- Email: [email protected]
Serum iron indices as a measure of iron deposits in chronic hepatitis C
Clinica Chimica Acta 304 (2001) 49–55 www.elsevier.com / locate / clinchim
Serum iron indices as a measure of iron deposits in chronic hepatitis C Carlo Fabris, Pierluigi Toniutto, Cathryn A. Scott, Edmondo Falleti, Claudio Avellini, Monica Del Forno, Martina Mattiuzzo, Barbara Branca, Mario Pirisi* Dipartimento di Patologia e Medicina Sperimentale e Clinica ( CF, PT, EF, MDF, MM, BB, MP); Istituto di Anatomia Patologica ( CAS, CA); University of Udine, 1 Piazzale Santa Maria della Misericordia, 33100 Udine, Italy Received 17 May 2000; received in revised form 22 September 2000; accepted 2 October 2000
Abstract Serum iron indices are believed to be elevated in patients with hepatitis C virus (HCV) infection in connection to the presence of hepatic inflammation, though this hypothesis has never been formally tested. We studied 69 consecutive, unselected anti HCV antibody positive patients, aged 14 to 70 years. Iron, transferrin saturation and ferritin were measured in fasting serum samples. Histologically detectable iron (HDI) as well as histologic grading and staging were estimated semiquantitatively in liver biopsy samples. The median values for serum iron, transferrin saturation and serum ferritin were 24 mmol / l (range, 8–61), 29 percent (range, 6–77) and 170 mg / l (range, 1–954), respectively. At univariate analysis, all three serum iron indices were positively correlated with grading and staging scores, as well as with HDI in the liver; only serum iron was positively correlated with transaminases. At multivariate analysis, independent associations were found between serum iron and the grading score; ferritin and sinusoidal and portal HDI; transferrin saturation and total hepatic HDI. In conclusion, in hepatitis C, serum iron reflects the degree of current hepatic inflammation and necrosis, whereas the extent of progressive deposition of iron in sites of fibrosis is best reflected by serum ferritin. Transferrin saturation is the best predictor of the status of hepatic iron deposits. 2001 Elsevier Science B.V. All rights reserved. Keywords: Hepatitis C; Iron; Ferritin; Transferrin saturation; Fibrosis
1. Introduction During the past few years, much attention has been given to the possible role of iron in modulating the course of viral hepatitis [1]. In particular, patients with lesser amounts of hepatic iron have been shown
*Corresponding author. Tel.: 139-432-559-801; fax: 139-43242-097. E-mail address: [email protected] (M. Pirisi).
to respond better to interferon therapy for chronic hepatitis C virus (HCV) infection than those with larger amounts [2,3]. The generally accepted gold standard for evaluating iron status in the liver has been the direct measurement of hepatic iron concentrations on liver biopsy specimens [4]. When hemochromatosis is ruled out, semiquantitative assessment of the amount and distribution of iron noted microscopically on liver biopsy sections stained for iron is a suitable alternative [5]. In fact, the latter method offers the advantage of enabling the evaluation of the lobular
0009-8981 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 00 )00397-1
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
50
and cellular distribution of histologically detectable iron (HDI). Indirect methods of assessing iron status include the measurement of serum ferritin and of serum iron with calculation of transferrin saturation. They have been shown to be predictive of non-response to interferon in chronic hepatitis C and have the great advantage of being non-invasive and easily repeatable [6,7]. Unfortunately, these methods are considered not reliable for estimating hepatic or total body iron stores in the presence of inflammation [8]. However, studies addressing the issue of how, in patients with hepatitis C, the elevation of serum iron indices relates to the zonal distribution of hepatic iron, to the degree of portal and acinar inflammation, and to the development of fibrosis, are lacking. To provide this information, we performed the present study.
2. Methods
2.1. Patients Sixty-nine consecutive, unselected, patients, aged 14 to 70 years, referred to our Institution for complete diagnostic work-up because of suspected liver disease, were studied retrospectively. All were anti HCV antibody positive and underwent needle biopsy of the liver. The demographic and clinical characteristics of these patients are presented in Table 1. No patient had evidence of other causes of liver disease (viruses other than HCV, autoimmunity, toxins and drugs, genetic), secondary iron overload Table 1 Demographic and clinical characteristics of the study population a N Age (years) Male:female, N Previous alcohol consumption (.10 g / d) no yes Serum HCV RNA positive Genotype 1 Genotype 2 Genotype 3 a
69 48.0612.1 42:27 23 46 59 38 18 3
Data are presented as frequencies (categorical variables), mean6S.D. (continuous variables with normal distribution).
(dyserythropoietic states, increased oral intake of iron, porphyria cutanea tarda), parenteral iron overload (transfusions, hemodyalisis) and none had received interferon therapy prior to biopsy. Patients with past history of alcohol consumption were not excluded; however, all patients had remained abstinent for at least 6 months before liver biopsy and biochemical measures were performed.
