Risk factors for 14–21 week abortions: a case-control study in Europe

Pierre-Yves Ancel1,4, Marie-Josèphe Saurel-Cubizolles1, Gian Carlo Di Renzo2, Emile Papiernik3, Gérard Bréart1 and The Europop Group*

1 Unité de Recherches Epidémiologiques en Santé Périnatale et Santé des Femmes (INSERM, Unité 149), 16 avenue Paul Vaillant-Couturier 94807 Villejuif cedex, France, 2 Institute of Obstetrics and Gynecology, University of Perugia, Perugia, Italy and 3 Université René Descartes, Service de Gynécologie Obstétrique de Port-Royal, 123 boulevard Port-Royal 75014 Paris, France


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Data from a case–control survey in Europe, carried out between 1994 and 1997, were used to investigate the risk factors for spontaneous abortions at 14–21 weeks (late abortions), according to the vital status of the fetus before the onset of labour. Late abortions included 62 involving a fetus alive before the onset of labour, 216 late abortions of a fetus already dead, together with 4592 control pregnancies at term (>=37 weeks) from seven countries. Histories of induced abortion, spontaneous abortion and preterm birth were more closely associated with late abortion of a live fetus than with late abortion of a dead fetus. Women aged >=35 years and women living alone had a much higher risk of late abortions than women aged 20–24 years and married women, regardless of the vital status of the fetus before labour. These results provide evidence that obstetric history and socio-demographic factors are risk factors for late abortions but differences are observed according to the vital status of the fetus before labour.

Key words: case-control study/European survey/fetal status/late abortion/risk factors


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Late abortions are rare events, accounting for ~10-30% of spontaneous abortions (Strobino et al., 1986Go; Lindbohm and Hemminki, 1988Go; Regan et al., 1989Go), that is ~1-2% of all pregnancies. However, they are traumatic and are a determinant risk factor for the outcome of subsequent pregnancies (Reginald et al., 1987Go; Thom et al., 1992Go). Most studies have considered spontaneous abortions as a whole and few have focused on late (second trimester) spontaneous abortions.

Several causes of spontaneous abortion have been identiified: uterine defects, immunological problems, hormonal imbalances, infections and fetal abnormalities (Strobino et al., 1986Go; Hay et al., 1994Go). Physiological mechanisms have not been thoroughly evaluated, but several factors are known to be associated with a high risk of spontaneous abortion: adverse obstetric history, previous pelvic inflammatory disease, advanced maternal age, specific occupations or exposure at work, and smoking habits (McDonald et al., 1978Go; Crenn Hebert et al., 1986Go; Goulet and Theriault, 1987Go; Lindbohm and Hemminki, 1988Go; Risch et al., 1988Go; Heisterberg et al., 1993Go). Although there have been few studies of late abortions, it has been established that chromosomal abnormalities are less frequent among late abortions than among first trimester abortions (Warburton et al., 1980Go), suggesting that the cause may be mainly of maternal origin.

It is unclear whether there are common risk factors for late spontaneous abortion and early preterm delivery. Knowledge of the risk factors for late abortion and early preterm birth may contribute to a better understanding of the mechanisms involved in early deliveries. It would also make it possible to test whether there is a continuous effect of several risk factors during pregnancy. Different types of aetiological mechanism may be involved in late abortion. We classified potential mechanisms into two broad categories. First, maternal disorders (uterine, vascular or infectious), similar to those reported to be risk factors for early preterm birth, may be involved in late abortion, decreasing the capacity of the uterus to support the fetus. Second, late abortion may have its origin in fetal disorders. We distinguished between these two types of situation by classifying late abortions according to the vital status of the fetus just before the onset of labour (alive or dead).

