ORIGINAL RESEARCH. Objective Nucleoside reverse transcriptase (NRTI) sparing is a favourable option for patients with NRTI failure or toxicity.

Transcription

1 HIV Medicine (2005), 6, ORIGINAL RESEARCH r 2005 British HIV Association Indinavir/ritonavir 800/100 mg bid and efavirenz 600 mg qd in patients failing treatment with combination nucleoside reverse transcriptase inhibitors: 96-week outcomes of HIV-NAT 009 MA Boyd, 1,2 U Siangphoe, 1 K Ruxrungtham, 1,3 CJ Duncombe, 1,2 M Stek, 4 JMA Lange, 1,5 DA Cooper 1,2 and P Phanuphak 1,3 1 The HIV Netherlands Australia Thailand Research Collaboration, the Thai Red Cross AIDS Research Centre, Bangkok, Thailand, 2 National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, Sydney, Australia, 3 Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, 4 Merck & Co. Pty. Ltd, Whitehouse Station, NJ, USA, and 5 International AIDS Therapy Evaluation Centre and Department of Infectious Diseases, Tropical Medicine and AIDS, Academic Medical Centre, Amsterdam, the Netherlands Objective Nucleoside reverse transcriptase (NRTI) sparing is a favourable option for patients with NRTI failure or toxicity. Methods Patients judged to be failing NRTI therapy were enrolled in a single-arm, open-label study of indinavir/ritonavir (IDV/r) 800/100 mg twice a day (bid) 1 efavirenz (EFV) 600 mg once a day (qd). The primary endpoint was the change in time-weighted average HIV RNA from baseline. The initial 48-week protocol was extended to 96 weeks by a single amendment. Analysis was by intention to treat. Results Sixty-one patients (23 female) were enrolled in the study. Baseline median inter-quartile range (IQR) NRTI exposure was 4.4 ( ) years; baseline median viral load was 4.09 log 10 HIV-1 RNA copies/ ml (range log 10 copies/ml); baseline median CD4 count was 169 cells/ml (range cells/ml). The mean (SD) change in time-weighted average HIV RNA from baseline at 48 and 96 weeks was 2.1 (0.7) and 2.1 (0.8) log 10 copies/ml respectively, resulting in 87% and 69% of patients with HIV RNA o50 copies/ml. Sixteen per cent of patients permanently ceased therapy and 26% underwent temporary drug interruptions because of study drug-related adverse events. Fastedlipid values rose significantly over the 96 weeks of study, as did median blood glucose and median serum creatinine levels. Twelve (20%) patients underwent IDV dose reduction, mainly because of nephrotoxicity (nine of 12 patients). Blood pressure values deteriorated following switch, but markers of nucleoside toxicity improved. Conclusions IDV/r 800/100 mg bid 1 EFV 600 mg qd gave a potent, durable response in these NRTI failures and was reasonably well tolerated. However, we observed adverse effects on renal, metabolic and blood pressure parameters. Lower doses of boosted IDV might improve toxicity while maintaining efficacy, and this possibility warrants further investigation. Keywords: antiretroviral toxicity, boosted protease inhibitor, non-nucleoside reverse transcriptase inhibitor, nucleoside failure, nucleoside sparing Received: 23 December 2004, accepted 21 April 2005 Correspondence: Dr Mark Boyd, Room 5D304.1, Department of Microbiology and Infectious Diseases, Level 5, Flinders Medical Centre, Flinders Drive, Bedford Park, 5042, Australia. Tel: ; fax: ; mark.boyd@fmc.sa.gov.au 410

2 IDV/r with EFV in NRTI failure 411 Introduction The current standard of antiretroviral therapy care for patients with HIV infection is generally combinations of two nucleoside reverse transcriptase inhibitors (NRTIs) with either a protease inhibitor (PI), often pharmacokinetically boosted with low-dose ritonavir, or a non-nucleoside reverse transcriptase inhibitor (NNRTI) [1]. In a number of countries in which the availability of antiretroviral therapy (ART) is limited, and where individuals often have to bear the costs of their medication and care, it is not uncommon for people with HIV infection to be treated with combinations of NRTIs (single, dual or triple) in order to gain some degree of benefit, even though full virological suppression to undetectable levels may not be achieved. This strategy almost inevitably leads to the development of NRTI resistance, potentially reducing the response to the same or other NRTIs in the future because of the development of broad cross-resistance within the class. All the currently available antiretroviral drugs are associated with toxicities and the development of adverse events, both short- and long-term, has become an increasing concern since the advent of potent combination therapy and the resultant improved life expectancy of people living with HIV infection. The NRTIs alone are associated with a wide range of tissue-specific toxicities, and these are thought to be related to the observation that the NRTIs exhibit cross-inhibition of human polymerases, in particular the human mitochondrial polymerase g, leading to a reduction in mitochondrial numbers and cellular dysfunction. There is, therefore, interest in the use of PI and NNRTI combinations as NRTI-sparing regimens for use either as first-line therapy or in situations in which NRTI-based regimens have failed because of toxicity or lack of potency. By the year 2000, a number of patients participating in early HIV Netherlands Australia Thailand Research Collaboration (HIV-NAT) studies examining the use of combination NRTIs were beginning to show evidence