2.2. Histology All liver biopsy specimens were obtained by the Menghini technique. Biopsies were at least 15 mm long, and included at least 4 portal tracts; they were fixed in formalin and embedded in paraffin. Slides were stained with hematoxylin and eosin, PAS with and without diastase predigestion, Gomori’s reticulin stain and Perl’s Prussian blue method. The slides were evaluated by two experienced pathologists (CAS, CA) without prior knowledge of clinical and laboratory data of patients. Presence of HDI deposits was estimated according to the system of Barton et al. [5] which evaluates iron separately in hepatocytes, sinusoids and portal tracts. Hepatocyte and sinusoidal scoring is based on the separate estimation of presence, but not quantity, of HDI deposits in each liver zone and on the extent (none, less than 1 / 3, between 1 / 3 and 2 / 3, more than 2 / 3) of total zones involved and ranges from 0 to 12. Portal tract scoring is evaluated separately in connective tissue, vascular walls and bile duct cells on the basis of the relative proportion of portal areas involved (none, less than 1 / 3, between 1 / 3 and 2 / 3, more than 2 / 3); scores range from 0 to 12. Total iron score is given by the sum of each score area. Grading and staging scores were evaluated according to the method of Ishak et al. [9].
2.3. Biohumoral determinations Anti HCV antibodies were tested in serum by a third generation enzyme immunoassay (Ortho Diagnostics, Raritan, NJ). Positive test results were always confirmed by immunoblotting (HCV Matrix, Abbott, Abbott Park, IL). Iron was measured by a chromogenic assay (Instrumentation Laboratory SpA, Milano, Italy). The mean coefficients of variation (CV) for this assay are: intra-assay CV, 1.4
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
percent; inter-assay CV, 2.9 percent; reference values in our laboratory are 11.6–30.4 mmol / l for males and 8.9–30.4 mmol / l for females. Transferrin was measured by means of a nephelometric method (Dade Behring Marburg GmbH, Marburg, Germany). The mean CV for this assay are: Intra-assay CV, 2.7 percent; inter-assay CV, 2.3 percent; reference values in our laboratory are 2.0–3.6 g / l. Total iron binding capacity was calculated from transferrin according to the following formula [8]: Total iron binding capacity (mg / dl) 5 Serum transferrin (mg / l) 3 0.1397 Transferrin saturation was calculated according to the following formula: Serum iron (mg / dl) ]]]]]]]]]] 3 100 Total iron binding capacity (mg / dl) Ferritin was determined by a microparticle enzyme immunoassay (Ferritin, Axsym System, Abbott Laboratories, Abbott Park, IL). The mean CV for this assay are: Intra-assay CV, 3.8 percent; inter-assay CV, 2.8 percent; reference values in our laboratory are 18.7–323.0 mg / l for males and 6.9–282.5 mg / l for females.
2.4. Molecular analysis of missense mutations in the HFE gene The presence of the Cys282Tyr and His63Asp missense mutations in the HFE gene was verified in patients with transferrin saturation $50 percent, by means of restriction fragment length polymorphism of polymerase chain reaction (PCR) products. Details of the methods used can be found elsewhere [10].
2.5. Statistical analysis Statistical analysis of the data was performed by means of the BMDP/ Dynamic statistical software package, release 7.0 (Statistical Software Ltd, Cork, Ireland). Shapiro and Wilk’s W test was applied to test normality of continuous data. The associations between variables were analyzed by performing Pearson’s chi-square tests and by calculating Spearman’s rank correlation coefficients. Differences among continuous variables were evaluated by
51
Mann–Whitney tests. Logistic regression analysis with a stepwise forward approach was applied to determine the variables independently associated with elevation of serum iron indices. Values of P were considered significant when equal to or below 0.05, and were two-tailed.
3. Results In the studied population, the median value for serum iron was 24 mmol / l (range, 8–61), the median value for transferrin saturation was 29 percent (range, 6–77) and the median value for serum ferritin was 170 mg / l (range, 1–954). Male patients had significantly higher serum iron (median in male 27 mmol / l, range 15–61; median in female 21 mmol / l, range 8–43, P50.006), transferrin saturation (median in male 32 percent, range 18–77; median in female 24 percent, range 6–58, P50.004) and serum ferritin (median in male 229 mg / l, range 76–954; median in female 108 mg / l, range 1–851, P,0.001) than female patients (Fig. 1). There was no significant correlation between any of these 3 parameters and the age of patients. Eight patients had transferrin saturation values $50 percent; one of them was heterozygote for Cys282Tyr mutation, one was homozygote and two were heterozygotes for the His63Asp mutation. There were neither Cys282Tyr homozygotes nor Cys282Tyr / His63Asp compound heterozygotes. Table 2 shows the correlation coefficients between serum iron, transferrin saturation and serum ferritin levels on the one hand, and iron scores in different liver compartments, on the other. The correlations between serum iron indices and HDI in the hepatocytes and sinusoids, which were generally good, were very strict for hepatocytes and sinusoids located in zone 1 of the liver acinus. There was a positive correlation between the serum concentrations of iron and those of aspartate aminotransferase (Spearman’s correlation coefficient 0.25, P,0.05). There was no significant correlation between serum aspartate aminotransferase and either transferrin saturation or serum ferritin. Serum iron indices did not differ among patients grouped according to HCV genotype. Serum ferritin was higher in the 46 drinkers than in the 23 non-drinkers (median in drinkers 211 mg / l,
52
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
Fig. 1. Scatterplot of individual values of serum iron, transferrin saturation and ferritin in 69 patients with chronic hepatitis C, grouped according to gender. Bars indicate group medians; *P,0.01 and **P,0.001 by Mann–Whitney tests.
