Long-term benefits and risks of hormone replacement therapy

Surveillance proposal

This section of the guideline on long-term benefits and risks of hormone replacement therapy should not be updated.

We identified 66 studies looking at long term risks and benefits of HRT, usually compared with an inactive control (mostly placebo). The studies assessed different types of HRT (oestrogen-only, combined, and tibolone) and varying durations of use. Many abstracts did not include specific details about the type or duration of HRT. Additionally, dosage information was not reported in all abstracts so information on dosage was not considered in this surveillance unless the study specifically compared a single regimen at 2 different doses (1 such study identified).

Coronary heart disease

We identified 2 large cohort studies (60,61) that assessed the risk of coronary heart disease with intravaginal HRT use compared with no HRT (table 11 in the data tables appendix). The results were inconsistent, with one study finding no effect and another finding lower risk of coronary heart disease with HRT. One of the studies found lower risk of coronary heart disease mortality.

Stroke

We identified 7 cohort studies (61–67), 1 Cochrane review (68) and one other systematic review (69) that assessed risk of stroke (ischaemic or haemorrhagic) or stroke mortality (table 12 in the data tables appendix). Studies assessed various types and durations of HRT use, and differing lengths of time since stopping HRT.

The studies showed mixed findings:

6 analyses suggested increased risk of stroke with
- Oestrogen-only HRT compared with placebo (68,69).
- Combined oestrogen (cyclic or continuous) and progestogen HRT compared with placebo or no HRT (65,68).
- HRT compared with no HRT (63,66).
1 analysis suggested increased risk of stroke mortality in the first year after stopping HRT compared with no HRT or current HRT use (67).
5 analyses suggested no effect on stroke (mostly haemorrhagic) with combined or oestrogen-only HRT, including analyses of whether HRT was started up to 5 years after menopause or more than 5 years (62,65,66).
6 analyses suggested lower risk of stroke with oestrogen-only HRT, intravaginal oestrogen, and combined HRT, compared with no HRT, which was analysed by whether HRT was started up to 5 years after menopause or after 5 years (62,64,65).
2 analyses suggested improved stroke mortality with 3–5 years of intravaginal oestrogen and more than a year after stopping HRT (67).

Venous thromboembolism

We identified 2 cohort studies (63,70), 2 Cochrane reviews (10,68), and 1 other systematic review (69) that measured the risk of venous thromboembolism with HRT (table 13 in the data tables appendix). Both oestrogen-only and combined HRT were associated with increased risk of venous thromboembolism. One of the Cochrane reviews (10) suggested no effect of tibolone on risk of venous thromboembolism. However, this result was uncertain, with the confidence intervals indicating that tibolone may be associated with less than half the risk as no HRT but could also nearly double the risk of venous thromboembolism.

Diabetes

We identified one systematic review (69) that suggested a reduced risk of diabetes with either oestrogen-only or combined HRT compared with placebo (table 14 in the data tables appendix).

Other cardiovascular outcomes

We identified 5 cohort studies, 4 RCTs and 2 Cochrane reviews that assessed the effects of HRT on other cardiovascular outcomes (table 15 in the data tables appendix). Preparations of HRT varied across the studies, including oestrogen-only, combined and unspecified HRT. Additionally, studies analysed differing time points, such as stopping in the past year or more than a year ago, having started HRT in the past 3 years or more than 3 years ago. Results suggested possible inconsistent effects of HRT compared with no HRT including:

increased risk of acute coronary syndromes (63)
no effect or increased risk of cardiovascular events or coronary events (10,29,68,71)
no effect on cerebrovascular events (10) but reduced arterial thromboembolic events (72)
inconsistent effects on blood pressure (8,73)
improved blood pressure in one study of combined HRT and increased risk of hypertension in one study of oestrogen alone or combined HRT
improved blood lipid profile (74).

Cardiovascular mortality (61,67,75–77) showed inconsistent results. Across 9 analyses, 4 suggested a lower risk of cardiovascular mortality and 3 suggested no effect. Analyses suggested increased cardiovascular mortality in the first year after stopping HRT but no effects more than a year after stopping HRT.

A Cochrane review (10) suggested that, compared with combined HRT, tibolone had no effect on cardiovascular, cerebrovascular, or thromboembolic events.

Breast cancer

We identified 10 cohort studies (76,78–86), 2 RCTs, 2 Cochrane reviews (10,68) and 1 other systematic review (69) that addressed the risk of breast cancer (table 16 in the data tables appendix).

Lower risks of breast cancer were seen in 3 studies of oestrogen-only HRT or unspecified HRT (that is, no details about the preparation were reported in the abstract) compared with placebo or no HRT (68,87).

No effect on breast cancer risk was seen in 3 studies of oestrogen-only HRT or unspecified HRT compared with no HRT (79,85,86).

Higher risks of breast cancer were seen in 11 studies of HRT compared with placebo or no HRT including:

6 studies of combined HRT (68,69,78,85–87)
4 studies of unspecified HRT (76,79,80,84)
1 study of intrauterine progestogen (83).

Tibolone showed inconsistent effects on breast cancer with 2 analyses showing no association with breast cancer and 2 suggesting increased risk of breast cancer compared with placebo, no HRT, or combined HRT (10,85).

One study assessed the outcomes for women using HRT who were subsequently diagnosed with breast cancer. HRT was associated with lower breast cancer mortality and recurrence compared with not using HRT at diagnosis (82).

