|Home | About | Journals | Submit | Contact Us | Français|
Background The risk of pre-eclampsia is reduced for second and later births. The causes and mechanisms behind this reduction are unknown. The aim of the study was to estimate the risk of pre-eclampsia in primiparous women according to history of spontaneous and induced abortions, while controlling for several potentially confounding factors.
Methods The sample consisted of 20 846 primiparous women participating in the Norwegian Mother and Child Cohort Study (MoBa). Information on abortions and confounders were self-reported in postal questionnaires. The diagnosis of pre-eclampsia was retrieved from the Medical Birth Registry of Norway. Estimation and confounder control was performed with multiple, logistic regression.
Results One previous induced abortion reduced the risk moderately [odds ratio (OR) 0.84, 95% confidence interval (CI) 0.69–1.02]. Two or more induced abortions reduced the risk more significantly (OR 0.36, 95% CI 0.18–0.73). Adjustment for confounders did not change the estimates.
Conclusions The protective effect of two prior induced abortions was similar to what is commonly seen after one birth. Spontaneous abortions may to a larger extent than induced abortions be associated with other factors, such as infertility, that may increase the risk of pre-eclampsia. Normal pregnancies interrupted in early pregnancy may induce immunological changes that reduce the risk of pre-eclampsia in a subsequent pregnancy.
A previous birth offers substantial protection against pre-eclampsia. The protective effect seems to abate with long intervals between the two pregnancies.1–3 The mechanisms underlying these observations have not been identified. Whether previous abortions offer any protection against pre-eclampsia in primiparous women is unresolved. Better knowledge of this issue may add to our understanding of the reduced risk among multiparous women, and may thereby provide new insight into the immunological basis of pre-eclampsia. Most studies that have addressed this question have been limited by small numbers. Also, many studies4–8 have not differentiated between previous spontaneous and previous induced abortions, which from both a social and pathophysiological point of view are very different conditions. Among the studies that have differentiated between induced and spontaneous abortions9,10 the results are conflicting. The time interval from the abortion to the next pregnancy that leads to a birth depends on the type of abortion. One study has corrected for pregnancy interval, but found no significant effect on pre-eclampsia risk.11 No previous studies have had the opportunity to correct for time intervals, change in paternity as well as for smoking and socioeconomic background.
The aim of the study was to assess the risk of pre-eclampsia in primiparous women with a history of abortion as compared with primiparous women without such history using a large pregnancy cohort.
This study is a subproject in the Norwegian Mother and Child Cohort Study (MoBa).12 MoBa is a cohort that aims to include 100 000 pregnancies by 2008. Pregnant women are recruited to the study through a postal invitation after they have signed up for the routine ultrasound examination in their local hospital. A questionnaire is filled in at about week 15 of pregnancy. Questions are asked about the timing and outcomes of all previous pregnancies as well as questions about education, habits and health. The record in the Medical Birth Registry of Norway (MBRN)13 from the present pregnancy is included as part of the data set. The study has been approved by the Regional Committee for Ethics in Medical Research and the Data Inspectorate. The present sample comprises 20 846 singleton pregnancies to primiparous women.
The main outcome was pre-eclampsia in the present pregnancy as registered in MBRN, including the HELLP syndrome and eclampsia. The recommended diagnostic criteria for pre-eclampsia in Norway are blood pressure >140/90 after 20 weeks gestation, combined with proteinuria >+1 dipstick on at least two occasions.14
Information on parity (used to select the study sample), previous abortions (the main exposure) and other variables are taken from the questionnaire sent to the pregnant women at week 15 (http://www.fhi.no/morogbarn). Abortions were subdivided as induced and spontaneous. All pregnancies that ended prior to the 22nd gestational week were considered to be abortions. Extra-uterine pregnancies were treated as spontaneous abortions. Maternal age, pre-pregnancy body mass index (BMI), smoking in pregnancy, educational attainment (in years), the time interval since last pregnancy, infertility treatment prior to the index pregnancy and change in paternity were considered potentially confounding factors. Pre-pregnancy height and body weight were used to calculate BMI = kg/m2. Smoking in pregnancy was coded ‘no’, ‘occasionally’ or ‘daily’. The time interval between the pregnancies was calculated by subtracting the year of the previous pregnancy from the year of birth in the present pregnancy. Women with no previous abortions (n = 15 322) consequently had missing information on interval. Data on infertility treatment prior to the index pregnancy included information on surgical interventions, treatment for endometriosis, hormone therapy, vaginal insemination, in vitro fertilization and other methods. The variable was dichotomized and coded infertility treatment ‘no’ and ‘yes’. The MoBa database contains no direct information on paternity in pregnancies that ended in abortion. A proxy variable was constructed. The questionnaire contains information on duration (in months and years) of sexual cohabitation prior to conception. Women who reported their relationship with the current partner to have lasted longer than the time interval since the previous pregnancy were considered to have the index pregnancy by the same partner. The categorization of the confounders is shown in Table 1.
