XIVth International AIDS Conference, Barcelona, Spain
Reported by David Margolis, MD, University of Texas Southwestern Medical Center
1. when & how to start therapy: ACTG 384, CLASS Study, CDC CD4 Study
2. Tenofovir (Viread) vd d4T, firstline therapy
3. T-20 Studies
4. Once A Day Studies: Kaletra, nevirapine+ddI+Tenofovir
5. Kaletra: in advanced HIV (<90 CD4s); immune reconstitution
6. Nevirapine and nevirapine Liver Toxicity
7. Lipodystrophy: FRAMS study & FRAM study
8. HGH for fat accumulation
9. Pre-mature heart disease
10. Treatment interruption studies
What, when, how to start it?
The late-breaker results of ACTG 384 came as something of a surprise [LbOr20A]. This was an open-label strategy study of nearly 1000 patients that compared several different sequential plans of antiviral therapy. In the first phase of the study subjects received the NRTI combinations of ZDV/3TC or D4T/DDI, and either NFV, EFV, or both NFV and EFV. After failure of initial therapy (either for toxicity, intolerance, or virologic failure) in phase 2 subjects received the nucleosides and the third drug they had not received in the initial round. Subjects that received both NFV and EFV initially reached their study endpoint upon initial failure of the quadruple regimen. The primary endpoint of the study was reached when subjects failed phase 2 of therapy. The mean CD4 count in the group was 278/μl and VL 4.9 logs, making the group quite relevant given current therapy guidelines.
The findings of the study were somewhat complex, but enlightening. About 40% of subjects had failed their 2nd regimen by week 144 (2 years). But of the 441 endpoints reached, more than half left the study for various reasons before failing phase 2 of therapy. Only 138 subjects failed their 2nd regimen virologically, and 40 failed due to toxicity. Many who failed according to study protocol may have simply tired of the studyıs demands and left to take HAART off-protocol. Nevertheless, this finding emphasizes the difficulty of maintaining therapy and viral suppression over 2 years. Yet the failure of only a minority of study subjects seemed driven by drug resistance.
Overall there was no significant advantage for any of the drug sequences for time-to-2nd-failure. However, when the planned analysis of secondary endpoints was considered, differences emerged.
KEY FINDINGS: Looking at either time to first virological failure or time to first regimen failure (for any reason), ZDV/3TC was superior to DDI/D4T when paired with EFV but not when paired with NFV, and EFV was superior to NFV when paired with ZDV/3TC but not when paired with DDI/D4T. It seemed that overall the most "successful" 3-drug regimen in this study, whether for reasons of superior potency, tolerability, ease of adherence, lesser toxicity, or some combination of these factors, was ZDV/3TC/EFV. The initial use of the 4-drug regimen delayed time to first failure in comparison to any of the three-drug regimens except ZDV/3TC/EFV.
While newer drugs and boosted PIs might now replace some of the therapeutic options used in ACTG 384, the study provides some important guidance. Simplicity and long-term tolerability of the entire regimen are likely to drive long-term outcomes as therapy improves.
Bartlett (Duke) presented the results of the CLASS study, an open-label, randomized study of 291 treatment naive patients [TuOrB1189] . Subjects received NRTI therapy with ABC/3TC, and one of either efavirenz, amprenavir boosted by ritonavir, or d4T. In-class substitutions were allowed. Treatment groups were reasonably matched, with baseline VL 4.9 logs/80,000 (42% >100K) and mean CD4 roughly 300. At 48 weeks by ITT (m=f), equal results were obtained for VL <400 (ca. 80% all arms), but there was an advantage for the NNRTI (ITT) arm if baseline VL was > 100K (85% NNRTI vs 72-75% for <400; 80% vs 53-54% for <50); and there was an advantage for the NNRTI arm when VL< 50 was considered (ca. 76% NNRTI vs. 59-62%). Overall however, time to first treatment failure or first virological failure, CD4 counts, and AEs were not significantly different between the arms. Similar to 384, this study suggests that there may be some overall advantage to initial therapy with 2 NRTIs and an NNRTI, but does not provide insight as to which patient-specific factors drive this observation. Future analysis of the comparative outcomes of second stage salvage therapy may provide interesting information about drug sequencing.
