New Statin Use, Mortality and First Cardiovascular Events in Older US Veterans by Frailty Status

Abstract

Background:

Statins are part of long-term medical regimens for many older adults. Whether frailty modifies the protective relationship between statins, mortality, and major adverse cardiovascular events (MACE) is unknown.

Methods:

This was a retrospective study of US Veterans ≥65, without CVD or prior statin use seen in 2002-2012, followed through 2017. A 31-item frailty index was used. The co-primary endpoint was all-cause mortality or MACE (MI, stroke/TIA, revascularization, or cardiovascular death). Cox proportional hazards models were developed to evaluate the association of statin use with outcomes; propensity score overlap weighting accounted for confounding by indication.

Results:

We identified 710,313 Veterans (mean age (SD) 75.3(6.5), 98% male, 89% white); 86,327 (12.1%) were frail. Over mean follow up of 8(5) years, there were 48.6 and 72.6 deaths per 1000 person-years (PY) among non-frail statin-users vs nonusers (weighted Incidence Rate Difference (wIRD)/1000 person years (PY), −24.0[95% CI, −24.5 to −23.6]), and 90.4 and 130.4 deaths per 1000PY among frail statin-users vs nonusers (wIRD/1000PY, −40.0[95% CI, −41.8 to −38.2]). There were 51.7 and 60.8 MACE per 1000PY among non-frail statin-users vs nonusers (wIRD/1000PY, −9.1 [95% CI, −9.7 to −8.5]), and 88.2 and 102.0 MACE per 1000PY among frail statin-users vs nonusers (wIRD/1000PY, −13.8[95% CI, −16.2 to −11.4]). There were no significant interactions by frailty for statin users vs non-users by either mortality or MACE outcomes, p-interaction 0.770 and 0.319, respectively. Statin use was associated with lower risk of all-cause mortality (HR, 0.61 (0.60-0.61)) and MACE (HR 0.86 (0.85-0.87)).

Conclusions:

New statin use is associated with a lower risk of mortality and MACE, independent of frailty. These findings should be confirmed in a randomized clinical trial.

Introduction

The expanding population of older adults is at risk for both cardiovascular disease (CVD) and frailty. Frailty is a multidimensional state of vulnerability associated with increased risk of disease, disability, and death.¹ Older adults with frailty are at greatest risk of CVD, however they are often regarded as unable to tolerate or benefit from preventive therapies such as statins.^2-4 Concerns include myalgias and functional impairment, drug-drug interactions, polypharmacy, and cognitive changes.⁵ However, data on older adults from lipid altering clinical trials is limited, and even fewer individuals with frailty have been studied.⁶ The lack of data on real-world use of statins in older adults with frailty has created an evidence gap.^{7, 8} Furthermore, while older adults with frailty are at highest risk of CVD,³ there has been increased focus on deprescribing statins in older adults due a poorly substantiated perception that risks associated with treatment exceed benefit.^{2, 9} The 2018 AHA/ACC Cholesterol guidelines for those over age 75 specifically state that: “decisions to not initiate statins, or even to deprescribe them, are reasonable in older adults when aggregate risks outweigh potential for meaningful benefit.”² Yet, the risks are not specified, the absence of data does not signify harm, and it is possible that older adults with frailty may benefit the most from preventive treatment with statins. An observational study of 120,173 primary care patients in France reported that stopping statins taken for primary prevention at age 75 was associated with higher risk of CVD and mortality, even among those with frailty.¹⁰ Whether older adults with frailty benefit from statins for prevention of mortality and major adverse cardiovascular events (MACE) is unknown.

Using Veterans Health Administration (VHA) data, a Veterans Affairs Frailty Index (VA-FI) was developed and validated in a cohort of nearly 3 million Veterans.¹¹ Over a third of Veterans aged 65 and older were identified as frail, and those with frailty have significantly higher risk of mortality and CVD compared to those who are not frail.^{11, 12} In this study we hypothesized that frail Veterans newly prescribed statins for primary CVD prevention would have lower rates of all-cause deaths and MACE compared to those who are frail and do not receive statins.

