Summary: Two phases of CD4+ T-cell increases are seen soon after potent antiretroviral therapy (ART) is initiated. In this 72-week analysis of 101 subjects with sustained viral suppression, we estimate the inflection point between the 2 phases to be 10 weeks after treatment initiation. Higher pretreatment HIV-1 RNA levels were associated with steeper initial CD4+ T-cell increases, likely reflecting greater redistribution of cells from lymphoid tissue to the peripheral blood compartment.
From the *Center for Biostatistics in AIDS Research, Harvard School of Public Health, Boston, MA; †University of California at San Diego, San Diego, CA; ‡Case Western Reserve University, Cleveland, OH; and §Beth Israel Deaconess Medical Center, Boston, MA.
Received for publication December 20, 2005; accepted July 31, 2006.
Supported by the AIDS Clinical Trials Group; Stanford University Virology Support Laboratory; and grants AI-38855, AI-38858, AI-25879 and AI-27659 from the National Institutes of Health. Funding also provided by Bristol-Myers Squibb, GlaxoSmithKline, and Agouron Pharmaceuticals.
Reprints: Ronald J. Bosch, PhD, Center for Biostatistics in AIDS Research, Harvard School of Public Health, 651 Huntington Avenue, Boston MA, 02115 (e-mail: email@example.com).
After the start of suppressive ART in an immunosuppressed HIV-infected person, the rise in CD4+ T-cell counts characteristically exhibits a biphasic pattern. A steep initial phase likely reflects redistribution of cells from lymphoid tissue over the first few months of treatment as suppression of viral replication reduces immune activation.1-4 Subsequently, and at a diminished slope, CD4 cell counts continue to rise at an average (median) rate of 35 to 75 cells/mm3 per year if viral suppression is maintained.5-9 Considering the second phase as beginning 12 weeks after starting ART, Hunt et al6 recently reported that greater second-phase increases were associated with younger age, lower pretreatment CD4 cell count, and higher pretreatment viral load, whereas only pretreatment viral load predicted CD4 changes from week 0 to 12. The precise inflection between the first and second phase increases is not clear, because other studies have used inflections between 3 and 12 weeks after initiation of ART in their analyses.1,10-12 In this report, we examine the slopes of CD4 cell count increases over 72 weeks in 101 subjects with frequent CD4 cell count evaluations and successful virologic responses to ART to estimate more precisely the time point soon after ART initiation that best delineates this change in the slope of CD4 cell count measurements. This contrasts with a recent investigation of much longer term (>2 years) responses to ART13 and the question of whether CD4 cell count increases reach a plateau after multiple years of treatment.
Data for this analysis are drawn from the AIDS Clinical Trials Group (ACTG) 364 study, in which 196 nucleoside analogue reverse transcriptase inhibitor (NRTI)-experienced subjects with screening HIV-1 RNA levels >500 copies/mL were randomized to potent ART with 2 NRTIs plus nelfinavir, efavirenz, or nelfinavir and efavirenz.14 Subjects were followed up to 144 weeks.15 CD4 cell counts were obtained after consensus methodology, and plasma HIV-1 RNA levels were quantified by reverse transcription polymerase chain reaction (RT-PCR; Amplicor HIV-1 Monitor; Roche Diagnostic Systems, Branchburg, NJ) at pre-entry; entry; weeks 2 (CD4 cell count only), 4, 8, 12, 16, and 24; and, subsequently, every 8 weeks. Baseline was the average of pre-entry and entry values, on the log10 scale for HIV-1 RNA level.14 In addition, an earlier CD4 cell measurement (on average, 10 months earlier; highly correlated with baseline CD4 cell count, r = 0.81), while subjects were taking their same prestudy NRTI regimen,14 was evaluated as a predictor. This analysis is based on 101 subjects with virologic control through 72 weeks of ART, defined as HIV-1 RNA level <500 copies/mL at week 16 and no confirmed HIV-1 RNA level ≥500 copies/mL from weeks 16 through 72. Virologic failure (HIV-1 RNA level ≥500 copies/mL) was the main reason why the other ACTG 364 subjects were not included in this analysis. Follow-up through week 72 was chosen to have 1 year of follow-up after the initial redistribution phase; requiring longer term viral suppression further reduced the number of evaluable subjects. Selected subjects each contributed 10 or more CD4 cell count time points; the average of the pre-entry and entry measurements was considered as a single week 0 time point.
