Secondary Adrenal Insufficiency in a Patient with Metastatic Melanoma Treated with Nivolumab

Nivolumab is a fully humanized IgG4 programmed death 1 (PD-1) immune checkpoint inhibitor antibody with a good safety profile, higher response rate, and durable efficacy for malignant melanoma [1]. Objective response rate was 35% and the 3-year survival rate was 43.5% in previously untreated Japanese patients with stage IV or recurrent malignant melanoma [2]. Among immune-related adverse events due to immune checkpoint blockades, endocrine side reactions were reported in cases of 7–24% [3-5].

Of endocrine-related complications, “hypophysitis” due to anti-PD-1 antibodies has been reported. “Hypophysitis” refers to the inflammation of the pituitary gland that can result in primary dysfunction of the gland, particularly the anterior pituitary, leading to deficiency in one or more hormones secreted by the anterior pituitary gland [6]. Secondary adrenal insufficiency is one of the clinical manifestations due to “hypophysitis” [6]. The incidence of “hypophysitis” with ipilimumab was reported to be 4–11% [4, 5], and that with nivolumab was 0.5–2% [3, 4], suggesting that there is a difference. Here, we will present a patient who was treated with nivolumab and suffered from secondary adrenal insufficiency as endocrine side effects.

An 83-year-old man with acral lentiginous melanoma on the right sole visited our hospital in March 2017. The primary site was excised, and inguinal lymph node dissection was performed in June 2017 in the department of dermatology in another hospital. He was followed up at our hospital from November 2017. During the follow-up, one subcutaneous nodule (7 × 7 mm) on the right knee that appeared in July 2018 was resected. Pathologically, it was diagnosed as in-transit metastasis of malignant melanoma. Genetic analysis of dissected specimen revealed no BRAF V600 mutation.

We started nivolumab 3 mg/kg every 2 weeks from September 2018 as adjuvant therapy. In April 2019, after 17 cycles of nivolumab (33 weeks after the initiation of nivolumab), he suddenly developed severe anorexia, nausea, and vomiting. The patient visited the emergency department due to continued anorexia and nausea. He was admitted to our hospital in May 2019 (day X).

Since the abdominal ultrasound and upper gastrointestinal endoscopy were normal, we suspected a non-gastrointestinal disease. We suspected adrenal insufficiency due to nivolumab and consulted an endocrinologist. The levels of cortisol (9.16 μg/dL; normal range, 6.24–18 μg/dL) and adrenocorticotropic hormone (ACTH) (15.7 pg/mL; normal range, 7.2–63.3 pg/mL) were within normal ranges (Table 1). As we did not get a definitive diagnosis, corticotrophin-releasing hormone loading test was performed (day X+6) (Fig. 1).

Cortisol levels before loading test were below the normal range and loading response was impaired (Fig. 1). Though ACTH responded, both ACTH levels before loading test and peak levels of ACTH were low (Fig. 1). Therefore, we thought that the pituitary gland was impaired partially. We diagnosed partial isolated ACTH deficiency.

We took a brain MRI, and no enlargement of the pituitary gland was noted (day X+8) (Fig. 2). Hydrocortisone of 30 mg per day was started from day X+8. On the following day, nausea disappeared, and he was able to eat. We decreased to 20 mg from day X+19, and 15 mg from day X+26, and he was discharged. At this time, radiation therapy was performed for in-transit metastases on the front of the right side of the femoral region for 14 days from day X+12 to X+25.

After discharge, in-transit metastases on the right lower leg developed and we resected them twice. Radiation therapy was performed for in-transit metastases on the back of the right side of the femoral region in July 2019. Nivolumab was resumed after 11 weeks from the absence of nivolumab administration. However, in August 2019, liver, pancreatic, and right peritoneal metastases were newly pointed out by CT scan. We decided to change from nivolumab monotherapy to combination therapy of nivolumab plus ipilimumab every 3 weeks from September to October 2019.

