Sage Journals: Discover world-class research

Abstract

Background:

Immune checkpoint inhibitors (ICIs) improve cancer survival but can provoke new or worsening autoimmune disease.

Objectives:

To explore the association of ICIs with multiple sclerosis (MS) using the U.S. Food and Drug Administration’s Adverse Event Reporting System (FAERS) database, and to compare the disproportionality signal of MS with that of other autoimmune neurological adverse events.

Design:

Secondary analysis of the FAERS database.

Methods:

We performed a disproportionality analysis of FAERS between 2003-Q4 and 2024-Q2. Outcomes were MS, myasthenia gravis (MG), Guillain–Barré syndrome (GBS), and autoimmune encephalitis (AIE). A signal required ⩾3 reports, proportional reporting ratio ⩾2, and χ² ⩾ 4.

Results:

There were 48 reports of MS or MS relapse associated with ICIs. The reporting odds ratio (ROR) for MS or MS relapse was 0.09 (95% confidence interval (CI): 0.068–0.12). In comparison, RORs were 21.05 (95% CI: 19.287–22.974) for MG, 8.075 (95% CI: 6.677–9.766) for GBS, and 29.03 (95% CI: 23.564–35.764) for AIE.

Conclusion:

We found no significant safety signal for MS with ICIs, in contrast to MG, GBS, and AIE. This heterogeneity underscores the need for continued pharmacovigilance and mechanistic research.

Plain language summary

What FDA reports reveal about MS and immune checkpoint inhibitors

Immune checkpoint inhibitors (ICIs) are a type of cancer treatment that help the immune system recognize and attack cancer cells. While these therapies have significantly improved survival for many cancer patients, they can sometimes cause the immune system to mistakenly attack healthy tissue—a group of side effects known as immune-related adverse events (irAEs). Some of these irAEs affect the nervous system. This study looked at whether ICIs are linked to multiple sclerosis (MS), a chronic autoimmune disease that affects the brain and spinal cord. Researchers used a large government database—the FDA Adverse Event Reporting System (FAERS)—to investigate whether MS or MS relapses are reported more often in people receiving ICIs, compared to other neurological autoimmune conditions like myasthenia gravis (MG), Guillain-Barré syndrome (GBS), and autoimmune encephalitis (AIE). The results showed that MS was not reported more frequently than expected after ICI treatment. In fact, the reporting rate for MS was much lower than for MG, GBS, or AIE. This suggests that ICIs are unlikely to significantly increase the risk of MS flare-ups or onset.

Keywords

disproportionality analysis FDA adverse event reporting system immune checkpoint inhibitors multiple sclerosis

Introduction

Immune checkpoint inhibitors (ICIs) have transformed the therapeutic landscape for various malignancies, offering improved survival for patients with advanced cancers.¹ By unleashing the immune system to recognize and destroy tumor cells, ICIs—such as anti-programmed cell death protein 1 (PD-1), anti-programmed death-ligand 1 (PD-L1), and anti-cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) antibodies—have demonstrated significant efficacy across a broad range of tumors.² However, their immune-enhancing mechanisms can also provoke unintended immune-related adverse events (irAEs), including the development or exacerbation of autoimmune diseases.³

While the association of ICIs with autoimmune neurological conditions—such as myasthenia gravis (MG), Guillain–Barré syndrome (GBS), and autoimmune encephalitis (AIE)—has been recognized,⁴ the risk of multiple sclerosis (MS) onset or relapse in the context of ICI therapy remains unclear. Given the potential for MS to be influenced by systemic immune activation, understanding whether ICIs exacerbate MS or contribute to its development is of clinical and mechanistic interest. To date, the relationship between ICIs and MS disease activity remains incompletely understood, though findings from a limited number of small case series suggest that the risk of disease exacerbation is low in most patients.⁵ A recent systematic review and meta-analysis found that while relapses were infrequent, neuroradiological disease activity occurred in a subset of patients, particularly those who had discontinued disease-modifying therapies prior to ICI initiation.⁵

In this study, we leveraged the U.S. Food and Drug Administration’s Adverse Event Reporting System (FAERS) to assess whether ICIs are associated with increased reporting of MS and MS relapses. Additionally, we compared the disproportionality of MS-related adverse events (AEs) with that of other autoimmune neurological conditions to better understand the relative safety profile of ICIs across different neurological irAEs.

