Lopinavir/ritonavir (LPV/r, Kaletra) is a protease inhibitor (PI) licensed for use as an antiretroviral agent in adults and children infected with HIV-1. The coformulation of LPV with a fixed dose of ritonavir, acting as a pharmacokinetic enhancer, results in substantially increased LPV drug exposure, potentially providing a pharmacologic barrier to the emergence of viral resistance. A pediatric formulation (oral solution) is also available. A pivotal randomized study in 653 antiretroviral-naive adults demonstrated superior antiviral efficacy of LPV/r compared with nelfinavir, with no evidence of development of resistance to LPV during the study duration (108 weeks) in the LPV/r arm. The use of LPV/r was also more often associated with a significant change in triglyceride concentrations, however.1 LPV/r has not been formally compared with other PIs in naive patients, including those that can be boosted by combination with a low dose of ritonavir. This drug is widely used as a second or subsequent line of treatment, often making it possible, in combination, to re-establish control over viral replication.2,3 Genotypic analysis of viral isolates from PI-experienced patients revealed 10 to 13 different codons in the protease gene associated with reduced sensitivity to LPV.4,5 Algorithms for interpreting viral genotypes have suggested that the efficacy of LPV/r treatment is significantly reduced by the presence of 6 or more mutations conferring resistance to LPV (www.hivfrenchresistance.org, www.hivdb.stanford.edu).
To date, only 1 noncomparative pediatric study of LPV/r has been conducted in 100 treatment-naive and -experienced children.8 Virologic results with LPV/r were clearly better in comparison to the results of other trials or observational pediatric studies already published. Although LPV/r is now widely used in children, no data are available concerning the predictive factors of virologic success.
The aim of this single-center observational study was to analyze and determine genotypic and pharmacologic predictive factors of virologic success. A prospective genotypic resistance analysis was also performed in patients who experienced virologic failure.
PATIENTS AND METHODS
From March 2000 to April 2001, the French Drug Agency authorized the prescription of LPV/r to patients with no other treatment option available through a temporary authorization for use (ATU). We analyzed the data collected from 69 consecutive children followed at Necker Hospital (Paris, France) who received LPV/r for at least 3 months during precommercialization use. Patients were considered to have failed treatment if the plasma HIV RNA exceeded 1000 copies/mL at least twice after at least 3 months of LPV/r treatment. Plasma HIV-RNA was quantified using the HIV RNA Cobas Amplicor kit (Roche Diagnostics, Meylan, France) at baseline, month 1, month 3, and every 3 months thereafter. The reverse transcriptase (RT) and protease gene sequences were determined from plasma samples with the Agence Nationale de Recherches sur le SIDA (ANRS) consensus method, as previously described.9 Genotypic resistance was analyzed before treatment with LPV/r and at the time of virologic failure. For each patient, the LPV mutation score was determined at baseline according to the ANRS resistance algorithm published in 2003 (available at: http://www.hivfrenchresistance.org). This score indicated the number of the following mutations in the protease gene: L10F/I/R/V, K20M/R, L24I, L33F, M46I/L, I50V, F53L, I54M/L/T/V, L63P, A71I/L/V/T, V82A/F/S/T, I84V, and L90M.
The plasma concentration of LPV was measured around the maximal concentration (Cmax) 3 to 5 hours after administration by high-performance liquid chromatography (HPLC).10 Only the first measurement obtained for each child was included in the analysis.
