- Email: [email protected]
Is there a threshold of hepatic iron concentration that leads to cirrhosis in C282Y hemochromatosis?
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2001 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 96, No. 2, 2001 ISSN 0002-9270/01/$20.00 PII S0002-9270(00)02265-6
Is There a Threshold of Hepatic Iron Concentration That Leads to Cirrhosis in C282Y Hemochromatosis? Paul C. Adams, M.D. London Health Sciences Centre, University of Western Ontario, London, Ontario, Canada
OBJECTIVE: The relationship between the degree of iron overload and the presence of cirrhosis has not been clearly established in hemochromatosis. Severe iron overload occurs without cirrhosis and moderate iron overload can occur with cirrhosis. Previous estimates may have overestimated the problem because of the overdiagnosis of hemochromatosis in patients with alcoholic liver disease and chronic viral hepatitis. The objective of this study was to determine if a threshold for hepatic iron concentration leads to the development of cirrhosis in C282Y hemochromatosis. METHODS: This study included only hemochromatosis patients who were homozygotes for the C282Y mutation of the HFE gene and had undergone liver biopsy with hepatic iron concentration. Analysis of the thresholds for cirrhosis were determined using receiver operating characteristic (ROC) curve analysis. RESULTS: Data were available on 100 C282Y homozygotes (62 men, 38 women; mean age 51, range 18 –74 yr). ROC curve analysis showed an area under the curve for hepatic iron concentration versus cirrhosis of 0.85 (95% CI ⫽ 0.75– 0.96). The threshold for the prediction of cirrhosis was 283 mol/g. At that threshold, the sensitivity was 85% and specificity 84%. CONCLUSIONS: From this analysis, it appears that a hepatic iron concentration ⬎283 mol/g is associated with cirrhosis. However, the low sensitivity of this threshold suggests that other cofactors contribute to the development of cirrhosis in hemochromatosis. Early diagnosis is encouraged to initiate iron depletion before the development of cirrhosis. (Am J Gastroenterol 2001;96:567–569. © 2001 by Am. Coll. of Gastroenterology)
holic siderosis and chronic viral hepatitis. The discovery of the hemochromatosis gene (HFE) in 1996 (5) has allowed for a genotypic definition of hemochromatosis, which is independent of the degree of iron overload (Homozygotes for the C282Y mutation of the HFE gene). The goal of this study was to determine whether there is a threshold of iron overload that leads to cirrhosis in C282Y homozygotes for hemochromatosis.
MATERIALS AND METHODS This study included only hemochromatosis patients who were homozygotes for the C282Y mutation of the HFE gene and had undergone liver biopsy with hepatic iron concentration. This study excluded C282Y homozygotes with concomitant alcohol abuse (n ⫽ 5, ⬎80 g ethanol/day men, ⬎60 g ethanol/day women) and chronic viral hepatitis (n ⫽ 1). There were no cases diagnosed with concomitant nonalcoholic steatohepatitis (NASH). Hepatic iron concentration was determined by atomic iron absorption spectrophotometry (6). The hepatic iron concentration was determined from liver tissue removed from the paraffin block. Before analysis, the sample had been examined to be sure that it was representative liver tissue with uniform iron distribution. Analysis of the thresholds of hepatic iron concentration for cirrhosis were determined using receiver operating characteristic (ROC) curve analysis on software developed at this medical center (courtesy of Tom Pellar) (7). The advantage of ROC curve analysis over discriminant analysis of thresholds is that it assesses the operating characteristics of the test (hepatic iron) over the entire range of levels rather than at an individual threshold (8). Cirrhotics and noncirrhotics were compared using the Student’s t test.
INTRODUCTION Although iron overload is a classical feature of hemochromatosis, it has been difficult to demonstrate that excess iron directly leads to cirrhosis of the liver. Experimental iron overload in animals has been inconsistent in producing cirrhosis. Iron overload has been demonstrated to lead to increased collagen gene expression, prolyl hydroxylase activity, and lipid peroxidation products (1– 4). Previous studies on iron overload in hemochromatosis have included patients with other risk factors, such as alco-
RESULTS Data were available on 100 C282Y homozygotes (62 men, 38 women; mean age 51, range 18 –74 yr). A comparison of cirrhotics and noncirrhotics is shown in Table 1. Hepatic iron concentration was significantly higher in cirrhotics compared to noncirrhotics (p ⬍ 0.0001) (Fig. 1). ROC curve analysis showed an area under the curve for hepatic iron concentration versus cirrhosis (yes/no) of 0.85 (95% CI ⫽ 0.75– 0.96) (Fig. 2).