The aim of this study was to analyse risk factors for second trimester abortions, using data from a large European case–control study. We assessed the independent effects of obstetric history, socio-economic status, marital status, smoking habits, and maternal age on the occurrence of second trimester abortion according to the vital status of the fetus (alive or dead) immediately before the onset of labour.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Study design and subject recruitment
An unmatched case–control study, the Europop study, was conducted between 1994 and 1997 in 17 European countries, with a single common design used in each country (Saurel-Cubizolles et al., 1997Go). This survey included two case groups: all consecutive single preterm deliveries (22–36 completed weeks of amenorrhoea) and all consecutive single late abortions (14–21 completed weeks of amenorrhoea), and one unmatched control group (>=37 weeks of amenorrhoea) corresponding to one in 10 consecutive full-term single births. Classification was based on obstetric estimation of gestational age using ultrasound examination and the timing of the first day of the last menstrual period. Depending on the country, the duration of the survey in a given maternity unit was defined as the point at which data from 200 consecutive preterm births had been collected, or 12 months.

Only the late abortion and control groups were included in this analysis. Information about the vital status of the fetus before abortion was collected from medical records and was available for 278 (96.5%) of the eligible cases (n = 288). This analysis focused on spontaneous abortions, so those in which labour was induced with a live fetus that died during labour or immediately after birth were excluded. We retained the seven countries with 15 or more late abortions with a known mode of delivery (spontaneous or induced labour) and a known vital status of the fetus; the numbers of spontaneous late abortions by country were as follows: 19 in the Czech Republic, 34 in Germany, 20 in Hungary, 51 in Italy, 90 in Poland, 34 in Romania and 30 in Slovenia. These numbers were too small for comparisons to be made between countries. The analysis included 62 late abortions of a fetus alive before the spontaneous onset of labour, 216 late abortions of a dead fetus before the onset of labour (regardless of the mode of delivery: spontaneous or induced labour), and 4592 controls at term (>=37 completed weeks).

Data collection
Data were collected using the same questionnaire in all countries. The questionnaire, which has been published elsewhere (Saurel-Cubizolles et al., 1997Go), had two parts. The first was completed by interviewing the women after the delivery, during their stay in the maternity unit. It dealt with the demographic characteristics, the educational level, occupation and working conditions, marital status, and obstetric history of the women. The second part was completed from medical records and dealt with complications during pregnancy, cervical incompetence, uterine malformation, management of the delivery, and gestational age at the end of the pregnancy. Data concerning fetal malformation were also collected from medical records; detailed information about the procedures for ascertaining malformation was not available in all countries.

Obstetric history was classified as primigravida, multigravida with no previous adverse pregnancy outcome (abortion or preterm birth) (reference value), multigravida with previous spontaneous first trimester abortion, multigravida with previous spontaneous second trimester abortion, multigravida with previous preterm delivery, and multigravida with previous induced abortion. These groups were mutually exclusive; women with more than one previous adverse pregnancy outcome were assigned to groups in the following order of preference: previous spontaneous second trimester abortion, previous preterm birth, previous spontaneous first trimester abortion, previous induced abortion. Uterine malformations, cervical incompetence and fetal malformations were defined using dichotomous variables: 0 (absent) and 1 (present).

Three social factors were considered in this analysis: women's educational level, social class of the household and marital status. Educational level was determined from age at the end of schooling. The social class of the household recorded was that corresponding to the occupation of the woman or her partner (whichever was higher); in cases of single women, only the occupation of the woman was considered. In all countries, occupation was codified using the International Labour Office classification (ILO) (ISCO-88, 1991), five groups were defined: professionals (ILO code 1–2 for the man or 1–3 for the woman) or intermediate occupations (ILO code 3 for the man or 4 for the woman), service workers (ILO code 4–5 for the man or 5 for the woman), industrial workers (ILO code 6–9 for the man or the woman) and unemployed, meaning that neither the woman nor her partner had a job. Marital status was scored as married, unmarried and living with the child's father, or unmarried and living without a partner.

Mothers were assigned to one of three categories of tobacco use, on the basis of self-reported smoking habits in the last 3 months of pregnancy: non-smokers, 1–14 cigarettes per day and 15 or more cigarettes per day.

