Iron Levels in Hepatocytes and Portal Tract Cells Predict Progression and Outcomes of Patients With Advanced Chronic Hepatitis C

T

V

S

aO

i[[woaPadatbt0cccasdnhitcdi

KC

CLIN

ICA

L–LIVER

,PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

GASTROENTEROLOGY 2011;140:1490–1500

CLINICAL—LIVER, PANCREAS, AND BILIARYTRACT

Iron Levels in Hepatocytes and Portal Tract Cells Predict Progressionand Outcomes of Patients With Advanced Chronic Hepatitis C

RICHARD W. LAMBRECHT,* RICHARD K. STERLING,‡ DEEPA NAISHADHAM,§ ANNE M. STODDARD,§

THOMAS ROGERS,� CHIHIRO MORISHIMA,¶ TIMOTHY R. MORGAN,# HERBERT L. BONKOVSKY,*,** and the HALT-CRIAL GROUP

*Department of Medicine, University of Connecticut Health Center, Farmington, Connecticut; ‡Hepatology Section, Virginia Commonwealth University, Richmond,irginia; §New England Research Institutes, Watertown, Massachusetts; �Department of Pathology, University of Texas Southwestern Medical Center, Dallas, Texas;

¶Virology Division, Department of Laboratory Medicine, University of Washington, Seattle, Washington; #Gastroenterology Service, VA Long Beach Healthcare

c

BACKGROUND & AIMS: Iron may influence severitynd progression of non-hemochromatotic liver diseases.ur aim was to assess the relationship of iron and HFE

genetic variations to progression and outcomes in theHALT-C Trial and whether PegIFN therapy influencediron variables. METHODS: Participants were random-zed to receive long-term PegIFN [n � 400] or no therapyn � 413] for 3.5 y, with follow-up for up to 8.7 ymedian 6.0 y]. Associations of patient characteristicsith iron variables at baseline and over time were carriedut using Kaplan-Meier analyses, Cox regression models,nd repeated measures analysis of covariance. RESULTS:articipants who developed clinical outcomes [CTP � 7,scites, encephalopathy, variceal bleeding, SBP, HCC,eath] had significantly higher baseline scores for stain-ble iron in hepatocytes and in portal tract cells thanhose without. There were significant direct correlationsetween stainable iron in portal triads and lobular andotal Ishak inflammatory and fibrosis scores [P �.0001]. Iron in triads at baseline increased risk of out-omes (HR � 1.35, P � 0.02). Stainable iron in hepato-ytes decreased, whereas that in portal stromal cells in-reased significantly [P � 0.0001] over time. Serum ironnd TIBC fell significantly over time [P � 0.0001], as diderum ferritin [P � 0.0003]. Chronic PegIFN treatmentid not affect stainable iron. HFE genetic variations didot correlate with outcomes, including development ofepatocellular carcinoma. CONCLUSIONS: Stainable

ron in hepatocytes and portal tract cells is a predic-or of progression and clinical and histological out-omes in advanced chronic hepatitis C. Chronic low-ose PegIFN therapy did not improve outcomes, nor

ron variables.

eywords: Liver Disease; Hepatitis C Virus; Treatment;linical Trial; HCC.

Iron is a cofactor that influences the severity and pro-gression of nonhemochromatotic liver diseases, espe-

ially steatohepatitis and viral hepatitis.1–7 The mostcommon known genetic factor that leads to iron loadingis genetic variation in the gene that is mutated in classicalhereditary hemochromatosis (HFE). Normally, the pro-tein product of this gene plays a key role in regulating theabsorption of iron from the gut. Three genetic variationsin HFE, ie, C282Y, H63D, and S65C, have been associatedwith hemochromatosis. Adverse effects of hepatic iron(with or without HFE mutations) and HFE mutations(with or without elevated iron) on survival after livertransplantation have been described.8 In order to pro-duce an iron-overload phenotype in the absence of othergenetic or environmental factors, both HFE alleles mustbe affected, eg, homozygosity for C282Y (the major mu-tation); compound heterozygosity for C282Y and H63D;or compound heterozygosity for C282Y and S65C. Evenin the presence of homozygosity for C282Y, most people,especially women, do not develop pathological iron over-load, indicating the importance of additional factors.9 –14

In an earlier report,15 we described significant positiveassociations among several measures of iron loading,HFE gene mutations, and severity of chronic hepatitis C(CHC) in the HALT-C (Hepatitis C Anti-Viral Long-TermTreatment to Prevent Cirrhosis) Trial cohort. Yet, theminor (H63D) HFE genetic variation led to impro-ved virological responses to pegylated interferon alfa-2a

Abbreviations used in this paper: CHC, chronic hepatitis C; HCC,hepatocellular carcinoma; HFE, gene that is mutated in classical he-reditary hemochromatosis; IFN, interferon; PegIFN, pegylated inter-feron alfa-2a; TIBC, total iron binding capacity.

© 2011 by the AGA Institute0016-5085/$36.00

doi:10.1053/j.gastro.2011.01.053

1ecelttS

s

tdoPcps

sIfisim

CLI

NIC

AL–

LIV

ER,

PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

May 2011 IRON, HFE, AND OUTCOMES OF CHC 1491

(PegIFN) � ribavirin lead-in therapy. Defects in immunefunctions have been described by others in persons car-rying HFE genetic variations, and interferon (IFN) ther-apy has been said to lead to decreases in hepatic iron.16

The aim of this study was to assess the relationship ofiron and HFE genetic variations to disease progressionand outcomes in the HALT-C Trial and to assess whetherlong-term IFN therapy influenced iron variables. Ourhypotheses were that greater amounts of hepatic and/ortotal body iron would be associated with greater progres-sion and adverse outcomes in participants with advancedCHC, and that long-term PegIFN therapy would be as-sociated with decreased hepatic iron.

Materials and MethodsHALT-C TrialThe HALT-C trial was a multicenter, prospective

study of the safety and efficacy of PegIFN treatment inpatients with advanced hepatitis C.17 It was conducted at

0 clinical sites between August 2000 and July 2006 withxtended observation through October 2009. Inclusionriteria included age older than 18 years, CHC withoutarlier decompensation (ie, no ascites, hepatic encepha-opathy, bleeding varices), and nonresponse to earlierreatment with interferon with or without ribavirin. De-ailed criteria for inclusion and exclusion are listed inupplementary Material.

Patients in whom serum hepatitis C virus RNA per-isted after 20 weeks of treatment with PegIFN, 180 �g

weekly and ribavirin, 1000/1200 mg/day, were random-ized at week 24 to either no treatment (control group) orto continue treatment with PegIFN, 90 �g subcutane-ously every week (lead-in patients). The protocol wasmodified after 2 years to permit randomization of pa-tients who had not responded to at least 24 weeks ofcombination treatment with PegIFN and ribavirin ad-ministered outside of the HALT-C Trial, provided theymet the same inclusion and exclusion criteria (expresspatients).17–21 Because genetic testing, quantitative he-patic iron concentration, and other iron variables werenot available on “express” patients, they were not in-cluded in this analysis. Furthermore, 73 participants didnot consent to genetic testing and were excluded fromanalysis of genetic data.

During the randomized phase of the study, partici-pants were seen every 3 months for interval history,physical examination, and laboratory testing to monitorthe effects of medical therapy, and to assess for clinicalend points and adverse events. The protocol called forrepeat liver biopsy 18 months after randomization (24months after enrollment for lead-in patients) and again42 months after randomization (48 months after enroll-ment for lead-in patients).

