Induced Abortion: Epidemiologic Aspects
Stanley K. Henshaw
Table Of Contents
Stanley K. Henshaw
THE GLOBAL PICTURE
ABORTION POLICY IN THE UNITED STATES
INCIDENCE OF LEGAL ABORTION
PERIOD OF GESTATION AND METHOD OF ABORTION
Induced abortion occurs at least occasionally in all societies and is common in those with low fertility. Even during the 19th century when abortion was illegal, its incidence in the United States and parts of Europe was high enough to be a factor in the declining fertility rates. It is only within the past 40 years, however, that abortion in Western countries has changed from being predominantly illegal and unsafe to being one of the most commonly performed and intensely studied surgical procedures in obstetric and gynecologic practice.
The two sources of national epidemiologic data on legal abortions in the United States are the Centers for Disease Control and Prevention (CDC) in Atlanta and The Alan Guttmacher Institute in New York City. The CDC publishes an annual survey on abortion based mainly on reports by state health authorities.1 This publication contains information on the characteristics of women obtaining abortions, including age, race, parity, marital status, number of previous induced abortions, weeks of gestation, and method of abortion. The CDC has also monitored abortion-related deaths since 1972. The Alan Guttmacher Institute has conducted periodic surveys of providers of abortion services that have identified larger numbers of abortions than the CDC and have provided information about abortion providers.2 To obtain a complete picture of legal abortion in the United States, it is necessary to generate estimates based on a combination of CDC and Alan Guttmacher Institute data. Such estimates have been used throughout this chapter.
|THE GLOBAL PICTURE|
Among the countries of the world, the legal status of induced abortion ranges from complete prohibition to elective abortion at the request of the pregnant woman.3 The situation as of the end of 2000 can be summarized as follows. Approximately one fourth of the world's 6 billion people lived in countries where abortion was prohibited without exception or where it was permitted only to save the life of the pregnant woman. These included most of the Muslim countries of Asia, almost two thirds of the countries of Latin America, a majority of the countries of Africa, and one country in Europe (Ireland). Approximately one tenth lived under statutes authorizing abortion on broader medical grounds, such as to avert a threat to the woman's physical health rather than to her life and sometimes on eugenic, or fetal, indication (known genetic or other impairment of the fetus or increased risk of such impairment) or juridical indication (e.g., rape, incest) as well.
A fourth of the world's population resided in countries where abortion is permitted to protect a woman's mental health as well as her physical health; where social factors, such as inadequate income, substandard housing, and unmarried status, could be taken into consideration in the evaluation of the threat to the woman's health (social-medical indication); or where adverse social conditions alone, without reference to health, could justify termination of pregnancy. Important countries in this group were Great Britain, India, and Japan.
Countries allowing abortion on request without specifying reasons—sometimes limited to the first trimester of pregnancy—accounted for two fifths of the world's people. Abortions on medical grounds are usually permitted beyond the gestational limit prescribed for elective abortions, and parental consent may be required if the pregnant woman is a minor. This category includes a heterogeneous list of countries: Austria, the People's Republic of China, Cuba, Denmark, France, Italy, The Netherlands, Norway, Singapore, the republics of the former Soviet Union, Sweden, Tunisia, the United States, Vietnam, and most of the formerly socialist republics of Eastern and Central Europe.
Several of the categories in the preceding paragraphs cover a range of situations. A statute authorizing abortion to avert a threat to the pregnant woman's mental health may be interpreted strictly or may allow most women to obtain abortions. Similarly, social indications may be defined or interpreted narrowly, as in Uruguay, or broadly, as in Great Britain, India, and Japan.