We analyzed absolute CD4 cell counts using a mixed-effects model, which incorporates individual variability (random effects) and explanatory parameters (fixed effects). Analyses first assessed whether the slope of the CD4 cell count increase changed after 8 and 12 weeks of ART, respectively,3 using a piecewise linear model. Random effects were used for each of the 3 model parameters (intercept, initial slope, and change in slope). There was significant evidence of an inflection in the slope of the CD4 cell count increase (P < 0.001 at week 8 or 12), so we proceeded to estimate the best change point by maximum likelihood. Profile likelihood methods were used to construct a 95% confidence interval for the change point16 and to test whether the change point differed between subgroups. With respect to the time point of the change in CD4 slope, similar estimates were obtained when restricting to CD4 cell measurements through week 52 or from analyses of square root-transformed CD4 cell counts (analyses not shown). The 101 subjects were mostly male (90%), with 22, 35, and 44 subjects from the nelfinavir, efavirenz, and nelfinavir plus efavirenz randomized treatment arms, respectively.14 Subjects had a median of 5.5 years of prior NRTI experience. The median age was 41 years (25th-75th percentiles: 36-47 years), the median baseline CD4 cell count was 380 cells/mm3 (25th-75th percentiles: 278-498 cells/mm3), and the median baseline HIV-1 RNA level was 3500 copies/mL (25th-75th percentiles: 1200-17,200 copies/mL).
Confirming previous observations,6 we found a significant (P < 0.001) reduction in the slope of the initial CD4 increase, which was apparent at week 8 and week 12 (Fig. 1). The reduction in CD4 slope was 3- to 4-fold. Within this 2-slope family of models, the best overall estimate for the change point was week 10 (95% confidence interval: 6 to 16).
Baseline HIV-1 RNA level had a pronounced effect on the slope of the initial CD4 cell count rise (see Fig. 1). The initial slope was 2- to 3-fold steeper in subjects with baseline HIV-1 RNA levels greater than 10,000 copies/mL (weeks 0-12: P < 0.001), although the slopes of subsequent CD4 cell count increases were similar (weeks 12-72: P = 0.21). There was no evidence that the time of the change point was different for subjects with baseline HIV-1 RNA levels greater than versus less than 10,000 copies/mL (P = 0.27). After adjusting for baseline HIV-1 RNA, neither prior CD4 cell count, age, nor randomized treatment arm significantly predicted CD4 slopes during weeks 0 through 12 or weeks 12 through 72 (P ≥ 0.30). The CD4 slope during weeks 0 through 12 also did not predict the CD4 slope during weeks 12 through 72 (P = 0.41).
Estimating this early CD4 change point using an antiretroviral-naive population with a broad range of pretreatment HIV-1 RNA levels and CD4 cell counts would be valuable to generalize our results. A limitation of our findings is that the analysis population had prior NRTI experience, with relatively low HIV-1 RNA levels and high CD4 cell counts at the time of initiating potent ART. In addition, our sample size may have limited our ability to show other factors related to CD4 cell count increases.
Our results suggest that analyses should examine CD4 cell count changes from 12 weeks6,12 or later after the initiation of potent ART so as to remove the influence of the steep initial slope of the CD4 cell count increase reflecting redistribution from lymphoid tissue to the periphery. We used a formal analytic method to estimate this early change point in CD4 slopes, in contrast to previous descriptive approaches.1,11,12 A recent analysis13 of long-term CD4 cell count increases did not exclude the time period corresponding to this steep initial phase, which might have influenced that study's estimates of when CD4 cell measurements reach a plateau after years of ART. Analyses that separate the initial phase of CD4 cell count increase may be needed, because different factors seem to influence this steep initial phase.6,12 In this type of analysis, it is important to note that a negative association between baseline CD4 count (CD40) and CD4 responses such as the change from baseline (CD4t − CD40) may partly be induced by the fact that the baseline CD4 cell count (CD40) is used to calculate the change. We addressed this issue by modeling with a separate earlier CD4 cell measurement. Prior CD4 cell count was not an independent predictor of CD4 cell count increases in our population, likely because few subjects had CD4 counts <200 cells/mm3. In contrast, other populations had a substantial fraction with pre-ART CD4 counts <200 cells/mm3 and showed an inverse association between pre-ART CD4 cell count and long-term CD4 cell count increases.6,12 Our finding that baseline viral load strongly influenced the initial (but not the subsequent) phase of the CD4 cell response to ART is consistent with the suggestion that relatively more lymphocytes, including recent thymic emigrants, are sequestered in lymphoid tissue in persons with higher viral loads, leading to greater CD4 cell redistribution after viral suppression.1,2,4,17,18
Drugs were provided by Bristol-Myers Squibb, GlaxoSmithKline, and Agouron Pharmaceuticals. The authors thank Marlene Smurzynski for constructive comments on the manuscript and especially thank the ACTG 364 study participants. The study was approved by the institutional review boards of the participating institutions, and all patients provided written informed consent.
