The aim of this study was to develop a mutation screening protocol for familial hypercholesterolaemia (FH) patients and to assess genotype/phenotype effects in terms of pre-treatment lipid profiles and presentation of tendon xanthomata (TX). A total of 158 families with clinical definitions of possible (120) or definite (38) FH were studied using a tiered screening protocol. Mutations were identified in 52 families, 44 families showing 23 different LDLR gene defects and eight families showing the common Apo B100 gene defect R3500Q. LDLR defects were detected in various regions of the gene with 56% in the LDL binding domain (exons 2-6) and 37% in the EGF precursor homology domain (exons 7-14). The most common mutations were D461N(7), C210X(5), 932delA(5), and C163Y(4). Frameshift mutations accounted for 20% with nonsense 13%, mis-sense 35%, splice 3%, Apo B 13% and 2% large deletion, 13% of clinically definite FH remained undefined. In conclusion, DNA based diagnosis is possible in 79% (30/38) of clinically definite FH families and of the 120 possible FH families at the start of the screening program, 18% (22/120) now have defined mutations. Overall 60 families from the original 158 meet the clinical and/or genetic criteria for definite FH. Tendon xanthomata were present in only 58% (30/52) of genetically defined FH families, thus limiting its use as a strict diagnostic criteria. Families with low density lipoprotein receptor (LDLR) defects present with higher total and LDL cholesterol levels and a higher incidence of TX than do those with the common Apo B variant, and frameshift mutations appear to have the most severe presentation.