Prospective case-control analysis of the aetiologies of acute undifferentiated fever in Vietnam

Le-Viet, Nhiem; Le, Viet-Nho; Chung, Hai; Phan, Duc-Tuan; Phan, Quang-Duong; Cao, Thanh-Van; Abat, Cédric; Raoult, Didier; Parola, Philippe

The codetection of O. tsutsugamushi and Leptospira in two patients could be considered a coinfection, as documented previously [7] . Almost all other cases of codetections of respiratory viruses and bacteria might also be coinfections. Unlike other studies that used serological tests, which might yield false-positive results due to cross-reacting antibodies or the nonspecific polyclonal immunoreactivity of different aetiologies [57] , we used a qPCR-direct method for pathogen detection and thus obtained strong evidence of the presence of the detected pathogens in the codetection cases. However, the results should be carefully interpreted because the presence of a microorganism in a febrile patient does not always indicate that it is a cause of the fever, and PCR cannot differentiate between infection and carriage or colonization. Thus, we first assessed every detected virus or bacterium in comparison to its frequency in AUFs and Controls, and we then evaluated the combination of positive clinical manifestations in the febrile patients and the presence of viruses or bacteria if their presence was not significantly different between the AUFs and Controls. Accordingly, although the detected frequency of various viruses and bacteria, particularly in throat swabs, was significant, we considered only 73 codetections, which consisted of coinfection with two or three pathogens, as causes or probable causes of AUF (Table 5 ).