Methods

Study Design and Population

This was a retrospective study that included all regular users of the VHA between 2002-2012 were eligible and identified through the VHA corporate data warehouse.¹³ A regular user had 2 primary care visits within 2 years prior to entry into the cohort; additionally, Veterans with healthcare appointments where a statin could not be prescribed, such as audiometry or prosthetics, were not included. Of these regular users, we included only those who were ≥65 and had no CVD diagnoses or statin prescriptions prior to their first qualifying visit.

We excluded persons with unlikely birth or death dates, and those with questionable demographic information, i.e. missing sex or race, were excluded. Prior CVD was defined as a history of myocardial infarction (MI), transient ischemic attack or stroke, peripheral arterial disease, or coronary revascularization. Additionally, those who died within 150 days of statin prescription were excluded, as in prior analyses because limited life expectancy is a contraindication to statin use.⁷ (Supplemental Figure 1) Finally, to ensure complete identification of statin use and outcomes, VHA data were linked to Centers for Medicare & Medicaid Services (CMS) data.

Exposure

To mitigate bias, a new-user design was employed.¹⁴ Individuals with any prior statin use in VHA pharmacy or CMS records were excluded. All statins approved for use in the US were identified in VHA pharmacy and CMS drug data,¹⁵ according to generic or trade name. (Supplemental Table 1) At entry, the entire cohort consisted regular VHA users who and had no documented prior statin prescriptions (i.e., nonusers). Index date was defined as date of visit when veteran turn 65. For those who were prescribed a statin after the first index date, a second index date was defined as date of first statin prescription identified in either VHA or CMS data. Statin dose was extracted to identify low and moderate/high intensity, as defined by national guidelines.² (Supplemental Table 2)

Frailty

Frailty was measured at time of entry into the cohort for the entire population and recalculated at time of new statin use for those who became statin users. Frailty was defined according to the validated VA-FI, based on the Rockwood cumulative deficit model of frailty.^{11, 16} This theory posits that over a lifetime individuals accumulate deficits across multiple health domains which can be captured in a comprehensive geriatric assessment or electronic claims data.¹⁷ These deficits can be summed and divided by a predetermined denominator to calculate a frailty score ranging from 0.0-1.0, with scores >0.2 considered frail.^{11, 18} The original VA-FI included 31 variables related to comorbidities (anemia, atrial fibrillation, cancer, stroke, coronary artery disease, diabetes, heart failure, hypertension, kidney disease, liver disease, lung disease, thyroid disease, osteoporosis, and incontinence), cognition and mood (dementia, anxiety, depression), physical function (arthritis, use of durable medical equipment, falls, fatigue, gait abnormality, Parkinson’s disease, peripheral vascular disease), nutrition (cachexia or debility), sensory impairment (hearing loss, peripheral neuropathy, vision loss), and general health status (chronic pain, failure to thrive, weight loss). As the current study includes participants free of CVD at baseline, the VA-FI was modified to remove coronary artery disease and stroke as has been done previously.¹²

In a sensitivity analysis we used an alternative definition of frailty that relies on 12 variables related to the physical manifestations of frailty developed by Figueroa and colleagues. Those with 0 items are non-frail, 1 are prefrail and 2+ are frail.¹⁹

Primary outcome

The co-primary outcomes were all-cause mortality and MACE. Mortality was extracted from the National Death Index.²⁰ Secondary outcomes were myocardial infarction (MI), ischemic stroke, revascularization with coronary artery bypass graft surgery or percutaneous coronary intervention, and composite CVD events (first occurrence of MI, ischemic stroke, or revascularization procedures).²¹ Both VHA and CMS claims were used to identify events (Supplemental Table 1).

