Simvastatin Improves Flow-Mediated Dilation but Reduces Adiponectin Levels and Insulin Sensitivity in Hypercholesterolemic Patients

We used a randomized, double-blind, placebo-controlled, and parallel study design. Age, sex, and BMI were matched among all subjects. Patients with hypercholesterolemia (LDL cholesterol levels ≥100 mg/dl and BMI ≥23.0 mg/m²) participated in this study. We excluded patients with overt liver disease, chronic renal failure, hypothyroidism, myopathy, uncontrolled diabetes, severe hypertension, stroke, acute coronary events, coronary revascularization within the preceding 3 months, or alcohol abuse. No patient had taken any lipid-lowering agent or antioxidant vitamin supplements or had undergone hormone replacement therapy during the 2 months preceding our study. Before and during the study period a dietitian educated patients on maintaining a low-fat diet. Activity levels of the subjects were not monitored before or during the study. Clinical characteristics of these patients are summarized in Table 1. Each of 32 patients in five groups was given placebo or 10, 20, 40, or 80 mg simvastatin, respectively, once daily during a 2-month treatment period. A research nurse counted pills at the end of treatment to monitor compliance. The patients were seen at least every 14 days during the study. To minimize side effects, we measured serum aspartate aminotransferase, alanine aminotransferase, creatine kinase, blood urea nitrogen, and creatinine before and after therapy. Two patients receiving 10 mg simvastatin withdrew from the study because they moved to other places and one patient each receiving 40 and 80 mg simvastatin experienced moderate elevations in serum liver enzyme or creatine kinase, respectively, and dropped out of the study. Thus, 32 patients receiving placebo and 30, 32, 31, and 31 patients receiving 10, 20, 40, and 80 mg simvastatin, respectively, finished the study. None of the patients were diabetic. No additional medications including aspirin or nonsteroidal anti-inflammatory drugs were allowed during the study period to avoid confounding effects of other drugs. Calcium channel or β-adrenergic blockers were withheld for ≥48 h before the study. The study was approved by the Gil Hospital Institute Review Board, and all participants gave written, informed consent.

Blood samples for laboratory assays were obtained at ∼8:00 a.m. after overnight fasting before and at the end of each 2-month treatment period. These samples were immediately coded so that investigators performing laboratory assays were blinded to subject identity or study sequence.

Assays for lipids, glucose, and plasma adiponectin were performed in duplicate by ELISA (R&D Systems, Minneapolis, MN) for high-sensitivity C-reactive protein (CRP) levels by latex agglutination [CRP-Latex(II); Denka-Seiken, Tokyo, Japan], and assays for plasma insulin levels were performed by immunoradiometric assay (INSULIN-RIABEADII; SRL, Tokyo, Japan) as described previously (15–19). The interassay and intra-assay coefficients of variation were <6%. The Quantitative Insulin-Sensitivity Check Index (QUICKI), a surrogate index of insulin sensitivity based on fasting glucose and insulin levels, was calculated as follows (insulin is expressed in microunits per milliliter and glucose in milligrams per deciliter): QUICKI = 1/[log(insulin) + log(glucose)] (20). Imaging studies of the right brachial artery were performed using an ATL HDI 3000 ultrasound machine (ATL-Philips, Bothell, WA) equipped with a 10 MHz linear-array transducer, based on a previously published technique (15–19).

Data are expressed as means ± SEM or median (range 25–75%). After testing data for normality, we used Student’s paired t test or the Wilcoxon signed-rank test to compare values between baseline and treatment at 2 months, as reported in Table 2. We used one-way ANOVA or Kruskal-Wallis ANOVA on ranks to compare baseline or treatment effects among treatment groups. Post hoc comparisons between different treatment pairs were made using the Student-Newman-Keuls multiple comparison procedures or Dunn’s method. Pearson or Spearman correlation coefficient analysis was used to assess associations between measured parameters. We calculated that 30 subjects would provide 80% power for detecting an absolute increase of 1.5% or greater in flow-mediated dilation (FMD) of the brachial artery between baseline and 20 mg simvastatin, with α = 0.05 based on our previous studies (18). The comparison of endothelium-dependent dilation was prospectively designated as the primary end point of the study. All other end points and comparisons were considered secondary. A value of P < 0.05 was considered to represent statistical significance.

Age, sex, and BMI were matched in all groups of subjects. There were no significant differences between groups for any of the baseline measurements (Table 2).

