Abstract
Background
Immune checkpoint inhibitors (ICIs) are associated with potentially severe immune-related adverse events (irAEs). Emerging clinical practice reports have suggested higher incidence of irAEs in real-world settings than initially observed in phase III clinical trials. Objectives were to determine the incidence of irAEs associated with ICIs in a clinical population, the Veterans Health Administration, characterize their time to onset, and explore potential risk factors.
Methods
This retrospective observational study included patients from eight Midwest VA medical centers who initiated an ICI between January 1, 2014, and June 30, 2022. Courses of incident prednisone therapy lasting at least seven days at a dose ≥ 20 mg/day were used to identify irAEs, within two years following ICI initiation. A multivariate Cox proportional hazards regression model was used to explore potential irAE risk factors.
Results
Of 1314 patients, the incidence of irAEs was 19.8%, with most (86.5%) occurring within one year of ICI initiation. Monthly incidence rates peaked three months following ICI initiation at 3.0% and decreased thereafter. Female gender (hazard ratio [HR] = 2.01, 95% confidence interval [CI]: 1.01–4.00) and combination therapy with ipilimumab and nivolumab (HR = 2.46, 95% CI: 1.44–4.21) were significantly associated with irAE incidence.
Conclusions
These findings are consistent with recent studies in clinical populations that demonstrate higher irAE incidence rates than originally reported in clinical trials. Our findings may enhance prompt recognition and treatment of irAEs for VA patients.
Introduction
Novel immunotherapies have revolutionized the treatment of advanced cancers, and patients are receiving increased opportunities to delay the progression of severe disease. Immune checkpoint inhibitors (ICIs) are immunotherapies that include antibodies against cytotoxic T lymphocyte antigen-4 (CTLA-4), programmed cell death protein 1 (PD-1), and PD-ligand 1 (PD-L1). Immune checkpoint inhibitors are becoming commonly used for several advanced cancers, some including melanoma, renal cell carcinoma, non-small cell lung cancer, and hepatocellular carcinoma. 1
A significant benefit to using less cytotoxic agents is avoiding adverse effects that make it difficult for patients to persist with therapy. However, immunotherapies carry their own risk profiles. Immune checkpoint inhibitors provoke activation of the body's immune response that leads to an antitumor response. In turn, inflammation of any organ or organ system can occur following ICI administration and require treatment discontinuation and/or systemic steroids for symptom treatment. These events are known as immune-related adverse events (irAEs). 1 As more literature from clinical practice experience becomes available, it has been observed that the incidence of irAEs may be higher in clinical practice than initially shown in phase III clinical trials, specifically for the CTLA-4 inhibitor ipilimumab. 2 Early phase III trials for ipilimumab showed incidences of 10–20% for grades 3–4 irAEs, whereas more recent clinical trials and observational studies have shown incidences of up to 45%.2–5 A meta-analysis of ICIs trials found a higher incidence of irAE with CTLA-4 inhibitors than with PD-1 and PD-L1 inhibitors. 6
Onsets of toxicities occur at different times for various known and unknown reasons. For example, dermatologic toxicities are often the earliest adverse effects to develop between two and five weeks. 1 Toxicities generally occur most frequently in the first three months. 7 Further, little is known about risk factors for ICI irAEs. Suggested risk factors include age less than 60 years, high body mass index (BMI), chronic smoking, a tumor proportion score of greater than or equal to 90%, and adenocarcinoma histology.8,9 The ability to understand which patients might be at greater risk for developing irAEs will also increase the likelihood of identifying and treating toxicities early.
Although the bank of literature surrounding the implications of immunotherapy continues to grow, there is much that remains unknown, specifically to various patient populations. With ICI use projected to increase, it is imperative for practitioners to better understand treatment responses in their patient population. Therefore, the primary objective of this study was to determine the incidence of irAEs associated with ICIs in the veteran population. Because of high variability reported in toxicity onset and the older age of the veteran population, secondary objectives were to determine time to onset of irAEs and identify possible risk factors for developing irAEs. Additional information regarding ICI irAE incidence, time to onset, and risk factors in specific populations may allow for more accurate patient education and improved recognition and treatment of irAEs.
Methods
Data sources
This retrospective observational study used administrative data extracted from the IV package of VA Vista Chemotherapy Manager (VCM) for eight Midwest VA medical centers to identify patients receiving ICIs. Additional administrative data for these sites, including pharmacy dispensing information and international classification of diseases codes (ICD-9 and ICD-10) from medical encounters, were obtained from the VA Corporate Data Warehouse (CDW), accessed via the VA Informatics and Computing Infrastructure, and linked to VCM data using patient identifiers common to both datasets. Vista Chemotherapy Manager data were used as the primary source of dispensing information for ICIs because medications dispensed by outpatient infusion in the VA health care system are not reliably captured at all sites across time in the standard pharmacy data dispensing information archived in the CDW. This study was approved by the University of Iowa Institutional Review Board and the Iowa City Veterans Administration Research and Development Committee.
