A clinical and electrophysiological study was performed in 119 Type 1A Charcot-Marie-Tooth disease (CMT1A) patients with proven 17p11.2 duplication. Onset of the first functional manifestations was in the first decade in 50% of cases and before the age of 20 years in 70% of cases. The predominant clinical signs were muscle weakness and wasting in the lower limbs. None of the patients was normal on clinical examination and all presented at least pes cavus or ankle jerk areflexia. Motor nerve conduction velocity (MNCV) was uniformly reduced in all nerves, and was < or = 33 m/s in the median nerve for all patients. Sensory potentials were abnormal in all cases, even where there was no clinical sensory loss. Needle electromyography recruitment was reduced in distal muscles for all patients. MNCV slowing was fully consistent with the presence of duplication even in clinically asymptomatic individuals or in children, confirming the complete electrophysiological penetrance of 17p11.2 duplication and making median nerve MNCV a reliable tool for screening affected at-risk individuals. Functional disability was mild. Ninety-six percent of patients were autonomous; 25% were asymptomatic and diagnosed by systematic family investigation especially on the basis of median nerve MNCV reduction. Early age at onset and greatly reduced median nerve MNCV were predictive of a more severe disease course; the earlier the onset the more reduced the median nerve MNCV and the higher the functional disability tended to be after an equivalent disease duration. Cross-sectional analysis of neurological deficit, functional deficit and MNCV according to disease duration showed that, regardless of age at onset, CMT1A disease with 17p11.2 duplication is a clinically progressive disorder. Neurological deficit and functional disability increased, whereas median nerve MNCV and compound muscle action potential (CMAP) amplitude did not change with disease course. Intrafamilial phenotype variation between parents and children and between siblings was studied in large families. Functional disability and neurological deficit differed widely and the highest range of median nerve MNCV within a family reached 23 m/s. Clinical and electrophysiological data were compared with those of CMT1B patients with peripheral myelin P0 protein point mutation. CMT1A patients were found to be more severely affected with more prolonged distal motor latency and more reduced CMAP amplitude, whereas MNCV did not significantly differ, indicating that peripheral myelin P0 protein point mutation is not always associated with a severe phenotype. The same genetic defect (17p11.2 duplication) results in variable expression within the phenotype, even in siblings with variations in age at onset, clinical severity and MNCV slowing. This phenotypic variation could be due to additional genetic factors related to peripheral myelin protein 22 expression as well as to other endogenous or environmental factors.