| |
Long-term effects of intermittent interleukin-2 therapy in chronic HIV-infected patients (ANRS 048-079 Trials)
|
| |
| |
AIDS:Volume 21(14)September 2007p 1887-1897
Durier, Christinea; Capitant, Catherinea; Lascaux, Anne-Sophieb; Goujard, Cecilec; Oksenhendler, Ericd; Poizot-Martin, Isabellee; Viard, Jean-Paulf; Weiss, Laurenceg; Netzer, Emmanuellea; Delfraissy, Jean-Francoisc,h; Aboulker, Jean-Pierrea; Levy, Yvesb
From the aINSERM SC10, Villejuif
bHopital Henri Mondor, Universite Paris 12, INSERM U 841, Creteil, Paris
cHopital Bicetre, Kremlin-Bicetre, Paris
dHopital Saint-Louis, Paris
eHopital Sainte-Marguerite, Marseille, Paris
fHopital Necker, Paris, France
gHopital Europeen Georges-Pompidou, Paris, France
hANRS, Paris, France.
Abstract
Objective: Interleukin (IL)-2 therapy leads to significant CD4 cell increases in HIV-infected patients. Since phase III trials are ongoing, studies supporting the long-term feasibility of this strategy are needed.
Methods: We studied the long-term outcomes of 131 patients treated with IL-2 in two studies initiated either before (ANRS 048) or following (ANRS 079) the advent of HAART.
Results: At the last assessment (median follow-up 3.4 years), these patients experienced a gain of 428 cells/μl and a decrease in plasma HIV RNA to 1.70 log10 copies/ml. In both studies, high CD4 cell counts were maintained with a median of ten 5-day cycles of subcutaneous IL-2. Median time since the last cycle was 2 years. At last assessment, 59% of 048 patients maintained a non-HAART regimen. Detailed analysis at week 170 showed that median CD4 cell counts were 856 (048) and 964 (079) cells/μl. This corresponded to a gain from baseline of 515 (048) and 627 (079) cells/μl. The median viral load decreases from baseline and corresponded to 1.70 (048) and 1.88 (079) log10 copies/ml. Comparisons across the studies showed that CD4 gains and viral load changes were similar whether HAART or non-HAART was used. The frequency of cycling, but not CD4 cell counts, viral loads or antiviral regimen at baseline, was predictive of long-term CD4 gain (P = 0.03).
Conclusion: Altogether, these observations support IL-2 as a long-term therapeutic strategy in HIV infection.
Introduction
The natural course of HIV infection is characterized by a progressive loss of CD4 T cells, which leads to increased susceptibility to opportunistic infections and malignancies and death. The advent of combined antiretroviral regimens has changed the prognosis of HIV infection by reducing AIDS-related morbidity and mortality [1]. This clinical benefit is related to the limitation of the immunological damage that is caused by HIV replication, and to the restoration of CD4 T-lymphocyte counts and specific responses against pathogens.
Intermittent interleukin (IL)-2 therapy has been proposed and evaluated as a strategy to increase the CD4 T-cell pool in HIV-infected patients. The initial rationale for the use of IL-2 in patients with AIDS came from the observation that it restored in-vitro cytomegalovirus-specific cytotoxic activity of T cells from immunodeficient patients [2]. Several phase II studies [3-24] have demonstrated that, when administered in a cyclic fashion, IL-2 leads in a large majority of patients to a significant and sustained increase in CD4 T-cell count beyond that seen in patients treated with antiviral drugs alone. The potential clinical benefit of this cytokine is currently under investigation in two large phase III international trials covering a wide range of CD4, SILCAAT and ESPRIT trials [25]. Besides their key clinical end points, these two trials are also addressing the question of the management of IL-2 therapy in the context of a long-term strategy.