Brooks presented a study [TuOrB1141] from the CDC that examined the relationship between CD4 count at HAART initiation and the likelihood of durable virologic response. Failure was defined as ≥0.5 log10 increase from VL nadir after HAART initiation followed by a 2nd equally or more elevated VL or by a switch in ART. These 470 therapy-naive patients from 11 US cities started HAART after 1996. In a mutivariate analysis, failure of therapy occurred earlier in persons with a baseline CD4 < 350 cells/mL than ≥350 cells/mL. The estimated probability of failure at by 6 months was 10.7% CD4 count 0-199, half that (5.1%) if CD3 was 200-349, and very low (0.3%) for persons with ≥ 350 cells/mL. These results confirm those of Phillips et al (JAMA 2001), and add a strong tug in the pro-therapy direction in the ongoing when-to-start debate for patients in the gray zone between CD4 200-350/μl.
Tenofovir: debutante drug
As expected, the recently approved nucleotide RTI tenofovir celebrated a major coming-out at the Barcelona meeting. Gilead 903, a phase III, randomized, double-blind, placebo-controlled study in 600 naive patients compared d4T to TDF when given with 3TC and EFV. Subjects were well-matched and representative (24% female, 36% nonwhite, VL 4.9 log, mean CD4 279/μl). At 48 weeks the arms performed identically (RNA <400c/mL by ITT 87%, HIV RNA <50 c/mL by ITT 82%). Declines in viral load, increases in CD4 count, AEs, and dropouts were all equivalent. The only striking difference noted in this trial was a significant increase in serum triglycerides from baseline in the D4T arm; in the TDF arm these did not change. Although this study provides rationale for the first-line use of TDF, there was no evidence of superior potency of TDF as might have been expected. Perhaps even better results will be obtained when TDF is tested in a QD format without the burden of BID placebos. An important strategic question to be answered will be the effect of the use of initial TDF on NRTI therapy options in subsequent therapy, should salvage be needed.
The activity of TDF in patients with NRTI experience was illustrated in the Gilead 907 study (WeOrB1266). This was a phase III, randomized, double blind, placebo controlled study in which in ART-experienced patients on a variety of stable 3- or 4-drug ART regimens added TDF as a single agent to their regimens. Patients could only enroll if they had low-level viremia with VL 400-10,000 cpm. 550 subjects with a mean VL of 2340 copies/ml enrolled. TDF 300 mg or placebo, (2:1 TDF:placebo) was added for 24 weeks, subjects were then unblinded and placebo subjects could switch to TDF. At baseline mean CD4 was 447/μl, 58% were on PI, 48% on NNRTI, and 94% had at least one NRTI resistance mutation. VL dropped 0.8 logs 2 weeks after TDF was added, and this decline persisted for the 48 weeks of the study. Mean decline of viral load on placebo was -0.03 compared to -0.61 on TDF at week 24. Decline on TDF was -.57 at week 48. Placebo arm patients also had a -0.7 log viral load decline 24 weeks after switching to the active arm. 6% of subjects in both arms reported any adverse event, and grade III/IV adverse events were equal between arms. There was no difference in serum creatinine or phosphate rises between arms, as had been seen with the prior nucleotide NRTI, adefovir.
Data on TDF resistance from these trials was presented [ThOrB1390]. In brief, at wk24 similar HIV RNA were observed for patients without TAMs (-0.80 log, n=97) or with 1-2 TAMs (-0.66 log, n=88). For ≥3 TAMs with M41L or L210W, responses were diminished (-0.21 log, n=86) but superior to PLB (p=0.013); pts with ≥3 TAMs without M41L or L210W had -0.67 log responses (n=61, p<0.001). When examining responses to TDF as predicted by phenotypic testing, patients with >4-fold reductions in TDF susceptibility (n=19, 9%) showed diminished responses of -0.22 log. As detailed at the Seville Resistance meeting (see NATAP summaries) the presence of M41L or L210W with at least two other TAMS, or more than 4-fold resistance by phenotypic assay predicts a markedly diminished response to TDF.
