Sites of ectopic pregnancy: a 10 year population-based study of 1800 cases

J. Bouyer1,4, J. Coste1, H. Fernandez2, J.L. Pouly3 and N. Job-Spira1

1 INSERM U569-IFR69, Hôpital de Bicêtre, 82 avenue du Général Leclerc, 94276 Le Kremlin Bicêtre, 2 Hôpital Antoine Béclère, Service de Gynécologie Obstétrique, 92141 Clamart and 3 CHU Hôtel-Dieu, Service de Gynécologie Obstétrique, BP 69-63003 Clermont-Ferrand, France


    Abstract
 Top
 Abstract
 Introduction
 Population and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: Several risk factors for ectopic pregnancy (EP) have been identified, but the site of implantation of EP has been little studied. METHODS: A total of 1800 surgically treated EP was registered between January 1992 and December 2001 in the Auvergne EP register and the women concerned were followed up. In this large population-based sample, we studied the distribution of EP sites, immediate complications, determining factors, and subsequent fertility. RESULTS: EP sites were interstitial (2.4%), isthmic (12.0%), ampullary (70.0%), fimbrial (11.1%), ovarian (3.2%) or abdominal (1.3%). No cervical pregnancies were observed. Complications and treatment depended on the site of EP. In multivariate analysis, the only risk factor associated with EP site was current use of an intrauterine device (IUD), which was more frequent in distal EP. The 2 year cumulative rate of subsequent spontaneous intrauterine pregnancy (IUP) increased progressively from interstitial to ovarian EP. Fair concordance (weighted {kappa} = 0.31) was observed between the sites of two successive EP if they were homolateral. CONCLUSION: In addition to providing an accurate description of the sites of implantation of EP, this study shows that current IUD use ‘protects’ against interstitial pregnancies, which are the most difficult to manage. It shows that subsequent fertility tends to be higher in women with distal EP.

Key words: ectopic pregnancy/fertility/localization/population-based/recurrence


    Introduction
 Top
 Abstract
 Introduction
 Population and methods
 Results
 Discussion
 Acknowledgements
 References
 
During the 1980s and 1990s, the incidence of ectopic pregnancy (EP) in developed countries increased by a factor of three to four (Weström et al., 1981Go; Atrash et al., 1986Go; Chow et al., 1987Go; Mäkinen, 1987Go; Anonymous, 1995Go) to reach 100 to 175 per 1000000 women aged 15–44 years. A decline in EP incidence has recently been observed in certain countries (Mäkinen, 1996Go; Coste et al., 2000Go), but it remains unclear whether this trend also applies elsewhere, and could be attributed to decreases in genital infection rates (Egger et al., 1998Go) or changes in contraceptive behaviour (Coste et al., 2000Go).

Several risk factors for EP have been identified (Chow et al., 1987Go; Coste et al., 1991Go; Fernandez et al., 1991aGo; Job-Spira et al., 1993Go): pelvic inflammatory disease (PID)—especially for infections involving Chlamydia trachomatis—smoking, previous pelvic surgery, previous EP, past and current use of an intrauterine device (IUD).

The site of implantation of EP has been little studied, at least in large population-based samples. Only a few articles, published >15 years ago, report descriptive results concerning the site of EP in large samples either from populations defined geographically, (Hallatt, 1982Go; Hallatt and Grove, 1985Go; Atrash et al., 1986Go; Herbertsson et al., 1987Go) or from hospital-based populations (Martin et al., 1988Go; Al-Meshari et al., 1993Go). The other studies on this subject are case reports or deal with very small series, with a review of previous studies. Moreover, all these papers focus on a specific site, none considering implantation of EP as a whole. To our knowledge, only one article has suggested a possible link between EP site and subsequent fertility (Pouly et al., 1991Go) or EP recurrence, although improvements in the diagnosis and management of EP have shifted clinicians’ concerns away from the immediate health of the woman, towards preserving her subsequent fertility.

The site of EP implantation merits study as it affects the severity of the condition, and the immediate and delayed side-effects. In addition, the effect of the site of implantation on subsequent fertility should be investigated.

The aim of this study was to investigate the distribution of EP sites in a population-based sample, and its variation over time. We then aimed to study the immediate complications and factors determining the site of EP. Finally, we determined the rates of subsequent fertility and EP recurrence.


