Introduction A panel of experts met in Paris on 18 April to present research and practice protocols, and to discuss topics of current interest related to the treatment of HIV/hepatitis C virus (HCV) co-infection. The information presented focussed on the following main aspects: The magnitude of the problem of HIV/HCV co-infection The differences and similarities between HIV and HCV disease paradigms, thereby allowing extrapolation of the lessons learned in HIV on the care of patients with HCV or HIV/HCV co-infections Strategies for managing HIV in HCV-co-infected patients using antiretroviral drugs Current standards for HCV treatment and ongoing management Strategies for treating HCV in HIV-co-infected patients using pegylated interferon (peg-IFN) plus ribavirin, and the management of possible adverse effects Special challenges in HIV/HCV co-infection, including non-responders to IFN/ribavirin, patients with cirrhosis, extrahepatic manifestations, and hepatitis B virus co-infection. The bulk of the meeting was devoted to a discussion of the specifics of HIV/HCV co-infection treatment, answering the questions: why? how? who? when? After presentations and discussions, a consensus of opinion regarding general treatment strategy was formulated. Hepatitis C virus treatment in co-infected patients: why? AIDS-related morbidity and mortality in HIV-infected patients continue to decrease as a result of effective antiretroviral therapy and prophylaxis for traditional opportunistic infections [1]. HIV-infected patients now have hope for a prolonged AIDS-free survival. Concurrently, however, the morbidity and mortality from co-morbid HCV infection within this population is on the increase. The magnitude of the co-infection problem becomes clear when its prevalence and impact on morbidity and mortality are considered [2–9]. Prevalence of hepatitis C virus–HIV co-infection In the United States, it is estimated that 30% of the 800 000 HIV-infected living individuals are co-infected with HCV [10,11]. Similar rates (33%) have been estimated for western Europe, although the number of HIV-infected individuals is less well defined [12]. However, the magnitude of the problem is alarming in countries such as Spain, where at least half of the 130 000 HIV-infected patients are estimated to be HIV/HCV co-infected [12]. In fact, among some sub-groups of HIV-infected patients, such as injection drug users, the prevalence of co-infection is as high as 70–90% [11–13]. Hepatitis C virus and clinical progression of HIV disease The Swiss Cohort Study demonstrated that HCV accelerates the progression of HIV disease [14]. This prospective study of patients starting highly active antiretroviral therapy (HAART) found that HCV was independently associated with an increased risk of progression to AIDS and death. This finding was not related to a lower usage or much poorer tolerance of antiretroviral drugs among individuals with hepatitis C, which is in agreement with findings from other European groups [15]. Therefore, hepatitis C might be considered to be a co-factor for HIV disease progression. On the other hand, the Johns Hopkins Cohort Study found that co-infected patients who had a baseline CD4 cell count of between 50 and 200 cells/mm3 progressed to death more quickly than their HIV-mono-infected counterparts [16]. This observation probably highlighted the fact that HCV-positive patients, most of whom were intravenous drug addicts, had significantly less exposure to HAART, with a delay in treatment until CD4 cell counts dropped below 50 cells/mm3. In contrast with this potential deleterious effect of hepatitis C on HIV disease progression, recent reports [17,18] have pointed out that the hepatitis G virus, an agent closely related to HCV, seem to exert a protective effect on the course of HIV disease. As treatment of HCV with IFN is equally effective against hepatitis G, its clearance might negatively influence HIV infection. Hepatitis C virus and response to highly active antiretroviral therapy The Swiss Cohort Study also demonstrated that HCV may impair immune reconstitution after effective HAART [14]. HCV-positive individuals were less likely to achieve a CD4 cell increase of at least 50 cells/mm3 at one year after the start of HAART compared with HCV-negative individuals. This observation has not, however, been confirmed by others [19], and warrants further studies. HIV and acceleration of hepatitis C virus liver disease HIV accelerates HCV-related liver disease [20–25]. Progression that typically takes up to 30 years or longer in HCV-mono-infected individuals has been shown to take less than half that time in co-infected individuals. An early study by Martin et al. [26] identified the development of cirrhosis within 3 years after HCV diagnosis in three co-infected patients. In 1993, Eyster et al. [27] found that liver failure was accelerated by HIV in HCV-infected haemophiliac individuals. The following year, Telfer et