Objectives: To establish the incidence, risk factors and correlation with survival of thrombocytopenia and thrombocytosis (T/T) among children with HIV infection (CWH).Design: A retrospective nested case control study of patients 0-18 years in five Baylor International Pediatric AIDS Initiative (BIPAI) centers in sub-Sahara Africa, 2004Africa, -2014 Methods: Clinical and laboratory variables including complete blood counts (CBC) were extracted from the BIPAI electronic medical record system. Incident cases of T/T were identified and frequency-matched on follow-up time with controls with normal platelets. We calculated the prevalence and incidence density of T/T and used conditional logistic regression to evaluate their association with selected clinical variables. We constructed Kaplan-Meier curves and a Cox proportional hazards model to evaluate the impact of T/T on survival.Results: Two thousand, one hundred and nine children were sampled. The incidence density of thrombocytopenia was 1 per 57.9 (95% confidence interval [CI] 50.3-66.8) CWHyears. Thrombocytopenia was higher in children with WHO Stage III/IV, lower in children on zidovudine, and had no association with use of lamivudine or nevirapine, CD4 þ suppression, age, and nutrition status. Thrombocytopenia was independently associated with 2.2-fold higher mortality (95% CI 1.62-3.08). The incidence density of thrombocytosis was 1 per 11.4 (95% CI 10.7-12.1) CWH-years. Thrombocytosis was associated with higher CD4 þ cell count, younger age, and use of lamivudine or nevirapine, and did not impact survival.Conclusions: Platelet count is a clinically valuable biomarker of HIV clinical progression and mortality. Laboratory studies are necessary to elucidate the mechanisms of T/T.