The relative risks of pre-eclampsia according to abortion history were calculated as crude odds ratios (cOR) with 95% CI. Adjustment for confounding was obtained by estimating adjusted odds ratios (aOR) in multiple logistic regression analyses (SPSS Inc, version 11, Chicago, IL, USA).
A total of 5524 women (26%) reported one or more previous abortions (Table 1). One or more spontaneous abortions and no induced abortions were reported by 12.8%, while 11.3% reported one or more induced abortions and no spontaneous ones. Of the women, 2.3% reported both spontaneous and induced abortions. Women with a history of abortions were slightly older and tended to have a higher pre-pregnancy BMI than women with no history of abortion (Table 1). Women with induced abortions were more likely to be daily smokers in their current pregnancy, while women with spontaneous abortions more often reported treatment for infertility. Women with spontaneous abortions had shorter intervals from time of abortion to birth and were less likely to have a new partner as compared with women with induced abortions.
The risk of pre-eclampsia was 5.4% (1121/20 846), while the risk of pre-eclampsia in primigravid women was 5.5% (850/15 322) (Table 2). The risk of pre-eclampsia in women with one previous abortion (spontaneous or induced) was 5.1% (216/4199). The risk of pre-eclampsia in women with two or more abortions (spontaneous or induced) was 4.2% (55/1325).
No significant difference in risk of pre-eclampsia was observed in women with spontaneous abortions as compared with women without spontaneous abortions (Table 2). Adjustment for maternal age, smoking in pregnancy, infertility treatment, pre-pregnancy BMI and education did not change the estimates. The mean time interval from the last spontaneous abortion to birth was 2.2 years [standard error of the mean (SEM) 0.05].
A slightly reduced risk of pre-eclampsia was observed in women with one induced abortion, as compared with women without induced abortions (aOR 0.84, 95% CI 0.69–1.02). For women with two or more induced abortions the risk of pre-eclampsia was reduced by >60% (aOR 0.36, 95% CI 0.18–0.73) as compared with women without induced abortions (Table 2). The mean time interval from the last induced abortion to birth was 6.8 years (SEM 0.09).
Analyses of women with previous abortions only (excluding primigravid women) allowed us to include time interval since the last abortion and change in paternity between the pregnancies to the list of potentially confounding factors. After adjustment for these variables, in addition to the adjustment for maternal age, smoking, infertility treatment, BMI and education, women with one spontaneous abortion had the same risk of pre-eclampsia as women with one induced abortion (Table 3). The adjustment did not change the low risk of pre-eclampsia (aOR 0.47) in women with two or more induced abortions.
Additional analyses were performed to explore the impact of changed paternity on pre-eclampsia risk (Table 4). When compared with primigravid women, one previous induced abortion tended to reduce the risk of pre-eclampsia for women with same paternity pregnancies but not for women with new paternity pregnancies, however the CIs were wide due to small numbers. For both same paternity and new paternity pregnancies two or more previous induced abortions tended to reduce the risk of pre-eclampsia. The risk appeared to be lower in same paternity as compared with new paternity pregnancies (Table 4). For women with two or more previous spontaneous abortions, a change in paternity tended to lower the estimated risk of pre-eclampsia as compared with spontaneous abortions with unchanged paternity, and compared with primigravid women, but again the CIs were wide (Table 4). Adjustment for potential confounders did not change the estimates.
The present study sample included 20 846 primiparous women selected from a study population of 51 960 women. To our knowledge, no study of this magnitude has previously been conducted to explore the effect of prior abortions on pre-eclampsia risk.
