Discordant responses to HAART in HIV-1 patients: the need to focus on intervention

HAART has achieved relevant success in the treatment of individuals infected with HIV-1 and still holds great promise of improvement in view of the relatively long pipeline of drugs under study and of new drug classes that are or will soon become available. New classes of integrase inhibitors, entry inhibitors and maturation inhibitors, coupled with improved molecules belonging to the presently available classes will provide future modular options for the treatment of patients with virologic failure of multiple lines of treatment.

In everyday clinical practice, however, success of HAART, defined as suppression of detectable plasma virus and recovery/increase in CD4⁺ cell numbers after 6–12 months, does not exceed 40–60% of treated patients, mostly irrespective of the chosen regimen [1–4]. Attention to adherence, resistance testing and pharmacokinetic monitoring has considerably helped to curb the number of patients who fail to control virus replication and to increase CD4⁺ cells on a new HAART regimen. In addition, a relevant proportion (20–30%) of patients on a new HAART regimen show discordant immunovirological responses, either with successful viral load control but weak or no CD4⁺ cell increase (immunological discordant responses [IDRs], 10–17% of cases) or with failing control of viral load but with appropriate CD4⁺ cell increase (virological discordant responses [VDRs], 10–23% of cases) to HAART [1–4]. In early observational studies, a predominance of drug-experienced patients entering a protease inhibitor (PI)-containing regimen was included [1–3]. However, recent reports show that discordant responses are also readily registered in previously HAART-naive patients [5,6].

The phenomenon of discordant responses is not limited to economically advanced areas. Similar proportions of patients showing IDRs and VDRs have also been reported in resource-poor areas of the world [7]. Overall, some level of variability in the proportion of discordant patients has been observed between reports. This is mainly due to several factors, including lack of agreement on a unifying definition for discordance, absence of homogeneous selection of the observed patient groups, variability in the treatment history and in HAART regimens used at the time of study entry. Future efforts should be aimed at reaching a shared international definition that will improve patient care and provide more homogeneity and comparability among the cohorts. With these limitations in mind, it is important to note that a fairly constant and relevant risk (15–30%) of achieving an incomplete response may be attributed to any new patient, or to any new regimen started after the initial HAART failure. Therefore, awareness and attention to this clinical evolution on treatment are required in HIV outpatient clinics, and there is clearly a need for the study of possible lines of intervention and management.

The observation of discordant responses to HAART bears more relevant clinical consequences than a purely theoretical argument. Apart from the obvious clinical questions that everyone faces periodically in everyday practice – ‘should we change the HAART regimen in this IDR patient, in order to increase the CD4⁺ cells to more than 200/µl and discontinue opportunistic infection prophylaxis?’ and ‘how do we cope with the anxiety of the patient (and of the physician) that is not satisfied by lack of CD4⁺ cell increase or by persistence of virus replication?’ – more relevant clinical risks have been identified in these groups. Patients with discordant responses are clearly at higher risk of death or of AIDS-defining events and, therefore, this condition does not represent a relatively benign and harmless response to HAART that needs only watchful clinical observation [2,3,5,6]. In an early prospective cohort of patients beginning HAART with first-generation PIs, after 30 months of observation, AIDS-defining events or death were 21% in IDRs and 14% in VDRs compared with 2% in responders [3]. Another large study showed that, compared with full responders, IDR had a 3.38 mean risk of progressing to AIDS or death after a median of 18 months of HAART [2]. A more recent study on responses to HAART in over 1500 patients confirmed that discordant responses were independently associated with increased mortality. In this study, IDR was observed in 15.4% of the patients and had a 1.87-fold increased risk of mortality, and VDR patients accounted for 11.7% of the cohort and showed a 2.47-fold increased risk [5]. Thus, when viewed from a chronological perspective, improved treatment strategies and more potent regimens/drugs over the past few years have not significantly affected the relative risk of death for patients with discordant responses.

