Simultaneous immune checkpoint blockade-associated diabetes mellitus and thyrotoxicosis

Abstract

Immune checkpoint inhibitors (ICIs) have become an increasingly prominent therapeutic option in the management of a variety of cancers. As their clinical use expands, so too should our understanding of immune-related adverse events (irAEs) associated with these agents. By enhancing immune activation, ICIs can cause off-target inflammation in various tissues, including endocrine. We present a case illustrating a rare but life-threatening manifestation of ICI toxicity: simultaneous severe diabetes mellitus and thyrotoxicosis after just a single dose of ICI. This case highlights the importance of prompt recognition and timely intervention of acute, life-threatening endocrine irAEs. We discuss the pathophysiology, management, and clinical course of ICI-induced diabetes mellitus and thyroid disease, with brief discussion of endocrine irAEs as a surrogate of therapeutic efficacy.

Introduction

Recent emergence of innovative therapies in oncology, immune checkpoint inhibitors (ICIs), has led to improved patient outcomes. Tumor cells evade immune detection by exploiting immune checkpoints, cytotoxic T lymphocyte-associated protein 4 (CTLA-4), programmed cell death protein 1 (PD-1), programmed cell death protein ligand, and others. ICIs are monoclonal antibodies targeting immune checkpoints, restoring cell-mediated immunity and enabling cancer cell destruction [1]. First successful in treating malignant melanoma and lung cancer, ICIs are now indicated in 55.5% of cancer patients with response in 20.1% of patients in 2023 [1, 2].

Despite their success, ICIs can cause off-target inflammation and autoimmunity against normal tissues, manifesting as immune-related adverse events (irAEs) including colitis, hepatitis, arthritis, dermatitis, pneumonitis, and endocrine toxicities such as hypophysitis, diabetes mellitus, and thyroid disease [1, 3-5].

We present the case of a previously healthy female presenting with simultaneous diabetes mellitus and thyrotoxicosis in setting of ICI use. We explore implications of these novel therapies on severe endocrine irAEs while stressing importance of early detection and prompt treatment to mitigate severe complications. We also address implications that ICI-induced irAEs may be a marker of therapeutic efficacy in cancer.

Case presentation

A 47-year-old female was transported by first responders to the emergency department with coma. Medical history included treated left breast ductal carcinoma in remission and stage III (rcT1A, cN1b, cM0) malignant melanoma (BRAF V600E+) from an upper back primary diagnosed and excised 5 years prior, with recent recurrence for which she was initiated on ipilimumab/nivolumab. Fourteen days following her first infusion of ipilimumab/nivolumab, she developed confusion and episodic nonbloody, nonbilious emesis. In addition, she had polydipsia and polyuria. These symptoms progressed until the next day, when her family found her unconscious in her home. She was promptly brought to the hospital for further evaluation.

Diagnostic assessment

In the emergency department, she was hypotensive and tachycardic, and her Glascow Coma Scale score was 9. Head imaging was negative for acute intracranial abnormalities. Initial laboratory work was notable for leukocytosis, elevated creatinine, and multiple electrolyte derangements (Table 1). Notably, initial glucose level was >1500 mg/dL (SI: >83.3 mmol/L) (reference range, 70-140 mg/dL [SI: 3.9-7.8 mmol/L]), lactate 4.7 mmol/L (0.5-2.2 mmol/L), and β-hydroxybutyrate 10.2 mmol/L (<0.4 mmol/L), concerning for diabetic ketoacidosis. She was initiated on intravenous crystalloid fluids, followed by regular insulin. Norepinephrine was required for hypotension (82/52 mm Hg) and empiric antibiotics were started based on initial concern for septic shock. Initial venous blood gas was pH 7.05 (reference range, 7.31-7.41), bicarbonate of 35 mmol/L (reference range, 22-28 mmol/L), and partial pressure of oxygen 57 mm Hg (SI: 7.6 kPa) (reference range, 30-50 mm Hg [SI: 4.0-6.7 kPa]). These values improved the following day to a pH of 7.40, bicarbonate 26 mmol/L, and partial pressure of oxygen of 42 mm Hg (SI: 5.6 kPa). She was admitted to the medical intensive care unit for ongoing management.

Table 1.

