Abstract Background Most studies have focused on women’s intimate partner violence (IPV) victimization but perpetration and bi-directional violence remain understudied. Methods Using linked criminal justice and population registers in Manitoba, Canada, we assessed the risk of death according to the role in IPV incidents (accused-only, victim-only, bidirectional, none). In this retrospective matched cohort study, we assembled a cohort of 212,068 adults who were followed from April 2004 to March 2023 to assess IPV incidents and subsequent death. Those involved in an IPV incident were 1:3 matched to persons with no history of IPV based on birth year, sex and marital status at the time of the incident. Results Men comprised 85% of accused-only, 21% of victim-only, and 50-51% in the other two groups. Overall, compared to those without IPV involvement, the adjusted Hazard Ratios for all-cause mortality were 1.38 [95% Confidence Intervals (CI): 1.28, 1.49] for accused-only, 1.39 (1.29, 1.50) for victim-only and 1.45 (1.19, 1.77) for bidirectional IPV. The associations were stronger among women, particularly that of bidirectional violence [1.24 (0.97, 1.60) among men and 1.92 (1.39, 2.66) among women]. Similar patterns were found for intentional mortality [0.96 (0.56, 1.64) among men and 2.43 (1.27, 4.65) among women]. Conclusions There are clear sex inequities in IPV involvement. Any type IPV involvement is associated with higher risk of death, particularly among women, who comprise most of the victims. Key messages • Any type IPV involvement is associated with higher risk of death, particularly among women, who comprise most of the victim. • Use of linked criminal justice data allows studying the consequences of IPV victimization, perpetration and bidirectional violence and monitor gender inequities over time.