Spontaneous fertility and pregnancy outcomes amongst 480 women with Turner syndrome Free

Abstract

Introduction

Turner syndrome (TS), affecting 1/2000 to 1/2500 live born girls (Nielsen and Wohlert, 1991; Stochholm et al., 2006), is a chromosomal aberration with a total or partial loss of one of the X chromosomes. Short stature and primary ovarian insufficiency are classic features of this syndrome. Haploinsufficiency of X pseudo-autosomal genes leads to an acceleration of follicular atresia (Modi et al., 2003) and, in most cases, to gonadal failure with primary amenorrhoea. In 5–20% of girls with TS, follicles are still present and allow for spontaneous menarche (Pasquino et al., 1997; Hovatta, 1999; Hreinsson et al., 2002), however few studies have evaluated the rate of spontaneous pregnancies (Hovatta, 1999; Birkebaek et al., 2002; Bryman et al., 2011; Hadnott et al., 2011). Oocyte donation (OD) represents, since the 1980s (Lutjen et al., 1984), an option for treating infertility in patients with TS. The success rate of pregnancies after OD is high but these pregnancies are associated with many complications (Bodri et al., 2006; Alvaro Mercadal et al., 2011; Chevalier et al., 2011). A study performed in 82 patients with TS evaluated pregnancy outcome after OD. In this cohort, only 40.2% of patients completed their pregnancy without complications. Two maternal deaths occurred because of aortic root rupture, one patient had eclampsia with an acute cardiac failure in a twin pregnancy and 31 patients (37.8%) had pregnancy-induced hypertensive disorders (PIHD) (Chevalier et al., 2011). Therefore cardiovascular complications represent a major concern during pregnancies in TS patients. Recommendations have been proposed for the management of patients with Turner syndrome before and during pregnancy (Bondy, 2007; Cabanes et al., 2010).

Another option in order to achieve pregnancies in TS patients can be fertility preservation. Different techniques have been suggested, such as ovarian tissue cryopreservation or, more recently, oocyte freezing (Borgström et al., 2009; Hewitt et al., 2013). Borgström et al. have searched for predictive factors of the presence of follicles in ovarian biopsies (Borgström et al., 2009). Mosaicism, signs of spontaneous puberty and normal FSH and AMH concentrations are positive and statistically significant prognostic factors for finding follicles. So far, no pregnancy has been reported in TS women after ovarian or oocyte preservation. A single case of a live birth has been reported after allografting of ovarian cortex between monozygotic twins with TS (45X/46XX mosaicism) and discordant ovarian function (Donnez et al., 2011). In order to give patients with Turner syndrome better counselling, we performed a study evaluating the prevalence and the outcome of spontaneous pregnancies (SP) in a large cohort of women with TS.

Methods

The French Ministry of Health set up a Reference Centre for Rare Growth Disorders (CRMERC) in 2006. This network of specialized tertiary care centres was established to improve and harmonize the care of patients with rare diseases during the transition from adolescence to adulthood. Patients’ files are recorded, after informed consent, in a web database called CEMARA. This database has been declared to the French national data protection agency (CNIL). We studied the cohort of adult patients followed up in seven endocrine units (Saint-Antoine, Pitié-Salpétrière, Bicêtre, Lyon, Marseille, Brest, Reims Hospitals) belonging to this centre between January 1999 and January 2014. The study was approved by the Institutional Review Board of our faculty. Adult patients with a standard peripheral leukocyte karyotype with more than 5% of cells displaying a total X or partial loss of the Xp chromosome (Bryman et al., 2011) were eligible. We considered the following data for each patient: age at diagnosis of TS, age at the time of the study, final height, cardiac and metabolic comorbidities and karyotype. Reproductive history was collected: occurrence and age of spontaneous menarche, age at pregnancy, outcome of spontaneous pregnancies including maternal and fetal complications, and delivery data. Those data were compared with a reference population issued from the French general population, as the outcomes of pregnancies are collected, using a database obtained from the Directorate General for Health belonging to the French Health Ministry (Blondel and Kermarrec, 2011).

Statistical analysis

The characteristics of the patients were expressed as frequencies and percentages for categorical variables and as median (range) for continuous variables. Relationships between categorical variables were assessed with χ² tests. A P-value of less than 0.05 was considered as statistically significant.

