Correspondence

Induced Abortion and the Risk of Breast Cancer

N Engl J Med 1997; 336:1834-1835June 19, 1997DOI: 10.1056/NEJM199706193362514

Article

To the Editor:

Whereas most studies have reported positive findings,1 Melbye et al. (Jan. 9 issue)2 reported an overall null association between induced abortion and breast cancer, on the basis of their record-linkage analysis of all Danish women born between 1935 and 1978. Although their prospective, computerized data base is, by its nature, free of possible recall bias, their study is nonetheless methodologically flawed.

Since the study encompasses such a wide age range, women who had induced abortions are concentrated in the younger end of the total cohort, resulting in considerably less average follow-up time for them than for women without induced abortions (9.6 vs. 20.7 years). Such a statistical selection bias can mask a significant relative risk, a difficulty that could have been avoided by birth-cohort matching for the subcohorts of women with induced abortions and without induced abortions.

Instead, Melbye et al. compounded this error in attempting to correct for it, by adjusting for the birth cohort. Ewertz and Duffy3 recently described a marked effect of the birth cohort (in five-year intervals) on the incidence of breast cancer in these same Danish women, an effect they admittedly could not adequately explain (although they did not consider abortion). The pattern of change in exposure to induced abortion over time in the Danish population4 is strikingly similar to that in the age-adjusted incidence of breast cancer,3 suggesting that induced abortion may be at least partly responsible for the observed cohort effect. However, by adjusting their crude overall rate ratio of 1.44 for the effect of birth cohort, Melbye et al. factored out any effect of the very factor they were investigating, thus obtaining a null result (relative risk = 1.00).

The authors' understated admission that they “might have obtained an incomplete history of induced abortions for some of the oldest women in the cohort” is misleading. Induced abortion has been legal (and on record) in Denmark for reasons other than medical necessity since 1939 and was only most recently liberalized in 1973.4 Since Melbye et al. used abortion data only from 1973 onward, it is certain that more than 30,000 women in the study cohort who had abortions were misclassified as having had no abortions,4 another source of error tending toward an underestimation of the relative risk.

Therefore, contrary to Hartge's5 reassuring endorsement of the “clear central finding that there is no overall risk,” in his editorial accompanying the report by Melbye et al., we believe that a proper analysis of the Danish cohort data will instead confirm a significant, positive overall association between induced abortion and breast cancer.

Joel Brind, Ph.D.
Baruch College–City University of New York, New York, NY 10010

Vernon M. Chinchilli, Ph.D.
Pennsylvania State University College of Medicine, Hershey, PA 17033

5 References

1. 1

   Brind J, Chinchilli VM, Severs WB, Summy-Long J. Induced abortion as an independent risk factor for breast cancer: comprehensive review and meta-analysis. J Epidemiol Community Health 1996;50:481-496
   CrossRef | Web of Science | Medline

2. 2

   Melbye M, Wohlfahrt J, Olsen JH, et al. Induced abortion and the risk of breast cancer. N Engl J Med 1997;336:81-85
   Free Full Text | Web of Science | Medline

3. 3

   Ewertz M, Duffy SW. Incidence of female breast cancer in relation to prevalence of risk factors in Denmark. Int J Cancer 1994;56:783-787
   CrossRef | Web of Science | Medline

4. 4

   Osler M, David HP, Morgall J, Rasmussen NK. Family planning services delivery: Danish experience. Dan Med Bull 1990;37:95-105
   Web of Science | Medline

5. 5

   Hartge P. Abortion, breast cancer, and epidemiology. N Engl J Med 1997;336:127-128
   Full Text | Web of Science | Medline

To the Editor:

We do not understand why Melbye et al. did not include in the Results section of their abstract the most interesting and important result: in women who had induced abortions after 18 weeks of gestation, the relative risk of breast cancer was 1.89 (95 percent confidence interval, 1.11 to 3.22).

The authors do report this result in Table 1 and in one line in the Results section of the text, and they discuss it in one paragraph in the Discussion section. In that section they say, “The fact that such an increase did not affect the overall result clearly indicates that it is based on small numbers and therefore requires cautious interpretation.” Although it is true that the number of women who had abortions after 18 weeks of gestation is smaller than the numbers in most of the other gestational-age categories, 14,000 patient-years is hardly a small number, even if only 14 breast cancers were found, and the 95 percent confidence intervals speak for themselves.

