Neurotoxicity associated with immune checkpoint inhibitors: Systematic review and meta-analysis.

e15096 Background: Immune checkpoint inhibitors (ICI) are associated with multiple immune related adverse events (irAE). Moderate to severe neurotoxicity is reported in less than 1% of all patients treated with ICI. We performed a systematic review and meta-analysis to assess neurotoxicity associated with anti-PD-1/PD-L1 and anti-CTLA4 therapy. Methods: The Embase, Ovid, Pubmed and Scopus database were comprehensively searched by two independent reviewers, from inception to 2019, to include Phase II and III clinical trials reporting neurotoxicity with the combination of, or monotherapy with anti-PD-1/PD-L1 and/or anti-CTLA4 therapies. Our primary outcome was assessment of neurologic irAE of all grades. We divided neurotoxicity into two groups: peripheral nervous system (PNS) including peripheral neuropathy (PeN), fatigue, Guillain-Barre syndrome (GBS), myelitis; and central nervous system (CNS) including stroke, myasthenia gravis (MG), encephalopathy/encephalitis, altered mental status (AMS), decreased appetite and headache. Statistical heterogeneity was quantified using I2 statistics and publication bias was assessed with Eggers regression test. The estimates were reported as odds Ratio (OR) with 95% confidence intervals (CI) using random effect model. Results: A total of 2,876 full text articles retrieved in the initial database search were analyzed according to PRISMA guidelines. The final analysis included 39 trials, evaluating 10,595 patients in the control arm and 13,110 patients in the immunotherapy arm. All-grade neurotoxicity events with ICI were significantly lower compared to placebo or standard of care (OR 0.55, 95% CI 0.38-0.81, I2 96.82%, P < 0.001). Only significant CNS irAE observed was fatigue (OR 0.82, CI 0.71-0.96, I2 81.75%, P < 0.001). Other CNS irAE including AMS (OR 1.318, 95% CI 0.86-2.0, I2 0%, P = 0.442), encephalopathy/encephalitis (OR 1.36, 95% CI 0.38-4.96, I2 0%, P = 0.82) and seizures (OR 1.96, 95% CI 0.64-6.02, I2 0%, P = 0.56) had statistically non-significant increased risk in ICI group. Only one case of myelitis was reported in placebo group. The ICI arm had statistically significant lower rates of PeN (OR 0.28, CI 0.16-0.49, I2 87.31, P < 0.001). One trial reported 1 patient with MG and 5 trials reported 1 patient each of GBS in the ICI arm. Conclusions: Our meta-analysis concludes that use of ICI is associated with lower rate of neurotoxicity overall. Fatigue and PeN are lower with ICI compared to standard of care or placebo. Although, CNS irAE are increased with ICI, they do not reach statistical significance.