The design of this trial (Figure 1) and the major clin-

phase of therapy have been reported.20,21 Outcomes forhe randomized trial included death from any cause,efinite or presumed hepatocellular carcinoma, evidencef clinical decompensation (increase in Child-Turcotte-ugh score to �7, development of ascites, hepatic en-ephalopathy, variceal bleeding, spontaneous bacterialeritonitis), and/or a 2-point increase in Ishak fibrosiscore. There was no benefit of therapy.17 In fact, among

those without cirrhosis at baseline, the risk of deathduring the randomized phase and during extended fol-low-up was higher in those treated with low-dose Pe-gIFN.22

After completion of the randomized phase of thestudy, participants were invited to take part in a phase ofextended observation in which they were asked to returnevery 6 months for interval history, physical examination,laboratory studies, and hepatic/abdominal ultrasoundexamination until October 2009. After the initial ran-domization to low-dose PegIFN or no treatment, themedian duration of follow-up was 6.0 years (range, 0 – 8.7years).

Two methods were used for histologic iron assessmentperformed on a 4-�m section from each biopsy specimentained with Prussian blue at the central Armed Forcesnstitute of Pathology histopathology laboratory. Therst, hereafter called the global hepatocytic iron score, as-igned by consensus of the hepatic pathologists involvedn the HALT-C Trial, employed the semi-quantitative

ethod described by Scheuer and Lefkowitch,18 with ascore of 1 assigned to biopsies with hemosiderin in rareperiportal hepatocytes or at the periphery of regenerativeparenchymal nodules, a score of 2 to biopsies with he-

Figure 1. Schematic of the HALT-C (Hepatitis C Anti-Viral Long-TermTreatment to Prevent Cirrhosis) Trial and participants studied. **Eachpatient counted more than once.

mosiderin in numerous periportal hepatocytes in zone 1

hlitphsb

tsLci

l

gsw

wofmtvMbl

vsgso

tig

CLIN

ICA

L–LIVER

,PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1492 LAMBRECHT ET AL GASTROENTEROLOGY Vol. 140, No. 5

of hepatic acini or extensively around the periphery ofregenerative parenchymal nodules, a score of 3 to biop-sies with hemosiderin extensively in zones 1 and 2 ofhepatic acini or more extensively in regenerative paren-chymal nodules, and a score of 4 to biopsies with hemo-siderin throughout all acinar zones or throughout regen-erative parenchymal nodules.

In addition, a more detailed assessment of hepatic ironstaining was performed by 1 of 3 hepatopathologists onthe HALT-C pathology panel based on a semi-quantita-tive scoring method used previously.23 Hemosiderin in

epatocytes was scored separately from that in reticu-oendothelial (Kupffer) cells, and results were recorded asron absent, iron present in �50%, or �50% of each cellype. For each biopsy specimen, the total numbers ofortal tracts and proportion of portal tracts with anyemosiderin staining were recorded. Whether there wastainable iron in endothelial cells, portal stromal cells, orile ducts was also recorded.

Hepatic iron concentrations were measured on por-ions of 144 liver biopsy specimens from 3 of the clinicalites (University of California-Irvine Medical Center andong Beach VA Medical Center, University of Connecti-ut Health Center, and University of Massachusetts Med-cal Center) using a spectrophotometric procedure.15,24

Hepatic iron indices were calculated as �mol Fe/g dryiver/age (years).

Three variations in the nucleotide sequence of the HFEene that cause changes in the resulting amino acidequence of the HFE protein (C282Y, H63D, and S65C)ere assessed (see Supplementary Material).

Data AnalysisAll data were analyzed by the data coordinating

center (New England Research Institutes). Most contin-uous variables were normally distributed, and means,standard deviations, and ranges were reported. Due tothe left-skewed distribution of serum ferritin concentra-tions, for statistical analysis we performed a log10 trans-formation on these results. To clarify interpretation ofresults, values for serum ferritin in Figures and Tableswere converted back to their antilogs. For time to firstclinical outcomes, we used survival analysis methods,including Kaplan-Meier product limit analysis and Coxproportional hazards regression analysis. P values �.05

ere considered statistically significant. For comparisonf the iron variables between the 2 treatment groups andor the outcomes groups, over time, we used repeated

easures analysis of covariance. These were done con-rolling for the baseline iron measures. The responseariables were assessed at baseline, 24, and 48 months.

easures of iron status were not collected systematicallyeyond month 48, although clinical outcomes and other

aboratory results were.

ResultsSummary Overview of Participants StudiedFigure 1 provides a summary overview of the

HALT-C Trial and the participants studied for this anal-ysis. Among the 813 who agreed to participate in thisstudy, 714 completed this phase. Most (659 [92%]) ofthese also completed the extended follow-up phase.There were no significant differences in numbers of par-ticipants with outcomes or other events, comparingthose treated with long-term low-dose PegIFN vs thosenot treated.

Selected Baseline Demographic, Laboratory,and Histological FeaturesAs for the entire HALT-C cohort,17 none of the

ariables compared (Table 1) showed any statisticallyignificant differences between the treatment vs controlroups with the exception of age, which was slightly butignificantly higher in the IFN-treated group (average agef 50.8 years vs 49.5 years for control group; P � .014).

HFE GenotypesThree single nucleotide polymorphisms in the

HFE gene have been previously implicated in the devel-opment of hereditary hemochromatosis. These 3 singlenucleotide polymorphisms are C282Y (which has beenidentified as the most severe mutation for causing hered-itary hemochromatosis), H63D (which is the most fre-quent of these genetic variations), and S65C (which israre in all populations studied to date). The frequenciesof these single nucleotide polymorphisms are shown inTable 2. As expected, there were no significant differencesbetween the control and PegIFN-treated groups (data notshown).

As described here, participants with a phenotype ofiron overload at baseline were excluded from the trial.Only 1 participant proved to be homozygous for themajor mutation (C282Y�/�) most often associatedwith hereditary hemochromatosis. Ten had compoundheterozygosity for C282Y�/� and H63D�/�, a genotypehat exhibits a lower penetrance for iron overload, whichs usually mild,1,2,4,25 and 1 each was compound heterozy-ous for H63D and S65C or for C282Y and S65C.

Measures of Iron Status at Baseline andClinical OutcomesWe observed significant associations between the

risk and rate of development of first clinical outcomesand the presence of higher grades of stainable iron inhepatocytes (Figure 2A) and higher percentages of portaltriads with stainable iron (Figure 2B). The associationscontinued to be present also for the liver biopsies per-formed at 24 and 48 months; those who experiencedclinical outcomes had significantly higher scores for hep-atocellular stainable iron and percentages of portal triads

.

S

R

ABR

H

Ai l.

