Abstract
The widely reported news that a daily low-dose aspirin tablet will probably do more harm than good in the healthy elderly will have come as a shock to many of them. For although there is good evidence that aspirin is useful in the secondary prevention of heart attack or stroke, or in subjects at high risk of them, aspirin's value as primary prevention – especially in the elderly – has never been clarified with strong evidence.
At least, not until two study reports in the New England Journal of Medicine in September indicated that not only is the cardiovascular ‘preventive' use of aspirin unlikely to lower risk, but its continued use will actually increase the risk of significant bleeding in the digestive tract and brain. 1 , 2
Both reports were based on the ASPirin in Reducing Events in the Elderly (ASPREE) study, a randomised placebo-controlled trial partially supported by the US National Institutes of Health and performed in almost 20,000 older subjects in Australia and the USA. The study began in 2010 and enrolled healthy participants (male and female) aged 70 and older who were followed up for an average of 4.7 years. At the end of follow-up 90% of the aspirin users (100 mg per day) were still fit and well, as were 90% on placebo. Rates of physical disability were similar, and rates of dementia were almost identical in both groups. But the study also found that aspirin was associated with a significantly increased risk of bleeding, primarily in the gastrointestinal tract and brain.
So what do the results mean for the healthy elderly taking aspirin in the hopes of fending off a cardiovascular event? An editorial in the same issue of the NEJM suggested that aspirin's preventative place – ‘in the post-statin era' – would remain largely in secondary prevention, as indicated a decade or more ago in several huge cohort studies; in the Women's Health Study of 2005, for example, a modest protective effect of aspirin was only found in high-risk (and older patients). 3 , 4 Interestingly in the latter study, aspirin significantly lowered the risk of stroke but had no benefit on MI or cardiovascular death. ‘Cardiovascular events in women,' said the investigators at the time, ‘are more likely to be strokes.'
Another aspirin trial making headlines in 2018 has also put a question mark over primary cardiovascular prevention. In the Aspirin to Reduce Risk of Initial Vascular Events (ARRIVE) trial, men aged 55 or older and women aged 60 or older without diabetes were recruited from primary care in six European countries and the USA, and randomised to 100 mg aspirin or placebo once daily for a median follow-up of five years. But once again, fewer than 5% of participants had an event during the study, with no difference between the two groups.
Among those commenting on the trials and their implications for primary care, Professor Peter Rothwell, Director of the Centre for Prevention of Stroke and Dementia at the University of Oxford described the results as ‘the most reliable evidence' on the risk-benefit balance of starting aspirin after the age of 70 in healthy individuals, and adding that ‘there is no overall benefit in terms of mortality, disability or dementia of starting aspirin and taking it for a period of about five years in healthy individuals aged over 70 years'.
The strongest case for aspirin in primary cardiovascular prevention – albeit with many caveats – came two years ago from the US Preventive Services Task Force (USPSTF) guideline, which recommended that adults in their 50s start low-dose aspirin if they have a 10% or greater 10-year cardiovascular event risk, do not have bleeding risk factors, and are willing to take aspirin for at least 10 years.
In the UK NICE, in its guidance on antiplatelet therapy last revised in 2015, warns against the routine use of antiplatelet treatments for the primary prevention of cardiovascular disease but does recommend use in secondary prevention.
Footnotes
Physical activity (or lack of) not implicated in cause of premature menopause
Female ballet dancers and marathon runners are well known to be at risk of infertility. Intense physical activity has been shown to disrupt regular ovarian cycles (via the hypothalamic–pituitary–ovarian axis), leading to luteal phase insufficiency, anovulation and amenorrhea. However, whether this same mechanism, associated with a disrupted frequency of ovulation and slower diminution of the follicle pool, has an effect on premature menopause is not known.
Now, the ever-active Nurses Health Study, whose first recruits were enrolled in 1976 (and in 1989 for a second cohort, Nurses Health Study II), has analysed this second cohort (of more than 100,000 subjects) to correlate menopausal status with level of exercise. From questionnaire answers dating back to 1989, the analysis identified 2786 study members who experienced menopause before the age of 45. But after adjustment for age, smoking and other factors, no association was found between adulthood physical activity and early menopause. Moreover, neither moderate nor strenuous activity in adolescence and young adulthood were related to risk, which did not vary according to ranges of body mass index or smoking status.
