The results of the Strategic Timing of Antiretroviral Treatment (START) trial demonstrated that among HIV-positive participants with a CD4+ cell count more than 500 cells/μl initiating antiretroviral therapy (ART) immediately resulted in lower rates of a composite outcome of AIDS events, serious non-AIDS events, or death compared with waiting until an individual's CD4+ cell count dropped to 350 cells/μl . The relative clinical benefit of immediate compared with deferred ART was consistent for a number of baseline subgroups . Immediate minus deferred ART absolute risk differences for the primary composite outcome were greater for older participants, those with a lower CD4+ : CD8+ ratio, and for those with a higher baseline viral load .
One subgroup of interest which has not been considered in START until now is duration of HIV infection at the time of randomization. Although the study is not powered to investigate clinical outcomes across subgroups of infection duration, CD4+ levels can be evaluated. Previous studies have demonstrated that the viral reservoir is diminished for individuals treated during the acute phase of HIV  and early after seroconversion . Also, in observational cohort studies, greater CD4+ cell count recovery has been reported for participants who initiate ART closer in time to acquiring HIV infection [5–9].
In this article, we take advantage of the randomized design of START to assess whether differences in follow-up CD4+ and CD8+ cell counts between the immediate and deferred ART groups vary by estimated duration of infection. To carry out this subgroup analysis, stored baseline specimens were tested using a multiassay algorithm (MAA) which had been previously used to estimate population level incidence of HIV [10,11]. This assay was used in combination with self-reported data on date of infection and date of diagnosis to identify 373 participants in START who were recently infected with HIV (≤6 months before randomization) . We compared differences between the immediate and deferred ART group in CD4+ and CD8+ cell counts and the CD4+ : CD8+ ratio for these recently infected participants with those who had been infected for a longer period of time.
The START trial randomized 4684 HIV+ adults with CD4+ cell counts more than 500 cells/μl to immediate or deferred initiation of ART. Individuals in the deferred ART group were to initiate ART when their CD4+ cell count declined to 350 cells/μl or they developed AIDS. Participants had to have two CD4+ cell counts more than 500 cells/μl at least 2 weeks apart within the 60 days prior to randomization . All participants provided informed consent to participate in the trial, and additional consent for storage of specimens for future research was also obtained.
At study entry, in addition to demographic characteristics, CD4+ cell count, CD8+ cell count, and HIV RNA level were measured, and the date the participant was first diagnosed with HIV as documented in the participant's record was recorded, and, if known, the date the participant believed they were first infected with HIV was also recorded. As previously described, stored baseline plasma samples were tested using a MAA, which incorporates both serologic and nonserologic markers of recent infection to identify participants who were recently infected at the time of randomization .
For the current study, we classified participants in the START trial into three groups based on their estimated duration of infection at baseline: Group 1 was considered infected for 6 months or less based on the MAA or self-reported date of infection; Group 2 was considered infected for 6–24 months and consisted of those participants for whom a) the MAA failed to confirm more recent infection, and the infection date was unknown, and the diagnosis date was 6 months or less before randomization, or b) their HIV diagnosis date was 6–24 months before randomization; and, Group 3 was considered infected for 24 months or longer if their HIV diagnosis date was 24 months or longer before randomization.
Baseline characteristics were compared between groups using Chi-square tests for proportions and Wilcoxon rank sum tests for medians. Longitudinal regression with repeated measures analyses were used to compare CD4+ cell count, CD8+ cell count, and CD4+ : CD8+ ratio levels during follow-up between the immediate and deferred ART groups among Groups 1–3, adjusting for baseline levels. Treatment differences (immediate minus deferred) for the three duration of infection groups were compared with 2 degree of freedom (df) interaction tests. In our previous report that compared self-reported date of infection with the MAA for START participants, we showed that participants recently infected were younger and had higher baseline viral load levels compared with those who were not recently infected . Based on this and the finding that absolute risk differences in the START composite primary endpoint varied by age and viral load , for CD4+ cell count and the CD4+ : CD8+ ratio we also carried out analyses to assess three-way interactions between treatment group, duration of infection group, and age and viral load. Two-way interactions between treatment group and age and viral load were assessed to determine treatment differences irrespective of duration of infection.
