Combination Therapy With Ezetimibe/Simvastatin Versus Statin ...

Clinical Therapeutics/Volume 33, Number 2, 2011

Combination Therapy With Ezetimibe/Simvastatin Versus Statin Monotherapy for Low-Density Lipoprotein Cholesterol Reduction and Goal Attainment in a Real-World Clinical Setting Howard S. Friedman, PhD, MMS1; Srinivasan Rajagopalan, PhD2; Jaime P. Barnes, SM3; and Hal Roseman, MD, MPH, FACC, FACP4 1

Analytic Solutions, LLC, New York, New York; 2Med Data Analytics, Inc, Williamsville, New York; 3 Schering-Plough Corporation, Kenilworth, New Jersey; and 4Cardiology Wellness Center, Nashville, Tennessee ABSTRACT Background: Randomized studies have demonstrated improved attainment of target low-density lipoprotein cholesterol (LDL-C) values when ezetimibe is combined with statins for the treatment of hypercholesterolemia. However, the efficacy of an intervention in randomized controlled clinical studies may not correlate with its efficacy in a real-world (community practice) setting. Data to support the LDL-C lowering efficacy of ezetimibe/simvastatin (EZE/SMV) outside of controlled clinical studies are currently lacking. Objective: Using patient data from a large, national, administrative claims database, this retrospective observational study evaluated the efficacy of EZE/SMV in lowering LDL-C values and achieving LDL-C target values in newly initiating users and compared the results with those from several statin monotherapies. Methods: Patients with hypercholesterolemia who had filled their first prescription for EZE/SMV or statin monotherapy between July 1, 2004, and December 31, 2007, and were ⱖ18 years old were evaluated. Pertinent data taken from the database included lipid-lowering drug name, date of first prescription, dose, serum lipid levels, and sample dates. Data on cardiovascular history, diabetes, and other concurrent diseases and therapies were also available. Following propensity score matching, multivariate regression models were constructed to estimate the impact of the choice of therapy on key treatment outcomes including the reduction of LDL-C (absolute and percent) as well as the percent of patients achieving goal LDL-C levels as defined by the updated American Heart Association/ American College of Cardiology (AHA/ACC) Guidelines for 2006.

212

Results: In the total population, both the percent decrease and the absolute reduction in LDL-C values were significantly greater with the use of EZE/SMV than with the 3 statin monotherapies (P ⬍ 0.005 for all comparisons). A significantly greater percentage of patients achieved AHA/ACC LDL-C goals in the first 3 months of EZE/SMV treatment compared with those using each statin monotherapy (P ⬍ 0.05 for all comparisons). Among matched patients with diagnosed diabetes, the percent reduction in LDL-C was also higher with the use of EZE/SMV than with each statin monotherapy (P ⬍ 0.005 for all comparisons). Conclusion: In a real-world setting, EZE/SMV appeared to be more effective than simvastatin, atorvastatin, or rosuvastatin monotherapy for attaining therapeutic goals set forth by national and professional societies in patients being initiated on statin-based lipid-lowering therapy. (Clin Ther. 2011;33:212–224) © 2011 Elsevier HS Journals, Inc. All rights reserved. Key words: ezetimibe, hypercholesterolemia, lowdensity lipoprotein cholesterol (LDL-C), managed care, simvastatin.

INTRODUCTION Death from coronary artery disease (CAD) in the United States claimed 2400 individuals daily in 2004, an average of 1 death every 37 seconds. Data from 2005 indicate that deaths from cardiovascular disease accounted for 35.2% of all deaths, or 1 in every 2.8 deaths in the United States, and that the rate is declinAccepted for publication February 22, 2011. doi:10.1016/j.clinthera.2011.02.011 0149-2918/$ - see front matter © 2011 Elsevier HS Journals, Inc. All rights reserved.

Volume 33 Number 2

H.S. Friedman et al. ing.1 This decline in cardiovascular-related deaths can be credited to both treatments and changes in risk factors, including lipid-lowering therapies, with 24% of this change attributed to reductions in total cholesterol levels.2 The primary etiology of cardiovascular disease is atherosclerosis, an inflammatory disease process predicated upon the deposition of cholesterol into the walls of affected arteries.3–5 Studies suggest not only that is there a causal relationship between blood lipid levels and cardiovascular disease but also that lowering lipid levels reduces cardiovascular mortality and morbidity.6 The cardiovascular disease risk incurred by elevated low-density lipoprotein cholesterol (LDL-C) is continuous and rises with higher levels of cholesterol in a curvilinear or log-linear fashion.7 Patients with low LDL-C levels and a high rate of LDL-C goal achievement can be expected to experience fewer cardiovascular events than those with LDL-C levels above current nationally recommended targets. In an analysis using data from 10 prospective (cohort) studies, 3 international studies in different communities, and 28 randomized controlled trials, a reduction of ischemic heart disease of 50% at age 40, 40% at age 50, 30% at age 60, and 20% at age 70 was expected from a 10% reduction in serum cholesterol concentration.8 Furthermore, the benefit can be realized quickly, with greater benefit after 2 years and the full benefit after 5 years. Aggressively reducing LDL-C has also been shown to prevent both primary and secondary cardiovascular events.9 –17 A 1998 meta-analysis of 38 trials demonstrated that every 10% reduction in total cholesterol is associated with a decrease in cardiovascular mortality by 15% and a decrease in allcause mortality by 11% regardless of treatment modalities.18 Several landmark trials have revealed that dyslipidemia associated with the diabetic state contributes significantly to cardiovascular mortality and morbidity. For example, in the United Kingdom Prospective Diabetes Study, a logistic regression model used to estimate the association between potential risk factors measured 3 to 4 months after diagnosis of diabetes and incident of cardiovascular disease identified LDL-C as the most significant factor, suggesting that every 39 mg/dL (1 mmol) decrease in LDL-C predicts a 36% risk reduction of CAD.19 The Collaborative Atorvastatin Diabetes Study was the first prospective trial to study statin therapy (atorvastatin [ATV] 10 mg daily)