CLI

NIC

AL–

LIV

ER,

PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

May 2011 IRON, HFE, AND OUTCOMES OF CHC 1493

to 24 months and from 24 to 48 months. In contrast, thelikelihood of development of clinical outcomes was notsignificantly related to the presence of stainable iron inreticuloendothelial cells or portal stromal cells. By Ka-plan-Meier analyses, there were also trends (albeit notsignificant at the 5% level) for higher serum ferritin (P �.16) or transferrin saturation (P � .11) at baseline to beassociated with development of outcomes. There were noassociations among HFE genotypes and risks or rates ofdevelopment of clinical outcomes. With use of Cox re-gression analysis of all time-varying iron variables, con-trolling for baseline levels and stratifying by treatmentgroup, the baseline iron variables that were significantpredictors of primary clinical outcomes were serum iron,total iron binding capacity (TIBC), transferrin saturation(higher values directly correlated with outcomes; P �

Table 1. Selected Demographic, Laboratory, and Histological

Variables Control gro

n 413ex, %Male 71Female 30

ace, %Non-Hispanic, white 70Black 16Hispanic (white) 10Other 3

ge, y 49.5ody mass index 30.0esults of blood testsMean serum AST, U/L 89.8Mean serum AST (ratio to ULN) 2.14Mean serum ALT, U/L 115Mean serum ALT (ratio to ULN) 2.3Mean serum alkaline phosphatase (ratio to ULN) 0.84Mean serum albumin, g/dL 3.86Mean serum globin, g/dL 3.64Mean serum total bilirubin, mg/dL 0.78Mean INR 1.04Mean serum iron, �g/dL 132Mean total iron binding capacity, �g/dL 372Mean serum iron saturation, % 37Mean serum ferritin, ng/mL 366

Findings at baseline liver biopsyCirrhosis (Ishak fibrosis 5–6), % 40Positive for iron in hepatocytes, % 37Positive for iron in R-E Kupffer cells, % 26Positive for iron in stromal cells, % 36Positive for iron in triads, % 59Average number of triads per biopsy 14.7

HCV variablesMean log10 serum HCV RNA level, IU/L 6.46

CV genotype, %Genotype 1 93Genotype 2 3Genotype 3 3Others (4, 5, 6) 1

LT, alanine aminotransferase; AST, aspartate aminotransferase; HCnterferon alfa-2a; R-E, reticuloendothelial; ULN, upper limit of norma

0001, P � .0001, and P � .025, respectively) and the

percent of hepatocytes with stainable iron (higher valuescorrelated with outcomes; P � .015). The minor differ-ences in significant results by the 2 methods of analysesare likely due to the use of tertiles for Kaplan-Meier vscontinuous variables for Cox regression and the differingvariables that were controlled for and correlated withbaseline results. The hazard ratio for an outcome wassignificantly increased in subjects with stainable iron inportal triads at baseline (hazard ratio � 1.35; P � .02).

Temporal Trends and Effects of Low-DosePegIFN Therapy on Findings at Liver BiopsyComparisons of the control and the PegIFN-

treated groups for hepatic fibrosis score, necroinflamma-tory score, global fat score, and global hepatocytic ironscore for the 3 liver biopsies (baseline, 24, and 48

tures at Baseline

PegIFN-treated group P values, control vs PegIFN-treated

400.63

7228

.16722072

50.8 .01429.8 .65

87.6 .602.11 .80

106 .112.15 .21.83 .85

3.89 .303.66 .540.81 .341.04 .93

134 .63370 .61

38 .53350 .56

40 .9036 .8129 .4031 .1961 .6215.1 .54

6.47 .60.16

95122

patitis C virus; INR, international normalized ratio; PegIFN, pegylated

Fea

up

V, he

months) are presented in Table 3. Table 3 also presents

c

s0

WH

CCSSC

T

CLIN

ICA

L–LIVER

,PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1494 LAMBRECHT ET AL GASTROENTEROLOGY Vol. 140, No. 5

detailed iron scoring (hepatocytic iron [% positive], re-ticuloendothelial iron [% positive], stromal cell iron [%positive], the proportion of portal triads that were posi-tive for iron [% positive]), and the hepatic iron concen-tration and the hepatic iron index. At baseline, there wereno significant differences between the control and Pe-gIFN-treated groups for any of these variables. However,as described previously for the entire HALT-C cohort,17

there were significant decreases in the necroinflammatoryscores at 24 months from 7.49 for control to 6.73 for thePegIFN-treated group (P � .0001), and at 48 monthsfrom 7.14 for control to 6.33 for the PegIFN-treatedgroup (P � .0001).

For the entire cohort, the global hepatocytic ironscores for stainable hepatic iron showed a significantdecrease with time (P � .0001). The percent of biopsieswith stainable iron in hepatocytes also decreased signif-icantly (P � .0005) with time, but did not vary betweenthe 2 treatment groups (P � 0.66). Conversely, the per-cent of biopsies with stainable iron in portal stromal cellsincreased significantly (P � .0001) with time, and indi-cated a strong trend toward increased frequency of stain-able iron in these cells for the treated group (P � 0.06)ompared to the control group (Figure 3).

The global hepatocytic iron score at 48 months wasignificantly lower (P � .024) for the control group (score �.34) than for the PegIFN-treated group (score � 0.46).

Both at 24 and 48 months the control group had asignificantly lower frequency of reticuloendothelial cellsthat were positive for iron (28% vs 38% at 24 months; 28%vs 35% at 48 months).

By definition, none of the 813 participants in theoriginal cohort had had any primary clinical outcomes atbaseline. By 24 months, 60 participants had experienced

Table 2. HFE Genotypes of Participants Who Consented toGenetic Testing (n � 740)

Genotype

Total

n %

T(�/�) at all 3 loci 510 6963D�/� only 146 20

H63D�/� only 16 2282Y�/� only 44 6282Y�/� only 1 �165C�/� only 9 165C�/� only 1 �1ompound heterozygotesC282Y�/�, H63D�/�, S65C�/� 10 1C282Y�/�, H63D�/�, S65C�/� 1 �1C282Y�/�, H63D�/�, S65C�/� 1 �1

otal 739a 100

HFE, the gene that is mutated in classical hereditary hemochr-omatosis.aOne participant who provided consent for genetic testing had missinggenotype data.

a primary clinical outcome, whereas 715 participants had

not. At 48 months, 84 additional participants (a total of144) had had a primary clinical outcome and 586 hadnot. At 72 and 96 months, 215 and 250 (38%) hadexperienced primary clinical outcomes. Global hepato-cytic iron scores were higher at months 24 and 48, com-pared to baseline, for those participants with a clinicaloutcome compared to those without (P � .032). Partici-pants who did not develop primary clinical outcomesshowed no change in portal triads positive for iron,whereas participants who did develop primary clinicaloutcomes had significantly increased frequency of stain-able iron in their portal triads (P � .038).

In liver biopsies from all 3 time points, there werepositive and highly significant correlations between thepercent of portal triads that were positive for stainableiron and the Ishak inflammatory scores (both for lobularinflammation and the total Ishak score). In addition,there were significant associations (P � .0014 –.0184)between steatosis scores and stainable iron at all biopsy

Figure 2. Measures of iron status at baseline and development ofclinical outcomes. (A) Time to first clinical outcome by global hepato-cytic iron score. (B) Time to first clinical outcome by percentage of portaltriads that contained stainable iron. Clinical outcomes were defined asdescribed previously.17 Stainable iron was assessed by a panel of hepa-topathologists, as described in Materials and Methods, who had no

tbcp

B

CLI

NIC

AL–

LIV

ER,

PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

May 2011 IRON, HFE, AND OUTCOMES OF CHC 1495

time points, although the presence of stainable iron ac-counted for only a small percent of the variation insteatosis (0.6 –1.5%). Both the percent of portal triadsthat were positive for iron and the global hepatocyticiron scores were also directly correlated with hepaticfibrosis scores, with significant time effects (P � .047 forhe interaction term of the fibrosis and iron scores ataseline, 24 months, and 48 months). Similarly, the per-entage of portal triads that stained positive for ironroduced an overall P value for interaction of time and

portal fibrosis of .020. In both cases, the relationshipchanged with time and became slightly weaker at 48months (results not shown). In summary, therefore,higher degrees of stainable iron in hepatocytes and portal

Table 3. Hepatic Histopathology in Participants Studied

iopsy at baselinenFibrosis score, mean (SD)Necroinflammatory score, mean (SD)Global fat score , mean (SD)Global hepatocytic iron score, mean (SD)Detailed iron scoring, n (%)