Methods

FAERS is a publicly available database that houses data on AEs submitted to the U.S. Food and Drug Administration (FDA). We conducted a disproportionality analysis of FAERS, spanning from the fourth quarter of 2003 to the second quarter of 2024, using the OpenVigil 2.1-MedDRA-v24 software⁶ to identify associations between ICIs and select neurological irAEs. OpenVigil 2.1 is an established pharmacovigilance tool that extracts, harmonizes, and standardizes reports from the publicly available FAERS database. It performs internal validation to ensure data quality and provides disproportionality analysis metrics.⁶ Disproportionality analysis is a validated pharmacovigilance method for detecting potential safety signals by evaluating whether specific drug–event combinations are reported more frequently than expected.

The ICIs included in our study comprised PD-1 inhibitors (pembrolizumab, nivolumab, cemiplimab, and dostarlimab), PD-L1 inhibitors (atezolizumab, durvalumab, and avelumab), and CTLA-4 inhibitors (ipilimumab and tremelimumab). For the purpose of analysis, all ICIs were grouped as a single drug class, due to the relatively limited number of reports for individual agents for some AEs.

The neurological irAEs examined included the following MedDRA preferred terms: “multiple sclerosis” (or “multiple sclerosis relapse”), “myasthenia gravis” (or “myasthenia gravis crisis” or “immune-mediated myasthenia gravis”), “Guillain–Barré syndrome,” and “encephalitis autoimmune.” MedDRA preferred terms are standardized terms used to group related clinical concepts to ensure consistency in AE reporting. To ensure inclusion of the most relevant reports, analyses were restricted to cases where ICIs were identified as the “primary suspect” drugs for the reported AEs. Duplicates were removed using OpenVigil 2 with verification against individual safety reports and case numbers.

Disproportionality was quantified using the reporting odds ratio (ROR) and its 95% confidence interval (CI). The ROR represents the odds of a specific AE (in this case, certain neurological irAE) occurring with the drug(s) of interest compared to the odds of the same event occurring with all other drugs in the database (Supplemental Table 1). A safety signal was defined based upon the Evans criteria,⁷ which require the following thresholds to be met: ⩾3 reports, a proportional reporting ratio (PRR) ⩾2, and a chi-squared value ⩾4. The PRR is another disproportionality metric that compares the proportion of a specific AE among all events reported for a given drug with the corresponding proportion for all other drugs in the database. Unlike the ROR, which is expressed as an odds ratio, the PRR is expressed as a risk ratio of reporting proportions (Supplemental Table 1).

Results

The distribution of reports for neurological irAEs included in the analysis is shown in Table 1. There were 48 reports of MS or MS relapse in association with ICIs. Among these MS-related reports, the average patient age was 56.1 ± 14.7 years; 24 were women, 18 were men, and 6 of unknown gender. The average patient age and gender breakdown for other irAEs were as follows: for MG, 71.8 ± 10.2 years (male: 345, female: 161, unknown: 80); for GBS, 67.8 ± 11.4 years (male: 61, female: 36, unknown: 17); and for AIE, 64.1 ± 12.7 years (male: 58, female: 39, unknown: 11). The number of reports with missing age values was as follows: 15 for MS, 145 for MG, 37 for GBS, and 24 for AIE. The ROR for MS or MS relapse was 0.09 (95% CI: 0.068–0.12). In contrast, significantly elevated RORs were observed for other autoimmune neurological AEs: 21.05 (95% CI: 19.29–22.97) for MG, 8.08 (95% CI: 6.68–9.77) for GBS, and 29.03 (95% CI: 23.56–35.76) for AIE (Figure 1). Based on Evans criteria, no significant safety signal was detected for MS or MS relapse in association with ICIs, in contrast to the strong signals observed for other autoimmune disorders. The strongest signal was observed for AIE.

Table 1.

Reports of select autoimmune neurological AEs associated with immune checkpoint inhibitors versus other drugs based on the data from the FDA Adverse Event Reporting System database.