In accordance with the long established fact that a previous birth is protective of pre-eclampsia in later pregnancies, the results from the present study show that this is true also for previous induced abortions in primiparous women. We found a >50% reduced risk of pre-eclampsia in primiparous women with two or more induced abortions as compared with all other subgroups, including women with one induced abortion. When compared with all women with singleton pregnancies in the study population (n = 47 304), the risk of pre-eclampsia in primiparous women with two prior induced abortions was similar to the pre-eclampsia risk in women with one prior birth (2.1% vs 2.5%). The pre-eclampsia risk in primiparous women was 5.4%.
The risk of pre-eclampsia in women with two or more induced abortions tended to be lower in both same paternity and new paternity pregnancies, as compared with primigravid women (1.4% and 2.5% vs 5.5%, Table 4). Stratified analyses to further explore the impact of changed paternity and gestational length at the time of the abortion on the pre-eclampsia risk were restricted by small numbers.
The strength of the present study, in addition to the large study sample, is the distinction between induced and spontaneous abortions and the ability to adjust for several confounders, including time interval since the last abortion and change in paternity. No previous study has corrected for pregnancy interval. The present study used a prospective cohort design where information on previous pregnancies, smoking habits and other exposures were collected before pre-eclampsia had developed clinically.
Misclassification due to under- or over-reporting of prior abortions by the women in the study may represent a potential weakness. However, the proportion of women in the study reporting one or more previous spontaneous abortions (15.1%) is similar to the proportion of women registered in the MBRN (15.0%) with previous spontaneous abortion(s). Nevertheless, the information on abortions both in MoBa and in the MBRN may be subject to underreporting leading to an underestimation of the effect of abortions on the pre-eclampsia risk.
Twice as many women in the induced abortion group were daily smokers as compared with primigravid women or women with previous spontaneous abortions. Smoking is associated with a reduced risk of pre-eclampsia,15 thus residual confounding due to lack of control for the amount of cigarettes smoked by women with induced abortions could be argued to explain the reduced risk in this group. Additional analyses adjusting for the number of cigarettes smoked did, however, not change the estimates.
For information on change in paternity between the last abortion and the index pregnancy, a proxy variable based on the duration of cohabitation was used. This represents an inaccuracy that may potentially influence the results. In Table 4, we aimed to assess whether the reduced risk after two induced abortions was seen both for women with the same partner and with a new partner. Random misclassification of paternity would tend to reduce any differences between these two groups.
Some previous studies on the effect of abortions on later pre-eclampsia risk have been published. Using a dataset including 2947 healthy women from a clinical trial of aspirin, Sibai et al.4 reported that primiparous women with no earlier abortions had a pre-eclampsia risk of 5.7%, while women with one prior abortion had a risk of 4.7% and women with two earlier abortions had a risk of 1.8%. They did not differentiate between spontaneous and induced abortions. In a similar trial, the Calcium for Preeclampsia Prevention (CPEP) trial, Sibai et al.5 studied 4314 primiparous women and reported that the risk of pre-eclampsia was 8.0%, 6.5% and 5.7% for women with none, one and two earlier abortions, although these differences were not statistically significant. Again, no differentiation between induced and spontaneous abortions was made. Saftlas et al.6 also used the CPEP data set, restricting the sample to 3242 primigravidas and 822 women with one previous abortion, and reported that a substantial pre-eclampsia risk reduction (from 8.0% to 4.3%) occurred for the subgroup of women with one abortion (without reporting effects of induced vs spontaneous abortions) and the same father, whereas the risk was not reduced if a change of partner had taken place. The mean maternal age in the CPEP study was 22 years, limiting the possibility of studying long pregnancy intervals. Xiong et al.7 used a large Canadian data set based on hospital records and found that the risk of pre-eclampsia was 2.9% in primigravid women and 2.6% in primiparous women with a previous abortion, but could not differentiate between induced and spontaneous abortions. Thus, all these studies suggest a certain reduction in risk if a primiparous woman has experienced any abortion.
Eras et al.,9 using records from obstetric practices for 2739 births, reported a significant reduction in pre-eclampsia risk both for women who had experienced one spontaneous abortion and one induced abortion as compared with primigravidas. Beck11 reviewed case records of 220 women who had had an abortion before the first term pregnancy. He found that the incidence of pre-eclampsia was significantly lower when there had been a previous induced abortion as compared with no previous abortion or previous spontaneous abortion. Also, the interval between the first and second pregnancy was longer in women having had a previous induced abortion as compared with a previous spontaneous abortion, but there were no difference in the mean time interval between normal pregnancies and pregnancies complicated by pre-eclampsia. In a sample of 9771 women pregnant for the first or second time, Seidman and colleagues10 found a protective effect of a prior induced abortion after adjustment for maternal age, social class, ethnic origin and smoking, but no protective effect of a previous spontaneous abortion. Campbell et al.16 found that the incidence of pre-eclampsia after early abortion (<13 weeks), either spontaneous or induced, was similar to the population incidence in a first pregnancy. Thus, our results are consistent with the findings of Beck and with the findings of Seidman and colleagues, but not with Campbell's observation.