Shouldn't it be emphasized that women who have an abortion after the 18th week of gestation probably have an increased risk of breast cancer? These data may also explain the different results in previous reports, especially those that do not give detailed information on the week of gestation at the time of abortion.

Richard E. Senghas, M.D.
Michael F. Dolan, M.D.
558 South Ave., Weston, MA 02193

Author/Editor Response

The authors reply:

To the Editor: Brind and Chinchilli describe the prospective design used in our study as free of bias but postulate that the analysis is methodologically flawed. We find their argument self-contradictory and based on fundamental misconceptions about the cohort design.

They claim that a selection bias is introduced because the average follow-up time for women with induced abortions is shorter than that for women without induced abortions. Such an objection can stem only from lack of insight into the design and analysis of a cohort study. For each woman entering the cohort, we calculated the follow-up time (person-years) and allocated this follow-up time according to the abortion history. The calculation of breast-cancer rates (cases per person-years) thus takes into account differences in follow-up time for women with abortions and women without abortions. This is a fundamental feature of the cohort design.1,2

Brind and Chinchilli advocate the use of birth-cohort matching. In our study, we take birth-cohort differences into account, although more efficiently than with the method they propose, by simultaneous adjustment for age and calendar period.1,2 We find it self-contradictory that they later argue against taking birth-cohort differences into account. Their own calculation of the crude relative risk of 1.44 (which we do not mention in our report) is next to meaningless, considering the strong confounding effect of age. With adjustment for age, the relative risk is reduced to 0.99 (95 percent confidence interval, 0.93 to 1.04), showing that there is very little effect of other potentially confounding variables (including birth cohort). It is a surprise to us that Brind and Chinchilli put any emphasis on an ecologic observation from Denmark, considering that we have a much more powerful methodology for addressing this question (i.e., a large cohort study with high-quality individual data).

As we noted in our article, computerized information exists on induced abortions since 1973. If the absence of a history of induced abortion before then had caused a substantial underestimation of the relative risk, we would have expected higher relative risks among younger women, because they would have been less likely to have had a previous abortion and would have been minimally misclassified. However, the adjusted relative risks (and 95 percent confidence intervals) for induced abortion according to the age of the women in 1973 are as follows: 35 years or older, 1.09 (0.93 to 1.27); 30 to 34 years, 1.04 (0.94 to 1.16); 25 to 29 years, 0.92 (0.83 to 1.02); 20 to 24 years, 0.97 (0.84 to 1.13); and less than 20 years, 1.06 (0.87 to 1.29). In view of these data, an important influence of underestimation due to misclassification is unlikely.

Senghas and Dolan argue that we should have emphasized the result for women with induced abortions after 18 weeks of gestation. Although we found this result interesting and in line with the hypothesis of Russo and Russo,3 the small number of cases of cancer in women in this category of gestational age prompted us not to overstate the finding.

Mads Melbye, M.D.
Jan Wohlfahrt, M.Sc.
Per Kragh Andersen, Ph.D.
Danish Epidemiology Science Center, DK-2300 Copenhagen S, Denmark

3 References

1. 1

   Clayton D, Hills M. Statistical models in epidemiology. Oxford, England: Oxford University Press, 1993.
   

2. 2

   Breslow NE, Day NE. Statistical methods in cancer research. Vol. 2. The design and analysis of cohort studies. Lyon, France: International Agency for Research on Cancer, 1987. (IARC scientific publications no. 82.)
   

3. 3

   Russo J, Russo IH. Susceptibility of the mammary gland to carcinogenesis. II. Pregnancy interruption as a risk factor in tumor incidence. Am J Pathol 1980;100:497-512
   Web of Science | Medline

Citing Articles (3)

Citing Articles

1. 1

   Tim Davidson. (2001) Abortion and breast cancer: a hard decision made harder. The Lancet Oncology 2:12, 756-758
   

2. 2

   Mads Melbye, Jan Wohlfahrt, Per Kragh Andersen. (2000) Induced Abortion and Risk of Breast Cancer. Epidemiology 11:2, 235
   

3. 3

   Mary K. Anglin. (1998) Dismantling the master's house: Cancer activists, discourses of prevention, and environmental justice. Identities 5:2, 183-217