AE of interest	Number of reports of AE of interest associated with ICIs^a	Number of reports of events other than AE of interest associated with ICIs	Number of reports of AE of interest associated with all other drugs	Number of reports of events other than AE of interest associated with all other drugs	Signal detected^b
Multiple sclerosis and/or multiple sclerosis relapse	48	96,048	69,074	12,470,945	No
Myasthenia gravis, myasthenia gravis crisis, and/or immune-mediated myasthenia gravis	586	95,510	3654	12,536,365	Yes
Guillain–Barre syndrome	113	95,983	1828	12,538,191	Yes
Autoimmune encephalitis	108	95,988	486	12,539,533	Yes

The ICIs included in this analysis were PD-1 inhibitors (pembrolizumab, nivolumab, cemiplimab, and dostarlimab), PD-L1 inhibitors (atezolizumab, durvalumab, and avelumab), and CTLA-4 inhibitors (ipilimumab and tremelimumab).

A safety signal was defined based on the Evans criteria, which require the following thresholds to be met: ⩾3 reports, a PRR ⩾2, and a chi-squared value ⩾4.

AE, adverse event; CTLA-4, cytotoxic T-lymphocyte-associated protein 4; FDA, U.S. Food and Drug Administration; ICI, immune checkpoint inhibitor; PD-L1, programmed death-ligand 1; PRR, proportional reporting ratio.

Figure 1.

Disproportionality analysis exploring the association between immune checkpoint inhibitors and autoimmune neurological adverse events based on the data from the FDA Adverse Event Reporting System database. The reporting odds ratio indicates the odds of a certain adverse event occurring with the drug(s) of interest, compared to the odds of the same adverse event occurring with all other drugs in the database. Autoimmune neurological adverse events that met Evans criteria for a safety signal are marked with an asterisk.

Discussion

Our analysis of FAERS data suggests that immune checkpoint blockade does not generate the same pharmacovigilance signal for MS as it does for other neurologic irAEs. While significant safety signals were detected for MG, GBS, and AIE, the ROR for MS showed no evidence of an increased signal. Taken together with published case series^8–11 that describe only sporadic ICI-associated MS relapses—most of which responded to high-dose steroids or ICI discontinuation—our findings reinforce the impression that checkpoint inhibition rarely precipitates fulminant MS activity in the real-world oncology setting. A recent meta-analysis showed a low relapse rate (5.45/100 patient-years), with events occurring within the first 6 months after ICI initiation.⁵ Relapses were not described as having unusual clinical features compared to typical MS relapses. Neuroradiological activity (24.9/100 patient-years) was more frequent than clinical relapses, again within the first 6 months. Most patients who developed new MRI lesions had discontinued diease modifying therapies (notably natalizumab in some cases, raising the possibility of rebound).⁵ We found no report of unusual lesion morphology or topography compared with conventional MS.

Several mechanistic explanations are plausible, although these remain hypothesis-generating rather than definitive. It is well-established that ICIs boost effector T-cell activity and enhance antibody generation by lifting inhibitory checkpoints on B-cell activation.¹ While MG, GBS, and AIE are acute, antibody-driven disorders that can flare rapidly when peripheral checkpoints are lifted, MS is a chronic, compartmentalized, T- and B-cell-mediated disease that evolves over months—often behind a partially restored blood–brain barrier. Lesion formation in MS generally requires activation of microglia,¹² the resident macrophages of the central nervous system (CNS), so a brief surge of peripheral adaptive immunity may not be enough to trigger overt clinical disease, particularly in patients whose CNS inflammation has long been quiescent. This view is reinforced by clinical data showing that neutralizing the IL-12/23 p40 subunit with ustekinumab—an intervention that blunts the same Th1/Th17 axis ICIs amplify—failed to reduce new gadolinium-enhancing lesions or clinical relapses in relapsing–remitting MS, despite clear efficacy in other antibody-mediated autoimmune diseases.¹³ Complementary work in p40-deficient mice demonstrates that, although p40 is abundantly expressed in secondary lymphoid organs and the CNS throughout experimental autoimmune encephalomyelitis, its deletion does not hinder disease propagation once encephalitogenic T cells are established,¹⁴ highlighting the relative independence of chronic CNS pathology from peripheral cytokine cues. Together, these observations suggest that checkpoint blockade preferentially unmasks conditions already primed for rapid, antibody-mediated injury—explaining the strong signals detected for MG, GBS, and AIE but not for the more slowly evolving, compartmentalized pathology of MS (Supplemental Figure 1). In addition, many cancer patients eligible for ICIs are older than the median age of MS onset; immunosenescence and cancer-associated immunosuppression could further blunt CNS-directed autoimmunity in this group. These pathophysiologic links remain speculative and warrant further study. Emerging data from other immunomodulatory approaches further underscore the complexity of immune signaling in MS. For instance, anti-CD40L therapy (Frexalimab^®) significantly reduced new gadolinium-enhancing brain lesions in a phase II trial.¹⁵ CD4-L is considered a key molecule in the CD40/CD40L costimulatory pathway, which regulates both adaptive (T and B cells) and innate (macrophages and dendritic cells) immune responses. Alternative explanations are also possible—for example, MS relapses may be under-recognized or under-reported in oncology settings, particularly when neurological symptoms are attributed to cancer progression or treatment-related toxicities.