February 2011

specifically in patients with type 2 diabetes without clinically evident cardiovascular disease but with at least 1 other additional risk factor. This trial demonstrated a 37% relative risk reduction in cardiovascular events following a decrease in the average LDL-C level from 118 mg/dL to 72 mg/dL.20 The importance of aggressive lipid lowering in diabetic patients with established coronary heart disease (CHD) was shown in the Treating to New Targets study. In this study, LDL-C in diabetic patients was reduced from a mean of 160 mg/dL to 96.2 mg/dL during an open-label period of ATV 10 mg. Whereas continued ATV 10 mg resulted in a 3% increase in LDL-C levels to 99 mg/dL, diabetic patients given ATV 80 mg saw a further reduction of 19% to 77 mg/dL (P ⬍ 0.0001).21 A metaanalysis of 14 randomized statin trials has demonstrated similar proportional effects of reducing LDL-C in patients with and without diabetes.22 Among the 18,686 diabetics in this meta-analysis, a 21% reduction in major vascular events (P ⬍ 0.0001) and a 9% reduction in all-cause mortality (P ⫽ 0.02) were observed for every 39 mg/dL reduction in LDL-C. Important to note was that the proportional effects of statin therapy in diabetic patients were independent of a history of vascular disease or other baseline characteristics. Finally, the results of the Heart Protection Study implied that regardless of baseline LDL-C levels, treatment with simvastatin (SMV) (40 mg daily), which achieves an LDL-C reduction of ⱖ30%, should be considered for all diabetic patients over the age of 40 years with a total cholesterol level ⱖ135 mg/dL.12 Several observational studies have shown that lipid treatment goals are not being achieved, even in individuals with CHD who may or may not be treated actively for hypercholesterolemia.23,24 For example, national survey studies conducted in Europe (eg, EUROASPIRE I and II25) and in the United States (eg, National Cholesterol Education Program Evaluation ProjecT Utilizing Novel E-Technology II [NEPTUNE II]26 and the Lipid Treatment Assessment Project [L-TAP]27,28) have found that most CHD patients were still not achieving target lipid goals. The NEPTUNE II project, using data collected in 2003, reported that only 62% of the 1322 patients with known CHD attained LDL-C levels of less than 100 mg/dL.26 In the most recent update from L-TAP, only 67% of high-risk patients were achieving LDL-C levels ⬍100 mg/dL and only 30% of patients with existing CAD (with ⱖ2 risk factors) were achieving the optional level of ⬍70 mg/dL.28

213

Clinical Therapeutics Interestingly, in both US studies, the patients at most risk were found to be the least compliant with national guidelines. Furthermore, a retrospective chart review of 386 patients (average age 64) from a large urban cardiology center indicated that even when cardiologists directed lipid management, the majority of their male patients were not even taking statins.29 The compliance of the diabetic patient with nationally recognized lipid guidelines appears to be equally disappointing and possibly worse. The third US National Health and Nutrition Examination Survey (NHANES: 1988 –1994) showed only 42% of patients with diabetes achieved an LDL-C goal ⬍130 mg/dL.30 The previously mentioned NEPTUNE II trial reported only 55% of the 1033 patients studied with diabetes and without recognized CAD had an LDL-C level ⬍100 mg/dL.27 In the Veterans Affairs Diabetes Trial in advanced type 2 diabetes, only 44% of the 1742 veterans met the LDL-C goal of ⬍100 mg/dL at the time of enrollment. Of the enrollees, 58% were receiving statin monotherapy and 11% were receiving combination therapy. In the setting of an academic family practice, only 47% of the 218 diabetic patients studied from abstracted chart information taken from their office visits between 1999 and 2001 attained the lipid goal of an LDL-C level of ⬍100 mg/dL as set forth at that time by the American Diabetes Association.31 The first available selective cholesterol absorption inhibitor, ezetimibe (EZE), has been shown to provide significant incremental reductions in LDL-C levels when coadministered with statins. It has been argued that combination therapy is a requisite to attain the recently recommended and more aggressive LDL-C goals.32 A meta-analysis of 5 randomized clinical trials of ezetimibe, involving 5039 patients, reported an additional 23.6% (95% CI, ⫺25.6 to ⫺21.7; P ⬍ 0.0001) relative reduction in LDL-C when added to ongoing statin therapy.33 Furthermore, several clinical trials have confirmed that this dual approach to lipid lowering is more effective for achieving greater LDL-C reductions and lipid goals than therapy with SMV,34 –36 ATV,37,38 and rosuvastatin (RSV) alone.39 At least one meta-analysis suggested there was a 3.4 times greater chance (95% CI, 2.0 –5.6; P ⬍ 0.0001) of being in compliance with guideline targets using a combination approach.31 For example, the Ezetimibe Add-on to Statin for Effectiveness (EASE) trial of 5802 subjects from 299 community-based practices showed that 71.0% of patients receiving ezetimibe added to a

