Severity scoring in community-acquired pneumonia caused by Streptococcus pneumoniae: a 5-year experience.

Multiple severity scoring systems have been devised and evaluated in community-acquired pneumonia (CAP), but a simplified set of prognostic indicators has not yet been developed. Streptococcus pneumoniae is the most frequent aetiological agent of CAP. Our aim was to characterise the outcome in the light of different severity scoring systems and to compare the predictive values of different sets of clinical parameters, using available clinical data for pneumococcal CAP patients. This is a case series retrospective analysis that included consecutive adult pneumococcal CAP patients admitted to Danbury Hospital between 1 January 1996 and 31 December 2000. The aetiology was confirmed by positive sputum and/or blood cultures. The severity assessment included the Pneumonia Outcome Research Trial (PORT) and British Thoracic Society (BTS) scoring systems and other additional parameters. Primary end-points were in-hospital CAP-attributable deaths and length of hospitalisation. N = 151 patients with S. pneumoniae CAP were identified. The mean (+/- standard deviation) age at the time of diagnosis was 68 (+/-15) years. Thirty-three patients (22%) were admitted to the medical intensive care unit. The mean (median) hospitalisation duration was 7.5 (+/-5) days. Door-to-antibiotic mean (median) administration time was 3.7 (2) hours. Most frequent antibiotics used initially were cephalosporins plus/minus macrolides or fluoroquinolones. The mean (+/- standard deviation) PORT score was 105 (+/-37). The observed CAP-related mortality was 9/151 (5.9%, 95% confidence interval: 3-9%). The mortality rate in ICU was 18% (6/33). Sixty-nine patients (45%) had S. pneumoniae bacteraemia an admission. The bacteraemic and non-bacteraemic patients had similar PORT scores (107 vs. 104, P = 0.66), length of hospitalisation (8 vs. 7 days, P = 0.41) and mortality rates (9% vs. 4%, P = 0.30). In conclusion, patients admitted with pneumococcal CAP, although severe and with multiple co-morbidities had low in-hospital mortality rates and lengths of hospitalisation. Neither prior antimicrobial use (or failure) nor antimicrobial resistance contributed to an adverse outcome. S. pneumoniae bacteraemia failed to correlate with need for ICU, length of stay, higher morbidity index or fatal outcome. Low rates of empirical antibiotic use for non-bacterial infections in the local community, implementation of an emergency department protocol for CAP therapy, early recognition of higher risk patients and placement in ICU, use of broad spectrum antibiotics, infectious disease approval or critical pathway restriction for admission orders, could all have combined to effect a good outcome for these patients.