Follow-up time was defined as time from entry into the cohort until date of death recorded in National Death Index, MACE occurred, or an individual in the control group switched to statin use for primary outcomes. If an event was observed the same day as a new statin prescription, it was counted as an event in the statin nonuse group. For secondary outcomes, follow-up time was defined as time from entry into cohort to time of event, or loss of follow-up, or switch to statin use.

Covariates

At index date, information on age, sex, race, ethnicity, body mass index (BMI), and region of the country was collected. Comorbidities associated with both exposure (statin use) and outcome (mortality, CVD events) were identified according to diagnosis codes from the VHA corporate data warehouse and CMS claims at entry into the cohort and updated for statin users at the time of switch to statin use.²² These included hyperlipidemia, hypertension, diabetes, liver disease, non-statin cholesterol-lowering medications, smoking status, anemia, cancer, heart failure, atrial fibrillation, chronic kidney disease, substance abuse, and mental health disorders (Supplemental Table 1). Additionally, variables specific to aging such as arthritis, dementia, polypharmacy, fatigue, and gait abnormality were included to capture reasons statins might or might not have been prescribed.

Smoking status at index date was defined using a validated algorithm.²³ Blood pressure medication use was defined as an active prescription 180 days prior to within 30 days of index date. Polypharmacy was defined as use of ≥5 drug classes.

Statistical Analysis:

As is recommended, to address confounding by indication, or reasons why a statin might be prescribed or not prescribed, a propensity score was developed using the above variables and interaction terms for age, sex, and race.^{24, 25} The overlap weighting method of using the propensity score was employed to minimize influence of extreme propensity scores on model output and ensure the entire population contributed to the analysis.²⁶

Cox proportional hazards models were fit to evaluate association of statin use with outcomes, using overlap weighting. All models were stratified by frailty. To account for potential bias due to the immortal time prior to first statin use, we included the study period prior to initiating statin for statin users as separate observations in the analysis. Robust covariance estimators were used in the model to account for intra-cluster correlation. The proportional hazards assumption was assessed using interaction between follow-up time (nature log scale) with statin use. Weighted generalized linear models with time to event as offset and repeated measurements were used to estimate event rates comparing statin users with nonusers.

In secondary analyses, prespecified subgroups were examined: age (10-year increments from 65 to ≥85 years), sex, race, prevalent diabetes, hyperlipidemia, dementia, arthritis, polypharmacy, and statin intensity (low vs moderate/high). We conducted several sensitivity analyses: first we restricted to those with ≥2 statin prescriptions during follow up, then we restricted to those who had a cholesterol level available in the 6 months before or 7 days after a statin prescription and stratified by quartile of total cholesterol and LDL. Finally, we repeated the main analysis using 0.05 increments of the FI from 0.00 to >0.35. Tests of interaction by frailty status were run for all subgroups. The propensity score of the statin use versus no use was used to re-create the overlap weighting. Then the Cox proportional hazard model was fit with the new weights, and the interaction term was added for testing. Because of the potential for type I error due to multiple comparisons, secondary analyses and endpoints should be interpreted as exploratory.

All analyses were conducted in SAS Enterprise Guide 7.1 (SAS Inc). P<.05 (2-sided) was considered statistically significant.

Ethics

This study was approved, and the requirement for obtaining patient informed consent was waived, by the VHA Boston Institutional Review Board.

Role of the Funding source

Funding for this work was obtained from the U.S. Veterans Administration and National Institute on Aging. The funders had no role in the study’s design, conduct, and reporting.

Results

Cohort Characteristics

There were 4,425,545 Veterans aged ≥65 seen at VHA between 2002-2012, 1,212,105 had no prior statin prescription, 285,459 were excluded due to missing demographic data and 209,572 due to prior CVD events. The resulting total of 710,313 Veterans were free of CVD and had no prior statin use. Among these, 86,327 (12.2%) were frail at baseline. Frail Veterans were more likely to have a statin prescription during follow up (27.9%) vs non-frail Veterans (23.6%).

