Résumé
BackgroundSerological tests play a crucial role to diagnose gambiense human African trypanosomiasis (HAT) by preselecting individuals for microscopic examination, and, in the near future, by directly identifying patients for treatment. Variability in reported specificities, the introduction of new rapid diagnostic tests (RDT) and the hypothesis that malaria decreases RDT specificity, led us to evaluate the specificity of 5 HAT screening tests.MethodsVenous blood samples from 1095 individuals from Côte d’Ivoire and Guinea were tested with commercial (Bioline HAT 2.0, HAT Sero-K-SeT, CATT/T.b. gambiense) and experimental (HAT Sero-K-SeT 2.0, DCN) HAT screening tests and with a malaria RDT. Individuals negative with all 5 HAT tests were considered HAT free, while positives underwent microscopy. HAT case definition was based on trypanosome detection by microscopy.ResultsOne HAT case was detected. Test specificities (n=1094) were: CATT/T.b. gambiense [98.9% (98.1–99.4%), p<0.0001] > HAT Sero-K-SeT [86.7% (84.5–88.5%), p<0.002] > Bioline HAT 2.0 [82.1% (79.7–84.2%), p=0.0113] > HAT Sero-K-SeT 2.0 [78.5% (76.0–80.9%)] and DCN [78.2% (75.7–80.6%)]. Bioline HAT 2.0 and DCN include 2 test lines, and specificities of line 1 [respectively 83.7% (81.4–85.8%) and 80.6% (78.2–82.9%)], corresponding to ISG-65, were significantly lower (p 0.0001) than with line 2 [respectively 95.8% (94.4–96.8%) and 94.5% (93.0–95.7%)]. The ISG-65 line therefore significantly decreased overall test specificity. Although all the HAT tests were less specific in malaria positive than in malaria negative individuals, differences (p values >0.08) were not significant.ConclusionCATT/T.b. gambiense is more specific than HAT RDTs. The HAT Sero-K-SeT is more specific than second generation RDTs which all contain ISG-65, either as a separate test line (Bioline HAT 2.0 and DCN) or within a single “mixed antigen” test line (HAT Sero-K-SeT 2.0). To improve specificity, removing ISG-65 from experimental RDTs or ignoring the ISG-65 line should be considered, if test sensitivity is not significantly impacted.