Salvage therapy with amprenavir, lopinavir and ritonavir 200 mg/d or 400 mg/d in HIV-infected patients in virological failure
note from Jules Levin: GSK recently made an announcement about amprenavir availability. I think availability will be limited to 50 mg pills and amprenavir liquid.
Study authors concluded:
"...The combination of amprenavir and lopinavir with 400 mg/d ritonavir had substantial antiviral efficacy and good tolerability in these heavily pretreated patients, despite a pharmacokinetic interaction between lopinavir and amprenavir and extensive genotypic and phenotypic resistance at baseline. The durable efficacy of the combination is reflected by the high proportion of patients (61% in the intent-to-treat analysis) in whom plasma HIV-1 RNA load was below 50 copies/ml after 26 weeks of treatment..." SEE BODY OF ARTICLE FOR FULL REVIEW OF ANTIVIRAL ACTIVITY and ADVERSE EVENTS.
From Jules Levin: Of note patients with >5x ULN ALT were excluded, the number of IDUs in study was small, and patients with 'hepatocellualr deficiency' were excluded:
"...Grade III/IV increases in serum triglyceride levels occurred in 22% (n=4) of patients receiving the 400 mg/d ritonavir dose (group 2) and in 37% (n=7) of patients receiving the 200 mg/d ritonavir dose (group 1) (P=0.33). Most grade III/IV triglyceride elevations (9/11) persisted at this level until week 26. Elevations of serum transaminase levels were observed in a minority of patients..."
"...clinicians testing new combinations of antiretroviral agents for which pharmacological data are not available should measure drug concentrations in order to detect possible pharmacokinetic interactions prior to routine use...the clinical significance of a given drug--drug interaction cannot be determined simply from the magnitude of the change in drug concentrations. The latter should always be determined in clinical studies..."
Antiviral Therapy 9:615--625 August 2004
Gilles Raguin1,2*, Geneviève Chêne3, Laurence Morand-Joubert4, Anne-Marie Taburet5, Cécile Droz3, Clotilde Le Tiec5, François Clavel6, Pierre-Marie Girard1 and the Puzzle 1 Study Group
1Service de Maladies Infectieuses, Hôpital Saint-Antoine, Assistance Publique-Hôpitaux de Paris, Paris, France
2Departement de Medecine, Hôpital Croix-St-Simon, Paris, France
3Inserm U593, Université Victor Segalen Bordeaux 2, Bordeaux, France
4Service de Bactériologie--Virologie, Hôpital St Antoine, Université Paris VI, Paris, France
5Service de Pharmacie, Hôpital Bicêtre, Le Kremlin-Bicêtre, France
6Inserm U552, Hôpital Bichat, Assistance Publique-Hôpitaux de Paris, Paris, France
Objectives: To compare the antiviral efficacy of a salvage therapy combining lopinavir and amprenavir with 200 mg/d or 400 mg/d ritonavir, together with nucleoside reverse transcriptase inhibitors, over a 26-week period in HIV-infected patients in whom multiple antiretroviral regimens had failed.
Design: Phase IIb, randomized, open-label, multicentre trial. Patients were eligible if they had <500 CD4+ cells/mm3 and >4 log10 copies/ml HIV-RNA after treatment with at least two protease inhibitors (PIs) and one non-nucleoside reverse transcriptase inhibitor.
Results: At baseline (n=37), the median CD4+ cell count was 207/mm3 and the median plasma HIV-1 RNA level was 4.7 log10 copies/ml; the median number of PI mutations was seven and the median decrease in phenotypic susceptibility to lopinavir and amprenavir was 9.7 and 2.6, respectively. The mean number of antiretrovirals received prior to randomization was 7.7.
The fall in the median HIV-1 RNA level at week 26 was --1.4 log10 copies/ml in the 200 mg/d ritonavir group and --2.5 log10 copies/ml in the 400 mg/d group (P=0.02). Viral load fell below 50 copies/ml in 32% and 61% of patients, respectively (P=0.07).
