- Email: [email protected]
A new syndrome of liver iron overload with normal transferrin saturation
THE LANCET
may be because it was a defective, cell-associated form such as that seen in subacute sclerosing panencephalitis. The pathological findings and short duration of this illness are, however, quite different from those of that disorder. Although recent evidence suggests that native Australian fruitbats may be the natural host of EMV,5 the origins of the virus remain obscure. Three human beings infected with this virus have been reported—the two patients who died and another patient who had an influenza-like illness and survived.1 All three patients had extensive contact with horses infected with the virus. Our patient cared for horses and assisted at their necropsies without gloves, mask, or protective eyewear. Since others involved in necropsies on infected horses in the previously described cluster of cases did not become infected, and since there was a history of close contact between the horses involved and the human cases, direct contact with respiratory secretions of infected animals seems to be necessary for transmission.
We thank P W Selleck, Australian Animal Health Laboratories, for permission to use unpublished results; H Samaratunga and A Tannenburg, Pathology Department, Royal Brisbane Hospital, for pathological assistance; and J Sheridan and L Selvey, Communicable Diseases Branch of Queensland Health, Brisbane, for epidemiological assistance.
References 1 Selvey LA, Wells RM, McCormack JG, et al. Infection of humans and horses by a newly described morbillivirus. Med J Aust 1995; 162: 642–45. 2 Murray K, Selleck P, Hooper P, et al. A morbillivirus that caused fatal disease in horses and humans. Science 1995; 268: 94–97. 3 Gould AR. Comparison of deduced matrix and fusion protein sequences of equine morbillivirus with cognate genes of the Paramyxoviridae. Virus Res 1996; 43: 17–31. 4 Hooper PT, Gould AR, Mitchell G, et al. Retrospective diagnosis of a second outbreak of equine morbillivirus disease. Aust Vet J 1996; 74: 244–45. 5 Young PL, Halpin K, Selleck PW, et al. Serological evidence for the presence in pteropus bats of a paramyxovirus related to equine morbillivirus. Emerg Infect Dis 1996; 2: 239–40.
A new syndrome of liver iron overload with normal transferrin saturation Romain Moirand, Abdel Majid Mortaji, Olivier Loréal, François Paillard, Pierre Brissot, Yves Deugnier
Summary Background We investigated patients who had unexplained hepatic iron overload and normal transferrin saturation. Methods 65 patients with a median liver iron concentration of 85 µmol/g dry weight of liver (normal <36 µmol/g), hyperferritinaemia (566 µg/L; normal <400 µg/L), and normal transferrin saturations (32%) were compared with genetic haemochromatosis (GH) controls including homozygous (matched for sex and serum ferritin concentration) and heterozyogus individuals. Relatives of patients who had ratios of liver iron concentration to age greater than 1·9 were also studied. Findings The 65 patients were significantly older and had significantly less hepatic iron overload than individuals with genetic haemochromatosis. The frequency of HLA-A3 antigen was significantly lower in these patients than in individuals with homozygous (p<0·0001) or heterozygous (p<0·0002) GH. Five HLA-identical siblings of the patients had normal serum ferritin concentrations. Most of the patients (95%) had one or more of the following conditions; obesity, hyperlipidaemia, abnormal glucose metabolism, or hypertension. Interpretation We have found a new non-HLA-linked ironoverload syndrome which suggests a link between iron excess and metabolic disorders. The current diagnostic criteria for genetic haemochromatosis should be reviewed.
Lancet 1997; 349: 95–97 See Commentary page 71 Clinique des Maladies du Foie, Hôpital Pontchaillou, 35033 Rennes, France (R Moirand MD, A M Mortaji MD, Prof P Brissot MD, Prof Y Deugnier MD); INSERM U49, CHRU Pontchaillou (O Loréal PhD, Prof P Brissot, Prof Y Deugnier) ; and Service de Cardiologie A, CHRU Hôtel Dieu, Rennes (F Paillard MD) Correspondence to: Prof Yves Deugnier
Vol 349 • January 11, 1997
Introduction Genetic haemochromatosis (GH) is an autosomal recessive disease and the likely gene, situated on chromosome 6 close to the HLA A locus, has been identified by Feder et al.1 Before the potential gene was identified, the diagnosis of GH in a given individual—ie, outside family studies—was based on the finding of hepatic parenchymal iron overload in the absence of any other known cause.2,3 The assertion of putative homozygosity relied upon a ratio of liver iron concentration to age, the hepatic-iron index (HII),4 greater than 1·9.5 Edwards et al6 have emphasised that transferrin saturation is usually increased in GH patients. We investigated patients with liver iron overload who had normal transferrin saturation unexplained by any of the usual causes, and compared them with individuals with GH and their relatives.