214

number of statins met their LDL-C goal compared with 20.6% of those receiving statin therapy alone (P ⬍ 0.001).40 The Examination of Potential Lipid-Modifying Effects of Rosuvastatin in Combination with Ezetimibe Versus Rosuvastatin Alone (EXPLORER) trial of 469 subjects demonstrated that a significantly higher proportions of patients at elevated CHD risk who received EZE and RSV achieved the consensus LDL-C target of ⬍100 mg/dL (94% vs 79%, respectively; P ⬍ 0.001) or the optional target of ⬍70 mg/dL (80% vs 35%, respectively; P ⬍ 0.001) than did those on statin monotherapy.41 And finally, the Vytorin Vs Atorvastatin (VYVA) trial of 1902 subjects found that 85.4% of patients with CHD or CHD risk equivalents achieved the LDL-C level of ⬍100 mg/dL using EZE/ SMV combination therapy compared with 70.0% of those taking atorvastatin alone (P ⬍ 0.001).37 The objective of the present study was to compare the real-world (outside of a controlled, clinical study environment) effectiveness of EZE/SMV in newly initiating users and newly initiating diabetic users with that of 3 statin monotherapies for LDL-C lowering and LDL-C goal achievement. Patients were categorized by risk level according to the updated American Heart Association/American College of Cardiology (AHA/ ACC) 2006 guidelines, and only those in moderate-risk and high-risk categories were evaluated. Our intention was to gain a real-world perspective on this combination compared with that of available statin monotherapies for the purpose of informing clinicians and managed care organizations on what may be the most effective treatment option for patients with hypercholesterolemia who are aiming to reach currently recommended LDL-C goals.

PATIENTS AND METHODS Data Source A large, national, administrative claims database (Ingenix, Eden Prairie, Minnesota) that includes demographic, pharmacy, and medical claims information for employed, commercially insured patients with dependents covering ⬎15 million lives per year was used. No patient in this commercially available database is uniquely identifiable. Approximately 7.5 million patients in the complete database have at least 1 laboratory measurement. Underlying geographic information is diverse across the United States. Additional details concerning the database may be found at the ISPOR International

Volume 33 Number 2

H.S. Friedman et al. Digest of Databases (http://www.ispor.org/DigestOfIntDB/ Default.aspx?rcd⫽393).

Study Population The objective of the study was to compare newly initiating users of both EZE/SMV and statin monotherapy. To be included in the study, patients had to fill their first prescription for EZE/SMV or statin monotherapy between July 1, 2004, and December 31, 2007, and be ⱖ18 years of age. The index date was defined as the date on which this first prescription for one of the therapies of interest was filled. Patients could not have filled a prescription for any lipid-lowering medication (including statins, ezetimibe, bile acid sequestrants, fibrates, and prescription-strength niacin) during the 6 months before the index date. Continuous enrollment in the database throughout the predefined 9-month study period (from 6 months before the index date to 3 months after) was required, along with at least 1 LDL-C measurement before the index date (from 6 months before) and 1 LDL-C measurement after the index date (1- to 3-month period after). When patients had multiple laboratory test results before the index date, the most recent results were taken as the preindex (baseline) value. The mean absolute change in LDL-C and percent change in LDL-C between the baseline LDL-C value and the postindex LDL-C value was calculated. LDL-C goal attainment was also assessed. When patients had multiple laboratory test results following the index date, the first LDL-C measurement during that period was used to determine the LDL-C change (absolute and percent), and the minimum measurement during that period was used to determine whether the patient achieved their LDL-C goal. Patients were excluded if there were any changes in their lipid-lowering therapy regimen (eg, dose titration, switch from one medication to another, or addition of another lipid-lowering therapy). Because liver disease may affect the synthesis of LDL-C, patients with liver disease were excluded from the study. Patients who had received drugs that are known to interact with the effectiveness of the index drugs during the study period were excluded. The list of interacting drugs is available on request and was generated using a proprietary database maintained by the study sponsor. LDL-C goals from the updated 2006 AHA/ACC guidelines42 were used for all patients (Table I). The AHA/ACC 2006 guideline update states that the target LDL-C goals for all patients with CHD and other clin-

February 2011

Table I. Patient risk stratification by the updated AHA/ACC LDL-C treatment goals.42 Risk Category Very high risk High risk Moderately high risk Moderate risk

LDL-C Goal (mg/dL) ⬍70* ⬍100 ⬍100† ⬍130

ACC ⫽ American College of Cardiology; AHA ⫽ American Heart Association; LDL-C ⫽ low-density lipoprotein cholesterol. *Optional goal, standard goal is 100 mg/dL. † Optional goal, standard goal is 130 mg/dL.

ical forms of atherosclerotic disease should be ⬍100 mg/dL and that it is reasonable to treat such patients to a goal of ⬍70 mg/dL. Patients were categorized by risk level according to the updated AHA/ACC 2006 guidelines, and only those in moderate-risk and high-risk categories were included. Patients in the high-risk category were further classified as very high risk if they had a previous myocardial infarction, stroke, or peripheral vascular disease in addition to 1 or more of the following: diabetes mellitus, angina, intermediate coronary syndrome, acute or other coronary occlusion, or baseline triglyceride ⱖ200 mg/dL with non-HDL-C ⱖ130 mg/dL and HDL-C ⬍40 mg/dL. Patients who were at or below their target AHA/ACC 2006 LDL-C goal at baseline (period up to and including index date) were also excluded. Treatment adherence was measured by calculating the percent of days covered by filled prescriptions for index lipid-lowering therapy between the index date and the first postindex laboratory measurement. For patients with multiple LDL-C laboratory values available for the time period 1 to 3 months after the index date, the value from the first test following the index date was used for the adherence calculation. For patients meeting the study criteria, all pharmaceutical and medical claims information and laboratory test results were extracted from the claims database. Data extracted from the database included year of birth, gender, lipid-lowering drug name, date of prescription, dose, package size, serum lipid levels, and sample dates. Data on cardiovascular history, risk factors, other concurrent diseases, and therapies were identified using International Classification of Diseases, Ninth Revision, Clinical Mod-

215

Clinical Therapeutics ification (ICD-9-CM) diagnostic codes and Current Procedural Terminology (CPT) codes. Patient comorbidities, including diabetes, were also identified using the ICD-9 codes from the medical records recorded before the index date.