Statistical analysis
Late abortions of live and dead fetuses were compared to controls in terms of obstetric history, uterine malformations, cervical incompetence, fetal malformations, maternal age, social factors and smoking habits during pregnancy. Crude associations were assessed by odds ratios (OR), using polychotomous logistic regression models. Late abortions are rare events, so OR are good estimates of relative risks. We carried out a multivariate polychotomous logistic regression analysis including: maternal age, obstetric history, social class of the household, marital status, and tobacco use. For statistical reasons, in the multivariate analysis, maternal age was scored as: >20, 20–24, 25–34, >=35; and obstetric history was classified as primigravida or multigravida with no previous adverse pregnancy outcome (reference value), multigravida with previous spontaneous abortion (first or second trimester) or preterm delivery (class 1), and multigravida with previous induced abortion (class 2).

Odds ratios for late abortions of a live and dead fetuses were compared using the Wald test (Hosmer and Lemeshow, 1989Go), defined by: , where: bl and bd are the parameters estimated for late abortion of a live fetus and of a dead fetus, respectively, from the polychotomous logistic regression. If every parameter difference is zero, the Wald statistic follows a {chi}2 distribution (Hosmer and Lemeshow, 1989Go).

Odds ratios were systematically calculated after adjustment for the country of survey. All P values are for two-tailed tests and the significance level was set at 0.05. Sample sizes differed slightly between variables due to missing data. SAS (SAS Institute, 1990) and BMDP (Brown et al., 1990Go) statistical software packages were used.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The distribution of late abortions in our sample, as a function of gestational age, was: 14–15 completed weeks of amenorrhoea (18.3%), 16–17 weeks (25.5%), 18–19 weeks (26.6%) and 20–21 weeks (29.5%). The distributions for late abortion of live and dead fetuses did not differ significantly. The clinical context differed between late abortions and controls. Specifically, uterine malformation was associated with higher risk of late abortion of a live fetus and the association between uterine malformation and late abortion of a dead fetus was almost significant (Table IGo). Fetal malformations appeared to be more strongly associated with late abortion of a dead fetus (OR = 8.3, 95% CI, 4.6–14.9), than with late abortion of a live fetus (OR = 3.0, 95% CI, 0.7–12.9), but the difference in the OR was not significant. Cervical incompetence was strongly associated with late abortion of a live fetus (OR = 12.3, 95% CI, 6.2–24.2) but not with late abortion of a dead fetus. Previous adverse pregnancy outcomes were strongly associated with both late abortion groups. The OR for the association with previous late spontaneous abortion (P value for heterogeneity in OR <0.05) and previous induced abortion (P value for heterogeneity in OR <0.20) were higher for late abortion of a live fetus than for late abortion of a dead fetus.


View this table:
[in this window]
[in a new window]
 
Table I. Late abortions, medical factors and obstetric history
 
Women aged >=40 years had a much higher risk of late abortions of a live fetus (OR = 8.8, 95% CI, 2.5–30.7) or of a dead fetus (OR = 6.8, 95% CI, 3.7–12.4) than women aged 20–24 years (Table IIGo). Women living alone were at least three times more likely than married women to suffer late abortion, independent of the vital status of the fetus, and unemployment was associated with a doubling in the risk of late abortion. No difference was observed in OR associated with maternal age, marital status and social class of the household according to the vital status of the fetus before labour. A low level of education was associated with a trebling in the risk of late abortion of a live fetus, whereas no association was observed with late abortion of a dead fetus (P value for heterogeneity in OR = 0.05). The association between smoking >=15 cigarettes per day with late abortion of a live fetus (OR = 2.7, 95% CI, 0.9–7.9) was not quite significant, whereas that with late abortion of a dead fetus was not significant (OR = 1.7, 95% CI, 0.8–3.8); however, no significant difference was observed between these two OR.


