Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Effects of First-Line Use of Nucleoside Analogues, Efavirenz, and Ritonavir-Boosted Protease Inhibitors on Lipid Levels

Effects of First-Line Use of Nucleoside Analogues, Efavirenz, and Ritonavir-Boosted Protease... Effects of First-Line Use of Nucleoside Analogues, Efavirenz, and Ritonavir-Boosted Protease Inhibitors on Lipid Levels 1 2 3 Andrew Hill, Will Sawyer, and Brian Gazzard 1 2 Pharmacology Research Laboratories, University of Liverpool, Liverpool, United Kingdom; MetaVirology Ltd, London, United Kingdom; Chelsea and Westminster Hospital, London, United Kingdom Introduction: Treatment with ritonavir-boosted protease inhibitors or efavirenz and nucleoside analogues leads to rises in lipids, which might contribute to cardiovas- cular risk. Method: A MEDLINE search for clinical trials of fi rst-line HAART with standardized 48-week lipids data available for combinations of two nucleoside analogues identifi ed 13 with additional boosted protease inhibitor (PI/r) regimens (n = 5,281) and two with efavirenz (n = 1,087). Inverse-variance weighting was used to provide estimates of combined 48-week elevations in each lipid parameter. Results: The trials were well balanced for mean baseline total cholesterol (155 mg/dL), trig- lycerides (125 mg/dL), LDL (95 mg/dL), and HDL (37 mg/dL). The PIs showed two dif- ferent types of lipid elevations: Group 1: saquinavir/r, atazanavir/r, and darunavir/r; and Group 2: lopinavir/ritonavir and fosamprenavir/r. There were greater elevations in total cholesterol and triglycerides for Group 2 compared to Group 1 but no differ- ences in LDL or HDL between Group 1 and Group 2. Patients treated with efavirenz showed similar rises in total cholesterol and LDL compared with the PIs in Group 2 but showed smaller rises than in Group 1. In addition, patients treated with abacavir/ lamivudine, zidovudine/lamiuvudine, or stavudine/lamivudine showed signifi cantly higher elevations in all four lipid parameters, compared with patients given tenofovir/ emtricitabine. Conclusion: There is a wide range of lipid elevations during 48 weeks of fi rst-line HAART, which depends on the choice of antiretrovirals used. Key words: cholesterol, HIV, lipids, meta-analysis, nucleoside analogues, protease inhibitors here is evidence from cohort studies that high baseline lipids. In addition, the population of antiretroviral treatment might increase the treated patients is ageing over time, with median 1,2 Trisk of cardiovascular disease, although the age at entry to trials in the late 30s and early 40s. In SMART study showed an increased cardiovascular one large US cohort study of managed care, rates of risk from stopping antiretrovirals. International hospitalization for myocardial infarction were sig- treatment guidelines recommend fi rst-line anti- nifi cantly higher for HIV-infected people relative to retroviral treatment with the combination of uninfected people. two nucleoside analogues plus either the non- When antiretroviral treatments have been com- nucleoside efavirenz or a ritonavir-boosted pro- pared directly in randomized clinical trials of naïve tease inhibitor. and pretreated patients, the mean lipid profi les Antiretroviral treatment can raise levels of have differed both between and within antiret- lipids, and elevated levels of certain lipid param- roviral drug classes. Ritonavir-boosted protease eters are known to elevate cardiovascular risk inhibitors (PIs) tend to show greater rises in lipids 5 10,11 for HIV-infected people. Therefore, the effect of than unboosted PIs. Of the ritonavir-boosted antiretrovirals on lipids might be an indicator of their associated cardiovascular risk. Some cardio- Address for correspondence: Andrew Hill, Pharmacology vascular risk equations include a combination of Research Laboratories, University of Liverpool, 70 Pembroke high-density lipoprotein (HDL) and total choles- Place, Liverpool L69 3GF, United Kingdom. Email: microhaart@ terol, whereas others use a combination of HDL, aol.com 6,7 low-density lipoprotein (LDL), and triglycerides. HIV Clin Trials 2009;10(1):1–12 People infected with HIV in Western countries © 2009 Thomas Land Publishers, Inc. www.thomasland.com tend to have a high incidence of pre-existing risk factors, such as a high incidence of smoking or doi: 10.1310/hct1001-1 1 2 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 PIs, lopinavir/ritonavir has caused greater rises IAS Pathogenesis Conference), and International in total cholesterol and triglycerides than either Conference on Drug Therapy in HIV Infection. 12,13 darunavir/ritonavir or saquinavir/ritonavir. To be included in the analysis, clinical trials By contrast, the KLEAN trial showed similar lipid needed to have (a) at least 50 antiretroviral-naïve profi les for fosamprenavir/ritonavir and lopina- patients studied for at least 48 weeks, (b) use of a vir/ritonavir. The choice of nucleoside analogues ritonavir-boosted PI (using a total daily ritonavir or non-nucleosides used can also affect lipid pro- dose of 100–200 mg) or efavirenz in combination 15–20 fi les. Efavirenz tends to cause greater rises in with a fi xed pair of nucleoside analogues, and lipids than either nevirapine or the investigational (c) data available on mean or median levels of total non-nucleoside reverse transcriptase inhibitor cholesterol, triglycerides, LDL, and HDL at base- (NNRTI) TMC278 (rilpivirine). In several studies, line and at Week 48. the nucleoside analogues abacavir, stavudine, or For each of the 15 trials included in the meta- zidovudine have caused greater rises in lipids than analysis, the mean levels of total cholesterol, trig- 15–17 tenofovir. lycerides, LDL, and HDL were recorded at baseline Given the wide range of different measures used, and after 48 weeks of treatment. We used the stan- it can be diffi cult for clinicians to compare lipid dard deviation of the mean lipid parameters from elevations from reports of different clinical trials, the two arms of the ARTEMIS trial to estimate the including either the same or different treatments. 95% confi dence intervals (CI) of the mean lipid In clinical trials of pretreated patients, changes in parameters for the other trials when estimates of lipids from baseline can also depend on the antiret- variability were unavailable. Although data on rovirals taken in screening and whether these were mean lipid levels were available for most trials, stopped or continued into the randomized phase of there were few standardized data on the percentage the trials. There also tends to be more use of lipid- of patients with lipid elevations above predefi ned lowering drugs in trials of pretreated patients, and cutoff levels, for example, the National Cholesterol these can be continued or stopped either at baseline Evaluation Program (NCEP) thresholds. or during the trial, with no consistent methods to The change from baseline in each lipid parameter adjust for this in the analysis. was calculated as the mean 48-week value minus the baseline value. The median and mean rises of each lipid parameter were found to be similar METHOD overall. The summary changes for each trial were The aim of this analysis was to estimate the effect presented graphically, with a combined estimate of nucleoside analogues, ritonavir-boosted PIs, and of change from all trials of each boosted PI and of efavirenz on lipids in naïve patients, using a stan- efavirenz. For these combined analyses, inverse- dardized format to report lipid elevations. A similar variance weighting was used. The changes in lipids approach was used in an overview of HIV RNA sup- were presented by trial and treatment arm with pression rates across clinical trials in naïve patients. associated 95% CIs. The trials were divided by the We conducted a search for prospective clini- use of different nucleoside reverse transcriptase cal trials of ritonavir-boosted PI treatment in inhibitor (NRTI) backbones (tenofovir/emtricit- antiretroviral-naïve patients, using MEDLINE and abine vs. either abavavir, stavudine, or zidovudine separate searches of PI drug labels and conference used with lamivudine). In separate meta-regression abstracts. The online literature searches used the analyses for each lipid parameter, the relative names of each antiretroviral, followed by “clini- effects of PIs or efavirenz versus the nucleoside cal trial.” In addition, we searched the US Food analogues were assessed. Models included base- and Drug Administration (FDA) product labels line measures for each study arm of mean age, for registration trials of each approved antiretro- proportion of males, proportion of Caucasian race, viral and for abstracts on clinical trials presented mean CD4 count, and mean HIV RNA, together at the following conferences: Annual Conference with three binary variables for efavirenz, PI group, Retroviruses and Opportunistic Infections, Inter- and NRTI group. A backwards stepwise regression national Conference on Antimicrobial Agents and method retaining the treatment group variables Chemotherapy (ICAAC), European AIDS Clinical was used to reduce the variables in the models. Society, International AIDS Conference (including All analyses were performed using SAS version EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 3 9.1 (SAS Institute, Cary, North Carolina, USA). Table 1. ACTG 5142 trial: median percent change Predicted values based on these models were used from baseline to Week 96 in total cholesterol and as estimates of mean lipid levels within each com- triglycerides bination of PI, efavirenz, and NRTI group. Treatment arm 2NRTI + 2NRTI + EFV + RESULTS EFV LPV/r LPV/r The search identifi ed 15 clinical trials with a n = 169 n = 158 n = 168 total of 6,368 patients. Thirteen trials included Total cholesterol +33% +32% +57% two nucleoside analogues and a ritonavir-boosted Triglycerides +19% +46% +62% 12 13 14 PI. Five trials (GEMINI, ARTEMIS, KLEAN, 23 24 ALERT, CASTLE ) were direct head-to-head Note: NRTI = nucleoside reverse transcriptase inhibitor; comparisons of ritonavir-boosted PIs. Three trials EFV = efavirenz; LPV/r = ritonavir-boosted lopinavir. 10 11 25 (BMS-089, Abbott 863, and SOLO ) compared ritonavir-boosted PIs with unboosted PIs. The 26 27 Abbott 418 and Abbott 730 trials were compari- once daily, or 1400/100 mg once daily; and 3,148 sons of twice-daily versus once-daily lopinavir/ patients taking lopinavir/ritonavir either at the ritonavir. The HEAT trial evaluated lopinavir/ 400/100 mg bid dose or the 800/200 mg once-daily ritonavir with two different NRTI backbones. The dose. All of the data on lopinavir/ritonavir were REDUCE trial compared two boosting doses of for patients using the soft-gelatin capsule formu- ritonavir in combination with fosamprenavir once lation, except for the Abbott 730 trial, where the daily. Finally the SHARE trial was an open-label tablet formulation was used. In addition, data were evaluation of fosamprenavir/ritonavir once daily. available for 1,087 patients taking efavirenz at the For the Abbott 863 trial, data on total cholesterol standard 600 mg once-daily dose. and triglycerides were included, but the data Overall, the NRTI backbone used was tenofovir/ on HDL and LDL were not available. Two trials emtricitabine for 3,698 patients, abacavir/lamvu- included two nucleoside analogues and efavirenz dine for 1,705 patients, stavudine/lamivudine for 15,16 (Gilead 903 and Gilead 934). 