What is the incidence and clinical significance of dry eye disease in patients treated with immune checkpoint inhibitors? A systematic review and meta-analysis of ocular immune-related adverse events.
Kai-Yang Chen, Hoi-Chun Chan, Chi-Ming Chan
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis assessed dry eye disease (DED) in adults treated with immune checkpoint inhibitors (ICIs), drawing on 13 studies. The pooled incidence was 2% (95% CI: 1%-3%), but estimates varied markedly across studies (0.2%-65%), likely reflecting differences in surveillance and diagnostic approaches. Combination CTLA-4 plus PD-1 therapy and lung cancer cohorts had numerically higher observed rates, though subgroup comparisons warrant caution. Most reported DED events were mild and treated topically; a minority required systemic immunomodulation.
Key findings
- Pooled DED incidence among ICI-treated patients was 2% (95% CI: 1%-3%), with study estimates ranging from 0.2%-65%.
- Combination CTLA-4 plus PD-1 therapy had a pooled incidence of 25%, versus 5% for CTLA-4, 13% for PD-1, and 15% for PD-L1 monotherapy; subgroup differences were not statistically significant (p = 0.18).
- Observed DED rates were 41% in lung cancer, 4% in melanoma, and 32% in renal cancer cohorts (p = 0.05), with potential influence from surveillance and referral differences.
- Abnormal Schirmer's test results were reported in 62% of symptomatic cases; most DED events were mild and managed with topical therapy.
Why it matters
DED is a potential ocular immune-related adverse event during ICI treatment, and symptoms may affect quality of life and treatment adherence. The wide variation in reported incidence means pooled estimates should not be assumed to apply uniformly across settings. Recognising the possibility of ocular surface symptoms and the role of assessment may help contextualise adverse-event reporting and support consistent monitoring.
What to keep in mind
The pooled estimate combines heterogeneous study designs and ascertainment approaches, with marked variation in reported incidence. Regimen- and tumour-specific estimates came from limited subsets and may be affected by differences in surveillance, ophthalmic access, and referral.