Overall, mean age (standard deviation [SD]) was 75.3 (6.5) years, range (65-107), 97.9% were male, and 89.1% were white. The most commonly prescribed statin was simvastatin (84.8%), followed by lovastatin (10.3%), pravastatin (3.1%), fluvastatin (1.2%), and then atorvastatin, rosuvastatin and pitavastatin. Table 1 shows demographics at index date, before and after propensity score weighting, by baseline frailty. Veterans prescribed statins were more likely to have CVD risk factors and chronic conditions, polypharmacy, and use of other non-statin lipid lowering drugs. Those not prescribed a statin were more likely to have dementia and liver disease. Additionally, among the frail, those with anemia, atrial fibrillation, and heart failure were less likely to receive a statin. Following propensity score overlap weighting, all characteristics were balanced between frailty groups, with standardized differences all <11% (Supplemental Figure 2).

Table 1.

Demographics of 710,313 Veterans ≥65, free of CVD before and after propensity score overlap weighting

	Before weighting				After Propensity Score weighting
	Frail N=86,327		Not Frail N=623,986		Frail N=86,327		Not Frail N=623,986
Variable	No statin N=62,234	Statin N=24,093	No statin N=476,659	Statin N=147,327	No statin N=62,234	Statin N=24,093	No statin N=476,659	Statin N=147,327
Age, mean (SD), y	78.3 (6.6)	77.3 (6.0)	75.0 (6.6)	74.5 (5.6)	77.8 (3.0)	77.3 (4.5)	75.1 (2.5)	74.4 (4.6)
BodyMass Index,mean, (SD)	27.2 (5.4)	28.9 (5.3)	27.3 (4.8)	28.5 (4.7)	28.0 (2.6)	28.5 (4.0)	28.0 (1.9)	28.3 (3.8)
Sex (%)
Male	96.3	96.4	98.1	98.3	96.8	96.2	98.3	98.3
Female	3.7	3.6	1.9	1.7	3.2	3.8	1.8	1.7
Race/ethnicity, %
White	89.8	87.9	89.3	88.7	89.1	88.1	88.9	88.9
Black/African American	8.7	10.4	8.9	9.3	9.2	10.2	9.2	9.2
Other**	1.5	1.7	1.8	1.9	1.6	1.7	1.9	1.9
Hispanic	3.3	4.9	4.0	5.6	3.4	5.0	3.8	5.6
Smoking status (%)
Current	9.8	10.8	11.8	12.7	10.7	10.5	13.5	12.4
Former	71.7	64.2	68.8	60.7	66.0	67.3	62.4	63.1
Never	18.5	25.0	19.4	26.6	23.3	22.2	24.5	24.5
Comorbidities (%)
Hypertension	81.7	90.9	59.4	75.9	87.4	89.2	70.8	72.7
Hyperlipidemia	17.6	58.1	14.7	56.6	40.0	42.6	38.3	45.3
Arthritis	59.3	67.9	31.1	38.7	63.3	66.1	36.5	36.8
Cancer	53.2	58.8	31.2	37.5	57.2	57.0	36.2	35.8
Diabetes	22.1	44.0	9.9	25.8	32.4	38.5	17.4	21.9
Anemia	42.8	39.1	10.0	10.5	40.9	39.5	10.4	10.3
Depression	36.2	38.9	10.4	12.6	36.8	38.6	11.9	12.1
Atrial Fibrillation	19.5	16.2	5.2	5.1	17.8	17.3	5.3	5.1
Dementia	19.5	13.0	4.7	3.0	14.9	14.4	3.8	3.2
Fatigue	17.6	16.9	3.1	3.4	16.9	17.2	3.4	3.4
Congestive Heart Failure	16.4	13.4	2.2	2.2	14.5	14.4	2.3	2.3
Gait abnormality	13.9	13.6	1.7	1.8	13.3	13.5	1.8	1.7
Sleep Apnea	6.1	9.1	2.5	3.6	7.6	8.3	3.3	3.3
Post-Traumatic Stress Disorder	5.0	7.1	2.1	2.3	5.5	7.0	2.4	2.2
Chronic Kidney Disease	4.3	5.5	0.5	0.9	5.3	5.0	0.8	0.8
Schizophrenia	1.7	1.6	0.9	0.7	1.4	1.7	0.8	0.7
Liver Disease	2.0	0.8	0.5	0.2	1.1	1.1	0.3	0.2
Substance Abuse	0.4	0.3	0.2	0.1	0.3	0.3	0.2	0.1
Medication use (%)
Polypharmacy (>5 drug classes)	64.7	85.4	30.7	55.3	76.4	81.7	47.9	48.9
Diuretics	36.0	44.2	24.8	35.4	41.0	42.7	32.1	32.9
ACE Inhibitor Use	30.7	46.3	23.4	38.3	38.0	43.2	32.2	34.8
Alpha Blocker	25.7	27.7	20.2	21.4	26.6	27.9	21.7	21.1
Calcium Channel Blocker	24.4	31.5	17.9	24.4	27.3	30.7	22.5	22.9
Beta Blocker	19.8	28.1	14.0	21.0	23.7	26.8	19.1	19.2
Angiotensin Receptor Blocker	4.9	8.8	3.4	5.7	6.6	7.9	4.9	5.1
Non-statin Lipid Lowering Drug	3.4	7.9	2.5	5.9	5.8	7.3	4.8	5.5