It is important to note that the FAERS database has several limitations. Because it relies on spontaneous and voluntary reporting, the database is subject to underreporting, reporting bias, and potential misclassification. The database lacks denominator data and a control group; therefore, true incidence rates cannot be calculated. Clinical factors and drug–drug interactions cannot be adjusted for. As such, while disproportionality analyses are validated approaches for safety signal detection, the results should be interpreted as reflecting reporting patterns rather than establishing or refuting causality.¹⁶ In addition, the comparator group of “other drugs” in FAERS is highly heterogeneous, encompassing all non-ICI agents. This may influence the magnitude of the disproportionality estimates, and thus, the results should be interpreted with caution. Additionally, we were unable to perform subgroup analyses due to the small number of MS reports, which would have produced unstable estimates. As such, our findings should be viewed as overall reporting patterns rather than subgroup-specific results. Another limitation is that, due to the relatively small number of MS-related reports, we aggregated all ICIs (PD-1, PD-L1, and CTLA-4 inhibitors) into a single analytical category. While this approach was necessary to preserve statistical stability, it may have masked differential risk profiles across subclasses, particularly since CTLA-4 inhibitors are known to carry a higher risk of irAEs than PD-1/PD-L1 inhibitors. Furthermore, our use of strict MedDRA preferred terms (“multiple sclerosis” or “multiple sclerosis relapse”) may have increased diagnostic specificity but reduced sensitivity, potentially excluding some relevant cases coded under broader demyelination-related terms. Finally, because multiple neurological irAEs were analyzed, the possibility of false positives due to multiple comparisons cannot be excluded. We did not apply formal corrections, as such methods may increase the risk of false negatives in exploratory signal-detection analyses. Nonetheless, the large effect sizes observed for MG, GBS, and AIE make it unlikely that these signals are explained by random variation alone.

Prospective cohort studies and linked registry analyses are needed to validate our findings, investigate mechanisms underlying the differential safety profiles of ICIs across neurologic irAEs, and identify predictors of irAEs associated with ICIs.

Conclusion

Within the constraints of spontaneous report data, we found no evidence that ICIs substantially elevate the risk of MS onset or relapse, in contrast to the well-described association with other autoimmune neurologic syndromes. This distinction highlights the heterogeneity of checkpoint-related irAEs and underscores the need for continued surveillance and mechanistic research to elucidate the pathways that drive their variable impact on preexisting or de novo autoimmune neurological disorders. Clinically, while our findings may provide some reassurance to oncologists, careful monitoring of patients with preexisting MS or those who develop new neurologic symptoms during ICI therapy remains important, given the limitations of FAERS.

Supplemental Material

sj-docx-2-tan-10.1177_17562864251392211 – Supplemental material for Reporting patterns of multiple sclerosis with immune checkpoint inhibitors: analysis of the FDA Adverse Event Reporting System

Supplemental material, sj-docx-2-tan-10.1177_17562864251392211 for Reporting patterns of multiple sclerosis with immune checkpoint inhibitors: analysis of the FDA Adverse Event Reporting System by Afsaneh Shirani, Samuel A. Goldlust and Olaf Stuve in Therapeutic Advances in Neurological Disorders

Supplemental Material

sj-docx-3-tan-10.1177_17562864251392211 – Supplemental material for Reporting patterns of multiple sclerosis with immune checkpoint inhibitors: analysis of the FDA Adverse Event Reporting System

Supplemental material, sj-docx-3-tan-10.1177_17562864251392211 for Reporting patterns of multiple sclerosis with immune checkpoint inhibitors: analysis of the FDA Adverse Event Reporting System by Afsaneh Shirani, Samuel A. Goldlust and Olaf Stuve in Therapeutic Advances in Neurological Disorders