For each child, 2 factors were established. First, the score of “genotypically” susceptible drugs (or genotypic sensitivity score [GSS]) coadministered with LPV/r, according to the results of baseline genotypic resistance, was interpreted using the 2003 ANRS French algorithm: susceptible drugs (1), possibly resistant drugs (0.5), and resistant drugs (0).11 Next, the maximum genotypic inhibitory quotient (GIQmax), calculated as the ratio of the Cmax LPV to the baseline LPV mutation score, was calculated.12
Quantitative variables are expressed as means ± SD. The Student t test was used to compare the means of 2 groups, and the Spearman coefficient was used to determine correlation. The definition used for virologic failure before 18 months of follow-up is based on the Kaplan-Meier method. All comparisons between groups concerning this criterion were made with the log-rank test. The χ2 test or Fisher exact test was used for all other comparisons between groups. All P values are 2-tailed, and the threshold of significance was set at 0.05. Hazard ratios (HRs) concerning the relations between different factors and virologic failure, with their 95% confidence interval, were modeled using the Cox regression method in univariate and multivariate models. Analyses were performed with SAS software (SAS Institute, Cary, NC).
Sixty-nine children (41 boys and 28 girls, mean age: 10.3 ± 4.8 years) followed in Necker Hospital received LPV/r for at least 3 months between April 2000 and November 2001. The mean (SD) period of treatment with LPV/r was 16.5 ± 8.3 months (median = 13.9 months, range: 4.3–34.9 months). Six children had never received any antiretroviral agent, 15 had received multiple therapy without PIs, and 48 were PI experienced (Table 1). In this last group, 21 children had only received 1 PI (mostly nelfinavir) and 27 had received at least 2 PIs. As shown in Table 1, baseline virologic and immunologic characteristics were similar in PI-naive and -experienced children.
All PI-naive children received LPV/r combined with 2 nucleoside reverse transcriptase inhibitors (NRTIs). For PI-experienced children, LPV/r was associated with 2 NRTIs (n = 33) or 2 NRTIs plus a nonnucleoside reverse transcription inhibitor (NNRTI, n = 15). The mean dose of LPV/r was 10.2 ± 4.0 mg/kg administered twice per day. Twenty-two children received the oral solution, and 47 received the soft elastic capsule formulation.
The HIV-1 resistance genotypes were determined for 64 of the 69 children (19 of 21 PI-naive children and 45 of 48 PI-experienced children). The prevalence of baseline protease- and RT-associated resistance mutations is summarized in Figure 1. For PI-naive and -experienced children, the median baseline LPV mutation number was 1 (range: 0–3) and 2.5 (range: 0–8), respectively (see Fig. 1A). The median number of NRTI resistance mutations was 3 (range: 0–8) for all children, and there was at least 1 NNRTI mutation for 31 (48%) of 64 children (see Fig. 1B). In accordance with these initial genotyping data, the GSS for the associated drugs was ≥2 for 35 children, ≥1 and <2 for 20 children, and <1 for 9 children.
At month 6, 62% of children treated had a viral load of <400 copies/mL and 52% had a viral load <50 copies/mL. After 12 months, the respective proportions were 63% and 57%, and after 18 months, the proportions were 57% and 49%. Figure 2 details the evolution of viral load below 400 and 50 copies/mL in PI-naive and -experienced children. The cumulative risk of virologic failure was used to identify predictive factors of virologic response. During the follow-up, 25 (36%) children met the failure criteria, 4 of them after 9 months of treatment. Age, sex, baseline CD4 cell percentage and HIV-1 RNA, and formulation type (solution or capsule) were not associated with the risk of failure (data not shown). In contrast, previous treatment with another PI was strongly associated with the risk of virologic failure (P < 0.004, log-rank test).
Lopinavir Resistance Mutations and Risk of Virologic Failure
The relation between the score of LPV resistance mutations observed at baseline and the risk of virologic failure is shown in Figure 3. Virologic efficacy was not affected, or only slightly affected, by the presence of 1 to 3 mutations, although the small number of children in each group did not allow for statistically valid comparisons (see Fig. 3A). By combining data for children with <4 or ≥4 mutations, however, virologic efficacy was significantly different between the 2 groups (HR = 10.50, range: 4.29–25.70; P < 10−3; see Fig. 3B).
