Sex Differences in the Use of Statins in Community Practice

Introduction

See Editorial by Khazanie et al

Statin therapy reduces cardiovascular risk in both women and men.^1–4 Nevertheless, sex differences in statin treatment and adherence to guideline-recommended lipid management are well documented.^5–9 The reasons underlying lower statin utilization in women remain poorly understood, and the degree to which these sex differences may be attributable to differences in patient perceptions, side effects, and differences in physician prescribing patterns merits further investigation.

The Patient and Provider Assessment of Lipid Management (PALM) registry is a US registry of 7938 individuals with atherosclerotic cardiovascular disease (ASCVD) or at high risk for ASCVD treated at 140 cardiology, primary care, and endocrinology outpatient practices. In addition to clinical, socioeconomic, and lipid data for all enrolled patients, the PALM Registry uniquely captured patient-reported side effects, patient perceptions and beliefs, and provider characteristics—all of which influence treatment use and adherence.¹⁰ In this study, we (1) evaluated whether statin treatment differed between women and men; (2) assessed potential reasons underlying sex differences in statin treatment including side effects, beliefs, and provider characteristics; and (3) identified whether sex differences in guideline-recommended statin treatment persisted after adjustment for demographics, clinical characteristics, socioeconomic status, education, patient beliefs, and provider characteristics.

Methods

Data Description—The PALM Registry

We set out to investigate sex differences in statin treatment using the PALM Registry—a nationwide registry of individuals (n=7938) with ASCVD or at high risk for ASCVD enrolled between May 2015 and November 2015 in 140 US community practices. The design and rationale for the PALM Registry have been described previously.¹⁰ Briefly, data regarding sex, socioeconomic status, education level, patient beliefs and perceptions, and statin use were acquired via patient surveys administered on an iPad in PALM providers’ offices. Providers enrolling patients into the PALM Registry also completed surveys before beginning enrollment. Study coordinators at each site abstracted patient clinical data including demographics, medical history, and current statin dosing intensity. Lipid levels were measured in all patients at a core laboratory. Patients who met a guideline recommendation for statin therapy based on the 2013 American College of Cardiology (ACC)/American Heart Association (AHA) Cholesterol Guideline¹ and who completed the survey at enrollment (95.3% response rate) were included (n=5693). The data, analytic methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure. All study sites obtained institutional review board approval for participation in the PALM Registry, and all study participants provided informed consent.

Data Definitions

ASCVD was defined as having a history of coronary artery disease (including prior myocardial infarction, coronary artery disease, coronary artery bypass grafting, or percutaneous coronary intervention), cerebrovascular disease (including prior stroke and transient ischemic attack), and peripheral artery disease (including prior abdominal aortic aneurysm, peripheral arterial disease, carotid artery stenosis, and noncoronary arterial revascularization). Patients were recommended for high-intensity statin therapy if they met ≥1 of the following criteria: (1) history of ASCVD and aged ≤75 years old; (2) LDL-C (low-density lipoprotein cholesterol) ≥190 mg/dL; or (3) adults aged 40 to 75 years with diabetes mellitus and 10-year ASCVD risk ≥7.5% based on the pooled cohort equation with LDL-C ≥70 mg/dL (if not on a statin) or irrespective of LDL-C if already on a statin at the time of enrollment.¹ Patients not meeting criteria for a high-intensity statin qualified for a moderate-intensity statin if they met ≥1 of the following: (1) history of ASCVD and >75 years old; (2) adults aged 40 to 75 years without diabetes mellitus with ASCVD risk ≥7.5% and LDL-C >70 mg/dL or already on a statin; or (3) adults aged 40 to 75 years with diabetes mellitus and 10-year ASCVD risk <7.5% with LDL-C ≥70 mg/dL or already on a statin.

