Beruflich Dokumente
Kultur Dokumente
The Author 2008; all rights reserved. Advance Access publication 21 October 2008 doi:10.1093/ije/dyn167
REPRODUCTIVE HEALTH
1333
1334 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
paternity as well as for smoking and socioeconomic therapy, vaginal insemination, in vitro fertilization and
background. other methods. The variable was dichotomized and
The aim of the study was to assess the risk of pre- coded infertility treatment no and yes. The MoBa
eclampsia in primiparous women with a history of database contains no direct information on paternity
abortion as compared with primiparous women with- in pregnancies that ended in abortion. A proxy vari-
out such history using a large pregnancy cohort. able was constructed. The questionnaire contains
information on duration (in months and years) of
sexual cohabitation prior to conception. Women who
Materials and methods reported their relationship with the current partner to
Subjects have lasted longer than the time interval since the
previous pregnancy were considered to have the index
This study is a subproject in the Norwegian Mother pregnancy by the same partner. The categorization of
and Child Cohort Study (MoBa).12 MoBa is a cohort the confounders is shown in Table 1.
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 Statistical analyses
routine ultrasound examination in their local hospital. The relative risks of pre-eclampsia according to
A questionnaire is filled in at about week 15 of preg- abortion history were calculated as crude odds ratios
nancy. Questions are asked about the timing and (cOR) with 95% CI. Adjustment for confounding was
outcomes of all previous pregnancies as well as ques- obtained by estimating adjusted odds ratios (aOR)
tions about education, habits and health. The record in multiple logistic regression analyses (SPSS Inc,
in the Medical Birth Registry of Norway (MBRN)13 version 11, Chicago, IL, USA).
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 Results
Data Inspectorate. The present sample comprises
20 846 singleton pregnancies to primiparous women. A total of 5524 women (26%) reported one or more
previous abortions (Table 1). One or more sponta-
neous abortions and no induced abortions were
Variables reported by 12.8%, while 11.3% reported one or
The main outcome was pre-eclampsia in the present more induced abortions and no spontaneous ones.
pregnancy as registered in MBRN, including the Of the women, 2.3% reported both spontaneous and
HELLP syndrome and eclampsia. The recommended induced abortions. Women with a history of abortions
diagnostic criteria for pre-eclampsia in Norway are were slightly older and tended to have a higher pre-
blood pressure 4140/90 after 20 weeks gestation, pregnancy BMI than women with no history of
combined with proteinuria 41 dipstick on at least abortion (Table 1). Women with induced abortions
two occasions.14 were more likely to be daily smokers in their current
Information on parity (used to select the study pregnancy, while women with spontaneous abortions
sample), previous abortions (the main exposure) and more often reported treatment for infertility. Women
other variables are taken from the questionnaire sent with spontaneous abortions had shorter intervals
to the pregnant women at week 15 (www.fhi.no/ from time of abortion to birth and were less likely
morogbarn). Abortions were subdivided as induced to have a new partner as compared with women with
and spontaneous. All pregnancies that ended prior induced abortions.
to the 22nd gestational week were considered to be
abortions. Extra-uterine pregnancies were treated as Risk of pre-eclampsia
spontaneous abortions. Maternal age, pre-pregnancy
body mass index (BMI), smoking in pregnancy, edu- The risk of pre-eclampsia was 5.4% (1121/20 846), while
cational attainment (in years), the time interval since the risk of pre-eclampsia in primigravid women was
last pregnancy, infertility treatment prior to the index 5.5% (850/15 322) (Table 2). The risk of pre-eclampsia
pregnancy and change in paternity were considered in women with one previous abortion (spontaneous
potentially confounding factors. Pre-pregnancy height or induced) was 5.1% (216/4199). The risk of pre-
and body weight were used to calculate BMI kg/m2. eclampsia in women with two or more abortions
Smoking in pregnancy was coded no, occasionally (spontaneous or induced) was 4.2% (55/1325).
