Guidelines for inpatients with community-acquired pneumonia (CAP) propose to use respiratory fluoroquinolone (RFQ) and/or third-generation cephalosporins (Ceph-3). However, broad-spectrum antibiotic therapy is associated with the emergence of drug-resistant bacteria. We established a guideline in which RFQ and Ceph-3 were excluded as a first course. Our aim was to evaluate the impact of our therapeutic choices for CAP on the length of hospital stay (LOS) and patient outcome. This was a cohort study of patients with CAP from July 2005 to June 2014. We compared patients benefiting from our guideline established in 2008 to those receiving non-consensual antibiotics. Disease severity was evaluated through the Pneumonia Severity Index (PSI). The empirical treatment for PSI III to V was a combination therapy of amoxicillin-clavulanic acid (AMX-C) + roxithromycin (RX) or AMX + ofloxacin. Adherence to guidelines was defined by the prescription of one of these antibiotic agents. Requirement for intensive care or death defined unfavorable outcome. Among 1,370 patients, 847 were treated according to our guideline (61.8 %, group 1) and 523 without concordant therapy (38.2 %, group 2). The mean PSI was similar: 82 vs. 83, p > 0.5. The mean LOS was lower in group 1: 7.6 days vs. 9.1 days, p < 0.001. An unfavorable outcome was less frequent in group 1: 5.4 % vs. 9.9 %, p = 0.001. In logistic regression models, concordant therapy was associated with a favorable outcome: adjusted odds ratio (AOR) [95 % confidence interval (CI)] 1.85 [1.20-2.88], p = 0.005. CAP therapy without RFQ and Ceph-3 use was associated with a shorter LOS and fewer unfavorable outcomes.