The Women’s Health Study (WHS) was a randomised trial of aspirin 100mg every other day and vitamin E in the primary prevention of cancer and cardiovascular disease (CVD). Previous analyses found no effect of aspirin at this dose on the incidence of cancer, over a mean follow-up of ten years. The participants were invited for further annual follow-up after the active intervention ended in 2004, and the current report presents the results of this (through March 2012).
The WHS included female health professionals aged ≥45 years with no history of cancer, CVD or other major chronic illness. It was conducted entirely by mail (study drugs posted out) and participants were sent questionnaires at 6 months, 12 months and then every year, seeking information on clinical endpoints, adherence and adverse effects. A medical record review was undertaken for all recorded cases of cancer and cardiovascular events, and all relevant information was reviewed by a blinded endpoints committee.
A total of 39,876 women were originally randomised to aspirin or placebo, and 33,682 continued observational follow-up (88.6% of survivors). Among those who opted for this, baseline characteristics did not differ between randomised groups. The median follow-up was 17.5 years.
Although the total cancer incidence was similar for the aspirin and placebo groups, there was an imbalance for colorectal cancer, with a 16-year incidence of 0.011 in the aspirin group and 0.014 in the placebo group (hazard ratio (HR) 0.80; 95% CI 0.67-0.97; p=0.021). Survival curves suggest that the difference emerged after ten years. Rates of self-reported gastrointestinal bleeding (0.085 vs. 0.075) and peptic ulcer (0.077 vs. 0.065) were however higher in the aspirin group (these differences were restricted to the trial period).
For CVD, the reduction in stroke seen in the trial period was reduced and there was no difference in major CVD or CVD death. The authors comment that this attenuation of the randomised effect of aspirin once it was discontinued is consistent with effects on the aggregation of platelets, which have a high turnover.
In the discussion, the authors note recent literature that supports a delayed effect of aspirin on cancer, and they say that their hazard ratio for colorectal cancer is similar to that reported by a meta-analysis which focused on daily aspirin use. The delayed differences may indicate an effect in the early stages of carcinogenesis; one alternative explanation is that increased bleeding leads to more investigation and early polyp removal.
The analysis has a number of limitations that need to be considered when interpreting the results:
• Information on gastrointestinal bleeding, peptic ulcers and polyps collected after the study finished in 2004 was from self-report only
• Participation in post-trial follow-up was not mandatory (those who remained in the study were found to have some differences to those who didn’t, despite there being little difference between the randomised groups)
• Adherence to aspirin throughout the study was imperfect
• Despite the use of analyses to overcome this, post-trial observational aspirin use could still lead to unmeasured confounding
The authors conclude that “whether the new results about long-term benefits for cancer will tip the balance in favour of aspirin [for primary prevention] remains to be determined”.