Statistical Analyses As a retrospective data study, the treatment assignment was not random. To correct for potential selection bias between treatment groups, propensity score matching using a 1:1 greedy algorithm43 was performed between SMV and EZE/SMV, RSV and EZE/ SMV, and ATV and EZE/SMV, in which each match was performed as a 1:1 pairwise match. Descriptive statistics were then performed to ensure match quality. General linear models (GLM) assuming gamma distributed residuals were performed to examine the impact of each therapy on the absolute and percent change in LDL-C level. Owing to the limited size of the diabetic subcohort, multiple linear regression (MLR) was used to compare LDL-C changes among matched diabetic groups. Conditional logistic regression models, designed to work with matched pairs, were used to assess whether there was a difference in probability of achieving goals between different therapies. Predictor variables included in these regression models included demographics, adherence, normalized strength of lipid-lowering therapy, and year of index date. The normalized strength measure was included as a way to control for dosage differences within each statin. The normalized strength was computed by first identifying the individual strength of each of the different statin doses. This strength was then normalized so that the normalized strength for each statin had a population mean of 0 and standard deviation of 1. Including the normalized strength variable in the model accounted for the fact that patients received different doses of each statin.

RESULTS Of the 843,794 patients in the database who were initiated on statin therapy between July 1, 2004, and December 31, 2007, a total of 5790 met all of our study inclusion criteria (Figure 1). Baseline characteristics of patients in each treatment cohort were not significantly different and are shown in Table II. Almost 60% of the study participants was male, and the average male and female participants were in their fifties. The most com-

216

mon concurrent conditions were hypertension and diabetes. More than 70% of the patients in each group had hypertension, a known risk factor for CHD. Of the 5790 who met the overall inclusion criteria, 1795 patients had a prior diagnosis of diabetes. Of note is that the difference in baseline obesity in diabetic patients between the drug groups (RSV, 11.3%; ATV, 5.9%; EZE/SMV, 6.5%; and SMV, 7.2%) was statistically significant (P ⫽ 0.0292). In addition, a minority of subjects had established cardiovascular or cerebrovascular disease.

Change in Lipid Values From Baseline to Postindex Date To control for variability between study cohorts, matching was performed as described in the methods section. The results for all pairwise comparisons in matched populations are shown in Table III. Baseline LDL-C levels were similar between the EZE/SMV cohort and other treatment cohorts following matching. The absolute change in LDL-C was significantly greater with EZE/ SMV versus RSV (–72.7 mg/dL vs – 64.7 mg/dL, respectively; P ⬍ 0.0001 for gamma GLM), EZE/SMV versus ATV (–75.7 mg/dL vs – 66.1 mg/dL, respectively; P ⫽ 0.0033 for gamma GLM), and EZE/SMV versus SMV (69.3 mg/dL vs 53.5 mg/dL, respectively; P ⬍ 0.0001 for gamma GLM) (Figure 2). Likewise, the percent change in LDL-C was greater with EZE/SMV versus RSV (44.8% vs 39.1%, respectively), EZE/SMV versus ATV (45.5% vs 39.7%, respectively), and EZE/SMV versus SMV (44.2% vs 33.8%, respectively). Pairwise comparisons of matched patients between EZE/SMV and any other group were statistically significant (P ⬍ 0.0001 for gamma GLM) (Figure 3). In matched comparisons among the subset of newly initiated users with diagnosed diabetes, those taking EZE/SMV still had a significantly greater percent reduction in LDL-C than those taking statin monotherapy (– 43.0% vs –37.9% for RSV [P ⫽ 0.0071 by MLR]; – 42.8% vs –32.4% for SMV [P ⬍ 0.0001 by MLR]; and – 43.1% vs –37.4% for ATV [P ⬍ 0.0001 by MLR]) (Figure 4). The absolute change in LDL-C was also significantly greater for those taking EZE/ SMV than for those taking SMV (– 64 mg/dL vs – 48 mg/dL, respectively; P ⬍ 0.0001 by MLR) and for users of EZE/SMV over users of ATV (– 67 mg/dL vs –58 mg/dL, respectively; P ⬍ 0.0001 by MLR).

Volume 33 Number 2

H.S. Friedman et al.

Figure 1. Patient disposition and selection for the study. ATV ⫽ atorvastatin; EZE ⫽ ezetimibe; LDL-C ⫽ lowdensity lipoprotein cholesterol; RSV ⫽ rosuvastatin; SMV ⫽ simvastatin.

February 2011

217

Clinical Therapeutics

Table II. Baseline characteristics of unmatched patients by study medication. Variable

EZE/SMV (n ⫽ 1028)

SMV (n ⫽ 957)

ATV (n ⫽ 2962)

RSV (n ⫽ 843)

Female patients, % Patient age, y (SD) Diabetes, % MI, % Other acute and subacute ischemic heart disease, % Blood disorder, % Familial hypercholesterolemia,% Angina, % Stroke, % Hypertension, % Alcohol, % Family history ischemic disease, % Menopause, % Chronic ischemic heart disease, % Heart failure, % Cancer, % HIV, % Arthritis, % Renal failure, % Depression % Infectious disease, % Mycoses, % Obesity, % Other endocrine disease, % Other mental disease, % Pregnancy, % Peripheral Vascular Disease (PVD), % Renal, % Respiratory disease, % Smoking, % Risk category (category 1 ⫽ highest) Tuberculosis, % Carotid, % Hepatic, %

32.7 54.0 (8.8) 28.6 1.3 2.7 8.5 38.9 6.9 0.0 71.9 0.5 2.1 3.7 14.7 0.3 5.2 0.0 7.1 1.2 8.2 7.3 3.2 4.6 17.1 3.4 0.1 2.1 1.2 11.1 6.3 1.9 0.3 0.0 0.0

