ABSTRACTBACKGROUNDLong-term ozone (O3) exposure could lead to a series of non-communicable diseases and increase the mortality risks. However, cohort-based studies were still rather rare, and inconsistent exposure metrics might impair the credibility of epidemiological evidence synthetisation. To provide more accurate meta-estimation, this review updated the systematic review with inclusion of recent studies and summarised the quantitative associations between O3 exposure and cause-specific mortality risks based on unified exposure metrics.METHODSResearch articles reporting relative risks between incremental long-term O3 exposure and causes of mortality covering all-cause, cardiovascular diseases, respiratory diseases, chronic obstructive pulmonary disease, pneumonia, ischaemic heart diseases, ischaemic stroke, congestive heart failure, cerebrovascular diseases, and lung cancer, estimated from cohort studies were identified through systematic searches in MEDLINE, Embase and Web of Science. Cross-metric conversion factors were estimated linearly by decadal of observations during 1990-2019. DerSimonian and Laird random effect meta-regression was applied to pool the relative risks.RESULTSA total of 20 studies involving 97,766,404 participants were included in the systematic review. After linearly adjusting the inconsistent O3 exposure metrics into congruity, the pooled relative risks (RR) associated with every 10 nmol mol-1 (ppbV) incremental O3 exposure, by mean of warm-season daily maximum 8-hour average metric, was: 1.010 with 95% confidence interval (CI) ranging 1.005–1.015 for all-cause mortality; 1.027 (95% CI: 1.004–1.049) for respiratory mortality; 1.061 (95% CI: 1.006– 1.119) for COPD mortality; 1.028 (95% CI: 1.001–1.058) for cardiovascular mortality; and 1.102 (95% CI: 1.046–1.162) for congestive heart failure mortality. Positive but insignificant mortality risk associations were found for ischaemic heart diseases, stroke, pneumonia, and lung cancer.DISCUSSIONSThis review covered up-to-date studies, expanded the O3-exposure associated mortality causes into wider range of categories, and firstly highlighted the issue of inconsistency in O3 exposure metrics. Non-intercept linear regression-based cross-metric RR conversion was another innovation, but limitation lay in the observation reliance, indicating further calibration with more credible observations available. Large uncertainties in the multi-study pooled RRs would inspire more future studies to corroborate or contradict the results from this review.CONCLUSIONSAdjustment for exposure metrics laid more solid foundation for multi-study meta-analysis, the results of which revealed unneglectable cardiopulmonary hazards from long-term O3 exposure.REGISTRATIONThe review was registered in PROSPERO (CRD42021270637).FUNDINGThis study is mainly funded by UK Natural Environment Research Council, UK National Centre for Atmospheric Science, Australian Research Council and Australian National Health and Medical Research Council.HighlightsUpdated evidence for O3-mortality associations from 20 cohorts has been provided.Adjusted various O3exposure metrics can provide more accurate risk estimations.Long-term O3-exposure was associated with increased mortality from all-causes, respiratory disease, COPD, cardiovascular disease and congestive heart failure.