There has been a controversy in the last 15 years on the correct management of brain stem cavernomas. We have reviewed our experience of the last 10 years in a single Institution and reviewed related literature published in the last 15 years. We recorded the demographics, clinical presentation, rebleeding episodes, incidence of neurological events and outcome assessed by recording the change of the modified Rankin scale in 21 cases. Univariate analysis was applied to test the effect of demographics, and presentation on the incidence and timing of rebleeding, chance of having a new neurological event, the number of subsequent neurological events and outcomes. Six cases were treated with surgery and 15 cases were managed conservatively. We obtained follow-up data in 20 patients (95%). Mean follow-up period was 79.7 months (range: 6-244, median 70 months). There were 0.05 rebleeding events per patient-year and 0.1 episodes of neurological deterioration per patient-year. No mortality was noted in either the surgical or the non-surgical group. Three of the six surgical cases had a reoperation. The outcome was improved in one patient, unchanged in 1, and worse in 3 surgical patients. In the case of conservative management the outcome was improved in two patients, unchanged in five patients, and worse in eight patients. Outcome was worse in the case of multiple cavernomas (p = 0.012). Our findings suggest that conservative management may be appropriate in individual cases when compared with surgery, but this difference was not statistically significant enough in order to support a change in practice. The natural history of brain stem cavernomas appears more benign than previously thought.