Proinflammatory cytokine changes in clinically stable, virologically suppressed, HIV-infected patients switching from protease inhibitors to abacavir.

To the Editor: Excess cardiovascular risk has been observed in HIV-infected patients, driven by traditional risk factors such as smoking, hypertension, diabetes, and aging, although therapy-induced hyperlipidemia and HIV-induced proinflammatory state and immune activation seem to play a role. A recent report from the Data Collection on Adverse Events of AntiHIV Drugs (DAD) Study, one of the largest cohorts evaluating cardiovascular risk in HIV-infected patients, assessed the associations between nucleoside analogues and myocardial infarction. Contrary to the hypothesis expecting a role for thymidine analogues because of their negative effects on plasma lipids, insulin sensitivity, and limb fat loss, DAD investigators unexpectedly found that the recent use of didanosine and abacavir increased the risk of myocardial infarction by 50% and 90%, respectively. These findings were even more surprising as abacavir has a neutral lipid profile, does not interfere with insulin sensitivity, is not associated with limb fat loss in naive patients, and even promotes a fat increase in patients with lipoatrophy switching to abacavir from thymidine analogues. Although the DAD investigators adjusted the effect of abacavir as much as possible, the possibility of bias cannot definitively be ruled out. However, if abacavir really plays a role in promoting cardiovascular disease, other mechanisms apart from lipid changes, insulin sensitivity, and body fat alterations would have to be involved. Several studies have shown that changes in inflammatory cytokines such as adiponectin and tumor necrosis factor (TNF)-alpha may be related to increased cardiovascular risk in the general population. More recently, this association has been suggested in HIV-infected patients, mainly those with uncontrolled viral replication before initiating or after discontinuing antiretroviral therapy. The Nevirapine, Efavirenz and Abacavir (NEFA) trial assessed the relative merits of abacavir, efavirenz, or nevirapine as substitutes for the protease inhibitor (PI) component in stable virologically suppressed patients treated with a first-generation, unboosted PI and willing to simplify their regimen. In an intensive metabolic substudy of NEFA, we found favorable lipid changes in patients switching to any of the 3 options, namely a decrease in total cholesterol in the abacavir group, an increase in HDL cholesterol, and a decrease in the ratio of total cholesterol (TC) and high-density lipoprotein (HDL) cholesterol (TC:HDL) in the nonnucleoside reverse transcriptase inhibitor (NNRTI) groups. In addition, adiponectin and soluble tumor necrosis factor alpha receptor type 2 (sTNFR2) were assessed in baseline and 96-week serum samples adequately stored at 80(C for further laboratory analysis, contemplated in the initial design before the substudy. Plasma adiponectin was determined by radioimmunoassay (Linco Research, St. Charles, MO) with a detection threshold of 1 ng/ mL and plasma sTNFR2 by a solid-phase enzyme immunoassay with amplified reactivity (Biosource Europe, Fleurus, Belgium) and a detection threshold of 0.1 ng/mL. In samples from 54 of 90 patients included in the metabolic substudy who continued with the assigned drug up to the end of follow-up (18 abacavir, 19 nevirapine, and 17 efavirenz), an increase in adiponectin levels and a decrease in sTNFR2 were observed in the overall population 96 weeks after the PI switch. We now present an analysis of the changes observed in these adipokines according to the antiretroviral regimen received in the NEFA metabolic substudy. Groups were comparable for baseline characteristics, including age {abacavir 38 years [IQR (interquartile range) 34–47], efavirenz 37 (35–45), and nevirapine 41 (37–46)}, proportion of men (abacavir 79%, efavirenz 84%, and nevirapine 83%), risk groups (drug users: abacavir 34%, efavirenz 41%, and nevirapine 41%; homosexuals: abacavir 21%, efavirenz 25%, and nevirapine 17%; heterosexuals: abacavir 38%, efavirenz 25%, and nevirapine 41%), AIDS (abacavir 34%, efavirenz 50%, and nevirapine 45%), and CD4 count [abacavir 456 cells/mL (IQR 308–728); efavirenz 549 (274–674), nevirapine 338 (249–552)], with the exception of stavudine use (abacavir 82.8%, efavirenz 62.1%, nevirapine 87.5%; P = 0.043). Although a trend to an increase in adiponectin levels was observed in patients switched to nevirapine [median (IQR) 5.5 (4.2–9.8) to 8.4 (4.9–10.7) mg/ mL; P = 0.1] and there were no changes in the efavirenz arm [4.4 (3.4–7.8) to 4.1 (2.8–7.8) mg/mL); P = 0.3], a significant decrease in adiponectin was seen in abacavir patients [8.7 (2.9–16.1) to 5.2 (2.6–7.7) mg/mL; P < 0.05] (Fig. 1A). Changes in adiponectin levels differed between the NNRTI and abacavir groups (P < 0.001). Conversely, a reduction in sTNFR2 was observed in the NNRTI arms, which was significant in the nevirapine group [6.8 (5.1–8.4) to 5.7 (4.2–7.5) ng/mL; P < 0.05] and efavirenz group [6.9 (5.6–7.9) to 6.5 (5.4–7.5) ng/mL;P= 0.3], whereas a nonsignificant increase was found in abacavir patients [5.2 (5.0–9.8) to 6.7 (5.3–8.9) ng/mL; P = 0.6] (Fig. 1B). After adjusting for several variables, including sex, age, Dtriglyceride, DTC/HDL, CD4, viral load, stavudine use, and lipodystrophy, in a multivariate linear regression model, abacavir remained significantly associated with a decrease in adiponectin levels (b = 6.001, P = 0.004) and a trend to an increase in sTNFR2 (b = 1.267, P = 0.146). Adiponectin is an adipokine with metabolic and anti-inflammatory effects. Hypoadiponectinemia is a significant predictor of endothelial dysfunction in peripheral vasculature and coronary