Patients and methods Selection of patients 65 patients who had been referred to our unit with suspected iron overload were retrospectively selected on the following criteria: serum transferrin saturation 45% or less on at least two occasions; liver iron overload defined as a liver iron concentration (LIC)7 greater than the upper limit of normal (normal <36 µmol/g dry liver weight)8 on liver biopsy sample (n=50), magnetic resonance imaging (n=15),9 or both (n=26). Exclusion criteria were: chronic alcohol use either in the past or present (alcohol consumption of >60 g/day in men or 40 g/day in women); biochemical or histological features of alcoholism on liver biopsy; excess oral intake of iron or ascorbic acid; repeated blood transfusions; haematological disorders; porphyria cutanea tarda (past or current skin lesions, or uroporphyrin inclusions on liver biopsy sample); chronic hepatitis; or an inflammatory syndrome (increased erythrocyte sedimentation rate or C-reactive protein). Patients were questioned about a history of blood donations, drug consumption, hyperlipidaemia, hypertension, or diabetes. 24 patients had an oral glucose tolerance test. HLA typing, by standard microlymphocytotoxicity test, was available for 62. Blood samples were taken between 0700 and 0800 h after the patient had fasted overnight. Patients were classified as hyperlipidaemic if either total fasting plasma cholesterol was more
95
THE LANCET
Patients (n=65)
Matched GH controls (n=65)
Age (years)
52 (36–76)
42 (18–73)*
Male/female
60/5
60/5
2
Body-mass index (kg/m ) Red-blood-cell indices Mean corpuscular volume (fL) Mean corpuscular haemoglobin content (g/L) Biochemistry Asparate aminotransferase ( N) Alanine aminotransferase ( N) glutamyltranspeptidase ( N) Serum iron (µmol/L) Transferrin saturation (%) Transferrin (g/L) Serum ferritin (µg/L) LIC (µmol/g) HII Iron removed (g)
26·7 (21–36)
23·5 (19–35)*
90 (85–107) 340 (318–361)
94 (85–103)* 340 (324–359)
0·4 (0·2–0·9) 0·5 (0·1–1·6) 0·7 (0·3–2·6) 21 (10–29) 32 (14–45) 2·6 (1·7–4·9) 566 (357–1730) 85 (44–273) 1·6 (0·8–4·0) 2·5 (1·2–10·0)
0·4 (0·1–1·0) 0·4 (0·1–2·7) 0·5 (0·2–1·9)* 37 (23–52)* 78 (44–98)* 1·9 (1·2–3·0)* 558 (336–1725) 227 (68–523)* 5·2 (2·2–11·4)* 5·0 (2·0–20·0)*
Data are given as median (range). *p<0·05 between groups. N=compared with values from established normal range. Normal ranges: serum iron 12·5–25·0 µmol/L; transferrin saturation 25–45%; transferrin 2–4 g/L; serum ferritin 30–400 µg/mL; LIC <36 µmol/g.
Table 1: Clinical and biochemical data
than 6·2 mmol/L, or triglycerides more than 1·7 mmol/L, or if they received hypolipidaemic treatment. Patients were deemed to have abnormal glucose metabolism if they required antidiabetic therapy or satisfied WHO criteria for impaired glucose tolerance or diabetes mellitus. Venesected patients had 350–400 mL of blood removed once a week. Iron depletion was defined when the serum ferritin was less than 50 ng/mL. The amount of removed iron was based on the calculation that 1 L blood contains 0·5 g iron. All available first-degree and second-degree relatives of patients who had an HII greater than 1·9 were HLA typed.1 Three different control groups were constituted from our GH database, all of whom were diagnosed on current criteria and had increased transferrin saturations. Matched controls were 65 homozygous GH subjects matched for age and serum ferritin to each patient. Confirmed homozygote controls were HLA-identical GH siblings of GH controls with an HII greater than 1·9. Heterozygote controls were selected from the families of confirmed GH homozygotes, and consisted of all siblings sharing one haplotype with the proband and all children of probands, unless there was a suspicion of homozygote-heterozygote parentage. 2, Mann-Whitney, and Spearman’s rank correlation tests were used as appropriate.
Results Of the 65 patients, 53 were referred because of an unexplained increased serum ferritin, and others for nonspecific symptoms such as fatigue (n=6) and arthralgias (n=6). The patients were significantly older than matched controls and also had significantly lower serum iron concentrations, LIC, HII, and amount of iron removed (table 1). Serum ferritin concentrations were more than 1000 µg/L in six patients. On liver biopsy samples, iron was always found within periportal hepatocytes and, less frequently, within Kupffer cells. 21 patients had an LIC greater than 100 µmol/g and in five others the LIC was above 150 µmol/g. The HII was more than 1·9 in 22
Body-mass index >25 kg/m2 Hyperlipidaemia Abnormal glucose metabolism Hypertension
Patients
Matched GH controls
p
72% (43/60) 65% (39/60) 43% (16/37) 19% (12/64)
35% (17/49) 17% (9/52) 8% (3/39) 12% (7/59)
0·0002 0·0001 0·0004 NS
Table 2: Prevalence of metabolic disorders
96
Patients with HII 1·9 (n=43)
Patients with HII >1·9 (n=22)
Matched GH controls* (n=22)
Age(years)
53 (36–76)
49 (37–68)
42 (28–65)‡
Male/female
38/5
22/0
22/0
Biochemistry Serum iron Transferrin saturation Serum ferritin LIC HII Iron removed (g)
20 (10–28)† 29 (14–44)† 539 (357–1730) 75 (44–120)† 1·4 (0·8–1·9)† 2·0 (1·2–2·4)†
22 (18–29) 36 (22–45) 606 (442–1651) 119 (75–273) 2·4 (1·9–4·0) 3·1 (1·8–10·0)
36 (25–52)‡ 79 (47–98)‡ 624 (398–1691) 220 (68–342)‡ 4·9 (2·2–8·9)‡ 4·2 (2·6–20·0)‡
Data are given as median (range). Normal ranges given in table 1. *Matched for sex and serum ferritin with patients with HII >1·9. †p<0·05 between patients with HII 1·9 and patients with HII >1·9. ‡p<0·05 between all patients and matched GH controls.