of both virological and immunological failure [2,3]. Most of these patients could not afford to fund their own PI-based salvage therapy at that time, and we elected to roll those consenting into a single-arm study designed to investigate the use of ritonavir-boosted indinavir (IDV/r) and efavirenz (EFV) without NRTIs in patients fulfilling agreed criteria for NRTI failure. Patients and methods Inclusion criteria were as follows: 18 years of age or older; documented and confirmed HIV-1 positive; ability to give informed consent; willing to use adequate contraception if of childbearing potential; and judged to have failed current or previous combination NRTI therapy, defined as follows: 1. a documented HIV viral load 5000 HIV-1 RNA copies/ml on combination NRTI therapy, or 2. a documented HIV viral load 1000 copies/ml on combination NRTI therapy, in addition to one or more of the following: (i) CD4 count baseline CD4 count at commencement of initial combination NRTI therapy; (ii) CD4 count 200 cells/ml; (iii) clinical treatment failure, with progression to AIDS or the presence of at least two Centers for Disease Control and Prevention (CDC) category B symptoms after a minimum of 3 months on combination NRTI therapy. The following laboratory values had to be obtained within 28 days of the baseline visit: haemoglobin 48.0 g/ dl, absolute neutrophil count 850 cells/ml, platelet count /mL, serum creatinine o2 times the upper limit of normal, total serum bilirubin within normal limits, aspartate aminotransferase (AST) and/or alanine aminotransferase (ALT) o5 times the upper limit of normal. Patients must have taken prior combination NRTI therapy continuously for a minimum of 3 months in order to be eligible for enrolment. Exclusion criteria were the need, or anticipated need, to use a number of specifically listed, contraindicated concomitant medications that had the potential to lead to drug drug interactions; use of other investigational agents, immunomodulatory agents or immunosuppressive therapy; an active opportunistic infection or malignancy not under adequate clinical control; significant hypersensitivity or other contraindication to any of the components of the study drugs; substance abuse which would interfere with patient compliance or safety; or any condition or history of any illness that might confound the results of the study, or pose additional risks in administering the study drugs. Screening was performed for the presence of tuberculosis by clinical examination, purified protein derivative (PPD) skin test and chest X-ray. Those with a positive PPD skin test without evidence of active tuberculosis on examination and chest X-ray were offered primary prophylaxis with isoniazid 300 mg daily for 9 months. Patients with a CD4 count o200 cells/ml or with HIV-related symptomatic disease received prophylaxis for Pneumocystis jirovecii pneumonia. HIV-NAT 009 was a single-arm open-label clinical trial. The Ethics Committee of the Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, approved the study, and written informed consent was obtained from all patients prior to enrolment. The study was initially designed as a 48-week trial, but was subsequently extended by means of a single-protocol amendment until week 96. All patients received indinavir (IDV) 400 mg, 2 capsules

3 412 MA Boyd et al. twice a day (bid) taken with ritonavir (RTV) 100 mg, 1 capsule bid and EFV 200 mg, 3 capsules once a day (qd) (at night) at the outset of the trial. The primary endpoint was the time-weighted average change in HIV RNA from baseline to weeks 48 and 96. Secondary endpoints were the proportion of participants with undetectable HIV RNA at weeks 48 and 96, the timeweighted average change in CD4-cell counts from baseline to weeks 48 and 96, and the development of drug-related toxicities sufficiently severe to warrant study-drug interruption or permanent discontinuation. After screening and study entry, patients had a clinical assessment at week 2 (without scheduled blood draw) to assess the tolerability of the regimen and reinforce adherence to medication. Following this, further visits were scheduled at weeks 4, 8 and 12 and then every 12 weeks until week 96. At these time-points clinical findings, adverse events, and haematological, biochemical, immunological and virological parameters were evaluated. Serum creatinine was monitored at screening, at week 36 and at every study visit thereafter. Concentrations of fasted, total cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, estimated low-density lipoprotein (LDL) cholesterol (calculated using the Friedewald equation and excluding patients if triglyceride concentration at that visit was mmol/l) and blood glucose were assessed at each study visit. Patients had serum lactate measured at baseline, week-48 and week-96 visits according to a standardized procedure. Patients were instructed to fast for at least 8 h prior to each study visit. Plasma HIV RNA was assessed to 50 copies/ml (Chiron branched DNA assay version 3; Chiron Corporation, Emeryville, CA, USA). There was no allowance in the approved protocol for dose modification of IDV. However, in December 2002 an amendment was passed which allowed for pharmacokinetically guided dose optimizations of IDV in cases of nephrotoxicity. Patients were eligible for IDV dose optimization if they had HIV RNA o50 copies/ml, had