The abortion statutes of many countries are not strictly enforced, and occasional abortions on medical grounds are probably tolerated in almost all countries. It is well known that in some countries with restrictive laws, abortions can be obtained openly and without interference from the authorities when performed by private physicians, as in Korea and parts of South America. Abortions may even be performed in public hospitals, as in Cuba before their legalization in 1979. Conversely, legal authorization of abortion on request does not guarantee that the procedure is actually available to all women who may want their pregnancies terminated. Lack of medical personnel and facilities and conservative attitudes among physicians may effectively curtail access to abortion, especially for economically or socially deprived women, as in parts of Austria, Germany, Ghana, India, Italy, and the United States.
The worldwide trend toward liberalization of abortion laws that was evident in the 1960s and 1970s has continued in recent years, although at a slower pace. From 1987 through 1997, laws were liberalized in the following countries with populations of 1 million or more: Albania, Belgium, Botswana, Bulgaria, Burkina Faso, Cambodia, Canada, Hungary, Malaysia, Mongolia, Pakistan, Romania, South Africa, and the Soviet Union. During this period, new restrictions were imposed in Chile, Poland, and in certain states of the United States.
Major reasons advanced by advocates of less-restrictive legislation in matters of abortion, and especially of abortion on request, have been considerations of public health (to combat illegal abortion with its associated morbidity and mortality); social justice (to give poor women access to abortion previously available only to the well-to-do); and women's rights (to secure a postulated right of all women to control their own bodies and reproduction). A desire to curb population growth, in the interest of economic and social development, has been an explicit reason for the adoption of nonrestrictive abortion policies in a few countries, such as Singapore, Tunisia, and China, and may have been an underlying reason for policy changes in some other countries, including India and Bangladesh. The majority of countries permitting abortion at the request of the pregnant woman or on broadly interpreted social indications, however, have low birth rates, and some of them actively pursue pronatalist population policies.
Opposition to the liberalization of abortion laws has come traditionally from conservative groups, mainly on moral and religious grounds, with the Roman Catholic Church the most vigorous and articulate but by no means the only opponent. The desire for higher birth rates has led to restrictive legislation in a few instances, especially in Eastern and Central Europe during the 1960s and 1970s. Abortion has become one of the most emotional and divisive political issues in a number of countries in recent years.
|ABORTION POLICY IN THE UNITED STATES|
In the United States, the tradition of the British common law that abortion was not a crime before the fetus quickened and that the aborted woman was immune from prosecution prevailed into the mid-19th century. Connecticut was the first state to deal with the subject by legislation by including it in its revised criminal code in 1821. In 1845, Massachusetts was the first state to enact a law dealing separately and exclusively with abortion. Of the 20 states with statutory provisions on abortion in 1860, all but three retained the common law immunity of women and approximately half incorporated the common law principle that abortion was a crime only after quickening. The other 13 states had no abortion statutes at that time.
After 1860, the medical profession began well-organized anti-abortion campaigns. These campaigns, as well as public reaction to explicit and flamboyant advertisements by abortionists and apothecaries and to sensationally reported abortion cases in the courts, caused legislatures to review the subject and to enact restrictive laws, which remained in force with little change for the next 100 years. In most states, a threat to the life of the pregnant woman was the sole legal ground on which abortion could be performed; in a few states, a serious threat to the woman's health was included.
A more liberal type of legislation was proposed by the American Law Institute in 1962. Following the example set by the countries of northern Europe, the relevant paragraph of the Institute's Model Penal Code would have permitted abortion if a licensed physician “believes there is substantial risk that continuance of the pregnancy would gravely impair the physical or mental health of the mother or that the child would be born with grave physical or mental defect” or if the pregnancy resulted from rape, forcible or statutory, or incest. Beginning with Colorado in 1967, approximately a dozen states adopted legislation based on the Model Penal Code.
In 1970, the legislatures of three states—Alaska, Hawaii, and New York—enacted laws authorizing abortion on request. In the state of Washington, the same result was achieved by popular referendum and in the District of Columbia as a result of a court decision.
