Long-term influence of oestrogen plus progestin vs oestrogen alone on breast cancer incidence in post-menopausal women

Prof Rowan Chlebowski - City of Hope National Medical Center, Duarte, USA

Thank you and these are my conflicts. The background goes back over half a century - conjugated equine oestrogen introduced into US clinical practice in 1942. After half a century, however, the effects of oestrogen alone or oestrogen plus progestin on breast cancer still remain controversial. I’m citing these two recent Lancet publications, one of which is a meta-analysis of 58 observational studies and the Million Women study which entered 37% of postmenopausal women in the United Kingdom and both suggest that both oestrogen plus progestin and oestrogen alone significantly increase breast cancer incidence and significantly increase deaths from breast cancer. Against this background we’ll present the results from the two Women’s Health Initiative hormone therapy randomised trials. These trials entered patients between 1993 and 1998, women between the ages of 50 and 79 years of age, no prior breast cancer with a mammogram not suggestive of malignancy. Postmenopausal women with prior hysterectomy were randomised to conjugated equine oestrogens or a placebo. Women who had intact uterus at that time it was known that endometrial cancer risk was associated with oestrogen use in women so the progestin blocked that so we used conjugated equine oestrogens plus medroxyprogesterone acetate versus placebo. The primary monitoring outcomes for the study were coronary heart disease for benefit and invasive breast cancer for harm and you’ll see paradoxical results, at least in the oestrogen alone trial.

So this is a randomised placebo-controlled double-blind study implemented at 40 US clinical centres. Annual mammograms were protocol mandated, breast cancers were verified by medical record review and deaths were verified by central death certificate review. Importantly, we used the National Death Index which captures 98-99% of deaths regardless of women’s consent to continue to be followed or dropout. So our mortality findings are really absolutely complete.

Here are some of the baseline characteristics. I’ll call your attention to the fact that over 20% of the women were entered between the ages of 70-79, they’re a little heavier than the population at that time. Currently the US population 40% of postmenopausal women in the US are obese and you can see it was a little higher in the CE alone trial. About half the women had never used hormone therapy previously.
Here’s a summary of the results and what we can see is the solid blue line is the oestrogen alone outcome for breast cancer incidence and the faded blue line is the control group who received placebo. The faded red line is the control group for the oestrogen plus progestin trial and this is the oestrogen plus progestin finding. So you can see that CEE alone after use for 7.2 years now with 19.2 years follow-up resulted in a 23% reduction in breast cancer incidence which was statistically significant whereas the CEE plus MPA use ended up increasing breast cancer by 29%, again statistically significant.

We like these curves because they show each dot is a cumulative hazard ratio, the conference intervals are the vertical bars and the standard deviations are indicated on the right. So it looks like oestrogen plus progestin was reducing breast cancer for the first five years but we have subsequently evaluated and published in peer reviewed journals and JAMA that that actually was a delay in diagnosis interfering with breast cancer mammography. But you can see after five years there is an elevated increase in breast cancer. The use of the drug was only for 5.6 years but you can see the increased use above the level is continuing for up to 20 years. So a woman takes oestrogen plus progestin for five years and she is exposed to a 20 year risk of increasing breast cancer risk and you can see it doesn’t seem to be levelling off. So one could speculate it will be a lifetime risk for short-term use.

Here’s the comparable plot looking at CEE alone and breast cancer incidence. Here you can see this is the incidence goes down quite rapidly and stays down. The use here was for 7.2 years, that’s the green segment. What happens is after that use stopped you can see it went up just a little and it’s staying significantly lower now through 20 years of follow-up. So use of oestrogen alone in this population, again which is an older population than normally used for vasomotor symptom management, results in a statistically significant reduction in breast cancer incidence.

Now we’ll look at some subgroup analysis. I’ll direct your attention to the right hand panel which is the CEE plus MPA trial looking at breast cancer characteristics. You can see regardless of histology, receptor status, that in all categories CEE plus MPA resulted in an increased number of breast cancers. You can see the deaths from breast cancers were increased in the bottom line by 45%, that was of borderline statistical significance, the deaths after breast cancer were statistically significant.

If we move to the right panel we can see interestingly that oestrogen, CEE alone, use significantly decreased by 55% the frequency of oestrogen receptor positive progesterone receptor negative cancers. These are poor risk cancers.

Then we see in the bottom we can see that deaths from cancer were significantly decreased. So we show this in the breast cancer mortality findings expanded here. We look first at the left hand panel – deaths from breast cancer for CEE plus MPA we talked about 45% increase of borderline significance but CEE alone use a 44% reduction in deaths from breast cancer. I should call your attention to the fact that none of the approved agents for breast cancer risk reduction, tamoxifen, raloxifene or the agents that have randomised trial support for their use, aromatase inhibitors, have been able to demonstrate a reduction in deaths from breast cancer. So this is a very unique finding.

Then we look at the deaths after breast cancer, this is deaths from all causes after breast cancer, and it was increased in the CEE plus MPA use but again decreased in the CEE alone use.

So this plot follows a number of things together. The blue line is the year to year influence of oestrogen alone use on breast cancer incidence. The open circle is the year to year use on deaths from breast cancer, you can see they very closely track. Here we’re going to look at the breast cancer incidence on E P which you can see it goes up and stays elevated and then these are the deaths from breast cancer – you can see they’re elevated but many more of the cancers in oestrogen plus progesterone were not associated with mortality risk.

Our conclusions, again summarising these 27,000 women followed for 19.2 years in less than seven minutes, are that use of CEE alone and CEE plus MPA have opposite effects on breast cancers. Pretty clear when you saw that last slide that they are both going in opposite directions. Surprisingly, effects of short term, 7.2 or 5.6 years of use, seem to have effects that are maintained through at least 20 years of follow-up. CEE alone significantly decreases breast cancer incidence and deaths from breast cancer which persist over decades after discontinuing use. CEE plus MPA significantly increases breast cancer incidence and associated mortality which persists over a decade after discontinuing use. These findings, in conjunction with other hormone therapy effects and clinical outcomes, should inform clinical decision making. Thank you.