or daily. The time interval between the pregnancies
was calculated by subtracting the year of the previous Risk of pre-eclampsia according to
pregnancy from the year of birth in the present previous spontaneous abortions
pregnancy. Women with no previous abortions No significant difference in risk of pre-eclampsia was
(n 15 322) consequently had missing information observed in women with spontaneous abortions as
on interval. Data on infertility treatment prior to the compared with women without spontaneous abor-
index pregnancy included information on surgical tions (Table 2). Adjustment for maternal age, smok-
interventions, treatment for endometriosis, hormone ing in pregnancy, infertility treatment, pre-pregnancy
PREVIOUS ABORTIONS AND RISK OF PRE-ECLAMPSIA 1335
Pre-pregnancy BMI
<20 73.9 11.9 12.1 2.0 2748
2024 73.7 12.5 11.4 2.3 11 382
2529 71.6 14.0 11.7 2.6 4094
30 71.3 15.3 10.5 2.9 1822
No response 83.4 8.3 7.1 1.3 800
Smoking in pregnancy
Never 74.4 13.0 10.5 2.2 18 015
Occasionally 65.3 13.3 18.2 3.3 829
Daily 64.1 11.8 19.9 4.2 1417
No response 80.9 10.1 7.5 1.5 585
Infertility treatment
No 73.4 12.2 12.0 2.3 18 102
Yes 69.0 20.8 7.1 3.0 1318
No response 79.0 12.8 6.3 1.9 1426
Change in paternity
No 73.8 14.9 11.4 2615
Yes 24.5 68.8 6.7 2795
No response 48.2 46.5 5.3 114
Education (years)
<12 70.2 13.4 13.3 3.1 4261
12 70.0 13.0 14.2 2.8 2753
1316 75.0 12.8 10.2 2.0 8732
17 75.8 12.4 9.9 1.9 4477
No response 74.3 10.9 11.7 3.0 623
Distribution (percentages) of maternal age, pre-pregnancy BMI, smoking in pregnancy, pregnancy interval, infertility treatment,
change in paternity and length of education according to abortion history in 20 846 primiparous women in the MoBa.
1336 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Pre-eclampsia
No Yes Percentage cOR 95% CI aOR 95% CI
Number of previous abortions
0 14 472 850 5.5 1.0 1.0
1 3983 216 5.1 0.92 0.791.08 0.93 0.801.09
2 1270 55 4.2 0.74 0.560.97 0.73 0.550.97
Number of previous spontaneous abortions
0 16 731 956 5.4 1.0 1.0
1 2423 133 5.2 0.96 0.801.16 0.94 0.781.13
2 571 32 5.3 0.98 0.681.41 0.91 0.631.32
Number of previous induced abortions
0 16 955 997 5.5 1.0 1.0
1 2353 116 4.7 0.84 0.691.02 0.87 0.711.06
2 377 8 2.1 0.36 0.180.73 0.39 0.190.80
Total 19 725 1121 5.4
Relative risks (estimated as cOR and aOR with 95% CI) of pre-eclampsia according to the number of previous abortions among
20 846 primiparous women in the MoBa. Adjustment was made for maternal age, smoking in pregnancy, infertility treatment,
pre-pregnancy BMI and education.
Table 3 Risk of pre-eclampsia according to abortions controlled for pregnancy interval and paternity change
Pre-eclampsia
No Yes Percentage cOR 95% CI aOR 95% CI
Type and number of previous abortions
One induced 1952 99 4.8 1.0 1.0
Two or more induced 307 7 2.3 0.45 0.210.98 0.47 0.211.03
One spontaneous 2031 117 5.4 1.14 0.861.50 1.05 0.731.51
Two or more spontaneous 492 30 5.7 1.20 0.791.83 1.02 0.611.71
Total 4782 253 5.0
Relative risks (estimated as cOR and aOR with 95% CI) of pre-eclampsia according to type of abortion for 5035 primiparous women
participating in the MoBa who had previous abortions. Women with both spontaneous and induced abortions (n 489) were
excluded from the analyses. Adjustment was made for maternal age, smoking in pregnancy, infertility treatment, time interval
since last abortion, change in paternity, pre-pregnancy BMI and education.
