In countries in which induced abortions are legal, as in Denmark, pregnancy terminations are done by qualified medical doctors during the first trimester. Histories of induced abortions have been associated with increased risk of preterm delivery in some studies1–3 but not in all.4–8 In a causal association, we would expect that risk to depend upon interpregnancy intervals.5,9
In 1995, there were 70,089 births in Denmark, and the perinatal mortality was 0.75%.10 First-trimester induced abortion was made legal in Denmark in 1973, thus induced abortions by unqualified persons are extremely uncommon. In 1995, 17,720 pregnancies were terminated legally, about 96.8% in the first trimester.11 The aim of this study was to examine the association between an induced abortion and the duration of the subsequent pregnancy, focusing on preterm or post-term deliveries.
The present study was based on cohorts identified by three Danish national registries in a population of 1.1 million women between 15 and 44 years of age.11 The registries were the Medical Birth Registry, established in 1973, the Hospital Discharge Registry, established in 1978, and the Induced Abortion Registry that contains computerized data on all induced abortions since 1973. All pregnancy events were identified and linked by unique 10-digit personal identification numbers given to all Danish citizens since 1968.
We identified all primigravidas who had their first registered pregnancies (index pregnancies) from 1980 to 1982. Women who had first-trimester induced abortions formed the abortion cohort identified by International Classification of Diseases, 8th Revision, codes from 640.00 to 642.99. Women were excluded if first pregnancies were terminated later than the first trimester. The control cohort consisted of women whose first registered pregnancies from 1980 to 1982 were not terminated.
All subsequent pregnancies up to 1994 were identified by outcome registries. Live and stillbirths were registered in the Medical Birth Registry, spontaneous abortions leading to hospitalization were registered in the Hospital Discharge Registry, and induced abortions were registered in the Induced Abortion Registry and in the Hospital Discharge Registry. Women who had at least two pregnancies between 1980 and 1994 were extracted (15,727 in the abortion cohort, 46,026 in the control cohort). Subcohorts of the abortion cohort were defined by number of first-trimester induced abortions before the first nonterminated pregnancy (informative pregnancy). Subcohorts of the control cohort were defined according to number of pregnancies before the first induced abortion (informative pregnancy). We then compared the risk of preterm delivery (before 37 weeks' gestation) and post-term delivery (42 or more weeks' gestation) in the two cohorts.
Analyses were further stratified by gravidity and interpregnancy interval to study risk as a function of time since induced abortion. We used logistic regression analyses to adjust for potential confounders such as number of previous spontaneous abortions and number of previous preterm or post-term deliveries. The Medical Birth Registry contains data on the best estimate of gestational age based on last menstrual period (LMP), corrected by ultrasound estimates.
Table 1 shows the main maternal and neonatal characteristics in the cohorts with respect to outcome of first pregnancies (index pregnancies). In controls, 4.1% of the live births were preterm, and 9.1% were post-term. Women in the abortion cohort were younger and more often lived in Copenhagen or other large cities. During follow-up, the highest gravidity and parity in controls were 11 and nine, respectively. Of 103,064 pregnancies, 12,780 ended in induced abortions; 18,915 occurred after those abortions (31,695), which were excluded. The 71,369 that remained in the control cohort were the informative pregnancies.
In the abortion cohort, 2755 women had 3432 repeat first-trimester induced abortions after index pregnancies, and consequently, 15,727 women had 19,159 first-trimester induced abortions before informative pregnancies. The cohort had 15,727 informative pregnancies during the 12-year follow-up.
In Table 2, the abortion cohort is stratified into three subcohorts according to number of induced abortions before informative pregnancies. In all subcohorts, the most commonly used method of abortion was dilation with vacuum aspiration. About 60% of induced abortions were done after 8 weeks' gestation.
Table 3 shows that preterm singleton live births were more frequent in all three abortion subcohorts than in controls of similar gravidity, and that women whose first pregnancies were terminated had higher risk of post-term deliveries.
Separate logistic regression analyses were used to adjust for maternal age and residence at the time of informative pregnancies, number of previous spontaneous abortions, number of previous preterm deliveries, and interpregnancy intervals. Table 4 shows the odds ratios (ORs) and 95% confidence interval (CI) for pre-term singleton live births stratified by number of previous induced abortions before the informative pregnancy, method of abortion, interpregnancy interval, and gravidity. Logistic regression analyses showed that ORs of preterm deliveries after induced abortions were significantly higher than controls, particularly in women whose first two pregnancies were terminated by dilation and evacuation. We further stratified analyses according to interpregnancy intervals and found an elevated OR of preterm delivery in women with interpregnancy intervals of more than 12 months.
