| |
Treatment of Acute HIV-1 Infection: Is It Coming of Age?
|
| |
| |
The Journal of Infectious Diseases Sept 15, 2006;194:725-733
Sabine Kinloch-de Loes
Royal Free Centre for HIV Medicine, Department of Infection and Immunity, Royal Free and University College Medical School, London, United Kingdom
(See the article by Hecht et al., and the brief report by Streeck et al. below)
The optimal time to start antiretroviral therapy for HIV-infected patients remains unknown [1-4]. Leaving public health considerations regarding transmission aside, if we are indeed able to diagnose HIV infection in patients at a very early stage, should we intervene with antiviral agents and, if so, for how long [5-7]? What is the potentially achievable long-term impact of highly active antiretroviral therapy (HAART) initiated around the time of HIV seroconversion, if such treatment is subsequently discontinued: lowering of the viral set point, preservation of CD4+ T cells, a decrease in rates of disease progression, long-term control of HIV viremia, or even viral eradication [8-10]? In contrast, one may argue that very early initiation of HAART may still be provided too late to have a major impact, because virus-induced immunopathogenesis has already taken hold [11-13]. This realization would justify a more conservative treatment approach and the initiation of HAART only during chronic infection, because this approach still allows for substantial immune reconstitution in terms of the CD4+ T cell count, as well as a dramatic decrease in morbidity and mortality, while preventing the potential development of early resistance to antiviral agents and reducing both drug toxicity and cost [14]. The question of the timing of HAART initiation is not trivial, particularly because such studies as the Strategies for Management of Anti-Retroviral Therapy study suggest that HAART initiated during chronic infection cannot be routinely discontinued without substantial risk [1-4, 15].
Important advances have been achieved in increasing the understanding of the events associated with the initial encounter between HIV and the immune system, particularly at the mucosal level, and of the impact of the virus in later infection [11-13, 16]. Virological parameters, such as the level of HIV DNA in peripheral blood mononuclear cells, have been shown to have a predictive value in terms of the viral set point after discontinuation of HAART [17]. Early initiation of HAART has also been associated with a decrease in cell-associated HIV-1 DNA, which seems to be more substantial than that noted when antiretroviral treatment is initiated during chronic infection [18-21]. Immunological studies have better defined the type of cellular immune responses that are preserved by early initiation of HAART and those that are potentially associated with viral control [22-24]. We also know that overall HIV plasma viremia often remains undetectable in treated, acutely infected subjects who are adherent to HAART and that a rapid rate of progression to AIDS was decreased by early intervention with HAART in one cohort [10, 25].
Previous experience in the treatment of acute HIV infection demonstrated that, with the advent of powerful HAART, what seemed to be an easy task proved to be more complicated than initially was thought [26]. Virus eradication is not around the corner, and neither is long-term control of viral replication after termination of HAART, even with the addition of immunomodulators [10]. Although an impact of early treatment on the CD4+ T cell count and on Centers for Disease Control and Prevention grade B conditions was noted in the seminal randomized trial of zidovudine monotherapy, cohort studies of seroconverters have not unequivocally shown a positive influence of early initiation of HAART on the CD4+ T cell count and viral load after discontinuation of treatment [9, 27-29]. In the absence of randomized, placebo-controlled studies of HAART, observational cohort studies have been the main source of information to guide treatment recommendations.
