Current Management and Outcome of Pregnancies in Women With Adrenal Insufficiency: Experience from a Multicenter Survey

Abstract

Context

Appropriate management of adrenal insufficiency (AI) in pregnancy can be challenging due to the rarity of the disease and lack of evidence-based recommendations to guide glucocorticoid and mineralocorticoid dosage adjustment.

Objective

Multicenter survey on current clinical approaches in managing AI during pregnancy.

Design

Retrospective anonymized data collection from 19 international centers from 2013 to 2019.

Setting and Patients

128 pregnancies in 113 women with different causes of AI: Addison disease (44%), secondary AI (25%), congenital adrenal hyperplasia (25%), and acquired AI due to bilateral adrenalectomy (6%).

Results

Hydrocortisone (HC) was the most commonly used glucocorticoid in 83% (97/117) of pregnancies. Glucocorticoid dosage was increased at any time during pregnancy in 73/128 (57%) of cases. In these cases, the difference in the daily dose of HC equivalent between baseline and the third trimester was 8.6 ± 5.4 (range 1-30) mg. Fludrocortisone dosage was increased in fewer cases (7/54 during the first trimester, 9/64 during the second trimester, and 9/62 cases during the third trimester). Overall, an adrenal crisis was reported in 9/128 (7%) pregnancies. Cesarean section was the most frequent mode of delivery at 58% (69/118). Fetal complications were reported in 3/120 (3%) and minor maternal complications in 15/120 (13%) pregnancies without fatal outcomes.

Conclusions

This survey confirms good maternal and fetal outcome in women with AI managed in specialized endocrine centers. An emphasis on careful endocrine follow-up and repeated patient education is likely to have reduced the risk of adrenal crisis and resulted in positive outcomes.

Addison disease, congenital adrenal hyperplasia, pregnancy, mineralocorticoid, glucocorticoid, adrenal crisis, miscarriage

Adrenal insufficiency (AI) is a potentially life-threatening condition that can be caused by destruction of adrenocortical cells by inborn alterations of steroidogenesis or by impairment of pituitary or hypothalamic stimulation of the adrenal cortex (1). AI is associated with an increased mortality risk and a reduced quality of life, as well as reduced ability to work (2). As AI predominantly affects women of reproductive age, the course and outcome of pregnancy may be affected. While the exact prevalence of AI in pregnancy is unknown, a recent population-based study from the United States reported an increase in prevalence of Addison disease (AD) from 5.6 per 100 000 in 2003 to 9.6 per 100 000 in 2011 in a cohort with 7.7 million births (3).

Prior to the introduction of glucocorticoid replacement therapy, AD was associated with high maternal mortality (4). However, modern glucocorticoid replacement therapy and improved obstetric care have both contributed immensely in reducing maternal as well as fetal morbidity and mortality. In patients with primary AI, replacement of both glucocorticoid and mineralocorticoid are usually necessary, while in secondary or tertiary AI only glucocorticoid substitution is required. However, it might be necessary to replace other deficient hormones in both primary AI, if it occurs as a part of polyglandular autoimmune syndrome (5), and secondary AI, if other pituitary axes are affected (6). In particular, women with secondary AI may need assisted reproduction techniques for conception (7). A population-based study on patients with primary AI from Sweden has indicated a persistently reduced parity, increased rate of preterm birth, and low birth weight (8). While studies have offered many important insights into a population basis, they fail to provide information on therapeutic strategies currently applied in this patient group.

Pregnancy can be considered as a state of hypercortisolism due to the upregulation of the whole hypothalamic–pituitary–adrenal axis. Furthermore, bound cortisol levels are increased due to increased corticosteroid-binding globulin under the high estrogen state of pregnancy (9). Several biochemical parameters, such as morning cortisol, adrenocorticotropic hormone, plasma sodium, and renin, change physiologically in pregnancy and are therefore not reliable in dosage titration of replacement therapies. Diagnosing new onset AI in pregnant women is further complicated, for many indicators suggestive of AI could be potentially mistaken as symptoms of pregnancy. Nevertheless, clinical information such as general wellbeing, appropriate weight gain during the course of pregnancy, blood pressure, and potassium levels may help in adjusting glucocorticoid and mineralocorticoid dosage. Ongoing follow-up during pregnancy, ideally from an endocrinologist, is expected to facilitate the appropriate management of AI to ensure self-adjustment of glucocorticoid dosage in response to potential stress events (10). Furthermore, improved obstetric care and increased awareness among physicians regarding the importance of individualized dosage adjustment of glucocorticoid and or mineralocorticoid replacement have likely contributed in improving outcomes in this rare disease. The clinical guidelines from the Endocrine Society suggest an empirical increase in hydrocortisone (HC) dosage by 20% to 40% in the last trimester (11, 12). However, this suggestion is based on a low level of evidence. Therefore, the aim of our multicenter retrospective survey was to analyze the current management approaches of dosage adjustment in pregnant women with AI in specialized centers and to assess the maternal and fetal outcome.