range 33–954; median in non-drinkers 127 mg / l, range 1–851; P50.006); serum iron and transferrin saturation had a trend to be higher in drinkers than in non-drinkers, but this did not reach statistical significance. Table 3 presents the analysis of the associations between the serum iron indices and the grading and staging scores of patients. Serum iron, transferrin saturation and serum ferritin were all positively
correlated with grading and staging scores. The parameter with the strongest association with grading was serum iron; the parameter with the strongest association with staging was serum ferritin. Multivariate analysis (including the following eight predictor variables: Age, gender, drinking habits, grading and staging scores, HDI in hepatocytes, in sinusoids and in portal triads) was performed to identify the factors independently associ-
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55 Table 2 Correlations between patterns of histologically detectable iron in the liver and serologic indices of iron status a Serum iron
Transferrin saturation
Serum ferritin
Portal triads Connective tissue Vascular walls Total
0.39 B 0.01 D 0.31 C
0.48 A 0.09 D 0.42 A
0.46 A 0.16 D 0.47 A
Hepatocytes Zone 1 Zone 2 Zone 3 Total
0.45 A 0.43 A 0.35 B 0.46 A
0.55 A 0.51 A 0.37 B 0.55 A
0.58 A 0.55 A 0.39 B 0.60 A
Sinusoids Zone 1 Zone 2 Zone 3 Total
0.37 B 0.39 A 0.34 B 0.43 A
0.35 B 0.37 B 0.28 C 0.40 A
0.56 A 0.53 A 0.44 A 0.60 A
53
ated with a high serum iron concentration (.24 mmol / l), a high transferrin saturation (.29%) and a high ferritin value (.170 mg / l). The results are presented in Table 4.
Iron in:
4. Discussion
a
Data are expressed as Spearman r coefficients. A5P,0.001; B5P,0.005; C5P,0.05; D5P.0.10.
Table 3 Correlations between the grading and staging scores (according to Ishak et al. [9]) and the biochemical parameters of iron status
Serum iron Transferrin saturation Ferritin
Grading
P value
Staging
P value
0.325 0.254 0.285
0.006 0.035 0.018
0.262 0.275 0.325
0.030 0.022 0.006
The present study has been performed on a consecutive, unselected series of patients with clinical, histologic and demographic features of hepatitis C in keeping with those commonly found in Western countries. In agreement with what has been noted by others [11–13], we confirm that the degree of elevation of iron indices in chronic hepatitis C is mild to moderate, rarely reaching the hemochromatotic range. Differences related to gender and to alcohol consumption are maintained, whereas there is no clear correlation between serum iron indices and aging. In patients with hepatitis C, both ferritin and transferrin saturation are excellent markers for presence of HDI in hepatocytes located in zone 1; this correlation becomes progressively less strict in zone 2 and zone 3. The process of iron accretion in the liver involves the progressive recruitment of storage cells from zone 1 to zone 3 [14]. Thus, the zonal distribution of hepatocyte iron indicates that, not only in normal subjects but also in hepatitis C patients, HDI reflects the storage function of the organ, to which corresponds a proportional increase of the serum iron indices. Besides being in the hepatocytes, HDI in the liver appears to be located
Table 4 Multivariate analysis of the factors associated with high serum iron, high transferrin saturation and high serum ferritin in patients with chronic hepatitis C
Age Gender Drinking habits Grading score a Staging score a HDI in hepatocytes HDI in sinusoids HDI in portal triads a
Serum iron .24 mmol / l
Transferrin saturation .29 percent
Serum ferritin .170 mg / l
NE NE NE 0.014 NE 0.003 NE NE
NE NE NE NE NE ,0.001 NE NE
NE 0.001 NE NE NE NE ,0.001 0.047
According to Ishak et al. [9]. HDI5histologically detectable iron; NE5not entered.
