When Does the CD4+ Cell Count Plateau? Evidence from Subjects Treated with a Lopinavir/ritonavir-based Regimen for Up to 7 Years M King, BA Da Silva, F McMillan, K Niemi, GJ Hanna Abbott, Abbott Park, IL Suppressive combination antiretroviral therapy leads to a rapid increase in CD4+ T-cell counts during the first several weeks of therapy. Thereafter, the rate of CD4+ T-cell increase gradually decreases. For various reasons, it remains difficult to determine how long CD4+ T-cell counts can continue to increase: • Large intra-individual variability • A relatively small number of studies with long-term data available • Methodology for determining whether a plateau has been reached: graphical inspection is attractive for practical purposes but relatively subjective, and segmented linear models (analyzing week A to week B, week B to week C, etc., separately) are dependent on the choice of segments and may suffer from small sample sizes within each segment. Previous studies have generally suggested a CD4+ T-cell count plateau 3–4 years after treatment initiation. We present an analysis using data from a 7-year study in antiretroviral-naοve subjects (Study 720) receiving a lopinavir/ritonavir (LPV/r)-based antiretroviral regimen evaluating whether CD4+ T-cell counts plateau in subjects receiving potent, suppressive antiretroviral therapy. Background Methods 46th Annual ICAAC • 27-30 September, 2 0 0 6 • S a n F r a n c i s c o , CA Abstract H-1401 We analyzed CD4+ T-cell count changes from baseline in 100 antiretroviral-naοve subjects treated with LPV/r + stavudine + lamivudine for up to 7 years (360 weeks). Nonlinear regression was used to fit the following model to the CD4+ T-cell count data: a + bt + ct2 if t < tp CD4 = p if t ³ tp This segmented model fits a quadratic relationship up to time tp followed by a plateau (p) after time tp (i.e., tp is the time of plateau). The model fit requires the curve to be continuous (the sections meet at tp) and smooth (the first derivatives or slopes are the same at tp); therefore, the value of tp is determined by the model fit (i.e., the model provides least-squares estimates of plateau time and plateau level). Separate models were fit by stratum of baseline CD4+ T-cell count (<50 cells/mm3, 50–199 cells/mm3, ³200 cells/mm3). To assess the variability of the estimated time of plateau, bootstrap replications (n=4000) were performed for each stratum of baseline CD4+ T-cell count. The sample size of each replication was equal to the original sample size, with observations from the original data set selected randomly with replacement. The model-fitting procedure was repeated for each bootstrap replicate, and the distribution of resulting bootstrap estimates of plateau times was summarized by box-and-whisker plots. { Results • 100 subjects were included in the analysis. 17 had baseline CD4+ T-cell count <50 cells/mm3, 19 had baseline between 50 and 199 cells/mm3, and 64 had baseline CD4+ count ³200 cells/mm3. • Mean CD4+ T-cell count values over time are shown in Figure 1 by baseline CD4+ T-cell count category. Mean increases in CD4+ T-cell count from baseline to year 7 were 532, 476, and 495 cells/mm3 for subjects with baseline CD4+ T-cell count <50, 50–199, and ³200 cells/mm3. • Individual CD4+ T-cell count data across the entire study are displayed in Figure 2. At week 360 (or final visit, for subjects discontinuing prior to week 360), 91% of subjects had CD4+ T-cell counts above 350 cells/mm3, 75% above 500 cells/mm3, and the overall median was 740 cells/mm3. 0 200 400 600 800 1000 0 50 100 150 200 250 300 350 ³200 (n=64) 50–199 (n=19) <50 (n=17) Weeks Cells/mm3 Baseline CD4 Count (cells/mm3) Figure 1. Mean CD4+ Cell Count Values Over Time in Study 720 Weeks Cells/mm3 0 500 1000 1500 2000 2500 0 50 100 150 200 250 300 350 Baseline CD4 ³200 Baseline CD4 50–199 Baseline CD4 <50 Figure 2. Individual CD4+ Cell Count Values Over Time in Study 720 • Fitted models are displayed in Figure 3. For subjects with baseline CD4+ T-cell counts <50 cells/mm3, a plateau was observed beginning at week 371 at a level of 583 cells/mm3. For subjects with baseline CD4+ T-cell counts between 50–199 or ³200, corresponding values were week 336 at 633 cells/mm3 and week 312 at 957 cells/mm3, respectively. Weeks Cells/mm3 0 200 400 600 800 1000 0 50 100 150 200 250 300 350 Baseline CD4 ³200 Baseline CD4 50–199 Baseline CD4 <50 Estimated time of plateau Baseline CD4 (cells/mm3) ³200 50–199 <50 Plateau time (weeks) 312 336 371 Plateau level (cells/mm3) 957 633 583 Figure 3. Fitted CD4 Cell Count Model with Estimated Plateau Time Sensitivity analyses provided results similar to the primary analysis: • Analyzing CD4+ T-cell count change from baseline instead of CD4+ T-cell count value: estimated plateau times were 302, 348, and 373 weeks for the ³200, 50–199, and <50 groups, respectively. • Analyzing only those subjects who completed the study: estimated plateau times were 310, 369, and 360 weeks for the ³200, 50–199, and <50 groups, respectively. • Bootstrap replicates (see Methods) were used to assess variability of estimated