Performance evaluation of automated depolarization analysis for detecting clinically unsuspected malaria in endemic countries.

This prospective study evaluated the efficiency of automated depolarization analysis for recognition of unsuspected malaria by haemozoin detection during routine full blood count (FBC) screening of 676 randomly selected out-patients in a malaria hypoendemic area of Senegal. An additional 123 patients with clinically suspected malaria were studied for comparison. Of the 799 samples, 648 (81.1%) were categorized as malaria-negative, 83 (10.4%) as malaria-positive, and 68 as treated (early convalescence) or subclinical malaria (indirect evidence of infection). At a discrimination level of one or more atypical pigment-containing monocytes (PCM), negative and positive agreement was found to be 95.6% and 91.6% respectively for all malaria-negative and parasite-positive samples combined. Increasing the discriminator to two or more PCM events improved the overall agreement to 97.5%. Multivariate analysis showed that the only significant risk factor for the presence of PCM (odds ratio>200) was malaria infection. In the randomly selected group of 676 patients, 41 unsuspected cases of malaria infection were detected using the panel of reference diagnostic tests, and 37 (90.2%) of these had atypical PCM. The detection of clinically unrecognized malaria infection as part of a routine FBC procedure is a potentially useful extended application for laboratories in countries with endemic malaria.