Primary Outcomes

Veterans were followed for a mean of 7.6(4.5) years during which there were 381,387 all-cause deaths and 266,874 MACE events. Before propensity score adjustment, there were 52.54 and 79.91 all-cause deaths per 1000 person-years (PY) among statin users and nonusers, respectively (weighted incidence rate difference [wIRD]/1000 PY, −27.4 [95% CI, −27.8 to −26.9]). In the frail group, there were 90.4 and 130.4 all-cause deaths per 1000PY among statin users and nonusers, respectively (wIRD/1000 PY, −40.0 [95% CI, −41.8 to −38.2]), compared to the non-frail group, where there were 48.6 and 72.6 all-cause deaths per 1000PY among statin users and nonusers, (wIRD/1000PY, −24.0 [95% CI, −24.5 to −23.6]) (Table 2). Before propensity score adjustment, there were 54.0 and 65.6 MACE events per 1000PY among statin users and nonusers, respectively, wIRD −10.6 (−11.1 to −10.0).

Table 2.

Association of statin use, mortality, and MACE in 710,313 US Veterans ≥65 years by frailty

Outcome	Statin User (N=171,420) Weighted rate/1000 PY	No Statin (N=710,313) Weighted rate/1000 PY	Weighted Incidence Rate Difference/1000 PY (95% CI)	Hazard Ratio (95% CI)	p-value	p- interaction
All-Cause Mortality n=381,387	52.54	79.91	−27.37 (−27.82, −26.92)	0.61 (0.60,0.61)	<0.001	…
Frail N=62,373	90.37	130.36	−40.00 (−41.78, −38.22)	0.63 (0.62,0.64)	<0.001	0.770
Not frail N=319,014	48.62	72.63	−24.01 (−24.46, −23.55)	0.61 (0.61,0.62)	<0.001
4-pt MACE*n=266,874	54.01	65.58	−10.56 (−11.14, −9.98)	0.86 (0.85,0.87)	<0.001
Frail N=37,684	88.18	101.98	−13.80 (−16.19, −11.41)	0.90 (0.88,0.92)	<0.001	0.319
Not frail N=229,190	51.68	60.75	−9.08 (−9.66, −8.49)	0.87 (0.86,0.88)	<0.001

^*Cardiovascular death, myocardial infarction, stroke or transient ischemic attack, or revascularization (coronary artery bypass graft surgery or percutaneous coronary intervention)

MACE = Major Adverse Cardiac Events

In the frail group where there were 88.2 and 102.0 MACE per 1000PY among statin users and nonusers, respectively (wIRD/1000PY, −13.8 [95% CI, −16.2 to −11.4]), compared with the non-frail group in which there were 51.7 and 60.8 MACE per 1000 PY among statin users and nonusers, (wIRD/1000PY, −9.1 [95% CI, −9.7 to −8.5]).