Statistical Analysis

We assessed sex differences in statin treatment patterns including statin use and high-intensity statin use in the overall population, as well as within primary and secondary prevention subgroups. Statin intensity was defined according to definitions from the 2013 ACC/AHA Cholesterol Guideline.¹ We described sex differences in baseline characteristics, lipid levels, provider characteristics, patient beliefs about statins and cardiovascular disease, patient-reported side effects, and willingness to try statin therapy. We compared women and men by statin treatment overall and guideline-recommended statin intensity, as well as frequency of not being on a statin because of: (1) prior discontinuation, (2) patient preference (ie, patient declined statin therapy), or (3) never being offered a statin. Categorical variables were presented as frequencies and continuous variables presented as medians with interquartile range. We assessed differences by sex in categorical variables using the Pearson χ² test and assessed differences by sex in continuous variables using the Wilcoxon rank-sum test. A 2-sided P<0.05 was considered statistically significant.

We evaluated the relationship between sex and statin use (both overall and at guideline-recommended intensity) using logistic regression analysis. We adjusted for the following potential confounders: demographics including age and race; clinical, socioeconomic, and educational factors including prior coronary artery disease, cerebrovascular disease and peripheral arterial disease, diabetes mellitus, obesity, smoking, hypertension, heart failure, yearly income, insurance status, education level, patient numeracy, pertinent patient beliefs, and perceptions including worry about heart disease, physician trust, patient’s statin beliefs about safety, effectiveness, and the link between high cholesterol and heart attack risk; and provider factors including cardiologist versus noncardiologist, use of 2013 ACC/AHA Guideline, urban versus rural setting, and provider time in practice. We measured patient numeracy using the subjective numeracy score—a previously validated instrument.¹¹ The generalized estimating equation was used to account for clustering of patients within sites. We also evaluated the unadjusted and adjusted odds ratio (95% CI) of guideline-recommended statin treatment to examine the sex difference in outcomes in the following subgroups: (1) patients with a primary prevention indication, (2) patients with a secondary prevention indication, (3) patients <75 or ≥75 years old, (4) education college or above, (5) income >$100 000/year, (6) income <$35 000/year, (7) patients treated by cardiologists, and (8) patients treated by a provider who reports using the ACC/AHA Guideline. Income was self-reported, and missing income data were imputed using 2014 median household income based on patient residence zip code or enrolling site zip code. Multiple imputation was used for all other variables with fully conditional specification approach to impute missing values for covariates with missing data (maximum missing, 11.7%).¹² Five imputed datasets were used for multivariable analysis. All analyses were performed using SAS, version 9.4 (SAS Institute, Inc, Cary, NC).

Results

Baseline Characteristics

Among 5693 patients with a 2013 ACC/AHA Cholesterol Guideline indication for statin treatment, 2460 (43%) were women. Baseline characteristics differed significantly between female and male patients (Table 1). Compared with men, women were more frequently black, had higher body mass index, were less likely to have a history of ASCVD, had lower 10-year risk among those without ASCVD, had lower tobacco use, and had higher lipid levels than male patients. Women also more frequently had Medicare and Medicaid with less private insurance, a lower yearly income, and were less likely to be seen by a cardiologist than men (51.9% versus 65.2%; P<0.001).

Statin Use and Dosing Intensity

Compared with men, women were significantly less likely to be on a statin (67.0% versus 78.4%; P<0.001) or, if on a statin, were less likely to be on guideline-recommended statin intensity (36.7% versus 45.2%; P<0.001; Table 2; Figure 1). Women were more likely to report never having been offered a statin (18.6% versus 13.5%; P<0.001), previously discontinuing a statin (10.9% versus 6.1%; P<0.001), or previously declining a statin (3.6% versus 2.0%; P<0.001; Figure 1). Trends were similar among patients eligible for statin therapy based on a primary or secondary prevention indication. Sex differences in statin use were identified in 3 of the 4 primary statin treatment groups identified in the guideline: individuals with diabetes mellitus, those with 10-year ASCVD risk ≥7.5%, and those with ASCVD (Table I in the Data Supplement). We did not observe significant sex differences in statin use in the limited sample of individuals with LDL-C ≥190 mg/dL. When considering only untreated patients (33.0% of women, 21.6% of men), the relative distribution of reasons for not being treated were as follows: untreated women reported never being offered therapy less frequently than untreated men (56.2% versus 62.6%; P=0.01), with similar rates of declining statins (10.9% versus 9.2%, P=0.25) and discontinuing statins (32.8% versus 28.2%; P=0.05).

