Prospective observational study of bacteremic pneumococcal pneumonia: Effect of discordant therapy on mortality.

Abstract

To evaluate the effect of discordant empirical therapy on outcome in bacteremic pneumococcal community-acquired pneumonia. Prospective observational study. A 600-bed teaching hospital with a reference area of 400,000 inhabitants. All patients aged > or =18 yrs with a diagnosis of community-acquired pneumonia whose blood cultures, obtained within the first 48 hrs of hospitalization, demonstrated growth of Streptococcus pneumoniae were included in the study. Discordant therapy was defined as failure to administer an antibiotic with in vitro activity against the isolated strain within 24 hrs of hospital admission. The 2002 breakpoints recommended for respiratory infections by the National Committee for Clinical Laboratory Standards were used to classify therapy. A total of 100 patients with bacteremic pneumococcal pneumonia were identified. Penicillin- and macrolide-resistant strains were identified in 29 and 18 cases, respectively. Only two strains had minimum inhibitory concentrations of >2 microg/mL for cephalosporins. Discordant therapy was documented in ten patients, five of whom died. Mortality in patients receiving concordant therapy was 14% (13 of 90). Nursing home residence (odds ratio [OR] = 14.8) and immunocompromise (OR = 11.5) were independently (p <.05) associated with discordant therapy. Risk of discordant therapy was significantly higher (p <.05) when empirical therapy did not include cefotaxime or ceftriaxone (OR = 10.4). Discordant therapy (OR = 27.3), multilobar involvement (OR = 14.2), underlying chronic obstructive pulmonary disease (OR = 9.1), and hospitalization during the previous 12 wks (OR = 7.9) were independently associated (p <.05) with death. The excess mortality for initial discordant therapy was estimated to be 35.6% (95% confidence interval, 3.73-67.4). Survival in patients with bacteremic community-acquired pneumococcal pneumonia can be improved by avoiding suboptimal therapy. Using the 2002 breakpoints, it is very unlikely that discordant therapy would be given with ceftriaxone or cefotaxime. Clinical outcome is worse in those patients receiving antimicrobial therapy that in vitro testing suggests would be ineffective.