Lipid Abnormalities in HIV-Infected Patients Are Not Correlated With Lopinavir Plasma Concentrations

To the Editor: The possibility of increasing protease inhibitor (PI) plasma concentration through association of subtherapeutic doses of ritonavir represents a significant advancement in antiretroviral therapy. In particular, a lopinavir/ritonavir (LPV/r) combination showed higher efficacy than nelfinavir in a prospective randomized trial, 1 and it is widely used in clinical practice. LPV/r use is frequently associated with a significant increase in triglyceride and/or cholesterol levels soon after starting therapy, however. 1 It is not clear whether dyslipidemia is a direct effect of the PI or a more complex scenario involving genetic, immunologic, and pharmacologic factors. Recently, it has been described that higher LPV concentrations are associated with higher risk of grade 3 concomitant elevations in total cholesterol and triglyceride levels. 2,3 Our research question was aimed at investigating whether lipid changes are associated with LPV plasma concentrations, taking into account several concomitant factors that also can modulate lipid levels. Patients enrolled in an ongoing trial of concentration-controlled therapeutic drug monitoring (TDM) in clinical practice (Resistance and Dosage Adapted Regimens [RADAR]) were selected for this analysis according to the following selection criteria: (1) LPV/r-containing regimens started, (2) triglyceride and total cholesterol levels available at baseline and at 2 points of the follow-up up to week 12, (3) no alcohol abuse admitted by patient interview and, (4) not assuming hypolipidemic drugs in the 6 months before and for the entire duration of the study. LPV/r was given at standard doses (400/100 mg twice daily or 533/133 mg twice daily according to the absence or presence of nonnucleoside reverse transcriptase inhibitors [NNRTIs] in the regimen, respectively). LPV trough plasma concentrations (defined as predose concentration before intake of the morning dose) were determined by a validated high performance liquid chromatography (HPLC) assay as previously described. 4 Plasma through concentrations were determined 1, 4, and 12 weeks after LPV/r treatment initiation, and in those cases in which they were under a cutoff of 4 μg/mL, prescribed doses of LPV/r were modified on the basis of expert advice. All the data are expressed as mean ± SD. Normal distribution was tested by the Shapiro-Wilke test. Comparison of total cholesterol and triglyceride serum levels between baseline and follow-up visit was done by the Student paired t test. Repeat-measure MANOVA was used to test for statistically significant changes over time for total cholesterol and triglyceride serum levels and to reveal correlations between LPV plasma concentration at each time point of the follow-up and the change over time for total cholesterol and triglyceride serum levels. The χ2 test was used to identify the association between LPV plasma concentration above a predefined cutoff and elevation of triglyceride and total cholesterol levels. Finally, a summary measure of LPV concentrations was calculated as the area under the curve (AUC) as well as total cholesterol and triglyceride level variations of the percentage change (Δ%AUC) during the period of observation. A simple linear regression analysis was carried out to identify factors predictive of change in total cholesterol or triglyceride serum levels over time, including the following variables: gender, LPV AUC, hepatitis C virus (HCV)–antibody (Ab) positivity, presence of stavudine in the current antiretroviral regimen, high total cholesterol baseline levels (>200 mg/dL), and high triglyceride baseline levels (>180 mg/dL). Analyses were performed with STATA (StataCorp 2000, release 7.0; College Station, TX) and with STATISTICA for Windows (StatSoft, 2000 STATISTICA for Windows [computer program manual], Tulsa, OK). Fifty-five patients with a mean age of 39 years (SD = 7) were included in this study. All patients were white, 29 (64.44%) were men, and 25 (55.56%) had drug addiction as a risk factor for HIV acquisition. The rates of HCV and hepatitis B virus coinfections were 29 of 45 (64.44%) and 2 of 45 (4.44%) patients, respectively. At any point of the follow-up, LPV plasma concentrations were highly variable, ranging from undetectable to 16 μg/mL. Mean LPV plasma concentrations (SD) at weeks 1, 4, and 12 of follow-up were as follows, respectively: 7.16 μg/mL (3.35), 6.73 μg/mL (3.51), and 6.10 μg/mL (3.32). Mean triglyceride levels (SD) at weeks 1, 4, and 12 of follow-up were as follows, respectively: 168 mg/dL (97), 256 mg/dL (196), and 276 mg/dL (234). Variations in triglyceride levels were significant at week 4 (P = 0.006) and week 12 (P = 0.005) with respect to baseline. A mean of 8.5-mg/dL increase in triglyceride levels for each week of follow-up has been found (95% confidence interval [CI]: 1.7–15.2; P = 0.014). By contrast, total cholesterol levels did not change significantly (from 185 mg/dL [53] at baseline to 206 mg/dL [59]; P = 0.09). Linear regression analysis