Results

Cohort characteristics

Our cohort included 480 adult women with TS. The clinical characteristics of these patients are displayed in Table I. TS was diagnosed at a median age of 10 years (range: 0–64). The median age at the time of the study was 30.5 years (18–72). These patients had short stature with a median adult height at 151 cm (127–172) and primary amenorrhoea in the majority of cases. Only 95 patients (19.8% of the cohort) had spontaneous menarche. Karyotype characteristics are reported in Table II. A monosomy was present in 181 of the 480 patient (37.7%). Cardiovascular comorbidities at the time of evaluation were hypertension in 54 patients (11.2%), diabetes in 42 (8.8%), aortic dilatation defined as one aortic root diameter exceeding 32 mm or 20 mm/m² (Bondy, 2007; Cabanes et al., 2010) in 51 (10.6%) and bicuspid aortic valve in 71 patients (14.8%). Among our cohort, 179 patients (37.3%) had hypothyroidism and/or positive anti-thyroperoxydase (anti-TPO) antibodies.

Maternal and obstetrical outcomes

Of the cohort of 480 women, 27 (5.6%) had a total of 52 spontaneous pregnancies. Characteristics of the SPs are described in Table II. The numbers of SPs per patient was one (n = 12 patients), two (n = 8) and three (n = 5). One patient had a total of four pregnancies including two full-term deliveries and two miscarriages. Another patient had a total of five SPs. However she had two miscarriages, two medical interruptions and one fetal death related to severe growth retardation and therefore no full-term delivery. Issues during all pregnancies from our cohort included legal abortion for two pregnancies (3.8%) and miscarriage for 16 pregnancies (30.8%), while medical interruption was performed for three pregnancies (5.8%), two of them because of fetal chromosomal abnormalities (one trisomy 21 and one trisomy 13) and the last one because of premature rupture of membranes at 20 weeks gestation (wg). A case of pregnancy termination occurred because of fetal death related to severe growth retardation. For 30 pregnancies (57.7%), the patients delivered at full term.

Maternal complications occurred in 6 of those 30 pregnancies (20%), with pregnancy-induced hypertensive disorders (PHDs) in four cases (13.3%), including two cases of mild pre-eclampsia (6.7%). One of the patients with pre-eclampsia had aortic bicuspidy. Concerning other maternal complications, one patient had gestational diabetes. Another patient had diabetes before being pregnant, no complication occurred during her pregnancy. One patient had intrahepatic cholestasis of pregnancy. Neither aortic dissection nor cardiac complication was observed in any of the pregnant women during pregnancy or post-partum. For five out of the 30 pregnancies, no cardiac evaluation was performed before or during pregnancies, as the diagnosis of Turner syndrome occurred after the spontaneous pregnancies.

Caesarean sections were performed in 14 cases (46.7%). This rate is higher than in the French general population, respectively 46.7% versus 21% (P < 0.001). The reasons for performing a Caesarean section in our series were failure of induced labour (n = 1), breech and transverse fetal position (n = 2), pelvic disproportion (n = 4), eclampsia (n = 1) and acute fetal distress during vaginal delivery (n = 2). The reason remains unknown for four patients.

Neonatal outcomes

There were 18 patients who gave birth to 30 healthy infants, 13 boys and 17 girls. No specific neonatal complication was reported. The median birthweight was 3030 g (range 2020–3460). Karyotyping was performed in 0/13 boys and 11/17 newborn girls. Two of these 11 newborns were diagnosed with TS, with the following karyotypes: 45X,46Xr(X) and 47,X,der(X)t(X;Y)x2/46,X,der(X)t(X;Y). The first one is now 8-year-old. She has repeated otitis and some difficulties in learning, probably due to her ring X. Her height is 118.5 cm (−1.5 SD) and growth hormone treatment has been initiated. The latter had the transmission of a maternal translocation (45,X/46,X,der(X)t(X;Y)(p11.4;p11.2) with an unexpected different mosaicism (Portnoï et al., 2012). She is now 3-year-old. Her height is 83.5 cm (−0.7 SD) and she has no clinical signs of Turner syndrome. In the remaining six newborn girls, karyotyping was not performed as no phenotype of TS was present at birth. No cardiac or brain abnormalities have been observed in any of the 30 children. There 18 babies who were breastfed for a median duration of 3 months (0.5–8).