Elizabeth Bertone-Johnson, who directed the research, said Several previous well designed studies have found suggestions that more physical activity is associated with older age at menopause, but even in those studies the size of the effect was very small. Our results, in conjunction with other studies, provide substantial evidence that physical activity is not importantly associated with early menopause.
However, another investigator in the study, Mingfei Zhao, in commenting on the somewhat counterintuitive findings, said: While our results do not suggest that more physical activity is associated with lower risk of early menopause, we would encourage premenopausal women to be physically active, as exercise is associated with a range of health benefits. Our results in no way suggest that premenopausal women should not be physically active.
Healthy lifestyle advice in premature ovarian insufficiency is always emphasised because of its increased risk of lower life expectancy, mainly as a result of cardiovascular risk factors.
The Nurses study identified ‘early' menopause as occurring before the age of 45. This, however, is at some variance with the latest European guidelines, which firstly as a definition favour ‘premature ovarian insufficiency' (POI) over premature or early menopause, and define the threshold as 40 years. 2
A clear explanation for POI remains elusive, with most recent studies suggesting a substantial genetic contribution (maybe 40%). Other known causes have been classified as autoimmune, environmental, and iatrogenic (chemo- and radiotherapy). In her comment on the Nurses study Bertone-Johnson said: Our work has suggested that environmental factors are associated with early menopause. We found higher intake of calcium and vitamin D from dairy foods to be associated with lower risk. Higher intake of vegetable protein was associated with lower risk as well, though animal protein was not. Cigarette smoking is associated with higher risk, as is being underweight. We are currently investigating other factors as well.
Another obesity paradox found in assessing risks for premenopausal breast cancer
Within the complex aetiology of breast cancer, adiposity (usually presented by body mass index) has an ill-defined place. Increased adiposity in childhood and before menopause has been reported to be inversely associated with breast cancer risk, whereas increased adiposity after menopause has been positively associated with risk. Moreover, because incidence rates are lower among premenopausal than postmenopausal women, individual studies have had only limited ability to investigate the association of BMI with any risk of premenopausal breast cancer. Nevertheless, most studies (albeit observational) indicate that increased physical activity at all ages is associated with lower breast cancer risk and better outcomes.
Now, however, a huge study population of 758,592 premenopausal women from 19 prospective cohorts has been analysed by The Premenopausal Breast Cancer Collaborative Group to investigate with some authority the association of BMI with premenopausal breast cancer risk, in particular according to age, risk factors for breast cancer, and tumour characteristics. 1 The group, which included the Institute of Cancer Research in London, pooled data from US, European, Asian and Australian studies, with follow-up ending at breast cancer diagnosis, menopause or age 55 – a mean of 9.3 years.
The results of the study suggest that increased adiposity is associated with a reduced risk of premenopausal breast cancer – and at a greater magnitude than previously shown and across the entire distribution of BMI. The strongest associations of risk were observed for BMI in early adulthood, with hazard ratios per 5 kg/m2 more pronounced in the 18–24 age band than in the 45–54. However, while hazard ratios did not vary appreciably between strata of other breast cancer risk factors, BMI associations were found stronger for estrogen receptor–positive and/or progesterone receptor–positive tumours than for hormone receptor–negative breast cancer at every age group.
Commenting on the results, the investigators write: ‘The stronger inverse associations of risk with BMI at younger than older ages suggest that adiposity in young adulthood or earlier, if adiposity at approximately 20 years of age is a proxy marker for adiposity in childhood, is the critical factor.' Their calculation from the results found an estimated 12–23% reduction in premenopausal breast cancer risk per 5 kg/m2 BMI difference depending on age.
Despite the somewhat surprising findings, the authors make no prophylactic claims for a high BMI before the menopause. ‘Obesity has many adverse effects on general health,' they write, ‘and we do not advocate weight gain as a preventative measure against premenopausal breast cancer. However, understanding the mechanistic action underlying the inverse association of premenopausal adiposity with breast cancer risk could potentially identify modifiable pathways.'