Several previous meta-analyses show a consistent inverse association between nut consumption and all-cause mortality, but the associations with cause-specific mortality remain uncertain. A recent meta-analysis on nut consumption and multiple health outcomes combined incidence and mortality outcomes across most of the analyses, which may have introduced heterogeneity across studies. We conducted an updated meta-analysis to evaluate the nut-mortality association. We searched PubMed and EMBASE and we contacted authors for additional data. The final analyses included 18 prospective studies. The random-effects summary RRs for high compared with low nut consumption were 0.81 (95% CI: 0.78-0.84) for all-cause mortality (18 studies with 81 034 deaths), 0.75 (95% CI: 0.71-0.79) for CVD mortality (17 studies with 20 381 deaths), 0.73 (95% CI: 0.67-0.80) for CHD mortality (14 studies with 10 438 deaths), 0.82 (95% CI: 0.73-0.91) for stroke mortality (13 studies with 4850 deaths) and 0.87 (95% CI: 0.80-0.93) for cancer mortality (11 studies 21 353 deaths). These results were broadly consistent within subgroups according to various study and population characteristics and within sensitivity analyses that took into account potential confounders. Peanut (5 studies) and tree nut (3 studies) consumption were similarly associated with mortality risks. Dose-response analyses suggested evidence for nonlinear associations between nut consumption and mortality (P-nonlinearity <0.001 for all outcomes except cancer mortality), with mortality risk levelling off at the consumption of about 3 servings per week (12 g d-1). Our findings suggest that nut consumption is associated with reduced all-cause and cause-specific mortality, with the strongest reduction for CHD mortality. Both tree nuts and peanuts may lower mortality and most of the survival benefits may be achieved at a relative low level of nut consumption.