35.7 54.4 (9.7) 34.8 2.0 4.0 8.9 34.3 5.9 0.0 70.6 0.7 3.0 4.1 13.4 0.2 5.1 0.1 6.7 1.3 8.8 7.9 4.6 4.7 16.9 2.4 0.3 2.6 1.3 11.5 7.2 3.1 0.0 0.0 0.0

32.2 54.0 (9.1) 31.1 2.3 3.3 8.4 37.4 5.6 0.0 72.4 0.7 2.9 3.9 15.2 0.4 4.2 0.2 7.0 1.4 8.7 8.4 3.8 4.7 16.3 3.3 0.1 2.5 1.4 12.8 7.5 2.4 0.0 0.0 0.0

30.1 53.7 (9.5) 29.4 1.1 2.1 8.5 36.8 5.3 0.0 73.1 0.5 3.0 2.5 15.9 0.5 4.7 0.4 7.8 1.7 9.4 8.8 3.7 6.9 17.9 3.4 0.5 3.7 1.7 11.5 7.6 2.6 0.0 0.0 0.0

ATV ⫽ atorvastatin; EZE ⫽ ezetimibe; MI ⫽ myocardial infarction; RSV ⫽ rosuvastatin; SMV ⫽ simvastatin.

LDL-C Goal Attainment At least two thirds of patients in all of the matched populations achieved their LDL-C goal within the time period (1–3 months) between baseline and postindex measurements (Figure 5 and Table III). A significantly

218

higher percent of patients achieved their LDL-C goal during the postindex period with the EZE/SMV combination than with the 3 statin monotherapies (81.8% for EZE/SMV vs 73.6% for RSV, P ⫽ 0.0245; 80.5% for EZE/SMV vs 73.3% for ATV, P ⫽ 0.0013; and

Volume 33 Number 2


Table III. Change in LDL-C and percent goal achievement for all pair-wise comparisons. EZE/SMV vs RSV

Age, y Female, % Baseline LDL-C, mg/dL LDL-C change, mg/dL LDL-C change, % Goal achievement, %

EZE/SMV vs ATV

EZE/SMV vs SMV

EZE/SMV (n ⫽ 840)

RSV (n ⫽ 840)

EZE/SMV (n ⫽ 1026)

ATV (n ⫽ 1026)

EZE/SMV (n ⫽ 784)

SMV (n ⫽ 784)

54.2 32.0 158.6 –72.7* –44.8* 81.8‡

53.7 30.1 160.7 –64.7 –39.1 73.6

54.0 32.8 162.8 –75.7† 45.5* 80.5§

53.2 30.3 163.2 –66.1 –39.7 73.3

54.3 33.2 153.4 –69.3* 44.2* 82.8¶

54.3 35.6 154.3 –53.5 –33.8 70.5

ATV ⫽ atorvastatin; EZE ⫽ ezetimibe; LDL-C ⫽ low-density lipoprotein cholesterol; RSV ⫽ rosuvastatin; SMV ⫽ simvastatin. *P ⬍ 0.0001. † P ⫽ 0.0033 (P values by gamma analysis; EZE/SMV vs statin monotherapy). ‡ P⫽ 0.0245. § P ⫽ 0.0013. ¶ P ⬍ 0.0001 (P values by conditional logistic regression analysis; EZE/SMV vs statin monotherapy).

82.8% for EZE/SMV vs 70.5% for SMV; P ⬍ 0.0001 for EZE/SMV). All comparisons were by conditional logistic regression analysis.

DISCUSSION Although similar analyses have compared efficacy among statin monotherapies, to our knowledge this is the first study that has compared the effectiveness of treatment initiation with EZE/SMV and other statin monotherapies in a real-world setting. The results demonstrate that EZE/SMV is superior to statin monotherapy (SMV, ATV, or RSV) with regard to both LDL-C lowering and therapeutic goal attainment in patients in a managed care population who have hypercholesterolemia and are at risk for cardiovascular events. Of importance is the fact that EZE/SMV combination therapy was also superior when our analyses were restricted only to newly initiated patients with diabetes. Starting patients on, or switching to, highefficacy LDL-C–lowering therapy has been advocated as a means of enhancing the achievement of treatment guidelines in real-world settings.44 The present study confirmed and corroborated the findings of clinical trials that a dual approach to lipid lowering is more effective for LDL-C reduction and goal attainment than that of statin therapy alone. The main benefit of the use of retrospective observational studies of real-world populations is that it al-

February 2011

lows understanding of real-world patient care. If a claims database contains accurate and reliably identified patient demographics and treatments, a retrospective examination of a large claims database permits insights into the usual care of patients and can achieve substantive valid analysis of chosen end points. Although randomization of the study population is not performed in the usual fashion seen in the beginning of prospective clinical trials, observational studies can be corrected retrospectively for differences in patient characteristics associated with nonrandom real-world treatment assignment using techniques such as propensity score matching. Although randomized clinical trials are useful in establishing efficacy of medications, patient care in the real world almost always differs from that of trial settings, which impose stringent inclusion criteria, firm adherence to treatment protocols, and close supervision and monitoring of treatment and outcomes. The discrepancy between clinical trial and real-world experiences of lipid treatment goal attainments have been explained, at least partially, by both physician45 and patient46 barriers. In particular, patient adherence to medical regimens has been identified as a significant health policy concern by the National Cholesterol Education Program.47 In clinical trials, the exclusion criteria eliminate patients with comorbidities and other conditions that may reduce their representative nature

219


Figure 2. Mean absolute reduction in LDL-C from baseline to postindex measurement among matched populations. ATV ⫽ atorvastatin; EZE ⫽ ezetimibe; LDL-C ⫽ low-density lipoprotein cholesterol; RSV ⫽ rosuvastatin; SMV ⫽ simvastatin. *P ⬍ 0.0001; †P ⬍ 0.0033. P values generated by general linear models assuming gamma distributed residuals.

