Abstract
KEY POINTS
Thyroid immune-related adverse events are the most common endocrine toxicity of immune checkpoint inhibitors (ICIs), with hypothyroidism and destructive thyroiditis being the most frequent clinical scenarios; for this reason, national oncologic guidelines provide monitoring parameters for thyroid function during therapy.
In patients with preexisting Graves’ disease, thyrotoxicosis during ICI therapy may more often reflect underlying autoimmune thyrotoxicosis rather than destructive thyroiditis; however, current evidence is insufficient to define the optimal role or timing of thyrotropin receptor antibody testing, including baseline assessment.
SUMMARY
Background
Immune checkpoint inhibitor (ICI) therapies are increasingly used in oncology to enhance antitumor immunity by blocking inhibitory signals such as anti–programmed cell death protein/ligand 1 (PD-1/PDL-1, and cytotoxic T-lymphocyte associated protein 4 (CTLA-4). 1 Although these therapies have substantially improved clinical outcomes in multiple malignancies, they have also been associated with a spectrum of immune-related adverse events (irAEs), including endocrine dysfunction.2,3 Thyroid dysfunction is among the most common endocrine irAEs, with both hypothyroidism and thyrotoxicosis from a destructive thyroiditis being the most common manifestations of this endocrine irAE.2,3 In contrast, endogenous hyperthyroidism from ICIs is rare, and risk factors for developing Graves’ disease following ICI therapy remain unclear. 4 The present study 5 aims to clarify the relationship between preexisting autoimmune thyroid disease and the development of hyperthyroidism among patients receiving ICI therapy.
Methods
This was a prospective case–control study of 19 individuals with preexisting Graves’ disease (GD) being treated with any ICI regimen at a university hospital in Japan between 2015 and 2024. The primary cohort group (GD-ICI) was compared with two control cohorts: Control 1, patients who had ICI therapy without preexisting Graves’ disease; and Control 2, patients with GD and without ICI therapy.
The Control 1 cohort was prospectively collected over the same study period and matched 1:5 with the GD-ICI cohort based on age, sex, baseline thyroid antibody status, and ICI type; exclusion criteria were prior ICI therapy, combination therapy with tyrosine kinase inhibitors, history of total or subtotal thyroidectomy, history of radioiodine therapy, and insufficient laboratory follow-up. The control 2 group was established retrospectively and consisted of outpatients with Graves’ disease who had not received ICI therapy and were either in remission or on low-dose thiamazole (≤10 mg/day), serving as a comparator to estimate the natural one-year exacerbation rate of Graves’ disease without ICI exposure; exclusion criteria were unconfirmed GD based on Japan Thyroid Association diagnostic criteria, 6 history of ICI therapy, thyroidectomy or radioiodine treatment, and insufficient laboratory follow-up.
Clinical and biochemical data were collected, including thyroid function testing and baseline thyroid antibody testing, with additional testing at onset of thyrotoxicosis in the GD-ICI and Control 1 groups. The reference standard for GD was considered thyrotropin receptor antibody (TRAb) positivity with thyrotoxicosis, with the secondary definitive standard being increased uptake on thyroid scintigraphy. Statistical analyses included descriptive statistics, univariate analyses, and logistic-regression modeling for contributing factors to hyperthyroidism.
Results
The study cohort comprised 19 individuals in the GD-ICI group, 95 in the propensity-matched Control 1 group and 269 in the Control 2 group. The GD-ICI cohort had a median age of 69 years (range, 60–70), and about half were female (47%); of the 19 patients, 13 had GD in remission and 6 were on low-dose thiamazole without overt thyroid dysfunction. Only one patient was positive for TRAbs.
Overall thyroid irAEs occurred at similar rates between patients with Graves’ disease and controls receiving ICIs (26.3% [5 of 19] vs. 29.5% [28 of 95]; 95% CI, 9.1–51.2 vs. 20.6–39.7); however, the type of thyroid irAE varied appreciably, with TRAb-positive being significantly more frequent in the GD-ICI group than in the Control 1 group (15.8% [3 of 19]; 95% CI, 3.4–39.6 vs. 0% [0 of 95]; 95% CI, 0–3.8; p = 0.004). Rates of destructive thyroiditis (10.5% vs. 15.8%, p = 0.734) and isolated hypothyroidism (0% vs. 13.7%, p = 0.121) were not significantly different between the GD-ICI and Control 1 groups. If comparing by type of thyrotoxicosis, 40% in the GD-ICI group had destructive thyroiditis as compared with 100% of the thyrotoxic control 1 cohort.