al. [28] published a retrospective study, which found that the median time from first exposure to HCV to clinical demise was only 16.5 years in co-infected haemophiliac individuals. In a large study of HCV-mono-infected and HIV/HCV co-infected individuals, Sánchez-Quijano et al. [29] found that within 15 years of initial HCV infection, 25% of those who were co-infected with HIV developed cirrhosis compared with only 6.5% of those without HIV infection. Similar data were obtained by Soto et al. [30], who followed a large group of HCV-mono-infected and co-infected patients. In the first 10 years, 14.9% of co-infected patients developed cirrhosis compared with only 2.6% of HCV-mono-infected patients. Overall, cirrhotic HIV-infected patients with HCV do very poorly. In a study conducted by Di Martino and colleagues [31], HIV/HCV-co-infected patients with cirrhosis were more likely to decompensate and die than patients who were HIV negative. Interestingly, HCV treatment with IFN plus ribavirin appeared to be protective in most instances, with protection extending to those with HIV. This suggests retrospectively that although HCV treatment is less effective in cirrhotic patients from an antiviral perspective, it may delay decompensation [32]. Hepatocellular carcinoma in co-infected patients Hepatocellular carcinoma (HCC) appears to occur at a younger age and after a shorter duration of HCV infection in co-infected individuals. This was the finding in a case–control study in which seven co-infected individuals with HCC were analysed [33]. The mean age at HCC diagnosis was 42 years in the co-infected group compared with 69 years in the control group. The estimated mean length of HCV infection before HCC diagnosis was 18 years in the co-infected group, compared with 28 years in the control group. End-stage liver disease mortality in co-infected patients A number of studies have demonstrated the association of HIV co-infection with an increased risk of morbidity and mortality caused by end-stage liver disease (ESLD) (Table 1) [2–9]. Iribarren et al. [34] reviewed the causes of death among a population of 1600 co-infected patients in a Spanish hospital over a 21 month period. Of the 44 total deaths, liver disease was responsible for up to 25%. Recent studies from Italy [3] and Spain [2,4] have compared the percentage of total in-hospital deaths caused by ESLD before 1995 with in-hospital deaths occurring within a time period after 1995. Although the total number of in-hospital deaths declined from the first time period to the later, the percentage of deaths caused by ESLD increased from 13 to 35% in the Italian study, and from 5 to 45% in the Spanish studies.Table 1: Mortality caused by end-stage liver disease among HIV-infected individuals. Similar results were seen in a US study [6]. In 1991, 11% of deaths in the studied HIV population were caused by ESLD. By 1998, ESLD was the leading cause of death, causing 50% of deaths (Table 1). Of those patients who died in 1998 of ESLD, 90% were HCV positive. In France, Cacoub et al. [8] documented a fivefold increase in deaths caused by liver disease in the time period before 1995 compared with after 1997, despite an overall declining death rate among co-infected individuals (Table 1). In 1997, Darby et al. [35] published a cohort study demonstrating the impact of co-infection among young men with haemophilia. HIV-infected patients, regardless of the severity of haemophilia, were found to be approximately sevenfold more likely to die of liver disease. Lesens et al. [36] also demonstrated a sevenfold increased risk of death in a 1999 prospective study of 147 HCV-positive haemophiliac individuals. Interestingly, co-infected patients with genotype 1 may have a more rapid progression of liver disease than individuals carrying other HCV genotypes [37]. Immune status influences hepatitis C virus liver disease Rockstroh et al. [38] looked at the association between immune function and the development of HCV liver disease. The study concluded either that immunosuppression accelerates the progression of liver disease or that once liver failure begins, deterioration with respect to AIDS also begins, progressing to death more rapidly. The finding of a greater severity of HCV liver disease as the immunodeficiency progresses has been confirmed by others [39]. Accordingly, the 1999 US Public Health Service/Infectious Diseases Society of America guidelines recognized HCV co-infection as an important opportunistic pathogen among HIV-infected patients [36,40]. Impact of highly active antiretroviral therapy on hepatitis C virus liver disease The impact of HAART on the progression of HCV liver disease is controversial. One possibility is that antiretroviral therapy could increase hepatic necroinflammatory activity and thereby accelerate the progression of HCV-related liver disease. Vento et al. [41] reported an increased mean Knodell score (from 8 to 13) in patients after starting