    Population and methods
 Top
 Abstract
 Introduction
 Population and methods
 Results
 Discussion
 Acknowledgements
 References
 
Study population
The methodology of the Auvergne EP register has been described elsewhere (Coste et al., 1994Go). Briefly, a register was established in January 1992 in three départements (districts) of the Auvergne region in the centre of France (Allier, Cantal and Puy-de-Dôme). All women between 15 and 44 years of age who live permanently in the target area and who had had either surgical or medical treatment for EP were registered and followed up until the age of 45 years, all reproductive outcomes being recorded. In each medical centre in the area (15 public or private maternity hospitals and 12 surgical units) a trained investigator, either a midwife or a physician, was responsible for case identification, follow-up and data collection. The basic information collected for each woman included socio-demographic characteristics, sexual, gynaecological, reproductive and surgical history, conditions at conception (use of contraception, ovulation induction, IVF), smoking habits, results of serological tests for Chlamydia trachomatis, characteristics of the EP (site, tubal rupture, haemoperitoneum), and the treatment given. The completeness of the register has been estimated to be ~90% (Coste et al., 1994Go, 2000Go).

The women were interviewed about reproductive events, by telephone, every 6 months after EP for 2 years, and then once per year until the age of 45 years. We asked questions about the desire for a new pregnancy, cumulative period of trying to become pregnant, pregnancies achieved, treatments for infertility and use of contraception.

This study was conducted on the 1933 EPs registered between January 1992 and December 2001. Women with medical treatment only (n = 133) were excluded from the study because the site of implantation could not be determined with certainty. Among the remaining 1800 subjects, the site of EP was determined by the operating surgeon and was known for 1679 subjects (93%), providing precise information concerning the distribution of EP implantation for this population-based sample. Six sites were recorded: interstitial part of the tube, isthmus, ampulla, fimbrium, ovary and abdomen.

For studies of subsequent fertility, we excluded the EP cases recorded in 2001 because we did not have enough time to record all the data for the first follow-up call for these cases. We also excluded 159 women who had undergone sterilization for contraception or therapeutic bilateral salpingectomy (with no desire for IVF). In addition, 133 women were lost to follow-up (8.7%). Thus, the analysis concerned 1228 women.

Factors studied and statistical methods
As very little is known about the factors determining the site of EP, we investigated the association between EP site and most of the known risk factors for EP. The treatment of EP was defined as radical if the tube in which the pregnancy occurred was removed, i.e. salpingectomy. Treatment was otherwise considered to be conservative. All French clinicians use similar guidelines when deciding between radical and conservative treatment (Pouly et al., 1991Go). Haemoperitoneum of >100 ml was classified as profuse (Fernandez et al., 1991cGo). We tested possible associations between EP site and these factors by means of {chi}2-tests and, when appropriate, {chi}2-tests for trend (Armitage and Berry, 1987Go) to take into account the ordered nature of the sites. Multivariate analyses were performed by multinomial logistic regression (Hosmer and Lemeshow, 2000Go).

Two reproductive outcomes were used to evaluate fertility: the recurrence of EP and intrauterine pregnancy (IUP). In both cases, survival analysis methods were used with person-time being the time to pregnancy. This is the cumulative period of time during which a woman was trying to become pregnant until she became pregnant or was censored. Cumulative rates and confidence intervals (CI) were calculated by the Kaplan–Meier estimator (Breslow and Day, 1987Go) for each of the sites. The curves obtained were compared by log-rank tests for univariate analysis, and by Cox regression (Cox and Oakes, 1990Go) to take into account confounding variables. Ties in time to pregnancy data were handled using a published method (Efron, 1977Go). The confounding factors had been identified as being associated with subsequent fertility either in our previous analyses (Job-Spira et al., 1996Go; Bernoux et al., 2000Go) or by other teams (Thorburn et al., 1988Go; Yao and Tulandi, 1997Go): age, educational level, prior tubal damage, history of infertility, tubal rupture and treatment (conservative or radical). For women with several successive pregnancies, the time to pregnancy was calculated until the first occurrence of the type of pregnancy studied (EP or IUP). Only ‘natural’ fertility was studied, and follow-up was censored if a woman began an IVF programme or had declared that she was not trying to become pregnant again (Bernoux et al., 2000Go). In terms of methodology, there is no entirely satisfactory way to take IVF into account because this event is not independent of a woman’s fertility (IVF is prescribed for less fertile women). The removal of women undergoing IVF from the sample analysed would therefore result in an overestimate of fertility. Moreover, such exclusion is always incomplete because we cannot identify women who will undergo IVF treatment in the future. In our study, we censored follow-up only when the woman actually underwent IVF. Thus, we did not exclude the entire follow-up period for women undergoing IVF, limiting the overestimation of infertility.

