| |
HAART Treatment During Acute HIV-Infection Not Recommended
|
| |
| |
"Is antiretroviral treatment of primary HIV infection clinically justified on the basis of current evidence?"
AIDS: Volume 18(5) 26 March 2004 pp 709-718Smith, Don Ea; Walker, Bruce Db; Cooper, David Aa; Rosenberg, Eric Sb; Kaldor, John Ma
From the aNational Centre for HIV Epidemiology and Clinical Research, University of New South Wales, Sydney, Australia and the bPartners AIDS Research Center and Howard Hughes Medical Institute, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.
The study authors conclude: ".......Where does this leave the clinician who has identified a patient as acutely infected with HIV? In some regards, the question of when to treat in PHI is similar to those patients identified with established infection; however, at the start of the disease course there is the possibility of proportionally larger benefits, making this an important question to answer. Based on the currently published data, there is no clear evidence that patients with access to antiretroviral therapy have any greater clinical benefit if therapy is introduced immediately during or prior to their seroconversion illness. ...... nor are there comparative data to suggest that short-term use of HAART during PHI can alter future disease progression........ there is currently no evidence from these studies to suggest that therapy during PHI results in a reduction in clinical progression compared with use of effective therapy in later disease.......Despite numerous studies, none has been appropriately powered and controlled to answer this question and we conclude that a randomized controlled trial of early short-term HAART versus deferred HAART therapy is both ethically justifiable and necessary..."
FIRST, here is David Margolis, MD, report he wrote for NATAP's CROI coverage on Bruce Walker's presentation at the Retrovirus Conference in Feb 2004.
Treating Acute HIV Infection and Treatment Interruptions to Boost Immune Control: an experiment comes to an end
Walker (Abstract 24) presented long-term follow-up of a select cohort of patients in whom symptomatic early HIV infection had been detected, and ART given followed by supervised treatment interruptions. Initial reports generated great excitement in the field as it was suggested that this strategy resulted in improved immune control of viremia. Initially, 7 of 8 patients maintained stable CD4 counts and low viral loads without therapy. However, 14 subjects have now been enrolled and treated in this protocol for up to 5.3 years with sobering results. Subjects have undergone initial ART and subsequently subjected to up to 4 supervised treatment interruptions. Re-treatment was required if viral load remained above 5000 copies for longer than 3 consecutive weeks, or was in excess of 50,000 copies on any single occasion. 11 of 14 subjects were able to maintain control of viremia by these criteria for at least 90 days, but only 6 of 14 at 1 year, and 3 of 14 at 3 years. Over the course of the study, a gradual increase in viremia and decline in CD4 cell counts was observed in most individuals, with a significant decline of CD4 cell count in 1 of the remaining 3 patients with controlled viremia. This pattern was not influenced by treatment interruptions. The total magnitude and breadth of HIV-specific CD8 T-cell responses increased during the first, second, and third supervised treatment interruptions, as the intact immune system was exposed to repeated to HIV antigen. However, the extent of these increases in anti-HIV immune response did not correlate with subsequent viral control. Other factors that may correlate with disease progression, such as HLA type, baseline viral load, chemokine receptor alleles, extent of HIV antibody production, did not correlate with viral control in this protocol. These findings were echoed in European studies (Abstracts 396 and 397), which also did not demonstrate a beneficial effect on HIV-specific immune control or level of viremia after early therapy or treatment interruption. These data indicate that despite initial control of viremia, durable immune control in persons following treated acute infection occurs infrequently. Unfortunately, it appears that the immune system, even when protected by ART, does not learn how to better control HIV on its repeated encounters with the virus. These data highlight the challenges to current efforts to develop an AIDS vaccine designed to retard disease progression rather than prevent infection.
FULL TEXT OF AIDS ARTICLE
Introduction
When antiretroviral therapy first became available, it was hypothesized that the modest clinical benefits seen in patients with advanced disease would be more pronounced in acutely infected individuals. Use of highly active antiretroviral therapy (HAART) in chronically infected individuals has now been shown to be effective in suppressing viral replication and preventing immunological decline, as well as in significantly influencing clinical progression events. As the duration and severity of the acute retroviral syndrome and the level of viral replication 6-12 months following resolution of the acute infection (viral set-point) are both strong predictors of long-term disease progression rates, initiation of HAART at this stage may influence disease progression. Indeed, although recent guidelines have suggested delaying treatment in persons with chronic infection, the same guidelines suggest that persons identified in acute infection may differ from other HIV-infected patients, and form a group where immediate treatment could be considered.
