If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Clinical guidelines recommend statins for patients with atherosclerotic cardiovascular disease (ASCVD), but many remain untreated. The goal of this study was to assess the impact of statin use on recurrent major adverse cardiovascular events (MACE). This study used medical records and insurance claims from 4 health care systems in the United States. Eligible adults who survived an ASCVD hospitalization from September 2013 to September 2014 were followed for 1 year. A multivariable extended Cox model examined the outcome of time-to-first MACE, then a multivariable joint marginal model investigated the association between post-index statin use and nonfatal and fatal MACE. There were 8,168 subjects in this study; 3,866 filled a statin prescription ≤90 days before the index ASCVD event (47.33%) and 4,152 filled a statin prescription after the index ASCVD event (50.83%). These post-index statin users were younger, with more co-morbidities. There were 763 events (315/763, 41.3% terminal) experienced by 686 (8.4%) patients. The adjusted overall MACE risk reduction was 18% (HR 0.82, 95% CI 0.70 to 0.95, p = 0.007) and was more substantial in the first 180 days (HR 0.72, 95% CI 0.60 to 0.86, p <0.001). There was a nonsignificant 19% reduction in the number of nonfatal MACE (rate ratio 0.81, 95% CI 0.49 to 1.32, p = 0.394) and a 65% reduction in the risk of all-cause death (HR 0.35, 95% CI 0.22 to 0.56, p <0.001). In conclusion, we found a modest increase in statin use after an ASCVD event, with nearly half of the patients untreated. The primary benefit of statin use was protection against early death. Statin use had the greatest impact in the first 6 months after an ASCVD event; therefore, it is crucial for patients to quickly adhere to this therapy.
Millions of adults in the United States (US) are affected by cardiovascular events each year, reducing their quality of life and increasing their risk for death.
Lipid-lowering therapy is a cornerstone of secondary ASCVD prevention, but many patients remain untreated. The near-term consequences of medication underutilization after an acute event are not well understood. Although there is a panoply of lipid-lowering therapies, we chose to limit this analysis to the American Heart Association guideline-recommended first-line medication of statins.
Hence, the purpose of this report is to study the association between statin use/nonuse and recurrent major adverse cardiovascular events (MACE) in a cohort of patients across the US who had a recent acute ASCVD event.
This study is a collaboration between 4 US health care systems (Baylor Scott & White [BSW], Texas; Henry Ford, Michigan; Geisinger, Pennsylvania; and Marshfield Clinic, Wisconsin) participating in the Health Care Systems Research Network (HCSRN). The HCSRN maintains data standards between health care organizations to create a common data model to assemble pooled data sets to answer multicenter research questions.
For this project, BSW developed and distributed code to collaborators to extract the minimum necessary care, administrative, and claims data to yield a deidentified dataset. This study received approval from the Baylor Scott & White Research Institute's institutional review board through expedited review and a waiver of informed consent. Henry Ford and Marshfield ceded to the Baylor Scott & White Research Institute's institutional review board with a reliance agreement, and Geisinger's institutional review board determined that the study did not involve human subjects and was not subject to their oversight.
We used a retrospective cohort design to answer our research question. Eligible adults survived an index ASCVD hospitalization from September 30, 2013 to September 30, 2014 and were followed up to 1 year. We extracted demographics (gender, age, race, Hispanic ethnicity, and insurance type), and used International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes and prescription fills during the year prior to the index to identify co-morbidities (type II diabetes mellitus, chronic kidney disease, hyperlipidemia, and hypertension). We recorded statin use 90 days before and after the index. The primary outcome was MACE (acute myocardial infarction, ischemic stroke, revascularization, unstable angina, acute presentation of chronic ASCVD, all-cause death) within 1 year of index.