722 patients, and zidovudine/lamivudine for 243 Two large trials of fi rst-line ritonavir-boosted PIs patients. The mean baseline CD4 count ranged from could not be included. The ACTG 5142 trial had 153 to 279 cells/µL, whereas mean baseline HIV been conducted in naïve patients but was excluded RNA levels ranged from 4.6 to 5.2 log copies/mL. from this analysis as a range of NRTI backbones The predominantly male patients had mean ages at had been used and only 96-week data were avail- baseline of between 36 and 40 and between 44%–78% able. However, summary data from ACTG 5142 were Caucasian. The majority of patients in each were used to cross-validate the results from this trial showed HIV RNA reductions below 50 copies/ analysis and are shown in Table 1. In addition, mL in intent-to-treat (ITT) analysis (Table 2). the BI 1182.33 trial (tipranavir/ritonavir vs. lopi- Mean levels of the four lipid parameters at base- navir/ritonavir in naïve patients) was excluded line and Week 48 are shown in Table 3. At base- because data on mean levels of lipids were not line, the mean levels of each lipid parameter were available. There were several trials of efavirenz- similar across the trials, with mean baseline levels based HAART that could not be included because as follows: total cholesterol, mean 155 (range, 147– lipid data had only been reported in terms of 168 mg/dL); triglycerides, mean 125 (range, graded toxicities and not as mean values. 105–166 mg/dL); LDL, mean 95 mg/dL (range, 89– Table 2 shows a summary of the 26 arms from 104 mg/dL); and HDL, mean 37 mg/dL (range, the 15 clinical trials included in the analysis. Data 31–42 mg/dL). Use of lipid-lowering drugs was were available for 166 patients taking saquinavir/ reported in only four of the trials. The percentage ritonavir 1000/100 mg bid; 343 patients on daruna- of patients using lipid-lowering drugs was 4.5% in 30 14 vir/ritonavir 800/100 mg once daily; 700 patients the SHARE trial, 11% in the KLEAN trial, 5% in 12 26 taking atazanavir/ritonavir 300/100 mg once the GEMINI trial, and 2.6% in the ABT-418 trial. daily; 924 patients taking fosamprenavir/ritonavir In the regression analyses, there were no sig- at doses of either 700/100 mg bid, 1400/200 mg nifi cant differences in lipid elevations between 4 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 Table 2. Baseline characteristics of trials included in analysis Baseline data and summary 48-week effi cacy CD4 count HIV RNA (log HIV RNA <50 Trial [ref] N Age %Male %Caucasian (cells/µL) copies/mL) Week 48 GEMINI [12] SQV/r TDF/FTC 166 38 81% 49% 156 5.2 65% ARTEMIS [13] DRV/r TDF/FTC 343 36 70% 40% 228 4.9 84% ALERT [23] ATV/r TDF/3TC 53 40 89% 49% 188 4.9 83% SHARE [29] ATV/r ABC/3TC 111 38 91% 54% 208 5.1 77% BMS-089 [10] ATV/r d4T/3TC 95 nd 73% 53% 201 4.8 75% CASTLE ATV/r TDF/FTC 440 34 69% nd 205 5.0 78% ALERT [23] FPV/r TDF/3TC 53 40 79% 64% 161 4.9 75% SOLO [24] FPV/r ABC/3TC 322 36 70% 51% 166 4.8 69% KLEAN [14] FPV/r ABC/3TC 434 38 78% 61% 188 5.1 66% REDUCE [28] FPV/r ABC/3TC 57 40 81% 46% 179 4.9 63% REDUCE [28] FPV/r ABC/3TC 58 39 81% 36% 259 4.7 67% ARTEMIS [13] LPV/r TDF/FTC 346 35 70% 44% 218 4.8 78% GEMINI [12] LPV/r TDF/FTC 171 37 77% 44% 153 5.2 64% ABT 418 [25] LPV/r TDF/FTC 75 39 81% 56% 214 4.8 64% ABT 418 [25] LPV/r TDF/FTC 115 38 79% 51% 232 4.6 70% ABT 730 [26] LPV/r TDF/FTC 333 nd 80% 78% 216 4.9 77% ABT 730 [26] LPV/r TDF/FTC 331 nd 77% 73% 215 5.1 76% HEAT [27] LPV/r TDF/FTC 286 39 80% 50% 193 4.8 67% CASTLE LPV/r TDF/FTC 443 36 69% nd 204 5.0 76% HEAT [27] LPV/r ABC/3TC 278 38 84% 52% 214 4.9 68% KLEAN [14] LPV/r ABC/3TC 444 37 78% 56% 194 5.1 65% ABT 863 [11] LPV/r d4T/3TC 326 38 80% 56% 232 5.0 67% Gilead-903 EFV TDF/3TC 299 36 76% 64% 279 4.9 82% Gilead-934 EFV TDF/FTC 244 38 86% 59% 245 5.0 80% Gilead-903 EFV d4T/3TC 301 36 76% 64% 279 4.9 81% Gilead-934 EFV ZDV/3TC 243 38 86% 59% 245 5.0 71% Note: SQV/r = ritonavir-boosted saquinavir; TDF = tenofovir; FTC = emtricitabine; DRV/r = ritonavir-boosted darunavir; ATV/r = ritonavir-boosted atazanavir; ABC = abacavir; 3TC = lamivudine; d4T = stavudine; FPV/r = ritonavir-boosted fos-amprenavir; LPV/r = ritonavir-boosted lopinavir; EFV = efavirenz; AZT = zidovudine. ritonavir-boosted fos-amprenavir (fAPV/r) using only 100 mg ritonavir per day. LPV/r 800/200 mg once daily. All other trials used LPV/r dosed at 400/100 mg bid. saquinavir, darunavir, or atazanavir when boosted In almost all cases, none of the baseline covari- with ritonavir, and so these PIs were grouped ates in the meta-regression analyses were retained (Group 1) to improve the power of the analyses. in the models after the backwards stepwise pro- Likewise, there were also no signifi cant differences cedures. Predicted values of the mean lipid levels in lipid elevations between fosamprenavir/rito- based on PI and NRTI grouping are summarized navir and lopinavir/ritonavir and these too were in Table 4. grouped together (Group 2). For the nucleoside Use of atazanavir/r, darunavir/r, or saquinavir/r analogues, there was no signifi cant difference in (PI Group 1) led to smaller rises in total cholesterol lipid elevations seen between abacavir, zidovu- than use of lopinavir/r and fosamprenavir/r (PI dine, or stavudine when given with lamivudine. Group 2) (p = .0390). The rises in total cholesterol Figures 1A to 1D show the mean rises in the four for PI Group 1 were comparable for trials using efa- lipid parameters, for the combined data for each virenz (p = .154). The use of stavudine, zidovudine, ritonavir-boosted PI and for efavirenz, stratifi ed for or abacavir as NRTIs also led to higher elevations the use of either tenofovir/emtricitabine or other in total cholesterol relative to the use of tenofovir nucleoside analogues. (p < .0001). EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 5 Table 3. Rises in mean lipid levels from baseline to Week 48 during clinical trials of boosted PIs or efarvirenz in naïve patients (mg/dL) Trial PI/r NRTI TCHOL TRIG LDL HDL GEMINI SQV/r TDF/FTC 157 → 183 (+26) 117 → 131 (+14) 89 → 107 (+18) 35 → 44 (+9) ARTEMIS DRV/r TDF/FTC 156 → 177 (+21) 131 → 143 (+12) 92 → 105 (+13) 39 → 44 (+5) ALERT ATV/r TDF/3TC 153 → 181 (+28) 123 → 133 (+10) 97 → 102 (+5) 38 → 48 (+10) CASTLE ATV/r TDF/FTC 147 → 164 (+17) 110 → 124 (+17) 91 → 102 (+11) 36 → 45 (+9) SHARE ATV/r ABC/3TC 165 → 198 (+33) 120 → 174 (+54) 103 → 115 (+12) 31 → 46 (+15) BMS-089 ATV/r d4T/3TC 158 → 183 (+25) 139 → 174 (+35) 96 → 119 (+23) 38 → 49 (+11) ALERT FPV/r TDF/3TC 160 → 179 (+19) 120 → 167 (+47) 95 → 99 (+4) 35 → 43 (+8) SOLO FPV/r ABC/3TC 162 → 214 (+52) 125 → 190 (+65) 96 → 123 (+27) 36 → 44 (+8) KLEAN FPV/r ABC/3TC 156 → 217 (+61) 112 → 179 (+67) 96 → 124 (+28) 33 → 46 (+13) REDUCE FPV/r ABC/3TC 166 → 197 (+31) 120 → 149 (+29) 104 → 108 (+4) 39 → 50 (+11) REDUCE FPV/r ABC/3TC 166 → 209 (+43) 120 → 172 (+52) 101 → 126 (+25) 37 → 46 (+9) ARTEMIS LPV/r TDF/FTC 156 → 188 (+32) 123 → 192 (+69) 93 → 105 (+12) 40 → 47 (+7) GEMINI LPV/r TDF/FTC 152 → 183 (+31) 117 → 172 (+55) 89 → 105 (+16) 38 → 49 (+11) ABT418 LPV/r TDF/FTC 168 → 195 (+27) 136 → 212 (+76) 102 → 115 (+13) 42 → 48 (+6) ABT418 LPV/r TDF/FTC 159 → 186 (+27) 137 → 219 (+82) 92 → 106 (+14) 40 → 43 (+3) ABT-730 LPV/r TDF/FTC 155 → 184 (+29) 151 → 196 (+45) 98 → 105 (+7) 40 → 45 (+5) ABT-730 LPV/r TDF/FTC 154 → 189 (+35) 156 → 220 (+64) 95 → 102 (+7) 39 → 48 (+9) HEAT LPV/r TDF/FTC 159 → 184 (+23) 134 → 189 (+38) 92 → 97 (+5) 35 → 47 (+12) CASTLE LPV/r TDF/FTC 147 → 185 (+38) 110 → 168 (+58) 91 → 108 (+17) 35 → 46 (+11) HEAT LPV/r ABC/3TC 159 → 202 (+43) 122 → 215 (+93) 93 → 105 (+12) 36 → 48 (+12) KLEAN LPV/r ABC/3TC 157 → 210 (+53) 117 → 195 (+78) 97 → 120 (+23) 34 → 48 (+14) ABT-863 LPV/r d4T/3TC 158 → 211 (+53) 166 → 291 (+125) — — Gilead-903 EFV TDF/3TC 150 → 175 (+25) 105 → 113 (+8) 94 → 106 (+12) 38 → 44 (+6) Gilead-934 EFV TDF/FTC 160 → 180 (+20) 105 → 113 (+8) 100 → 110 (+10) 36 → 42 (+6) Gilead-903 EFV d4T/3TC 149 → 204 (+55) 129 → 183 (+54) 96 → 122 (+18) 37 → 45 (+8) Gilead-934 EFV ZDV/3TC 157 → 191 (+34) 122 → 136 (+14) 98 → 116 (+18) 36 → 45 (+9) Note: Numbers in parentheses show absolute change from baseline to Week 48. PI/r = ritonavir-boosted protease inhibitor; NRTI = nucleoside reverse transcriptase inhibitor; TCHOL = total cholesterol; TRIG = triglycerides; LDL = low-density lipoprotein; HDL = high-density lipoprotein; SQV/r = ritonavir-boosted saquinavir; TDF = tenofovir; FTC = emtricitabine; DRV/r = ritonavir-boosted darunavir; ATV/r = ritonavir-boosted atazanavir; ABC = abacavir; 3TC = lamivudine; FPV/r = ritonavir-boosted fos-amprenavir; LPV/r = ritonavir-boosted lopinavir; d4T = stavudine; EFV = efavirenz; ZDV = zidovudine. fAPV/r using 100 mg ritonavir per day. All other trials used fAPV/r 1400/200 mg once daily or 700/100 mg bid (KLEAN). LPV/r 800/200 mg once daily. All other trials used LPV/r dosed at 400/100 mg bid. For triglycerides, the mean rise from use of rise in LDL (10.0 mg/dL; 95% CI 4–17; p = .001) for atazanavir/r, darunavir/r, or saquinavir/r was patients taking abacavir/lamivudine, stavudine/ 39 mg/dL (95% CI 33–44) lower than for trials of lamivudine, or zidovudine/lamivudine relative to lopinavir/r and fosamprenavir/r (p < .0001) and was those using tenofovir/emtricitabine (Table 4). 10 mg/dL (95% CI 3–16) higher than for trials with For HDL, there was no evidence of an effect of efavirenz (p = .005). The mean rise in triglycerides was antiretroviral treatment on absolute changes from 23 mg/dL higher (95% CI 19–27; p < .001) for patients baseline. There was evidence to suggest that per- receiving either abacavir/lamivudine, stavudine/ centage change from baseline was less for those lamivudine, or zidovudine/lamivudine compared studies with efavirenz as opposed to a boosted PI to those receiving tenofovir/emtricitabine. (−10%; 95% CI −20% to −1%; p = .032) and greater For LDL, there was no evidence for differences for those using tenofovir as opposed to other between the boosted PIs or with efavirenz. The NRTIs (11%; 95% CI 4% to 15%; p < .001). confi dence intervals around estimates of change Table 4 shows summary results on lipid elevat- were wide for all of the PIs. There was a greater ions from the ACTG 5142 study , which randomized 6 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n = 1048) EFV TDF (n = 543) non-TDF (n = 544) 0 102030 40506070 mean change from baseline to week 48 (mg/dL) SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n =1048) EFV TDF (n = 543) non-TDF (n = 544) –20 0 20 40 60 80 100 120 mean change from baseline to week 48 (mg/dL) Figure 1. Mean absolute change from baseline to Week 48 (±95% confi dence intervals). (A) Total cholesterol (mg/dL); (B) triglycerides (mg/dL); (C) LDL (mg/dL); (D) HDL (mg/dL). SQV/rtv = ritonavir-boosted saquinavir; TDF = tenofovir; DRV/rtv = ritonavir-boosted darunavir; ATV/rtv = ritonavir-boosted atazanavir; fAPV/rtv = ritonavir-boosted fos-amprenavir; LPV/rtv = ritonavir-boosted lopinavir; EFV = efavirenz. EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 7 SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n = 722) EFV TDF (n = 543) non-TDF (n = 544) –10 –5 0 5 10 15 20 25 30 35 mean change from baseline to week 48 (mg/dL) SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n = 722) EFV TDF (n = 543) non-TDF (n = 544) 0 2 4 6 8 10 1214 1618 mean change from baseline to week 48 (mg/dL) Figure 1. (Continued). 