View this table:
[in this window]
[in a new window]
 
Table II. Late abortions and socio-demographic factors
 
In multivariate analysis, the factors independently associated with late abortion after adjustment were: previous spontaneous abortion or preterm birth, advanced maternal age (>=35 years) and living alone (Table IIIGo). The OR for the association of late abortion of a live fetus with previous abortion, preterm birth and previous induced abortion were higher than those for associations of these factors with late abortion of a dead fetus (P value for heterogeneity in OR < 0.05). Low socio-economic status, defined as a service worker, industrial worker or unemployed household, tended to be associated with a doubling in the risk of late abortion of a live fetus, but this association was significant only for industrial workers. No difference in OR was observed for maternal age, marital status and social class of the household. Smoking habits were not associated with late abortion after adjustment for covariates.


View this table:
[in this window]
[in a new window]
 
Table III. Late abortions, obstetric history and socio-demographic factors: multivariate analyses
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Few data have been published concerning risk factors for late abortions. The advantages of the data set used here are that it is large, for such rare outcome, with standardized recording of social and medical characteristics for the case group and the controls (at term births). This analysis provides new results, comparing two different groups of late abortions defined in terms of vital status of the fetus just before the onset of labour. As expected, obstetric history was a strong predictor of late abortion. We found that the associations with obstetric history were stronger for late abortion of a live fetus than for a late abortion of a dead fetus. Advanced maternal age and living without a partner were strongly associated with late abortion, regardless of the vital status of the fetus before labour. Low socio-economic status was associated with a higher risk of late abortion, particularly for abortion of a live fetus. Heavy smoking was not significantly associated with late abortion after adjustment for covariates, regardless of what the vital status of the fetus was.

Few late abortions were recorded in most of the 17 participating countries, suggesting that case recruitment was incomplete in these countries. These countries also included larger numbers of controls. We decided to exclude countries with fewer than 15 abortions in order to prevent case selection and over-representation of the controls in the sample. Abortions of live fetuses following the induced onset of labour were excluded because this study concerned the aetiology of spontaneous abortions. Early abortions are under-represented in most hospital-based studies because the symptoms are not always recognized (Lindbohm and Hemminki, 1988Go). In contrast, women who suffer abortions between 14 and 21 weeks of gestation are more likely to be treated in hospital because of the marked symptoms at later stages of gestation. In our sample, fewer abortions were reported between 14 and 15 weeks than were reported later in the pregnancy, whereas a decrease in the true prevalence with gestational age might have been expected (Modvig et al., 1990Go). This suggests that early second trimester abortions may be under-represented in our sample. In addition, the participating maternity units were included on a voluntary basis and thus the controls may not be representative of the general population. More pregnancy complications than expected from the general population may be included if these maternity units tend to recruit patients at risk. However, it is difficult to determine whether these selection biases (for cases and controls), if indeed they exist, modified the associations between late abortions and the various risk factors. Lastly, due to small sizes of the samples from individual countries, differences between countries were not considered in this analysis. For methodological reasons, all estimates were adjusted for country. This adjustment made it possible to produce unbiased associations by including possible confounding effects due to differences in demographic and social structures between countries.

Cervical incompetence was highly associated with late abortion of live fetuses. Cervical incompetence is a well-known proximal risk factor for spontaneous abortion and preterm birth (Strobino et al., 1980Go; Berkowitz and Papiernik, 1993Go). The lack of association with late abortion of a dead fetus was expected because cervical incompetence does not directly cause fetal death. Lastly, we found a strong association between late abortion of a dead fetus and fetal abnormalities. These results suggest that cervico-uterine abnormalities are strongly involved in late abortions in which the fetus is alive before the onset of labour, whereas fetal malformation is a major risk factor for fetal death. We did not include these two factors in the multivariate analysis, because they are more proximal in the causal pathway and may be consequences of some of the other factors. For instance, cervical incompetence would have been diagnosed after previous late abortion or early preterm birth. The risk of overadjustment was therefore high and we chose to estimate associations with general predictors without taking these clinical factors into account.

