Ribavirin is a broad-spectrum, oral, purine nucleoside analog antiviral agent that is similar in structure to guanosine. It is widely used in an aerosol formulation to treat respiratory syncytial virus infections in children [7]. In preliminary trials, an oral preparation of ribavirin was found to improve serum aminotransferase levels in approximately one third of patients with hepatitis C [8-10]. However, a 6-month course of ribavirin therapy was associated with little change in liver histopathology and serum HCV RNA levels. We evaluated the effect of ribavirin administered for a longer period in a randomized, double-blind, placebo-controlled trial.

Methods

Top Methods Results Discussion Author & Article Info References

Study Sample

Patients enrolled in our trial had compensated chronic hepatitis C and no evidence of other liver disease. Entry criteria included 1) persistent elevations in serum alanine aminotransferase levels that averaged more than twice the upper limit of normal on three measurements made within the previous 6 months; 2) the presence of antibody to HCV [anti-HCV] and HCV RNA in serum; and 3) chronic hepatitis determined by liver biopsy. Exclusion criteria included therapy with other agents, such as interferon, during the previous 6 months; decompensated liver disease; pregnancy or inability to practice birth control; preexisting anemia; serologic evidence of ongoing hepatitis B, hepatitis D, or HIV infection; other significant medical illness; or alcohol abuse (> 60 g/d).

Study Design and Treatment

After a preliminary period of screening assessments, patients were admitted to the Clinical Center of the National Institutes of Health for medical evaluation and liver biopsy. They were then randomly assigned (using random numbers derived from published tables in blocks of 6, 8, or 10) to receive either ribavirin or placebo. Ribavirin (Virazole, ICN Pharmaceuticals, Costa Mesa, California) was administered as three 200-mg capsules twice daily. Placebo was administered in capsules that were identical in appearance to and given in the same manner as the ribavirin capsules. Therapy was continued for 48 weeks. After starting therapy, patients were evaluated in an outpatient setting once a week for the first month, once every 2 weeks for the second month, then once monthly until 6 months, and bimonthly thereafter. At each visit, patients were interviewed and examined and had blood drawn for routine biochemical and hematologic tests and tests for viral markers. Patients and caregivers were blinded to treatment assignments. The blood test results were also withheld from patients and caregivers but were monitored by an investigator who had no contact with patients during the trial. Compliance was monitored by evaluating the regularity of attendance at clinic visits, by reviewing patient diaries, and by pill count. Maintenance of the blinded nature of the study was assessed by asking patients at the end of the study whether they believed they had received ribavirin or placebo.

During the last 2 weeks of treatment, patients again had medical evaluations and liver biopsies, after which the treatment code was broken. Patients who had received placebo were offered treatment with ribavirin in an open trial. Patients who had received ribavirin were followed at 1- to 2-month intervals for at least 6 months.

Routine blood tests included complete blood counts and biochemical tests of liver and renal function. Selected serum samples from each patient were tested for anti-HCV (HCV EIA 2.0, Abbott Laboratories, North Chicago, Illinois). They were also tested for HCV RNA by reverse transcription polymerase chain reaction (PCR) using "nested" primers from the 5' noncoding region of the viral genome [11] and for amount of HCV by the branched DNA (bDNA) signal amplification assay [12] (Quantiplex, Chiron, Emeryville, California). The bDNA assay has a sensitivity of 3.5 x 10⁵ genome-equivalents per mL, whereas the PCR assay can detect as few as 500 HCV RNA molecules. Genotypes of HCV were determined by reverse hybridization of HCV complementary DNA amplified from serum by reverse transcription PCR [13, 14] (Inno-Lipa, Innogenetics, Brussels, Belgium). Liver biopsy specimens taken before and after therapy were evaluated as a group and read in random order by a single investigator who was blind to the treatment status of each patient. The effect of therapy was assessed by evaluating the degree of necrosis, inflammation, and fibrosis semiquantitatively using the histology activity index [15, 16].