Risk factors for the development of either discordance condition have been elucidated partially over the last few years. Besides the possibility of alterations in adherence to therapy and in individual pharmacokinetic variations, patients with IDR are more likely to be older, have more advanced immune compromise with lower CD4⁺ T-cell nadirs, and are likely to have increased thymic exhaustion [1,3,8]. Earlier observations support these recent clinical associations. Long-term suppressive HAART may not result in the recovery of CD4⁺ T-cell counts to preinfection levels [9] as the degree of impairment at baseline may condition immune reconstitution [10]. In some cases, patients on a tenofovir–didanosine backbone combination may experience decreases in absolute CD4⁺ cell counts [11]; however, long-term treatment with this combination does not lead to increased IDR in drug-experienced patients [12,13]. Although the majority of patients with IDR are administered PIs, this probably represents a selection/treatment bias rather than true PI-associated effects. Prospective, randomized drug comparisons on the risk of developing discordant responses are needed to shed light on the role played by HAART regimens. In addition, patients with IDR have low CD4⁺ T-cell numbers with a hyperactivated/apoptotic phenotype and an increased size of the HIV reservoir as a possible expression of incomplete virus control, while their thymic function is reduced only partially [14]. Therefore, based on these findings, it remains to be evaluated whether HIV-associated alterations in innate immunity with derangement of natural killer (NK) cell phenotype and function in IDR patients [15–17] are involved in altered monocyte and dendritic cell editing [18–20] with impaired recovery of CD4⁺ cell numbers and function. In this case, potential interventions specifically addressing innate immunity [21] could contribute to improve the response to HAART.

Recent studies show that patients with VDR are younger, have higher viral load and are more likely to be on a PI-including treatment [22,23]. Some effects of HAART regimen components (PIs in this case) may result in immune effects that could contribute to the genesis of discordant responses. Unfortunately, there are no randomized prospective trials or cohorts comparing the effect of different treatment regimens on the probability of developing discordant responses to HAART for either VDR or IDR patients. The possibility exists that a clinical selection bias could explain the higher frequency of patients treated with PI-containing regimens, as opposed to non-nucleoside reverse transcriptase inhibitor (NNRTI)-containing regimens. The higher starting viral load in these patients and the inclination of caring clinicians to provide treatments with high potency and high genetic barriers could have contributed to increased prescription of a PI-containing regimen [22]. Reduced fitness of replicating virus could also explain sustained CD4⁺ cell numbers in the presence of continuing viral replication in these VDR patients [24,25]. Protease inhibitors interact with proteasome activity and nuclear factor-κB nuclear translocation [26,27], thus, potentially contributing to reduced T-cell activation and increased CD4⁺ T-cell survival. Observations directly comparing the risk of developing VDR based on either the different PI used (e.g., lopinavir/ritonavir vs atazanavir) or the choice between NNRTI and PI are required to clarify this point.

There is a cogent need for an improved and rational approach to the clinical management of patients with discordant responses to HAART, particularly in view of their risk of disease progression/death, and of the considerable burden in terms of quality of life. In this regard, among possible options for therapeutic intervention, immunoadjuvant, immunosuppressive or antiretroviral therapy intensification strategies should be considered.

Among immunoadjuvants, the administration of subcutaneous or intravenous recombinant interleukin (rIL)-2 in addition to HAART results in significant increases in CD4⁺ cell counts [28–30]. In these patients, CD4⁺ T-cell recovery is indirectly correlated with nadir CD4⁺ cell. The presently studied dosing regimens still cause relevant general toxicities, including fever, rash, anorexia, headache and myalgias, for daily administrations of 4.5 MU every 12 h for 5 days every 51 days and also for 5 MU/day [29–31]. However, most of the presently published trials of rIL-2 were set to evaluate the possibility of improving control of viral replication in unselected patient groups or to evaluate the possibility of sparing HAART in treatment interruption designs. No trial has ever been performed addressing the usefulness of rIL-2 for the recovery of CD4⁺ cells in IDR patients, or in order to decrease the proportion of discordant responses to HAART. In addition, lower doses may show improved tolerability and still target NK cells and innate immunity, possibly improving the editing of CD4⁺- or CD8⁺-mediated immune responses [32]. Pathways for T-cell reconstitution in HIV-1-infected patients include stimulation of thymic differentiation, support of extrathymic differentiation or peripheral expansion. IL-7 has the potential to support replication of CD4⁺ cells at both thymic and peripheral levels [33,34]. Ongoing trials are evaluating its safety in dose-escalating Phase I trials in healthy volunteers (NCT00099671, NIH). In view of the strong inverse correlation between serum IL-7 levels and CD4⁺ T-cell counts [35] with already significantly elevated IL-7 plasma levels in IDR patients [14], its potential use in IDR patients remains to be determined.