• Download figure
• Download PowerPoint

Patient Perceptions and Beliefs

Women and men differed in terms of their beliefs and perceptions surrounding statins and cardiovascular disease (Table 3). Women more frequently stated that they either occasionally or often worry about heart attack or stroke (45.7% versus 34.4%; P<0.001), yet were less likely to believe that people with high cholesterol are more likely to have a heart attack (75.4% versus 82.1%; P<0.001). Women were less likely than men to agree with the statements that statins are effective (68.0% versus 73.2%; P<0.001) and statins are safe (47.9% versus 55.2%; P<0.001). Women were also more likely to report believing that statins can cause diabetes mellitus, muscle symptoms, and liver damage (Table 3). Physician trust was similar between women and men (65.9% versus 64.5%; P=0.29).

Patient-Reported Symptoms and Willingness to Reattempt Statin

Many women (50.1%) and men (43.1%) currently using statins reported some type of adverse symptoms associated with statin use. Relatively few (5.3%) reported stopping their statin because of side effects, but women were more likely to discontinue statin therapy than men (7.9% versus 3.6%; P<0.001; Table 4). Among former statin users, women most frequently listed side effects as a reason for stopping (150 of 267 former users, 56.2%). Among current statin users, those on high-intensity statins, and former statin users, women were more likely than men to report previously experiencing a side effect when on statins (Table 4). Compared with men, women previously on a statin were less willing to try another statin.

Statin Use After Multivariable Adjustment

After adjustment for relevant demographic-, clinical-, socioeconomic-, belief-, and provider-related confounders, women remained less likely to receive any statin (odds ratio, 0.70; 95% CI, 0.61–0.81; P<0.001) or a guideline-recommended statin (odds ratio, 0.82; 95% CI, 0.73–0.92; P<0.001; Figure 2). Adjusted odds ratios for all variables included in the final models are provided in Tables II and III in the Data Supplement. When evaluated within key subgroups, including primary prevention, secondary prevention, age <75 and ≥75 years, patients with college education or above, income ≥$100 000/year, income <$35 000/year, those treated by cardiologists, and those treated by providers following the ACC/AHA guideline, women remained less likely to be treated with a statin than men in all subgroups (Figure 2). Women were also less likely to receive guideline-recommended statin dosing compared with men in all subgroups on unadjusted analysis, although these trends were no longer statistically significant after adjustment in the primary prevention and high-income (≥$100 000) subgroups (Figure 2).

• Download figure
• Download PowerPoint

Discussion

In a large sample of US adults seen in contemporary community practice, we found that women were less likely than men to receive guideline-recommended statin therapy. There appear to be several reasons for these sex-related differences in statin use: women were less likely to report having been offered statin therapy, more likely to decline statin therapy when offered, and more likely to discontinue statin therapy after starting. These sex differences in statin treatment in PALM were consistent by a number of different subgroups that were analyzed, including indication (primary versus secondary prevention), education, income, and provider type.

Prior studies have demonstrated similar sex-related differences in statin treatment across a variety of study populations.^{5–7,13–15} However, these prior studies had not investigated the underlying causes for these care differences. We found that the biggest contributor to sex differences in statin use was the difference in the proportion of patients offered a statin by their physician. Among those recommended for statins, 18.6% of women reported that they had never been offered a statin by their physician versus 13.5% of men; clinical differences did not explain this gap. Although female patients had less ASCVD and slightly lower 10-year calculated ASCVD risk scores, all of those included in the analysis met guideline indications for statin therapy.¹ Furthermore, sex differences in statin utilization persisted even after adjusting for clinical characteristics and when we reanalyzed our results by those with ASCVD versus not.

Some have suggested that differences in statin use may be explained by insurance status or other socioeconomic factors.¹³ In the PALM Registry, we found that female patients were less likely to be privately insured, had lower annual household incomes, and were less likely to be seen by a cardiologist than male patients. However, even after adjusting for sociodemographic characteristics such as age, insurance status, education, numeracy, and income, female patients remained significantly less likely to receive statin treatment and ACC/AHA Guideline–recommended statin intensity.

