2024 Volume 71 Issue 10 Pages 1003-1011
A 38-year-old Japanese woman with a history of abnormal thyroid function of non-autoimmune origin, pituitary endocrine tumor, and untreated diabetes mellitus was referred to our outpatient clinic when she became pregnant with twins. Physical findings consistent with Cushing’s syndrome (CS) were absent at the time of presentation. Although baseline plasma adrenocorticotropic hormone, serum cortisol, and 24-hour urinary free cortisol excretion levels were above the upper limits of normal non-pregnant reference ranges, we could not exclude a physiological increase associated with pregnancy. No medical or surgical intervention for hypercortisolism was performed during pregnancy. Spontaneous vaginal delivery resulted in the normal delivery of live twins. A diagnosis of Cushing’s disease (CD) was established when papery skin developed postpartum. Transsphenoidal surgery was performed and the hypercortisolism partially resolved post-operatively. The patient’s abnormal thyroid function also resolved. Pregnancy in women with endogenous CS is rare, with less than 300 cases reported. Most reported cases of CS during pregnancy are of adrenal origin. Only two cases of twin pregnancies with CD have been reported. Therefore, we reported the third case of CD in a twin pregnancy and reviewed the diagnostic and therapeutic challenges associated with CD during pregnancy.
Endogenous Cushing’s syndrome (CS) is a rare disorder with an incidence of 1.8 to 3.2 per million people per year and is more common in women than in men [1]. Pregnancy is rare in women diagnosed with CS, with fewer than 300 cases being reported to date [2-29]. This is because physiologically, excessive cortisol secretion tends to interfere with ovulation through hypogonadotropic hypogonadism, making it difficult to conceive [9, 30-32]. The etiology of CS in pregnancy differs from that in the non-pregnant state, with 70% of non-pregnant cases being Cushing’s disease (CD) of pituitary origin; whereas only 28% of cases are CD in pregnancy and more than half of cases are of adrenal origin [2]. Regardless of the etiology of CS, maternal and fetal mortality is high, with an overall fetal loss rate of approximately 25% in pregnancies [2]. A previous study indicated that pregnant patients who received treatment had better fetal outcomes than those who did not [2].
We recently encountered a case of CD in a twin pregnancy complicated with diabetes mellitus (DM). To the best of our knowledge, only two cases of CD with twin pregnancies have been reported [5, 25]; hence, our case is the third. CD was diagnosed during pregnancy, and treatment with metyrapone (1 g/day) was initiated in the first twin [5]. One twin died one day after delivery, whereas the other survived with intensive care support. The second patient became pregnant after 18 months of mitotane therapy for CD, and mitotane was discontinued [25]. Two healthy babies of 2,180 g and 2,550 g each, with normal adrenal function, were delivered by emergency C-section at 38 weeks. Herein, we discuss the challenges in diagnosing and appropriately treating CD during pregnancy.
A 38-year-old Japanese woman was referred to a local physician by her obstetrician/gynecologist for hypothyroidism during infertility treatment in year X. Blood tests revealed a thyroid stimulating hormone (TSH) level of 0.436 μIU/mL, free triiodothyronine (FT3) level of 1.96 pg/mL, and free thyroxine (FT4) level of 0.86 ng/mL. Tests for antithyroglobulin and anti-thyroid peroxidase (TPO) antibodies yielded negative results. Magnetic resonance imaging (MRI) of the pituitary gland (Fig. 1) showed a T1WI low-signal and T2WI high-signal nodule 10 mm in diameter on the anterior pituitary. A nonfunctioning pituitary tumor was suspected. However, the patient did not undergo surgery because the tumor was far from the optic nerve. On thyroid releasing hormone provocation test, pretest TSH was 0.366 μIU/mL and peak TSH was 4.59 μIU/mL. Supplemental levothyroxine treatment was initiated in May of Year X. She was also found to have DM (glycated hemoglobin: 6.5%); however, no medication was administered. Another MRI was performed in October X, with stable findings, and the patient was kept under observation.
