S. RICHARD*, H. H. BILLETT
Division of Haematology, Department of Medicine, *New York Presbyterian Hospital, New York, USA and The Albert Einstein College of Medicine,
Monte®ore Medical Center, Bronx, New York, USA
Summary We investigated the prevalence of positive viral hepatitis titres in sickle cell disease (SCD) and the relationship of abnormal liver function tests (LFTs) to transfusions and ferritin levels. Charts from 141 patients with SCD were reviewed and recent laboratory data on serum ferritin, hepatitis serology, units of packed red blood cells transfused and LFTs were collected. Hepatitis B core antibodies were positive in 14% of patients (12/86) and Hepatitis C viral antibody titres were positive in 16.5% (15/91). There was a relationship of positive serologies to transfusion for hepatitis C virus (HCV), but not for hepatitis B virus (HBV). Hepatitis C antibody negative (HCVAb)) patients had fewer packed red blood cells (pRBC) transfused than Hepatitis C antibody positive (HCVAb+) (6.4 vs. 20.3,P0.08). Patients with ferritins < 500 ng/ml compared to those with > 1000 ng/ml also showed a difference in units transfused (P< 0.003). Steady state LFTs, with the exception of alkaline phosphatase, had no relationship to serum ferritin or hepatitis serologies. Males were twice as likely to have positive serology as females but more females had elevated ferritin levels. Paired analysis of LFTs in steady state and crisis failed to demonstrate deterioration during crisis. We conclude that: (1) there is a relationship of positive Hepatitis C serology, but not Hepatitis B serology, to transfusion; (2) ferritin levels correlate with transfusion number but not with LFTs; (3) in our population, LFTs in SCD are usually normal and do not increase in vaso-occlusive crises.
Keywords Sickle cell, liver function, transfusion, iron-overloaded
Introduction
There is a paucity of recent data on liver function in sickle cell disease (SCD). While liver histology studies and autopsy series have demonstrated pathology secondary to sickle cell disease, the liver pathology abnormalities do not correlate well with symptoms or with abnormal liver enzymes (Green, Conley & Berthrong, 1953; Song, 1957; Rosenblate, Eisenstein & Holmes, 1970; Bauer, Moore & Hutchins, 1980). Abnormalities in liver function tests (LFTs) have been reported to be common and relatively mild in sickle cell patients in steady state. Data have also suggested that these abnormalities are more severe and nearly universal in vaso-occlusive crises, with haemoglobin SC (SC) patients faring much better than those with haemoglobin SS (SS) (Johnson et al., 1985; Ballaset al., 1982; Schubert, 1986). Much of
the information on sickle cell liver disease antedates the availability of current more sophisticated hepatitis testing. Recent articles on hepatitis C serologies in sickle cell disease have found a prevalence of positive serologies (10±20%) that is similar to a comparable nonsickle population (DeVaultet al., 1994; Hasanet al., 1996). A relationship of hepatitis C positivity with transfusions, as well as an association with increased LFTs has also been postulated (Devaultet al., 1994; Hasanet al., 1996). Studies looking at serum ferritin have found variable results in correlating the level of ferritin with the number of packed red blood cells (pRBC) transfused but have found signi®cant differences in these levels in steady state and crisis. (Petersonet al., 1975; Hussainet al., 1978; Vichinskyet al., 1981; Johnsonet al., 1985; Brownell, Lowson & Brozovic, 1986).
In view of the discrepancies noted and the more speci®c testing available, we decided to re-examine liver function test abnormalities, serum ferritin, hepatitis serologies and transfusion number in an effort to further clarify these issues.
Accepted for publication 2 October 2001
Correspondence: Henny Billett, Ullmann 903 A, Albert Einstein College of Medicine, 1300 Morris Park Ave, Bronx, NY 10461.
Materials and methods
Patient population
All charts and laboratory data on adult patients with sickle cell disease who were considered to have their primary care at the Bronx Comprehensive Sickle Cell Center at Monte®ore Medical Center, Bronx, New York, were reviewed. Steady state laboratory parameters were monitored during scheduled outpatient visits. Sickle cell crisis parameters were investigated during painful vaso-occlusive episodes requiring either: (a) a visit to the day hospital associated with the Bronx Comprehensive Sickle Cell Center (BCSS), designed to deal with milder episodes of pain; or (b) inpatient hospitalization. Crisis laboratory parameters were assessed within the ®rst three days of presentation.
