SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Anemia may increase the likelihood of achieving a sustained virological response (SVR) during pegylated interferon and ribavirin treatment of hepatitis C virus (HCV) infection. To determine whether hemoglobin decline is associated with SVR, we retrospectively evaluated the CHARIOT study of 871 treatment-naïve HCV genotype 1 patients. Anemia (serum hemoglobin <100 g/L) occurred in 137 (16%) patients, of whom only 14 (10%) received erythropoietin. Hemoglobin decline >30g/L from baseline occurred in 76% of patients overall, including 526 patients who did not become anemic. Virological responses were higher in anemic patients compared with those who did not develop anemia (end of treatment, 80% versus 65%, P = 0.003; SVR, 61% versus 50%, P = 0.02); these differences remained significant when patients receiving erythropoietin were excluded from analysis. SVR was also higher in patients with hemoglobin decline >30 g/L compared with patients without a similar decline. In multiple logistic regression analyses with treatment group and baseline characteristics, the odds ratio for SVR was 1.97 (95% confidence interval, 1.08-3.62) for anemia and 2.17 (95% confidence interval, 1.31-3.62) for hemoglobin decline >30 g/L. Patients who first developed a hemoglobin decline >30 g/L during weeks 5-12 and 13-48 were more likely to achieve SVR than those who first developed such changes in weeks 0-4 or who never experienced them. Conclusion: Patients with HCV genotype 1 infection who develop anemia or experience a hemoglobin decline >30 g/L during weeks 5-48 of therapy achieve higher virological responses to pegylated interferon and ribavirin therapy that are unrelated to erythropoietin use. (HEPATOLOGY 2011;)

Anemia frequently develops during antiviral therapy with pegylated interferon (PEG-IFN) and ribavirin for chronic hepatitis C virus (HCV) infection, affecting up to 30% of patients. Low hemoglobin levels may result from ribavirin-induced hemolysis or from interferon-induced bone marrow suppression. Significant anemia may lead to dosage reduction and, in some cases, treatment discontinuation resulting in suboptimal treatment outcomes. Hematopoietic growth factors such as erythropoietin have been used to maintain hemoglobin concentrations during antiviral therapy and have been shown to improve quality of life but not sustained virological response (SVR) rates.1

Retrospective analysis of a recent large study of PEG-IFN and ribavirin in treatment-naïve HCV genotype 1 patients revealed that the magnitude of hemoglobin decline during therapy was associated with SVR.2 However, in that study, approximately 50% of the patients with protocol-defined anemia received erythropoietin, which may have confounded the association between hemoglobin decline and SVR. We therefore explored possible associations between virological response and extent of hemoglobin decline and anemia in patients enrolled in a treatment-naïve HCV genotype 1 study that examined induction versus standard PEG-IFN dosing during the first 12 weeks of combination antiviral therapy and in which there was limited use of hematopoietic growth factors.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

CHARIOT Study Population.

The CHARIOT trial methods and patient population have been described in detail.3 Eligible subjects included treatment-naïve adults aged 18-75 years with chronic HCV genotype 1 infection and compensated liver disease (Child-Pugh score <7). Standard clinical and laboratory exclusion criteria were used, including neutrophil count <1,500 cells/mm3, platelet count <90,000 cells/mm3, and hemoglobin concentration <120 g/L in women or <130 g/L in men. Patients meeting screening eligibility criteria were randomly assigned 1:1 to receive PEG-IFN alfa-2a either in an induction dose or standard dose regimen. The induction regimen consisted of 360 μg of PEG-IFN alfa-2a weekly for the first 12 weeks followed by 180 μg of PEG-IFN alfa-2a weekly for 36 weeks. The standard dose regimen involved 180 μg of PEG-IFN alfa-2a weekly for 48 weeks. Patients received ribavirin concomitantly for 48 weeks with dosing based on body weight (1,000 mg/day if <75 kg; 1,200 mg/day if ≥75 kg). Patients in both arms without an early virological response at week 12 continued therapy to week 24; patients with detectable HCV RNA at week 24 ceased therapy.