Table 3: Characteristics of patients according to HII and of matched homozygous GH controls
patients. Follow-up information was available for 36 patients. Weight reduction (n=4) did not decrease the serum ferritin concentration. Iron depletion was achieved in 26 patients (table 1), and the amount of removed iron correlated with LIC (r=0·85, p<0·0001) but not with the serum ferritin. The majority (95%) of patients were overweight, hyperlipidaemic, or hypertensive, or had an abnormal glucose metabolism (table 2). The frequency of HLA A3 in patients (17/62 [28%]) was similar to that of the general population of Brittany, but differed significantly from that of the heterozygote controls (82/142 [58%], p=0·0002) and the matched GH controls (51/65 [78%], p=0·0001). When patients and matched GH controls with an HII greater than 1·9 were compared, a similar pattern was observed (5/22 [24%] vs 18/22 [82%], p=0·0002). Patients with an HII above 1·9 did not have any special clinical features but had significantly higher serum and LICs than patients with an HII of 1·9 or less. Serum ferritin did not differ between the two groups (table 3). The prevalence of HLA A3 in patients with HII above 1·9 differed significantly from that of all control groups but not from patients with an HII of 1·9 or less (p=0·49). Four male and one female HLA-identical siblings of patients were identified in five different families; four had normal serum iron concentrations and one, a 66-year-old alcoholic man, had increased serum iron and transferrin saturation with a normal serum ferritin concentration (286 µg/L)—which does not support a diagnosis of homozygous GH.
Discussion Our study describes a new, non-HLA-linked, iron-overload syndrome in which transferrin saturation is normal and there is close association with various metabolic disorders. In most iron-overloaded patients, a raised serum iron concentration and transferrin saturation precede the increase in serum ferritin concentration, whatever the mechanism of iron excess.6,12 We do not think that serum iron concentrations or transferrin saturations were falsely underestimated in our patients because both the haemoglobin concentration and inflammatory markers were normal. However, for an identical serum ferritin concentration, LIC and removed iron were 2·5 times lower in these patients than in matched controls, which suggests an iron excess much less than that predicted by the serum ferritin. Screening for iron overload by transferrin saturation alone would misdiagnose this type of overload, and
Vol 349 • January 11, 1997
THE LANCET
therefore only direct methods8,13 (liver histology2 or magnetic-resonance imaging9), should be used to quantify liver iron stores. The iron-overload syndrome described here cannot be placed within the current classification of iron-overload disorders.3,14,15 Common causes of secondary iron overload were carefully excluded in these patients—especially sideroblastic anaemia, chronic haemolysis, porphyria cutanea tarda, chronic hepatitis C,16 and steatohepatitis.17 In addition, weekly venesection was well tolerated by most treated patients. Finally, chronic alcoholism was excluded; not only by history but also on chemical and histological data. The slight increase in γ-glutamyl transpeptidase (GGT) observed in some of our patients was probably related to obesity.18 Although 34% of these patients had HII above 1·9, we do not think that the GH gene is involved, since none of the confirmed GH homozygote controls had a normal transferrin saturation. The prevalence of HLA A3 (which in a given population may be used as a reference for the GH gene)19 did not differ between our patients and the general population and was significantly lower than in both homozygote and heterozygote controls. Also, none of the five HLA-identical siblings had clinical or biochemical features of iron excess. In GH families, such features can arise because of genetic recombination, but Powell et al20 have shown that this is rare. Thus, it is unlikely that genetic recombination could account for all our patients. The identification of this non-HLA-linked iron-overload syndrome has important implications. Because this condition is likely to be misdiagnosed as early or underexpressed GH, its recognition should help to avoid costly and potentially disruptive family-screening procedures and, in GH gene research programmes, consideration of people with atypical phenotypes. Finally, our findings suggest that the prevalence of the GH gene,19 which relies heavily on the prevalence of phenotypic features of iron overload, may be overestimated. This new condition is a homogeneous entity, characterised by a normal transferrin saturation, hyperferritinaemia, and mild to moderate iron overload. Furthermore, almost all patients had a concomitant metabolic disorder (table 2) associated with the insulin resistance syndrome.21,22 Thus, it is tempting to add iron excess to this cluster of metabolic aberrations. Support for this theory includes the finding of increased serum ferritin concentrations in diabetes23 and people who eat a lot of meat24 and the current debate about the putative role of body-iron excess as a risk factor for coronary heart disease.25 Iron excess was usually mild to moderate in our patients. Is this excess clinically important and should it be treated? Experimental,26 epidemiological,27 and clinical studies28 suggest a role of increased body-iron stores in human carcinogenesis. Thus, treatment of iron excess, whatever its severity, may be an important goal for general health. The description of this new syndrome suggests that the diagnostic criteria of iron-overload syndromes should be modified. Raised serum ferritin concentration and transferrin saturation are both needed to detect all types of iron overload. Just as idiopathic iron overload does not always correspond to GH, an HII cut-off point of 1·9 does not discriminate between the different iron-overload conditions. Also, the data suggest that there may be a link between increased liver-iron stores and various common metabolic disorders. The exact nature and significance of this link
Vol 349 • January 11, 1997
remains to be clarified. This study was supported by grants from the Association pour la Recherche contre le Cancer and the Programme Hospitalier de Recherche Clinique 1994.
