OBJECTIVE : The purpose of this study was to systematically investigate the reporting of selection criteria and outcome measures, and to examine definitions of complications used in venoarterial extracorporeal membrane oxygenation studies (V-A ECMO). METHODS : Medline, EMBASE and the Cochrane central register were searched for V-A ECMO studies from January 2005 to July 2017. Studies with ≤99 patients or without patient centered outcomes were excluded. Two reviewers independently assessed search results and undertook data extraction. RESULTS : Forty-six studies met the inclusion criteria, and all were retrospective, observational studies. Inconsistent reporting of selection criteria, ECMO management and outcome measures was common. In-hospital mortality was the most common primary outcome (41% of studies), followed by 30-day mortality (11%). Bleeding was the most frequent complication reported, most commonly defined as "bleeding requiring transfusion" (median ≥ 2 Units/day). Significant variation in reporting and definitions was also evident for vascular, neurological renal and infectious complications. CONCLUSIONS : This systematic review provides clinicians with the most commonly reported selection criteria, outcome measures and complications used in ECMO practice. However non-standardized definitions and inconsistent reporting limits their ability to inform practice. New consensus driven definitions of complications and patient centred outcomes are urgently needed.