GRAPHICAL ABSTRACT

Highlights

• ICI-T1DM occurred in 0.46% (32/6,956) of ICI-treated patients.

• Pre-existing diabetes doubled, thyroid dysfunction quadrupled the risk.

• Combination PD-1/PD-L1 + CTLA-4 and ≥12-week ICI raised risk 3–5-fold.

• ICI-T1DM was associated with a 78% reduction in all-cause mortality.

• Early glucose monitoring is recommended for high-risk ICI recipients.

INTRODUCTION

The advent of immune checkpoint inhibitors (ICIs) has revolutionized the treatment of various malignancies, particularly metastatic solid tumors. ICIs target molecules such as programmed death-1 (PD-1), programmed death-ligand 1 (PDL1), and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), enhancing immune response against cancer cells. Although pivotal in cancer therapy, these agents can cause unintended autoimmune endocrine disorders as immune-related adverse events (irAEs), including hypothyroidism, hyperthyroidism, hypophysitis, primary adrenal insufficiency, and type 1 diabetes mellitus (T1DM) [1].

ICI-induced T1DM (ICI-T1DM) is characterized by the rapid destruction of pancreatic β-cells, potentially within 5 days of starting ICI therapy or even several months after discontinuation [2-4]. Managing ICI-T1DM is challenging due to its lifelong insulin requirement profoundly impacting patient care and quality of life. Although less frequent, ICI-T1DM occurs in approximately 0.2% to 1.9% patients and poses significant life-threatening risks [2,5,6], leading to severe hyperglycemic events such as diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state (HHS).

While overall mortality rate in patients with DKA is <1%, it exceeds 5% in older or critically ill patients [7]. Mortality rate in patients with HHS is higher, ranging from 5% to 20%, worsening with coma, hypotension, or severe comorbidities such as cancer [7]. Considering that most ICI recipients are older adults with metastatic cancer, effective strategies for managing acute hyperglycemia in ICI-T1DM are essential.

ICI-T1DM has a distinct pathophysiology compared with classic T1DM [8]; however, prior studies mainly focused on clinical and lab data associated with classic T1DM, such as human leukocyte antigen (HLA) genotypes and autoantibodies, which are often unavailable at ICI initiation [2,9,10]. To address this gap, some studies have explored clinical risk factors for ICI-T1DM, but their findings require further validation. Research on how ICI-T1DM impacts survival in ICI-treated patients also remains limited.

This study analyzed data from one of Korea’s largest cancer databases to assess ICI-T1DM incidence, clinical characteristics, and its survival impact on a large cohort. We identified the risk and survival factors that can inform pre-treatment evaluations aimed at enhancing screening protocols, minimizing side effects, and maximizing ICIs’ therapeutic benefits.

METHODS

Ethics statement

This study was approved by the Institutional Review Board of Asan Medical Center (approval no. 2023-0434), and this study was conducted in accordance with the Declaration of Helsinki. The need for written informed consent was waived owing to the retrospective nature of this study.

Database and data extraction

This study included 6,956 patients on ICI therapy at a tertiary care center between January 1, 2017, and February 28, 2023. The treatments involved anti–PD-1 inhibitors (nivolumab, pembrolizumab), PD-L1 inhibitors (atezolizumab, durvalumab, avelumab), and anti–CTLA-4 inhibitors (ipilimumab). Patients received either anti–PD-1/PD-L1 inhibitors alone or in combination with anti–CTLA-4 inhibitors, administered concurrently or sequentially. No patients received anti–CTLA-4 inhibitors alone.

Data on patient characteristics included age, sex, tumor types, ICI types, prior diabetes status, glucocorticoid (GC) use, and body mass index (BMI). Age was recorded as a continuous variable and categorized as <65 or ≥65 years. BMI was calculated as weight in kilograms divided by the square of height in meters (kg/m²) and categorized as <18.5 (underweight), 18.5–22.9 (normal), 23–24.9 (overweight), and ≥25 (obese).

Tumor types comprised lung, hepatocellular, gastric, renal cell, urothelial, head and neck, melanoma, breast, colorectal cancers, and others. Remaining cancers were grouped as ‘other,’ and unidentified primary tumors were grouped as ‘malignancy of unknown origin.’ ICIs were classified as ‘anti–PD-1/PD-L1’ or ‘both anti–PD-1/PD-L1 and anti–CTLA-4.’ Prior diabetes was determined via medical records, anti-diabetic medication use, or pre-treatment glycosylated hemoglobin (HbA1c) ≥6.5% and fasting glucose ≥126 mg/dL. ICI treatment duration was categorized as <12 or ≥12 weeks, calculated from the first to the last ICI administration.