Selected analyses were also carried out within START treatment groups. In the deferred group, the rate of decline to CD4+ less than 350 cells/μl (censored at last CD4+ measurement) or AIDS (threshold for initiating ART; censored at last known alive date) was compared among the three groups using proportional hazards regression analysis, with and without censoring for ART initiation. Hazard ratios adjusted for age, sex, race, geographic location, CD4+ cell count, and HIV RNA level at entry are cited with 95% confidence intervals (CIs). In the immediate ART group, the three groups were compared for virologic suppression (HIV RNA ≤ 200 copies/ml) at 1, 4, and 8 months following randomization using separate adjusted logistic regression models for each time point with baseline covariates corresponding to age, sex, race, geographic location, CD4+ cell count, and HIV RNA level. For this analysis, we restricted the cohort to immediate ART group participants who initiated ART within 1 week of randomization.
Statistical analyses were performed using SAS, version 9.3 (SAS Institute, Cary, North Carolina, USA). All P values cited are two-sided and not adjusted for multiple comparisons.
Group 1 includes 373 participants (167 participants who self-reported being infected in the 6 months prior to enrollment and 206 participants who did not know the duration of their infection, but who were diagnosed with HIV in the 6 months before randomization and were also classified as recently infected by the MAA) . Groups 2 and 3 include 2634 and 1605 participants, respectively. Seventy-two participants who would have been classifiable in either Group 1 or 2 did not have a baseline sample available for MAA testing and are excluded from this analysis.
Baseline characteristics according to estimated duration of infection at randomization
At entry, individuals in the recently infected group (Group 1) were younger than those in Groups 2 and 3 (31 vs. 34 vs. 40 years, P < 0.001) and more likely to be MSM (75 vs. 59 vs. 44%, P < 0.001). Group 1 also had a higher median baseline HIV RNA and a smaller percentage of participants with viral load less than 3000 copies/ml than the other two groups (27 199 vs. 13 553 vs. 9744 copies/ml; P < 0.001) and (11.5 vs. 23.6 vs. 29.8%; P < 0.001), a moderately higher baseline CD4+ cell count (660 vs. 654 vs. 644 cells/μl; P = 0.007), a moderately lower baseline CD8+ cell count (1018 vs. 1031 vs. 1070 cells/μl; P = 0.003), and a higher CD4+ : CD8+ ratio (0.70 vs. 0.67 vs. 0.63; P < 0.001) (Table 1). Among participants who reported their presumed date of HIV infection (overall 40%), the median time from infection to randomization for Groups 1, 2, and 3 were 4.8, 14.3, and 58.0 months, respectively.
Treatment differences in use of antiretroviral therapy
In Group 1, ART was used for 95% of follow-up time in the immediate ART group and 33% of follow-up time in the deferred ART group; for Group 2, these percentages were 95 and 30; and for Group 3, the percentages were 93 and 26.
Treatment differences in CD4+ and CD8+ cell counts and the CD4+ : CD8+ ratio
As previously reported, over an average follow-up period of 3 years, the mean CD4+ cell count was 194 cells/μl higher in the immediate compared with the deferred ART group . The average follow-up CD4+ difference between the immediate and deferred ART groups varied significantly according to estimated duration of HIV infection (231 vs. 202 and 172 cells/μl; P < 0.001 for interaction). The mean CD4+ cell counts increased over follow-up in the immediate ART group for all three subgroups, with the greatest increase in the recent infection group (188, 171, and 151, respectively, for Groups 1, 2, and 3; P = 0.002) (Fig. 1). For the deferred ART group, the average CD4+ cell count declined over follow-up by 42, 31, and 21 cells/μl for Groups 1, 2, and 3, respectively (P = 0.18) (Fig. 1).
The mean CD4+ : CD8+ ratio difference (immediate − deferred) over follow-up also varied according to estimated duration of infection at entry (0.52 vs. 0.40 vs. 0.34; P < 0.001 for interaction) (Fig. 2). The mean CD4+ : CD8+ ratio increased over follow-up for all three subgroups in the immediate ART group, and the increase varied according to duration of infection at entry (P < 0.001). For the CD4+ : CD8+ ratio in the deferred ART group, there was little change from baseline and the differences among the three subgroups did not vary significantly (P = 0.57).
The mean CD8+ difference (immediate − deferred) did not vary significantly over follow-up according to estimated duration of infection (P = 0.33 for interaction) (Supplemental Fig. 1, http://links.lww.com/QAD/B469).