References 1
2
3
4
5
6
7 8
9
10
11
12 13
14
15 16
17
18
19 20
21 22 23
24
25 26
27 28
Feder JN, Gnirke A, Thomas W, et al. A novel MHC class I-like gene is mutated in patients with hereditary haemochromatosis. Nat Genet 1996; 13: 399–408. 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: 2050–59. Brissot P, Deugnier Y. Genetic haemochromatosis. In: McIntyre N, Benhamou J, Bircher J, Rizzetto M, Rodes J, eds. Oxford textbook of clinical hepatology. Oxford: Oxford University Press. 1991: 948–58. 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: 24–29. Summers KM, Halliday JW, Powell LW. Identification of homozygous hemochromatosis subjects by measurement of hepatic iron index. Hepatology 1990; 12: 20–25. Edwards CQ, Griffen LM, Goldgar D, Drummond C, Skolnick MH, Kushner JP. Prevalence of hemochromatosis among 11 065 presumably healthy blood donors. N Engl J Med 1988; 318: 1355–62. Barry M, Sherlock S. Measurement of liver concentration in needlebiopsy specimens. Lancet 1971; i: 100–03. Brissot P, Bourel M, Herry D, et al. Assessment of liver iron content in 271 patients: a reevaluation of direct and indirect methods. Gastroenterology 1981; 80: 557–65. Gandon Y, Guyader D, Heautot JF, et al. Hemochromatosis: diagnosis and quantification of liver iron with gradient-echo MR imaging. Radiology 1994; 193: 533–38. Rolland-Cachera MF, Cole TJ, Sempi M, Tichet J, Rossignol C, Charraud A. Body mass index variations: centiles from birth to 87 years. Eur J Clin Nutr 1991; 45: 13–21. Simon M, Bourel M, Genetet B, Fauchet R, Edan G, Brissot P. Idiopathic hemochromatosis and iron overload in alcoholic liver disease: differentiation by HLA phenotype. Gastroenterology 1977; 73: 655–58. Peto TEA, Pippard MJ, Weatherall DJ. Iron overload in mild sideroblastic anaemias. Lancet 1983; i: 375–78. Bonkovsky HL, Slaker DP, Bills EB, Wolf DC. Usefulness and limitations of laboratory and hepatic imaging studies in iron-storage disease. Gastroenterology 1990; 99: 1079–91. Tavill AS, Sharma BK, Bacon BR. Iron and the liver: genetic hemochromatosis and other hepatic overload disorders. Prog Liver Dis 1990; 9: 281–305. Bacon BR. Causes of iron overload. N Engl J Med 1992; 326: 126–27. Di Bisceglie AM, Axiotis CA, Hoofnagle JH, Bacon BR. Measurements of iron status in patients with chronic hepatitis. Gastroenterology 1992; 102: 2108–13. Bacon BR, Farahvash MJ, Janney CG, Neuschwander-Tetri BA. Nonalcoholic steatohepatitis: an expanded clinical entity. Gastroenterology 1994; 107: 1103–09. Nilssen O, Forde OH. Seven-year longitudinal population study of change in gamma-glutamyltransferase: the Tromso Study. Am J Epidemiol 1994; 139: 787–92. Simon M, Brissot P. The genetics of haemochromatosis. J Hepatol 1988; 6: 116–24. Powell LW, Summers KM, Board PG, Axelsen E, Webb S, Halliday JW. Expression of hemochromoatosis in homozygous subjects: implications for early diagnosis and prevention. Gastroenterology 1990; 98: 1625–32. Reaven G. Role of insulin resistance in human disease. Diabetes 1988; 37: 1495–507. Després JP, Marette A. Relation of components of insulin resistance syndrome to coronary disease risk. Curr Opin Lipidol 1994; 5: 274–89. Dinneen SF, O’Mahony MS, O’Brien T, Cronin CC, Murray DM, O’Sullivan DJ. Serum ferritin in newly diagnosed and poorly controlled diabetes mellitus. Ir J Med Sci 1992; 161: 636–38. Leggett BA, Brown NN, Bryant SJ, Duplock L, Powell LW, Halliday JW. Factors affecting the concentration of ferritin in serum in a healthy Australian population. Clin Chem 1990; 36: 1350–55. Conrad ME. Excess iron and catastrophic illness. Am J Hematol 1993; 43: 234–36. Hann HWL, Stahlhut MW, Rubin R, Maddrey WC. Antitumor effect of deferoxamine on human hepatocellular carcinoma growing in athymic nude mice. Cancer 1992; 70: 2051–56. Stevens RG, Jones DY, Micozzi MS, Taylor PR. Body iron stores and the risk of cancer. N Engl J Med 1988; 319: 1047–52. Turlin B, Juguet F, Moirand R, et al. Increased liver iron stores in non-
97
may be because it was a defective, cell-associated form such as that seen in subacute sclerosing panencephalitis. The pathological findings and short duration of this illness are, however, quite different from those of that disorder. Although recent evidence suggests that native Australian fruitbats may be the natural host of EMV,5 the origins of the virus remain obscure. Three human beings infected with this virus have been reported—the two patients who died and another patient who had an influenza-like illness and survived.1 All three patients had extensive contact with horses infected with the virus. Our patient cared for horses and assisted at their necropsies without gloves, mask, or protective eyewear. Since others involved in necropsies on infected horses in the previously described cluster of cases did not become infected, and since there was a history of close contact between the horses involved and the human cases, direct contact with respiratory secretions of infected animals seems to be necessary for transmission.
