Distinctive seizure signature in the first video case-control study of a naturally-occurring feline autoimmune encephalitis model.

Background and objectiveAutoimmune encephalitis (AE) is a form of brain inflammation where pathogenic autoantibodies bind surface proteins. In humans, AE is at least as common as infective encephalitis, and seizures are a prominent manifestation. The most common adult human AE is associated with antibodies to leucine-rich glioma-inactivated 1 (LGI1-Ab-E). AE in non-human mammals is also recognised, notably the polar bear 'Knut', diagnosed with N-methyl D-aspartate receptor antibody encephalitis. LGI1-Ab-E is an emerging cause of spontaneously-arising AE in domestic cats. Our objective was to phenotype the seizure profile of feline LGI1-Ab-E and probe parallels to its human counterpart.MethodsWe characterised seizures in naturally-occurring feline LGI1-Ab-E. Three veterinary and two human neurologists independently blind-rated 35 LGI1-antibody positive and negative feline seizure videos from 24 cats (16 LGI1-Ab-E positive, 8 negative). Data analysed included seizure frequency, semiologies and their co-occurrence, localisation, inter-rater agreement, and predictive factors.ResultsThe mean number of daily seizures at peak was significantly higher in LGI1-antibody positive compared to LGI1-antibody-negative cats (12.6 vs. 1.9/day, pcorr = 0.011). Semiologies statistically significantly enriched in LGI1-Ab-E observations included orofacial automatisms (88/120, 73 % vs. 26/55, 47 %, pcorr = 0.024), salivation (87/120, 73 % vs. 23/55, 42 %, pcorr = 0.004); and mydriasis (79/120, 66 % vs 19/55, 35 %, pcorr = 0.004), and almost exclusively seen in LGI1-Ab-E were circling (39/120, 33 % vs. 1/55, 2 %, pcorr=<0.001) and aggression (14/120, 12 % vs. 0/55, 0 %, non significant after correction). A temporal lobe onset was proposed in 67 % (80/120) of seropositive ratings, compared to 28 % (15/55) LGI1-Ab-E negative (p SignificanceFeline LGI1-Ab-E represents a clinically distinctive seizure disorder. Our findings highlight the value of studying naturally-occurring, biologically representative animal models which closely mimic human diseases. This bidirectional translational approach confers benefits across species and unites human and veterinary neurology.