568
Adams
AJG – Vol. 96, No. 2, 2001
Table 1. Comparison of Cirrhotic and Noncirrhotic C282Y Homozygotes Cirrhosis (n ⫽ 22) Age Gender Hepatic iron Ferritin Transferrin saturation AST ALT
Noncirrhotic (n ⫽ 78)
p
57 (34–74) 20 men, 2 women 448 mol/g (193) 2692 g/L (1350) 82% (16)
49 (18–72) NS 42 men, 36 women 202 mol/g (115) ⬍0.0001 1116 g/L (1008) ⬍0.0001 75% (17) 0.09
89 IU/L (108) 104 IU/L (167)
56 IU/L (64) 45 IU/L (59)
0.07 0.01
Data reported as mean (standard deviation). Reference range hepatic iron concentration: 0 –35 [mol/g dry weight, 0 –1935 (g/g)]. Reference ranges: serum ferritin [15–200 (g/L women); 30 –300 (g/L men); transferrin saturation (20 –55%); AST, ALT ⫽ 0 – 40 IU/L].
The threshold for the prediction of cirrhosis was 283 mol/g (15,803 g/g). At that threshold, the sensitivity was 85%, specificity 84%, positive likelihood ratio 5.2, and negative likelihood ratio 0.18. Hepatic iron concentration was significantly greater in men (277 ⫾ 186 mol/g, mean ⫾ SD) compared to women (176 ⫾ 103 mol/g, p ⬍ 0.001).
DISCUSSION From this analysis, it appears that a hepatic iron concentration of greater than 283 mol/g is associated with cirrhosis. This is approximately eight times the upper limit of the reference range for hepatic iron concentration. It is a much higher iron concentration than has been implicated as a cofactor in the production of hepatic fibrosis in hepatitis C (9) or NASH (10 –12). Previous studies in hemochromatosis (before the use of genetic testing) suggested that the threshold for cirrhosis was 400 mol/g in an Australian study (6). Studies from France have studied hepatic iron concentrations as compared to fibrosis stages (0 to 4) and have not
Figure 1. Hepatic iron concentration was significantly increased in cirrhotics (n ⫽ 22) compared to noncirrhotic (n ⫽ 78) C282Y homozygotes (p ⬍ 0.0001). Solid bars represent the mean hepatic iron concentration. The dotted line represents the upper limit of normal of the reference range for hepatic iron concentration (35 mol/g dry weight).
Figure 2. Receiver operating characteristic (ROC) curve analysis of hepatic iron concentration versus presence of cirrhosis in 100 C282Y homozygotes. Area under the curve was 0.85 (95% confidence interval ⫽ 0.75– 0.96). The threshold for the prediction of cirrhosis was 283 mol/g.
been able to demonstrate a discrete threshold. The mean hepatic iron concentration in their cirrhotic patients was 371 mol/g ⫾ 138 (n ⫽ 28) (13, 14). Our study also does not demonstrate a discrete threshold, as can be seen in Figure 1. The ROC analysis always generates an optimal threshold based on (sensitivity/1 ⫺ specificity). If there is an excellent threshold, the area under the ROC curve is usually ⬎0.95, which was not seen in this analysis. The observation that there are many cirrhotics with iron concentrations ⬍283 mol/g and noncirrhotics with iron concentrations ⬎283 mol/g suggests that other factors contribute to the development of cirrhosis in hemochromatosis. The lower threshold in our study may be related to the inclusion of younger patients and earlier diagnosis. The ROC curve methodology always generates a threshold, so the inclusion of milder cases will decrease the threshold. The area under the curve is an indicator of the power of the threshold, which was not evaluated by the same techniques in the other studies. Previous studies have suggested that alcoholism, age, and sex are cofactors (13), and in this study, cirrhosis was most common in older men. Alcoholism and viral hepatitis have been excluded at entry in this study. The low prevalence of alcohol abuse in this study is consistent with our previous pedigree studies (15), and older studies have likely included alcoholic siderosis patients as having hemochromatosis. Genetic susceptibility to cirrhosis is another possibility, as has been postulated in alcoholic liver disease. Early diagnosis and iron depletion by venesection has been demonstrated to prevent the development of cirrhosis in hemochromatosis (16, 17). Iron depletion in cirrhotics can arrest further progression, which may explain the very low rate of liver transplantation reported in C282Y homozygotes. Hepatocellular carcinoma can develop in the cirrhotic liver, despite iron depletion. Cirrhosis remains a preventable life-threatening complication in a common genetic disease. Early diagnosis remains the key to improving long-term survival,
AJG – February, 2001
and this will require more aggressive case detection, including population screening (18).
