Since 1979, the Centers for Disease Control and Prevention's (CDC) Pregnancy Mortality Surveillance System has conducted surveillance of deaths from all pregnancy outcomes, including abortion (both induced and spontaneous). The Pregnancy Mortality Surveillance System identifies cases through a variety of sources, including state health departments, vital records data from CDC's National Center for Health Statistics, hospitals and clinics, organizations, individuals, and the media. Deaths from spontaneous abortion are investigated and coded by experienced clinicians, who review death certificates, hospital and other medical records, and autopsy reports to determine whether the death was abortion related, whether the abortion was induced or spontaneous, and the gestational age and cause of death. An abortion-related death is defined as a death resulting from a direct complication of an abortion, an indirect complication caused by a chain of events initiated by the abortion, or the aggravation of a preexisting condition by the physiologic effects of the abortion. 1
We calculated spontaneous abortion case-fatality rates for the years 1981 to 1991, the latest year with information available on both the number of deaths due to spontaneous abortion and the number of spontaneous abortions. We defined the case-fatality rate as the number of spontaneous abortion–related deaths per 100,000 spontaneous abortions. For the numerator, we used all pregnancy-related deaths in the Pregnancy Mortality Surveillance System for which the outcome of pregnancy was a spontaneous fetal loss at less than 20 weeks' gestation. For the denominator, we used published estimates of the number of spontaneous losses of clinically recognized pregnancies at less than 20 weeks' gestation. 5 We elected to use the number of spontaneous abortions as the denominator, instead of using live births, to estimate more accurately the risk of mortality in the event of a spontaneous abortion. On the basis of the literature, 2 we applied gestational age–specific distribution of spontaneous abortions to our estimate of 9,279,100 spontaneous abortions into those occurring in the first 12 weeks (87%) and those occurring 13–15 weeks (9%), and from 16–19 weeks (4%).
Because of the increasing risk of the occurrence of a spontaneous abortion with increasing maternal age, 6,7 we stratified woman's age as less than 30 years, 30–34 years, and 35 years and more. We categorized woman's race as white and black/other race because denominator data for the early 1980s were grouped into these categories. Gestational age was defined as the number of completed weeks from the date of the onset of the last menstrual period. If the exact gestational age was not stated, attempts were made to categorize the abortion as occurring in the first or second trimester, based on all the available data. Gestational age was divided into less than 13 weeks' and 13 weeks' or later; the latter group was further stratified into 13–15 weeks' and 16–19 weeks'. We categorized the causes of death as infection (eg, peritonitis, endometritis, and septicemia), hemorrhage (vaginal or intra-abdominal), embolism (pulmonary or amniotic fluid), complications from anesthesia, and other causes. For each death, the presence or absence of disseminated intravascular coagulation (DIC) 8 was determined on the basis of bleeding times, clinical history, and/or pathologic diagnosis. We calculated risk ratios (RR) for death by the woman's age, race, and gestational age. We used the χ2 test of trend to examine trends in annual case-fatality rates. Data were analyzed using SAS 6.11 for Windows (SAS Institute, Cary, NC) and 95% confidence intervals (CI) were determined using Poisson regression.
A total of 62 spontaneous abortion–related deaths were reported to the Pregnancy Mortality Surveillance System for the years 1981–1991. The overall case-fatality rate for the 11-year period was 0.7 per 100,000 spontaneous abortions. The case-fatality rate was relatively constant in these years; the high was 1.0 per 100,000 in 1985; the low 0.3 per 100,000 in 1989 (Figure 1). No significant trend was noted (χ2 for trend = 5.43; P = .71).
Women 35 years of age and older had a 70% greater risk of spontaneous abortion–related death compared to women under 30, although this increased risk was not statistically significant (Table 1). Women of black and other races had nearly a 3.8-fold greater risk of dying after a spontaneous abortion than did white women. When we stratified by age and race simultaneously, we found that among women less than 30 years and women 35 years of age and older, women who were of black or other race were significantly more likely to die than white women (Table 2).
Compared with women who had a spontaneous abortion in the first 12 weeks of pregnancy, women who had a spontaneous abortion after 12 weeks' gestation had a statistically significantly eightfold higher risk of dying (Table 1). While this increased risk was five-fold higher for women who had a spontaneous abortion in weeks 13–15 compared with women with a miscarriage before 13 weeks, it was 13.7 for women at 16–19 weeks' gestation (Table 1).
From 1981 through 1991, the leading cause of spontaneous abortion–related death was infection (59%), followed by hemorrhage (18%), embolism (13%), complications from anesthesia (5%), and other causes (5%). In 48% of the cases in which DIC was not the primary cause of death, DIC was found to be an associated condition. Among these DIC-associated cases, infection was the primary cause of death in 70%, followed by hemorrhage in 17% and embolism in 13%.