evidence of nephrotoxicity defined as a syndrome of clinical nephrolithiasis (renal colic with or without haematuria) and/or a raised creatinine level above the upper reference limit of normal (i.e. 142 mmol/l), and fulfilled protocoldefined criteria for the peak IDV concentration (C max ) associated with nephrotoxicity in our experience ( 10 mg/ L), with a concomitant IDV trough concentration (C min )of 0.25 mg/l [4]. Patients underwent an IDV dose reduction to 600 mg bid initially, and pharmacokinetic levels (C min /C max ), serum creatinine and HIV RNA were rechecked 4 weeks later. If dose-reduction criteria were once again met on this lower IDV dose, a second dose reduction to IDV (400 mg) was allowed and IDV pharmacokinetic levels, serum creatinine and HIV RNA were rechecked 4 weeks later. If pharmacokinetic results did not completely fit dose-optimization criteria but an adjustment was thought necessary to fully optimize IDV pharmacokinetics, the clinician was at liberty to adjust the RTV dose to 200 mg bid, with follow-up monitoring as per protocol following dose adjustment. In all cases patients were instructed to ingest the IDV together with RTV. Adherence was monitored at every study visit. Patients were asked by the study physician to estimate the number of missed study-drug doses over the preceding week and since the previous, scheduled study visit. Efficacy analysis was based on the intention-to-treat (ITT) population using a last value carried forward (LOCF) imputation for continuous variables and a missing equals failure (M 5 F) imputation made for dichotomous variables. Time-weighted average change from baseline in viral load (HIV-1 RNA log copies/ml) and change in CD4 count to weeks 48 and 96 were compared by paired t-tests. Median-absolute laboratory result changes from baseline were evaluated using the Wilcoxon signed rank test. Correlations between blood pressure and serum creatinine were assessed by calculating Spearman rank-correlation coefficients. Adherence evaluation was assessed as an average rate with standard deviation (SD) over the week before the study visit or since the prior scheduled study visit. All tests and confidence intervals (CIs) were considered to be significant at a P-value of 0.05 (twotailed test). Statistical analysis was performed using SPSS version 9.0 (SPSS, Chicago, IL, USA). Results Between June and October 2001, 69 Thai HIV-seropositive patients were screened, of whom 61 were enrolled to commence treatment. All 61 patients commenced treatment and are included in this analysis. Baseline characteristics of the cohort are described in Table 1. The study cohort was two-thirds male with a mean (SD) age of 36.4 (6.5) years, a mean (SD) weight of 56.5 (9.7) kg, a median (IQR) CD4 count of 169 (60 277) cells/mm 3 and a median (IQR) log 10 plasma HIV-RNA copies/ml of 4.09 ( ). Fifty-six per cent of patients had symptomatic HIV disease, 20% had a prior AIDS diagnosis and 25% were asymptomatic. The median duration of prior NRTI exposure was 4.4 years (IQR ). The NRTI combinations to which patients had been exposed included zidovudine/lamivudine (ZDV/3TC), zidovudine/ didanosine (ZDV/ddI), stavudine/didanosine (d4t/ddi), ZDV/ddI/3TC, d4t/ddi/3tc and stavudine/lamivudine/abacavir (d4t/3tc/abc). Thirty-four patients (55.7%) had been exposed to more than one combination of NRTIs. The disposition of patients over the course of the study, and the reasons for permanent study withdrawal, are presented in

4 IDV/r with EFV in NRTI failure 413 Table 1 Baseline characteristics for the 61 patients included in analyses Characteristic Sex (female/male) [n (%)] 23/38 (38/62) Age (years) [mean (SD)] 36.4 (6.5) Weight (kg) [mean (SD)] 56.5 (9.7) BMI (kg/m 2 ) [mean (SD)] 21.6 (2.9) BMI o22 kg/m 2 [n (%)] 25 (41.0) CD4 count (cells/ml) [median (IQR)] 169 (60 277) Plasma viral load (log 10 HIV-1 RNA copies/ml) [median (IQR)] 4.09 ( ) CDC stage [n (%)] A 15 (24.6) B 34 (55.7) C 12 (19.7) Exposure category [n (%)] Heterosexual contact 57 (94) Homosexual contact 2 (3) Unknown 2 (3) Duration of NRTI exposure (years) [median (IQR)] 4.4 ( ) Number of NRTIs [n (%)] Two 22 (36) Three 7 (12) Four 30 (49) Five 2 (3) Number of prior regimens [n (%)] One 27 (44) Two 30 (49) Three 4 (7) Last therapy [n (%)] Dual 57 (93) Triple 4 (7) BMI, body mass index; CDC, Centers for Disease Control and Prevention; IQR, interquartile range; SD, standard deviation. Fig. 1. Ten patients (16.4%) permanently withdrew from the study over the 96-week period of the study. Fourteen patients (23%) had to temporarily discontinue therapy on 20 separate occasions as a result of adverse events; 10 patients interrupted on one occasion, two interrupted twice and two others interrupted three times. The reasons for study interruption were nephrotoxicity in six cases, gastrointestinal intolerance in six cases, central nervous system side effects in five cases and hepatotoxicity in three cases. Of the five patients withdrawn for virological failure, four demonstrated the K103N mutation associated with high-level resistance to all NNRTI agents; one patient demonstrated an L63P mutation in the protease gene in the absence of any resistance mutation associated with resistance to NNRTIs. This L63P mutation is reported to confer minor resistance to multiple PIs when found in combination with a major protease mutation. However, no primary