Treatment differences in CD4+ and CD4+ : CD8+ ratio by HIV RNA level and age
The mean follow-up CD4+ cell count and the CD4+ : CD8+ ratio were greater in the immediate vs. deferred ART groups among those with higher baseline HIV RNA levels (Fig. 3a and c; HIV-RNA × treatment interactions P ≤ 0.001 for both CD4+ cell count and CD4+ : CD8+ ratio). These trends did not differ by duration of infection group (three-way interaction P = 0.39 and 0.34). For example, treatment differences in CD4+ cell count between Groups 1 and 3 were 46, 31, and 45 cells/μl for the three baseline viral load strata. For CD4+ : CD8+ ratio these Group 1 vs. Group 3 mean differences were 0.09, 0.20, and 0.20 for the three viral load strata. Similarly, differences in mean CD4+ cell count and CD4+ : CD8+ ratio differences between treatment groups for the three duration of infection groups did not vary by age (Fig. 3b and d; three-way interactions P = 0.71 and 0.19).
Rate of CD4+ decline to less than 350 cells/μl or AIDS in the deferred antiretroviral therapy group
In the deferred ART group, the rate of CD4+ decline to less than 350 cells/μl or development of AIDS was greater among those recently infected compared with those who were not recently infected (Fig. 4a). Rates per 100 person years were 15.6, 12.7, and 10.5 for Groups 1, 2, and 3, respectively (P = 0.002). Hazard ratios adjusted for age, sex, race, geographic location, CD4+ cell count, and HIV RNA level at entry for Group 1 vs. 3 and for Group 2 vs. 3 were 1.38 (95% CI: 1.03–1.84) and 1.26 (95% CI: 1.06–1.49), respectively (P value for corresponding 2 df likelihood ratio test <0.001). A stronger trend (P < 0.001) was observed when follow-up was censored at ART initiation (Fig. 4b). Adjusted hazard ratios for Group 1 vs. 3 and for Group 2 vs. 3 were 1.52 (95% CI: 1.14–2.05) and 1.36 (95% CI: 1.14–1.63). The median (25th, 75th percentile) CD4+ cell count at ART initiation for Groups 1, 2, and 3 were 386 (300, 509), 399 (316, 551), and 440 (335, 595) cells/μl, respectively.
Rates of CD4+ decline to less than 350 cells/μl or development of AIDS within each infection group separately by sex and within subgroups defined by race/ethnicity and region of enrollment are presented in Supplemental Table 1, http://links.lww.com/QAD/B469. Predictors of CD4+ decline to less than 350 cells/μl or development of AIDS within each of the infection groups are presented in Supplemental Table 2, http://links.lww.com/QAD/B469.
Virologic suppression in the immediate antiretroviral therapy group
Supplemental Table 3, http://links.lww.com/QAD/B469 compares the percentage with virologic suppression at 1, 4, and 8 months in the immediate ART group by duration of infection group. At 1 month fewer participants in Group 1 had a viral load 200 copies/ml or less than in Groups 2 and 3 (P ≤ 0.001 for unadjusted 2 df test for difference among subgroups). By 4 and 8 months the percentage with viral load 200 copies/ml or less was similar for the three subgroups defined by duration of HIV infection (P ≥ 0.10 for at both visits for unadjusted 2 df comparisons). When considering adjusted comparisons, the difference in suppression rates at month 1 was no longer significant (P = 0.07) after including baseline viral load in the model.
The primary purpose of this study was to determine whether differences in follow-up levels of CD4+ cell count between the immediate and deferred ART groups in START varied by estimated duration of HIV infection at the time of randomization. We found that the treatment difference in follow-up CD4+ cell count was greatest for those infected for less than 6 months at the time of randomization.
Since an early study in 16 individuals of the initiation of combination ART during primary HIV infection reported that CD4+ cell count increased to levels similar to those of HIV-negative controls and also appeared to prevent permanent loss of important subsets of CD4+ T lymphocytes , several other larger observational studies have reported that an enhanced CD4+ cell count recovery was evident among participants who initiated ART closer to the time of seroconversion [5,7–9]. Findings for the immediate ART group in START are consistent with the findings of these studies but have the advantage of having an appropriate control group with which to compare the CD4+ cell count changes during follow-up resulting from immediate ART. Further follow-up of this cohort is necessary to determine the implications of the CD4+ cell count differences between treatment groups we observed after an average of 3 years. Follow-up of START participants is ongoing and planned to continue through 2021.