In France, studies of IL-2 therapy in HIV infection began more than 11 years ago, on the initiative of the ANRS (French National Agency for AIDS Research). The first study was launched in 1995 and compared the biological effect of IL-2 administered either by continuous intravenous infusion (CIV) or subcutaneously (SC) to antiviral drug-naive patients, who started a combination of nucleoside reverse transcriptase inhibitors (NRTI). The second study started in 1997 and compared the biological effects of IL-2 combined with a more potent combination of antiretrovirals (HAART) to HAART alone. We report here the extensive long-term follow-up of these two studies.
Discussion
We report here the ANRS experience of the long-term management of IL-2 therapy in HIV infection through the long-term follow-up of two multicenter and randomized IL-2 studies initiated in the 1990s. In all, 210 patients were enrolled in the initial randomized phases of these studies. Ninety percent of these patients participated in the extended phase of these studies, during which patients from the control groups were allowed to start SC IL-2 therapy. Finally, 131 patients who participated in the long-term follow-up and who received IL-2 in the initial or extended phases of these studies were followed. Patients from the 048 study (median follow-up 5.1 years) experienced a 90% increase in CD4 T-cell counts from baseline. Similarly, patients from the 079 study maintained an increase of up to 180% in their CD4 T-cell counts (median follow-up 3.3 years). Interestingly, the magnitude of these CD4 T-cell increases was not significantly different to that seen at the end of the randomized phases of these studies [11,23]. These results demonstrated that the initial increase in CD4 T cells could be maintained over the long-term.
Given the toxicity of IL-2, we were concerned to provide clear guidelines to physicians for the continuation of IL-2 cycling beyond the end of the randomized phases of our studies. Empirically, it was proposed to patients to receive an additional IL-2 cycle when their CD4 T-cell counts dropped to 25% below the plateau. This uniform schedule provided a good opportunity to evaluate IL-2 requirements for maintaining CD4 T-cell counts in the long-term. A maximum of four cycles was needed during the follow-up of the 048 study, whereas no cycles were given to patients from the 079 study, to maintain high CD4 cell counts (Table 4). Our results extend the long-term experience of the National Institute of Health (NIH) cohort of IL-2-treated patients [26].
Taken together, these data demonstrated that the impact of IL-2 on the quantitative restoration of CD4 T-cell counts is a durable effect. In a clinical care setting, our results suggest that administration of a new cycle of IL-2 once CD4 T-cell counts drop 25% below the plateau, is a manageable clinical strategy to maintain high CD4 T-cell counts with infrequent cycles. These results, added to the recent observation that patients' health perception of the impact of IL-2 on quality of life improves in the long-term [27], may contribute to a better acceptance of IL-2 therapy as a long-term strategy.
Mechanistically, recent data have shown that IL-2 administration to HIV-infected patients leads to an increase in the rates of lymphocyte proliferation and apoptosis in the short-term [28-32]. In the long-term, it appears that the major mechanism by which IL-2 induces an expansion of CD4 T cells is a prolongation of their survival. This was clearly demonstrated in an elegant study using in-vivo labeling techniques [33]. Mathematical models showed that IL-2 increases the half-life of CD4 T cells. This effect may explain why a significant number of cycles is usually necessary to achieve a stable plateau of CD4 T cells and why a low frequency of IL-2 cycles over time is required to sustain CD4 T-cell counts.
Previous studies have emphasized that the magnitude of the increase of CD4 T cells in IL-2-treated patients directly relates to the dose of IL-2, the nadir of CD4 T-cell counts, and HIV viral loads [5,6,8,10,34-36]. Most of these studies evaluated the impact of these factors on short-term CD4 responses. In the present study, we investigated factors predictive of a sustained CD4 T-cell count from baseline to week 170. We were not able to look at the impact of the nadir of CD4 T cells since we did not collect this information. We found only a trend to a greater CD4 T-cell gain in patients with higher baseline CD4 counts. High values of CD4 T-cell counts at enrollment may, however, have precluded our ability to discern an impact of baseline CD4 T-cell counts on long-term CD4 responses. The strongest predictors of CD4 response were the total number of cycles given during the study period, and a time interval since the last cycle shorter than 2 years. This information would help to define the optimal use of IL-2 therapy as a long-term strategy. It would also contribute to the acceptance of a regular cycling schedule in long-term, ongoing, multicenter, large phase III trials (SILCAAT and ESPRIT, [25]) exploring whether these IL-2 biological effects translate into a clinical benefit in the long-term.