Supplemental Material

sj-pdf-1-tan-10.1177_17562864251392211 – Supplemental material for Reporting patterns of multiple sclerosis with immune checkpoint inhibitors: analysis of the FDA Adverse Event Reporting System

Supplemental material, sj-pdf-1-tan-10.1177_17562864251392211 for Reporting patterns of multiple sclerosis with immune checkpoint inhibitors: analysis of the FDA Adverse Event Reporting System by Afsaneh Shirani, Samuel A. Goldlust and Olaf Stuve in Therapeutic Advances in Neurological Disorders

Footnotes

Acknowledgements

The authors would like to thank the Clinical Scholars Program at Saint Luke’s Marion Bloch Neuroscience Institute for its support.

Declarations

ORCID iDs

Afsaneh Shirani

Olaf Stuve

Supplemental material

Supplemental material for this article is available online.

References

Pardoll

DM.

The blockade of immune checkpoints in cancer immunotherapy. Nat Rev Cancer 2012; 12(4): 252–264.

Topalian

Drake

Pardoll

DM.

Immune checkpoint blockade: a common denominator approach to cancer therapy. Cancer Cell 2015; 27(4): 450–461.

Postow

Sidlow

Hellmann

MD.

Immune-related adverse events associated with immune checkpoint blockade. N Engl J Med 2018; 378(2): 158–168.

Haugh

Probasco

Johnson

DB.

Neurologic complications of immune checkpoint inhibitors. Expert Opin Drug Saf 2020; 19(4): 479–488.

Gelibter

Saraceno

Susani

, et al. Do immune checkpoint inhibitors affect the course of multiple sclerosis? A systematic review and meta-analysis. Eur J Neurol 2025; 32(1): e16547.

Böhm

von Hehn

Herdegen

, et al. OpenVigil FDA—inspection of U.S. American adverse drug events pharmacovigilance data and novel clinical applications. PLoS One 2016; 11(6): e0157753.

Evans

Waller

Davis

Use of proportional reporting ratios (PRRs) for signal generation from spontaneous adverse drug reaction reports. Pharmacoepidemiol Drug Saf 2001; 10(6): 483–486.

Androdias

Noroy

Psimaras

, et al. Impact of immune checkpoint inhibitors on the course of multiple sclerosis. Neurol Neuroimmunol Neuroinflamm 2024; 11: e200202.

Nylander

Rowles

Poole

, et al. Clinical outcomes after use of checkpoint inhibitor immunotherapies in people with multiple sclerosis. Mult Scler J Exp Transl Clin 2024; 10: 20552173241252563.

10.

Conway

Pua

DKA

Holroyd

, et al. Neurologic disease activity in people with multiple sclerosis treated with immune checkpoint inhibitors. Mult Scler J 2023; 29: 471–474.

11.

Quinn

Rajarajan

Kopinsky

, et al. Multi-institutional study of neurologic outcomes in people with multiple sclerosis who are treated with immune checkpoint inhibitors for oncologic indications (PL4.004). Neurology 2024; 102: 6774.

12.

Yong

VW.

Microglia in multiple sclerosis: protectors turn destroyers. Neuron 2022; 110(21): 3534–3548.

13.

Segal

Constantinescu

Raychaudhuri

, et al. Repeated subcutaneous injections of IL-12/23 p40 neutralising antibody, ustekinumab, in patients with relapsing-remitting multiple sclerosis: a phase II, double-blind, placebo-controlled, randomised, dose-ranging study. Lancet Neurol 2008; 7(9): 796–804.

14.

Cravens

Hussain

Miller-Little

, et al. IL-12/IL-23p40 is highly expressed in secondary lymphoid organs and the CNS during all stages of EAE, but its deletion does not affect disease perpetuation. PLoS One 2016; 11(10): e0165248.

15.

Vermersch

Granziera

Mao-Draayer

, et al. Inhibition of CD40L with Frexalimab in multiple sclerosis. N Engl J Med 2024; 390(7): 589–600.

16.

Sakaeda

Tamon

Kadoyama

, et al. Data mining of the public version of the FDA Adverse Event Reporting System. Int J Med Sci 2013; 10(7): 796–803.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB

0.22 MB