National Cancer Institute Fact Sheet Analysis
The Abortion – Breast Cancer Connection (ABC Link)
National Physicians Center for Family Resources (NPC)
• Overall tone of denial of abc link: “…it appears that there is no overall association between spontaneous or induced abortion and breast cancer risk, …”, even though, to date, 28 out of 37 studies worldwide and 13 out of 15 in the US report a positive association.
• Confusion of induced and spontaneous abortion: These two terms appear together repeatedly, as if they are equivalent. Never is the proper equation of spontaneous abortion and miscarriage made; in fact, the word miscarriage never appears. Yet paradoxically, the “inability to separate induced from spontaneous abortions” is offered as a criticism of earlier studies on the abc link.
• Misrepresentation of the published medical literature on the abc link: a) A key study on American women which relied on prospective medical records (Howe et al., 1989), and which found a significant abc link (overall RR = 1.9), is not cited at all, even though much more weight is given to data “from studies that collected data on abortion history before the breast cancers occurred”, and even though it is presented as a weakness that “Most of the early studies relied on self-reports of induced abortion”.b)
The study specifically funded by the NCI to examine the abc link (Daling et al., 1994), which study reported a significant overall link (RR = 1.5) and much higher risks for teenagers (RR = 2.5) and actually reported “RR = ∞” for teenagers with any family history of breast cancer, is not cited at all.c) The only comprehensive review and meta-analysis on the abc link (Brind et al., 1996), which reported a positive association in 18 out of 23 studies worldwide (9 out of 10 in the US), is not cited at all.
• Reliance on flawed studies which do not show an ABC link, merely because they are based on prospective data, namely:
a) the study by Melbye et al., 1997, even though it misclassified 60,000 abortion-positive women as abortion negative, used breast cancer records which antedated abortion records, and included over 350,000 women under age 25, among other flaws;
b) the study by Goldacre et al., 2001, even though it misclassified over 90% of the abortion-positive women in the study as abortion-negative;
c) the null studies by Newcomb and Mandelson (2000) and Lazovich (2000), even though both are so small (23 and 26 patients with induced abortion, respectively) and of such low statistical power that neither could even detect a RR as low as 1.5 with statistical significance.
• Inclusion of inaccurate statements, i.e.:
a) “In three of the (four) studies, information was based on medical records rather than on the woman’s self-report;”. In fact, this is true of only two (i.e., Goldacre and Newcomb & Mandelson) of the four studies referred to.
b) “The strength of this study (Melbye 1997) include … the ability to account for breast cancer risk factors that may differ between those women who have had abortions and those who have not,”. In fact, the lack of such data on potential confounders was a weakness of the Melbye study, which only adjusted for age and age at first term pregnancy. Most studies also adjust for age at menarche, age at menopause, etc.
c) “Most of the early studies necessarily relied on self-reports of induced abortion, which have been shown to differ between breast cancer patients and other women.” In fact, the opposite is true. Even the only study cited on the fact sheet which examined this question reported: “The authors’ data do not suggest that controls are more reluctant to report a history of induced abortion than are women with breast cancer.”
• Disguising the established breast cancer risk factor that is directly affected by abortion in a substantial proportion of abortion patients, i.e.: “Well established breast cancer risk factors include … a late age at the time of the first birth of a full-term baby”. Abortion, which, in childless girls and women, necessarily delays the first full-term pregnancy, is not mentioned at all in this context.