753 naïve patients to three arms: (a) two nucleoside ritonavir caused similar elevations in total cho- analogues plus efavirenz, (b) two nucleoside ana- lesterol to Week 96, whereas the combination arm logues plus lopinavir/ritonavir, and (c) efavirenz of efavirenz plus lopinavir/ritonavir caused the plus lopinavir/ritonavir. In this study, the arms greatest rises. Elevations in triglycerides were of 2 NRTIs + efavirenz and 2 NRTIs + lopinavir/ greater for 2 NRTIs + lopinavir/ritonavir than for 8 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 Table 4. Meta-regression analysis of absolute and percentage changes in lipids from baseline to Week 48 ATV/r, SQV/r, or DRV/r LPV/r or fAPV/r EFV ABC/3TC or ABC/3TC or d4T/3TC or TDF/FTC d4T/3TC TDF/FTC d4T/3TC TDF/FTC ZDV/3TC Mean absolute change (mg/dL) Total cholesterol +19 (9–27) +38 (25–49) +32 (26–38) +52 (44–59) +25 (15–34) +44 (36–54) Triglycerides +20 (15–25) +44 (38–49) +59 (55–62) +82 (78–86) +10 (6–15) +34 (28–39) LDL +11 (4–18) +22 (13–30) +11 (6–16) +22 (17–27) +11 (4–19) +22 (15–29) HDL +9 (2–16) +12 (4–20) +8 (3–13) +12 (5–19) +5 (–2–14) +9 (1–17) Mean percentage change (%) Total cholesterol +12% (10–14%) +24% (22–27%) +21% (19–22%) +33% (31–34%) +16% (14–19%) +29% (26–31%) Triglycerides +17% (11–23%) +35% (28–42%) +46% (42–50%) +64% (60–68%) +9% (2–16%) +27% (20–34%) LDL +12% (11–13%) +23% (22–24%) +12% (11–13%) +23% (22–24%) +12% (10–13%) +23% (21–24%) HDL +24% (18–31%) +35% (27–43%) +23% (9–28%) +34% (13–34%) +14% (6–22%) +25% (18–34%) Note: Effects of boosted protease inhibitors (PIs), efavirenz, and nucleoside reverse transcriptase inhibitors (NRTIs) summarized using predicted values (95% confi - dence intervals in parentheses). Conversion factors from mg/dL to mmol/L: Total cholesterol, LDL, and HDL: mg/dL/38.6 = mmol/L. Triglycerides: mg/dL/88.5 = mmol/L. ATV/r = ritonavir-boosted atazanavir; SQV/r = ritonavir-boosted saquinavir; DRV/r = ritonavir-boosted darunavir; LPV/r = ritonavir-boosted lopinavir; fAPV/r = ritonavir-boosted fos-amprenavir; EFV = efavirenz; TDF = tenofovir; FTC = emtricitabine; ABC = abacavir; 3TC = lamivudine; d4T = stavudine; ZDV = zidovudine. Models fi t for each lipid parameter separately. EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 9 2 NRTIs + efavirenz, and also highest in the com- Trials Group (ACTG) Grade 1-4 thresholds or NCEP 20 22 bination arm of efavirenz + lopinavir/ritonavir. thresholds are reported. Unfortunately there are The nucleoside analogue combinations used in several different threshold levels used, and there is this trial included over 50% using stavudine or not consistent reporting of these data, including the zidovudine. NCEP thresholds, across HIV clinical trials. This analysis was performed to determine the relative effects of PIs and efavirenz versus nucleo- DISCUSSION side analogues on lipid elevations in naïve patients This meta-analysis of 48-week data from 15 clini- in a systematic way. This type of analysis has sev- cal trials of fi rst-line HAART in 6,368 antiretroviral- eral limitations. The nucleoside backbones were naïve patients suggest that the choice of PI, NNRTI, used in different proportions of patients taking and nucleoside backbone can affect lipid eleva- different PIs, and there are not data on all potential tions. Patients receiving lopinavir/ritonavir or combinations. Stavudine is no longer routinely fosamprenavir/ritonavir tended to show greater used in Europe and North America, but the results rises in total cholesterol and triglycerides than were included for completeness. Even so, when the those using the other three ritonavir-boosted trials including stavudine were removed, the over- proteases: saquinavir, darunavir, or atazanavir. all conclusions remained very similar. There were There was no signifi cant difference between efa- no data available from fi rst-line studies of zidovu- virenz and ritonavir-boosted saquinavir, darunavir, dine used with boosted PIs, except the Gilead 934 or atazanavir for rises in total cholesterol or LDL; trial of zidovudine/lamivudine/efavirenz. The rises in triglycerides were higher for the PI-treated results are based only on published summary data patients, and rises in HDL were slightly higher for on mean changes; individual patient level data efavirenz-treated patients. The use of abacavir/lami- were not available. vudine, zidovudine/lamivudine, or stavudine/ The results from this meta-analysis might be lamivudine also led to signifi cantly greater rises affected by biases, such as differences between tri- in all four lipid parameters compared to the use of als in how lipids are measured or summary results tenofovir/emtricitabine. reported, differential use of lipid-lowering drugs, There are several advantages of comparing and different patient populations. However, the lipid elevations in clinical trials of treatment-naïve baseline mean lipid levels were similar between the patients. Samples are collected by standardized trials, and the lipid elevations appeared to be similar protocols and are measured at central laboratories. for patients taking fi xed combinations of nucleoside The use of lipid-lowering agents appears to be analogues and PIs in different trials. Also, the results low for treatment-naïve patients for the four trials appear concordant with those from other random- where this was reported. The baseline lipid levels ized trials not included in this analysis. The ACTG were similar among the 12 trials included. A high 5142 trial showed similar elevations in total choles- proportion of patients showed full virological sup- terol for lopinavir/ritonavir and efavirenz after 96 pression by 48 weeks, which minimizes switches weeks and differences in lipid elevations between in treatment and the potential for “attrition bias.” the NRTIs used, with tenofovir showing the most Most important, the antiretrovirals received are neutral lipid profi le. Also in ACTG 5142, the rises in standardized compared to trials in experienced lipids were highest for patients taking both lopina- patients where a range of “optimized background” vir/ritonavir and efavirenz. The more neutral lipid treatments may have varying effects on lipids, profi le for tenofovir/emtricitabine has been seen which could be diffi cult to control for. For trials of in the BICOMBO (versus abacavir/lamivudine). fi rst-line HAART, there is no carry-over effect on There was no signifi cant difference in lipid eleva- lipids from antiretrovirals previously taken. tions between darunavir/ritonavir and atazanavir/ There are several different methods to report lipid ritonavir in an intensive metabolic study of healthy abnormalities. Most trials report mean or median volunteers. There was no signifi cant difference values over time (either absolute or percentage in lipid elevations between saquinavir/ritonavir change). Trials can be reported in units of either 2000/100 mg once daily and atazanavir/ritonavir mg/dL or mmol/L. In addition, the percentage of 300/100 mg once daily in the BASIC trial of naïve patients with values above predefi ned AIDS Clinical patients, reported recently at Week 24. 10 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 The associations between antiretrovirals and one of several drug toxicities that could affect sur- lipid elevations were highly statistically signifi cant. vival during lifelong treatment for HIV infection However the clinical signifi cance of these results and should not be considered in isolation. for the risk of cardiovascular disease is still unclear. In cardiovascular risk equations, a wide range of factors infl uence the long-term risks of cardiovas- REFERENCES cular disease. Other markers of cardiovascular risk 1. The D:A:D Study Group. Class of antiretroviral drugs are not included in this meta-analysis but may also and the risk of myocardial infarction. N Engl J Med. be important, such as glucose levels, insulin sensi- 2007;356:1723–1735. tivity, infl ammatory markers, changes in the QTc 2. Sabin C, Worm S, Weber R, Reiss P, El-Sadr W, Thiebaut R, interval, carotid intima thickness, and blood pres- et al. Recent use of abacavir and didanosine, but not thymidine analogues, is associated with risk of myocardial sure. The mechanisms by which different antiret- infarction. In: Program and abstracts of the 15th Confer- roviral treatment classes affect lipids are not well ence on Retroviruses and Opportunistic Infections (CROI); understood and may differ between the classes of February 2008; Boston, MA. Abstract 957c. antiretrovirals currently used. Although all of the 3. Bozzette S, Ake C, Tam H, Phippard A, Cohen D, treatments in this meta-analysis raised lipids to Scharfstein D, Louis T. Long-term survival and serious cardiovascular events in HIV-infected patients treated some extent, new combinations have led to more with highly active antiretroviral therapy. JAIDS. 2008;47(3): neutral lipid profi les: in the TMC278-C204 trial, 338–341. use of the NNRTI rilpivirine (TMC278) caused no 4. Hammer S, Eron J, Reiss P, Schooley R, Thompson M, signifi cant rises in lipids, while the efavirenz con- Walmsley S, et al. Antiretroviral treatment of Adult HIV trol arm raised lipid levels signifi cantly by Week infection. 2008 recommendations of the International AIDS Society-USA Panel. JAMA. 2008;300(5):555–570. 48. In the STARTMRK trial, the rises in lipids were 5. Law M, Friis-Moller N, El-Sadr W, Weber R, Reiss P, signifi cantly lower for fi rst-line use of the integrase Arminio-Monforte A, et al. The use of the Framingham inhibitor raltegravir compared with efavirenz. In equation to predict myocardial infarction in HIV-infected the STARTMRK trial, the background nucleosides patients: Comparison with observed events in the D.A.D used were tenofovir/emtricitabine. study. HIV Med. 2006;7:218–230. 6. Anderson K, Odell P, Wilson P, Kannel W. Cardiovascular Analysis of large cohort studies of HIV-infected risk profi les. Am Heart J. 1990;121:293–298. people has suggested that there may be an increased 7. Assmann G, Cullen P, Schulte H. Simple scoring scheme risk of myocardial infarction associated with HIV for calculating the risk of acute coronary events based infection, particularly for those with lower CD4 on the 10-year follow up of the Prospective Cardiovas- counts. The SMART study showed an increased cular Munster (PROCAM) Study. Circulation. 