On January 22, 1973, two landmark decisions invalidating the abortion laws of most states were handed down by the Supreme Court. In one of these, Roe v Wade, the Court ruled seven to two that during the first trimester, “the abortion decision and its effectuation must be left to the medical judgment of the pregnant woman's attending physician,” in consultation with the pregnant woman. After the first trimester, “the State… may, if it chooses, regulate the abortion procedure in ways that are reasonably related to maternal health.” After the fetus has reached viability, “the State… may, if it chooses… proscribe abortion except where necessary… for the preservation of the life or health of the mother.” In the other case, Doe v Bolton, the Supreme Court struck down a number of procedural provisions, such as the requirements that the termination of pregnancy be authorized by an abortion committee and that the woman seeking abortion must be a resident of the state in which the operation is to be performed. The Court also stressed that the attending physician's “medical judgment may be exercised in the light of all factors—physical, emotional, psychological, familial, and the woman's age—relevant to the well-being of this patient.” This statement echoes the World Health Organization's definition of health as “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity.”
The reception of these decisions ranged from enthusiastic support by those committed to women's “right to choose” to emphatic rejection by an increasingly well-organized “right-to-life” movement. In response to anti-abortion sentiment, most states have attempted to institute some form of restriction on abortion services. Between 1976 and 1989, most of these restrictions were found unconstitutional by the Supreme Court, including the limitation of second-trimester abortions to hospitals; mandatory waiting periods after counseling before the procedure could be performed; required husband's consent; and the requirement that the woman be given certain specific information that may not be relevant to her particular circumstances. The court found that parents could not exercise a veto over a minor daughter's decision to terminate her pregnancy, but approved parental consent requirements for minors provided that a court or other mechanism be available in which a young woman can show that she is sufficiently mature to make an informed decision or that an abortion would be in her best interest. Parental consent or notification requirements are in effect in many states and are enjoined in others because they do not meet the Supreme Court's requirements.
Another type of restriction permitted by court decisions is the exclusion of abortion coverage from Medicaid and other governmental health-funding programs. As of 2001, the federal government and all but 19 states provide no public funding of abortions except in cases in which the woman's life would be endangered by continuation of the pregnancy or the pregnancy resulted from rape or incest. This is perhaps the most severe governmental limitation on women's access to abortion services; one study found that approximately one third of women who would have had publicly funded abortions instead carried their pregnancies to term when public funding was unavailable.4
In 1989, the Supreme Court changed direction in Webster v Reproductive Health Services. In this decision and in Casey v Planned Parenthood of SE Pennsylvania, the court, although reaffirming that the state may not place an “undue” burden on a woman's right to abortion, indicated that restrictions such as waiting periods and required specific content of preabortion counseling are permissible. Laws requiring waiting periods, usually 24 hours, and specific counseling are now in effect in many states. The information required to be provided generally includes the gestational age of the fetus; characteristics of the fetus at that point in development; risks of the abortion procedure; risks of continuing the pregnancy; and assistance that may be available if the woman were to have the baby and keep it or place it for adoption. Although laws in Guam and Louisiana that would have prohibited almost all abortions were not allowed to stand, the Court has not yet clarified the extent of restrictions that would be permitted under the “undue burden” standard.
|INCIDENCE OF LEGAL ABORTION|
The number of pregnancies terminated each year by induced abortion throughout the world cannot be known exactly. A recent estimate is that approximately 20 million illegal and 26 million legal abortions take place annually, for a worldwide total of 46 million.5 Because there are approximately 130 million live births a year, there is approximately one induced abortion for each four known pregnancies (pregnancies ending in birth or induced abortion).
In the United States, relatively few legal abortions were performed before 1960. After 1960, the attitude of the medical profession began to change; mental health gained acceptance as a valid reason for the termination of pregnancy and more legal abortions were performed. This trend was greatly accelerated after the passage of more liberal abortion laws in a number of states and continued after the legalization of abortion by the Supreme Court (Table 1). Many of the legal abortions replaced abortions that would have been performed illegally if the law had not changed.