A twin pregnancy was discovered in January of the following year, and she was referred to our outpatient clinic for the management of DM complicated by pregnancy at 14 weeks of gestation. At the time of her initial presentation to our hospital, she was 158.9 cm tall, weighed 47.4 kg, and had a body mass index (BMI) of 18.8 kg/m2. Her blood pressure was 118/75 mm Hg and she also complained of hyperemesis gravidarum. There were no visual field deficits, headaches, or physical findings typical of CD, such as a moon face or central obesity. At the time of referral to our hospital, she was taking levothyroxine at a dosage of 75 μg/day. After nutritional counseling, the patient began self-monitoring of blood glucose levels and self-administration of insulin lispro at a dosage of 2 units before breakfast and 2 units before dinner.
The presence of a pituitary tumor led us to measure adrenocorticotropic hormone (ACTH) and cortisol levels in April X + 1 (at 21 weeks of gestation), although she had no signs of CD. A possible diagnosis of CD was suggested by elevated plasma ACTH levels of 72.2 pg/mL (reference range 7.2–63.3), cortisol levels of 42.1 μg/dL (7.07–19.6), and 24-hour urinary free cortisol excretion (UFC) levels determined by chemiluminescent immunoassay (CLIA) of 658 μg/day (4.3–176). Other anterior pituitary hormones at baseline were as follows: growth hormone 9.54 ng/mL, insulin-like growth factor-1 157 ng/mL, FT4 0.94 ng/dL, FT3 2.10 pg/mL, TSH less than 0.003 mIU/mL, prolactin 116 ng/mL, luteinizing hormone 0.13 mIU/mL, and follicle stimulating hormone 0.05 mIU/mL at 21 weeks of gestation. Estradiol and progesterone levels were not measured. A dexamethasone suppression test with 1 mg dexamethasone was performed in May year X + 1 (at 23 weeks’ gestation), which showed no suppression of either ACTH or cortisol, with an ACTH level of 93.8 pg/mL and a cortisol level of 37.6 μg/dL. Another MRI was performed in June X + 1 (at 25 weeks’ gestation); however, no change in size was observed, and the patient was followed-up. On August X + 1 (37 weeks and 6 days of gestation), the patient gave birth to twins by spontaneous vaginal delivery, while the mother was receiving 20 units of insulin lispro daily. The female newborns weighed 2,080 g and 1,944 g, respectively, and received short-term care in the neonatal intensive care unit after delivery. Subsequent developments were unremarkable.
Postpartum ACTH was 41.1 pg/mL, cortisol was 29.7 μg/dL, and UFC was 91.8 μg/day, consistent with CD. Post-dexamethasone suppression test (with 0.5 mg dexamethasone) ACTH and cortisol levels were 55.5 pg/mL and 13.1 μg/dL, respectively. Papery skin appeared on November X + 1. A corticotropin-releasing hormone (CRH) provocation test revealed that ACTH levels increased from 66.0 pg/mL to 162.0 pg/mL after 30 min, as shown in Table 1. A suppression test with 8 mg dexamethasone resulted in an ACTH level of 12.9 pg/mL and a cortisol level of 1.29 μg/dL the next morning, leading to a diagnosis of CD.
| Time (min) | 0 | 30 | 60 | 90 | 120 |
|---|---|---|---|---|---|
| ACTH (pg/mL) | 66.0 | 162.0 | 83.3 | 48.2 | 17.6 |
| Cortisol (μg/dL) | 18.0 | 26.9 | 23.6 | 18.0 | 14.4 |
The patient underwent transsphenoidal pituitary tumor resection at the Department of Neurosurgery at our hospital on May X + 2. Pathology confirmed CD with positive immunostaining for ACTH and TPIT (Fig. 2; TPIT data not shown). Postoperative blood samples showed decreased ACTH and cortisol levels at 60.3 pg/mL and 12.1 μg/dL, respectively. However, remnant disease was suggested, as ACTH was not suppressed post-operatively. She was weaned off levothyroxine and hydrocortisone 2.5 months after surgery, and her thyroid function has remained normal since then. Her glucose tolerance also normalized. Laboratory data 7 months after surgery were as follows: TSH 1.081 μIU/mL, FT3 2.87 pg/mL, and FT4 level of 1.02 ng/mL without thyroid hormone replacement.