2002;105: 310–315. risk of myocardial infarction in patients stopping 8. Blick G, Greiger-Zanlugo P, Plasencia V, Ulrich G, HAART versus those continuing treatment, which Heiman M, Alaie D, et al. STRESS: A study of cardiovascular was correlated with increases of the infl ammatory risk evaluation by Framingham risk score and technetium 36,37 marker D-dimer. The D:A:D cohort showed Tc99 sestembi exercise stress testing (MIBI) comparing evidence for a higher risk of myocardial infarc- HIV-AIDS to non-HIV-infected patients. In: Program and abstracts of the 4th IAS Conference on HIV Pathogenesis, tion from the use of HAART, particularly if PIs are Treatment and Prevention; July 2007; Sydney, Australia. used; more recently the same cohort also suggests Abstract TUPEB062. an increased cardiovascular risk from use of the 9. Klein D, Hurley L, Silverberg M, Horberg M, Quesenberry C, 1,2 nucleoside analogues abacavir or didanosine. Sidney S. Surveillance data for MI hospitalizations among It is unclear the extent to which these apparently HIV+ and HIV- Northern Californians. In: Program and abstracts of the 14th Conference on Retroviruses and increased risks are due to the effects of antiretrovi- Opportunistic Infections (CROI); February 2007; Los ral treatment on lipids. Angeles, CA. Abstract 807. In summary, given the caveats of meta-analysis, 10. Malan N, Krantz E, Neal D, Kastango K, Frederick D, there is evidence that the choice of both the rito- Mathew M. Efficacy and safety of atazanavir with and navir-boosted PIs, efavirenz, and the nucleoside without ritonavir in antiretroviral-naïve subjects. BMS089: 48-week results. In: Program and abstracts of the 13th analogues can infl uence lipid elevations during Conference on Retroviruses and Opportunistic Infections; fi rst-line HAART. More research is required to February 5–8, 2006; Denver, CO. Abstract 107LB. show whether these lipid elevations, or other mark- 11. Walmsley S, Bernstein B, King M, et al. Lopinavir-ritonavir ers, signifi cantly alter cardiovascular risk for the versus nelfi navir for the initial treatment of HIV infection. different antiretrovirals. Cardiovascular disease is N Engl J Med. 2002;346(26):2039–2046. EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 11 12. Walmsley S, Ruxrungtham K, Slim J, Ward D, Larson P, patients: 48 week safety/effi cacy results from COL103952 Raffi F. The Gemini Study saquinavir/r (SQV/r) vs lopinavir/r (ALERT). In: Program and abstracts of the Fourth IAS Con- (LPV/r) plus emtricitabine/tenofovir (FTC/TDF) as ini- ference on HIV Pathogenesis, Treatment and Prevention; tial therapy in HIV-1 infected patients. In: Program and July 2007; Sydney, Australia. Abstract WEPEB023. abstracts of the European AIDS Clinical Society Confer- 24. Molina J, Andrade-Villanueva J, Echevarria J, ence; October 2007; Madrid, Spain. Abstract PS1/4. Chetchotisakd P, Corral J, David N, et al. Efficacy and 13. DeJesus E, Ortiz R, Khanlou H, Voronin E, vanLunzen J, safety of once-daily atazanavir-ritonavir compared to Andrade-Villanueva J, et al. Efficacy and safety of twice daily lopinavir-ritonavir, each in combination with darunavir/ritonavir versus lopinavir/ritonavir in ARV treat- tenofovir and emtricitabine in ARV-naïve HIV-1 infected ment naive HIV-1 infected patients at Week 48: ARTEMIS. subjects: The CASTLE study, 48 week results. In: Program In: Program and abstracts from ICAAC; September 2007; and abstracts of the 15th Conference on Retroviruses and Chicago, IL. Abstract H-718b. Opportunistic Infections (CROI); February 2008; Boston, 14. Eron J, Yeni P, Gathe J, Estrada V, DeJesus E, MA. Abstract 37. Staszewski S, et al. The KLEAN study of fosamprenavi- 25. Gathe J, Ive P, Wood R, Schurmann D, Bellos N, DeJesus E, ritonavir versus lopinavir-ritonavir, each in combination et al. SOLO: 48-week efficacy and safety comparison of with abacavir- lamivudine, or initial treatment of HIV infec- once-daily fosamprenavir/ritonavir versus twice-daily nelfi navir tion over 48 weeks: A randomised non-inferiority trial. in naïve HIV-1 infected patients. AIDS. 2004; 18:1529–1537. Lancet. 2006;368:476–482. 26. Johnson MA, Gathe JC, Podzamczer D, Molina J, Naylor 15. Gallant J, DeJesus E, Arribas J, Pozniak A, et al. Tenofovir C, Chiu Y, et al. A once-daily lopinavir/ritonavir based DF, emtricitabine and efavirenz vs zidovudine, lamivudine regimen provides noninferior antiviral activity compared and efavirenz for HIV. N Engl J Med. 2006;354:251–260. with a twice-daily regimen. J Acquir Immune Defi c Syndr. 16. Gallant JE, Staszewski S, Pozniak AL, et al. Effi cacy and 2006;43:153–160. safety of tenofovir DF vs stavudine in combination therapy 27. Gathe J, daSilva B, Loufty M, Podzdamczer D, Rubio R, in antiretroviral-naive patients: A 3-year randomized trial. Gibbs S, et al. Study M05-730 primary efficacy results JAMA. 2004;292(2):191–201. at week 48: Phase 3, randomized, open-label study of 17. Martinez E, Arranz J, Podzdamczer D, Ribera E, Knobel H, lopinavir-ritonavir (LPV/r) tablets once-daily (OD) versus Roca V, et al. Effi cacy and safety of NRTI’s switch to teno- twice-daily (BID), co-administered with tenofovir (TDF) + fovir plus emtricitabine versus abacavir plus lamivudine in emtricitabine (FTC) in antiretroviral naïve (ARV) HIV-1 patients with virologic suppression receiving a lamivudine infected subjects. In: Program and abstracts of the 15th containing HAART: The BICOMBO study. In: Program and Conference on Retroviruses and Opportunistic Infections abstracts of the Fourth IAS Conference on HIV Patho- (CROI); February 2008; Boston, MA. Abstract 775. genesis, Treatment and Prevention; July 2007; Sydney, 28. Smith K, Fine D, Patel P, Bellos N, Sloan L, Lackey P, et al. Australia. Abstract WESS101. Effi cacy and safety of abacavir/lamivudine compared to 18. van Leth F, Phanuphak P, Stroes E, Gazzard B, Cahn P, tenofovir/emtriticitabine in combination with one-daily et al. Nevirapine and efavirenz elicit different changes in lopinavir-ritonavir through 48 weeks in the HEAT Study. In: lipid profi les in antiretroviral-therapy-naive patients infected Program and abstracts of the 15th Conference on Retro- with HIV-1. PLoS Med. Vol. 1, No. 1, e19. doi:10.1371/ viruses and Opportunistic Infections (CROI); February journal.pmed.0010019 2008; Boston, MA. Abstract 774. 19. Ruxrungtham K, Bellos N, Morales-Ramirez J, Timerman A, 29. Hicks C, De Jesus E, Wohl D, Liao Q, Pappa K, Lancaster T. Madruga J, Katabira E, et al. The metabolic profile of Once-daily fosamprenavir boosted with either 100 mg or TMC278, an investigational non-nucleoside reverse tran- 200 mg of ritonavir along with abacavir/lamivudine: 48 scriptase inhibitor. In: Program and abstracts of the Fourth week safety and efficacy results from COL100758. In: IAS Conference on HIV Pathogenesis, Treatment and Pre- Program and abstracts of the 11th European AIDS Confer- vention; July 2007; Sydney, Australia. Abstract TUAB105. ence; October 2007; Madrid, Spain. Abstract PS 7.01. 20. Haubrich R, et al. Metabolic outcomes of ACTG 5142: 30. Elion R, DeJesus E, Sension M, Berger D, Towner W, A prospective randomised phase III trial of NRTI, PI- and Richmond G, et al. Once-daily abacavir/lamivudine NNRTI-sparing regimens for initial treatment of HIV-1 infec- (ABC/3TC) and boosted atazanavir (ATV/RTV) in antiretro- tion. In: Program and abstracts of the 14th Conference on viral-naive HIV-1 infected subjects: 48-week results from Retroviruses and Opportunistic Infections; February 2007; COL102060 (SHARE). Poster exhibition: 4th IAS Confer- Los Angeles, CA. Abstract 38. ence on HIV Pathogenesis, Treatment and Prevention; July 21. Grundy S, Cleeman J, Bairey N, Merz A, Brewer H, Clark L. 2007; Sydney, Australia. Abstract WEPEB033. Implications of recent clinical trials for the National 31. Cooper D, et al. Effi cacy and safety of two doses of tipra- Cholesterol Education Program Adult Treatment Panel III navir/ritonavir versus lopinavir/ritonavir-based therapy in Guidelines. Circulation. 2004;110:227–239. antiretroviral-naïve patients: Results of BI 1182.33. In: Pro- 22. Bartlett J, Fath M, DeMasi R, Hermes A, Quinn J, Mondou E. gram and abstracts of the Eighth International Congress An updated systematic overview of triple combination on Drug Therapy in HIV Infection; 2006; Glasgow. Abstract therapy in antiretroviral-naive HIV-infected adults. AIDS. PL13.4. 2006;20(16):2051–2064. 32. Tomaka F, Lefebvre E, Sekar V, van Baelen B, 23. Smith K, Weinberg W, DeJesus E, et al. Once-daily ritona- Vangeneugden T, Vendervoorde A, et al. Similar changes in vir (100 mg) boosting of fosamprenavir or atazanavir with metabolic parameters of darunavir (TMC114) and atazana- tenofovir /emticitabine in antiretroviral naïve HIV-infected vir, each coadmininstered with low-dose ritonavir in healthy 12 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 volunteers. HIV DART: Frontiers in Drug Development of versus efavirenz based combination therapy in treatment- Antiretroviral Therapies; December 2006; Cancun, Mexico. naïve HIV-1 infected patients. In: Program and abstracts Abstract 85. of the ICAAC-IDSA 2008 Conference; October 2008; 33. Vrouenraets S, Garcia E, Jackson A, Raffi F, Jayaweera D, Washington, DC. Abstract H896a. Katlama C, et al. Saquinavir versus atazanavir once daily, 36. Kuller L, et al. Elevated levels of interleukin-6 and D-dimer each boosted with ritonavir 100 mg and combined with are associated with an increased risk of death in patients tenofovir/emtricitabine, result in comparable lipid changes with HIV. In: Program and abstracts of the 15th Conference after 24 weeks in treatment naïve HIV infected patients. on Retroviruses and Opportunistic Infections; February Antiviral Ther. 2008;13(Suppl 4):A46. 2008; Boston, MA. Abstract 139. 34. Hense H. Observations, predictions and decisions – 37. Calmy A, et al. HIV activates markers of cardiovascular risk assessing cardiovascular risk assessment. Int J Epidemiol. in a randomized treatment interruption trial: STACCATO. 2004;33:235–239. In: Program and abstracts of the 15th Conference on 35. Lennox J, DeJesus E, Lazzarin A, Pollard R, Madruga J, Retroviruses and Opportunistic Infections; February 2008; Zhao J, et al. Safety and efficacy of raltegravir-based Boston, MA. Abstract 140. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png HIV Clinical Trials Taylor & Francis

Effects of First-Line Use of Nucleoside Analogues, Efavirenz, and Ritonavir-Boosted Protease Inhibitors on Lipid Levels

HIV Clinical Trials , Volume 10 (1): 12 – Feb 1, 2009

Loading next page...
 
/lp/taylor-francis/effects-of-first-line-use-of-nucleoside-analogues-efavirenz-and-6y938wWF50

References (39)

Publisher
Taylor & Francis
Copyright
© 2009 Thomas Land Publishers, Inc.