After adjustment for the ritonavir dose, a smaller number of PI mutations was the only baseline characteristic associated with a better virological response at week 26.
Amprenavir concentrations were significantly lower in presence of lopinavir. The lopinavir inhibitory quotient at week 6 correlated weakly with the change in the HIV-RNA level at week 26.
Conclusion: Combination of amprenavir, lopinavir and 400 mg/d ritonavir shows significant virological efficacy without increased toxicity in HIV-infected patients in whom multiple antiretroviral regimens have failed.
Sponsorship: supported by ANRS (National Agency for AIDS Research), Paris, France.
Combination antiretroviral therapy often fails to suppress HIV replication durably. Viral breakthrough occurs in an estimated 20% to 50% of patients within the first year of triple-drug therapy, owing to poor adherence (often related to adverse effects) and the emergence of drug-resistant genetic variants. The risk of virological failure increases markedly with successive antiretroviral regimens. For patients who are in virological failure after several antiretroviral regimens, current treatment guidelines recommend the use of new drug combinations, including investigational drugs when possible. Most clinical trials in this setting have evaluated the impact of salvage regimens containing only one investigational drug, which rarely drive viral load below the detection limit of current assays. One potential problem with combinations of new protease inhibitors (PIs) is the risk of negative pharmacokinetic interactions.
Lopinavir is a novel HIV-1 protease inhibitor that is approximately 10 times more potent than ritonavir. Its combination with low doses of ritonavir results in substantially increased lopinavir exposure. In these conditions, it is assumed that lopinavir can erect a higher barrier to resistance, by virtue of higher and more stable plasma concentrations.
Amprenavir is a PI with a unique in vitro resistance profile and limited cross-resistance to saquinavir and indinavir in vitro, raising the possibility that viruses resistant to these latter drugs may still be susceptible to amprenavir. Co-administration of ritonavir with amprenavir significantly increases amprenavir exposure. Preliminary data from small, short-term, retrospective, non-comparative clinical trials suggest that lopinavir/ritonavir could reduce the high amprenavir concentrations observed during ritonavir boosting. The effect of an additional dose of 200 mg/d ritonavir is unknown.
The objective of the Puzzle 1-ANRS 104 study was to compare virological suppression during 26 weeks of salvage therapy combining lopinavir and amprenavir with either 200 mg/d or 400 mg/d ritonavir, along with nucleoside reverse transcriptase inhibitors (NRTIs), in patients carrying multidrug-resistant isolates.
The Puzzle 1-ANRS 104 study was a prospective, randomized, open-label, multi-centre clinical trial. Patients were randomized centrally to receive a total ritonavir dose of 200 mg/d (group 1) or 400 mg/d (group 2), using a computer-generated random-number list. The randomization ratio was 1:1 and was stratified according to the clinical centre. In each of the two arms, patients were further randomized to either amprenavir (1200 mg/d) or lopinavir (800 mg/d) for 2 weeks before the second PI was added and the NRTIs were optimized on the basis of viral genotyping results and previous antiretroviral exposure. All PIs were administered twice daily. A pharmacokinetic-based objective of the study was to compare the trough levels of lopinavir and amprenavir when given alone (week 2 data) or with the other PI (data at weeks 6, 14 and 26). The study protocol was approved by the Ethics Committee of Hôpital Saint-Antoine, Université Paris VI.
