We investigated the relationships between changes in CD4 lymphocytes counts over 24 weeks after the initiation of therapy with indinavir at dosages of > or = 2.4 g/day (n = 15) in human immunodeficiency virus-positive patients and compared them to the baseline values. Starting CD4 count were linked to the time-weighted average CD4 cell count (return) through a nonlinear effect model. The diminution of destruction of CD4 cells after the initiation of indinavir therapy was estimated by fitting simultaneous differential equations to the data by using a linked lymph node (LN)-blood (BL) (two-compartment) system in which there is a constant rate of generation (R), first-order transfer rate constants (KLN-BL and KBL-LN) of compartment exchange, and first-order rate constants of CD4 destruction in the absence and presence of indinavir (KLN-OUT1 and KLN-OUT2). The half-life of CD4 lymphocytes was calculated from the rate constants by standard two-compartment methods. The CD4 lymphocyte counts at the start and return were linked in a sigmoid-Emax model were the maximal effect (Emax) was at 574.6 cells/microliters and 50% of the effect occurred at 157.1 cells/microliters (r2 = 0.94; P < 0.001). The mean +/- standard deviation (median) KLN-OUT2 was 0.574 +/- 0.202 (0.589), indicating that indinavir decrease the destruction of CD4 cells by circa 41 to 42%. The mean (median) CD4 half-life was 11.5 +/- 5.72 day (10.3 days). In multivariate analysis, KLN-OUT2 was significantly correlated with starting the CD4 cells count and the change in the CD4 cell count on therapy. The relationship between CD4 lymphocyte half-life and the starting CD4 lymphocyte count was hyperbolic, with a rapid increase in half-life as the CD4 count decreased. On the basis of the calculated half-life, the average production (destruction) of CD4 lymphocytes was approximately 3 x 10(9) cells/day, with an individual variation of 44-fold. These findings suggest that (i) the CD4 lymphocyte cell count at the start is significantly correlated to both the decrease in the destruction rate of CD4 cells and the degree of change in the CD4 lymphocytes on therapy, (ii) the lower the initial CD4 lymphocyte count, the higher the amount of CD4 lymphocyte turnover and the lower the ability of the immune system to increase absolute CD4 lymphocyte levels after viral suppression, consistent with a decreased regenerative capacity with progression of disease; and (iii) the increase in CD4 lymphocytes is likely secondary to the expansion of proliferating pool of cells since our determinations are based on 24 weeks of effect.