ACKNOWLEDGMENTS The author acknowledges the support of Tom Pellar, who developed the computer software, and the late Leslie Valberg, who contributed many of the patients to this study. Reprint requests and correspondence: Paul C. Adams, M.D., Department of Medicine, London Health Sciences Centre—University Campus, 339 Windermere Rd., London, Ontario, Canada N6A 5A5. Received Apr. 6, 2000; accepted Aug. 25, 2000
REFERENCES 1. Park CH, Bacon BR, Brittenham GM, et al. Pathology of dietary carbonyl iron overload in rats. Lab Invest 1987;57: 555– 63. 2. Weintraub LR, Goral A, Grasso J, et al. Pathogenesis of hepatic fibrosis in experimental iron overload. Br J Haematol 1985;59:321–31. 3. Gualdi R, Casalgrandi G, Montosi G, et al. Excess iron into hepatocytes is required for activation of collagen type I gene during experimental siderosis. Gastroenterology 1994;107: 1118 –24. 4. Pietrangelo A, Gualdi R, Casalgrandi G, et al. Enhanced hepatic collagen type I mRNA expression into fat-storing cells in a rodent model of hemochromatosis. Hepatology 1994;19: 714 –21. 5. Feder JN, Gnirke A, Thomas W, et al. A novel MHC class I-like gene is mutated in patients with hereditary hemochromatosis. Nat Genet 1996;13:399 – 408. 6. Bassett ML, Halliday JW, Powell LW. Value of hepatic iron measurements in early hemochromatosis and determination of the critical iron level associated with fibrosis. Hepatology 1986;6:24 –9.
Hepatic Iron Concentration and Hemochromatosis
569
7. Pellar TG, Leung FY, Henderson AR. A computer program for rapid generation of receiver operating characteristic curves and likelihood ratios in the evaluation of diagnostic tests. Ann Clin Biochem 1988;25:411– 6. 8. Henderson AR. Assessing test accuracy and its clinical consequences: A primer for receiver operating characteristic curve analysis. Ann Clin Biochem 1993;30:521–39. 9. Bonkovsky H, Banner B, Rothman A. Iron and chronic viral hepatitis. Hepatology 1998;25:759 – 68. 10. Bonkovsky HL, Jawaid Q, Tortorelli K, et al. Non-alcoholic steatohepatitis, and iron. Increased prevalence of mutations of the HFE gene in non-alcoholic steatohepatitis. J Hepatol 1999; 31:421–9. 11. George DK, Goldwurm S, MacDonald GA, et al. Increased hepatic iron concentration in nonalcoholic steatohepatitis is associated with increased fibrosis. Gastroenterology 1998;114: 311– 8. 12. George DK, Powell LW, Losowsky MS. The haemochromatosis gene: A co-factor for chronic liver diseases? J Gastroenterol Hepatol 1999;14:745–9. 13. Deugnier YM, Loreal O, Turlin B, et al. Liver pathology in genetic hemochromatosis: A review of 135 homozygous cases and their bioclinical correlations. Gastroenterology 1992;102: 2050 –9. 14. Loreal O, Deugnier Y, Moirand R, et al. Liver fibrosis in genetic hemochromatosis — respective roles of iron and noniron-related factors in 127 homozygous patients. J Hepatol 1992;16:122–7. 15. Adams PC, Agnew S. Alcoholism in hereditary hemochromatosis revisited: Prevalence and clinical consequences among homozygous siblings. Hepatology 1996;23:724 –7. 16. Niederau C, Fischer R, Purschel A, et al. Long-term survival in patients with hereditary hemochromatosis. Gastroenterology 1996;110:1107–19. 17. Adams PC, Speechley M, Kertesz AE. Long-term survival analysis in hereditary hemochromatosis. Gastroenterology 1991;101:368 –72. 18. Adams PC. Population screening for haemochromatosis. Gut 2000;46:301–3.