We thank P W Selleck, Australian Animal Health Laboratories, for permission to use unpublished results; H Samaratunga and A Tannenburg, Pathology Department, Royal Brisbane Hospital, for pathological assistance; and J Sheridan and L Selvey, Communicable Diseases Branch of Queensland Health, Brisbane, for epidemiological assistance.
References 1 Selvey LA, Wells RM, McCormack JG, et al. Infection of humans and horses by a newly described morbillivirus. Med J Aust 1995; 162: 642–45. 2 Murray K, Selleck P, Hooper P, et al. A morbillivirus that caused fatal disease in horses and humans. Science 1995; 268: 94–97. 3 Gould AR. Comparison of deduced matrix and fusion protein sequences of equine morbillivirus with cognate genes of the Paramyxoviridae. Virus Res 1996; 43: 17–31. 4 Hooper PT, Gould AR, Mitchell G, et al. Retrospective diagnosis of a second outbreak of equine morbillivirus disease. Aust Vet J 1996; 74: 244–45. 5 Young PL, Halpin K, Selleck PW, et al. Serological evidence for the presence in pteropus bats of a paramyxovirus related to equine morbillivirus. Emerg Infect Dis 1996; 2: 239–40.
A new syndrome of liver iron overload with normal transferrin saturation Romain Moirand, Abdel Majid Mortaji, Olivier Loréal, François Paillard, Pierre Brissot, Yves Deugnier
Summary Background We investigated patients who had unexplained hepatic iron overload and normal transferrin saturation. Methods 65 patients with a median liver iron concentration of 85 µmol/g dry weight of liver (normal <36 µmol/g), hyperferritinaemia (566 µg/L; normal <400 µg/L), and normal transferrin saturations (32%) were compared with genetic haemochromatosis (GH) controls including homozygous (matched for sex and serum ferritin concentration) and heterozyogus individuals. Relatives of patients who had ratios of liver iron concentration to age greater than 1·9 were also studied. Findings The 65 patients were significantly older and had significantly less hepatic iron overload than individuals with genetic haemochromatosis. The frequency of HLA-A3 antigen was significantly lower in these patients than in individuals with homozygous (p<0·0001) or heterozygous (p<0·0002) GH. Five HLA-identical siblings of the patients had normal serum ferritin concentrations. Most of the patients (95%) had one or more of the following conditions; obesity, hyperlipidaemia, abnormal glucose metabolism, or hypertension. Interpretation We have found a new non-HLA-linked ironoverload syndrome which suggests a link between iron excess and metabolic disorders. The current diagnostic criteria for genetic haemochromatosis should be reviewed.
Lancet 1997; 349: 95–97 See Commentary page 71 Clinique des Maladies du Foie, Hôpital Pontchaillou, 35033 Rennes, France (R Moirand MD, A M Mortaji MD, Prof P Brissot MD, Prof Y Deugnier MD); INSERM U49, CHRU Pontchaillou (O Loréal PhD, Prof P Brissot, Prof Y Deugnier) ; and Service de Cardiologie A, CHRU Hôtel Dieu, Rennes (F Paillard MD) Correspondence to: Prof Yves Deugnier
Vol 349 • January 11, 1997
Introduction Genetic haemochromatosis (GH) is an autosomal recessive disease and the likely gene, situated on chromosome 6 close to the HLA A locus, has been identified by Feder et al.1 Before the potential gene was identified, the diagnosis of GH in a given individual—ie, outside family studies—was based on the finding of hepatic parenchymal iron overload in the absence of any other known cause.2,3 The assertion of putative homozygosity relied upon a ratio of liver iron concentration to age, the hepatic-iron index (HII),4 greater than 1·9.5 Edwards et al6 have emphasised that transferrin saturation is usually increased in GH patients. We investigated patients with liver iron overload who had normal transferrin saturation unexplained by any of the usual causes, and compared them with individuals with GH and their relatives.