ing HDL-C levels.49 Before attention is directed to these secondary goals, however, it should be emphasized again that LDL-C goals set forth by national and professional society guidelines are rarely met in real-world clinical settings. Since adherence to medical regimens and attention to guidelines appear to diminish after initiation of therapy, it has been argued that using the most effective lipid-lowering agent in the beginning of treatment may reduce the efficacy differences seen between clinical practice and research trials.50 It is likely that high-efficacy statins or high-potency combination therapy may be needed to minimize treatment failure and to improve cardiovascular outcomes optimally. In the long term, the best pharmacologic approach not only will improve patient outcomes but also will reduce overall patient health care costs. When managed care organizations consider policies to lower the burden of cardiovascular disease in their members, they must ensure formulary access to appropriate treatment alternatives. The results of this study should help managed care organizations understand the health care implications of making available combination lipid-lowering therapy for their own patient pop-

to patients in clinical practice. Close surveillance of subjects in clinical trials results in greater compliance and enforcement of the prescribed medical regimens than would be expected in real-world situations. Consequently, such scrutiny may lead subjects to change their behavior to approximate the expectations of the observing investigators. Furthermore, patients recruited in such trials may already be predisposed to a better outcome, which is reflected by their willingness to participate. Their interest in engaging in research may signify an elevated level of health knowledge and of confidence in the planned intervention. Thus, results of observational studies of subjects drawn from large, managed care databases may more closely approximate the patient experiences seen in the usual care setting than those available through clinical trials. Despite impressive effects on cardiovascular outcomes in patients treated with statin therapy in clinical trials, a significant residual risk of major cardiovascular events persists.48 Therefore, it has been argued that the focus of lipid therapy should be directed not only toward reducing LDL-C levels but also toward other lipid goals, such as reducing triglycerides and increas-

Figure 3. Mean percent reduction in LDL-C from baseline to postindex measurement among matched populations. ATV ⫽ atorvastatin; EZE ⫽ ezetimibe; LDL-C ⫽ low-density lipoprotein cholesterol; RSV ⫽ rosuvastatin; SMV ⫽ simvastatin. *P ⬍ 0.0001. P values generated by general linear models assuming gamma distributed residuals.

220

Volume 33 Number 2


Figure 4. Percent reduction in LDL-C among matched diabetic patients for each drug group. ATV ⫽ atorvastatin; EZE ⫽ ezetimibe; LDL-C ⫽ low-density lipoprotein cholesterol; RSV ⫽ rosuvastatin; SMV ⫽ simvastatin. *P ⫽ 0.0071; †P ⬍ 0.0001. P values generated by general linear models assuming gamma distributed residuals.

the results, these analyses were limited to patients in the very high risk, high-risk, and moderate-risk categories. (2) Despite the large pool of available patient data, the sample size for analysis was small, partially owing to limited availability of LDL-C laboratory data. (3) Adherence in these analyses was measured based on prescription refill activity rather than patient reporting, and any data regarding medication samples used by patients were not available. (4) Sample biases involved in the claims database may weaken the generalizability of the results to the entire US population. For example, our study was based on data on patients who were gainfully employed and commercially insured, thus representing a relatively stable and homogeneous population. (5) This study used propensity score matching, which is limited in its ability to match only variables that are accessible. Although this study has sought to identify relevant variables related to illness severity, it is recognized that other variables that may have been useful were not available. For example, examination of the association or extent of illness, severity of comorbid diseases, or the factors that affected the choice of index drug selection were not possible

ulations. It is anticipated that intensive lipid lowering beyond currently achieved treatment goals will translate into more lives and health care resources saved. Outcome data from ongoing trials are necessary to confirm what is reported here. Although requiring confirmation, the managed care population evaluated by this study probably reflects those covered by other managed care organizations. The findings of this study should be interpreted in light of several limitations, most of which are associated with retrospective claims data analyses in general. (1) Risk categories were assigned based on data provided by administrative claims. As such, some patients without CHD or CHD risk equivalents may have been assigned inadvertently to a risk level lower than appropriate for their true medical condition owing to the underreporting of certain risk factors, such as family history of premature CHD or smoking. This would have no impact on the observed LDL-C change and percent LDL-C change but would result in a lower assigned LDL-C goal and a higher percent of patients achieving that goal. To avoid this potential skewing of the analysis and confounding of the interpretation of

Figure 5. American Heart Association/American College of Cardiology LDL-C goal attainment among matched populations. ATV ⫽ atorvastatin; EZE ⫽ ezetimibe; LDL-C ⫽ low-density lipoprotein cholesterol; RSV ⫽ rosuvastatin; SMV ⫽ simvastatin. *P ⫽ 0.0245; †P ⫽ 0.0013; ‡P ⬍ 0.0001. P values generated by conditional logistic regression models.

February 2011

221

Clinical Therapeutics because these elements were not completely captured by claims data. Our estimates of goal achievement in this study are higher than those in previously reported observational studies. This difference can be explained in part by selection differences between study populations and study designs. First of all, our study examined treated patients, whereas other observational studies included both untreated as well as treated patients. In addition, there is a lack of randomness associated with actively monitored patients who were required to have LDL-C values both before and after therapy initiation. We expect the information from this study will be helpful especially for managed care organizations that implement population health improvement strategies to reduce cardiovascular disease burden in their respective population of covered patients.

CONCLUSION The results of this study, based on data emanating from a real-world environment, suggest that the EZE/SMV combination is the most effective lipid-lowering treatment for achieving consensus therapeutic goals set forth by national and professional societies.