Characteristics of the patients with spontaneous pregnancies

Characteristics of the 27 patients with SP were compared with the non-pregnant TS patients of the cohort. Results are displayed in Table III. The median age at TS diagnosis for the 27 spontaneously pregnant patients was 20 years (0–45). It is significantly older than the median age of diagnosis for the non-pregnant TS patients (P < 0.01). Predictive factors of SP were spontaneous menarche occurrence (92.6% versus 15.4%, P < 0.001) and 45X, 46XX mosaicism (70.4 versus 24.5%, P < 0.001). Among the 27 patients with SP, 22 had regular cycles. Their preconception cycle length was lower than 27 days, between 27–35 days, and from 35 to 45 days for three, fourteen and five patients, respectively. Two patients were put on hormonal replacement therapy (HRT) before their spontaneous menarche. Their SP occurred as they were on HRT. One patient started her pregnancy while she was on oral combined contraceptive pill. Data about cycle length before conception were not available for two patients.

Among the 27 patients with SP, the primary presenting factor leading to karyotype evaluation was growth retardation during infancy in 11 (41%) and infertility in 10 (37%). In such cases, karyotype evaluation was performed because of repeated miscarriages (n = 5) or diminished ovarian reserve associated with short stature (n = 5). One patient was diagnosed because of familial screening. In five cases (19%), diagnosis of TS was performed after their SP pregnancy, as those patients had already delivered and then presented primary ovarian insufficiency.

Concerning cardiovascular comorbidities, two patients had bicuspidy and one had mild aortic coarctation. These three patients presented with mild aortic insufficiency. No aortic root dilatation was observed after pregnancy in any of the 27 patients. Among them, 10 (37%) had hypothyroidism and/or positive anti-thyroperoxydase antibodies. This percentage is statistically not different from the non-pregnant patients with TS.

The mean age at first SP was not statistically different in women with TS compared with the general French population, respectively 27.5 versus 28.1 years (Pison, 2010). The median delay to conception was 6 months (0–84) versus 3 months in French general population (Pison, 2010).

Discussion

We report here a cohort of 480 adult patients with TS, in whom we evaluated the prevalence and outcome of spontaneous pregnancies. The prevalence of spontaneous pregnancies in this French cohort was 5.6%. There were 18 patients (3.8%) who had at least one live born child.

So far in the literature, only three studies have evaluated pregnancy rates in women with TS (Birkebaek et al., 2002; Bryman et al., 2011; Hadnott et al., 2011). The first one, from Denmark (Birkebaek et al., 2002) was based on a national cytogenetic central register, including women born after 1942 and over 18 years in December 1993. The rate of spontaneous pregnancies reported in this study was 7.6%, but the patients' phenotypes and clinical indications for karyotyping were not available. The second study from Sweden (Bryman et al., 2011), evaluated both spontaneous pregnancies and pregnancies issued from assisted reproductive techniques. The prevalence of SP was 4.8%, in accordance with our results. The last study, from the United States (Hadnott et al., 2011) found a lower prevalence of SP, reaching only 1.8%. The percentage of SP in our cohort is higher, although it is probably underestimated. Indeed, the median age in our cohort was 30.5 years at the time of evaluation. Therefore, the youngest women may not have yet wished to become pregnant. Another potential reason for underestimating the pregnancy rate in TS patients might be that women with TS are known to have lower self-esteem, to be more often single and less sexually active than women of the general population (Carel et al., 2006). Indeed, information about sexual activity and/or whether the patient is in a couple and seeking a pregnancy were available for 307 patients of our cohort, and only 118 patients (38.4%) mentioned living in a couple and desiring a pregnancy. Interestingly, our study shows that mean age at first pregnancy in the 27 TS patients with SP is not different from that in the general population. Furthermore, the median delay to conception was only 6 months in our cohort of TS pregnant patients. This figure differs only slightly from the general population with a median delay of 3 months.

Concerning cytological characteristics of the patients in our cohort, the karyotypes were in good agreement with several other cohorts of TS patients (Pasquino et al., 1997; Sybert and McCauley, 2004). We interestingly report the occurrence of SP in two women with X monosomy. Although some cases have already been described in the literature (Hovatta, 1999; Bryman et al., 2011), they raise the issue of potential different karyotypes between the ovaries and the peripheral blood (Varela et al., 1991; Tarani et al., 1998). Indeed it is well known that the percentage of mosaicism can differ from one tissue to another (Lespinasse et al., 1998). We also originally report the occurrence of four SP including two deliveries at full term in a patient with a mosaicism containing Y fragment. This patient had a spontaneous puberty and had refused the prophylactic gonadectomy. Current guidelines recommend gonadectomy in TS women with detectable Y chromosomal material due to an elevated risk of developing gonadal neoplasia (Michala et al., 2008). However, our patient illustrates the fact that gonadectomy might be delayed in some cases. After informed consent, if gonadectomy is delayed, patients should be regularly monitored, every year, by pelvic ultrasounds.