Material and Methods

Patients and pregnancies

We collected data on 128 pregnancies in 113 women with AI from 19 centers, with varying number of pregnancies per center (range 1-28 per center). Data collection was restricted to pregnancies between 2013 and 2019. Minimal clinical annotations for inclusion in the study were patient’s age, etiology of AI, maternal and fetal outcome, presence of adrenal crisis during pregnancy, and mode of delivery. Predefined data points were collected by each center for each patient. Upon anonymization, data were transferred for central analysis. The central ethics Committee of the University of Zurich approved this study and provided a waiver of consent to collect retrospective data from different centers in an anonymized form. Where required, centers obtained additional local ethics approval for data collection.

We assessed clinical and biochemical parameters before pregnancy and during the 3 trimesters (first trimester weeks 1-13, second trimester weeks 14-27, and third trimester weeks 28-40 of pregnancy). The main clinical parameters included weight, height, body mass index (BMI), and blood pressure, and the main biochemical parameters were serum sodium and potassium. Maternal outcome was classified as good, with complications, or fatal. Likewise, fetal outcome was defined as good, with complications, or fatal.

We converted glucocorticoid medication to the equivalent dose of immediate-release HC dosage as follows: 1 mg of dual-release HC as 1 mg of HC equivalent, 1 mg of prednisone and prednisolone as 4 mg of HC, 1 mg of cortisone acetate as 0.8 mg of HC, and 1 mg of dexamethasone as 25 mg of HC (13). The mineralocorticoid potency of 100 µg of fludrocortisone was defined as 1 mineralocorticoid unit (MCU). For the glucocorticoids, we calculated 1 mg of immediate and dual-release HC as a mineralocorticoid potency of 0.054 MCU, 1 mg of prednisone or prednisolone as 0.013 MCU, 1 mg of cortisone acetate as 0.0432 MCU, while dexamethasone was regarded as 0 MCU (14).

Statistical analysis

Each pregnancy was considered as an independent case, including ones occurring in the same patient. The retrospective, case-based form of the study resulted in missing data for different variables that could be explained either by lack of documentation or incomplete follow-up of the patient by the participating centers. Further missing data across all pregnancies were explained by miscarriages or ongoing pregnancies. Data were gathered from participating centers in an anonymized form. Statistical analysis was performed using IBM SPSS Statistics for Windows (released 2017, Version 25.0., IBM Corp, Armonk, NY). Graphs were generated using GraphPad Prism 5 (GraphPad Software, La Jolla, CA). Variables were assessed for normality by evaluation of histograms and by the Kolmogorov–Smirnov test. Differences between groups were assessed using Fisher’s exact test for categorical variables, the Mann–Whitney U-test for quantitative non-normally distributed variables, and Student’s t-test for normally distributed variables. Overall, comparisons of continuous variables between groups were carried out with analysis of variance (ANOVA) or the Kruskal–Wallis test. Correlations were assessed by Pearson’s or Spearman’s correlation coefficient (r). For paired data, the Wilcoxon test was used. A probability value of P < .05 was considered statistically significant for all tests. Unless stated otherwise, P-values are denoted as follows: *P < .05; **P < .01; ***P < .001.

Results

Description of patients with AI

The cause of AI was AD in 56 pregnancies (44%), secondary AI in 32 pregnancies (25%), and congenital adrenal hyperplasia (CAH) in 32 pregnancies (25%). In 8 pregnancies (6%), the cause was bilateral adrenalectomy performed for treatment of bilateral pheochromocytoma or therapy-resistant adrenocorticotropic hormone-dependent Cushing syndrome.