ISSN
1528-4336
eISSN
1945-5771
DOI
10.1310/hct1001-1
pmid
19362991
Publisher site
See Article on Publisher Site

Abstract

Effects of First-Line Use of Nucleoside Analogues, Efavirenz, and Ritonavir-Boosted Protease Inhibitors on Lipid Levels 1 2 3 Andrew Hill, Will Sawyer, and Brian Gazzard 1 2 Pharmacology Research Laboratories, University of Liverpool, Liverpool, United Kingdom; MetaVirology Ltd, London, United Kingdom; Chelsea and Westminster Hospital, London, United Kingdom Introduction: Treatment with ritonavir-boosted protease inhibitors or efavirenz and nucleoside analogues leads to rises in lipids, which might contribute to cardiovas- cular risk. Method: A MEDLINE search for clinical trials of fi rst-line HAART with standardized 48-week lipids data available for combinations of two nucleoside analogues identifi ed 13 with additional boosted protease inhibitor (PI/r) regimens (n = 5,281) and two with efavirenz (n = 1,087). Inverse-variance weighting was used to provide estimates of combined 48-week elevations in each lipid parameter. Results: The trials were well balanced for mean baseline total cholesterol (155 mg/dL), trig- lycerides (125 mg/dL), LDL (95 mg/dL), and HDL (37 mg/dL). The PIs showed two dif- ferent types of lipid elevations: Group 1: saquinavir/r, atazanavir/r, and darunavir/r; and Group 2: lopinavir/ritonavir and fosamprenavir/r. There were greater elevations in total cholesterol and triglycerides for Group 2 compared to Group 1 but no differ- ences in LDL or HDL between Group 1 and Group 2. Patients treated with efavirenz showed similar rises in total cholesterol and LDL compared with the PIs in Group 2 but showed smaller rises than in Group 1. In addition, patients treated with abacavir/ lamivudine, zidovudine/lamiuvudine, or stavudine/lamivudine showed signifi cantly higher elevations in all four lipid parameters, compared with patients given tenofovir/ emtricitabine. Conclusion: There is a wide range of lipid elevations during 48 weeks of fi rst-line HAART, which depends on the choice of antiretrovirals used. Key words: cholesterol, HIV, lipids, meta-analysis, nucleoside analogues, protease inhibitors here is evidence from cohort studies that high baseline lipids. In addition, the population of antiretroviral treatment might increase the treated patients is ageing over time, with median 1,2 Trisk of cardiovascular disease, although the age at entry to trials in the late 30s and early 40s. In SMART study showed an increased cardiovascular one large US cohort study of managed care, rates of risk from stopping antiretrovirals. International hospitalization for myocardial infarction were sig- treatment guidelines recommend fi rst-line anti- nifi cantly higher for HIV-infected people relative to retroviral treatment with the combination of uninfected people. two nucleoside analogues plus either the non- When antiretroviral treatments have been com- nucleoside efavirenz or a ritonavir-boosted pro- pared directly in randomized clinical trials of naïve tease inhibitor. and pretreated patients, the mean lipid profi les Antiretroviral treatment can raise levels of have differed both between and within antiret- lipids, and elevated levels of certain lipid param- roviral drug classes. Ritonavir-boosted protease eters are known to elevate cardiovascular risk inhibitors (PIs) tend to show greater rises in lipids 5 10,11 for HIV-infected people. Therefore, the effect of than unboosted PIs. Of the ritonavir-boosted antiretrovirals on lipids might be an indicator of their associated cardiovascular risk. Some cardio- Address for correspondence: Andrew Hill, Pharmacology vascular risk equations include a combination of Research Laboratories, University of Liverpool, 70 Pembroke high-density lipoprotein (HDL) and total choles- Place, Liverpool L69 3GF, United Kingdom. Email: microhaart@ terol, whereas others use a combination of HDL, aol.com 6,7 low-density lipoprotein (LDL), and triglycerides. HIV Clin Trials 2009;10(1):1–12 People infected with HIV in Western countries © 2009 Thomas Land Publishers, Inc. www.thomasland.com tend to have a high incidence of pre-existing risk factors, such as a high incidence of smoking or doi: 10.1310/hct1001-1 1 2 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 PIs, lopinavir/ritonavir has caused greater rises IAS Pathogenesis Conference), and International in total cholesterol and triglycerides than either Conference on Drug Therapy in HIV Infection. 12,13 darunavir/ritonavir or saquinavir/ritonavir. To be included in the analysis, clinical trials By contrast, the KLEAN trial showed similar lipid needed to have (a) at least 50 antiretroviral-naïve profi les for fosamprenavir/ritonavir and lopina- patients studied for at least 48 weeks, (b) use of a vir/ritonavir. The choice of nucleoside analogues ritonavir-boosted PI (using a total daily ritonavir or non-nucleosides used can also affect lipid pro- dose of 100–200 mg) or efavirenz in combination 15–20 fi les. Efavirenz tends to cause greater rises in with a fi xed pair of nucleoside analogues, and lipids than either nevirapine or the investigational (c) data available on mean or median levels of total non-nucleoside reverse transcriptase inhibitor cholesterol, triglycerides, LDL, and HDL at base- (NNRTI) TMC278 (rilpivirine). In several studies, line and at Week 48. the nucleoside analogues abacavir, stavudine, or For each of the 15 trials included in the meta- zidovudine have caused greater rises in lipids than analysis, the mean levels of total cholesterol, trig- 15–17 tenofovir. lycerides, LDL, and HDL were recorded at baseline Given the wide range of different measures used, and after 48 weeks of treatment. We used the stan- it can be diffi cult for clinicians to compare lipid dard deviation of the mean lipid parameters from elevations from reports of different clinical trials, the two arms of the ARTEMIS trial to estimate the including either the same or different treatments. 95% confi dence intervals (CI) of the mean lipid In clinical trials of pretreated patients, changes in parameters for the other trials when estimates of lipids from baseline can also depend on the antiret- variability were unavailable. Although data on rovirals taken in screening and whether these were mean lipid levels were available for most trials, stopped or continued into the randomized phase of there were few standardized data on the percentage the trials. There also tends to be more use of lipid- of patients with lipid elevations above predefi ned lowering drugs in trials of pretreated patients, and cutoff levels, for example, the National Cholesterol these can be continued or stopped either at baseline Evaluation Program (NCEP) thresholds. or during the trial, with no consistent methods to The change from baseline in each lipid parameter adjust for this in the analysis. was calculated as the mean 48-week value minus the baseline value. The median and mean rises of each lipid parameter were found to be similar METHOD overall. The summary changes for each trial were The aim of this analysis was to estimate the effect presented graphically, with a combined estimate of nucleoside analogues, ritonavir-boosted PIs, and of change from all trials of each boosted PI and of efavirenz on lipids in naïve patients, using a stan- efavirenz. For these combined analyses, inverse- dardized format to report lipid elevations. A similar variance weighting was used. The changes in lipids approach was used in an overview of HIV RNA sup- were presented by trial and treatment arm with pression rates across clinical trials in naïve patients. associated 95% CIs. The trials were divided by the We conducted a search for prospective clini- use of different nucleoside reverse transcriptase cal trials of ritonavir-boosted PI treatment in inhibitor (NRTI) backbones (tenofovir/emtricit- antiretroviral-naïve patients, using MEDLINE and abine vs. either abavavir, stavudine, or zidovudine separate searches of PI drug labels and conference used with lamivudine). In separate meta-regression abstracts. The online literature searches used the analyses for each lipid parameter, the relative names of each antiretroviral, followed by “clini- effects of PIs or efavirenz versus the nucleoside cal trial.” In addition, we searched the US Food analogues were assessed. Models included base- and Drug Administration (FDA) product labels line measures for each study arm of mean age, for registration trials of each approved antiretro- proportion of males, proportion of Caucasian race, viral and for abstracts on clinical trials presented mean CD4 count, and mean HIV RNA, together at the following conferences: Annual Conference with three binary variables for efavirenz, PI group, Retroviruses and Opportunistic Infections, Inter- and NRTI group. A backwards stepwise regression national Conference on Antimicrobial Agents and method retaining the treatment group variables Chemotherapy (ICAAC), European AIDS Clinical was used to reduce the variables in the models. Society, International AIDS Conference (including All analyses were performed using SAS version EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 3 9.1 (SAS Institute, Cary, North Carolina, USA). Table 1. ACTG 5142 trial: median percent change Predicted values based on these models were used from baseline to Week 96 in total cholesterol and as estimates of mean lipid levels within each com- triglycerides bination of PI, efavirenz, and NRTI group. Treatment arm 2NRTI + 2NRTI + EFV + RESULTS EFV LPV/r LPV/r The search identifi ed 15 clinical trials with a n = 169 n = 158 n = 168 total of 6,368 patients. Thirteen trials included Total cholesterol +33% +32% +57% two nucleoside analogues and a ritonavir-boosted Triglycerides +19% +46% +62% 12 13 14 PI. Five trials (GEMINI, ARTEMIS, KLEAN, 23 24 ALERT, CASTLE ) were direct head-to-head Note: NRTI = nucleoside reverse transcriptase inhibitor; comparisons of ritonavir-boosted PIs. Three trials EFV = efavirenz; LPV/r = ritonavir-boosted lopinavir. 10 11 25 (BMS-089, Abbott 863, and SOLO ) compared ritonavir-boosted PIs with unboosted PIs. The 26 27 Abbott 418 and Abbott 730 trials were compari- once daily, or 1400/100 mg once daily; and 3,148 sons of twice-daily versus once-daily lopinavir/ patients taking lopinavir/ritonavir either at the ritonavir. The HEAT trial evaluated lopinavir/ 400/100 mg bid dose or the 800/200 mg once-daily ritonavir with two different NRTI backbones. The dose. All of the data on lopinavir/ritonavir were REDUCE trial compared two boosting doses of for patients using the soft-gelatin capsule formu- ritonavir in combination with fosamprenavir once lation, except for the Abbott 730 trial, where the daily. Finally the SHARE trial was an open-label tablet formulation was used. In addition, data were evaluation of fosamprenavir/ritonavir once daily. available for 1,087 patients taking efavirenz at the For the Abbott 863 trial, data on total cholesterol standard 600 mg once-daily dose. and triglycerides were included, but the data Overall, the NRTI backbone used was tenofovir/ on HDL and LDL were not available. Two trials emtricitabine for 3,698 patients, abacavir/lamvu- included two nucleoside analogues and efavirenz dine for 1,705 patients, stavudine/lamivudine for 15,16 (Gilead 903 and Gilead 934). 