We performed urinary antigen tests for pneumococcus and Legionella for patients with community-acquired pneumonia (CAP), to prescribe a documented antibiotic therapy. We report the efficiency of low-spectrum antibiotic treatment, illustrating the inappropriateness of bacteriological respiratory sampling.
Patients and methods: Patients with CAP were enrolled from three different units; the pneumonia severity index was used to assess the disease. Respiratory samples were also listed. Low-spectrum antibiotic therapy was amoxicillin for pneumococcal infection, and macrolides or non-anti-pneumococcal fluoroquinolone for legionellosis.
Results: Six hundred and seventy-five CAP were diagnosed during the study period,, 150 with positive urinary antigen tests (23%), among which 108 pneumococcal infections (73%), 40 legionellosis (26%), and two mixed infections. The pneumonia severity index was 106+/-38. Amoxicillin was prescribed in 108 cases, fluoroquinolone in 24 cases, macrolide in 18 cases. The outcome was favourable for 138 patients (92%). Eighty three respiratory samples allowed identification of a bacterium for 58 patients (39%), among which 24 strains were not in the antibiotic spectrum: Haemophilus influenzae and Pseudmomonas aeruginosa in six cases, Staphylococcus aureus in five cases, Klebsiella pneumoniae in two cases, and another Gram negative bacillus in five cases. These strains were resistant in vitro to the prescribed treatment in 19/24 cases (79%). One out of 12 patients who died had a respiratory sample positive for Enterobacter spp strain resistant to the ongoing antibiotic treatment.
Conclusion: The low-spectrum antibiotic therapy based on urinary antigen tests is efficient, and demonstrates respiratory tract colonisation with bacteriological strains usually considered as pathogenic.