Hepatocytic ironR-E ironStromal cell ironTriads with iron

Hepatic iron concentration, ug/g dry wt, mean (SD)Hepatic iron index, mean (SD)

Biopsy at 24 monthsnFibrosis score, mean (SD)Necroinflammatory score, mean (SD)Global fat score, mean (SD)Global hepatocytic iron score, mean (SD)Detailed iron scoring, n (%)

Hepatocytic ironR-E ironStromal cell ironTriads with iron

Hepatic iron concentration, ug/g dry wt, mean (SD)Hepatic iron index, mean (SD)

Biopsy at 48 monthsnFibrosis score, mean (SD)Necroinflammatory score, mean (SD)Global fat score, mean (SD)Global hepatocytic iron score, mean (SD)Detailed iron scoring, n (%)

Hepatocytic ironR-E ironStromal cell ironTriads with iron

Hepatic iron concentration, ug/g dry wt, mean (SD)Hepatic iron index, mean (SD)

NOTE. Fewer participants underwent liver biopsies at 24 and 48 montdeclined, participant had died, participants had undergone liver tranbelieved biopsy should not be performed.IFN, interferon; R-E, reticuloendothelial; SD, standard deviation.

triads at baseline correlated with more fibrosis and with

the occurrence of clinical outcomes, and those who ex-perienced clinical outcomes had increases in stainableiron compared to those without. There were no particu-lar outcomes that were more frequent in those withstainable iron; rather, all outcomes occurred with greaterfrequencies and in about the same proportions (Supple-mentary Table 3).

HFE Genetic Variations and Clinical Courseand Outcomes

HFE genetic variations (Table 2) did not correlatewith primary clinical outcomes, including developmentof hepatocellular carcinoma, nor did they correlate with

Control IFN-treated P values

413 4004.13 (1.27) 4.08 (1.24) .567.39 (2.01) 7.50 (2.09) .431.33 (0.89) 1.37 (0.95) .580.51 (0.72) 0.49 (0.69) .72

149 (37) 141 (36) .81143 (36) 121 (31) .19105 (26) 112 (29) .40237 (59) 236 (61) .62479 (763) 1344 (794) .440.53 (0.28) 0.48 (0.31) .46

346 3393.92 (1.47) 3.92 (1.47) .997.49 (2.03) 6.73 (2.33) �.00011.31 (0.83) 1.33 (0.80) .770.36 (0.63) 0.38 (0.63) .72

101 (30) 99 (29) .8795 (28) 127 (38) .009

156 (46) 169 (50) .32191 (57) 194 (58) .82368 (605) 1334 (748) .820.51 (0.25) 0.48 (0.27) .57

287 2834.00 (1.60) 4.07 (1.60) .617.14 (2.19) 6.33 (2.30) �.00011.14 (0.69) 1.23 (0.76) .150.34 (0.59) 0.46 (0.70) .0237

71 (25) 85 (30) .1779 (28) 100 (35) .0505

142 (50) 158 (56) .15166 (59) 172 (61) .57521 (809) 1360 (650) .360.61 (0.63) 0.49 (0.25) .28

cause of a variety of reasons. The major reasons included participanttation, participant had withdrawn from the study, and investigators

1

1

1

hs besplan

greater increases in stainable iron in liver biopsies at 24

et

si

CLIN

ICA

L–LIVER

,PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1496 LAMBRECHT ET AL GASTROENTEROLOGY Vol. 140, No. 5

or 48 months (data not shown). There were no detectableeffects of iron status or HFE genotypes on several mark-rs of immunological functions previously described inhis cohort26,27 (intrahepatic cytotoxic T-lymphocyte

function or peripheral blood lymphocyte proliferation;data not shown).

Salient features of the 11 participants with HFE geno-types associated with hereditary hemochromatosis aresummarized in Supplementary Table 1. Participant no. 1was C282Y�/� homozygous and 10 (participant nos.2–11) were C282Y�/� and H63D�/� compound heterozy-gous. At baseline, although 6 had elevated serum ferritinlevels and 5 had elevated serum transferrin saturations,only 1 (no. 5) had hepatic iron overload. Two (nos. 4 and10), 1 of which at baseline had no stainable hepatic iron,experienced a clinical outcome; and another (no. 8) ex-perienced a histopathological outcome (progression offibrosis score from 3 to 6). There was no evidence of app-reciable iron loading in these participants. Worthy ofnote is that 4 of 10 participants carrying the H63Dgenetic variation experienced complete virological re-sponses during lead-in therapy. Previously, we showedthat participants with this genetic makeup are signifi-cantly more likely to achieve complete virological re-sponse and sustained viral response than those with-

Figure 3. Effect of time and treatment on hepatic stainable iron. Rignificantly over time (P � .0001). There was significant decrease in

ncreases in the percent of biopsies with stainable iron in portal stromal cgroup (P � .06).

out.15

Temporal Trends In and Effects of Low-DosePegIFN Therapy on Serum Measures of IronStatusThe effects of time and treatment group on serum

iron levels, total iron binding capacity, the log of serumferritin levels, and the serum transferrin saturation scoresare shown in Figure 4. Serum iron levels (Figure 4A)decreased significantly with time (P � .0001) and withthe treatment group (the PegIFN-treated group had alarger decrease in serum iron levels than the controlgroup; P � .0304). The P value for the interaction of thetime and treatment groups for serum iron levels was alsosignificant (P � .012). Analysis of the serum iron bindingcapacity (Figure 4B) and ferritin values (Figure 4D) indi-cated similar, statistically significant decreases with time(P � .0001 and 0.0003, respectively) and with treatmentgroup (P � .0004 and .0475). The P values for the inter-action of time and treatment group were �.0001 for boththe iron binding capacity and serum ferritin values.

Prediction of Disease Progression inAdvanced CHCModels to predict disease progression from base-

line results of readily available laboratory tests (serumalbumin, alanine aminotransferase, aspartate amino-transferase, total bilirubin, and platelet count), not in-

are mean � standard error. The global hepatocytic iron score fellercent of biopsies with stainable iron in hepatocytes (P � .0005) and� .0001), with a strong trend to be greater in pegylated interferon �-2a

esultsthe pells (P

cluding measures of iron status, were recently described

td

CLI

NIC

AL–

LIV

ER,

PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

May 2011 IRON, HFE, AND OUTCOMES OF CHC 1497

by Ghany et al28 for the HALT-C cohort. As describedhere, those more likely to experience outcomes also hadhigher serum irons and ferritins and lower total iron-binding capacities at baseline. Addition of the baselinemeasures of iron status to this model revealed that lowerTIBC at baseline was a significant additional predictor ofoutcomes (P � .005), whereas serum iron, ferritin, orglobal hepatocytic iron scores were not (details shown inSupplementary Table 2).