In cases of EP recurrence, we investigated whether the sites of the two successive EP were identical or close to one another, by calculating weighted {kappa}-statistics (Fleiss, 1981Go). {kappa}-statistics are parameters of agreement that take chance agreement into account (agreement or concordance between the sites of the two EP in this study). The weighted {kappa}-statistic is based on the idea that if two sites differ by more than one category, then the disagreement should be given more weight than if they differ by only one category. Quadratic (dis)agreement weights were used. Values of {kappa} >0.8 indicate almost perfect agreement, values of 0.6–0.8 indicate substantial agreement, with 0.4–0.6 indicating moderate agreement, 0.2–0.4 fair agreement, and <0.2 poor agreement (Landis and Koch, 1977Go).

Statistical analyses were performed with STATA software (StataCorp, 2001Go).


    Results
 Top
 Abstract
 Introduction
 Population and methods
 Results
 Discussion
 Acknowledgements
 References
 
Distribution of EP sites
The distribution of the sites of ectopic pregnancy is given in Table IGo. Only 4.5% extratubal EPs (ovarian and abdominal) were observed and about three-quarters of the tubal pregnancies (1175/1603 = 73%) were ampullary. No cervical pregnancies were observed. The 95% confidence interval (CI) of the proportion of EPs accounted for by cervical pregnancies is thus 0–0.0022. Thus, cervical EPs are probably very rare, accounting for <=1 in 455 EP. The distribution of implantation sites varied over time (P = 0.04). This variation was accounted for mostly by increases in the proportion of interstitial and isthmic EP. These increases were also accounted for by a decrease in the number of women using IUD at the time of EP (as shown below, IUD use is associated with EP site). Similar results were obtained if we excluded the period 2000–2001, in which the proportion of medical treatments was higher (and thus the proportion of undetermined sites of EP was higher).


View this table:
[in this window]
[in a new window]
 
Table I. Distribution of the sites of ectopic pregnancy (EP): Auvergne register (France) 1992–2001
 
Clinical features and management
The proportion of cases of EP diagnosed before 6 weeks gestation increased significantly from interstitial to abdominal EP sites (Table IIGo). This trend remained significant after adjustment for tubal rupture or haemoperitoneum, the frequency of which also depended on the site of EP. Adjustment for haemoperitoneum also made it possible to take into account, at least partly, possible errors in gestational age due to metrorrhagia early in the pregnancy. Concentrations of hCG, an indicator of pregnancy activity, decreased progressively from interstitial to abdominal EP sites, and this trend remained significant after adjustment for pregnancy term at the time of diagnosis.


View this table:
[in this window]
[in a new window]
 
Table II. Clinical features and management of ectopic pregnancy (EP) according to site: Auvergne EP register (France) 1992–2001
 
The proportion of EP initially managed conservatively increased progressively from interstitial to abdominal EP sites. Moreover, the increase over time in the proportion of EP treated conservatively was larger for interstitial sites than for any other location. The first treatment was considered to have failed if the clinician decided to treat the woman again for the same EP, either medically or surgically. Differences in the rates of failure of the first treatment were accounted for mostly by a high proportion of failure in interstitial (14.6%) and fimbrial (11.3%) EP.

Results similar to those in Table IIGo were obtained if the main characteristics of the women (such as age, smoking, previous sexually transmitted disease, current IUD use and educational level) were taken into account as potential confounders.

Determinant factors
The relationship between EP site and the characteristics of the woman is shown in Table IIIGo. The association with current use of an IUD was particularly strong: the proportion of IUD users increased from interstitial to ovarian sites (in the population studied, about one woman in three was a current IUD user). However, the distribution of EP sites was similar in women who had used an IUD only in the past and women who had never used an IUD.