Vol. 96, No. 2, 2001 ISSN 0002-9270/01/$20.00 PII S0002-9270(00)02265-6
Is There a Threshold of Hepatic Iron Concentration That Leads to Cirrhosis in C282Y Hemochromatosis? Paul C. Adams, M.D. London Health Sciences Centre, University of Western Ontario, London, Ontario, Canada
OBJECTIVE: The relationship between the degree of iron overload and the presence of cirrhosis has not been clearly established in hemochromatosis. Severe iron overload occurs without cirrhosis and moderate iron overload can occur with cirrhosis. Previous estimates may have overestimated the problem because of the overdiagnosis of hemochromatosis in patients with alcoholic liver disease and chronic viral hepatitis. The objective of this study was to determine if a threshold for hepatic iron concentration leads to the development of cirrhosis in C282Y hemochromatosis. METHODS: This study included only hemochromatosis patients who were homozygotes for the C282Y mutation of the HFE gene and had undergone liver biopsy with hepatic iron concentration. Analysis of the thresholds for cirrhosis were determined using receiver operating characteristic (ROC) curve analysis. RESULTS: Data were available on 100 C282Y homozygotes (62 men, 38 women; mean age 51, range 18 –74 yr). ROC curve analysis showed an area under the curve for hepatic iron concentration versus cirrhosis of 0.85 (95% CI ⫽ 0.75– 0.96). The threshold for the prediction of cirrhosis was 283 mol/g. At that threshold, the sensitivity was 85% and specificity 84%. CONCLUSIONS: From this analysis, it appears that a hepatic iron concentration ⬎283 mol/g is associated with cirrhosis. However, the low sensitivity of this threshold suggests that other cofactors contribute to the development of cirrhosis in hemochromatosis. Early diagnosis is encouraged to initiate iron depletion before the development of cirrhosis. (Am J Gastroenterol 2001;96:567–569. © 2001 by Am. Coll. of Gastroenterology)
holic siderosis and chronic viral hepatitis. The discovery of the hemochromatosis gene (HFE) in 1996 (5) has allowed for a genotypic definition of hemochromatosis, which is independent of the degree of iron overload (Homozygotes for the C282Y mutation of the HFE gene). The goal of this study was to determine whether there is a threshold of iron overload that leads to cirrhosis in C282Y homozygotes for hemochromatosis.
MATERIALS AND METHODS This study included only hemochromatosis patients who were homozygotes for the C282Y mutation of the HFE gene and had undergone liver biopsy with hepatic iron concentration. This study excluded C282Y homozygotes with concomitant alcohol abuse (n ⫽ 5, ⬎80 g ethanol/day men, ⬎60 g ethanol/day women) and chronic viral hepatitis (n ⫽ 1). There were no cases diagnosed with concomitant nonalcoholic steatohepatitis (NASH). Hepatic iron concentration was determined by atomic iron absorption spectrophotometry (6). The hepatic iron concentration was determined from liver tissue removed from the paraffin block. Before analysis, the sample had been examined to be sure that it was representative liver tissue with uniform iron distribution. Analysis of the thresholds of hepatic iron concentration for cirrhosis were determined using receiver operating characteristic (ROC) curve analysis on software developed at this medical center (courtesy of Tom Pellar) (7). The advantage of ROC curve analysis over discriminant analysis of thresholds is that it assesses the operating characteristics of the test (hepatic iron) over the entire range of levels rather than at an individual threshold (8). Cirrhotics and noncirrhotics were compared using the Student’s t test.
INTRODUCTION Although iron overload is a classical feature of hemochromatosis, it has been difficult to demonstrate that excess iron directly leads to cirrhosis of the liver. Experimental iron overload in animals has been inconsistent in producing cirrhosis. Iron overload has been demonstrated to lead to increased collagen gene expression, prolyl hydroxylase activity, and lipid peroxidation products (1– 4). Previous studies on iron overload in hemochromatosis have included patients with other risk factors, such as alco-
RESULTS Data were available on 100 C282Y homozygotes (62 men, 38 women; mean age 51, range 18 –74 yr). A comparison of cirrhotics and noncirrhotics is shown in Table 1. Hepatic iron concentration was significantly higher in cirrhotics compared to noncirrhotics (p ⬍ 0.0001) (Fig. 1). ROC curve analysis showed an area under the curve for hepatic iron concentration versus cirrhosis (yes/no) of 0.85 (95% CI ⫽ 0.75– 0.96) (Fig. 2).