HAART during primary HIV infection (PHI) results in improvements in surrogate markers of disease progression similar to that seen in chronic infection and, importantly, seems to preserve HIV-specific immune responses. In addition, recovery of the various subpopulations of CD4 T cells occurs faster and appears to be more complete if antiretroviral therapy is commenced in early stages of HIV-1 infection. These considerations provided the theoretical rationale for the consideration of early initiation of HAART during primary HIV-1 infection.
The preservation of HIV-specific cellular immune responses may increase the likelihood that viral suppression will be maintained if treatment is stopped. In primate studies, early use of even suboptimal antiretroviral therapy can substantially delay disease progression compared with that with no therapy. Vaccination studies in primates also suggest that initial stimulation and subsequent antigen boosting facilitate expansion of functional cytotoxic T lymphocyte responses (with appropriate T cell help) to a level such that immunological control over on-going viral replication is achieved. Reports of some newly infected patients using this immunization strategy suggest that their progression rates may have been reduced.
Despite these promising reports, a number of key questions still face clinicians who have identified a patient with acute HIV infection and, here, the literature to some extent is confusing. Should patients be advised to commence therapy or should they wait until they have recovered physically and emotionally from PHI in order to determine how effectively, or ineffectively, their immune system controls the virus? Although it appears to be unreasonable to delay antiretroviral therapy until a severe and possibly irreversible immunodeficiency has occurred, it is not known whether HIV-1 progression can be altered by an immediate and time-limited therapeutic intervention. Will initiation of therapy during PHI yield greater benefits than initiating it later in the disease course? This review summarizes the work to date evaluating the clinical benefits of initiating antiretroviral therapy during primary HIV infection.
Methods
A Medline search was undertaken to identify all published peer-reviewed human studies of the clinical effect of antiretroviral therapy used during primary HIV infection. The abstracts obtained in this search were used to determine the suitability of the reported study for inclusion in the review. The definition of primary HIV infection was allowed to include patients with symptomatic seroconversion illness and incomplete antibody responses to HIV, as well as asymptomatic subjects with a positive antibody test who had a previously negative test within the preceding 6-12 months. These groups are often referred to as acute and early HIV infection, respectively.
Key search words were HIV, therapy/treatment and acute/primary. Studies were considered acceptable for inclusion in this review if they described the patient population as acutely or newly infected, listed details of treatment regimens used and described responses in terms of circulating HIV RNA levels, CD4 cell numbers or HIV-related clinical progression events.Three study designs were reviewed: those describing randomized clinical trials, those describing a treated cohort and a contemporaneously identified or historical comparator group and open label cohorts describing treated PHI subjects with no comparator group.
Virological responses could be recorded as either a log10 copies/ml decline in plasma virus or the proportion achieving an undetectable viral load (by the assay reported as used). Time taken to achieve an undetectable viral load was also considered as an end-point if available. Moreover, it was required that studies reported outcomes at least 6 months after the diagnosis of acute infection.
The key research questions addressed were:
Can differences in efficacy be distinguished amongst various treatments used in PHI patients?
Does treatment during PHI result in a significant improvement in virological and immunological markers compared with treatment initiated later, during chronic HIV infection?
Do long-term treatment outcomes improve if therapy is initiated during PHI?
Does transient therapy during PHI alter the viral set-point or disease progression?
Results
Thirty-one papers were identified that met the review criteria; including four that described randomized interventions. These four randomized trials plus six non-randomized comparative studies and 17 uncontrolled studies addressed the differential efficacy of therapies during PHI (question 1). Three non-randomized comparative studies addressed the relative efficacy of therapy during PHI compared with therapy in chronic infection (question 2). Two non-randomized studies described the long-term effects of initiating therapy during PHI (question 3) and two of the randomized trials plus three uncontrolled studies addressed the efficacy of transient therapy during PHI (question 4).
Can differences in efficacy be distinguished amongst various treatments used in PHI patients?