Initial laboratory values in emergency department

Lab	Value (SI units)	Reference range (SI units)
Hemoglobin	10.5 g/dL (105 g/L)	11.6-15.0 g/dL (116-150 g/L)
Leukocytes	34.2 × 103/µL (34.2 × 109/L)	3.4-9.6 × 103/µL (3.4-9.6 × 109/L)
Platelet count	374 × 103/µL (374 × 109/L)	157-371 × 103/µL (157-371 × 109/L)
Sodium	119 mEq/L (119 mmol/L)	135-145 mEq/L (135-145 mmol/L)
Potassium	7.1 mEq/L (7.1 mmol/L)	3.6-5.2 mEq/L (3.6-5.2 mmol/L)
Anion gap	38 mEq/L (38 mmol/L)	7-15 mEq/L (7-15 mmol/L)
Blood urea nitrogen	77 mg/dL (27.5 mmol/L)	6-21 mg/dL (2.1-7.5 mmol/L)
Creatinine	2.92 mg/dL (258 µmol/L)	0.59-1.04 mg/dL (52-92 µmol/L)
Glucose	>1500 mg/dL (>83.3 mmol/L)	70-140 mg/dL (3.9-7.8 mmol/L)
Venous pH	7.05	7.32-7.43
Lactate	4.7 mmol/L	0.5-2.2 mmol/L
β-hydroxybutyrate	10.2 mmol/L	<0.4 mmol/L

Further workup in the medical intensive care unit was performed for infectious, neurologic, and oncologic etiologies. Pertinent findings included lipase of 1385 U/L (reference range, 13-60 U/L) with radiographic evidence of acute pancreatitis. Additional laboratory results revealed low TSH at 0.08 mIU/L (reference range, 0.3-4.2 mIU/L) and <0.01 mIU/L on repeat testing, free T4 >7.7 ng/dL (SI: >99.1 pmol/L) (reference range, 0.9-1.7 ng/dL [SI: 11.6-21.9 pmol/L]), free T3 15.9 pg/mL (SI: 24.4 pmol/L) (reference range, 2.0-4.4 pg/mL [SI: 3.1-6.8 pmol/L]), negative thyrotropin receptor antibody, hemoglobin A1c 6.0% (reference range, 4.0-5.6%), undetectable C-peptide levels with simultaneous blood glucose level 250 mg/dL (SI: 13.9 mmol/L), and elevated glutamic acid decarboxylase 65 antibodies at 2.73 nmol/L (reference range, ≤0.02 nmol/L). Other antibodies associated with diabetes mellitus were negative: insulinoma-associated antigen 2 antibody, insulin, and zinc transporter 8 antibody antibodies. Ultimately, she was diagnosed with new ICI-induced diabetes mellitus (ICIDM) presenting with diabetic ketoacidosis (DKA) and simultaneous thyrotoxicosis.

Treatment

Her encephalopathy and DKA gradually resolved with insulin infusion and normalization of blood glucose levels and acid-base status. She was weaned off blood pressure support. Shortly thereafter, she was transitioned to a basal-bolus insulin regimen for ongoing glycemic control and treated with propranolol for symptomatic palpitations.

ICIDM and thyrotoxicosis were determined to be secondary to recent ICI administration. She was eventually stable for discharge to home, with outpatient endocrinology and oncology follow-up.

Outcome and follow-up

For ICIDM, she was followed closely by endocrinology and remained on basal-bolus insulin. Notably, at 1 month, her hemoglobin A1c rose to 8.3%, compared to 6.0% at the time of her hospitalization. Given simultaneous hyperglycemia and undetectable C-peptide levels during hospitalization and on repeat 3 weeks later (<0.1 ng/mL (SI: <0.033 nmol/L) (reference range, 1.1-4.4 ng/mL [SI: 0.37-1.46 nmol/L]), it was confirmed she had fulminant islet β-cell destruction, attributed to ICI therapy. Follow-up thyroid laboratory values showed progression of thyroiditis from the hyperthyroid phase (TSH <0.01 mIU/L (reference range, 0.3-4.2 mIU/L), free T4 >7.7 ng/dL (SI: >99.1 pmol/L) (reference range, 0.9-1.7 ng/dL [SI: 11.6-21.9 pmol/L]), free T3 15.9 pg/mL (SI: 24.4 pmol/L) (reference range, 2.0-4.4 pg/mL [SI: 3.1-6.9 pmol/L]) to hypothyroidism (TSH 34.7 mIU/L, free T4 <0.3 ng/dL (SI: <3.9 pmol/L)) after approximately 2 months. She was subsequently started on thyroid hormone replacement.