An article [30] and a brief report [31] in the current issue of the Journal of Infectious Diseases address the potential influence of HAART on virological and immunological surrogate markers in cohorts of HIV seroconverters. Hecht et al. [30] have used the Acute Infection and Early Disease Research Program observational cohort of American and Canadian subjects who were enrolled within 1 year of HIV seroconversion. The subjects chose whether to receive HAART or not, and, in the present analysis, subjects who had seroconversion within the past 6 months were selected for analysis on the basis of the following criteria: the treatment group had to initiate HAART (HAART duration, >12 weeks; >3 antiviral agents were used) either within 6 months of receiving a negative or indeterminate result of an HIV antibody test or while a less-sensitive ELISA antibody test revealed a standardized optical density of <0.75. The untreated subjects met the same criteria, did not receive HAART, and were monitored for at least 6 months after enrollment. Treated subjects were classified into 2 groups: the "acute treatment group" (i.e., subjects receiving treatment within 2 weeks of seroconversion) and the "early treatment group" (i.e., subjects receiving treatment between 2 weeks and 6 months after seroconversion). The study outcomes were the plasma HIV-1 RNA levels and CD4+ T cell counts determined at 24, 48, and 72 weeks of untreated observation after treatment cessation for the intervention group, compared with the same outcomes determined at the same periods of observation after enrollment for the nonintervention group.
A total of 337 subjects in the untreated group, as well as 58 subjects in the treated group (13 of whom received acute treatment and 45 of whom received early treatment), met the selected criteria for inclusion. No significant demographic or clinical differences were noted between the 2 groups, except for the differences in both the initial viral load between the untreated and acute treatment groups and the number of weeks from infection between the acute treatment and early treatment groups. HAART was given for a median of 1.5 years; undetectable viral loads (<500 HIV-1 RNA copies/mL) were achieved in all subjects receiving early treatment and in 11 of 13 subjects receiving acute treatment.
CD4+ T cell count and viral load benefits were noted in the treated group, compared with observations noted in the untreated group, at 24 weeks after treatment discontinuation. The longer-term viral load benefit (up to 72 weeks) was noted in the acute treatment group only when using adjusted analysis for baseline CD4+ T cell count and viral load. However, the CD4+ T cell count benefit persisted over 72 weeks for the early treatment group, in both adjusted and unadjusted analyses, whereas it persisted only in adjusted analyses for the acute treatment group.
Not surprisingly, subjects experienced viral rebound, and control of HIV viremia after discontinuation of HAART did not occur in the majority of subjects [10]. However, the benefit observed with regard to surrogate markers is consistent with our understanding of the rapid onset of HIV immunopathogenesis and the potential impact of HAART initiated during acute infection [11-13, 22, 23]. These encouraging results have to be interpreted within the context of the limitations of this study that the authors clearly point out, such as the absence of randomization, possible biases associated with the reasons for stopping HAART, the small number of subjects, and heterogeneous treatment regimens and durations. The potential impact of early HAART suggested in the article by Hecht et al. [30] was mirrored in a recent analysis from the CASCADE cohort, when initiation of HAART during the first year of infection was evaluated. A median duration of HAART of 1.1 years was associated with a persistent effect on the viral load and CD4+ T cell count for up to 5 years after seroconversion [32].
In contrast to the results of the study of Hecht et al. [30], no major influence of early initiation of therapy was noted in the study of Streeck et al. [31], when a short-term HAART intervention was used, in terms of the viral load and the CD4+ T cell count noted at 24 weeks after termination of therapy. This cohort recruited in Berlin, Germany, included 20 consecutively recruited, symptomatic, acutely infected subjects, as determined by clinical and laboratory markers (detectable HIV-1 viremia and a negative result of ELISA or a positive result of ELISA with <3 bands noted on Western blot analysis), who chose (or did not choose) to initiate HAART for 24 weeks. This period was followed by discontinuation of treatment and an additional follow-up of 6 months, when CD4+ T cell counts and viral loads were compared between the treated and untreated groups. Analysis of HIV-1-specific CD8+ T cell responses was performed before and after discontinuation of treatment, by use of an interferon (IFN)-y enzyme-linked immunospot assay and intracellular cytokine and surface staining. CD107a expression was used as a surrogate marker for T cell degranulation.
All of the 12 subjects who chose to initiate HAART had an undetectable viral load and showed improvement in their CD4+ T cell counts by week 24, compared with the findings noted for untreated subjects. The HIV-1-specific CD8+ T cell responses of the treated subjects also showed differentiation from effector memory cells toward an effector cell phenotype by week 24, as well as higher levels of HIV-1-specific CD8+ T cell responses at week 48. Levels of secretion of interleukin-2 (IL-2) by HIV-1-specific CD8+ T cells, however, remained low.