According to the univariate analysis, protease mutations at positions 10I/F, 46I, 54V/L, 71V/I, 82 A/F/S/T, and 90M were individually associated with the risk of failure (P < 0.01 each, log-rank test; Table 2). Conversely, PI mutations at positions 20M, 36I, 63P, and 77I were not associated with a higher risk of virologic failure (data not shown).
Antiretroviral Treatment Associated With Lopinavir/r
The GSS, corresponding to the number of active drugs coadministered with LPV/r according to the initial resistance genotype, was not predictive; the risk of virologic failure appeared to be slightly higher for a GSS less than 2, but the difference was not significant (P < 0.12). The GSS was not, however, independent of the number of LPV resistance mutations. Indeed, 65% of the children with 4 or more LPV resistance mutations had a GSS of <2 compared with 35% of children with <4 mutations (P < 0.03). Among the 48 PI-experienced children, 15 also received an NNRTI (9 received nevirapine and 6 received efavirenz), but no significant difference was evidenced with NNRTI use (P < 0.2).
Plasma Lopinavir Concentration
In 53 cases, the plasma LPV Cmax was measured during the first 6 months of treatment. The mean plasma concentration for all children did not differ according to the results of viral load at month 6. Stratification according to prior treatment revealed a significant difference. Indeed, in children previously treated with a PI, the plasma LPV Cmax was significantly higher in children with a viral load <400 copies/mL at month 6 (12.7 ± 6.4 vs. 7.7 ± 5.9 mg/L; P < 0.007). Based on the cumulative risk of virologic failure in the Cox model, the same difference for the predictive value of plasma LPV Cmax was correlated with previous PI experience. The observed relation for PI-pretreated children did not reach statistical significance, however (HR = 0.92, range: 0.85–1.01; P < 0.063; see Table 2).
Genotypic Inhibitory Quotient
The GIQmax was defined as the ratio of LPV Cmax to the baseline LPV mutation score. The mean value was 4.9 ± 4.4 ng/mL (range: 0–16 ng/mL). For all children as a group, a lower GIQmax was significantly associated with virologic failure (HR = 0.82, range: 0.70–0.96; P < 0.014; see Table 2). The relation was even stronger in the pretreated subgroup (HR = 0.63, range: 0.48–0.83; P < 0.001).
Multivariate Analysis of Factors Predictive of Virologic Failure
In a first model including all children, among parameters that included previous PI treatment, LPV mutation score, and GIQmax (Table 3), only the number of LPV resistance mutations was significantly related to virologic failure (HR = 8.29, range: 2.80–24.50; P < 10−3). For the group of children pretreated with PIs, however, the LPV mutation score and the GIQmax were independently associated with the virologic outcome (HR = 4.26, range: 1.14–15.92; P < 0.03 and HR = 0.71, range: 0.53–0.96; P < 0.02, respectively; see Table 3).
Evolution of the Protease Genotype in Patients Failing Treatment on Lopinavir/r
Resistance genotype was determined for 22 children (2 PI-naive children and 20 PI-experienced children) at the time of virologic failure. The median time-to-resistance analysis at failure was 38 weeks (range: 24–104 weeks). Five patients (2 PI-naive children and 3 PI-experienced children) exhibited no change in protease gene between baseline and the time of testing. The most frequently selected protease mutations were at codons 10 (27%), 46 (23%), and 54 (18%). Moreover, acquisition of major protease mutations at codons 46, 50, 82, 84, and 90 was observed in 7 of 22 (32%) PI-experienced children. There was a correlation between the accumulation of protease mutations and the number of weeks of viral replication greater than 1000 copies/mL on therapy (P = 0.03, r = 0.48). This result was also observed for 6 children with viral replication less than 10,000 copies/mL.