DiscussionIn this article, we provide long-term follow-up

information on those participants in the HALT-C trialwho did not experience a sustained viral response tolead-in therapy and who agreed to enter the randomizedphase of the trial. Our major findings in this large (n �813) cohort, followed for up to 8.7 years (median, 6 years)are that, regardless of treatment group, (1) the presenceand degree of stainable iron (global hepatocytic ironscore; percent triads positive for iron) were predictive ofadverse clinical outcomes; (2) stainable iron in hepato-cytes decreased significantly over time; (3) stainable ironin portal stromal cells increased significantly over time;(4) long-term low-dose IFN therapy was not associatedwith less stainable iron in the liver (in contrast to earlierfindings with full-dose standard IFN)16; (5) there were nosignificant associations between HFE genetic variations

Figure 4. Effect of time and treatment on serum measures of iron stacapacity fell significantly, and progressively over time (P � .0001) and sigP � .0004). There were no significant changes in transferrin saturationPegIFN group (P � .048).

and either clinical outcomes or measure of iron status;

and (6) lower levels of TIBC contributed significantly toa model for predicting outcomes of CHC based on gen-erally available liver tests (Supplementary Table 2).28

Our results extend a number of previously publishedreports of the association of hepatic iron with adverseoutcomes in CHC, as well as in other liver diseases,especially fatty liver/steatohepatitis.1– 6,8,15,29 –32 The im-portance of iron as a comorbid factor in CHC is empha-sized by several recent reports of greater fibrosis andgreater risks of HCC development with more hepaticiron.30,32 Conversely, iron reduction (usually by therapeu-tic phlebotomies) has consistently been associated withreductions in serum alanine aminotransferase levels, inseverity of hepatic necroinflammation, and risk of devel-opment of HCC.1,5,31 Iron reduction has also led to im-proved responses to IFN-based therapy of CHC.29 Amonghe possible reasons for improvement of CHC by stan-ard IFN29 is an effect to decrease hepatic iron.16 This

effect is diminished or abolished by ribavirin, which leadsto iron loading due to the hemolytic anemia that attendsits use. In this large study, we observed that long-termlow-dose Peg-IFN did not lead to decreased hepatic ironor to any improvement in clinical outcomes or hepaticfibrosis (Table 3).17

Strengths of this study include a large, well-character-ized cohort with prospective and extended follow-up(Figure 1). There was little drop-out of participants dur-

Results are mean � standard error. Serum iron and total iron bindingntly more in the pegylated interferon �-2a (PegIFN) group (P � .03 andm ferritin also fell significantly over time (P � .0003), but less so in the

tus.nifica

. Seru

ing the lead-in and randomized phases, and, even during

g

toAfspao

amtTawout

aatchtnitgoigtirrH

hdoeoHufra

(nitp(isaogi

opts

1

CLIN

ICA

L–LIVER

,PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1498 LAMBRECHT ET AL GASTROENTEROLOGY Vol. 140, No. 5

the extended follow-up phase, continued patient fol-low-up was 70% for at least 6.5 years and 55% for at least8.7 years after randomization.22

Our study, although the largest undertaken to dateand with the longest and best follow-up achieved so faramong longitudinal studies of CHC,17 has limitations.First, enrollment was limited to persons with advancedCHC who had failed at least one (often more than one)course of adequate IFN � ribavirin therapy. Second, wetook pains to exclude participants who, at baseline, hadiron overload. The reason was that we did not want toconfound results by introducing the variable of pre-exis-tent iron loading. We were highly successful in excludingparticipants with hemochromatosis from the trial. In-deed, only 1 of 1050 (participant no. 1 of SupplementaryTable 1) who later had HFE genetic testing carried the

enotype (C282Y�/�) strongly associated with hemochro-matosis, and this participant did not express an iron-overload phenotype (Supplementary Table 1).15 In long-erm follow-up, this participant did not develop clinicalr histopathological outcomes and did not develop HCC.nother limitation is that we do not have full long-term

ollow-up data on participants who experienced the moreevere outcomes (eg, liver decompensation, liver trans-lantation) because they did not continue in the studynd because their iron status at the time of these seriousutcomes was usually unknown.

We speculate that this drop-out of subjects with moredvanced and/or rapidly progressing liver disease is theain reason why the frequency of stainable iron in hepa-

ocytes decreased over time in this study (Figure 3A, B).he rise in stainable iron in portal stromal cells suggestsn intrahepatic redistribution of iron from hepatocytes,hich, when iron-loaded, may be more susceptible toxidative damage and death, with the iron being takenp by macrophages and eventually moving into the por-al tract stromal cells (Figure 3C).

Iron, the most abundant element of the whole of earth,nd fourth most abundant element of the earth’s crust, isn essential element for virtually all forms of life on ourerrestrial home. Despite its abundance, during theourse of evolution and throughout most of humanistory, organisms, including humans, have had difficul-ies absorbing sufficient amounts of iron to meet theireeds for formation of heme, hemoproteins, and other

ron-containing proteins. Several proteins have developedhat can help to increase the absorption of iron from theastrointestinal tract, whereas there are no well-devel-ped physiological processes for the shedding of excess

ron. The liver is the primary site for iron storage, andenetic or acquired defects leading to excessive iron up-ake and storage are characterized by gradual hepaticron deposition, development of hepatic fibrosis and cir-hosis, and, at times, of HCC. Among subjects with cir-hosis due to hemochromatosis, there is a lifetime risk of

The iron-loading that can complicate chronic viralepatitis, alcoholic liver disease, nonalcoholic fatty liverisease (including nonalcoholic steatohepatitis), andther hepatic disorders is of multifactorial origin. How-ver, one key mechanism is an inappropriately low levelf hepcidin production and/or secretion by the liver.epcidin has emerged as the primary hormone that reg-lates iron metabolism. Hepcidin is made by and released

rom many cell types, but especially by hepatocytes. Itselease is increased by proinflammatory cytokines, suchs interleukin-1, interleukin-6, tumor necrosis factor–�

(hepcidin is an acute phase reactant), and by HFE, HFE2hemojuvelin), transferrin receptor-2, bone morphoge-etic proteins, and other factors that respond to hepatic

ron status, to hypoxia, and to the degree of adequacy ofissue oxygen delivery. Hepcidin production is sup-ressed by the transmembrane serine protease TMPRSS6

also known as matriptase-2), which is required to senseron deficiency33–35 (see reviews36,37). In the case of CHC,erum hepcidin levels have been found to be inappropri-tely low for the degree of iron loading38 and expressionf hepatitis C virus genes has led to decreased hepcidinene expression due to hepatitis C virus–mediated defectsn histone deacetylase activity in infected hepatocytes.39

The complexity of the interplay between iron and CHCis underscored by our recent finding that mutations inHFE, particularly the common H63D variation, whilebeing associated with higher hepatic and total body iron,is also associated with significantly higher likelihood ofcomplete and sustained responses to IFN � ribavirintherapy.15

An important and still unresolved question is whetherlong-term, sustained iron reduction to near iron deple-tion can help to prevent or forestall disease progressionand/or development of HCC in CHC. This has beenreported in a long-term nonrandomized prospectivestudy from Japan.31 In view of the relative ease and safety

f iron reduction and adherence to a low-iron diet, arospective, randomized, controlled trial of iron reduc-ion in persons who cannot tolerate or have not re-ponded to PegIFN � ribavirin therapy would be wel-

come.

Supplementary Material

Note: To access the supplementary materialaccompanying this article, visit the online version ofGastroenterology at www.gastrojournal.org, and at doi:

0.1053/j.gastro.2011.01.053.

References

1. Alla V, Bonkovsky HL. Iron in nonhemochromatotic liver disorders.Semin Liver Dis 2005;25:461–472.

2. Bonkovsky HL, Banner BF, Lambrecht RW, Rubin RB. Iron in liverdiseases other than hemochromatosis. Semin Liver Dis 1996;16:65–82.

3. Bonkovsky HL, Banner BF, Rothman AL. Iron and chronic viral

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3CLI

NIC

AL–

LIV

ER,

PA

NCREA

S,A

ND

BIL

IARY

TRA

CT

May 2011 IRON, HFE, AND OUTCOMES OF CHC 1499

4. Bonkovsky HL, Lambrecht RW. Iron-induced liver injury. Clin LiverDis 2000;4:409–429.

5. Bonkovsky HL, Lambrecht RW, Shan Y. Iron as a co-morbid factorin nonhemochromatotic liver disease. Alcohol 2003;30:137–144.