54
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
commonly in the cells lining the sinusoids and in the connective tissue of portal tracts. The correlations between serum iron indices and HDI do not hold in other liver compartments, such as bile ducts. In the absence of overt iron overload, the interpretation traditionally offered for the elevation of serum iron indices in hepatitis C has been that they reflect the release of iron from damaged liver cells [1]. However, this issue is open to scrutiny. In fact, to our knowledge, there has been to date a single study in which HDI in the liver, evaluated semiquantitatively, has been related to serum iron indices in patients with HCV infection [13]. Unfortunately, the method which was used by these authors is relatively insensitive for non-hemochromatotic states of iron overload and no attempt was made to verify zonal distribution of iron. Furthermore, grading and staging scores appeared not to correlate with HDI, a finding that does not seem to give support to the hypothesis that, in these patients, iron is released from the hepatocytes because of inflammatory damage. In contrast, we observed a good correlation between the levels of serum iron indices and the measures of inflammatory activity (grading) and of progression of fibrosis (staging). In particular, serum iron bears a strict relationship with the inflammatory activity, whereas serum ferritin is strictly related to staging. Moreover, we found a significant correlation between the serum concentrations of aspartate aminotransferase and those of iron, but not those of ferritin. In multivariate analysis, ferritin is the only index independently associated with gender and with sinusoidal and portal HDI, while serum iron appears the only index independently associated with the grading score. In other words, in hepatitis C serum iron levels may reflect the degree of current hepatic inflammation and necrosis, whereas ferritin levels may reflect the extent of progressive deposition of iron in sites of scarring, where iron deposition itself may be contributing to the development of fibrosis [15]. According to this scenario, interface hepatitis with its sequelae (sinusoidal capillarization and microshunting) is the major factor leading to iron deposition and fibrotic changes in chronic hepatitis C: It parallels the distribution of HDI and induces the mild elevation of serum iron indices. In contrast, transferrin saturation appears to be independently associated only with HDI in the liver, therefore
resulting the best predictor of the status of hepatic iron deposits in these patients. Measures of serum iron, transferrin saturation and serum ferritin are commonly included in the evaluation of patients with chronic liver disease due to hepatitis C. Hopefully, better knowledge of the associations between specific aspects of the morphologic picture and individual iron indices in this condition may make the latter more informative clinically.
Acknowledgements Supported in part by grants from Italian Government Institutions (Ministero della Universita` e della Ricerca Scientifica) and from Udinese per la vita onlus.
References [1] Bonkovsky HL, Banner BF, Rothman AL. Iron and chronic viral hepatitis. Hepatology 1997;25:759–68. [2] Van Thiel DH, Friedlander L, Fagiuoli S, Wright HI, Irish W, Gvaler JS. Response to interferon alpha therapy is influenced by the iron content of the liver. J Hepatol 1994;20:410–5. [3] Clemente MG, Congia M, Lai ME et al. Effect of iron overload on the response to recombinant interferon-alfa treatment in transfusion-dependent patients with thalassemia major and chronic hepatitis C. J Pediatr 1994;125:123–8. [4] Kowdley KV, Trainer TD, Saltzman JR et al. Utility of hepatic iron index in American patients with hereditary hemochromatosis: a multicenter study. Gastroenterology 1997;113:1270–7. [5] Barton AL, Banner BF, Cable EE, Bonkovsky HL. Distribution of iron in the liver predicts the response of chronic hepatitis C infection to interferon therapy. Am J Clin Pathol 1995;103:419–24. [6] Arber N, Moshkowitz M, Konikoff FM. Elevated serum iron predicts poor response to interferon treatment in patients with chronic HCV infection. Dig Dis Sci 1995;40:2431–3. [7] Banner BF, Barton AL, Cable EE, Smith L, Bonkovsky HL. A detailed analysis of the Knodell score and other histologic parameters as predictors of response to interferon therapy in chronic hepatitis C. Mod Pathol 1995;8:232–8. [8] Wallach J. Interpretation of diagnostic tests: a synopsis of laboratory medicine, 5th ed, Boston: Little, Brown and Company, 1992. [9] Ishak K, Baptista A, Bianchi L et al. Histological grading and staging of chronic hepatitis. J Hepatol 1995;22:696–9.
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55 [10] Pirisi M, Toniutto P, Uzzau A et al. Carriage of HFE mutations and outcome of surgical resection for hepatocellular carcinoma in cirrhotic patients. Cancer 2000;89:297– 302. [11] Di Bisceglie AM, Axiotis CA, Hoofnagle JH, Bacon BR. Measurements of iron status in patients with chronic hepatitis. Gastroenterology 1992;102:2108–13. [12] Arber N, Konikoff FM, Moshkowitz M et al. Increased serum iron and iron saturation without liver iron accumulation distinguish chronic hepatitis C from other chronic liver diseases. Dig Dis Sci 1994;39:2656–9.
55
[13] Haque S, Chandra B, Gerber MA, Lok AS. Iron overload in patients with chronic hepatitis C: a clinicopathologic study. Hum Pathol 1996;27:1277–81. [14] Tavill AS. Hemochromatosis. In: Schiff L, Schiff ER, editors, Diseases of the liver, 7th ed, Philadelphia: J.B. Lippincott Company, 1993, pp. 669–91. [15] Pietrangelo A. Iron, oxidative stress, and liver fibrogenesis. J Hepatol 1998;28:8–13.
