The incidence of all stroke and stroke subtype in the United Kingdom, 1985 to 2008: A systematic review
Bhatnagar P., Scarborough P., Smeeton NC., Allender S.
Background. There is considerable geographic variation in stroke mortality around the United Kingdom (UK). Whether this is due to geographical differences in incidence or case-fatality is unclear. We conducted a systematic review of high-quality studies documenting the incidence of any stroke and stroke subtypes, between 1985 and 2008 in the UK. We aimed to study geographic and temporal trends in relation to equivalent mortality trends. Methods. MEDLINE and EMBASE were searched, reference lists inspected and authors of included papers were contacted. All rates were standardised to the European Standard Population for those over 45, and between 45 and 74 years. Stroke mortality rates for the included areas were then calculated to produce rate ratios of stroke mortality to incidence for each location. Results. Five papers were included in this review. Geographic variation was narrow but incidence appeared to largely mirror mortality rates for all stroke. For men over 45, incidence (and confidence intervals) per 100,000 ranged from 124 (109-141) in South London, to 185 (164-208) in Scotland. For men, premature (45-74 years) stroke incidence per 100,000 ranged from 79 (67-94) in the North West, to 112 (95-132) in Scotland. Stroke subtype data was more geographically restricted, but did suggest there is no sizeable variation in incidence by subtype around the country. Only one paper, based in South London, had data on temporal trends. This showed that there has been a decline in stroke incidence since the mid 1990 s. This could not be compared to any other locations in this review. Conclusions. Geographic variations in stroke incidence appear to mirror variations in mortality rates. This suggests policies to reduce inequalities in stroke mortality should be directed at risk factor profiles rather than treatment after a first incident event. More high quality stroke incidence data from around the UK are needed before this can be confirmed. © 2010 Bhatnagar et al; licensee BioMed Central Ltd.