Efavirenz induces a striking and generalized increase of HDL-cholesterol in HIV-infected patients

Abstract

This study analyses in depth lipid and lipoprotein changes in 20 HIV-infected patients on an efavirenz-containing regimen, either as initial therapy or as a substitute for a protease inhibitor. Total plasma lipids and the distribution of subclasses of lipoproteins were analysed at baseline and after 48 weeks by nuclear magnetic resonance spectroscopy, and showed a mean increase of 20% in HDL-cholesterol in both regimens. Up to 85% of patients significantly improved their atherogenic index (LDL/HDL ratio). Highly active antiretroviral therapy (HAART) has been associated with lipid disturbances in HIV-infected patients, resulting in hypercholesterolemia, hypertriglyceridemia and low HDL levels, leading to an increase in cardiovascular risk. These abnormalities have been mostly attributed to the protease inhibitor (PI) antiretroviral class, appearing in more than half of the cases after 2 years of continuous therapy [1,2]. This dyslipidemia is currently treated with diet, exercise and statins, fibrates or a combination of both. Nevertheless, it has been proposed that switching the PI to a non-nucleoside transcriptase reverse inhibitor (NNRTI) would be of benefit in the control of hyperlipidemia. In fact, nevirapine has been shown to exert significant benefits on the lipoprotein profile when the PI is switched to this drug [3,4], whereas the second commercialized NNRTI, efavirenz, has achieved more controversial results in clinical trials [5]. The purpose of this trial was to analyse in detail lipid and lipoprotein changes in 20 HIV-infected patients after 48 weeks of an efavirenz-containing regimen, either as initial therapy or as a substitute for a PI. Therefore, patients were classified into two groups: antiretroviral-naive subjects (naive group; n = 10) and PI-treated subjects (PI group; n = 10). All patients in the naive group started didanosine and stavudine together with efavirenz, whereas individuals on PI continued the same nucleosides and replaced the PI by efavirenz. Seven out of 10 subjects (70%) in the PI group presented evident lipodystrophy at baseline. None of these subjects had taken lipid-lowering drugs during the study. In addition, all of them maintained the same antiretroviral regimen during the 48-week treatment. The distribution of the total plasma lipids and lipoprotein subclasses was analysed at baseline and after 48 weeks by nuclear magnetic resonance (NMR) spectroscopy (NMR LipoProfile; Raleigh, USA). The subclasses of lipoproteins quantified were six discrete subclasses of VLDL according to their increase in molecular weight, three subclasses of LDL, IDL, and HDL divided into five subclasses, as well as the determination of lipid content, the number of particles and the average particle size. The clinical details of patients and lipid changes between baseline and after 48 weeks on efavirenz treatment are shown in Table 1.Table 1: Baseline epidemilogical, clinical and laboratory aspects of patients in both groups.In both groups, 48-week treatment with efavirenz significantly increased HDL-cholesterol, from 38.5 (15.5) to 52.5 (16.0) mg/dl in the naive group [95% confidence interval (CI) for the increment, 8.4–19.4 mg/dl; P < 0.0001] and 41.8 (15.5) to 48.0 (14.6) mg/dl in the PI group (95% CI, 2.4–10.0 mg/dl; P = 0.005) (pair-wise Student's t-test) (Table 1). Moreover, when patients of both groups were studied individually it was observed that HDL was maintained or elevated in all 20 subjects regardless of plasma triglyceride variations. The HDL increase was confined to the HDL2 subfraction and its independence of plasma triglyceride changes was confirmed by multivariate analysis of variance. Regarding the prevalence of low HDL-cholesterol (< 40 mg/dl), it was present at the beginning of the study in 70% of patients in the naive group and 50% in the PI group. Fifty per cent of patients with low HDL-cholesterol at baseline increased their level above 40 mg/dl in the naive group. In the PI group, 90% of subjects increased their HDL-cholesterol levels and 20% of patients with low HDL-cholesterol at the start reached HDL-cholesterol levels above 40 mg/dl after treatment. Triglyceride levels significantly decreased only in the PI group, from 174.6 mg/dl (73) to 138.9 mg/dl (85) (95% CI for the difference −25.3–96.6 mg/dl; P = 0.019). No statistically significant changes were seen in total cholesterol or LDL-cholesterol levels. The HDL-cholesterol changes observed in all 20 subjects treated with efavirenz resulted in a strikingly significant reduction of the LDL/HDL cholesterol ratio from 3.34 to 2.91 (P = 0.028). The results clearly show that efavirenz has an HDL-raising effect in both situations, naive and PI-treated individuals. Moreover, HDL elevations occurred even in those patients who experienced an increase in plasma triglyceride levels, which suggests an HDL-specific action of this drug independent of triglyceride lipolisis. This effect is quite remarkable because HDL increased by 26.6% in the naive group and by 12.9% in the PI group. These elevations are superior compared with the average elevations reached with statins (2–8%) or with fibrates (up to 20%) in intervention studies that have obtained significant reductions in cardiovascular disease mortality and morbidity. In the PI-replaced group efavirenz showed a clear trend towards lower plasma triglyceride levels, although this effect is probably the result of PI withdrawal, and plasma total and LDL-cholesterol were maintained. However, the increase in the protective HDL fraction in both treatment groups significantly lowered the LDL/HDL cholesterol ratio, and therefore resulted in a better lipid profile. An improvement in the atherogenic index was observed in 85% of patients despite the LDL increase and regardless of triglyceride changes. The fact that all these effects are observed in both regimens suggests a direct action of this drug. This beneficial effect on the lipoprotein profile has also been described in nevirapine-treated individuals [3,4]. The mechanism behind this HDL-raising effect is not known, but its surprising independence of triglyceride changes suggests that it might involve some degree of interference with the process of triglyceride–cholesteryl ester exchange mediated by the cholesterol ester transfer protein, which is known to interact with water-insoluble drugs such as efavirenz. Therefore, it could be speculated that this protective hyperalphalipoproteinemic effect might be a class property, common to both efavirenz and nevirapine. As PI may have an opposite effect on the lipid profile by inducing insulin resistance and interfering with adipocyte differentiation and maturation, switching PI to NNRTI could be of interest to treat hyperlipidemia in this population. Further studies are needed to clarify the mechanism responsible for the significant increase in HDL-cholesterol.