After propensity score overlap weighting was applied, statin use was significantly associated with a lower risk of overall all-cause mortality (HR, 0.61 [95% CI, 0.60 to 0.61]) and MACE (HR, 0.86 (0.85-0.87), with no significant interaction according to frailty group, p-interaction 0.770 and 0.319, respectively (Table 2).

Secondary Outcomes

There were 162,546 revascularization events, 49,598 MIs, and 63,718 strokes. For all outcomes, there were more events among the frail cohort. After adjustment, HR for revascularization events among the non-frail statin users vs non-users was 0.82 (95% CI, 0.81-0.83), and 0.87 (95% CI 0.84-0.90) among the frail statin users vs non-users. Corresponding HRs for MI were 1.02 (95% CI, 0.99 to 1.03) and 1.04 (95% CI 0.98-1.10) and for stroke 0.92 (95% CI 0.90-0.94) and 0.93 (0.88-0.97) for the non-frail and frail statin users vs non-users, respectively. All p-interactions >0.05. (Table 3).

Table 3.

Association of statin use and MACE outcomes in 710,313 US Veterans ≥65 years by frailty

Outcome	Statin User (N=171,420) Weighted rate/1000 PY	No Statin (N=710,313) Weighted rate/1000 PY	Weighted Incidence Rate Difference*/ 1000PY (95% CI)	Hazard Ratio (95% CI)	p-value	p- interaction
Major cardiovascular events
Revascularization N= 162,546	31.49	39.85	−8.36 (−8.79,−7.92)	0.82 (0.81,0.83)	<0.001	…
Frail N=22,695	49.29	60.49	−11.21 (−12.95,−9.46)	0.87 (0.84,0.90)	<0.001	0.286
Not frail N=139,851	29.63	37.08	−7.45 (−7.89,−7.01)	0.82 (0.81,0.83)	<0.001
MI N= 49,598	10.29	10.34	−0.05 (−0.28,0.17)	1.01 (0.99,1.03)	0.326	…
Frail N=6,860	15.15	14.76	0.39 (−0.46, 1.23)	1.04 (0.98,1.10)	0.193	0.368
Not frail N=42,738	9.75	9.74	0.01 (−0.22,0.25)	1.02 (0.99,1.04)	0.234
Stroke N= 63,718	12.60	14.19	−1.59 (−1.85,−1.33)	0.91 (0.90,0.93)	<0 .001	…
Frail N=9,652	20.07	22.64	−2.57 (−3.59,−1.54)	0.93 (0.88,0.97)	0.002	0.435
Not frail N=54,066	11.79	13.05	−1.26 (−1.52,−1.00)	0.92 (0.90,0.94)	<0.001

^*Cardiovascular death, myocardial infarction, stroke or transient ischemic attack, revascularization (coronary artery bypass graft surgery or percutaneous coronary intervention)

MACE = Major Adverse Cardiac Events

Sensitivity analysis

Results remained consistent for both mortality and MACE when the alternative physical frailty definition was used. (Supplemental Table 3).

Subgroup analyses

In subgroup analysis, the association between statin use and lower risk of mortality and MACE remained statistically significant across all age groups, with greatest absolute benefit for mortality in those >85 years who were frail (wIRD −47.98 (−55.1 to −40.85)). (Figure 1) For mortality, there was a statistically significant interaction for statins by frailty, comparing users vs non-users, in the 65-75 age group (p<0.01), however the HRs were similar, 0.60 (0.58-0.63) and 0.63 (0.62-0.64) for frail and non-frail groups, respectively. There were no meaningful interactions according to frailty status among those 75-85 and 85 and older, p=0.612 and 0.159. Across all age groups, there were no statistically significant interactions for MACE, all p>0.05. (Figure 2) Additionally, results remained consistent for both mortality and MACE when granular levels of the FI (0 to >0.35) were considered, results not shown.