Since the 048 and 079 studies were conducted in patients before and after the availability of HAART, we took advantage of this cohort to compare the outcomes of patients treated with IL-2 with or without HAART, a question that was never investigated in IL-2 trials. First of all, the antiviral regimen initiated at entry of the 048 and 079 studies (NRTIs or HAART) did not influence the CD4 T-cell response to IL-2 in the short or long term. Patients from the 048 study, however, needed more IL-2 cycles to maintain high CD4 T-cell counts than patients treated initially with IL-2 and HAART. It may be that, as recently reported [37], the level of slow ongoing viral replication or the persistence of a degree of immune activation blunts the long-term effect of IL-2.
Patients treated with IL-2 either with NRTIs (048) or HAART (079) experienced a similar decrease of plasma viral load from baseline. At the last assessment, however, the percentage of patients with plasma HIV RNA less than 500 copies/ml was 22% greater in the 079 study in comparison with the 048 (80 and 58%). This difference reflects the higher baseline viral loads in patients from the 048 as compared to the 079 study. It is interesting to note that physicians have kept a majority of patients from the 048 study under the same combination of NRTIs that were initiated at entry of the study. Up to 67% of IL-2-treated patients were still on NRTIs at the last assessment of this cohort. Although the level of virus replication was low in these patients (median plasma viral load, 2.3 log10 copies/ml), this incompletely suppressed viremia could lead to the development of drug resistance mutations over time. As expected, the virological outcome of HAART-alone patients was significantly better than that of patients treated with IL-2 and NRTIs. Finally, there was a trend towards a better outcome for HAART-alone patients as compared with patients receiving IL-2 with HAART.
Altogether, these data showed that a majority of patients treated with IL-2 maintained a substantial increase of CD4 T cells despite, for some of them, a low level of virus replication, mostly in patients treated with a non-optimal regimen of antiretroviral drugs. Our study was not designed to demonstrate whether patients benefit clinically from the increase of CD4 T cells induced by IL-2 therapy in the setting of incompletely suppressed viremia. This slow ongoing virus replication may hamper the potential clinical benefit of IL-2. This question will be addressed in ongoing long-term large phase III clinical studies, which investigate whether the quantitative impact of IL-2 may help to prevent disease progression in patients with CD4 cell counts below (SILCAAT study) or above (ESPRIT study) 300 cells/μl.
Globally, our results indicate that intermittent IL-2 therapy leads to a long-term increase in CD4 T-cell counts. It is sustainable and does not require a high frequency of IL-2 cycles. Our data help inform physicians and patients on the long-term feasibility and management of intermittent IL-2 therapy in HIV infection.
Methods
Patients and design
ANRS 048 and ANRS 79 were open-label, randomized, multicenter, controlled clinical trials comparing the efficacy and safety of IL-2 in combination with antiretroviral therapy with that of antiretroviral therapy alone. Detailed descriptions of the trials designs and results have been published [11,23].
In ANRS 048, 94 symptom-free HIV-1 adult patients, with CD4 counts of 250-550 cells/μl, naive to antiretroviral drugs, were initially randomized to receive antiretroviral therapy [zidovudine (ZDV) + didanosine] alone (control, n = 26) or in combination with IL-2 SC (3 MIU/m2 twice daily for 5 days; n = 24), PEG (2 MIU/m2 intravenous bolus; n = 22) or CIV (12 MIU/day; n = 22) cycles given every 8 weeks. At week 56, patients were allowed to start, continue, or switch to IL-2 SC (3 MIU/m2 twice daily for 5 days) in a 4-year, long-term follow-up until week 258.