It is uncommon for patients with CS to get pregnant due to the ovulatory disturbances associated with impaired gonadotropin secretion [9, 30-32]. To date, less than 300 cases of CS complicated by pregnancy have been reported [2-29]. The number of patients with CD is even lower. A literature search revealed 28 women with 33 pregnancies, including the current case, from 2000 to 2024 [5-29] (Table 2). Cabezón et al. reported a case of twin pregnancy in 1999 [5]. To the best of our knowledge, only two cases of CD with twin pregnancies have been reported [5, 25], and this is the third reported case. Subclinical CD is a subtype of CD, lacking specific symptoms of CD. Subclinical CD is frequently discovered as a pituitary incidentaloma [33] as in our case. Hence, we may say that the current case was diagnosed as subclinical CD. If so, this is the first case of subclinical CD complicated with pregnancy.
| Author | Year | Reference | Note | Age | Timing of onset/diagnosis of CD | Treatment during pregnancy | Pregnancy outcome | Foetal outcomes |
|---|---|---|---|---|---|---|---|---|
| Cabezón C* | 1999 | [5] | 27 | At 22 weeks gestation | Metyrapone (1 g/d) | SVD at 28 weeks | A nonvirilized female (830 g) with hyaline membrane disease and severe respiratory depression was delivered and died 1 day later. A normal male (1,020 g) with hypoglycemia, hyponatremia, and hyaline membrane disease was also delivered and survived after intensive care support. | |
| Dobrovsky V | 2003 | [6] | 32 | 1 year before pregnancy, followed by pituitary apoplexy resulting in hypopituitarism thereafter | None | Spontaneous abortion | Died | |
| Verdugo C | 2004 | [7] | 26 | 6 months before pregnancy | TSS at 23 weeks gestation | SVD at 39 weeks | A 3,300-g normal female baby | |
| Chiodini I | 2004 | [8] | 1st pregnancy | 24 | TSS × 2 + GKRS before pregnancy | None | CS at 34 weeks | A 2,080-g healthy male baby |
| 2nd pregnancy | 25 | None | SVD | A normal female baby | ||||
| Lindsay JR | 2005 | [9] | Case 1 | 24 | Before pregnancy | Metyrapone 500 mg daily before TSS at 18 weeks gestation | SVD at 34 weeks | A 1,712-g healthy male baby |
| Case 2 | 34 | At 5 weeks gestation | TSS at 14 weeks gestation with remnant on the left cavernous sinus | Intrauterine death at 33 weeks | Died | |||
| Case 3 | 32 | At 9 weeks gestation | TSS at 10 weeks + 5 days gestation | SVD | A healthy full-term infant without complication | |||
| Case 4 | 33 | Not described | TSS at 17 weeks gestation | CS at 24 weeks | Weighed 440 g and died 5 days later. | |||
| Miyoshi T | 2005 | [10] | 19 | At 4 months postpartum | None | SVD at 36 weeks | A 3,250-g healthy male baby | |
| Yawar A | 2007 | [11] | 1st pregnancy | 28 | 1-year history of symptoms and signs; diagnosed at postpartum | None | Normal delivery | A normal baby |
| 2nd pregnancy | 31 | 4-year history of symptoms and signs and 3 yr before the 2nd pregnancy | None | Normal delivery | A normal baby | |||
| Casonato A | 2010 | [12] | 29 | TSS 2 years before pregnancy | Ketoconazole during the 1st & 2nd trimester | Induced labor at 38 weeks | not described | |
| Triay J | 2010 | [13] | 33 | At the 3rd trimester | None | Emergent CS at 34 weeks | not described | |
| Ghodki P | 2011 | [14] | 26 | At full term | None | CS at full term | A 2.5-kg baby with an APGAR score of 9 | |
| Boronat M | 2011 | [15] | 26 | Before pregnancy in treatment with ketoconazole waiting for TSS | Ketoconazole (200 mg bid) until 13th week, TSS at 16th week, Metyrapone 250–500 mg bid after 20th week | Induced labor at 34 weeks | A 2,480-g male baby | |