Figure 1:

Association of Statin Use and Mortality Among 710,313 US Veterans ≥65, Free of CVD, According to Frailty Status, Stratified by Sex, Race, Age, Diabetes, Dementia, Hyperlipidemia, Polypharmacy, BMI, and LDL LDL quartile cutoffs in mg/dL <101, 101-120, 121.0-141.3 and >=141.4; min: 25 max: 993

Figure 2:

Association of Statin Use and MACE Among 710,313 US Veterans ≥65, Free of CVD, According to Frailty Status, Stratified by Sex, Race, Age, Diabetes, Dementia, Hyperlipidemia, Polypharmacy, BMI, and LDL LDL quartile cutoffs in mg/dL <101, 101-120, 121.0-141.3 and >=141.4; min: 25 max: 993

There were 14,699 women included, and results remained unchanged when stratified by sex for mortality and MACE outcomes.(Figures 1 and 2). Results additionally remained consistent when stratified by for race, dementia, diabetes, BMI, polypharmacy, hyperlipidemia diagnosis, and LDL cholesterol levels (Figures 1 and 2). We further considered statin intensity, additional race categories, ethnicity, 2+ statin prescriptions in follow up, and total cholesterol (Supplemental Tables 4–8) and results remained similar. A significant interaction for Asians according to frailty for mortality, p=0.006, suggesting greater benefit for statins among those of Asian descent (Supplemental Table 5).

Discussion

In this study of 710,313 US Veterans aged ≥65 and free of CVD, frailty did not modify the expected association of statins on mortality or MACE outcomes. Rather, statins were associated with the same or greater reduction in mortality and first MACE events among frail Veterans compared to those not prescribed statins. In fact, Veterans with frailty prescribed statins had fewer events compared to those without frailty, with greater absolute risk reduction. Across multiple subgroups, a consistent pattern of lower risk for mortality and MACE remained consistent for those with frailty compared to those without. Moreover, when an alternate definition of frailty that relies on physical manifestations of frailty was used results were nearly identical. To our knowledge, this report is the first to examine the role of statins on mortality and CVD prevention, specifically focusing on the role of frailty.

Current risk guidelines for statins in primary prevention encourage clinicians to begin with a CVD risk assessment. Traditional risk factors such as smoking, hyperlipidemia, diabetes, and hypertension are reviewed alongside age, sex, and race, non-modifiable risk factors. Greater chronological age is the strongest CVD risk factor, but biologic age also deserves assessment.²⁷ Some adults who have lived 80 years are physiologically similar to 60-year-olds, and vice versa. The spectrum of rates of biological aging relates directly to heterogeneity in health status seen in the aging population and leads to the difficulty of making a single recommendation for all older adults. Frailty provides critical perspective regarding biologic or physiologic, age.¹

Epidemiologic studies suggest older adults with frailty are at the highest risk of CVD and mortality.⁴

Since older adults with frailty have the highest risk of mortality and CVD it may not be surprising that they had greatest associated benefit from statins in this study. In fact, similar findings have been reported in recent data from heart failure trials that tested sacubitril/valsartan and dapagliflozin and found that those with frailty had the greatest reduction in risk of outcomes, including hospitalizations and death, because it is the frail older adult who is at the highest risk of having these events.^{28, 29}

Because frailty and CVD have shared pathophysiology, in part through inflammatory pathways, it is possible these findings may be driven by the anti-inflammatory properties of statins.³⁰ This may also explain, in part, why our study showed a greater reduction in mortality risk compared to CVD events, similar to prior work by our group.⁷ The ongoing STAtins in Reducing Events in the Elderly (STAREE) trial in Australia and Pragmatic Evaluation of Events and Benefits of Lipid-Lowering in Older Adults (PREVENTABLE) trial in the US, both primary prevention trials of atorvastatin vs placebo for adults ≥70 and 75 respectively, will definitively elucidate the role of statins for primary prevention among older adults, including those with frailty.^{31, 32}