Deaths from spontaneous abortion are rare; less than 5% of pregnancy-related deaths are due to spontaneous abortion. 9 For the years 1981–1991, we found the rate of death from a spontaneous abortion to be 0.7 per 100,000 estimated spontaneous abortions. This rate is similar to the mortality rate for legal induced abortion, which was 0.7 deaths per 100,000 induced abortions from 1981 to 1987. 10
Sixty-two reported deaths took place in the 11 years from 1981 through 1991—an average of six deaths per year. In the 9 years from 1972 through 1980, 122 spontaneous abortion deaths were reported, including 101 non-intrauterine device (IUD)–related deaths and 21 IUD-related deaths—an average of 11 non-IUD–related deaths per year. 4 The case fatality rate of 0.7 found for 1981–1991 is lower than that estimated for 1972–1980, 1.0 per 100,000 spontaneous abortions. While the methods used to calculate the number of spontaneous abortions for the former period were different from those used in the current analysis and did not take into account differences in abortion rates by maternal age, both the annual number of deaths and the estimated death rates indicate a decrease in the risk of death from spontaneous abortion between the two periods. Among all pregnancy-related deaths, the percentage of deaths due to complications of anesthesia and to hemorrhage decreased from the period 1979–1986 to the period 1987–1990. 9 Thus, improvements in medical practice might have decreased the risk for a spontaneous abortion–related death. On the other hand, the 1970s was a time of heightened awareness of IUD-associated spontaneous abortion–related deaths. Hence, spontaneous abortion–related deaths may have been better publicized in the 1970s, and their identification more rigorous than in the 1980s.
A dramatic decrease in the use of IUDs as a method of contraception occurred between 1973 and 1988. 11,12 From 1970 to 1974, the Dalkon Shield, an IUD with a multifilament tail, was marketed in the United States but was withdrawn by the manufacturer from the market in 1974 as a result of an association between IUDs and pregnancy-related complications. 13 In 1975, IUD use as a risk factor for spontaneous abortion–related mortality was reported; the Dalkon Shield carried a higher risk (RR = 5) of death compared to the other IUDs, with an increased risk (RR = 22) among long-term IUD users. 14 In 1980, the manufacturer advised physicians to remove the Dalkon Shield from asymptomatic women because of the risk of infection by Actinomyces, and, in 1983, the Food and Drug Administration recommended that women with a Dalkon Shield have it removed because of the increased risk (OR = 5) of pelvic inflammatory disease compared with other IUD types. 15,16 During 1972–1980, 21 of the 122 deaths (17%) were associated with use of an IUD; during 1981–1991, no such IUD-associated deaths were reported in our study.
We identified several demographic characteristics that place women at increased risk of dying from a spontaneous abortion. Women 35 years and older are not only more likely to have a pregnancy end in a spontaneous abortion than younger women, 2 but they also have an increased risk of a spontaneous abortion–related death. As seen with pregnancy-related mortality in general, 9 nonwhite women are almost four times more likely to die from a spontaneous abortion. When examining the racial disparity often noted in reproductive health outcomes, while race may be a marker for poverty, it may also represent unmeasured social and physical factors, such as access to health care or care-seeking behaviors. 17 While this study does not allow us to examine the reason behind the racial disparities, further research should be done to explore the disparity noted in pregnancy outcomes.
For both spontaneous and induced abortions, the risk of death is higher at 13–19 weeks' gestation than in the first 12 weeks of pregnancy. We found an RR of 8.0 for spontaneous abortion death in the second versus first trimester; Lawson et al 18 found an RR of 5.8 for legal induced abortion at 13 to 20 weeks gestation, compared with the first 12 weeks. Within the second trimester, an increase in risk with increasing gestational age was also found. The RR for death after an induced abortion at 13–15 weeks was 5.0 and at 16–20 weeks, 12.9. 18 We found an RR of death related to spontaneous abortion of 5.0 at 13–15 weeks and 13.7 at 16–19 weeks. For the years 1972–1980, Berman et al 4 found an RR of 8.8 for spontaneous abortion death at 12–19 weeks compared with less than 12 weeks. Thus, in the first 20 weeks of pregnancy, the risk of death from an induced abortion is similar to the risk of death from a spontaneous abortion. As gestational age increases, the increased blood volume, increased uterine size, and increased fetal and placental tissue could increase the risk of complications in both types of abortions.
The percent of spontaneous abortion–related deaths from 1981 through 1991 due to infection were slightly greater than the percent from 1972 through 1980 (59% vs. 48%). It is unclear whether this reflects a true increase or a relative increase, ie, a smaller percentage of deaths due to anesthesia and hemorrhage. In addition, we found DIC was associated with many deaths due to infection, hemorrhage, and embolism. Because the results of clinical or pathologic tests for DIC were not available in the records of all cases, the prevalence of DIC in spontaneous abortion–related deaths may be underestimated. Disseminated intravascular coagulation should be considered during assessment of women with spontaneous abortions in order to anticipate this potential complication.