722 patients, and zidovudine/lamivudine for 243 Two large trials of fi rst-line ritonavir-boosted PIs patients. The mean baseline CD4 count ranged from could not be included. The ACTG 5142 trial had 153 to 279 cells/µL, whereas mean baseline HIV been conducted in naïve patients but was excluded RNA levels ranged from 4.6 to 5.2 log copies/mL. from this analysis as a range of NRTI backbones The predominantly male patients had mean ages at had been used and only 96-week data were avail- baseline of between 36 and 40 and between 44%–78% able. However, summary data from ACTG 5142 were Caucasian. The majority of patients in each were used to cross-validate the results from this trial showed HIV RNA reductions below 50 copies/ analysis and are shown in Table 1. In addition, mL in intent-to-treat (ITT) analysis (Table 2). the BI 1182.33 trial (tipranavir/ritonavir vs. lopi- Mean levels of the four lipid parameters at base- navir/ritonavir in naïve patients) was excluded line and Week 48 are shown in Table 3. At base- because data on mean levels of lipids were not line, the mean levels of each lipid parameter were available. There were several trials of efavirenz- similar across the trials, with mean baseline levels based HAART that could not be included because as follows: total cholesterol, mean 155 (range, 147– lipid data had only been reported in terms of 168 mg/dL); triglycerides, mean 125 (range, graded toxicities and not as mean values. 105–166 mg/dL); LDL, mean 95 mg/dL (range, 89– Table 2 shows a summary of the 26 arms from 104 mg/dL); and HDL, mean 37 mg/dL (range, the 15 clinical trials included in the analysis. Data 31–42 mg/dL). Use of lipid-lowering drugs was were available for 166 patients taking saquinavir/ reported in only four of the trials. The percentage ritonavir 1000/100 mg bid; 343 patients on daruna- of patients using lipid-lowering drugs was 4.5% in 30 14 vir/ritonavir 800/100 mg once daily; 700 patients the SHARE trial, 11% in the KLEAN trial, 5% in 12 26 taking atazanavir/ritonavir 300/100 mg once the GEMINI trial, and 2.6% in the ABT-418 trial. daily; 924 patients taking fosamprenavir/ritonavir In the regression analyses, there were no sig- at doses of either 700/100 mg bid, 1400/200 mg nifi cant differences in lipid elevations between 4 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 Table 2. Baseline characteristics of trials included in analysis Baseline data and summary 48-week effi cacy CD4 count HIV RNA (log HIV RNA <50 Trial [ref] N Age %Male %Caucasian (cells/µL) copies/mL) Week 48 GEMINI [12] SQV/r TDF/FTC 166 38 81% 49% 156 5.2 65% ARTEMIS [13] DRV/r TDF/FTC 343 36 70% 40% 228 4.9 84% ALERT [23] ATV/r TDF/3TC 53 40 89% 49% 188 4.9 83% SHARE [29] ATV/r ABC/3TC 111 38 91% 54% 208 5.1 77% BMS-089 [10] ATV/r d4T/3TC 95 nd 73% 53% 201 4.8 75% CASTLE ATV/r TDF/FTC 440 34 69% nd 205 5.0 78% ALERT [23] FPV/r TDF/3TC 53 40 79% 64% 161 4.9 75% SOLO [24] FPV/r ABC/3TC 322 36 70% 51% 166 4.8 69% KLEAN [14] FPV/r ABC/3TC 434 38 78% 61% 188 5.1 66% REDUCE [28] FPV/r ABC/3TC 57 40 81% 46% 179 4.9 63% REDUCE [28] FPV/r ABC/3TC 58 39 81% 36% 259 4.7 67% ARTEMIS [13] LPV/r TDF/FTC 346 35 70% 44% 218 4.8 78% GEMINI [12] LPV/r TDF/FTC 171 37 77% 44% 153 5.2 64% ABT 418 [25] LPV/r TDF/FTC 75 39 81% 56% 214 4.8 64% ABT 418 [25] LPV/r TDF/FTC 115 38 79% 51% 232 4.6 70% ABT 730 [26] LPV/r TDF/FTC 333 nd 80% 78% 216 4.9 77% ABT 730 [26] LPV/r TDF/FTC 331 nd 77% 73% 215 5.1 76% HEAT [27] LPV/r TDF/FTC 286 39 80% 50% 193 4.8 67% CASTLE LPV/r TDF/FTC 443 36 69% nd 204 5.0 76% HEAT [27] LPV/r ABC/3TC 278 38 84% 52% 214 4.9 68% KLEAN [14] LPV/r ABC/3TC 444 37 78% 56% 194 5.1 65% ABT 863 [11] LPV/r d4T/3TC 326 38 80% 56% 232 5.0 67% Gilead-903 EFV TDF/3TC 299 36 76% 64% 279 4.9 82% Gilead-934 EFV TDF/FTC 244 38 86% 59% 245 5.0 80% Gilead-903 EFV d4T/3TC 301 36 76% 64% 279 4.9 81% Gilead-934 EFV ZDV/3TC 243 38 86% 59% 245 5.0 71% Note: SQV/r = ritonavir-boosted saquinavir; TDF = tenofovir; FTC = emtricitabine; DRV/r = ritonavir-boosted darunavir; ATV/r = ritonavir-boosted atazanavir; ABC = abacavir; 3TC = lamivudine; d4T = stavudine; FPV/r = ritonavir-boosted fos-amprenavir; LPV/r = ritonavir-boosted lopinavir; EFV = efavirenz; AZT = zidovudine. ritonavir-boosted fos-amprenavir (fAPV/r) using only 100 mg ritonavir per day. LPV/r 800/200 mg once daily. All other trials used LPV/r dosed at 400/100 mg bid. saquinavir, darunavir, or atazanavir when boosted In almost all cases, none of the baseline covari- with ritonavir, and so these PIs were grouped ates in the meta-regression analyses were retained (Group 1) to improve the power of the analyses. in the models after the backwards stepwise pro- Likewise, there were also no signifi cant differences cedures. Predicted values of the mean lipid levels in lipid elevations between fosamprenavir/rito- based on PI and NRTI grouping are summarized navir and lopinavir/ritonavir and these too were in Table 4. grouped together (Group 2). For the nucleoside Use of atazanavir/r, darunavir/r, or saquinavir/r analogues, there was no signifi cant difference in (PI Group 1) led to smaller rises in total cholesterol lipid elevations seen between abacavir, zidovu- than use of lopinavir/r and fosamprenavir/r (PI dine, or stavudine when given with lamivudine. Group 2) (p = .0390). The rises in total cholesterol Figures 1A to 1D show the mean rises in the four for PI Group 1 were comparable for trials using efa- lipid parameters, for the combined data for each virenz (p = .154). The use of stavudine, zidovudine, ritonavir-boosted PI and for efavirenz, stratifi ed for or abacavir as NRTIs also led to higher elevations the use of either tenofovir/emtricitabine or other in total cholesterol relative to the use of tenofovir nucleoside analogues. (p < .0001). EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 5 Table 3. Rises in mean lipid levels from baseline to Week 48 during clinical trials of boosted PIs or efarvirenz in naïve patients (mg/dL) Trial PI/r NRTI TCHOL TRIG LDL HDL GEMINI SQV/r TDF/FTC 157 → 183 (+26) 117 → 131 (+14) 89 → 107 (+18) 35 → 44 (+9) ARTEMIS DRV/r TDF/FTC 156 → 177 (+21) 131 → 143 (+12) 92 → 105 (+13) 39 → 44 (+5) ALERT ATV/r TDF/3TC 153 → 181 (+28) 123 → 133 (+10) 97 → 102 (+5) 38 → 48 (+10) CASTLE ATV/r TDF/FTC 147 → 164 (+17) 110 → 124 (+17) 91 → 102 (+11) 36 → 45 (+9) SHARE ATV/r ABC/3TC 165 → 198 (+33) 120 → 174 (+54) 103 → 115 (+12) 31 → 46 (+15) BMS-089 ATV/r d4T/3TC 158 → 183 (+25) 139 → 174 (+35) 96 → 119 (+23) 38 → 49 (+11) ALERT FPV/r TDF/3TC 160 → 179 (+19) 120 → 167 (+47) 95 → 99 (+4) 35 → 43 (+8) SOLO FPV/r ABC/3TC 162 → 214 (+52) 125 → 190 (+65) 96 → 123 (+27) 36 → 44 (+8) KLEAN FPV/r ABC/3TC 156 → 217 (+61) 112 → 179 (+67) 96 → 124 (+28) 33 → 46 (+13) REDUCE FPV/r ABC/3TC 166 → 197 (+31) 120 → 149 (+29) 104 → 108 (+4) 39 → 50 (+11) REDUCE FPV/r ABC/3TC 166 → 209 (+43) 120 → 172 (+52) 101 → 126 (+25) 37 → 46 (+9) ARTEMIS LPV/r TDF/FTC 156 → 188 (+32) 123 → 192 (+69) 93 → 105 (+12) 40 → 47 (+7) GEMINI LPV/r TDF/FTC 152 → 183 (+31) 117 → 172 (+55) 89 → 105 (+16) 38 → 49 (+11) ABT418 LPV/r TDF/FTC 168 → 195 (+27) 136 → 212 (+76) 102 → 115 (+13) 42 → 48 (+6) ABT418 LPV/r TDF/FTC 159 → 186 (+27) 137 → 219 (+82) 92 → 106 (+14) 40 → 43 (+3) ABT-730 LPV/r TDF/FTC 155 → 184 (+29) 151 → 196 (+45) 98 → 105 (+7) 40 → 45 (+5) ABT-730 LPV/r TDF/FTC 154 → 189 (+35) 156 → 220 (+64) 95 → 102 (+7) 39 → 48 (+9) HEAT LPV/r TDF/FTC 159 → 184 (+23) 134 → 189 (+38) 92 → 97 (+5) 35 → 47 (+12) CASTLE LPV/r TDF/FTC 147 → 185 (+38) 110 → 168 (+58) 91 → 108 (+17) 35 → 46 (+11) HEAT LPV/r ABC/3TC 159 → 202 (+43) 122 → 215 (+93) 93 → 105 (+12) 36 → 48 (+12) KLEAN LPV/r ABC/3TC 157 → 210 (+53) 117 → 195 (+78) 97 → 120 (+23) 34 → 48 (+14) ABT-863 LPV/r d4T/3TC 158 → 211 (+53) 166 → 291 (+125) — — Gilead-903 EFV TDF/3TC 150 → 175 (+25) 105 → 113 (+8) 94 → 106 (+12) 38 → 44 (+6) Gilead-934 EFV TDF/FTC 160 → 180 (+20) 105 → 113 (+8) 100 → 110 (+10) 36 → 42 (+6) Gilead-903 EFV d4T/3TC 149 → 204 (+55) 129 → 183 (+54) 96 → 122 (+18) 37 → 45 (+8) Gilead-934 EFV ZDV/3TC 157 → 191 (+34) 122 → 136 (+14) 98 → 116 (+18) 36 → 45 (+9) Note: Numbers in parentheses show absolute change from baseline to Week 48. PI/r = ritonavir-boosted protease inhibitor; NRTI = nucleoside reverse transcriptase inhibitor; TCHOL = total cholesterol; TRIG = triglycerides; LDL = low-density lipoprotein; HDL = high-density lipoprotein; SQV/r = ritonavir-boosted saquinavir; TDF = tenofovir; FTC = emtricitabine; DRV/r = ritonavir-boosted darunavir; ATV/r = ritonavir-boosted atazanavir; ABC = abacavir; 3TC = lamivudine; FPV/r = ritonavir-boosted fos-amprenavir; LPV/r = ritonavir-boosted lopinavir; d4T = stavudine; EFV = efavirenz; ZDV = zidovudine. fAPV/r using 100 mg ritonavir per day. All other trials used fAPV/r 1400/200 mg once daily or 700/100 mg bid (KLEAN). LPV/r 800/200 mg once daily. All other trials used LPV/r dosed at 400/100 mg bid. For triglycerides, the mean rise from use of rise in LDL (10.0 mg/dL; 95% CI 4–17; p = .001) for atazanavir/r, darunavir/r, or saquinavir/r was patients taking abacavir/lamivudine, stavudine/ 39 mg/dL (95% CI 33–44) lower than for trials of lamivudine, or zidovudine/lamivudine relative to lopinavir/r and fosamprenavir/r (p < .0001) and was those using tenofovir/emtricitabine (Table 4). 10 mg/dL (95% CI 3–16) higher than for trials with For HDL, there was no evidence of an effect of efavirenz (p = .005). The mean rise in triglycerides was antiretroviral treatment on absolute changes from 23 mg/dL higher (95% CI 19–27; p < .001) for patients baseline. There was evidence to suggest that per- receiving either abacavir/lamivudine, stavudine/ centage change from baseline was less for those lamivudine, or zidovudine/lamivudine compared studies with efavirenz as opposed to a boosted PI to those receiving tenofovir/emtricitabine. (−10%; 95% CI −20% to −1%; p = .032) and greater For LDL, there was no evidence for differences for those using tenofovir as opposed to other between the boosted PIs or with efavirenz. The NRTIs (11%; 95% CI 4% to 15%; p < .001). confi dence intervals around estimates of change Table 4 shows summary results on lipid elevat- were wide for all of the PIs. There was a greater ions from the ACTG 5142 study , which randomized 6 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n = 1048) EFV TDF (n = 543) non-TDF (n = 544) 0 102030 40506070 mean change from baseline to week 48 (mg/dL) SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n =1048) EFV TDF (n = 543) non-TDF (n = 544) –20 0 20 40 60 80 100 120 mean change from baseline to week 48 (mg/dL) Figure 1. Mean absolute change from baseline to Week 48 (±95% confi dence intervals). (A) Total cholesterol (mg/dL); (B) triglycerides (mg/dL); (C) LDL (mg/dL); (D) HDL (mg/dL). SQV/rtv = ritonavir-boosted saquinavir; TDF = tenofovir; DRV/rtv = ritonavir-boosted darunavir; ATV/rtv = ritonavir-boosted atazanavir; fAPV/rtv = ritonavir-boosted fos-amprenavir; LPV/rtv = ritonavir-boosted lopinavir; EFV = efavirenz. EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 7 SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n = 722) EFV TDF (n = 543) non-TDF (n = 544) –10 –5 0 5 10 15 20 25 30 35 mean change from baseline to week 48 (mg/dL) SQV/rtv TDF (n = 166) DRV/rtv TDF (n = 343) ATV/rtv TDF (n = 493) non-TDF (n = 207) fAPV/rtv TDF (n = 53) non-TDF (n = 871) LPV/rtv TDF (n = 2100) non-TDF (n = 722) EFV TDF (n = 543) non-TDF (n = 544) 0 2 4 6 8 10 1214 1618 mean change from baseline to week 48 (mg/dL) Figure 1. (Continued). 