ACKNOWLEDGMENTS We thank Sean Gregory for providing writing and editorial assistance on behalf of the authors as well as assistance with the electronic submission. The study, and assistance with the manuscript, was funded by Merck/Schering-Plough Pharmaceuticals, North Wales, Pennsylvania. All authors were involved in the data analysis and development of the manuscript, and all gave final approval. Dr. Friedman contributed to the study design and performed the statistical analyses; Dr. Rajagopalan was responsible for principal study design and summary review; Mrs. Barnes contributed to study design and summary review; and Dr. Roseman contributed to study design, provided strategy for analysis of diabetic patients, and contributed to summary review. Dr. Rajagopalan is a full-time employee of Med Data Analytics, Inc. Mr. Barnes is a full-time employee of Schering-Plough. Dr. Roseman is the owner of Analytic Solutions, LLC, a consulting firm that works in the pharmaceutical industry with companies including Merck/ScheringPlough Pharmaceuticals. These data have not been presented previously at a scientific meeting. The authors

222

have indicated that they have no other conflicts of interest regarding the content of this article.

REFERENCES 1. Rosamond W, Flegal K, Furie K, et al. Heart disease and stroke statistics—2008 update: A report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2008;117:e25– e146. 2. Ford ES, Ajani UA, Croft JB, et al. Explaining the decrease in US deaths from coronary disease, 1980 –2000. N Engl J Med. 2007;356:2388 –2398. 3. Ross R. Atherosclerosis—an inflammatory disease. N Engl J Med. 1999;340:115–126. 4. Steinberg D. Atherogenesis in perspective: hypercholesterolemia and inflammation as partners in crime. Nat Med. 2002;8:1211–1217. 5. Berliner JA, Navab M, Fogelman AM, et al. Atherosclerosis: Basic mechanisms. Oxidation, inflammation, and genetics. Circulation. 1995;91:2488 –2496. 6. Gotto AM Jr. Evolving concepts of dyslipidemia, atherosclerosis, and cardiovascular disease: The Louis F. Bishop Lecture. J Am Coll Cardiol. 2005;46:1219 –1224. 7. Stamler J, Wentworth D, Neaton JD. Is relationship between serum cholesterol and risk of premature death from coronary heart disease continuous and graded? Findings in 356,222 primary screenees of the Multiple Risk Factor Intervention Trial (MRFIT). JAMA. 1986;256:2823–2828. 8. Law MR, Wald NJ, Thompson SG. By how much and how quickly does reduction in serum cholesterol concentration lower risk of ischaemic heart disease? BMJ. 1994;308:367– 372. 9. The Lipid Research Clinics Coronary Primary Prevention Trial results. I. Reduction in incidence of coronary heart disease. JAMA. 1984;251:351–364. 10. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: The Scandinavian Simvastatin Survival Study (4S). Lancet. 1994;344:1383–1389. 11. The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. N Engl J Med. 1998;339:1349 –1357. 12. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: A randomised placebo-controlled trial. Lancet. 2002;360:7–22. 13. Cannon CP, Braunwald E, McCabe CH, et al. Intensive versus moderate lipid lowering with statins after acute coronary syndromes. N Engl J Med. 2004;350:1495–1504. 14. Downs JR, Clearfield M, Weis S, et al. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels: Results of AFCAPS/

Volume 33 Number 2


15.

16.

17.

18.

19.

20.

21.

22.

TexCAPS. Air Force/Texas Coronary Atherosclerosis Prevention Study. JAMA. 1998;279:1615–1622. LaRosa JC, Grundy SM, Waters DD, et al. Intensive lipid lowering with atorvastatin in patients with stable coronary disease. N Engl J Med. 2005; 352:1425–1435. Sacks FM, Pfeffer MA, Moye LA, et al. The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events Trial investigators. N Engl J Med. 1996;335:1001– 1009. Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. West of Scotland Coronary Prevention Study Group. N Engl J Med. 1995;333:1301–1307. Gould AL, Rossouw JE, Santanello NC, et al. Cholesterol reduction yields clinical benefit: Impact of statin trials. Circulation. 1998;97: 946 –952. Turner RC, Millns H, Neil HA, et al. Risk factors for coronary artery disease in non–insulin dependent diabetes mellitus: United Kingdom Prospective Diabetes Study (UKPDS: 23). BMJ. 1998;316:823– 828. Colhoun HM, Betteridge DJ, Durrington PN, et al. Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): Multicentre randomised placebo-controlled trial. Lancet. 2004;364:685– 696. Shepherd J, Barter P, Carmena R, et al. Effect of lowering LDL cholesterol substantially below currently recommended levels in patients with coronary heart disease and diabetes: The Treating to New Targets (TNT) study. Diabetes Care. 2006;29: 1220 –1226. Kearney PM, Blackwell L, Collins R, et al. Efficacy of cholesterol-lowering therapy in 18,686 people with diabetes in 14 randomised trials of

February 2011

23.

24.

25.

26.

27.

28.

29.