This study is unique because it used national data on the number of spontaneous abortion deaths and the number of spontaneous abortions to calculate case-fatality rates. This approach allowed us to examine risk ratios for a variety of characteristics of women who had a spontaneous abortion and to compare case-fatality rates of deaths due to other pregnancy outcomes. Because we limited our estimates to clinically recognized pregnancies, we underestimated the total number of pregnancy losses. However, the impact of preclinical pregnancy on the risk of death from fetal loss is unknown, and the identification of very early, preclinical pregnancy loss is, in general, limited to research studies. We were not able to examine the effect of marital status or geographical location, as was done by Berman et al, 4 because region- and marital-specific rates of spontaneous abortions are not available. We also were not able to examine potential behavioral and reproductive risk factors for spontaneous abortions, such as cigarette smoking, alcohol use, or previous abortion history, 6 as these data were not available from the death certificates or medical charts.
While this study found a low risk of death from spontaneous abortion, which appears to have decreased since the 1970s, nevertheless, surveillance of spontaneous abortion–related deaths should continue to monitor these events, as spontaneous abortion remains one of the most common outcomes of pregnancy.
1. Centers for Disease Control. Abortion surveillance report 1981. Atlanta, GA: Centers for Disease Control, 1985.
2. Wilcox A. Surveillance of pregnancy loss in human populations. Am J Ind Med 1983;4:285–91.
3. Mills JL, Simpson JL, Driscoll SG, Jovanovic-Peterson L, Van Allen M, Aarons JH, et al. Incidence of spontaneous abortion among normal women and insulin-dependent diabetic women whose pregnancies were identified within 21 days of conception. N Engl J Med 1988;319:1617–23.
4. Berman S, MacKay H, Grimes D, Binkin N. Deaths from spontaneous abortion in the United States. JAMA 1985;253:3119–23.
5. Saraiya M, Berg CJ, Shulman H, Green CA, Atrash HK. Estimates of the number of clinically recognized pregnancies in the United States, 1981–1991. Am J Epidemiol 1999;149:1025–9.
6. Kline J, Stein Z, Susser M. Environmental factors, miscarriages and karyotypes. In: Kline J, Stein Z, Susser M, eds. Conception to birth. New York: Oxford University Press, 1989:118–45.
7. Kline J, Stein Z. Spontaneous abortion (miscarriage). In: Bracken M, ed. Perinatal epidemiology. New York: Oxford University Press, 1984:23–51.
8. Sipes S, Weiner C. Coagulation disorders in pregnancy. In: Reece E, Hobbins J, Mahone M, Petrie R, eds. Medicine of the fetus and mother. Philadelphia: JB Lippincott Company, 1992.
9. Berg C, Atrash H, Koonin L, Tucker M. Pregnancy-related mortality in the United States, 1987–1990. Obstet Gynecol 1996;88:161–7.
10. Koonin LM, Smith JC, Ramick M, Strauss LT, Hopkins FW. Abortion surveillance—United States, 1993 and 1994. MMWR CDC Surveill Summ 1997;46:37–98.
11. Mosher W, Pratt W. Contraceptive use in the United States, 1973–1988. Hyattsville, MD: National Center for Health Statistics, Advance Data, 1990, vol. 182; no. 2.
12. Atrash H, Frye A, Hogue C. Incidence of morbidity and mortality with IUD in situ in the 1980s and 1990s. In: Bardin CW, Mishell DR, eds. Fourth International Conference on IUDS. Boston: Butterworth-Heinemann, 1994:76–87.
13. Centers for Disease Control. IUD safety: Report of a nationwide physician survey. MMWR Morb Mortal Wkly Rep 1974;23:226–31.
14. Cates W, Ory HW, Rochat R, Tyler C. The intrauterine device and deaths from spontaneous abortions. N Engl J Med 1976;295:1155–9.
15. Lee NC, Rubin Gl, Ory HW, Burkman RT. Type of intauterine device and the risk of pelvic inflammatory disease. Obstet Gynecol 1983;62:1–6.
16. Centers for Disease Control. Current trends. IUD safety: Report of a nationwide physician survey. MMWR Morb Mortal Wkly Rep 1997;46:969–73.
17. Rowley D, Hogue C, Blackmore C, Ferre CD, Hatfield-Timajchy K, Branch P, et al. Preterm-delivery among African-American women: A research strategy. Am J Prev Med 1993;96:1–6.
18. Lawson HA, Frye A, Atrash HK, Smith JC, Shulman HB, Ramick M. Abortion mortality, United States, 1972 through 1987. Am J Obstet Gynecol 1994;171:1365–72.