Previous pathological events in the patient's obstetric history may have been more accurately reported by cases than by controls. Our results are consistent with those reported in a previous prospective study (Regan et al., 1989Go). As expected, previous adverse pregnancy outcomes were highly associated with late abortion, and previous late spontaneous abortion itself was the most predictive factor. It has been reported that recurrent spontaneous abortions frequently concern chromosomally normal fetuses (Strobino et al., 1986Go; Kline et al., 1989aGo). Thus, the stronger association with late abortion of a live fetus (more frequently chromosomally normal) than with late abortion of a dead fetus (more frequently chromosomally abnormal) was not unexpected. Moreover, in our data, the association between late abortion of a live fetus and previous spontaneous abortion or preterm birth decreased after adjusting for cervical incompetence (data not shown), suggesting a possible causal relationship between cervical abnormalities and the recurrence of late abortions. Previous induced abortion was significantly associated with late abortion of a live fetus whereas no association was observed with late abortion of a dead fetus. It has been suggested that induced abortion increases the risk of subsequent spontaneous abortion, by damaging the cervix or uterus (Kline et al., 1989bGo), but there is still some debate over its precise role (Rowland Hogue et al., 1982Go).

We found a significant association between maternal age and late abortion, independent of the vital status of the fetus, as previously reported for spontaneous abortions as a whole (Rish et al., 1988). Ageing may affect the quality and viability of the embryo. As the frequency of abortion with chromosome abnormalities rises with maternal age (Stein, 1985Go), the higher risk of late abortion for older women in our series may partly reflect fetal chromosome abnormalities resulting in fetal death. An impairment in uterine function, in the capacity of the uterus to support the fetus (Stein, 1985Go) or involving uteroplacental hypoperfusion (Naeye, 1983Go), may also increase the risk of late abortion. However, women prone to spontaneous abortion have to become pregnant more often than low-risk women to obtain a live child, so they are more likely to be pregnant when they are older. This fertility selection results in an apparent increase in risk with age, whereas the excess risk may actually be due to specific processes independent of age (Modvig et al., 1990Go). Nevertheless, in our analysis, the effect of maternal age was well adjusted for previous adverse pregnancy outcomes. Thus, our results are consistent with advanced maternal age contributing to the occurrence of late abortion.

The relationship between social status and spontaneous abortion is unclear. A low level of education has been reported to be associated with spontaneous abortion (Parazzini et al., 1997Go), but some studies have reported no association with socio-economic status (Rachootin and Olsen, 1982Go; Parazzini et al., 1991Go). Early spontaneous abortions are probably more frequently medically identified in women with earlier access to medical care, due to earlier detection of the pregnancy, and these women are probably in higher socio-economic categories. In our analysis restricted to late abortions, the risk was higher for women in low socio-economic groups. Moreover, a low level of education and low socio-economic status seemed to be more closely associated with late abortions of a live fetus than with late abortion of a dead fetus. Although unemployment was associated with a doubling in the risk of late abortion of a dead fetus in the univariate analysis, this association was not significant after adjustment. This suggests that marital status, maternal age and obstetric history partly account for the univariate association. Further analyses, with greater statistical power, focused on behaviour, living and working conditions, are required to explain this link. These results show that poor social conditions are a risk factor, albeit non-proximal for late abortions.

Unmarried women living without a partner had a much higher risk of late abortion than other women, regardless of the vital status of the fetus before the onset of labour. Stress due to the social conditions of most of these women (poverty, unwanted pregnancy, loneliness, physical abuse) may be a pathogenic factor. Misclassification is also possible in this group: women may themselves induce the abortions and then present at the hospital, where obstetric staff diagnose spontaneous abortion. No data are available concerning this issue.