753 naïve patients to three arms: (a) two nucleoside ritonavir caused similar elevations in total cho- analogues plus efavirenz, (b) two nucleoside ana- lesterol to Week 96, whereas the combination arm logues plus lopinavir/ritonavir, and (c) efavirenz of efavirenz plus lopinavir/ritonavir caused the plus lopinavir/ritonavir. In this study, the arms greatest rises. Elevations in triglycerides were of 2 NRTIs + efavirenz and 2 NRTIs + lopinavir/ greater for 2 NRTIs + lopinavir/ritonavir than for 8 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 Table 4. Meta-regression analysis of absolute and percentage changes in lipids from baseline to Week 48 ATV/r, SQV/r, or DRV/r LPV/r or fAPV/r EFV ABC/3TC or ABC/3TC or d4T/3TC or TDF/FTC d4T/3TC TDF/FTC d4T/3TC TDF/FTC ZDV/3TC Mean absolute change (mg/dL) Total cholesterol +19 (9–27) +38 (25–49) +32 (26–38) +52 (44–59) +25 (15–34) +44 (36–54) Triglycerides +20 (15–25) +44 (38–49) +59 (55–62) +82 (78–86) +10 (6–15) +34 (28–39) LDL +11 (4–18) +22 (13–30) +11 (6–16) +22 (17–27) +11 (4–19) +22 (15–29) HDL +9 (2–16) +12 (4–20) +8 (3–13) +12 (5–19) +5 (–2–14) +9 (1–17) Mean percentage change (%) Total cholesterol +12% (10–14%) +24% (22–27%) +21% (19–22%) +33% (31–34%) +16% (14–19%) +29% (26–31%) Triglycerides +17% (11–23%) +35% (28–42%) +46% (42–50%) +64% (60–68%) +9% (2–16%) +27% (20–34%) LDL +12% (11–13%) +23% (22–24%) +12% (11–13%) +23% (22–24%) +12% (10–13%) +23% (21–24%) HDL +24% (18–31%) +35% (27–43%) +23% (9–28%) +34% (13–34%) +14% (6–22%) +25% (18–34%) Note: Effects of boosted protease inhibitors (PIs), efavirenz, and nucleoside reverse transcriptase inhibitors (NRTIs) summarized using predicted values (95% confi - dence intervals in parentheses). Conversion factors from mg/dL to mmol/L: Total cholesterol, LDL, and HDL: mg/dL/38.6 = mmol/L. Triglycerides: mg/dL/88.5 = mmol/L. ATV/r = ritonavir-boosted atazanavir; SQV/r = ritonavir-boosted saquinavir; DRV/r = ritonavir-boosted darunavir; LPV/r = ritonavir-boosted lopinavir; fAPV/r = ritonavir-boosted fos-amprenavir; EFV = efavirenz; TDF = tenofovir; FTC = emtricitabine; ABC = abacavir; 3TC = lamivudine; d4T = stavudine; ZDV = zidovudine. Models fi t for each lipid parameter separately. EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 9 2 NRTIs + efavirenz, and also highest in the com- Trials Group (ACTG) Grade 1-4 thresholds or NCEP 20 22 bination arm of efavirenz + lopinavir/ritonavir. thresholds are reported. Unfortunately there are The nucleoside analogue combinations used in several different threshold levels used, and there is this trial included over 50% using stavudine or not consistent reporting of these data, including the zidovudine. NCEP thresholds, across HIV clinical trials. This analysis was performed to determine the relative effects of PIs and efavirenz versus nucleo- DISCUSSION side analogues on lipid elevations in naïve patients This meta-analysis of 48-week data from 15 clini- in a systematic way. This type of analysis has sev- cal trials of fi rst-line HAART in 6,368 antiretroviral- eral limitations. The nucleoside backbones were naïve patients suggest that the choice of PI, NNRTI, used in different proportions of patients taking and nucleoside backbone can affect lipid eleva- different PIs, and there are not data on all potential tions. Patients receiving lopinavir/ritonavir or combinations. Stavudine is no longer routinely fosamprenavir/ritonavir tended to show greater used in Europe and North America, but the results rises in total cholesterol and triglycerides than were included for completeness. Even so, when the those using the other three ritonavir-boosted trials including stavudine were removed, the over- proteases: saquinavir, darunavir, or atazanavir. all conclusions remained very similar. There were There was no signifi cant difference between efa- no data available from fi rst-line studies of zidovu- virenz and ritonavir-boosted saquinavir, darunavir, dine used with boosted PIs, except the Gilead 934 or atazanavir for rises in total cholesterol or LDL; trial of zidovudine/lamivudine/efavirenz. The rises in triglycerides were higher for the PI-treated results are based only on published summary data patients, and rises in HDL were slightly higher for on mean changes; individual patient level data efavirenz-treated patients. The use of abacavir/lami- were not available. vudine, zidovudine/lamivudine, or stavudine/ The results from this meta-analysis might be lamivudine also led to signifi cantly greater rises affected by biases, such as differences between tri- in all four lipid parameters compared to the use of als in how lipids are measured or summary results tenofovir/emtricitabine. reported, differential use of lipid-lowering drugs, There are several advantages of comparing and different patient populations. However, the lipid elevations in clinical trials of treatment-naïve baseline mean lipid levels were similar between the patients. Samples are collected by standardized trials, and the lipid elevations appeared to be similar protocols and are measured at central laboratories. for patients taking fi xed combinations of nucleoside The use of lipid-lowering agents appears to be analogues and PIs in different trials. Also, the results low for treatment-naïve patients for the four trials appear concordant with those from other random- where this was reported. The baseline lipid levels ized trials not included in this analysis. The ACTG were similar among the 12 trials included. A high 5142 trial showed similar elevations in total choles- proportion of patients showed full virological sup- terol for lopinavir/ritonavir and efavirenz after 96 pression by 48 weeks, which minimizes switches weeks and differences in lipid elevations between in treatment and the potential for “attrition bias.” the NRTIs used, with tenofovir showing the most Most important, the antiretrovirals received are neutral lipid profi le. Also in ACTG 5142, the rises in standardized compared to trials in experienced lipids were highest for patients taking both lopina- patients where a range of “optimized background” vir/ritonavir and efavirenz. The more neutral lipid treatments may have varying effects on lipids, profi le for tenofovir/emtricitabine has been seen which could be diffi cult to control for. For trials of in the BICOMBO (versus abacavir/lamivudine). fi rst-line HAART, there is no carry-over effect on There was no signifi cant difference in lipid eleva- lipids from antiretrovirals previously taken. tions between darunavir/ritonavir and atazanavir/ There are several different methods to report lipid ritonavir in an intensive metabolic study of healthy abnormalities. Most trials report mean or median volunteers. There was no signifi cant difference values over time (either absolute or percentage in lipid elevations between saquinavir/ritonavir change). Trials can be reported in units of either 2000/100 mg once daily and atazanavir/ritonavir mg/dL or mmol/L. In addition, the percentage of 300/100 mg once daily in the BASIC trial of naïve patients with values above predefi ned AIDS Clinical patients, reported recently at Week 24. 10 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 The associations between antiretrovirals and one of several drug toxicities that could affect sur- lipid elevations were highly statistically signifi cant. vival during lifelong treatment for HIV infection However the clinical signifi cance of these results and should not be considered in isolation. for the risk of cardiovascular disease is still unclear. In cardiovascular risk equations, a wide range of factors infl uence the long-term risks of cardiovas- REFERENCES cular disease. Other markers of cardiovascular risk 1. The D:A:D Study Group. Class of antiretroviral drugs are not included in this meta-analysis but may also and the risk of myocardial infarction. N Engl J Med. be important, such as glucose levels, insulin sensi- 2007;356:1723–1735. tivity, infl ammatory markers, changes in the QTc 2. Sabin C, Worm S, Weber R, Reiss P, El-Sadr W, Thiebaut R, interval, carotid intima thickness, and blood pres- et al. Recent use of abacavir and didanosine, but not thymidine analogues, is associated with risk of myocardial sure. The mechanisms by which different antiret- infarction. In: Program and abstracts of the 15th Confer- roviral treatment classes affect lipids are not well ence on Retroviruses and Opportunistic Infections (CROI); understood and may differ between the classes of February 2008; Boston, MA. Abstract 957c. antiretrovirals currently used. Although all of the 3. Bozzette S, Ake C, Tam H, Phippard A, Cohen D, treatments in this meta-analysis raised lipids to Scharfstein D, Louis T. Long-term survival and serious cardiovascular events in HIV-infected patients treated some extent, new combinations have led to more with highly active antiretroviral therapy. JAIDS. 2008;47(3): neutral lipid profi les: in the TMC278-C204 trial, 338–341. use of the NNRTI rilpivirine (TMC278) caused no 4. Hammer S, Eron J, Reiss P, Schooley R, Thompson M, signifi cant rises in lipids, while the efavirenz con- Walmsley S, et al. Antiretroviral treatment of Adult HIV trol arm raised lipid levels signifi cantly by Week infection. 2008 recommendations of the International AIDS Society-USA Panel. JAMA. 2008;300(5):555–570. 48. In the STARTMRK trial, the rises in lipids were 5. Law M, Friis-Moller N, El-Sadr W, Weber R, Reiss P, signifi cantly lower for fi rst-line use of the integrase Arminio-Monforte A, et al. The use of the Framingham inhibitor raltegravir compared with efavirenz. In equation to predict myocardial infarction in HIV-infected the STARTMRK trial, the background nucleosides patients: Comparison with observed events in the D.A.D used were tenofovir/emtricitabine. study. HIV Med. 2006;7:218–230. 