We created a multivariable extended Cox model with robust sandwich estimates to investigate the association between post-index statin use (assessed through prescription fill) and time-to-first recurrent MACE while accounting for clustering (by healthcare system), demographics (gender, age, race), and co-morbidities (type II diabetes mellitus, chronic kidney disease, hyperlipidemia, hypertension, ASCVD history). After analyzing the time-to-first MACE, we considered an alternate analytic framework using the marginal joint model of Huang and Wang
This model implies that the subject-specific event rate is positively correlated with the risk of terminal event (i.e., subjects who survive longer tend to have lower event rates); although, this model has also been shown to yield unbiased estimates for independent processes.
Herein, we distinguished MACE as being nonfatal (recurrent) or fatal (terminal). Because the model could not provide stable estimates for the third or fourth events owing to the small event counts, we included information only through the second MACE. We used the same covariates as the first model but dichotomized age and race to improve computational efficiency. We used a nonparametric bootstrap method for clustered data by repeatedly sampling the subjects with replacement to estimate standard errors and obtain 95% confidence intervals (CIs) and p-values. Analyses were performed in SAS version 9.4 (Cary, North Carolina) and the ‘reReg’ R package.
There were 8,168 patients in this study, with the 3 leading causes of entry being acute presentation of chronic ASCVD (2,337, 28.6%), acute myocardial infarction (2,156, 25.4%), and ischemic stroke (1,458, 17.9%). There were 3,866 patients (47.33%) who filled a statin prescription in the 90 days before the index event; statin users increased to 4,152 within 90 days after index (50.83%). These post-index statin users were younger with more co-morbidities than nonusers (Table 1). Of the pre-index statin users, 3,274 continued treatment after the index event (84.69%); 878 of pre-index nonusers (20.41%) were initiated on statins after index. There were 763 events experienced by 686 patients (8.4%) within 1 year (Table 2). Of all events, 315 were terminal (41.3%); most deaths (284, 90.2%) occurred as the first MACE in follow-up.
Table 1Patient characteristics of statin users and non-users
The unadjusted effect of post-index statin use was associated with a 20% reduction in the risk of 1-year MACE (HR: 0.80, 95% CI: 0.69 to 0.93, p = 0.0043) and was similar after adjusting for demographics and co-morbidities (adjusted HR: 0.82, 95% CI: 0.70 to 0.95, p = 0.0074). We observed a time-dependent effect of statin use (interaction term between time and statin use: p <0.0001) and achieved the proportional hazards assumption by dichotomizing the follow-up time as ≤180 and >180 days. This unadjusted model showed a 29% risk reduction (HR: 0.71, 95% CI: 0.59 to 0.85, p = 0.0003) in the initial 180 days after the index for those who used a statin versus those who did not. However, it was not associated with the MACE outcome after 180 days (HR: 1.04, 95% CI: 0.80 to 1.35, p = 0.7663). Similarly, the adjusted model showed a 28% risk reduction (HR: 0.72, 95% CI: 0.60 to 0.86, p = 0.0004) in the initial 180 days after the index event for those who used a statin compared with those who did not (Figure 1). All-cause death was a substantial driver of the MACE risk difference, with 93 post-index statin users (2.24%) dying versus 191 nonusers (4.76%).
In the alternate analytic framework, which distinguished MACE as nonfatal (recurrent) and terminal, we found that taking a statin after index was associated with a (nonsignificant) 20% reduction in the number of nonfatal MACE recurrences (rate ratio = 0.80, 95% CI 0.46 to 1.39, p = 0.429) and a 67% reduction in the risk of all-cause death (HR 0.33, 95% CI 0.23 to 0.49, p <0.001). The effect of taking a statin was similar after adjusting for demographics and co-morbidities (Table 3). From this multivariable joint model, we found that taking a statin was associated with a (nonsignificant) 19% reduction in the number of nonfatal MACE recurrences (rate ratio = 0.81, 95% CI 0.49 to 1.32, p = 0.394) as well as a 65% reduction in the risk of all-cause death (HR 0.35, 95% CI 0.22 to 0.56, p <0.001). Being older than the median age of this cohort (i.e., >73.79 years) more than doubled the risk of all-cause death and being White was associated with more than a 2-fold increase in risk of recurrent (nonfatal) MACE. Figure 2 depicts the patients under observation and their MACE according to statin use.
