Deaths Associated With Abortion Compared to Childbirth
A Review of New and Old Data and The Medical and Legal Implications
A Closer Look at the Difference in Rates for Deaths from Homicide
Finland's record linkage study found homicide accounted for five percent of pregnancy-associated deaths between 1987 and 1994.95 Most of these deaths occurred among women who had undergone an abortion. As shown in Figure 5, the risk of dying from homicide for post-abortive women was more than four times greater than the risk of homicide among the general population. Over the eight-year period examined in the mortality study of Medicaid eligible women in California, after controlling for age and psychiatric history, homicide deaths among aborting women were ninety-three percent higher (OR = 1.93, 95% CI = 1.11 to 3.33).96 These findings are consistent with another study of pregnancy-associated deaths in Maryland that showed that homicide is the leading cause of pregnancy-associated deaths.97
Further investigation is necessary to more fully understand the association between abortion and increased risk of death from homicide. Increased risk-taking and substance abuse may play some role in this association, but it is also possible that many of these deaths are related to domestic violence. Women who become pregnant in a violent or potentially violent domestic situation may choose abortion to prevent a child from being born into an abusive situation. If this abortion is secretly obtained against an abusive male's wishes, subsequent discovery or disclosure of the abortion may result in violence and even death. Alternatively, if an abusive male partner is unwilling to accept or tolerate the birth of a child, the woman may become a victim of verbal or physical abuse aimed at compelling her to submit to an unwanted abortion.98
According to one study of battered women, the target of battery during their pregnancies shifted from their face and breasts to their pregnant abdomens,99 which suggests hostility toward the women's fertility. In one study of violent deaths among pregnant women, three out of every four were killed during their first 20 weeks of pregnancy.100
Research indicates that pregnant women are at higher risk of being abused.101 Once the pattern of violence has escalated, it may not naturally recede simply because the woman submits to the unwanted abortion. Negative post-abortion reactions may instead make matters worse.
Figure 5
Whether a woman is covertly or overtly coerced into an unwanted abortion, any post-abortion reaction - on either the part of the woman or man - that includes grief, resentment, or anger on the woman's part may increase the frequency and intensity of subsequent hostility and domestic conflicts.102 This hypothesis is supported by clinical experience with abused women and at least one survey of women participating in post-abortion programs. In the Elliot Institute survey of 260 women, fifty-nine percent agreed with the statement that after their abortion, "I started losing my temper more easily," and forty-eight percent agreed that "I became more violent when angered."103 In this same sample, fifty-six percent reported experiencing suicidal feelings, with twenty-eight percent actually attempting suicide one or more times.104 Approximately thirty-seven percent described themselves as "self-destructive" with another thirteen percent "unsure" (that is, unwilling to rule out that they had become self-destructive).105 Further analysis of this data found that increased post-abortion levels of self-hatred, hatred of the male, and hatred of men in general, were all significantly correlated to each other. In addition, suicidal tendencies and self-destructive behavior were statistically associated with shorter tempers and increased levels of anger and violence (p < .00001). In turn, short tempers and self-destructive behavior were also significantly associated with feeling less in touch with one's emotions, feeling unable to grieve, faking displays of happiness, and feeling less control over one's life. Women who are angry and self-destructive following an abortion may be less inclined to avoid violent confrontations. Some, who may be unable to commit a direct act of suicide, may even gravitate toward abusive males who may do the job for them. According to one post-abortive woman,
One night during a drunken spree, he held a knife to my chest. I told him to kill me, that I wanted to die. I had nothing. No parents, no husband, really, no baby, and no self-respect. How could he respect me? I had killed our child. How could I look at myself in the mirror every day? I was a murderer. I truly wanted to die.106
This and similar self-reports suggest that post-abortion reactions may aggravate or precipitate domestic violence.107
Future investigations of the association between abortion and homicide may require examination of police records and interviews with domestic partners, friends, or relatives. Some cases may be convoluted. For example, in 1999 a Pennsylvania couple, Michael Oravec and Rhonda Jo Reller, allegedly entered into a suicide pact a month after an abortion which resulted in profound regret and depression.108 Oravec survived, subsequently pleaded, and was sentenced for the murder of Reller.109
Overall risk of Death from Violence
As the discussion above suggests, self-destructive tendencies may play an important part in any deaths resulting from violence, including deaths attributed to accidents and homicide. For this reason, a general assessment of all violent deaths associated with pregnancy outcome may also be instructive. Researchers in Finland did such an assessment. During the period examined, deaths from violent causes accounted for fifty-five percent of the deaths among women who had been pregnant in the previous year. Women who gave birth had only forty-seven percent of the risk of death from violence (suicide, accident, or violence) as women who had not been pregnant in the prior year (OR = 0.47; 95% CI = 0.30 to 0.74), while women who had abortions had 181 percent of the risk of death as women who had not been pregnant (OR = 1.81, 95% CI = 1.31 to 2.50).110 The ninety-five percent confidence limits (CI) for the latter indicates that after allowing for chance variance in the sample population, it is ninety-five percent likely that the true odds ratio of death from violence lies somewhere between 1.31 and 2.50 times higher relative risk for women who had an abortion in the prior year compared to non-pregnant women.
