A national French survey on the use of growth factors as adjuvant treatment of chronic hepatitis C


  • This study was conducted on behalf of the: Association Nationale des Hépato-Gastroentérologues des Hôpitaux généraux (National Association of General Hospital Hepatogastroenterologists) ANGH and the Association Française de l'Etude du Foie (French Association for the study of the Liver) AFEF.

  • Potential conflict of interest: Nothing to report.


We conducted a national retrospective survey on hospital practitioners to evaluate the magnitude of erythropoietin (EPO) or granulocyte colony-stimulating factor (G-CSF) prescriptions in patients treated for chronic hepatitis C. Four hundred seventy-one questionnaires were sent, and 274 practitioners (58.2%) responded. Forty-six percent of practitioners used EPO, and 31% used G-CSF. The total number of HCV-infected patients receiving antiviral therapy per year was estimated at 6,630 patients, of whom 8.8% and 4% received EPO and G-CSF, respectively. EPO-β was the main EPO molecule prescribed at a median dose of 30,000 IU/wk (range: 2,000-80,000). The indications for prescribing EPO varied greatly, including “fragile patients” (34%), “low” Hb level (8-11 g/dL) (19%), “rapid decline” in Hb level (2-5 g/dL during the first month of therapy) (12%), and symptomatic anemic patients (7%). G-CSF was mainly prescribed for a “low” level of neutrophils ranging from 400 to 750 neutrophils/mm3. In multivariate analysis, independent predictors of EPO and G-CSF prescription were age of practitioner less than 45 years (EPO: OR = 1.96, P = 0.03; G-CSF: OR = 2.27, P = 0.004), practice in university hospital (EPO: OR = 5.89, P < 0.0001; G-CSF: OR = 2.39, P = 0.003), and the high number of CHC treated/year (EPO: OR = 6.18, P < 0.0001; G-CSF: OR = 2.58, P = 0.002). Conclusion: Our survey reveals an important rate of EPO and G-CSF prescriptions but with considerable disparity in the schedule of injections, the molecules used, and above all the indications. The suitable role of EPO and G-CSF as complements to HCV therapy urgently needs to be clarified. (HEPATOLOGY 2007;45:377–383.)

Hepatitis C virus infection is a major cause of morbidity and mortality, affecting 170 million people worldwide1, 2 and a total of 368,000 people in France in 2004, of whom 221,000 were RNA positive.3 Persistent infection over many years is associated with an increased risk of complications including cirrhosis, hepatic decompensation, and hepatocellular carcinoma. The current and more effective treatment for HCV infection consists of a combination of pegylated interferon with ribavirin, providing a sustained virological response rate of 40% to 45% in patients with HCV genotype 1 and of 80% in patients with HCV genotypes 2 or 3 in randomized controlled trials.4, 5 However, the sustained viral response rate decreases dramatically when adherence to treatment is not optimal.6 Patients may not comply with the treatment regularly because of the multiple side effects encountered, especially a dose-limiting hemolytic anemia with ribavirin that often results in early withdrawal. Neutropenia is another concern, but to a lesser extent. Adverse hematological events may require treatment disruption or result in a 10% to 20% reduction in virological response.4, 5, 7 Because the primary goal of HCV therapy is to achieve a sustained viral response and ultimately to avoid long-term complications of HCV infection, it seems important to maximize therapeutic compliance. Evolving experience and recent clinical trials indicate that the use of hematopoietic growth factors (HGF), namely the granulocyte colony–stimulating factor (G-CSF) and erythropoietin (EPO), in addition to the treatment of HCV infection, may allow maintenance of the optimal dose and duration of ribavirin or pegylated interferon (PEG-IFN) while improving the quality of life.8–10 Currently, no official guidelines exist for treating antiviral therapy–associated anemia and neutropenia, and many issues remain unresolved. Nevertheless, different commercialized HGF seem to be prescribed more and more in France, although there no data are available about the real burden of these prescriptions. Our study aims to evaluate the magnitude of HGF prescriptions in France and to specify the schedules of these prescriptions.


ANGH, Association Nationale des Hépato-Gastroentérologues des Hôpitaux généraux de France; EPO, erythropoietin; G-CSF, granulocyte colony-stimulating factor; HGF, hematopoietic growth factors; PEG-IFN, pegylated interferon; QoL, quality of life.

Materials and Methods

Study Design.

