National trends in stroke hospitalisations, mortality, and recurrent stroke in Australia and New Zealand

Robust data on trends in stroke hospitalisation rate and outcomes are needed to improve stroke care, yet are sparse and show conflicting results. We investigated the trends in hospitalisation rate, survival, and risk of stroke recurrence following an acute stroke using population-wide data from Australia and New Zealand over a 10-year period. We included adults aged ≥ 18 years hospitalised with an acute stroke from 2008 to 17. Age- and sex-standardized trends of stroke hospitalisation were calculated. Flexible parametric survival models were used to estimate and assess trends in survival and cumulative incidence of stroke recurrence at 30 days, 1 year and 5 years. 331,016 patients were included (mean age 73.2 ± 14.5 years, 48.0% female), with ischaemic stroke (185,800, 56.1%) being the most common type, followed by haemorrhagic stroke (81,877, 24.7%) and unspecified stroke (63,339, 19.1%). The overall age- and sex-standardised stroke hospitalisation rate decreased by 13.5% (179.3/100,000 person-years in 2008 to 155.1/100,000 person-years in 2017, P for trend < 0.01). However, the decline was limited to those ≥ 65 years, with the hospitalisation rate rising in those aged 55-64 years and stagnant in those aged 18-54 years. The adjusted survival probability at 30-day (83.4% in 2008/2009 to 85.9% in 2016/2017), 1-year (70.5% to 74.3%), and 5-year (51.4% to 53.0%) following a stroke has improved over time (P for trend < 0.01). The adjusted risk of stroke recurrence at 30 days and 5 years was unchanged over time, whereas the 1-year risk increased from 9.8% in 2008-2009 to 10.8% in 2016 (P for trend < 0.01). In patients hospitalised for acute stroke in Australia and New Zealand, the overall stroke hospitalisation rate and mortality declined over 10 years. However, this was accompanied by a static or rising stroke hospitalisation rate in younger patients and a concerning lack of improvement in the risk of stroke recurrence.