New-onset primary adrenal insufficiency in pregnancy associated with a unilateral adrenal infarction: a case report

Abstract

Summary

New-onset primary adrenal insufficiency is rare in pregnancy. The symptoms of adrenal insufficiency such as nausea, vomiting and dizziness may be attributed to the pregnancy itself, which can lead to a delay in the diagnosis. The presence of hypotension, hypoglycemia or hyperkalemia should raise the suspicion for adrenal insufficiency. We report the case of a 25-year-old woman who presented with tachycardia, left flank pain and vomiting at 36 weeks’ gestation. She was found to have primary adrenal insufficiency and started on hydrocortisone and fludrocortisone with resolution of the vomiting and tachycardia. MRI of the abdomen revealed an acute nonhemorrhagic infarct of the left adrenal gland. The contralateral adrenal gland was normal. Autoimmune and infectious etiologies of primary adrenal insufficiency were ruled out and the adrenal insufficiency was attributed to the unilateral adrenal infarction. Adrenal insufficiency persisted after delivery and then resolved at approximately 16 months post partum. This case highlights the need to test women with unilateral adrenal infarction in pregnancy for the presence of primary adrenal insufficiency.

Learning points

• Adrenal insufficiency should be considered when a pregnant woman develops nausea, vomiting and dizziness in association with hypotension or hypoglycemia. Hypovolemic hyponatremia related to vomiting can occur in pregnancy, but the failure to correct hyponatremia despite adequate IV hydration should raise the suspicion for adrenal insufficiency.

• Adrenal infarction should be in the differential diagnosis for unilateral flank pain in pregnancy. Other common etiologies for flank pain in pregnancy include nephrolithiasis, pyelonephritis and acute cholecystitis.

• Unilateral adrenal infarction in pregnancy can lead to the development of primary adrenal insufficiency. Following delivery, these patients need to be monitored for the resolution of the adrenal insufficiency.

Introduction

Adrenal insufficiency (AI) is characterized by deficient or absent cortisol production. It is classified as primary AI when there is destruction of the adrenal cortex or defects in adrenal steroidogenesis and secondary AI when damage occurs at the level of the pituitary gland or hypothalamus. AI can be insidious in onset and lead to symptoms such as fatigue, anorexia, nausea, vomiting and unintentional weight loss or it may present acutely as an adrenal crisis (1). The symptoms and signs of adrenal crisis include hypotension, tachycardia, impaired cognitive function and abdominal pain. Hyponatremia, hyperkalemia and hypoglycemia are frequently noted during an adrenal crisis (2). AI is rare in pregnancy with a prevalence of 5.5/100 000 (3). Pregnant women with AI are at increased risk of preterm delivery and delivery by cesarean section (3, 4). The patients included in these studies had AI that preceded the pregnancy. It is unclear whether these risks also apply to patients who develop new-onset AI in pregnancy.

We present the case of a 25-year-old woman who developed new-onset primary AI during pregnancy leading to the development of an adrenal crisis. Investigation into the etiology of the AI led to the finding of a subacute non-hemorrhagic infarct of the left adrenal gland (NHAI). No alternative etiologies for the primary AI were found. This case highlights the need to consider AI in pregnant women who present with a unilateral adrenal infarction.

Case presentation

A 25-year-old woman, who was at 36 weeks’ gestation of her second pregnancy, presented to labor and delivery triage with a 2-day history of left flank pain and 1-day history of nausea, vomiting and dizziness. The flank pain radiated to the front of her abdomen and was 7/10 in severity. She denied dysuria, vaginal bleeding, contractions and decrease in fetal movement. She had no past medical history and her first pregnancy resulted in an uncompleted spontaneous vaginal delivery 18 months ago. There was no history of aspirin, anti-coagulant or illicit drug use. Vital signs showed a blood pressure of 111/65 mm Hg, heart rate of 129 bpm, respiratory rate of 24/min, SpO₂ of 100% and temperature of 98.3°F. She had tenderness of the left flank. Her abdomen was gravid and there was no abdominal tenderness or contractions. Her cervix was closed, thick and high. No cutaneous or mucosal hyperpigmentation was noted.