These results show the very good antiviral effect on viremia of a 24-week course of HAART initiated at primary HIV infection [20]. However, control of viremia did not lead to a major influence on the CD4+ T cell counts and viral loads noted 6 months after discontinuation of HAART, compared with the findings for untreated subjects. At treatment discontinuation, this early HAART intervention of short duration was associated with HIV-1-specific CD8+ T cell responses higher than those noted in the untreated group, but it was not associated with persistence of the HIV-1-specific CD8 T effector cell phenotype that has been suggested to be associated with viral control [8, 33]. It is unclear whether the chosen treatment duration and its potential impact on HIV-1-specific T cell responses is the major factor associated with the absence of effect on HIV surrogate markers after cessation of therapy [34, 35]. To further clarify, in a larger, randomized study, the potential impact of a short period of early intervention with HAART on the time to a CD4+ cell count of <350 cells/L, we will need to wait for the results of the ongoing Short Pulse Anti-Retroviral Therapy At HIV sero-Conversion trial, which compares no treatment with treatment given for 3 or 12 months within 6 months of infection [36].
In conclusion, there is a strong case for testing interventions early after infection, in an attempt to alter the course of the disease and decrease the overall duration of exposure to drugs [11-13, 18-20, 22, 23, 25, 32]. It will be important to further assess the role of cell-associated DNA, T cell activation markers, and immune responses, in addition to viral loads and CD4+ T cell counts, in determining which subjects among the heterogeneous acutely infected population are most likely to benefit from early treatment initiation [9, 10, 17, 20, 37-39]. Data from cohort studies of persons with HIV seroconversion and from clinical trials should further clarify the type and duration of therapy to be tested in future randomized trials [9, 10, 20, 28, 29, 36, 37]. The present public health effort to rapidly identify early HIV infection should be a major contributor to this effort.
A Multicenter Observational Study of the Potential Benefits of Initiating Combination Antiretroviral Therapy during Acute HIV Infection
The Journal of Infectious Diseases 2006;194:725-733
Frederick M. Hecht,1 Lei Wang,4 Ann Collier,4 Susan Little,2 Martin Markowitz,6 Joseph Margolick,7 J. Michael Kilby,8 Eric Daar,3 Brian Conway,9 and Sarah Holte,4,5 for the AIEDRP Network
1University of California, San Francisco, San Francisco, 2University of California, San Diego, San Diego, and 3University of California, Los Angeles, Los Angeles; 4Fred Hutchinson Cancer Research Center and 5University of Washington, Seattle, Washington; 6Aaron Diamond AIDS Research Center, New York, New York; 7Johns Hopkins University School of Public Health, Baltimore, Maryland; 8University of Alabama at Birmingham, Birmingham; 9University of British Columbia, Vancouver, Canada
Abstract
Objective. Uncontrolled studies have suggested a benefit, after treatment discontinuation, of initiating highly active antiretroviral therapy (HAART) during primary human immunodeficiency virus (HIV) infection. We assessed whether initiation of HAART within 2 weeks of (acute treatment) or between 2 weeks and 6 months after (early treatment) HIV seroconversion was associated with improvements in the viral load and the CD4+ T cell count after discontinuation of treatment in an observational cohort.
Methods. Subjects from the multicenter Acute Infection and Early Disease Research Program cohort were enrolled in the present study within 6 months of HIV seroconversion and self-selected whether to initiate HAART. Subjects who received acute (n = 13) or early (n = 45) treatment received HAART for at least 12 weeks and then subsequently stopped treatment, whereas untreated subjects (n = 337) declined treatment. HIV RNA levels and CD4+ T cell counts at 24, 48, and 72 weeks after treatment cessation in the 2 treatment groups were compared with those noted in the untreated group during the same periods of observation after enrollment.