In the California study, there was not a comparison group to nonpregnant women. Instead, after controlling for age and prior psychiatric history, the researchers found that over the full eight year period examined, aborting women had a risk of death from violence that was 178 percent of the risk of death for delivering women (OR = 1.78; 95% CI = 1.28 to 2.47).111 Figure 6 represents a graph of the rate of death per 100,000 women by pregnancy outcome for this group. It shows that the disparity in deaths from violent causes is greatest nearest the event of abortion or delivery and rapidly declines over four years, after which differences in risk of death from violent causes are no longer statistically significant. This time-based effect would seem to support the hypothesis that abortion has a causal direct or indirect impact on risk of violent deaths at least within the first four years after an abortion. The decline in risk over four years may be explained by the healing effects of time as women process their grief and overcome selfdestructive tendencies that were caused or aggravated by their abortions.
Figure 6
While more research is clearly warranted, these two large record-based studies have established that abortion is at least a marker, if not a causal factor, for increased risk of death from violence. A causal interpretation is supported by other research, clinical experience, and the self-reports of post-abortive women. The latter is especially important in regard to understanding the causes of death from violence. Together, the preponderance of evidence clearly refutes the claim that mortality rates associated with abortion are lower than those associated with childbirth.
While more research is needed, the potential impact of abortion on deaths from violent causes can be estimated by using the ninety-five percent confidence interval identified in the California study. It is ninety-five percent likely that the true difference in relative risk lies between 1.28 and 2.47, at least among low-income citizens of California. Assuming that this range is similar to all 1.4 million women undergoing abortions each year, we can estimate that between 766 and 4,021 deaths from violent causes each year may be related to or aggravated by a prior abortion.112
A Closer Look at Differences in Rates of Death from Natural Causes
In the Finland study, deaths from natural causes accounted for forty-five percent of the deaths among the recently pregnant women.113 As seen in Figure 7, the age adjusted odds ratio of dying from natural causes within a year following any pregnancy was lower than that of non-pregnant women. This finding suggests that women who are capable of becoming pregnant are simply healthier and less likely to die of natural causes than women who cannot or do not become pregnant. Conversely, the women who are most likely to die from natural causes may be least likely to become pregnant in the last year of their lives.