Thyroid dysfunction was defined as hypothyroidism (thyroid-stimulating hormone [TSH] ≥6.8 μIU/mL or use of levothyroxine therapy) or hyperthyroidism (TSH <0.6 μIU/mL or use of anti-thyroid medication such as methimazole or propylthiouracil). Thyroid autoantibodies assays were considered positive if the levels exceeded the laboratory’s normal reference range: anti-thyroglobulin antibody (anti-Tg Ab) and anti-thyroid peroxidase antibody (anti-TPO Ab) >60 U/mL, and thyrotropin-binding inhibitory immunoglobulin (TBII) >1.5 IU/L. Adrenal dysfunction was assessed based on serum cortisol and aldosterone-renin ratio (ARR) levels. Adrenal insufficiency was defined as a baseline cortisol <3.0 μg/dL or peak cortisol <18 μg/dL during an adrenocorticotropic hormone stimulation test. Hypercortisolism was suspected in cases where an overnight dexamethasone suppression test showed cortisol ≥1.8 μg/dL, indicating insufficient suppression. Aldosterone excess was defined as ARR ≥30 and aldosterone ≥15 ng/dL.

GC administration was defined as administration before or during the first 90 days of ICI therapy, including prednisolone, methylprednisolone, hydrocortisone, and dexamethasone. This 90-day timeframe is commonly used to assess the immunosuppressive effects of GCs, which can persist and influence the ICI-induced immune response [2,11].

No clear international criteria exist for ICI-T1DM classification. In this study, ICI-T1DM was defined as no prior insulin requirement but persistent insulin use post-ICI therapy with a C-peptide level <1.0 ng/mL. Patients were initially included if they had a diabetes mellitus (DM) diagnosis (newly diagnosed or pre-existing), persistent insulin dependency post-ICI therapy, and a post-ICI C-peptide level <1.0 ng/mL. To ensure accurate classification, exclusion criteria were applied. Patients with pre-existing insulin dependency or a pre-ICI C-peptide level <1.0 ng/mL were excluded, as were those whose C-peptide levels dropped below 1.0 ng/mL post-ICI therapy without requiring persistent insulin therapy. Additionally, we excluded patients with confounding diagnoses such as acute pancreatitis that independently explain β-cell dysfunction. Finally, patients whose C-peptide levels recovered to ≥1.0 ng/mL during follow-up were also excluded. The stepwise inclusion and exclusion process is detailed in Fig. 1.

The ICI-T1DM onset date was set as the earliest date possible when these criteria were satisfied. If a patient experienced DKA or HHS before meeting these criteria post-ICI therapy, the onset date was set as the date of the DKA/HHS diagnosis. DKA was diagnosed based on a blood glucose level >250 mg/dL, elevated blood ketones (increased β-hydroxybutyrate), and acidosis (pH <7.3, HCO3− <15 mEq/L). HHS was diagnosed based on a blood glucose level >600 mg/dL, significant hyperosmolarity (effective plasma osmolality >320 mOsm/kg), and severe dehydration [12,13]. In this study, ‘both’ refers to cases where DKA and HHS occur simultaneously in a single patient, meeting the diagnostic criteria for both conditions. Although pathophysiologically distinct, severe hyperglycemic crises, often triggered by dehydration or stress, may show features of both [7]. The term ‘both’ describes a single event with overlapping features, not repeated occurrences of either DKA or HHS.

For survival analysis, data of the deceased patients were extracted from in-hospital deaths and the national cancer registry. For surviving patients, the observation endpoint was set as the last documented visit before May 2024. Mortality data were obtained from the hospital’s electronic medical record system and Korea Central Cancer Registry, ensuring comprehensive inclusion of both in-hospital and external deaths.

Statistical analysis

Categorical variables were summarized as frequencies, and continuous variables were summarized as medians. The normality of continuous variables was assessed, and non-parametric tests were used owing to the small number of patients with ICI-T1DM. Statistical significance was defined as P<0.05.

Univariate logistic regression was used to evaluate the associations between ICI-T1DM and potential risk factors (age, sex, tumor types, ICI types, ICI duration, pre-existing diabetes, GC administration, adrenal dysfunction, thyroid dysfunction, and BMI). Variables showing significant associations underwent multivariable logistic regression to adjust for confounders. The results were reported as hazard ratios (HRs) with 95% confidence intervals (CIs).

Survival analysis was conducted using Kaplan–Meier curves to visualize survival probabilities over time, and Cox proportional hazards models were used to assess the effects of ICIT1DM and other variables on mortality. To minimize confounding, patients with multiple tumor types were excluded from the analysis. Data linkage was performed using unique patient identifiers.

All statistical analyses were conducted using SPSS version 25.0 (IBM, Armonk, NY, USA), and Kaplan–Meier curves were created in R version 4.3.0 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Pre-existing diabetes mellitus pre-ICI therapy

Of the 6,956 patients who received ICI therapy, 1,488 had pre-existing diabetes based on a history of diabetes medications (oral hypoglycemic agents, insulin, or glucagon-like peptide-1 receptor agonists). This figure excludes 275 non-diabetic patients who temporarily used insulin for other reasons. Additionally, 192 patients were newly identified as diabetic after reviewing medical records, and six satisfied the fasting glucose (≥126 mg/dL) and HbA1c (≥6.5%) levels criteria. In total, 1,686 patients had pre-existing diabetes prior to ICI therapy. None required multiple daily insulin injections for classic T1DM or pancreatectomy-related diabetes (Table 1).