In ANRS 079, 116 symptom-free HIV-1 adult patients, with CD4 counts of 200-550 cells/μl, naive to antiretrovirals or naive to protease inhibitors (PI) were randomized to start HAART [stavudine + lamivudine + indinavir) either alone (HAART, n = 58) or combined with SC IL-2 (5 MIU twice daily for 5 days) every 4 weeks for three cycles, and then every 8 weeks for seven cycles (HAART + IL-2, n = 58). After week 74, patients could continue IL-2, and control patients who had not achieved a 50% increase in their CD4 T-cell count were offered the possibility to receive SC IL-2 (4.5 MIU twice daily for 5 days) until week 170.
Patients who initiated IL-2 therapy at randomization were referred to as the 'immediate IL-2 group' and those who did so during the long-term follow-up as the 'deferred IL-2 group'. All patients gave their informed consent. The long-term studies were approved by the ethics review board (Creteil, France).
Statistical methods
Baseline characteristics of the two clinical trials were tabulated. Changes in CD4 cell counts, plasma HIV RNA values, antiretroviral therapy from baseline to last follow-up, and number of IL-2 cycles administered to patients were described.
Data from the two clinical trials were then pooled for analysis. Combining the data was appropriate because the patients' eligibility criteria and the guidelines for IL-2 administration in the long-term were similar in the two studies. Data from patients enrolled in the 048 study were censored at week 170, to match with the ultimate time point of the ANRS 079 long-term follow-up.
The analysis of predictors of CD4 cell count changes (in cells/μl) was performed using linear model analyses and restricted to the patients who had been primarily treated with IL-2. Independent variables were gender, body mass index, age, CD4 cell counts and HIV RNA levels at baseline, antiviral regimen, number of IL-2 cycles during the initial and long-term follow-up phases, total dose of IL-2, interval between IL-2 cycles.
The 'immediate IL-2 groups' from both studies were compared at baseline and week 170 of the follow-up. Comparisons were made between patients treated with IL-2 either with NRTIs or HAART and HAART alone (Wilcoxon and Fisher tests). Two-sided P values less than 0.05 were considered statistically significant. Analyses were completed using SAS 8.2 software (SAS Institute Inc., Cary, North Carolina, USA).
Results
Patient follow-up
A total of 210 patients was originally analyzed in the ANRS 048 (n = 94) and 079 (n = 116) studies [11,23]. Their baseline characteristics are listed in Table 1. During the controlled phase of these studies, two patients died (accident and suicide) and 12 patients (six in each study) were lost to follow-up (Fig. 1).
IL-2 therapy was administered to 131 patients during the initial phase or the follow-up, of the 048 (n = 69) and 079 (n = 62) studies (Table 2).
The median follow-up from baseline to last assessment was 5.1 years (range, 2.0-5.4) and 3.3 years (range, 1.6-3.5) for ANRS 048 and 079, respectively. Up to 90% (n = 166) of patients were still on active follow-up at the end of the extension phases.
Immunological, virological and treatment outcomes of interleukin-2-treated patients
The median follow-up of 131 IL-2-treated patients was 3.4 years (Table 2). At the last assessment, median CD4 T-cell counts and plasma HIV RNA were 791 cells/μl and 2.3 log10 copies/ml for a gain of +428 CD4 T cells and a decrease in plasma viral load of 1.70 log10 copies/ml. Detailed analysis of the cohort of 69 patients who received IL-2 therapy in the 048 study showed that the median CD4 cell count increased from 376 cells/μl (range, 251 to 345) at baseline to 726 cells/μl (176 to 1609) at week 258. This corresponded to an increase from baseline of +297 cells/μl (range, 109 to 1068) and +404 cells/μl (range, -120 to +697) for patients from the 'immediate' and 'deferred' IL-2 groups, respectively (Fig. 2). Plasma HIV RNA values decreased from 4.55 log10 copies/ml (range, < 1.7 to 5.46) at baseline to 2.65 log10 copies/ml (range, < 1.3 to 5.09) at week 264. Fifty two and 26% of those patients had plasma HIV RNA values below 500 and 50 copies/ml at week 258, respectively. At the last assessment, 28 patients (41%) switched from an NRTI combination to a triple combination including a PI or a non-nucleoside reverse transcriptase inhibitor (NNRTI).