Laboratory determinations
The steady statevs. crisis haematological parameters that were studied were Hb, MCV and red cell distribution width (RDW). LFTs similarly analysed were: gamma glutamyl transpeptidase (GGT), albumin, alkaline phosphatase (AP), serum aspartate transferase (AST), serum alanine transferase (ALT), and total, direct and indirect bilirubin (TB, DB and IB respectively). Ferritin levels were obtained in steady state only. Hepatitis serologies were performed by ELISA in the routine virology laboratory. The serologies tested were hepatitis B surface antigen (HBsAg), hepatitis B surface antibody (HBsAb), hepatitis B core antibody (HBcAb) and hepatitis C antibody (HCV Ab). With their consent, all patients seen in the Sickle Cell Center who had not had certain prior laboratory determinations and were available to be tested had those laboratory tests performed. RBC transfusions
The number of packed red blood cell (pRBC) units transfused were assessed by retrieving data on the number of units issued over the entire association, usually the lifetime, of the patient at Monte®ore Medical Center (MMC). No outside hospital transfusion data were inclu-ded. No data on chelation was obtained.
Statistical analysis
Data was entered into Microsoft Excel data sheets and analysed statistically using paired and unpairedt-tests and ANOVA, where applicable.
Results
Patient demographics
The 141 patients included in the study ranged in age from 19 to 77 years (mean 32.9 9.11 years) and included 64 females and 77 males. There were 97SS, 17SC, seven HbSb+, ®ve HbSb0 and one HbSOArab. There were 14 patients whose electrophoresis demonstrated a single S band but were not further genotyped; these were included only in the virology and in some ferritin studies and are noted in the appropriate results below. Sixty-®ve patients were examined in crisis, and steady state and paired analyses were performed.
Hepatitis serologies
These data are depicted in Table 1. A total of 93 patients had some or all hepatitis serology studies: 47 males and 46 females; seventy-one patients were SS and eight SC, four Sb+, ®ve Sb0, one SOAraband four were not further genotyped. All 91 patients who had HBsAg studies done were negative. Of the 91 patients who were tested for HCV, 15 (16.5%) were positive. Of these 15 HCV Ab+ patients, 10 were male and ®ve were female with a mean age of 32.7 years. HBcAb serologies were performed on 86 patients: 12 (14%) were positive and of these 12, eight were male and four female, with a mean age of 37.2 years (range 27±48 years). Eleven of the patients were SS and one was SC. Of 84 patients with HBsAb serologies, 17 (20.2%) were positive: 11 male and six female, with a mean age of 29.6 years (range 21±45 years), 14 of whom were SS, one SC, one SO Arab and one SS phenotype (not further studied). If any positive hepatitis B serology was considered an indication of previous infection, i.e. HBsAb and/or HBcAb+, then 23/89 were positive (25.8%). However, of the 17 HBsAb positive patients, 11 were HBcAb). As patients often could not remember and we could not differentiate past exposure from immunization on the basis of these serologies, these HBsAb+ HBcAb) patients were not included, except where stated, in the study correlations. There were twice as many males as females with hepatitis serology positivity, although almost equal numbers of males and females were studied. The ratio of SS : SC positivity was higher than the ratio of SS : SC patients (Table 1).
Only two of the patients were both HBcAb+ and Hepatitis C antibody positive (HCVAb+) but 25/93 (26.9%) had either HCVAb+ and/or HBcAb+. If those with HBsAb positivity are included as well, the number increased to 35/93 (37.6%).
Relationship of hepatitis serologies to transfusions Table 2 depicts hepatitis serologies and number of units transfused at MMC. Of the 12 patients who were HBcAb+, ®ve (41.6%) had received > 10 U. Similarly, of the 15 HCVAb+ patients, nine (60%) had received > 10 U. Of the 91 patients with both transfusion and HCV Ab testing, nine of the 34 patients who had received > 10 U pRBC (26.4%) were HCVAb+ in comparison to six of the 57 (10.7%) who had received < 10 U transfusion. This magnitude of difference was not observed for the 86 patients with both transfusion and HBcAb data: ®ve (15.6%) of the 32 heavily transfused patients were positive for HBcAb, similar to the ®gure of 13.2% HbcAb+ (7/54) who received fewer than 10 U.
Patients who had HCVAb+ had an average of 20.3+ 26.0 number of units of pRBC transfused, whereas the patients who had HBcAb+ had 9.3 + 9.4 number of units of pRBC transfused. Patients with completely
negative serologies had a mean of 6.4 + 10.6 units of pRBC (Table 2).