Comparing abortion to birth, however, the risk of death from natural causes was sixty percent higher for women who had abortions compared to women who gave birth. One possible explanation would be that the women who died after an abortion were already in ill health before the abortions and sought the abortion to protect their health. But the STAKES researchers rejected this hypothesis when an examination of abortion registry records showed that only a single woman in this group had her abortion for reasons of maternal health.114
Similarly, the comparative mortality study using low-income women in California showed that over the eight years following a pregnancy outcome, aborting women were forty-four percent more likely to die from natural causes than women who had delivered (OR = 1.44; 95% CI = 1.08 to 1.91).115
The findings would appear to support the view that induced abortion produces an unnatural physical and psychological stress on women that can result in a negative impact on their general health. This theory is also supported by studies that have examined the amount of health care sought by women before and after induced abortions. In a review of the records of a group general practice in northwest London treating about 10,000 patients, researchers discovered that on average there was as much as an eighty percent increase in requests for health care services in the year following an abortion compared to the year prior to an abortion.116
Figure 7
A number of questionnaire-based studies have also reported an increased rate of health problems post-abortion. Following clinical trials of RU-486 on 145 women, 7.6 percent reported increased health problems two weeks after their medical abortions. This figure rose to 13.8 percent by six to eight weeks post-abortion.117
Another study examined health ratings compiled by 1,428 patients chosen at random from office visits to sixty-nine general practitioners. The validity of these self-assessments were checked against ratings by their physicians and an independent physician's review of patients' medical records. The investigators found that women with a history of pregnancy loss, especially abortion, had significantly lower general health ratings than other women. The more pregnancy losses a woman had suffered, the more negative her general health score. Loss of woman's most recent pregnancy was more strongly associated with lower health than were losses followed by successful deliveries.118 While the researchers found that miscarriage was also associated with a lower health score, induced abortion was more strongly associated with a lower health assessment and more frequently identified by women as the cause of their reduced level of health. More than twenty percent of the women participating in the study expressed a moderate to strong need for professional help to resolve their loss. From these data, the psychiatrist who led the research team concluded that pathological grief after the loss of an unborn child, whether by miscarriage or abortion, has a detrimental effect on the psychological and physical health of women. He proposed several possible reasons for this: (1) aborting women may be hesitant to discuss feelings of loss or to seek professional help because of the moral, familial, and political controversies surrounding abortion; (2) losses that are not mourned may lead to pathological grief which is associated with depression, and depression is associated with a suppression of the immune system,119 increasing the risk of infections and cancers; (3) psychological conflict may consume energy that would otherwise be spent in more healthy ways; and (4) prolonged or unresolved mourning may distract the woman from taking care of other health needs or confuse her interpretation of crisis situations.120
In addition to these factors, a history of abortion has been linked to heightened anxiety,121 sleeping disorders,122 eating disorders,123 and promiscuity,124 all of which can have a direct negative impact on a woman's health. Other unhealthy behaviors that have been linked to abortion are increased alcohol consumption,125 drug abuse,126 and smoking.127 Heavier smoking has been correlated to higher levels of anxiety among women with a history of abortion128 and is a major cause of respiratory diseases and death. Since lung cancer develops slowly, however, one would not expect an association between lung cancer and abortion to be detected in a study examining only eight years of death certificates associated with pregnancy outcome, as was done in the California study. A review of the literature on elevated smoking levels following abortion, however, has concluded that even the lowest estimate of a two percent increased smoking rate following abortion would lead to 4,310 additional cancer cases in the lifetime of the 1.4 million women having an abortion each year-of whom, at current mortality rates for lung cancer, 3,750 would die from this disease.129 If all smoking-related deaths were taken into account, a two percent increase in smoking rates among women who have had abortions would lead to 11,250 additional deaths annually.130
Heart disease is another major cause of death that may be impacted by a history of abortion. In the California study, among women with only one known pregnancy, during the eight years following their pregnancies those who had abortions were nearly three times more likely to die from circulatory diseases (OR = 2.87; 95% CI = 1.68 to 4.89) and over five times more likely to die of cerebrovascular disease (OR = 5.46.; 95% CI = 1.60 to 18.65) compared to women who delivered.131 The impact of abortion on subsequent substance abuse, eating disorders, smoking, and substance abuse may explain part of this finding. Also, since it is known that women who abort are at higher risk of long term clinical depression than women who carry unintended pregnancies to term,132 and depression is an independent risk factor for death from heart disease,133 an association between abortion and heart diseases may be mediated by depression.