mutation was found. A total of 435 specifically reported adverse events were documented during the study that were considered to be at least possibly related to the study medications (Fig. 2). Overall, 85% of patients experienced central nervous symptom adverse events, 82% experienced cutaneous adverse events, 64% experienced gastrointestinal adverse events, and 16% experienced musculoskeletal adverse events, all of which were of mild to moderate [AIDS Clinical Trial Group (ACTG) grade 1 2] severity. Fifty-nine per cent of patients experienced urogenital adverse events, which were reported as serious (ACTG grade 3 4) in three patients because of one or more episodes of renal colic requiring hospitalization for management. Four patients had to stop therapy because of renal colic; one patient on one occasion, another patient on two occasions and two patients on three separate occasions each. Twenty-four patients (39%) had laboratory adverse events, 22 of which were classified as serious (grade 3 4). The majority of these serious laboratory adverse events (15 of 24; 63%) were represented by raised, fasted triglycerides to a level 48.5 mmol/l. After 48 and 96 weeks of study, the changes in timeweighted average viral load from baseline were 2.1 log 10 copies/ml (SD 0.7) and 2.1 log 10 copies/ml (SD 0.8) (week 48 versus week 96, P ), respectively, and these changes resulted in 87% and 69% of the cohort with viral loado50 copies/ml (Fig. 3). By weeks 48 and 96, there were changes in timeweighted average CD4 count from baseline of 73.9 and cells/ml respectively (Po0.001 for changes compared with baseline) (Fig. 4). Table 2 shows the median (IQR) fasting-serum lipid and blood glucose values at baseline and at the week-96 timepoint. Median total cholesterol increased by 72% (IQR 45 90%) (Po0.001); the HDL-cholesterol fraction increased by 29% (IQR %) (Po0.001); the estimated LDLcholesterol fraction increased by 71% (IQR %) (Po0.001); the total cholesterol/hdl cholesterol (TC/HDLc) ratio increased by 16.1% (IQR 2.4 to 37.9%) (Po0.01); triglycerides increased by 151% (IQR %) (Po0.001); blood glucose increased by 5.3% (IQR %) (Po0.001), but no patient was diagnosed with newonset frank diabetes during the study. Utilizing guidelines for initiation of lipid-lowering therapy in HIV-infected patients, and using as a minimum requirement at least two fasted-lipid specimens collected on separate occasions, we found that 45 of 61 (77%) of the initially enrolled patients would qualify for lipid-lowering therapy if this were their only risk factor for cardiovascular disease [5]. Figure 5 shows renal function in three cohorts in the study. The cohort as a whole (left-hand panel) showed a gradual and significant deterioration in renal function from a median screening value of 78.7 mmol/l (IQR mmol/l) to 90.6 mmol/l (IQR mmol/l) at week 48 (Po0.01), rising further to 94.6 mmol/l (IQR mmol/l) (Po0.01) at week 96. The central panel

5 414 MA Boyd et al. week 0 week 24 week 48 week 72 Screened (n = 69) Enrolled (n = 61) IDV/RTV/EFV (n = 61) On study (n = 60) IDV/RTV/EFV (n = 60) On study (n = 58) IDV/RTV/EFV (n = 58) On study (n = 52) IDV/RTV/EFV (n = 52) Excluded (n = 8) - No viral failure with NRTI combination (n = 4) - No CD4 failure with NRTI combination (n = 4) Discontinued (n = 1) - HIV disease progression (n = 1) Discontinued (n = 2) - Virological failure (n = 2) Discontinued [n (n = 6] 6) - Virological failure [n (n = 2] 3) - Gastrointestinal intolerance (n [n = 2) 2] - Lost to follow up[n (n = 1] 1) week 96 On study (n = 51) IDV/RTV/EFV (n = 51) Discontinued (n = 1) - Lost to follow up (n = 1) Fig. 1 Study disposition. IDV, indinavir; RTV, ritonavir; EFV, efavirenz; NRTI, nucleoside reverse transcriptase inhibitor. shows median (IQR) serum creatinine values in the 12 patients who underwent dose optimization during the second 48 weeks of the study. In contrast to the whole cohort, these 12 patients demonstrated no significant rise in serum creatinine between the week-48 and week-96 time-points following dose optimization (P for the difference between the week-48 and week-96 results). The right-hand panel shows median (IQR) serum creatinine values in those patients who did not meet the criteria for dose optimization and therefore did not undergo IDVguided dose optimizations. These patients demonstrated a gradual decline in renal function over the first 48 weeks of the study, which continued to week 96 [screening creatinine 78.7 mmol/l (IQR mmol/l); week-48 creatinine 84.9 mmol/l (IQR mmol/l); Po0.01 compared with the screening value; week-96 creatinine 91.9 mmol/l (IQR mmol/l); P compared with the week-48 value]. At baseline, the median systolic and diastolic blood pressures in the cohort were 110 mmhg (IQR mmhg) and 70 mmhg ( mmhg) respectively. At week 96, the median systolic blood pressure was still 110 mmhg but with an interquartile range of mmhg, representing an overall median change from baseline of 10 mmhg (0 20 mmhg) (Po0.001). At week 96, the median diastolic blood pressure was 80 mmhg, representing a change from baseline of 10 mmhg (0 10 mmhg) (Po0.001). We found that the rise in systolic blood pressure over the 96 weeks correlated with changes in creatinine over the same time frame [correlation coefficient (95% confidence intervals) ( ); P ]. However, we did not find a similar correlation when changes in diastolic blood pressure were correlated with changes in creatinine [correlation coefficient (95% confidence intervals) ( ); P ]. We observed an improvement in markers of NRTI toxicity following switch. Haemoglobin levels showed a significant improvement from a median baseline value of 13 g/dl (IQR g/dl) to 14 g/dl (IQR g/dl) at 96 weeks (P ). ALT levels also showed a decline after switch