Only a few reports have addressed the role of statins in frail populations. A retrospective study using French national health insurance claims data followed 120,173 adults free of CVD who had reached 75 years and had two years of prior statin use to evaluate the association of stopping statins after age 75 on CVD events. There were significant increases in CVD events following statin cessation, among those with frailty and without, HR 1.26 (1.07-1.48) and 1.42 (1.18-1.70), respectively.¹⁰ This practice of discontinuing statins in older adults is increasingly common and is based on the premise that there is limited evidence to justify statin use after age 75.² An Australian cohort of 1,665 men ≥70, mostly with CVD, examined the relationship between statin use, long term institutionalization and death and reported no difference in outcomes among those prescribed statins among those who were frail vs non frail, with a HR for mortality of 1.24 (0.71-2.17).³³

Our findings fill an important gap in the literature as older individuals living with frailty are a growing population who have not been included in statin trials.¹² While evidence suggests older adults with limited life expectancy should not be prescribed a statin for primary prevention,³⁴ this recommendation is only relevant to a small minority of older adults. Those with frailty may have shorted estimated life expectancy compared to age matched peers, but frailty itself is not a life limiting condition.¹¹ In fact, among frail older adults, most have a life expectancy within the 2-5 year timeframe needed to derive benefit from statins.^{11, 35} Prevention of CVD events could significantly improve quality of life in this vulnerable population, though must be balanced with the potential for increased risk of adverse reactions and drug-drug interactions that are more common in those with frailty.⁵

This study has important strengths. The cohort is derived from the largest single payer healthcare system in the US, including both electronic health records and insurance claims data to provide near complete capture of exposure and events. Availability of the entire VHA enabled identification of large numbers of new statin users with important geriatric conditions often underrepresented in clinical trials. Validated methods identified frailty, and rigorous statistical methodology reduced confounding by indication. Older Veterans are also a population with endemic multimorbidity, polypharmacy, and socioeconomic challenges, such that utility of statins in this population may have important generalizability to typical patients who are often very different from study populations in clinical trials.

This study has important limitations. This study is observational and retrospective, and although attempts were made to carefully adjust for confounding and bias, residual bias likely remains. Similarly, while a new-user design allows for a clean cohort without prior statin use, this may introduce additional selection bias. Additionally, because frailty is identified using claims data those who are frail may also have greater use of the healthcare system and therefore more preventive care opportunities, resulting in confounding and reverse causation. Most of the current cohort are men, limiting generalizability, although 15,000 women included. Additionally, relying on claims may underestimate conditions such as dementia. Information on adverse events such as myalgias, diabetes, polypharmacy, and drug-drug interactions were unavailable. Finally, the most prescribed statin in this study was simvastatin, a medication that no longer reflects current practice and less potent than more common statins in use today such as atorvastatin and rosuvastatin.

In conclusion, new statin use is associated with a significantly lower risk of all-cause mortality and MACE, regardless of frailty status, among US Veterans ≥65. These findings support the use of statins as prescribed in older adults, including those with frailty, and should be confirmed in a randomized clinical trial.

Key Points:

• The role of statin use for prevention of mortality and major cardiovascular events among frail older adults is unknown.

• In this electronic health record study of 710,313 US Veterans ≥65 years and free of CVD, of whom 12% were frail, statin therapy was associated with a 39% lower risk of mortality and 14% lower risk of first major cardiovascular events, independent of frailty.

• Among those who were frail, statin use was associated with the same or greater reduction in mortality and cardiovascular events compared to those who were not frail.

Why does this matter?

These findings support the use of statins for primary prevention in older adults, including those with frailty.