A crude, though not significant, association, was observed between heavy smoking during pregnancy and late abortion for fetuses alive before the onset of labour. This is consistent with previous studies reporting a significant effect of high tobacco consumption (>=20 cigarettes per day), but a weak association or no association for moderate levels of consumption (Armstrong et al., 1992Go; Windham et al., 1992Go; Kline et al., 1995Go). However, heavy smoking was not associated with late abortion after adjustment for covariates. These findings may be interpreted in several ways. First, the crude association may be partly accounted for by other risk factors. Second, previous studies provide some evidence of an association between smoking during pregnancy and abortion of chromosomally normal fetuses (Kline et al., 1995Go), suggesting that the effect of smoking is probably easiest to detect with viable fetuses (Kline et al., 1989bGo). This is consistent with our results, because the frequency of fetal abnormalities was lower in the live fetuses aborted late than in the group with dead fetuses. Assuming a causal link with smoking, low levels of exposure in this population may not have shown the expected moderately higher risk due to a lack of statistical power. The data analysed were obtained by retrospective interview. It is possible that women who experienced late abortion under-reported their smoking habits more than controls, leading to an underestimation of the effect of smoking.

Our findings are consistent with late abortion and early preterm birth having common risk factors, as advanced maternal age and previous adverse pregnancy outcomes have also been reported to increase the risk of early preterm birth (Hoffman and Bakketeig, 1984Go; Ancel et al., 1999Go). Furthermore, previous late abortions are considered to be an independent risk factor for early preterm birth (Hoffman and Bakketeig, 1984Go; Ancel et al., 1999Go). This suggests that common underlying mechanisms may be involved in both these outcomes of pregnancy.

Thus, our results provide evidence that obstetric history and social factors are risk factors for late abortion. We were also able to determine the strength of the associations and to show differences according to the vital status of the fetus just before the onset of labour. We are aware that this classification is somewhat arbitrary, but, as few reports have been published concerning the risk factors for late abortion, our results provide valuable new information. The roles of defective uteroplacental perfusion and infections require further investigation to improve our understanding of theses two types of late abortion. The strong association between marital status and late abortion suggests that underlying factors, such as poor social or psychological conditions, may be involved in the observed relationship, and suggest that clinicians should take such results into account in their preventive strategies.

Appendix. Composition of the Europop group
Members of the steering committee
Project leader: Di Renzo, G.C., Perugia, Italy.

Bréart, G., MD, Paris, France.

Papiernik, E., MD, Paris, France.

Patel, N., MD, Dundee, United Kingdom.

Saurel-Cubizolles, M.J., PhD, Villejuif, France.

Taylor, D., MD, Leicester, United Kingdom.

Todini, S., MSc, Perugia, Italy.

Members of the national staffs
Czech Republic: Kudela, M., MD, Vetr, M., MD, in Olomouc.

Finland: Heikkilä, A., MD, Erkkola, R., MD, Forström, J., MD, in Turku.

France: Papiernik, E., MD, Lucidarme, P., midwife, in Paris; Tafforeau, J., MD, in Brussels.

Germany: Künzel, W., MD, Herrero-Garcia, J., MD, in Giessen; Dudenhausen, J., MD, Henrich, W., MD, in Berlin.

Greece: Antsaklis, A., MD, Haritatos, G., MD, in Athens.

Hungary: Kovacs, L., MD, Nyari, T., MD, Bartfai, G., MD, in Szeged.

Ireland: O'Herlihy, C., MD, Murphy, J., MD, Stewart, H., in Dublin.

Italy: Di Renzo, G.C., MD, Bruschettini, P.L., MD, Moscioni, P., MD, in Perugia; Cosmi, E., MD, Spinelli, A., MSc, Donati, S., MD, in Rome.

Poland: Breborowicz, G.H., MD, Anholcer, A., MD, in Poznan.

Romania: Stamatian, F., MD, in Cluj.

Russia: Mikhailov, A.V., MD, in St Petersburg.

Slovenia: Pajntar, M., MD, Pirc, M., MD, Verdenik, I., MD, in Ljubljana.

Spain: Escribà-Aguir, V., MD, in Valencia; Carrera, J.M., MD, in Barcelona.

Sweden: Marsal, K., MD, Stale, H., MD, in Malmö.

The Netherlands: Buitendijk, S., Ph.D, van der Pal, K., MSc, in Leiden; van Geijn, H., MD, in Amsterdam.