The mean age of women at the time of AI diagnosis differed significantly among the subgroups, with 26 years (range 14-35, n = 48) for primary AI, 26 years (range 8-44, n = 31) for secondary AI, 9 years (range 0-38, n = 29) for CAH, and 30 years (range 23-37, n = 6) for other causes of AI (P < .001, Table 1). In contrast, the mean age at the time of pregnancy was similar among different groups with 33 ± 4.5 years in women with AD, 32 ± 5.5 in women with secondary AI, 32 ± 6.1 in CAH, and 35 ± 3.8 in AI due to other reasons (P = .470, Table 1). In most pregnancies (125/128, 98%), the diagnosis of AI had been established before pregnancy. In 2 cases, AI was diagnosed during the third trimester (in the 28th and 34th weeks of gestation).

Table 1.

Baseline characteristics of pregnant patients with adrenal insufficiency.

	AD (n = 56)	Secondary AI (n = 32)	CAH (n = 32)	Other reason for AI (n = 8)	P-value
Age at diagnosis (years)	26 (14-35)	26 (8-44)	9 (0-38)	30 (23-37)	<.001  a
Age at pregnancy (years)	32.7 ± 4.5	32.0 ± 5.5	31.8 ± 6.1	34.9 ± 3.8	.470b
Baseline BMI (kg/m2)	24.3 ± 4.5	27.5 ± 6.9	25.7 ± 4.6	24.3 ± 6.0	.143a

Values are given as median (range) or mean ± standard deviation.

Abbreviations: AD, Addison disease; AI, adrenal insufficiency; CAH, congenital adrenal hyperplasia; BMI, body mass index;

^aKruskal–Wallis test.

^bOne-way ANOVA.

Among women with secondary AI, 45% (14/31) had at least 1 additional hormonal deficiency. Thyroid dysfunction was the most commonly reported additional comorbidity, with 46.8% (52/111) having hypothyroidism. Hypothyroidism coexisted in primary AI as primary hypothyroidism in 55.3% (26/47) and as secondary hypothyroidism in women with secondary AI in 58.6% (17/29) of cases. In contrast, fewer patients with CAH had hypothyroidism (21%). Some women had other diseases of the thyroid, for example, multinodular goiter (n = 1), thyroid nodules (n = 2), hyperthyroidism that was not otherwise specified (n = 1), Graves disease (n = 1), and papillary thyroid carcinoma (n = 2).

Reproductive history of patients with AI

Of the women, 28.3% (36/127) had at least 1 miscarriage in their past medical history. The percentage of miscarriages was significantly higher among patients with CAH (43.8%, 14/32) than among patients with AD (27.3%, 15/55) and was considerably lower in patients with secondary AI (15.6%, 5/32) or due to other diseases (25%, 2/8) (all P < 0.001). The exact week of miscarriage was recorded in 28 cases, with 75% of miscarriages occurring in the first trimester and 25% in the second trimester. From 2013 onwards, 3 miscarriages were reported: 3 in the ninth week and 1 in the 15th week of gestation.

Management of glucocorticoid and mineralocorticoid substitution therapy during pregnancy

Glucocorticoid substitution

At baseline, immediate-release HC was the most commonly used glucocorticoid (67/100, 67%) with considerably fewer patients treated with another single medication (18/100, 18% dual-release HC; 8/100, 8% prednisolone; 2/100, 2% dexamethasone, 2/100, 2% cortisone acetate), or combinatory therapies (2/100, 2% immediate-release HC with dexamethasone and 1/100, 1% immediate-release HC with prednisolone). Treatment data at baseline were not available for 26 pregnancies, while in a further 2 cases AI was first diagnosed during pregnancy. Of note, in 1 case, immediate-release HC was given continuously through a subcutaneous pump throughout pregnancy (15).

During the first trimester, in 80/117 (68.4%) pregnancies the reported glucocorticoid was immediate-release HC, in 15/117 (12.8%) dual-release HC, in 2/117 (1.7%) a combination of both, in 11/117 (9.4%) prednisolone, in 7/117 (6.0%) cortisone acetate, and in 2/117 (1.7%) a combination of prednisolone with HC. In contrast, no cases of dexamethasone treatment were reported. The dosage was split for the immediate-release HC formulation to 2 times per day in 28/80 (35.0%) cases, 3 times per day in 51/80 (63.7%) of the pregnancies, and 4 times per day for 1 pregnancy (1/80, 1.3%). Of note, in 2 pregnancies with dual-release HC (2/15, 13.3%), the treatment scheme was 20 mg in the morning supplemented with another 5 mg in the afternoon.

The applied glucocorticoid during the further course of pregnancy remained comparable with the first trimester (data not shown). Of note, dexamethasone was added as the only glucocorticoid in 1 patient with CAH during week 17 and in 1 patient with CAH during week 32 of pregnancy.

