BACKGROUND: Several studies and meta-analyses have shown that mortality in people with schizophrenia is higher than that in the general population but have used relative measures, such as standardised mortality ratios. We did a systematic review and meta-analysis to estimate years of potential life lost and life expectancy in schizophrenia, which are more direct, absolute measures of increased mortality. METHODS: We searched MEDLINE, PsycINFO, Embase, Cinahl, and Web of Science for published studies on years of potential life lost and life expectancy in schizophrenia. Data from individual studies were combined in meta-analyses as weighted averages. We did subgroup analyses for sex, geographical region, timing of publication, and risk of bias (estimated with the Newcastle-Ottawa Scale). FINDINGS: We identified 11 studies in 13 publications covering all inhabited continents except South America (Africa n=1, Asia n=1, Australia n=1, Europe n=7, and North America n=3) that involved up to 247 603 patients. Schizophrenia was associated with a weighted average of 14·5 years of potential life lost (95% CI 11·2-17·8), and was higher for men than women (15·9, 13·8-18·0 vs 13·6, 11·4-15·8). Loss was least in the Asian study and greatest in Africa. The overall weighted average life expectancy was 64·7 years (95% CI 61·1-71·3), and was lower for men than women (59·9 years, 95% CI 55·5-64·3 vs 67·6 years, 63·1-72·1). Life expectancy was lowest in Asia and Africa. Timing of publication and risk of bias had little effect on results. INTERPRETATION: The effects of schizophrenia on years potential life lost and life expectancy seem to be substantial and not to have lessened over time. Development and implementation of interventions and initiatives to reduce this mortality gap are urgently needed. FUNDING: None.