Results. The acute treatment group had lower mean HIV RNA levels at 24 weeks without therapy (-0.48 log10 copies/mL [95% confidence interval {CI}, -0.82 to -0.13 log10 copies/mL]) and higher mean CD4+ T cell counts (112 cells/uL [95% CI, 20-205 cells/uL]), compared with the untreated group at 24 weeks. The differences in the laboratory values for the acute treatment group versus the untreated group at 72 weeks without therapy were as follows: for the HIV RNA level, -0.35 log10 copies/mL (95% CI, -0.91 to 0.21 log10 copies/mL) and, for the CD4 T+ cell count, 112 cells/uL (95% CI, -15 to 213 cells/uL). The early treatment group had lower HIV RNA levels at 24 weeks than did the untreated group, but differences were no longer apparent by week 48; CD4+ T cell counts were higher in the early treatment group at week 24 (116 cells/L [95% CI, 75-157 cells/uL]) and week 72 (70 cells/uL [95% CI, 2-138 cells/uL]).
Conclusions. Initiation of HAART within 2 weeks of antibody seroconversion was associated with viral load and CD4+ T cell count benefits for 24 weeks after termination of HAART, with there being trends toward a longer-term benefit. Later initiation of HAART was associated with a persistent but decreasing CD4+ T cell count benefit and a loss of the viral load benefit by week 72 after discontinuation of treatment.
Introduction
The long-term course of HIV infection can be accurately predicted based on the HIV RNA level that is established soon after infection occurs [1, 2]. The viral load typically becomes high during primary HIV infection and then decreases under the pressure of immune responses [3-5]. Individuals with plasma viral loads that remain elevated during the first year of HIV infection have a high risk of disease progression, whereas those with low viral loads have a more favorable course of disease [1]. This finding implies that there are key early events in the pathogenesis of HIV infection that determine the long-term pace of disease progression.
A critical question is whether it is possible to intervene early in the course of HIV infection to improve the subsequent course of disease. One reason why more-effective immune responses do not develop may be that HIV targets CD4+ T lymphocytes and preferentially infects activated CD4+ T cells [6]. To the extent that CD4+ T cells are important in immune responses to HIV, the virus may disrupt the early development of these responses by directly or indirectly triggering the death of the CD4+ T cells that most effectively respond to the virus. If antiretroviral therapy (ART) is initiated during early infection, the disruption caused by HIV may be minimized, allowing better development of these early CD4+ T cell anti-HIV responses.
One randomized, controlled trial of zidovudine monotherapy given for 6 months to persons with primary HIV infection found that, compared with the placebo group, the treatment group maintained higher CD4+ T cell counts for 12 months after cessation of treatment [7]. However, HIV RNA levels did not appear to be altered by zidovudine monotherapy once treatment was stopped. A pilot study of 8 persons who had HAART initiated during primary HIV infection and then stopped receiving HAART according to a structured treatment interruption protocol found that 5 of 8 persons maintained an HIV RNA level of <500 copies/mL at a median of 6.5 months after cessation of HAART [8]. In contrast, only 4 of 109 persons in a historical comparison group achieved such low viral loads. At longer-term follow-up, however, this level of control of viremia was not maintained in most persons [9]. Despite the provocative findings of these studies, it remains unclear whether the long-term findings regarding viral loads and CD4+ T cell counts achieved with this treatment approach are better than the findings that would have been achieved had the treatment not been given.
A randomized, controlled trial comparing HAART with no treatment would be the strongest methodological means of evaluating the benefit of a limited course of HAART in individuals with primary HIV infection. Such a trial is difficult to perform, however, because many potential subjects are not willing to have the treatment decision assigned by chance, because many physicians are not at equipoise given the existing data, and because recruitment would be challenging. To further explore this question, we analyzed data from the Acute Infection and Early Disease Research Program (AIEDRP) cohort, a multicenter observational cohort study of acute and early HIV infection. Although subjects were nonrandomized and were heterogeneous with respect to treatment regimens, evaluation of these subjects may nonetheless offer insights regarding the potential benefits of HAART initiated during acute and early HIV infection.
