Appropriate treatment for community acquired pneumonia caused by Streptococcus pneumoniae leads to improved clinical outcomes. Unfortunately diagnosis is delayed as culture of clinical samples takes at least 24 hours and due to its fastidious nature, S. pneumoniae may not always be isolated by culture. In response to the need for more efficient means of diagnosis we evaluated the BinaxNOW Streptococcus pneumoniae lateral flow urinary antigen test, using fresh unconcentrated urine.
Urinary antigen results were retrieved retrospectively from a 5 year period. Respiratory and blood culture results were then manually extracted for each of these patients. Sixteen duplicate urinary antigen tests were all seen to give the same result and so removed from further analysis. As none of the three sample types can be considered a gold standard a Bayesian latent class analysis was performed in R with the LCCA package to estimate performance characteristics of each test.
Twenty-nine of 474 urinary antigen tests were positive. 425 of these had at least one culture sample of which 14 were positive. Urinary antigen was detected but S. pneumoniae not isolated by culture in 19 patients. Four had a negative urinary antigen result but positive cultures for S. pneumoniae, including one blood culture. The estimated sensitivity with 95% confidence intervals was 73% (46-100) for urinary antigen and estimated specificity was 97% (95-99). The estimated negative predictive value was 98% and the positive predictive value was 56%.
The evaluation demonstrates that when used on fresh unconcentrated urine the BinaxNOW Streptococcus pneumoniae lateral flow urinary antigen test has a moderate sensitivity and high specificity which is comparable to previously published literature. Analysis of this data has highlighted a need to assess and improve the testing algorithms used within the Fiona Stanley Hospital Network for the management of community acquired pneumonia.