Bereavement by suicide and other unnatural causes is associated with suicide but evidence regarding risk of substance misuse is inconsistent. This may be due to heterogeneity in patterns of alcohol or drug use after traumatic bereavement; some increasing use to cope with the loss and others reducing use. To highlight the problems of focussing on diagnostic thresholds when investigating substance use after traumatic loss, we aimed to test whether people bereaved by suicide or other unnatural causes are more likely to reduce or stop their substance use than people bereaved by sudden natural causes. Using multivariable logistic regression and data from an online survey of 1854 UK-based bereaved adults, we tested the association between bereavement by suicide and other unnatural causes and post-bereavement reduction/cessation in (i) alcohol and (ii) drug use. There were no group differences in the proportions who reduced/stopped alcohol use, but a significantly greater proportion of people bereaved by sudden unnatural causes reduced/stopped drug use post-bereavement than people bereaved by sudden natural causes (AOR = 2.61; 95% CI = 1.44-4.71; = 0.001; 4.1% versus 1.7%). In sub-group analyses this applied separately to people bereaved by suicide and non-suicide unnatural causes. Research into post-bereavement substance use should accommodate apparent divergent sub-diagnostic patterns.