The effect of transient zidovudine monotherapy in acute infection has been addressed in two randomized, placebo-controlled studies, with viral load reductions from baselines of 1.55 and 1.91 log10 copies/ml reported in drug recipients compared with 0.85 and 1.85 log10 copies/ml declines in placebo recipients, respectively. These viral load reductions did not achieve statistical significance in either study. In addition to the lack of clear benefit for monotherapy, three open-label studies involving dual therapies also showed variable reductions in viral load from baseline. Zidovudine plus didanosine led to a 1.3 log10 copies/ml reduction (not dissimilar to the previously described untreated patients) in one group of 41 patients yet a 4.63 log10 copies/ml reduction in a smaller group of five patients with a reported median viral load of 4.42 log10 copies/ml prior to therapy. Treatment with zidovudine and an experimental non-nucleoside reverse transcriptase inhibitor (L-697661) led to 50% of patients having undetectable viral load (< 200 copies/ml) at 18 months.
Triple therapy regimens have been associated with a wide range of suppressive effects: from only 27% of 15 subjects achieving < 20 copies/ml, from a baseline of 5.4 (± 0.2) log10 copies/ml, with one triple nucleoside regimen, to all of 10 subjects achieving < 50 copies/ml from a mean pretreatment level of 4.9 (± 0.8) log10 copies/ml on a protease inhibitor- containing regimen. Two other studies using the same triple nucleoside combination achieved complete suppression to < 20 or < 500 copies/ml in all patients studied.
In studies reporting results with quadruple therapy, undetectability of viral RNA in plasma appears to have been achieved more consistently: with 80, 86 and 100% of subjects reaching < 50, < 50 and < 200 copies/ml, respectively, although none of these studies had comparator groups that were untreated or treated with triple combination therapy.
The time taken for subjects to achieve an undetectable viral load result was additionally described in 13 papers, with the median time to undetectable levels varying from 9.5 days to 180 days. Most of these studies did not have frequent sampling points after the first month on therapy. Therefore, medians exceeding 1 month can generally only be estimated imprecisely and generally occurred between the visits at 3 and 6 months. In addition, some studies reported efficacy only for adherent patients rather than for all subjects initiated on therapy.
CD4 cell numbers are reported to normalize through the use of therapy compared with those in subjects who received no therapy, with an improvement from 654 to 763 x 106 cell/l versus a decline from 576 to 459 x 106 cell/l reported in one study. In the two randomized trials involving zidovudine monotherapy for 6 months, the change in CD4 cell count from baseline was +173 x 106 cell/l (+6 x 106 cell/l in placebo group) after 15 months in one study, compared with -13 x 106 cell/l (-283 x 106 cell/l in the placebo group) reported by Niu and colleagues. The open-label studies of combination antiretroviral therapy reported increases ranging from an additional 73 x 106 cell/l to more than 500 x 106 cell/l. In two recent studies involving combination therapy with antiretroviral drugs plus immunomodulatory drugs, greater increases in CD4 cell numbers were observed. The addition of ciclosporin resulted in an increase of approximately 850 x 106 cell/l compared with approximately 275 x 106 cell/l increase in control subjects treated with combination antiretroviral drugs alone. The use of interleukin-2 with antiretroviral drugs resulted in an increase of 1460 x 106 cell/l compared with an increase of 132 x 106 cell/l with antiretroviral drugs alone.
In summary, these studies indicate that even monotherapy regimens result in significantly greater CD4 cell recovery compared with no therapy and suggest that continuous triple or quadruple therapy in acute HIV infection gives a high likelihood of short- to medium-term viral suppression, although the relative potency of different therapies has not been clearly defined.
Does treatment during primary HIV infection result in a significant improvement in virological and immunological markers compared with treatment initiated later, during chronic HIV infection?
No randomized trials address this question. Three observational studies compared the treatment effects in subjects treated during PHI and those treated for the first time in chronic infection. Yerly et al. used a more sensitive viral assay with a detection limit of only 3 copies/ml and reported greater virological suppression in PHI, with 75% of these patients achieving this response compared with only 8% of chronically infected patients. However, when using the commercial assay with a higher limit of detection (50 copies/ml), Kaufmann and colleagues did not find a significant difference in viral suppression (89% in PHI and 72% in chronic infection; P > 0.05) but did note that recovery of peripheral CD4 T lymphocytes appeared to be greater with intervention during PHI (288 x 106 cell/l increase in PHI versus 106 x 106 cell/l increase in chronic infection; P < 0.001), although the baseline CD4 cell counts were significantly different between the two groups (470 and 203 x 106 cell/l, respectively).