ICI-T1DM incidence and presentation

From a cohort of 6,956 patients who received ICI therapy between January 1, 2017, and February 28, 2023, we identified 95 patients with a diagnosis of DM (new or pre-existing), persistent insulin dependency following ICI therapy, and post-ICI Cpeptide levels <1.0 ng/mL. To ensure accurate classification of ICI-T1DM, we applied specific exclusion criteria. Forty-five patients were excluded due to pre-existing insulin dependency or pre-ICI C-peptide levels <1.0 ng/mL. Additionally, 15 patients with transient hyperglycemia who did not require sustained insulin therapy and one patient with concurrent acute pancreatitis at the time of diagnosis were excluded. Finally, two patients whose initial C-peptide levels <1.0 ng/mL normalized to ≥1.0 ng/mL during follow-up were also excluded. After these exclusions, 32 patients were classified as having ICI-T1DM, representing 0.46% of the cohort. The median period, defined from time of ICI initiation to ICI-T1DM classification, was 41 weeks (range, 9 to 1,359 days). Among them, 22 patients (62.5%) had acute hyperglycemic complications, such as DKA (three patients, 9.4%), HHS (three patients, 9.4%), and both (14 patients, 43.8%) (Table 1). Among the 32 patients with ICI-T1DM, glutamic acid decarboxylase antibody (GAD-Ab) testing was performed in 20 patients (62.5%), and only one patient (5% of those tested) was GAD-Ab positive (≥1.0 U/mL).

ICI-T1DM risk factors

The univariate analysis revealed that patients with pre-existing DM prior to ICI therapy (odds ratio [OR], 2.774; 95% CI, 1.382 to 5.566) and those treated with both anti–CTLA-4 and anti–PD-1/PD-L1 ICIs (OR, 5.979; 95% CI, 2.280 to 15.681) were significantly more likely to develop ICI-T1DM. Additionally, patients with prolonged ICI treatment duration (≥12 weeks; OR, 6.781; 95% CI, 2.643 to 17.863) and those with renal cell carcinoma were more likely to develop ICI-T1DM (OR, 3.225; 95% CI, 1.235 to 8.422). Thyroid dysfunction (OR, 5.371; 95% CI, 2.539 to 11.362) and adrenal dysfunction (OR, 2.588; 95% CI, 1.115 to 6.006) were significantly associated with an increased risk of ICI-T1DM (Table 1).

The multivariate logistic regression analysis identified significant risk factors for ICI-T1DM after adjusting for potential confounders. The model demonstrated that patients who received both anti–CTLA-4 and anti–PD-1/PD-L1 ICIs (HR, 3.666; 95% CI, 1.224 to 10.979) and those with prolonged ICI treatment duration (≥12 weeks; HR, 4.789; 95% CI, 1.806 to 12.701) were at a significantly increased risk of developing ICIT1DM. Additionally, pre-existing DM prior to ICI therapy (HR, 2.352; 95% CI, 1.140 to 4.854) and thyroid dysfunction (HR, 4.027; 95% CI, 1.847 to 8.779) remained significant risk factors in the multivariate analysis (Table 2).

Survival outcomes

Overall, 6,956 patients received ICI therapy, of whom 3,242 patients died. The median follow-up period was 245 days (interquartile range, 84 to 580). The Cox proportional hazards model revealed that ICI-T1DM incidence was associated with 78% reduction in mortality (HR, 0.224; 95% CI, 0.093 to 0.539). Kaplan–Meier survival curves show survival differences between patients with and without ICI-T1DM (Fig. 2). Additionally, risk factors such as receiving both ICI types (HR, 0.671; 95% CI, 0.512 to 0.880) significantly reduced mortality rates. Thyroid dysfunction is another significant risk factor associated with improved survival (HR, 0.616; 95% CI, 0.566 to 0.670). Conversely, steroid use within 90 days of initiating ICI therapy increased mortality risk (HR, 1.505; 95% CI, 1.401 to 1.617). Furthermore, BMI influenced survival outcomes for patients who are overweight (HR, 0.770; 95% CI, 0.701 to 0.845) and obese (HR, 0.733; 95% CI, 0.672 to 0.801), reducing their mortality, whereas underweight patients faced a higher mortality risk (HR, 1.505; 95% CI, 1.339 to 1.688).

Additionally, the analysis revealed that certain cancer types significantly influenced survival outcomes in patients receiving ICI therapy. Specifically, the HRs for mortality were 1.258 (95% CI, 1.112 to 1.410) for lung cancer, 1.304 (95% CI, 1.146 to 1.483) for hepatocellular carcinoma, 1.371 (95% CI, 1.173 to 1.602) for gastric cancer, 1.365 (95% CI, 1.113 to 1.674) for renal cell carcinoma, 1.652 (95% CI, 1.381 to 1.977) for urothelial carcinoma, and 1.553 (95% CI, 1.244 to 1.938) for head and neck cancer. Breast cancer was the only type associated with a significant survival benefit, with an HR of 0.643 (95% CI, 0.445 to 0.930). Other cancer types also demonstrated increased mortality risk, with an HR of 1.567 (95% CI, 1.381 to 1.778). The detailed results are presented in Table 3.