Glucocorticoid dosage was not adapted uniformly, in line with existing guidelines. Instead, the data revealed that individualized dosage adaptation during the different trimesters did not follow a certain pattern, but there were also cases where the dosage was not changed during the whole course of pregnancy (Figure 1). No occurrence of an adrenal crisis was reported in any of the cases with a decrease in the dosage. During the last trimester, 5/109 patients (4.6%) had a decrease in daily dosage of 5 to 10 mg of HC equivalent due to the development of hypertension. Among the different subgroups, those with AD tended to receive the highest substitution dosage (Table 2).

Table 2.

Hydrocortisone equivalent dosage and total mineralocorticoid factor before pregnancy and per trimester divided in different subgroups of patients with adrenal insufficiency.

	AD	Secondary AI	CAH	Other reasons of AI	P-valuea
Before pregnancy
Daily HC equivalent dosage (mg)	22.4 (15.0-31.4) n = 39	18.1 (7.5-40.0) n = 25	21.8 (6.3-50.0) n = 31	18.5 (15-20) n = 5	.012
Total MCU	2.0 (0.8-3.4) n = 39	0.9 (0.0-2.2) n = 26	1.5 (0.1-4.2) n = 29	1.9 (1.1-2.5) n = 5	<.001
First trimester
Daily HC equivalent dosage (mg)	24.0 (15-40) n = 53	18.8 (7.5-40) n = 25	21.8 (7.5-45) n = 30	23.8 (20-30) n = 8	.02
Total MCU	2.1 (0.0-4.2) n = 56	0.7 (0.0-2.2) n = 32	1.3 (0.0-4.2) n = 31	2.3 (1.4-3.1) n = 8	<.001
Second trimester
Daily HC equivalent dosage (mg)	27.4 (15-70) n = 50	21.6 (10-60) n = 28	23.9 (7.5-45) n = 30	25 (20-30) n = 7	.24
Total MCU	2.4 (0.9-5.8) n = 51	1.0 (0.5-2.2) n = 28	1.4 (0.1-4.2) n = 29	2.6 (1.4-4.6) n = 7	<.001
Third trimester
Daily HC equivalent dosage (mg)	26.6 (10-40) n = 50	23.0 (10-50) n = 27	25.6 (7.5-50) n = 28	25.7 (20-35) n = 7	.053
Total MCU	2.4 (0.9-4.1) n = 50	1.1 (0.5-2.2) n = 27	1.8 (0.1-5.6) n = 26	2.7 (1.4-4.4) n = 7	<.001

Values are given as median (range) or mean ± standard deviation.

Abbreviation: AD, Addison disease; AI, adrenal insufficiency; CAH, congenital adrenal hyperplasia; HC, hydrocortisone; MCU, mineralocorticoid unit.

^aKruskal–Wallis test

Figure 1.

Dosage adaptation for daily glucocorticoid (upper panel) and mineralocorticoid (middle panel) substitution therapy and calculated total mineralocorticoid potency (lower panel) during pregnancy for women with adrenal insufficiency. TM, trimester; HC, hydrocortisone. Asterisks denote significant differences comparison with baseline (*P < .05; **P < .01; ***P < .001).

Mineralocorticoid substitution

Before pregnancy, 95% (37/39) of patients with AD, 42% (13/31) of patients with CAH, and 80% (4/5) of patients with AI due to other diseases were receiving fludrocortisone treatment. As expected, none of the patients (0/32) with secondary AI had mineralocorticoid substitution. During the whole course of pregnancy, fludrocortisone dosage was adapted in some cases but to a lesser extent compared with HC adaptation (Figure 1). The reason for the decrease in fludrocortisone dosage was not recorded; no correlation with increased blood pressure or onset of pre-eclampsia was documented.

Clinical course and adrenal crisis during pregnancy

Blood pressure measurements as well as serum sodium and potassium remained in the normal range during the course of pregnancy in the majority of cases (Table 3 and Figure 2). Hypertensive values (defined as values above 140/90 mmHg) were reported in 6/94 (6.4%) at baseline, in 2/102 (2%) in the first trimester, in 2/102 (2%) during the second trimester, and in 7/100 (7%) in the third trimester. No relevant differences were noted among the subgroups. Similarly, no significant differences in systolic and diastolic blood pressure were evident during pregnancy comparison with baseline levels (Figure 2). In contrast, sodium concentrations decreased significantly compared with baseline (139.0 ± 3.1; range 125-145 mmol/L) in the first trimester (136.3 ± 2.7; range 130-144 mmol/L, P < .001), second trimester (136.7 ± 2.9; range 132-145 mmol/L, P < .001), and third trimester (137.7 ± 2.8; range 129-145 mmol/L, P = .0036). Low serum sodium concentrations (<135 mmol/L) were reported in 3/79 (3.8%) at baseline, in 21/94 (22.3%) in the first trimester, in 22/91 (24.2%) in the second trimester, and in 9/82 (11.0%) in the third trimester.

