BACKGROUND: The aim was to compare characteristics and outcome after cardiac arrest where cardiopulmonary resuscitation was attempted outside and inside hospital over 12 years. METHODS: All out-of-hospital cardiac arrests (OHCAs) in Göteborg between 1994 and 2006 and all in-hospital cardiac arrests (IHCAs) in 1 of the city's 2 hospitals for whom the rescue team was called between 1994 and 2006 were included in the survey. RESULTS: The study included 2,984 cases of OHCA and 1,478 cases of IHCA. Patients with OHCA differed from those with an IHCA; they were younger, included fewer women, were less frequently found in ventricular fibrillation, and were treated later. If patients were found in a shockable rhythm, survival to 1 month/discharge was 18% after OHCA and 61% after IHCA (P < .0001). Corresponding values for a nonshockable rhythm were 3% and 21% (P < .0001). Survival was higher on daytime and weekdays as compared with nighttime and weekends after IHCA but not after OHCA. Among patients found in a shockable rhythm, a multivariate analysis considering age, gender, witnessed status, delay to defibrillation, time of day, day of week, and location showed that IHCA was associated with increased survival compared with OHCA (adjusted odds ratio 3.18, 95% CI 2.07-4.88). CONCLUSION: Compared with OHCA, the survival of patients with IHCA increased 3-fold for shockable rhythm and 7-fold for nonshockable rhythm in our practice setting. If patients were found in a shockable rhythm, the higher survival after IHCA was only partly explained by a shorter treatment delay. The time and day of CA were associated with survival in IHCA but not OHCA.