Clot risk from HRT pills and patches

“HRT pills DOUBLE the risk of getting blood clots – but patches are safer, say experts,” reads the headline in the Daily Mail today. It reports on research that has investigated the risk of blood clots from two different forms of hormone replacement therapy (HRT). The newspaper says, “One million women are currently taking HRT, with an estimated three out of four using pills.”

The findings come from a review of 17 studies of women taking HRT. It is already well known that HRT is a risk factor for blood clots, but this new study provides valuable evidence about the size of the risk and gives some idea of the difference between using patches or pills. However, women taking HRT should not be overly alarmed; the actual risk is still relatively small. It is too early to conclude that patches are safer than pills as there were a smaller number of studies which looked at their use. Much further research, preferably randomised trials of patches compared with pills, will be needed to confirm whether HRT patches are safer.

Where did the story come from?

Marianne Canonico and colleagues from Inserm Cardiovascular Epidemiology Section and Université Paris-Sud, Villejuif Cedex, France and the University of Glasgow carried out this research. Individual researchers received funding from Inserm, Assistance Publique des Hopitaux de Paris and the University of Glasgow. It was published in the peer-reviewed British Medical Journal .

What kind of scientific study was this?

This was a systematic review and meta-analysis in which the researchers combined results from several studies; some were observational studies and some randomised controlled trials. These studies looked at the risk of venous thromboembolism (a blood clot in the vein, either in the location that it was formed – thrombosis – or that has travelled to another vein in the body – embolism) in women taking hormone replacement therapy.

The researchers carried out a search of the electronic database Medline, for all English language studies published between 1974 and 2007 that included any keywords relating to HRT (e.g. oestrogen replacement or oestrogen therapy) in combination with those relating to venous thromboembolism. Each of the identified observational or experimental study types was assessed for study quality. If the study was considered suitable for inclusion, the researchers collected the relevant information on the type of HRT used (e.g. the type of hormones used, route of administration and duration of treatment) and the characteristics of venous thromboembolism (e.g. deep vein thrombosis or pulmonary embolism, how it was diagnosed, whether there was another suspected cause).

The data from the observational studies and the randomised trials was combined separately and the researchers carried out statistical tests to see whether there were any significant differences between the methods and results of the individual studies that may affect the reliability of the combined results.

What were the results of the study?

The initial search identified 1,890 articles which were filtered to give a final seven case-control studies (four of which involved HRT patches as well as tablets), nine randomised controlled trials and one cohort study. All of the studies were considered to be of high quality, with most of them investigating a first episode of venous thromboembolism that did not have any identified provoking risk factors (idiopathic).

All individual studies apart from one found a consistent trend for increased risk of venous thromboembolism with use of HRT. The combined results of the eight observational studies (the case-control studies and the cohort study) found that oral HRT significantly increased risk of thromboembolism by 2.5 times compared with placebo. The nine randomised controlled trials also found significant risk from oral HRT, but the size of the risk was slightly less, at 2.1 times. However, the four observational studies investigating HRT given via a patch found that, although there was still a trend for increased risk versus placebo, this was not significant.

The researchers then carried out a separate analysis of the trials to look at other characteristics of HRT use that may affect risk. They found that previous use of HRT did not significantly increase risk compared with first time users. There was no difference in the size of risk whether oestrogen was used alone or in combination with progestogen. However, the length of therapy seemed to have an effect, with use of HRT for less than one year significantly increasing the risk by four times, compared with a double increased risk for women who used HRT for more than one year. They also found that was an even greater risk if women had an additional genetic condition that increases their blood’s tendency to clot, or if they were overweight.

What interpretations did the researchers draw from these results?

The researchers conclude that “current use of oral oestrogen increases the risk of venous thromboembolism by twofold to threefold” and this may be even greater during the first year of use or in women with other risk factors. They say that HRT given via a patch may be safer but that further research is needed.

What does the NHS Knowledge Service make of this study?

Hormone therapy is already well-recognised as one of the risk factors for venous blood clots, but this new study provides valuable evidence about the size of the risks and gives some idea of the difference between patches and pills. However, there are several limitations, which should be considered:

It should not be assumed from this research that it is unsafe to take HRT in pill form while patches are safe. Only four observational studies followed women using HRT patches while eight observational studies and nine randomised controlled trials – the most reliable research method - investigated oral HRT. Although the combined results of the four observational studies of HRT patches did not find significantly increased risk of venous thromboembolism, many more studies, ideally randomised controlled trials, will be needed to confirm that this is the case.
The HRT used in the studies differed in terms of the type of oestrogen used, the dose and whether or not it was combined with a progestogen hormone (although the researchers did not find this that this affected the risk). Trials were also of different lengths and used different populations of women, e.g. postmenopausal women with a healthy uterus or women who had undergone hysterectomy. These things may all affect venous thromboembolism risk. In addition, it is unclear what other risk factors women may have had (besides weight and the clotting disorders, which the researchers considered) and whether these differed between trials.
The actual size of the risk from oral HRT remains small. The researchers say that while one thromboembolism could be expected in 1,000 woman of this age over a year, an additional 1.5 would be expected to be seen in 1,000 women taking oral HRT for one year. These absolute risks are comparable to those seen in other studies of oral HRT but cannot be compared with any risks calculated for women using patches because no significant increase in risk was demonstrated for HRT patches.
Only one electronic database was used and while Medline is a reliable source citing a large quantity of published research, some studies may have been missed that could have been identified through other search methods.
Not all of the studies were identified to study venous thromboembolism as a primary outcome; in several it was a secondary outcome as part of studies that were designed to investigate the incidence of other things, e.g. coronary heart disease or stroke. The use of secondary outcomes for meta-analysis may also affect the reliability of the results.

An accompanying editorial in the British Medical Journal suggests that while waiting for the results of further trials, healthy menopausal women aged 50–59 should be reassured that the risk of thromboembolism when taking HRT is low and that the risks may be lower with lower doses of hormones. Women with previous venous thromboembolism or a mutation affecting prothrombin should be offered alternatives to oestrogen.