ObjectiveThe main aim of this network meta-analysis is to identify the empiric antibiotic (Em-ATB) with the highest probability of being the best (HPBB) in terms of (1) cure rate and (2) mortality rate in hospitalised patients with community acquired pneumonia (CAP) .MethodInclusion criteria: (1) adult patients (>16 years old) diagnosed with CAP that required hospitalisation; (2) randomised to at least two different Em-ATBs, (3) that report cure rate and (4) are written in English or Spanish. Exclusion criteria: (1) ambiguous antibiotics protocol and (2) published exclusively in abstract or letter format. Data sources: Medline, Embase, Cochrane and citation reviews from 1 January 2000 to 31 December 2018. Risk of bias: Cochrane’s tool. Quality of the systematic review (SR): A MeaSurement Tool to Assess systematic Reviews-2. Certainity of the evidence: Grading of Recommendations Assessment, Development and Evaluation. Statistical analyses: frequentist method performed with the ‘netmeta’ library, R package.Results27 randomised controlled trials (RCTs) from the initial 41 307 screened citations were included. Regarding the risk of bias, more than one quarter of the studies presented low risk and no study presented high risk in all domains. The SR quality is moderate. For cure, two networks were constructed. Thus, two Em-ATBs have the HPBB: cetaroline 600 mg (two times a day) and piperacillin 2000 mg (two times a day). For mortality, three networks were constructed. Thus, three Em-ATBs have the HPBB: ceftriaxone 2000 mg (once a day) plus levofloxacin 500 (two times a day), ertapenem 1000 mg (two times a day) and amikacin 250 mg (two times a day) plus clarithromycin 500 mg (two times a day). The certainity of evidence for each results is moderate.ConclusionFor cure rate, ceftaroline and piperaciline are the options with the HPBB. However, for mortality rate, the options are ceftriaxone plus levofloxacin, ertapenem and amikacin plus clarithromycin. It seems necessary to conduct an RCT that compares treatments with the HPBB for each event (cure or mortality) (CRD42017060692).

No abstract