1. Pakker NG, Notermans DW, de Boer RJ, et al. Biphasic kinetics of peripheral blood T cells after triple combination therapy in HIV-1 infection: a composite of redistribution and proliferation. Nat Med
2. Bucy RP, Hockett RD, Derdeyn CA, et al. Initial increase in blood CD4(+) lymphocytes after HIV antiretroviral therapy reflects redistribution from lymphoid tissues. J Clin Invest
3. Lederman MM. Immune restoration and CD4+ T-cell function with antiretroviral therapies. AIDS
. 2001;15(Suppl 2):S11-S15.
4. Diaz M, Douek DC, Valdez H, et al. T cells containing T cell receptor excision circles are inversely related to HIV replication and are selectively and rapidly released into circulation with antiretroviral treatment. AIDS
5. Kaufmann GR, Bloch M, Finlayson R, et al. The extent of HIV-1-related immunodeficiency and age predict the long-term CD4 T lymphocyte response to potent antiretroviral therapy. AIDS
6. Hunt PW, Deeks SG, Rodriguez B, et al. Continued CD4 cell count increases in HIV-infected adults experiencing 4 years of viral suppression on antiretroviral therapy. AIDS
7. Gulick RM, Meibohm A, Havlir D, et al. Six-year follow-up of HIV-1-infected adults in a clinical trial of antiretroviral therapy with indinavir, zidovudine, and lamivudine. AIDS
8. Viard JP, Burgard M, Hubert JB, et al. Impact of 5 years of maximally successful highly active antiretroviral therapy on CD4 cell count and HIV-1 DNA level. AIDS
9. Smith K, Aga E, Bosch RJ, et al. Long-term changes in circulating CD4 T lymphocytes in virologically suppressed patients after 6 years of highly active antiretroviral therapy. AIDS
10. Powderly WG, Landay A, Lederman MM. Recovery of the immune system with antiretroviral therapy: the end of opportunism? JAMA
11. Lederman MM, McKinnis R, Kelleher D, et al. Cellular restoration in HIV infected persons treated with abacavir and a protease inhibitor: age inversely predicts naïve CD4 cell count increase. AIDS
12. Smith CJ, Sabin CA, Youle MS, et al. Factors influencing increases in CD4 cell counts of HIV-positive persons receiving long-term highly active antiretroviral therapy. J Infect Dis
13. Chu H, Gange SJ, Yamashita TE, et al. Individual variation in CD4 cell count trajectory among HIV-infected men and women on long-term highly active antiretroviral therapy: an application using Bayesian random change-point model. Am J Epidemiol
14. Albrecht MA, Bosch RJ, Hammer SM, et al. Nelfinavir, efavirenz, or both after the failure of nucleoside treatment of HIV infection. N Engl J Med
15. Katzenstein DA, Bosch RJ, Hellmann N, et al. Phenotypic susceptibility and virological outcome in nucleoside-experienced patients receiving three or four antiretroviral drugs. AIDS
16. Hall CB, Lipton RB, Sliwinski M, et al. A change point model for estimating the onset of cognitive decline in preclinical Alzheimer's disease. Stat Med
17. Mosier DE. HIV results in the frame. CD4+ cell turnover. Nature
18. Nokta MA, Li XD, Al-Harthi L, et al. Entrapment of recent thymic emigrants in lymphoid tissues from HIV-infected patients: association with HIV cellular viral load. AIDS