Detection by Nasopharyngeal Swabs Alone Underestimates Respiratory Syncytial Virus–Related Hospitalization Incidence in Adults: The Multispecimen Study's Final Analysis

BACKGROUND: Most epidemiologic studies and clinical testing use single nasal/nasopharyngeal swab (NPS) for respiratory syncytial virus (RSV) detection. Studies document that RSV detection improves if another specimen is added to NPS, but the impact of using multiple specimen types has not been assessed. We quantified RSV detection increase using multiple-specimen testing versus NPS alone. METHODS: Hospitalized adults aged ≥40 years with acute respiratory illness were prospectively enrolled in 7 US/Canadian hospitals. NPS, saliva, sputa, and acute/convalescent sera were collected and tested. RESULTS: Among 3669 participants, 100% had NPS, 97.7% saliva, 33.0% sputum, and 33.4% paired serology. RSV detection was 112% higher (95% CI, 86%-141%) using all specimen types as compared with NPS alone. Serology test sensitivity was the highest (73.0%; 95% CI, 65.1%-80.8%), followed by sputum (70.1%; 95% CI, 62.1%-78.0%), saliva (61.4%; 95% CI, 55.4%-67.5%), and NPS (47.2%; 95% CI, 41.1%-53.4%). Among those with congestive heart failure exacerbations, additional specimens increased detection by 267% (95% CI, 85%-625%), and saliva detected more infections than NPS. Among 1013 participants with paired NPSs from different time points, specimens collected on average 1 day later detected 30% fewer RSV infections. CONCLUSIONS: RSV detection increased >100% using 4 specimen types, suggesting a 2-fold correction factor is appropriate for incidence and prevalence studies relying on NPS alone. Saliva was more sensitive than NPS, warranting further study, particularly in cardiac cases.