Table 3.

Development of blood pressure and serum electrolytes during pregnancy in patients with adrenal insufficiency

	AD	Secondary AI	CAH	Other reasons of AI	P-valuea
Before pregnancy
Systolic blood pressure (mmHg)	113.8 ± 13.4 n = 39	115.4 ± 16.7 n = 25	110.9 ± 11.2 n = 26	107.5 ± 12.6 n = 4	.558
Diastolic blood pressure (mmHg)	72.4 ± 9.6 n = 39	70.7 ± 10.3 n = 25	70.0 ± 9.2 n = 26	68.3 ± 6.2 n = 4	.690
Serum sodium (mmol/L)	138.5 ± 3.2 n = 40	139.4 ± 3.7 n = 25	139.5 ± 2.2 n = 25	138.0 ± 1.4 n = 5	.449
Serum potassium (mmol/L)	4.3 ± 0.5 n = 40	4.1 ± 0.4 n = 25	4.2 ± 0.4 n = 25	4.1 ± 0.1 n = 5	.342
First trimester
Systolic blood pressure (mmHg)	109.7 ± 11.8 n = 46	114.7 ± 9.7 n = 24	110.0 ± 12.4 n = 26	116.8 ± 17.4 n = 6	.225
Diastolic blood pressure (mmHg)	71.2 ± 9.93 n = 46	70.4 ± 8.8 n = 24	71.0 ± 8.6 n = 26	75.0 ± 18.8 n = 6	.790
Serum sodium (mmol/L)	135.2 ± 2.1 n = 44	137.5 ± 2.9 n = 15	137.6 ± 2.6 n = 23	136.1 ± 2.7 n = 7	<.001
Serum potassium (mmol/L)	4.2 ± 0.3 n = 43	4.1 ± 0.5 n = 15	4.0 ± 0.4 n = 23	4.3 ± 0.2 n = 7	.076
Second trimester
Systolic blood pressure (mmHg)	108.4 ± 11.1 n = 45	114.9 ± 15.3 n = 23	111.1 ± 9.4 n = 27	120.6 ± 12.9 n = 7	.034
Diastolic blood pressure (mmHg)	66.0 ± 8.3 n = 45	71.9 ± 8.1 n = 23	69.1 ± 7.7 n = 27	78.2 ± 10.4 n = 7	.001
Serum sodium (mmol/L)	135.9 ± 2.7 n = 36	138.1 ± 2.5 n = 18	136.3 ± 3.9 n = 24	136.7 ± 2.8 n = 6	.052
Serum potassium (mmol/L)	4.1 ± 0.4 n = 40	3.9 ± 0.5 n = 25	4.0 ± 0.3 n = 25	4.1 ± 0.4 n = 5	.415
Third trimester
Systolic blood pressure (mmHg)	111.4 ± 12.3 n = 43	117.0 ± 16.7 n = 25	118.7 ± 16.4 n = 25	112.4 ± 12.0 n = 7	.487
Diastolic blood pressure (mmHg)	74.1 ± 11.9 n = 43	75.7 ± 14.1 n = 25	70.2 ± 8.6 n = 25	71.7 ± 6.2 n = 7	.359
Serum sodium (mmol/L)	137.2 ± 2.9 n = 35	138.6 ± 2.5 n = 19	137.7 ± 2.7 n = 22	137.8 ± 3.1 n = 6	.376
Serum potassium (mmol/L)	4.1 ± 0.4 n = 35	4.0 ± 0.4 n = 19	4.0 ± 0.4 n = 22	4.1 ± 0.3 n = 6	.653

Values are given as median (range) or mean ± standard deviation.

Abbreviations: AD, Addison disease; AI, adrenal insufficiency; CAH, congenital adrenal hyperplasia; BMI, body mass index.

^aANOVA

Figure 2.

Systolic and diastolic blood pressure (upper panel) and serum electrolytes with sodium (middle panel) and potassium (lower panel) during pregnancy for women with adrenal insufficiency. TM, trimester. Asterisks denote significant differences in comparison to baseline (*P < .05; **P < .01; ***P < .001).