568
Adams
AJG – Vol. 96, No. 2, 2001
Table 1. Comparison of Cirrhotic and Noncirrhotic C282Y Homozygotes Cirrhosis (n ⫽ 22) Age Gender Hepatic iron Ferritin Transferrin saturation AST ALT
Noncirrhotic (n ⫽ 78)
p
57 (34–74) 20 men, 2 women 448 mol/g (193) 2692 g/L (1350) 82% (16)
49 (18–72) NS 42 men, 36 women 202 mol/g (115) ⬍0.0001 1116 g/L (1008) ⬍0.0001 75% (17) 0.09
89 IU/L (108) 104 IU/L (167)
56 IU/L (64) 45 IU/L (59)
0.07 0.01
Data reported as mean (standard deviation). Reference range hepatic iron concentration: 0 –35 [mol/g dry weight, 0 –1935 (g/g)]. Reference ranges: serum ferritin [15–200 (g/L women); 30 –300 (g/L men); transferrin saturation (20 –55%); AST, ALT ⫽ 0 – 40 IU/L].
The threshold for the prediction of cirrhosis was 283 mol/g (15,803 g/g). At that threshold, the sensitivity was 85%, specificity 84%, positive likelihood ratio 5.2, and negative likelihood ratio 0.18. Hepatic iron concentration was significantly greater in men (277 ⫾ 186 mol/g, mean ⫾ SD) compared to women (176 ⫾ 103 mol/g, p ⬍ 0.001).
DISCUSSION From this analysis, it appears that a hepatic iron concentration of greater than 283 mol/g is associated with cirrhosis. This is approximately eight times the upper limit of the reference range for hepatic iron concentration. It is a much higher iron concentration than has been implicated as a cofactor in the production of hepatic fibrosis in hepatitis C (9) or NASH (10 –12). Previous studies in hemochromatosis (before the use of genetic testing) suggested that the threshold for cirrhosis was 400 mol/g in an Australian study (6). Studies from France have studied hepatic iron concentrations as compared to fibrosis stages (0 to 4) and have not
Figure 1. Hepatic iron concentration was significantly increased in cirrhotics (n ⫽ 22) compared to noncirrhotic (n ⫽ 78) C282Y homozygotes (p ⬍ 0.0001). Solid bars represent the mean hepatic iron concentration. The dotted line represents the upper limit of normal of the reference range for hepatic iron concentration (35 mol/g dry weight).
Figure 2. Receiver operating characteristic (ROC) curve analysis of hepatic iron concentration versus presence of cirrhosis in 100 C282Y homozygotes. Area under the curve was 0.85 (95% confidence interval ⫽ 0.75– 0.96). The threshold for the prediction of cirrhosis was 283 mol/g.
been able to demonstrate a discrete threshold. The mean hepatic iron concentration in their cirrhotic patients was 371 mol/g ⫾ 138 (n ⫽ 28) (13, 14). Our study also does not demonstrate a discrete threshold, as can be seen in Figure 1. The ROC analysis always generates an optimal threshold based on (sensitivity/1 ⫺ specificity). If there is an excellent threshold, the area under the ROC curve is usually ⬎0.95, which was not seen in this analysis. The observation that there are many cirrhotics with iron concentrations ⬍283 mol/g and noncirrhotics with iron concentrations ⬎283 mol/g suggests that other factors contribute to the development of cirrhosis in hemochromatosis. The lower threshold in our study may be related to the inclusion of younger patients and earlier diagnosis. The ROC curve methodology always generates a threshold, so the inclusion of milder cases will decrease the threshold. The area under the curve is an indicator of the power of the threshold, which was not evaluated by the same techniques in the other studies. Previous studies have suggested that alcoholism, age, and sex are cofactors (13), and in this study, cirrhosis was most common in older men. Alcoholism and viral hepatitis have been excluded at entry in this study. The low prevalence of alcohol abuse in this study is consistent with our previous pedigree studies (15), and older studies have likely included alcoholic siderosis patients as having hemochromatosis. Genetic susceptibility to cirrhosis is another possibility, as has been postulated in alcoholic liver disease. Early diagnosis and iron depletion by venesection has been demonstrated to prevent the development of cirrhosis in hemochromatosis (16, 17). Iron depletion in cirrhotics can arrest further progression, which may explain the very low rate of liver transplantation reported in C282Y homozygotes. Hepatocellular carcinoma can develop in the cirrhotic liver, despite iron depletion. Cirrhosis remains a preventable life-threatening complication in a common genetic disease. Early diagnosis remains the key to improving long-term survival,
AJG – February, 2001
and this will require more aggressive case detection, including population screening (18).