| Gopal RA | 2012 | [16] | 22 | At 7th month gestation | None | SVD at 34 weeks | A 2,090-g female baby with respiratory distress and septic shock due to transient adrenal insufficiency in the neonatal period | |
| Ragonese M | 2012 | [17] | 28 | At 6 weeks gestation | None | SVD at 39 weeks | A 2,865-g healthy female baby | |
| Ferraù F | 2012 | [18] | Case 1 | 34 | GKRS 13 months before pregnancy | None | Spontaneous delivery at 38 weeks | A 2,120-g healthy male baby |
| Case 2-1st pregnancy | 25 | TSS, GKRS 3 years before the 1st pregnancy followed by ketoconazole 400mg until 4 months before the 1st pregnancy | None | Spontaneous abortion at 10 weeks | Died | |||
| 2nd pregnancy | 25 ? | None | Placental abruption at 17 weeks | Died | ||||
| 3rd pregnancy | 28 | None | Spontaneous delivery at 38 weeks | A 2,600-g healthy female baby | ||||
| 4th pregnancy | 29 ? | None | SVD at 39 weeks | A 3,150-g healthy male baby | ||||
| Woo I | 2013 | [19] | 29 | TSS 3 years before pregnancy | Cabergoline 4.5–5.0 mg twice weekly | Forceps-assisted SVD at 38 weeks | A 3,350-g healthy female baby | |
| Abbassy M | 2015 | [20] | 38 | Recurrent CD after 8 years of remission | 2nd TSS at 18th week | SVD at 39 weeks | A 3,600-g healthy baby | |
| Costenaro F | 2015 | [21] | 39 | A prior diagnosis of CD at age of 25 years | Ketoconazole until 7th week and resumed at 16th week (100 mg qds), Methyldopa 250 mg bid | SVD at 36 weeks | A 2,770-g healthy female baby | |
| Nakhleh A | 2016 | [22] | 29 | TSS at 3 years before with cavernous sinus remnant | Cabergoline started before pregnancy at 3.5 mg/week and titrated down from to 2 mg/week at 7th week | Elective CS at 40 weeks | A 3,370-g healthy female baby | |
| Sek KSY | 2017 | [23] | 36 | Prior CD; TSS and GKRS | Cabergoline 0.25 mg twice weekly throughout the pregnancy | SVD at 40 weeks | A 3,195-g healthy male baby | |
| Dogansen S | 2017 | [24] | 27 | Recurrent pregnancy loss and diabetes mellitus, 9-month history of weight gain, menstrual irregularity and hirsutism before pregnancy | TSS at 1st trimester | SVD at 38 weeks | A 2,700-g healthy female baby | |
| Magkou D* | 2018 | [25] | Case 1 | 20 | 2 years before pregnancy | Mitotane terminated on pregnancy | Emergent CS at 38 weeks | A 2,180-g and a 2,550-g healthy twin babies with normal adrenal functions |
| Jolly K | 2019 | [26] | 30 | TSS 6 years before with recurrence; diagnosed at 13 weeks gestation | TSS at 23rd week | Emergent CS at 38 weeks | Baby died 36 hours later due to complications of congenital diaphragmatic hernia | |
| Kim HB | 2020 | [27] | Case 2 | 8 months after delivery | None | Elective CS at 37 weeks | A 2,200-g female infant | |
| Sridharan K | 2021 | [28] | 26 | At 16 weeks gestation | TSS at 20 weeks gestation | SVD at 40 weeks | No signs of neonatal adrenal insufficiency or other neonatal complications | |
| Murillo NB | 2023 | [29] | 29 | 1 year before pregnancy | Cabergoline 0.5 mg weekly terminated on pregnancy and restarted on the 2nd trimester | CS at 38 weeks | A 3,140-g healthy female baby | |
| Asai H* | 2024 | current case | 38 | PitNET detected 11 months before pregnancy; diagnosed 3 months after delivery | None | SVD at 36 weeks + 6 days | 2,080-g and 1,944-g healthy female babies |
Case reports on Cushing’s disease associated with pregnancy between 2000–2024 are listed.
