OBJECTIVE: To evaluate the frequency of concussion in Swedish ice hockey and to establish a uniform grading and treatment model for concussions of different severity. METHODS: Frequency of concussion was investigated in two studies, one retrospective and one prospective. In the retrospective study, all Swedish elite ice hockey players (n = 265) were asked to answer a questionnaire on the number and treatment of previous concussions. Only concussions diagnosed by a doctor were recorded. The questionnaire was completed by 227 players (86%). In the prospective study, all injuries including concussions occurring during game and practice in the Swedish Elite League (n = 12 teams) were recorded during four years. The causes of injury, referees judgements, diagnosis, treatment, and time absent from ice hockey were registered on special cards. RESULTS: In the retrospective study, 51 out of 227 players (22%) in the Swedish Elite League reported at least one concussion. In the prospective study, 52 concussions were reported. The incidence of a concussion is at least one concussion every year/team or a yearly risk of about 5% for a player to sustain a concussion. Most concussions occurred during league play (81%). Body contact (checking or boarding) was the most common cause of concussions. The players were absent from full training and play on a mean of 6 d. CONCLUSIONS: As this injury is potentially dangerous it must be treated seriously according to a simple treatment model presented. In cases of repeated concussions during the same season, a longer period of time away from play is suggested. In players who have sustained several concussions over the years a thorough medical examination including EEG, CT/MRI, and neuropsychological tests should be performed. If any of these is pathological the player should be advised to give up ice hockey.