*1970-1972: Abortions reported to the Centers for Disease Control and Prevention.
†1973-1997: Abortions reported to The Alan Guttmacher Institute.
After 1973, the number of abortions rose to 1.6 million in 1980, remained at that level until 1990, and then fell significantly during the 1990s. Both the abortion rate (abortions per 1000 women aged 15 to 44 years) and the abortion ratio (abortions per 100 births and abortions) peaked around 1981 and have declined since 1985. In 1997, the abortion rate was 22.2 per 1000, and 25% of pregnancies (excluding miscarriages) ended in abortion. At 1997 age-specific rates, a cohort of 100 women would have 71 abortions by the time they reach menopause. At 1993 rates, 43% of women would have at least one abortion by age 45; rates have fallen since 1993, so the percentage would be somewhat lower now.6 The relatively large number of abortions reflects an even larger number of unintended pregnancies. In 1994, an estimated 48% of all pregnancies were unintended; of these, 47% ended in induced abortion, 40% in birth, and the rest in spontaneous abortion or stillbirth.6 The recent decline in the abortion rate may be attributable in part to a shift toward use of more effective contraceptives, especially long-acting hormonal methods.
Compared with Western European countries where abortion is available either at the request of the pregnant woman or on broadly interpreted social or social-medical indications and where reasonably complete statistics have been published for recent years, the United States is at the high end of the range (Table 2). The abortion rate in the United States is three to four times as high as in The Netherlands and Germany and also higher than in Britain and in Scandinavia. However, it is lower than in some of the countries of Eastern and Central Europe and the former Soviet Union. No numbers and rates are shown for France, Italy, and Spain because the published statistics are known to be seriously incomplete. Legal abortion rates in developing countries range from much lower than in the United States (Tunisia) to much higher (Cuba, Romania, Vietnam).
† Reporting incomplete
Abortion occurs among women in all population subgroups, as listed in Table 3. It is most common among women aged 18 to 29 years (67%) and among women who are not currently married (81%). A majority of abortion patients have children, even though many are unmarried. Minorities are over-represented: 42% are nonwhite and approximately 17%, most of whom are classified as white, are Hispanic. Approximately 41% are non-Hispanic white women.
*For age <15. Denominator is women age 14; for 40+, denominator is women aged 40–44
† Denominator is abortions plus live births 6 months later (to match time of conception with abortions). Births and abortions are adjusted to age of woman at time of conception.
‡ Separated women are included with married
§ Weeks since onset of last menstrual period
#Includes hysterotomy, hysterectomy, some combination procedures, and procedures reported as “other”
Table 3 illustrates the differences between the two major approaches to the statistical evaluation of pregnancy termination: abortion rates (per 1000 women of reproductive age) and abortion ratios (per 100 pregnancies, excluding spontaneous abortions). Age-specific abortion rates start at a low level in the youngest age group, rise to a peak at 20 to 24 years (when most women are still unmarried but are sexually active), then decline steadily to a low level among women in their 40s. The pattern of abortion ratios per 100 pregnancies is quite different. Ratios are fairly high among the youngest women, decline progressively to 30 to 34 years of age when most women are married and many are building their families, then rise to a higher level among the relatively few pregnancies occurring after age 40.
Whether measured by rates or ratios, abortion is more common among unmarried women than among married women, even when age is taken into account. Rates are highest among unmarried cohabiting women. Because approximately half of women having abortions are nulliparous and 66% intend to have children in the future, it is of highest priority in providing abortion services to minimize risk to the woman's subsequent reproductive potential.7
Although white women, including Hispanics, made up a majority of women having abortions in 1997, the abortion rate for black and other women (48 per 1000) was three times as high as the rate for white women (16 per 1000). This difference reflects a high percentage of unwanted and mistimed pregnancies among economically and socially disadvantaged women rather than a greater propensity to terminate such pregnancies by abortion. Compared with white women, nonwhite women have a higher rate of unintended pregnancies; these result in more unplanned births as well as in more abortions.
