Objective: To identify factors available to rescuers at the scene of a drowning that predict favourable outcomes. Design: Systematic review and meta-analysis. Data sources: PubMed, Embase and Cochrane Library were searched (1979-2015) without restrictions on age, language or location and references lists of included articles. Study selection: Cohort and case-control studies reporting submersion duration, age, water temperature, salinity, emergency services response time and survival and/or neurological outcomes were eligible. Two reviewers independently screened articles for inclusion, extracted data, and assessed quality using GRADE. Variables for all factors, including time and temperature intervals, were categorized using those used in the articles. Random effects meta-analyses, study heterogeneity and publication bias were evaluated. Results: Twenty-four cohort studies met the inclusion criteria. The strongest predictor was submersion duration. Meta-analysis showed that favourable outcome was associated with shorter compared to longer submersion durations in all time cutoffs evaluated: ≤5-6min: risk ratio [RR]=2.90; (95% confidence interval [CI]: 1.73, 4.86); ≤10-11min: RR=5.11 (95% CI: 2.03, 12.82); ≤15-25min: RR=26.92 (95% CI: 5.06, 143.3). Favourable outcomes were seen with shorter EMS response times (RR=2.84 (95% CI: 1.08, 7.47)) and salt water versus fresh water 1.16 (95% CI: 1.08, 1.24). No difference in outcome was seen with victim’s age, water temperatures, or witnessed versus unwitnessed drownings. Conclusions: Increasing submersion duration was associated with worse outcomes. Submersion durations 25min were invariably fatal. This information may be useful to rescuers and EMS systems deciding when to perform a rescue versus a body recovery.