Background: Immune checkpoint inhibitors (ICIs) have become a standard of care in relapsed or refractory Hodgkin lymphoma (HL), offering improved survival and durable responses. However, pivotal clinical trials have historically excluded patients with pre-existing autoimmune diseases (AI) due to concerns over heightened immune-related toxicity, resulting in a critical evidence gap regarding real-world safety and efficacy in this population. As ICIs are increasingly used in broader clinical practice, it is essential to understand the risks and outcomes for HL patients with underlying autoimmune conditions.

Methods: We conducted a retrospective cohort study using the TriNetX Global Collaborative Network. Adult patients (≥18 years) with HL who received at least one dose of nivolumab or pembrolizumab were identified and stratified by presence (HL + AI + ICI; n=299) or absence (HL + ICI w/o AI; n=299) of autoimmune diagnoses prior to ICI initiation. Propensity score matching was performed on demographics, comorbidities, and prior HL therapies. The median follow-up after matching was 479 days for the AI cohort and 572 days for the non-AI cohort. Outcomes were compared using Kaplan-Meier analysis, reporting hazard ratios (HRs) and log-rank p-values.

Results: Over the follow-up period, patients with pre-existing autoimmune disease receiving ICIs experienced a significantly higher risk of septic shock (HR 2.29, 95% CI 1.12–4.68, p=0.019) compared to those without autoimmune disease, while the risk of pneumonia was similar between groups (HR 1.10, p=0.665). The occurrence of marked systemic inflammation, as defined by CRP ≥30 mg/L, did not differ significantly (HR 1.12, p=0.707). All-cause mortality was modestly higher in the AI cohort (HR 1.16, p=0.351), though not statistically significant. Incidence of diarrhea or colitis was numerically lower in the AI group (HR 0.73, p=0.167), whereas hospitalization risk was similar (HR 1.11, p=0.720). Outpatient visits were rare and not analyzable due to low event counts.

Rates of peripheral neuropathy trended higher in patients with autoimmune disease (HR 2.70, p=0.06), though not reaching conventional significance. There were no significant differences in initiation of systemic corticosteroids (HR 0.60, p=0.328) or elevated liver enzymes (HR 1.37, p=0.226) between groups. Adrenal insufficiency and hypophysitis (10 cases, 0 in control) occurred exclusively in the autoimmune cohort, with the latter reaching statistical significance (log-rank p=0.031). Hypothyroidism (HR 1.16, p=0.583) and hyperthyroidism (HR 1.31, p=0.687) were comparable across groups. Risks of pneumonitis (HR 1.56, p=0.446), hepatitis (HR 1.60, p=0.418), and myositis (HR 4.35, p=0.151) were not significantly different. Guillain-Barré syndrome (HR 1.05, p=0.974) and hemolytic anemia (HR 1.57, p=0.618) were infrequent, with no significant differences observed.

A composite immune-related adverse event endpoint showed a numerically higher but statistically non-significant risk in the AI group (HR 1.34, p=0.243). Incidence of nephritis, myocarditis, pericarditis, and encephalitis was extremely low and did not differ meaningfully between cohorts. For several rare outcomes, such as nephritis and encephalitis, no events occurred in one or both cohorts, precluding meaningful hazard ratio estimation.

Conclusions: In this real-world, propensity-matched study, patients with pre-existing autoimmune disease who received ICIs for Hodgkin lymphoma had a higher risk of septic shock and hypophysitis, with trends toward increased peripheral neuropathy and all-cause mortality, but broadly similar risks for most other infectious, immune, and healthcare utilization outcomes compared to those without autoimmune disease. These data suggest that, with careful monitoring, ICIs can be considered in selected HL patients with controlled autoimmune disease, though clinicians should remain vigilant for potentially severe immune and infectious complications unique to this population. Further research is warranted to refine risk stratification and management strategies for this growing and vulnerable patient subgroup.

This content is only available as a PDF.
2025
Sign in via your Institution