TORO, TORO but no matador:
Two large phase III studies of the investigational fusion inhibitor T-20 were presented [LbOr19A, LbOr19B]. TORO-1 was performed in the US and TORO-2 in Europe. Both studies treated subjects with more than 3 months of HAART experience, and 3-drug class exposure and resistance, with salvage therapy consisting of optimized background therapy (OB) alone vs. OB and T-20. Both studies were quite similar in patient population and T-20 response. Subjects had an entry VL of ca. 5 logs, mean CD4 of 80-100, exposure to a mean of 12 drugs in the past, and therapy for an average of 7 years or more. By last-obervational-carried-forward, intent-to-treat, missing-equals-failure analysis OB resulted in 0.65-0.76 log drop in VL while OB and T-20 yielded a 1.43-1.70 decline in VL at week 24 of study. CD4 counts rose more than 50 cells/μl. AEs were frequent but equal in both arms, and injection site reactions limited therapy in only 35 of these motivated subjects. Although, as expected in this population, complete suppression was uncommon, more than twice as many (up to 20%) of subject who received T-20 achieved VL of < 5 copies at week 24. This study provided proof-of-concept that T-20 is a potent agent that can be tolerated by subjects with very advanced disease. While resistance is expected to develop when T-20 is used without other new or potent agents, further follow-up will reveal if reasonable clinical benefit can be conferred in "deep salvage" by fusion inhibitors. Access to T-20 is likely to increase later this year. An expanded access program is planned for Fall by Roche & Trimeris for 1200 patients worldwide of which 600 will be in the USA.
A handful once-a-day
In a small study of 38 ARV-naive patients, Feinberg [TuPeB4445] suggested that Kaletra could succeed as a once-a-day PI. LPV/r (Kaletra) 800/200mg QD was compared to 400/100mg BID with d4T/3TC BID. A reasonably mixed patient population with median baseline VL 4.7 log10 copies/mL (50,000) and CD4 cell count 264 cells/mm3 were studied for 72 weeks. Tolerability and AEs were comparable between the two arms. At Week 72, 83% and 82% of patients randomized to LPV/r QD and BID, respectively, had VL<50 copies/mL (on treatment) and 74% and 58%, respectively, had VL <50 copies/μL (intent-to-treat [NC=F], p=0.31). Although the pill burden (6 Kaletra capsules at once) may present a challenge, if this boosted PI is used in initial therapy this study provides proof-of-concept for once-a-day therapy.
A randomized, multicenter study of patients who had been on therapy for more than 9 months and achieved VL <80 cpm for at least 6 months, were switched to NVP + ddI + TDF (dosed together with breakfast) or continued their original regimen [LbPe9017]. 159 patients were randomized, and a planned interim analysis of the 65 (41%) who had reached 24 weeks of study presented. VL was still <80 cpm in 63/65, with no significant change in CD4. Triglycerides decreased slightly in the qd arm (mean decrease 61 mg/dl). Self-reported adherence was 94% in the qd arm and 92% for continued original therapy. At this early time in the study, despite the PK data suggesting that ddI levels are significantly increased by TDF, no obvious increase had yet been seen in ddI toxicity.
News from the field on Kaletra
Used BID, Kaletra was successful in a cohort of naive patients with advanced disease and high-level viremia [TuPeB4447] followed in routine clinical practices in Germany. 86 patients (84% male, 16% female; median age 41 years) with baseline median HIV-1 RNA 321,561 copies/μL and CD4 cells 90/μL were treated with LPV and two NRTIs. 69% (23/33) achieved HIV- RNA <400 copies/mL at month 3 and 87% at month 6 (13/15; on- treatment analysis). As expected increased triglycerides (from median 150mg/mL to 190mg/mL) and total cholesterol (from median 160mg/mL to 190mg/mL) was observed at month 6.