6. Bonkovsky HL, Troy N, McNeal K, et al. Iron and HFE or TfR1mutations as comorbid factors for development and progressionof chronic hepatitis C. J Hepatol 2002;37:848–854.

7. Bonkovsky H, Lambrecht R. Hemochromatosis, iron overload,and porphyria cutanea tarda. In: Barton J, Edwards C, eds. Hemo-chromatosis: genetics, pathophysiology, diagnosis and treat-ment. Cambridge, UK: Cambridge University Press, 2000:453–467.

8. Kowdley KV, Brandhagen DJ, Gish RG, et al. Survival after livertransplantation in patients with hepatic iron overload: the na-tional hemochromatosis transplant registry. Gastroenterology2005;129:494–503.

9. Adams PC, Reboussin DM, Barton JC, et al. Hemochromatosisand iron-overload screening in a racially diverse population.N Engl J Med 2005;352:1769–1778.

0. Beutler E, Felitti VJ, Koziol JA, Ho NJ, Gelbart T. Penetrance of845G¡A (C282Y) HFE hereditary haemochromatosis mutation inthe USA. Lancet 2002;359:211–218.

1. Bulaj ZJ, Ajioka RS, Phillips JD, et al. Disease-related conditionsin relatives of patients with hemochromatosis. N Engl J Med2000;343:1529–1535.

2. Fargion S, Mandelli C, Piperno A, et al. Survival and prognosticfactors in 212 Italian patients with genetic hemochromatosis.Hepatology 1992;15:655–659.

3. Loreal O, Deugnier Y, Moirand R, et al. Liver fibrosis in genetichemochromatosis. Respective roles of iron and non-iron-relatedfactors in 127 homozygous patients. J Hepatol 1992;16:122–127.

4. Allen KJ, Gurrin LC, Constantine CC, et al. Iron-overload-relateddisease in HFE hereditary hemochromatosis. N Engl J Med 208;358:221–230.

5. Bonkovsky HL, Naishadham D, Lambrecht RW, et al. Roles of ironand HFE mutations on severity and response to therapy duringretreatment of advanced chronic hepatitis C. Gastroenterology2006;131:1440–1451.

6. Boucher E, Bourienne A, Adams P, Turlin B, Brissot P, Deugnier Y.Liver iron concentration and distribution in chronic hepatitis Cbefore and after interferon treatment. Gut 1997;41:115–120.

7. Di Bisceglie AM, Shiffman ML, Everson GT, et al. Prolongedtherapy of advanced chronic hepatitis C with low-dose peginter-feron. N Engl J Med 2008;359:2429–2441.

8. Scheuer P, Lefkowitch J. Liver biopsy interpretation. Philadelphia:Saunders Co., 2000.

9. Ishak K, Baptista A, Bianchi L, et al. Histological grading andstaging of chronic hepatitis. J Hepatol 1995;22:696–699.

0. Lee WM, Dienstag JL, Lindsay KL, et al. Evolution of the HALT-CTrial: pegylated interferon as maintenance therapy for chronichepatitis C in previous interferon nonresponders. Control ClinTrials 2004;25:472–492.

1. Shiffman ML, Di Bisceglie AM, Lindsay KL, et al. Peginterferonalfa-2a and ribavirin in patients with chronic hepatitis C who havefailed prior treatment. Gastroenterology 2004;126:1015–1023;discussion 947.

2. Dienstag JL, Bonkovsky HL, Kim H-Y, et al. Progression of ad-vanced chronic hepatitis C (CHC): further follow-up results fromthe HALT-C Trial. Abstract of Digestive Disease Week, New Or-leans, LA, 2010; abstract no. 237.

3. Barton AL, Banner BF, Cable EE, Bonkovsky HL. Distribution ofiron in the liver predicts the response of chronic hepatitis Cinfection to interferon therapy. Am J Clin Pathol 1995;103:419–

4. Vinogradov S. Spectrophotometric determination of iron in hemeproteins. Methods Enzymol 1986;123:320–323.

5. Feder JN, Gnirke A, Thomas W, et al. A novel MHC class I-likegene is mutated in patients with hereditary haemochromatosis.Nat Genet 1996;13:399–408.

6. Morishima C, Polyak SJ, Ray R, et al. Hepatitis C virus-specificimmune responses and quasi-species variability at baseline areassociated with nonresponse to antiviral therapy during ad-vanced hepatitis C. J Infect Dis 2006;193:931–940.

7. Rothman AL, Morishima C, Bonkovsky HL, et al. Associationsamong clinical, immunological, and viral quasispecies measure-ments in advanced chronic hepatitis C. Hepatology 2005;41:617–625.

8. Ghany MG, Lok AS, Everhart JE, et al. Predicting clinical andhistologic outcomes based on standard laboratory tests in ad-vanced chronic hepatitis C. Gastroenterology 2010;138:136–146.

9. Desai TK, Jamil LH, Balasubramaniam M, Koff R, Bonkovsky HL.Phlebotomy improves therapeutic response to interferon in pa-tients with chronic hepatitis C: a meta-analysis of six prospectiverandomized controlled trials. Dig Dis Sci 2008;53:815–822.

0. Erhardt A, Maschner-Olberg A, Mellenthin C, et al. HFE mutationsand chronic hepatitis C: H63D and C282Y heterozygosity areindependent risk factors for liver fibrosis and cirrhosis. J Hepatol2003;38:335–342.

1. Kato J, Miyanishi K, Kobune M, et al. Long-term phlebotomy withlow-iron diet therapy lowers risk of development of hepatocellularcarcinoma from chronic hepatitis C. J Gastroenterol 2007;42:830–836.

2. Smith BC, Gorve J, Guzail MA, et al. Heterozygosity for hereditaryhemochromatosis is associated with more fibrosis in chronichepatitis C. Hepatology 1998;27:1695–1699.

3. Du X, She E, Gelbart T, et al. The serine protease TMPRSS6 isrequired to sense iron deficiency. Science 2008;320:1088–1092.

4. Finberg KE, Heeney MM, Campagna DR, et al. Mutations inTMPRSS6 cause iron-refractory iron deficiency anemia (IRIDA).Nat Genet 2008;40:569–571.

5. Folgueras AR, de Lara FM, Pendas AM, et al. Membrane-boundserine protease matriptase-2 (Tmprss6) is an essential regulatorof iron homeostasis. Blood 2008;112:2539–2545.

6. Muckenthaler MU. Fine tuning of hepcidin expression by positiveand negative regulators. Cell Metab 2008;8:1–3.

7. Nemeth E, Ganz T. Regulation of iron metabolism by hepcidin.Annu Rev Nutr 2006;26:323–342.

8. Fujita N, Sugimoto R, Takeo M, et al. Hepcidin expression in theliver: relatively low level in patients with chronic hepatitis C. MolMed 2007;13:97–104.

9. Miura K, Taura K, Kodama Y, Schnabl B, Brenner DA. Hepatitis Cvirus-induced oxidative stress suppresses hepcidin expressionthrough increased histone deacetylase activity. Hepatology 2008;48:1420–1429.

Received September 10, 2010. Accepted January 20, 2011.

Reprint requestsAddress requests for reprints to: Herbert L. Bonkovsky, MD,

Cannon Research Center, Carolinas Medical Center, 1542 GardenTerrace, Charlotte, North Carolina 28203. e-mail: [email protected]; fax: (704) 355-7648.

