Statin Use Over 65 Years of Age and All-Cause Mortality

A 10-Year Follow-Up of 19 518 People

Sophia Eilat-Tsanani, MD; Elad Mor, MD; Yochai Schonmann, MD, MSc


J Am Geriatr Soc. 2019;67(10):2038-2044. 

In This Article


There were 37 230 members aged 65 years or older in the northern district of CHS on January 1, 2004. After excluding those with preexisting ASCVD and those who had died before cohort entry (January 1, 2007), 19 518 older CHS members remained eligible for primary prevention and were included in our analyses (Figure 1). At baseline, 5270 (27%) of the cohort members were statin users. Statin users had lower LDL-C levels compared with nonusers: mean = 113 (standard deviation [SD] ±32) mg/dl and 120 (SD ±33) mg/dl, respectively; P < .0001. Statin users were less likely to have a waiver from social security payments and more likely than nonusers to have cardiovascular risk factors, regardless of their age. Smoking, obesity, and diabetes were more common among those younger than 75 years; diagnoses of hypertension and hyperlipidemia were slightly more common in those older than 75 years (Table 1).

Figure 1.

Study population.

Cohort participants contributed 141 400 person-years for analysis: median = 9.7; interquartile range = 4.4-10.0 years. One-quarter of statin users younger than 75 years and 13% of those older than 75 years fully adhered to statin therapy throughout the follow-up (ie, had claimed enough doses of a statin to cover ≥80% of their follow-up). Adherence rates were somewhat higher when calculated by summing the absolute number of pills claimed, regardless of their dose or strength (43% of those younger than 75 years and 29% of those older had adhered to therapy). Over the 10 years of follow-up, 5151 (26%) of the participants had died, and 37% had an ASCVD (fatal or nonfatal) event. Supplementary Table S1 shows age-group-specific data on follow-up, statin use, and crude mortality and ASCVD rates among statin users and nonusers.

We explored the crude and the adjusted association between adherence to statins and mortality and ASCVD (Figure 2, Supplementary Table S2) After adjusting to age, sex, sociodemographic characteristics, time-updated cardiovascular risk, and time-updated cardiovascular morbidity, adherence to statins was associated with lower all-cause mortality (hazard ratio [HR] = .66; 95% confidence interval [CI] = .56-.79; P < .0001) and with lower ASCVD rates (HR = .80; 95%CI = .71-.90; P < .0001). The associations remained robust in stratified analyses regardless of age group or sex. The association between statin adherence and reduced all-cause mortality was stronger among those older than 75 years compared with those younger (HR = .56; 95% CI = .44-.71 vs HR = .80; 95% CI = .64-1.00; P[interaction] = .046). Similarly, the association between adherence to statins and reduced ASCVD rates was also stronger among those older than 75 years (P[interaction] = .012). There was no statistical evidence for a differential effect of sex on the association between statin adherence and mortality or ASCVD (P[interaction] = .558 for mortality, and P[interaction] = .927 for ASCVD) (Supplementary Table S2).

Figure 2.

Association between statin adherence*, all-cause mortality, and atherosclerotic cardiovascular disease (ASCVD)**.
Hazard ratios were estimated from a Cox regression model with current age as the underlying time scale with 95% confidence intervals in brackets. P values are <.0001 and derived from likelihood-ratio tests unless specifically specified otherwise (a- P = .082).
*Adherence was categorized as proportion of days covered (PDC) more than 80%. PDC was calculated for each participant by dividing the total amount of medication claimed from the pharmacy (expressed in multiples of the World Health Organization daily defined doses) by the total number of follow-up days. Cumulative PDC was calculated for each calendar year separately and modeled as a time-updated variable.
**ASCVD was defined as the first occurrence of ischemic heart disease or stroke (captured in the Clalit Health Services [CHS] chronic disease register) or all-cause mortality, after excluding patients who had a registered malignancy diagnosis (censured on the date of diagnosis). Because malignancy and cardiovascular disease are the leading causes of death among older people, this provides an approximation of cardiovascular mortality.
Model 1: Crude; adjusted for age (as the underlying time scale) and sex.
Model 2: Minimally adjusted; additionally adjusted for calendar period (2007-2011 vs 2012-2016), ethnicity, eligibility for social security waiver, and marital status.
Model 3: Fully adjusted; additionally adjusted for smoking status, obesity, hypertension, hyperlipidemia, diabetes, cerebrovascular accidents, ischemic heart disease, carotid artery disease, peripheral vascular disease, congestive heart failure, cardiomyopathy, valvulopathy, chronic renal failure, and malignancy.
All additional diagnoses in this model were obtained from the CHS chronic disease register, and they were modeled as time-updated variables.

Adherence to statins remained associated with lower rates of all-cause mortality when we defined adherence in narrower bands of statin use (P[trend] < .0001 with increased PDC), as well as when we used the pooled number of pills instead of DDDs to define cumulative adherence. The findings remained robust when we excluded patients with diabetes and renal failure from the analyses. Finally, the results remained similar when the analysis was limited to those who had used statins at baseline (Supplementary Tables S3-S7).