This was a retrospective study conducted by practitioners belonging to general or university hospitals. In November 2005, a first anonymous questionnaire was sent to 296 department heads of general hospitals belonging to the “Association Nationale des Hépato-Gastroentérologues des Hôpitaux généraux de France” (ANGH) covering 95 out of the 100 metropolitan and overseas administrative divisions (“departments”) that make up France. The addresses of the ANGH's members were taken from the 2005 yearbook. Four months later, the same questionnaire was mailed again to the department heads of those general hospitals who had not answered. This survey was then extended to the university hospitals, and 175 questionnaires were sent to hepatogastroenterologists practicing in these settings. There was no follow-up letter for university hospitals. All of the assistants of each department head were encouraged to fill out the questionnaire.

Survey and Data Record.

The single-page questionnaire was mailed with a cover letter describing the purpose of the survey and its confidential nature. The following data were recorded: age, sex, department area and/or town, specialized area of medicine (hepato-gastroenterologist, internist, or infectious diseases specialist), number of HCV-positive patients treated in the previous 12 months, and number of patients treated with EPO or with G-CSF during that period. Regarding HGF prescriptions, we asked practitioners the following questions: (1) what molecule, dose, and number of injections per week were prescribed for the 3 available erythropoietic agents [epoetin alfa (Eprex), epoetin beta (Neorecormon), and darbepoetin alfa (Aranesp)] and the 3 available G-CSF [filgrastim (Neupogen), lenograstim (Granocyte), and pegfilgrastim (Neulasta)]? (2) If you did not use EPO (or G-CSF), do youprefer to decrease the ribavirin dose to 600 mg/day when hemoglobin (Hb) level is less than 10 g/dL and to stop ribavirin below an Hb level less than 8.5 g/dL according to manufacturer's recommendations? (3) If you did not use EPO (or G-CSF), do you prefer to adapt antiviral therapy to the patients' clinical tolerance? Finally, we also inquired about the adverse effects observed with these molecules.

Statistical Analysis.

Responses were collected by 2 investigators (T.T. for general hospitals and F.L.B. for university hospitals), and data were analyzed by a third investigator (V.D.M.) with the NCSS package for Windows (Hintze J, 2004, Kaysiville, UT). Quantitative variables were expressed as means ± SE or as medians (95% confidence intervals, CI) in case of abnormal distribution. The relationship between physician characteristics and HGF prescription was assessed using a chi-square or Student t test for univariate analyses. Multivariate analyses were performed using logistic regression analysis on the basis of univariate analyses results and absence of colinearity. A P value less than 0.05 was considered statistically significant.


Study Population.

Of 471 questionnaires sent, we collected 274 responses (58.2%), including 205 of 296 responses (69.3%) from ANGH members and 69 of 175 responses (39.4%) from university hospitals. Nearly 75% of all responses came from general hospitals. ANGH practitioners from 84 departments (88.4%) responded. Only ANGH practitioners from 13 hospitals representing 10 metropolitan and one overseas ‘departments’ did not respond despite a follow-up letter.

The total number of HCV-infected patients receiving antiviral therapy per year as declared by the responding practitioners was estimated at 6,630 patients. Table 1 shows the characteristics of responding practitioners from general hospitals and university hospitals. In general hospitals, practitioners were older (48 ± 8 versus 45 ± 8 years, P = 0.013), predominantly male (83% versus 57%, P < 0.0001) and declared a lower number of HCV-infected patients treated per year. Considering the whole population of practitioners, 90% and 74% said that they had treated fewer than 50 and 30 cases of CHC respectively over the previous 12 months.

Table 1. Characteristics of Practitioners
ParameterOverallUniversity HospitalsGeneral HospitalsP Value
  1. P value indicates difference between university and general hospitals.

Age (years)47 ± 845 ± 848 ± 80.013
Sex (% males)76%56.7%83.1%<0.0001
Hepatogastroenterologists (%)96%98.6%95.1%NS
Number of cases of hepatitis C treated/year
 >30 cases/year (%)25.9%56.5%15.6%<10−6
 >50 cases/year (%)9.8%26.1%4.4%<10−6

EPO Prescription.