Placebo significantly reduced total and LDL cholesterol levels from baseline. Simvastatin doses of 10, 20, 40, and 80 mg significantly reduced total cholesterol (mean changes 27, 25, 37, and 38%), LDL cholesterol (39, 38, 52, and 54%), and apolipoprotein B levels (24, 30, 36, and 42%) from baseline (all P < 0.001 by paired t test) after 2 months of administration. These effects of four different doses of simvastatin were also significant compared with placebo (P < 0.001 by ANOVA) (Fig. 1). No significant changes in other lipid profiles were noted with simvastatin therapy at any of the doses evaluated compared with placebo.

Placebo did not significantly improve the flow-mediated dilator response to hyperemia relative to baseline measurements. Simvastatin doses of 10, 20, 40, and 80 mg significantly improved FMD (mean changes 68, 40, 49, and 63%) after 2 months of therapy compared with baseline (P < 0.001 by paired t test). All of these effects were also significant compared with placebo (P < 0.001 by ANOVA) (Fig. 1). Brachial artery dilator responses to nitroglycerin were not significantly different among any of the therapies. In addition, no significant changes in high-sensitivity CRP were noted with any of the therapies (P = 0.363 by ANOVA on ranks).

Placebo did not significantly change insulin or glucose levels from baseline. Simvastatin at a dose of 80 mg significantly increased glucose levels (mean change 7%) after 2 months of administration compared with baseline (P = 0.011 by paired t test). The effects of four different doses of simvastatin on glucose were not significant compared with placebo (P = 0.501 by ANOVA). Simvastatin doses of 10, 20, 40, and 80 mg tended to increase insulin levels (mean changes 67, 160, 93, and 96%) after 2 months of therapy compared with baseline (P = 0.096, P = 0.007, P = 0.005, and P = 0.268 by paired t test, respectively). The effects of four different doses of simvastatin in terms of raising fasting insulin levels were significant compared with placebo (P = 0.006 by ANOVA on ranks) (Fig. 1). Post hoc comparison demonstrated a significant difference between placebo and 20 mg simvastatin (P < 0.05). We observed significant correlations between baseline HDL cholesterol levels and baseline adiponectin levels (r = 0.354 before placebo; r = 0.540 before 10 mg simvastatin; r = 0.323 before 20 mg simvastatin; r = 0.376 before 40 mg simvastatin; and r = 0.537 before 80 mg simvastatin). Placebo did not significantly change plasma adiponectin levels and insulin sensitivity (determined by QUICKI) relative to baseline measurements. Simvastatin at doses of 10, 20, 40, and 80 mg significantly decreased plasma adiponectin levels (mean changes 4, 12, 5, and 10%) and insulin sensitivity (5, 8, 6, and 6%) compared with baseline (all P < 0.05 by paired t test). Moreover, these effects of all four doses of simvastatin were significant compared with placebo (P = 0.011 for adiponectin and P = 0.034 for QUICKI by ANOVA) (Fig. 1). The magnitude of percent changes in FMD, adiponectin, and QUICKI were not significantly different among the four different doses of simvastatin despite dose-dependent changes in reduction of apolipoprotein B levels.

We investigated whether changes in percent flow-mediated dilator response to hyperemia, plasma levels of adiponectin, or insulin resistance were related to changes in lipoprotein levels. There were no significant correlations between changes in these parameters and changes in lipoprotein levels after any of the simvastatin doses. There were inverse correlations between percent changes in adiponectin levels and percent changes in insulin (r = −0.414, P = 0.018 after placebo; r = −0.273, P = 0.152 after 10 mg simvastatin; r = −0.211, P = 0.276 after 20 mg simvastatin; r = −0.263, P = 0.152 after 40 mg simvastatin; and r = −0.366, P = 0.043 after 80 mg simvastatin) after therapies. There were correlations between percent changes in adiponectin levels and percent changes in QUICKI (r = 0.543, P = 0.001 after placebo; r = 0.366, P = 0.051 after 20 mg simvastatin; r = 0.441, P = 0.011 after 20 mg simvastatin; r = 0.234, P = 0.205 after 40 mg simvastatin; and r = 0.380, P = 0.035 after 80 mg simvastatin) after therapies.