Patients
The patient population included all veterans in the Veterans Integrated Service Network (VISN) 23 who received a first-time administration of an ICI between January 1, 2014, and June 30, 2022. Immune checkpoint inhibitors received by veterans during this time frame included pembrolizumab, nivolumab, atezolizumab, durvalumab, cemiplimab, avelumab, and ipilimumab. Eight patients could not be linked between VCM and CDW data sources and were excluded.
Outcome
The primary outcome was conceptually defined as an irAE of grades 3–4 toxicity and operationally defined as a new course of prednisone therapy of ≥7 days duration at a dose ≥20 mg per day, referred to hereafter as a prednisone-treated irAE (pt-irAE). Alternative definitions requiring 14-day and 28-day prednisone courses were examined as secondary outcomes. Prednisone-treated-irAEs were examined for up to two years following ICI initiation and events occurring within six months after ICI discontinuation were included as pt-irAEs. As direct evidence and grading of irAEs cannot be easily determined from administrative health care records, prednisone exposure was used as a proxy as it is the primary treatment for grades 3–4 irAEs according to the Common Terminology Criteria for Adverse Events scale. 10
Analysis
Incident pt-irAEs following ICI initiation and the time to first event from initiation were determined. Monthly pt-irAE incidence rates following ICI initiation were calculated by the number of events occurring each month, assessed as consecutive 30-day intervals, divided by the number of patients persisting on ICI therapy at the beginning of that month. Multivariate Cox proportional hazards regression was used to explore risk factors for experiencing pt-irAEs, including patient demographics, BMI, ICI medication type, cancer type, use of concurrent chemotherapy or a vascular endothelial growth factor (VEGF) inhibitor, and preexisting autoimmune disease. 11
Results
A total of 1314 patients in VISN 23 initiated an ICI between January 1, 2014, and June 30, 2022. Patients were predominantly male (97.6%) and White race (89.4%) (Table 1). Approximately half of patients (55.6%) were between 65 and 74 years of age. The most frequently used ICI at index was pembrolizumab (49.9%), and the most common cancer type was respiratory cancer (46.4%). Additionally, 27.2% of patients were also treated with chemotherapy or a VEGF inhibitor during ICI treatment course.
Patient characteristics (N = 1314).
ICI: immune checkpoint inhibitor.
The incidence of pt-irAEs was 19.8% (Table 2). Alternative definitions requiring longer prednisone treatment courses produced lower incidence estimates. Incidence was highest in the five months following ICI initiation (Figure 1). The majority of pt-irAEs (86.5%) occurred within 365 days of ICI initiation, and 66.2% of pt-irAEs occurred within 180 days (Table 3).
Incidence rate curve of prednisone-treated immune-related adverse events following initiation of an immune checkpoint inhibitor.
Incidence of prednisone-treated immune-related adverse events (pt-irAE), defined by an observed course of varying required durations.
Time to onset of prednisone-treated immune-related adverse event (pt-irAE) (N = 260).
Factors associated with prednisone-treated immune-related adverse event (pt-irAE) risk following initiation of an immune checkpoint inhibitor, using cox proportional hazards regression.
CI: confidence interval; HR: hazard ratio; ICI: immune checkpoint inhibitor; irAE: immune-related adverse event; VEGF: vascular endothelial growth factor.
In a multivariate Cox proportional hazards regression model, statistically significant risk factors associated with pt-irAEs included female gender and combination therapy with ipilimumab and nivolumab (Table 4). Women were twice as likely (hazard ratio [HR] = 2.01, 95% confidence interval [CI]: 1.01–4.00) to have a pt-irAE compared to men, and patients who initiated combination therapy with ipilimumab and nivolumab were nearly 2.5 times as likely (HR = 2.46, 95% CI: 1.44–4.21) relative to pembrolizumab monotherapy. Some other potential risk factors failed to show statistical significance but may have clinical significance if the point estimates of risk are accurately estimated. Patients with preexisting autoimmune disorders were 30% more likely to have pt-irAEs (HR = 1.30, 95% CI: 0.74–2.28). Patients who received combination therapy with an ICI and chemotherapy were 21% less likely to have irAEs (HR = 0.79, 95% CI: 0.55, 1.15). Other factors of note that failed to show statistical significance and any evidence of clinical significance included age, BMI, and race.