Patients were recruited from 16 French clinical AIDS units. Adults with baseline CD4+ cell counts <500/mm3 and baseline plasma HIV-1 RNA >4 log10 copies/ml (10,000 copies/ml) after treatment with at least two PIs and one non-nucleoside reverse transcriptase inhibitor (NNRTI) and whose antiretroviral regimen had not been changed during the previous 3 months were eligible for the study. Noninclusion criteria included prior exposure to amprenavir or lopinavir, ongoing substance abuse, haemoglobin <9 g/dl, absolute neutrophil count <1000 × 106/l, platelet count <50000 ×106/l, alanine transaminase (ALT) or aspartate transaminase (AST) activity >5 times the upper limit of normal, and bilirubin, creatinine or lipase more than twice the upper limit of normal. Nonpregnant and non-lactating women were eligible if they agreed to use barrier birth control methods. Patients were not eligible in case of alcohol abuse, fructose intolerance, hepatocellular deficiency or a history of pancreatitis. Patients being treated for an acute opportunistic infection or a malignancy, and patients taking hydroxyurea, oestrogen, progesterone, rifampin or bepridil were also ineligible. All the patients signed an Ethics Committee-approved informed consent form.
Follow-up and laboratory measurements
Study visits were planned at weeks 2, 4 and 6, then monthly for the subsequent 20 weeks. Plasma HIV-1 RNA was assayed locally at weeks --2, 0, 2, 4, 6, 14 and 26, by using the Roche Amplicor HIV-1 Monitor kit (Roche, France; limit of quantitation (LOQ) 200 copies/ml) or the Quantiplex™ HIV-RNA 3.0 assay (Bayer Diagnostics, France; LOQ 50 copies/ml). Viral genotyping was done at weeks --2, 2, 6 and 26, based on complete sequencing of the reverse transcriptase and protease coding regions. The genotype was interpreted for each inhibitor by using the 2001 update of the French National Agency for AIDS Research (ANRS) algorithm. Viruses with definite or possible resistance to a given antiretroviral drug were considered resistant to that drug. A previously described lopinavir mutation score was also calculated for each patient. Resistance phenotyping was done at baseline in all the patients and at week 26 in 20/37 patients, using a recombinant virus assay (Phenoscript™; VIRalliance, France). CD4 and CD8 cell counts, and haematological and clinical chemistry measurements were performed at each study visit. Plasma trough concentrations (Cmin) of lopinavir, amprenavir and ritonavir were measured before the morning dose in all the patients at weeks 2, 4, 6, 14 and 26. Amprenavir, lopinavir and ritonavir were measured using a validated reverse-phase highperformance liquid chromatography method and UV detection after liquid--liquid extraction from alkalinized plasma. Amprenavir was assayed using a method initially described for indinavir. For simultaneous ritonavir and lopinavir assay, a washing step with hexane was included to allow measurements of low concentrations. The lower limits of amprenavir, lopinavir and ritonavir quantification were 40, 75 and 25 ng/ml, respectively. Day-to-day coefficients of variation, determined with quality controls (QC) included in each analytical run, were below 7% for high, medium and low QCs, with the exception of the low ritonavir QC (9.8%). Both assays were linear up to 10 000 ng/ml. The protein binding-corrected inhibitory quotient (IQ) was calculated as the steadystate unbound Cmin/IC90 ratio. The unbound Cmin was the product of the total plasma concentration and average unbound plasma fractions (0.10 and 0.01 for amprenavir and lopinavir, respectively).
The main comparison in this trial was between the patients receiving 200 mg/d ritonavir (group 1) and patients receiving 400 mg/d ritonavir (group 2). It was calculated that 26 subjects were required in each arm to detect a difference of at least 0.5 log10 copies/ml with 95% power and a type I error of 5% (two-sided). Enrolment started in November 2000 and ended in July 2001, when no patients had been randomized for 2 months owing to widespread access to amprenavir and lopinavir/ritonavir-based therapy in France. After a total of 40 patients had been randomized, the Scientific Committee decided to halt recruitment after consultation with the Data Safety and Monitoring Board. All patients who received at least one dose of the study drugs were included in the analysis. Missing data were considered to represent treatment failure in the intention-to-treat analysis. We used the Mann--Whitney two-sample statistics to compare changes in plasma HIV-1 RNA levels from baseline to week 26, and to compare other quantitative variables between arms. The Wilcoxon matched-pairs signed-ranks test was used for paired observations. HIV-RNA changes were studied between randomization and last visit, using a three-phase linear mixed model (week 0 to week 2, week 2 to week 6, week 6 to week 26). The effect of each potential determinant (age, baseline CD4+ cell count, AIDS stage, number of previous antiretrovirals, baseline number of genotypic resistance mutations up to 5 vs >5, lopinavir mutation score up to 5 vs >5, ritonavir dosage, initial PI received and plasma PI concentrations) was tested on the first slope (short-term response before 2 weeks with one new PI, whatever the ritonavir dose), the second slope (midterm response before 6 weeks and after the addition of a second PI and optimized NRTIs), and the third slope (long-term response after 6 weeks in the trial). All reported P values are based on two-sided tests.