6 IDV/r with EFV in NRTI failure 415 % of patients with adverse events Neurological Cutaneous Gastrointestinal Urogenital Laboratory Grades I II Grades III IV Musculoskeletal Fig. 2 Adverse events (grades 1 2 and grades 3 4). % of patients with HIV RNA < 50 copies/ml On study Weeks 68.8% Fig. 3 Percentage of patients with undetectable HIV RNA (intention-to-treat; missing equals failure). from a median baseline value of 31 U/L (IQR U/L) to 17 U/L (IQR U/L) at 96 weeks (Po0.01). Median serum lactate measurements, however, did not significantly change: week 0, 1.2 mmol/l (IQR mmol/l); week 96, 1.3 mmol/l (IQR mmol/l; P 5 0.5). The mean rates of patient self-reported adherence for all weeks preceding the study visit and for the intervening periods since the previous study visit week were 99% (SD 1.1%) and 99% (SD 0.9%) respectively. Discussion This study confirms that the use of IDV 800 mg bid, pharmacokinetically boosted with RTV 100 mg bid, and combined with EFV 600 mg qd, provides a robust and durable response in patients with prolonged prior exposure to various combinations of NRTIs. The virological results at weeks 48 and 96 (87% and 68% of the ITT population with viral load o50 copies/ml) compare favourably with the results of published trials of the use of a combination of an unboosted PI and an NNRTI with at least one NRTI in ARTexperienced patients. The percentage of patients with undetectable viral load at week 96 in our study was equal to the week-24 result recorded in the study of Haas et al. (68%) and the week-48 result of Albrecht et al. (67%) [6,7]. Moreover, the HIV-NAT 009 study used a stricter 50 copies/ ml limit to define the level of complete viral suppression. The results of our study therefore provide strong evidence that the combination of a boosted PI and an NNRTI is sufficiently potent and durable to be considered as a reasonable alternative to either a PI or NNRTI in combination with NRTIs for patients in whom the use of NRTIs is considered undesirable or contraindicated. However, the use of both PI and NNRTI classes following failure

7 416 MA Boyd et al. Time-weighted average change in CD4 (cells/µl) On study Weeks Fig. 4 Time-weighted average change in CD4 count over time (intention-to-treat; last value carried forward). Table 2 Median lipid and glucose values at weeks 0 and 96 Variable Week 0 (n 5 61) Week 96 (n 5 51) Change over 96 weeks (n 5 51) % change over 96 weeks (n 5 51) P-value for change over 96 weeks Type of lipid Total cholesterol (mmol/l) 4.27 ( ) 7.85 ( ) 2.87 ( ) (45 90) o0.001 HDL cholesterol (mmol/l) 0.91 ( ) 1.24 ( ) 0.39 ( ) ( ) o0.001 Estimated LDL cholesterol* 2.73 ( ) 4.65 ( ) 1.70 ( ) ( ) o0.001 (mmol/l) (n 5 60) (n 5 30) (n 5 30) (n 5 30) TC/HDL-c ratio 4.77 ( ) 6.02 ( ) 0.90 ( ) ( ) o0.001 Triglycerides (mmol/l) 1.41 ( ) 4.01 ( ) 2.38 ( ) ( ) o0.001 Blood glucose (mmol/l) 5.00 ( ) 5.44 ( ) 0.28 (0 0.72) 5.34 ( ) o0.001 Median values with interquartile range in parentheses are shown. HDL, high-density lipoprotein; HDL-c, high-density lipoprotein cholesterol; LDL, low-density lipoprotein; TC, total cholesterol. *Estimated using the Friedwald equation: [LDL] 5 [TC] [HDL] ([TG]/2.2), excluding patients if triglycerides (TG) at that visit mmol/l. of NRTIs exposes patients to both remaining oral forms of currently available ART and therefore raises concerns about the potential for induction of multiclass resistance if patients were to fail the regimen. In our study, of the five patients judged to have virologically failed over the 96 weeks, four demonstrated a typical NNRTI mutation (K103N) in the absence of PI mutations; the other patient had secondary PI mutations present but primary mutations were not demonstrated. Therefore, in the course of a routine 3-monthly clinical follow-up for patients using this combination, the development of major resistance to the PI class seems unlikely. Another salvage option which has recently been investigated in patients with multiclass ART failure is the use of double-boosted PIs, and their use in the setting of NRTI failure may be more favourable so as to spare exposure to the NNRTI class. Combination PI therapy is limited by the complex and unpredictable pharmacokinetic interactions that occur between some agents as well as the risk of increased toxicity, particularly dyslipidaemia. However, studies performed to date with select agents have shown good pharmacokinetic profiles and potent virological responses despite baseline resistance PI mutations, and in the absence of prior PI exposure would probably fully suppress viral replication if tolerated [8 10]. To date, however, no studies of double-boosted PIs in NRTI failures alone have been published. Apart from the intrinsic potency of the regimen, there are, possibly, three other factors that may have contributed to the high rate of complete virological suppression in this study. Firstly, analysis of patient-reported adherence in this study suggested that patients were adhering to therapy at a level of 99% over the course of the study. Although patient-reported adherence probably represents an overestimate of actual medication adherence, the excellent virological responses observed on study combined with the patient-estimated adherence rates suggest that patients were generally adhering at a level of 95% or greater, which has been found to be associated with successful treatment outcomes [11]. Secondly, a pharmacokinetic sub-study

