Objective: A first study to compare suicides by missing persons with other suicide cases. Design: Retrospective cohort study for the period 1994–2007. Geographical location: Queensland, Australia. Population: 194 suicides by missing persons and 7545 other suicides were identified through the Queensland Suicide Register and the National Coroners Information System. Main outcome measure: χ2 statistics and binary logistic regression were used to identify distinct characteristics of suicides by missing persons. Results: Compared with other suicide cases, missing persons significantly more often died by motor vehicle exhaust gas toxicity (23.7% vs 16.4%; χ2=7.32, p<0.01), jumping from height (6.7% vs 3.2%; χ2=7.08, p<0.01) or drowning (8.2% vs 1.8%; χ2=39.53, p<0.01), but less frequently by hanging (29.4% vs 39.9%; χ2=8.82, p<0.01). They were most frequently located in natural outdoors locations (58.2% vs 11.1%; χ2=388.25, p<0.01). Persons gone missing were less likely to have lived alone at time of death (OR 0.45, 95% CI 0.26 to 0.76), yet more likely to be institutionalised (OR 3.12, 95% CI 1.28 to 7.64). They were less likely to have been physically ill (OR 0.64, 95% CI 0.43 to 0.95) or have a history of problematic consumptions of alcohol (OR 0.52, 95% CI 0.31 to 0.87). In comparison to other suicide cases, missing persons more often communicated their suicidal intent prior to death (OR 1.58, 95% CI 1.13 to 2.22). Conclusions: Suicides by missing persons show several distinct characteristics in comparisons to other suicides. The findings have implications for development of suicide prevention strategies focusing on early identification and interventions targeting this group. In particular, it may offer assistance to police in designing risk assessment procedures and subsequent investigations of missing persons. ;