Decreased recovery of CD4 lymphocytes in older HIV-infected patients beginning highly active antiretroviral therapy

Among virological responders, the area under the curve of the CD4 count minus baseline (AUCMB) after 3, 9, 15 and 18 months of highly active antiretroviral therapy (HAART) was less in individuals 55 years or older (P< 0.05). Fewer older individuals achieved increases of 50, 100, or over 150 CD4 cells/μl. A random quadratic time course model estimated that the AUCMB decreased 35 cells/year for each 10 years of additional age during the first 12 months after HAART (P< 0.005). The rate of CD4 cell replenishment, the most important mediator of the clinical response to highly active antiretroviral therapy (HAART) [1,2] is related to the pretreatment absolute CD4 cell count, the baseline HIV load and suppression and subsequent suppression thereof, and hepatitis C virus (HCV) infection [3–5]. Age-related decreases in thymopoiesis, and slower CD4 cell recovery in older individuals after the receipt of cytotoxic chemotherapy suggest that CD4 reconstitution after HAART may be affected by increasing age [6,7]. As there are sparse data regarding the rate of CD4 cell recovery in individuals receiving HAART who are over the age of 55 years [5,8–10], we sought to compare the rates and magnitude of the repletion of CD4 cells in older versus younger HIV-infected patients who have shown a successful virological response to HAART. Data were retrospectively retrieved from routinely collected, comprehensive electronic medical records covering all individuals receiving care at the Veteran's Administration Greater Los Angeles Medical Center between March 1996 and December 1999. To be eligible for study, we required subjects to have one or more CD4 cell count and viral load measurement before HAART and two or more subsequent measurements. A baseline HIV viral load of 10 000 copies/ml or more (Roche Molecular Systems, Branchburg, NJ, USA) must have been followed by a viral load measurement of less than 1000 copies/ml or a more than 2 log decrease below baseline within 3 months of beginning HAART. Subjects were censored from the analyses if their viral load was subsequently 1000 copies/ml or greater or upon loss to follow-up. HAART was defined as the combination of one or more nucleoside reverse transcriptase inhibitor plus either a non-nucleoside reverse transcriptase inhibitor or one or more protease inhibitors. For descriptive analyses, we compared individuals 55 years of age or older with younger individuals. In multivariate analyses, age was treated as a continuous variable. We assessed both the raw change in CD4 lymphocyte count from baseline and the area under the curve of CD4 cell measurements minus the baseline measurement (AUCMB) [11]. Laboratory values obtained 45 days before or after 90 day intervals were assigned to the median time-point. Descriptive analyses were used to explore the raw CD4 cell count data and the AUCMB for CD4 cell counts over time. The Kruskal–Wallis test was used to examine differences in continuous measurements between groups. To account for the varying degrees of precision in the raw AUCMB calculations and to control for differences in baseline characteristics between age groups, time courses of CD4 cell count were modelled using mixed effects linear models. Of these, a random quadratic time course controlling for baseline viral load and CD4 cell count, age, ethnicity, and HCV status was the simplest model that gave an adequate fit. All computations were carried out using the MIXED procedure in the SAS system [12]. We identified 80 men and no women, 13 of whom were 55 years old or older, who met the inclusion criteria for this study. Overall, the mean, minimum and maximum ages were 47, 27 and 70, respectively. The baseline viral load (mean ± SD log10 HIV-RNA copies/ml) and CD4 cell count (lymphocytes/μl ± SD) were 5.0 ± 0.5 and 209 ± 155, respectively. The mean duration of follow-up was 502 days, and on average, six CD4 and viral load tests were performed. Five subjects were Hispanic, 34 were African-American, and 41 were Caucasian. Fifteen subjects were HCV seropositive, 59 were seronegative, and the status of six was unknown. We detected no significant inter-group differences by age in terms of demographics, baseline laboratory values, and HCV seropositivity or follow-up intensity. The change in raw CD4 cell count and AUCMB was less in men 55 years of age or older than in younger men; differences were statistically significant at months 3, 9, 15 and 18 for the CD4 cell AUCMB (P < 0.05 for all comparisons). Significant differences were found in the proportion of patients who achieved absolute increases of 50, 100, and over 150 CD4 cells/μl after 3 and 9 months of HAART (Fig. 1). On average, after one year of HAART, the CD4 cell count increased by 100 cells/Ul for younger patients versus 50 cells/μl for older patients.Fig. 1.: Proportional distribution of absolute CD4 cell count increases over baseline after beginning highly active antiretroviral therapy. The left and right panels show results in men less than 55 and over 55 years of age, respectively. The P values (Kruskal–Wallis test) for the differences in the distributions of CD4 cell count increases at 3, 6, 9 and 12 months were 0.013, 0.057, 0.009 and 0.177, respectively. HAART, Highly active antiretroviral therapy. ▪ < 50 CD4 cells; ▒ 50–100 CD4 cells; ▓ 101–150 CD4 cells; ░ > 150 CD4 cells.The random quadratic time course model estimated that for each 10 years of additional chronological age, the AUCMB, and thus CD4 cell recovery, was 35 cells fewer during the first 12 months after HAART (P < 0.005). This conclusion was consistent for a variety of alternative models that we explored (e.g. categorical treatment of age, linear trajectories only, and controlling for other baseline values). These data demonstrate that despite excellent virological responses during the 24 months after the initiation of HAART, the rate of CD4 cell increase and the proportion of individuals with an increase of 50, 100, or 150 cells/μl or more were less in men 55 years of age or older. This difference persisted after controlling for the effect of baseline HIV viral load, HCV seropositivity, and patient ethnicity. For every 10 additional years of age there was a replenishment of 35 fewer CD4 cells/μl per year, as measured by the AUCMB. Age-related decreases of thymic-naive CD4 cell production is a particularly appealing explanation for these results [7,13,14]. Others have also found a negative association between age and CD4 cell count recovery for individuals who do [5,10] and do not sustain a robust virological response to HAART [8,9]. In particular, our data substantiate earlier reports and more clearly evaluate the magnitude of the age-related decreases in CD4 cell recovery. Ours is an ethnically diverse patient population and our outcome persisted for more than one year. Studies before and since HAART [15–17] found that increased age is associated with a more rapid loss of CD4 cells after primary infection and faster clinical disease progression. We did not adjust for the effects of the duration of HIV infection, the use of other antiretroviral therapies before HAART, or other co-morbid conditions. Nonetheless, age-related differences in CD4 cell recovery could have implications for considerations regarding the timing of antiretroviral therapy in older HIV-infected individuals. These considerations will become increasingly relevant as the HIV-infected patient population ages as a result of the life prolonging effects of HAART. Matthew Bidwell Goetza W. John Boscardinb Dorothy Wileyc Salam Alkasspoolesa