Patients and methods Selection of patients 65 patients who had been referred to our unit with suspected iron overload were retrospectively selected on the following criteria: serum transferrin saturation 45% or less on at least two occasions; liver iron overload defined as a liver iron concentration (LIC)7 greater than the upper limit of normal (normal <36 µmol/g dry liver weight)8 on liver biopsy sample (n=50), magnetic resonance imaging (n=15),9 or both (n=26). Exclusion criteria were: chronic alcohol use either in the past or present (alcohol consumption of >60 g/day in men or 40 g/day in women); biochemical or histological features of alcoholism on liver biopsy; excess oral intake of iron or ascorbic acid; repeated blood transfusions; haematological disorders; porphyria cutanea tarda (past or current skin lesions, or uroporphyrin inclusions on liver biopsy sample); chronic hepatitis; or an inflammatory syndrome (increased erythrocyte sedimentation rate or C-reactive protein). Patients were questioned about a history of blood donations, drug consumption, hyperlipidaemia, hypertension, or diabetes. 24 patients had an oral glucose tolerance test. HLA typing, by standard microlymphocytotoxicity test, was available for 62. Blood samples were taken between 0700 and 0800 h after the patient had fasted overnight. Patients were classified as hyperlipidaemic if either total fasting plasma cholesterol was more
95
THE LANCET
Patients (n=65)
Matched GH controls (n=65)
Age (years)
52 (36–76)
42 (18–73)*
Male/female
60/5
60/5
2
Body-mass index (kg/m ) Red-blood-cell indices Mean corpuscular volume (fL) Mean corpuscular haemoglobin content (g/L) Biochemistry Asparate aminotransferase ( N) Alanine aminotransferase ( N) glutamyltranspeptidase ( N) Serum iron (µmol/L) Transferrin saturation (%) Transferrin (g/L) Serum ferritin (µg/L) LIC (µmol/g) HII Iron removed (g)
26·7 (21–36)
23·5 (19–35)*
90 (85–107) 340 (318–361)
94 (85–103)* 340 (324–359)
0·4 (0·2–0·9) 0·5 (0·1–1·6) 0·7 (0·3–2·6) 21 (10–29) 32 (14–45) 2·6 (1·7–4·9) 566 (357–1730) 85 (44–273) 1·6 (0·8–4·0) 2·5 (1·2–10·0)
0·4 (0·1–1·0) 0·4 (0·1–2·7) 0·5 (0·2–1·9)* 37 (23–52)* 78 (44–98)* 1·9 (1·2–3·0)* 558 (336–1725) 227 (68–523)* 5·2 (2·2–11·4)* 5·0 (2·0–20·0)*
Data are given as median (range). *p<0·05 between groups. N=compared with values from established normal range. Normal ranges: serum iron 12·5–25·0 µmol/L; transferrin saturation 25–45%; transferrin 2–4 g/L; serum ferritin 30–400 µg/mL; LIC <36 µmol/g.
Table 1: Clinical and biochemical data
than 6·2 mmol/L, or triglycerides more than 1·7 mmol/L, or if they received hypolipidaemic treatment. Patients were deemed to have abnormal glucose metabolism if they required antidiabetic therapy or satisfied WHO criteria for impaired glucose tolerance or diabetes mellitus. Venesected patients had 350–400 mL of blood removed once a week. Iron depletion was defined when the serum ferritin was less than 50 ng/mL. The amount of removed iron was based on the calculation that 1 L blood contains 0·5 g iron. All available first-degree and second-degree relatives of patients who had an HII greater than 1·9 were HLA typed.1 Three different control groups were constituted from our GH database, all of whom were diagnosed on current criteria and had increased transferrin saturations. Matched controls were 65 homozygous GH subjects matched for age and serum ferritin to each patient. Confirmed homozygote controls were HLA-identical GH siblings of GH controls with an HII greater than 1·9. Heterozygote controls were selected from the families of confirmed GH homozygotes, and consisted of all siblings sharing one haplotype with the proband and all children of probands, unless there was a suspicion of homozygote-heterozygote parentage. 2, Mann-Whitney, and Spearman’s rank correlation tests were used as appropriate.
Results Of the 65 patients, 53 were referred because of an unexplained increased serum ferritin, and others for nonspecific symptoms such as fatigue (n=6) and arthralgias (n=6). The patients were significantly older than matched controls and also had significantly lower serum iron concentrations, LIC, HII, and amount of iron removed (table 1). Serum ferritin concentrations were more than 1000 µg/L in six patients. On liver biopsy samples, iron was always found within periportal hepatocytes and, less frequently, within Kupffer cells. 21 patients had an LIC greater than 100 µmol/g and in five others the LIC was above 150 µmol/g. The HII was more than 1·9 in 22
Body-mass index >25 kg/m2 Hyperlipidaemia Abnormal glucose metabolism Hypertension
Patients
Matched GH controls
p
72% (43/60) 65% (39/60) 43% (16/37) 19% (12/64)
35% (17/49) 17% (9/52) 8% (3/39) 12% (7/59)
0·0002 0·0001 0·0004 NS
Table 2: Prevalence of metabolic disorders
96
Patients with HII 1·9 (n=43)
Patients with HII >1·9 (n=22)
Matched GH controls* (n=22)
Age(years)
53 (36–76)
49 (37–68)
42 (28–65)‡
Male/female
38/5
22/0
22/0
Biochemistry Serum iron Transferrin saturation Serum ferritin LIC HII Iron removed (g)
20 (10–28)† 29 (14–44)† 539 (357–1730) 75 (44–120)† 1·4 (0·8–1·9)† 2·0 (1·2–2·4)†
22 (18–29) 36 (22–45) 606 (442–1651) 119 (75–273) 2·4 (1·9–4·0) 3·1 (1·8–10·0)
36 (25–52)‡ 79 (47–98)‡ 624 (398–1691) 220 (68–342)‡ 4·9 (2·2–8·9)‡ 4·2 (2·6–20·0)‡
Data are given as median (range). Normal ranges given in table 1. *Matched for sex and serum ferritin with patients with HII >1·9. †p<0·05 between patients with HII 1·9 and patients with HII >1·9. ‡p<0·05 between all patients and matched GH controls.