Among other demographic characteristics associated with abortion incidence are income (in 1994, the abortion rate of women with family income less than $15,000 was four times that of women whose family income was $60,000 or more); Hispanic origin (the abortion rate of Hispanic women is between that of white and nonwhite women); place of residence (incidence is relatively low among residents of nonmetropolitan counties); and religion (abortion rates are below average among Protestant women, close to average among Catholics, and above average among women with no religious identification).7
The issue of repeat abortion is a matter of concern, especially for those who feel that abortion is unacceptable as a primary method of fertility regulation and should be used only as a backup measure when contraception has failed. Others fear that even minor adverse effects on the outcome of later pregnancies would be cumulated by multiple abortion experiences. The percentage of repeat abortions has increased over the years since legalization and now amounts to 49%; 21% are third or higher order abortions. The increase does not reflect a progressive change from contraception to abortion as the primary method of fertility regulation. Rather, it reflects the growing number of women who have had a first legal abortion and, therefore, are at risk of having a repeat abortion. Most women who have had an abortion are at risk of a repeat unintended pregnancy and abortion because they are sexually active, able to become pregnant, have difficulty using contraceptives effectively, are willing to end an unintended pregnancy by abortion, and are concentrated in subgroups with high rates of unintended pregnancy. It is therefore important for clinicians to pay special attention to the contraceptive needs of abortion patients.
Women usually have a number of reasons for seeking abortions, more than three reasons on average according to one study. The reasons most commonly reported were that a baby would interfere with work, school, or family responsibilities (76%); lack of financial resources to support a child (68%); and problems in the relationship with her partner or desire to avoid single parenthood (51%). Health reasons were reported by 7% and fear of a possible fetal defect by 13% (although only 1% had been advised by a physician of such a risk). One percent of the respondents said the pregnancy was the result of rape or incest.8 More than half of women having abortions practice contraception during the month they become pregnant, although not necessarily correctly and consistently.7
|PERIOD OF GESTATION AND METHOD OF ABORTION|
One of the most important factors in the evaluation of morbidity and mortality associated with induced abortion is the period of gestation at which the pregnancy is terminated. Although the traditional division has been between abortions in the first trimester and those in the second trimester, experience indicates that this dichotomy is not sufficient because morbidity and mortality increase with the progress of gestation, even within each trimester.
In the United States, more than half (55%) of the abortions take place at 8 weeks from the last menstrual period or earlier, and an increasing proportion (18% in 1997) are performed before 7 weeks. Medical methods of early abortion by use of mifepristone or methotrexate in combination with the prostaglandin misoprostol are now available, and early vacuum aspiration, either manual or electric, is now accepted as safe. Contributing to these developments is the availability in clinics and physicians' offices of sensitive vaginal ultrasound equipment for assessing pregnancy and ensuring that the uterus is empty after the procedure.
In 1997, 12% of abortions occurred after 12 weeks, including 1% after 20 weeks. Under extraordinary circumstances, abortions may be performed after 24 weeks; a reasonable estimate is that there are on the order of 1000 such abortions a year. Many of these follow the discovery of fetal abnormalities.
Abortions past 12 weeks occur most frequently among the youngest women, as shown in Figure 1. The strong inverse association of period of gestation and a woman's age probably reflects the inexperience of the very young in recognizing the symptoms of pregnancy, their unwillingness to accept the reality of their situation, their ignorance about where to seek advice and help, and their hesitation to confide in adults. Economic considerations and, in many states, regulations prohibiting abortions for minors without parental consent or notification or a court order also contribute to delays.