Investigation

Initial laboratory investigation revealed hyponatremia to 125 mmol/L (reference range: 135–145 mmol/L), hypoglycemia to 54 mg/dL and low serum bicarbonate of 14 mmol/L (reference range: 22–32 mmol/L) (Table 1). Her potassium and creatinine were in the normal range. Thyroid hormone level was in the normal range with a TSH of 2.51 µIU/mL (reference range: 0.34–5.60 µIU/mL) and free T4 of 1.13 ng/dL (reference range: 0.58–1.64 ng/dL). Urinalysis showed moderate leukocyte esterase and three to five white blood cells per high power field (WBC/HPF).

Table 1.

Serial laboratory data during hospitalization.

Laboratory test (normal reference range)	At admission	Day 2 of hospitalization	Day 3 of hospitalization
Sodium (135–145 mmol/L)	125 mmol/L	133 mmol/L	136 mmol/L
Potassium (3.5–5.1 mmol/L)	3.8 mmol/L	3.9 mmol/L	3.6 mmol/L
Chloride (98–107 mmol/L)	102 mmol/L	106 mmol/L	110 mmol/L
Bicarbonate (22–32 mmol/L)	14 mmol/L	12 mmol/L	17 mmol/L
Glucose (70–110 mg/dL)	54 mg/dL	48 mg/dL	152 mg/dL
ACTH (7.2–63.3 pg/mL)		220.8 pg/mL
Cortisol (5–23 µg/dL)		7.7 µg/dL

Hyponatremia and hypoglycemia were noted at admission which persisted despite IV hydration. Her hyponatremia and hypoglycemia were corrected after the administration of stress dose hydrocortisone. Bold text indicates lab results that are outside the normal range.

ACTH, adrenocorticotropic hormone.

A CT chest with pulmonary embolism (PE) protocol was obtained due to the tachycardia and tachypnea which showed no central or segmental PE. An indeterminate 1.9 cm left adrenal mass was incidentally noted (Fig. 1). As part of the workup for the incidentally noted adrenal mass, adrenocorticotropic hormone (ACTH) level was obtained which was elevated at 220.8 pg/mL (reference range: 7.2–63.3 pg/mL) and the AM cortisol was low for pregnancy at 7.7 µg/dL (non-pregnancy reference range: 5–23 µg/dL). A cosyntropin stimulation test was not done. MRI of the abdomen with and without contrast was obtained 15 days later following delivery which showed that the left adrenal gland had a central T2 hyperintensity with a rim of T2 hypointensity and was nonenhancing (Fig. 2). The findings were reported to be consistent with subacute NHAI. The right adrenal gland appeared normal. The 21-hydroxylase antibody, T-spot tuberculosis test and urinary Histoplasma antigen were negative.

Figure 1.

CT chest with PE protocol showing the incidental finding of a 1.9 cm left adrenal hypodensity (red arrow). The green arrow shows the right adrenal gland which is normal in morphology.

Figure 2.

MRI abdomen showing T2 hyperintensity in the left adrenal gland with a hypoechoic rim. The red arrow shows the left adrenal gland.

Treatment

She was started on normal saline infusion at 125 mL/h for IV hydration and ondansetron 4 mg q6h for nausea. One injection of 12.5 mg 50% dextrose was given IV for the hypoglycemia. The patient was then able to tolerate the intake of clear liquids but had intermittent vomiting and continued to experience hypoglycemia. Her IV fluids were changed to 5% dextrose with ½ normal saline at 125 mL/h. She had persistent hypotension despite the IV hydration with the lowest BP reading of 88/59 mm Hg due to which she received an additional 1 L normal saline bolus with a transient improvement in BP. Cephalexin 2 gram q24h was started as empiric treatment for a urinary tract infection (UTI).

