Abstract
Nasal pathogen detection sensitivities can be as low as 70% despite advances in molecular diagnostics. This may be linked to the choice of sampling method. A diagnostic test accuracy review for sensitivity was undertaken to compare sensitivity of swabbing to the nasopharynx and extracting nasal aspirates, using the PRISMA protocol, Cochrane rapid review methodology, and QUADAS-2 risk of bias tools, with meta-analysis of included studies. Sensitivities were calculated by a consensus standard of positivity by either method as the 'gold standard.' Insufficient sampling methodology, cross sectional study designs, and studies pooling samples across anatomical sites were excluded. Of 13 subsequently eligible studies, 8 had 'high' risk of bias, and 5 had 'high' applicability concerns. There were no statistical differences in overall sensitivities between collection methods for eight different viruses, and this did not differ with use of PCR, immunofluorescence, or culture. In one study alone, Influenza H1N1(2009) favored nasopharyngeal swabs, with aspirates having 93.3% of the sensitivity of swabs (p > 0.001). Similarly equivocal sensitivities were noted in reports detecting bacteria. The chain of sampling, from anatomical site to laboratory results, features different potential foci along which sensitivity may be lost. A fair body of evidence exists that use of a different sampling method will not yield more respiratory pathogens.