1. Top of page
  2. Abstract
  6. Acknowledgements

Adherence to antiviral therapy is essential to achieve sustained virological responses in patients treated for hepatitis C. An important limitation to use of appropriate doses of ribavirin is development of anemia. The aim of this study is to identify risk factors associated with anemia in liver transplant recipients undergoing treatment for recurrent hepatitis C virus (HCV). Retrospective analysis was performed on 115 adult liver transplantation (LT) recipients who received antiviral treatment. Anemia was defined as hemoglobin of <10 gm/dL or the use of erythropoietin replacement therapy. Variables found to be significant in univariate analysis were further studied in multivariate analysis. The mean (± standard deviation [SD]) age of our cohort was 52.1 (± 8.8) yr. Anemia developed in 44 patients (38.3%). Mean (± SD) onset of anemia was 8.9 (± 6.8) weeks after initiation of antiviral therapy. A total of 30 patients (26%) required erythropoietin replacement, at a mean (± SD) of 7.9 (± 6.0) weeks after start of antiviral treatment. A total of 27 patients (24%) required ribavirin dose reduction, at a mean (± SD) time to dose reduction of 8.1 (± 6.3) weeks. In univariate analysis, body mass index (BMI) (P < 0.01), mycophenolate mofetil use (P = 0.05), trimethoprim-sulfamethoxazole (P = 0.02), and age (P = 0.02) were statistically significant. In conclusion, in multivariate analysis, BMI (P < 0.01) and age (P = 0.02) were found to be independent predictors of anemia. Anemia is common in liver transplant recipients treated for recurrent HCV. Special vigilance is required for older patients and patients with a low BMI. Liver Transpt 13:1032–1038, 2007. © 2007 AASLD.

Hepatitis C virus (HCV) is the most common indication for liver transplantation (LT). Histologic recurrence occurs in 75 to 80% of patients after LT, with 30% progressing to cirrhosis within 5 to 9 yr.1–4 As a result of this accelerated course of HCV infection, long-term graft and patient survival may be reduced in liver transplant recipients.5, 6

Ribavirin with interferon (IFN) or pegylated IFN is the most effective therapies to treat recurrent HCV in transplant recipients.7, 8 Sustained virological response ranges between 10 and 30%.7, 9, 10 However, treatment is often limited by the development of multiple adverse effects, such as anemia, which occurs in approximately 43 to 76% of transplant recipients.11–16

A number of potential mechanisms are believed responsible for the development of anemia such as IFN-associated myelosuppression, increased production of inflammatory cytokines such as hepcidin, and HCV interference with erythropoietin production.17–20 However, the most common reason for the development of anemia is ribavirin dose-dependent hemolysis.21, 22 Accumulation of phosphorylated ribavirin in the red blood cell leads to oxidative injury and consequently hemolytic anemia.23–26

Anemia can impact quality of life and treatment compliance, and can be associated with cardiovascular events.11, 27–32 Anemia is one of the leading causes of dose reduction and early cessation of antiviral therapy for recurrent HCV in transplant recipients.11–16 Anemia can also reduce the probability of achieving a sustained virological response. In nontransplant patients, ribavirin adherence has been demonstrated to be essential to achieve a virologic response.33, 34

Because of the clinical and viral impact of treatment-associated anemia, we sought to identify its predictive factors. Recognition of these factors can lead to increased vigilance and perhaps the institution of preeminent steps to avoid this adverse effect. The goal of the current study was to identify the causes of anemia in transplant recipients undergoing treatment for hepatitis C.


  1. Top of page
  2. Abstract
  6. Acknowledgements

All adult (age >18 yr) recipients transplanted at the University of California Los Angeles Medical Center between 1998 and 2005 and treated for recurrent hepatitis C with IFN and ribavirin, alone or in combination, were studied. Criteria for beginning antiviral therapy included detectable HCV ribonucleic acid, a fibrosis score of 2 or more, and/or elevation of alanine aminotransferases at least twice the upper limit of normal. Some of the recipients may have been reported in other series from our institution.35, 36

Data was obtained by medical chart review and the University of California Los Angeles Liver Transplant database. Demographic data, comorbid conditions, liver pathology results, immunosuppressant type, and laboratory data was recorded. Hematologic and biochemical data was collected prior to initiation of antiviral therapy and for the first 24 weeks of therapy. Information regarding dose reduction and use of growth factors and blood transfusions was also recorded. All data was collected by Health Insurance Portability and Accountability Act–certified, study-authorized individuals, in accordance with institutional review board requirements. Collected data was input into a large database that did not include any patient identifiers.