With regard to serum potassium levels, only minor changes were observed between baseline (4.2 ± 0.4, range 3.0-5.5 mmol/L) and second trimester (4.0 ± 0.4, range 3.0-5.0 mmol/L, P = .0234; Figure 2). Low serum potassium levels (<3.5 mmol/L) were reported in 2/79 (2.5%) at baseline and in 3/94 (3.2%), 5/91 (5.5%), and 5/82 (6.1%) during the first, second, and third trimesters, respectively.

No cases of hypernatremia (>145 mmol/L) or hyperkalemia (>5.5 mmol/L) were reported at any time point. When taking into consideration the mineralocorticoid substitution and serum electrolytes in the third trimester of pregnancy, a negative correlation between the total mineralocorticoid potency and serum sodium (R = −0.217, P = .047) and a positive correlation for serum potassium were observed (R = 0.281, P = .011 Figure 3). In contrast, no significant correlation was found at any other time point before or during pregnancy (Figure 3).

Figure 3.

Relation between mineralocorticoid potency and serum electrolytes at baseline and during the 3 trimesters of pregnancy. AD, Addison disease; AI, adrenal insufficiency; CAH, congenital adrenal hyperplasia.

In 9/128 (7%) of pregnancies, an adrenal crisis was reported without any relationship to the etiology of AI (Table 4). In 1 case, an adrenal crisis was precipitated by an influenza A infection, which was successfully managed. One patient experienced 4 recurring episodes of adrenal crisis due to hyperemesis gravidarum during the first trimester. In 1 case, the trigger was gastroenteritis during the first trimester. One patient was diagnosed with new AI when she presented with adrenal crisis in the third trimester. In the other 5 cases, no further data regarding onset and cause of adrenal crisis were available.

Table 4.

Course and outcome of pregnancies in patients with adrenal insufficiency

	AD (n = 56)	Secondary AI (n = 32)	CAH (n = 32)	Other reason for AI (n = 8)	P-value
Spontaneous abortions (%)	15/55 (27.3)	5/32 (15.6)	14/32 (43.8)	2/8 (25.0)	<.001  a
Adrenal crisis during pregnancy (%)	5/56 (8.9)	2/32 (6.3)	2/32 (6.3)	0/8 (0)	.607a
Cesarean section (%)	24/50 (48.0)	20/29 (69.0)	20/31 (64.5)	5/8 (62.5)	.224 a
Week of labor	37.9 ± 2.2	38.7 ± 1.8	37.0 ± 3.3	37.5 ± 2.2	.075b
Preterm births (%)	8/50 (16.0)	6/28 (21.4)	10/31 (32.3)	1/8 (12.5)	.335a
First APGAR score	8.3 ± 1.5	8.6 ± 1.8	8.7 ± 0.8	8.3 ± 1.0	.739b
Child height (cm)	48.7 ± 3.3	51.0 ± 5.1	48.1 ± 4.7	51.5 ± 0.7	.337b
Child weight (g)	3261 ± 606	3322 ± 674	2934 ± 696	2814 ± 344	.063b

Values are given as median (range) or mean ± standard deviation.

Abbreviations: AD, Addison disease; AI, adrenal insufficiency; CAH, congenital adrenal hyperplasia.

^aKruskal–Wallis test.

^bOne-way ANOVA.

Delivery management and outcome

Overall, 25/117 (21.4%) deliveries were reported as preterm without significant differences between the subgroups. Among those preterm births in patients with CAH 1/10 (10%) was defined as extremely preterm (25th week) in the context of chorioamnionitis. A further early preterm birth (28th week) in a patient with AD was reported, but no further details of the etiology were available. The remaining preterm births were moderate to late preterm (28th to 32nd week).

Information regarding administration of a stress dosage of glucocorticoid during delivery was available in 90/128 cases (70.3%); of these, glucocorticoids were administered intravenously in 83/90 (92.2%) cases. In the 7 cases where no stress dosage was provided, no adverse events were reported. In 54/90 cases, the exact dosage of substitution medication was available. The majority of these cases (36/54, 66.7%) received 100 mg of HC, but overall a wide range of dosages of HC supplementation were used ranging from 50 to 300 mg.

Mode of delivery was cesarean section (CS) in 69/118 (58%) of the cases. However, only a few cases had an emergency CS (n = 4) due to pre-eclampsia. Vaginal delivery was slightly more common in patients with AD (52%, 26/50). CS was more common in patients with secondary AI (20/29, 69%), patients with CAH (20/31, 65%), and patients with other types of AI (5/8, 63%). The exact reason for CS was not documented in these cases.