Anemia was defined in the study protocol as serum hemoglobin concentration <100 g/L. There was limited access to hematopoietic growth factors at clinical sites during the trial, although their use was permitted if deemed necessary on clinical safety grounds. Ribavirin dose was reduced if hemoglobin was <100 g/L and was withheld if <85 g/L. Dose modification of daily ribavirin dose was performed in decrements of 200 mg. PEG-IFN alfa-2a dose was reduced for neutrophil counts <750 cells/mm3 and platelet counts <50,000 cells/mm3 and was withheld for absolute neutrophil counts <500 cells/mm3 and platelet counts <25,000 cells/mm3. Dose modifications of weekly PEG-IFN alfa-2a were made by decremental adjustments of 180 μg to 135 μg, 90 μg, and 45 μg in patients receiving standard dose and 360 μg to 270 μg, 180 μg, 135 μg, 90 μg, and 45 μg in those receiving induction dose based on the severity of adverse events.

Cumulative exposure to PEG-IFN alfa-2a and ribavirin was determined by calculation of the percentage of planned dose received through week 4, 8, 12, 24, and 48. Reductions from maximum dose occurred through both clinician-directed dose modification and patient nonadherence, with adherence assessed via recording the injections and doses of PEG-IFN alfa-2a and ribavirin at each visit according to the patient's detailed statements and via documentation of drugs dispensed through pharmacy records.

Assessments and Efficacy Endpoints.

Clinical and laboratory safety and efficacy assessments were performed during the treatment period every 4 weeks during the first 24 weeks, then 6 weekly through week 48; and after 4 weeks (week 52), 12 weeks (week 60), and 24 weeks (week 72) of follow-up. Quantitative serum HCV RNA levels were measured at baseline and at weeks 4, 8, 12, 24, 48, and 72. HCV RNA assessments were performed in two central laboratories using the Roche Ampliprep/Cobas TaqMan HCV Test with a detection limit of 15 IU/mL. Treatment response endpoints were defined as undetectable HCV RNA, including the primary endpoint at week 72, and were based on Taqman results below the level of detection (both undetectable and <15 IU/mL).

Liver biopsies were taken within 36 months of treatment and scored by local pathologists. Fibrosis stage was classified according to the Metavir system as F0 (none), F1 (minimal), F2 (moderate), F3 (severe), and F4 (cirrhosis).4 Patients without liver biopsies were considered to have missing fibrosis stage with no inclusion of clinical or noninvasive methods of disease staging.

Statistical Analysis.

The intention-to-treat analysis population was defined as all randomized patients who received at least one dose of study medication. Percentages were calculated for binary parameters. Means were calculated for continuous variables. Multiple logistic regression analysis was performed with SVR as the outcome variable. Explanatory variables included baseline demographics, treatment group, viral and liver disease characteristics, and laboratory parameters, including anemia (based on the protocol definition of hemoglobin <100 g/L) and maximum hemoglobin decline >30 g/L from baseline values. Anemia and maximum hemoglobin decline during treatment were fitted in separate multiple logistic regressions because of their high negative correlation. The locally weighted scatter plot smoothing (LOWESS) method was used to explore the relationship between SVR and changes in serum hemoglobin values during therapy. Due to the exploratory nature of these analyses, no alpha-adjustment was applied to account for multiple significance testing. Data were analyzed with SAS version 8.2 (SAS Institute, Cary, NC).

Study Conduct.

An Australian-based protocol steering committee oversaw the study. The clinical trial was registered with both the National Institutes of Health (NCT00192647) and the Australian Therapeutic Goods Administration (ACTRN12605000488606).

Results

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

A total of 896 patients were enrolled in the study between September 2004 and February 2007; of these, 871 received at least one dose of study drug and constituted the intention-to-treat population. Anemia, defined as serum hemoglobin <100 g/L at any time during treatment, occurred in 137 (16%) patients; of these, 14 (10%) patients received erythropoietin. A maximal hemoglobin decline >30 g/L from baseline occurred in 661 patients (76%) in total, including all but two of the anemic patients, plus 526 patients who did not become anemic. A maximum hemoglobin decline ≤30 g/L occurred in 205 patients. Data were missing from five patients.

