Abstract OBJECTIVE: To modelize the rate of CD4 cell count decline and its determinants after cessation of combination antiretroviral therapy (cART) started during primary HIV infection (PHI) and compare it with never-treated patients. METHODS: Kinetics of CD4 counts were analyzed on the square root scale by using a mixed-effects model in 170 patients who received cART during PHI from the Primary Infection (PRIMO) cohort and 123 never-treated patients from the Seroconverters (SEROCO) cohort. RESULTS: CD4 cell count fell rapidly during the first 5 months and more slowly thereafter. The timing of treatment initiation had no influence on the rate of CD4 cell decline. In contrast, a larger increase in CD4 cell counts during cART was associated with a steeper decline and a larger loss of CD4 cells after treatment interruption. The mean CD4 cell loss 3 years postinterruption was 383 cells per microliter. In the SEROCO cohort, the CD4 T-cell decline was less steep (3-year CD4 loss 239 cells/microL). As a result, the mean CD4 cell counts were similar (416 cells/microL) 3 years after cART interruption (PRIMO) or after infection (SEROCO). CONCLUSIONS: These data question the benefit of a limited course of cART even when initiated within 3 months after PHI diagnosis.