More than one-third of Americans and > 20% of veterans have obesity with a body mass index (BMI) ≥ 30 kg/m2.1,2 It is well documented that patients with obesity have altered lipid metabolism, drug distribution, and drug clearance.3-5 As many as 8.2 million Americans may receive statin (3-hydroxymethylglutaryl coenzyme A reductase inhibitors) prescriptions if the American College of Cardiology/American Heart Association 2013 Cholesterol Guidelines are followed; therefore, it is important to examine how the efficacy of these drugs is altered in patients with obesity.6
Multiple studies have examined the benefits of statin therapy through lowering low-density lipoprotein cholesterol (LDL-C); however, few have examined the impact of obesity on statin efficacy. For example, only 18% of subjects in the Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) trial were classified as having obesity, and subjects in the Scandinavian Simvastatin Survival Study (4S) trial had a mean BMI of only 26 kg/m2.7,8 Though statins decreased mortality in both of these studies, it is unknown whether the lipid-lowering effects were the same for participants with and without obesity. The Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS) demonstrated a decrease in major cardiovascular events and all-cause mortality with atorvastatin 10 mg daily therapy in a sample where more than one-third of subjects had obesity.9 However, the mean baseline BMI of subjects in both study groups was only 28 kg/m2, and outcomes for those with and without obesity were not compared.9
Studies that have examined statin efficacy in those with and without obesity include the Heart Protection Study (HPS), a post hoc analysis of the West of Scotland Coronary Prevention Study (WOSCOPS), and a meta-analysis by Blassetto and colleagues. The HPS examined the event rate of vascular events with simvastatin 40 mg daily in patients with diabetes mellitus (DM).10 Though these subgroups were compared in HPS, no statistical difference was demonstrated between these groups for the rate of vascular events among those with and without DM.10 However, the obesity subgroup’s event rate ratios were consistently higher than were those for the nonobese group.10
A post hoc analysis of WOSCOPS examined obesity as a factor for change in LDL-C with pravastatin 40 mg therapy.11 Though the authors found that no significant difference was present between those with and those without obesity, the data supporting this claim were not disclosed, which makes drawing clinical conclusions from this analysis difficult.11 A meta-analysis by Blassetto and colleagues examined the association between rosuvastatin’s efficacy in lowering LDL-C among the subgroups of hypertension, atherosclerosis, type 2 DM, and obesity.12 Though these subgroups were not compared statistically, the obesity subgroup had the lowest mean percent change in lowering LDL-C. Moreover, patients without obesity were not examined as a subgroup.12
With the expected increase in statin therapy and a significant portion of the U.S. population having obesity, it is necessary to determine if obesity alters the efficacy of statins. This study was conducted to determine the effect of obesity on the percent change in LDL-C with statin therapy within a veteran population.
This study was a retrospective review examining follow-up data from January 1, 2009 to July 1, 2014 from the VA Midsouth Healthcare Network. This network services more than 350,000 patients each year in Tennessee, Kentucky, and West Virginia. Data were gathered and analyzed on the VA Informatics and Computing Infrastructure (VINCI) servers. Patients were included in this study if they were aged ≥ 18 years with a new filled prescription for simvastatin 20 mg or simvastatin 40 mg daily. Simvastatin was chosen because it was the formulary statin during the study period. This study was approved by the James H. Quillen VAMC/East Tennessee State University Institutional Review Board.
Patients were excluded if they had received treatment for hyperlipidemia (niacin, colestyramine, colestipol, colesevelam, other statins, gemfibrozil, fenofibrate, omega-3 ethyl esters, ezetimibe) during the 6 weeks prior to the initial fill date of the statin prescription. Patients whose simvastatin therapy did not span the follow-up period from the time of filling to the follow-up lipid panel were excluded, as were those who had not filled a simvastatin prescription within 30 days of their baseline lipid panel. Also excluded were patients who were newly established at the VA, pregnant, or receiving concomitant antihyperlipidemia agents, dialysis, or interacting medications (tacrolimus, cyclosporine, atazanavir, darunavir, nelfinavir, saquinavir, ritonavir, indinavir, lopinavir, tipranavir, fosamprenavir, fluconazole, voriconazole, itraconazole, voriconazole, posaconazole, amiodarone, or colchicine). Patients with a BMI < 18 kg/m2, hepatic failure as measured by an aspartate transaminase/alanine transaminase (AST/ALT) ratio > 3 times the upper limit of normal, hepatitis, a history of alcoholism, any change in statin dose prior to follow-up cholesterol values, or no follow-up LDL-C values also were excluded.
The baseline data collected included age, sex, weight, height, BMI, hemoglobin A1c, LDL-C, ALT/AST, and serum creatinine (SCr). All other laboratory results were required to be within 270 days of the time the lipid panel was obtained. The index date was set as the date the initial prescription was filled between February 1, 2009 and April 1, 2014. Follow-up levels for LDL-C were obtained 40 to 95 days after the index date. Direct LDL-C values were preferred unless only calculated values were available. Calculated LDL-C values were determined by using the Friedewald equation. An audit of 150 patient charts was conducted to ensure the integrity of data pulled from the database.
The percent changes in LDL-C were calculated for those with and without obesity for both simvastatin 20 mg daily and simvastatin 40 mg daily. The primary outcome was the percent change in LDL-C from baseline. All laboratory values were compared using independent 2-tailed t tests with α set to .05. To have an 80% chance of detecting a 5% difference in percent change in LDL-C between the experimental and control groups, 129 patients were required. To determine whether an association was present, a correlation between BMI and percent change in LDL-C was conducted. All statistics were conducted using SAS software (Cary, North Carolina).