Alan Landay [TuPeB4439] presented 3 year follow-up on immune reconstitution in antiretroviral-naive patients treated with Kaletra. Baseline median CD4 count was 391 cells/μl. At week 156, 78/100 pts remained on study drug. At week 156 the median CD4 cell count had risen to 596 cells/μl. Of 17 patients with CD4 < 50 cells/μl, 14 were > 200 cells/μl, 10 > 350 cells/μl, 3 > 500 cell/μl at week 156 or their last observed value. Of 19 patients with 50-199 CD4s at baseline 32% (6) achieved >500 CD4 count. As we have seen before, considerable immune reconstitution is possible if viremia is durably suppressed.
Nevirapine fares well in Spanish studies
The intruigingly-named SENC trial (Spanish Efavirenz vs. Nevirapine Comparison) found no difference in the overall outcomes of patients [TuPeB4441]. NVP and D4T/ddI was compared with EFV and D4T/ddI in 67 antiretroviral naive HIV+ patients in this open-label trial. Baseline CD4s were >100 and plasma HIV-RNA >500 and < 5.0 logs. At 48 weeks, 23/31 (74%) in the EFV group and 23/36 (64%) in the NVP group had <50 HIV-RNA copies/μL (intent-to-treat analysis, noncompleter = failure). Adherence, CD4 count gains, and drop-outs for AEs were similar in the 2 arms. Of note there were few dropouts in the study overall, striking given the fact that all subjects received D4T/DDI. In ACTG 384, significantly more dropouts and AEs were attributed to D4T/DDI.
Another Spanish study compared NVP to NFV using as background NRTIs D4T/DDI in a population of similar characteristics, and reached similar conclusions [TuPeB1186]. Of 140 patients evaluable patients at 12 months, the proportions of patients with plasma HIV-1 RNA below 200 copies/mL were about 90% (on-treatment) and 70% (intent-to-treat) in both arms. The mean increase in CD4+ count at 12 months was similar in both groups. By 18 months 70% of the subjects had suffered AEs, and 41% had switched therapy (reportedly largely due to AEs). In the authorıs opinion, while this study demonstrates that NVP can perform well in some settings and is therefore an important option for therapy, the population studied was small and the large number of switches may have obscured the ability to detect differences in the regimens.
Complications of HIV and antiretroviral therapy
Nevirapine-related liver toxicity: some guilt by association?
Mayers of Boerhinger-Ingelheim presented clinically relevant data on nevirapine (NVP) hepatotoxicity. Mayers proposed that the "hepatotoxicity spectrum" was comprised of a few cases of true fulminant liver damage, more cases of symptomatic but reversible cases of drug-related liver toxicity, and many cases of asymptomatic elevations of liver enzyme blood levels, all of which occur in the background of concurrent viral hepatitis, alcohol use, and other drug liver toxicity. Gleaned from a clinical database of over 18,000 patients in controlled clinical trials, two independent case-selection rules were used to include all "potential" clinical hepatic events. All hepatic and biliary adverse-event reported terms, and all extra-hepatic or systemic reported terms consistent with liver disease (within ħ14 days of elevated ALT/AST ≥3X ULN) were collected from case report forms from these industry and NIAID-sponsored trials. All cases were cross-reviewed by internal and external experts blinded to treatment.
Of note, toxicity events across trials occurred more frequently in the NVP arm only during the initial 6 weeks of treatment. Rates were comparable among trials and similar to active control arms not receiving NVP after 6 weeks. Predictive risk factors were found to be baseline ALT/AST >2.5X ULN (RR = 4.3;P < .01), and HCV/HBV co-infection (RR = 5.7; P < .01). Gender was not a risk factor, while race and baseline CD4+ were weak and inconsistent risk factors. NVP-related events occurred at similar rates to other ARVs, and the most frequent drug-related hepatic events were asymptomatic elevations of ALT/AST. Of note, 46% of symptomatic hepatic events were associated with rash, and all except one occurred in the first 4-6 weeks.
