January 22, 2015

expert reaction to paper on hormonal contraceptives and brain tumours

A study published in the journal of Clinical Pharmacology has looked at the use of hormonal contraceptives and a link with brain tumours. The researchers report an association between ever using these contraceptives and incidence of glioma, which increase with duration of use, though they recognise potential confounding factors.

Prof. Paul Pharoah, Professor of Cancer Epidemiology, University of Cambridge, said:

“This study has used national data from prescribing and cancer incidence in Denmark to study the association between the use of the oral contraceptive pill and risk of brain cancers. They report a weak association with oral contraceptive pill use being associated with a small increase in risk of brain tumours – a fifty per cent relative increase. It is important to remember that association does not necessarily imply causation. This type of study is particularly prone to bias as it relies on routine data and information on other possible risk factors was not available. It may be that pill use is simply correlated with another important risk factor.

“Moreover, the findings are somewhat contrary to the findings of other studies which have reported either no association or a small reduction in risk of brain tumours associated with oral contraceptive pill use.

“Even if the results are taken at face value the findings have limited implications for women who take the pill. The chance of a woman developing a brain cancer in he lifetime is approximately 5 in 1000.  This would be increased to just 8 in a thousand.

“The oral contraceptive pill has many other beneficial effects. For example a 20 year old woman who takes the oral contraceptive for ten years will reduce her chance of getting ovarian cancer by about 15 in 1000. A much bigger reduction than any possible increase in the risk of brain cancer.

“Women should not base their decision on whether or not to take the contraceptive pill on the results of this study.”

Prof. Sir David Spiegelhalter, Winton Professor of the Public Understanding of Risk, University of Cambridge, said:

“These results are interesting but need to be kept in perspective. Glioma is fortunately very rare. Even if oral contraceptives did increase the risk to the extent suggested by this study, it would only mean one extra glioma each year for every 50,000 women taking the pill.  Suppose, however, all these women changed to a less effective form of contraception and 10,000 of them got pregnant: we would then expect one extra mother and 40 extra babies to die.”

Prof. Kevin McConway, Professor of Applied Statistics, The Open University, said:

“I make no bones about concentrating my comments on the reasons why we shouldn’t worry about these findings. Previous scares about possible increased health risks from contraceptive pills have had bad consequences for public health. On hearing that taking the Pill might lead to an increased risk of certain diseases, some women switched away from the Pill to less reliable forms of contraception, or none at all. That increased numbers of abortions, and exposed the women to health risks from pregnancy and abortion that were generally far greater than any reduction in risk from the diseases in the scare. These are among the standard examples in examining how the public deals with risk; we’ve really got to be careful in reporting such things. That is why Dr Gaist, in the press release, says that “a risk-benefit evaluation would still favour the use of hormonal contraceptives in eligible users.” If someone stops the Pill purely as a result of this new study, they will be putting themselves at greater risk than if they continued.

“Dr Gaist also points out that these brain tumours are pretty rare in women of an age that might be taking the Pill – just 5 per 100,000 women, annually. Given how often we see media stories about young people with brain tumours, this rarity might come as a surprise to some people, but one can’t judge how common a disease is from media stories about individuals. Brain tumours are not at all common. So even if long term pill use does double the risk, well, twice a very small risk is still a very small risk.

“Though this is a carefully conducted and well reported study, it’s therefore particularly important to look at its limitations, and the study report makes these admirably explicit. It is an observational study, so it’s difficult to sort out what might be causing what. The women who had a tumour (the cases) and those who didn’t (the controls) are likely to differ in other ways than their level of use of hormonal contraceptives, and maybe these other differences are the real reason for why some had a glioma and others did not. The researchers can, and did, try to allow for some such differences in their statistical analysis, but they can’t allow for things on which they have no data, and the study report points out several of these. For instance, there is (equivocal) evidence from previous studies that there may be an association between obesity and glioma risk – the researchers could not allow for that because they had no measurement of obesity in the women they studied. The researchers point out several other limitations, and that’s what they conclude only that “long term hormonal contraceptive use MAY increase the risk of glioma.” Or it may not – what they are really calling for is more research.

“But that research won’t be easy, and that again goes back to the fact that these tumours are pretty rare. Despite studying the entire female population of Denmark of reproductive age, that’s well over a million women, over a period of ten years, this study found only 317 gliomas, and that is a fairly small number when it comes to trying to sort out what might be causing what. The researchers did their best with that they had, but it’s the number of cases of the disease that matters in this kind of study, and there just isn’t enough evidence to justify strong conclusions.”

‘Hormonal contraceptive use and risk of glioma among younger women: a nationwide case-control study’ by Andersen et al. published in the British Journal of Clinical Pharmacology on Thursday 22^nd January 2015. 

Declared interests

None declared