Trends in Incidence Rates of Invasive Lobular and Ductal Breast Carcinoma
Abstract & Commentary
The rationale for the present investigation was that a growing number of studies have reported that the risk of breast cancer associated with use of combined hormone replacement therapy (CHRT) differs by histological type. Specifically, 5 separate studies have shown that ever use and current use of CHRT are associated with 2.0-fold to 3.9-fold increased risks of invasive lobular breast carcinoma, the second most common histological type of breast cancer. These same studies show little effect on risk of the most common histological type, invasive ductal breast cancer. Li and colleagues note that these 5 studies also found that use of unopposed estrogen replacemnet therapy (ERT) was not strongly associated with risk of either type. The 2 studies that had sufficient power to examine the relationship between duration of CHRT use and invasive lobular breast cancer found a positive correlation. Li et al also note that invasive ductal breast cancer is more likely to be hormone receptor-positive and to have a better prognosis than invasive ductal breast carcinoma. However, invasive lobular breast cancer is less likely to be detected by mammography and clinical breast examination (Li CI, et al. JAMA. 2003;289:1421-1424).
Given the above considerations, Li et al examined 9 cancer registries that participate in the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute to ask about incidence trends in types of breast cancers. These registries cover the regions of Atlanta, Detroit, San Francisco-Oakland, Seattle, Connecticut, Hawaii, New Mexico, and Utah. A total of 190,458 women were included who were 30 years or older and who had had breast cancer. There was no increase in breast cancer rate from 1987 to 1999 (about 210 cases per 100,000). Rates of lobular carcinoma increased 1.52-fold and those of ductal-lobular 1.96-fold, for a combined increase of 1.65-fold (19.8/100,000 to 33.4/100,000). Rates of invasive ductal breast cancer remained at 154/100,000. Thus, the percentage of invasive breast cancers that was lobular increased from 9.5% to 15.6% for the 12-year interval. The registry does not allow for ascertainment of hormone use, but Li et al note that the data are consistent with the hypothesis that CHRT is associated with an increased risk of invasive lobular breast cancer.
Comment by Sarah L. Berga, MD
I chose to review this study because it illustrates a number of important concepts and because the results suggest ways to refine our prescription of postmenopausal hormone therapy. A methodological strength of the study is that the analysis is based on a very large number of cases of known invasive breast cancer. There were more than 190,000 cases of breast cancer in which the histological type was known! In contrast, the conclusions of the Women’s Health Initiative were based on only 290 cases of invasive breast cancer, and the histological type was not considered in that analysis. The conceptual strength is that the present study is based on the premises that not all breast cancers are the same and that not all hormone regimens are the same. While these caveats might seem intuitively obvious, most studies ask if "hormone use causes breast cancer," and the underlying assumptions are that hormone use is homogeneous and that all breast cancers are the same with regard to causality (and that all women have the same host characteristics). These assumptions have always struck me as unjustified simplifications, but it is nice to have data that reveal the error of this reductionistic question and analytic approach.
The current study design would have been immeasurably strengthened if actual hormone use could have been ascertained, but there are other data in which hormone use has been determined, and these studies also suggest that it is the progestin rather than estrogen component that increases the risk of breast cancer. As Li et al point out in the introduction, in the Women’s Health Initiative, a strength of which is that hormone use is known, combined hormone use was associated with "a statistically nonsignificant 26% increase in risk of breast cancer after 5.2 years." If one assumes that the interpretation suggested by the present study results is valid, namely, that it is combined estrogen-progestin use that increases the risk of lobular breast cancer, then this might well change prescribing practices. One prescribing response would be to give very low doses of unopposed estrogen to women with intact uteri while monitoring the endometrium. Another strategy would be to minimize circulating progestin levels by prescribing the most minimal dose of progestin possible or by confining the progestin exposure to the uterus by inserting a progestin-containing intrauterine device (which gives trivial circulating levels of progestin). Based on extant data, it is reasonable to hope that strategies that minimize or eliminate progestin exposure also might confer better brain (mood and cognition) and cardiovascular outcomes as well.
Questions that are raised by the study include whether the type of progestin matters (Does medroxyprogesterone acetate confer greater risk than norethindrone or progesterone?) and whether there is a difference in risk between cyclic and continuous progestin regimens. The study by Marchbanks et al (New Engl J Med. 2002;346:2025-2032 and reviewed in the August 2002 issue of OB/GYN Clinical Alert) regarding risk of breast cancer among oral contraceptive users found no increase in risk from combined estrogen-progestin exposure, but both the estrogen and the progestin component in oral contraceptives differs from the most commonly used estrogen and progestin preparations used for postmenopausal hormone use during the study period. The present study provides the rationale for additional studies that address these points. Given the paucity of data that address these questions, however, it is too soon to draw conclusions or become dogmatic.
The current study illustrates another key point, which is that the best conclusions are those that reconcile competing data and conclusions by refining both the questions and the answers. In this case, the question "Does hormone use cause breast cancer?" now becomes "Does a particular regimen of postmenopausal hormone use increase the risk of any given type of breast cancer?" This would not be the end of the refinements, however, because one still wants to know about the "host characteristics" that also modify risk. It is a statement of the obvious that not all women are the same, but most epidemiological studies have done little to help us delineate which women might be at higher risk of getting lobular breast cancer from extended combined hormone use.
My final parting shot is to ask where was the news coverage about this article? It was in JAMA, and most of the time articles in JAMA about breast cancer are widely reported. There might be a number of explanations, but I wonder if the conclusions are not simple enough for the newspaper audience. In other words, the findings are too "gray" and we all know that the best news is "black and white."
Dr. Berga is Professor and Director, Division of Reproductive Endocrinology and Infertility, University of Pittsburgh.