We observed that simvastatin significantly improved endothelium-dependent dilation but reduced adiponectin levels and insulin sensitivity in hypercholesterolemic patients independent of dosage. Despite dose-dependent effects of simvastatin on apolipoprotein B reduction, the magnitude of these percent changes (FMD, adiponectin, and QUICKI) were not significantly different among four different doses of simvastatin.

Because of reciprocal relationships between endothelial dysfunction and insulin resistance (2,3), we hypothesized that improvements in endothelial dysfunction may be accompanied by simultaneous improvement in metabolic parameters. However, simvastatin significantly reduced adiponectin levels and insulin sensitivity despite improvement in endothelium-dependent dilation in hypercholesterolemic patients. Further, there were no significant correlations between endothelial dysfunction and metabolic parameters. By contrast, in previous studies with fenofibrate, ramipril, angiotensin II receptor blockers, or efonidipine, improvement in endothelial dysfunction was accompanied by simultaneous improvement in insulin sensitivity and increased adiponectin levels (15–19). Taken together, these results suggest that not all mechanisms for improving endothelial dysfunction are tightly coupled to metabolic homeostasis. Alternatively, potential improvements in insulin sensitivity and adiponectin levels caused by improvement in endothelial function after simvastatin therapy may be masked by other endothelial-independent effects of simvastatin that worsen insulin resistance and lower adiponectin levels.

Adiponectin is an adipose-derived factor that augments and mimics metabolic and vascular actions of insulin (14). In our study, simvastatin therapy significantly decreased adiponectin levels without a corresponding change in BMI. This finding is consistent with a recent article (21). Thus, it is possible that statin therapy is directly altering adiponectin levels independent of adiposity. In 3T3-L1 adipocytes pravastatin increases expression of adiponectin mRNA and enhances adiponectin secretion into conditioned media. These effects correspond to increased plasma levels of adiponectin and enhanced insulin sensitivity in C57BL/6J mice without changes in body weight. In contrast, simvastatin does not increase production of adiponectin in 3T3L1 adipocytes (22). Thus, the lipophilic status of particular statins may determine, in part, beneficial or detrimental effects on adiponectin levels and insulin sensitivity. Decreasing adiponectin levels is predicted to worsen insulin sensitivity by multiple mechanisms (14). In the current study, there were correlations between percent changes in adiponectin levels and percent changes in insulin or QUICKI after simvastatin therapy.

There may be additional mechanisms to reduce insulin sensitivity and adiponectin levels after simvastatin therapy. Lipophilic statins, particularly at high doses, may cause unfavorable pleiotropic effects such as reduction of insulin secretion and exacerbation of insulin resistance (7–9). Lipophilic statins inhibit the glucose-induced elevation of free Ca²⁺ level in cytoplasm, leading to suppressed insulin secretion. Glucose-induced elevation in the intracellular Ca²⁺ level is attributable to inflow of Ca²⁺ after activation of the L-type Ca²⁺ channel in β cells. Simvastatin suppresses glucose-induced elevation of intracellular Ca²⁺ level in a dose-dependent manner in an experiment using the patch-clamp method. When the influence of statins on glucose-stimulated insulin secretion was evaluated by direct measurement, simvastatin exerted significant suppression (7). The possibility of lipophilic statins reducing sensitivity to insulin is suggested by an experiment using rats with streptozotocin-induced diabetes. Impaired glucose tolerance is observed after 6 weeks of treatment with atorvastatin compared with that in the control group (9). Moreover, lovastatin treatment downregulates expression of the insulin-responsive glucose transporter GLUT4 and upregulates GLUT1 in 3T3-L1 adipocytes. These effects are associated with marked inhibition of insulin-stimulated glucose transport. Under these conditions, lovastatin had no effect on cell cholesterol levels, but its metabolic effects were reversed by mevalonate. This result suggests that inhibition of isoprenoid biosynthesis causes insulin resistance in 3T3-L1 adipocytes (8).