Ferritin and transfusions
There was a consistent increase in ferritin levels and number of transfused units (Table 3). Patients with ferritin in the 0±200 ng/ml range had a mean of 5.4 + 8.9 U of transfusions and those with ferritin levels 201±500 ng/ml had 7.3 + 12.3 of PRBC.
In contrast, patients with ferritins of 501±1000 ng/ml had an average of 28.1 + 46.1 U transfused and those with ferritin > 1000 ng/ml had 36.3 + 37.0 U transfused. The P-value comparing the mean number of transfu-sions for patients with either low ferritin levels or 201±500 ng/ml with those > 1000 ng/ml was highly signi®cant (P0.002 and 0.003, respectively). Patients with ferritin 501±1000 ng/ml and with > 1000 ng/ml could not be differentiated from each other.
Ferritin and hepatitis serology
Only nine patients had both hepatitis serology and serum ferritin data available. Patients who were HCVAb+ were more likely to be in the high ferritin groups (Table 3). This did not hold for the six patients who were HBcAb+ and had ferritin data available. Ferritin levels were higher in patients with any positive serology compared with those in completely negative serology, 1424.8 + 1479.6 ng/ml (n23) vs. 674.1 + 1321.3 ng/ml (n42), respect-ively,P0.05.
Table 2. Hepatitis serologies of sickle cell disease patients divided into low transfusion (< 10 U pRBC) and high transfusion number and their relationship to positive and negative hepatitis serologies
HCVAb + (%) HBcAb + (%) Both negative < 10 U transfused 6/57 (10.7%) 7/54 (13.2%) 39/50 > 10 U transfused 9/34 (26.4%) 5/32 (15.6%) 11/30 Total 15/91 (16.5%) 12/86 (14%) 50/80 No. Units tx'ed (mean SD) 20.3 26.0 9.3 9.4 6.4 10.6
Table 3. Strati®ed ferritin levels as related to patient demographic data, transfusion number and hepatitis serology Ferritin (ng/ml) (n) 0±200 (31) 201±500 (22) 501±1000 (12) > 1000 (20)
M : F 17 : 14 11 : 11 9 : 3 6 : 14
SS : SC 19 : 8 17 : 1 9 : 0 18 : 0
Units tx'ed (mean SD) 5.4 8.9* 7.3 12.3* 28.1 46.1 36.3 37.0*
Units tx'ed (Range) 0±37 0±47 0±52 0±116
HCVAb+ (n9) 1 1 3 4
HBcAb+ (n6) 2 2 1 1
*Pbetween > 1000 ng/ml and other groups < 0.003.
Table 1. Hepatitis serologies of MMC sickle cell patients. Results of serologies performed on patients with sickle cell disease in steady state
No.
tested %Positive M : F SS : SC Age(years)
HBsAg 91 0 ± ± ±
HBcAb 86 14.0 2 : 1 11 : 1 37.2 HBsAb 84 20.2 1.8 : 1 14 : 1 29.6 HCV 91 16.5 2 : 1 13 : 1 32.7
Comparison of SS and SC
Results from patients with SS vs. SC disease were compared by an unpaired Student's t-test (Table 4). As expected, differences in the CBC parameters of Hb, MCV and RDW were statistically signi®cant. The Hb in SC was an average of 3.3 g/dl higher than SS. Mean MCV in SS was 89.6¯ while in SC it was 80.5¯ (P0. 0002). The RDW was signi®cantly higher in SS than SC (21.1 vs. 16.6; P< 0.0001). Signi®cant differences in LFTs were noted between SS and SC in levels of GGT, AST, TB, DB and IB. The GGT in SSvs. SC was 89.1 + 122.0 IU/l vs. 27.0 + 13.8 IU/l, respectively (P0.005). AST in SS was 35.9 + 16 IU/l and in SC was 22.3 + 10.3 IU/l (P 0.0002). The total bilirubin in SS was 3.4 + 2.2 mg/dl and 1.4 + 0.4 mg/dl in SC (P< 0.0001) and mean indirect bilirubin in SS was 2.9 + 1.9 mg/dl and in SC was 1.1 + 0.3 mg/dl (P< 0.0001) (Table 4). When direct bilirubin was separately analysed, the mean level in SS patients was 0.6 + 0.9 mg/dl, while in SC it was 0.2 + 0.1 mg/dl (P0.006). GGT was also increased and was abnormal in SS, whereas it remained normal in SC, but the standard deviation was very large.