The PALM Registry was also unique in its ability to examine the degree to which patient beliefs may have contributed to sex differences in statin utilization. Our findings are consistent with what was found in the USAGE (Understanding Statin Use in America and Gaps in Patient Education) internet-based survey, which found in 2011 that female patients were more likely to report stopping or switching their statin than male patients, frequently because of muscle-related complaints.¹⁵ This work extends the finding of USAGE as it included chart reviews for clinical data, core lab lipid panels, and included patients seen in routine clinical practice. In PALM, we found that women were more likely than men to decline statin therapy. While a small sample size and significant missingness limited our ability to consider patient-reported reasons for declining statin therapy by sex, a recent analysis of the overall PALM population demonstrated that patients declining statin therapy frequently cited worries about side effects (36.8%), preference for diet/exercise (25.0%), and a preference for natural remedies (16.0%) among reasons for declining statins.¹⁶ When surveyed, we also found that women were less likely to believe statins were safe or effective compared with men. Similarly, women were more likely to experience perceived statin-associated side effects and discontinue therapy because of side effects than men. Male patients, both among those patients previously on a statin and among statin naive patients, were more willing to try a statin. Nonetheless, differences in rates of adults declining statins or discontinuing statins explained only about half of the 11% absolute difference in statin utilization.

Importantly, we examined reasons for lack of statin utilization among all potentially eligible patients, not just those who were untreated. Had we only evaluated patients who were not on a statin, sex-based differences would have appeared artificially attenuated. Among those not on statins, women were less likely to report never being offered a statin than men, with no difference by sex in the relative proportions of untreated men and women who discontinued or declined a statin. However, at a population level, among those eligible for statins, women were more likely than men to report all 3 reasons for nontreatment (never being offered, discontinued, and declining a statin), contributing to large population-level differences in overall statin use.

Because it appears that both provider and patient factors contribute to sex-related differences in statin use, addressing these differences will likely require a multipronged approach. On the provider side, national performance metric systems have been demonstrated to improve the use of evidence-based medications in several conditions. While we found greater statin underuse in women than in men, care gaps existed for both. Therefore, we believe that overall quality improvement efforts could bridge these gaps, regardless of patient sex. In addition to provider-directed efforts, our study importantly identified differences in patient beliefs that need to be addressed. For example, women had less belief than men in the safety and effectiveness of statins, which likely contributed to higher rates of patient refusal and discontinuation in women. Consequently, focused campaigns that emphasize the risk of ASCVD in women, as well as the need for effective prevention, could be helpful. The AHA Go Red for Women campaign represents one past highly successful effort, focusing on cardiovascular disease prevention in female patients.^17–19 While the success of this initiative reassures us that progress is possible, the sex differences in statin treatment observed in our study emphasize that more work remains to be done.

Our study had some limitations. First, the PALM Registry was observational in nature; therefore, we cannot determine direct causality based on the associations detected. Second, our study did not capture provider reasoning for statin prescribing (or nonprescribing) in specific patients. Finally, the history of being offered a statin and the reasons for statin discontinuation were based on patient report and, as a result, could be limited by recall bias. This being said, there are no reasons to believe recall bias should differentially affect women versus men.

Conclusions

We observed that sex differences in statin treatment continue to persist in contemporary practice. Women were less likely to receive statins or guideline-recommended statin intensity, even after correcting for demographic, socioeconomic, and clinical confounders. These care differences appear to be attributable to women reporting being offered a statin by their physician less commonly and women reporting refusing or discontinuing their statin more frequently. Sex-related differences in patient beliefs about statins and cholesterol may have further contributed to these care differences. Patients and providers alike must be educated on the safety and efficacy of statin therapy to optimize therapeutic efforts.

Acknowledgments

We thank Erin Campbell, MS, for her editorial contributions to this manuscript. Erin Campbell did not receive compensation for her contributions, apart from her employment at the institution where this study was conducted.

Footnotes