Awaiting the verdict: statins and the road ahead for primary prevention in older adults

Cardiovascular disease (CVD) is a common cause of morbidity in developed countries and the leading cause of death worldwide. In the United States, CVD was responsible for approximately 929,000 deaths in 2020, accounting for 1 in every 3.6 deaths.¹ Beyond its impact on mortality, CVD significantly contributes to disability among older adults, resulting in an approximate six-year reduction in ‘healthspan’ — defined as life with good health and function — for people aged 65 and older.^1,2 It is estimated that 70% of individuals over the age of 70 will develop CVD, implying that primary prevention strategies, such as the use of statins for lipid-lowering, are critical even in patients of advanced age.² Despite the extensive evidence supporting the use of statins for primary prevention in adults up to age 75, beyond that point, there is limited data from randomized trials, leading to IIb recommendations in the current guidelines that it may be reasonable to both initiate, or potentially deprescribe a statin, depending on frailty, multimorbidity, functional decline, or other factors that may limit life-expectancy.³ This uncertainty about the role of statins in primary prevention represents a significant public health gap, particularly considering the heterogeneity of older adults, where clinicians encounter a spectrum from robust, independent individuals to those who are frail or dependent on assistance for self-care activities.^4,5

There are two ongoing large-scale placebo-controlled randomized trials designed to address this evidence gap. In Australia, the STAtins in Reducing Events in the Elderly (STAREE) trial has enrolled 9,971 community-dwelling adults ≥70 years without a history of cardiovascular disease, diabetes, or dementia, and will compare randomization to atorvastatin 40 mg versus placebo on the co-primary outcomes of survival free of dementia and disability and the incidence of major CVD events.⁶ In the US, the Pragmatic Evaluation of evENTs And Benefits of Lipid-lowering in oldEr adults (PREVENTABLE) trial plans to enroll 20,000 community-dwelling adults ≥75 years free of cardiovascular disease, and is similarly comparing atorvastatin 40 mg versus placebo on a primary outcome of survival free of dementia and persistent disability, with incident CVD being a secondary outcome.⁷ The results from STAREE and PREVENTABLE are highly anticipated, yet are not expected for several years. STAREE has recently completed enrollment and expects at least three more years of follow-up. However, less than half of participants in STAREE are 75 years or older, and so will have more limited power in that subgroup.⁶ Meanwhile, PREVENTABLE is still enrolling, with results not likely to be available until 2027 or later.⁷ In the interim, the words of William Shakespeare seem relevant, albeit a bit dramatic, “I am to wait, and waiting so be hell.”

Of course, as a field, sitting on our hands while we wait on definitive evidence is not our only available option, given the rise of big data and the growing maturity of causal inference.⁸ In this issue of the Journal of the American Geriatrics Society, Orkaby et al. present the results of a large, well-designed observational study investigating the association of initiating a statin on the first occurrence of a major cardiovascular event and all-cause mortality in a retrospective cohort of 710,313 adults aged 65 years and older in the US Veterans Affairs (VA) system.⁹ Using a previously validated frailty index (VA-FI) based on the theory of deficit accumulation, Orkaby et al. also examine heterogeneity in the relationship between initiating a statin and the same set of outcomes. The study population was mean age of 75.3 years, 12.1% were categorized as frail by the VA-FI, and the cohort was followed for a mean of 7.6 years. In propensity score weighted analyses, the initiation of a statin was associated with a 39% reduction in mortality risk, and a 14% lower incidence of major adverse cardiovascular events (Figure 1). The results for frail patients that initiated a statin were similar in relative terms, though associated with larger absolute benefits, with a reduction of 40 deaths per 1000 person-years and 13.8 major cardiovascular events per 1000 person-years compared to those not prescribed statins. In subgroup analyses by age, the results for mortality were largely consistent inclusive of adults over 75 or even 85 years or age. However, there was some indication of a weakening effect on CVD, consistent with competing risks becoming more relevant with increasing age, with numerically smaller relative reductions observed in patients >85 years of age versus those 85 years or less (Figure 1).⁹

Figure 1.

Effects of new statin use on mortality and first cardiovascular events in older adults: a visual summary of Orkaby et al., 2023

The study of Orkaby et al. represents a beautifully conducted epidemiologic analysis of an important clinical problem.⁹ It has several notable strengths, including a very large sample size, considerable follow-up for incident CVD and all-cause mortality, careful attention paid to statistical adjustments to address confounding by indication and other biases, the use of a validated approach to frailty, and numerous sensitivity analyses. In addition, the focus on initiating statin use is consistent with a target trial approach, and should mitigate several biases that can confound analyses of prevalent use.¹⁰ The authors also linked VA data with Medicare claims, which is as good as one can do to pragmatically ascertain incident CVD in the US.