ACKNOWLEDGMENTS The author acknowledges the support of Tom Pellar, who developed the computer software, and the late Leslie Valberg, who contributed many of the patients to this study. Reprint requests and correspondence: Paul C. Adams, M.D., Department of Medicine, London Health Sciences Centre—University Campus, 339 Windermere Rd., London, Ontario, Canada N6A 5A5. Received Apr. 6, 2000; accepted Aug. 25, 2000
REFERENCES 1. Park CH, Bacon BR, Brittenham GM, et al. Pathology of dietary carbonyl iron overload in rats. Lab Invest 1987;57: 555– 63. 2. Weintraub LR, Goral A, Grasso J, et al. Pathogenesis of hepatic fibrosis in experimental iron overload. Br J Haematol 1985;59:321–31. 3. Gualdi R, Casalgrandi G, Montosi G, et al. Excess iron into hepatocytes is required for activation of collagen type I gene during experimental siderosis. Gastroenterology 1994;107: 1118 –24. 4. Pietrangelo A, Gualdi R, Casalgrandi G, et al. Enhanced hepatic collagen type I mRNA expression into fat-storing cells in a rodent model of hemochromatosis. Hepatology 1994;19: 714 –21. 5. Feder JN, Gnirke A, Thomas W, et al. A novel MHC class I-like gene is mutated in patients with hereditary hemochromatosis. Nat Genet 1996;13:399 – 408. 6. Bassett ML, Halliday JW, Powell LW. Value of hepatic iron measurements in early hemochromatosis and determination of the critical iron level associated with fibrosis. Hepatology 1986;6:24 –9.
Hepatic Iron Concentration and Hemochromatosis
569
7. Pellar TG, Leung FY, Henderson AR. A computer program for rapid generation of receiver operating characteristic curves and likelihood ratios in the evaluation of diagnostic tests. Ann Clin Biochem 1988;25:411– 6. 8. Henderson AR. Assessing test accuracy and its clinical consequences: A primer for receiver operating characteristic curve analysis. Ann Clin Biochem 1993;30:521–39. 9. Bonkovsky H, Banner B, Rothman A. Iron and chronic viral hepatitis. Hepatology 1998;25:759 – 68. 10. Bonkovsky HL, Jawaid Q, Tortorelli K, et al. Non-alcoholic steatohepatitis, and iron. Increased prevalence of mutations of the HFE gene in non-alcoholic steatohepatitis. J Hepatol 1999; 31:421–9. 11. George DK, Goldwurm S, MacDonald GA, et al. Increased hepatic iron concentration in nonalcoholic steatohepatitis is associated with increased fibrosis. Gastroenterology 1998;114: 311– 8. 12. George DK, Powell LW, Losowsky MS. The haemochromatosis gene: A co-factor for chronic liver diseases? J Gastroenterol Hepatol 1999;14:745–9. 13. Deugnier YM, Loreal O, Turlin B, et al. Liver pathology in genetic hemochromatosis: A review of 135 homozygous cases and their bioclinical correlations. Gastroenterology 1992;102: 2050 –9. 14. Loreal O, Deugnier Y, Moirand R, et al. Liver fibrosis in genetic hemochromatosis — respective roles of iron and noniron-related factors in 127 homozygous patients. J Hepatol 1992;16:122–7. 15. Adams PC, Agnew S. Alcoholism in hereditary hemochromatosis revisited: Prevalence and clinical consequences among homozygous siblings. Hepatology 1996;23:724 –7. 16. Niederau C, Fischer R, Purschel A, et al. Long-term survival in patients with hereditary hemochromatosis. Gastroenterology 1996;110:1107–19. 17. Adams PC, Speechley M, Kertesz AE. Long-term survival analysis in hereditary hemochromatosis. Gastroenterology 1991;101:368 –72. 18. Adams PC. Population screening for haemochromatosis. Gut 2000;46:301–3.