View this table:
[in this window]
[in a new window]
 
Table III. Distribution of the potential determinant factors of the site of ectopic pregnancy (EP): Auvergne EP register (France) 1992–2001
 
A history of damaged tubes (as shown by a history of EP or tubal surgery) was more frequent in women with proximal implantation than in those with distal implantation. This association was similar regardless of the side of the body on which previous EP had occurred (contralateral or homolateral). The other factors studied, including those not shown in Table IIIGo (educational level, previous induced abortion, history of infertility and induction of the pregnancy), were not associated with EP site.

In multivariate analysis, the only factor that remained significantly associated with EP site was current use of an IUD, although prior EP was also close to the significance threshold.

Subsequent fertility
For the 1228 women followed up, mean follow-up time was 3.2 years, and 693 women had tried to become pregnant again. It was not possible to study subsequent fertility among women with an abdominal pregnancy because only nine of these women had tried to become pregnant again. Of the remaining 684 women, 78 experienced another EP. The 2 year cumulative rate of recurrent EP was 0.22 (95% CI 0.17–0.28). No significant differences in EP recurrence rates were observed according to the site of the current EP (Table IVGo).


View this table:
[in this window]
[in a new window]
 
Table IV. Subsequent fertility according to the site of the index ectopic pregnancy (EP) (among women trying to become pregnant again): Auvergne EP register (France) 1992–2001
 
The 2 year cumulative rate of IUP was 0.76 overall [95% CI (0.71–0.80)], and increased progressively from interstitial to ovarian EP sites (Table IVGo and Figure 1Go), with a test for trend giving a P-value close to the threshold of statistical significance (P = 0.08). However, this crude trend was partly accounted for by confounding (adjusted test for trend P = 0.38), in particular by the larger proportion, among cases of ovarian and fimbrial EP, of women with an IUD, whose subsequent fertility is known to be better (Bernoux et al., 2000Go).



View larger version (14K):
[in this window]
[in a new window]
 
Figure 1. Cumulative intrauterine pregnancy rates according to the site of the ectopic pregnancy (EP) (among women trying to become pregnant again): Auvergne EP register (France) 1992–2001.

 
If we considered all the women (whether or not they had tried to become pregnant), 113 had had another EP. The sites of the two successive pregnancies were known for 99 women and two of these pairs of EP included one abdominal EP. Fair concordance was observed between the sites of implantation if the two EPs were homolateral (weighted {kappa} = 0.31; P < 0.01; Table VGo). Conversely, no concordance was observed if the two EPs were contralateral (weighted {kappa} = –0.03; P = 0.58). The women with homolateral recurrent EP were slightly older than those with contralateral recurrent EP. However, we found no significant differences for any of the other characteristics of the women, including infectious history.


View this table:
[in this window]
[in a new window]
 
Table V. Sites of the two successive ectopic pregnancies (EP) among women with recurrent EP: Auvergne EP register (France) 1992–2001
 

    Discussion
 Top
 Abstract
 Introduction
 Population and methods
 Results
 Discussion
 Acknowledgements
 References
 
Almost all the women living in the Auvergne region who were treated for EP between January 1992 and June 2001 were included in this study, the completeness of the Auvergne EP register being estimated to be ~90% (Coste et al., 1995Go, 2000Go). The characteristics of the index EP were collected prospectively. Women were also interviewed every 6 months after EP about their seeking to become pregnant again and fertility outcome, and few (8.7%) were lost to follow-up. Thus, selection and recall biases, which frequently occur in retrospective studies with hospital recruitment, were probably very limited. The proportion of medical treatments increased with time. Therefore, the number of EP for which the site was not determined also increased, and was especially large in the last period studied (2000–2001). However, we obtained similar results if we excluded this period.

Some abdominal pregnancies are not primary abdominal pregnancies, but result instead from ruptured tubal or ovarian pregnancies (Hallatt and Grove, 1985Go; Costa et al., 1991Go; Dover and Powell, 1995Go). Even if the primary site of implantation of such pregnancies was known (for instance fimbrium), we classified them as abdominal because they could not be considered to be similar to other fimbrial EP: even if only secondarily, their implantation was abdominal (Martin et al., 1988Go). We therefore retained the ‘abdominal’ category and studied the characteristics of EP for all six sites.