Depression is also a risk factor for development of several forms of cancer.134 Breast cancer135 and cervical cancer136 are also significantly associated with abortion and/or delayed childbirth. Women who abort also lose the protective effect of childbirth, which reduces the risk of cancers of the breast, cervix, colon and rectum, ovaries, endometrium, and liver.137 Projecting the increased relative risk for contracting just three of these cancers (breast, ovarian, and endometrium) on mortality rates associated with each type of cancer, as many as 32,000 cancer deaths each year may be attributable to negative effects of abortion on maternal health.138 Record linkage studies examining pregnancy outcomes with death certificates over periods of twenty to forty years will be required to better identify the actual risk.
The California study also found that abortion was significantly associated with an elevated risk of death from AIDS (OR = 2.18; 95% CI = 1.10 to 4.31).139 This finding is consistent with previous research identifying higher rates of HIV-1 infection among women who have abortions compared to those who deliver.140 Abortion may be a contributing factor in AIDS since pelvic inflammatory disease, which is a relatively common complication of abortion,141 may increase the risk of HIV transmission.142 Increased levels of substance abuse and promiscuity following abortion may also contribute to a higher risk of HIV infection and death from AIDS.143
It is also known that induced abortion is associated with a subsequent risk of placenta previa and premature delivery.144 Increased rates of genital tract infection, pelvic inflammatory disease, endometritis, ectopic pregnancy, retained placenta, preeclampsia, and other complications of pregnancy and delivery in subsequent pregnancies have also been identified in the literature.145 All of these complications are associated with higher risk of maternal and neonatal death.146 Even if these deaths are actually traceable to latent abortion morbidity (scarring of the uterus, for example), these deaths would be classified as maternal deaths rather than abortion-related deaths, and would therefore confound the comparison of mortality rates between abortion and delivery.
Abortion is also associated with a subsequent increased need for treatments for mental illness compared to delivery.147 In the California study, after controlling for prior mental illness, researchers found that women who had abortions were three times more likely to die from causes attributed to mental disease than women who carried to term.148
Conclusions
In arriving at the conclusion that abortion's mortality rates are lower than those of childbirth in Roe v. Wade, Justice Blackmun relied on the studies and opinions of population control advocates Christopher Tietze, Malcolm Potts, and Lawrence Lader, all of whom were zealous promoters of liberalized abortion laws.149 The studies they relied on, however, had many methodological problems, including very limited access to patients for follow-up, no control group of delivering women, and lack of an objective standard for comparing mortality rates of delivering and aborting women. The focus of these abortion advocates appeared to be limited to identifying the risk of death from short-term complications of abortion such as septic infection or therapeutic misadventure. But subsequent experience has shown that abortion can have both subtle and profound effects on women's psychological and physical wellbeing. In the 1960s and early 1970s, abortion advocates erroneously believed, without any recourse to supporting data, that the risk of death from suicide after an abortion was negligible while the risk of suicide among women with unintended pregnancies was high. Similarly, in making their estimates of abortion mortality rates, abortion advocates did not consider the impact of abortion morbidity on longevity. Death arising from conditions created or aggravated by abortion complications, such as ectopic pregnancy, pelvic inflammatory disease, depression, and breast cancer, should also be considered in any comparison of mortality rates. Nor did their appraisals recognize nor account for the protective effects of early and frequent childbirth against a variety of cancers and other ailments.