Following the finding of elevated ACTH on the second day of hospitalization, hydrocortisone 100 mg IV q8h was started. Her hypotension resolved after the initiation of steroids and there was resolution of her nausea and vomiting over the subsequent 24-h period. The hydrocortisone dose was weaned to 20 mg in the morning and 10 mg in the afternoon over the following 2 days and she was also started on fludrocortisone 0.05 mg daily. The patient was discharged home on this regimen on the fourth day of hospitalization.

Outcome and follow-up

The patient had a scheduled induction of labor at 38 + 5 weeks’ gestation. Hydrocortisone 100 mg IV q8h was administered during labor. She had an uncomplicated vaginal delivery and the infant was healthy. Hydrocortisone and Fludrocortisone were continued after discharge.

She had her first appointment in the endocrinology clinic 2 months after delivery at which time the hydrocortisone dose was lowered to 15 mg in the morning and 5 mg in the afternoon and the fludrocortisone 0.05 mg daily was continued. Labs showed persistent elevation of ACTH to 151.9 pg/mL along with a low dehydroepiandrosterone sulfate (DHEAS) of 9 µg/dL (reference range: 65–380 µg/dL), which was evidence of ongoing adrenal insufficiency. The patient discontinued the fludrocortisone on her own after this appointment and denied the development of dizziness following discontinuation. She was normotensive with normal renin activity of 2.9 ng/mL/h (reference range: 0.5–4.0 ng/mL/h) and normal potassium level due to which the fludrocortisone was not resumed. Serial ACTH and DHEAS were obtained which showed elevated ACTH and low DHEAS.

The ACTH level was noted to have normalized 16 months after delivery but the DHEAS level remained persistently low. The patient felt well and was eager to stop the hydrocortisone because she felt that the hydrocortisone was making it difficult for her to lose weight. A cosyntropin stimulation test was then obtained which showed a peak post-cosyntropin cortisol level of 14. 5 µg/dL using the Access cortisol immunoassay. This was considered an adequate response given the lower cutoff for newer cortisol assays (5). Following shared decision-making, the decision was made to stop hydrocortisone. The patient was educated about the possible symptoms of recurrence of adrenal insufficiency. The patient did well after stopping the hydrocortisone and did not have recurrence of nausea, vomiting or dizziness.

Discussion

Pregnancy is associated with activation of the hypothalamic–pituitary–adrenal axis. There is increase in the plasma total cortisol due to the increased hepatic production of cortisol-binding globulin (CBG) under the effect of estrogen and increased free cortisol due to the displacement of cortisol from CBG by progesterone (6). The plasma total cortisol level rises in parallel to gestation and ranges from 16.3–55 µg/dL in the third trimester (7). Our patient’s cortisol level was low for the third trimester at 7.7 µg/dL. Given that the patient was critically ill, we favored giving stress dose of hydrocortisone immediately over performing a cosyntropin stimulation test. The placenta is a source of ACTH and the ACTH level rises through pregnancy (7). In one study, a 3-fold elevation in the ACTH level was found in the third trimester compared to the first trimester (8). The ACTH level reaches a peak during delivery and normalized by 24 h after delivery (8). In our patient, the ACTH level was almost 4-fold elevated in the third trimester and remained elevated for over a year post partum. This supports the diagnosis of primary AI.

The most common etiology of primary AI is autoimmune adrenalitis (Addison’s disease). Elevated levels of 21-hydroxylase antibody are highly specific for the diagnosis of Addison’s disease and are found in over 90% of individuals with this diagnosis (9). Infectious etiologies of primary AI include tuberculosis, fungal infections and HIV. Infiltration of the adrenal glands by metastatic cancers, primary adrenal lymphoma, sarcoidosis or amyloidosis can lead to primary AI, which can be apparent on imaging (10, 11, 12, 13). Bilateral adrenal hemorrhage can lead to the development of new-onset primary AI in pregnancy (14). The adrenal glands are each drained by a single adrenal vein, and venous congestion associated with pregnancy can be a risk factor for the development of adrenal hemorrhage and infarction. The hypercoagulable state associated with pregnancy is also a risk factor for the development of NHAI (15).