Overall, 136 practitioners (49.6%) used HGF: 52 only EPO, 74 both EPO and G-CSF, and 10 only G-CSF. There were 126 (46%) who prescribed EPO (EPO-α, 20%; darbepoetin-α, 30%; EPO-β, 38%; and not determined in 12%). The number of EPO-treated patients was 581 (8.8% of the HCV-infected population receiving antiviral therapy). It ranged from 1 to 30 (median, 3 patients) per practitioner. Only 21.4% of practitioners prescribed EPO for more than 5 patients; most (78%) work in university hospitals. To compare the dose administered by each practitioner, we assumed that 200 IU EPO was equivalent to 1 μg darbepoetin-α, and we used the mean value between the lowest and the highest declared doses. The median dose was 30,000 IU/week (range, 2,000-80,000 IU/wk) with 1 to 3 injections/week. The reported indications for prescribing EPO varied greatly among practitioners. They could be categorized according to patient status as follows: “fragile” patients, patients with a rapid decline of Hb level, patients with a “low” Hb level, and patients with symptomatic anemia. The “fragile patients” represented 34% of the main indications of EPO prescription. It concerned HIV-HCV-coinfected patients, hemodialyzed patients, or patients with end-stage renal disease, liver transplant recipients, and patients with cirrhosis. The patients with so-called “low” Hb level represented the second indication of EPO prescription (19%), but the threshold of “low” Hb level extended from 8 to 11 g/dL. A rapid decline in Hb level concerned 12% of EPO prescription. This decline in Hb level was widely scattered from 2 to 5 points during the first month of therapy. Symptomatic anemic patients (7%) were the last indication. In univariate analysis (Table 2), practitioner characteristics associated with EPO prescription were as follows: practice in university hospital (84% versus 33%, P < 0.0001), age younger than 45 years (59% versus 41%, P = 0.005), female gender (61% versus 42.5 %, P = 0.009), and more than 30 chronic hepatitis C treated/year (82% versus 33.5%, P < 0.0001). In multivariate analysis (Table 2), 3 independent predictors of EPO prescription were assessed: age younger than 45 years (OR = 1.93, P = 0.03), practice in university hospital (OR = 5.89, P < 0.0001), and high number of CHC treated/year (OR = 6.18, P < 0.0001).

Table 2. Multivariate Analysis of Factors Associated with EPO Prescription
ParameterUnivariate AnalysesLogistic Regression Analysis
% EPO PrescriptionPOR95% CI ORP Value
University hospitals (vs. general hospitals)84.1 vs. 33.2<0.00015.892.64-13.12<0.0001
Hepatogastroenterologists (vs. others)46.2 vs. 36.4NS0.830.20-3.35NS
Age < 45 years (vs. ≥45 years)58.8 vs. 410.0051.931.05-3.570.035
Female gender of practitioner (vs. male gender)61.3 vs. 42.50.0091.070.51-2.27NS
≥30 cases (vs. <30 cases) of hepatitis C treated/year81.7 vs. 33.5<0.00016.182.79-13.67<0.0001
Use of G-CSF (vs. non use)88.1 vs. 27<10−5

G-CSF Prescription.

G-CSF was prescribed by 84 (31%) practitioners, 74 of whom used both EPO and G-CSF. The number of G-CSF–treated patients was 268 (4% of the HCV-infected population receiving antiviral therapy). One to 20 patients (median, 2 patients) were treated by each of these practitioners. Only 10.8% of practitioners prescribed G-CSF for more than 5 patients; most (67%) work in university hospitals. When only one molecule of G-CSF was used (84.5%), it was filgrastim in 43 cases (61%), lenograstim in 21 cases (30%), and pegfilgrastim in seven cases (9%). The G-CSF molecule was not mentioned in 7% of additional cases. Several molecules of G-CSF were used in 8.5%. The schedule of lenograstim administration was either 1 or 2 injections per week. Seventy-four percent of practitioners prescribed high doses of lenograstim (Granocyte 34), whereas the others used Granocyte 13. Concerning pegfilgrastim, the prescription schedule was undetermined in three cases, and the medication was administered at 6 mg/2 weeks in 2 cases and at 6 mg/3 weeks in the remaining two cases. Filgrastim was the most prescribed molecule (n = 43) at a mean dose of 43 millions ± 20 (range, 15-90) IU/week. Most practitioners prescribed filgrastim once weekly (73%), and a minority (9%) used a thrice-weekly schedule. The reported indication for prescribing G-CSF was mainly a “low” level of neutrophils (65%), but the threshold of prescription ranged from 400 to 750 neutrophils/mm3. The other indications were related to patient condition: patients with cirrhosis, HIV-HCV-coinfected patients, hemodialyzed patients, and liver transplant recipients, or aimed to preserve a good anti-viral response, especially in patients with a high fibrosis score (F2 or greater according to the Metavir scoring system). In univariate analysis (Table 3), practitioner characteristics associated with G-CSF prescription were as follows: practice in university hospital (63% versus 20%, P < 0.0001), age younger than 45 years (41% versus 27%, P = 0.018), female sex (43.5 % versus 28 %, P = 0.023), high number of chronic hepatitis C treated/year (>30/year) (60% versus 21%, P < 0.0001). The use of EPO was also more frequent for practitioners using G-CSF (59% versus 7%, P < 0.0001). In multivariate analysis (Table 3), the 3 independent predictors of G-CSF prescription were: age younger than 45 years (OR = 2.27, P = 0.004), practice in university hospital (OR = 2.40, P = 0.003), and high number of chronic hepatitis C treated per year (OR = 2.59, P = 0.002).