To identify whether the increased ORs of preterm delivery were due to abortion method or late timing of induced abortion, we stratified data into induced abortions within 8 weeks' gestation and after 8 weeks' gestation. Within 8 weeks' gestation, previous abortion was associated with an increased OR of preterm delivery when time to informative pregnancy was longer than 12 months. One or two dilations with vacuum aspiration were related to an increased OR of preterm delivery when the time to pregnancy was longer than 6 months. After 8 weeks' gestation, women with one or two previous abortions had an increased risk of preterm delivery, particularly those whose time to informative pregnancy was longer than 12 months. The OR of preterm delivery after two dilations with evacuation was 27.35 (95% CI 8.48, 88.21). Very preterm birth (before 34 weeks' gestation) was more frequent in women who had induced abortions before the pregnancy under study. After adjusting for potential confounders and stratifying by gravidity, the ORs of very preterm singleton live births in women with one or more previous induced abortions were 1.99 (95% CI 1.64, 2.43), and 2.03 (95% CI 1.36, 3.04), respectively.
Table 5 shows the association between number of previous induced abortions and post-term singleton live births. In the logistic regression analyses, we included the number of previous post-term deliveries instead of the number of previous spontaneous abortions and preterm deliveries. In general, higher risk of post-term deliveries followed one, two, or more previous induced abortions. When data were stratified by abortion method, interpregnancy interval, and one or two vacuum aspirations, an increased risk of post-term delivery remained. Three or more induced abortions were associated with high risk of post-term delivery, particularly in women whose interpregnancy intervals were longer than 12 months.
We further stratified the risk of post-term delivery by time of last abortion. When the last induced abortion was within 8 weeks' gestation, a high risk of post-term delivery followed one or two induced abortions but only in women whose interpregnancy intervals were longer than 12 months. When the last induced abortion occurred after 8 weeks' gestation, the number of previous induced abortions was associated with a high risk of post-term delivery.
We restricted the cohorts to women who were certain of their LMPs according to reporting midwives. We found that adjusted ORs of preterm singleton live births in women with one, two, or more previous induced abortions were 1.89 (95% CI 1.67, 2.13), 2.61 (95% CI 1.98, 3.43), and 2.23 (95% CI 1.34, 3.72), respectively. Adjusted ORs of post-term singleton live births in women with one, two, or more previous induced abortions were 1.32 (95% CI 1.22, 1.43), 1.41 (95% CI 1.16, 1.71), and 1.54 (95% CI 1.03, 2.30), respectively.
We found an increased risk of preterm and post-term delivery after induced abortion but only in women with interpregnancy intervals of more than 12 months for preterm birth. An association between the number of abortions and preterm or post-term deliveries was found after evacuation. The post-term results were unexpected and it might be a chance association.
The main advantages of our study were its size and the fact that it was based on data from a uniformly organized health care system. It has a population-based design without differential recruitment or incomplete follow-up. The completeness of induced abortion recordings in Denmark is more than 90%.12 Weaknesses included lack of control for some confounding factors. Estimates of gestational ages were uncertain, regardless of method, but they were made early postpartum and without information on past induced abortion. Any misclassifications should be random and of similar magnitude in the two cohorts. When restricting the study to women who reported certain LMP dates, 83.8% and 85.9% in the abortion and control cohorts, respectively, estimates of the association between induced abortion and preterm or post-term birth were similar to those found for the entire cohorts.
Lang et al1 found an association with ORs of preterm singleton live births of 1.1, 1.9, and 3.6, respectively, following one, two, or more induced abortions after adjusting for 22 potential confounders. In one case-control study,2 previous induced abortion was associated with a modest, not statistically significant increased risk of preterm singleton live births (relative risk [RR] 1.6). That risk increased significantly with increasing previous induced abortions. Three reviews4–6 concluded that dilation and evacuation increased the risk of preterm delivery.
A short pregnancy interval is a risk factor for preterm delivery but only after a birth,5,9 and results stratified by interpregnancy interval are difficult to interpret. A short interval could indicate higher fecundity, which might be related to a low risk of preterm delivery; however, the interval depends on contraceptive method, and it is likely that controls used more effective methods than women in the abortion cohort (as indicated by the unwanted pregnancies). A short interpregnancy interval might also be a risk factor in itself. Depletion of essential nutrients might be more profound after a birth than after an induced abortion, and the risk of infection and other complications might vary in the two cohorts, so we chose to show interpregnancy interval-specific results because it modified the preterm but not the post-term deliveries.
