The diagnosis of human African trypanosomiasis (HAT) due to Trypanosoma brucei gambiense relies on an initial serologic screening with the card agglutination test for trypanosomiasis (CATT) for T. b. gambiense, followed by parasitologic confirmation in most endemic areas. Unfortunately, field parasitologic methods lack sensitivity and the management of serologically suspected individuals (i.e., individuals with a positive CATT result but negative parasitology) remains controversial. In Kajo-Keji County in southern Sudan, we prospectively collected sociodemographic and laboratory data of a cohort of 2,274 serologically suspected individuals. Thirty-three percent (n = 749) attended at least one follow-up visit and HAT was confirmed in 64 (9%) cases. Individuals with lower initial CATT-plasma (CATT-P) end-dilution titers had lowest risks (10.4 and 13.8/100 person-years for 1:4 and 1:8 titers, respectively) that significantly increased for higher dilutions: relative risks = 5.1 (95% confidence interval [CI] = 2.6-9.5) and 4.6 (95% CI = 2.8-9.8) for 1:16 and 1:32 titers, respectively. The cumulative yearly risk was also high (76%) in individuals found with 11-20 cells in the cerebrospinal fluid, but this involved only eight patients. Adjustment for potential confounders did not affect the results. In conclusion, treatment with pentamidine should be considered for all serologically suspected individuals with a CATT-P end-dilution titer >/= 1:16 in areas of a moderate to high prevalence of HAT.