plateau times. The distributions of estimated plateau times are displayed in Figure 4. The differences between the distributions of values for the <50 group compared to the ³200 group suggest that CD4+ T-cell counts continued increasing for a longer period of time in those who started with very low CD4+ T-cell counts. Plateau time (weeks) Baseline CD4+ cell count (cells/mm3) 0 100 200 300 400 500 600 <50 50–199 >200 75th percentile Median 25th percentile Figure 4. Box-and-Whisker Plots of Bootstrap Estimates of Plateau Times by Baseline CD4+ Count Determining the time at which CD4+ T-cell count is no longer increasing is challenging due to large within-subject variability and the gradually slowing rate of CD4+ T-cell count increase that is typically observed. Unlike methods of graphical inspection and segmented linear models, our model provides an explicit estimate of the time of CD4+ T-cell count plateau. We used it to evaluate CD4+ T-cell counts in subjects treated with a LPV/r-based regimen for up to 7 years in study 720. Subjects with very low baseline CD4+ T-cell counts (<50 cells/mm3) appeared to demonstrate increases throughout the entire 7-year treatment period, while those with higher baseline CD4+ T-counts reached a plateau between years 6 and 7. Notably, however, the group with lowest baseline CD4+ T-cell counts did not “catch up” to the other groups, and whether all subjects in that group will be able to achieve “normal” CD4+ T-cell counts, even with continued suppressive therapy, remains unclear. At the least, it would apparently require several more years of therapy. Results of this study compare favorably with others analyzing long-term CD4+ T-cell count trajectories (Table 1) in subjects who have started on their first antiretroviral regimen. Compared to the only study that identified continued CD4+ T-cell count increases beyond 5 years (ATHENA cohort, Gras, et al. 2006), the current study using a LPV/r-based regimen compared favorably: mean increases from baseline to year 7 for subjects with baseline CD4+ T-cell counts of <50, 50–199, and ³200 were 532, 476, and 495, respectively, for the current study, and 449, 429, and approximately 412, respectively, for the ATHENA cohort. The overall increases in the current study are consistent with reports that ritonavir-boosted PI regimens result in higher CD4+ T-cell increases compared to other regimens (Bartlett 2005) and that LPV/r-based regimens had larger annual increases compared to other boosted PI regimens (Mocroft 2006). Discussion 06ABTB752-2 Conclusions References • Based on a model that explicitly identifies a time of CD4+ T-cell count plateau, antiretroviral-naοve subjects treated with a LPV/r-based regimen had increases for at least 6–7 years. • Subjects with lower baseline CD4+ T-cell counts had longer durations of CD4+ T-cell count increases (plateau after 7 years for subjects with baseline CD4+ <50 cells/mm3, compared to a plateau at 6 years in subjects with baseline values of ³200 cells/mm3). • However, CD4+ T-cell count values did not “catch up” to those who started with higher CD4+ T-cell counts at baseline. • Results from this analysis of LPV/r-treated subjects compare favorably to previous studies that have suggested a plateau for CD4+ T-cell increases at earlier timepoints, lower final CD4+ T-cell counts, or both. Bartlett JA, et al. 12th CROI 2005, Abstract 586. Esteve A, et al. 13th CROI 2006, Abstract 611. Garcia F, et al. JAIDS 2004;36:702–13. Gras L, et al. 13th CROI 2006, Abstract 530. Kaufmann GR, et al. AIDS 2002;16:359–67. Kaufmann GR, et al. Arch Intern Med 2003;163:2187–95. Keruly JC, et al. 13th CROI 2006, Abstract 529. Le Moing V, et al. 13th CROI 2006, Abstract 609. Mocroft A, et al. AIDS 2006;20:1141–50. Tarwater PM, et al. JAIDS 2001;27:168–75. Acknowledgements We gratefully thank the subjects who enrolled in Study M97-720 and the M97-720 Study Group: M Albrecht, C Benson, J Eron, M Glesby, R Gulick, C Hicks, H Kessler, R Murphy, M Thompson, AC White, P Wolfe. © 2006 Abbott Follow-up time, Reference Study No. of subjects Summary of CD4+ T-cell count findings Tarwater 2001 MACS cohort 31/2 years Plateau at 2–31/2 years after treatment initiation n=314 Kaufmann 2002 Observational cohort 4 years Plateau during the third year of treatment (Australia) n=95 Kaufmann 2003 Swiss cohort 4 years Plateau after 2–3 years n=2235 Garcia 2004 Observational cohort ~5 years Plateau after 4–5 years (Barcelona) n=861 Keruly 2006 Observational cohort 5 years Plateau after 1–3 years (Baltimore) n=261 Le Moing 2006 APROCO cohort 5 years Plateau after 3 years n=870 Esteve 2006 PISCIS cohort 5+ years Plateau after 4 years except subjects with baseline n=1452 <100 cells/mm3 Gras 2006 ATHENA cohort 7 years Small increases between years 5–7 for subjects with n=529 baseline <500 cells/mm3 Current analysis Study 720 7 years Plateau between years 6 and 7 (baseline ³50 cells/mm3) n=100 or after year 7 (baseline <50 cells/mm3) Table 1. Comparison with Other Studies Evaluating Long-term CD4+ T-cell Trajectories in Subjects Receiving Combination Antiretroviral Therapy.