Table 3Results from the joint marginal model for recurrent and terminal major adverse cardiovascular events
In this multicenter study of 8,168 patients, we found that approximately half of the cohort did not fill a statin prescription within 90 days after an ASCVD event. About 8% had ≥1 MACE within 1 year after their index ASCVD event, but post-index statin users had approximately 19% less risk of 1-year MACE. Statin protection was highest (28% risk reduction) in the first 180 days of hospitalization. Considering recurrent and terminal event processes jointly, the primary benefit of statin use was protection against early death.
Overall, statin use in this cohort was low, although the results were similar to those from a large database representative of patients with ASCVD in the US.
This exceeds the percent of statin users in our study, although it is likely a function of different underlying sociodemographic characteristics in VA samples, which tend to include fewer women and minorities (who are less likely to receive statin therapy).
found a 43% adherence rate for patients 2 years after admission for coronary heart disease.
Because the effect of post-index statin use decreased over time, we hypothesize that some patients who filled their initial prescription did not adhere to it long-term. A prior study showed that patients with ASCVD who were 50% adherent to statins had a 30% higher mortality risk than similar patients with a 90% adherence.
Such patients typically require more physician contact to find the best type and dose of statin. Patients in a VA study with more than the median number of annual primary care visits (i.e., >3) were more likely to adhere to a statin regimen.
found that several types of pharmacist-led interventions improved statin adherence, with the most successful type of intervention being between pharmacists and providers. In one such intervention, pharmacists contacted physicians of patients recently admitted for coronary heart disease, and the rates of statin use were 72% for patients in the intervention group versus 43% for patients in the control group at 2 years. Similarly, George and coworkers
suggest that collaboration between physicians and pharmacists is needed to produce patient-directed interventions including counseling, education, removing barriers to care, and medication reminders. Yao et al
even detected a better statin adherence rate for patients who were the same gender as their prescribing physician. Taken together, it is clear that combating the problem of nonadherence requires an investment of time, trust, and education from several parties.
Statins are not only indicated, but also confer the greatest benefit for secondary prevention.
who found low statin adherence, high statin discontinuation, and a high 2-year rate of ASCVD hospitalizations in a cohort of high risk patients. Another study considered primary and secondary prevention, finding that 20.9% and 43.0% of patients had at least 1 cardiovascular event in a 2 years follow-up, respectively. They also found widespread underuse of statins.
Randomized clinical trials (RCTs) identify an isolated treatment effect by eliminating extraneous variability (including behavior/adherence). Results from real-world data analyses, such as ours, are important because they represent a constellation of additional factors that occur in patients’ lives (unobserved in RCTs).
Hence, the results observed in this study may also reflect other health behaviors. Due to the nature of our study, patients could not receive a monetary incentive for adhering to a given study protocol, as occurs in RCTs.
The biggest limitation of this study is its retrospective design. Our data were limited to insurance claims; hence, if a patient filled a prescription without insurance, it was not recorded. Further, script fills are indirect measures of adherence. Similarly, we do not know the rate at which providers prescribed drugs or if patients received drug counseling. Due to the definition of the primary variable of interest, immortal time bias may be present.
Strengths of this study include the use of previously published ICD and Current Procedural Terminology code sets, its multicenter design, and robust modeling approach.
In conclusion, this study demonstrated a modest increase in statin treatment after an initial ASCVD event, with nearly half of the patients remaining undertreated. When jointly modeling nonfatal and fatal events, we observed that the primary benefit of statin use was protection against early death. Furthermore, because we observed that statin use may have the greatest impact in the first 6 months after an ASCVD event, it is crucial for patients to quickly adhere to therapy.
The authors have no conflicts of interest to declare.
This work was funded by a grant from Amgen Inc. to the Baylor Scott & White Research Institute.