Table 3. Multivariate Analysis of Factors Associated with G-CSF Prescription
ParameterUnivariate AnalysesLogistic Regression Analysis
% GCSF prescriptionPOR95% CI ORP Value
University hospitals (vs. general hospitals)63.2 vs. 20<0.00012.401.33-4.340.004
Hepatogastroenterologists (vs. others)31 vs. 27.3NS0.660.15-2.84NS
Age < 45 yrs (vs. ≥45 yrs)40.7 vs. 26.60.0182.271.29-3.950.004
Female gender of practitioner (vs. male gender)43.5 vs. 28.10.0231.320.69-2.49NS
≥30 cases (vs. <30 cases) of hepatitis C treated/year60 vs. 20.7<0.00012.591.40-4.780.002
Use of EPO (vs. non use)59.2 vs. 6.8<0.0001

Reported Side Effects.

Side effects of EPO or G-CSF were reported by only 13% of practitioners. For EPO, 13 adverse events were reported: “allergy” with pruritus (n = 5), headache (n = 2), arterial hypertension (n = 2), pure red-cell aplasia (n = 2), and myalgia (n = 2). For G-CSF, only eight side effects were observed: arthralgia (n = 2) and bone pain (n = 6).


The main result provided by the current study was that EPO and G-CSF were widely used by 46% and 31% of practitioners, respectively, and for 8.8% and 4% of HCV-infected patients, respectively. Encouraging patient compliance to antiviral therapy and limiting dose disruptions or decreases provoked by side effects may be useful in achieving viral clearance. To maximize sustained viral response rate in HCV-positive patients receiving treatment, it appears critical to maintain ribavirin dose, especially during the first 12 weeks,11 and adequate doses and duration of PEG-IFN.6 Toward this, HGF may be a promising approach to counteract adverse effects of both ribavirin and interferon, including ribavirin-induced hemolysis and PEG-IFN–related neutropenia, which have resulted in the need for dose reduction or discontinuation of therapy in up to 30% of patients.12 This “optimized therapy” could be of particular interest in “difficult-to-treat” patients, including HIV- and HCV-coinfected patients, patients with end-stage renal disease, liver transplant recipients, and patients with cirrhosis.13 Despite growing evidence suggesting that EPO and G-CSF may preserve HCV-combination therapy, there is no license to date for their prescription in the setting of chronic hepatitis C. Indeed, health authorities are still waiting for randomized controlled trials showing that HGF actually increases the rates of sustained viral response. Moreover, numerous issues such as optimal dosage, time of initiation, duration of therapy, and efficacy of this strategy are still unresolved.

We present herein the first survey providing a reliable picture assessing the clinical use of HGF as an adjuvant therapy of chronic hepatitis C in a representative panel of French practitioners treating more than 6,500 HCV-infected patients during the past year. Indeed, the questionnaire was anonymous, and the responding practitioners were informed of the scientific nature of this study conducted on behalf of 2 scientific societies (i.e., AFEF and ANGH) totally independent of the firms commercializing EPO and G-CSF. The 60% response rate closely correlates with the percentages (49%-94%) noted in other French epidemiological surveys on viral hepatitis.14–17

The profile of a practitioner prescribing EPO or GCSF was typically a young female doctor from a university hospital in charge of a large number of HCV-infected patients. Despite the wide-scale use of these drugs and the reproducible profile of practitioners prescribing them, the responses reflected considerable disparity in terms of the circumstances of prescription, the molecules used, and the schedules followed. For instance, the administered dose of EPO varied from 2,000 to 80,000 IU/week, with one to three injections/week similarly in general or university hospitals. Surprisingly, most practitioners used EPO-β as the main molecule, whereas trials in chronic hepatitis C have preferred EPO-α.8, 9, 18–23 Few studies used EPO-β21, 24 or darbepoetin-α.25, 26 Regarding EPO indications, large variations of the Hb threshold (8-11 g/dL) or the magnitude of Hb decline (1-5 points) were observed during the first month of therapy. These results clearly show the lack of harmonization for EPO prescription in most trials, especially those related to treating recurrent hepatitis C after liver transplantation where EPO was started at a Hb level between 8 and 10 g/dL.22, 23, 26, 27