The third paper, although describing suppression effects in HAART-treated PHI subjects, reported only comparisons in proviral DNA levels between PHI and chronically infected patients treated with HAART therapy and, therefore, yielded no additional comparative data to address this question.
Together these data suggest that, during first antiretroviral therapy, recovery of total CD4 cell numbers, but not CD4 percentage, is possibly greater in PHI patients, who initiated therapy with higher baseline CD4 cell counts than in chronically infected patients. There is a suggestion that viral suppression is also greater with therapy during PHI, but with differences only detected when using a very sensitive research assay. However, as these studies were all non-randomized comparisons, there could be relevant differences between those who start early versus those who deferred therapy.
Do long-term treatment outcomes improve if therapy is initiated during primary HIV infection?
Although the first randomized trial of 6 months of zidovudine monotherapy versus placebo had suggested a significant reduction in progression to early symptomatic disease (one versus seven events), longer follow-up of this cohort failed to confirm any delay in progression to AIDS, suggesting that monotherapy initiated during PHI does not impact on future disease progression.
A more recent non-randomized study of two PHI cohorts compared disease progression rates in 47 subjects identified at PHI but not initiated on treatment at that point with 20 subjects treated with zidovudine, lamivudine and indinavir. Amongst the untreated patients, 13% had progressed to AIDS over 78 weeks compared with none of the treated subjects (P < 0.01). Significantly, fewer PHI-treated patients also developed either Centers for Disease Control and Prevention (CDC) category B- or C-defining infections (5 versus 21%; P = 0.02). Among the 47 untreated comparator subjects, seven did eventually receive some form of later-stage antiretroviral therapy, but this was triple therapy in only two subjects and is unlikely to have influenced the overall conclusions.
HIV disease progression was reported in seven of the studies describing the impact of HAART intervention. All reported no progression to symptomatic disease, with the exception of the more recently reported randomized trial of HAART alone or HAART with hydroxyurea, where one third of the subjects on the hydroxyurea arm progressed to CDC category B or C disease during the 12-month observation period.
Together these data support the notion that initiation of continuous combination antiretroviral therapy during PHI slows disease progression rates compared with progression in those taking no or suboptimal therapy. However, early treatment compared with no treatment is not of relevance to the research question, as there is no suggestion of therapy being withheld. Rather does HAART impart any greater effect if initiated early? Delay in progression to AIDS has been clearly proven for antiretroviral therapy initiated during chronic infection and, therefore, it remains unanswered whether there is any greater benefit in therapy initiated during PHI versus chronic infection.
Does transient therapy during primary HIV infection alter the viral set-point or disease progression?
Both randomized montherapy studies of short courses of zidouvudine failed to show any long-term benefits in delaying progression to AIDS. In one open-label study of eight subjects, which involved therapy for 383 to 1081 days, apparent control of viraemia (to < 5000 copies/ml) was noted in three subjects after their first treatment interruption and in another three subjects after their second interruption. After a median of 6.5 months off therapy, five of these eight maintained low levels viral replication, which was significantly different to the result expected from the MACS natural history cohort. A small French cohort study of nine subjects treated for 1 year suggested that viral loads remained low (500-12 395 copies/ml) in four, even at 18 months off therapy.
However, two more recent open-label studies of transient therapy suggested that treating PHI does not influence viral set-points. Markowitz and colleagues treated 16 patients for more prolonged periods (931-1822 days), with 11 of these patients also receiving therapeutic vaccinations before discontinuing therapy. They reported that viral loads stabilized < 5000 copies/ml in only four subjects and that the viral load distribution of their patients was identical to that seen in the MACS natural history cohort. Their patients however, also differed from those of Rosenberg et al. in that they had a longer period between start of symptoms and commencement of therapy. Similar results were reported from a UK group, where 37 patients received a short course of therapy (3 months or until < 50 copies/ml) before stopping. These patients were also more likely to have had their HIV infection longer (median of 40 days from first symptoms to diagnosis) and were treated for shorter periods (3 months or until viral loads fell to < 50 copies/ml). After 48 weeks off therapy, their mean viral load was 4.25 log10 copies/ml, which was comparable with a mean viral load of 4.3 log10 copies/ml in untreated seroconverters from the CASCADE cohort.
