New oral contraceptives and breast cancer

a recent article, highly publicized in the press, found a significant risk of breast cancer in women on contemporary (ie lower dose) hormonal contraception (see DOI: 10.1056/NEJMoa1700732). study was industry-supported.

Details:

--nation-wide prospective cohort study of all women in Denmark between 15-49 yo, who had not had cancer or venous thromboembolism.

--1.8 million  women followed average of 10.8 years

--fully adjusted models (below) included: level of education, parity, polycystic ovary syndrome, endometriosis, family history of breast/ovarian cancer; and BMI, smoking status, and age at first delivery when women were parous (these last 3 were available only in women who had a kid)

Results:

--11,517 cases of breast cancer occurred overall

--relative risk of breast cancer:

    --current or recent hormonal contraception users: 20% increase, RR 1.20 (1.14-1.26)

        --users for <1 year: 9% increase, RR 1.09 (0.96-1.23), nonsignificant

        --users for >10 years: 38% increase, RR 1.38 (1.26-1.51), p=0.002

--after discontinuation of hormonal contraception:

    --users of >5 yrs were at greater risk than nonusers, though this was only significant in the subgroup of women who used hormonal contraception for 5-10 years (where there was a 33% increase if stopped <1 yr, then 16% if stopped 1-5 years since recent use, and 30% if stopped 5-10 years since recent use), and in the group using hormonal contraceptives >10yrs (where there was a 52% increase if <1 yr since recent use, a nonsignificant increase if 1-5 yrs since recent use, and insufficient data for 5-10 years after stopping).  And, no increase in breast cancer in women who used hormonal contraception for <5 years. 

--women with current or recent use of progestin-only intrauterine devices (vs never-users): 21% increase, RR 1.21 (1.11-1.33)

--overall adjusted absolute risk of breast cancer among current and recent users of any hormonal contraceptive: 13 per 100,000 person-years, or about 1 extra breast cancer for every 7690  women using hormonal contraception for 1 year. specifically:

    --used < 1yr: 9% increased risk, 3 breast cancers per 100,000 person-yrs

    ​--used 1-5yr: 18% increased risk, 9 breast cancers per 100,000 person-yrs

    ​--used 5-10yr: 24% increased risk, 14 breast cancers per 100,000 person-yrs

    ​--used >10yr: 38% increased risk, 19 breast cancers per 100,000 person-yrs​

--no significant difference if take triphasic vs monophasic oral contraceptives

--similar breast cancer risk in estrogen/progestin combinations if 50mcg estrogen vs 20-40 mcg

--in breakdown by progestins in the oral contraceptives (controlling for estrogen dosage), breast cancer risk was elevated only using gestodene and if used >5 years.

--in patients using the progestin-only IUD (503,441 person-yrs), there was a 21% increased risk, similar to those on oral levonorgestrel alone (10,541 person-yrs)​. Very few breast cancer events in those on progestin-only implants (42,217 person-yrs) and depot medroxyprogesterone acetate injections (19,308 person-yrs). 

--subgroup analyses: not much difference if include women who had used hormonal contraception for >5 years before this analysis, though:

    --women <35 yo on combo meds and on IUD had more risk (overall RR 24% for current or recent use)

    --for women starting at age 15 at entry to this study, RR 1.99 if current or recent use (though 2,761,715 person-yrs and 89 cases)

Commentary:

--a 2013 report to the United Nations found that 140 million women use hormonal contraception.  in Denmark, the % of women has increased from 1995 to 2012, from 24% to 39%

--the role of estrogen as a breast cancer promoter has been clear in many studies. though the role of progesterone is equivocal, the addition of progestins to estrogens in postmenopausal women seems to increase the breast cancer risk. [there are a paucity of data on progestin-only therapies overall and especially with the newer progestins such as desogestrel, gestodene or drospirenone. Gestodene is included in femodene, femodette, gynera, harmonet, meliane, minesse, minulet; i believe that none of these are available in the US at this time)​.