Of the 40 patients randomized, 37 started treatment. The other three patients abandoned the trial between pre-inclusion and inclusion, and their decision was blinded to the randomization group. We therefore considered the analysis of the 37 patients who started treatment as an intention-to-treat analysis . Nineteen patients were allocated to a regimen containing a total daily ritonavir dose of 200 mg (group 1) and 18 to a regimen containing a total daily ritonavir dose of 400 mg (group 2). The two arms did not differ in terms of baseline characteristics. Overall, the mean CD4 cell count was 207 × 106/l and the mean baseline plasma HIV-1 RNA level was 4.7 log10 copies/ml. The median length of prior exposure to antiretroviral therapy was 94 months. In addition to NRTI, 35 patients had been exposed to at least two PIs and one NNRTI, and two patients had been exposed to two PIs only, with median exposure times of 53 months for PIs, 28 months for NNRTI and 94 months for NRTI.
The median number of NRTIs, NNRTIs and PIs administered was 5, 1 and 3, respectively. The mean number of baseline PI and reverse transcriptase mutations was 7 and 6.9, respectively. Table 2 below summarizes resistance mutations and proportions of patients with them. The proportion of patients with baseline genotypic resistance, using the 2001 ANRS-AC11 algorithm, were as follows: NRTIs: zidovudine (97%), lamivudine (65%), zalcitabine (43%), didanosine (89%), stavudine (89%) and abacavir (89%); NNRTIs: nevirapine (81%) and efavirenz (78%); and PIs: indinavir (73%), ritonavir (70%), nelfinavir (70%), saquinavir (65%), amprenavir (43%) and lopinavir (35%). The median decrease in phenotypic susceptibility was 9.7 (range 0.2; 95.3) for lopinavir and 2.6 (0.5; 24.3) for amprenavir.
Table 1. Baseline characteristics of the patients; Puzzle 1-ANRS 104 Study
Total, n=37 ; Group 1 (ritonavir 200 mg/d), n=19 ; Group 2 (ritonavir 400 mg/d), n=18
Median age in years (range) 42 (27--65) 43 (27--61) 41 (31--65)
Males, n (%) 33 (89) 18 (95) 15 (83)
HIV transmission category, n (%)
Heterosexuals 7 (18.9) 4 (21.0) 3 (17.0)
Men who have sex with men 25 (67.6) 12 (63.1) 13 (72.0)
Intravenous drug users 3 (8.1) 1 (5.3) 2 (11.0)
Others or unknown 2 (5.4) 2 (11.6) 0
CDC clinical stage, n (%)
A 16 (43) 5 (26) 11 (61)
B 7 (19) 6 (32) 1 (6)
C 14 (38) 8 (42) 6 (33)
Median CD4+ cells/mm3 (range) 207 (3--495) 220 (3--388) 192 (102--495)
Median HIV1-RNA log10 copies/ml (range) 4.7 (3.8--5.7) 4.9 (4.0--5.7) 4.6 (3.8--5.6)
Median number of antiretrovirals taken prior to inclusion (range) 7.8 (4--12) 7.6 (4--12) 8.0 (4--12)
Median number of antiretrovirals taken prior to and still in use at inclusion (range) 9.0 (7--12) 9.0 (7--12) 10.0 (8--12)
Median number of PIs taken prior to inclusion (range) 3.0 (2--5) 3.0 (2--4) 3.0 (2--5)
Median number of NNRTIs taken prior to inclusion (range) 1.0 (0--2) 1.0 (0--2) 1.0 (0--2)
Median number of NRTIs taken prior to inclusion (range) 5.0 (4--7) 5.0 (4--7) 5.0 (4--7)
Median (range) duration of prior exposure to PI in months 53 (28--71) 53 (42--57) 55 (28--71)
Median (range) duration of prior exposure to NNRTI in months 28 (0--39) 26 (0--37) 29 (0--39)
Median (range) duration of prior exposure to NRTI in months 94 (51--152) 78 (51--127) 100 (52--152)
Median number of protease mutations (range) 7.0 (1--10) 7.0 (1--9) 6.5 (1--10)