Overall, symptomatic hepatic events in NVP clinical trials were rarely serious and improved on withdrawal of study drugs. Fulminant hepatic events were rare in clinical trials (0.1%). These findings supported the widely used clinical strategy of permanent discontinuation of NVP for any symptomatic hepatic event, interruption of NVP for asymptomatic ALT/AST >5X ULN and evaluation for new or underlying liver diseases, and reintroduction of NVP on a case-by-case basis with heightened clinical and laboratory vigilance based on clinical needs and judgment.
Little insight into lipodystrophy
Murphy presented the results of FRAMS II, a substudy of body shape changes and metabolic parameters from within the Atlantic study [WeOrB1306]. This well-known study compared 3TC, a NVP, and IDV in the background of ddI/d4T-based regimens. Baseline measurements for the substudy were taken at week 72 weeks of the parent Atlantic study, using a questionnaire, DEXA scans, CT scans, and multiple lab markers. 49% of substudy subjects reported lipoatrophy, and 25% reported fat accumulation. At week 144, visceral adipose fat was increased in the IDV group, although subcutaneous fat was unchanged. In the arms that received 3TC or NVP there was an increase in visceral fat, as the levels were low at baseline the change did not reach significance. This was of interest in view of the findings reported by Grunfeld, which seemed to suggest that increases in visceral fat were due to overall weight gain rather than a specific HIV-related syndrome. Overall there was no difference between the groups in DEXA measurements or body impedence plethysmography (another sensitive test of overall body fat mass).
In a surprise presentation of new data in an overview session, Carl Grunfeld compared study populations in the FRAM to those in the CARDIA Database. CARDIA is a prospective study of cariovascular disease, and the FRAMs are HIV+ patients (+/- ART) randomly selected from clinic databases. Both cohort were aged 33-45 yrs. Subjects were evaluated by self-report questionnaire and physical exam. Grunfeld reported a positive association between HIV infection and lipoatrophy, but no association between HIV infection and central adiposity or buffalo hump. As discussed in detail by David Alain Wohl on this website, these findings were unique and fly in the face of common perception. This author suspects that methodological issues, and weight gain not related to HIV drug toxicity may have driven this result.
Don Kotler presented preliminary findings from the HIV adipose redistribution syndrome (HARS) study [LbOr18]. Recombinant human growth hormone (HGH) was given to HIV+ subjects with lipodystrophy. Patients could enroll if they had an abnormal waist-to-hip ratio or waist circumfrance. Enrollees were treated with 12 weeks of placebo injection or rHGH (4mg) were given qd or qod. Patients with diabetes or receiving glucose-lowering therapy were excluded. The primary study endpoint was DEXA scan, an x-ray test that accurately measures fat mass. At 12 weeks the visceral adipose tissue measured cross-sectionally in the midabdomen was unchanged in the placebo group, had decreased 20 cm2 in the every other day group, and 30 cm2 in the daily group. However, there was no change in limb fat, at least as could be measured in this study at 12 weeks of therapy. Total cholesterol and non-HDL lipid (the "bad" cholesterol) fell, although those declines may not have been clinically significant. No severe adverse events were seen, although some subjects chose to stop rHGH therapy. However, fasting blood sugar increased and glucose tolerance decreased, consistent with the diabetogenic effect of HGH that has been observed in the past. This expensive therapy therefore appeared somewhat effective in treating visceral fat accumulation in a short study, although perhaps at the cost of another metabolic complication, diabetes.
Is angina looming on the HAART horizon?
Several studies of cardiovascular (CV) complications occurring in HIV+ individuals receiving HAART were presented. Although there has been significant concern that metabolic complications induced by HAART and/or HIV will lead to a higher incidence of CV complications, a large VA database analysis by Bozette last fall did not detect a change in the incidence of CV complications. However, studies presented in Barcelona suggested that somewhat increased rates may be found with longer periods of observation.