Richard W. Lambrecht, PhD, is currently at the University ofMassachusetts Medical School, Worcester, MA.

AcknowledgmentsThis is publication no. 50 of the HALT-C Trial. The HALT-C Trial

was registered with clinicaltrials.gov (#NCT00006164).

CLIN

ICA

L–LIVER

,PA

NCREA

S,A

ND

BILIA

RY

TRA

CT

1500 LAMBRECHT ET AL GASTROENTEROLOGY Vol. 140, No. 5

We thank H. James Norton for help with statistical analyses andsummaries.

In addition to the authors of this article, the following individualswere instrumental in the planning, conduct and/or care of patientsenrolled in this study at each of the participating institutions asfollows: University of Massachusetts Medical Center, Worcester, MA:(Contract N01-DK-9-2326) Gyongyi Szabo, MD, Barbara F. Banner,MD, Maureen Cormier, RN, Donna Giansiracusa, RN. University ofConnecticut Health Center, Farmington, CT: (Grant M01RR-06192)Gloria Borders, RN, Michelle Kelley, RN, ANP. Saint Louis UniversitySchool of Medicine, St Louis, MO: (Contract N01-DK-9-2324) AdrianM. Di Bisceglie, MD, Bruce Bacon, MD, Brent Neuschwander-Tetri,MD, Elizabeth M. Brunt, MD, Debra King, RN. Massachusetts GeneralHospital, Boston, MA: (Contract N01-DK-9-2319, GrantM01RR-01066; Grant 1 UL1 RR025758-01, Harvard Clinical andTranslational Science Center) Jules L. Dienstag, MD, Raymond T.Chung, MD, Andrea E. Reid, MD, Atul K. Bhan, MD, Wallis A.Molchen, David P. Lundmark. University of Colorado Denver, Schoolof Medicine, Aurora, CO: (Contract N01-DK-9-2327, Grant M01RR-00051, Grant 1 UL1 RR 025780-01) Gregory T. Everson, MD,Thomas Trouillot, MD, Marcelo Kugelmas, MD, S. Russell Nash, MD,Jennifer DeSanto, RN, Carol McKinley, RN. University of California-Irvine, Irvine, CA: (Contract N01-DK-9-2320, Grant M01RR-00827)John C. Hoefs, MD, John R. Craig, MD, M. Mazen Jamal, MD, MPH,Muhammad Sheikh, MD, Choon Park, RN. University of TexasSouthwestern Medical Center, Dallas, TX: (Contract N01-DK-9-2321,Grant M01RR-00633, Grant 1 UL1 RR024982-01, North and CentralTexas Clinical and Translational Science Initiative) William M. Lee,MD, Peter F. Malet, MD, Janel Shelton, Nicole Crowder, LVN, RivkaElbein, RN, BSN, Nancy Liston, MPH. University of SouthernCalifornia, Los Angeles, CA: (Contract N01-DK-9-2325, GrantM01RR-00043) Karen L. Lindsay, MD, MMM, SuganthaGovindarajan, MD, Carol B. Jones, RN, Susan L. Milstein, RN.University of Michigan Medical Center, Ann Arbor, MI: (Contract N01-DK-9-2323, Grant M01RR-00042, Grant 1 UL1 RR024986, MichiganCenter for Clinical and Health Research) Anna S. Lok, MD, Robert J.Fontana, MD, Joel K. Greenson, MD, Pamela A. Richtmyer, LPN,CCRC, R. Tess Bonham, BS. Virginia Commonwealth UniversityHealth System, Richmond, VA: (Contract N01-DK-9-2322, GrantM01RR-00065) Mitchell L. Shiffman, MD, Melissa J. Contos, MD, A.

Diseases Branch, National Institute of Diabetes and Digestive andKidney Diseases, National Institutes of Health, Bethesda, MD: MarcG. Ghany, MD, T. Jake Liang, MD, David Kleiner, MD, PhD, YoonPark, RN, Elenita Rivera, RN, Vanessa Haynes-Williams, RN.National Institute of Diabetes and Digestive and Kidney Diseases,Division of Digestive Diseases and Nutrition, Bethesda, MD: JamesE. Everhart, MD, MPH, Leonard B. Seeff, MD, Patricia R. Robuck,PhD, Jay H. Hoofnagle, MD, Elizabeth C. Wright, PhD. University ofWashington, Seattle, WA: (Contract N01-DK-9-2318) David R.Gretch, MD, PhD, Minjun Chung Apodaca, BS, ASCP, Rohit Shankar,BC, ASCP, Natalia Antonov, M. Ed. New England Research Institutes,Watertown, MA: (Contract N01-DK-9-2328) Kristin K. Snow, MSc,ScD, Teresa M. Curto, MSW, MPH, Margaret C. Bell, MS, MPH. InovaFairfax Hospital, Falls Church, VA: Zachary D. Goodman, MD, PhD,Fanny Monge, Michelle Parks. Data and Safety Monitoring BoardMembers: (Chair) Gary L. Davis, MD, Guadalupe Garcia-Tsao, MD,Michael Kutner, PhD, Stanley M. Lemon, MD, Robert P. Perrillo, MD.

Conflicts of interestThis author discloses the following:Dr Sterling is a consultant and receives research support from

Hoffman-La Roche, Inc.The remaining authors disclose no conflicts.

FundingThis study was supported by the National Institute of Diabetes

and Digestive and Kidney Diseases (contract numbers are listed inAcknowledgments). Additional support was provided by the NationalInstitute of Allergy and Infectious Diseases (NIAID), the NationalCancer Institute, the National Center for Minority Health and HealthDisparities, and by General Clinical Research Center and Clinicaland Translational Science Center grants from the National Centerfor Research Resources, National Institutes of Health (grantnumbers are listed in the Acknowledgments).

The content is solely the responsibility of the authors and doesnot necessarily represent the official views of the National Centerfor Research Resources or the National Institutes of Health.

Additional funding to conduct this study was supplied byHoffmann-La Roche, Inc., (now Genentech) through a CooperativeResearch and Development Agreement (CRADA) with the National

calywt

btaI

May 2011 IRON, HFE, AND OUTCOMES OF CHC 1500.e1

Supplementary Material

Detailed criteria for inclusion and exclusion wereas follows: exclusion criteria included a Child–Turcotte–Pugh score �7, a platelet count �50,000/uL, blood he-moglobin �11 g/dL, serum creatinine �1.5 mg/dL, al-pha-fetoprotein �200 ng/mL, or a mass suggestive ofHCC on imaging. Other exclusions were known hemo-chromatosis or histologic evidence of hepatic iron over-load (Scheuer iron grade 3– 4�),18 human immunodefi-iency virus infection, autoimmune diseases, substancebuse within the past 2 years, malignancy (other thanocalized squamous cell cancer of skin) within the past 5ears, immunosuppressive treatment, Coumadin useithin the past 6 months, or any serious medical condi-

ion likely to compromise participation for 4 years.To be eligible for entry, patients must have had a liver

iopsy performed within 12 months of enrollment intohe HALT-C study. Liver biopsies were graded for activitynd fibrosis using the Ishak Modified Hepatic Activityndex Scoring System19 by consensus face-to-face vote of

the 10 clinical site hepatopathologists and Z. Goodman,the coordinating pathologist from the Armed Forces In-

Ishak fibrosis score of �3 (bridging fibrosis or cirrhosis)on the baseline biopsy, or an Ishak score of 2 on thebaseline biopsy, provided that an earlier liver biopsy hadhad an Ishak fibrosis score of �3.