Median number of reverse transcriptase mutations (range) 6.9 (0--12) 7.3 (1--12) 6.6 (0--11)
Median total cholesterol (range)* 5.8 (3.0--7.6) 6.0 (4.4--7.6) 5.1 (3.0--6.8)
Median total triglyceride (range)* 1.8 (0.7--21.3) 1.6 (0.7--21.3) 1.8 (1.1--8.8)
NRTIs, nucleoside reverse transcriptase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor.
*Fasting value available for 10 patients in group 1 and for seven patients in group 2.
Antiviral activity and determinants of the virological response
A rapid decline in plasma HIV-1 RNA was observed in both treatment arms within the first 6 weeks, and persisted through week 26. The mean observed reduction from baseline was significantly different between the group receiving 400 mg/d ritonavir and the group receiving 200 mg/d ritonavir (--2.5 log10 copies/ml and --1.4 log10 copies/ml, respectively at week 26; P=0.02). At week 6, the mean decrease from baseline was --2.2 log10 copies/ml in patients receiving lopinavir+amprenavir+ritonavir 400 mg/d (group 2) and --1.5 log10 copies/ml in patients receiving lopinavir+amprenavir+ritonavir 200 mg/d (group 1). At 26 weeks, HIV-1 RNA was below the detection limit (<50 copies/ml) in 11/18 patients (61%) in the 400 mg/d ritonavir arm (group 2) and 6/19 patients (32%) in the 200 mg/d ritonavir arm (group 1) (P=0.07). A robustness analysis in which treatment was considered to have failed in the three patients who abandoned the trial before starting treatment (one in the 400 mg/d ritonavir arm and two in the 200 mg/d ritonavir arm), did not modify this conclusion (data not shown).
At week 2, the mean observed reduction from baseline was --0.9 log10 copies/ml in the group receiving lopinavir+ritonavir 200 mg/d (group 1a), --0.4 log10 copies/ml in the group receiving amprenavir+ritonavir 200 mg/d (group 1b), --1.3 log10 copies/ml in the group receiving lopinavir+ritonavir 400 mg/d (group 2a) and --1.2 log10 copies/ml in the group receiving amprenavir+ ritonavir 400 mg/d (group 2b). The decrease in HIV-1 RNA from baseline to week 2 was not significantly different between patients who started on lopinavir+ritonavir and those who started on amprenavir+ ritonavir (P=0.30). Based on a three-phase linear model of HIV-1 RNA changes (up to week 2, week 2 to week 6, and week 6 to week 26), the dose of ritonavir had a significant effect on the first slope, up to week 2 (P=0.04): predicted reduction in HIVRNA of --0.59 log10 copies/ml/week with amprenavir+ ritonavir 400 mg/d (group 2b) and --0.75 with lopinavir+ritonavir 400 mg/d (group 2a) versus --0.28 log10 copies/ml with amprenavir+ritonavir 200 mg/d (group 1b) and --0.44 with lopinavir+ritonavir 200 mg/d (group 1a). The predicted decline in HIV-RNA from week 2 to week 6 was also significantly different from zero but not as strong as in the initial 2 week period, suggesting an additional but less marked effect after the introduction of the second PI and optimized NRTIs. Finally, patients with more than five protease resistance mutations or a lopinavir mutation score higher than 5 at baseline had a significantly less steep slope than other patients (P=0.04 and P=0.006, respectively). Other characteristics such as age, baseline CD4+ cell count, AIDS stage, the number of prior antiretrovirals and plasma PI concentrations were not associated with changes in HIV-RNA. (p=0.92).
