Introduction
Physiologically, pregnancy comes with biological hyperandrogenism [
1]. Many mechanisms have been described to explain this hyperandrogenism. First the increase of total testosterone results from the increase of Sex Hormone Binding Globulin (SHGB). The increase of free testosterone serum concentration may be the consequence of the stimulation by human Chorionic Gonadotrophin (hCG), an adrenal influence and a reduced renal clearance of testosterone during pregnancy [
2]. Thanks to many protective mechanisms, maternal virilization during pregnancy is uncommon. PCOS is the major cause of hirsutism in women, but is rarely responsible on its own for a maternal virilization during pregnancy [
3]. Luteinic cysts and luteoma during pregnancy are the two other benign ovarian etiologies of maternal virilization. Potential other causes are ovarian tumors and adrenal etiologies such as adrenal tumor, 21-hydroxylase deficiency and Cushing syndrome [
4,
5]. In a recent review of the literature, only 30% of the patients with luteinic cysts were virilized and recurrences were anecdotal [
6,
7] as well as with luteoma [
8]. Up to date, only few cases of recurrent maternal virilization during pregnancy have been published [
7,
9‐
14], including one with PCOS [
4]. We report two cases of recurrent virilization during successive pregnancies, in women having PCOS and gestational diabetes.
Discussion
The originality of these two cases is the recurrence of a virilizing syndrome in successive pregnancies, complicated with a gestational diabetes. The origin remains imprecise although the spontaneous regression suggests a benign etiology. The investigation of the adrenal and somatotrope axes excluded an adrenal cause and acromegaly. An ovarian abnormality was then most likely to be responsible for these cases of virilization.
Luteinic pregnancy cysts can cause maternal virilization. They occur preferentially in young, Caucasian women, mostly primiparous, with multiple or molar pregnancies with high levels of hCG. The literature describes co-morbidities such as PCOS, dysthyroidism or gestational diabetes [
6]. Only 30% of patients with luteinic pregnancy cysts have clinical signs of virilization [
5]. High androgen concentrations without fetal virilization is usually observed [
5,
6]. At U/S, luteinic cysts are often bilateral, anechogenic and their largest diameter varies from 1 to 3 cm [
16]. The anatomopathological exam describes luteinization and hypertrophy of the internal layer of the ovarian theca [
7]. In our first case, U/S and MRI found an anechogenic cyst of 3 cm, associated with polycystic ovaries. In our second case, only the aspect of polycystic ovaries was described on the MRI. In addition, none of our patients experienced multiple or molar pregnancies. Therefore, the diagnosis of luteinic cysts seems unlikely in our patients.
Luteoma is among the most common benign ovarian tumors that may be associated with maternal hyperandrogenism. It is described on U/S as an ovarian anechogenic mass with net limits, and a thick capsule. At histopathology, large polyhedral and eosinophilic cells with few or no lipids are found [
5,
17]. Biological hyperandrogenism is described to be up to seventy-fold higher than normal. In the literature, 25 to 35% of mothers with luteoma are virilized as well as 2/3 of their new-born daughters [
5]. Generally, luteoma is characterized by a spontaneous resolution of hyperandrogenism within weeks after delivery, although some symptoms like the deep voice and the clitoromegaly may remain [
14]. As the consequence of the inhibitory effect of androgens on the mammary gland, lactation may be delayed [
18]. Several data in our first case suggest luteoma as the potential cause of virilization. First, luteoma occurs mostly in Afro-Caribbean women with pre-existing PCOS [
5,
13]. The appearance of clinical symptoms, the persistence of a deep voice and clitoromegaly also supports this diagnosis. Unfortunately, neither the imaging nor the anatomopathological analysis of the resected ovary was contributive because of the ovarian necrosis secondary to the adnexal torsion. Nevertheless, luteoma was the most likely diagnosis and our first case would be the 6th reported observation of recurrent luteoma in pregnancy [
4,
9,
12‐
14].
If it were not the case, and in our second patient, an aggravation of the pre-existing PCOS during pregnancy could be suggested. Serum testosterone levels during pregnancy are higher in women with PCOS than in controls [
19‐
21]. However, in the majority of cases, this has no clinical repercussion. However, few cases of virilization in pregnant PCOS patients were described in literature [
3‐
5] and they occurred during the first trimester of pregnancy, while in our patients, the onset of virilization occurred during the 2nd trimester of pregnancy and it was rapid, which is different from the PCOS cases previously reported. The cause of PCOS aggravation remains enigmatic. “LH-like” action of hCG was evoked in the pathophysiology of benign ovarian hyperandrogenism of pregnancy. The cells of the ovarian theca are sensitive to hCG and its “LH-like” effect, which lead to a reactional ovarian hyperplasia [
4]. However, the peak of hCG appears in the first trimester of pregnancy whereas the symptoms seem to appear mainly in the 2nd or 3rd trimester of pregnancy.
Other authors discuss the role of insulin and/or IGF1, which together with LH, increase androgen biosynthesis by ovarian theca cells [
4,
22]. Accordingly, in our cases, the excess of testosterone was associated with overweight and gestational hyperglycemia. The worsening effect of hyperinsulinism in PCOS is well known [
23], due to insulin resistance that increases during pregnancy. However, this occurs in many cases of PCOS without resulting in clinical hyperandrogenism. Interestingly, however, recent studies reported that insulin resistance associated to PCOS could contribute to hyperandrogenism by inhibiting placental aromatase and decreasing conversion of androgens to estrogens [
24,
25]. Such a phenomenon could be amplified in some patients, for unknown reason(s). It has been hypothesized that pregnancy hCG, that persistently overstimulates PCOS ovaries in pregnancy, can be associated with CYP19A1 (aromatase) deficiency and placental inability for androgen conversion [
26]. Alternatively, in PCOS, the Anti-Müllerian Hormone (AMH) is high [
27] and it is known to inhibit aromatase activity [
28]. Even if AMH tends to decrease during pregnancy, its serum level remains higher than in controls [
29]. We can therefore suspect an inhibition of placental aromatase by excessive production of AMH in our cases. Unfortunately, in our patients, we have no data about serum AHM during pregnancy and placental aromatase could not be measured.
Finally, there could be the same mechanisms that are still not understood as those that occur in ovarian hyperstimulation in patients with PCOS, involving growth factors such as vascular endothelium Growth Factor [
30].
Conclusion
Even though PCOS is frequent, virilization during pregnancy is rare. The physiopathology remains to be elucidated. Hyperplasia of ovarian theca cells seems to be the most likely explanation, which would suggest that PCOS belongs to a spectrum of abnormal reactivity of the ovary to hCG stimulation along with luteoma and luteinic cyst of pregnancy [
4]. Insulin resistance and low rate of SHBG could worsen hyperandrogenism but is not enough to explain virilization. Finally, a placental dysregulation with partial aromatase deficiency can be suggested. The question of how avoiding recurrence in a future pregnancy remains open. A biological follow-up and U/S maternal monitoring could help preventing complications and avoid deleterious adnexectomy. As far as the child prognosis is concerned, based on literature data [
31] and the fact that hyperandrogenism occurred rather lately, the risk of virilization of a female fetus is likely very low. Conversely, caution should be paid to the risk of developing PCOS and metabolic syndrome at adolescence. Recently, Filippou et al. [
32] suggested that elevated maternal testosterone levels during pregnancy could predict an ovarian dysfunction in adult daughter of PCOS women. Therefore, treatment should focus on protecting the fetus from possible virilization as well as its mother, but also on preserving the subsequent fertility in both.