Few articles describing the distribution of all EP implantation sites have been published. Moreover, the distribution is often cited with respect to all pregnancies, rather than with respect to EP, and may therefore vary according to the proportion of EP among pregnancies. Most of the remaining papers are devoted to specific sites (mainly interstitial, ovarian or abdominal EP), often studied from hospital-based data. However, the published figures are consistent with those reported here: 3–11% for interstitial EP (Fernandez et al., 1991bGo; Kabukoba and de Courcy-Wheeler, 1992Go; Khalifa et al., 1994Go; Tulandi et al., 1995Go); 1–6% for ovarian EP (Gérin-Lajoie, 1951Go; Boronow et al., 1965Go; Hallatt, 1982Go; Grimes et al., 1983Go; Steiner et al., 1983Go; Herbertsson et al., 1987Go; Sandvei et al., 1987Go; Cabero et al., 1989Go; Raziel et al., 1990Go; Al-Meshari et al., 1993Go; Chahtane et al., 1993Go); 0.9–1.4% for abdominal EP (Radman, 1978Go; Delke et al., 1982Go; Atrash et al., 1987Go).

An increase in the incidence of ovarian EP over time has been reported (Steiner et al., 1983Go; Sandvei et al., 1987Go; Cabero et al., 1989Go; Raziel et al., 1990Go). Some authors account for this increase by improvements in diagnosis (Cabero et al., 1989Go; Raziel et al., 1990Go). This is consistent with our observations because the incidence of ovarian EP increased until 1997, with the proportion of ovarian EP remaining fairly steady thereafter. Another possible interpretation is based on the role of IUD in interfering with the transport of ova from the ovary into the tube and variations in IUD use over time (Herbertsson et al., 1987Go). This interpretation is not supported by our observations because IUD use has decreased in the Auvergne region since 1992 (Coste et al., 2000Go).

We observed that the increase over time in the proportion of EP treated conservatively was greater for interstitial EP than for EP at other sites. This is consistent with the fact that before the 1990s, interstitial pregnancy was treated by cornual resection or hysterectomy probably due to late diagnosis in many cases (Parente et al., 1983Go; Tulandi et al., 1995Go). Since then, advances in transvaginal sonography and in medical treatment with methotrexate have opened up new avenues for conservative treatment by laparoscopy (Fernandez et al., 1991bGo).

In multivariate analysis, the only factor significantly associated with the site of EP was current contraception with an IUD, which was more frequent in distal EP cases. This is consistent with previous findings (McMorries et al., 1979Go; Hallatt, 1982Go; Sandvei et al., 1987Go; Raziel et al., 1990Go). The IUD may cause mild inflammation, resulting in deciliation of the endosalpinx and therefore delayed ovum transport, leading to ectopic implantation (Herbertsson et al., 1987Go). This finding both confirms the role of IUD in the aetiology of EP (Bouyer et al., 2000Go), and suggests that there may be tubal factors involved in the aetiology of ovarian pregnancy (Sandvei et al., 1987Go). IUD seem to play a role only in the short term, because no association with the site of EP was found if the woman had used an IUD only in the past.

The higher proportion of women with markers of tubal damage, such as previous EP or previous tubal surgery, that we found in proximal EP cases is consistent with previous reports (Boronow et al., 1965Go; Herbertsson et al., 1987Go; Cabero et al., 1989Go; Raziel et al., 1990Go; Yarali et al., 1994Go; Lau and Tulandi, 1999Go), but this association was found to be weaker after adjustment for other determinant factors. Moreover, similar results were obtained regardless of the site of the previous EP (homo- or contralateral), suggesting that tubal surgery per se did not determine the location of a subsequent EP.

The crude difference in subsequent fertility observed between sites was close to the threshold of statistical significance and became non-significant if conservative treatment was considered. It is therefore possible that subsequent fertility does not depend on the EP site. However, as treatment was more frequently conservative in cases of distal EP, we cannot rule out the possibility that treatment plays the role of an intermediate factor between EP site and subsequent fertility. This would imply that the site itself plays a role, as previously suggested for ovarian pregnancies (Boronow et al., 1965Go; Hallatt, 1982Go; Raziel et al., 1990Go).

The concordance between the sites of recurrent homolateral EP, but not between the sites of contralateral EP, is interesting. As the infectious histories of women with homo- and contralateral EP were similar, this result suggests that surgical treatment of the tube may induce recurrence at the same location. In cases of conservative management of EP, this, in turn, may be an argument for favouring medical treatment whenever possible.