In one retrospective study, NHAI was present in 1.3% of abdominal MRIs obtained in pregnant women for the evaluation of abdominal or flank pain (16). MRI features of NHAI include unilateral adrenal enlargement, increased T2 signal intensity without T1 hyperintensity and retroperitoneal edema near the affected adrenal gland. CT findings of NHAI consist of enlargement and hypoenhancement of the adrenal gland (17). Urinary tract calculi should be considered in the differential diagnosis for acute flank pain in pregnancy. In one study, urinary calculi was diagnosed in 13 out of 20 pregnant women who underwent CT of the abdomen and pelvis for the evaluation of acute, refractory flank pain (18).

Unilateral NHAI typically does not lead to AI if the contralateral adrenal gland is healthy (19). The possible underlying pathophysiologic mechanism for the development of AI in our patient is that the healthy adrenal gland could not keep up with the demand for increased cortisol production in pregnancy. A literature review of all published cases reports of unilateral NHAI in pregnancy is provided in Table 2. The majority of cases of NHAI in pregnancy presented in the third trimester, with abdominal or flank pain and vomiting being the most common presenting symptoms. The right adrenal gland was more commonly affected. Only two previous cases of unilateral NHAI in pregnancy leading to AI were noted (20, 21). The left adrenal gland was involved in both cases, similar to our patient. Flank pain was the presenting clinic manifestation in both cases, but in neither of the two cases did the patient have hypotension or hypoglycemia. The first patient had spontaneous rupture of membranes during day 2 of admission and the AI was diagnosed post-partum based on a low random serum cortisol (20). In contrast to our patient, she did not develop adrenal crisis and only required oral hydrocortisone replacement. The second patient was diagnosed with AI based on a cosyntropin stimulation test that showed a mid-normal ACTH level and peak cortisol of 11 µg/dL (21). That patient stopped the hydrocortisone during pregnancy but went on to have a spontaneous vaginal delivery at 39 weeks and was given stress dose IV hydrocortisone during labor. Our patient is unique in having developed adrenal crisis as a result of the NHAI. Recovery of adrenal function occurred around 9 months post partum in the first patient, whereas the second case report did not provide information about the assessment of adrenal function following delivery. The longer duration required for recovery of adrenal function in our patient is likely related to having had a more extensive infarction.

Table 2.

Literature review of case reports of unilateral NHAI in pregnancy.

Number	Author (year of publication)	Age (years)	Gestational age (weeks)	Presenting symptoms	Imaging modality	Laterality	Adrenal function status	Pregnancy and neonatal outcome
1	Ornaghi et al. (2023) (19)	21	32	Right upper quadrant and flank pain, nausea and vomiting	CT with contrast	Right	Normal	Uncomplicated delivery, healthy neonate
2	Shah et al. (2022) (20)	25	32	Left flank pain, vomiting	CT with contrast	Left	Adrenal insufficiency diagnosed post-partum	Preterm spontaneous rupture of membranes, neonate had breathing difficulties but made a full recovery
3	Warda et al. (2021) (21)	24	30	Left upper quadrant abdominal pain radiating to the back	MRI without contrast	Left	Adrenal insufficiency diagnosed based on suboptimal response on cosyntropin stimulation test	Uncomplicated delivery, viable neonate
4	Jerbaka et al. (2021) (22)	36	36	Left abdominal pain	CT with contrast	Left	Normal	Oligohydramnios, healthy neonate
5	Padilla et al. (2021) (23)	25	Second trimester	Right upper quadrant abdominal pain, nausea	MRI without contrast	Right	Not reported	Not reported
6	Chasseloup et al. (2019) (24)	30	32	Right abdominal pain	CT with contrast	Right	Normal	Preterm labor, healthy neonate
7	Hynes et al. (2019) (25)	21	29	Right upper quadrant abdominal pain, nausea, vomiting	MRI without contrast and CT with contrast	Right	Not reported	Uncomplicated pregnancy, neonatal outcome not mentioned
8	Aljenaee et al. (2017) (26)	29	24	Right upper quadrant abdominal pain	CT with contrast	Right	Not reported	Uncomplicated delivery, neonatal outcome not mentioned
9	Reichman et al. (2016) (27)	28	38	Right upper quadrant and flank pain	MRI without contrast	Right	Normal	Not reported
10	Sormunen-Harju et al. (2016) (28)	31	38	Right upper abdominal pain	CT with contrast	Right	Normal	Uncomplicated delivery, healthy neonate
11	Green et al. (2013) (29)	25	28	Right upper quadrant abdominal pain, nausea, vomiting	CT with contrast	Right	Normal	Preterm labor, healthy neonate
12	Schmitt et al. (2010) (30)	29	36	Right upper quadrant and flank pain	CT with contrast	Right	Normal	Emergency cesarean section for uncontrolled pain, neonate with moderate respiratory distress

