Abortion mortality is low; anesthesia-related deaths are of rising importance

Abortion-related deaths have become increasingly rare since induced abortion became legal, with the case-fatality rate falling by about 90% between 1972 and 1987. Over that period, nonwhite women and those in the mid-to-late second trimester of their pregnancy are more likely than others to have died during or following an induced abortion. Among the various abortion methods, curettage and evacuation procedures are associated with the lowest mortality risk; complications related to the anesthesia used have become the most common cause of abortion-related deaths. These are among the findings of an analysis of abortion mortality surveillance data collected by the Centers for Disease Control and Prevention (CDC).(1)

The CDC data reflect deaths resulting from a direct complication of abortion, an indirect complication or the aggravation by an induced abortion of a preexisting condition. Such deaths are reported to the CDC through a variety of sources, ranging from state health departments and hospital maternal mortality committees to the news media and private citizens. Such reports are then investigated by medical epidemiologists, who determine if the deaths were abortion-related, and if so, whether they resulted from legal abortions, illegal abortions or spontaneous abortions.

Between 1972, the last full year before Roe v. Wade, and 1987, the most recent year for which data are available, a total of 667 deaths possibly related to abortion were reported to the CDC. Of these, 240 could definitely be attributed to legal induced abortions, 172 were associated with spontaneous abortions, 154 were judged not to have been abortion-related, 88 were caused by illegal induced abortions and 13 were abortion-related but could not be classified further. The case-fatality rate for legal induced abortion over the entire 1972-1987 period (1.3 deaths per 100,000 legal abortion procedures) masks the precipitous decline in abortion-related mortality, from 4.1 deaths per 100,000 abortion procedures to 0.4 per 100,000.

A number of maternal characteristics were associated with the risk of abortion-related death for the period as a whole. The mortality rate was significantly higher among nonwhite women than among white women (2.3 vs. 0.9 deaths per 100,000 procedures), and was significantly lower among women aged 19 or younger (1.0 deaths per 100,000) than among women aged 40 and older (3.1 per 100,000). The rate also increased with parity, although the elevation in risk was statistically significant only among women who had had three or more children (2.5 deaths per 100,000).

The timing of the procedure and the abortion method used also appear to affect the mortality risk. The mortality risk for induced abortions performed at 16-20 weeks or at more than 20 weeks of gestation (9.3 deaths and 12.0 deaths per 100,000 procedures, respectively) was substantially higher than the risk associated with procedures done at eight weeks or earlier (0.4 per 100,000). Methods such as first-trimester curettage (either suction or sharp) and second-trimester evacuation were associated with the lowest mortality rates (0.5 deaths per 100,000 and 3.7 deaths per 100,000, respectively); hysterectomy and hysterotomy, although rarely used to induce abortion, were associated with the highest mortality rate (51.6 deaths per 100,000), while instillation of prostaglandin or saline was associated with a rate of 7.1 deaths per 100,000. Problems such as infection, embolism, hemorrhage or complications associated with anesthesia each accounted for about 20% of all abortion-related deaths that occurred between 1972 and 1987.

To examine trends over time, the analysts compared data for the periods 1972-1976, 1977-1982 and 1983-1987. They found that the rate of abortion-related deaths declined markedly between the first two intervals (from 3.0 deaths per 100,000 procedures to 1.0 per 100,000), then fell slightly (to 0.7 per 100,000) in the most recent period.

Although a decline in abortion-related mortality occurred among both nonwhite and white women, nonwhite women continued to be at greater risk than white women: In the period 1972-1976, the rate of abortion-related mortality among nonwhite women was more than twice as great as that among white women (5.3 deaths per 100,000 procedures vs. 2.0 per 100,000 procedures). By 1983-1987, nonwhite women were at about twice the mortality risk of white women (rates of 1.1 per 100,000 and 0.5 per 100,000 respectively).

The risk associated with age also narrowed: In 1972-1976, mortality rates were more than twice as high among women aged 30 and older (4.3 per 100,000) as among those younger than 20 (1.9 per 100,000), and there was a similar difference between these age-groups in 1977-1982 (1.8 per 100,000 vs. 0.8 per 100,000). By 1983-1987, however, no significant differences by age could be detected.

The mortality rate associated with each abortion method declined throughout the period; by 1983-1987, the difference in the mortality rates for evacuation abortions and instillation abortions had narrowed considerably and was no longer statistically significant.

Likewise, the mortality rate associated with each cause of death decreased between 1972 and 1987. In addition, the distribution of deaths by cause shifted substantially: Complications of anesthesia, which had represented the smallest proportion of deaths (16%) in the period 1972-1976, became an increasingly important risk factor in 1977-1982 and were the most common cause of abortion-related mortality in 1983-1987, accounting for 29% of all deaths.

In discussing their findings, the researchers note that, on average, white women obtain abortions earlier in pregnancy than do nonwhite women, a characteristic that may explain the apparent significance of race as a risk factor. Their data, however, did not permit multivariate analysis, which could have assessed the interaction of race and gestational age. As for the increased risk associated with methods such as hysterectomy and hysterotomy, the analysts observe that women having these procedures may have other risk factors that make them "less than optimal operative candidates." They also suggest that the narrowing of the difference in risk between evacuation and instillation procedures may reflect greater use of the evacuation procedures in midtrimester abortions.

The growth in the share of deaths attributable to anesthesia complications is a matter of particular concern, the researchers comment, and underscores the need for "greater caution...in choosing appropriate anesthesia and analgesia for abortion procedures." In particular, they add, "the majority of deaths in 1983 through 1987 were associated with general anesthesia during the first trimester, a time [otherwise] usually associated with the lowest overall risk." Providers should be equipped to manage anesthesia complications, the investigators add, since "most, if not all, anesthesia deaths should be preventable."

The analysts point out that their data have several limitations. First, although the CDC relies on multiple reporting sources (thus increasing the likelihood that all abortion-related deaths are recorded), completeness of reporting cannot be guaranteed. Additionally, because the analysts estimated some denominators, bias is possible if some deaths were missed and the characteristics of women whose deaths were not reported differed from those of women for whom data were available.

Finally, the investigators contend that the interval between abortion-related deaths and their reporting to the CDC has been growing. They encourage providers and government agencies to report such deaths promptly, to "improve the collection of meaningful data needed to understand how to prevent pregnancy-related morbidity and mortality."

Reference

1. H.W. Lawson et al., "Abortion Mortality, United States, 1972 Through 1987," American Journal of Obstetrics and Gynecology, 171:1365-1372, 1994.

Copyright The Alan Guttmacher Institute Jul 1995
Provided by ProQuest Information and Learning Company. All rights Reserved