Overall virologic outcome in this noncomparative observational study was similar to that reported in the only pediatric study available with LPV/r.8 It favorably compared with those of previous reports of other PI-based multitherapy in children.13 Sustained undetectable viral replication was observed in all children who had not been pretreated and in nearly all those who had previously received treatment without a PI. The virologic result was stable over time, with a similar proportion of children with an undetectable viral load between 12 and 18 months. In contrast, viral replication was not controlled in half of the PI-experienced children, some of whom had already been heavily pretreated. As expected, baseline resistance genotype appeared to be predictive of virologic response. Individually, protease mutations at codons 10I/F, 46I, 54V/L, 71V/I, 82 A/F/S/T, and 90M were particularly described to be associated with a risk of failure. These data slightly differed from those presented in a previous study conducted in heavily antiretroviral-experienced adults, where the baseline mutation at codon 90 had a smaller effect on the rate of virologic response, whereas a mutation at position 36 seemed to influence the virologic response to a greater extent.14 The threshold number of LPV mutations associated in this group of children with significantly poorer antiviral activity of LPV was approximately 4. Recent algorithms for the interpretation of resistance genotypes based on data for adult patients suggest that a lower virologic response was observed when 6 or more protease mutations were present before treatment initiation.4–7,15,16 The apparent difference is, however, probably a result of the definition used for virologic success or failure. These results, as previously described in adults, showed that there was an intermediate zone between full antiviral activity (0–3 mutations) and complete loss of activity (≥6 mutations). Thus, after 3 mutations, each consecutive mutation is associated with decreasing efficacy. A recent study that modeled viral genotype-phenotype correlations for more than 1000 viral strains also reported a threshold of 3 mutations for a loss of phenotypic efficacy.17 In situations of intermediate resistance, particular attention must be given to optimizing treatment. Our study strongly suggested that control of the plasma LPV concentration was particularly important in PI-experienced children. When expressed in the form of the GIQmax, it was independently associated with the virologic outcome. This confirmed that the inhibitory concentration of LPV for 90% of viruses (IC 90%) carrying mutations conferring resistance was higher than that of wild-type viruses, meaning that the plasma concentration must be higher for effectiveness. A maximal or trough LPV concentration can be used to guide therapy, because similar results were recently obtained for trough concentration in adults, probably as a result of the pharmacokinetic moiety of LPV.18,19 These results are consistent with those obtained with boosted amprenavir in PI-experienced patients, for which the GIQmax was a better predictor of virologic response than the virologic and pharmacologic variables measured alone.12
Finally, in children failing an LPV-containing regimen, accumulation of additional PI-associated resistance mutations was evidenced only in patients with prior PI experience. The cutoff value and duration of persistent viral replication requiring modification or interruption of an ongoing regimen have not yet been defined in children. Accordingly, these findings suggest that a cutoff lower than 10,000 copies/mL might be more appropriate, as suggested by most treatment recommendations for adults.
In conclusion, this study confirmed the significant and sustained virologic response achieved with the use of an LPV/r-containing regimen in naive- and PI-experienced children. Baseline LPV resistance score and GIQmax can predict the virologic response, particularly in previously PI-treated children. LPV plasma levels should be optimized to achieve sufficient plasma concentrations to overcome the resistance level.
The authors thank S. Brun and I. Cohen-Codar for their valuable comments.
1. Walmsley S, Bernstein B, King M, et al. Lopinavir
-ritonavir versus nelfinavir for the initial treatment of HIV
infection. N Engl J Med
2. Cvetkovic RS, Goa KL. Lopinavir
/ritonavir: a review of its use in the management of HIV
3. Badaro R, DeJesus E, Lazzarin A, et al. Efficacy and safety of atazanavir (ATV) with ritonavir (RTV) or saquinavir (SQV) versus lopinavir
/ritonavir (LPV/r) in combination with tenofovir (TFV) and one NRTI in patients who have experienced virologic failure to multiple HAART regimens: 16-week results from BMS AI424-045 [abstract 118]. Presented at the 2nd IAS Conference on HIV
Pathogenesis and Treatment, Paris, July 2003.
