Statin Use in Outpatients with Obstructive Coronary Artery Disease

Abstract

Background

Clinical trials have shown that statin therapy reduces cardiovascular morbidity and mortality in patients with coronary artery disease (CAD), even among patients with low-density lipoprotein (LDL) levels <100mg/dL. We sought to determine the extent to which patients with obstructive CAD in routine outpatient care are treated with statins, nonstatins, or no lipid lowering therapy.

Methods and Results

Within the American College of Cardiology's Practice Innovation and Clinical Excellence (PINNACLE) outpatient registry, we examined rates of treatment with statin and non-statin medications in 38,775 outpatients with obstructive CAD (history of myocardial infarction or coronary revascularization) and without documented contraindications to statin therapy. Among these patients, 30,160 (77.8%) were prescribed statins, 2042 (5.3%) were treated only with non-statin lipid lowering medications, and 6573 (17.0%) were untreated. Lack of medical insurance was associated with no statin treatment, while male sex, coexisting hypertension, and a recent coronary revascularization were associated with statin treatment. Among those not on any lipid lowering therapy, LDL levels were available for 51.2% (3365/6573). Among these untreated patients, LDL levels were <100 mg/dL in 1794 (53.3%) patients and ≥100 mg/dL in 1571 (46.7%) patients.

Conclusion

Despite robust clinical trial evidence, a substantial number of patients with obstructive CAD remain untreated with statins. A small proportion were treated with non-statin therapy, while 1 in 6 patients were simply untreated, of which half had LDL values <100mg/dl. These findings illustrate important opportunities to improve lipid management in outpatients with obstructive CAD.

Although randomized clinical trials^1-5 have consistently demonstrated that secondary prevention treatment with statins substantially reduces recurrent cardiovascular events and mortality in patients with obstructive coronary artery disease (CAD), the extent to which these high-risk patients are treated with statin, non-statin, or no lipid lowering therapy in the outpatient cardiology setting is largely unknown. This is because prior studies have primarily examined rates of statin treatment for dyslipidemia or non-specific diagnoses of CAD in population-based or health insurance databases^6-10. However, statin treatment rates among outpatients with documented obstructive CAD (i.e., patients with prior myocardial infarction or coronary revascularization who are at high-risk of recurrent events) may be different.

Moreover, among untreated patients, it is unclear if patients are untreated because they have normal-range low-density lipoprotein (LDL) levels. Although current guidelines recommend that high-risk patients be treated to LDL levels <100 mg/dL, implicit within these guidelines are that patients are “treated” with a lipid lowering agent (preferentially statins).¹¹ In addition, despite the guideline recommendations of LDL targets, clinical trials have found that the benefits of statin therapy are consistent across all subgroups, including those with LDL levels <100 mg/dL.^{1, 12, 13} Therefore, non-treatment with statins because of a “normal-range” LDL level in patients at high-risk for recurrent cardiovascular events may represent a gap in the quality of CAD care.

Recently, the American College of Cardiology's National Cardiovascular Disease Registry (NCDR) initiated the Practice INNovation And CLinical Excellence (PINNACLE Registry™ ) program, the first prospective registry of cardiovascular outpatient care in the U.S. The emergence of this outpatient registry provides a unique opportunity to better understand lipid therapy in outpatients with CAD. Specifically, it affords the opportunity to quantify the proportion of patients with known obstructive CAD who are not treated with statins or any lipid lowering medications. It can also describe the prevalence of under-treatment with statins among those with LDL levels of <100 and ≥100.

METHODS

Study Population

The NCDR PINNACLE Registry was launched in 2008 by the American College of Cardiology and represents the first, national, prospective, office-based, quality improvement registry of cardiac patients in the U.S.^{14, 15} Within participating practices, longitudinal patient data were collected at the point of care and included patients’ symptoms, vital signs, medications, laboratory values, and recent hospitalizations using a standard data collection tool with written definitions, uniform data entry and transmission requirements, and data quality checks. For the purposes of this study, we assessed clinical data from 179,608 patients enrolled from 30 practices between July 1, 2009, and June 30, 2010. We included only those adult patients with documented obstructive CAD (n=39,601), which was defined as a prior myocardial infarction, percutaneous coronary intervention or coronary artery bypass graft surgery. If a patient had more than one clinic visit during the study period, demographic and comorbidity data from only the index encounter were used to minimize overrepresentation by patients with multiple visits. For patients with multiple visits during the study period, they were considered to be on lipid-lowering medication if this was indicated on any of the visits over the study period. Of the 30 practices enrolled in PINNACLE, we excluded 5 practices with fewer than 10 CAD patients and 1 practice with inconsistent data entry (total patients excluded=311). We also excluded 515 patients with documented contraindications to statins (determined by the treating physician as a medical, patient, or system reason for exclusion). Our final study sample thus included 38,775 patients from 24 practices representing 111 practice locations in 18 states across the U.S. (Figure 1).