Anemia was defined as hemoglobin level less than 10 gm/dL and/or the use of erythropoietin replacement therapy.18, 37 Fibrosis was scored as: 0, no fibrosis; 1, enlarged fibrotic portal tracts; 2, periportal or portal-portal septa, but intact architecture; 3, architectural distortion but no obvious cirrhosis; and 4, probable or definite cirrhosis as previously described.38, 39 Creatinine clearance was estimated using the Cockcroft and Gault40 formula.

The immunosuppression regimens after LT consisted of combinations of cyclosporine, prednisone, tacrolimus, sirolimus, and mycophenolate mofetil. According to our protocol, patients initially received 1 gm of intravenous methylprednisolone the day of transplantation; this was then tapered to 20 mg/day over 1 week. Oral prednisone was subsequently started at 20 mg/day and tapered as tolerated. All patients were treated with trimethoprim-sulfamethoxazole for 1 yr. Our protocol for posttransplantation target immunosuppression levels has been previously described.41

Initially, we used IFN therapy at a dose of 3 million units subcutaneously three times weekly. When pegylated IFN became available, pegylated IFN alfa 2a 90 μg or alfa 2b 0.5 mg/kg was administered subcutaneous once every week, and dose increased as tolerated to 180 μg or 1.5 mg/kg subcutaneously once every week, respectively. Ribavirin was started at 400 mg per oral twice daily, and the dose was increased as tolerated to a goal dose of 12 mg/kg. Previously, ribavirin was used as monotherapy for its proposed antiinflammatory effects. When IFN use became acceptable in transplant recipients and when ribavirin use was not found to be efficacious, the use of ribavirin use as monotherapy was discontinued.

Clinic visits and laboratory studies were performed 2 weeks after initiating therapy, and every month after. Dose reduction of ribavirin occurred when hemoglobin dropped below 10 gm/dL, and treatment stopped when hemoglobin fell below 8.5 gm/dL. Blood transfusions were used if hemoglobin fell below 8.0 gm/dL and/or if the patient became symptomatic.

Continuous variables are presented throughout as means (± standard deviation [SD]). Categorical variables are expressed as a percentage. Logistic regression analysis was used to evaluate predictors of anemia with a hemoglobin cutoff of 10 gm/dL. Predictors included demographic data, baseline laboratory data, immunosuppressant and other medications, cytomegalovirus exposure, fibrosis stage, and genotype.

Variables found to be significant in univariate analyses were further studied in a multivariate analysis. Stepwise backward techniques were used to select variables in the final model with probability of 0.05. Standard model checking techniques were used and model fit was verified.42–45 A P value of 0.05 or less in a 2-tailed test was considered statistically significant. Statistical analyses were performed using STATA (STATA Corporation, College Station, TX).


  1. Top of page
  2. Abstract
  6. Acknowledgements


A total of 115 liver transplant recipients received antiviral therapy for recurrent HCV during the study period (Table 1). Most patients were men, and were recipients of deceased donors. Less than one-third of the patients had concomitant diabetes. None of the patients in this study were on dialysis. Of the total study population, 94 (82%) had prior exposure to cytomegalovirus. A total of 59 (51%) of the patients were on trimethoprim-sulfamethoxazole. Tacrolimus and cyclosporine were used in 86.1% (n = 99) and 13.9% (n = 16) of patients, respectively. Mean number of weeks of prednisone after LT was 58.0 (± 102.9) weeks. Mycophenolate mofetil was used as part of the immunosuppressant regimen in 42 patients (36.5%). Baseline laboratories and disease severity are shown in Table 2. Data was complete for all variables except liver biopsy fibrosis stage in 5 patients (4%) and HCV genotype in 43 patients (37%).

Table 1. Patient Characteristics (n = 115)
Age, mean (± SD) (yr)52.1 (± 8.8)
Body mass index, mean (± SD)27.5 (± 4.9)
Diabetes, number (%)32 (28)
Gender, male, number (%)84 (73)
Race, number (%) 
 Latino24 (21)
 Caucasian77 (67)
 Native American2 (1.7)
 Asian8 (7.0)
 African American4 (3.5)
Table 2. Baseline Laboratory and Disease Severity
Baseline laboratory 
  1. Abbreviation: SD, standard deviation.