8 IDV/r with EFV in NRTI failure Whole cohort IDV-down dosers IDV-nondown dosers Median serum creatinine [mmol/l (IQR)] ** NS ** ** ** * *P < 0.05 **P < 0.01 Upper limit of normal range 0 Week sc sc sc On study Fig. 5 Serum creatinine changes. The IDV-down dosers box and whisker graphs (central panel) indicate the results in those patients who underwent indinavir (IDV) dose optimization in the second 48 weeks of study. The IDV-nondown dosers box and whisker graphs (far right panel) indicate the results in all patients in the study who continued using IDV 800 mg for the full 96 weeks of study. sc, screening visit; IQR, interquartile range. performed in 20 study participants suggested that, despite the addition of EFV, the pharmacokinetics of the RTVboosted IDV and EFV were more than adequate to provide adequate antiretroviral potency [12]. Thirdly, patients may have demonstrated a hypersusceptible response to the EFV. This phenomenon has been reported in association with multiple NRTI mutations, including the typical ZDV mutations, which often confer broad cross-resistance within the NRTI class [13]. While we did not formally assess NRTI resistance in patients at baseline, it is likely that the majority had significant degrees of NRTI resistance, given that the median time of prior NRTI exposure was 4.3 years (IQR ) and that all were failing NRTIs at a viral load of copies/ml. The study combination was reasonably well tolerated, although we did observe a number of adverse events over the course of the 96 weeks, including nephrotoxicity and the development of a metabolic profile associated with increased risk of cardiovascular disease. Four patients (7%) were forced to interrupt therapy on at least one occasion because of disabling renal colic. Nephrotoxicity has been one of the major drawbacks to the use of IDV. In a previous randomized trial of RTV-boosted IDV dosed at 800/100 mg bid in comparison with 800 mg three times a day (tid) in combination with ZDV and 3TC conducted in a similar population of HIV-infected and NRTI-exposed subjects in Thailand, clinical nephrolithiasis was reported in 30% of patients (15 of 50) in the boosted IDV arm over a period of 112 weeks [14]. We had expected to see a significant reduction in clinical nephrolithiasis in the current study, as we predicted that the EFV would induce the hepatic cytochrome-mediated metabolism of IDV and result in lower pharmacokinetic exposure to the drug and thereby less nephrotoxicity. However, 7% of patients experienced episodes of nephrolithiasis and 20% underwent IDV dose reductions as a result of nephrotoxicity. A recent report has demonstrated that IDV can be given at doses lower than those currently recommended by the manufacturer, with responses to therapy at least comparable to that seen with recommended doses, little nephrotoxicity, and adequate pharmacokinetic parameters [15]. A study of IDV/r 400/ 100 mg plus d4t and 3TC in antiretroviral-naive patients in Thailand with low baseline CD4 counts and high viral loads has also demonstrated impressive early clinical outcomes and adequate pharmacokinetic parameters in the majority of patients [16]. The regimen was associated with a significant deterioration in metabolic parameters including an increase in the TC/HDL-c ratio, an indicator that is considered to be a powerful predictor for future vascular events [17]. The majority of patients (77%) qualified for institution of lipidlowering therapy according to recent international guidelines [5]. The increase in HDL cholesterol is likely to be related to an effect of EFV [18], although it might also be

9 418 MA Boyd et al. accounted for by improved virological control [19]. In addition to these lipid profile changes, the rise in glucose levels probably reflects insulin resistance, which is a known complication of IDV therapy, and which also confers increased risk of atherosclerosis [20]. A prospective observational study has recently demonstrated that combination ART is associated with a 26% relative increase in the rate of myocardial infarction per year of exposure to ART over the first 6 years of exposure [21]. The investigators found that higher total serum-cholesterol levels and higher triglyceride levels were associated with an increased incidence of myocardial infarction. Recent, further analyses of the same database suggest that the increased risk for myocardial infarction is mirrored by an increased risk for all cardiovascular events, and that the risk of an event increases with more time spent on combination ART [22,23]. The Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) Study Group has also presented an analysis that suggested that hypertension was not independently associated with the use of ART [24]. We observed a significant rise in both diastolic and systolic blood pressures which, in the absence of a known association with ART itself, is probably explained by the nephrotoxicity of IDV and the