HAART. Conversely, other studies have suggested that the use of HIV protease inhibitors (PI) may be associated with an improvement in liver histology with respect to those without PI [42,43]. This benefit may probably equally be seen with other potent antiretroviral regimens without PI. Prospective studies involving paired liver biopsies are needed to address the impact of HAART and immune reconstitution on HCV-related liver disease. Hepatotoxicity of highly active antiretroviral therapy in co-infected patients The association of chronic HCV with hepatotoxicity during HAART is well established [44,45]. Hepatitis C is an independent risk factor for hepatotoxicity with HAART (Table 2) [46–52]. Overall, significant liver enzyme elevations are seen in approximately 15% of individuals receiving antiretroviral drugs. Severe hepatotoxicity, however, leading to drug discontinuation, occurs in less than 10% of cases. Two mechanisms have been involved (Table 3), the first of which represents a hypersensitivity reaction, often affecting the skin and other organs, and occurring a few days to weeks after beginning antiretroviral therapy. A second mechanism with delayed onset (typically appearing several months after beginning therapy) is limited to the liver, and represents an intrinsic toxic effect of the drugs in use, and therefore is dose related [52]. Drugs such as nevirapine can produce liver toxicity by both mechanisms; whereas abacavir tends to involve just the first mechanism, often in the context of a multiorganic reaction. Drugs such as stavudine may cause liver toxicity through a cumulative effect.Table 2: Major studies assessing the risk of severe hepatic damage after beginning antiretroviral therapy. Table 3: Mechanisms of liver toxicity using antiretroviral drugs [32]. More rarely, in HCV chronic carriers experiencing a dramatic CD4 cell increase after beginning antiretroviral therapy, increases in transaminases can reflect an immune reconstitution syndrome [53], resembling what has been described in individuals with latent cytomegalovirus or mycobacterial infections. Further research is needed to determine the mechanism by which HCV infection or HCV-related liver disease increases the risk of HAART-associated liver injury. Hepatitis C virus-RNA dynamics in HIV infection and impact of highly active antiretroviral therapy Overall, serum HCV-RNA titres are 1.5 to twofold higher in HIV/HCV co-infected individuals with respect to individuals with single HCV infections [54–56], probably reflecting an impairment in the control of HCV replication in the setting of immunodeficiency. Whether this increase in HCV viral burden contributes to explaining the greater liver injury noticed in HIV/HCV co-infected patients is unknown, although there is no clear correlation between the extent of liver fibrosis and the level of HCV RNA. In individuals who begin potent antiretroviral therapy, serum HCV-RNA levels tend to increase during the first 3 months [57–59], decreasing slowly thereafter, first returning to baseline levels and even decreasing much later [60] (Fig. 1). Antiretroviral therapy may thus indirectly benefit the prognosis of HCV-related liver disease, reducing HCV replication on the long term. However, the relationship of HCV load and the progression of liver disease is uncertain.Fig. 1.: Dynamics of serum hepatitis C virus RNA in HIV infection and impact of antiretroviral treatment.Hepatitis C virus treatment in co-infected patients: how? The primary goal of HCV treatment is to achieve a sustained virological response that permits fibrosis regression, the disappearance of extrahepatic manifestations, and a reduction of the risk of transmission [61]. Moreover, in patients without sustained virological response, the progression of fibrosis could be ameliorated through suppressive maintenance therapy [32,62]. Until recently, IFN plus ribavirin combination therapy was the standard of care for the treatment of HCV infection [11,61]. Peg-IFN plus ribavirin combination therapy is, however, currently the preferred option, as it allows one to achieve the highest virological response rates to date: 41–42% for genotype 1 and 76–80% for genotypes 2 and 3 (Table 4) [63,64]. In addition to inducing virological response, peg-IFN/ribavirin also allows fibrosis regression in viral sustained responders [64,65]. In non-responders, in whom the first goal is not achieved, viral eradication, the second goal, slows fibrosis progression, and the prevention of clinical outcomes (ESLD, HCC, and death) [32] might be attained with maintenance therapy using peg-IFN monotherapy.Table 4: Sustained virological response to pegylated interferons in HIV-negative individuals with chronic hepatitis C. Factors associated with sustained virological response The achievement of sustained response depends on host and viral factors. Poynard et al. [65] identified five independent predictors of sustained response to IFN/ribavirin. Genotype 2 