DISCUSSION

In our study, ICI-T1DM was identified in 32 patients (0.46%) with a median onset of 41 weeks (range, 9 to 1,359 days post-ICI). Among them, 62.5% (n=14) experienced acute hyperglycemic complications, such as DKA and/or HHS. Significant risk factors included combined anti–CTLA-4 and anti–PD-1/PD-L1 therapy (HR, 3.666; 95% CI, 1.224 to 10.979), prolonged ICI treatment (≥12 weeks; HR, 4.789; 95% CI, 1.806 to 12.701), and pre-existing diabetes (HR, 2.352; 95% CI, 1.140 to 4.854). ICI-T1DM was notably associated with a 78% reduction in mortality (HR, 0.224; 95% CI, 0.093 to 0.539), suggesting its potential as a marker of enhanced immune activation. Receiving both ICI types and being overweight/obese were associated with significantly lower mortality, whereas GC administration increased mortality (HR, 1.505; 95% CI, 1.401 to 1.617).

Among the identified cases, one patient showed delayed-onset ICI-T1DM, diagnosed 1,359 days after starting nivolumab and 379 days after adding ipilimumab. This patient continued ICI therapy throughout, with no other causes for β-cell dysfunction identified. Prolonged ICI exposure and combination therapy likely led to cumulative immune activation, triggering delayed-onset ICI-T1DM. Although rare, such cases emphasize the necessity for extended monitoring of irAEs during long-term ICI therapy. Similar cases in prior studies highlight the variability in ICI-T1DM onset timing [9,14-17].

Owing to their different etiologies and pathophysiology, distinguishing classic T1DM from ICI-T1DM is crucial [8]. Classic T1DM involves slow β-cell destruction through autoimmune mechanisms, often marked by autoantibodies. In contrast, ICI-T1DM develops rapidly post-treatment, driven by enhanced ICI-induced immune responses [8]. A comprehensive review highlighted that traditional T1DM-associated autoantibodies, including GAD-Ab, are infrequent in ICI-T1DM, suggesting distinct autoimmune responses [18]. Similarly, in this study, among the 20 patients with ICI-T1DM who underwent GAD-Ab testing only one (5%) was positive (1.2 U/mL), with levels declining during follow-up. This underscores the need for alternative biomarkers or diagnostic criteria tailored to its unique pathophysiology.

Previous studies have indicated that anti–PD-1 therapy is associated with the risk of developing ICI-T1DM [10]. Patients receiving both anti–PD-1 and anti–CTLA-4 had a significantly higher risk of developing ICI-T1DM compared with those on anti–PD-1/PD-L1 alone, as confirmed by other studies [2,10]. Experimental and clinical research suggests that the absence of PD-1 on T-cells or PD-L1 on pancreatic β-cells is a major cause of classic T1DM. In contrast, the absence or blockade of CTLA-4 is not significantly associated with T1DM development [19]. Additionally, anti-TNF and anti-IFN interventions can prevent ICI-T1DM development in anti–PD-L1–treated models, highlighting the protective role of PD-L1 expression against immune-mediated stress in pancreatic β-cells [20]. Therefore, the impaired PD-1/PD-L1 interactions are critical for ICI-T1DM development. Anti–PD-1/PD-L1 therapy blocks these interactions, increasing T-cell infiltration and β-cell destruction, which are further amplified by the additional anti–CTLA-4 therapies.

Survival analysis showed that patients treated with both anti–PD-1/PD-L1 and anti–CTLA-4 had significantly lower mortality rates (HR, 0.671; 95% CI, 0.512 to 0.880) than those treated with anti–PD-1/PD-L1 alone. Combined ICI therapy leads to dual immune activation, resulting in a strong anticancer effect; this is consistent with clinical studies and meta-analyses reporting high response rates and prolonged progression-free survival [21,22].

Prolonged ICI treatment (≥12 weeks) was significantly associated with a higher risk of developing ICI-T1DM (HR, 4.789; 95% CI, 1.806 to 12.701). The 12-week cutoff is commonly used in immuno-oncological studies as a meaningful threshold for evaluating responses and toxicities [23,24]. Extended ICI exposure likely amplifies immune activation by boosting antitumor efficacy and increasing the risk of immune-mediated β-cell destruction. This aligns with studies showing that longer ICI durations intensify immune responses, especially in highly immunogenic tumors, such as melanoma and renal cell carcinoma [5,25]. The heightened ICI-T1DM development risk underscores the need for vigilant monitoring and early intervention for irAEs during extended treatment. Future research should explore how prolonged ICI therapy impacts irAE development and identify biomarkers predictive of ICI-T1DM to enable personalized management strategies [9,26,27].