Serum iron indices as a measure of iron deposits in chronic hepatitis C Carlo Fabris, Pierluigi Toniutto, Cathryn A. Scott, Edmondo Falleti, Claudio Avellini, Monica Del Forno, Martina Mattiuzzo, Barbara Branca, Mario Pirisi* Dipartimento di Patologia e Medicina Sperimentale e Clinica ( CF, PT, EF, MDF, MM, BB, MP); Istituto di Anatomia Patologica ( CAS, CA); University of Udine, 1 Piazzale Santa Maria della Misericordia, 33100 Udine, Italy Received 17 May 2000; received in revised form 22 September 2000; accepted 2 October 2000
Abstract Serum iron indices are believed to be elevated in patients with hepatitis C virus (HCV) infection in connection to the presence of hepatic inflammation, though this hypothesis has never been formally tested. We studied 69 consecutive, unselected anti HCV antibody positive patients, aged 14 to 70 years. Iron, transferrin saturation and ferritin were measured in fasting serum samples. Histologically detectable iron (HDI) as well as histologic grading and staging were estimated semiquantitatively in liver biopsy samples. The median values for serum iron, transferrin saturation and serum ferritin were 24 mmol / l (range, 8–61), 29 percent (range, 6–77) and 170 mg / l (range, 1–954), respectively. At univariate analysis, all three serum iron indices were positively correlated with grading and staging scores, as well as with HDI in the liver; only serum iron was positively correlated with transaminases. At multivariate analysis, independent associations were found between serum iron and the grading score; ferritin and sinusoidal and portal HDI; transferrin saturation and total hepatic HDI. In conclusion, in hepatitis C, serum iron reflects the degree of current hepatic inflammation and necrosis, whereas the extent of progressive deposition of iron in sites of fibrosis is best reflected by serum ferritin. Transferrin saturation is the best predictor of the status of hepatic iron deposits. 2001 Elsevier Science B.V. All rights reserved. Keywords: Hepatitis C; Iron; Ferritin; Transferrin saturation; Fibrosis
1. Introduction During the past few years, much attention has been given to the possible role of iron in modulating the course of viral hepatitis [1]. In particular, patients with lesser amounts of hepatic iron have been shown
*Corresponding author. Tel.: 139-432-559-801; fax: 139-43242-097. E-mail address: [email protected] (M. Pirisi).
to respond better to interferon therapy for chronic hepatitis C virus (HCV) infection than those with larger amounts [2,3]. The generally accepted gold standard for evaluating iron status in the liver has been the direct measurement of hepatic iron concentrations on liver biopsy specimens [4]. When hemochromatosis is ruled out, semiquantitative assessment of the amount and distribution of iron noted microscopically on liver biopsy sections stained for iron is a suitable alternative [5]. In fact, the latter method offers the advantage of enabling the evaluation of the lobular
0009-8981 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 00 )00397-1
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
50
and cellular distribution of histologically detectable iron (HDI). Indirect methods of assessing iron status include the measurement of serum ferritin and of serum iron with calculation of transferrin saturation. They have been shown to be predictive of non-response to interferon in chronic hepatitis C and have the great advantage of being non-invasive and easily repeatable [6,7]. Unfortunately, these methods are considered not reliable for estimating hepatic or total body iron stores in the presence of inflammation [8]. However, studies addressing the issue of how, in patients with hepatitis C, the elevation of serum iron indices relates to the zonal distribution of hepatic iron, to the degree of portal and acinar inflammation, and to the development of fibrosis, are lacking. To provide this information, we performed the present study.
2. Methods
2.1. Patients Sixty-nine consecutive, unselected, patients, aged 14 to 70 years, referred to our Institution for complete diagnostic work-up because of suspected liver disease, were studied retrospectively. All were anti HCV antibody positive and underwent needle biopsy of the liver. The demographic and clinical characteristics of these patients are presented in Table 1. No patient had evidence of other causes of liver disease (viruses other than HCV, autoimmunity, toxins and drugs, genetic), secondary iron overload Table 1 Demographic and clinical characteristics of the study population a N Age (years) Male:female, N Previous alcohol consumption (.10 g / d) no yes Serum HCV RNA positive Genotype 1 Genotype 2 Genotype 3 a
69 48.0612.1 42:27 23 46 59 38 18 3
Data are presented as frequencies (categorical variables), mean6S.D. (continuous variables with normal distribution).
(dyserythropoietic states, increased oral intake of iron, porphyria cutanea tarda), parenteral iron overload (transfusions, hemodyalisis) and none had received interferon therapy prior to biopsy. Patients with past history of alcohol consumption were not excluded; however, all patients had remained abstinent for at least 6 months before liver biopsy and biochemical measures were performed.