In 15/120 (12.5%) cases, maternal complications were reported: anemia (n = 2), severe postpartum hemorrhage (n = 4), mild to severe pre-eclampsia (n = 7), and chorioamnionitis (n = 1). However, no adverse fetal outcomes were reported in these cases. In three cases, miscarriage was documented, 2 in ninth week and 1 in the 15th week of pregnancy. Two pregnancies were ongoing at the time of study completion. In 3 cases, fetal complications occurred, with 2 premature babies requiring intensive care and 1 having macrosomia and shoulder dystocia in a woman with type 1 diabetes. No maternal or fetal fatalities occurred.

Discussion

To our knowledge, this is the first retrospective multicenter survey with a detailed analysis of course, management, and outcome of a large number of pregnancies in women with AI of different etiologies. As these women had regular follow-up at specialized endocrine units during the course of their pregnancy, it was possible to collect and analyze information that sheds light on the current treatment and management of this rare clinical condition and to provide insights on the outcome of affected mothers and their offspring.

Miscarriages were relatively prevalent in the past medical history of included women, being highest among those with CAH. A high rate of miscarriage among patients with CAH has previously been described in a German cohort that included 39 patients with either the classical or the nonclassical form of CAH (16) and was attributed to a combination of androgen excess and cortisol insufficiency. A Swedish study reported increased rates of therapeutic abortions among 62 patients with CAH, while the rate of spontaneous abortions was not higher than in a control population (17).

Interestingly, in the current dataset, women with secondary AI had the lowest rate of spontaneous abortion, but this could be explained by the use of assisted reproduction methods for conception often with frequent follow-up during pregnancy. Overall, the reasons of reduced fertility in women with secondary AI are believed to be caused by additional hormonal deficiencies such as hypothyroidism and or secondary hypogonadism (7). Almost 1 out of 3 patients with AD reported a miscarriage in their past medical history, but the reasons could not be determined in the context of the present study. However, overall, this rate is within the range (15-30%) of the normal population (18, 19).

In the past, many cases of primary AI were reported to result in maternal and fetal death, but recent reports correlate adverse events mostly with poor compliance (20). In the present study, the rate of miscarriage among patients with AD, CAH, and AI due to other etiologies was comparable with the reported rate in a similar study of patients with primary AI in the German population (16). Of note, in that study, all miscarriages were reported among patients that suffered from primary AI in the context of autoimmune polyendocrine syndrome type 2. This information was not available in our dataset and we could not access the incidence of miscarriages in parallel with other manifestations of the syndrome, for example, in combination with hypothyroidism.

The objective of following up a woman with AI during pregnancy is to keep replacement therapy at levels that avoid the effects of over-treatment (eg, gestational diabetes, excessive weight gain, arterial hypertension) and of under-treatment (adrenal crisis, electrolyte imbalance). In keeping with the general recommendation from the Endocrine Society (11), glucocorticoids were in most instances increased during the course of pregnancy in a range between 5 and 25 mg of HC equivalents. However, adjustment of glucocorticoid dosage varied significantly among patients and centers, emphasizing that current practice is a rather individualized approach.

Likewise, fludrocortisone dosage was adjusted in almost a quarter of patients, without a clear relationship with potassium concentrations or blood pressure measurements. Electrolyte balance during pregnancy is affected by the mineralocorticoid antagonistic action of progesterone (21, 22). To counterbalance the effects of increased progesterone levels during pregnancy, higher doses of a mineralocorticoid receptor agonist are required (22, 23). Proper adjustment of mineralocorticoid dosage is accomplished by evaluating blood pressure and assessment of blood electrolytes, as active renin cannot be used as a reliable indicator for treatment during pregnancy (14). Despite the observed tendency of increasing mineralocorticoid potency during the course of pregnancy in the majority of cases, serum sodium showed an initial decrease in the first trimester. Afterwards, there was no further change in serum sodium, and serum potassium remained stable during the course of pregnancy. Our findings support an earlier report that pregnant women with primary AI have an increasing requirement for fludrocortisone substitution as pregnancy advances to maintain blood pressure and serum potassium in the normal range (24). However, during normal pregnancy despite increased sodium retention the increase in plasma volume results in mildly reduced serum sodium concentrations. Accordingly, reference intervals for pregnant women are approximately 2 to 5 mmol/L lower than those outside of pregnancy (25). Therefore, the observed reduced mean serum sodium levels during pregnancy could be explained by a combination of physiological effects of pregnancy and modulation by mineralocorticoid action implemented by the substitution treatment. It is interesting to note that the current data for the third trimester indicated a negative correlation between the total mineralocorticoid potency and serum sodium, but it was positive for serum potassium levels. This pattern provides indirect evidence that changes in electrolytes were the trigger for an increase in substitution therapy but were less likely to be explained by steroid overdosage.

