BackgroundLafora disease (LD) is a fatal form of progressive myoclonic epilepsy caused by biallelic pathogenic variants in EPM2A or NHLRC1. With a few exceptions, the influence of genetic factors on disease progression has yet to be confirmed. We present a systematic review and meta-analysis of the known pathogenic variants to identify genotype–phenotype correlations.MethodsWe collected all reported cases with genetically-confirmed LD containing data on disease history. Pathogenic variants were classified into missense (MS) and protein-truncating (PT). Three genotype classes were defined according to the combination of the variants: MS/MS, MS/PT, and PT/PT. Time-to-event analysis was performed to evaluate survival and loss of autonomy.Results250 cases described in 70 articles were included. The mutated gene was NHLRC1 in 56% and EPM2A in 44% of cases. 114 pathogenic variants (67 EPM2A; 47 NHLRC1) were identified. The NHLRC1 genotype PT/PT was associated with shorter survival [HR 2.88; 95% CI 1.23–6.78] and a trend of higher probability of loss of autonomy [HR 2.03, 95% CI 0.75–5.56] at the multivariable Cox regression analysis. The population carrying the homozygous p.Asp146Asn variant of NHLRC1 genotype was confirmed to have a more favourable prognosis in terms of disease duration.ConclusionsThis study demonstrates the existence of prognostic genetic factors in LD, namely the genotype defined according to the functional impact of the pathogenic variants. Although the reasons why NHLRC1 genotype PT/PT is associated with a poorer prognosis have yet to be fully elucidated, it may be speculated that malin plays a pivotal role in LD pathogenesis.