SUBJECTS AND METHODS
Subjects. The AIEDRP cohort study is a multicenter, observational cohort study of persons enrolled within 1 year of having HIV antibody seroconversion. Participants in the cohort self-selected whether or not to start receiving HAART. For this analysis, we recruited persons enrolled at 10 of the original study sites located in the United States and Canada. Participants in the cohort had to meet one of the following criteria for acute or early HIV infection at enrollment: (1) a negative or indeterminate result of an antibody test and an HIV RNA load of >5000 copies/mL, (2) a documented negative result of an HIV antibody test performed within 12 months of enrollment of a participant in whom the current HIV antibody test result is positive, or (3) a history consistent with recent HIV infection and an optical density of <1.0, as determined by a less-sensitive EIA antibody test [10, 11]. A history consistent with recent HIV infection meant that a subject had no prior positive result of HIV antibody tests or had not previously received treatment for HIV infection, had a CD4+ T cell percentage of >14%, and had self-reported a recent negative HIV antibody test result or a recent illness consistent with an acute retroviral syndrome.
For this analysis, participants had to meet additional criteria specific to inclusion in either the treated or untreated groups. The treated subjects had to initiate HAART either within 6 months of receiving a negative or indeterminate result of an HIV antibody test or while a less-sensitive EIA antibody test revealed a standardized optical density of <0.75 (a finding consistent with a duration of <6 months from antibody seroconversion). The untreated subjects met the same study-entry criteria as did the treated subjects, but the untreated subjects did not receive HAART and had to be monitored for at least 6 months after enrollment. HAART was defined as treatment with >3 antiretroviral agents. The treated subjects had to maintain treatment for at least 12 weeks but then stop treatment for least 4 weeks. Subjects elected whether or not to start HAART and typically had knowledge of at least one viral load measurement and CD4+ T cell count when making that decision. HAART regimens were either site-specific protocols or regimens chosen by a subject's physician. The subjects in the study were enrolled in protocols approved by the institutional review boards of each of the participating study centers.
Analysis. The key end points of the study were defined as the HIV RNA levels and the CD4+ T cell counts determined at 24, 48, and 72 weeks of untreated observation. For the untreated subjects, this was the time since enrollment in the study, with censoring done at the time that HAART was initiated, if this occurred. For the treated subjects, the observation period started 4 weeks after treatment was stopped, to allow return of viremia after treatment discontinuation. The time receiving treatment or any time occurring before treatment was not counted for subjects in the treatment group; one result of this approach is that the treatment group had HIV infection for a longer duration than did the untreated group at each time point in the analysis. The treated group was divided into 2 subgroups according to predetermined cutoff periods: subjects who started receiving HAART within 2 weeks of antibody seroconversion (i.e., the acute treatment group) and subjects who started receiving HAART between 2 weeks and 6 months after antibody seroconversion (i.e., the early treatment group).
To compare viral loads and CD4+ T cell counts at 24, 48, and 72 weeks, all data collected for study subjects were analyzed using natural splines with 2 df to allow for nonlinear trajectories of viral load measurements and CD4+ T cell counts over time. We used this approach because the data are unbalanced and because definition of a specific value at week 24, 48, or 72 was not possible for many study subjects. The spline coefficients were treated as random effects, and population-level parameters were obtained using linear random-effects models with the following outcomes: a log10 viral load or a CD4+ T cell count modeled as a natural spline on time with 2 df [12]. This method accounts for any differences in the frequency of measurements between groups. A fixed effect for treatment cohort status was included to make comparisons between the 3 cohorts. Differences in log10 viral loads and CD4+ T cell counts at 24, 48, and 72 weeks after day 0 of the study were calculated using linear combinations of the estimated model parameters with spline values for those time points. Confidence intervals were obtained using linear combinations of the variance/covariance for the estimated spline coefficients.