In the current study, the reported rate of adrenal crisis of around 7% is within the expected range, considering the high incidence of adrenal crisis even among educated populations based on recent studies (10, 26). To avoid development of an adrenal crisis during delivery, the endocrinologist should provide the obstetric team with a written therapeutic plan regarding intravenous glucocorticoid coverage. This should be started before the active phase of labor with a recommended initial bolus of 100 mg of HC followed by continuous infusion or bolus dosages every 6 hours (27). However, there is no universal agreement regarding the timing and dosage adaptation of glucocorticoid substitution peri- and post partum. In fact, in the vast majority of cases in the current evaluation, glucocorticoid in stress dosage was administered intravenously. However, a wide variety of HC dosage was evident during labor, and individual cases reported of missing or delayed stress coverage.

A further surprising finding was the predominantly high rate of CS as the delivery mode among patients with secondary AI. Because of the retrospective nature of data retrieval, the indications for CS remain uncertain. In accordance with previous findings, the rate of CS among patients with CAH was high, for which some percentage can be assumed to be related to earlier genital surgery (16, 20, 27). Furthermore, the increased rate of CS could also be related to the overall frequency of CS in large tertiary care hospitals (28).

Independent of the mode of delivery, most deliveries were without complications. Likewise, our study revealed good maternal and fetal outcomes for the overall patient cohort. Children born preterm were reported in all groups but were more frequent among patients with CAH. Similarly, the mean weight and height of these children tended to be lower. In a comparable study from Germany, children born to patients with CAH weighed significantly less and had a tendency to be smaller than the general population (16). As our study lacked information regarding the sex of the children, we were unable to report the number of children born small for gestational age. Furthermore, considering the multicentric nature of the study with data from different countries, a clear comparison with the respective general population was not possible.

Limitations of our study include possible missing data, over- or under-reporting of some aspects of baseline or follow-up parameters, and selection bias. In particular, it should be noted that those centers that took part in the current study are specialized in the diagnosis and treatment of patients with AI and are more likely to be integrated into multidisciplinary teams as well as providing patient management based on their clinical experience in addition to the existing guidelines. Therefore, it is possible that data gathered from expert centers will not be representative for the overall patient population. The strengths of our study include the availability of an extensive and recent dataset reflecting the current clinical management of pregnancies in women with AI from a variety of countries.

Taken together, this survey confirms good maternal and fetal outcome of pregnancies in AI treated in specialized centers. These data provide reassuring grounds for counselling of women with known AI and motivation for regular endocrine follow-up to adjust glucocorticoid and mineralocorticoid dosage during the course of pregnancy. Regarding the etiology and the prevention of miscarriages among patients with AI, further prospective studies are required.

Abbreviations

Abbreviations
 
• AI

  adrenal insufficiency

 
• AD

  Addison disease

 
• ANOVA

  analysis of variance

 
• CAH

  congenital adrenal hyperplasia

 
• BMI

  body mass Index

 
• HC

  hydrocortisone

 
• MCU

  mineralocorticoid unit

 
• CAH

  congenital adrenal hyperplasia

 
• CS

  cesarean section

Acknowledgments

Financial Support: The study was supported by the Deutsche Forschungsgemeinschaft (DFG) within the CRC/Transregio 205/1 “The Adrenal: Central Relay in Health and Disease” to N.R., S.H., and F.B. This research was supported by the James A. Ruppe Career Development Award in Endocrinology (I.B.) and the Catalyst Award for Advancing in Academics from Mayo Clinic (I.B.). This research was partly supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the National Institutes of Health (NIH) USA under award K23DK121888 (to I.B).

Additional Information

Disclosure Statement: F.B. serves as an associate editor for the Journal of Clinical Endocrinology & Metabolism. I.B. reports advisory board participation with Corcept, HRA Pharma, and ClinCor outside the submitted work. The views expressed are those of the author(s) and not necessarily those of the National Institutes of Health USA. All other authors have nothing to disclose. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Data Availability: Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.