Changes in serum hemoglobin concentration before, during, and after combination antiviral therapy are shown in Fig. 1A. In most patients, a decline in hemoglobin >30 g/L occurred during the first 12 weeks of therapy in both treatment arms (Fig. 1B), whereas the development of anemia (hemoglobin <100 g/L) occurred gradually over the course of treatment (Fig. 1C).

thumbnail image

Figure 1. Hemoglobin concentrations before, during, and after therapy are shown in the total study population. (A) The lowest levels occurred during therapy and returned to baseline levels after the end of therapy. (B) Most patients experienced a hemoglobin decline >30 g/L during the first 12 weeks of therapy, with no difference observed between patients treated with the standard or induction regimens. (C) Anemia developed slowly over the course of treatment in both groups.

Download figure to PowerPoint

The baseline demographics of patients who developed anemia compared with those who did not are shown in Table 1. Patients who developed anemia were more likely to be female and significantly older with lower body weight, body mass index, creatinine clearance, hemoglobin levels, white cell counts and platelet counts than patients who did not become anemic. Patients with hemoglobin decline >30 g/L were more likely to be older, female, and with lower body weight and higher baseline hemoglobin than patients with a maximal hemoglobin decline ≤30 g/L (data not shown). The allocated and mean dosages received for PEG-IFN and ribavirin at weeks 12, 24, and 48 of therapy are shown in Table 2. At baseline, more patients who became anemic were allocated a lower dose of ribavirin (1,000 mg versus 1,200 mg) than patients who did not become anemic (61% versus 44%; P = 0.0002). The mean daily ribavirin dosage was significantly lower in patients who developed anemia compared with those who did not become anemic at week 12 (998 ± 143 mg/day versus 1,052 ± 152 mg/day; P = 0.0001) and week 24 (967 ± 169 mg/day versus 1030 ± 210 mg/day; P = 0.0002); there was no significant difference in ribavirin exposure at week 48. The mean weekly PEG-IFN dosage at week 48 was significantly lower in patients who did not become anemic compared with anemic patients for both standard and induction therapy arms; there was no significant difference in PEG-IFN exposure at earlier times. Similar outcomes were observed when PEG-IFN and ribavirin exposure were analyzed as a percentage of planned target dose (data not shown).

Table 1. Baseline Demographics in Patients with and Without Anemia
 Hemoglobin<100 g/L (n = 137)Hemoglobin ≥100 g/L (n = 732)P Value
  1. Abbreviation: AST, aspartate aminotransferase.

Male sex, n (%)70 (51)512 (70)<0.001
Age, years, mean ± SD48 ± 8.842.6 ± 9.2<0.001
Race, n (%)  0.002
 Caucasian106 (77)612 (84) 
 Asian31 (23)85 (12) 
 Other35 (5) 
Weight, kg, mean ± SD71.2 ± 13.979.23 ± 16.6<0.001
Body mass index, kg/m2, mean ± SD25.3 ± 4.626.72 ± 4.70.0013
Mean HCV RNA, log10 IU/mL, mean ± SD6.14 ± 0.766.19 ± 0.740.44
HCV RNA, n (%)  0.58
 <400,00030 (22)129 (18) 
 400,000 to <800,00016 (12)88 (12) 
 >800,00090 (66)503 (69) 
 Missing1 (<1)12 (2) 
Metavir score, n (%)  0.11
 F06 (4)47 (6) 
 F128 (20)181 (25) 
 F238 (28)196 (27) 
 F317 (12)80 (11) 
 F410 (7)20 (3) 
 Missing38 (28)208 (28) 
Hemoglobin, g/L, mean ± SD148 ± 11157 ± 11<0.001
White blood cells, ×109/L, mean ± SD6.79 ± 1.797.47 ± 2.090.001
Platelets, ×109/L, mean ± SD224.9 ± 64.5242.8 ± 66.50.004
Serum albumin g/L, mean ± SD42.9 ±3.243.4 ± 3.40.11
Bilirubin, log10 μmol/L, mean ± SD0.91 ± 0.190.92 ± 0.210.69
Serum AST, log10 U/L, mean ± SD1.49 ± 0.251.48 ± 0.250.67
Serum creatinine, μmol/L, mean ± SD76.46 ± 13.8676.87 ± 13.420.76
Creatinine clearance, mL/minute, mean ± SD101.37 ± 27.67119.96 ± 30.09<0.0001
Table 2. Allocated and On-Treatment PEG-IFN and Ribavirin Dosing
Study MedicationMean DosageHemoglobin <100 g/LHemoglobin ≥100 g/L
InductionStandardInductionStandard
  • *

    P = 0.045 versus anemic patients (induction PEG-IFN group).