The pathophysiology of type 2 diabetes mellitus (T2DM) begins with insulin resistance. Initially, the pancreas compensates by increasing insulin secretion to maintain normal glucose levels, but sustained resistance stresses β-cells, causing dysfunction and death [28,29]. This reduces β-cell mass and insulin secretion capacity over time, leading to fasting and postprandial glucose levels meeting diabetes diagnostic criteria [30-32]. In this study, as no patients required multiple daily insulin injections due to classic T1DM or total pancreatectomy, most classified as having ‘pre-existing DM before ICI treatment,’ are likely T2DM, marked by insulin resistance and gradual β-cell mass decline. These patients showed a higher risk of developing ICI-T1DM than those without diabetes (HR, 2.352; 95% CI, 1.140 to 4.854), as β-cells already damaged in T2DM are more vulnerable to additional destruction, often leading to insulin dependency. However, patients who are overweight or obese did not show a significantly higher risk, suggesting that although insulin resistance is common in T2DM and obesity, it does not independently drive ICI-T1DM development.

GC administration within 90 days prior to ICI therapy did not significantly affect the risk of developing ICI-T1DM but was associated with increased mortality (HR, 1.505; 95% CI, 1.401 to 1.617). This suggests that GC-induced immunosuppression may reduce anti-cancer efficacy, leading to higher mortality rates, especially in patients with severe underlying conditions [2,33-35]. Similarly, potent immunosuppressants, including disease-modifying antirheumatic drugs (DMARDs), biological DMARDs, and high-dose GCs, have shown reduced ICI-T1DM risk but increased mortality, likely reflecting their effects on immune suppression and advanced disease states. Patients on GCs often have severe comorbidities, further explaining their elevated mortality rates. Interpreting these results requires caution, as higher mortality in these patients may reflect the severity of their underlying conditions rather than reduced ICI efficacy [2]. Among cancer types, breast cancer was the only malignancy with a significant survival benefit from ICI therapy (HR, 0.643; 95% CI, 0.445 to 0.930), consistent with findings on PD-L1–expressing triple-negative breast cancer showing improved progression-free and overall survival [36,37]. In contrast, cancers such as lung and hepatocellular carcinomas did not exhibit similar benefits, possibly due to coexisting malignancies or greater tumor burden. Herein, 808 patients had multiple cancers, including 557 with lung cancer, potentially confounding survival analysis. These findings underscore the need for tailored interpretations and refined analyses, particularly for patients with complex disease profiles or multiple malignancies, to better understand ICI benefits across diverse populations [38,39].

One notable finding is that the patients who developed ICI-T1DM demonstrated a significant survival benefit compared with those who did not (HR, 0.224; 95% CI, 0.093 to 0.539), consistent with studies showing higher survival rates in patients with irAEs [33,34,40]. This suggests that ICI-T1DM may serve as a positive indicator of enhanced immune activation and effective ICI response. In contrast, a large-sized, retrospective study reported that ICI-T1DM occurrence did not significantly impact the survival of patients receiving ICI therapy [2]. This study included a broad population of patients from various regions and races with different health conditions, as it utilized data from a diverse group of insured patients in the United States. The varied treatment environments across multiple healthcare institutions could have contributed to variability in survival outcomes. However, our study collected data from a single tertiary care institution, which likely provided more consistent treatment and monitoring, resulting in less variability in survival outcomes. Additionally, differences in tumor type distribution between populations could have influenced survival outcomes, as cancer type significantly affects ICI efficacy and survival [41,42].

In this study, thyroid dysfunction was significantly associated with an increased risk of ICI-T1DM (HR, 4.027; 95% CI, 1.847 to 8.779) and improved survival (HR, 0.616; 95% CI, 0.566 to 0.881). These results suggest it may serve as a biomarker for heightened immune activation during ICI therapy, reflecting robust antitumor responses [5].

In this study, 62.5% of patients with ICI-T1DM experienced severe hyperglycemic complications such as DKA or HHS, but no deaths were reported. Considering that the mortality rate for DKA in older adults exceeds 5% and mortality rate for HHS is 5% to 20% [7], with a median patient age of 63 years and metastatic tumors, the absence of death suggests effective medical access and acute interventions. This likely maximized the survival benefits observed with ICI-T1DM. Long-term mortality for DKA is 8% for 60 to 69 years of age, increasing to over 27% for patients exceeding 70 years, whereas for HHS, it is 10% for patients below 75 years, increasing to over 19% for patients aged 75 years [43]. DKA/HHS occurrence increases long-term mortality risk by 2.8-fold, and recurring episodes increase it by 4.5-fold. Cancer further increases mortality risk by 3.1-fold [43]. Therefore, preventing the initial occurrence of acute hyperglycemic complications is crucial for improving long-term survival rates in patients with ICI-T1DM. However, once DKA/HHS has occurred, preventing further occurrences is critical to reduce long-term mortality risk. Therefore, the process of identifying risk factors and actively monitoring high-risk groups for early detection and intervention of ICI-T1DM before patients develop DKA or HHS is essential. Hence, patients suspected of having ICI-T1DM should be promptly referred to an endocrinologist or diabetes specialist.