For diarrhoea due to nivolumab plus ipilimumab, the patient required hospitalization for approximately 1 month. Because of diarrhoea, we stopped nivolumab plus ipilimumab. Response to the combination treatment of nivolumab plus ipilimumab was poor and metastatic lesions had grown. The patient entered the hospital for palliative care and died in August 2020.

In our country, nivolumab has been used for patients with unresectable malignant melanoma since September 2014 and patients with adjuvant therapy since August 2018. It is well known that nivolumab induces side effects different from conventional cytotoxic agents. Among immune-related adverse events, endocrine side reactions were reported in cases of 7–24% [3-5]. The incidence of “hypophysitis” with ipilimumab was reported to be 4–11% [4, 5], and that with nivolumab was 0.5–2% [3, 4], suggesting that there is a difference.

Nine studies including Japanese patients with secondary adrenal insufficiency due to nivolumab reported several features (Table 2) [7-14]. There were more male than female patients who suffered from this complication (7 vs. 2). The most common symptoms were fatigue and anorexia, seen in 5 of 9 cases. Adrenal insufficiency with hypothyroidism was seen in 3 of 9 cases [7, 12, 14].

Of interest was that there could be differences in the frequency of endocrinological side effects, especially secondary adrenal insufficiency due to anti-PD-1 antibody (nivolumab or pembrolizumab) and anti-CTLA-4 antibody (ipilimumab) [15].

Our main findings were as follows. First, it is difficult to diagnose secondary adrenal insufficiency due to immune checkpoint blockades because symptoms of “hypophysitis” are typically nonspecific [6]. In emergency department, the diagnosis was not established although anorexia and nausea had persisted before hospital admission.

Second, initial cortisol and ACTH levels were within the normal range when symptoms appeared. This had made our correct diagnosis delayed slightly and delayed start of treatment. With the help of endocrinologists, corticotropin-releasing hormone loading test after admission established secondary adrenal insufficiency in this case (Fig. 1).

Third, our patient developed adrenal insufficiency after 17 cycles of treatment from the initiation of nivolumab (33 weeks after the initiation of nivolumab). In a previous study, a difference on the timing of onset from administration was revealed to be longer in anti-PD-1 antibody group than in ipilimumab group (median time is 25.8 weeks vs. 9.3 weeks) [15]. In other reports, onset time from the start of treatment was reported between 3 and 15 months [7-14]. Adrenal insufficiency in our case developed late from the start of nivolumab.

Fourth, treatment with hydrocortisone improved immediately clinical symptoms. Since it is said that the adrenal functional impairment is irreversible, administration of hydrocortisone is needed for life.

Fifth, the pituitary gland was not enlarged on MRI at the time of diagnosis (Fig. 2). Faje et al. [15] reported that among 22 patients with “hypophysitis” who received anti-PD-1 antibodies, 13 of 18 patients who had MRI did not show enlarged pituitary gland, while among 64 patients who received ipilimumab, 60 of 61 patients who had MRI showed enlarged pituitary gland. Our case reflects entity of previous reports.

In conclusion, since adrenal insufficiency develops suddenly, and characteristics of clinical symptoms are nonspecific, we should alert to monitor symptoms and signs. If needed, immediate consultation to subspecialties is crucial.

We would like to thank the patient’s family members, nurses, and staff members of the department of dermatology in Asahi General Hospital.

Study approval statement was not required for this study in accordance with local/national guidelines. Written informed consent was obtained from the parent/legal guardian of the patient for publication of the details of their medical case and any accompanying images.

The authors have no conflicts of interest to declare.

This study did not receive any funding.

Wataru Omata: investigation, writing – original draft. Satoko Nakamura, Chie Urasaki, Hiromichi Morita, and Hiroki Funaishi: writing – review and editing. Kazuki Kobayashi and Hisashi Koide: writing – review and editing, and supervision. Arata Tsutsumida and Hiroyuki Matsue: supervision.

All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.