For melanoma, she followed with medical oncology every 3 months and had significant response to a single dose of ipilimumab/nivolumab. Whole-body, fluorine-18 fluorodeoxyglucose positron emission tomography-computed tomography scan 3 months later revealed marked improvement in metastatic lymphadenopathy and resolution of a likely metastatic pulmonary nodule (Fig. 1). She later underwent malignant inguinal and iliac lymph node dissection and resumed ICI therapy perioperatively with no further adverse events. At last follow-up, she had no evidence of disease.

Figure 1.

Positron emission tomography computed tomography comparison between (A) initial scan revealing right inguinal lymphadenopathy (red arrow) corresponding to known malignancy and (B) follow-up scan 3 months after ICI initiation revealing resolution of lymphadenopathy (red arrow).

Discussion

ICIs have become a mainstay of treatment in malignant melanoma, resulting in increasing incidence of irAEs [6-8]. We present the case of a young female who received a single dose of ICI therapy, resulting in life-threatening DKA and thyrotoxicosis 14 days later.

ICIDM was described previously in the literature with incidence 0.2% to 1.4% [3, 9, 10]. It is caused by severe β-islet cell inflammation and thought to closely resemble pathophysiology of type 1 diabetes mellitus, as both are mediated by PD-1/programmed cell death protein ligand axis inhibition and less by the CTLA-4 pathway [9, 10]. Median time to diagnosis of ICIDM is about 15 weeks (8-26.5 weeks); the majority of patients are diagnosed after an average 4.5 treatment cycles, contrasted from our patient who presented after just 1 dose [9]. ICIDM typically presents with acute, severe hyperglycemia and DKA with low hemoglobin A1c and C-peptide. Based on a meta-analysis, Wu et al proposed the following diagnostic criteria for ICIDM: 1. Previous ICI exposure, 2. new-onset hyperglycemia (blood glucose level at least 11.1 mmol/L and/or hemoglobin A1C at least 6.5), and 3. insulin deficiency, as measured by depleted C-peptide <0.4 or DKA [10].

There are no gold standard markers to stratify patients at risk for developing ICIDM. Typically, patients with positive antibodies are diagnosed earlier in disease course, although our patient had a single positive antibody and presented after 1 dose [5, 10]. In a small cohort analysis, those with ICIDM had pancreatic atrophy similar to that seen in patients with ICI-induced pancreatitis [5]. They had significantly lower measured pancreatic volume compared to ICI-treated controls [5]. This suggests patients with lower pancreatic volume and reserve may have predisposition to developing ICIDM; however, further studies with higher sample size will be needed to evaluate any direct correlation. HLA DR4 has been reported in 76% of patients with ICIDM in a study of 27 patients [11]. Ruiz-Esteves and colleagues developed a type 1 diabetes polygenic risk score to identify individuals at risk for ICIDM in a retrospective study [12]. These results are promising for future prospective efforts developing a risk stratification model. Regardless given increasing incidence of ICIDM, it is recommended to routinely monitor glucose levels before each cycle of ICI therapy for early detection [5, 9].

Interestingly, there is also evidence of possible ICI-associated pancreatitis, or exocrine injury, in addition to ICIDM. This is suggested by the elevated lipase and imaging findings our patient had on presentation. Although less common than ICIDM, it is still a recognized adverse event, mediated by T-cell inflammation of pancreatic tissue following loss of peripheral immune tolerance [13]. In certain cases, both endocrine and exocrine pancreatic toxicities may coexist, suggesting a spectrum of immune-mediated pancreatic injury rather than a single unified pathogenic process.