Approximately 98% of abortions are performed by instrumental evacuation. During the first trimester, vacuum aspiration with or without subsequent check curettage is used in all but a small percentage of cases, when dilation and curettage or early medical (nonsurgical) methods are used. Between 16 and 20 weeks, dilation and evacuation is used in 94% of abortions, and after 20 weeks in at least 85%. Prostaglandins are often used for cervical preparation with dilation and evacuation. Intrauterine saline instillation and uterine surgery (hysterectomy and hysterotomy) are rarely used.1 It is expected that an increasing number of early abortions will be performed non-surgically by use of mifepristone and misoprostol or other drugs.
Experience has shown that both first- and second-trimester abortions can be performed safely in clinics and physicians' offices. The proportion of abortions performed in hospitals has declined from more than half in 1973 to 7% in 1996. In that year, only approximately 1% of all abortions were hospital inpatient procedures, the remainder being performed on an outpatient basis or in a nonhospital facility.2 The number of hospitals where abortions are known to be performed has dropped from a peak of 1687 in 1976 to 703 in 1996, and among hospitals offering abortion services, the number of abortions per hospital has fallen. An important factor in the shift of abortion services to nonhospital settings is the greater cost of using hospital facilities to provide the service.
Abortion services are increasingly concentrated in high-volume facilities. In 1996, 79% of abortions took place in the 442 clinics and 12 hospitals that performed 1000 or more of the procedures during the year, and 70% took place in clinics where half or more of patient visits were for abortion services. Such facilities are generally located in large metropolitan areas, with the result that many women in small cities and rural areas must travel long distances for services. The concentration of abortions in high-volume clinics probably increases the experience and competence of the professionals involved and, thereby, the quality of the services. Conversely, follow-up care is fragmented, with possible unfavorable effects on the patient.
Complication rates for abortion are difficult to define and to measure. For example, blood loss that would be considered unproblematic to one investigator might be considered hemorrhage by another. Complications recorded on state-mandated abortion-reporting forms may be underreported, when infection and other problems may appear after the day the procedure was performed and the reporting form completed. Because only approximately one third of abortion patients on average return for routine follow-up examination, delayed complications are not always known to the abortion facility.
The most recent study of a large number of U.S. abortion patients covered 170,000 consecutive first-trimester abortions performed by vacuum aspiration between 1971 and 1987. Follow-up information was obtained either at the clinic or by return of a physician's note on all but 8%. Only 121 patients had complications requiring hospitalization, a rate of 0.07%. Minor complications were reported for 0.85%; the most common of these were mild infection and need for resuctioning.9 The largest database of abortion complications in the United States is maintained by the National Abortion Federation, an association of abortion providers. In 1999, 112 members reported information on 207,646 abortions at all gestational ages; follow-up information was available for 24%. Complications requiring hospitalization were reported for 193 patients for a rate of 0.09% of all patients and 0.39% of those with follow-up. Infections were reported for 126 patients (0.06% of all patients), continued pregnancies for 110 (0.05%), and perforations for 52 (0.03%).10 The complication rate for all patients represents a minimum estimate of postabortal complications. Because patients with complications are more likely to return for care than those without complications, the rate for patients with follow-up represents a maximum estimate.
The most comprehensive large-scale investigations of early medical complications of legal abortions were undertaken in the United States by the Joint Program for the Study of Abortion, sponsored by The Population Council during 1970/197111 and by the CDC in 1971/1975 and again in 1975/1978.12 These large-scale studies, based on 73,000, 80,400, and 84,300 abortions, respectively, permit assessment of the interaction of variables that determine levels of morbidity, including age, parity, socioeconomic and health status, period of gestation, operative procedure, and concurrent sterilization.
The three studies found a decline in the rate of major complications, probably a consequence of the increasing experience of medical professionals with abortion techniques. As indicated by the studies cited above, rates have almost certainly declined further since 1978.
The numbers of women with one or more major complications per 100 women undergoing abortions in 1975/1978 appear in Table 4. The rates shown are based on three categories of adverse experiences: pelvic infection with fever of 38°C or more for 3 or more days; hemorrhage requiring blood transfusion; and unintended major surgery (laparotomy, hysterotomy, or hysterectomy).
