ֱ

Is Cholesterol Denialism Pseudoscience or Appropriate Medical Care?

— Calling an end to debate rarely serves the advancement of knowledge

MedpageToday
A photo of statin pills forming a question mark.
Abramson is a retired family medicine physician and has served as expert in pharmaceutical litigation. Kaplan is a researcher and an expert in clinical research standards.

Recently, a cardiologist in Medscape arguing that physicians and scientists who challenge the putative clinical benefit of statin medications are pseudoscientists. He claimed, "...somewhere along the way, cholesterol deniers stopped questioning the scientific evidence and started denying it." The benefit of statin medications, he argued, is "settled science."

We agree that evidence supports the use of statins for secondary prevention in people with established risk factors or with evidence of atherosclerosis. But the benefits, especially for people with low cardiovascular (CV) risk, are often modest.

A Review of the Evidence

Data from the National Health and Nutrition Examination Survey (NHANES) show using statins are not at high risk. Meanwhile, the that was commissioned by the U.S. Preventive Services Task Force (USPSTF) included 12 studies of statin therapy for primary prevention. None demonstrated a statistically significant reduction in CV mortality. Further, the aggregate reduction in CV deaths across the studies was not statistically significant.

Among the 18 trials in the USPSTF review that evaluated the effects of statin therapy for primary prevention on death from any cause, only two reported a significant benefit. One, , was a small study and the confidence interval for all-cause mortality was within one percentage point of being non-significant.

The far larger JUPITER trial was when cardiovascular morbidity and mortality amongst patients who received rosuvastatin fell below the rate for those assigned to placebo. However, the of the JUPITER trial did not find sufficient evidence to justify an indication for preventing cardiovascular death. Reduction of neither cardiovascular nor all-cause mortality is included as an indication in the current for rosuvastatin.

The USPSTF review is consistent with several other investigations. For example, in 2011 the internationally respected non-profit Cochrane Library that considered a subset of the studies included in the later USPSTF analysis. The review concluded: "Only limited evidence showed that primary prevention with statins may be cost effective and improve patient quality of life." And further: "Caution should be taken in prescribing statins for primary prevention among people at low cardiovascular risk."

The 2011 Cochrane review , with a strikingly different conclusion: "Reductions in all-cause mortality, major vascular events and revascularizations were found with no excess of adverse events among people without evidence of CVD [cardiovascular disease] treated with statins." Similarly, in 2013 the American College of Cardiology and the American Heart Association (ACC/AHA) issued that recommended lowering the threshold to initiate statin therapy for primary prevention to a ≥7.5% 10-year risk of CV disease. This of American adults without CV disease for whom statin therapy is recommended by more than 8 million, from 26.3 million to 34.5 million.

What changed between 2011 and 2013 to justify such a radical expansion of the number of low CV risk Americans for whom statins were recommended?

A Different Interpretation of the Evidence

The was established in 1994 with the goal of aggregating data across major statin clinical trials. Pooling data from all qualifying trials would amplify the statistical power by adding information about important subgroups like women and the elderly, as well as increasing the potential for relatively small clinical benefits to achieve statistical significance.

Investigators in these studies (all but one of which were funded by pharmaceutical makers, according to our review) provided confidential individual-level trial data. But there was a catch: before receiving the patient-level data, the CTT agreed that the data would be "" and would not be used in any publication "without the permission of the responsible trialists." In practical terms, this meant that nobody outside the collaborative would be given the opportunity to verify and replicate its results.

In 2012, the CTT that addressed the effects of statin therapy "in people at low risk of vascular disease." This meta-analysis concluded statins are safe and effective for low-risk primary prevention patients (10-year CV risk of <20%) and urged reconsideration of guidelines and recommendations written before 2012. The 2012 CTT meta-analysis was acknowledged in the 2013 Cochrane Review, which noted statin therapy "demonstrated a consistent 20% relative risk reduction in major vascular events with statins regardless of baseline risk." A year later, the 2013 ACC/AHA guidelines cited the 2012 CTT meta-analysis as "strong evidence that statins reduce total mortality in primary prevention."

The CTT's prohibition on disclosing patient-level data came into stark relief after one of its leaders, Professor Rory Collins, called on the British Medical Journal (BMJ) to retract co-authored by one of us (Abramson). The paper had re-analyzed the data presented in the CTT's 2012 meta-analysis and came to a very different conclusion: for people at low-risk, statins did not significantly reduce all-cause mortality and 140 low-risk people had to be treated with a statin for 5 years to prevent one non-fatal major CV event. The BMJ appointed an external panel of six experts, including two statisticians, to adjudicate the demand for retraction. After thorough review, the that Collins' demand did "not meet any of the criteria for retraction."

In the process of adjudicating this dispute, the BMJ editors noted the absurdity of treating millions of people worldwide with statins while the data underlying these recommendations remained inaccessible for independent analysis. The editors wrote to the lead investigators of 32 major statin trials and asked them to make their study data available for independent analysis. One year later, despite follow-up e-mails and phone calls, only seven of 32 investigators had even bothered to respond. The BMJ editors then for statin researchers to make their data available. Even now, 9 years later, the data remain inaccessible.

The Need for Accessible Data

In conclusion, there are two problems with the statin denialism argument. First, even in the age of , peer reviewers, guideline writers, and independent analysts have been denied access to the data that might resolve differences of opinion. We are not denying the evidence: we just need access to it to determine if the science is settled. Calling an end to debate rarely serves the advancement of knowledge.

Second, the age of medical paternalism is (or should be) coming to an end. Shared medical decision-making requires disclosures about the absolute risk reduction for specific categories of patients expressed in terms of number needed to treat, uncertainty, and assessment of patient preferences.

To be clear, we recognize that statins reduce cardiovascular events and reduce all-cause mortality for secondary prevention patients. But, even when statins are used for secondary prevention of CV disease, about 80 people must be treated with a statin for 5 years to prevent a single death and about 30 people must be treated for 5 years to prevent one heart attack or stroke.

Given these numbers, many people -- ourselves included -- would choose to take a statin for secondary prevention. For primary prevention in people with less than 20% 10-year risk, between 100 and 140 people must be treated to prevent a single non-fatal major CV event. Furthermore, as just published in JAMA Internal Medicine, even those numbers needed to treat are actually much higher because the 2013 pooled cohort equations used to calculate estimated 10-year risk erroneously inflated CV risk estimates, potentially leading to inappropriate statin eligibility for about 17.3 million Americans.

Importantly, the preferences of those who choose to reject statin therapy must be respected. Doctors who engage in high-quality informed shared decision-making are delivering optimal patient care, not harmful pseudoscience.

practiced as a family doctor for 22 years. He served on the faculty of Harvard Medical School for 25 years, as an unpaid consultant to the Department of Justice and FBI, and has written two books about the commercial takeover of the knowledge doctors must rely on; the more recent is, . is a senior scholar at Stanford University's Clinical Excellence Research Center, a former associate director of the NIH, and a former chief science officer for the Agency for Health Care Research and Quality.