Parity is associated with a preterm delivery,13 and gravidity is associated with outcome.14 Our study started with primigravidas, so all women had the same gravidity. We compared outcomes with similar gravidity, but most controls were of para 1, and all in the abortion group were of para 0. We did this because our para 0 group among controls by definition had subsequent pregnancies; therefore, an unbiased comparison based upon parity could not be done.
We could not control for prepregnancy weight,2,15 hypertension in pregnancy,16 maternal asthma,17 paternal occupation,18 smoking during pregnancy,17,19 mother's own small for gestational age birth,20 maternal race, socioeconomic status, placental abnormalities, gestational bleeding, and cervical incompetence,5 and it is possible that those factors fully or partly account for the outcomes reported.
1. Lang JM, Lieberman E, Cohen A. A comparison of risk factors for preterm labor and term small-for-gestational-age birth. Epidemiology 1996;7:369–76.
2. De Haas I, Harlow BL, Cramer DW, Frigoletto FD Jr. Spontaneous preterm birth: A case-control study. Am J Obstet Gynecol 1991; 165:1290–6.
3. Lekea-Karanika V, Tzoumaka-Bakoula C, Golding J. Previous obstetric history and subsequent preterm delivery in Greece. Eur J Obstet Gynecol Reprod Biol 1990;37:99–109.
4. Kramer MS. Determinants of low birth weight: Methodological assessment and meta-analysis. Bull World Health Organ 1987;65:663–737.
5. Berkowitz GS, Papiernik E. Epidemiology of preterm birth. Epidemiol Rev 1993;15:414–43.
6. Atrash HK, Hogue CJ. The effect of pregnancy termination on future reproduction. Baillieres Clin Obstet Gynaecol 1990;4:391–405.
7. Frank PI, Kay CR, Lewis TL, Parish S. Outcome of pregnancy following induced abortion. Report from the joint study of the Royal College of General Practitioners and the Royal College of Obstetricians and Gynaecologists. Br J Obstet Gynaecol 1985;92:308–16.
8. Meirik O, Bergstrom R. Outcome of delivery subsequent to vacuum-aspiration abortion in nulliparous women. Acta Obstet Gynecol Scand 1983;62:499–509.
9. Rawlings JS, Rawlings VB, Read JA. Prevalence of low birth weight and preterm delivery in relation to the interval between pregnancies among white and black women. N Engl J Med 1995;332:69–74.
10. The Danish National Board of Health. Sundhedsstatistikken [health statistics]. Copenhagen: DNBH, 1997:6.
11. The Danish National Board of Health. Sundhedsstatistikken [health statistics]. Copenhagen: DNBH, 1997:1.
12. Krebs L, Johansen AM, Helweg-Larsen K. Reporting of induced abortions in 1994. A comparison between the data in the Registry of Legally Induced Abortions and the National Patient Registry. Ugeskr Laeger 1997;159:1607–11.
13. Lumley J. The epidemiology of preterm birth. Baillieres Clin Obstet Gynaecol 1993;7:477–98.
14. Hogue CJ, Cates W Jr, Tietze C. Impact of vacuum aspiration abortion on future childbearing: A review. Fam Plann Perspect 1983;15:119–26.
15. Hickey CA, Cliver SP, McNeal SF, Goldenberg RL. Low pregravida body mass index as a risk factor for preterm birth: Variation by ethnic group. Obstet Gynecol 1997;89:206–12.
16. Ananth CV, Peedicayil A, Savitz DA. Effect of hypertensive diseases in pregnancy on birthweight, gestational duration, and small-for-gestational-age births. Epidemiology 1995;6:391–5.
17. Kelly YJ, Brabin BJ, Milligan P, Heaf DP, Reid J, Pearson MG. Maternal asthma, premature birth, and the risk of respiratory morbidity in schoolchildren in Merseyside. Thorax 1995;50:525–30.
18. Savitz DA, Brett KM, Dole N, Tse CK. Male and female occupation in relation to miscarriage and preterm delivery in central North Carolina. Ann Epidemiol 1997;7:509–16.
19. Wisborg K, Henriksen TB, Hedegaard M, Secher NJ. Smoking during pregnancy and preterm birth. Br J Obstet Gynaecol 1996; 103:800–5.
20. Klebanoff MA, Meirik O, Berendes HW. Second-generation consequences of small-for-dates birth. Pediatrics 1989;84:343–7.