The variable treatment difference in CD4+ cell count observed according to duration of infection is also generally consistent with a randomized trial of short-course ART in primary HIV infection. In the Short Pulse Anti-Retroviral Therapy at Seroconversion trial, 366 participants were randomized (1 : 1 : 1) to ART for 48 weeks, ART for 12 weeks, or no ART and followed for a median of 4.2 years for a composite primary endpoint of CD4+ less than 350 cells/μl or initiation of ART. Median CD4+ cell count at entry was 559 cells/μl and the median time from seroconversion at randomization was estimated at 12 weeks. A trend for a greater effect of 48 weeks of ART (followed by ART discontinuation) compared with no ART on the primary endpoint was found for those randomized sooner after seroconversion. Hazard ratios (48 weeks of ART vs. no ART) for CD4+ less than 350 cells/μl or initiation of ART were 0.41 for those randomized within 4 weeks of seroconversion, 0.65 for those randomized with 12 weeks and 1.28 for those randomized within 12 weeks (P = 0.09 for trend) .
In our study, follow-up CD8+ differences between treatment groups did not differ by duration of infection, and as a consequence when the CD4+ : CD8+ ratio was considered, a pattern similar to that seen for CD4+ cell count was observed. In a small study comparing CD4+ and CD8+ cell count recovery following ART initiation in individuals with primary and chronic HIV infection, there was decline in CD8+ cell count among those with primary infection similar to what we observed but an increase in those with chronic infection after 1 month. However, by 12 months, CD8+ cell count had declined in both groups. Like our study, those with primary infection had much greater CD4+ cell count increases than those with chronic infection .
Among the immediate ART group, the recently infected subgroup of participants tended to have a lower rate of virologic suppression at month 1 compared with those not recently infected. By month 4, however, rates of suppression were higher and comparable with the rates in the other two groups. This slower time to reach suppression is not surprising given the higher baseline viral load among those recently infected participants. Among participants in the deferred ART group, we observed a higher rate of CD4+ cell count decline to less than 350 cells/μl or progression to AIDS within the recently infected subgroup compared with those infected longer. This difference in rates could also be attributed to the higher viral load in the recent infection group. The study eligibility criteria may have also played a part. The study required two CD4+ cell counts above 500 cells/μl within 60 days before randomization and these criteria coupled with the likelihood that many sites may have had participants under care who could readily be screened and consented likely led to the enrollment of individuals with more stable CD4+ levels and lower viral load. This in turn led to slower rates of CD4+ cell count decline following randomization among many participants in START, particularly those who have been infected longer.
There were some limitations to our study. As noted earlier, we cannot determine whether the greater CD4+ cell count differences for the recent infection group as compared with the other two groups led to differences in clinical outcomes. This may not even be possible to establish with the planned follow-up through 2021. In addition, there is likely some misclassification of participants in the groups, particularly Groups 2 and 3 which are largely based on the date of diagnosis and not the date of infection or the MAA. Although we can be fairly confident that the three subgroups differ substantially in duration of infection at randomization, the magnitude of the difference is uncertain because of misclassification and the large percentage of participants not knowing the date of infection. Finally, it is possible that duration of infection is a marker for another factor. Although treatment is randomized in START, duration of infection is not. Those with recent infection differed from those with later infection in many respects.
In summary, the subgroup of recently infected individuals from the START trial constitutes the largest randomized study of immediate vs. deferred ART for those with recent infection. If differences in CD4+ cell count and the CD4+ : CD8+ ratio between the immediate and deferred groups for those with recent infection result in greater clinical benefit than those infected longer at randomization, then the overall benefit of immediate ART in START may be underestimated because the great majority of participants in START had been infected for longer than 6 months, with many longer than 2 years likely as a result of screening and eligibility criteria.
We wish to thank the participants and clinical staff of the START trial. See Initiation of Antiretroviral Therapy in Early Asymptomatic HIV Infection. New England Journal of Medicine 2015;373(9):795–807 for the complete list of START investigators.
Author contributions: Contributions of the authors are study conception (J.D.N., O.L.), statistical analysis (S.S.), drafting of the article (J.D.N., O.L., S.S.), MAA laboratory work (K.E.S., O.L.), revisions to article (P.d.l.T., G.T., R.D., A.N.P., J.A.M.).
The study was supported in part by the National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health (NIH, grants UM1-AI068641 and UM1-AI120197), National Institutes of Health Clinical Center, National Cancer Institute, National Heart, Lung, and Blood Institute, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institute of Mental Health, National Institute of Neurological Disorders and Stroke, National Institute of Arthritis and Musculoskeletal and Skin Diseases, Agence Nationale de Recherches sur le SIDA et les Hépatites Virales (France), National Health and Medical Research Council (Australia), National Research Foundation (Denmark), Bundes ministerium für Bildung und Forschung (Germany), European AIDS Treatment Network, Medical Research Council (United Kingdom), National Institute for Health Research, National Health Service (United Kingdom), and University of Minnesota. Antiretroviral drugs were donated to the central drug repository by AbbVie, Bristol-Myers Squibb, Gilead Sciences, GlaxoSmithKline/ViiV Healthcare, Janssen Scientific Affairs, and Merck. Additional support was provided by the HIV Prevention Trials Network sponsored by NIAID, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institute of Drug Abuse, National Institute of Mental Health, and the Office of AIDS Research, of the NIH DHHS (UM1 AI068613), and NIAID (R01 AI095068).
