Misleading communication of benefits of long-term statin treatment
In an editorial in BMJ, Gigerenzer et al1 chastised researchers who committed “sins” against transparent reporting, such as presenting relative risk (which inflates the appearance of interventions) without presenting absolute risk (which provides a more appropriate assessment of the benefit of an intervention). In a BMJ “Research News” article2, Jacqui Wise committed this sin in miscommunicating the benefits of statin treatment reported in a follow-up study of participants from the WOSCOPS trial3. According to Wise, the study demonstrated that treatment with statins “reduced deaths from coronary heart disease by 28% in men with very high levels of low density lipoprotein (LDL) cholesterol over a 20 year follow-up period”. A 28% reduction in risk of coronary-related death appears impressive to unsophisticated readers, but this form of data presentation is what Gigerenzer et al. called “incomplete and misleading risk information”.
The 28% effect reported by Wise, as well as in the abstract of the publication by Vallejo-Vaz et al., is referred to as relative risk reduction, which, when presented in isolation, is an amplification of the appearance of a benefit to a treated versus untreated group. The actual difference in rates of coronary death to the population (the absolute risk) reported by Vallejo-Vaz et al. were 9% in placebo and 6.7% in treated individuals. Therefore, the clinically relevant difference in the rate of coronary death between treated and untreated groups was only 2.3%, not 28%.
According to Gigerenzer et al., the routine use of relative risk, rather than absolute risk, is appealing to researchers and journalists because ”big numbers make better headlines”. Diamond and Ravnskov4 addressed this same issue in a review of how statin researchers have used deceptive statistics to amplify the minuscule beneficial effects of statins on coronary events. Thus, both Vallejo-Vaz et al. and Wise have misinformed readers by focusing solely on the more dramatic, but misleading, relative risk statistic.
The problem with Wise’s report, as well as the paper by Vallejo-Vaz et al., however, goes deeper than just the sin of selective reporting of relative risk. Wise included mention of a fatal methodological flaw in the Vallejo-Vaz et al study that renders its outcomes uninterpretable. According to Wise, “Five years after the initial trial finished, around one third of the men who were originally assigned pravastatin or placebo were taking statins.” Thus, the majority of people on a statin during the study stopped taking their medication, and to make matters worse, one third of the people that had been on placebo began taking statins after termination of the trial. Hence, any control the investigators had during the 5 year study in which participants were on the drug versus placebo was lost with its termination.
To make matters worse, there was no monitoring of which subjects were on a statin following the termination of the 5 year study. According to Vallejo-Vaz et al, “No later data on the proportion of individuals taking statin therapy were available for the subsequent years of follow-up.” The only thing we can be reasonably certain of is that the majority of all people in the study from both groups were not taking a statin for the final 15 years of the follow-up period. Hence, the follow-up report of outcomes 15 years after the original trial was completed was so flawed as to render its findings uninterpretable.
Why was it that 15 years later, there was a small but significant reduction in coronary events in the group that had been treated during the original trial? The answer to this question is a matter of speculation. It is perhaps attributable to a chance event or perhaps individuals in the original statin-treated group incorporated beneficial changes in their lifestyle, such as participating in smoking cessation or weight loss programs. It is known that people who adhere to statin treatment have been described as more health seeking than comparable patients who do not remain adherent. In one such study, Dormuth et al.5 demonstrated that individuals that remained adherent to statin treatment were less likely to have auto and workplace accidents than less adherent patients. These authors urged caution in interpreting “analyses that attribute surprising protective effects to preventive medications” because “patients who adhere to statins are systematically more health seeking than comparable patients who do not remain adherent”.
Overall, despite the sensational headlines of Wise’s report and the news reports this follow-up study generated, the study provided nothing of value in terms of guidance for clinicians seeking high caliber research on long-term effects of statins on coronary and non-coronary outcomes.
David Diamond, Ph.D.
Departments of Psychology, Molecular Pharmacology & Physiology
University of South Florida
Tampa, Florida, USA
Malcolm Kendrick, MD
East Cheshire Trust, Macclesfield District General Hospital,
Macclesfield, Cheshire, England
Luca Mascitelli, MD
Medical Service, Comando Brigata Alpina “Julia”/Multinational Land Force,
1. Gigerenzer G, Wegwarth O, Feufel M. Misleading communication of risk. BMJ 2010;341:c4830.
2. Wise J. Long term study backs statins for patients with high LDL and no other risk factors. BMJ 2017;358:j4171. doi: 10.1136/bmj.j4171
3. Vallejo-Vaz AJ, Robertson M, Catapano AL, et al. LDL-Cholesterol Lowering for the Primary Prevention of Cardiovascular Disease Among Men with Primary Elevations of LDL-Cholesterol Levels of 190 mg/dL or Above: Analyses from the WOSCOPS 5-year Randomised Trial and 20-year Observational Follow-Up. Circulation 2017 doi: 10.1161/CIRCULATIONAHA.117.027966
4. Diamond DM, Ravnskov U. How statistical deception created the appearance that statins are safe and effective in primary and secondary prevention of cardiovascular disease. Expert Rev Clin Pharmacol 2015;8(2):201-10.
5. Dormuth CR, Patrick AR, Shrank WH, et al. Statin adherence and risk of accidents: a cautionary tale. Circulation 2009;119(15):2051-7. doi: 10.1161/CIRCULATIONAHA.108.824151
Competing interests: No competing interests