Barbarini [WeOrB1307] presented the findings of a multicenter study performed in Italy in which 1915 patients starting ART were approached for randomization of a PI-based or NNRTI based regimen. 67 subjects refused entry, and 297 were excluded due to a long list of reasons including CV comorbidities, cocaine use. The data and background were very rapidly presented, and we look forward to careful review of published findings. 1551 subjects were randomized, and complete data was available on 587 of 776 NNRTI-treated enrollees, and 621 of 775 PI-treated subjects. The groups were well-matched for CD4 and VL (ca. 350/μl and 5 log copies/ml). Of note, 87% were said to be "smokers ," although in Italy a non-smoker is said to be someone who smokes less than 15 cigarettes per day. If patients switched therapies, they were "censored" at that point (i.e. if they had been observed for 3 years, the data up to the switch was included, and was not after the switch or if they had not been observed for 3 years). By Kaplan-Meier analysis, CV events had occurred in 10% more of the PI than NNRTI group at 30 months (p < 0.01). However, 3-year data was yet available from only 128 of the 1551 subjects. The absolute number of event observed was quite small. In the PI arm there were 10 subjects with occlusion of 2 heart vessels found at catheterization, and 7 subjects with 3-vessel disease. In the NNRTI arm, 1 subject had 2-vessel disease. In the PI group there were 23 CV events (12 subjects with heart attacks, and 11 with angina), but only one heart attack and one case of angina in the NNRTI group.
The study suggested that with longer follow-up more CV events will be seen on PI-containing therapies. Given concerns about the methodology of this study, one would be well to await published results before drawing definite conclusions.
Along these lines, a study of cardiovascular mortality within the HOPS Study was presented by Holmberg [TuPeB4496]. A longitudinal cohort study including 5675 HIV+ patient from clinics across the US, the database now includes 21 heart attack (MI) endpoints. The incidence of MI appears to be increasing with a 3-yr lag time after the initiation of HAART. MI was associated with PI use, with an increased odds ratio of 4.92 after adjustment for the presence diabetes, smoking, hypertension, age, gender, and elevated lipids. No data was yet available on the risk of non-PI containing HAART.
At risk for anal cancer despite HAART
Palefsky [LbOr21] presented findings of importance to men living longer with HIV infection. Prior to HIV, the incidence of anal cancer in MSMs was 35 in 100,000. Palefsky reported that since 1998, this risk had more than doubled, and now exceeded the risk of cervical cancer in women. In a survey of more than 350 men in San Francisco, anal swabs and anoscopy found early neoplasia in 52% of the population. Unexpectedly, this risk was independent of CD4 count and increased by HAART. Although it was unclear if the increase was statistically significant, risk was clearly not significantly decreased. These findings served to re-emphasize the need to screen MSMs for anal neoplasia, as has been called for over many years. Although sensitive and simple to perform, anal Pap smears are different than cervical Pap smears, and in this authorıs experience the effective use of this test requires close collaboration with a pathologist trained in the technique.
Stopping and SCARTing
A trial aimed at preserving HIV specific CD4+ T-helper responses (Short course of antiretroviral therapy or SCART) added yet another catchy acronym to the annals HIV infection [TuOrB1185]. Fidler and colleagues from London and other centers across England treated 60 patients with ZDV/3TC/NVP or ZDV/3TC/ABC/EFV identified during primary infection. The primary endpoint of the study was HIV specific CD8 and CD4 T-cell responses as measured by the ability of these patientıs cells to make interferon gamma when exposed to fragments of HIV. It seemed that the intent of the study was to treat patients for 3 months, although some appeared to be treated for longer. Nevertheless, HAART was given for much shorter times than in prior acute infection studies. The average pVL more than 21 weeks off SCART was significantly lower than baseline (mean drop 0.5 log HIV RNA copies RNA/ml p=0.05). However, "baseline" VLs during acute infection may be higher than in chronic infection. Of 52 patients who received and then stopped SCART, 6 restarted therapy within 2 years for progression of disease, a proportion that does not seem lower than average. The percentages of cells that responded in the interferon gamma assay was not higher after receiving SCART. The authors could only conclude that SCART appeared safe and that a larger, randomized and controlled study would be needed.
