Objectives: To evaluate the nature, prevalence, and severity of chronic memory deficit in patients resuscitated after cardiac arrest outside hospital and to determine whether such deficits are related to duration of cardiac arrest., Design: Case-control study., Subjects: 35 survivors of cardiac arrest outside hospital and 35 controls matched for age and sex who had had acute myocardial infarction without cardiac arrest., Main Outcome Measures: Subjects assessed at least two months after index event for affective state (hospital anxiety and depression scale), premorbid intelligence (national adult reading test), short term recall (digit recall test), and episodic long term memory (Rivermead behavioural memory test)., Results: Cases and controls showed no difference in short term recall. Cases scored lower on Rivermead test than controls (mean (SD) score out of 24 points: 17.4 (5.4) v 21.8 (2.0), P < 0.001), particularly in subtests relating to verbal and spatial memory. Moderate or severe impairment was found in 37% of cases and in no controls. Severity of impairment of memory correlated significantly with measures of duration of cardiac arrest. This deficit was not significantly associated with subjects' age, interval from index event to assessment, occupation, measures of comorbidity, social deprivation, anxiety or depression scores, or estimated premorbid intelligence., Conclusions: Clinically important impairment of memory was common after cardiac arrest outside hospital. Improvement in response times of emergency services could reduce the severity of such deficits. With an increasing numbers of people expected to survive cardiac arrest outside hospital, rehabilitation of those with memory deficit merits specific attention.