Objectives, To assess the relationship between haematocrit and risk of
stroke.Design. Prospective study of a cohort of men followed up for 9
.5 years. Setting. General practices in 24 towns in England, Scotland
and Wales (British Regional Heart Study). Subjects. A total of 7735 me
n aged 40-59 years at screening, selected at random from one general p
ractice in each of 24 towns. Main outcome measures. Fatal and non-fata
l strokes. Results. During a follow-up period of 9.5 years for all men
there were 123 stroke events (33 fatal) in the 7346 men in whom the h
aematocrit level had been determined. In the cohort as a whole, risk o
f stroke was significantly raised at haematocrit levels greater than o
r equal to 51% (relative risk [RR] = 2.5; 95% confidence intervals [CI
] 1.2-5.0) after adjustment for age, social class, smelting, body mass
index, physical activity, presence of diabetes and pre-existing ischa
emic heart disease. Further adjustment for systolic blood pressure did
not attenuate this association (RR = 2.4; 95% CI 1.2-4.9). A raised h
aematocrit was associated with an increase of stroke only in men with
hypertension (systolic blood pressure greater than or equal to 160 mmH
g or diastolic blood pressure greater than or equal to 90 mmHg or on r
egular antihypertensive treatment). No increased risk of stroke was se
en at the higher haematocrit level (greater than or equal to 51%) in n
ormotensive men. At haematocrit levels below 51%, hypertension was ass
ociated with a three-fold increase in risk of stroke compared with nor
motension (RR = 3.4, 95% CI 2.3, 5.1). At haematocrit levels greater t
han or equal to 51%, hypertension was associated with a nine-fold incr
ease in risk of stroke compared with normotension (RR = 9.3; 95% CI 4.
2, 21.0). Exclusion of men receiving regular antihypertensive therapy
did not alter the strong associations seen. Conclusion. The data sugge
st that an elevated haematocrit is an independent risk factor for stro
ke and that it interacts synergistically with elevated blood pressure.