This study has several limitations, including its retrospective design, single-center setting, small sample size for ICI-T1DM cases, and lack of long-term follow-up data, all of which may introduce selection bias and limit generalizability. With only 32 ICI-T1DM cases, the small sample size limits statistical power, particularly for subgroup analyses. Despite these limitations, we employed multivariable regression analyses to adjust for confounders and enhance the robustness of our findings. Detailed inclusion and exclusion criteria improved the precision and reliability of ICI-T1DM classification. The single-center set-up, encompassing 6,956 ICI-treated patients, ensured consistency in treatment protocols and standardized data collection, reducing confounding variables typical in multi-institutional studies. This set-up allowed us to attribute observed effects more clearly to specific treatments and patient characteristics, strengthening internal validity. Our study also addresses the limitations of diagnostic code-based research, which can lead to misclassification and underreporting. By establishing classification criteria based on clinical scenarios and biochemical markers, such as decreased C-peptide levels, our research enhances the precision and reliability of patient identification. This methodological rigor enabled a more accurate evaluation of ICI-T1DM epidemiology and facilitated the identification of key risk factors and survival outcomes, providing valuable insights for disease management. Future research should validate these findings in larger, multicenter studies with extended follow-up to improve comprehension of ICI-T1DM’s long-term impacts and refine diagnostic criteria. Incorporating autoimmune biomarkers and glucose-stimulated C-peptide testing could further enhance diagnostic specificity and sensitivity, refining therapeutic strategies and improving patient outcomes.

In conclusion, this study revealed an ICI-T1DM incidence rate of 0.46%, with dual therapy of anti–PD-1/PD-L1 and anti–CTLA-4, prolonged ICI treatment (≥12 weeks), pre-existing diabetes, and thyroid dysfunction identified as major risk factors. Moreover, the improved survival rates observed in patients with ICI-T1DM imply ICI-enhanced immune activation. This study underscores the importance of identifying high-risk patients based on these risk factors and emphasizes the need for rigorous blood glucose monitoring and early intervention to manage ICI-T1DM effectively, ultimately contributing to maximizing the survival rates in patients undergoing ICI therapy.

NOTES

CONFLICTS OF INTEREST

No potential conflict of interest relevant to this article was reported.

AUTHOR CONTRIBUTIONS

Conception and design: Y.K.C., E.H.K.

Acquisition, analysis, interpretation of data: S.H.G., Y.K.C.

Drafting the work or revising: S.H.G., Y.K.C.

Final approval of the manuscript: all authors.

FUNDING

None

ACKNOWLEDGMENTS

None

Fig. 1.

Flowchart for immune checkpoint inhibitor-induced type 1 diabetes mellitus (ICI-T1DM) classification. aPersistent insulin requirement was defined as ongoing insulin therapy at the last follow-up or during subsequent management.

Fig. 2.

Kaplan–Meier curves for overall survival (immune checkpoint inhibitor-induced type 1 diabetes mellitus [ICIT1DM] is time-varying covariate).