Turkey: Gökmen, O., MD, Güler, C., MD, Caglar, T., MD, in Ankara.

United Kingdom: Owen, P., MD, in Dundee.

Members of the epidemiological analysis staff
Ancel, P.-Y., MD, France.

Bréart, G., MD, France.

Lelong, N., MSc, France.

Saurel-Cubizolles, M.-J., PhD, France.


    Acknowledgments
 
We would like to thank Monique Kaminski for helpful comments, the interviewers for data collection and the technical teams for computer input of data in all countries. This European study was funded by the European Union (contract BMH1-CT94-1041). Dr Pierre-Yves Ancel is funded by the National Research Agency in France.


    Notes
 
* For details of members of Europop Group see appendix Back

4 To whom correspondence should be addressed at: INSERM Unité 149, 16 avenue Paul Vaillant-Couturier, 94807 Villejuif cedex, France. E-mail: ancel{at}vjf.inserm.fr Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Ancel, P.Y., Saurel-Cubizolles, M.J., Di Renzo, G.C. et al. (1999) Social differences of very preterm birth in Europe: Interaction with obstetric history. Am. J. Epidemiol., 149, 908–915.[Abstract]

Armstrong, B.G., McDonald, A.D. and Sloan, M. (1992) Cigarette, alcohol, and coffee consumption and spontaneous abortion. Am. J. Public Health, 82, 85–87.[Abstract]

Berkowitz, G.S. and Papiernik, E. (1993) Epidemiology of preterm birth. Epidemiol. Rev., 15, 414–443.[ISI][Medline]

Brown, M.R., Engelman, L. and Jenrich, R.I. (1990) BMDP Statistical Software Manual, Vol. 2. University of California Press, Berkeley.

Crenn Hebert, C., Bouyer, J., Collin, D. et al. (1986) Spontaneous abortion and interpregnancy interval. Eur. J. Obstet. Gynecol. Reprod. Biol., 22, 125–132.[ISI][Medline]

Goulet, L. and Theriault, G. (1987) Association between spontaneous abortion and ergonomic factors. A literature review of the epidemiologic evidence. Scand. J. Work Environ. Health, 13, 399–403.[ISI][Medline]

Hay, P.E., Lamont, R.F. and Taylor-Robinson, D. (1994) Abnormal bacterial colonisation of the genital tract and subsequent preterm delivery and late miscarriage. Br. Med. J., 308, 295–298.[Abstract/Free Full Text]

Heisterberg, L. (1993) Factors influencing spontaneous abortion, dyspareunia, and pelvic pain. Obstet. Gynecol., 81, 594–597.[Abstract]

Hoffman, H.J. and Bakketeig, L.S. (1984) Risk factors associated with the occurrence of preterm birth. Clin. Obstet. Gynecol., 27, 539–552.[ISI][Medline]

Hosmer, D.W. and Lemeshow, S. (1989) Applied Logistic Regression. Wiley, New York.

International Standard Classification of Occupations (1991) ISCO-88. International Labour Office, Geneva.

Kline, J., Stein, Z. and Susser, M. (eds) (1989a) The recurrence of miscarriage and chromosomal aberrations. In Conception to Birth. Epidemiology of Prenatal Development. Oxford University Press, New York, pp. 146–161.

Kline, J., Stein, Z. and Susser, M. (eds) (1989b) Environmental Factors, miscarriage and karyotypes. In Conception to Birth. Epidemiology of Prenatal Development. Oxford University Press, New York, pp. 118–145.

Kline, J., Levin, B., Kinney, A. et al. (1995) Cigarette smoking and spontaneous abortion of known karyotype. Am. J. Epidemiol., 141, 417–427.[Abstract]

Lindbohm, M.L. and Hemminki, K. (1988) Nationwide data base on medically diagnosed spontaneous abortions in Finland. Int. J. Epidemiol., 17, 568–573.[Abstract]

McDonald, A.D., McDonald, J.C., Armstrong, B. et al. (1978) Occupation and pregnancy outcome. Br. J. Indust. Med., 44, 521–526.