For the cohort of 62 patients who received IL-2 therapy in the ANRS 079 study, the median CD4 cell count increased from 349 cells/μl (range, 202 to 549) at baseline to 981 cells/μl (23 to 2614) at week 170. The increase from baseline was +627 cells/μl (range, -46 to +2137) and +788 cells/μl (range, -123 to +2120) for patients from the 'immediate' and 'deferred' IL-2 groups, respectively (Fig. 2). Plasma HIV RNA values decreased from 3.98 log10 copies/ml (range, < 1.7 to 5.69) at baseline to 1.7 log10 copies/ml (range, < 1.3 to 4.91) at week 170. About 80% of these patients reached plasma viral load values < 50 copies/ml 12 weeks following HAART initiation, and this percentage remained constant until week 170. At week 170, 33 of 57 (58%) patients changed their initial HAART regimen: seven stopped therapy, six were treated with NRTIs only, 10 were treated with NRTIs and NNRTIs, and 10 switched from indinavir to another protease inhibitor.
During the extension phase periods, four patients developed a class B or C AIDS-defining event: two patients from the 048 (one cervical cancer and one zoster infection) and two from the 079 study (one zoster and one tuberculosis).
Interleukin-2 exposure of patients during study follow-up
A total of 690 cycles of IL-2 (220 and 470 during the controlled phase and the 4-year extended phase, respectively) were given in patients from the 048 study, in which a median of 10 IL-2 cycles per patient was given (range, 1 to 20) corresponding to 11 cycles (range, 1 to 19) for patients from the 'immediate IL-2 group', and nine cycles (range, 1 to 20) for those of the 'deferred IL-2 group'. The median time interval since the last IL-2 cycle was 2 years (range, 0 to 5) and was similar for 'immediate IL-2' and 'deferred IL-2' recipients (2.1 and 1.8 years, respectively). The number of IL-2 recipients per 6-month period decreased gradually from 51 at the end of the initial phase of the 048 study to 13 patients in the last 2 years of follow-up (Fig. 3a). The ratio of IL-2 cycles per patient and per 6-month interval decreased from 2.7: 1 at study entry to < 0.5: 1 at month 66 of follow-up (Fig. 3b).
In the ANRS 079 study, 510 IL-2 cycles (418 and 92 during the initial period and the 2-year extended phase, respectively) were given. Patients from the 'immediate IL-2 group' received a median of 10 cycles (range, 1 to 14) while those from 'deferred IL-2 group' received a median of 4 cycles (range, 2 to 10) (Fig. 3a). The median interval since the last IL-2 cycle was 1.9 and 0.3 years for the 'immediate' and 'deferred IL-2' groups. The ratio of IL-2 cycles per patient and per 6-month interval decreased from 3.7: 1 at study entry to < 0.1: 1 at month 48 of the follow-up (Fig. 3b)
Predictors of CD4 cell count changes in patients initially treated with interleukin-2 combined either with nucleoside reverse transciptase inhibitors or HAART
Body mass index, sex, and age did not predict the CD4 cell count changes from baseline to 170 weeks. CD4 cell count or viral load values prior to IL-2 were not significantly correlated with CD4 increases. A trend to a higher magnitude of CD4 T-cell gain was, however, observed in patients with higher baseline CD4 cell counts (+673 versus +494 cells/μl, P = 0.094) (Table 3). Interestingly, the antiviral regimen initiated at entry of the studies (NRTIs or HAART) did not influence the CD4 T-cell response to IL-2 therapy. The number of cycles during the controlled or the follow-up phases, and a time interval since the last cycle shorter than two years were, however, associated with the CD4 T-cell gain.