This study is the first to describe the sites of implantation of EP in a large population-based sample. It confirms that the complications of EP vary considerably according to site. It shows that current IUD use ‘protects’ against interstitial pregnancies, which are the most difficult type of EP to manage. Finally, it shows that subsequent fertility tends to be higher for distal and extratubal EP and that there is fair concordance between the sites of two successive EP if they are homolateral.


    Acknowledgements
 Top
 Abstract
 Introduction
 Population and methods
 Results
 Discussion
 Acknowledgements
 References
 
We thank Julie Sappa for her careful review of the English version of this paper. This study was supported by the National Register Committee (Comité National des Registres–INSERM–InVS), France.


    Notes
 
4 To whom correspondence should be addressed. E-mail: bouyer{at}vjf.inserm.fr Back


    References
 Top
 Abstract
 Introduction
 Population and methods
 Results
 Discussion
 Acknowledgements
 References
 
Al-Meshari, A.A., Chowdhury, N. and Adelusi, B. (1993) Ovarian pregnancy. Int. J. Gynecol. Obstet., 41, 269–272.[ISI][Medline]

Anonymous (1995) Ectopic pregnacy—United States, 1970–1992. Morbid. Mortal. Wk. Rep., 44, 46–48.

Armitage, P. and Berry, G. (1987) Statistical Methods in Medical Research. Blackwell, Oxford.

Atrash, H.K., Hughes, J.M. and Hogue, C.J.R. (1986) Ectopic pregnancy in the United States, 1970–1983, CDC surveillance summaries. Morbid. Mortal. Wk. Rep., 35, 29–37.

Atrash, H.K., Friede, A. and Hogue, C.J.R. (1987) Abdominal pregnancy in the United States. Frequency and maternal mortality. Obstet. Gynecol., 69, 333.[Abstract]

Bernoux, A., Job-Spira, N., Germain, E., Coste, J. and Bouyer, J. (2000) Fertility outcome after ectopic pregnancy and use of an intrauterine device at the time of the index ectopic pregnancy. Hum. Reprod., 15, 1173–1177.[Abstract/Free Full Text]

Boronow, R., McElin, T.W., West, R.H. and Buckingham, J.C. (1965) Ovarian pregnancy. Report of four cases and a thirteen-year survey of the English literature. Am. J. Obstet. Gynecol., 91, 1095–1106.[ISI][Medline]

Bouyer, J., Rachou, E., Germain, E., Fernandez, H., Coste, J., Pouly, J.L. and Job-Spira, N. (2000) Risk factors for extrauterine pregnancy in women using an intrauterine device. Fertil. Steril., 74, 899–908.[ISI][Medline]

Breslow, N.E. and Day, N.E. (1987) Statistical Methods in Cancer Research. Volume II. The Design and Analysis of Cohort Studies. Oxford University Press, London.

Cabero, A., Laso, E., Lain, J.M., Manas, C., Escribano, I. and Calaf, J. (1989) Increasing incidence of ovarian pregnancy. Eur. J. Obstet. Gynecol. Reprod. Biol., 31, 227–232.[ISI][Medline]

Chahtane, A., Dehayni, M., Rhrab, B., Kharbach, A., El Amrani, S. and Chaoui, A. (1993) La grossesse ovarienne. A propos de 4 observations avec revue de la littérature [Ovarian pregnancy: based upon four cases, with a review of the literature]. Rev. Fr. Gynécol. Obstet., 88, 35–38.[Medline]

Chow, W.H., Daling, J.R., Cates, W. and Greenberg, R.S. (1987) Epidemiology of ectopic pregnancy. Epidemiol. Rev., 9, 70–94.[ISI][Medline]

Costa, S.D., Presley, J. and Bastert, G. (1991) Advanced abdominal pregnancy. Obstet. Gynecol. Surv., 46, 515–525.[Medline]

Coste, J., Job-Spira, N., Fernandez, H., Papiernik, E. and Spira, A. (1991) Risk factors for ectopic pregnancy: a case–control study in France, with special focus on infectious factors. Am. J. Epidemiol., 133, 839–849.[Abstract]

Coste, J., Job-Spira, N., Aublet-Cuvelier, B., Germain, E., Glowaczover, E., Fernandez, H. and Pouly, J.L. (1994) Incidence of ectopic pregnancy. First results of a population-based register in France. Hum. Reprod., 9, 742–745.[Abstract]

Coste, J., Aublet-Cuvelier, B., Bouyer, J., Germain, E. and Job-Spira, N. (1995) Evaluation de l’exhaustivité du registre des grossesses extra-utérines d’Auvergne par la méthode capture-recapture [Evaluation of the completeness of the Auvergne register of ectopic pregnancy with the capture–recapture method]. Rev. Epidemiol. Santé Publique, 43 (Suppl. 1), 10.