LFTs in steady state vs. crisis
Paired analyses of LFTs in steady state and crisis were investigated for each patient of each Hb subgroup and
there were no signi®cant differences in levels for any of the LFTs analysed (AP, GGT, albumin, DB, IB, TB, AST, ALT) in the non-SS groups, but the numbers were small. In SS disease (Table 5), however, GGT and AST showed a difference during vaso-occlusive crisis with the GGT in steady state being 91.9 + 76.8 IU/l and in crisis 80.6 + 72.7 IU/l (P0.007), while the AST appeared to increase during crisis from 35 + 17 IU/l in steady state to 44 + 26 IU/l in crisis (P0.01). For unpaired HbSS patients, ALT was normal in 72/83 patients in steady state and in 70/80 patients in crisis, direct bilirubin was normal in 81/83 in steady state and 75/80 during crisis and alkaline phosphatase was normal in 62/76 patients in steady state and in 60/76 patients in crisis. In addition, as indirect bilirubin can be used as an index of dense cells and thus indirectly, of both the degree of anaemia and of sickle cell turnover, it is notable that there was no correlation of indirect bilirubin with LFTs (Billett, Nagel & Fabry, 1988).
Steady state LFTs vs. ferritin in SS
LFTs were compared with ferritin values. Only alkaline phosphatase showed increased levels as ferritin increased; the average levels of AP in patients with ferritin < 500 ng/ml were 88.3 IU/lvs. 140.9 IU/l for those with ferritin > 1000 ng/ml (P0.02). SS SC P-value Hb (g/dl) 8.1 1.3 11.4 1.2 < 0.0001 MCV (¯) 89.6 9.8 80.5 6.4 0.0002 RDW 21.1 3.8 16.6 1.9 < 0.0001 GGT (IU/l) 89.1 122 27 13.8 0.005 AST (IU/l) 35.9 16 22.3 10.3 0.0002 TB (mg/dl) 3.4 2.2 1.4 0.4 < 0.0001 DB (mg/dl) 0.6 0.9 0.2 0.1 0.006 IB (mg/dl) 2.9 1.9 1.1 0.3 < 0.0001 AP (IU/l) 98.8 61.2 77.6 35.2 0.08 ALT (IU/l) 21.4 15.7 14.9 10.9 0.07
Table 4. Comparison of haematological parameters and LFTs in HbSS and HbSC patients in steady state
Steady state Crisis P-value GGT (IU/l) 91.9 76.8 74.3 70.8 0.007 AST (IU/l) 35 17 44 26 0.01 ALT (IU/l) 21.8 14.7 28.4 26.9 NS AP (IU/l) 99.0 56.4 93.8 38.1 NS TB (mg/dl) 3.14 2.04 4.2 4.7 NS DB (mg/dl) 0.47 0.24 0.62 0.69 NS IB (mg/dl) 2.92 2.2 3.81 3.63 NS NS, not signi®cant.
Table 5. A paired analysis of LFTs in steady state and vaso-occlusive crisis in patients with SS disease
Steady state LFTs vs. hepatitis serologies
Analysis of steady state LFTs vs. hepatitis serologies in patients whose serologies were positive showed no differ-ences in any of the eight LFTs measured; although the ALT was 25.1 + 13.3 IU/l in HCVAb+, and was 21.5 + 9.3 IU/l in HBcAb+ and in patients with negative serologies was 19.3 + 9.2 IU/l, the difference was not signi®cant (P> 0.05 (Table 6).
Discussion
There have been two articles published recently on hepatitis C serologies in sickle cell disease. Hasan et al. (1996) found a hepatitis C prevalence of 10.1% in 99 patients (85 SS, 8 SC and 6 Sb0or Sb+). De Vaultet al. (1994) studied 121 patients and found 20.7% to be HCV Ab+. In our study, 91 patients had hepatitis C serologies done, and we found a hepatitis C positivity rate of 16.5% (15/91), approximately in between the two ®gures previ-ously demonstrated. As all of these patients in our adult sickle cell population were over 21 years of age and had typically been followed throughout their lives by the BCSS, most patients had been transfused prior to the widespread use of third generation HCV ELISA testing after 1995.
Fewer data have been published on hepatitis B serolo-gies in sickle cell disease. Johnsonet al. (1985) found the prevalence of HBsAg, sAb or cAb to be19% in both SS and SC patients. In our study group of patients all 91 had negative HBsAg, but 14% had HBcAb+ and 20.2% had HBsAb+. We found an overall prevalence of 27% of hepatitis B or C serology positivity in the study population. Interestingly, males had twice the incidence of hepatitis serology positivity as females, although approximately equal numbers of men and women were studied (47 and 46, respectively). This male/female discrepancy has not been previously noted. Also, the ratio of SS to SC with positive hepatitis serology was 10±15 : 1, which is higher than expected from the number studied (approximately 8 : 1), but probably represents the decreased transfusion need of SC.