Table 2. Baseline genotypic resistance patterns: key positions of definite and possible resistance to each antiretroviral drug, using the latest ANRS-AC11 algorithm (2001); Puzzle 1-ANRS 104 Study
Number of patients 37
Key positions of reverse transcriptase mutations
Nucleoside reverse transcriptase inhibitor
| ||n (%) |
|M41L ||24 (65) |
|D67N ||25 (68) |
|T69SS ||2 (5) |
|K70R ||14 (38) |
|L74V ||11 (30) |
|V75A/M ||3 (8) |
|Y115F ||1 (3) |
|Q151M ||2 (5) |
|M184V ||21 (57) |
|L210W ||16 (43) |
|T215F/Y ||32 (87) |
|K219E/Q ||19 (51) |
Non-nucleoside reverse transcriptase inhibitor
|L100I ||6 (16) |
|K101E ||1 (3) |
|K103N ||15 (41) |
|V106A ||1 (3) |
|Y181C/I ||10 (27) |
|G190A/E/S ||8 (22) |
|L10F/I/R/V ||25 (68) |
|K20M/R ||10 (27) |
|L24I ||1 (3) |
|D30N ||3 (8) |
|V32I ||2 (5) |
|L33F ||1 (3) |
|M36I ||18 (49) |
|M46I/L ||14 (38) |
|F53L ||3 (8) |
|I54V ||12 (32) |
|L63P ||30 (81) |
|A71I/L/T/V ||20 (54) |
|G73S ||5 (14) |
|V77I ||8 (22) |
|V82A/F/S/T ||15 (41) |
|I84V ||10 (27) |
|N88D/S ||3 (8) |
|L90M ||18 (49) |
CD4 cell count changes
At week 26 the median increase in the CD4 cell count was 123/mm3 (inter quartile range 48; 142) in the 400 mg/d ritonavir arm (group 2) and 115/mm3 (inter quartile range 0; 244) in the 200 mg/d arm (group 1) the same range as those measured at week 6, although with greater variability. At baseline, the median lopinavir IC90 for resistant isolates was 49.2 nM (30.9 ng/ml) with a range of 1.2 nM to >500 nM. The median protein binding-corrected lopinavir IQ at week 6 was 1.8 (0.02--157.5) (n=34). The median IC90 of amprenavir was 80.5 nM (40.7 ng/ml) with a range of 16.3 nM to 990.7 nM, and the median protein binding-corrected amprenavir IQ at week 6 was 2.44 (0.2--13) (n=34). We also examined the relationship of viral susceptibility, drug exposure and the combination of the two, that is, the IQ with virological outcome. The lopinavir IQ at week 6 correlated weakly with the change in viral load at week 26 (R2=0.19; P=0.01). The correlation between the amprenavir IQ and viral load at week 26 was even weaker and non significant (R2=0.05; P=0.23).
