BACKGROUND: It is uncertain if improvements in long-term cardiovascular (CV) mortality have occurred in both men and women with ischemic and non-ischemic forms of heart failure (HF). METHODS: The Western Australia Hospital Morbidity Database was used to identify all index (first-ever) hospitalizations for HF between 1990 and 2005. Patients were followed until death attributed to cardiovascular causes or censored on December 31, 2006 to determine 5-year survival. Cox proportional hazards models were used to compare the adjusted mortality hazard ratio (HR) during the study follow-up (4-year periods). RESULTS: A total of 21,507 patients (mean age 73.9 years, 49.1% women) were identified. Women were significantly older than men, and less likely to have ischemic HF (38.8% versus 46.1%). Over the period, age-standardized incidence of first HF hospitalization declined but with the least decline in women with non-ischemic HF (-13.3%) compared to other subgroups. Risk-adjusted 5-year CV mortality declined over the study period, with HR 0.64 (95% CI 0.60-0.68) for patients admitted in 1998-2001 compared to 1990-1993, with significant improvement in both forms of HF, and in both sexes and across age groups. However, overall total HF hospitalizations increased (+26.7%) over the period, particularly for non-ischemic HF (+43.7%), of which elderly women formed the predominant group. CONCLUSIONS: Risk-adjusted long-term survival improved similarly in men and women, including the elderly, with ischemic and non-ischemic forms of HF during 1990-2005 in Western Australia. However, there was a growing burden of HF hospitalizations particularly for HF of non-ischemic aetiology. Copyright © 2011 Elsevier Ireland Ltd. All rights reserved.