adjusted for time (MANOVA) revealed lack of an association between LPV plasma concentrations and changes over time for either total cholesterol or triglyceride serum levels. Moreover, no association between LPV plasma concentrations above a predefined cutoff and elevation of triglyceride and total cholesterol levels has been found. For instance, using an LPV plasma concentration of >8 μg/mL and concomitant grade 3 hyperlipidemia or higher (triglycerides ≥750 mg/dL and total cholesterol ≥300 mg/dL) as cutoff values, 3 although 18 (40%), 15 (33.33%), and 7 (15.56%) patients (at weeks 1, 4, and 12, respectively) reached a LPV plasma concentration higher than the proposed cutoff, none of them developed grade 3 hyperlipidemia or higher. Six patients had grade 2 hyperlipidemia (triglycerides ≥400 mg/dL and total cholesterol ≥240 mg/dL) at least once over the follow-up period, and these cases were not related to LPV plasma concentrations. Finally, none of the variables that were analyzed were associated with either a triglyceride or total cholesterol increase above predefined cutoff levels. To more fully understand the possible relationships between overall LPV drug exposure and variation of lipid levels as continuous measures over time, AUC values were used as synthesis measures. With this approach, we were not able to demonstrate statistically significant correlations of LPV AUC with either total cholesterol Δ%AUC (r = 0.1, P = 0.49) or Δ%AUC of triglyceride levels (r = 0.01, P = 0.092). Univariate simple regression analysis demonstrated a significant negative association between baseline normal cholesterol level (≤200 mg/dL) and total cholesterol Δ%AUC increase during the follow-up period (coefficient: −27.06, 95% CI: −48.10 to −6.02; P = 0.013). The LPV/r combination is able to suppress HIV replication durably in naive patients, increased LPV exposure obtained in vivo through the combination with ritonavir may also overcome a certain grade of resistance but not all, and the ability of LPV to inhibit HIV-1 from patients failing other PIs depends on the plasma concentration reached and the number of mutations that have accumulated. 5 Recent studies have described concentration-dependent relationships between LPV and total cholesterol and triglyceride levels. 2,3 Based on these data, it has been suggested that the LPV/r dosage should be reduced in an attempt to lower lipid elevation. A limitation to all these studies is that none of the other variables possibly implicated have been assessed. In fact, stavudine use in the current antiretroviral regimen, baseline lipid levels, and HCV coinfection 6,7 may also have an impact and may therefore act as confounding factors. By means of analysis using a design identical to that of Gutierrez et al, 2 employing LPV and lipid concentrations as categoric variables, we were not able to confirm their results in our cohort. Moreover, a more stringent analysis using AUC value as an indicator of overall exposure to LPV and lipid variations over time provided confirmation. Interestingly, when the patient population is divided according to HCV-Ab reactivity, this factor did not appear to be correlated with protection against an increase in cholesterol in contrast to previous observations 6,7 in patients not exposed to LPV/r-containing regimens. Therefore, the present data suggest that the influence of LPV on lipid variations is caused by an intrinsic pharmacodynamic drug effect independent of pharmacokinetic variability. Also, this effect appeared to be particularly intense on triglyceride levels; however, it was able to overcome the HCV protective effect on cholesterol increase. Taken together, these results may indicate that it is premature to suggest an LPV/r dose reduction based on plasma drug concentrations to prevent or treat abnormal lipid elevations. Further studies are necessary to confirm these results and to clarify further the biologic mechanisms for these findings. ACKNOWLEDGMENTS The following are members of the RADAR study group: G. Carosi, G. Cadeo, F. Castelnuovo, D. Bella, F. Moretti, P. Nasta, A. Patroni, E. Quiros-Roldan, and V. Tirelli (Brescia); S. Lo Caputo, P. Pierotti, and F. Mazzotta (Florence 1); F. Leoncini and G. Sterrantino (Florence 2); A. De Luca, S. Di Giambenedetto, and R. Cauda (Rome 1); A. Antinori and D. Zinzi (Rome 2); N. Ladisa, P. Maggi, L. Monno, and G. Pastore (Bari); A. Cargnel and V. Micheli (Milan); A. Orani and P. Perini (Lecco); D. Santoro and R. Signori (Como); L. Minoli, P. Sacchi, and C. Zocchetti (Pavia); F. Ghinelli and L. Sighinolfi (Ferrara); M. Regazzi-Bonora and P. Villani (Clinical Pharmacology Unit, IRCCS S. Matteo, Pavia); N. Manca and F. Gargiulo (Institute of Microbiology, University of Brescia, Brescia) and C. Tinelli (Statistical Unit, IRCCS S. Matteo, Pavia). Carlo Torti, MD Eugenia Quiros-Roldan, MD,PhD Mario Regazzi-Bonora, MD Andrea De Luca, MD Sergio Lo Caputo, MD Simona Di Giambenedetto, MD Andrea Patroni, MD Paola Villani, BSc Valeria Micheli, BSc Giampiero Carosi, MD RADAR (Resistance and Dosage Adapted Regimens) Study Group of the MASTER Cohort