statins: A meta-analysis. Lancet. 2008;371:117–125. Andrade SE, Walker AM, Gottlieb LK, et al. Discontinuation of antihyperlipidemic drugs— do rates reported in clinical trials reflect rates in primary care settings? N Engl J Med. 1995;332:1125–1131. Majumdar SR, Gurwitz JH, Soumerai SB. Undertreatment of hyperlipidemia in the secondary prevention of coronary artery disease. J Gen Intern Med. 1999;14:711–717. EUROASPIRE Study Group. A European Society of Cardiology survey of secondary prevention of coronary heart disease: Principal results. EUROASPIRE Study Group. European Action on Secondary Prevention through Intervention to Reduce Events. Eur Heart J. 1997;18:1569 –1582. Davidson MH, Maki KC, Pearson TA, et al. Results of the National Cholesterol Education (NCEP) Program Evaluation ProjecT Utilizing Novel E-Technology (NEPTUNE) II survey and implications for treatment under the recent NCEP Writing Group recommendations. Am J Cardiol. 2005;96:556 –563. Pearson TA, Laurora I, Chu H, Kafonek S. The lipid treatment assessment project (L-TAP): A multicenter survey to evaluate the percentages of dyslipidemic patients receiving lipid-lowering therapy and achieving low-density lipoprotein cholesterol goals. Arch Intern Med. 2000;160:459 – 467. Waters DD, Brotons C, Chiang CW, et al. Lipid treatment assessment project 2: A multinational survey to evaluate the proportion of patients achieving low-density lipoprotein cholesterol goals. Circulation. 2009; 120:28 –34. Aliyu ZY, Yousif SB, Plantholt K, et al. Assessing compliance of cardiologists with the national cholesterol education program (NCEP) III guidelines in an ambulatory care setting. Lipids Health Dis. 2004;3:9.

30. Resnick HE, Foster GL, Bardsley J, Ratner RE. Achievement of American Diabetes Association clinical practice recommendations among US adults with diabetes, 1999 – 2002: The National Health and Nutrition Examination Survey. Diabetes Care. 2006;29:531–537. 31. Putzer G, Roetzheim R, Ramirez AM, et al. Compliance with recommendations for lipid management among patients with type 2 diabetes in an academic family practice. J Am Board Fam Pract. 2004;17:101–107. 32. McKenney JM. Pharmacologic options for aggressive low-density lipoprotein cholesterol lowering: Benefits versus risks. Am J Cardiol. 2005; 96:60E– 66E. 33. Mikhailidis DP, Sibbring GC, Ballantyne CM, et al. Meta-analysis of the cholesterol-lowering effect of ezetimibe added to ongoing statin therapy. Curr Med Res Opin. 2007; 23:2009 –2026. 34. Bays HE, Ose L, Fraser N, et al. A multicenter, randomized, doubleblind, placebo-controlled, factorial design study to evaluate the lipidaltering efficacy and safety profile of the ezetimibe/simvastatin tablet compared with ezetimibe and simvastatin monotherapy in patients with primary hypercholesterolemia. Clin Ther. 2004;26:1758 –1773. 35. Davidson MH, McGarry T, Bettis R, et al. Ezetimibe coadministered with simvastatin in patients with primary hypercholesterolemia. J Am Coll Cardiol. 2002;40:2125–2134. 36. Goldberg AC, Sapre A, Liu J, et al. Efficacy and safety of ezetimibe coadministered with simvastatin in patients with primary hypercholesterolemia: A randomized, double-blind, placebo-controlled trial. Mayo Clin Proc. 2004;79:620 – 629. 37. Ballantyne CM, Abate N, Yuan Z, et al. Dose-comparison study of the combination of ezetimibe and simvastatin (Vytorin) versus atorvastatin in patients with hypercholesterolemia: The Vytorin Versus

223


38.

39.

40.

41.

42.

43.

44.

45.

46.

Atorvastatin (VYVA) study. Am Heart J. 2005;149:464 – 473. Ballantyne CM, Blazing MA, King TR, et al. Efficacy and safety of ezetimibe co-administered with simvastatin compared with atorvastatin in adults with hypercholesterolemia. Am J Cardiol. 2004;93:1487–1494. Catapano AL, Davidson MH, Ballantyne CM, et al. Lipid-altering efficacy of the ezetimibe/simvastatin single tablet versus rosuvastatin in hypercholesterolemic patients. Curr Med Res Opin. 2006;22:2041–2053. Pearson TA, Denke MA, McBride PE, et al. A community-based, randomized trial of ezetimibe added to statin therapy to attain NCEP ATP III goals for LDL cholesterol in hypercholesterolemic patients: The ezetimibe add-on to statin for effectiveness (EASE) trial. Mayo Clin Proc. 2005;80:587–595. Ballantyne CM, Weiss R, Moccetti T, et al. Efficacy and safety of rosuvastatin 40 mg alone or in combination with ezetimibe in patients at high risk of cardiovascular disease (results from the EXPLORER study). Am J Cardiol. 2007;99:673– 680. Smith SC Jr, Allen J, Blair SN, et al. AHA/ACC guidelines for secondary prevention for patients with coronary and other atherosclerotic vascular disease: 2006 update: Endorsed by the National Heart, Lung, and Blood Institute. Circulation. 2006; 113:2363–2372. Dehejia R, Wahba S. Propensity score-matching methods for nonexperimental causal studies. Rev Econ Stat. 2002;84:151–161. Davidson MH. Differences between clinical trial efficacy and real-world effectiveness. Am J Manag Care. 2006; 12:S405–S411. Cabana MD, Rand CS, Powe NR, et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999; 282:1458 –1465. Chaudhry HJ, McDermott B. Recognizing and improving patient nonad-

224

herence to statin therapy. Curr Atheroscler Rep. 2008;10:19 –24. 47. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002;106:3143–3421. 48. Baigent C, Keech A, Kearney PM, et al. Efficacy and safety of cholesterol-lowering treatment: Prospective meta-analysis of data from

90,056 participants in 14 randomised trials of statins. Lancet. 2005;366:1267–1278. 49. Gotto AM. High-density lipoprotein cholesterol and triglycerides as therapeutic targets for preventing and treating coronary artery disease. Am Heart J. 2002;144:S33–S42. 50. Caspard H, Chan AK, Walker AM. Determinants of the differences in LDL-cholesterol after initiation of statin treatment. Ann Pharmacother. 2006;40:21–26.

Address correspondence to: Howard S. Friedman, PhD, Analytic Solutions, LLC, 65 East 3rd Street, Suite 1R, New York NY 10003. E-mail: [email protected]

Volume 33 Number 2