Role of the funding source. AIEDRP is funded by the US National Institute of Allergy and Infectious Diseases (NIAID). NIAID staff provided input regarding the design of network data collection, which was used for the present study, and they reviewed the final manuscript without suggesting changes. Otherwise, sponsors of the study had no role in the study design, data analysis, data interpretation, or writing of the report. The corresponding author (F.M.H.) had full access to all the data and full responsibility for the decision to submit the report for publication.
RESULTS
There were 337 persons in the untreated group and 58 persons in the treatment groups (13 of whom were considered to be in the acute treatment group and 45 of whom were considered to be in the early treatment group). There were no significant demographic differences between these groups (table 1). The subjects in the acute treatment group had a higher mean viral load and tended to have a lower mean CD4+ T cell count at study entry than did the subjects in either the untreated or early treatment groups, but the differences were not statistically significant.
On the basis of our selection criteria, all treated subjects received >3 antiretroviral agents. For 12 (92%) of the 13 participants in the acute treatment group and for 35 (80%) of the 45 participants in the early treatment group, treatment was initiated with regimens containing a protease inhibitor. One of the subjects in the acute treatment group and 4 of the subjects in the early treatment group who were receiving a protease inhibitor also received a nonnucleoside reverse-transcriptase inhibitor as part of their regimen. Of the remaining subjects, 4 (9%) of those in the early treatment group received a nonnucleoside reverse-transcriptase-based regimen, and 1 participant (8%) in the acute treatment group and 5 participants (11%) in the early treatment group received an abacavir-based regimen. All of the participants in the early treatment group had HIV-1 RNA levels of <500 copies/mL, as did 11 of 13 participants in the acute treatment group. The median duration of ART in both treatment groups was 1.5 years (table 1).
In the untreated group, viral loads initially decreased (figure 1A) and then slowly increased over time. The initial decrease is consistent with the viral load decrease that is expected very early in untreated HIV infection, which presumably occurs as developing immune responses control viremia. In both treatment groups, the viral load initially increased rapidly after ART was stopped, and it then increased slowly over time (figures 1B and 1C). At 24 weeks of posttreatment observation, both treatment groups had HIV-1 RNA levels that were about 0.5 log10 copies/mL lower in unadjusted analyses than those in the untreated group (table 2). After adjustment for the viral loads and CD4+ T cell counts at baseline, the difference in the viral load, when compared with that in the untreated group, remained similar in the early treatment group but was 0.69 log10 copies/mL lower in the acute treatment group. In the acute treatment group, the viral load remained 0.35 log10 copies/mL lower than that in the untreated group, in unadjusted analysis after 72 weeks of untreated observation; however, this difference was no longer statistically significant. In adjusted analysis, however, the viral load advantage for the acute treatment group at 72 weeks (0.68 log10 copies/mL) was similar to that noted at 24 weeks (0.69 log10 copies/mL), and the difference continued to be statistically significant. In contrast, the mean viral load in the early treatment group became similar to that in the untreated group after 24 weeks (figure 1D), and there was no statistically significant difference in the viral load between the 2 groups at 48 and 72 weeks, in either the unadjusted or adjusted analysis (table 2).
CD4+ T cell counts decreased gradually in the untreated and treated groups (figure 2). CD4+ T cell counts remained >100 cells/L higher in the acute treatment group than in the untreated group throughout the period of observation (table 3), in both unadjusted and adjusted analyses. The acute treatment group had a statistically significant difference in the CD4+ T cell count in both unadjusted and adjusted analyses at 24 weeks. The difference remained statistically significant only in the adjusted analyses at weeks 48 and 72, because there were fewer observations; however, the magnitude of difference remained similar in both adjusted and unadjusted analyses (table 2).