The mechanisms behind the parity effect are not understood. A relatively new observation is that the inter-pregnancy interval may be a major determinant. If the time since the previous birth is 11 years of more, the risk of pre-eclampsia in second births appears to be as high as in a first birth.1–3 In our study, the mean time interval from an induced abortion to birth was about 6 years, which may dilute a protective effect. Women with prior induced abortions had a much longer time interval between the abortion and the subsequent delivery than women with prior spontaneous abortions. Nevertheless, adjustment for the time interval between the abortion and the subsequent birth did not change the pre-eclampsia risk among women with same paternity pregnancies, nor in new paternity pregnancies. Thus, the increased risk of pre-eclampsia with increasing interval between the first and second birth demonstrated in a previous study1 could not be demonstrated for abortions and subsequent pre-eclampsia risk in the present study. The lack of association may perhaps be explained by differences in gestational age, however small numbers restricted us from further analyses.
Recent studies have reported that both maternal and paternal genes influence the risk of pre-eclampsia. Mothers and fathers themselves born from pre-eclamptic pregnancies contribute to an increased risk in the next generation.17–19 Although there was a tendency that women with changed paternity were slightly less protected by two or more previous induced abortions, and slightly more protected by two or more previous spontaneous abortions as compared with women with no change in paternity, these findings were not significant and caution should be taken to draw conclusions. Thus the impact of changed paternity on the pre-eclampsia risk remains unresolved. However, it seems likely that a change in paternity may increase or decrease the risk of pre-eclampsia, depending on the paternal genes transmitted to the foetus and their interaction with maternal genes. This is opposed to the immune based primipaternity hypothesis suggesting the first pregnancy of any father will increase pre-eclampsia risk.20 We have not been able to control for the genetic predisposition to pre-eclampsia in the parenting men and women in this study.
Although genetic factors inevitably are important in the development of pre-eclampsia, they are unlikely to explain the reduced risk in second and later pregnancies. We speculate whether any pregnancy to a woman induces a ‘memory’ which reduces the risk of pre-eclampsia in a subsequent pregnancy, regardless of the parental genetic contributions which may also alter the risk between pregnancies. This ‘memory’ may be hypothesized to decrease with increasing time since the pregnancy, as reflected by the increased risk observed with long interpregnancy intervals.1 An induced abortion can in most instances be assumed to be a normal pregnancy in contrast to a pregnancy that ends in a spontaneous abortion. The underlying pathology behind a spontaneous abortion may in itself predispose to pre-eclampsia.21 We found that women with spontaneous abortions more often had been treated for infertility. Basso et al.22 have suggested that subfecundity and pre-eclampsia may share the same etiology. Thus, the protective effect of a spontaneous abortion on pre-eclampsia risk in a subsequent pregnancy may be overshadowed by risk-increasing factors, and should be studied further.
In summary, we have found a significant protective effect of having two or more induced abortions prior to the first birth on the risk of pre-eclampsia. The protective effect of two prior induced abortions is similar to the protective effect of one prior birth. The time interval between the abortion and the subsequent birth did not influence the results. The impact of changed paternity remains unresolved. We speculate that normal pregnancies that are interrupted in early pregnancy may induce immunological changes that reduce the risk of pre-eclampsia in a subsequent pregnancy, regardless of partner change.
This study was supported by The Norwegian Foundation for Health and Rehabilitation (2002/2/0088 to L.T.); The Norwegian Research Council (151918/S10); National Institute of Environmental Health Sciences (N01-ES–85433); National Institute of Neurological Diseases and Stroke (1 UO1 NS 047537); The sixth Research Framework of the European Union (EARNEST).
The efforts of all staff members in the Norwegian Mother and Child Cohort Study and at the Medical Birth Registry of Norway are greatly appreciated.
Conflict of interest: None declared.