Gallant reported the observation of 101 patients in the Hopkins clinic [ThOrB1439] who chose to initiate a treatment interruption (TI) after response to HAART. These patients planned to restarted HAART based on undefined CD4 and HIV RNA criteria. Pre-HAART VL levels correlated well with rebound levels 6 weeks after TI (r2 = 0.59). At the latest timepoint for which data was available 68 of the 101 subjects remained off HAART after TI. These 68 subjects had been off therapy for a mean of 68 weeks (mean VL 43,000), while those subjects that had restarted therapy had done so after a mean of 34 weeks (mean VL 177,000). CD4 counts had dropped more than 400 cells to a mean of 316/μl in those subjects who restarted HAART, while CD4 counts had dropped only about 100 cells to a mean of 508/μl in those who remained off HAART.
Reinitiation (RI) of therapy was predicted by a lower nadir CD4 count prior to first HAART. Those with nadirs < 200 cells 7-fold more likely to reinitiate, and those with CD4 counts 200-350/μl had a 4-fold more likely, those with nadirs > 350 cells/μl. The rate of CD4 decline and the pre-ART VL did not predict RI. 2 subject had thrombocytopenia after TI, and 3 had a retroviral rebound syndrome (RI in 2 of these). No Ois were seen; thus far subjects who underwent RI had resuppression of HIV RNA. Overall, Gallant stated that two-thirds of subjects remained off therapy at 74 weeks, and that the best prediction of continued TI was nadir CD4, followed by whether the subjects pre-HAART CD4 and VL met current guidelines (CD4< 350, VL > 50,000 or not). Gallant summed it all up in a manner worthy of Yogi Berra: "If they didnıt need therapy then, they probably donıt need it now." These findings are reassuring, and worthy of confirmation in other, broader cohorts. The discovery of surrogate markers which might predict successful and durable TI would be useful.
Treatment during primary HIV infection (PHI) has been hypothesized to preserve antiviral immune response. Miro [ThOrB1437] reported findings in 12 subjects treated during PHI. Subjects initiated HAART within 90 days of infection (mean 9 weeks), and had mean CD4 counts of 590/μl and mean viral loads were ca. 300,000. HAART was given for at least 1 year (mean 2 yrs.), with successful suppression to < 20 copies/ml. STIs were then started, stopping therapy for 2 months, restarting for 2 months, stopping for 2-4 months, then restarting for 2 months. Genotypic resistance was not detected by standard methods after interruptions, and increases in CD8 and CD4 antiviral responses were seen in 2/3rds of subjects after the 3rd interruption. However, only 4 of 12 subjects contained replication with VL < 5000/ml after the 3rd interruption, and only 3 of 12 had viral rebound after the 3rd interruption that was ≥ 0.5 logs lower than seen after earlier interruptions. These findings were in contrast to the Rosenberg study in which 7 of 8 patient controlled viremia (although those subject were treated earlier, prior to seroconversion), but similar to the Garcia (4/10 "controllers") and Hirsch (24/133 "controllers") studies presented earlier this year. Overall, most studies do not find benefit of treatment in PHI followed by STI, and in ultrasensitive assays minor resistant species can be detected following interruptions (as presented at the Seville meeting).
IL-2 has long been studied as a possible adjunctive therapy to aid immune reconstitution and boost CD4 cell counts. Multinational clinical endpoint studies to examine the clinical benefit of adjunctive IL-2 are underway, but will not be completed for several years. Valdez [ThOrA1481] presented preliminary results from ACTG 5046s, a substudy of immune response nested within an ACTG IL-2 study, that were somewhat disappointing. Subjects were vaccinated with the Remmune inactivated HIV vaccine, tetanus, and hepatitis A and B vaccines after many months of HAART +/- IL2. Through a quirk of substudy recruitment, the substudy subjects that received HAART and IL-2 had higher percentages of activated CD4 and CD8 cell, and higher VL at baseline. This may have affected the study outcome. Nevertheless, cell mediate immune responses were not as great in the group that received HAART/IL-2 compared with HAART alone, and fewer IL-2 subjects developed antibodies to Hepatitis A after vaccination.