*: twin pregnancy. A report on twin pregnancy case in 1999 is also included.
CD: Cushing’s disease, CS: cesarean section, GKRS: gamma-knife radiosurgery, SVD; spontaneous vaginal delivery, TSS: trans-sphenoidal surgery
CS during pregnancy is a frequent cause of both maternal and fetal complications, including maternal hypertension, hyperglycemia, preeclampsia, prematurity, stillbirth, spontaneous abortion/intrauterine death, and intrauterine growth retardation of the fetus [9]. Seven babies were lost and 29 survived in the literature (Table 2). No information was available for two babies [12, 13].
Therefore, the European Society of Endocrinology states that pregnancy should be avoided in the presence of CS of any etiology [30]. Comparing pregnancy loss and global morbidity between active and resolved CS, both are much higher in the former (23.6% vs. 8.5% and 33.3% vs. 4.9%, respectively) [30]. This suggests a negative effect of undiagnosed hypercortisolism long before CS is diagnosed, as delayed diagnosis of CS is common in non-pregnant women [3]. There were no typical signs of CD during pregnancy in our case, and impaired glucose tolerance was the only clinical sign observed. Plasma ACTH and cortisol levels may not have been measured if the previous physician had not detected a pituitary tumor on imaging, and a subsequent diagnosis would not have been possible.
Diagnosis of CD during pregnancy is challenging [9, 30-32]. In fact, the diagnosis of CD was established before pregnancy in 23 (65.7%), during pregnancy in eight (22.9%), postpartum in three (8.6%), and not specified in one (2.9%) out of the 35 pregnancies reported to date (Table 2). The placenta produces CRH, which increases the secretion of ACTH and cortisol [9, 30-32]. In addition, estrogen-induced hepatic production of corticosteroid-binding globulin (CBG) is increased during pregnancy. As a result, UFC levels increase to almost three times the normal levels [30]. The UFC level in our case (658 μg/day by CLIA kits [UFC-CLIA]) was 3.7 times the upper limit of the reference range for non-pregnant normal individuals (176 μg/day) at a commercial laboratory. UFC values more than 3-fold are required for a confident diagnosis [30]. We wondered if this was also applicable to twin pregnancies because the total placental volume was expected to be larger in twin pregnancies; hence, we hesitated to diagnose CD, although there are no supporting data on further elevation of either cortisol or CBG levels in this specific condition. However, more than 70 μg/day of UFC by radioimmunoassay (RIA) kits (UFC-RIA) is regarded as high in non-pregnant normal individuals, and patients with CD mostly have UFC over 100 μg/day (UFC-RIA), according to the recently published 2023 edition of “Guidelines for the Practice of Diencephalon-Pituitary Dysfunction and Congenital Nephrogenic Diabetes Insipidus and Related Disorders” by the Japan Endocrine Society [33]. UFC-CLIA is lower than UFC-RIA: UFC-CLIA = 0.832 × UFC-RIA – 4.23 [33]. According to this formula, UFC-CLIA in the current case corresponds to 795 μg/day on UFC-RIA. The UFC value far exceeded the upper limit of the normal reference range for non-pregnant women. In this context, we may have missed the opportunity to diagnose CD during pregnancy.
The Japan Endocrine Society recommends the use of a 0.5 mg dexamethasone dose for the dexamethasone suppression test as a screening test for CD in the general population to reduce the incidence of false-negative results [33]. Subclinical CD is suspected if overnight serum cortisol is greater than 3 mg/dL and overt CD is suspected if overnight serum cortisol is greater than 5 mg/dL [33]. However, the false positive rate is very high, at about 80% in normal pregnancies; hence, this test is not appropriate during pregnancy [30]. A high-dose dexamethasone suppression test has been previously recommended [9]. However, exposure to excessively high glucocorticoid levels during development may increase the risk of preterm birth, long-term psychological developmental and neurosensory damage, serious infections in the first year, and obesity, insulin resistance, and immune abnormalities in adulthood [34-38]. In addition, massive doses of exogenous steroids may exacerbate hypercoagulability in CD [39] and worsen DM as well. Measurement of late night salivary cortisol levels may help diagnose CS during pregnancy, as diurnal cortisol variation is restored during pregnancy [30]. However, salivary cortisol measurement is not available in Japan. Midnight serum cortisol levels may be an alternative; however, this was not determined in our case because our patient was managed on an outpatient basis throughout the course of her pregnancy. Moreover, we did not measure the midnight serum cortisol levels after delivery. As mentioned previously, the diagnosis of CS during pregnancy is challenging and frequently delayed until 18–26 weeks of gestation [2, 9].
