Cefepime versus Ceftriaxone for Empiric Treatment of Hospitalized Patients with Community-Acquired Pneumonia

We read with interest the report of the clinical trial by Zervos et al. (5) in which cefepime was compared with ceftriaxone for empiric treatment of hospitalized patients with community-acquired pneumonia. Although that paper presented some interesting findings, there are several methodological flaws with the study that we would like to comment on.

The first problem relates to patient eligibility. The criteria for hospitalization of patients (3, 4) were not clearly defined. It appears that patients were included in the trial without radiological evidence of pneumonia, even though it has been suggested that a chest film showing infiltrates is necessary to establish the diagnosis of pneumonia (1). Moreover, the authors apparently also considered as a criterion for patient eligibility the presence of bacteria susceptible to cefepime and ceftriaxone. This seems rather risky since it has been shown that it is seldom possible to reach a specific etiologic diagnosis by roentgenography alone (4). We also found that in Table 2, which describes the characteristics of pneumonia presented by the patients at the time of entry in the study, the severity of pneumonia is classified as mild, moderate, or severe, but definitions for these categories are not provided.

A second difficulty encountered by the reader is that, in evaluating efficacy, the authors define the category “improvement” as “improvement of one or more signs and symptoms of pneumonia without a complete resolution but a lack of progression of radiographic findings,” and consider it a favorable clinical response. This would mean that if a patient improved in just one symptom, say for example, cough, this would be regarded as a “favorable clinical response.” We feel that the category “improvement” makes little medical sense in dealing with pneumonia. In addition, in the General Guidelines for the Evaluation of New Anti-Infective Drugs for the Treatment of Respiratory Tract Infections (2), this category is not included in the definition of clinical response. Moreover, it is not clear how the category “improvement” was determined for patients which were included in the study without a previous chest film.

Thirdly, we are also concerned about the fact that in this study some patients were treated concomitantly with several other antibiotics. Although treatment with erythromycin and metronidazole can be justified in cases where atypical pathogens or anaerobes are suspected, there were seven patients who received six additional antibiotics. Despite the fact that the authors suggest that antibiotic treatment had little impact on the efficacy of the drugs studied, it is difficult to classify these patients with respect to clinical response. Besides, the authors never mention when these additional antibiotics where added or what the criterion was for adding them. We feel that the criteria for patient evaluation in this study are debatable. For example, what exactly do the authors mean when they say that “Seven additional patients who received additional antimicrobial agents were approved by the sponsor as evaluable”?

Finally, this study also lacks a clearly stated hypothesis and, thus, a sample size calculation.