Cardiovascular Risk and Statin Use in the United States

The overall purpose of this article was to examine statin use in multiple contexts, both in patients with a a diagnosis of hyperlipidemia and those with other cardiovascular-risk states. Previous articles have shown the benefit of statins in reducing cardiovascular events in populations exhibiting other cardiovascular risk factors and not just in the context of hyperlipidemia. This article could possibly help change the guidelines for statin use and broaden the population it is prescribed to, in connection with the recent American Heart Association (AHA) and American College of Cardiology (ACC) change in recommendations of when to prescribe statins to decrease overall cardiovascular risk.

This study used a cross-sectional analysis of data from a 2010 Medical Expenditure Panel Survey (MEPS) study. MEPS uses a complex study design, with over-sampling of minority groups so that end result is representative of the population as a whole. Exclusion criteria included people with contraindications to statin therapy. Independent variables measured were cardiac risk factors (HTN, diabetes, etc.) and demographics (race, age, poverty). One of the issues identified with the methodology is that the study the assumption that participant's blood pressure was 120/80 without a specific diagnosis of HTN. This assumption may have underestimated actual blood pressure. Another problem with the methodology is that diagnoses were self-reported.

The sample was stratified into high and low cardiovascular risk groups. Two sets of logistical regression models were run, one with the whole sample (all-eligible analysis) and one with the sample restricted to those without a diagnosis of diabetes or heart disease. The dependent variable was statin use and the independent variables included age, sex, race, insurance status, poverty, tobacco use, hyperlipidemia, hypertension (HTN), peripheral arterial disease and history of a cerebrovascular accident. Analysis of the demographic data shows that the study sample was representative of the population. Results noted that the age variation is wide. In the overall sample 38.4% had a diagnosis of hyperlipidemia and 46.9% of these individuals were using a stain. Individuals with a diagnosis of coronary artery disease (CAD) with no hyperlipidemia, only 15.5% were on a statin. Patients had a high likelihood of being on a statin if they had a diagnosis of hyperlipidemia or diabetes mellitus (DM), but not if other factors were present without these conditions. It seems that more weight placed by physician on these diagnoses over others to place patient on statin. Cerebrovascular disease and peripheral arterial disease are considered to be coronary artery disease equivalents, but they are not treated as such by physicians when prescribing statins.

In the overall all-eligible model, when stratified by a diagnosis of hyperlipidemia, age, CAD, DM and HTN were significant predictors of statin use. Sex was not a significant predictors of whether a statin was prescribed or not. Individuals with CAD without hyperlipidemia had the lowest probability of being on a statin (0.11) as compared to 0.32 for individuals with only a hyperlipidemia diagnosis and 0.44 for individuals with both hyperlipidemia and CAD. When holding all other variables constant, sex and age are significant, meaning that older patients and men more likely to be on statin with all other variables equal. In Figure 3, we see statin use by Framingham risk index, which demonstrates that clinicians are not taking Framingham risk score into account independent of a diagnosis of hyperlipidemia when prescribing statins.

In the Discussion and Conclusions, we noted that this study was not designed to assess reasons for these disparities in prescribing statins to patients. The AHA-ACC guidelines changed recently, so many physicians have not changed their practice accordingly as of yet. Physicians can be slow to update their practice. We need to be addressing high cardiovascular risk group overall when making decisions whether or not to prescribe a statin, seeing as we are not doing a good job treating cardiovascular risk factors other than hyperlipidemia. We are not doing a good job of secondary prevention and potential long-term benefits.

Limitations of this study included self-report data, physician compliance with new guidelines, and the fact that this article did not talk about side-effects and contraindications to statin use. This article did not assess if patients were on different statins and had adverse reactions to them and what the physician decided to do. We made note that rosuvastatin and atorvastatin have lower risk of myositis compared to other statins, so patients can have a trial on a low-dose if they had bad reaction previously.

Patients are also not aware of new guidelines and can be resistant to new medication that they perceive is not for them because they might not have hyperlipidemia. Education is important that mortality benefit is independent of lipid levels. No marketing push for other uses besides hyperlipidemia.

Another limitation of this study is that it is a retrospective study, so the authors must use odds ratios and not relative risk to analyze the results. A benefit of this study design is in the large sample size obtained, which can only be done in retrospective study. Prospective studies are not always feasible since they are resource intensive in terms of time and money.

Overall, we felt that perhaps this study can be done in another 5 years to assess how new guidelines affect practice. Another future question to assess is why does this problem exist? Perhaps another study can use hospital-based medical records to obtain data on patient diagnoses, instead of relying on self-report. The practical implication of this article is important and applicable to our future practice. This article gives baseline data to refer to later if a future study looks at this same subject in light of the changing AHA-ACC guidelines.

Competing interests: None declared