resultant significant rise in creatinine levels over the course of the study. An analysis of changes in serum creatinine level and blood pressure demonstrated a significant relationship between rises in creatinine level and changes in systolic (but not diastolic) blood pressure. There is a continuous, graded relation between blood pressure and the risk of cardiovascular disease [25]. The observed increases in systolic and diastolic blood pressures in addition to the presence of a potentially pro-atherogenic metabolic profile would combine to form a relatively high risk of premature atherosclerosis compared with the presence of either risk factor alone in patients using this antiretroviral regimen. Therefore, the routine monitoring of metabolic profile and resting blood pressure is advisable [5]. It is worth noting that newer agents available in both the PI (e.g. atazanavir) and NRTI (e.g. abacavir and tenofovir) classes appear to have less association with metabolic derangement and mitochondrial toxicity, and therefore the potential now exists to construct antiretroviral regimens with much lower potential for long-term adverse outcomes. Anaemia improved in the cohort after switch, and this may reflect improved well-being following immune reconstitution or a reversal of mitochondrial toxicity following withdrawal of NRTIs. Prior to switching, 42 of the 61 patients (68%) enrolled in the study were taking a ZDV-containing regimen and therefore it is not surprising, given the known association between the use of ZDV and anaemia, that haemoglobin levels demonstrated an increase from the baseline established while on combined NRTI therapy. The improvement in alanine aminotransferase (ALT) is probably related to the association between hepatitis and mitochondrial toxicity. Lactate measurements, however, did not significantly change from baseline, and this demonstrates that the routine measurement of lactate has poor sensitivity for mitochondrial toxicity [26]. Our study is limited by the small sample size, which does not allow an exhaustive investigation of factors predicative of virological response. It was also not randomized, and therefore we are unable to comment on the relative performance of this regimen in NRTI failures compared with other reasonable strategies such as a boosted PI with NRTIs, or a dual-boosted PI approach. The currently favoured first-line strategy for antiretroviral-naive patients is the use of an NNRTI in combination with two NRTIs [1], therefore making a boosted PI and NNRTI regimen an unlikely candidate for salvage therapy following virological failure because of possible NNRTI resistance. However, this regimen may be favourable for those who have failed to tolerate NRTIs because of short- or long-term adverse effects in the absence of virological failure. We may, therefore, conclude that, while the use of RTVboosted IDV and EFV provides a robust and durable virological and immunological response in patients with extensive prior experience with NRTIs, there is a price to be paid in terms of toxicity. Patient metabolic profiles became significantly deranged and 20% of the cohort experienced nephrotoxicity of a magnitude sufficient to warrant IDVdose reduction. In addition, the regimen had an adverse effect on both systolic and diastolic blood pressures, thereby creating, in combination with the metabolic disturbances, a markedly increased risk for cardiovascular disease. These adverse events limit the utility of this antiretroviral drug combination, at least at the doses initially used, and particularly in patients who have other risk factors for cardiovascular disease. In addition, given the need for dose reductions of IDV over the course of the study, our findings provide further evidence that IDV is best used in situations in which therapeutic drug monitoring is available [27]. However, in settings and situations in which therapeutic drug monitoring is not available, the controlled investigation of lower doses of IDV in this regimen may help to determine the optimal dose of IDV for any given setting or population. Lower doses of IDV would lower the cost of combination ART in both developed and less developed countries, would improve the toxicity profile of the drug, and may ultimately provide a favourable option for the provision of low-cost PIcontaining combination ART in the future.

10 IDV/r with EFV in NRTI failure 419 Acknowledgements The trial was supported by a grant from Merck & Co. Pty. Ltd, Whitehouse Station, NJ, USA. We acknowledge the contribution of all HIV-NAT staff, in particular T. Chuenyam, C. Ungsedhapand, S. Ubolyam, P. Srasuebkul, N. Chomchey, T. Methanukroh and T. Jupimai. We acknowledge the advice and assistance of P. Reiss from the Department of Infectious Diseases, Tropical Medicine and AIDS, Academic Medical Centre, Amsterdam, the Netherlands, and E. Hassink from the International Antiviral Therapy Evaluation Centre (IATEC), Amsterdam, the Netherlands. The help of C. Tan of Merck & Co. was appreciated, as were the contributions of S. Kitsiripornchai, D. Suwanchaiyong, W. Kunasirirat, S. Chatchawalchonteera and J. Supakitanakun from BLH Trading Co. Pty. Ltd, Thailand. Finally, we thank the participating study patients for their outstanding contribution. References 1 Yeni PG, Hammer SM, Hirsch MS et al. Treatment for Adult HIV Infection Recommendations of the