Taken together these studies suggest that, although individual patients may improve control of HIV, transient treatment of PHI does not significantly influence viral replication levels compared with cohorts of seroconverters from earlier natural history cohorts. However, both the natural history cohorts used in these comparisons contained more asymptomatic seroconvertors, who would be expected to have lower viral loads than symptomatic seroconverters, who made up the majority of the treated cohorts reviewed here.
Discussion
From this review, it can be determined that antiretroviral therapy initiated during PHI has significant short-term immunological and virological efficacy, compared with no therapy. Although only one paper has evaluated this question, not surprisingly, there is a suggestion of delayed clinical progression by remaining on PHI-initiated therapy compared with never being treated. More limited data suggest significantly greater virological suppression and greater immune recovery if therapy is initiated during PHI rather than in chronic infection. However, there is currently no evidence from these studies to suggest that therapy during PHI results in a reduction in clinical progression compared with use of effective therapy in later disease, nor are there comparative data to suggest that short-term use of HAART during PHI can alter future disease progression. A possible benefit has been suggested in a recent cohort analysis, which compared viral loads and CD4 cell counts in 134 untreated patients with those 24 weeks after treatment interruption in 61 patients transiently treated during PHI. Only after adjustment for baseline differences in viral load and date of infection was there a significant benefit found in viral load, but not CD4 counts, for patients who received early therapy.
While HAART seems to suppress viral replication effectively and abate the rapid immunological decline that typically is observed in acute infection, no long-term clinical benefit has yet been shown. This is in marked contrast to the clinical benefits of even suboptimal therapy seen in primates (possibly because the disease course in untreated monkeys is much more rapid and the differences become apparent much earlier).
Determining the clinical benefit of antiretroviral therapy during PHI from the published literature is limited by a number of factors. First, the only randomized trials occur at both ends of the HAART treatment era, with two comparing suboptimal zidovudine monotherapy with placebo, and the most recent two comparing the addition of immune modulators to triple therapy. There has not been a randomized trial comparing effective therapy initially with the same therapy that has been deferred. Secondly, the types of patient described as PHI subjects differ across studies. Some studies describe only individuals initiated on therapy during their seroconversion illness, while others include subjects 3-6 months after seroconversion. The viral loads at the time of initiation of therapy can give an indication as to how acute these patients are; most studies have mean RNA viral loads of far less than 100 000 copies/ml when these data are provided, which is substantially different from mean viral loads > 10 x 106 copies/ml reported in some cohorts of acute infection. In many ways, there is little justification for grouping responses in patients treated prior to seroconversion (who have very high levels of circulating virus, narrowly focused cellular immune responses, incomplete antibody responses, but better preserved HIV-specific T helper responses), with patients who are known to have seroconvertered within a 6-12 month window period (who are less likely to be symptomatic, have lower viral loads, depleted cellular immune responses but greater breadth and magnitude of cellular and humoral immune responses). The latter group are also less likely to have had a severe seroconversion illness and would thus be a better prognostic group than those identified by their symptomatic seroconversion. Ideally, these groups should be studied separately, as they are likely to have differing AIDS progression rates and therapy would be expected to achieve different changes in immune responses.
In assessing the value of commencing therapy during acute illness, an intention-to-treat analysis (where patients are analysed according to their original treatment assignment) is the most appropriate. However, some studies present data only on patients remaining adherent to therapy. This restricts the results to the maximal therapeutic effect seen rather than its impact across a group of patients.
As the follow-up of subjects is usually limited, the effect of more recent and effective therapies cannot be assessed in terms of progression to AIDS. Only one paper attempts to assess this, by using an earlier untreated cohort as the control group. Unfortunately, this control group included patients treated with suboptimal therapies, making the comparison one of initial HAART versus no (or inadequate) therapy, rather than the ideal comparison of immediate treatment of PHI versus initiating treatment in chronically infected subjects. If these identified but untreated individuals continue to remain under clinical care, it is assumed that they would also be able to access similar, or newer, therapies by deferring treatment.
Despite the higher levels of circulating virus noted during acute infection, rapid reduction in circulating plasma HIV levels is frequently described, with a viral decay rate the same as seen during treatment of established infection [43]. From the studies reported, it seems that treatment-induced viral suppression in PHI is similar to, if not greater than, that during chronic infection. However, it is worth noting that viral load declines of approximately 1 log10 copies/ml or greater have been noted following induction of cytotoxic T lymphocyte responses during the acute retroviral syndrome, and that the suppressive effects noted with therapy may not be related solely to the potency of the medications.