Modvig, J., Schmidt, L. and Damsgaard, M.T. (1990) Measurement of total risk of spontaneous abortion: the virtue of conditional risk estimation. Am. J. Epidemiol., 132, 1021–1038.[Abstract]

Naeye, R.L. (1983) Maternal age, obstetric complications, and the outcome of pregnancy. Obstet. Gynecol., 61, 210–216.[Abstract]

Parazzini, F., Chatenoud, L., Tozzi, L. et al. (1997) Determinants of risk of spontaneous abortions in the first trimester of pregnancy. Epidemiology, 8, 681–683.[ISI][Medline]

Parazzini, F., Bocciolone, L., Fedele, L. et al. (1991) Risk factors for spontaneous abortion. Int. J. Epidemiol., 20, 157–161.[Abstract]

Rachootin, P. and Olsen, J. (1982) Prevalence and socioeconomic correlates of subfecundity and spontaneous abortion in Denmark. Int. J. Epidemiol., 11, 245–249.[Abstract]

Regan, L., Braude, P.R. and Trembath, P.L. (1989) Influence of past reproductive performance on risk of spontaneous abortion. Br. Med. J., 299, 541–545.[ISI][Medline]

Reginald, P.W., Beard, R.W., Chapple, J. et al. (1987) Outcome of pregnancies progressing beyond 28 weeks gestation in women with a history of recurrent miscarriage. Br. J. Obstet. Gynaecol., 94, 643–648.[ISI][Medline]

Risch, H.A., Weiss, N.S., Clarke, E.A. et al. (1988) Risk factors for spontaneous abortion and its recurrence. Am. J. Epidemiol., 128, 420–430.[Abstract]

Rowland Hogue, C.J., Cates, W. and Tietze, C. (1982) The effects of induced abortion on subsequent reproduction. Epidemiol. Rev., 4, 66–94.[ISI][Medline]

SAS (1990) SAS/STAT User's Guide, Version 6, 4th edn, Vol. 1. SAS Institute Inc., Cary, NC.

Saurel-Cubizolles, M.J., Di Renzo, G.C., Bréart, G. and the EUROPOP Group (1997) Women's work and preterm birth: epidemiological knowledge and description of a European project. Prenat. Neonat. Med., 2, 161–180.

Stein, Z.A. (1985) A woman's age: childbearing and child rearing. Am. J. Epidemiol., 121, 327–342.[ISI][Medline]

Strobino, B.R., Kline, J., Shrout, P. et al. (1980) Recurrent spontaneous abortion: definition of a syndrome. In Porter, I.H. and Hook, E.B. (eds), Human Embryogenic and Fetal Death. Academic Press, New York, pp. 315–329.

Strobino, B., Fox, H.E., Kline, J. et al. (1986) Characteristics of women with recurrent spontaneous abortions and women with favorable reproductive histories. Am. J. Public Health, 76, 986–991.[Abstract]

Thom, D.H., Nelson, L.M., Vaughan, T.L. et al. (1992) Spontaneous abortion and subsequent adverse birth outcomes. Am. J. Obstet. Gynecol., 166, 111–116.[ISI][Medline]

Warburton, D., Stein, Z., Kline, J. et al. (1980) Chromosome abnormalities in spontaneous abortion: data from the New York City Study. In Porter, I.H. and Hook, E.B. (eds), Human Embryogenic and Fetal Death. Academic Press, New York, pp. 261–287.

Windham, G.C., Swan, S.H. and Fenster, L. (1992) Parental cigarette smoking and the risk of spontaneous abortion. Am. J. Epidemiol., 135, 1394–1403.[Abstract]

Submitted on April 10, 2000; accepted on July 31, 2000.





This Article
Abstract
FREE Full Text (PDF )
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (3)
Request Permissions
Google Scholar
Articles by Ancel, P.-Y.
Articles by Europop Group, T.
PubMed
PubMed Citation
Articles by Ancel, P.-Y.
Articles by Europop Group, T.