Figure 1. Flowchart of PINNACLE patients included in the analyses.

Abbreviations: CAD=coronary artery disease; LDL=low density lipoprotein

Study Outcomes and Variables

There were 2 principal outcomes for this study. First, we examined the proportion of patients with obstructive CAD who were treated with a statin, a non-statin, or no lipid lowering therapy. This was determined by whether a patient was documented to be on a statin or non-statin agent, and these rates include treatment decisions made during the office encounter. Second, among patients on no lipid lowering therapy, we further examined the distribution of LDL levels in this population and quantified the proportion of patients with an LDL level <100 and ≥100.

Within PINNACLE, we also collected patient information on age (mean and categorized as <60, 60 to 79, and ≥80), sex, race (white vs. other), health insurance (private [fee-for-service or health-maintenance-organization], public [Medicare, Medicaid, Veterans Administration], or none), prior myocardial infarction, percutaneous coronary intervention or coronary artery bypass graft surgery, heart failure, diabetes mellitus, hypertension, dyslipidemia, peripheral arterial disease, prior stroke, tobacco use (current, former, never), body mass index (categorized as <30, 30-34.9, and ≥35 kg/m²), and atrial fibrillation. As PINNACLE is a practice improvement registry that captures data from routine clinical care in outpatient cardiology practices, all comorbidity variables in PINNACLE were physician-reported as part of the medical record.

Statistical Analyses

Demographic and clinical characteristics were compared between those treated and not treated with statins using chi square tests for categorical variables and t-tests for continuous variables. To determine independent predictors for statin non-treatment, hierarchical log Poisson regression models, with a random effect for site to adjust for clustering by site, were constructed. Patient-level covariates were entered as fixed effects in the multivariable model and included the following: age, sex, body mass index, diabetes mellitus, hypertension, prior myocardial infarction, recent percutaneous coronary intervention or coronary artery bypass graft surgery within the past 12 months, current tobacco use, and insurance status (none vs. public vs. private). Race was not included in the models due to a high (51.4%) missing rate. Because the rate of statin non-treatment exceeded 10%, we estimated a relative rate (RR) directly by using log-Poisson or modified Poisson regression models at all steps (as opposed to logistic regression) to avoid overestimation of estimates of effect.^{16, 17} Non-linearity between the covariates and outcomes was tested, and cubic spline terms were added to the model as appropriate, which was required for age (eTable 1, eFigure 1) and body mass index (eTable 2, eFigure 2).

To better characterize patients who were not treated with any lipid lowering therapy, we examined the distribution of LDL levels in this group. Furthermore, the proportion of untreated patients with an LDL level of <100 and ≥100 mg/dL was determined.

Besides race, data were missing for 17% of patients (7% were missing 1 value, 5% were missing 2 values, and 4% were missing 3 or more values; the highest missing rate for any single variable was body mass index at 10%). Missing covariate data were imputed with a single imputation dataset using IVEWare.¹⁸ For each analysis, the null hypothesis was evaluated at a 2-sided significance level of 0.05 with 95% confidence intervals (CIs) calculated. All analyses were performed with SAS version 9.2 (SAS Institute, Cary, North Carolina), R version 2.7.0 (Foundation for Statistical Computing, Vienna, Austria), and IVEWare (Institute for Social Research, University of Michigan).

RESULTS

Of 38,775 patients with obstructive CAD, 30,160 (77.8%) were treated with statins, 2042 (5.3%) were treated with only non-statin lipid lowering medications, 6441 (16.6%) were treated with both statin and non-statin lipid lowering medications, and 6573 (17.0%) were untreated (Figure 2). Patients not treated with statins (n=8615; 22.2%) were more likely to be younger, female, and have a history of prior myocardial infarction, prior coronary artery bypass grafting, and atrial fibrillation. Patients not treated with statins were also less likely to have prior percutaneous coronary interventions and to have coexisting heart failure, diabetes mellitus, hypertension, and peripheral artery disease (Table 1; p<0.001 for all).

Figure 2. Flowchart of lipid-lowering therapy in patients with obstructive CAD in the PINNACLE registry.

Abbreviations: CAD=coronary artery disease; LDL=low density lipoprotein

Table 1.