 Creatinine, mean (± SD) (mg/dL)1.1 (±0.3)
 Creatinine clearance, mean (± SD) (mL/minute)88.6 (±33.9)
 Aminotransferase, mean (±SD) 
  Aspartate aminotransferase (IU/mL)136.8 (±105.4)
  Alanine aminotransferase (IU/mL)188.9 (±139.6)
 Hemoglobin, mean (± SD) (gm/dL)13.1 (±1.6)
 Total bilirubin, mean (± SD) (mg/dL)1.5 (±2.0)
Disease severity 
 Fibrosis stage, number (%) 
  021 (18)
  121 (18)
  249 (43)
  313 (11)
  46 (5)
  Unknown5 (4)
 Hepatitis C genotype, number (%) 
  152 (45)
  26 (5)
  310 (9)
  42 (2)
  62 (2)
  Unknown43 (37)

Mean (± SD) baseline hemoglobin and starting ribavirin dose was 13.2 (± 1.6) gm/dL and 626.7 (± 234.8) mg, respectively. Combination therapy (ribavirin with pegylated-IFN or IFN) and ribavirin monotherapy was used in 57 (49.6%) and 55 (47.8%) patients, respectively. A total of 84 patients (73%) were on antiviral therapy for at least 24 weeks.


Thirty-eight percent (n = 44) of our study cohort developed anemia with a mean (± SD) onset of 8.9 (± 6.8) weeks from the start of antiviral therapy. The mean (± SD) hemoglobin at the onset of anemia was 9.0 (± 0.7) gm/dL when first detected. The mean (± SD) ribavirin dosage at the onset of anemia was 666.7 (± 264.9) mg. Using the lowest recorded hemoglobin, the mean (± SD) drop in hemoglobin in anemic patients was 3.9 (± 1.5) gm/dL, which represented a 31% decrease from baseline.

Twenty-four percent (n = 27) of the total study population underwent ribavirin dose reduction. Mean (± SD) time to dose reduction for these patients after starting therapy was 8.1 (± 6.3) weeks. Three patients (2.6%) required blood transfusions at a mean (± SD) time of 6.0 (± 3.0) weeks from the start of antiviral therapy. A total of 30 patients (26%) were started on erythropoietin. Mean (± SD) hemoglobin at the start of erythropoietin use was 9.8 ± 1.6 gm/dL. However, 4 of the 30 patients were started on erythropoietin at a hemoglobin greater than 10 gm/dL; mean (± SD) hemoglobin of 10.3 ± 0.4 gm/dL. The mean (± SD) ribavirin dosage at the time erythropoietin was started was 640.0 ± 264.2 mg. Of those on erythropoietin, 14 patients (47%) also underwent ribavirin dose reduction. Conversely, of the 27 patients that underwent ribavirin dose reduction, 14 patients (52%) required erythropoietin. The incidence of anemia in patients on combination therapy (ribavirin with pegylated-IFN or IFN) and ribavirin monotherapy anemia was 31% and 40%, respectively, which did not represent a significant difference.

Univariate and Multivariate Analysis

In the univariate analysis, the following factors were found to be statistically significant and associated with anemia: body mass index (BMI) (odds ratio [OR], 0.97; 95% confidence interval [CI], 0.95-0.99; P < 0.01), mycophenolate mofetil use (OR, 2.17; 95% CI, 0.99-4.74; P = 0.05), trimethoprim-sulfamethoxazole use (OR, 2.58; 95% CI, 1.18-5.64; P = 0.02), and age (OR, 1.07; 95% CI, 1.01-1.12; P = 0.02). Specific antiviral regimen, starting ribavirin dosage, and duration of treatment were not significantly associated with anemia (Table 3). In the multivariate model, BMI (OR, 0.97; 95% CI, 0.95-0.99; P < 0.01) and age (OR, 1.07; 95% CI, 1.01-1.13; P = 0.02) were found to be independent predictors of anemia.

Table 3. Variables Influencing Anemia in Univariate Analysis
VariableP valueOR (95% CI)
  • *

    Statistical significance in univariate analysis (P < 0.05).

