Cancer specific risk and mortality due to autoimmune hemolytic anemia (AIHA) associated with immune checkpoint inhibitors (ICI) Free

Background

Immune checkpoint inhibitors (ICI) are cornerstone of treatment of majority of the solid cancers with increasingly being used in first line settings. They can cause significant immune mediated adverse events, one of them being autoimmune hemolytic anemia (AIHA). In this study we evaluate the risk of AIHA in specific cancer types and then impact of AIHA on mortality in those specific cancers.

Methods

To identify patients at risk for ICI-induced AIHA, we utilized the TriNetX Research Network to include patients ≥18 from 2012–2024 who received ICIs. Patients were limited to those with cancer types for which ICI use is indicated. Those with lymphoid or hematopoietic cancers were excluded. AIHA was confirmed either through ICD-10 code or a combination of haptoglobin ≤40 mg/dL with a positive direct antiglobulin test (DAT). Patients were then stratified by cancer type or ICI treatment type.

To assess the absolute risk increase (ARI) of AIHA from ICI use, we calculated the proportion of patients who developed AIHA among those treated with ICIs and compared this to the proportion in patients not receiving ICIs. The ARI was defined as the difference in AIHA incidence between ICI-treated and untreated groups. Statistical significance was assessed using a chi-square test of independence. Survival following ICI treatment was evaluated using Kaplan-Meier analysis at 30 days, 90 days, and 5 years. Differences in survival were compared using log-rank tests.

Results

Among 344 patients with solid tumors, AIHA was significantly more common in those treated with immune checkpoint inhibitors (ICI) compared to non-ICI patients (0.98% vs. 0.24%; ARI: 0.74%; p < 0.0001). The increased risk was most notable in melanoma (0.36% vs. 0.13%; ARI 0.24%; p < 0.0001) and with PD-1 inhibitors (0.29% vs. 0.24%; p = 0.0079), while PD-L1 inhibitors did not increase AIHA risk (0.25% vs. 0.25%; p = 1.0). Given the elevated risk observed in melanoma, we also evaluated patients treated with the combination of nivolumab and ipilimumab, a common regimen for this population. The risk of AIHA in this group was 0.46% compared to 0.25% in non-ICI patients, corresponding to an ARI of 0.21% (p < 0.0001).

Across all cancers, patients with AIHA had significantly higher 30-day mortality (15% vs. 4%; p < 0.0001) and lower 30-day survival (85% vs. 96%; HR 4.2 [95% CI: 2.3–7.5]). Lung cancer patients with AIHA had higher 30-day mortality (17% vs. 10%; p = 0.1865), and markedly lower survival (83% vs. 98%; HR 8.9 [2.0–39]; p = 0.0004). GI cancers showed similar trends (25% vs. 17% mortality; survival 75% vs. 93%; HR 4.2 [1.4–13]; p = 0.0058). GU and melanoma groups did not show statistically significant differences in 30-day mortality, though GU patients had decreased survival (86% vs. 96%; HR 4.2 [0.9–20]; p = 0.0462).

At 60 days, AIHA was associated with significantly higher mortality across all cancers (23% vs. 7.8%; p < 0.0001) and reduced survival (76% vs. 92%; HR 3.4 [2.2–5.2]). Lung cancer patients had 25% vs. 10% mortality (p = 0.0067) and worse survival (75% vs. 93%; HR 4.0 [1.7–9.2]; p = 0.0005). GI cancer patients had 34% vs. 19% mortality (p = 0.0697) and lower survival (65% vs. 81%; HR 2.2 [1.1–4.7]; p = 0.0311). GU cancers showed no significant mortality difference (22% vs. 17%; p = 0.4848), but survival was worse in AIHA (77% vs. 91%; HR 2.9 [1.02–8.0]; p = 0.0369). Melanoma showed no survival or mortality difference (HR 1.0; p = 0.984).

At 5 years, AIHA was associated with higher mortality in all cancers (60% vs. 46%; p = 0.0005) and lower survival (27% vs. 40%; HR 1.7 [1.3–2.0]; p < 0.0001). GU cancers demonstrated the largest 5-year survival gap (mortality 64% vs. 35%, p = 0.0016; survival 18% vs. 53%; HR 2.6 [1.5–4.4]; p = 0.0005). Lung cancer showed worse survival (22% vs. 35%; HR 1.6 [1.1–2.3]; p = 0.0162), although mortality difference did not reach significance (65% vs. 53%; p = 0.1031). GI and melanoma cancers showed no significant differences in long-term survival or mortality.

Conclusion

Despite its rarity, ICI associated AIHA is significantly more common in melanoma patients. The development of AIHA is associated with markedly increased short term and long-term mortality across multiple cancer subtypes with most pronounced impact on GU and lung cancers. The findings highlight the need for increased vigilance, early recognition and proactive management of AIHA in patients receiving ICI therapy.