BMI and education did not change the estimates. Adjusting for time interval and change in
The mean time interval from the last spontaneous paternity
abortion to birth was 2.2 years [standard error of the Analyses of women with previous abortions only
mean (SEM) 0.05]. (excluding primigravid women) allowed us to include
time interval since the last abortion and change in
paternity between the pregnancies to the list of
Risk of pre-eclampsia according to potentially confounding factors. After adjustment for
previous induced abortions these variables, in addition to the adjustment for
A slightly reduced risk of pre-eclampsia was maternal age, smoking, infertility treatment, BMI and
observed in women with one induced abortion, as education, women with one spontaneous abortion
compared with women without induced abortions had the same risk of pre-eclampsia as women with
(aOR 0.84, 95% CI 0.691.02). For women with one induced abortion (Table 3). The adjustment did
two or more induced abortions the risk of pre- not change the low risk of pre-eclampsia (aOR 0.47)
eclampsia was reduced by 460% (aOR 0.36, 95% CI in women with two or more induced abortions.
0.180.73) as compared with women without induced
abortions (Table 2). The mean time interval from the Additional analyses on changed paternity
last induced abortion to birth was 6.8 years Additional analyses were performed to explore
(SEM 0.09). the impact of changed paternity on pre-eclampsia
PREVIOUS ABORTIONS AND RISK OF PRE-ECLAMPSIA 1337
risk (Table 4). When compared with primigravid more induced abortions as compared with all other
women, one previous induced abortion tended to subgroups, including women with one induced abor-
reduce the risk of pre-eclampsia for women with tion. When compared with all women with singleton
same paternity pregnancies but not for women with pregnancies in the study population (n 47 304),
new paternity pregnancies, however the CIs were the risk of pre-eclampsia in primiparous women
wide due to small numbers. For both same paternity with two prior induced abortions was similar to the
and new paternity pregnancies two or more previous pre-eclampsia risk in women with one prior birth
induced abortions tended to reduce the risk of pre- (2.1% vs 2.5%). The pre-eclampsia risk in primiparous
eclampsia. The risk appeared to be lower in same women was 5.4%.
paternity as compared with new paternity pregnancies The risk of pre-eclampsia in women with two or
(Table 4). For women with two or more previous more induced abortions tended to be lower in both
spontaneous abortions, a change in paternity tended same paternity and new paternity pregnancies, as
to lower the estimated risk of pre-eclampsia as com- compared with primigravid women (1.4% and 2.5% vs
pared with spontaneous abortions with unchanged 5.5%, Table 4). Stratified analyses to further explore
paternity, and compared with primigravid women, the impact of changed paternity and gestational
but again the CIs were wide (Table 4). Adjustment for length at the time of the abortion on the pre-
potential confounders did not change the estimates. 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
Discussion for several confounders, including time interval since
The present study sample included 20 846 primiparous the last abortion and change in paternity. No previous
women selected from a study population of 51 960 study has corrected for pregnancy interval. The pres-
women. To our knowledge, no study of this magni- ent study used a prospective cohort design where
tude has previously been conducted to explore the information on previous pregnancies, smoking habits
effect of prior abortions on pre-eclampsia risk. and other exposures were collected before pre-
In accordance with the long established fact that eclampsia had developed clinically.
a previous birth is protective of pre-eclampsia in later Misclassification due to under- or over-reporting
pregnancies, the results from the present study show of prior abortions by the women in the study may
that this is true also for previous induced abortions in represent a potential weakness. However, the propor-
primiparous women. We found a 450% reduced risk tion of women in the study reporting one or more
of pre-eclampsia in primiparous women with two or previous spontaneous abortions (15.1%) is similar to
1338 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
the proportion of women registered in the MBRN risk if a primiparous woman has experienced any
(15.0%) with previous spontaneous abortion(s). abortion.