Tolerability was similar in the two arms (Table 6). During the 26 weeks of the study, 4/18 patients (22%) receiving 400 mg/d ritonavir discontinued treatment, compared with 4/19 patients (21%) receiving 200 mg/d ritonavir. The incidence of adverse events was also similar: 6/18 patients (33%) receiving 400 mg/d ritonavir had grade IV adverse events (n=7), compared with 9/19 (47%) patients (10 events) receiving 200 mg/d ritonavir. The proportion of events related to the study drugs were 6/7 and 6/10, respectively. The most common grade IV events were hypertriglyceridaemia (eight cases, four in each group), hospitalization (five cases) and neutropenia (two cases). At week 26, fasting serum triglyceride and total cholesterol levels were measured in 15 patients in each arm. Median total cholesterol and triglyceride levels (mmol/l) were similar in the two arms: 5.1 [197 mg/dL] (group 2; 400 mg/d ritonavir) versus 5.8 [224 mg/dL] (group 1; 200 mg/d ritonavir) (P=0.5), and 3.6 [318 mg/dL] (group 2) versus 3.8 [336 mg/dL] (group 1) (P=0.8). One patient discontinued the study drugs for 30 days because of high triglycerides levels. No deaths occurred during the study period. Eight patients developed HIV-related illnesses, comprising oral candidiasis (n=4), herpes simplex (n=2), peripheral neuropathy (n=1) and relapse of Kaposi's sarcoma (n=1). There was no significant difference in the frequency of such events between the two arms.
The combination of amprenavir and lopinavir with 400 mg/d ritonavir had substantial antiviral efficacy and good tolerability in these heavily pretreated patients, despite a pharmacokinetic interaction between lopinavir and amprenavir and extensive genotypic and phenotypic resistance at baseline. The durable efficacy of the combination is reflected by the high proportion of patients (61% in the intent-to-treat analysis) in whom plasma HIV-1 RNA load was below 50 copies/ml after 26 weeks of treatment.
Our results compare very favourably with those of other clinical trials of salvage regimens for patients in whom multiple antiretrovirals had failed. For example, ACTG 359 was a randomized comparison of six oral antiretroviral regimens based on combinations of saquinavir soft gel with ritonavir or nelfinavir, together with delavirdine and/or adefovir dipivoxil, in indinavir-experienced patients with virological failure. The most effective regimen yielded HIV-RNA values below 500 copies/ml at week 16 in 40% of patients. In the ACTG 398 trial, Hammer et al. examined whether adding a second PI improved the antiviral efficacy of a four-drug combination in patients in whom virological failure had occurred on a PI-containing regimen. The proportion of patients with viral load below 200 copies/ml at 24 weeks in the dual- PI arms was only 35%. In an uncontrolled trial, Montaner et al. tested a 'multiple drug rescue therapy' containing five to nine antiretroviral agents in 106 heavily pretreated patients. In the intention-to continue analysis, approximately 28% of patients had plasma HIV-RNA values below 50 copies/ml at 1 year. Simpler and better-tolerated regimens based on resistance testing and pharmacokinetic profiling may be preferable. In our study, salvage therapy combining two new PIs and 400 mg/d ritonavir yielded better virological responses, with viral replication being maximally suppressed (<50 copies/ml) in a high proportion of subjects (45%).
The antiviral activity of lopinavir and amprenavir, combined with either 200 or 400 mg/d ritonavir, was illustrated by the decline in plasma HIV-1 RNA during the first two weeks of treatment. In a context of large proportions of undetectable HIV-1 RNA measurements, our analysis of plasma viral load decline might have yielded conservative results of the treatment effect. However, since baseline mean HIV-1 RNA was rather high and most sensitive tests available were used, we believe that the bias towards a null difference was probably limited and that our results reflect a fairly robust estimation of the effect of 400 mg/d of ritonavir versus 200 mg/d in this context. The threephase linear model showed that most of the viral load decline occurred during the first 2 weeks and was therefore a result of treatment with a single boosted PI regimen (lopinavir/ritonavir or amprenavir/ritonavir), although the decline in patients receiving the 400 mg/d ritonavir boost was significantly larger. There was no significant difference whether the PI was lopinavir or amprenavir. Not surprisingly, baseline resistance to multiple drugs was frequent in our patients. Several studies have shown that baseline antiretroviral resistance genotyping is predictive of the virological response in pretreated patients. In contrast, the prognostic value of phenotypic resistance testing is controversial. In our study, the 400 mg/d ritonavir dose and a smaller number of protease mutations at baseline (<=5), were associated with a better virological response at week 26, contrary to baseline amprenavir and lopinavir phenotypic resistance indices. Both the total number of baseline protease mutations and the previously constructed lopinavir mutation score were associated with the virological response in two separate analyses, with a higher risk of virological failure with six or more mutations. Concerning the lopinavir mutations score, similar data were obtained in another study with a similar patient population.