  • P = 0.0001 versus anemic patients (standard PEG-IFN group).

  • P = 0.0002 versus nonanemic patients (ribavirin).

  • §

    P = 0.0001 versus nonanemic patients (ribavirin).

  • P = 0.0002 versus nonanemic patients (ribavirin).

PEG-IFN (μg/week)Allocated dose at baseline, % of patients53%47%49%51%
Up to week 12, mean (SD)342 (41)177 (9)335 (58)177 (21)
Up to week 24, mean (SD)257 (31)173 (16)239 (48)169 (30)
 Up to week 48, mean (SD)202 (41)164 (29)190 (54)*146 (48)
All Patients Treated (Combined PEG-IFN Groups)
Ribavirin (mg/day)Allocated dose at baseline (% of patients)1,000 mg/day (61%) or 1,200 mg/day (39%)1,000 mg/day (44%) or 1,200 mg/day (56%)
Up to week 12, mean (SD)998 (143)§1,052 (152)
Up to week 24, mean (SD)967 (169)1,030 (210)
Up to week 48, mean (SD)895 (232)918 (301)

Virological responses at the end of treatment (ETR) and at the end of follow-up (SVR) were significantly different between patients with hemoglobin <100 g/L at any time during treatment compared with those with hemoglobin ≥100 g/L (ETR, 80% versus 65%, respectively, P = 0.003; SVR, 61% versus 50%, respectively, P = 0.02). Relapse rates were similar, however (Fig. 2A). Similarly, ETR and SVR rates were significantly higher in patients with hemoglobin decline >30 g/L compared with those with hemoglobin decline ≤30 g/L. An ETR occurred in 72% of patients with a hemoglobin decline >30 g/L compared with 52% of those without a similar change in hemoglobin (P < 0.001). Similarly, a SVR occurred in 54% with a hemoglobin decline >30 g/L compared with 46% with a hemoglobin decline ≤30 g/L (P = 0.049). Relapse rates were similar (Fig. 2B). In separate multiple logistic regression analyses, both hemoglobin <100 g/L (protocol defined anemia) and maximum hemoglobin decline >30 g/L during treatment were significantly associated with SVR rate. The odds ratio estimate for SVR for hemoglobin <100 g/L was 1.97 (95% confidence interval, 1.08-3.62; P = 0.028). The odds ratio estimate for hemoglobin decline >30 g/L was 2.17 (95% confidence interval, 1.31-3.62; P = 0.003). We also examined virological responses earlier in the course of therapy (Table 3). A rapid virological response occurred more often in patients who became anemic compared with those who did not (42% versus 29%; P = 0.002). However, there was no difference between these two groups in regard to partial or complete early virological response, and there was no difference in any early virological response between patients who had a decline in hemoglobin > 30 g/L and those who did not. In order to determine whether erythropoietin use may have influenced virological outcomes, we performed separate analyses (data not shown) excluding the 14 patients who received erythropoietin and examined baseline demographics in patients with and without anemia, by time to develop hemoglobin decline >30g/L and by virological responses. No change in outcomes for any of the demographic, hematological, or virological responses was seen when patients who received erythropoietin were excluded.

Table 3. Early Virological Responses
Virological ResponseHemoglobin <100 g/LHemoglobin ≥100 g/LHemoglobin decline >30 g/LHemoglobin decline ≤30 g/L
  • Abbreviations: cEVR, complete early virological response; pEVR, partial early virological response; RVR, rapid virological response.

  • *

    P = 0.002 versus RVR in patients with hemoglobin <100 g/L.

RVR42%29%*37%41%
cEVR39%41%48%50%
pEVR15%9%13%9%
thumbnail image

Figure 2. End of treatment and SVRs were significantly greater in (A) patients with anemia compared with those who did not develop anemia (ETR, 80% versus 65%, P = 0.003; SVR, 61% versus 50%, P = 0.02, respectively) and in (B) patients with hemoglobin decline >30 g/L compared with those with hemoglobin decline ≤30 g/L (ETR, 72% versus 52%, P < 0.001; SVR, 54% versus 46%, P = 0.049, respectively). Relapse rates were similar in both groups.

