Sir—Ernest Beutler and co-workers (Jan 19, p 211) calculate that fewer than 1% of C282Y homozygotes develop frank clinical haemochromatosis, a penetrance much lower than previously thought. We believe their conclusion is flawed. The prevalence of symptomatic haemochromatosis is not the same as the penetrance of the C282Y/C282Y genotype. Penetrance is the proportion of people who have the genotype that develops the phenotype. Definition of the phenotype is central. Four stages are recognised in haemochromatosis; end-organ damage and severe fibrosis or cirrhosis define stage 4. In Beutler and co-workers’ study, only three of 119 patients had serum ferritin higher than 1000 g/L and raised aspartate aminotransferase; only one had bronzed diabetes. This number would underestimate the proportion of people with the C282Y/C282Y genotype who had stage 4 disease, and their data suggest that the proportion of patients with stage 2–3 is more than 1% (stage 1 is genetic predisposition without phenotypic abnormality). Stage 4 liver disease can be predicted by measurement of serum ferritin, aspartate aminotransferase, and presence of hepatomegaly. Beutler and coworkers misquote Guyarder’s conclusion as saying that only haemochromatosis patients with serum ferritin concentrations of more than 1000 g/L and with raised aspartate aminotransferase concentrations develop severe liver fibrosis. Guyarder and colleagues noted that 30 of 68 patients with serum ferritin higher than 1000 g/L and normal aspartate aminotransferase had severe hepatic fibrosis. No information on hepatic fibrosis is available from Beutler and co-workers’ study. However, 57% of patients had raised serum transferrin saturation, and 65% raised serum ferritin. 30% had been previously diagnosed with haemochromatosis. Most (28 of 45) of these patients were excluded from the analysis of symptoms, lowering the estimate of penetrance based on symptoms. If an agreed definition of phenotype is used, Beutler and coworkers’ study suggests a penetrance much higher than 1%, perhaps as high as 50% if based on serum transferrin saturation. We have found screening of liver clinic referrals with serum transferrin saturation to be a reliable predictor of haemochromatosis (figure). This approach identified 11 C282Y homozygous over a 3-year period, of whom nine had phenotypic haemochromatosis. On screening of people with late-onset diabetes mellitus, undiagnosed C282Y homozygosity is five times as common as in controls. All eight homozygotes detected had phenotypic haemochromatosis with cirrhosis (four with hepatocellular carcinoma). Beutler and co-workers note no loss of C282Y homozygotes from the population with increasing age, and conclude that this supports low penetrance. However, they took no account of demographic changes in southern California in the past few decades. Immigration of younger Hispanic and Asian populations, in which there is a lower prevalence of C282Y homozygosity, would have a confounding effect by diluting the prevalence in younger agegroups. Population screening for the mutation in the hereditary haemochromatosis (HFE) protein is controversial since the penetrance of the HFE mutation is unclear. We believe Beutler and co-workers underestimate the penetrance. Based on liver-biopsy findings of screendetected homozygotes, the penetrance of the HFE mutation is far higher than they suggest. *A Poullis, S J Moodie, J D Maxwell