Concerning pregnancy outcome, our study confirms a higher rate of spontaneous miscarriage in patients with TS as compared with the general French population (30.8% versus 15% P < 0.01). Different mechanisms could explain this elevated rate of pregnancy loss in TS patients, such as fetal chromosomal abnormalities or poor oocyte quality. Different studies have attributed the increased risk of miscarriage in TS patients to small uterine size and reduced endometrial thickness and receptivity (Yaron et al., 1996; O'Donnell et al., 2012). Another explanation could be a high prevalence of autoimmunity disorders in these patients (Abir et al., 2001). However, no difference in the percentage of thyroid autoimmunity disorders was identified between pregnant and non-pregnant women in our cohort. Therefore thyroid autoimmunity does not seem to play a major role in the rate of spontaneous miscarriages. Concerning fetal chromosomal abnormalities, in the Swedish study, karyotyping was not performed in the offspring (Bryman et al., 2011). In the American study (Hadnott et al., 2011), no chromosomal aberration was reported. In the Danish study, 25 children had a karyotype and six of them had chromosomal abnormities (Birkebaek et al., 2002). Our data confirm that the rate of chromosomal abnormalities is high in SP pregnancies in women with Turner syndrome. Two newborn girls were diagnosed with TS (45X,46Xr(X); 47,X,der(X)t(X;Y)x2/46,X,der(X)t(X;Y) and two medical interruptions were performed because of fetal chromosomal abnormalities (one trisomy 21 and one trisomy 13). Furthermore, karyotypes had not been systematically performed in all newborns, as they had not been performed in boys. This may have reduced the rate of chromosomal abnormalities. Interestingly, Balen et al. have evaluated oocyte quality after a cycle of ovarian stimulation, in a 28-year-old patient with a 45X,46XX karyotype (Balen et al., 2010). This cycle was performed to assess oocyte quality and genetics, before undertaking oocytes freezing in a fertility preservation programme. Eleven cumulus-enclosed oocytes were retrieved. Their cytogenetic analysis revealed that they all contained 23 chromosomes. No alteration in chromosome X number was detected in any of the oocytes analysed by FISH, and full karyotypes obtained from five MII oocytes were normal. This study may be reassuring considering oocyte quality obtained after ovarian stimulation in TS patients with 45X,46XX mosaicism. However according to previous studies, as well as ours, the rate of aneuploidy seems to be much higher in SP for TS patients. The mechanisms involved are not fully elucidated. First of all, one should distinguish X chromosomal structural abnormalities from 45X, 46XX mosaicism. In our two girls with Turner syndrome, issued from SP, both mothers had X chromosomal structural abnormalities. One had a ring and the other one, an X chromosome containing Y fragments (Portnoï et al., 2012). Both children inherited the abnormal X chromosome. In theory, the rate of inheritance should be 50%, it may however be lower after fertilization. Trisomies 21 and 13 might be related to poor oocyte quality related to Turner syndrome. Therefore, we recommend performing prenatal diagnosis in women with TS.

Although cardiac or brain abnormalities have been reported in newborns issued from SP in patients with TS, in previous studies (Birkebaek et al., 2002; Bryman et al., 2011; Hadnott et al., 2011), none has been observed among the 30 children of our cohort.

Maternal complications were essentially pregnancy-induced hypertensive disorders, present in 4 cases (13.3% of the 30 achieved pregnancies). Therefore, spontaneous pregnancies seem to be less complicated than pregnancies obtained after OD in TS women (Bodri et al., 2006; Chevalier et al., 2011). Those series have reported hypertensive disorders in 20–28% of cases. Mercadal et al. even reported a rate of 50% of maternal complications in pregnancies after oocyte donation in women with TS (Alvaro Mercadal et al., 2011). One hypothesis may rely on the younger age of TS patients in our cohort as compared with the mean age of pregnancies obtained in series after OD. Another hypothesis is that patients achieving SP probably have a less severe phenotype of the syndrome than those requiring OD, including less cardiovascular as well as renal phenotypes. Indeed, among the 27 patients with SP in our cohort, 37% were investigated due to primary infertility. The rate of 13.3% of pregnancy-induced hypertensive disorders in our cohort is lower than in TS pregnant patients after OD. However, it remains double that in the general population. This suggests that there is a risk factor inherent to the syndrome itself. Patients with TS have a higher risk of cardiovascular and renal abnormalities which could predispose to hypertension (Alvaro Mercadal et al., 2011). Indeed one of our patients with pre-eclampsia had aortic bicuspidy. Some other hypotheses explaining the higher rate of hypertensive disorders during pregnancy rely on vascular defects in the endometrium and/or placenta in Turner syndrome (Rogers et al., 1992; Biljan et al., 1995).