In contrast, the GD-ICI group had a significantly higher incidence of hyperthyroidism than the Control 2 group (15.8% vs. 3.7%, p = 0.046), and ICI exposure remained independently associated with hyperthyroidism after adjustment for sex (OR: 4.40; 95% CI, 1.06–18.3; p = 0.041).
Conclusions
This study highlights that ICIs can unmask or exacerbate autoimmune thyroid disease, which can change the etiology and management of thyrotoxicosis in individuals undergoing ICI therapy. The findings underscore the need for careful thyroid function monitoring before and during ICI therapy and suggest that preexisting autoimmune thyroid disease should be considered a risk factor for hyperthyroidism with ICI therapy.
COMMENTARY
The study by Suzuki et al. helps contextualize thyroid irAEs for individual patients undergoing ICI therapy and how underlying autoimmune thyroid conditions can impact the subtype of thyroid irAE. The observations from this study serve as an important reminder about the use of structured monitoring programs for thyroid irAEs as suggested by existing guidelines,2,3 and that preexisting autoimmune thyroid disease is a risk factor for immune-mediated ICI thyrotoxicosis, rather than more common destructive thyroiditis.
Thyroid irAEs are the most common endocrine irAE from ICI therapy, occurring in up to approximately 20% of patients. 2 A systematic review and meta-analysis of 101 studies encompassing nearly 20,000 patients 7 found the incidence of thyroid irAEs to be most common in inducing hypothyroidism (3.8–16.4%), followed by hyperthyroidism (1.4–10.4%) and thyroiditis (1.6–4.6%). Within these ranges, higher rates of thyroid irAEs are seen primarily with PD-1, PDL-1, and combination ICI therapies, and lower frequencies are seen with CTLA-4 therapies. Based on the frequency of these thyroid irAEs, National Comprehensive Cancer Network (NCCN) 2 guidelines, for example, recommend monitoring TSH and free T4 every 4 to 6 weeks while a patient is on ICI therapy, irrespective of thyroid antibody status. If thyrotoxicosis is detected, recommendations include initiation of beta-blockade and continued monitoring, with consideration of evaluation for Graves’ disease in 4–6 weeks if thyrotoxicosis persists.
The present study would argue that perhaps in patients with prior Graves’ disease, this evaluation can be expedited to the time of thyrotoxicosis since these are the patients found to be particularly at risk for immune-mediated hyperthyroidism. Checking TRAb status prior to ICI therapy could be considered; however, in this study, only 1 of the 19 GD-ICI patients had baseline positive TRAbs, and this cohort would be too small to make conclusions about the utility of obtaining baseline TRAb or thyroid-stimulating immunoglobulin levels prior to initiation of ICI in patients with preexisting disease. The timing of TRAb testing is clinically relevant, as transient TRAb positivity has been reported in a subset of patients with destructive thyroiditis, 8 potentially confounding the distinction from Graves’ disease. This overlap may result in additional and unnecessary imaging and follow-up testing for patients who already have a substantial burden of care, highlighting the need for studies to define the optimal timing and clinical utility of TRAb measurement in this setting.
In our clinical practice, we find that our oncology protocols for monitoring thyroid irAEs generally align with recommendations from the NCCN. The threshold for requesting endocrinology input varies and occurs more commonly for thyrotoxic patients rather than for those with primary hypothyroidism. Since most patients with thyrotoxicosis will have destructive thyroiditis, it may be most practical for oncologists to complete short-interval lab follow-up, with subsequent referral for progressive thyrotoxicosis. We would expect that most patients with stable Graves’ disease or GD in remission would already be following up with an endocrinologist, but early referral in this population would otherwise be reasonable.
As ICIs continue to reshape cancer treatment, studies such as this one provide critical guidance for integrating endocrinology and oncology care and for individualizing care for patients with preexisting autoimmune thyroid disease.