As expected, multivariate analysis demonstrated that the strongest independent predictors of EPO prescription were associated with the experience of “difficult-to-treat” HCV-positive patients (i.e., practice in university hospital and high number of HCV-infected patients treated per year). The third independent predictor was a practitioner age of younger than 45 years. Such “young” practitioners were probably more receptive to innovative treatments. Adverse events related to EPO were rare. However, pure red-cell aplasia was reported for 2 patients treated with epoetin-β. This uncommon disorder, associated with the presence of anti-EPO antibodies, has no well-identified causative factors.30 It has been described in hemodialyzed patients receiving subcutaneous injections of epoetin-α, and it has been reported in only one HCV-infected patient treated with peginterferon α-2a, ribavirin, and epoetin-α.31 Moreover, it was remarkable to discover these cases 3 years after the reported peak of incidence in 2001-2002, when improvements in storage, handling, EPO administration, and the monitoring of treated patients were supposed to prevent such disorders.32 Although there were not enough data to check the diagnosis of pure red-cell aplasia, we think that our two reported cases were valid because two cases of pure red-cell aplasia in French HCV-infected patients occurred within the same period as reported by one of the manufacturers (Roche, personal communication). One of the 2 reported cases by the Roche laboratory was doubtful.

Although a great disparity was reported in the doses administered and prescription indications of G-CSF, it was prescribed less frequently than EPO and mainly involved filgrastim, in concordance with studies employing G-CSF as an adjuvant therapy in chronic hepatitis C.21, 23, 26–29 The prescription of G-CSF covered a smaller proportion of practitioners (31% versus 46%) and only 4% of HCV-infected patients undergoing antiviral therapy. Two reasons may explain this finding. First, ribavirin-induced anemia is more common than IFN-induced neutropenia. In patients receiving PEG-IFN and ribavirin, an Hb drop of 3 g/dL or more occurred in 49% of cases whereas 34% of patients became anemic according to the WHO definition of anemia (Hb < 11 g/dl).33 In the pivotal trials, neutropenia accounting for dose reductions of PEG-interferon occurred in only 20% of patients.4, 5 Secondly, practitioners were probably less responsive to neutropenia than to anemia. Anemia is indeed more often symptomatic than neutropenia. Nevertheless, the use of G-CSF has revolutionized the field of dose-intensive chemotherapy by reducing neutropenia-related complications.34, 35 As regards chronic hepatitis C, no official guidelines have been set for the use of G-CSF, probably because of the absence of firm data resulting in an increased sustained viral response or reduced infections. There are conflicting results on neutropenia as a risk factor for infections during antiviral therapy. Three retrospective studies did not identify such a risk factor36–38; the more recent one demonstrated that infections did not correlate with the nadir of neutrophil count (<1,000 or >750/mm3) or with the magnitude of its decrease from baseline.37 Conversely, Puoti et al.39 reported that acute respiratory infections were associated with neutropenia and that pegylated interferon increased the risk of non-respiratory infections independent of neutropenia.39 Moreover, neutropenic HCV-infected patients with end-stage liver disease, liver transplant recipients, or HIV infection are more susceptible to bacterial infections. The latter conditions warrant further trials to clarify how and when G-CSF should be used.

HGF are not yet an official adjuvant therapy for chronic hepatitis C. Recent studies indicate that EPO makes it possible to maintain ribavirin dosage and improve quality of life during combination therapy,8–10 as in the case of reported patients with cancer.40 Moreover, EPO was shown to enhance platelet reactivity and platelet counts in patients with alcoholic cirrhosis, an interesting feature during antiviral therapy in thrombopenic HCV-positive patients with cirrhosis.41 For all these reasons, EPO will likely play a greater role in the treatment of HCV infection.

Finally, our French survey reveals an important rate of HGF prescriptions but with considerable disparity in the schedule of injections, the molecule used, and above all regarding the indications. We are aware that this study has several limitations. First, it does not provide an accurate estimation of the dose and duration of HGF actually received by the patients because it is based on self-reports by practitioners and cannot avoid recall biases. Second, some practitioners could have been reluctant to declare that they transgressed official recommendations by prescribing HGF. These drugs are costly, but when compared with standard care, the use of EPO has recently been shown to be cost-effective in managing HCV by increasing therapeutic compliance, improving quality of life, and avoiding complications of chronic liver disease.42 While awaiting new innovating molecules causing fewer hematological disorders,43 the main task of health care authorities and scientific societies will be to clarify the suitable role of HGF as a complement to HCV therapy. Moreover, all available EPO and G-CSF have different pharmacological properties, and trials comparing these molecules are also warranted. Finally, the anonymous, retrospective, and uncontrolled design of the current study did not allow us to provide reliable indications regarding the impact of adjuvant therapies on the sustained responses rates. Such crucial information would require further controlled trials.