No aortic dissection was reported in the cohort, either during pregnancy or in the post-partum period. This suggests that it is a minor phenotype of TS in patients with spontaneous pregnancies. Moreover, it is known that OD is an independent risk factor for pregnancy complications (Younis and Laufer, 2015). Despite the fact that there was no aortic dissection in our cohort, regular surveillance before, during and even in the weeks after a pregnancy is of paramount importance, as dissection is always a potential risk in TS pregnancies. All patients with Turner syndrome should be carefully monitored before and during their pregnancy, according to the American or the European recommendations (Bondy, 2007; Cabanes et al., 2010). Before conception, check-ups should include cardiovascular examination as well as ambulatory blood pressure monitoring, two-dimensional transthoracic ultrasound with colour Doppler imaging, and when available magnetic resonance angiography of the heart and aorta. During pregnancy, echocardiography should be performed at the end of the first and second trimester and every month in the third trimester. Moreover, as cardiovascular risk persists in the post-partum period, an echocardiography is recommended one month after the delivery.

Caesarean section was performed in 46.7% of the patients included in our cohort (versus 21% in French general population, P < 0.001) (Blondel and Kermarrec, 2011). The reasons could be that short stature predisposes for disproportion of the pelvis, and therefore Caesarean section is easily considered in TS patients. Moreover, patients with TS have a predisposition to cardiovascular diseases which can be revealed during labour (Chevalier et al., 2011). In our patients, the reasons for performing a Caesarean section were failure of induced labour (n = 1), breech and transverse fetal position (n = 2), pelvic disproportion (n = 4), eclampsia (n = 1) and acute fetal distress during vaginal delivery (n = 2). The reason remains unknown for four patients. Our study illustrates the fact that, at least in some cases, Caesarian section was performed for some reasons unrelated to Turner syndrome.

One objective of our study was to highlight predictive factors of SP. We found that age at TS diagnosis, spontaneous menarche occurrence and karyotype with mosaicism without Y fragment are predictive factors of SP. Therefore, a TS patient with regular menses has a higher probability of SP. However the risk of POI is very high in such patients, and careful counselling is needed. Oocyte freezing can be proposed in order to preserve their fertility. Some studies have reported uterine size and endometrium thickness as predictors of likelihood of pregnancy after oocyte donation (Remohí et al., 1997). However data concerning uterine size were not available in our population of TS. Furthermore, it would have been interesting to evaluate AMH levels and SP occurrence. Unfortunately, hormonal measurements were not available for some patients. In the literature, AMH level is correlated with karyotype in Turner syndrome and occurrence of spontaneous puberty (Hagen et al., 2010; Visser et al., 2013). Recently, Lunding et al. have reported AMH as predictor of premature ovarian insufficiency in 120 Turner syndrome patients (Lunding et al., 2015). However, no study so far has reported a link between AMH level and spontaneous pregnancy in TS.

In conclusion, we hope that our study will help physicians counselling patients with Turner syndrome, as well as their families. Considering pregnancy outcome, our study illustrates that spontaneous pregnancies are much less complicated than pregnancies obtained after oocyte donation in TS patients. Although two main different options for fertility preservation are available nowadays, such as cryopreservation of ovarian tissue in young girls and oocyte freezing in adolescents or young women after puberty, the rate of spontaneous pregnancies in TS should be taken into account. Predicting factors of spontaneous pregnancies, apart from karyotype and spontaneous menarche, are still lacking and should be further evaluated.

Authors' roles

S.C.-M. planned and supervised the study. V.B. collected and analysed the data and wrote the manuscript. B.D., D.Z., C.C., S.S., A.B.d.l.P., F.A., A.F., V.K., T.B., B.D., F.B.-C.; J.-C.C., P.C., J.L. and P.T. contributed to the acquisition of data. All authors were involved in revising the manuscript and approved the final version.

Funding

Funding was provided by the Association pour la recherche Claude Bernard, Paris, France.

Conflict of interest

None declared.

Acknowledgements

We are grateful to all of the centres who contributed to the data collected in the current study. We thank Jean Bouyer and Benjamin Bouyer for their help in the statistical analysis.

References