The cornerstone for the management of primary AI is adequate glucocorticoid and mineralocorticoid replacement. The Endocrine Society guidelines suggest using hydrocortisone over other formulations of glucocorticoids such as prednisone (31). The typical pre-pregnancy hydrocortisone dose is 15–25 mg per day in two to three divided doses. The hydrocortisone dose is typically increased by about 20–40% beginning from the 24th week onward to mimic the physiological increase in free cortisol during pregnancy. Our patient was given a total daily hydrocortisone dose of 30 mg to reflect the increased hydrocortisone dose used during pregnancy. Mineralocorticoid replacement is done using fludrocortisone. The typical dose of fludrocortisone is 0.05–0.1 µg per day. The dose is adjusted based on the blood pressure and serum potassium level. Progesterone exerts an anti-mineralocorticoid effect due to which the fludrocortisone dose may need to be increased in late pregnancy if the patient develops orthostatic hypotension and hyperkalemia.

Our patient presented with hypotension, intractable vomiting, hyponatremia and hypoglycemia which was consistent with a diagnosis of adrenal crisis. The subacute adrenal infarction coupled with a UTI likely triggered the adrenal crisis. The incidence of adrenal crisis in patients with preexisting AI in pregnancy is around 7% (32). The incidence of adrenal crisis in pregnancy without a history of prepregnancy AI has not been reported in the literature. The treatment of adrenal crisis in pregnancy consists of prompt administration of stress dose IV hydrocortisone 100 mg every 6–8 h (7). Normal saline is administered to correct the hypotension. Five percent dextrose is added to the IV fluids in case of hypoglycemia. Fludrocortisone is not administered during an adrenal crisis because high dose hydrocortisone itself exerts a mineralocorticoid effect. Stress dose IV hydrocortisone is also administered at the onset of the second stage of labor and the dose is repeated every 6–8 h until delivery.

Our patient’s AI persisted after delivery but later resolved at approximately 16 months post partum. The likely underlying mechanism behind the recovery of adrenal function is that the elevated ACTH level led to the hyperplasia of the contralateral adrenal gland which was then able to produce sufficient cortisol. Animal models have shown that unilateral adrenalectomy leads to increase in weight of the residual adrenal gland and that proliferation of cells occurs primarily in the zona fasciculata (33). Our case highlights that recovery of adrenal function is possible in the case of AI secondary to unilateral adrenal infarction in pregnancy.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the case study reported.

Funding

This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.

Patient consent

Written informed consent for publication of their clinical details and/or clinical images was obtained from the patient.

Author contribution statement

LM participated in the treatment of the patient and drafting the manuscript. DE, JM, PS and SS participated in literature review, reviewing and revising the manuscript. All authors have read and approved the final draft of the manuscript.