Thyroid dysfunction induced by ICIs can manifest as either hypothyroidism or thyroiditis, occurring after one dose or up to 3 years following ICI initiation [1]. This toxicity is most commonly associated with combination anti-PD-1 and anti-CTLA-4 agents, rather than monotherapy [1]. ICI-induced thyroiditis results from cytotoxic CD8+ and helper T-cell activation, followed by acute inflammation with release of cytokines IL-2 and IFN-γ [1]. This process leads to T-cell infiltration and thyroid gland destruction [1]. TSH-receptor antibodies are typically absent in this cell-mediated “destructive thyroiditis,” which contrasts from the antibody-mediated thyroiditis seen in Graves disease, although these mechanistically distinct conditions can both lead to thyrotoxicosis. The majority of destructive ICI-associated cases present as thyroiditis and if symptomatic at presentation, will present with thyrotoxicosis as observed in our patient [1]. Most cases of overt or subclinical thyrotoxicosis subsequently resolve to a euthyroid state or progress to hypothyroidism within weeks to months. Our patient's thyroid function has followed the same course, and she will likely require long-term thyroid hormone replacement.

Evaluation of TSH and free T4 before ICI initiation is recommended to rule out preexisting thyroid disorders, with routine monitoring every 4 to 6 weeks during treatment and until 4 to 6 weeks after treatment discontinuation [1]. Guidelines for managing irAEs have been published by consensus committees, including the American Society of Clinical Oncology [14]. If patients have grade 1 thyrotoxicosis based on the Common Terminology Criteria for Adverse Effects grading system, ICI therapy can proceed with regular laboratory result monitoring [14]. For grade 2-4 adverse events, therapy should be discontinued temporarily with initiation of a β blocker, as was the case with our patient [14]. For patients presenting with thyrotoxicosis, repeat testing should occur around 6 weeks after diagnosis to monitor for progression to subclinical hypothyroidism.

Graves disease should be suspected when there is elevated thyroid hormone, symptoms of hyperthyroidism, persistence of thyrotoxic symptoms for 6 weeks, or presence of characteristic signs such as a goiter or orbitopathy [1]. Diagnosis can be confirmed with positive thyroid-stimulating immunoglobulin and/or thyrotropin receptor antibodies, along with Doppler ultrasound and scintigraphy if suspicion is high [1]. Notably, our patient lacked these antibodies, an occurrence historically seen in roughly 3% to 5% of patients with Graves disease and 50% of patients with thyrotoxicosis from any cause [15, 16].

Our patient had significant tumor response to ICI therapy, and some studies propose irAEs from ICIs may lead to improved cancer efficacy [4, 17, 18]. A study of 91 patients with cancer who received atezolizumab and avelumab between 2016 and 2018 showed those who developed ICI-induced thyrotoxicosis had improved overall survival [13]. However, a recently performed large retrospective study of 62 randomized trials and a total of 42 247 patients with solid tumors who received ICI therapy showed no significant association between the development of irAEs and therapeutic efficacy in cancer [14]. For the melanoma subgroup, compared to a non-ICI based control arm, there was a medium-strength positive correlation between endocrine adverse events of any grade and overall survival; however, this was not statistically significant [19]. When considered together, there is not strong evidence that presence of IrAE are necessarily a surrogate of treatment response or serve as a positive prognostic indicator [19].

Although successful in improving patient survival, ICIs are associated with irAEs that have been increasing in incidence. In fact, because of high patient volumes at our comprehensive cancer center, our institution offers an ICI clinic that specializes in management of ICI-induced irAEs and is available to consult on acute management of irAEs in the inpatient setting. Prompt recognition and multidisciplinary treatment can lead to better outcomes.

Learning points

• Although ICIs have revolutionized cancer outcomes, they are associated with life-threatening irAEs affecting multiple endocrine organs.

• Close education and monitoring for endocrine irAEs is recommended before initiation, during, and even after treatment with ICIs.

• There are no validated markers to risk stratify individuals predisposed to ICIDM and thyroiditis; this is a topic that may be explored in future studies.

• Patient education and prompt recognition and treatment of endocrine irAEs may lead to favorable outcomes.

• There is no strong evidence that presence of irAEs is a surrogate of cancer response to ICIs.

Contributors

All authors made individual contributions to authorship. M.L. and M.F. were involved in conception and design of the work and acquisition, analysis, and interpretation of data. Y.C.K. was involved in management of this patient and supervision of drafting and final approval of this manuscript submission.

Funding

No public or commercial funding.

Disclosures

None declared.

Informed patient consent for publication

Signed informed consent obtained directly from the patient.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.