Other cancers

We identified 14 cohort studies (76,80,88–98), 5 RCTs (77,84,99–102), 3 Cochrane reviews (10,68,103) and 1 other systematic review (69) that addressed the risk of cancers other than breast cancer with HRT use compared with an inactive control (table 17 in the data tables appendix).

Overall, studies indicated that HRT use was associated with:

a generally consistent reduced risk of gastrointestinal cancers, including colorectal cancers.
an increased risk of ovarian cancer, melanoma and in any cancer
no effect on risk of non-Hodgkin's lymphoma or lung cancer.

Osteoporosis

We identified 2 cohort studies (104,105), 1 RCT (106), 1 Cochrane review (68) and one other systematic review (69) that addressed outcomes related to osteoporosis (table 18 in the data tables appendix). Overall 9 of 10 analyses showed a lower risk of fracture or increased bone mineral density with HRT use. The remaining analysis suggested no effect.

Dementia

We identified 1 cohort study (107), 4 RCTs (20,108–110), 1 Cochrane review (68) and 1 other systematic review (69) that assessed dementia and cognitive outcomes (table 19 in the data tables appendix). Overall, results were inconsistent across the 13 analyses:

6 analyses suggested worse cognitive outcomes with HRT; however, most analyses were of varying measures of cognitive function rather than diagnosis of dementia
6 analyses suggested no effect of HRT; again, most analyses were of varying measures of cognitive function rather than diagnosis of dementia
1 analysis suggested reduced risk of Alzheimer's disease with HRT.

Long-term risks not currently covered in the guideline

We identified 4 cohort studies (75,76,82,111), 1 RCT (77), and 1 Cochrane review (10), that measured the effects of HRT use on mortality (1 study specified the outcome as non-breast cancer mortality) (table 20 in the data tables appendix). In 7 analyses, HRT was associated with lower mortality, and 5 analyses found no effect. There was no indication that the results were dependent on population characteristics or type or duration of HRT use.

We identified 11 cohort studies, 4 RCTs, 2 systematic reviews and 3 Cochrane reviews that addressed other outcomes that were not considered in the guideline (table 21 in the data tables appendix). Results indicated:

HRT was associated with increased risk of faecal incontinence, fibroids, gallbladder disease and gallstones, hearing loss, joint swelling, rheumatoid arthritis, and urinary incontinence (68,69,97,112–116).
HRT was associated with improvements in albuminuria, anxiety, carpal tunnel syndrome, joint pain, lung function, and tinnitus(20,113,117–120).
There may be no association between HRT and sudden sensorineural hearing loss (120).
Inconsistent effects on intraocular pressure were seen, with improvement seen with conjugated equine oestrogen, but no effect seen with combined HRT.(121)

One cohort study (122) compared 2 doses of conjugated equine oestrogen – less than 0.625 mg daily and 0.625 mg daily. Progestogen was also used in both groups. The occurrence of global index events (coronary heart disease, breast cancer, stroke, pulmonary embolism, hip fracture, colorectal cancer, endometrial cancer, or death) was lower with the lower dose. For the higher dose, duration of treatment of 5 or more years was associated with higher rates of global index events than a duration of less than 5 years.

Intelligence gathering

No topic expert feedback was relevant to this section of the guideline.

Impact statement

Coronary heart disease

The guideline considered the effects of HRT on coronary heart disease. Both the guideline and the new evidence found no or reduced risk with oestrogen-only HRT. Therefore, the new evidence is consistent with current recommendations to explain that HRT with oestrogen alone is associated with no, or reduced, risk of coronary heart disease.

Stroke

Evidence identified in developing the guideline found possible increased risk of stroke with combined or oestrogen-only HRT. However, the effects were uncertain. The guideline recommends explaining to women that taking oral (but not transdermal) oestrogen is associated with a small increase in the risk of stroke and that the baseline population risk of stroke in women aged under 60 years is very low.

The new evidence was mixed, with some new evidence indicating an increased risk of stroke with HRT and other studies finding no effect or reduced risk of stroke. Therefore, the uncertain risks of stroke with HRT noted in the guidelines are unlikely to change substantively.

Venous thromboembolism

The guideline recommended explaining to women that oral HRT was associated with an increased risk of venous thromboembolism, but there was no increased risk for transdermal HRT. The new evidence also indicated an increased risk of venous thromboembolism with HRT, and an uncertain effect of tibolone on venous thromboembolism. Therefore, an update in this area is not necessary because the findings are consistent with the guidelines recommendations on oral HRT, and is is unclear whether tibolone has a different risk profile to oral HRT.

Diabetes

The new evidence of reduced risk of diabetes with HRT is consistent with evidence considered in the guideline. However, in developing the guideline, the protective effects of HRT on type 2 diabetes appeared to last only until HRT was stopped. The recommendations therefore noted there to be no increased risk of type 2 diabetes, rather than a reduced risk of diabetes. The new evidence did not inform whether the effect on diabetes continues after stopping HRT, thus no update in this area is needed.

Other cardiovascular outcomes

The guideline recommended that HRT does not increase cardiovascular disease risk when started in women aged under 60 years and does not affect the risk of dying from cardiovascular disease. The new evidence showed inconsistent effects on other cardiovascular outcomes and cardiovascular mortality. The new evidence is thus unlikely to substantively change the guideline's conclusions about risk of cardiovascular disease and mortality.

Overall impact on cardiovascular outcomes

Overall, the new evidence was generally consistent with the guideline's conclusions about cardiovascular risks associated with HRT use. There was no clear indicator that any additional cardiovascular risks need to be considered by the guideline.