Conflicts of interest
There are no conflicts of interest.
Data presented previously in part as a poster. Schlusser K, Sharma S, Metcalf JA, Pinto AN, Draenert R, de la Torre P, et al. Early treatment is likely more important than thought. 24th Conference on Retroviruses and Opportunistic Infections (CROI), 13–16 February 2017. [Abstract 472].
1. Lundgren JD, Babiker AG, Gordin F, Emery S, Grund B, Sharma S, et al. Initiation of antiretroviral therapy in early asymptomatic HIV infection
. N Engl J Med
2. Molina JM, Grund B, Gordin F, Williams I, Schechter M, Losso M, et al. Which HIV-infected adults with high CD4 T-cell counts benefit most from immediate initiation of antiretroviral therapy? A post-hoc subgroup analysis of the START trial
. Lancet HIV
3. Buzon MJ, Martin-Gayo E, Pereyra F, Ouyang Z, Sun H, Li JZ, et al. Long-term antiretroviral treatment initiated at primary HIV-1 infection affects the size, composition, and decay kinetics of the reservoir of HIV-1-infected CD4 T cells
. J Virol
4. Ostrowski M, Benko E, Yue FY, Kim CJ, Huibner S, Lee T, et al. Intensifying antiretroviral therapy with raltegravir and maraviroc during early human immunodeficiency virus (HIV) infection does not accelerate HIV reservoir reduction
. Open Forum Infects Dis
5. Le T, Wright EJ, Smith DM, He W, Catano G, Okulicz JF, et al. Enhanced CD4+ recovery with earlier HIV-1 antiretroviral therapy
. N Engl J Med
6. Okulicz JF, Le TD, Agan BK, Camargo JF, Landrum ML, Wright E, et al. Influence of the timing of antiretroviral therapy on the potential for normalization of immune status in human immunodeficiency virus 1-infected individuals
. JAMA Intern Med
7. Babiker A, Darbyshire J, Pezzotti P, Porter K, Prins M, Sabin C, et al. Short-term CD4 cell response after highly active antiretroviral therapy initiated at different times from seroconversion in 1,500 seroconverters
. J Acquir Immune Defic Syndr
8. Ding Y, Duan S, Wu Z, Ye R, Yang Y, Yao S, et al. Timing of antiretroviral therapy initiation after diagnosis of recent human immunodeficiency virus infection and CD4+ T-cell recovery
. Clin Microbiol Infect
9. Stirrup OT, Copas AJ, Phillips AN, Gill MJ, Geskus RB, Touloumi G, et al. Predictors of CD4 cell recovery following initiation of antiretroviral therapy among HIV-1 positive patients with well estimated dates of seroconversion
. HIV Med
10. Konikoff J, Brookmeyer R, Longosz AF, Cousins MM, Celum C, Buchbinder SP, et al. Performance of a limiting-antigen avidity enzyme immunoassay for cross-sectional estimation of HIV incidence in the United States
. PLoS One
11. Huynh D, Laeyendecker O, Brookmeyer R. A serial risk score approach to disease classification that accounts for accuracy and costs
12. Schlusser KE, Sharma S, de la Torre P, Tambussi G, Draenert R, Pinto AN, et al. Comparison of self-report to biomarkers of recent HIV infection: findings from the START trial
. AIDS Behav
13. Zaunders JJ, Cunningham PH, Kelleher AD, Kaufmann GR, Jaramillo AB, Wright R, et al. Potent antiretroviral therapy of primary human immunodeficiency virus type 1 (HIV-1) infection: partial normalization of T lymphocyte subsets and limited reduction in HIV-1 DNA despite clearance of plasma viremia
. J Infect Dis
14. The SPARTAC Trial Investigators. Short-course antiretroviral therapy in primary HIV infection
. N Engl J Med
15. Kaufmann GR, Zaunders JJ, Cunningham P, Kelleher AD, Grey P, Smith D, et al. Rapid restoration of CD4 T cell subsets in subjects receiving antiretroviral therapy during primary HIV-1 infection