Incidence of haemorrhagic stroke increased in the elderly; maybe related to aspirin use

The rate of haemorrhagic stroke in older people has not fallen over the past 25 years, in comparison to younger people in whom there has been a marked reduction: use of antithrombotic drugs seems to be a major factor associated with the difference according to a study published early online by the Lancet Neurology. Previous studies in the UK have concluded that rates of fatal haemorrhagic stroke have fallen for a number of years: this is consistent with better control of hypertension, the major risk factor. These studies have not, however, included people over 75 due to difficulties in reliable death certificate data. As the main causes may differ in this population, data cannot be extrapolated from the younger group. This paper reports an analysis of data from two separate studies of roughly the same population separated by two decades, investigating the incidence of stroke by age and risk factors.

The studies that provided the data were the Oxford Community Stroke Project (OCSP; 1981–86) and the Oxford Vascular Study (OXVASC; 2002–06): these covered substantially the same population, and data for the OSCP could be re-analysed to include the same practices involved in OXVASC. Analysis thus used data on 91,108 individuals from OXVASC and 87,861 from OCSP (both estimated mid-study values); the populations were similar, although the proportion of people aged over 75 increased by about a third with time. Both studies used the same definition for intracerebral haemorrhage, and CT imaging was used in both for the majority of cases (imaging, autopsy or both used in 89% for OCSP and 96% in OXVASC). Incidence rates for all and fatal intracerebral haemorrhages was calculated for both studies, and age specific rates were calculated for above and below 75 years. Incidences were also calculated for moderate to severe hypertension (both uncontrolled and controlled) and use of antithrombotic drugs (low-dose aspirin, clopidogrel, or warfarin).

There were 512 eligible strokes in the OXVASC population, and 557 in OCSP. In both studies, the rate of haemorrhagic stroke was 10% (52 and 55 respectively) - about half were fatal. Comparing the two populations, there was a suggestion of a decrease in rate overall (rate ration 0.72, 95% CI 0.49 to 1.05, p=0.08: this was accounted for by a significant reduction in the rate for people aged under 75 (RR 0.53, 95% CI 0.29 to 0.95; p=0.03), as there was no significant reduction in those aged over 75. There were fewer events associated with hypertension (RR 0.37, 95% CI 0.20 to 0.69; p=0.002), but more associated with antithrombotic use (RR 7.4, 95% CI 1.7 to 32; p=0.007): there was a significant rise in the proportion of people taking antithrombotic drugs before their stroke, from 4% to 41%. About one third of antithrombotic-associated strokes were preceded by warfarin use and two-thirds with aspirin or clopidogrel. In those aged over 75, there was also an increase in bleeds thought to be related to amyloid angiopathy.

The authors conclude that there has been a substantial fall in the incidence of haemorrhagic strokes due to hypertension over the past 25 years, however the incidence has not fallen overall. This is in part associated with antithrombotic use. While drug use would not have been directly causal in all these cases, estimates based on published data suggest that about 20% of haemorrhagic strokes in people over 75 are due to antithrombotic treatment. They note that despite the fall in younger patients, hypertension was still the most common cause in this age group.

This study has been widely reported in the media. Expert commentators note that the figures suggest that the risks from taking antithrombotic medications probably outweighs the benefit in healthy older people, however those prescribed these drugs to prevent stroke due to an underlying medical condition should continue to take them.

[Editor's comment: while the media reports concentrate on aspirin, given that there were probably many more people taking this (or clopidogrel) than warfarin, the absolute risk with warfarin will be significantly greater.]

Lancet Neurology, published early online 1 May 2007; DOI:10.1016/S1474-4422(07)70107-2 (link to abstract);
BBC News report;
the Stroke Association have issued a response