Portal de Periódicos da CAPES

Kinetics of CD4 cells after discontinuation of antiretroviral therapy in patients with virological failure and a CD4 cell count greater than 500 cells/μl

2002; Lippincott Williams & Wilkins; Volume: 16; Issue: 11 Linguagem: Inglês

10.1097/00002030-200207260-00013

1473-5571

Cristina Mussini, Roberto Bugarini, Carlo Federico Perno, Andrea Antinori, Vanni Borghi, Ada Bertoli, Roberta D’Arrigo, Andrea Bedini, Nicola Mongiardo, Andrea Cossarizza, Roberto Esposito,

HIV/AIDS Research and Interventions

After a median of 37 months on antiretroviral therapy, 16 patients were asked to discontinue treatment instead of changing it. After a median observation time of 10.5 months, most patients experienced a rapid and progressive decrease in their CD4 cell count, even without a high viral load rebound. This decline was unrelated to the CD4 cell count and HIV-RNA values at interruption, but was more profound in patients in whom the M184V mutation had disappeared after lamivudine discontinuation. The most obvious option for patients experiencing drug failure after antiretroviral therapy remains a switch to therapeutic regimens including new antiretroviral drugs. However, cross-resistance and toxicity can make this option difficult, particularly for patients with a high CD4 cell count, in whom the potential decrease in viral load induced by the utilization of new drugs has to be balanced by the limited number of future therapeutic options caused by the possible development of resistance. Under these conditions, other hypotheses, including the discontinuation of therapy, may be considered. Therefore, to evaluate the effect of discontinuation in these patients, and to determine how long they could remain off therapy without damaging the immune system, we studied HIV-positive patients mostly treated with suboptimal therapeutic regimens, with virological failure and a CD4 cell count greater than 500 cells/μl. This study includes 16 patients treated with the following antiretroviral regimens: 12 patients: two reverse transcriptase nucleoside inhibitors (NRTI); two patients: two NRTI plus one protease inhibitor (PI); and two patients: two NRTI plus one non-nucleoside reverse trascriptase inhibitor. All patients signed a written informed consent to discontinue antiretroviral therapy instead of changing therapy. Patients were advised of a possible higher risk of HIV transmission to sexual partners during the period of discontinuation of treatment. Inclusion criteria included: a nadir CD4 cell count greater than 350 cells/μl before starting antiretroviral therapy, a history of an HIV plasma viral load below 500 copies/ml, and an increase in HIV-RNA plasma level (confirmed in two different controls) to 3 log10 copies/ml or greater in the presence of a CD4 cell count greater than 500 cells/μl. The primary end-point was therapy reinitiation. The main criteria for re-starting antiretroviral therapy were patients who reached a CD4 cell count below or equal to 350 cells/μl on two different occasions, and patients who withdrew from the study. Because of the low level of plasma viraemia at discontinuation, the HIV genotype was evaluable before and after therapy interruption in only six patients. Genomic HIV-1 RNA was extracted from plasma samples using the Qiamp Viral Kit (Qiagen, Germany). HIV RNA was then reverse transcribed, amplified and sequenced by using the ViroSeq Genotyping Kit Version 2 (Applied Biosystems, USA) [1]. Using a mixed model with a random intercept and an autoregressive variance–covariance structure for each patient, we studied changes in the CD4 cell count and HIV-RNA plasma level over time from the beginning of antiretroviral therapy. To model different trends in the CD4 cell count, CD8 cell count and HIV-RNA plasma level over time a spline-regression was used. We considered the duration of therapy, length of period of undetectable viral load, value of CD4 lymphocytes and HIV-RNA level at discontinuation as covariates. We used the same model to study whether CD4 cell changes were different during the undetectable HIV-RNA level period. To evaluate whether there were factors associated with a different rate of change of both the CD4 cell count and HIV-RNA level, we included the CD4 cell count at baseline, HIV-RNA level at baseline, duration of previous therapy and duration of HIV-RNA undetectability in the models as covariates. In particular, we estimated whether there was an interaction between the rate of change (slope) and these parameters (e.g. if low viral load at discontinuation was predictive of a higher rate of change for the CD4 cell count over time). Slopes during the three periods (treatment, 6 months after discontinuation, and before the reinitiation of therapy) were compared (e.g. if a high CD4 cell increase during antiretroviral treatment induced a lower decrease in the CD4 cell count after discontinuation). Patients started antiretroviral therapy with a median CD4 cell count of 469.5 cells/μl (range 351–665). The median treatment period before discontinuation was 37 months (range 18–48 months). The viral load became undetectable after a median of 5 months of treatment (range 3–16), and remained undetectable for a median of 14 months (range 4–26). Before discontinuation HIV RNA was detectable in the plasma for a median time of 12 months (range 4–26). During the overall period on treatment, patients experienced a mean CD4 cell increase of 8.13 cells/μl per month [P < 0.001; 95% confidence interval (CI) 5.34–11.76]; in particular, during the period on treatment with a detectable HIV plasma viral load, a median increase of 9.02 cells/month was detected. Considering the trend after their first detectable HIV-RNA plasma level, we observed a monthly mean increase of HIV-RNA log10 0.008 (P = 0.35; 95% CI −0.011–0.047). At discontinuation, patients had a median CD4 cell count of 777.5 cells/μl (range 520–1125) and a median HIV-RNA plasma level of 3.59 log10 (range 3.0–4.6). The median observational time after discontinuation was 10.5 months (range 4–22 months). During the first 6 months of discontinuation, the mean CD4 cell count showed a decrease of 44.40 cells/μl per month (P < 0.001; 95% CI 35.56–69.21), then a decrease of 17.46 cells/μl per month (P < 