Download figure to PowerPoint

A further logistic regression analysis was conducted with the covariates hemoglobin <100 g/L (anemia), fibrosis score, treatment group, and the three two-way interactions. None of the two-way interactions were statistically significant, whereas the main effects of anemia (P = 0.0023) and fibrosis score (P < 0.001) were highly significant. Similarly, a logistic regression analysis was conducted with the covariates maximum hemoglobin decline >30 g/L, fibrosis score, treatment group, and the three two-way interactions. None of the two-way interactions were statistically significant. The main effects of maximum hemoglobin decline (P = 0.0062) and fibrosis score (P < 0.001) were highly significant. Given this outcome, we then used the LOWESS method to evaluate the local probabilities of SVR against ranges of values of lowest postbaseline hemoglobin (anemia) and ranges of values of decline in hemoglobin concentration for patients with and without cirrhosis (results are shown only for patients receiving the standard treatment regimen). When the estimated local probabilities of SVR were plotted against the values of the lowest postbaseline hemoglobin, the overall trend showed that the probability of SVR was higher in patients with a postbaseline hemoglobin ≤120 g/L and lower in those whose hemoglobin level remained >120 g/L (Fig. 3A). The trend toward increasing probability of SVR with a lower nadir in hemoglobin was apparent both in patients without cirrhosis (Fig. 3B) and in those with cirrhosis (Fig. 3C). The overall probability of SVR appeared to increase with greater maximum decreases in hemoglobin concentration up to approximately 30 g/L (Fig. 4A), after which the probability of SVR was relatively stable but declined steadily when hemoglobin levels decreased by more than 60 g/L. The trend was generally similar for patients with and without cirrhosis (Fig. 4B,C).

thumbnail image

Figure 3. LOWESS plot of SVR probability by lowest hemoglobin value is shown for patients receiving the standard treatment regimen only (intention-to-treat population). (A) For all patients, the probability of SVR was higher with a postbaseline hemoglobin ≤120 g/L and lower if hemoglobin levels remained >120 g/L. Similar responses occurred in patients (B) without cirrhosis and (C) with cirrhosis, although SVR rates were lower overall in patients with cirrhosis.

Download figure to PowerPoint

thumbnail image

Figure 4. LOWESS plot of SVR probability by hemoglobin decline is shown for patients receiving the standard treatment regimen only (intention-to-treat population). (A) For all patients, the probability of SVR increased with greater maximum decreases in hemoglobin concentration to approximately 30 g/L and was then stable, but it declined when hemoglobin levels decreased by >60 g/L. Similar responses occurred in patients (B) without cirrhosis and (C) with cirrhosis, although SVR rates were lower overall in patients with cirrhosis.

Download figure to PowerPoint

We then explored the possibility that virological response rates might vary by time of first occurrence of a hemoglobin decline >30 g/L. Most patients (n = 308) experienced the first occurrence of a hemoglobin decline >30 g/L during weeks 5-12, with a similar number developing this change during weeks 0-4 (n = 172) and weeks 13-48 (n = 181). Baseline demographic characteristics among these groups are compared in Table 4. Patients without a significant hemoglobin decline throughout treatment were younger with a higher body weight, less hepatic fibrosis, and lower baseline hemoglobin concentration and higher creatinine clearance than patients with hemoglobin declines >30 g/L. Patients with a rapid hemoglobin decline during weeks 0-4 were older, with higher baseline hemoglobin concentrations and lower platelet counts, and were more likely to have advanced hepatic fibrosis (F2-F4). Fewer patients with a rapid hemoglobin decline during weeks 0-4 or who did not experience a decline >30 g/L achieved SVR compared with higher SVR rates among patients with a >30 g/L decrease in hemoglobin first occurring in weeks 5-12 and 13-48 of therapy (P = 0.02) (Fig. 5).

Table 4. Baseline Patient Demographics by Time to Develop Hemoglobin Decline >30 g/L
 0-4 weeks (n = 172)5-12 weeks (n = 308)13-48 weeks (n = 181)No Hemoglobin Decline (n = 205)P Value
  1. Abbreviation: AST, aspartate aminotransferase.