The rationale for the use of immunosuppressive drugs is represented by the presence of CD4⁺, CD8⁺ and NK cell activation and of apoptotic activation in patients with IDR to HAART. The administration of either ciclosporin A (CsA) or anti-TNF monoclonal antibodies (etanercept) or of mycophenolate [34,36] could provide efficient control or excessive activation of the immune system. Etanercept 25 mg twice weekly reduces IL-6 and C-reactive protein levels in HIV-1-infected patients treated with subcutaneous rIL-2 [37] and improves CD4⁺ T-cell recovery in sputum-positive tuberculosis patients treated with antimycobacterial drugs [38]. Trials specifically targeting discordant patient cohorts are unfortunately lacking. CsA has so far been administered as adjunctive initial treatment in combination with HAART for 2 weeks to a small number of patients (nine) with primary HIV infection, with good responses at 48 weeks [39]. In a small randomized trial (ACTG A5138) in 20 chronically infected drug-naive patients treated for 2 weeks, trends in increased CD4⁺ and CD8⁺ T cells and decreased T-cell activation were observed [39,40]. However, caution should be used in interpreting these results. Since these trials had completely different primary objectives and patient selection/follow-up criteria, their results and conclusions cannot be transferred or applied to IDR patients. Specific trials designed to assess the usefulness of CsA in IDR are needed to validate or discard this treatment possibility.

An additional option that needs to be considered in these patients is represented by antiretroviral treatment intensification. Increasing levels of HIV DNA in CD4⁺ cells of IDRs, as opposed to full responders and to VDR patients, and the absence of significant decreases of T-cell receptor excision circles [14] suggest the presence of residual viral replication. This could support the option of adding a new drug class to the ongoing HAART. For example, the effects of either adding enfuvirtide to an existing PI-containing HAART, or switching to or adding an integrase inhibitor to a NNRTI regimen could be evaluated. Additional blocking of virus entry or inhibition of integration may provide improved control of low-level residual replication in IDR patients, leading to decreased cell activation and apoptosis with ultimate recovery of CD4⁺ cells. As far as patients with VDR are concerned, persistence of detectable virus replication suggests that they may also benefit from treatment intensification, provided that resistance testing rules out the presence of resistance mutations to components of the ongoing regimen. It should be noted that the above considerations only represent reasonable speculations. Without any support from clinical controlled studies encompassing pharmacokinetic and virology analyses, these strategies do not represent good clinical options. Unfortunately, the lack of published systematic studies or controlled trials in this area is secondary to a lack of studies at all and reflects the neglected condition of discordant responses to HAART. If these aspects could enter the agenda of international and national agencies and study groups, the data gathered through multicenter controlled studies would also benefit patients that are being treated in developing areas where discordant responses to HAART are present [7] and are likely to determine more marked adverse clinical effects.

In conclusion, there is a general need for the organization of controlled observational and intervention studies with regard to discordant responses to HAART. It is time to move on to more structured trials and analyses in the relevant fraction of patients that display discordant responses to HAART. Observations derived from a limited number of small reports in studies with different end points are not warranted, and new approaches using presently available immunoadjuvant, immunosuppressive, HAART intensification strategies, alone or in combination will result in improved clinical care for these patients.

Financial disclosure

Supported by grants from Ministero della Salute – Istituto Superiore di Sanità, V Progetto AIDS and VI Progetto AIDS.