Table 3: Characteristics of patients according to HII and of matched homozygous GH controls
patients. Follow-up information was available for 36 patients. Weight reduction (n=4) did not decrease the serum ferritin concentration. Iron depletion was achieved in 26 patients (table 1), and the amount of removed iron correlated with LIC (r=0·85, p<0·0001) but not with the serum ferritin. The majority (95%) of patients were overweight, hyperlipidaemic, or hypertensive, or had an abnormal glucose metabolism (table 2). The frequency of HLA A3 in patients (17/62 [28%]) was similar to that of the general population of Brittany, but differed significantly from that of the heterozygote controls (82/142 [58%], p=0·0002) and the matched GH controls (51/65 [78%], p=0·0001). When patients and matched GH controls with an HII greater than 1·9 were compared, a similar pattern was observed (5/22 [24%] vs 18/22 [82%], p=0·0002). Patients with an HII above 1·9 did not have any special clinical features but had significantly higher serum and LICs than patients with an HII of 1·9 or less. Serum ferritin did not differ between the two groups (table 3). The prevalence of HLA A3 in patients with HII above 1·9 differed significantly from that of all control groups but not from patients with an HII of 1·9 or less (p=0·49). Four male and one female HLA-identical siblings of patients were identified in five different families; four had normal serum iron concentrations and one, a 66-year-old alcoholic man, had increased serum iron and transferrin saturation with a normal serum ferritin concentration (286 µg/L)—which does not support a diagnosis of homozygous GH.
Discussion Our study describes a new, non-HLA-linked, iron-overload syndrome in which transferrin saturation is normal and there is close association with various metabolic disorders. In most iron-overloaded patients, a raised serum iron concentration and transferrin saturation precede the increase in serum ferritin concentration, whatever the mechanism of iron excess.6,12 We do not think that serum iron concentrations or transferrin saturations were falsely underestimated in our patients because both the haemoglobin concentration and inflammatory markers were normal. However, for an identical serum ferritin concentration, LIC and removed iron were 2·5 times lower in these patients than in matched controls, which suggests an iron excess much less than that predicted by the serum ferritin. Screening for iron overload by transferrin saturation alone would misdiagnose this type of overload, and
Vol 349 • January 11, 1997
THE LANCET
therefore only direct methods8,13 (liver histology2 or magnetic-resonance imaging9), should be used to quantify liver iron stores. The iron-overload syndrome described here cannot be placed within the current classification of iron-overload disorders.3,14,15 Common causes of secondary iron overload were carefully excluded in these patients—especially sideroblastic anaemia, chronic haemolysis, porphyria cutanea tarda, chronic hepatitis C,16 and steatohepatitis.17 In addition, weekly venesection was well tolerated by most treated patients. Finally, chronic alcoholism was excluded; not only by history but also on chemical and histological data. The slight increase in γ-glutamyl transpeptidase (GGT) observed in some of our patients was probably related to obesity.18 Although 34% of these patients had HII above 1·9, we do not think that the GH gene is involved, since none of the confirmed GH homozygote controls had a normal transferrin saturation. The prevalence of HLA A3 (which in a given population may be used as a reference for the GH gene)19 did not differ between our patients and the general population and was significantly lower than in both homozygote and heterozygote controls. Also, none of the five HLA-identical siblings had clinical or biochemical features of iron excess. In GH families, such features can arise because of genetic recombination, but Powell et al20 have shown that this is rare. Thus, it is unlikely that genetic recombination could account for all our patients. The identification of this non-HLA-linked iron-overload syndrome has important implications. Because this condition is likely to be misdiagnosed as early or underexpressed GH, its recognition should help to avoid costly and potentially disruptive family-screening procedures and, in GH gene research programmes, consideration of people with atypical phenotypes. Finally, our findings suggest that the prevalence of the GH gene,19 which relies heavily on the prevalence of phenotypic features of iron overload, may be overestimated. This new condition is a homogeneous entity, characterised by a normal transferrin saturation, hyperferritinaemia, and mild to moderate iron overload. Furthermore, almost all patients had a concomitant metabolic disorder (table 2) associated with the insulin resistance syndrome.21,22 Thus, it is tempting to add iron excess to this cluster of metabolic aberrations. Support for this theory includes the finding of increased serum ferritin concentrations in diabetes23 and people who eat a lot of meat24 and the current debate about the putative role of body-iron excess as a risk factor for coronary heart disease.25 Iron excess was usually mild to moderate in our patients. Is this excess clinically important and should it be treated? Experimental,26 epidemiological,27 and clinical studies28 suggest a role of increased body-iron stores in human carcinogenesis. Thus, treatment of iron excess, whatever its severity, may be an important goal for general health. The description of this new syndrome suggests that the diagnostic criteria of iron-overload syndromes should be modified. Raised serum ferritin concentration and transferrin saturation are both needed to detect all types of iron overload. Just as idiopathic iron overload does not always correspond to GH, an HII cut-off point of 1·9 does not discriminate between the different iron-overload conditions. Also, the data suggest that there may be a link between increased liver-iron stores and various common metabolic disorders. The exact nature and significance of this link
Vol 349 • January 11, 1997
remains to be clarified. This study was supported by grants from the Association pour la Recherche contre le Cancer and the Programme Hospitalier de Recherche Clinique 1994.
