IntroductionConventional genetic investigation fails to identify the F8 causal variant in 2.5%‐10% of haemophilia A (HA) patients with non‐severe phenotypes. In these cases, F8 deep intronic variants could be causal.AimTo identify pathogenic F8 deep intronic variants in genetically unresolved families with non‐severe HA analysed in the haematology laboratory of the Hospices Civils de Lyon.MethodsThe whole F8 was analysed by next generation sequencing. The pathogenic impact of candidate variants identified was assessed using both in silico analysis (MaxEntScan and spliceAI) and functional analysis (RNA or minigene assay).ResultsSequencing was performed in 49/55 families included for which a DNA sample from a male propositus was available. In total, 33 candidate variants from 43 propositi were identified. These variants corresponded to 31 single nucleotide substitutions, one 173‐bp deletion, and an 869‐bp tandem triplication. No candidate variant was found in six propositi. The most frequent variants found were the association of [c.2113+1154G>C and c.5374‐304C>T], identified in five propositi, and the c.2114‐6529C>G identified in nine propositi. Four variants had been previously described as HA‐causing. Splicing functional assay found a deleterious impact for 11 substitutions (c.671‐94G>A, c.788‐312A>G, c.2113+1154G>C, c.2114‐6529C>G, c.5999‐820A>T, c.5999‐786C>A, c.5999‐669G>T, c.5999‐669G>A, c.5999‐669G>C, c.6900+4104A>C, and c.6901‐2992A>G). The HA‐causing variant was identified in 33/49 (67%) cases. In total, F8 deep intronic variants caused 8.8% of the non‐severe HA among the 1643 families analysed in our laboratory.ConclusionThe results emphasise the value of whole F8 gene sequencing combined with splicing functional analyses to improve the diagnosis yield for non‐severe HA.