Coste, J., Bouyer J., Germain E., Ughetto S., Pouly J.L. and Job-Spira, N. (2000) Recent declining trend in ectopic pregnancy in France: evidence of two clinicoepidemiologic entities. Fertil. Steril., 74, 881–886.[ISI][Medline]

Cox, D.R. and Oakes, D. (1990) Analysis of Survival Data. Chapman & Hall, London.

Delke, I., Veridiano, N.P. and Tancer, M.L. (1982) Abdominal pregnancy: review of current management and addition of 10 cases. Obstet. Gynecol., 60, 200.[Abstract]

Dover, R.W. and Powell, M.C. (1995) Management of a primary abdominal pregnancy. Am. J. Obstet. Gynecol., 172, 1603–1604.[ISI][Medline]

Efron, B. (1977) The efficiency of Cox’s likelihood function for censored data. J. Am. Stat. Assoc., 76, 312–319.

Egger, M., Low, N., Smith, G.D., Linblom, B. and Herrmann, R. (1998) Screening for chlamydial infections and the risk of ectopic pregnancy in a county in Sweden: ecological analysis. Br. Med. J., 316, 1776–1780.[Abstract/Free Full Text]

Fernandez, H., Coste, J. and Job-Spira, N. (1991a) Controlled ovarian hyperstimulation as a risk factor for ectopic pregnancy. Obstet. Gynecol., 78, 656–659.[Abstract]

Fernandez, H., De Ziegler, D., Bourget, P., Feltain, P. and Frydman, R. (1991b) The place of methotrexate in the management of interstitial pregnancy. Hum. Reprod., 6, 302–306.[ISI][Medline]

Fernandez, H., Lelaidier, C., Thouvenez, V. and Frydman, R. (1991c) The use of a pretherapeutic, predictive score to determine inclusion criteria for the non-surgical treatment of ectopic pregnancy. Hum. Reprod., 6, 995–998.[Abstract]

Fleiss, J.L. (1981) Statistical Methods for Rates and Proportions. Wiley, New York.

Gérin-Lajoie, L. (1951) Ovarian pregnancy. Am. J. Obstet. Gynecol., 62, 920–929.[ISI][Medline]

Grimes, H.G., Nosal, R.A. and Gallagher, J.C. (1983) Ovarian pregnancy: a series of 24 cases. Obstet. Gynecol., 61, 174–180.[Abstract]

Hallatt, J.G. (1982) Primary ovarian pregnancy: a report of twenty-five cases. Am. J. Obstet. Gynecol., 143, 55–60.[ISI][Medline]

Hallatt, J.G. and Grove, J.A. (1985) Abdominal pregnancy: a study of twenty-one consecutive cases. Am. J. Obstet. Gynecol., 152, 444–449.[ISI][Medline]

Herbertsson, G., Magnusson, S. and Benediktsdottir, K. (1987) Ovarian pregnancy and IUCD use in a defined complete population. Acta Obstet. Gynecol. Scand., 66, 607–610.[ISI][Medline]

Hosmer, D.W. and Lemeshow, S. (2000) Applied Logistic Regression. Wiley, New York.

Job-Spira, N., Collet, P., Coste, J., Brémond, A. and Laumon, B. (1993) Facteurs de risque de la grossesse extra-utérine. Résultats d’une enquête cas-témoins dans la région Rhône-Alpes [Risk factors for ectopic pregnancy. Results of a case control study in the Rhone-Alpes region]. Contracept. Fertil. Sex. 21, 307–312.[ISI][Medline]

Job-Spira, N., Bouyer, J., Pouly, J.L., Germain, E., Coste, J., Aublet-Cuvelier, B. and Fernandez, H. (1996) Fertility after ectopic pregnancy: first results of a population-based cohort study in France. Hum. Reprod., 11, 99–104.[Abstract]

