Background Hospitalized patients with non-severe community-acquired pneumonia (CAP) are treated with a β-lactam plus either a macrolide or doxycycline. Limited data exist on the effectiveness of the latter combination. Therefore, we aimed to compare the combination of doxycycline vs. macrolide when either is combined with a β-lactam from effectiveness and safety perspectives. Methods This was a retrospective cohort study in CAP inpatients between December 2013 and November 2020. Patients were divided into BL-D (β-lactam plus doxycycline) and BL-M (β-lactam plus a macrolide [azithromycin or clarithromycin]) groups. The primary endpoint was time to clinical stability. Secondary endpoints included length of stay (LOS) and in-hospital mortality. Results Of 197 patients included, 57 were in the BL-D arm and 140 were in the BL-M arm. Patients were similar at baseline, except for the presence of leukocytosis, risk factors for drug resistance, and duration of therapy (P < 0.05 for all comparisons). No difference in clinical cure rate was observed (94.7% vs. 91.4%; P = 0.43). Time to clinical stability and LOS were similar in both groups at 4 (P = 0.82) and 7 days (P = 0.62), respectively. While only three patients died, only one (from the BL-M group) was due to sepsis. Liver enzymes elevation was more prominent in the BL-M group (21.4% vs. 5.3%; P = 0.01). A subgroup analysis showed shorter time to clinical stability with clarithromycin but higher cure rates with azithromycin. Conclusions Data on doxycycline use with a β-lactam are scarce. Our study showed that such regimen was comparable in effectiveness to regimens involving macrolides with a better safety profile.