International AIDS Society USA Panel. J Am Med Assoc 2004; 2: Ruxrungtham K, Kroon ED, Ungsedhapand C et al. A randomized, dose-finding study with didanosine plus stavudine versus didanosine alone in antiviral-naive, HIV-infected Thai patients. AIDS 2000; 14: Ungsedhapand C, Kroon ED, Suwanagool S et al. A randomized, open-label, comparative trial of zidovudine plus lamivudine versus zidovudine plus lamivudine plus didanosine in antiretroviral-naive HIV-1-infected Thai patients. J Acquir Immune Defic Syndr 2001; 27: Burger D, Boyd MA, Duncombe C et al. Pharmacokinetics and pharmacodynamics of indinavir with or without low-dose ritonavir. J Antimicrob Chemother 2003; 51: Schambelan M, Benson C, Carr A et al. Management of metabolic complications associated with antiretroviral therapy for HIV-1 infection: Recommendations of an International AIDS Society-USA Panel. J Acquir Immune Defic Syndr 2002; 31: Haas DW, Fessel WJ, Delapenha RA et al. Therapy with efavirenz plus indinavir in patients with extensive prior nucleoside reverse-transcriptase inhibitor experience: a randomized, double-blind, placebo-controlled trial. J Infect Dis 2001; 183: Albrecht MA, Bosch RJ, Hammer SM et al. Nelfinavir, efavirenz, or both after the failure of nucleoside treatment of HIV infection. N Engl J Med 2000; 345: La Porte CJ, Wasmuth JC, Schneider K et al. Lopinavir/ritonavir plus saquinvir in salvage therapy; pharmacokinetics, tolerability and efficacy. AIDS 2003; 17: Stephan C, Hentig N, Kourbeti I et al. Saquinavir drug exposure is not impaired by the boosted double protease inhibitor combination of lopinavir/ritonavir. AIDS 2004; 18: Boffito M, Kurowski M, Kruse G et al. Atazanavir enhances saquinavir hard-gel concentrations in a ritonavir-boosted once-daily regimen. AIDS 2004; 18: Paterson DL, Swindells S, Mohr J et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000; 133: Boyd MA, Aarnoutse RE, Ruxrungtham K et al. The pharmacokinetics of indinavir/ritonavir 800/100 mg in combination with efavirenz 600 mg in HIV-1 infected subjects. J Acquir Immune Defic Syndr 2003; 34: Shulman N, Zolopa AR, Passaro D et al. Phenotypic hypersensitivity to NNRTI in treatment-experienced HIVinfected patients: impact on virological response to efavirenz based therapy. AIDS 2001; 15: Boyd MA, Duncombe CJ, Ruxrungtham K et al. Indinavir TID vs indinavir/ritonavir BID in combination with AZT/3TC for HIV infection in nucleoside pretreated patients: HIV-NAT week follow-up. Program and Abstracts of the 9th Conference on Retroviruses and Opportunistic Infections. Seattle, WA, February 2002 [Abstract 422-W]. 15 Duvivier C, Myrto A, Marcelin AG et al. Efficacy and safety of ritonavir/indinavir 100/400 mg twice daily in combination with two nucleoside analogues in antiretroviral treatment-naive HIV-infected individuals. Antivir Ther 2004; 8: Boyd M, Piroon M, Burger D et al. Pharmacokinetics of reduced-dosed indinavir/ritonavir 400/100 mg twice-daily in HIV-1 infected Thai patients. Antivir Ther 2005; 10: Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. High density lipoprotein cholesterol as a protective factor against coronary artery disease. The Framingham Study. Am J Med 1977; 62: van Leth F, Hall D, Lange JMA, Reiss P. Increase in highdensity lipoprotein cholesterol when using nevirapine and/or efavirenz is maintained up to 2 years. Program and Abstracts of the XV International AIDS Conference. Bangkok, Thailand, July 2004 [Abstract WePpB2065]. 19 Grunfeld C, Pang M, Doerrler W et al. Lipids, lipoproteins, triglyceride clearance, and cytokines in human immunodeficiency virus infection and the acquired immunodeficiency syndrome. J Clin Endocrinol Metab 1992; 74: Noor MA, Seneviratne T, Aweeka FT et al. Indinavir acutely inhibits insulin-stimulated glucose disposal in humans: a randomized, placebo-controlled study. AIDS 2002; 29: F1 F8. 21 The DAD Study Group. Combination antiretroviral therapy and the risk of myocardial infarction. N Engl J Med 2003; 349:

11 420 MA Boyd et al. 22 Law MG, D Arminio Monforte A, Friis-Moller N et al. Cardio and cerebrovascular events and predicted rates of myocardial infarction in the DAD study. Program and Abstracts of the 11th Conference on Retroviruses and Opportunistic Infections. San Francisco, CA, February 2004 [Abstract 737]. 23 El-Sadr W, Reiss P, De Wit S et al. Relationship between prolonged exposure to combination ART and myocardial infarction: effect of sex, age, and lipid changes. Program and Abstracts of the 12th Conference on Retroviruses and Opportunistic Infections. Boston, MA, February 2005 [Abstract 42]. 24 Thiébaut R, El-Sadr W, Chenuc G et al. Predictors of hypertension and changes in blood pressure in HIV-infected patients in the DAD study. Program and Abstracts of the 11th Conference on Retroviruses and Opportunistic Infections. San Francisco, CA, February 2004 [Abstract 75]. 25 Stamler J Blood pressure and high blood pressure. Aspects of risk. Hypertension 1991; 18: Moyle GJ, Datta D, Mandalia S, Morlese J, Asboe D, Gazzard BG. Hyperlactataemia and lactic acidosis during antiretroviral therapy: relevance, reproducibility and possible risk factors. AIDS 2002; 16: Wasmuth JC, la Porte CJ, Schneider K, Burger DM, Rockstroh JK. Comparison of two reduced-dose regimens of indinavir (600 mg vs 400 mg twice daily) and ritonavir (100 mg twice daily) in healthy volunteers (COREDIR). Antivir Ther 2004; 9: