Response to a protease-inhibitor ( ritonavir )-containing combination antiretroviral regimen in HIV-infected children

The Division of Infectious Diseases, Department of Pediatrics, 1Hospital for Sick Children, Toronto, Ontario; 2Sainte Justine’s Hospital, University of Montreal, Montreal, Quebec; 3British Columbia Children’s Hospital, Vancouver, British Columbia Correspondence and reprints: Dr Upton Allen, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8. Telephone 416-813-8129, fax 416-813-8404, e-mail upton.allen@sickkids.ca Received for publication December 24, 2001. Accepted June 19, 2002 UD Allen, N Lapointe, SE Read, JC Forbes, SM King, S Wasfy, and the Canadian Pediatric AIDS Research Group. Response to a protease-inhibitor (ritonavir)-containing combination antiretroviral regimen in HIV-infected children. Can J Infect Dis 2003;14(2):89-93.

A ntiretroviral regimens for use among individuals who are failing antiretroviral therapy have largely been evaluated in adults.Protease-inhibitor-containing regimens have had profound beneficial effects in suppressing viral replication (1)(2)(3)(4).
Modification of therapy usually entails changing at least two agents, sometimes including a protease inhibitor (5)(6).While such regimens have also been shown to be beneficial in children (7)(8)(9)(10)(11), several questions remain unanswered from a pedi-©2003 Pulsus Group Inc.All rights reserved ORIGINAL ARTICLE atric perspective.In this context, it is important to document the effectiveness of various regimens in lowering viral load in infected children as well as to explore the reasons why, in some children, the success rates are suboptimal with proteaseinhibitor-containing antiretroviral therapy (11).
Ritonavir is one of the most widely used protease inhibitors in Canada.The present study evaluated ritonavir-containing combination regimens that were employed in a setting where patients were failing existing therapy by virtue of having persistently high viral loads.The objectives of the study were: first, to determine if a treatment strategy employing triple or quadruple combination antiretroviral therapy containing ritonavir would result in significant reductions in HIV-1 ribonucleic acid (RNA) levels in a population of previously treated HIV-infected children; second, to compare the characteristics of patients whose viral loads responded to modification of therapy with those who failed to respond; and third, to examine the effects of the above regimen on changes in CD4.

METHODS
An open nonrandomized intervention trial that involved three Canadian tertiary centres affiliated with the Canadian Pediatric AIDS Research Group (CPARG) was conducted.The study was conducted between March 1998 and March 2000 and subjects were followed for a maximum duration of 48 weeks.
HIV-infected children were eligible if they were younger than 18 years of age and were experienced with antiretroviral therapy, but were ritonavir-naive.Informed consent was obtained.They were eligible for enrolment if their physicians assessed them as failing their existing antiretroviral therapy by virtue of having persistently elevated HIV-1 RNA levels.
All subjects were seen initially and were then followed at an interval of approximately monthly for six months, and then every two months for six months at which time they had clinical, virological and immunological assessments.HIV-1 RNA levels and CD4 counts were performed.HIV-1 RNA levels were performed in regional references laboratories using the Chiron (Chiron, USA) B DNA assay (lower limit of detection less than 50 copies/mL).The planned antiretroviral study regimen consisted of two new agents to which the virus was presumed to be sensitive, one of which was ritonavir.

Statistical analyses
Using the binomial test of proportions, the sample size was based on an estimated success rate of 80% with lower and upper bounds of ±20% based on a 95% confidence interval.This resulted in a target sample size of 15 subjects.Wherever appropriate, means were compared with Student's t test and medians with a nonparametric procedure (Kruskal-Wallis).Proportions were compared using χ 2 or Fisher's exact test as appropriate.

RESULTS
The baseline characteristics of the patients enrolled are shown in Table 1.Among 15 children, the majority of whom had perinatally-acquired HIV, the median age was 7.9 years (range 1.6 to 14.8).The median CD4 count was 557 cells/mm 3 (range 57 to 1702), while the median viral load at enrolment was 72,600 copies/mL (range 3626 to 796,440).Patients were on antiretroviral therapy for a median of 2.3 years (range 1.2 to 5.8).The majority (73.3%) had experienced an opportunistic condition before enrolment.The classes of the antiretroviral agents that the patients were receiving at enrolment are shown in Table 1.In this regard, the majority of patients were either receiving two nucleoside reverse transcriptase inhibitors (NRTIs) plus a protease inhibitor other than ritonavir (46.7%) or two NRTIs (40%).Zidovudine and lamivudine were the most frequently used NRTIs.Overall, 60% had received a protease inhibitor other than ritonavir before enrolment.
Eleven patients followed the study regimen.Four additional patients were enrolled, but the only change in therapy was the introduction of ritonavir to replace another protease  inhibitor.The results were analyzed on all 15 patients on an intent-to-treat basis.An initial improvement in CD4 counts was observed in 73.3% of patients.As shown in Table 2, 53.3% of subjects have an increase of greater than 20% in the CD4 counts over the first 12 weeks, while 6.7% experienced a corresponding decrease in CD4 counts.For all patients, the median percentage change in CD4 was 23.7% (range -26.3% to 346%).For those patients who had an initial response over the first 12 weeks, the median percentage change was 30.0%(range 6% to 346%).The corresponding changes among patients who did not show an initial CD4 response were median -13.8% (range -26.3% to -7.7%).Two-thirds of the patients (10 of 15) experienced initial reductions in viral load of at least a 0.5 log 10 .In nine patients, this reduction occurred within the first 12 weeks, while one patient responded at 16 weeks.The proportions of patients who experienced specific degrees of reduction in viral load over the first 12 weeks are shown in Table 2.
We examined the extent to which changes in CD4 counts and viral loads were sustained.These are shown in Figures 1  and 2, respectively.Among subjects who showed initial improvements in CD4 counts, eight of 11 (72.7%) had sustained improvements through to 32 to 48 weeks.At weeks 32 to 48, the overall improvement in CD4 counts was 25% above the median baseline of 557 cells/mm 3 .
For the group as a whole, there was a reduction in median viral load of 1.2 to 1.6 logs between baseline and weeks 32 to 48.Eighty per cent of subjects who had initial reductions in viral loads had sustained improvements through to 32 to 48 weeks.
We examined the baseline characteristics of patients who experienced a significant initial reduction in viral load (0.5 log or greater) with those who did not have a significant decrease.These characteristics are shown in Table 3.There was no significant difference in median ages between responders and nonresponders (median 7.2 years; range 1.6 to 14.8 versus 8.8 years; range 5.2 to 12.8, respectively, P=0.33).While the median baseline viral load was higher in the nonresponders than in the responders, the difference was not significant (94,774 versus 69,405 copies/mL, respectively, P=0.71).However, responders had higher median baseline CD4 counts compared with nonresponders 614 cells/mm 3 (range 180 to 1702) versus 120 (range 57 to 558, P=0.014).The duration of previous antiretroviral therapy was not significantly different between responders and nonresponders (medians 2.2 versus 3.8 years, P=0.27).
The differences between responders and nonresponders were re-examined by excluding the one patient who received ritonavir for less than three days, which was not long enough to assess effectiveness.The results were similar to those in Table 3 (P=0.007for differences in CD4).
In 11 of 15 patients, the change in treatment regimen at enrolment resulted in the addition of at least two new agents, one of which was the protease inhibitor, ritonavir.Of these 11 patients, seven had a significant reduction in HIV-1 RNA levels of at least 0.5 log over the first 12 weeks.Four patients had a switch in the choice of protease inhibitor therapy as the only change.These four patients represented deviations from the planned protocol.Of these four patients, one had a reduction in viral load of at least 0.5 log over the first 12 weeks, but with no increase in CD4 counts.
Twenty per cent of the patients experienced no adverse events that were temporally associated with the use of the ritonavir-containing combination regimen.Twelve patients (80%) had the following events documented while they were on the study regimen: grade 1 toxicity with peripheral neuropathy (two of 12); anemia at a grade 1 or 2 level (three of 12); deranged liver function tests at a grade 3 or 4 level (five of 12); elevated serum lipase at a grade 3 level (one of 12) and neutropenia at a grade 2 level (one of 12).However, five of these 12 patients had evidence indicating that the onset of these abnormalities occurred before the initiation of the riton-  *Nine patients had a significant response within the first 12 weeks, while one patient responded at 16 weeks avir-containing regimen.While the majority of subjects tolerated ritonavir, taste aversions occurred in three subjects.

DISCUSSION
Among HIV-infected children, it has been shown that changes in CD4 counts and HIV-1 RNA levels are independent predictors of disease progression and survival (12).In this study involving HIV-infected children who were failing existing antiretroviral therapy, we demonstrated that a significant proportion respond to a revision of the antiretroviral agents, including a protease inhibitor (ritonavir) as part of a combination regimen.We documented that changes in CD4 counts and viral loads occurred within the first 12 weeks after therapy was modified.The majority of patients who responded had greater than a 1.5 log decrease in viral load.
While the initial response to a modification in treatment is important, it is also worthwhile to examine whether the improvements seen are sustained.We were able to show that the majority of patients who experienced improvements in CD4 counts and viral loads had sustained responses for up to 32 to 48 weeks after modification of their treatments.
In a large randomized trial, Nachman et al (11) demonstrated the beneficial effects of a ritonavir-containing antiretroviral regimen.The authors noted that triple therapy containing ritonavir was just greater than 50% effective in maximally suppressing the viral load (11).Our findings are consistent with these results.In our study, close to two-thirds of patients had an initial response to a switch in treatment to a ritonavir-containing regimen.However, to provide some insight into the reasons for the lack of response in some patients, we examined several factors, including HIV-1 RNA levels and CD4 counts.Others have shown that, in pediatric cohorts, the virological responses do not necessarily correlate with baseline HIV-1 RNA levels (13).Likewise, we found no significant difference in baseline viral loads between patients who had reductions on HIV-1 RNA levels compared with those who did not have reductions in such levels.However, other investigators have observed that among patients receiving protease inhibitor-based salvage therapy, lower plasma HIV-1 RNA levels at the time of switching treatment were associated with improved outcomes (11,14,15).
Low baseline CD4 counts have been shown to be predictive of virological failure in response to protease inhibitor-containing salvage therapy (14).In our study, we demonstrated that patients with lower CD4 counts were less likely to respond to changes in treatment.This may help to define a group of children who should be given priority for antiretroviral resistance testing, particularly if they are failing existing therapy.
We assumed that the level or resistance to ritonavir would be low in a ritonavir-naive population.Thus, we felt that lack of responses would not be expected to be primarily due to drug resistance affecting ritonavir.However, it is likely that there was resistance to other components of the combination regimen.We have previously demonstrated that, among patients in the CPARG cohort, resistance to the nucleoside analog zidovudine was associated with lower CD4 counts and a longer duration of therapy (16).
While the majority of patients tolerated ritonavir, taste aversion was noted.Deranged liver function tests were among the most common features of the adverse events profiles documented.In the patients in whom the onset of ritonavir therapy was followed by adverse events, these events were consistent with the known side effects of this drug (4,17).However, it should be noted that in this study we did not conclusively show that these events were due to ritonavir as opposed to concomitant antiretroviral agents.The study is limited by sample size.This is a feature of many pediatric studies involving antiretroviral agents.In addition, it would have been ideal to have obtained serial testing of drug resistance on all subjects.In this context, we performed resistance testing on a subset of patients, but data were insufficient to allow for adequate quantitative analyses.The study was not designed to separate out the therapeutic effects of ritonavir from other antiretroviral agents, because the plan was to change more than one agent at the same time.In a minority of cases where there was a deviation from the study protocol that resulted in a change in the protease inhibitor as the only treatment modification, the therapeutic effect over the first 12 weeks was suboptimal (no significant viral load reduction and/or declining CD4 counts).This is consistent with what is expected given the known disadvantages of adding a single new agent to a failing regimen (6,18).

CONCLUSIONS
In summary, a ritonavir-containing combination antiretroviral regimen was beneficial for the majority of children who were failing existing therapy.Because benefit was not observed for some children, further studies need to be done to examine the characteristics of the children who fail to respond to salvage regimens and the role of novel combinations of antiretroviral agents, including protease inhibitors in these patients.Such efforts will require the cooperation of several centres, using similar protocols in a manner similar to many pediatric oncology studies.