*Excluding abortions by uterine surgery as the primary method
† Weighted according to the distribution of abortions by weeks of gestation in 1978
‡ Suction or surgical curettage at 13 weeks
In analyzing the complications of abortion, a distinction must be made between complications associated with the termination of pregnancy and those associated with pre-existing pathology or with surgical procedures undertaken for other purposes, such as the achievement of permanent sterility or the removal of the uterus because of fibroid tumors. Therefore, complication rates are shown not only for the aggregate of all patients but also for presumably healthy women subject only to the risks associated with the abortion procedure (i.e., excluding all women with reported pre-existing complications and those undergoing sterilizing operations).
Complication rates are also shown for all patients and for those with follow-up. Follow-up was defined as any contact with the woman 14 or more days after the abortion, including contact by mail or by telephone either directly by the investigator or by the physician who referred her to the hospital or clinic or by the physician to whom she was referred for aftercare.
As listed in Table 4, the overall rate of major complications was 0.31 per 100 for all women in the study. The maximum estimate was 0.43 per 100 women with follow-up. Complication rates were higher at 6 weeks' gestation or earlier than they were at 7 to 10 weeks and reached their peaks at approximately 20 weeks. Between approximately 8 or 9 weeks and 20 weeks, the rate of major complications increased by approximately 20% for each additional week of gestation. The rates observed in the group of presumably healthy women exposed to the risk of abortion only were lower at each period of gestation.
Complication rates were lowest for abortions by vacuum aspiration at 12 weeks' gestation or earlier, followed in ascending order by dilation and evacuation (instrumental evacuation atk 13 weeks or later); medical induction by instillation of saline solution, prostaglandin or urea; and uterine surgery.
Because abortions are often obtained by young, nulliparous women, the effect of induced abortion on future reproductive capacity is an important concern. A series of reviews have examined the literature for studies that estimate the relative risk of adverse reproductive outcomes for women who have had an abortion compared with those who have not. Only studies that met basic criteria for methodologic adequacy were included. Following are some of the conclusions.13,14,15 Of nine studies identified, eight found no elevated risk of secondary infertility that was statistically significant. The one study that did find an elevated risk took place in Greece, where abortion was illegal and was often performed by dilation and curettage rather than vacuum aspiration.
None of 10 studies reviewed found a statistically significant association between induced abortion and subsequent ectopic pregnancy. Of seven studies of midtrimester spontaneous abortion, none found a significantly higher risk among women who have had one prior pregnancy that was terminated by vacuum aspiration compared with women pregnant for the first time. Three of four studies, however, found an association between dilation and curettage and subsequent midtrimester spontaneous abortion. Similarly, none of the seven studies examined found prematurity to be more common among women who had had an induced abortion than among women with no prior pregnancies. None of nine studies reviewed showed an increased risk of low birthweight after vacuum aspiration. Thus, although infection occasionally occurs after induced abortion and can undoubtedly impair fecundity in some women, adverse effects of vacuum aspiration are so rare as to be undetectable in epidemiologic studies. The studies as a group suggest that the true relative risk of reproductive impairment could be very close to one. The few available evaluations lead to the same conclusion with regard to the effect of abortion on complications of labor and delivery.
Studies of abortions performed by dilation and curettage have had less clear results. The effects of second-trimester abortions and multiple abortions have not been adequately studied for definite conclusions to be drawn, but if the effects were large, they would have been noted.