Nevertheless, the information on abortions both in Eras et al.,9 using records from obstetric practices for
MoBa and in the MBRN may be subject to under- 2739 births, reported a significant reduction in pre-
reporting leading to an underestimation of the effect eclampsia risk both for women who had experienced
of abortions on the pre-eclampsia risk. one spontaneous abortion and one induced abortion
Twice as many women in the induced abortion as compared with primigravidas. Beck11 reviewed case
group were daily smokers as compared with primi- records of 220 women who had had an abortion
gravid women or women with previous spontaneous before the first term pregnancy. He found that the
abortions. Smoking is associated with a reduced risk incidence of pre-eclampsia was significantly lower
of pre-eclampsia,15 thus residual confounding due to when there had been a previous induced abortion as
lack of control for the amount of cigarettes smoked compared with no previous abortion or previous spon-
by women with induced abortions could be argued taneous abortion. Also, the interval between the first
to explain the reduced risk in this group. Additional and second pregnancy was longer in women having
analyses adjusting for the number of cigarettes had a previous induced abortion as compared with
smoked did, however, not change the estimates. a previous spontaneous abortion, but there were no
For information on change in paternity between the difference in the mean time interval between normal
last abortion and the index pregnancy, a proxy vari- pregnancies and pregnancies complicated by pre-
able based on the duration of cohabitation was used. eclampsia. In a sample of 9771 women pregnant for
This represents an inaccuracy that may potentially the first or second time, Seidman and colleagues10
influence the results. In Table 4, we aimed to assess found a protective effect of a prior induced abor-
whether the reduced risk after two induced abortions tion after adjustment for maternal age, social class,
was seen both for women with the same partner ethnic origin and smoking, but no protective effect
and with a new partner. Random misclassification of of a previous spontaneous abortion. Campbell et al.16
paternity would tend to reduce any differences found that the incidence of pre-eclampsia after early
between these two groups. abortion (<13 weeks), either spontaneous or induced,
Some previous studies on the effect of abortions on was similar to the population incidence in a first
later pre-eclampsia risk have been published. Using a pregnancy. Thus, our results are consistent with the
dataset including 2947 healthy women from a clinical findings of Beck and with the findings of Seidman
trial of aspirin, Sibai et al.4 reported that primiparous and colleagues, but not with Campbells observation.
women with no earlier abortions had a pre-eclampsia The mechanisms behind the parity effect are not
risk of 5.7%, while women with one prior abortion understood. A relatively new observation is that the
had a risk of 4.7% and women with two earlier abor- inter-pregnancy interval may be a major determinant.
tions had a risk of 1.8%. They did not differentiate If the time since the previous birth is 11 years of
between spontaneous and induced abortions. In a more, the risk of pre-eclampsia in second births
similar trial, the Calcium for Preeclampsia Prevention appears to be as high as in a first birth.13 In our
(CPEP) trial, Sibai et al.5 studied 4314 primiparous study, the mean time interval from an induced abor-
women and reported that the risk of pre-eclampsia tion to birth was about 6 years, which may dilute
was 8.0%, 6.5% and 5.7% for women with none, one a protective effect. Women with prior induced abor-
and two earlier abortions, although these differences tions had a much longer time interval between the
were not statistically significant. Again, no differen- abortion and the subsequent delivery than women
tiation between induced and spontaneous abortions with prior spontaneous abortions. Nevertheless,
was made. Saftlas et al.6 also used the CPEP data set, adjustment for the time interval between the abortion
restricting the sample to 3242 primigravidas and 822 and the subsequent birth did not change the pre-
women with one previous abortion, and reported that eclampsia risk among women with same paternity
a substantial pre-eclampsia risk reduction (from 8.0% pregnancies, nor in new paternity pregnancies. Thus,
to 4.3%) occurred for the subgroup of women with the increased risk of pre-eclampsia with increasing
one abortion (without reporting effects of induced vs interval between the first and second birth demon-
spontaneous abortions) and the same father, whereas strated in a previous study1 could not be demon-
the risk was not reduced if a change of partner strated for abortions and subsequent pre-eclampsia
had taken place. The mean maternal age in the CPEP risk in the present study. The lack of association may
study was 22 years, limiting the possibility of studying perhaps be explained by differences in gestational age,
long pregnancy intervals. Xiong et al.7 used a large however small numbers restricted us from further
Canadian data set based on hospital records and analyses.