or 3 is the most important predictor. The remaining four predictors were: low viral load (< 3.5 million copies/ml), no or just portal fibrosis, female sex, and age below 40 years. Subsequent analyses demonstrated that female sex as a predictor was an issue of body mass index rather than sex. Ribavirin doses particularly need to be adequate to weight if optimal response rates are to be achieved [64]. For instance, when using adequate ribavirin doses, up to 48% of individuals with genotype 1 and up to 88% of those with genotypes 2/3 reached sustained response using peg-IFN plus ribavirin [64]. Additional predictive factors of response related to HIV include CD4 cell counts greater than 500 cells/mm3, plasma HIV-RNA levels below 10 000 copies/ml, and no alcohol consumption [23,66]. Treatment considerations related to CD4 cell counts Co-infected patients with CD4 cell counts greater than 500 cells/mm3 should be treated for HCV eradication. In individuals with CD4 cell counts of less than 500/mm3, treatment is less effective [66], but may be considered in order to reduce the risk of hepatotoxicity of antiretroviral drugs and the higher risk of progression to liver failure among patients with lower CD4 cell counts [38,39]. Overall, current HCV treatment in HIV/HCV co-infected patients can normalize alanine aminotransferase (ALT) levels and clean HCV RNA by 50%, decrease the progression of fibrosis by 60%, and decrease the risk of dying by 16% [20–25,42,43]. Weight-based dosing Dosing on the basis of the patient's body weight seems to be the key to optimized success with minimal side-effects using peg-IFN/ribavirin: peg-IFN alfa-2b (1.5 μg/kg per week) or peg-IFN alfa-2a (180 μg per week) plus ribavirin (> 10.6 mg/kg per day) represents the most effective HCV treatment option [63,64,67]. This is especially important when considering the weight variation among population groups. For example, Americans weigh an average of 10 kg more than Europeans. Weight-based dosing not only ensures that patients receive enough drug, it also ensures that they do not receive too much drug, thereby reducing the risk of adverse events that may result if a standard dose is given to a low-weight patient. An adequate ribavirin dose, particularly at the beginning of treatment, is linked to an increased likelihood of sustained virological response. In a recent multicentre trial [68], greater ribavirin use at week 4 of treatment was associated with a greater response rate at week 24. Considering these findings, an adequate dose of ribavirin, particularly at the start of therapy, should not be modified without first trying other strategies to increase tolerance, such as the use of epoetin alfa [69]. Concern about ribavirin use in co-infected patients There has been concern about the use of ribavirin in HIV/HCV co-infected patients because of dose-dependent anaemia and drug–drug interactions [23,70]. Ribavirin-induced anaemia may be more significant in co-infected patients [21,22]. This risk, however, should not preclude treatment, because it can quite often be successfully managed with erythropoietin [69] or by ribavirin dose reduction. Prospective studies examining the efficacy of this approach are ongoing. Concerns regarding interactions between ribavirin and antiretroviral drugs are more complex. Ribavirin, a guanosine nucleoside analogue, is known to inhibit the intracellular phosphorylation of zidovudine, stavudine and zalcitabine in vitro. There is concern that this may cause anti-HIV antagonism in vivo. However, to date, clinical data have not supported these in-vitro observations [70,71]. In addition, ribavirin enhances the phosphorylation of didanosine (Fig. 2), which may be of benefit in increasing the anti-HIV effect [72]. However, recent case reports have led to concern about the possible increased risk of pancreatitis and mitochondrial toxicity in patients taking ribavirin and didanosine [73–76]. Therefore, patients receiving ribavirin in combination with nucleoside analogues such as zidovudine or didanosine should be observed closely, and in some cases, consideration may be given to modyfing HAART to avoid the combination of these drugs. Another aspect that is still unclear regards the potential compromise in the effect of ribavirin on HCV as a result of the concomitant use of zidovudine or stavudine, because all these compounds share the same phosphorylation pathways. Prospective studies are underway to evaluate the clinical and pharmacological interactions of HAART and peg-IFN/ribavirin therapy.Fig. 2.: Metabolic pathways leading to the potentiation of didanosine by ribavirin. Ribavirin inhibits inosine monophosphate (IMP) dehydrogenase. This leads to an increase of the IMP pool, which acts as a phosphate donor for the conversion of didanosine (ddI) into dideoxy-IMP (ddIMP). This compound is then metabolized into the triphosphorylated metabolites dideoxy-adenosine monophosphate (ddAMP), dideoxy-adenosine 5'-diphosphate (ddADP) and dideoxy-adenosine triphosphate (ddATP). The increased concentrations of ddATP inhibits both HIV reverse transcriptase and mitochondrial DNA polymerase γ.Warning on the risk of lactic acidosis Lactic acidosis is part of the spectrum of diseases associated with mitochondrial toxicity [77]. Clinical manifestations of mitochondrial toxicity may not be reversible and include pancreatitis, hepatitis/hepatic steatosis, myopathy, peripheral neuropathy, and lactic acidosis. Nucleoside analogues vary not only in the type of organ affected more frequently, but in their potential for causing toxicity-related disease [78]. So far, there is no evidence of mitochondrial toxicity nor of lactic acidosis with the use of ribavirin without other nucleoside analogues. A 1995 study [79] observed 1836 patients over 5 years with any anti-HIV nucleoside, and found a rate of lactic acidosis of 1.3 cases per 1000 patient-years. A more recent French study [80] found 8.4 cases per 1000 patient-years among 876 patients observed over 18 months, with any nucleoside analogues. When only patients taking stavudine and didanosine were considered, the incidence increased to 15.6 cases per 1000 patient-years [80]. The Food and Drug Administration has reports of 106 cases of lactic acidosis associated with nucleoside use. Of them, 46 were associated with a single nucleoside and 60 with dual nucleosides [69]. Overall, stavudine and didanosine were the nucleosides most commonly associated with lactic acidosis, 77 and 27% of the cases, respectively. A mortality rate of 56% among these 106 cases reported to the US Food and Drug Administration underscores the severity of the condition. Ribavirin is a potent inhibitor of inosine monophosphate (IMP) dehydrogenase, which leads to elevated levels of dideoxyadenosine triphosphate, the metabolically active product of didanosine, a key culprit involved in mitochondrial toxicity (Fig. 2). In-vitro assessment of the efficiency of ribavirin as a substrate for DNA-polymerase gamma has not been made, as the development of ribavirin predated that concern. No cases of lactic acidosis with ribavirin have been described so far in individuals with HCV mono-infection. Thirteen cases of lactic acidosis or mitochondrial toxicity associated with ribavirin have been identified up to now, however, in HIV/HCV-co-infected individuals [81]. Only nine of them had data reported that included the specifics of what drugs the patients were on. Eight were on didanosine, a fact that suggests the possibility that interaction was the culprit. One was not receiving nucleoside analogues, and that individual was stable with low-grade acidaemia, and was able to continue with peg-IFN/ribavirin. Six of the nine were also on stavudine, and two of the nine were on abacavir [81]. Abacavir is a purine analogue, so it is involved in the same metabolic pathway as didanosine (Fig. 2); therefore, it could interact with ribavirin. However, current data suggest that abacavir has a very low potential for mitochondrial toxicity [77]. Physicians treating HIV/HCV co-infected patients need to be aware of the potential risk of lactic acidosis associated with the concomitant use of ribavirin and purine nucleoside analogues. Although this condition has so far been seen only in combination with didanosine, the risk may also exist in combination with other nucleoside analogues. Taking into account these data, the use of ribavirin with nucleosides, particularly with didanosine, requires careful clinical and laboratory monitoring. Clinicians should consider routine (monthly) monitoring of serum lactate and amylase levels in HIV/HCV co-infected patients exposed concomitantly to didanosine and ribavirin. Interferon/ribavirin in co-infected patients Until 2 years ago, IFN monotherapy was the only drug available for the treatment of chronic hepatitis C. Overall, response rates to IFN observed in co-infected patients were similar to those observed in HIV-negative patients (Table 5) [82–90]. However, response rates were generally significantly lower among HIV/HCV co-infected patients with low CD4 cell counts [66]. In 1999, IFN/ribavirin combination therapy became the standard of care for the treatment of chronic hepatitis C [61]. Information on the safety and efficacy of combination therapy in co-infected individuals is scarce (Table 6 and Table 7) [70,71,91–101]. Twenty co-infected patients treated with standard IFN/ribavirin were followed by Suciu et al. [93]; at 18 months follow-up, eight (40%) had a sustained virological response. Likewise, Pérez-Olmeda et al. [100] noticed a sustained response rate of 34% taking IFN/ribavirin in 18 HIV/HCV relapsers to a previous course of IFN monotherapy. Those studies and others [101] were a proof of principle to determine that, with adequate CD4 T cells, HCV could indeed be eradicated in a co-infected patient. Similarly, in a French prospective study, Landau et al. [70] followed 19 co-infected