OBJECTIVES::  There is an evident increase of publications of systematic reviews with network meta-analyses (NMA) to compare several medical interventions. However, since this recent technique still poses some challenges, its characteristics should be better investigated. We aimed to characterize the worldwide NMA publications on drug interventions. METHODS::  A systematic review of NMAs comparing any pharmacological intervention was performed. Searches in Medline and Scopus along with manual search were conducted. The main characteristics were collected: authors/year/country of publication, evaluated medical condition, evaluated drugs, and statistical analyses performed. RESULTS::  We identified 365 NMAs conducted in more than 30 different countries (2003-2016) and published by more than 200 different journals. United States (n=115), United Kingdom (n=86) and China (n=73) published the higher number of studies. Placebo was used as comparator by 230 NMAs (63%). Almost all studies involved randomized controlled trials and used Bayesian approach methods. NMAs evaluating different doses or drugs regimens sum 144 (39%). Cardiovascular diseases were evaluated by 98 NMAs, being aspirin and apixaban (41% and 26%, respectively) the most studied drugs. Breast cancer (9 NMA) and leukaemia (6 NMA) were part of the 50 studies on oncologic disorders. Adalimumab, infliximab and methotrexate were the main interventions for autoimmune disorders, each one being included in 18 NMAs. Mental health diseases were referred in 32 NMAs, being schizophrenia (7 studies) mostly evaluated with risperidone (34%). Pegylated-interferon was analyzed by 66% of the 32 NMAs on infectious diseases. Salmeterol and tiotropium (included in 14 NMA each) were analyzed for respiratory disorders, mainly chronic obstructive pulmonary disease. Other NMAs accounted for musculoskeletal disorders, pain, skin/eye disorders. CONCLUSIONS:: The map of the published NMAs emphasizes the potential of this tool and highlights the gaps on literature of healthcare evidence synthesis. The conduction and report of NMA should reflect transparency and reproducibility.