found that the risk of pre-eclampsia was 2.9% in Recent studies have reported that both maternal and
primigravid women and 2.6% in primiparous women paternal genes influence the risk of pre-eclampsia.
with a previous abortion, but could not differenti- Mothers and fathers themselves born from pre-
ate between induced and spontaneous abortions. eclamptic pregnancies contribute to an increased
Thus, all these studies suggest a certain reduction in risk in the next generation.1719 Although there was
PREVIOUS ABORTIONS AND RISK OF PRE-ECLAMPSIA 1339
a tendency that women with changed paternity suggested that subfecundity and pre-eclampsia may
were slightly less protected by two or more previous share the same etiology. Thus, the protective effect
induced abortions, and slightly more protected by two of a spontaneous abortion on pre-eclampsia risk in
or more previous spontaneous abortions as compared a subsequent pregnancy may be overshadowed by
with women with no change in paternity, these risk-increasing factors, and should be studied further.
findings were not significant and caution should be In summary, we have found a significant protective
taken to draw conclusions. Thus the impact of effect of having two or more induced abortions prior
changed paternity on the pre-eclampsia risk remains to the first birth on the risk of pre-eclampsia. The
unresolved. However, it seems likely that a change protective effect of two prior induced abortions is
in paternity may increase or decrease the risk of similar to the protective effect of one prior birth. The
pre-eclampsia, depending on the paternal genes time interval between the abortion and the subse-
transmitted to the foetus and their interaction with quent birth did not influence the results. The impact
maternal genes. This is opposed to the immune based of changed paternity remains unresolved. We specu-
primipaternity hypothesis suggesting the first preg- late that normal pregnancies that are interrupted in
nancy of any father will increase pre-eclampsia risk.20 early pregnancy may induce immunological changes
We have not been able to control for the genetic that reduce the risk of pre-eclampsia in a subsequent
predisposition to pre-eclampsia in the parenting men pregnancy, regardless of partner change.
and women in this study.
Although genetic factors inevitably are important in
the development of pre-eclampsia, they are unlikely to Funding
explain the reduced risk in second and later preg- This study was supported by The Norwegian Founda-
nancies. We speculate whether any pregnancy to a tion for Health and Rehabilitation (2002/2/0088 to
woman induces a memory which reduces the risk of L.T.); The Norwegian Research Council (151918/S10);
pre-eclampsia in a subsequent pregnancy, regardless National Institute of Environmental Health Sciences
of the parental genetic contributions which may also (N01-ES85433); National Institute of Neurological
alter the risk between pregnancies. This memory Diseases and Stroke (1 UO1 NS 047537); The sixth
may be hypothesized to decrease with increasing time Research Framework of the European Union
since the pregnancy, as reflected by the increased (EARNEST).
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 Acknowledgements
that ends in a spontaneous abortion. The underly- The efforts of all staff members in the Norwegian
ing pathology behind a spontaneous abortion may Mother and Child Cohort Study and at the Medical
in itself predispose to pre-eclampsia.21 We found Birth Registry of Norway are greatly appreciated.
that women with spontaneous abortions more often
had been treated for infertility. Basso et al.22 have Conflict of interest: None declared.
KEY MESSAGES
Having had two or more induced abortions halves the risk of pre-eclampsia in primiparous women.
The reduction in risk after two induced abortions is of the same magnitude as a previous full-term
pregnancy.
Having had spontaneous abortions does not change the risk.
References 4
Sibai BM, Gordon T, Thom E et al. Risk factors for
1 preeclampsia in healthy nulliparous women: a prospective
Trogstad LIS, Eskild A, Magnus P, Samuelsen SO,
Nesheim B-I. Changing paternity and time since last multicenter study. Am J Obstet Gynecol 1995;172:64248.
