No significant difference in recurrence rates among women with DCIS taking anastrozole or tamoxifen

Share :
Published: 11 Dec 2015
Views: 2031
Prof Jack Cuzick - Wolfson Institute of Prevention Medicine, London, United Kingdom

Prof Cuzick presents, at a press conference at SABCS 2015 about the results from the phase III IBIS-II DCIS clinical trial. 

The data showed postmenopausal women with ductal carcinoma in situ had similar outcomes with disease recurrence whether they took tamoxifen or the aromatase inhibitor anastrozole for five years after surgery, but women in the two groups had different side effects.

Read the news article.

We know that the third generation aromatase inhibitors are more effective than tamoxifen for treating women with invasive breast cancer but there’s really limited, in fact very little, data for these presentations. Also there’s a simultaneous publication that’s going to occur with this in The Lancet today on the B-35 trial from the NSABP about the role of aromatase inhibitors in DCIS. As I said, the NSABP reported at ASCO, and will now be available, a 27% reduction in recurrences with anastrozole compared to tamoxifen in this population. So that’s really the background.

Our trial was just under 3,000 patients, it was virtually everywhere except North America. 14 countries participated, 236 centres, total recruitment was 2,980 patients. This is for postmenopausal women with locally excised DCIS which was found to be oestrogen receptor positive and we had an amendment which allowed atypical hyperplasia in LCIS but that was very few patients in this study. Median follow-up 7 years.

The results were an indication of really very similar effects on overall recurrences, a slightly better effect for anastrozole, hazard ratio of 0.89, so 11% fewer recurrences overall but this was not significant. If one looks at the subgroups, this is just focussing on the invasive recurrences, we saw some indication that anastrozole might be working better for HER2 negative patients, again not significant, ER positive recurrences were actually more reduced by anastrozole. A subgroup, this was not pre-planned, this was a post-hoc evaluation, if we looked at ER positive HER2 negative tumours we did see a striking effect in that particular group of almost two-thirds reduction in recurrence. So some interest in subgroups but again all of that was post-hoc.

One of the important things which unfortunately isn’t in The Lancet paper, it wasn’t allowed but we will be showing it, is that we did do a simple meta-analysis with the B-35 trial. Our results are shown on the top line there, that 11% reduction; NSABP had a 27% reduction. If you put these two things together you end up with a 21% reduction in all breast cancer events associated with anastrozole in favour over tamoxifen. That is statistically significant if you put those two studies together.

Although, as I said, the overall efficacy differences were small we did see dramatic differences in side effect profiles, in particular other cancers where overall there was not a significant difference but the types of cancers were dramatically different. Gynaecologic cancers, particularly endometrial cancers, there were eleven in the tamoxifen arm and one in the anastrozole arm. That’s probably a combination of two effects, we do know that tamoxifen increases endometrial cancer by two to three fold and there is emerging evidence that the anti-oestrogenic properties of anastrozole actually reduce endometrial cancer. So together really quite a dramatic reduction in endometrial cancer. A somewhat newer finding was a pretty strong effect on ovarian cancer which has been seen, to a small extent, in some of the previous studies looking at anastrozole but very, very big effect here, five versus zero. So that’s interesting. We also saw big reductions in skin cancers, again that’s been suggested earlier, primarily the non-melanoma skin cancers. Other cancers were slightly in the other direction in favour of tamoxifen: colorectal cancer was ten versus five and small differences. So, overall, the effect was not that striking but there were clear differences for gynaecologic cancers and non-melanoma skin cancers.

In terms of fractures, again this is not unsuspected. We saw a higher fracture rate with the aromatase inhibitor, again this is probably a mixture of the fact that this does lower bone density and does increase fracture rates and evidence that tamoxifen actually does reduce fractures. So it’s probably a combination of two things going on there. Thromboembolic events were more common with tamoxifen, again that’s been seen previously, dramatic differences. Overall major thromboembolic events are reduced by about two-thirds with anastrozole compared to tamoxifen, or probably more accurately increased by a factor of three with tamoxifen compared to anastrozole.

Cardiovascular events overall, there were a little bit more on anastrozole, no differences in cardio infarction. A somewhat surprising effect on strokes, cardiovascular accidents, transient ischaemic attack events which were more common in anastrozole. This has not been seen previously, we don’t really have a good explanation for it, something that I think needs to be looked at more fully but it was significantly higher with the anastrozole arm.

In terms of other side effects musculoskeletal aches and pains are known to be more common with aromatase inhibitors and we saw that very, very clearly – 64% reporting that overall with anastrozole, 54% with tamoxifen so about a 10% absolute difference. Arthralgia, joint stiffness, paraesthesia, carpal tunnel syndrome, osteoporosis all more common with anastrozole; muscle spasms very much more common with tamoxifen, again this is a known finding.

On the other side we saw virtually the reverse effects for vasomotor and gynaecologic symptoms – hot flushes or hot flashes here, vaginal bleeding, discharge, all much more common with tamoxifen, again about a 10% absolute difference going in the other direction. Vaginal dryness was the only effect that we saw that was more common with anastrozole, again that’s been reported previously.

Overall that led to virtually identical adherences, really not bad five year compliance or adherence 67% and virtually identical between the two treatments. This is typical of what’s been seen in both the adjuvant studies for women with breast cancer and in the prevention studies, roughly a 65-70% five year adherence to treatment which isn’t too bad, and no difference between the arms.

So, in summary, no significant differences in recurrence between the two drugs in our study but if you look at all of the data from the B-35, the ATAC and from IBIS-2 there is evidence to support lower recurrence rates with anastrozole compared to tamoxifen. No overall effect on other cancers but there were very clear site-specific differences, endometrial cancer, ovarian cancer and skin cancer more common with tamoxifen. Some increase in gastrointestinal, lung and lymphatic cancers with anastrozole but not significant. Really we’ve reported the data on death but there’s really nothing to say yet, it’s too early to look at mortality in this study.

Expected differences in the side effect profiles, more fractures and joint related symptoms with anastrozole, more vasomotor and gynaecologic problems with tamoxifen. This unexpected increase in strokes, cerebral vascular accidents, with anastrozole wasn’t seen in previous studies so I think the significance of that is very unclear. Overall there’s no clear differences in efficacy, some slight favour in terms of anastrozole when you look at all of the data but there are significant differences in toxicity profiles and we feel that anastrozole is now another agent that could be considered for ER positive DCIS. Clearly there are different side effect profiles so this is yet another choice. Women who can’t take tamoxifen, for example because of prior deep vein thrombosis, the AIs are very attractive. If there are particular concerns about fractures then tamoxifen might be better. So this offers yet another choice.

I’d just like to acknowledge all the participants - this was a multinational, multi-centre study with many groups around the world being involved. The results are available online with the B-35 results today in The Lancet so the full results are available in print at this moment. Thank you very much.