Cardiovascular events 

Results: 

44% of 1154 respondents indicated that the impact of HRT on increasing or decreasing the risk of cardiovascular events was probably known. 

48% of respondents indicated that regarding cardiovascular events, HRT was likely to be neutral. 

There was a significant positive correlation between respondent’s level of certainty and their estimate of benefit.  (Spearman’s rho = 0.26  p<0.0001)  Those who felt more certain were more likely to believe that HRT had a beneficial effect on cardiovascular event rates.

Our comment: 

The impact of HRT on cardiovascular events is very uncertain but may be very harmful, in the short term but neutral for survivors in the medium term.

Explanation:

The hypothesis raised by observational studies that HRT would have cardiovascular benefits has been overturned by recent RCT results.[1]

We have identified four sources of RCT data.  Ann has had a myocardial infarction so the first 2 studies listed below are most relevant for her.  The last 2 are more relevant for women who do not have coronary disease.  We are concerned that Australian doctors have not been adequately informed about these studies.

1.  The HERS trial (August 1998) compared 4.1 years of oestrogen plus progestogen (n=1380) vs placebo (n=1383) for postmenopausal US women aged 55 to 79 with pre-existing coronary disease. [2]  The primary research question for the trial was:  Does HRT reduce cardiovascular events (coronary deaths plus non-fatal myocardial infarction)?

There was no significant difference in the rate of coronary deaths plus non-fatal myocardial infarction at 4.1 years. (12.7% in the placebo group vs 12.5% in the HRT group, p = 0.99) 

However there were significantly more coronary deaths plus non-fatal myocardial infarcts in the HRT group at 1 year. (2.7% in the placebo group vs 4.1% in the HRT group, p= 0.046), but this difference became non-significant after the first year. 

The graph below shows a trend for more non-fatal myocardial infarcts initially in the HRT group but that trend disappeared after the third year.  Please note that after 4.1 years the numbers of women followed drops to 113 so the suggestion of a late trend for benefit is not reliable. 

Incidence of non-fatal myocardial infarction in the HERS trial

___ Oestrogen-Progestogen                  ----- Placebo

The graph below shows a trend for more coronary deaths in the HRT group that did not disappear with time.  One interpretation of the two graphs is that some women on HRT had fatal infarcts rather than the non-fatal infarcts they would have had on placebo.

Incidence of coronary heart disease death in the HERS trial

___ Oestrogen-Progestogen                  ----- Placebo

2.  The ERA trial (August 2000) compared 3.2 years of treatment with oestrogen plus progestogen (n=104) vs oestrogen alone (n=100) vs placebo (n=105) for postmenopausal US women aged 42 to 80 with pre-existing coronary disease. [3] 

There was no significant difference in the progression of coronary atherosclerosis.

3.  The Hemminki and McPherson study (February 2000) pooled data from 28 RCTs that compared HRT (n=2206) vs a comparitor (n=1278) for up to 3 years.[4]  The aim was to support or question the reduced rate of cardiovascular events on HRT (an odds ratio of 0.7) seen in observational studies.[5] 

The pooled data showed an increased rate of cardiovascular events on HRT (odds ratio 1.78 95% confidence interval 0.7-4.52) that was not statistically significant.  This could be because the difference was only due to chance or because of small numbers due to inadequate reporting of adverse events.  However the findings do suggest that the chance of the odds ratio of 0.7 seen in observational studies being correct is very small (3 in 100). 

This result suggests that the findings of the observational studies arise from confounding eg the “healthy woman effect”.

4.  The ongoing Women’s Health Initiative trial is comparing oestrogen plus progestogen or oestrogen alone for women without a uterus (n>13,550) vs placebo (n>13,550) for healthy postmenopausal US women aged 55 to 79. 

In April 2000 the investigators informed the participants that “during the first two years there was a small increase in the number of heart attacks, strokes, and blood clots in women taking active hormones compared to inactive (placebo) pills.  Over time, these differences seem to get smaller…”[6] 

Events occurred in “less than 1%” of women and the difference did not reach statistical significance[7] so it may be only due to chance.  However in healthy women, if the difference is real, then even infrequent harm may be clinically important if it is severe eg heart attacks, strokes and pulmonary emboli.

Conclusions

All of the RCT evidence is consistent with initial cardiovascular harm during the first one or two years.  The harm is not frequent but is severe.  There is still uncertainty but it appears that harm is more likely for those like Ann who are at higher risk already.  The available evidence suggests that there is no cardiovascular benefit during three to four years of HRT.  Any benefit after that would have to be very large to make up for the initial harm.

Next: Deep venous thrombosis and pulmonary embolism