Baseline Characteristics of Patients Without Documented Exclusions to Statins

	On Statins n=30160	Not on Statins n=8615	p-value
Age, years, mean ± SD	68.0 ± 11.6	67.0 ± 14.5	<0.001
< 60	24.8%	30.1%
60 to 79	58.5%	50.0%
≥ 80	16.7%	20.0%
Male sex	67.5%	54.8%	<0.001
Race
White	89.0%	87.9%	0.05
Non-white	11.0%	12.1%
Missing	50.8%	55.5%
Health Insurance			0.08
Private	58.6%	59.9%
Public	35.0%	34.0%
None	6.4%	6.0%
Prior myocardial infarction	36.6%	39.9%	<0.001
Prior PCI within 12 months	32.8%	25.3%	<0.001
Prior CABG within 12 months	28.1%	31.7%	<0.001
Heart failure	18.1%	16.3%	<0.001
Diabetes mellitus	27.4%	20.5%	<0.001
Hypertension	75.6%	66.6%	<0.001
Dyslipidemia	80.7%	53.8%	<0.001
Peripheral arterial disease	8.1%	5.9%	<0.001
Prior stroke	5.7%	6.1%	0.16
Tobacco use
Current	13.6%	13.0%	<0.001
Former	45.7%	38.7%
Never	40.7%	48.3%
Body mass index, kg/m2, mean ± SD	29.7 ± 6.0	29.2 ± 6.6	<0.001
< 30	58.5%	61.0%
30 to 34.9	24.9%	22.2%
≥ 35	16.6%	16.8%
Atrial fibrillation	15.4%	19.8%	<0.001

All data are presented as % unless otherwise specified

PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting

In multivariable hierarchical modified Poisson regression models, adjusted for clustering of patients within sites, lack of medical insurance was associated with a lower likelihood of being treated with a statin (adjusted RR 0.93, 95% CI 0.88-0.99; p=0.03), while male sex (adjusted RR 1.10, 95% CI, 1.07-1.13; p<0.001), coexisting hypertension (adjusted RR 1.07, 95% CI 1.02-1.11, p=0.002), prior CABG (adjusted RR 1.09, 95% CI 1.05-1.13, p<0.001), and prior PCI (adjusted RR 1.11, 95% CI 1.06-1.16, p<0.001) were associated with a higher likelihood of being treated with a statin (Figure 3). Coexisting diabetes mellitus, current or former tobacco use, and a history of prior MI were weakly associated with a higher likelihood of statin therapy. Age and BMI were non-linearly associated with statin treatment, with patients aged 60-80 and BMI 30-35 having the highest rates of statin treatment (Appendix 1).

Figure 3. Patient factors associated with statin treatment in a multivariable model.

Age and BMI were included as nonlinear terms in the model using cubic splines (Appendix 1). Abbreviations: CABG=coronary artery bypass graft surgery; PCI=percutaneous coronary intervention; RR=rate ratio; CI=confidence interval

Among patients who were not on any lipid lowering therapy (n=6573; 17.0%), LDL levels were available for 3365 (51.2%) patients. Of those with LDL results, 1794 (53.3%) had LDL levels of <100 mg/dL and 644 (19.1%) had LDL levels <70 mg/dL. There were an additional 1571 (46.7%) patients who were not treated with any lipid lowering medications who had LDL levels of ≥100 mg/dL (Figure 4).

Figure 4. Distribution of LDL levels among patients with obstructive coronary artery disease untreated with any lipid lowering medications.

Abbreviations: LDL=low density lipoprotein

DISCUSSION

In this large, national, contemporary registry of cardiac outpatients, we found that more than 1 in 5 patients with obstructive CAD and without documented contraindications were not treated with statins, a class of medications with demonstrated efficacy in secondary prevention, while a small proportion were treated only with non-statin lipid lowering medications. Importantly, 17% of patients were not treated with any lipid lowering therapy at all. Notably, half of these untreated high-risk patients had LDL levels <100 mg/dL. Collectively, these findings suggest there are significant opportunities for improvement in the quality of lipid management for a substantial number of outpatients with obstructive CAD.

Our report extends the work of prior studies, which have reported statin treatment rates of 60% to 70% in outpatients with CAD.^{6-10, 19} Among 290 patients with CAD in a Maryland health maintenance organization, 69% were on lipid-lowering therapy.⁷ An evaluation of 108 patients with CAD in the 2003-04 National Health and Nutrition Examination Survey found that 63% were treated with statins.¹⁰ Similarly, among 1960 post-MI patients in a multicenter U.S. registry from 2003-04, approximately 70% of patients were on statins at 12-months after their MI.¹⁹ These studies, however, were limited by smaller study samples and the specificity of the diagnosis of CAD, and were all conducted before 2005, prior to the publication of current guidelines and recent studies.^{4, 11, 12} Our study extends these findings in a large, contemporary sample of patients with obstructive CAD. Additionally, we were able to examine the LDL profiles of patients not treated with any lipid lowering therapy and rates of treatment with other lipid lowering agents among those not treated with statins.

