Early-infantile developmental and epileptic encephalopathy: the aetiologies, phenotypic differences and outcomes-a...

In this study, we have evaluated the underlying aetiologies, yield of genetic testing and long-term outcomes in patients with early-infantile developmental and epileptic encephalopathies. We have prospectively studied patients with seizure onset before 3 months of age. Based on the clinical details, neuroimaging, metabolic testing and comprehensive genetic evaluation, patients were classified into different aetiological groups. The phenotypic differences between genetic/unknown groups and remaining aetiologies were compared. Factors that could affect seizure control were also assessed. A total of 80 children (M:F ratio-1.5:1) were recruited. The median seizure onset age was 28 days (range, 1-90 days). The aetiologies were confirmed in 66 patients (83%). The patients were further classified into four aetiological groups: genetic (50%), structural (19%), metabolic (14%; all were vitamin responsive) and unknown (17%). On comparing for the phenotypic differences between the groups, children in the 'genetic/unknown' groups were more frequently observed to have severe developmental delay (Odds Ratio = 57; P < 0.0001), autistic behaviours (Odds Ratio = 37; P < 0.0001), tone abnormalities (Odds Ratio = 9; P = 0.0006) and movement disorder (Odds Ratio = 19; P < 0.0001). Clonic seizures were more common in the vitamin responsive/structural groups (Risk Ratio = 1.36; P = 0.05) as compared to patients with 'genetic/unknown' aetiologies. On the contrary, vitamin responsive/structural aetiology patients were less likely to have tonic seizures (Risk Ratio = 0.66; P = 0.04). Metabolic testing was diagnostic in three out of 41 patients tested (all three had biotinidase deficiency). MRI was abnormal in 35/80 patients (malformation observed in 16/35; 19/35 had non-specific changes that did not contribute to underlying aetiology). A molecular diagnosis was achieved in 53 out of 77 patients tested (69%). Next-generation sequencing had a yield of 51%, while microarray had a yield of 14%. STXBP1 was the most common (five patients) single-gene defect identified. There were 24 novel variants. The mean follow-up period was 30 months (range, 4-72 months). On multivariate logistic regression for the important factors that could affect seizure control (seizure onset age, time lag of first visit to paediatric neurologist and aetiologies), only vitamin responsive aetiology had a statistically significant positive effect on seizure control (P = 0.02). Genetic aetiologies are the most common cause of early-infantile developmental and epileptic encephalopathies. Patients in the genetic/unknown groups had a more severe phenotype. Patients with vitamin responsive epilepsies had the best probability of seizure control.