even in higher-risk patients requiring primary prevention.

'If you take a male who is 50 years old, a smoker, with high blood pressure, who eats the worst diet in the world . . . then if I were an honest physician, I would tell him that maybe he should be taking a statin. And if he asked how much would that reduce his risk, I would have to tell him that it would reduce his risk by 2% over the next five years. If he understood that information, he would say, You're expecting me to take a pill everyday for five years? And it's going to cost me two dollars a day? You're crazy! I'm not going to do it."

If physicians were truly honest with their patients, Wright concludes, "I think there probably would be very few people being treated for primary prevention with a statin."



The CTT responds
In response, Dr Colin Baigent (Clinical Trial Service Unit [CTSU], University of Oxford, UK) co-primary investigator for the CTT, told heartwire he considers the CTT database a resource and said he was unaware that Wright and Abramson wanted access. "If people are interested in the data, they can write to us and we'll do our best."

He also stated that he has every intention of doing the analysis proposed by the Comment authors. "They've suggested we should also look at effects in people with no known coronary disease at baseline. We haven't had the opportunity to do that, but I do plan to do those analyses, and I will discuss them with my coprincipal investigators and we'll decide whether we want to formally respond in print [to the Lancet Comment]."

But Baigent also underscores the pitfalls of proving benefit—or lack of benefit—in a very low-risk population. "My suggestion would be that if we are looking at a group of patients with no history of vascular disease, what we should be doing is looking at that group of people and also looking at the effects in people with vascular disease and testing to see whether there are any differences between these groups; rather than just looking at primary prevention in isolation, we should use all the data and ask the question: is it clear that the effects of statins differ substantially in those without disease? Because inevitably what will happen is that we'll have many fewer events in people with no history of disease, so the risk is that we'll inappropriately conclude that statins don't work."

Baigent also dismissed the idea that ties to industry had any effect on the CTT collaborators' conclusions. The CTT, he reminded heartwire, is a joint collaboration of the CTSU at Oxford and the University of Sydney, Australia, made up of clinical trialists who have participated in major statin trials, many of which were publicly funded, without industry involvement. The CTSU itself has a policy of not accepting honoraria from industry, as do several of the collaborators individually, and travel funding for the CTT to meet during the annual AHA meeting is provided by government funding, he explained. Any industry representatives who attend the CTT meetings, like all other collaborators who are not on the writing group, have no influence or veto powers in the writing of the CTT's conclusions.

"I'm very disappointed to hear that they think there's in some way some conflict of interest, " Baigent summed up. "I don't think there's any justification in the implication that we have been in any way influenced by industry. We are independent research scientists."



A question of interpretation?

Also commenting on the Comment for heartwire, Dr Niteesh Choudhry (Harvard Medical School) took issue with the emphasis by Abramson and Wright on mortality outcomes. Choudhry and colleagues published their own meta-analysis of statins in primary prevention, as reported by heartwire, in November 2006. At that time, Choudhry told heartwire that the guidelines are supported by the data, particularly for higher-risk patients requiring primary prevention.

"[Abramson and Wright] are saying that statins reduce the risk of nonfatal events in a significant way but don't reduce the rate of fatal events, and that's basically what we found as well," Choudhry stated. "But the difference is in the interpretation of whether nonfatal events are important or not. These authors make the argument that because statins don't reduce mortality in a primary-prevention population maybe it's not worth the cost of providing them to patients who don't have proven vascular disease. The argument I would make in contrast to that is that statins do reduce nonfatal events—MIs and other vascular events—and those are important. Even if they don't reduce deaths [in primary prevention], which is debatable, preventing heart attacks is still an important outcome."



Magnitude of benefit vs money matters
Where both Baigent and Choudhry appear to agree with Abramson and Wright is that the magnitude of benefit in a low-risk primary-prevention population is unclear.

"Risk stratification is of course the key point, because we've got so much information now on statins and their effects: it's very clear that they are effective."

Some sort of risk score for primary-prevention patients to determine the absolute benefit of statins based on level of risk would be useful, he notes. Indeed, CTT collaborators have a paper "in gestation" examining the use of a risk score for diabetics, and such an index could also be devised for other, lower-risk patients, Baigent agreed.

"We're conscious of the fact that there is interest in looking to see how low the benefits go. But really, the evidence of benefit of statins is so immense that I would be very surprised if we find a group that clearly doesn't benefit, when appropriate allowance is made for the multiple comparisons that we would have to do. . . . If you were treating middle-aged women with no history of disease and their risk is very low, the number of women you'd have to treat to avoid one event would be enormous, and that would be very expensive," Baigent continued. "So society would tend to agree that that wasn't an appropriate use of money, but that wouldn't be the same as saying that statins don't work in those people. Very likely they do, but their benefits are so small that they're not worth it."

But Baigent is adamant: "I don't recall writing—or even thinking—that statins should be used in everyone. That's simply not something we advocate."

What price point would be deemed worthwhile is fodder for further study, says Choudhry. Until then, he believes clinicians should feel comfortable using statins in primary prevention. "For patients who are at high risk by virtue of their risk factors but don't have proven vascular disease, who have average or above-average cholesterol levels, they should receive statin therapy," Choudhry proposed. "I think the data on that are relatively clear. What I believe the data also support is that there is benefit for intermediate-risk patients, although maybe not in terms of mortality. The unresolved question that the data do not answer, and even data from the CTT won't answer, is whether the magnitude of benefit in those intermediate-risk folks is worth the cost of paying for these drugs for tens of millions of people."


http://www.theheart.org/article/767615.do
Tweet