REFERENCES

• 1. Postow MA, Sidlow R, Hellmann MD. Immune-related adverse events associated with immune checkpoint blockade. N Engl J Med 2018;378:158-68.ArticlePubMed
• 2. Chen X, Affinati AH, Lee Y, Turcu AF, Henry NL, Schiopu E, et al. Immune checkpoint inhibitors and risk of type 1 diabetes. Diabetes Care 2022;45:1170-6.ArticlePubMedPMCPDF
• 3. Marsiglio J, McPherson JP, Kovacsovics-Bankowski M, Jeter J, Vaklavas C, Swami U, et al. A single center case series of immune checkpoint inhibitor-induced type 1 diabetes mellitus, patterns of disease onset and long-term clinical outcome. Front Immunol 2023;14:1229823.ArticlePubMedPMC
• 4. Marsiglio J, Kovacsovics M, Wang X, Hardikar S, Young A, Doherty J, et al. 1251 Immune checkpoint inhibitor-induced type 1 diabetes mellitus is associated with long-term response and survival in patients with melanoma. J Immunother Cancer 2022;10(Suppl 2):A1298.Article
• 5. Barroso-Sousa R, Barry WT, Garrido-Castro AC, Hodi FS, Min L, Krop IE, et al. Incidence of endocrine dysfunction following the use of different immune checkpoint inhibitor regimens: a systematic review and meta-analysis. JAMA Oncol 2018;4:173-82.ArticlePubMedPMC
• 6. Husebye ES, Castinetti F, Criseno S, Curigliano G, Decallonne B, Fleseriu M, et al. Endocrine-related adverse conditions in patients receiving immune checkpoint inhibition: an ESE clinical practice guideline. Eur J Endocrinol 2022;187:G1-21.ArticlePubMedPMCPDF
• 7. Kitabchi AE, Umpierrez GE, Miles JM, Fisher JN. Hyperglycemic crises in adult patients with diabetes. Diabetes Care 2009;32:1335-43.ArticlePubMedPMCPDF
• 8. Cho YK, Jung CH. Immune-checkpoint inhibitors-induced type 1 diabetes mellitus: from its molecular mechanisms to clinical practice. Diabetes Metab J 2023;47:757-66.ArticlePubMedPMCPDF
• 9. Kotwal A, Haddox C, Block M, Kudva YC. Immune checkpoint inhibitors: an emerging cause of insulin-dependent diabetes. BMJ Open Diabetes Res Care 2019;7:e000591.ArticlePubMedPMC
• 10. Mourad D, Azar NS, Eid AA, Azar ST. Immune checkpoint inhibitor-induced diabetes mellitus: potential role of T cells in the underlying mechanism. Int J Mol Sci 2021;22:2093.ArticlePubMedPMC
• 11. Waljee AK, Rogers MA, Lin P, Singal AG, Stein JD, Marks RM, et al. Short term use of oral corticosteroids and related harms among adults in the United States: population based cohort study. BMJ 2017;357:j1415.ArticlePubMedPMC
• 12. Korean Diabetes Association. Clinical practice guidelines for diabetes. 8th ed. Seoul: Korean Diabetes Association; 2021.
• 13. Loscalzo J, Fauci A, Kasper D, Hauser S, Longo D, Jameson JL. Harrison’s principles of internal medicine. 21st ed. New York: McGraw-Hill Education; 2022.
• 14. Stamatouli AM, Quandt Z, Perdigoto AL, Clark PL, Kluger H, Weiss SA, et al. Collateral damage: insulin-dependent diabetes induced with checkpoint inhibitors. Diabetes 2018;67:1471-80.ArticlePubMedPMCPDF
• 15. Okamoto M, Okamoto M, Gotoh K, Masaki T, Ozeki Y, Ando H, et al. Fulminant type 1 diabetes mellitus with anti-programmed cell death-1 therapy. J Diabetes Investig 2016;7:915-8.ArticlePubMedPMCPDF
• 16. Clotman K, Janssens K, Specenier P, Weets I, De Block CE. Programmed cell death-1 inhibitor-induced type 1 diabetes mellitus. J Clin Endocrinol Metab 2018;103:3144-54.ArticlePubMed
• 17. Godwin JL, Jaggi S, Sirisena I, Sharda P, Rao AD, Mehra R, et al. Nivolumab-induced autoimmune diabetes mellitus presenting as diabetic ketoacidosis in a patient with metastatic lung cancer. J Immunother Cancer 2017;5:40.ArticlePubMedPMCPDF
• 18. Wu L, Tsang V, Menzies AM, Sasson SC, Carlino MS, Brown DA, et al. Risk factors and characteristics of checkpoint inhibitor-associated autoimmune diabetes mellitus (CIADM): a systematic review and delineation from type 1 diabetes. Diabetes Care 2023;46:1292-9.ArticlePubMedPDF
• 19. Ansari MJ, Salama AD, Chitnis T, Smith RN, Yagita H, Akiba H, et al. The programmed death-1 (PD-1) pathway regulates autoimmune diabetes in nonobese diabetic (NOD) mice. J Exp Med 2003;198:63-9.ArticlePubMedPMCPDF
• 20. Perdigoto AL, Deng S, Du KC, Kuchroo M, Burkhardt DB, Tong A, et al. Immune cells and their inflammatory mediators modify β cells and cause checkpoint inhibitor-induced diabetes. JCI Insight 2022;7:e156330.ArticlePubMedPMC
• 21. Zhang L, Sun L, Zhou Y, Yu J, Lin Y, Wasan HS, et al. Association of survival and immune-related adverse events with anti-PD-1/PD-L1 and anti-CTLA-4 inhibitors, alone or their combination for the treatment of cancer: a systematic review and meta-analysis of 13 clinical trials. Front Oncol 2021;11:575457.ArticlePubMedPMC