Hasanet al. (1996) found 23.3% of patients with > 10 U were HCVAb+ and 7.9% with < 10 U transfused were HCVAb+. None of the patients who had never been
transfused were HCVAb+. Devault et al. (1994) found 30.3% with > 10 U transfused were HCVAb+ and 8.6% of those who received < 10 U transfused were HCVAb+. This study agrees with these ®ndings in that 26.4% of patients who had received > 10 U transfusions were HCVAb+ and 10.7% with < 10 U transfusions were HCVAb+. In contrast, when HBcAb positivity was compared with number of transfusion units, no such association was made: 15.6% of patients who had received > 10 U transfusions were HBcAb+, while 13.2% of patients with < 10 U transfusion were HBcAb+.
Hepatitis serology was correlated with mean number of units transfused. Hepatitis C serologies showed a relation-ship with the mean number of units transfused while hepatitis B did not, again demonstrating lack of a relationship between hepatitis B serologies and transfu-sion.
We used steady state serum ferritins as there is some data documenting increases in serum ferritin with crisis (Brownell, Lowson & Brozovic, 1986). Of interest is that the > 1000 ng/ml ferritin group had more women than men although the ratio of males to females who transfused > 10 U was approximately 2 : 1. There have been some studies attempting to correlate serum ferritin with trans-fusions and the results have not been consistent, with some reports showing no correlation (Hussain et al., 1978; Peterson et al., 1975), while others show some relationship (Vichinsky et al., 1981). Of note is the study by Johnson et al. (1985) which found, as we did, signi®cantly higher ferritin levels in females than in males in SS, but these levels correlated with transfusion history which, unlike ours, was signi®cantly greater in women. In SC however, the same group found signi®cantly higher ferritin in males and no difference in transfusion history between the sexes. Our data supports the relationship of transfusions to ferritin and this difference is signi®cant between most groups. The reason for the gender difference is unclear at this stage.
Patients with HCVAb+, HBsAb+ or HBcAb+ tended to have higher ferritins than those with completely negative serologies. The mean ferritin level in the former group was 1424.8 + 1479.6 ng/ml, while the mean ferritin level in the negative serology group was 674.1 + 1321.3 ng/ml. However, the number of patients in each category was Table 6. Mean levels of LFTs in patients
with HbSS steady state divided according to hepatitis serology status
Hepatitis serology ALT (IU/l) DB (mg/dl) AP (IU/l) GGTP (IU/l) HCVAb+ 25.1 13.3 1.4 2.5 116 47.3 42.3 27.6 HBsAb+ 21.5 9.3 2.0 3.4 110.3 63.5 40.6 12.4 Negative serologies 19.3 9.2 0.3 0.1 101.5 116.4 77.4 103.9
small. HCVAb+ patients were likely to have higher ferritins, while HBcAb+ patients were more often found in the lower ferritin ranges, again supporting the rela-tionship of hepatitis C but not B serologies to the transfusions described above.
Liver function abnormalities have been assumed to be common and mild in asymptomatic SS patients and to become more severe in crisis. A paired analysis of LFTs in SS disease in steady state vs. crisis showed no real signi®cant differences in our study. This is in contrast to data we have previously published which had demonstra-ted a fall in indirect bilirubin that correlademonstra-ted with the decrease in dense cells during crisis (Billett, Nagel & Fabry, 1988). However, in that study, the data was collected daily whereas in this study, values within the ®rst three days were averaged and the signi®cant differences may not be obvious.
We found that the majority of SS patients had normal LFTs, in contrast to previously published data (Ballas et al., 1982). The LFTs of AST, TB, and IB were signi®cantly different in SS vs. SC but probably re¯ect haemolysis rather than liver disease. GGT and DB were the only LFTs not directly related to haemolysis that were signi®cantly higher in SS than in SC. The difference in direct bilirubin levels is small but has been noted previously ( Johnson et al., 1985) and may represent haemoglobin load. Although these are statis-tically signi®cant differences, the levels are probably not of clinical relevance as the vast majority still fell within normal range.