The original comparisons of reported abortion deaths to national maternal mortality rates relied upon in Roe were also flawed by the fact that even the deaths attributable to immediate complications of abortion occurred primarily among healthy women who had little or no risk of death from childbirth. In this sense, they were "extra" deaths. They were not simply unsuccessful attempts to save these women from dying during dangerous pregnancies and deliveries. In fact, there are no studies that have established when, if ever, abortion reduces a woman's risk of death compared to childbirth.150 In one of the few studies undertaken to determine if maternal deaths could have been avoided by abortion, it was concluded that therapeutic abortion would not have prevented any of the twenty-one maternal deaths which occurred among the 74,317 pregnancies examined.151
It is also noteworthy that as many as ninety percent of maternal deaths related to childbirth are associated with caesarean section deliveries which have a maternal mortality rate of approximately 100 per 100,000 c-section deliveries, compared to only 1.1 per 100,000 for vaginal deliveries.152 The hundred-fold higher mortality rate reported following c-section deliveries compared to vaginal deliveries may reflect a combination of the following three factors: (1) women who are ill or faced with higher-risk deliveries are more likely to be delivered by a c-section; (2) a coroner is less likely to miss the fact that a woman who has undergone a recent c-section was recently pregnant; and (3) there are significant surgical risks associated with c-sections and this procedure is arguably overused.153 Since a doctor should normally anticipate that a healthy, young pregnant woman is able have an uncomplicated vaginal delivery, a strong argument can be made that the most relevant comparison for healthy women would be a comparison of mortality rates associated with abortion compared to those associated with a vaginal delivery. Conversely, for women with known health problems, there is not yet any research showing that abortion is less dangerous for these women than childbirth;154 there is only the presumption that this may be true. Clearly, carefully designed casecontrol studies are needed to determine when, if ever, abortion is associated with a reduced mortality risk compared to delivery.
Medical intervention in a health natural process such as pregnancy should only be undertaken when there is clear medical evidence that the treatment produces clearly defined benefits that outweigh any related risks. As David Grimes, M.D., has noted, interventions based on theories that are not substantiated by research place patients at risk of injury: "Uncritical thinking has hurt women and children around the world. Notable examples include diethylstilbestrol, the Dalkon Shield, and bottle feeding. We cannot afford, either economically or ethically, to allow good intentions to dictate practice without the check of scientific controls. Our practices need dispassionate scientific scrutiny."155
While some medical experts will certainly continue to defend the opinion that abortion is a safe alternative to childbirth, this opinion can no longer be characterized as a "now-established fact."156 It is at best an unsubstantiated opinion, most likely a hope, and at worst, an ideological mantra. While "[d]octors often differ in their estimate of comparative health risks and appropriate treatment,"157 responsible differences of opinion must be reconcilable with empirical evidence. In the case at hand, it is clear that prior comparisons of mortality rates associated with abortion and childbirth have been crudely constructed on the basis of an incomplete and inaccurate reporting system. Using the standards developed for evidence-based medicine, the recent record-based case-control studies represent the best available medical evidence on this issue and supercede any "expert opinions" that diverge from this evidence.158
After thirty years of experience with legal abortion in the United States, it is now clear that mortality risks associated with abortion significantly exceed those associated with childbirth, both in the short term (under one year) and in the longer term. While statistical association is not proof of causation, it is clear that abortion is, at the very least, a marker for elevated mortality rates. In the context of the additional studies reviewed in this paper, it is also clear that the interpretation of a causal effect cannot be ruled out. It is therefore reasonable for legislators to conclude that abortion, at any stage of pregnancy, poses a significant risk to women's health. Since Roe established comparative mortality rates as the standard for determining when states can regulate abortion to protect the health interests of women, this new medical evidence would appear to be sufficient to establish a compelling state interest in regulating abortion throughout all stages of pregnancy.
*David C. Reardon, Ph.D., is a biomedical ethicist, the director of the Elliot Institute, and a widely published researcher on the psychological and physical effects of abortion on women.
†Thomas W. Strahan, J.D., died on November 13, 2003. He was a civil rights lawyer who practiced law in Minneapolis, Minnesota and editor of Detrimental Effects of Abortion: An Annotated Bibliography with Commentary.
‡John M. Thorp, Jr., M.D., is the McAllister Distinguished Professor of Obstetrics and Gynecology at the University of North Carolina School of Medicine.
**Martha W. Shuping, M.D., graduated from Wake Forest University Medical School, Winston-Salem, North Carolina, in 1984, and completed her psychiatry residency at North Carolina Baptist Hospital in 1988. She has been in private practice in Winston-Salem since 1991.
Next Page:
Endnotes:
1,
2, 3,
4,
5
Article copyrights are held solely by author.
[ Japan-Lifeissues.net ] [ OMI Japan/Korea ]