Our finding that half of untreated patients had an LDL of <100 mg/dL raises concerns about missed opportunities for intensive secondary prevention in patients with documented obstructive CAD. While appearing to have reached established LDL treatment targets for secondary prevention, these patients were not receiving any active lipid lowering treatment. Although we can only speculate as to the actual reasons for non-treatment of these patients, our data suggest that some physicians may be making decisions about lipid management for these patients based on baseline LDL levels alone. However, prior analyses from clinical trials have reported that there is a constant relative risk reduction for recurrent cardiac events of 20% with statins across the spectrum of baseline LDL levels.^{1, 12, 13} Others have posited that the only definitive inference from secondary prevention clinical trials of statin therapy is that empirical statin therapy in high-risk patients reduces cardiovascular events, independent of LDL levels. These investigators contend that, because these studies randomized patients to fixed dose statin therapy regardless of LDL level, a treat-to-target strategy has yet to be studied in a clinical trial.^20-22 Therefore, a strategy of initiating lipid lowering therapy (preferably statins) in patients with known CAD only when they have LDL levels ≥100 mg/dL may reflect a misapplication of the current evidence base and clinical guidelines. Further study is needed, however, to better understand the reasons for non-treatment of these patients at high risk of a recurrent cardiovascular event, including patients’ baseline LDL levels.

Finally, we identified many patients with CAD and elevated LDL levels who were untreated with any lipid lowering medications. The reasons as to why these patients remain untreated despite poor lipid control were not documented. If this practice were to be pervasive, it would mean that a significant number of patients are not even considered for lipid lowering therapy for their CAD.

Our study findings should be interpreted in light of the following limitations. First, the PINNACLE program enrolled patients from highly motivated practices dedicated to quality improvement. Rates of statin use among patients with CAD in other U.S. practices are likely to be lower than those reported in this study. Second, although we have data regarding patients’ LDL levels, we are unable to establish the underlying reasons as to why patients were not treated with statins. While a low untreated LDL level may be contributing to the decision not to treat these patients, further study is required to test this hypothesis. Third, the PINNACLE program allows clinicians to designate exclusions for statin therapy without adjudication, such that they are removed from the denominator of eligible patients. The number of patients with documented exclusions (n=515; 1.3%) was lower than the discontinuation rates that had been observed in clinical trials (ranging from 4%-33%^1-5) which may indicate that exclusions were under-documented in PINNACLE or possibly improved tolerance when the choice and dosing of statins is allowed to vary according to patient preference (i.e., a patient may have myalgias with one particular type or dose of statin but would tolerate a lower dose of the statin or a difference type). Regardless, if under-documentation of contraindications to statin therapy occurred, many patients still remained untreated with other non-statin lipid lowering agents. Fourth, LDL levels were not available for half of the patients who were not treated with lipid lowering medications. Nonetheless, we have no reason to believe that LDL data were differentially missing by LDL group (e.g, <100 mg/dL vs. ≥100 mg/dL). Fifth, we did not have specific information regarding statin type, statin dose, or type of non-statin medication used, which limited our ability to provide additional insights on how these patients were treated. Furthermore, we did not have information regarding prescription drug coverage or actual adherence to the selected lipid lowering strategy. Nonetheless, our study was interested in assessing physician treatment of patients, and therefore actual patient adherence, while of interest, would not affect our findings of physician under-treatment of patients with CAD at high-risk of recurrent cardiovascular events. Finally, while this study was not designed to examine the association of statin therapy with specific clinical outcomes, such as readmission and mortality, treatment with statins is a recognized performance measure in patients with obstructive CAD regardless of patient characteristics,²³ based upon the wealth of clinical trial and observational data supporting the use of statins in secondary prevention.

Conclusion

In a large contemporary practice improvement registry of outpatients with obstructive CAD, 78% of patients were receiving guideline-based treatment with statins, 5% were treated solely with non-statin lipid lowering medications, and 17% of patients were untreated. Among untreated patients, half had LDL levels <100 mg/dL while the other half were untreated despite LDL levels >100 mg/dL. These findings highlight important opportunities to improve the use of statin therapy in outpatients with obstructive CAD who are at high-risk for recurrent cardiovascular events.