6. Anderson K, Odell P, Wilson P, Kannel W. Cardiovascular Analysis of large cohort studies of HIV-infected risk profi les. Am Heart J. 1990;121:293–298. people has suggested that there may be an increased 7. Assmann G, Cullen P, Schulte H. Simple scoring scheme risk of myocardial infarction associated with HIV for calculating the risk of acute coronary events based infection, particularly for those with lower CD4 on the 10-year follow up of the Prospective Cardiovas- counts. The SMART study showed an increased cular Munster (PROCAM) Study. Circulation. 2002;105: 310–315. risk of myocardial infarction in patients stopping 8. Blick G, Greiger-Zanlugo P, Plasencia V, Ulrich G, HAART versus those continuing treatment, which Heiman M, Alaie D, et al. STRESS: A study of cardiovascular was correlated with increases of the infl ammatory risk evaluation by Framingham risk score and technetium 36,37 marker D-dimer. The D:A:D cohort showed Tc99 sestembi exercise stress testing (MIBI) comparing evidence for a higher risk of myocardial infarc- HIV-AIDS to non-HIV-infected patients. In: Program and abstracts of the 4th IAS Conference on HIV Pathogenesis, tion from the use of HAART, particularly if PIs are Treatment and Prevention; July 2007; Sydney, Australia. used; more recently the same cohort also suggests Abstract TUPEB062. an increased cardiovascular risk from use of the 9. Klein D, Hurley L, Silverberg M, Horberg M, Quesenberry C, 1,2 nucleoside analogues abacavir or didanosine. Sidney S. Surveillance data for MI hospitalizations among It is unclear the extent to which these apparently HIV+ and HIV- Northern Californians. In: Program and abstracts of the 14th Conference on Retroviruses and increased risks are due to the effects of antiretrovi- Opportunistic Infections (CROI); February 2007; Los ral treatment on lipids. Angeles, CA. Abstract 807. In summary, given the caveats of meta-analysis, 10. Malan N, Krantz E, Neal D, Kastango K, Frederick D, there is evidence that the choice of both the rito- Mathew M. Efficacy and safety of atazanavir with and navir-boosted PIs, efavirenz, and the nucleoside without ritonavir in antiretroviral-naïve subjects. BMS089: 48-week results. In: Program and abstracts of the 13th analogues can infl uence lipid elevations during Conference on Retroviruses and Opportunistic Infections; fi rst-line HAART. More research is required to February 5–8, 2006; Denver, CO. Abstract 107LB. show whether these lipid elevations, or other mark- 11. Walmsley S, Bernstein B, King M, et al. Lopinavir-ritonavir ers, signifi cantly alter cardiovascular risk for the versus nelfi navir for the initial treatment of HIV infection. different antiretrovirals. Cardiovascular disease is N Engl J Med. 2002;346(26):2039–2046. EFFECT OF FIRST-LINE HAART ON LIPID LEVELS � H ILL ET AL. 11 12. Walmsley S, Ruxrungtham K, Slim J, Ward D, Larson P, patients: 48 week safety/effi cacy results from COL103952 Raffi F. The Gemini Study saquinavir/r (SQV/r) vs lopinavir/r (ALERT). In: Program and abstracts of the Fourth IAS Con- (LPV/r) plus emtricitabine/tenofovir (FTC/TDF) as ini- ference on HIV Pathogenesis, Treatment and Prevention; tial therapy in HIV-1 infected patients. In: Program and July 2007; Sydney, Australia. Abstract WEPEB023. abstracts of the European AIDS Clinical Society Confer- 24. Molina J, Andrade-Villanueva J, Echevarria J, ence; October 2007; Madrid, Spain. Abstract PS1/4. Chetchotisakd P, Corral J, David N, et al. Efficacy and 13. DeJesus E, Ortiz R, Khanlou H, Voronin E, vanLunzen J, safety of once-daily atazanavir-ritonavir compared to Andrade-Villanueva J, et al. Efficacy and safety of twice daily lopinavir-ritonavir, each in combination with darunavir/ritonavir versus lopinavir/ritonavir in ARV treat- tenofovir and emtricitabine in ARV-naïve HIV-1 infected ment naive HIV-1 infected patients at Week 48: ARTEMIS. subjects: The CASTLE study, 48 week results. In: Program In: Program and abstracts from ICAAC; September 2007; and abstracts of the 15th Conference on Retroviruses and Chicago, IL. Abstract H-718b. Opportunistic Infections (CROI); February 2008; Boston, 14. Eron J, Yeni P, Gathe J, Estrada V, DeJesus E, MA. Abstract 37. Staszewski S, et al. The KLEAN study of fosamprenavi- 25. Gathe J, Ive P, Wood R, Schurmann D, Bellos N, DeJesus E, ritonavir versus lopinavir-ritonavir, each in combination et al. SOLO: 48-week efficacy and safety comparison of with abacavir- lamivudine, or initial treatment of HIV infec- once-daily fosamprenavir/ritonavir versus twice-daily nelfi navir tion over 48 weeks: A randomised non-inferiority trial. in naïve HIV-1 infected patients. AIDS. 2004; 18:1529–1537. Lancet. 2006;368:476–482. 26. Johnson MA, Gathe JC, Podzamczer D, Molina J, Naylor 15. Gallant J, DeJesus E, Arribas J, Pozniak A, et al. Tenofovir C, Chiu Y, et al. A once-daily lopinavir/ritonavir based DF, emtricitabine and efavirenz vs zidovudine, lamivudine regimen provides noninferior antiviral activity compared and efavirenz for HIV. N Engl J Med. 2006;354:251–260. with a twice-daily regimen. J Acquir Immune Defi c Syndr. 16. Gallant JE, Staszewski S, Pozniak AL, et al. Effi cacy and 2006;43:153–160. safety of tenofovir DF vs stavudine in combination therapy 27. Gathe J, daSilva B, Loufty M, Podzdamczer D, Rubio R, in antiretroviral-naive patients: A 3-year randomized trial. Gibbs S, et al. Study M05-730 primary efficacy results JAMA. 2004;292(2):191–201. at week 48: Phase 3, randomized, open-label study of 17. Martinez E, Arranz J, Podzdamczer D, Ribera E, Knobel H, lopinavir-ritonavir (LPV/r) tablets once-daily (OD) versus Roca V, et al. Effi cacy and safety of NRTI’s switch to teno- twice-daily (BID), co-administered with tenofovir (TDF) + fovir plus emtricitabine versus abacavir plus lamivudine in emtricitabine (FTC) in antiretroviral naïve (ARV) HIV-1 patients with virologic suppression receiving a lamivudine infected subjects. In: Program and abstracts of the 15th containing HAART: The BICOMBO study. In: Program and Conference on Retroviruses and Opportunistic Infections abstracts of the Fourth IAS Conference on HIV Patho- (CROI); February 2008; Boston, MA. Abstract 775. genesis, Treatment and Prevention; July 2007; Sydney, 28. Smith K, Fine D, Patel P, Bellos N, Sloan L, Lackey P, et al. Australia. Abstract WESS101. Effi cacy and safety of abacavir/lamivudine compared to 18. van Leth F, Phanuphak P, Stroes E, Gazzard B, Cahn P, tenofovir/emtriticitabine in combination with one-daily et al. Nevirapine and efavirenz elicit different changes in lopinavir-ritonavir through 48 weeks in the HEAT Study. In: lipid profi les in antiretroviral-therapy-naive patients infected Program and abstracts of the 15th Conference on Retro- with HIV-1. PLoS Med. Vol. 1, No. 1, e19. doi:10.1371/ viruses and Opportunistic Infections (CROI); February journal.pmed.0010019 2008; Boston, MA. Abstract 774. 19. Ruxrungtham K, Bellos N, Morales-Ramirez J, Timerman A, 29. Hicks C, De Jesus E, Wohl D, Liao Q, Pappa K, Lancaster T. Madruga J, Katabira E, et al. The metabolic profile of Once-daily fosamprenavir boosted with either 100 mg or TMC278, an investigational non-nucleoside reverse tran- 200 mg of ritonavir along with abacavir/lamivudine: 48 scriptase inhibitor. In: Program and abstracts of the Fourth week safety and efficacy results from COL100758. In: IAS Conference on HIV Pathogenesis, Treatment and Pre- Program and abstracts of the 11th European AIDS Confer- vention; July 2007; Sydney, Australia. Abstract TUAB105. ence; October 2007; Madrid, Spain. Abstract PS 7.01. 20. Haubrich R, et al. Metabolic outcomes of ACTG 5142: 30. Elion R, DeJesus E, Sension M, Berger D, Towner W, A prospective randomised phase III trial of NRTI, PI- and Richmond G, et al. Once-daily abacavir/lamivudine NNRTI-sparing regimens for initial treatment of HIV-1 infec- (ABC/3TC) and boosted atazanavir (ATV/RTV) in antiretro- tion. In: Program and abstracts of the 14th Conference on viral-naive HIV-1 infected subjects: 48-week results from Retroviruses and Opportunistic Infections; February 2007; COL102060 (SHARE). Poster exhibition: 4th IAS Confer- Los Angeles, CA. Abstract 38. ence on HIV Pathogenesis, Treatment and Prevention; July 21. Grundy S, Cleeman J, Bairey N, Merz A, Brewer H, Clark L. 2007; Sydney, Australia. Abstract WEPEB033. Implications of recent clinical trials for the National 31. Cooper D, et al. Effi cacy and safety of two doses of tipra- Cholesterol Education Program Adult Treatment Panel III navir/ritonavir versus lopinavir/ritonavir-based therapy in Guidelines. Circulation. 2004;110:227–239. antiretroviral-naïve patients: Results of BI 1182.33. In: Pro- 22. Bartlett J, Fath M, DeMasi R, Hermes A, Quinn J, Mondou E. gram and abstracts of the Eighth International Congress An updated systematic overview of triple combination on Drug Therapy in HIV Infection; 2006; Glasgow. Abstract therapy in antiretroviral-naive HIV-infected adults. AIDS. PL13.4. 2006;20(16):2051–2064. 32. Tomaka F, Lefebvre E, Sekar V, van Baelen B, 23. Smith K, Weinberg W, DeJesus E, et al. Once-daily ritona- Vangeneugden T, Vendervoorde A, et al. Similar changes in vir (100 mg) boosting of fosamprenavir or atazanavir with metabolic parameters of darunavir (TMC114) and atazana- tenofovir /emticitabine in antiretroviral naïve HIV-infected vir, each coadmininstered with low-dose ritonavir in healthy 12 HIV CLINICAL TRIALS � 10/1 � J AN-FEB 2009 volunteers. HIV DART: Frontiers in Drug Development of versus efavirenz based combination therapy in treatment- Antiretroviral Therapies; December 2006; Cancun, Mexico. naïve HIV-1 infected patients. In: Program and abstracts Abstract 85. of the ICAAC-IDSA 2008 Conference; October 2008; 33. Vrouenraets S, Garcia E, Jackson A, Raffi F, Jayaweera D, Washington, DC. Abstract H896a. Katlama C, et al. Saquinavir versus atazanavir once daily, 36. Kuller L, et al. Elevated levels of interleukin-6 and D-dimer each boosted with ritonavir 100 mg and combined with are associated with an increased risk of death in patients tenofovir/emtricitabine, result in comparable lipid changes with HIV. In: Program and abstracts of the 15th Conference after 24 weeks in treatment naïve HIV infected patients. on Retroviruses and Opportunistic Infections; February Antiviral Ther. 2008;13(Suppl 4):A46. 2008; Boston, MA. Abstract 139. 34. Hense H. Observations, predictions and decisions – 37. Calmy A, et al. HIV activates markers of cardiovascular risk assessing cardiovascular risk assessment. Int J Epidemiol. in a randomized treatment interruption trial: STACCATO. 2004;33:235–239. In: Program and abstracts of the 15th Conference on 35. Lennox J, DeJesus E, Lazzarin A, Pollard R, Madruga J, Retroviruses and Opportunistic Infections; February 2008; Zhao J, et al. Safety and efficacy of raltegravir-based Boston, MA. Abstract 140.

Journal

HIV Clinical TrialsTaylor & Francis

Published: Feb 1, 2009

Keywords: cholesterol; HIV; lipids; meta-analysis; nucleoside analogues; protease inhibitors

There are no references for this article.