2.2. Histology All liver biopsy specimens were obtained by the Menghini technique. Biopsies were at least 15 mm long, and included at least 4 portal tracts; they were fixed in formalin and embedded in paraffin. Slides were stained with hematoxylin and eosin, PAS with and without diastase predigestion, Gomori’s reticulin stain and Perl’s Prussian blue method. The slides were evaluated by two experienced pathologists (CAS, CA) without prior knowledge of clinical and laboratory data of patients. Presence of HDI deposits was estimated according to the system of Barton et al. [5] which evaluates iron separately in hepatocytes, sinusoids and portal tracts. Hepatocyte and sinusoidal scoring is based on the separate estimation of presence, but not quantity, of HDI deposits in each liver zone and on the extent (none, less than 1 / 3, between 1 / 3 and 2 / 3, more than 2 / 3) of total zones involved and ranges from 0 to 12. Portal tract scoring is evaluated separately in connective tissue, vascular walls and bile duct cells on the basis of the relative proportion of portal areas involved (none, less than 1 / 3, between 1 / 3 and 2 / 3, more than 2 / 3); scores range from 0 to 12. Total iron score is given by the sum of each score area. Grading and staging scores were evaluated according to the method of Ishak et al. [9].
2.3. Biohumoral determinations Anti HCV antibodies were tested in serum by a third generation enzyme immunoassay (Ortho Diagnostics, Raritan, NJ). Positive test results were always confirmed by immunoblotting (HCV Matrix, Abbott, Abbott Park, IL). Iron was measured by a chromogenic assay (Instrumentation Laboratory SpA, Milano, Italy). The mean coefficients of variation (CV) for this assay are: intra-assay CV, 1.4
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
percent; inter-assay CV, 2.9 percent; reference values in our laboratory are 11.6–30.4 mmol / l for males and 8.9–30.4 mmol / l for females. Transferrin was measured by means of a nephelometric method (Dade Behring Marburg GmbH, Marburg, Germany). The mean CV for this assay are: Intra-assay CV, 2.7 percent; inter-assay CV, 2.3 percent; reference values in our laboratory are 2.0–3.6 g / l. Total iron binding capacity was calculated from transferrin according to the following formula [8]: Total iron binding capacity (mg / dl) 5 Serum transferrin (mg / l) 3 0.1397 Transferrin saturation was calculated according to the following formula: Serum iron (mg / dl) ]]]]]]]]]] 3 100 Total iron binding capacity (mg / dl) Ferritin was determined by a microparticle enzyme immunoassay (Ferritin, Axsym System, Abbott Laboratories, Abbott Park, IL). The mean CV for this assay are: Intra-assay CV, 3.8 percent; inter-assay CV, 2.8 percent; reference values in our laboratory are 18.7–323.0 mg / l for males and 6.9–282.5 mg / l for females.
2.4. Molecular analysis of missense mutations in the HFE gene The presence of the Cys282Tyr and His63Asp missense mutations in the HFE gene was verified in patients with transferrin saturation $50 percent, by means of restriction fragment length polymorphism of polymerase chain reaction (PCR) products. Details of the methods used can be found elsewhere [10].
2.5. Statistical analysis Statistical analysis of the data was performed by means of the BMDP/ Dynamic statistical software package, release 7.0 (Statistical Software Ltd, Cork, Ireland). Shapiro and Wilk’s W test was applied to test normality of continuous data. The associations between variables were analyzed by performing Pearson’s chi-square tests and by calculating Spearman’s rank correlation coefficients. Differences among continuous variables were evaluated by
51
Mann–Whitney tests. Logistic regression analysis with a stepwise forward approach was applied to determine the variables independently associated with elevation of serum iron indices. Values of P were considered significant when equal to or below 0.05, and were two-tailed.
3. Results In the studied population, the median value for serum iron was 24 mmol / l (range, 8–61), the median value for transferrin saturation was 29 percent (range, 6–77) and the median value for serum ferritin was 170 mg / l (range, 1–954). Male patients had significantly higher serum iron (median in male 27 mmol / l, range 15–61; median in female 21 mmol / l, range 8–43, P50.006), transferrin saturation (median in male 32 percent, range 18–77; median in female 24 percent, range 6–58, P50.004) and serum ferritin (median in male 229 mg / l, range 76–954; median in female 108 mg / l, range 1–851, P,0.001) than female patients (Fig. 1). There was no significant correlation between any of these 3 parameters and the age of patients. Eight patients had transferrin saturation values $50 percent; one of them was heterozygote for Cys282Tyr mutation, one was homozygote and two were heterozygotes for the His63Asp mutation. There were neither Cys282Tyr homozygotes nor Cys282Tyr / His63Asp compound heterozygotes. Table 2 shows the correlation coefficients between serum iron, transferrin saturation and serum ferritin levels on the one hand, and iron scores in different liver compartments, on the other. The correlations between serum iron indices and HDI in the hepatocytes and sinusoids, which were generally good, were very strict for hepatocytes and sinusoids located in zone 1 of the liver acinus. There was a positive correlation between the serum concentrations of iron and those of aspartate aminotransferase (Spearman’s correlation coefficient 0.25, P,0.05). There was no significant correlation between serum aspartate aminotransferase and either transferrin saturation or serum ferritin. Serum iron indices did not differ among patients grouped according to HCV genotype. Serum ferritin was higher in the 46 drinkers than in the 23 non-drinkers (median in drinkers 211 mg / l,
52
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
Fig. 1. Scatterplot of individual values of serum iron, transferrin saturation and ferritin in 69 patients with chronic hepatitis C, grouped according to gender. Bars indicate group medians; *P,0.01 and **P,0.001 by Mann–Whitney tests.