Discussion
Documenting outcomes and experience from clinical oncology practice can provide vital complementary data to determine whether results from clinical trials, which are crucial for guiding clinical decision-making, can be generalized to real-world patients receiving ICI therapy. For example, a recent retrospective observational study of patients receiving ICIs in the VA healthcare system observed lower overall survival rates than reported in pivotal clinical trials. 12 In this retrospective observational study of eight VA medical centers, we observed a real-world incidence of pt-irAEs associated with ICIs of 19.8%. This is comparable to a recently published meta-analysis of ICI clinical trials, which found incidences of nearly 30% with CTLA-4 inhibitors and nearly 20% with PD-1 inhibitors and PD-L1 inhibitors. 6 For monotherapy, the majority of ICIs (96.3%) used in our study were PD-L and PD-L1 inhibitors, with pembrolizumab used most. This was not surprising as it is a guideline-recommended treatment for lung cancers, and respiratory cancers accounted for nearly half of cancers in our population. Given this information, it is reasonable that our overall irAE incidence was closer to 20% than to 30%. One prior retrospective chart review of VA patients at a single medical center, not among the eight sites included in the current study, reported an overall irAE rate of 56.7%. 13 Direct comparison with our findings is limited as this rate included all irAEs, and no stratification by toxicity grade was reported.
Meta-analyses of ICIs have shown that irAE incidence is higher with CTLA-4 inhibitors, used as monotherapy or in combination with a PD-1 or PD-L1 inhibitor than with PD-1 and PD-L1 inhibitors.6,14,15 All patients in our study who received a CTLA-4 inhibitor received it in combination with a PD-1 or PD-L1 inhibitor. We found a 38.3% incidence with ipilimumab/nivolumab combination, compared to a range of 19.5–22.3% with PD-1 and PD-L1 inhibitors. These findings are consistent with meta-analysis risk estimates and demonstrate greater caution may be needed with particular agents.
Our secondary objective looking at irAE onset found that 66% of pt-irAEs occurred within six months of ICI initiation, which aligns with the typical median onset of between 2 and 16 months found by a 2020 review. 16 In a single-site chart review of VA patients initiating ICIs, approximately half of irAEs occurred within three months and approximately 70–75% within six months. 12 Few pt-irAEs (13.5%) in our study occurred after 365 days following ICI initiation, however, the risk is still present, and clinicians and patients should continue to monitor for irAEs later in treatment course.
Additional findings of our study correlate with current literature. Although statistical significance was not demonstrated, these findings include a lower incidence of pt-irAEs when patients also received chemotherapy and a higher incidence in those with preexisting autoimmune disorders.17–19 It is important to note that patients with preexisting autoimmune disorders were excluded from phase III clinical trials, leaving risk estimates largely unknown for these patients. A 2023 retrospective study from a large cancer center demonstrated that approximately 25% of patients with existing autoimmune disorders developed a new irAE after initiating ICI therapy and approximately 15% of patients had flares of their autoimmune disorder. 19 Because our study found an approximate 33% increased irAE risk for these patients, we were likely underpowered to demonstrate statistical significance. The ability of oncology providers to more accurately describe safety risks to patients may help to make informed decisions regarding their care.
The factors examined in our study provide guidance for estimates of individualized irAE risk. Suggested irAE risk factors in current literature that our study did not find to be associated with greater irAE incidence include lower age and higher BMI. 9 This prompts the need for more studies that examine risk factors in patients receiving ICIs.
To our knowledge, this is one of the largest retrospective studies examining the topic of irAEs with ICIs. Also unique is the fact that our study population encompassed several types of cancer. However, there are limitations to our study. Similar, smaller studies identified exact grade of irAEs, whereas we did not have this information available and instead used a surrogate of prednisone treatment to identify irAEs. While we restricted outcome events in the primary analysis to prednisone courses of at least 20 mg/day for at least seven days, it is possible that prednisone was prescribed for indications other than an irAE. Additionally, inability to track prednisone prescriptions dispensed outside the VA, the largely male patient population, and irAEs treated with medications other than prednisone, may have led to underestimates in calculated irAE incidences. Lastly, without chart review of every patient, we were unable to obtain information regarding cancer histology, type of irAE, and tumor proportion score. This information would have provided more specifics to our data but would not have altered the primary outcomes.
Given that our study population differs from clinical trials and from other observational studies, our findings are beneficial for physicians, pharmacists, and other oncology specialists. Patients may be provided more accurate estimates of their likelihood of experiencing irAEs and the timing of them. Future studies that would add to evidence for clinical practice include patient response to steroids, the ability of patients to resume ICI therapy following an irAE, risk factors for the development of irAEs, and an expanded review of irAEs for those with preexisting autoimmune conditions.
Footnotes
Acknowledgements
This research was supported by the Iowa City VA Health Care System, Department of Pharmacy Services. Additional support was provided by the Health Services Research and Development Service, Department of Veterans Affairs (Dr Lund, CIN 13–412). None of these sponsors had any role in the study design, methods, analyses, and interpretation, or in preparation of the manuscript and the decision to submit it for publication. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs.
Authors’ Contributions
All authors contributed substantially to the conceptual development and design of the project. BL was responsible for data acquisition and analysis. CK wrote the first draft of the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