In one study (DeVaultet al., 1994) 82% of patients with increased ALT were HCVAb+ in contrast to only 14% with normal ALT. We found no such signi®cant association. More recent data from Seeff et al. (2000) suggests that HCVAb+ non-SS patients had lower liver related morbidity and mortality rates and may be at less risk for progressive liver disease than was previously thought. This agrees with our data demonstrating a lack of abnormal LFTs associated with high ferritin levels and suggests a lack of overt liver pathology in association with transfusion in the majority of our patients. It is unclear whether the criteria for haemosiderosis organ damage is the same for SS disease as for AA individuals.
In summary, we found the prevalence of hepatitis C to be 16.5% in our patients with sickle cell disease, and the prevalence of hepatitis B or hepatitis C positive patients to be 27%. Males had twice the incidence of hepatitis serology positivity as females. Hepatitis C serology showed a correlation with transfusion history in our patients, while hepatitis B serology did not, though patients with
completely negative serologies had the least number of transfusions.
Serum ferritin levels were related to number of trans-fusions, more women had ferritin levels > 1000 ng/ml, even though the ratio of males to females transfused > 10 U was approximately 2 : 1. Patients with positive hepatitis serologies tended to have higher ferritins, especially hepatitis C serology positivity. LFTs and haematological parameters were consistent with a higher degree of haemolysis in SS than in SC disease. The majority of SS patients, however, had normal LFTs in contrast to previously published data and there was no signi®cant association between ALT or other LFT abnormalities and HCVAb + in our patients.
We demonstrate here, then, that the majority of patients with sickle cell disease have normal liver function as measured by standard biochemical parameters, despite transfusion and exposure to transfusion transmitted viruses. This improvement is in contrast to previous studies and may represent the better health care delivery now available to our patients. With improved blood screening techniques and rapid treatment of sickle cell vaso-occlusive pathologies, liver pathology secondary to sickle cell disease may become an even more uncommon event.
References
Ballas S., Lewis C., Noone A., Krasnow S., Kamarulzaman E. & Burka E. (1982) Clinical, hematological, and biochemical features of HbSC disease.American Journal of Hematology13, 37±51.
Bauer T., Moore G. & Hutchins G. (1980) The liver in sickle cell disease: a clinicopathologic study of 70 patients. American Journal of Medicine69,833±837.
Billett H., Nagel R. & Fabry M. (1988) Evolution of laboratory parameters during sickle cell painful crisis: evidence compat-ible with dense red cell sequestration without thrombosis. American Journal of Medical Sciences296,293±298.
Brownell A., Lowson S. & Brozovic M. (1986) Serum ferritin concentration in sickle cell disease.Journal of Clinical Pathology 39,253±255.
DeVault K., Friedman L., Westerberg S., Martin P., Hosein B. & Ballas S. (1994) Hepatitis C in sickle cell anemia. Journal of Clinical Gastroenterology18,206±209.
Green T., Conley C. & Berthrong M. (1953) The liver in sickle cell anemia.Bulletin of the Johns Hopkins Hospital92,99±122. Hasan M., Marsh F., Posner G., Bellevue R., Dosik H., Suatengco
R. & Ramani N. (1996) Chronic hepatitis C in patients with sickle cell disease. American Journal of Gastroenterology 91, 1204±1206.
Hussain M., Davis L., Laulicht M. & Hoffbrand A. (1978) Value of serum ferritin estimation in sickle cell anemia. Archive of Disease in Childhood53,319±321.
Johnson C., Omata M., Tong M., Simmons J. Jr, Weiner J. & Tatter D. (1985) Liver involvement in sickle cell disease.Medicine64, 349±356.
Peterson C., Graziano J., de Ciutiis A., Grady R., Cerami A., Worwood M. & Jacobs A. (1975) Iron metabolism, sickle cell disease and response to cyanate.Blood46,583±590. Rosenblate H., Eisenstein R. & Holmes A. (1970) The liver in
sickle cell anemia. A clinical-pathologic study. Archives of Pathology90,235±245.
Schubert T. (1986) Hepatobiliary system in sickle cell disease. Gastroenterology90,2013±2021.
Seeff L., Miller R., Rabkin C., Buskell-Bales Z., Straley-Eason K., Smoak B., Johnson L., Lee S. & Kaplan E. (2000) 45-year fol-low-up of hepatitis C virus infection in healthy young adults. Annals of Internal Medicine132,105±111.
Song Y. (1957) Hepatic lesions in sickle cell anemia. American Journal of Pathology33,331±344.
Vichinsky E., Kleman K., Embury S. & Lubin B. (1981) The diagnosis of iron de®ciency anemia in sickle cell disease.Blood 58,963±968.