References 1
2
3
4
5
6
7 8
9
10
11
12 13
14
15 16
17
18
19 20
21 22 23
24
25 26
27 28
Feder JN, Gnirke A, Thomas W, et al. A novel MHC class I-like gene is mutated in patients with hereditary haemochromatosis. Nat Genet 1996; 13: 399–408. 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: 2050–59. Brissot P, Deugnier Y. Genetic haemochromatosis. In: McIntyre N, Benhamou J, Bircher J, Rizzetto M, Rodes J, eds. Oxford textbook of clinical hepatology. Oxford: Oxford University Press. 1991: 948–58. 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: 24–29. Summers KM, Halliday JW, Powell LW. Identification of homozygous hemochromatosis subjects by measurement of hepatic iron index. Hepatology 1990; 12: 20–25. Edwards CQ, Griffen LM, Goldgar D, Drummond C, Skolnick MH, Kushner JP. Prevalence of hemochromatosis among 11 065 presumably healthy blood donors. N Engl J Med 1988; 318: 1355–62. Barry M, Sherlock S. Measurement of liver concentration in needlebiopsy specimens. Lancet 1971; i: 100–03. Brissot P, Bourel M, Herry D, et al. Assessment of liver iron content in 271 patients: a reevaluation of direct and indirect methods. Gastroenterology 1981; 80: 557–65. Gandon Y, Guyader D, Heautot JF, et al. Hemochromatosis: diagnosis and quantification of liver iron with gradient-echo MR imaging. Radiology 1994; 193: 533–38. Rolland-Cachera MF, Cole TJ, Sempi M, Tichet J, Rossignol C, Charraud A. Body mass index variations: centiles from birth to 87 years. Eur J Clin Nutr 1991; 45: 13–21. Simon M, Bourel M, Genetet B, Fauchet R, Edan G, Brissot P. Idiopathic hemochromatosis and iron overload in alcoholic liver disease: differentiation by HLA phenotype. Gastroenterology 1977; 73: 655–58. Peto TEA, Pippard MJ, Weatherall DJ. Iron overload in mild sideroblastic anaemias. Lancet 1983; i: 375–78. Bonkovsky HL, Slaker DP, Bills EB, Wolf DC. Usefulness and limitations of laboratory and hepatic imaging studies in iron-storage disease. Gastroenterology 1990; 99: 1079–91. Tavill AS, Sharma BK, Bacon BR. Iron and the liver: genetic hemochromatosis and other hepatic overload disorders. Prog Liver Dis 1990; 9: 281–305. Bacon BR. Causes of iron overload. N Engl J Med 1992; 326: 126–27. Di Bisceglie AM, Axiotis CA, Hoofnagle JH, Bacon BR. Measurements of iron status in patients with chronic hepatitis. Gastroenterology 1992; 102: 2108–13. Bacon BR, Farahvash MJ, Janney CG, Neuschwander-Tetri BA. Nonalcoholic steatohepatitis: an expanded clinical entity. Gastroenterology 1994; 107: 1103–09. Nilssen O, Forde OH. Seven-year longitudinal population study of change in gamma-glutamyltransferase: the Tromso Study. Am J Epidemiol 1994; 139: 787–92. Simon M, Brissot P. The genetics of haemochromatosis. J Hepatol 1988; 6: 116–24. Powell LW, Summers KM, Board PG, Axelsen E, Webb S, Halliday JW. Expression of hemochromoatosis in homozygous subjects: implications for early diagnosis and prevention. Gastroenterology 1990; 98: 1625–32. Reaven G. Role of insulin resistance in human disease. Diabetes 1988; 37: 1495–507. Després JP, Marette A. Relation of components of insulin resistance syndrome to coronary disease risk. Curr Opin Lipidol 1994; 5: 274–89. Dinneen SF, O’Mahony MS, O’Brien T, Cronin CC, Murray DM, O’Sullivan DJ. Serum ferritin in newly diagnosed and poorly controlled diabetes mellitus. Ir J Med Sci 1992; 161: 636–38. Leggett BA, Brown NN, Bryant SJ, Duplock L, Powell LW, Halliday JW. Factors affecting the concentration of ferritin in serum in a healthy Australian population. Clin Chem 1990; 36: 1350–55. Conrad ME. Excess iron and catastrophic illness. Am J Hematol 1993; 43: 234–36. Hann HWL, Stahlhut MW, Rubin R, Maddrey WC. Antitumor effect of deferoxamine on human hepatocellular carcinoma growing in athymic nude mice. Cancer 1992; 70: 2051–56. Stevens RG, Jones DY, Micozzi MS, Taylor PR. Body iron stores and the risk of cancer. N Engl J Med 1988; 319: 1047–52. Turlin B, Juguet F, Moirand R, et al. Increased liver iron stores in non-
97