Based on the known prognostic value of a reduction in plasma HIV RNA, treatment of acute infection should delay disease progression while patients remain on therapy. However, whether this will result in these individuals ending up in a better viral load category should they discontinue therapy in the future is not clear. Both extremes have been described in individual case reports, with either a rapid return to very high levels or a plateau at low levels. Small cohort studies have suggested viral rebound to levels below that expected in untreated subjects, with a viral load < 12 395 copies/ml in five of nine subjects [31], and < 500 copies/ml in five of eight subjects. Two of the studies reviewed here do not support the notion of altering the viral set-point through early therapeutic intervention. Both studies noted similar viral loads off therapy to that seen in historical cohorts of untreated patients. Differences in treatment length do not appear to be an important factor in differentiating these outcomes, as patients in the most favourable outcome studies were treated for either 1 year or 1-3 years compared with 3 months to 3 years in the less-favourable studies. Similarly, all studies included both acute and early subjects, although there seems to have been more acutely infected patients in the study reporting the best outcomes.
In terms of virus-specific immune function, there are now multiple studies indicating an augmentation of individual patient's HIV-specific T helper cell responses following early treatment of acute infection. In contrast, at least one study indicates that cytotoxic T lymphocyte responses are blunted by early therapy. Unfortunately, there are accumulating data suggesting that this immunological advantage noted during temporary antiviral suppression may not persist, with evidence that the HIV-specific T helper responses generated are rapidly destroyed by on-going high-level viral replication and, from primate models, that the virus may subsequently mutate to escape the cytotoxic T lymphocyte control imposed by an intervention-induced immune response. Similarly, there is no evidence that commencing treatment at the early stages of chronic HIV infection produces weaker virological and immunological effects than achievable with HAART commenced during PHI, as recovery of HIV-specific responses still appears to occur if treatment is initiated once the acute stage has passed.
Arguments against the use of HAART during PHI have generally flowed from the belief that individuals will unnecessarily be on therapy for prolonged periods, thus increasing their exposure to long-term drug toxicities. This would certainly be the case if those identified during PHI were followed over the progression period to AIDS, an average of 10-11 years. However, it must be remembered that the subset of newly infected persons with more severe or prolonged PHI symptoms are more likely to be rapid progressors. Some of these individuals may have only a short period from their PHI illness before their immunological markers justify antiretroviral intervention. An increased risk of developing drug-resistant mutations is an additional concern with prolonged therapy. During PHI, viral loads are significantly higher than in chronic infection and may be more difficult to reduce with some regimens. It has been noted that evolution within the viral genome occurs in subjects on apparently suppressive therapy. While this does not appear to have allowed drug-resistance mutations to arise over the short to medium term, it is feasible that this could happen if patients are maintained on therapy for many years. The development of drug resistance is also a function of adherence. It is possible that compliance might be reduced in patients who have felt pressured into commencing therapy quickly, based on their PHI staging, as opposed to those who have had more opportunity to discuss compliance issues with their treating doctor.
Where does this leave the clinician who has identified a patient as acutely infected with HIV? In some regards, the question of when to treat in PHI is similar to those patients identified with established infection; however, at the start of the disease course there is the possibility of proportionally larger benefits, making this an important question to answer. Based on the currently published data, there is no clear evidence that patients with access to antiretroviral therapy have any greater clinical benefit if therapy is introduced immediately during or prior to their seroconversion illness. Despite numerous studies, none has been appropriately powered and controlled to answer this question and we conclude that a randomized controlled trial of early short-term HAART versus deferred HAART therapy is both ethically justifiable and necessary.
Acknowledgements
The National Centre in HIV Epidemiology and Clinical Research is affiliated with the University of New South Wales, Sydney. The National Center in HIV Epidemiology and Clinical Research and Partners AIDS Research Center received funding for this project from the Acute Infection, Early Disease Research Program (AIEDRP) of the National Institute of Health.The National Centre in HIV Epidemiology and Clinical Research is sponsored by the Commonwealth Department of Health and Ageing.
|
|
| |
| |
|
|
|