patients who were non-responders to IFN monotherapy at 3 months. After 6 months on IFN/ribavirin combination therapy, one third achieved normal ALT and undetectable HCV-RNA levels.Table 5: Treatment response to interferon monotherapy in HIV/hepatitis C virus co-infected patients. Table 6: Treatment response to interferon plus ribavirin in HIV/hepatitis C virus co-infected patients. Table 7: Response to interferon plus ribavirin in HIV/hepatitis C virus co-infected patients who relapsed or failed to respond after a previous course of interferon monotherapy. Preliminary data show that patients receiving daily IFN/ribavirin achieve a higher sustained response rate than patients receiving thrice-weekly IFN/ribavirin, especially in those with HCV genotype 1 [102,103]. In an ongoing study from the Hepatitis Resource Network [102], HCV-RNA became undetectable after 12 weeks of therapy in 35% of patients receiving IFN daily plus ribavirin compared with only 8% of those receiving standard IFN three times weekly plus ribavirin. This supports the hypothesis that consistent exposure to IFN, rather than three times weekly, is more effective for viral eradication. These findings that favour an induction with IFN are being addressed specifically by an ongoing large Spanish trial [97]. In patients who relapsed after a previous course of IFN monotherapy, re-treatment with dual IFN/ribavirin provides an overall rate of response of one third (Table 7) [97,100,101]. Pegylated interferon/ribavirin in co-infected patients The first report of the safety and efficacy of combination therapy with peg-IFN plus ribavirin in HIV-positive individuals has shown that the overall rate of end-of-treatment response was 65% [78]. The trial is being conducted in Spain, and so far 31 individuals have completed treatment. The rate of relapse for these patients should be available at the end of the year. The French RIBAVIC-ANRS HCO2 study [104] is a prospective multicentre randomized trial currently in progress. To date, 239 patients have been randomly assigned to receive either IFN 3 MIU three times a week plus ribavirin 800 mg/day or peg-IFN 1.5 μg/kg per week plus ribavirin 800 mg/day for 48 weeks. Up to 24 weeks after the onset of treatment, the HIV viral load has remained stable in both treatment groups. Mean CD4 cell counts, ALT, weight, haemoglobin, total white blood counts, lymphocyte counts, neutrophil counts, and platelet counts all declined after the onset of HCV treatment, with no difference between the groups. Twenty-two cases of severe adverse events have been documented: nine in the peg-IFN/ribavirin group, 12 in the IFN/ribavirin group, and one before randomization. Adverse events included various infections, trauma, depression, neuropathy, ascites, and detoxification for drug abuse. The definitive withdrawal of treatments occurred in 29 cases: 13 in the peg-IFN/ribavirin group and 16 in the IFN/ribavirin group. Of these 29 cases, 26 withdrawals were requested by the patient as a result of fatigue (20), depression (two), and a desire not to start the trial (four). Larger studies with peg-IFN/ribavirin in co-infected patients are ongoing. The ACTG 5071 trial plans to include 150 subjects and an industry-sponsored study (NR-15961) will include 741 patients. The first results of these studies are expected to be released in 2003. Duration of therapy The duration of IFN/ribavirin therapy should be tailored according to the patient's virological response at 24 weeks of treatment and the number of positive predictors of response [65]. If HCV RNA is positive by polymerase chain reaction (PCR), treatment should be discontinued and other therapeutic strategies considered. If PCR is negative, and the patient has fewer than four predictors of response, treatment should continue another 24 weeks. If the patient has four or more predictors, treatment may be discontinued at 24 weeks. Moreover, results from a recent trial [63] suggested that the persistence of detectable HCV RNA after 12 weeks on peg-IFN plus ribavirin is highly predictive of a lack of further sustained response. Therefore, treatment might be discontinued at this early timepoint, avoiding unnecessary drug exposure and toxicity. This observation is of particular relevance in HIV-positive patients, most of whom are already under other multiple drugs. Further studies are needed to clarify this aspect. Management of anaemia with erythropoietin The major side-effect of ribavirin is reversible haemolytic anaemia. Haemoglobin levels drop below 11 g/dl in 25 to 35% of patients receiving IFN/ribavirin therapy [105,106]. Although this is usually managed by ribavirin dose reduction or discontinuation, recombinant human erythropoietin therapy has proved to be an effective treatment [22,69]. In addition, the ribavirin dose is more likely to remain unmodified in patients with anaemia treated with erythropoietin. From what is known about the importance of maintaining adequate and cons