-- which brings me to the side issue of: it seems to be inappropriately easy for regulating agencies (FDA, EMA, etc) to approve new drugs based on variations on what has already been approved.  In this case, gestodene, one of the new-fangled potent progestins with some androgenic and antimineralocorticoid effects. And, yet again, sometimes newer is not better.  and sometimes these drugs are a lot worse (also happens with medical devices, which have even less scrutiny than drugs). it is a tad scary and dangerous how unregulated drugs become available, are aggressively marketed, and become popular in clinical medicine....

--absolute increases in breast cancer in the Danish study were small. A metaanalysis of observational studies from 2000-2012 found that combined oral contraceptives (older generation) were associated with an 8% increase in breast cancer, but a 27% decrease in ovarian cancer, 43% decrease in endometrial cancer and a 14% decrease in colorectal cancer (see Bassuk SS. Ann Epidemiol 2015;25: 193). so, even with the increase in breast cancers, there may be an overall protective effect. but important to remember that this not the same population being assessed as in the Danish study, so hard to draw firm conclusions. different potential risk factors in different populations

--this Danish study raised concern about residual increased risk in those who used hormonal contraception for at least 5 years who had stopped for at least 5 years. longer studies are needed to assess this more accurately (there were limited data on those stopping 10 years before, so the cohort needs to be followed for many more years to know)

--study limitations: large database, though missing data on age at menarche, breast-feeding, alcohol, physical activity. also, no data on women who might have been on hormonal contraceptives before the study, who were then off them and would therefore be classified as no hormonal contraceptives. also much of the increased risk in breast cancer from a short study like this one will be missed, since most cases occur well beyond the limited time of this study  when women are older (though this study does showcase the public health advantage of most other industrialized countries, where (unlike the US) there are high-quality, extensive  health databases including demographics, medications, diseases, comorbidities, mortality and other important clinical outcomes.)

--as a perspective issue, the 20% increased breast cancer risk in this study is actually not much different from large studies in the past with older hormonal contraceptives (eg the Nurses' Health Study)

--it is pretty unexpected that the progesterone IUDs did have such a significant increase in breast cancer risk:

    --there is only small systemic absorption, most of the hormonal effects are local paracrine effects, but about 5% of women do get acne, weight gain, depression, decreased libido, and ovarian cysts (see Beatty MN. Ther CLin Risk Manag 2009; 5: 561), and some women do get systemic levels similar to those using levonorgestrel-only pills

    --and, there did not seem to be any increased breast cancer risk with hormonal contraception that consistently increases progestin levels (implants or injected depot medroxyprogesterone acetate)

    ​--and, a 21% increase in an observational study (even such a large one) often is not really significant when repeated if a rigorously controlled RCT were done (which probably will never happen). as mentioned in prior blogs, there is frequently no really significant association in observational studies unless one finds relative risks more in the 2x range or higher.

--this all raises the utility of nonhormonal IUDs, eg the copper-containing ones. these copper IUDs would presumably not increase risk of breast cancer (i have not seen studies on this. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2763153/ does support copper IUDs in breast cancer patients, though without data), though there are the disadvantages of heavier, longer and more uncomfortable menses (especially early after insertion), there is no protection against upper genital tract infection (seems to be more risk of PID than the hormonal one, since the progestin may decrease risk), and there may be a higher failure rate than the progestin-releasing one. so, nonhormonal ones tend to be used less frequently. 

so, a few things as a perspective on this study:

--though there might be a small increase in absolute incidence of breast cancer, birth control pills also seem to decrease ovarian, endometrial and colon cancers, with likely net benefit

--and, i really think that the current push to prescribing long-acting reversible contraceptives LARCs (eg IUDs) still really makes a lot of sense, given the relative lack of systemic hormonal effects, despite this study. (see the caveats noted above). and, AAP and ACOG recommend IUDs even in teens, and AAP states these should be first-line contraceptives (in 2014)

--so, my general sense is that this article, though given wide play in the literature, should not really affect current practice. patients should be aware that even these new hormonal contraceptives with lower dose hormones can increase the risk of breast cancer modestly and may not be any better than the old contraceptives (given the general tendency to think newer is better and safer).  though i would personally avoid using any with gestodene....