Background Determining lower respiratory tract infection (LRTI) aetiology is complex. Culture-based methods are laborious with poor sensitivity. Molecular assays improve detection of potential pathogens, but incorrect interpretation of results may lead to inappropriate antimicrobial therapy. Methods The utility of the BioFire® FilmArray® Pneumonia Panel plus (FA-PP) to detect LRTI pathogens, and the potential impact on antimicrobial stewardship in a low-resource setting, were assessed. Routine LRT samples were included from adult patients with clinically suspected LRTI or with a concomitant blood culture at Groote Schuur Hospital and referring facilities. Culture and FA-PP results were compared, and pharmacy data analysed to determine appropriateness of antibiotic therapy. Results There was an 80% correlation between cultured LRTI pathogens and the FA-PP bin ≥107 results. Compared with culture, the FA-PP detected substantially more pathogens (86.6% versus 17.9%) and produced a combined 100% positive percent agreement, and 88% negative percent agreement. The FA-PP detected bacterial/viral coinfections in 27% of samples. Correlation of FA-PP results with pharmacy data (n = 69) indicated a potential antibiotic change in 75% of cases, but this is difficult to accurately characterize without a ‘gold standard’ for treatment or complete clinical data. Conclusions The FA-PP increased the number of positive samples with typical bacteria, but the semi-quantitative reporting algorithm does not describe the correlation between the different bin values and colonization versus infection. This complicates result interpretation and may lead to inappropriate antimicrobial treatment. This study highlights the potential positive impact of rapid molecular assays for routine care in lower-income settings, but also underscores the interpretive challenges associated with these tests.

Toxigenic Vibrio cholerae O1/O139 is causative of cholera, which is a well characterised potentially epidemic gastrointestinal disease. Less is known about the pathogenesis and clinical presentation of non-toxigenic V. cholerae non-O1/non-O139, although they are increasingly implicated in human disease globally, have been isolated from various South African water sources and can contaminate the environment. The authors describe a case of pseudo-bacteraemia with non-toxigenic V. cholerae non-O1/non-O139 in a neonate.