Proviral HIV-DNA predicts viral rebound and viral setpoint after structured treatment interruptions

In HIV-1-infected patients with long-term undetectable viraemia on highly active antiretroviral treatment (HAART), we found that pre-HAART plasma viraemia and the baseline proviral DNA level were significantly associated with the viraemia setpoint during scheduled treatment interruptions. In long-term treated patients, pre-HAART viraemia may not be available, and in these circumstances proviral DNA, measured at the time of scheduled treatment interruption, can help to identify patients likely to reach a low viraemia setpoint after treatment interruption.

The long-term toxicity of antiretroviral therapy [¹], the emergence of HIV-1 drug-resistance [^2,3] and the cost of long-term treatment have driven new strategies of antiretroviral treatment including the delay of initiation of treatment [⁴], structured treatment interruptions [^5–8] or simplified maintenance therapy [^9–11]. In several clinical trials intermittent or simplified treatment has failed in a relatively high proportion of patients, whereas others have benefitted [^5–11]. Such patients maintained stable CD4 cell counts and low viraemia off therapy. The predictors of virological failure in such studies were previous mono or bi-therapy, baseline drug resistance mutation(s) and high HIV-1-RNA levels [^5–11]. The selection of patients who may benefit from treatment interruption or simplified therapy may be improved by taking into account treatment history, the detection of baseline resistance mutation(s) and an evaluation of the pool of infected cells by the quantitation of proviral DNA.

We have previously reported that in the Swiss–Spanish intermittent treatment trial (SSITT), which enrolled chronically infected patients with undetectable viraemia, only a low proportion of patients (17%) achieved a sustained viraemia level of less than 5000 copies/ml after long-term treatment interruption [⁸]. In that investigation, the viraemia level before the introduction of highly active antiretroviral therapy (HAART) was the best predictor of virological response. We now assess whether baseline proviral DNA levels predict virological response in all Swiss patients included in the SSITT.

In the SSITT, enrolled patients had been on HAART without previous failure and had undetectable viraemia for at least 6 months. Their HAART was interrupted for 2 weeks and restarted for 8 weeks. After four of these cycles, treatment was definitively interrupted at week 40. Virological failure was defined for the 0–40 week period by HIV-1-RNA levels greater than 50 copies/ml after 8 weeks of re-treatment [^8,12]. The quantification of cell-associated proviral DNA was performed as previously described, with a detection limit of 3 copies/10⁶ peripheral blood mononuclear cells (PBMC) [¹³]. The Cox regression model was used to analyse the relative risk of virological failure.

Baseline PBMC samples were available at baseline for 87 out of 97 of the Swiss patients (90%) enrolled in the SSITT. At the time of initiation of HAART, the median CD4 cell count was 361 cells/mm³ (range 1–1035) and the median viraemia was 4.53 log₁₀ RNA copies/ml (range 2.23–6.18). The median time on HAART before inclusion in the protocol was 27 months (range 9–45). All patients had detectable proviral DNA with a median of 2.25 log₁₀ copies/10⁶ PBMC (range 0.48–2.96). The proviral DNA level was associated with the pre-HAART CD4 cell count and viraemia (Spearman correlation R = −0.43 and 0.42; P < 0.001, respectively), whereas no correlation was found with the duration of viral suppression on HAART or baseline CD4 cell counts.

During the first 40 weeks, 26 out of 87 patients (30%) failed to decrease viraemia to less than 50 copies/ml after 8 weeks of re-treatment and were considered to be in virological failure. In univariate analysis, virological failure was significantly associated with pre-HAART viraemia, pre-HAART CD4 cell count and proviral DNA at baseline, but not with the baseline CD4 cell count (P = 0.48). In multivariate analysis, pre-HAART viraemia was the only independent predictor of virological failure (Table 1). No discriminating proviral DNA threshold for virological failure was observed.

Table 1: Predictors of virological failure to week 40.

On the other hand, baseline proviral DNA was significantly associated with the peak of viraemia rebound during the short period of interruption at weeks 2, 12, 22 and 32 (Spearman R = 0.51, 0.42, 0.53 and 0.44, respectively; all P < 0.001). The viraemia rebounds at week 2, 12, 22 and 32 were also significantly associated with the pre-HAART CD4 cell count (R = −0.51, −0.44, −0.42; P < 0.001 and R = −0.33, P = 0.008, respectively) and pre-HAART viraemia (R = 0.50, 0.49, 0.51, 0.39, respectively; P = 0.001).

Among the 56 patients who completed the four cycles and underwent the long-term interruption starting at week 40, pre-HAART viraemia (Spearman R = 0.66; P < 0.001) and baseline proviral DNA (R = 0.41; P = 0.002) were associated with the viraemia level after 12 weeks of final treatment interruption.

In previous studies, proviral DNA has been found to be a predictor of disease progression and virological outcome on HAART [^14–19], and to predict the level of residual viraemia after HAART [^20,21]. In this investigation, we have assessed the predictive value of proviral DNA in the context of treatment interruption in patients with undetectable viraemia on HAART. We found that proviral DNA predicted the failure to reach undetectable viraemia after re-treatment during short-term interruptions or the level of viraemia after long-term interruption. However, in multivariate analysis the pre-HAART plasma HIV-1-RNA level had a better predictive value than baseline proviral DNA. In clinical practice pre-HAART plasma HIV RNA corresponding to the viraemia setpoint is not always available either in the context of long-term treated patients or in patients treated at the time of primary HIV infection. In these circumstances, proviral DNA levels can be used as a surrogate marker of plasma HIV-1 RNA, and can therefore help to identify patients likely to reach low viraemia setpoints after treatment interruption. Such patients may potentially represent a group in whom the strategy of simplified maintenance therapy might be revisited.

Swiss HIV Cohort Study

The members of the above study are: M. Battegay, E. Bernasconi, H. Bucher, Ph. Bürgisser, M. Egger, P. Erb, W. Fierz, M. Fischer, M. Flepp (Chairman of the Clinical and Laboratory Committee), P. Francioli (President of the SHCS, Centre Hospitalier Universitaire Vaudois, CH-1011- Lausanne), H.J. Furrer, M. Gorgievski, H. Günthard, P. Grob, B. Hirschel, L. Kaiser, C. Kind, Th. Klimkait, B. Ledergerber, U. Lauper, M. Opravil, F. Paccaud, G. Pantaleo, L. Perrin, J.-C. Piffaretti, M. Rickenbach (Head of Data Center), C. Rudin (Chairman of the Mother and Child Substudy), J. Schupbach, R. Speck, A. Telenti, A. Trkola, P. Vernazza (Chairman of the Scientific Board), Th. Wagels, R. Weber, S. Yerly.

Sponsorship: This work was supported by the Swiss National Research Foundation (grant no. 3345-64120.00) and by the Swiss HIV Cohort Study (Swiss National Science Foundation, grant no. 3345-062041).