Pharmacokinetics, drug metabolism and drug--drug interactions are also important factors when selecting combinations for patients with virological failure. Many combinations are designed to exploit drug--drug interactions between substrates and inhibitors of cytochrome P450-catalysed oxidative metabolism.
Unfortunately, pharmacokinetic data are rarely available for interactions among drugs used in antiretroviral combinations. We found that the median amprenavir Cmin was significantly lower after the addition of lopinavir, tending to confirm the existence of a pharmacokinetic interaction between the two drugs. This interaction has been described in more detail in a subgroup of patients who agreed to participate in a pharmacokinetic study. As previously reported, we found that amprenavir 1200 mg/d + ritonavir 200 mg/d yielded trough concentrations three to four times higher than the standard amprenavir 1200 mg twice daily regimen. Increasing the ritonavir dose from 200 mg/d to 400 mg/d had little effect on amprenavir exposure. The magnitude of the decrease in amprenavir concentrations was highly variable among the patients. However, the mean trough plasma concentrations of amprenavir, when this drug was administered with lopinavir/ritonavir, were still higher than the estimated in vivo trough concentrations calculated to yield 90% of the maximum antiviral effect (228 ng/ml) over 4 weeks. Furthermore, the median IQ was 2.44 and remained above 1 in most patients.
The mean trough plasma concentrations of lopinavir observed with the 200 mg/d dose of ritonavir were within the range of those previously reported with the standard 800/200 dose regimen, despite the capacity of amprenavir to decrease the lopinavir plasma concentrations slightly in patients receiving 200 mg/d of ritonavir. This effect could be due to an inducing action of amprenavir on CYP3A as recently reported with the amprenavir+delavirdine combination. Interestingly, increasing the ritonavir dose to 400 mg/d led to an increase in the lopinavir concentration, suggesting that the ritonavir inhibitory effect on lopinavir metabolism might outweigh the amprenavir inducing effect on lopinavir metabolism. As protease inhibitor levels correlate with the development of resistance, higher concentrations of lopinavir may overcome reduced susceptibility and explain, in part, the better virological efficacy observed with the higher dose of ritonavir.
Several lessons can be drawn from this study. Firstly, clinicians testing new combinations of antiretroviral agents for which pharmacological data are not available should measure drug concentrations in order to detect possible pharmacokinetic interactions prior to routine use. Secondly, the clinical significance of a given drug--drug interaction cannot be determined simply from the magnitude of the change in drug concentrations. The latter should always be determined in clinical studies, along with their potential impact on virological--immunological outcome.
The good tolerability of the amprenavir/lopinavir/ritonavir combination is reflected by the fact that only one patient discontinued a study drug because of a related adverse clinical or biological event during the 26 weeks of treatment. Grade III/IV increases in serum triglyceride levels occurred in 22% (n=4) of patients receiving the 400 mg/d ritonavir dose (group 2) and in 37% (n=7) of patients receiving the 200 mg/d ritonavir dose (group 1) (P=0.33). Most grade III/IV triglyceride elevations (9/11) persisted at this level until week 26. Elevations of serum transaminase levels were observed in a minority of patients.
In conclusion, our results show that the combination of amprenavir (1200 mg/d), lopinavir (800 mg/d) and ritonavir (400 mg/d) is durably potent and well tolerated in patients in whom multiple antiretrovirals have failed, yielding sustained virological responses in 60% of cases after 26 weeks of treatment.