The early treatment group had a mean CD4+ T cell count that was >100 cells/L higher than that noted in the untreated group at 24 and 48 weeks (figure 2 and table 3). There was a statistically significant CD4+ T cell count advantage in the early treatment group, compared with the untreated group, in both unadjusted and adjusted analyses at 24, 48, and 72 weeks. However, the magnitude of the CD4+ T cell count advantage decreased in both adjusted and unadjusted analyses after week 48 (table 1 and figure 2).
We also performed analyses of the viral load and CD4+ T cell count trajectories, dividing the treatment group into subjects receiving treatment for more or for less time than the median duration of ART. We found no differences associated with treatment duration (data not shown), but our power to detect differences is limited because of the small sample sizes resulting from subdividing the treatment group.
DISCUSSION
We found that initiation of HAART within 2 weeks of HIV antibody seroconversion was associated with sustained viral load and CD4+ T cell count benefits for up to 72 weeks after termination of therapy. The differences between the acute treatment group and the untreated group, however, were statistically significant only in analyses that adjusted for the CD4+ T cell count and viral load at baseline. Because study subjects typically decided whether to start receiving ART after they knew their viral load and CD4+ T cell count, the untreated group is likely to be enriched for persons who rapidly achieved low viral loads and maintained good CD4+ T cell counts. Although our statistical adjustment will account for this selection bias, it is likely to overcorrect for higher viral loads that are associated with earlier stages of acute HIV infection. Thus, the unadjusted analyses are more conservative, and the adjusted analyses may overstate the benefits of treatment during acute HIV infection.
Our results suggest that treatment given during acute HIV infection may modify the long-term course of disease; however, the results must be viewed cautiously because of the lack of statistical significance of the CD4+ T cell count and viral load benefits at 72 weeks, the small sample size, and the need for longer follow-up to assess the durability of benefit. The 0.48-log10 reduction in viral load observed at 72 weeks in unadjusted analyses would be clinically significant. In previous treatment studies, a >0.4-log reduction in viral load was biologically significant [13], and a two-thirds-log reduction in viral load, which was the magnitude of viral load reduction that we observed in adjusted analysis at 72 weeks, corresponded to an about 50% reduction in the risk of disease progression over 6 years [2].
In contrast to subjects in the acute treatment group, subjects who initiated HAART later but within 6 months of seroconversion appeared to receive more temporary benefits. After stopping therapy for 24 weeks, the early treatment group had a viral load that was lower than that in untreated subjects, but this advantage was lost by 48 weeks. There was a more sustained CD4+ T cell count advantage in the early treatment group, when compared with the untreated group. The magnitude of the CD4+ T cell count benefit, however, appeared to decrease 72 weeks after stopping treatment; this finding is potentially consistent with a delayed effect of the viral load rebound on CD4+ T cell counts. It is notable that the CD4+ T cell count benefit remained statistically significant in both adjusted and unadjusted analyses for 72 weeks; additional follow-up may help to resolve the question of whether a more modest CD4+ T cell count benefit persists with longer follow-up, despite the lack of a sustained difference in viral load between the early treatment and untreated groups.
There are both parallels and differences between our findings and those of Desquilbet et al. [14], who recently reported an observational comparison of persons who did and did not receive ART for acute or early HIV infection. Similar to the early treatment group in our study, the treatment group in the study of Desquilbet et al. [14] also had HIV-1 RNA levels that were initially lower than those in the untreated group; however, by 12 months after treatment, the HIV-1 RNA levels in the groups were similar. When Desquilbet and colleagues analyzed participants who met our criteria for acute infection separately, they did not note a difference in the HIV-1 RNA levels of these subjects, compared with the levels noted in the untreated subjects. However, their study included only 7 subjects with acute HIV infection and a smaller overall study population. The small sample size means that there is higher risk of a type II error of concluding that there is no benefit resulting from treatment initiated during acute infection, when, in fact, there is a benefit. It is also possible that subtle differences in the duration of infection before initiation of HAART in the 2 studies may influence results.
