|Association between hormone replacement therapy and subsequent stroke: a meta-analysis
|Bath PM, Gray LJ
This review assessed the risk of stroke following the use of hormone replacement therapy (HRT). The authors concluded that HRT could not be recommended for the primary or secondary prevention of stroke. Some caution is required when interpreting the results of this review as there was considerable clinical variation across the included studies.
To assess the risk of stroke following the use of hormone replacement therapy (HRT).
MEDLINE, EMBASE and the Cochrane Library were searched (up to 2004) for publications in the English language. Previous reviews (see Other Publications of Related Interest nos.1-4) and the reference lists of retrieved articles were also checked for relevant studies.
Study designs of evaluations included in the review
Completed randomised controlled trials (RCTs) were eligible for inclusion. The length of follow-up ranged from 0.7 to 10.0 years.
Specific interventions included in the review
Studies of HRT were eligible for inclusion. The included studies assessed oestrogen alone (conjugated equine oestrogen or oestradiol) and in combination with progesterone (desogestrel, medroxyprogesterone acetate, norgestrel, gestogene, cyproterone acetate, or northisterone acetate). Compliance with treatment ranged from 46.2% to 98%.
Participants included in the review
Studies of men or women receiving HRT were eligible for inclusion. The indication for treatment varied across the individual studies and included 'healthy' patients as well as those with cerebrovascular disease, coronary disease, those who had undergone a hysterectomy and other conditions. The average age ranged from 55 to 71 years.
Outcomes assessed in the review
Studies that reported cerebrovascular events were eligible for inclusion. In the included studies patients had experienced a fatal or nonfatal ischaemic or haemorrhagic stroke. Studies that did not make a distinction between stroke and transient ischaemic attack were excluded from the review.
How were decisions on the relevance of primary studies made?
The authors did not state how the papers were selected for the review, or how many reviewers performed the selection.
Assessment of study quality
The authors assigned a score to each study (1 being the lowest and 5 the highest) to assess the quality of randomisation, blinding, reporting of withdrawals, generation of random numbers and concealment of allocation. The authors did not state how many reviewers performed the quality assessment.
Two reviewers independently extracted the data from the included studies. Any discrepancies were resolved by discussion. Data on the occurrence of stroke events (fatal and nonfatal), type of stroke (ischaemic, haemorrhagic, or unknown) and the functional outcome (combined death and disability or dependency) were extracted from the individual studies and used to calculate an odds ratio (OR). Data on the number of transient ischaemic attacks were extracted, where reported. Data were extracted using an intention-to-treat format, where possible.
Methods of synthesis
How were the studies combined?
The results from the individual studies were combined using a random-effects meta-analysis. A pooled OR and 95% confidence interval (CI) were calculated separately for stroke, type of stroke (ischaemic, haemorrhagic), outcome (fatal, nonfatal, death or dependence) and transient ischaemic attack. Publication bias was assessed using the method of Egger.
How were differences between studies investigated?
The chi-squared test was used to assess heterogeneity. Sensitivity analyses were performed on the effect of HRT on total stroke for the following prognostic variables: phase of stroke prevention (primary or secondary), type of HRT (oestrogen alone or oestrogen plus progesterone), type of oestrogen (oestradiol or conjugated equine oestrogen), size of trial (less than or equal to 5,000 or more than 5,000 patients), length of follow-up (less than or equal to 3 years or more than 3 years), gender and trial quality (high or low, i.e. score of 5 versus less than 5). Interactions between subgroups and treatment effect were assessed.
Results of the review
Twenty-eight RCTs (n=39,769) were included in the review.
Fifteen studies scored 5 for methodological quality, six scored 4, three scored 3 and four scored 2.
HRT was associated with a significant increase in the likelihood of stroke (OR 1.29, 95% CI: 1.13, 1.47) compared with control, based on participants in 28 studies. It was estimated that for every 147 patients treated with HRT, one would experience a stroke.
HRT was associated with a significant increase in the likelihood of nonfatal stroke (OR 1.23, 95% CI: 1.06, 1.44) and a slight increase in the likelihood of fatal stroke (OR 1.28, 95% CI: 0.87, 1.88), although this was not statistically significant.
HRT was associated with a significant increase in the likelihood of poor outcome (combined death or disability) after stroke (OR 1.56, 95% CI: 1.11, 2.20). The likelihood of ischaemic stroke was increased in patients treated with HRT (OR 1.29, 95% CI: 1.06, 1.56), whereas the likelihood of haemorrhagic stroke (OR 1.07, 95% CI: 0.65, 1.75) or transient ischaemic attack (OR 1.02, 95% CI: 0.78, 1.34) was not significantly different between the treatment arms.
No evidence of statistical heterogeneity was found for stroke (P=0.56) or any of the subgroups. None of the prognostic factors assessed in the sensitivity analyses were found to influence the outcome of 'total stroke' and no evidence of publication bias was found (Egger's test, P=0.19).
HRT could not be recommended for the primary or secondary prevention of stroke.
The review addressed a clear research question and the inclusion criteria appeared appropriate. Several relevant sources were searched for English publications only; therefore, the potential for language bias could not be ruled out. Unpublished studies were not sought explicitly, although the authors did not find any evidence of publication bias among the studies included in the analysis. The methods used to select studies for inclusion were not reported, thus the possibility of selection bias could not be ruled out. However, methods were used to minimise reviewer error and bias in the abstraction of data.
There was considerable clinical differences between the individual studies. It may not, therefore, have been appropriate to combine the studies included in the review. The authors assessed statistical differences across the studies and investigated several prognostic factors that might have influenced the effect of HRT on stroke. However, the influence of patient age was not considered. Furthermore, two individual studies contributed the majority of the patients to the review. A detailed consideration of how clinically similar the patients and treatments used in these trials were in comparison with other trials, and how this might have influenced the findings of the review, would have been appropriate.
In conclusion, although this was a relatively well-conducted systematic review, the authors' conclusions were based on a diverse group of patients and different HRT treatments. Further consideration of which patients are at increased risk of stroke following HRT is required.
Implications of the review for practice and research
Practice: The authors stated that HRT could not be recommended for the primary or secondary prevention of stroke.
Research: The authors did not state any implications for further research.
Bath PM, Gray LJ. Association between hormone replacement therapy and subsequent stroke: a meta-analysis. BMJ 2005; 330: 342
Other publications of related interest
1. Wren BG. Megatrials of hormonal replacement therapy. Drugs Aging 1998;12:343-8. 2. Zec RF, Trivedi MA. Effects of hormone replacement therapy on cognitive aging and dementia risk in postmenopausal women: a review of ongoing large-scale, long-term clinical trials. Climacteric 2002;5:122-34. 3. Collins P. Clinical cardiovascular studies of hormone replacement therapy. Am J Cardiol 2002;90 Suppl:30-4F. 4. Salpeter SR, Walsh JM, Greyber E, Ormiston TM, Salpeter EE. Mortality associated with hormone replacement therapy in younger and older women. J Gen Intern Med 2004;19:791-804.
Subject indexing assigned by NLM
Aged; Data Collection /standards; Female; Hormone Replacement Therapy /adverse effects; Humans; Middle Aged; Postmenopause; Prognosis; Randomized Controlled Trials as Topic /standards; Risk Factors; Stroke /chemically induced /pathology /prevention & control
Database entry date
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.