High levels of mortality from legal abortion prevailed in the United States as elsewhere during the period of restrictive legislation, when a significant proportion of the women undergoing abortion experienced pre-existing complications that made them poor risks for any type of surgery. In 1972, the mortality rate was 4 per 100,000 legal terminations. The rate fell to 2 in 1973 to 1977, when elective abortion became more easily available to most women (Table 5). As abortion services improved with increasing experience and technical advances, mortality fell to 0.8 in 1978 to 1982 and to 0.7 in 1983 to 1987 and 1988 to 1992. Mortality from illegal abortion fell from 39 recorded deaths in 1972 to 2 in 1976. In the 1980s through the early 1990s, there was on average slightly less than one death per year from illegal abortions (i.e., abortions performed by someone other than a licensed physician or someone acting under the supervision of a licensed physician).
*Eight deaths with unknown gestation were distributed proportionately.
Information on abortion mortality is provided by the CDC, which investigates all reports of abortion-related deaths, whether or not abortion is recorded as a cause of death on the death certificate. The CDC reports include all deaths associated with abortion, including those with another primary cause, usually a pre-existing medical condition. The main causes of death associated with legal abortion are complications of anesthesia, hemorrhage, infection, and embolism. Since the early 1970s, the number of deaths has dropped sharply from all causes. 16 One of the main risk factors for abortion mortality is the period of gestation when pregnancies are terminated. In the United States between 1983 and 1987, mortality ranged from 0.2 per 100,000 abortions at 8 weeks or less to 15.9 per 100,000 for abortions at 21 weeks or more. Mortality increased by approximately 30% to 40% with each week of gestation past 8 weeks.
Age has little relation to abortion mortality, but mortality is distinctly higher among women of minority races than among white women. Mortality was higher during 1974 to 1977 for first-trimester abortions performed in hospitals than for abortions in nonhospital facilities. The difference disappeared when women with pre-existing complications or concurrent sterilization were excluded from the hospital data.17 Similar results were found for abortions performed at 13 to 15 weeks by dilation and evacuation between 1972 and 1978.18 These analyses have not been repeated for more recent years.
Abortion-related mortality may be compared appropriately with the risk to life associated with carrying a pregnancy to term. In the United States, maternal mortality attributed to complications of pregnancy and childbirth, excluding induced abortion, was 7.4 deaths per 100,000 live births in 1983 to 1987, based on data recorded on death certificates. Excluding deaths from spontaneous abortion and ectopic pregnancy, some of which may have resulted from pregnancies that would have ended in induced abortion if they had been normal, mortality was 6.1 per 100,000 live births. These statistics exclude deaths associated with but not attributed to pregnancy and childbirth. Thus, mortality from induced abortion before around 16 weeks of gestation is lower than maternal mortality attributed to childbirth, and mortality from abortion is not statistically significantly higher at any gestation. Overall, maternal mortality is at least 10 times mortality from induced abortion when deaths associated with pregnancy and childbirth are included.
The safety of abortion in the United States is comparable to that in other developed countries, even though most other countries have a smaller proportion of second-trimester abortions and count only the deaths attributed to abortion. As listed in Table 6, abortion mortality ranges from 0.1 to 1.0 death per 100,000 procedures. Because of the small numbers of deaths on which the rates are based, the differences among the countries are not statistically significant.
*Deaths were attributed to legal abortion on death certificates, except for England and Wales, where deaths are recorded on the abortion notification form, and the United States, where deaths include all those associated with abortion after investigation by the Centers for Disease Control and Prevention.
†95% confidence intervals in parentheses
In all countries for which we have data, there were sharp declines in the abortion mortality rates between the 1970s and the 1980s. Factors that appear to account for the declines include increasing skill of abortion practitioners as a result of having more years of experience; the nearly universal adoption of vacuum aspiration; improved treatment of complications; advances in abortion technology; and, in some countries, a reduction in the proportion of procedures performed during the second trimester. There may also have been improvements in the underlying health condition of patients.
12. Grimes DA, Schulz KF, Cates W et al: The Joint Program for the Study of Abortion. In Hern W, Andrikopoulos B (eds): Abortion in the Seventies, pp 41–46. New York, National Abortion Federation, 1977