The CD4+ T cell count benefit that we noted is consistent with that found by Kinloch-de Loes et al. [7] in an earlier trial of zidovudine in subjects with early HIV infection, although the magnitude of benefit that we noted was larger, possibly because of the greater potency of the regimens used in our study. Of note, the earlier study also found a CD4 + T cell count benefit resulting from early treatment, despite a lack of significant improvement in viral load. Although earlier studies of acute treatment of HIV infection have suggested that there are no dramatic, sustained virological benefits of early treatment of HIV infection [9, 15], they have not had significant power or an adequate comparison group to determine whether there is a virological benefit of the magnitude that we observed.
The exact mechanism of the sustained viral load and CD4+ T cell count benefit noted after initiation of HAART for acute HIV infection cannot be answered by the present study. The results, however, are consistent with the hypothesis that early initiation of ART improves immune responses to HIV infection, because HIV-1 RNA levels were persistently lower in the acute treatment group. Treatment also decreases T cell activation, which is high during acute HIV infection [16] and has been shown to be a strong predictor of CD4+ T cell depletion [16, 17]. It is possible that early treatment leads to a shorter period of high activation of T cells, potentially reducing damage to the immune system. Recent data from rhesus macaques with acute simian immunodeficiency virus infection indicate that there is far more widespread infection of memory CD4+ T cells than previously was appreciated, with 30%-60% of such cells infected and killed within weeks of infection [18] and with massive depletion of the memory CD4+ T cells in gut-associated lymphatic tissue [19], which may result in long-term damage to the immune system that influences disease progression. Early treatment may reduce initial depletion of CD4+ T cells and might be responsible for the sustained CD4+ T cell count benefit that we found.
The present study has several limitations. Treatment was not assigned randomly, and unmeasured differences in the study groups may account for our results. Treatment regimens were heterogeneous, the duration of treatment varied between participants, and the data on treated persons did not include assessment of the reasons for stopping therapy. Results from the present study characterize the impact of treatment initiated during acute HIV infection for only 72 weeks after discontinuation of therapy, and the benefits may wane, similar to our findings for the early treatment group. The acute treatment group was small, and the viral load and CD4+ T cell count differences remained significant at 72 weeks only in adjusted analyses. As noted, it is possible that the statistical adjustments overcorrected for differences at baseline. The results at 72 weeks depend, to some extent, on the nonparametric spline model and earlier data, rather than complete data for all study subjects at week 72 after discontinuation of treatment. Our definition of 2 weeks after seroconversion as the end of the period of acute HIV infection was arbitrary, and a more sustained treatment benefit may occur after this time point; however, we did not have adequate numbers of subjects to effectively compare smaller differences in the timing of HAART initiation.
Although the results of the present study support initiation of HAART during acute HIV infection, we believe that caution is indicated in making this recommendation on a clinical basis, because of both the limitations of observational studies and the lack of randomized, controlled studies; the lack of statistical significance of our unadjusted analyses after 24 weeks; and the need for longer-term assessment to further evaluate the durability of potential benefits. We believe that, on the basis of current evidence, patients need to be informed of the possible benefits of treatment, especially when HIV infection is diagnosed during the acute stage; they also need to be informed of the remaining uncertainties about the reliability of current data and the durability of benefit. Balanced against the possible benefits of treatment are the financial costs, the potential toxicities of treatment, and the risk of developing drug resistance.
The possible benefits of treatment during acute HIV infection may add to the rationale for diagnosis of acute HIV infection, on the basis of the public health benefits of early diagnosis in potentially preventing additional HIV transmission [20]. The typical symptoms of primary HIV infection and the role of HIV RNA testing in diagnosis have been well described elsewhere [21-25].
Our results suggest more limited benefit from treatment initiated longer than a few weeks after seroconversion. A randomized, controlled trial of treatment during this stage would be useful to better define the potential benefits of treatment during this period. Additional studies are also needed to determine whether the benefit that we observed in association with treatment initiated during acute HIV infection can be corroborated and to assess the optimal duration of HAART for persons with acute HIV infection.
|
|
| |
| |
|
|
|