5
pregnancy; the impact on pre-eclampsia risk. A study of Sibai BM, Ewell M, Levine RJ et al. Risk factors associated
547 238 women with and without previous pre-eclamp- with preeclampsia in healthy nulliparous women. Am J
sia. Int J Epidemiol 2001;30:131722. Obstet Gynecol 1997;177:100310.
2 6
Basso O, Christensen K, Olsen J. Higher risk of pre- Saftlas AF, Levine RJ, Klebanoff MA et al. Abortion,
eclampsia after change of partner. An effect of longer changed paternity and risk of preeclampsia in nulliparous
interpregnancy intervals? Epidemiology 2001;12:62429. women. Am J Epidemiol 2003;157:110814.
3 7
Skjrven R, Wilcox AJ, Lie RT. The interval between Xiong X, Fraser WD, Demianczuk NN. History of abortion,
pregnancies and the risk of preeclampsia. N Engl J Med preterm, term birth and risk of preeclampsia: a population-
2002;346:3338. based study. Am J Obstet Gynecol 2002;187:101318.
1340 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
8 16
Abi-Said D, Annegers JF, Combs-Cantrell D, Campbell DM, MacGillivray I, Carr-Hill R. Pre-eclampsia
Frankowski RF, Willmore LJ. Case-control study of the in second pregnancy. BJOG 1985;92:13140.
risk factors for eclampsia. Am J Epidemiol 1995;142:43741. 17
Esplin MS, Fausett MB, Fraser A et al. Paternal and
9
Eras JL, Saftlas AF, Triche E, Hsu CD, Risch HA, maternal components of the predisposition to preeclamp-
Bracken MB. Abortion and its effect on risk of preeclampsia sia. N Engl J Med 2001;344:86772.
and transient hypertension. Epidemiology 2000;11:3643. 18
Lie RT, Brunborg H, Gjessing HK, Lie-Nilsen E,
10
Seidman DS, Ever-Hadani P, Stevenson DK, Gale R. The Irgens LM. Fetal and maternal contributions to risk of
effect of abortion on the incidence of pre-eclampsia. pre-eclampsia: population-based study. Br Med J 1998;
Eur J Obstet Gynecol Repr Biol 1989;33:10914. 316:134347.
11 19
Beck I. Incidence of pre-eclampsia in first full-term Skjrven R, Vatten LJ, Wilcox AJ, Rnning T, Irgens LM,
pregnancies preceded by abortion. J Obstet Gynaecol Lie RT. Recurrence of pre-eclampsia across generations:
1985;6:8284. exploring fetal and maternal genetic components in
12
Magnus P, Irgens LM, Haug K et al. Cohort profile: The a population based cohort. Br Med J 2005;331:88791.
Norwegian Mother and Child Cohort Study (MoBa). 20
Robillard P-Y, Dekker GA, Hulsey T. Revisiting the
Int J Epidemiol 2006;35:114650. epidemiological standard of preeclampsia: primigravidity
13
Irgens LM. The Medical Birth Registry of Norway. or primipaternity? Eur J Obstet Gynecol Reprod Biol
Epidemiological research and surveillance throughout 30 1999;84:3741.
years. Acta Obstet Gynecol Scand 2000;79:43539. 21
Merviel P, Carbillon L, Challier JC, Rabreau M,
14
Dalaker K, Berle EJ. The Norwegian Society of Obstetrics and Beaufils M, Uzan S. Pathophysiology of preeclampsia:
Gynecology; Clinical Guidelines in Obstetrics. Oslo, Norway: links with implantation disorders. Eur J Obstet Gynecol
Norwegian Medical Association, 1999. Reprod Biol 2004;115:13447.
15 22
Hammoud AO, Bujold E, Sorokin Y, Schild C, Krapp M, Basso O, Weinberg CR, Baird DD, Wilcox AJ, Olsen J.
Baumann P. Smoking in pregnancy revisited: findings Subfecundity as a correlate of preeclampsia: a Study
from a large population-based study. Am J Obstet Gynecol within the Danish National Birth Cohort. Am J Epidemiol
2005;192:185662. 2003;157:195202.