Background An alcohol-induced memory blackout represents an amnesia to recall events but does not involve a loss of consciousness. Memory blackouts are a common occurrence among college drinkers, but it is not clear if a history of memory blackouts is predictive of future alcohol-related injury above and beyond the risk associated with heavy drinking episodes.

Objective To determine whether baseline memory blackouts can prospectively identify college students with alcohol-related injury in the next 24 months after controlling for heavy drinking days.

Methods Data were analysed from the College Health Intervention Project Study (CHIPS), a randomised controlled trial of screening and brief physician intervention for problem alcohol use among 796 undergraduate and 158 graduate students at four university sites in the USA and one in Canada, conducted from 2004 to 2009. Multivariate analyses used generalised estimating equations with the logit link.

Results The overall 24-month alcohol-related injury rate was 25.6%, with no significant difference between men and women (p=0.51). Alcohol-induced memory blackouts at baseline exhibited a significant dose–response on odds of alcohol-related injury during follow-up, increasing from 1.57 (95% CI 1.13 to 2.19) for subjects reporting 1–2 memory blackouts at baseline to 2.64 (95% CI 1.65 to 4.21) for students acknowledging 6+ memory blackouts at baseline. The link between memory blackouts and injury was mediated by younger age, prior alcohol-related injury, heavy drinking, and sensation-seeking disposition.

Conclusions Memory blackouts are a significant predictor of future alcohol-related injury among college drinkers after adjusting for heavy drinking episodes.