To evaluate scoring systems to predict Legionella pneumonia and therapeutic efficacy against Legionella pneumonia, the Japanese Society of Chemotherapy Legionella committee has collected data on cases of Legionella pneumonia from throughout Japan. We analyzed 176 patients with Legionella pneumonia and compared them with 217 patients with Streptococcus pneumoniae pneumonia and 202 patients with Mycoplasma pneumoniae pneumonia. We evaluated four scoring systems, the Winthrop-University Hospital score, Community-Based Pneumonia Incidence Study Group score, and Japan Respiratory Society score, but they demonstrated limited sensitivity and specificity for predicting Legionella pneumonia. Using six clinical and laboratory parameters (high fever, high C-reactive protein, high lactate dehydrogenase, thrombocytopenia, hyponatremia, and unproductive cough) reported by Fiumefreddo and colleagues, only 6% had Legionnella pneumonia when less than 2 parameters were present. The efficacy rates of antibiotics at the time of termination were 94.6% for intravenous antibiotics, including ciprofloxacin and pazufloxacin, and 95.5% for oral antibiotics, including ciprofloxacin, levofloxacin, garenoxacin, moxifloxacin, and clarithromycin. Our results suggested that the previously reported clinical scoring systems to predict Legionnella pneumonia are not useful, but 6 simple diagnostic score accurately ruled out Legionnella pneumonia, which may help to optimize initial empiric therapy. Quinolones and clarithromycin still showed good clinical efficacy against Legionella pneumonia.
【저자키워드】 Rapid diagnosis, community-acquired pneumonia, therapeutic efficacy, Legionella, clinical prediction rule,