CD was diagnosed after delivery when the dexamethasone suppression test was repeated, along with the CRH provocation test. There are reports of better maternal and fetal outcomes when CD is treated during pregnancy, in addition to a higher birth rate when it is implemented [9, 40], and the fetal mortality rate with treatment is 25% to 33% that of untreated cases, depending on whether medical or surgical treatment is administered. Transsphenoidal surgery should be performed before 24 weeks of gestation to minimize the risk of preterm delivery [31]. However, according to our literature search, 14 of 17 babies, excluding those with no fetal outcomes available [12, 13], survived when no treatment was given during pregnancy, whereas 15 of 19 counterparts survived with surgical and/or medical treatment during pregnancy (Table 2). All four babies including ours [10, 27] survived even if the diagnosis of CD was delayed after childbirth. Hypercortisolism was not anticipated when the patient first presented to our clinic at 14 weeks of gestation. Elevated baseline plasma ACTH and cortisol levels were observed at 21 weeks of gestation. Even if CD is identified during pregnancy in our case, it is recommended to postpone surgery [13, 14, 16].
Medical treatment during the first trimester is not recommended because of its teratogenic potential. Medication may be used in the second and third trimesters if surgical management is not indicated [9, 30-32]. However, only 10 cases of CD with medical treatment during pregnancy were found in the literature [2]. Metyrapone was the most commonly administered drug. However, blood pressure and potassium levels should be monitored regularly, as the drug may worsen hypertension and/or decrease potassium levels. A case of CD with twin pregnancies was treated with metyrapone [5]. One of the twins died after delivery, despite treatment. Ketoconazole has also been used to treat CD; however, there is limited data currently available on this form of treatment. Cabergoline has only been reported in four cases of CD [19, 22, 23, 29] with good maternal and fetal outcomes; however, breastfeeding is not feasible. Mitotane treatment was terminated in another twin with favorable fetal outcomes [25]. Pasireotide and osilodrostat have not been previously reported in pregnant women with CD. Mifepristone, a glucocorticoid receptor blocker available in some countries, is contraindicated because it is also a progesterone blocker used to terminate pregnancy [30].
The previous physician suspected central hypothyroidism prior to pregnancy in the current case; however, the FT4 level remained within the normal range compared with the low FT3 level. This may be due to excessive cortisol, which suppresses TSH secretion and the conversion of T4 to T3. Postoperative thyroid function normalized without thyroid hormone replacement. However, euthyroid sick syndrome secondary to hyperemesis gravidarum may have been present. In fact, the patient’s BMI was 18.8 kg/m2 on presentation at 14 weeks of gestation; hence, dietary restrictions for DM during pregnancy may have led to malnutrition. Nutritional status was expected to improve after delivery because dietary restriction was no longer required. Therefore, thyroid hormone replacement therapy may not have been necessary.
In conclusion, we report the third case of CD diagnosed in a twin pregnancy complicated by DM, along with a review of the literature. The patient had a twin pregnancy, which made the diagnosis of CD more difficult. It is important to confirm a diagnosis of CD as early as possible and provide the necessary preconception care and treatment to ensure a safe environment for the fetus. Further studies are necessary to establish the optimal pharmacotherapy of CD during pregnancy.
We are grateful to Prof. Oki Y for helpful suggestions on the management of the patient. A part of this paper was presented at the 249th Tokai Regional Meeting of the Japanese Society of Internal Medicine, February 19, 2023, Nagoya City, Japan.
None of the authors have any potential conflicts of interest associated with this case report.
Written informed consent was obtained from the patient for publication of this case report.