4. Kempf DJ, Isaacson JD, King MS, et al. Identification of genotypic changes in human immunodeficiency virus protease that correlate with reduced susceptibility to the protease inhibitor lopinavir
among viral isolates from protease inhibitor-experienced patients. J Virol
5. Masquelier B, Breilh D, Neau D, et al. Human immunodeficiency virus type 1 genotypic and pharmacokinetic determinants of the virological response to lopinavir
-ritonavir-containing therapy in protease inhibitor-experienced patients. Antimicrob Agents Chemother
6. Kempf DJ, Isaacson JD, King MS, et al. Analysis of the virological response with respect to baseline viral phenotype and genotype in protease inhibitor-experienced HIV
-1-infected patients receiving lopinavir
/ritonavir therapy. Antivir Ther
7. Bongiovanni M, Bini T, Adorni F, et al. Virological success of lopinavir
/ritonavir salvage regimen is affected by an increasing number of lopinavir
/ritonavir-related mutations. Antivir Ther
8. Saez-Llorens X, Violari A, Deetz CO, et al. Forty-eight-week evaluation of lopinavir
/ritonavir, a new protease inhibitor, in human immunodeficiency virus-infected children
. Pediatr Infect Dis J
9. Mouroux M, Yvon-Groussin A, Peytavin G, et al. Early virological failure in naive human immunodeficiency virus patients receiving saquinavir (soft gel capsule)-stavudine-zalcitabine (MIKADO trial) is not associated with mutations conferring viral resistance
. J Clin Microbiol
10. Keil K, Frerichs VA, DiFrancesco R, et al. Reverse phase high-performance liquid chromatography method for the analysis of amprenavir, efavirenz, indinavir, lopinavir
, nelfinavir and its active metabolite (M8), ritonavir, and saquinavir in heparinized human plasma. Ther Drug Monit
11. DeGruttola V, Dix L, D’Aquila R, et al. The relation between baseline HIV
and response to antiretroviral therapy: re-analysis of retrospective and prospective studies using a standardized data analysis plan. Antivir Ther
12. Marcelin AG, Lamotte C, Delaugerre C, et al. Genotypic inhibitory quotient
as predictor of virological response to ritonavir-amprenavir in human immunodeficiency virus type 1 protease inhibitor-experienced patients. Antimicrob Agents Chemother
13. Van Rossum AM, Fraaij PL, De Groot R. Efficacy of highly active antiretroviral therapy in HIV
-1 infected children
. Lancet Infect Dis
14. Isaacson J, Kempf D, Calvez V, et al. Quantitative estimate of the effect of individual baseline mutations in HIV
protease on the virologic response to lopinavir
/ritonavir therapy in heavily antiretroviral-experienced patients. Presented at the 9th Conference on Retroviruses and Opportunistic Infections, Seattle, February 24–28, 2002.
15. Monno L, Saracino A, Scudeller L, et al. HIV
-1 phenotypic susceptibility to lopinavir
(LPV) genotypic analysis in LPV/r-naive subjects with protease inhibitor experience. J Acquir Immune Defic Syndr
16. Barreiro P, Camino N, De Mendoza C, et al. Comparison of the efficacy, safety and predictive value of HIV
genotyping using distinct ritonavir-boosted protease inhibitors. Int J Antimicrob Agents
17. Wang D, Larder B. Enhanced prediction of lopinavir resistance
from genotype by use of artificial neural networks. J Infect Dis
18. Marcelin AG, Cohen-Codar I, King M, et al. Virological and pharmacological parameters predicting the response to lopinavir
/ritonavir [abstract 827]. Presented at the 2nd IAS Conference on HIV
Pathogenesis and Treatment, Paris, July 2003.
19. Gonzalez de Requena D, Gracia-Benayas T, Gallego O, et al. Predictive power of genotypic inhibitory quotient
plasma levels in both efficacy and lipid elevations of LPV/r-based salvage regimens at 48 weeks [abstract 620]. Presented at the 11th Conference on Retroviruses and Opportunistic Infections, San Francisco, February 8–11, 2004.