Kabukoba, J.J. and de Courcy-Wheeler, R.H.B. (1992) Hysteroscopy in the diagnosis of suspected interstitial pregnancy. Int. J. Gynecol. Obstet., 37, 121–126.[ISI][Medline]

Khalifa, Y., Redgment, C.J., Yazdani, N., Taranisi, M. and Craft, I.L. (1994) Intramural pregancy following difficult embryo transfer. Hum. Reprod., 9, 2427–2428.[Abstract]

Landis, J.R. and Koch, G.G. (1977) The measurement of observer agreement for categorical data. Biometrics, 33, 159–174.[ISI][Medline]

Lau, S. and Tulandi, T. (1999) Conservative medical and surgical managment of interstitial ectopic pregnancy. Fertil. Steril., 72, 207–215.[ISI][Medline]

Mäkinen, J.I. (1987) Ectopic pregnancy in Finland 1967–83: a massive increase. Br. Med. J., 294, 740–741.[ISI][Medline]

Mäkinen, J.I. (1996) Ectopic pregnancy falls in Finland. Lancet, 348, 129–130.

Martin, J.N., Sessums, J.K., Martin, R.W., Pryor, J.A. and Morrison, J.C. (1988) Abdominal pregnancy: current concepts of management. Obstet. Gynecol., 71, 549–557.[Abstract]

McMorries, K., Lofton, R.H., Stinson, J.C. and Cummings, R.V. (1979) Is the IUCD increasing the number of ovarian pregnancies? Contemp. Obstet. Gynecol., 13, 165–169.

Parente, J.T., Ou, C.S., Levy, J. and Legatt, E. (1983) Cervical pregnancy analysis: a review and report of five cases. Obstet. Gynecol., 62, 79–82.[Abstract]

Pouly, J.L., Chapron, C., Manhès, H., Canis, M., Wattiez, A. and Bruhat, M.A. (1991) Multifactorial analysis of fertility after conservative laparoscopic treatment of ectopic pregnancy in a series of 223 patients. Fertil. Steril., 56, 453–460.[ISI][Medline]

Radman, H.M. (1978) Abdominal pregnancy: pathogenesis, diagnosis, and treatment. South. Med. J. 71, 670.[ISI][Medline]

Raziel, A., Golan, A., Pansky, M., Ron-El, R., Bukovsky, I. and Caspi, E. (1990) Ovarian pregnancy: a report of twenty cases in one institution. Am. J. Obstet. Gynecol., 163, 1182–1185.[ISI][Medline]

Sandvei, R., Sandstad, E., Steier, J.A. and Ulstein, M. (1987) Ovarian pregnancy associated with the intrauterine contraceptive device. Acta Obstet. Gynecol. Scand., 66, 137–141.[ISI][Medline]

StataCorp (2001) Stata Statistical Software. Release 7.0. Stata Corporation, College Station, TX.

Steiner, E., Tabaste, J.L., Servaud, M., Catanzano, G., Boudinet, F., Dabir, P. and Ardilouze, J.L. (1983) Grossesses ovariennes: le point en 1982. A propos de 6 cas opérés, revue de la littérature [Ovarian pregnancy: six personal cases]. J. Gynecol. Obstet. Biol. Reprod., 12, 401–406.[Medline]

Thorburn, J., Philipson, M. and Linblom, B. (1988) Fertility after ectopic pregnancy in relation to background factors and surgical treatment. Fertil. Steril., 49, 595–601.[ISI][Medline]

Tulandi, T., Vilos, G. and Gomel, V. (1995) Laparoscopic treatment of interstitial pregnancy. Obstet. Gynecol., 85, 465–467.[Abstract/Free Full Text]

Weström, L., Bengtsson, L.P.H. and Mardh, P.-A. (1981) Incidence, trends and risks of ectopic pregnancy in a population of women. Br. Med. J., 282, 15–18.[ISI][Medline]

Yao, M. and Tulandi, T. (1997) Current status of surgical and nonsurgical management of ectopic pregnancy. Fertil. Steril., 67, 421–433.[ISI][Medline]

Yarali, H., McComb, P.F. and Lee, N.H. (1994) Evaluation of the sites of ectopic pregnancies subsequent to reconstructive tubal surgery. Hum. Reprod., 9, 90–91.[Abstract]

Submitted on April 23, 2002; resubmitted on July 5, 2002; accepted on August 8, 2002.