Male sex, n (%)111 (65)205 (67)113 (62)150 (73)0.13
Mean age, years46.643.942.940.4<0.0001
Race, n (%)    0.62
 Caucasian148 (86)251 (81)149 (82)167 (81) 
 Asian19 (11)43 (14)22 (12)32 (16) 
 Other5 (3)14 (5)10 (6)6 (3) 
Mean weight, kg77.676.277.681.30.007
Mean body mass index, kg/m226.626.026.527.20.06
Mean HCV RNA, log10 IU/mL6.216.206.126.180.61
HCV RNA, n (%)    0.80
 <400,00032 (19)51 (17)40 (22)34 (17) 
 400,000 to <800,00019 (11)34 (11)21 (12)30 (15) 
 >800,000119 (69)217 (70)117 (65)139 (68) 
 Missing2 (1)6 (2)6 (2)2 (<1) 
Metavir score n (%)    0.007
 F06 (3)17 (6)12 (7)18 (9) 
 F137 (22)70 (23)50 (28)52 (25) 
 F254 (31)89 (29)35 (19)55 (27) 
 F328 (16)33 (11)20 (11)15 (7) 
 F411 (6)10 (3)6 (3)2 (<1) 
 Missing36 (21)89 (29)58 (32)63 (31) 
Mean hemoglobin, g/L160156155151<0.001
Mean white blood cells, ×109/L7.387.187.327.670.07
Mean platelets, ×109/L2262362442540.0005
Mean serum albumin, g/L43.443.543.343.00.40
Mean total bilirubin, log10 μmol/L0.950.930.910.880.01
Mean serum AST, log10 U/L1.531.491.461.460.033
Mean serum creatinine, μmol/L77.976.676.176.80.66
Creatinine clearance, mL/minute, mean ± SD111.7 ± 30113.6 ± 28117.4 ± 29126.6 ± 32<0.0001
thumbnail image

Figure 5. SVR occurred in a greater proportion of patients who first experienced a hemoglobin decline >30 g/L during weeks 5-12 and 13-48 compared with those who did so in the first 4 weeks of therapy or who never experienced a hemoglobin decline ≤30 g/L during therapy (no hb [DOWNWARDS ARROW]).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

In a large population of HCV genotype 1 patients treated with PEG-IFN and weight-based ribavirin, we found that the odds for achieving SVR for patients whose lowest hemoglobin was <100 g/L or whose maximum hemoglobin decline was >30 g/L were about twice the odds of those whose maximum hemoglobin decline was ≤30 g/L or whose lowest hemoglobin during treatment was ≥100 g/L. Clinically relevant limits to these outcomes occurred in patients whose hemoglobin concentration remained >120 g/L, who developed hemoglobin declines >60 g/L, or who developed a decline >30 g/L during the initial 4 weeks of therapy, because they did not experience improved virological responses.

A similar relationship between hemoglobin decline and improved treatment response was noted from post hoc analysis of the IDEAL study of over 3,000 HCV genotype 1 patients treated with either PEG-IFN α2a or α2b plus weight-based ribavirin.2 In that study, 75% of patients experienced a decline in serum hemoglobin >30 g/L, of whom 37% developed anemia (similarly defined as <100 g/L). The probability of SVR was related to increasing decline in hemoglobin from baseline so that patients with >30 g/L hemoglobin decline achieved an SVR rate of 44% compared with 30% in patients with ≤30 g/L hemoglobin decline. Anemia occurred in 865 (29%) patients, and erythropoietin was given to 52% of this group. The use of erythropoietin was associated with significantly higher SVR rates among patients with the early onset of anemia (from 0-8 weeks), but not in those with later onset anemia. Patients developing anemia during the first 8 weeks of therapy who were treated with erythropoietin also experienced a lower treatment discontinuation rate than those who developed anemia at a later time.

The major distinction between the IDEAL and the CHARIOT studies was the level of hematopoietic growth factor use with only 14 patients (10% of those developing anemia) in our study receiving erythropoietin. We found no difference in virological outcomes when these 14 patients were excluded from analysis. Given that both studies demonstrated similar relationships between haemoglobin decline and SVR, it seems unlikely that erythropoietin use per se is the major factor contributing to the increased SVR rates seen in patients with significant therapy-induced hemoglobin decline. However, greater utilization of erythropoietin, particularly among those patients with hemoglobin declines >30 g/L during the initial 4 weeks of therapy, may have improved SVR rates in the CHARIOT study. Specific studies are required to examine the role of erythropoietin in this group of patients with a rapid hemoglobin decline.