range 33–954; median in non-drinkers 127 mg / l, range 1–851; P50.006); serum iron and transferrin saturation had a trend to be higher in drinkers than in non-drinkers, but this did not reach statistical significance. Table 3 presents the analysis of the associations between the serum iron indices and the grading and staging scores of patients. Serum iron, transferrin saturation and serum ferritin were all positively
correlated with grading and staging scores. The parameter with the strongest association with grading was serum iron; the parameter with the strongest association with staging was serum ferritin. Multivariate analysis (including the following eight predictor variables: Age, gender, drinking habits, grading and staging scores, HDI in hepatocytes, in sinusoids and in portal triads) was performed to identify the factors independently associ-
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55 Table 2 Correlations between patterns of histologically detectable iron in the liver and serologic indices of iron status a Serum iron
Transferrin saturation
Serum ferritin
Portal triads Connective tissue Vascular walls Total
0.39 B 0.01 D 0.31 C
0.48 A 0.09 D 0.42 A
0.46 A 0.16 D 0.47 A
Hepatocytes Zone 1 Zone 2 Zone 3 Total
0.45 A 0.43 A 0.35 B 0.46 A
0.55 A 0.51 A 0.37 B 0.55 A
0.58 A 0.55 A 0.39 B 0.60 A
Sinusoids Zone 1 Zone 2 Zone 3 Total
0.37 B 0.39 A 0.34 B 0.43 A
0.35 B 0.37 B 0.28 C 0.40 A
0.56 A 0.53 A 0.44 A 0.60 A
53
ated with a high serum iron concentration (.24 mmol / l), a high transferrin saturation (.29%) and a high ferritin value (.170 mg / l). The results are presented in Table 4.
Iron in:
4. Discussion
a
Data are expressed as Spearman r coefficients. A5P,0.001; B5P,0.005; C5P,0.05; D5P.0.10.
Table 3 Correlations between the grading and staging scores (according to Ishak et al. [9]) and the biochemical parameters of iron status
Serum iron Transferrin saturation Ferritin
Grading
P value
Staging
P value
0.325 0.254 0.285
0.006 0.035 0.018
0.262 0.275 0.325
0.030 0.022 0.006
The present study has been performed on a consecutive, unselected series of patients with clinical, histologic and demographic features of hepatitis C in keeping with those commonly found in Western countries. In agreement with what has been noted by others [11–13], we confirm that the degree of elevation of iron indices in chronic hepatitis C is mild to moderate, rarely reaching the hemochromatotic range. Differences related to gender and to alcohol consumption are maintained, whereas there is no clear correlation between serum iron indices and aging. In patients with hepatitis C, both ferritin and transferrin saturation are excellent markers for presence of HDI in hepatocytes located in zone 1; this correlation becomes progressively less strict in zone 2 and zone 3. The process of iron accretion in the liver involves the progressive recruitment of storage cells from zone 1 to zone 3 [14]. Thus, the zonal distribution of hepatocyte iron indicates that, not only in normal subjects but also in hepatitis C patients, HDI reflects the storage function of the organ, to which corresponds a proportional increase of the serum iron indices. Besides being in the hepatocytes, HDI in the liver appears to be located
Table 4 Multivariate analysis of the factors associated with high serum iron, high transferrin saturation and high serum ferritin in patients with chronic hepatitis C
Age Gender Drinking habits Grading score a Staging score a HDI in hepatocytes HDI in sinusoids HDI in portal triads a
Serum iron .24 mmol / l
Transferrin saturation .29 percent
Serum ferritin .170 mg / l
NE NE NE 0.014 NE 0.003 NE NE
NE NE NE NE NE ,0.001 NE NE
NE 0.001 NE NE NE NE ,0.001 0.047
According to Ishak et al. [9]. HDI5histologically detectable iron; NE5not entered.