0.001; 95% CI 2.81–21.54) in the following months (Fig. 1). CD8 cells showed a trend of decline similar to CD4 cells. For the HIV-RNA plasma level a mean increase of 0.115 log10 per month (P = 0.017; 95% CI 0.040–0.370) was estimated for the first 6 months of discontinuation, followed by a mean increase of 0.070 log10 per month (P = 0.031; 95% CI 0.023–0.150) (Fig. 2).Fig. 1.: Trends of CD4 cells during treatment, after 6 months off-therapy, and at the end of observation.Fig. 2.: Trends of HIV plasma viral load during treatment, after 6 months off-therapy, and at the end of observation.When adjusting for CD4 cell counts and HIV-RNA plasma levels at the time of discontinuation, we estimated a significant increase of 0.09 log10 HIV RNA per month (P = 0.03; 95% CI 0.007–0.17) for each more 1 log10 HIV-RNA level plasma at discontinuation. The CD4 cell count and HIV-RNA rate of change were not associated with the CD4 cell value at baseline. The length of period of an undetectable plasma HIV-RNA level and the duration of treatment do not seem to be associated with response variables. No significant correlation between trends of CD4 cell counts and HIV-RNA slopes in the three periods analysed was found. CD4 cell counts below 350 cells/μl developed in three patients, after 10, 16 and 4 months of interruption, respectively. Overall, no clinical event or HIV transmission to sexual partners occurred during the period of discontinuation. The viral genotype was analysed in six patients with a viral load greater than 3 log10 copies/ml both 1 month before and 6 months after discontinuation. Five patients treated with a lamivudine-containing regimen showed the reverse transcriptase mutation, M184V, at discontinuation, together with other mutations mainly associated with resistance to thymidine analogues. In four of them a second genotype performed 6 months later revealed the disappearance of both lamivudine- and thymidine-associated mutations (except mutation K70R, which remained in one patient) (Table 1). The sixth patient, not treated with lamivudine, had a genotype that was substantially unchanged before and after therapy.Table 1: HIV genomic sequences from the six patients tested before and after discontinuation of antiretroviral therapy. The four patients whose mutation M184V disappeared after the interruption of lamivudine-containing regimens had a substantial decrease in CD4 lymphocytes (mean 360 cells/μl) after the discontinuation of therapy, accompanied in two cases by an increase in the HIV-RNA level of approximately 1 log10 (mean 0.92). By contrast, the decrease in the CD4 cell count was quite limited in the last patient treated with lamivudine, as well as in the patient treated with a lamivudine-sparing regimen. The present study shows that a number of patients continued to increase their CD4 cell counts despite a long period of low but detectable viral loads in the presence of a suboptimal treatment. This phenomenon was previously reported in patients treated with highly active antiretroviral therapy [2]. In addition, after therapy discontinuation some patients could stay without treatment for a long period of time, whereas others showed a rapid decline in CD4 cells. In our case, most patients were on double nucleoside therapy, yet not only did they increase their CD4 cell count, but also maintained quite stable HIV-RNA plasma levels for a long period of time. The hypothesis that this suboptimal therapy is responsible for this immune ‘homeostasis’ is supported by the fact that after therapy withdrawal the decrease in the CD4 cell count was rapid, even in the absence of a high viral load rebound. This phenomenon of CD4 cell decline could be caused by a rapid re-distribution in peripheral blood CD4 cells after stopping therapy, corresponding to the relatively rapid increase in CD4 cell count after initiating therapy, which has been ascribed to the redistribution of cells from lymphoid tissue [3], and may reflect the redistribution of lymphocytes back into tissues. The decline in the CD4 cell count continues thereafter at a pace slower than in the previous months, but still faster than the natural decline of 60 cells/μl per year observed in the absence of therapy [4]. Like the CD4 cell decline, the rate of increase of plasma HIV RNA is at a maximum in the first 6 months, even if at levels lower than in previous studies [5,6]. In a previous study [6], very low replication during highly active antiretroviral therapy was associated with a severe impairment of the HIV-specific immune response, resulting in very high HIV-RNA rebound levels after stopping therapy. In our cases, a continued viral replication persisting for 1 year before discontinuation may have sustained HIV-specific cytotoxic T lymphocytes to increase viral control during rebound after interruption. Genotyping analysis indicated that five out of six patients lose their M184V mutation in reverse transcriptase after therapy interruption. This mutation, associated with therapy with lamivudine, alters the catalytic core of the reverse transcriptase enzyme [7], and therefore affects the processivity of the enzyme and its replication kinetics. In agreement with these observations, our data show a relatively low viral load before discontinuation, whereas the subsequent loss of M184V leads to a CD4 cell count decrease, mostly together with increased viraemia (approximately 1 log). This suggests that the decreased processivity of HIV reverse transcriptase in lamivudine-treated patients (almost invariably harbouring the M184V mutation after virological failure) may be one of the factors associated with clinical benefit, even in the case of limited but continuous virus replication. In conclusion, the results from this study suggest the importance of new trials aimed at assessing whether the continuation of treatment with a suboptimal regimen may be associated in some patients with sustained clinical benefit, andand therefore may be preferred to discontinuation. Cristina Mussinia Roberto Bugarinib Carlo Federico Pernoc Andrea Antinoric Vanni Borghia Ada Bertolic Roberta D'Arrigoc Andrea Bedinia Nicola Mongiardoa Andrea Cossarizzad Roberto Espositoa Acknowledgements The authors would like to thank Dr J. Aberg for helpful comments and discussion.