Table 3. Predictors of CD4 cell count changes in patients from the 'immediate IL-2' groups, and followed up for at least 170 weeks.
Comparison of long-term outcomes of patients treated initially with interleukin-2 combined with either nucleoside reverse transciptase inhibitors or HAART
Baseline characteristics (body mass index, gender, age, CD4 T-cell counts) were similar except that 53% of 079 patients were pre-exposed to antiviral drugs as compared with none from the 048 study. Accordingly, the median baseline plasma HIV RNA level was lower in the ANRS 079 group (3.9 log10 copies/ml) than in the ANRS 048 group (4.5 log10 copies/ml). During follow-up, 12 of 36 (33%) patients from the 048 switched to a PI or NNRTI regimen (Table 4).
At 170 weeks, the median CD4 cell counts were similar in patients from the 048 and 079 studies and were respectively 856 and 964 cells/μl. This corresponded to a gain from baseline of +515 and +627 cells/μl (P = 0.35) in the 048 and 079 studies, respectively (Fig. 2a). The median viral load decreases from baseline were similar in the two groups and corresponded to 1.70 and 1.88 log10 copies/ml (P = 0.75) in the 048 and 079, respectively (Fig. 2b). A higher percentage of patients (64 versus 39%), however, reached a plasma HIV RNA below 50 copies/ml in the 079 as compared to 048 study (P = 0.027) (Fig. 2c).
A median of 11 (range, 2 to 17) and 10 (range, 1 to 14) cycles of IL-2 were given throughout the follow-up in the 048 and 079 studies, respectively. Seventy-five and 24% of the 048 and 079 patients (P < 0.0001) needed to receive additional IL-2 cycles (median 5 in the 048 and 2 in the 079 study; P < 0.0001) to maintain their CD4 cell responses. At the last assessment, the median interval since the IL-2 cycle was 16 months (range, 1.5 to 37) and 24 months (range, 1 to 38) in ANRS 048 and ANRS 079, respectively (P = 0.033).
Comparison of the long-term outcomes of patients treated with interleukin-2 either with HAART or nucleoside reverse transciptase inhibitors to patients treated with HAART
In the ANRS 079 study (Fig. 2, Table 4), immediate IL-2 recipients experienced a median increase from baseline of +627 CD4 cells in comparison with +372 CD4 cells in patients treated with HAART alone (P = 0.002). Eighty percent (36/45) and 92% (34/37) of patients treated or not treated with IL-2 had plasma HIV RNA < 500 copies/ml, respectively (P = 0.13). These percentages were 64% (29/45) and 78% (29/37) for plasma HIV RNA < 50 copies/ml (P = 0.17). A trend to a higher decrease in plasma viral load was noted in patients treated with HAART alone in comparison with IL-2 recipients (median log10 viral load, -2.63 versus -1.88, P = 0.052).
Patients immediately treated with IL-2 in the 048 study experienced a faster and greater increase in CD4 T-cell counts, in comparison with patients who initiated HAART alone in the 079 study (Fig. 2a). The median increase in CD4 T cells was +372 cells/μl in the HAART-alone group as compared with +515 cells/μl in patients immediately treated with IL-2 in ANRS 048 (P = 0.042, Table 4). At the last assessment, 67% of patients from the 048 study were still on the NRTI regimen. Ninety-two percent (34/37) of patients treated with HAART alone, as compared to 58% (21/36) of patients treated with IL-2 and NRTIs, had a plasma HIV RNA values < 500 copies/ml (P < 0.001) (Fig. 2b and c).
|
|
| |
| |
|
|
|