Baseline creatinine (mg/dL)0.420.61 (0.18–2.06)
Baseline creatinine clearance (mg/dL)0.651.00 (0.99–1.01)
Baseline white blood cell (103/uL)0.860.99 (0.84–1.15)
Baseline hemoglobin (gm/dL)0.610.93 (0.74–1.19)
Baseline platelet (103/uL)0.130.99 (0.99–1.00)
Baseline glucose (mg/dL)0.741.00 (0.99–1.00)
Baseline aspartate aminotransferase (U/L)0.281.00 (0.99–1.00)
Baseline alanine aminotransferase (U/L)0.721.00 (0.99–1.00)
Baseline total bilirubin (mg/dL)0.340.89 (0.67–1.15)
Body mass index*<0.010.97 (0.95–0.99)
Age (yr)*0.021.07 (1.01–1.12)
Gender0.180.56 (0.24–1.30)
Race0.780.97 (0.78–1.21)
Weight (kg)0.260.99 (0.98–1.00)
Diabetes0.600.79 (0.34–2.86)
Mycophenolate mofetil*0.052.17 (1.00–4.74)
Tacrolimus0.951.03 (0.35–3.09
Cyclosporine0.950.96 (0.32–2.87)
Sirolimus0.731.63 (0.10– 26.71)
Trimethoprim-sulfamethoxazole*0.022.58 (1.18–5.64)
Fibrosis stage0.120.78 (0.57–1.07)
Hepatitis C genotype0.220.90 (0.77–1.06)
Hepatitis C viral load (IU/mL)0.341.01 (0.99–1.02)
Ribavirin0.861.24 (0.11–14.16)
Interferon0.162.91 (0.66–12.82)
Pegylated interferon0.671.18 (0.55–2.50)
Duration of antiviral treatment (weeks)0.471.03 (0.94–1.13)
Starting ribavirin dose (mg)0.370.88 (0.67–1.16)
Starting ribavirin dose (mg/kg)0.620.97 (0.87–1.09)

Analysis was also performed on age and BMI as categorical variables. Age > 55 yr was significantly associated with anemia (OR, 2.61; 95% CI, 1.14-5.97; P = 0.022). The percentage of patients with anemia by age group is shown in Figure 1. BMI < 25 was also significantly associated with anemia (OR, 3.03; 95% CI, 1.29-7.15; P = 0.011), as seen in Figure 2.

thumbnail image

Figure 1. Percentage of patients with anemia by age group, P = 0.022. Cohort size for each stratum: <35 yr, 4 patients; 35-44 yr, 13 patients; 45-54 yr, 59 patients; 55-64 yr, 32 patients; and >65 yr, 7 patients.

Download figure to PowerPoint

thumbnail image

Figure 2. Percentage of patients with anemia by BMI, P = 0.011. Cohort size for each stratum: 18.5-24.9, 34 patients; 25-29.9, 54 patients; 30-40, 24 patients, and > 40, 3 patients.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  6. Acknowledgements

Current regimens used to treat HCV can be effective in the LT population, however side effects such as anemia significantly impede treatment success and quality of life.11–16, 27, 29, 46, 47 This study identified independent predictors of anemia to help recognize patients at risk of developing anemia during antiviral treatment. Our study is one of the largest reviews of anemia in posttransplantation patients and represents the first to investigate these risk factors in the liver transplant population. Age and BMI were found to be independent predictors of anemia in transplant recipients.

In this study, we found that the incidence of posttransplantation anemia in patients undergoing HCV antiviral therapy was 40%. A total of 27 patients (27%) underwent ribavirin dose reduction, 30 patients (26%) were started on erythropoietin replacement, and 3 patients (2.6%) required blood transfusions. In those that developed anemia, decreases in hemoglobin occurred soon after starting antiviral therapy (mean 8.9 ± 6.8 weeks).