The effects of statins on adiponectin and insulin sensitivity in humans are controversial. Simvastatin and atorvastatin improve insulin sensitivity in some diabetic patients (4). However, others report that simvastatin either does not change or worsens insulin sensitivity in patients with metabolic syndrome (5) or type 2 diabetes (6). We reported that 20 mg simvastatin tended to reduce plasma levels of adiponectin and insulin sensitivity. However, these effects were not statistically significant. Simvastatin at 20 mg reduces plasma adiponectin levels from 4.5 (range 3.4–7.0) to 4.5 (2.9–6.0) μg/ml (P = 0.153) and QUICKI from 0.475 to 0.458 (P = 0.191) in hypercholesterolemic, hypertensive patients (18) and reduces plasma adiponectin levels from 3.8 (2.7–5.2) to 3.7 (2.2–5.1) μg/ml (P = 0.247) and QUICKI from 0.367 to 0.360 (P = 0.270) in patients with type 2 diabetes (19). In the current study, we observed that simvastatin significantly reduced adiponectin levels and insulin sensitivity in hypercholesterolemic patients independent of dose. Despite dose-dependent effects of simvastatin on apolipoprotein B reduction, the magnitude of these percent changes on adiponectin and QUICKI were not significantly different among different doses of simvastatin. The effects of simvastatin on glucose were not significant compared with placebo. However, the effects of simvastatin on insulin were significant compared with placebo. In the current study, the average BMI is larger than that in our previous studies (18,19). Thus, the metabolic effects of simvastatin may differ depending on BMI. Because measures of insulin resistance were considered secondary in the current study, we did not measure insulin sensitivity using the reference standard euglycemic glucose clamp technique. However, QUICKI, the surrogate measure of insulin sensitivity we used, is the most extensively validated and accurate surrogate index of insulin sensitivity currently available in humans. Indeed, QUICKI is the most appropriate surrogate measure of insulin sensitivity to use with our patient population and study design (20,23,24). In our current study we evaluated 2 months of therapy. However, others using longer durations of therapy have found similar results (6,21,25).

Simvastatin significantly increases serum insulin levels whereas a modified Mediterranean-type diet counteracts this effect of simvastatin (25). Indeed, recent large-scale clinical studies have demonstrated that lipophilic statins may increase rather than decrease the rate of onset of new diabetes (10–13). In the subgroup analysis of diabetic patients in the Anglo-Scandinavian Cardiac Outcomes Trial–Lipid Lowering Arm (ASCOT-LLA) study, the cumulative incidence of nonfatal myocardial infarction and fatal coronary heart disease was reduced in the 10 mg atorvastatin treatment group. However, onset of diabetes was seen more frequently in the atorvastatin treatment group than in the placebo group (12). Moreover, the onset of diabetes was seen more frequently in the 40 mg simvastatin treatment group than in the placebo group (11). A nested case-control study demonstrated adjusted odds ratios of 0.7 and 1.0 for pravastatin use alone and simvastatin use alone, respectively, compared with nonexposed individuals (10). Atorvastatin at 80 mg is associated with a statistically significant increase in the risk of developing A1C >6% both in nondiabetic subjects (adjusted hazard ratio [HR] 1.78) and in diabetic subjects (adjusted HR 2.36). The pooled adjusted HR was 1.84 (P < 0.0001) (13). These results emphasize the importance of combination therapy for patients at high risk or with diabetes. Such combinations may include the addition of thiazolidinediones (21), fenofibrate (15), ACE inhibitors (19), or angiotensin II type 1 receptor blockers (18) in addition to therapeutic lifestyle changes (25). Indeed, combination therapies improve both insulin resistance and endothelial function by multiple independent and interdependent mechanisms that improve the overall cardiometabolic profile to a greater extent than monotherapy (2,3,26).

We and others reported that simvastatin lowers CRP in hyperlipidemic coronary patients (27,28). In the current study, we observed that simvastatin tended to lower CRP in hyperlipidemic patients. However, these results did not achieve statistical significance. This may be due, in part, to very low baseline CRP levels. Interestingly, the magnitudes of percent changes in CRP levels were not significantly different among four different doses of simvastatin despite differential effects on apolipoprotein B reduction. Our observation is consistent with a previous study (28).

In summary, simvastatin significantly improved endothelium-dependent dilation, but reduced adiponectin levels and insulin sensitivity in hypercholesterolemic patients independent of dosage and the extent of apolipoprotein B reduction.

This study was partly supported by established investigator awards (2004-1 and 2005-1), Gachon University Gil Medical Center.

Parts of this study were presented in abstract form at the 57th Annual Scientific Session of the American College of Cardiology, Chicago, Illinois, 30 March 2008; at the American Heart Association Meeting, Orlando, Florida, 4–7 November 2007; and at the European Society of Cardiology Meeting, Vienna, Austria, 1–5 September 2007.