BACKGROUND: The effect of CD4 count on malaria incidence in HIV infected adults on antiretroviral therapy (ART) was assessed in the context of a randomized controlled trial on the effect of stopping cotrimoxazole (CTX). METHODS: This study presents a sub-analysis of the COSTOP trial (ISRCTN44723643) which was carried out among HIV-infected Ugandan adults stable on ART with CD4 counts ≥250 cells/µl. Participants were randomized (1:1) to continue CTX or stop CTX and receive matching placebo, and were followed up for a minimum of 1 year (median 2.5 years). CD4 counts were measured at baseline, 3 months and then every 6 months. Clinical malaria was defined as fever and a positive blood slide. First, the relationship between current CD4 count during follow-up and malaria among participants on placebo was examined; using random effects Poisson regression to account for repeated episodes. Second, the effect of CD4 count at enrolment, CD4 count at ART initiation, and CD4 count during follow-up on malaria, was assessed within each trial arm; to examine whether the effect of CD4 count differed by CTX use. RESULTS: 2180 participants were enrolled into the COSTOP trial. The incidence of clinical malaria was approximately four episodes/100 person years in the CTX arm and 14 episodes/100 person years in the placebo arm. There was no evidence of an association of current CD4 and clinical malaria incidence (P = 0.56), or parasitaemia levels (P = 0.24), in the placebo arm. Malaria incidence did not differ by CD4 count at ART initiation, enrolment or during follow up, irrespective of CTX use. When compared with participants in the lowest CD4 stratum, rate ratios within each trial arm were all close to 1, and P values were all above P = 0.30. CONCLUSIONS: The immune status of HIV infected participants who are stable on ART as measured by CD4 count was not associated with malaria incidence and did not modify the effect of stopping CTX on malaria. The decision of whether to stop or continue CTX prophylaxis for malaria in HIV infected individuals who are stable on ART should not be based on CD4 counts alone. COSTOP trial registration number ISRCTN44723643.