Comparison of operational criteria for treatment outcome in gambiense human African trypanosomiasis

To develop a simple and standard operational decision tool for the diagnosis of relapse after treatment for human African trypanosomiasis (HAT), by evaluating the performance of several criteria currently used by HAT control programs and research projects. We identified 10 different criteria for relapse, based on trypanosome presence and/or white blood cell count in cerebrospinal fluid, and compared their specificity, sensitivity and time to diagnosis on a data set containing 63 relapsed and 247 cured T.b. gambiense patients. At any time point, the criterion 'Trypanosomes present and/or a cerebrospinal white blood cell count >= 50/mu l' allowed accurate and timely detection of HAT relapse, irrespective of disease stage. This criterion was 13-25% more sensitive (P <= 0.013) than trypanosome detection alone and was > 97% specific. Lumbar punctures at the end of treatment and at 3-month post-treatment provided limited clinical information. Adequate detection of relapse was possible with a simple criterion but these findings should be validated in a prospective study before adoption in clinical practice.