[Assessment of five rapid tests and two algorithms for the HIV infection diagnosis in Niger].

Niger is a Sahelian country of 1,267,000 km2 and 11 M inhabitants (84% living in rural area). The national seroprevalence of the HIV infection is low (2.08% in urban area and 0. 64% in rural area). Because of limited resources, laboratories and qualified personnel are scarce, so the classical strategy using ELISA methods to diagnose HIV is available only in the capital Niamey. In order to define a strategy suitable to the Nigerian field conditions, we have determined the diagnostic value of five commercialised rapid assays (Génie II, InoCheck, Determine, Capillus and Doublecheck). A panel of plasma, whole blood and dried blood spot from 42 positive and 160 negative control individuals was tested and the sensitivity and specificity of each assay were calculated. The two best algorithms using two tests were selected: a algorithm discriminating HIV1 & 2 combining Determine and Genie II showing an excellent sensitivity and specificity (100%) but with many practical constraints; and a non discriminative algorithm combining Determine and Double-check less specific but workable on capillary blood and fully compatible with the field conditions. Its negative predictive value was 100% and its positive predictive value was > 97% for urban population and > 91% for rural population. Dried blood spots on filter paper proved to be very practical and efficient to use by the reference laboratory for the quality control of the peripheral diagnostic centres.