Autosomal dominant type IIa hypercholesterolaemia (ADH) is characterised by an elevation of total plasma cholesterol associated with increased LDL particles. Numerous different molecular defects have been identified in the LDL receptor (LDLR) and few specific mutations in the apolipoprotein B (APOB) gene resulting in familial hypercholesterolaemia and familial defective apoB-100 respectively. To estimate the respective contribution of LDLR, APOB and other gene defects in this disease, we studied 33 well characterised French families diagnosed over at least three generations with ADH through the candidate gene approach. An estimation of the proportions performed with the HOMOG3R program showed that an LDLR gene defect was involved in approximately 50% of the families (P = 0.001). On the other hand, the estimated contribution of an APOB gene defect was only 15%. This low estimation of ADH due to an APOB gene defect is further strengthened by the existence of only two probands carrying the APOB (R3500Q) mutation in the sample. More importantly and surprisingly, 35% of the families in the sample were estimated to be linked to neither LDLR nor APOB genes. These data were confirmed by the exclusion of both genes through direct haplotyping in three families. Our results demonstrate that the relative contributions of LDLR and APOB gene defects to the disease are very different. Furthermore, our results also show that genetic heterogeneity is, generally, underestimated in ADH, and that at least three major groups of defects are involved. At this point, the contribution of the recently mapped FH3 gene to ADH cannot be assessed nor its importance in the group of 'non LDLR/non APOB' families.