The aim of this double-blind, 2 parallel group, randomized, multicenter study was to compare the efficacy and the safety of pristinamycin (P), 1 g bid, versus amoxicillin-clavulanic acid (AAC), 500 mg q.i.d., for 10-14 days in the treatment of non severe community-acquired pneumonia in hospitalized adults. From December 1992 to July 1994, 180 patients were included: 92 in the group P and 88 in the group AAC. The both groups were similar on demographic, clinical and bacteriological criteria. 96 pathogens of which more than half were pneumococci, were isolated in 79/180 (44%) patients. The primary assessment was the global success rate defined as long-term (D40 +/- 7), clinical, radiological and bacteriological efficacy in the "per protocol" population (75 patients in the group P and 83 in the group AAC). The global success was obtained in 63/75 (84%) patients in the group P and 70/83 (84.3%) patients in the group AAC. At the end of treatment (D14 +/- 3), theses rates were respectively 85.4% and 84.3%. The both treatments were equivalents. Adverse events (mainly gastro-intestinal disorders) were reported by 55/92 (59.8%) patients in the group P and 49/87 (56.2%) patients in the group AAC.