Twenty-three RCTs were included (n=7,885); 11 were classed as high quality and 12 as low quality. All of the results reported here were based on intention-to-treat data unless otherwise stated.
Mortality (18 RCTs):
Mortality ranged from 0% to 7% in the fluoroquinolone group and from 0.5% to 8% in the comparator antibiotics. There was no difference between fluoroquinolone and comparator in mortality (OR 0.85, 95% CI 0.65 to 1.12). There was no statistically significant heterogeneity.
Treatment success (15 trials):
Treatment was successful for 84.2% of patients in the fluoroquinolone group and 82.2% in the comparator group. Significantly more patients in the fluoroquinolone group had treatment success than in the comparison group (OR 1.17, 95% CI 1.00 to 1.36). Similar results were obtained from the analysis based on patients for whom outcome data were available at the end of the study. There was no statistically significant heterogeneity.
Significantly more patients who received fluoroquinolones had treatment success than the comparator group where pneumonia was severe (OR 1.84, 95% CI 1.02 to 3.29; seven RCTs), where intravenous therapy was used (OR 1.44, 95% CI 1.13 to 1.85; 11 trials) and where patients were admitted to hospital (OR 1.30, 95% CI 1.04 to 1.61; 15 RCTs), but not in subgroups of studies of patients with mild to moderate pneumonia or moderate to severe pneumonia, trials that used oral therapy or that included outpatients. Significantly more patients who received fluoroquinolones had treatment success compared to combination therapy (OR 1.39, 95% CI 1.02 to 1.90; seven RCTs), but not β-lactam or macrolide monotherapy. Other subgroups where there was no significant difference in effectiveness between fluoroquinolone and comparator were the patients with Streptococcus pneumonia, bacteraemic pneumonia or bacteraemic pneumonia due to Streptococcus pneumonia.
Analyses based on study quality and source of funding found that fluoroquinolones were not more effective than comparator in the subgroup of studies funded by pharmaceutical companies and in the subgroup of high-quality studies; although in the latter these studies were of patients with milder pneumonia. The authors stated that they did not find evidence of publication bias.
Fluoroquinolones were associated with significantly fewer adverse outcomes than comparator antibiotics (OR 0.86, 95% CI 0.78 to 0.96), although withdrawal due to adverse effects was similar in both groups. Most adverse outcomes were mild to moderately severe disturbances of the gastrointestinal tract.