We identified several patient characteristics that were associated with on-treatment development of anemia. Anemia was more likely in women with lower body weight, older age, lower creatinine clearance, and lower baseline hemoglobin concentrations, white cell counts, and platelet counts (Table 1). Those who developed a hemoglobin decline >30 g/L were more likely to be female and older with lower body weight, but with a higher baseline hemoglobin level than those who never developed a similar fall in hemoglobin concentration (data not shown). When analyzed by time to first occurrence of a hemoglobin decline >30 g/L, we observed similar changes, because older patients with a lower body weight, lower creatinine clearance, and higher baseline hemoglobin level were more likely to develop a hemoglobin decline >30 g/L (Table 4). These findings are consistent with previous studies that have identified similar clinical risk factors for developing ribavirin-induced anemia.5-7 A predictive pharmacokinetic model that incorporates some of these factors has been reported,8 but the use of patient characteristics to predict ribavirin-associated hematological changes has not gained widespread clinical use.

The precise mechanisms underlying the higher SVR rates in patients with a decline in hemoglobin remain unclear. Given the well-known hemolytic effects of ribavirin, it would be reasonable to assume that this observation is related to an individual pharmacokinetic response to that drug. Pharmacokinetic studies have shown that ribavirin reaches steady state plasma concentrations after 3-12 weeks of continued dosing and that ribavirin clearance is determined principally by body weight and renal function.9 A study of 380 Caucasian male HCV patients of mixed genotypes with plasma sampled at steady state (weeks 8-48 of therapy) reported that lean body weight was the most important covariate affecting ribavirin clearance, which increased linearly with body weight.10 As discussed above, we found that patients who became anemic or developed a hemoglobin decline >30 g/L had significantly lower body weights than those who did not, suggesting that decreased ribavirin clearance relative to heavier patients might account for the difference in hemoglobin concentrations and, by extension, greater exposure to ribavirin. Plasma ribavirin determinations may help to resolve this. Variability in ribavirin dosage due to dose reduction or treatment adherence did not appear to be a confounding factor, because we identified favorable virological responses in anemic patients despite significantly lower mean ribavirin exposure during the first 24 weeks of therapy (Table 2). Individual pharmacokinetic responses to ribavirin may be related to recently described variants in the inosine triphosphatase (ITPA) gene that result in ITP deficiency and therefore protection against ribavirin-induced anemia.11 Precisely how ITP deficiency interacts with the mechanisms leading to ribavirin-induced anemia remains unclear. Interestingly, no association of ITPA variants with rapid or sustained virological response to PEG-IFN and ribavirin was identified by Fellay and colleagues,12 although a trend for increased SVR was observed when patients were stratified by interleukin-28b genotype, which is a strong predictor of treatment outcome.

Although we found significant relationships with both anemia and hemoglobin decline >30 g/L during therapy and higher SVR, the proportion of patients who developed a hemoglobin decline >30 g/L was considerably greater, suggesting that the absolute decline in hemoglobin may be more clinically relevant. In this regard, the identification of a subset of patients with rapid hemoglobin decline who do not benefit in terms of improved SVR provides useful information for prediction of outcome and potential opportunities for interventional strategies such as erythropoietin. Furthermore, the relationship between hemoglobin decline and treatment response remained highly significant following adjustment for fibrosis stage, with both factors being strongly associated with SVR in a multivariate model. Despite this, patients with cirrhosis had generally lower SVR rates than patients without cirrhosis as reported in the CHARIOT study, an outcome that did not appear to relate to lower ribavirin adherence.13

In conclusion, we have shown that the odds of achieving an SVR for patients with HCV genotype 1 infection who develop anemia or who experience a decline in hemoglobin >30 g/L, even if they do not become anemic, are approximately twice that of those who do not develop similar hematological changes. This relationship was identified with or without the inclusion of 14 patients who received erythropoietin. However, patients with hemoglobin concentrations >120 g/L, those with a >30 g/L decline within the initial 4 weeks of therapy, and those with decline >60 g/L from baseline during therapy do not achieve similar virological benefits.

References

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References