Despite these strengths, several aspects of the study and its’ results should provide some pause. First, as a VA study, the cohort predominantly consisted of men (98%) and was primarily white (89%), which limits generalization to the broader population of US adults free of cardiovascular disease.⁹ Second, the study time period for treatment (2002–2012) reflected prescriptions for simvastatin for the vast majority of participants (85%). Statins of different potencies, such as atorvastatin, which is being tested in the STAREE and PREVENTABLE trials at a dosage of 40 mg per day, might have yielded different results.^6,7 This includes variations in side effects, a critical aspect not evaluated in this study. Third, it is unfortunate that parallel analyses for the incidence of dementia and disability are unlikely to be informative, given the significant challenges associated with pragmatically ascertaining those conditions. Fourth, it is unexpected that the author’s observed a larger relative reduction associated with initiating a statin for all-cause mortality versus major cardiovascular events (39% versus 14% relative reduction overall). A stronger relative effect on all-cause mortality is not what has been observed in previous randomized primary prevention trials of lipid-lowering therapy (Table 1), and suggests the presence of residual confounding. While the authors leveraged the wealth of health data available in the VA and Medicare claims, indicators of educational attainment and individual-level measures of socioeconomic status are not commonly tracked within administrative systems. With that said, the potential impact of these factors in producing bias is unclear. Higher educational attainment is strongly associated with longer life expectancy,¹¹ there is limited data on the association of educational attainment with statin initiation, while having less than a high school education was not associated with prevalent statin use for primary prevention in adults (aged 40 to 75 years) from the National Health and Nutrition Examination Surveys.¹²

Table 1.

Lipid-lowering therapy effect size for cardiovascular events and all-cause mortality across selected randomized clinical trials in primary prevention.

Study (year, sample size)	Lipid-lowering therapy	Cardiovascular events Effect measure (95 % CI)	Cardiovascular events definition	All-cause mortality Effect measure (95 % CI)	Reference
CTT pooled meta-analysis (2019, N=14,483 >75 years)	Statins	RR = 0.92 (0.73–1.16)a	Coronary heart disease death, nonfatal myocardial infarction, coronary revascularization, and stroke.	RR = 0.93 (0.85–1.03)a	Cholesterol Treatment Trialists’ Collaboration. Lancet. 2019; 393(10170):407–415.
EWTOPIA 75 (2019, N=3411 ≥75 years)	Ezetimibe	HR = 0.66 (0.37–0.98)b	Coronary heart disease death, fatal and nonfatal myocardial infarction, fatal and nonfatal stroke.	HR = 1.09 (0.89–1.34)	Ouchi Y, Sasaki J, Arai H, et al. Circulation. 2019; 140(12):992–1003.
CLEAR-OUTCOMESb (2023, N=4206)	Bempedoic acid	HR = 0.70 (0.55–0.89)c	Death from cardiovascular causes, nonfatal myocardial infarction, nonfatal stroke, or coronary revascularization.	HR = 0.73 (0.54–0.98)	Nissen SE, Menon V, Nicholls SJ, et al. JAMA. 2023;330(2):131–140.
Orkaby et alc (2023, N=86,327)	Statins	HR = 0.90 (0.88–0.92)	Cardiovascular death, myocardial infarction, stroke, or transient ischemic attack	HR = 0.63 (0.62–0.64)	Orkaby AR, Lu B, Ho Y-L, et al. J Am Geriatr Soc. Published online December 6, 2023.

CTT = denotes Cholesterol Treatment Trialists; CI = Confidence Interval; HR= Hazard Ratio; RR = Relative Risk.

^aEffect per 1 mmol/L reduction in LDL cholesterol.

^bData by age groups not publicly available, mean age was 67.9 years.

^cIn Veterans classified as frail based on the Veterans Affairs Frailty Index.

The results of Orkaby et al. by frailty status are consistent with several positive cardiovascular trials that have examined treatment effect heterogeneity as a function of deficit accumulation.^13–15 Adults with higher FI scores exhibit higher absolute risk reductions as they are at higher risk of cardiovascular morbidity and mortality, but relative risk reductions are typically consistent, recognizing such analyses are invariably underpowered. The lack of heterogeneity in a relative sense highlights the need to not solely focus on hazards ratios for decision making, as absolute risk reductions are the more relevant summary metric.¹⁶ An additional hypothesis is that deficit accumulation or the construct of frailty itself, may not be the best lens with which to evaluate the suitability of pharmacologic interventions for cardiovascular disease in older adults. For example, in the Systolic Blood Pressure Intervention Trial, intensive control of blood pressure consistently reduced incident cardiovascular disease when examined across the spectrum of deficit accumulation, but there was no evidence of cardiovascular benefit amongst older participants with lower cognitive function at baseline.^14,17 As the authors discuss, the explanation for this finding may relate to the fact that frailty, while strongly associated with mortality, is often not life-limiting. Many frail adults in the community will have life-expectancy exceeding 5 years, and so have the potential to benefit from preventive cardiovascular therapies.^18,19 Both STAREE and PREVENTABLE are well-positioned to examine this issue with respect to statin use, as both trials will include frail participants, will have suitable data to build a FI, and are ascertaining cognitive function.^6,7

In summary, the results of Orkaby et al. are reassuring, suggesting that the primary prevention benefits of statins for CVD observed in middle age populations likely also extend to older adults, including those 75 years or older and those more likely to be frail.⁹ One must still be cautious with this interpretation, as it is conceivable that the observed differences reflect educational and socioeconomic differences between patients that did or did not initiate a statin. The results in patients >85 years also suggest an important hypothesis to be tested across STAREE and PREVENTABLE, addressing whether there is a point in the continuum of chronologic age and life expectancy where there ceases to be sufficient time to benefit from the primary prevention with statins.²⁰ Like many others in the field, we will continue to anxiously wait for the answer to these important questions.