• 22. Gu L, Khadaroo PA, Su H, Kong L, Chen L, Wang X, et al. The safety and tolerability of combined immune checkpoint inhibitors (anti-PD-1/PD-L1 plus anti-CTLA-4): a systematic review and meta-analysis. BMC Cancer 2019;19:559.ArticlePubMedPMCPDF
• 23. Seymour L, Bogaerts J, Perrone A, Ford R, Schwartz LH, Mandrekar S, et al. iRECIST: guidelines for response criteria for use in trials testing immunotherapeutics. Lancet Oncol 2017;18:e143-52.ArticlePubMedPMC
• 24. Wolchok JD, Hoos A, O’Day S, Weber JS, Hamid O, Lebbe C, et al. Guidelines for the evaluation of immune therapy activity in solid tumors: immune-related response criteria. Clin Cancer Res 2009;15:7412-20.ArticlePubMedPDF
• 25. Postow MA, Callahan MK, Wolchok JD. Immune checkpoint blockade in cancer therapy. J Clin Oncol 2015;33:1974-82.ArticlePubMedPMC
• 26. Michot JM, Bigenwald C, Champiat S, Collins M, Carbonnel F, Postel-Vinay S, et al. Immune-related adverse events with immune checkpoint blockade: a comprehensive review. Eur J Cancer 2016;54:139-48.ArticlePubMed
• 27. Brahmer JR, Lacchetti C, Schneider BJ, Atkins MB, Brassil KJ, Caterino JM, et al. Management of immune-related adverse events in patients treated with immune checkpoint inhibitor therapy: American Society of Clinical Oncology Clinical Practice Guideline. J Clin Oncol 2018;36:1714-68.PubMed
• 28. Cnop M, Welsh N, Jonas JC, Jorns A, Lenzen S, Eizirik DL. Mechanisms of pancreatic beta-cell death in type 1 and type 2 diabetes: many differences, few similarities. Diabetes 2005;54 Suppl 2:S97-107.PubMed
• 29. Dutta P, Ma L, Ali Y, Sloot PM, Zheng J. Boolean network modeling of β-cell apoptosis and insulin resistance in type 2 diabetes mellitus. BMC Syst Biol 2019;13(Suppl 2):36.ArticlePubMedPMCPDF
• 30. Montanya E. Insulin resistance compensation: not just a matter of β-Cells? Diabetes 2014;63:832-4.ArticlePubMedPDF
• 31. Shanik MH, Xu Y, Skrha J, Dankner R, Zick Y, Roth J. Insulin resistance and hyperinsulinemia: is hyperinsulinemia the cart or the horse? Diabetes Care 2008;31 Suppl 2:S262-8.PubMed
• 32. Taylor R. Insulin resistance and type 2 diabetes. Diabetes 2012;61:778-9.ArticlePubMedPMCPDF
• 33. Bruera S, Suarez-Almazor ME. The effects of glucocorticoids and immunosuppressants on cancer outcomes in checkpoint inhibitor therapy. Front Oncol 2022;12:928390.ArticlePubMedPMC
• 34. Maslov DV, Tawagi K, Kc M, Simenson V, Yuan H, Parent C, et al. Timing of steroid initiation and response rates to immune checkpoint inhibitors in metastatic cancer. J Immunother Cancer 2021;9:e002261.ArticlePubMedPMC
• 35. Yu Y. Repurposing glucocorticoids as adjuvant reagents for immune checkpoint inhibitors in solid cancers. Cancer Biol Med 2021;18:944-8.ArticlePubMedPMC
• 36. Cortes J, Rugo HS, Cescon DW, Im SA, Yusof MM, Gallardo C, et al. Pembrolizumab plus chemotherapy in advanced triplenegative breast cancer. N Engl J Med 2022;387:217-26.ArticlePubMed
• 37. Schmid P, Adams S, Rugo HS, Schneeweiss A, Barrios CH, Iwata H, et al. Atezolizumab and nab-paclitaxel in advanced triple-negative breast cancer. N Engl J Med 2018;379:2108-21.PubMed
• 38. Horn L, Mansfield AS, Szczesna A, Havel L, Krzakowski M, Hochmair MJ, et al. First-line atezolizumab plus chemotherapy in extensive-stage small-cell lung cancer. N Engl J Med 2018;379:2220-9.ArticlePubMed
• 39. Herbst RS, Soria JC, Kowanetz M, Fine GD, Hamid O, Gordon MS, et al. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients. Nature 2014;515:563-7.ArticlePubMedPMCPDF
• 40. van Not OJ, Verheijden RJ, van den Eertwegh AJ, Haanen JB, Aarts MJ, van den Berkmortel FW, et al. Association of immune-related adverse event management with survival in patients with advanced melanoma. JAMA Oncol 2022;8:1794-801.ArticlePubMedPMC
• 41. Meybodi SM, Farasati Far B, Pourmolaei A, Baradarbarjastehbaf F, Safaei M, Mohammadkhani N, et al. Immune checkpoint inhibitors promising role in cancer therapy: clinical evidence and immune-related adverse events. Med Oncol 2023;40:243.ArticlePubMedPDF
• 42. Wang C, Li J, Zhang Q, Wu J, Xiao Y, Song L, et al. The landscape of immune checkpoint inhibitor therapy in advanced lung cancer. BMC Cancer 2021;21:968.ArticlePubMedPMCPDF
• 43. Huang CC, Weng SF, Tsai KT, Chen PJ, Lin HJ, Wang JJ, et al. Long-term mortality risk after hyperglycemic crisis episodes in geriatric patients with diabetes: a national population-based cohort study. Diabetes Care 2015;38:746-51.ArticlePubMedPDF

Figure & Data

References

Citations

Citations to this article as recorded by  

About this article

Go Sh, Cho YK, Koh EH. Risk Factors and Survival Outcomes of Immune Checkpoint Inhibitor-Induced Type 1 Diabetes Mellitus: A Retrospective Cohort Study. Diabetes Metab J. 2026;50(1):115-126.

Received: Aug 05, 2024; Accepted: Apr 08, 2025

DOI: https://doi.org/10.4093/dmj.2024.0455.

Download citation ↓