Introduction Microbiological testing for atypical pathogens in patients attending hospital with community acquired pneumonia (CAP) is recommended for moderate or severe disease (NICE CG191 2014) or for patients failing to respond to treatment. Although it is unclear whether testing improves outcome even in severe disease, many patients have such tests performed regardless of severity. Having revised our pathways for assessment, treatment and documentation of patients with community acquired pneumonia we hypothesised that testing for atypical organisms has no impact on treatment decisions for these patients.

Method We retrospectively identified all patients with a diagnosis of CAP who had investigations for atypical microbiology, September 2013 to May 2014, via our pneumonia database. We assessed CURB-65 score, atypical microbiology results and laboratory costings. The notes for all patients with positive atypical microbiological results were reviewed.

Results 343 patients were identified for whom 329 were analysed.

329 patients generated 991 samples in total (825 serum, 165 urine antigen, 1 urine virology) at a laboratory cost of £5,594.29.

Five samples were positive, one for urine legionella antigen.

Greater than 50% of serological samples had no second (paired) sample sent.

There was no correlation between CURB-65 scores and requesting of atypical microbiology requesting.

One patient with positive legionella antigen had prolongation of treatment from 5 days to 14 days.

No other patients had treatment changes as a consequence of atypical microbiological testing.

Conclusion Atypical microbiological testing, in hospital, for CAP patients is commonly performed at significant cost with minimal clinical utility. We recommend that non-selective serological sampling is abandoned. The impact of legionella urinary antigen testing on outcome in moderate and severe cases requires a prospective study.