Methods for Assessing HFE MutationsFor the C282Y analysis, a TaqMan assay was used.

Briefly, DNA was amplified by polymerase chain reactionusing 2 flanking primers (5=ATT GGG GAT GGG ACCTAC CA and 3=TCA CAT ACC CCA GAT CAC ATT GAG)to produce a 126-base pair product, along with fluores-cently tagged interior primers (5=ATA TAC GTg/a CCAGGT GG) that distinguish between the 2 polymor-phisms. The H63D and S65C polymorphisms were ana-lyzed using a restriction fragment length polymorphismapproach. To detect the H63D mutation, DNA was am-plified using 5=TGG TCT TTC CTT GTT TGA AGC TTand 5=TTC TAC TGG AAA CCC ATG GAG TCC toproduce a 110-base pair product that was digested byBcl1. Bcl1 digests the wild-type product by not the mu-tant. For S65C mutational analysis, DNA was amplifiedusing the same primers as for H63D detection but diges-tion was with Hinf1, which digests the wild-type product

ex; HFE, gene that is mutated in classical hereditary hemochromatosis; M, male; ND,

SL

�7�

�7�

A

1500.e2 LAMBRECHT ET AL GASTROENTEROLOGY Vol. 140, No. 5

Supplementary Table 1. Selected Features of Participants w

Serum ferritin,ng/mL

Transferrinsaturation [%] Iron

No. Age Sex Group BL 24 48 BL 24 48 BL 24

C282Y�/� (C21 47 M NR 473 294 ND 44 13 ND 0 0

C282Y�/� and H63D�/

2 66 F NR 153 30 21 31 31 10 1 03 54 M BT/Rel 61 ND 216 78 ND 62 0 ND4 53 M NR 359 719 ND 33 58 ND 0 05 43 M BT/Rel 537 ND ND 83 ND ND 3 ND6 46 M NR 970 ND ND 60 ND ND 1 ND7 49 M NR 949 837 1150 71 75 46 1 28 44 M BT/Rel 431 65 98 51 24 27 2 19 47 M BT/Rel 90 106 45 22 17 19 0 0

10 49 F NR 197 268 296 34 69 27 1 011 43 M NR 199 571 516 42 35 26 0 0

BL, baseline; BT/Rel, breakthrough/relapse; F, female; HAI, hepatic activity indtest biopsy not done; NR, XXX.

Supplementary Table 2. Effects of Adding Measures of IronChronic Hepatitis C

Predictors in modeln � 813

Ghany’s disease progression model

Estimate (SE) HR P value Es

Log (AST/ALT) 0.76 (0.31) 2.145 .0125Total bilirubin, mg/dL 0.49 (0.16) 1.634 .0022Albumin, g/dL �1.21 (0.25) 0.298 �.0001 �Platelets 50,000/�L �0.009 (0.002) 0.991 �.0001 �0

erum iron �0og serum ferritin

Risk percentilesRisk of developing an outcome

in 3.85 y, %

75th percentile 85th to �90th percentile 3890th percentile 59

Predictors in modeln � 813

Ghany’s disease progressionmodel � iron

Estimate (SE) HR P value E

Log (AST/ALT) 0.81 (0.31) 2.252 .0086Total bilirubin, mg/dL 0.42 (0.17) 1.525 .0145Albumin, g/dL �1.18 (0.25) 0.306 �.0001 �Platelets 50,000/�L �0.0009 (0.002) 0.991 �.0001 �Iron saturation .57 (0.47) 1.763 0.23Iron binding (TIBC) �Iron grade

Risk percentilesRisk of developing an outcome

in 3.85 y, %

75th percentile 95th to �90th percentile 3490th percentile 62

ith HFE Genotypes Associated With Hemochromatosis

Hepatic histopathy

OutcomesFibrosis HAI Fat

48 BL 24 48 BL 24 48 BL 24 48 Clinical Histopathology

82Y homozygous)ND 5 5 ND 6 9 ND 2 3 ND None Already cirrhotic

� (compound heterozygous)0 3 3 2 6 5 4 1 1 1 None No0 4 ND 4 6 ND 6 0 ND 0 None No

ND 5 5 ND 4 7 ND 0 2 ND Yes Already cirrhoticND 3 2 ND 11 8 ND 3 1 ND None NoND 6 ND ND 6 ND ND 2 ND ND None Already cirrhotic1 3 3 3 8 8 9 2 1 1 None No1 3 1 6 9 3 8 2 0 2 None Yes0 3 3 3 12 8 8 1 1 2 None No0 6 4 6 8 9 7 3 2 1 Yes Already cirrhotic0 5 3 3 9 7 8 2 1 2 None Already cirrhotic

Status to the Ghany Model for Disease Progression in Advanced

Ghany’s disease progressionmodel � serum iron

Ghany’s disease progressionmodel � log serum ferritin

timate (SE) HR P value Estimate (SE) HR P value

0.75 (0.31) 2.118 .0147 0.85 (0.31) 2.328 .00720.51 (0.17) 1.659 .0024 0.46 (0.16) 1.588 .00491.22 (0.25) 0.296 �.0001 �1.19 (0.25) 0.304 �.0001.009 (0.002) 0.991 �.0001 �0.009 (0.002) 0.991 �.0001.001 (0.002) 0.999 .74

0.19 (0.18) 1.214 .28

Risk of developing an outcomein 3.85 y, %

Risk of developing an outcomein 3.85 y, %

8 938 3259 64

Ghany’s disease progressionmodel � TIBC

Ghany’s disease progressionmodel � global hepatocytic iron

score (iron-grade)

stimate (SE) HR P value Estimate (SE) HR P value

0.85 (0.30) 2.347 .0051 0.81 (0.31) 2.238 .00940.39 (0.16) 1.472 .0180 0.46 (0.17) 1.577 .00621.14 (0.25) 0.319 �.0001 �1.20 (0.25) 0.00 �.0001

0.009 (0.002) 0.991 �.0001 �0.009 (0.002) 0.991 �.0001

0.004 (0.001) 0.996 .00520.09 (0.11) 1.096 .4023

Risk of developing an outcomein 3.85 y, %

Risk of developing an outcomein 3.85 y, %

9 933 3560 62

LT, alanine aminotransferase; AST, aspartate aminotransferase; HR, hazard ratio; SE, standard error; TIBC, total iron binding capacity.

May 2011 IRON, HFE, AND OUTCOMES OF CHC 1500.e3

Supplementary Table 3. Clinical Outcomes in Subjects Studied According to the Presence and Degree of Stainable Iron inthe Liver

Outcomes

Stainable iron in hepatocytes (n � 787)% Portal triads positive for iron staining

(n � 787)

0–1�(n � 497)

2–4�(n � 290)

0–26%(n � 521)

�26%(n � 266)

n % n % n % n %

CTP �7 82 16.5 50 17.2 79 15.2 53 19.9Ascites 49 9.9 25 8.6 46 8.8 28 10.5SBP 3 0.6 2 0.7 3 0.6 2 0.8Encephalopathy 26 5.2 15 5.2 25 4.8 16 6.0HCC/PHCC 39 7.9 28 9.7 38 7.3 29 10.9Liver transplant 40 8.1 24 8.3 40 7.7 24 9.0Death, all causes 66 13.3 38 13.1 60 11.5 44 16.5Death, liver-related 32 6.5 27 9.4 35 6.8 24 9.2Two-point increase in Ishak fibrosis score 76 29.6 36 22.4 84 27.5 28 28.4

NOTE. None of the differences in percentages of subjects between the 2 groups in each major category is significant.