Our results confirm previous observations that age is an independent predictor of anemia48–51 (Fig. 1). A potential mechanism is increased erythrocyte membrane abnormalities and/or friability in elderly patients leading to greater susceptibility to ribavirin-induced oxidative damage.52 Another mechanism may be decreased ability of older patients to increase erythropoiesis in the presence of bone marrow suppression from IFN.37 Most transplant recipients are age 50 yr or older.53 With the increasing age of transplant recipients, anemia may be even more prevalent in the future and warrants additional study.54

We also found an important relationship between BMI and anemia (Fig. 2). The reason BMI may have been related to anemia in our study was perhaps because our lighter patients were “overdosed.” Patient weight is associated with sustained virological response, requiring weight-based ribavirin dosing.55, 56 Indeed, the efficacy and probability of achieving a sustained virological response with weight-based ribavirin must be considered along with the risk of hemolytic anemia.

A significant association was found between mycophenolate mofetil use and anemia in univariate analysis (P = 0.05), though it was not statistically significant in the multivariate model. In a series of 19 liver transplant patients, Hebert et al.57 found that 32% developed anemia within 4 weeks after the initiation of mycophenolate mofetil.

Hodo et al.58 demonstrated a potential interaction in a liver transplant recipient who developed pure red cell aplasia when treated with both mycophenolate mofetil and ribavirin. In our study, 46% of patients on both mycophenolate mofetil and ribavirin developed anemia. Mycophenolate mofetil may be associated with anemia by inhibition of inosine 5′ monophosphate dehydrogenase, which results in decreased synthesis of erythroid precursors.58

Our study did not find a significant association between creatinine clearance and anemia (P = 0.647). However, it is interesting to note that 2 of the variables used to calculate creatinine clearance40 relate to the 2 predictors found in our study to be independently associated with anemia: age and BMI. Perhaps, with larger sample size, an association between creatinine clearance and anemia would also have been shown. In clinical practice, ribavirin dosing is adjusted for creatinine clearance because of the known associations between decreased creatinine clearance and increased risk of hemolytic anemia.59–63 Thus, there may have been a bias in our center to use lower doses of ribavirin in patients with renal insufficiency.

Antiviral medications are well known to cause anemia. Our study did not find a significant association between specific antiviral medications and anemia as seen in Table 3. Moreover, this study was not designed to specifically evaluate the effect of different antiviral regimens on anemia, as that would require a control arm of patients receiving transplants who did not undergo antiviral treatment. A limitation of this study is that we did not collect laboratory data such as LDH, haptoglobin, reticulocyte count, peripheral smear, or serum erythropoietin levels that may have better elucidated the etiology of anemia. Anemia in this population is likely multifactorial in nature, characterized by dose-dependent hemolysis from ribavirin, bone marrow suppression from IFN, diminished endogenous serum erythropoietin production due to HCV itself, as well as anemia of chronic disease from chronic inflammation.17–22, 47, 50, 64, 65

The current study has other important limitations. The definition of anemia is somewhat controversial. Some studies have utilized an absolute value, while others a change in hemoglobin values. The World Health Organization defines anemia as a hemoglobin less than 12 gm/dL in women and less than 13 gm/dL in men66. Several hepatitis C treatment studies have followed the World Health Organization threshold.11, 67 However, our threshold is consistent with most hepatitis C studies utilizing a value of 10 gm/dL.49, 55, 68, 69 Using the World Health Organization definition would have resulted in mostpatients being labeled as anemic, since the mean hemoglobin at the start of antiviral therapy was 13 gm/dL in our cohort. Although several studies have used a mean change in hemoglobin as their indicator of anemia,37, 64, 70 we chose a threshold value since it reflects most clinical decisions for intervention with drug dose reduction or cessation, growth factors, and blood transfusions. Moreover, all patients on erythropoietin were included in our definition of anemia even though hemoglobin of 4 of the 30 patients on erythropoietin did not fall below 10 gm/dL. It was assumed that hemoglobin nadir of 10 gm/dL in those 4 patients (mean hemoglobin of 10.3 [± 0.4] gm/dL) was precluded by medication use.

Anemia is a common complication of HCV treatment in liver transplant recipients, and can be managed by dose reduction, use of erythropoietin, and blood transfusions. Our study identified age and BMI as independent predictors of anemia. Recognition can lead to early intervention and prevent ribavirin dose reduction.


  1. Top of page
  2. Abstract
  6. Acknowledgements

We thank Lisette Peralta, Jim Hoffman, and Eshan Taqcvi for data collection, and Gregg Kunder, Kevin King, and Anna Zafar for their excellent patient care.


  1. Top of page
  2. Abstract
  6. Acknowledgements