54
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55
commonly in the cells lining the sinusoids and in the connective tissue of portal tracts. The correlations between serum iron indices and HDI do not hold in other liver compartments, such as bile ducts. In the absence of overt iron overload, the interpretation traditionally offered for the elevation of serum iron indices in hepatitis C has been that they reflect the release of iron from damaged liver cells [1]. However, this issue is open to scrutiny. In fact, to our knowledge, there has been to date a single study in which HDI in the liver, evaluated semiquantitatively, has been related to serum iron indices in patients with HCV infection [13]. Unfortunately, the method which was used by these authors is relatively insensitive for non-hemochromatotic states of iron overload and no attempt was made to verify zonal distribution of iron. Furthermore, grading and staging scores appeared not to correlate with HDI, a finding that does not seem to give support to the hypothesis that, in these patients, iron is released from the hepatocytes because of inflammatory damage. In contrast, we observed a good correlation between the levels of serum iron indices and the measures of inflammatory activity (grading) and of progression of fibrosis (staging). In particular, serum iron bears a strict relationship with the inflammatory activity, whereas serum ferritin is strictly related to staging. Moreover, we found a significant correlation between the serum concentrations of aspartate aminotransferase and those of iron, but not those of ferritin. In multivariate analysis, ferritin is the only index independently associated with gender and with sinusoidal and portal HDI, while serum iron appears the only index independently associated with the grading score. In other words, in hepatitis C serum iron levels may reflect the degree of current hepatic inflammation and necrosis, whereas ferritin levels may reflect the extent of progressive deposition of iron in sites of scarring, where iron deposition itself may be contributing to the development of fibrosis [15]. According to this scenario, interface hepatitis with its sequelae (sinusoidal capillarization and microshunting) is the major factor leading to iron deposition and fibrotic changes in chronic hepatitis C: It parallels the distribution of HDI and induces the mild elevation of serum iron indices. In contrast, transferrin saturation appears to be independently associated only with HDI in the liver, therefore
resulting the best predictor of the status of hepatic iron deposits in these patients. Measures of serum iron, transferrin saturation and serum ferritin are commonly included in the evaluation of patients with chronic liver disease due to hepatitis C. Hopefully, better knowledge of the associations between specific aspects of the morphologic picture and individual iron indices in this condition may make the latter more informative clinically.
Acknowledgements Supported in part by grants from Italian Government Institutions (Ministero della Universita` e della Ricerca Scientifica) and from Udinese per la vita onlus.
References [1] Bonkovsky HL, Banner BF, Rothman AL. Iron and chronic viral hepatitis. Hepatology 1997;25:759–68. [2] Van Thiel DH, Friedlander L, Fagiuoli S, Wright HI, Irish W, Gvaler JS. Response to interferon alpha therapy is influenced by the iron content of the liver. J Hepatol 1994;20:410–5. [3] Clemente MG, Congia M, Lai ME et al. Effect of iron overload on the response to recombinant interferon-alfa treatment in transfusion-dependent patients with thalassemia major and chronic hepatitis C. J Pediatr 1994;125:123–8. [4] Kowdley KV, Trainer TD, Saltzman JR et al. Utility of hepatic iron index in American patients with hereditary hemochromatosis: a multicenter study. Gastroenterology 1997;113:1270–7. [5] Barton AL, Banner BF, Cable EE, Bonkovsky HL. Distribution of iron in the liver predicts the response of chronic hepatitis C infection to interferon therapy. Am J Clin Pathol 1995;103:419–24. [6] Arber N, Moshkowitz M, Konikoff FM. Elevated serum iron predicts poor response to interferon treatment in patients with chronic HCV infection. Dig Dis Sci 1995;40:2431–3. [7] Banner BF, Barton AL, Cable EE, Smith L, Bonkovsky HL. A detailed analysis of the Knodell score and other histologic parameters as predictors of response to interferon therapy in chronic hepatitis C. Mod Pathol 1995;8:232–8. [8] Wallach J. Interpretation of diagnostic tests: a synopsis of laboratory medicine, 5th ed, Boston: Little, Brown and Company, 1992. [9] Ishak K, Baptista A, Bianchi L et al. Histological grading and staging of chronic hepatitis. J Hepatol 1995;22:696–9.
C. Fabris et al. / Clinica Chimica Acta 304 (2001) 49 – 55 [10] Pirisi M, Toniutto P, Uzzau A et al. Carriage of HFE mutations and outcome of surgical resection for hepatocellular carcinoma in cirrhotic patients. Cancer 2000;89:297– 302. [11] Di Bisceglie AM, Axiotis CA, Hoofnagle JH, Bacon BR. Measurements of iron status in patients with chronic hepatitis. Gastroenterology 1992;102:2108–13. [12] Arber N, Konikoff FM, Moshkowitz M et al. Increased serum iron and iron saturation without liver iron accumulation distinguish chronic hepatitis C from other chronic liver diseases. Dig Dis Sci 1994;39:2656–9.
55
[13] Haque S, Chandra B, Gerber MA, Lok AS. Iron overload in patients with chronic hepatitis C: a clinicopathologic study. Hum Pathol 1996;27:1277–81. [14] Tavill AS. Hemochromatosis. In: Schiff L, Schiff ER, editors, Diseases of the liver, 7th ed, Philadelphia: J.B. Lippincott Company, 1993, pp. 669–91. [15] Pietrangelo A. Iron, oxidative stress, and liver fibrogenesis. J Hepatol 1998;28:8–13.