Ampicillin versus cefamandole as initial therapy for community-acquired pneumonia.

One hundred seven patients with community-acquired pneumonia thought to be of bacterial etiology by the admitting physician but whose initial sputum Gram stain was inadequate to direct specific therapy were randomized to receive either intravenous ampicillin or cefamandole as empiric therapy. Patients were excluded if the initial sputum Gram stain was highly suggestive of infection with Streptococcus pneumoniae, Staphylococcus aureus, or an enteric gram-negative bacillus. The two study groups had comparable demographic and presenting clinical features. The mean age of the patients evaluable for determination of clinical efficacy was 69 years, and greater than 75% had at least one serious underlying medical disorder. In the 90 evaluable patients, there were 11 therapeutic failures (12%), including 5 deaths (5%). Cefamandole, a broad-spectrum antibiotic, was not more efficacious than ampicillin in producing a satisfactory clinical response or in shortening the duration of parenteral therapy. Patients received an average of only 4 days of intravenous antibiotics before changeover to oral therapy and were hospitalized for a mean of 7 days. No patient experienced a relapse of pneumonia following successful completion of parenteral drug therapy. We conclude that cefamandole is not a more effective agent than ampicillin for empiric therapy of community-acquired bacterial pneumonia of uncertain etiology.