Statistical Analysis

Changes in serum aminotransferase levels during therapy were assessed by using repeated-measures analysis of variance. The two treatment groups at entry were compared for similarity using a two-sample Student t-test. In addition, patients were classified as responders, partial responders, or nonresponders on the basis of serum aminotransferase levels during therapy. A complete response was defined as having occurred if the mean of the six bimonthly values for alanine aminotransferase levels fell into the normal range (< 41 U/L [0.69 µkat/L]). A partial response was defined as having occurred if the mean of the alanine aminotransferase levels was between 1 and 1.5 times the upper limit of the normal range (41 to 62 U/L [0.69 to 1.04 µkat/L]) and was less than 50% of the pretreatment value. All patients who did not have a complete or a partial response were considered nonresponders. These criteria for a definition of response are different from those used in trials of interferon- for chronic hepatitis C [5, 6] but were based on previous experience with ribavirin [8] and were established before the start of the trial. The characteristics before treatment of responders, partial responders, and nonresponders were compared using the Fisher exact test on nominal data or Wilcoxon tests on ordered data.

Our trial was approved by the Institutional Review Board of the National Institute of Diabetes and Digestive and Kidney Diseases, and all patients gave written, informed consent.

Results

Top Methods Results Discussion Author & Article Info References

Between October 1991 and August 1992, we enrolled 58 patients in the trial (39 men and 19 women, ranging in age from 28 to 66 years [mean age, 44 years]). The known duration of hepatitis ranged from 1 to 26 years (mean duration, 6.3 years). The presumed source of hepatitis was intravenous drug use in 30 patients, blood transfusion in 19 patients, and unknown in 9 patients. Seventeen (29%) patients had previously been treated with interferon-, which had not produced a lasting beneficial effect. Before treatment, HCV RNA was detectable in the serum specimens of all patients by PCR and in the serum specimens of 51 patients (88%) by bDNA assay. The HCV genotype was 1a in 24 patients, 1b in 21 patients, 2a in four patients, 2b in three patients, 3 in four patients, and 4 in one patient. The serum HCV RNA level was not sufficient to allow genotyping in one patient.

Twenty-nine patients were randomly assigned to receive ribavirin treatment and 29 were assigned to receive placebo. The initial demographic characteristics and serum biochemical and serologic features of the two groups were similar in all important ways with the exception of the mean hematocrit, which was initially higher in the ribavirin-treated group (Table 1).

Therapy was continued for 48 weeks. All patients were followed for the full 48 weeks and all had repeated liver biopsy. Compliance was similar for patients receiving ribavirin and for controls. Thus, of the 13 visits scheduled during therapy for the 58 patients, only 15 (2%) were missed (1.9% for patients receiving ribavirin and 2.1% for controls). Pill counts also suggested excellent compliance, which was similar for patients receiving ribavirin (mean, 96.0%; range, 80.3% to 100%) and for controls (mean, 96.7%; range, 87.4% to 100%). At the time of the liver biopsy that was done at 1 year, patients were asked whether they believed they were receiving ribavirin or placebo. Seven (12%) insisted that they could not tell; among the 51 patients who would guess, 38 (75%) were correct (21 of 26 ribavirin-treated patients [81%] and 17 of 25 controls [68%]).

The mean serum alanine aminotransferase levels decreased among the 29 ribavirin-treated patients (P = 0.0001) but not among the 29 controls (Figure 1). The decrease in alanine aminotransferase levels seen with ribavirin therapy occurred predominantly during the first 2 months, and the levels thereafter remained constant at an average of 54% below baseline levels. The mean serum alanine aminotransferase levels (the average of the 6 bimonthly values) during therapy were below baseline levels in all ribavirin-treated patients (ranging from –87% to –9%), with the exception of 1 patient in whom the mean value increased by 49%. Alanine aminotransferase levels reached the normal range and remained within it (complete response) in 6 ribavirin-treated patients (21%) and came to within 1.5 times the upper limit of the normal range (partial response) in another 4 ribavirin-treated patients (14%). No controls had either a complete or a partial response. Thus, defined biochemical responses occurred in 10 of 29 ribavirin-treated patients (35% [95% CI, 18% to 54%]) and 0 of 29 controls (0% [CI, 0% to 12%]). Changes in aspartate aminotransferase levels generally paralleled those in alanine aminotransferase levels. Prothrombin times and serum albumin and total protein levels did not change appreciably in either treated patients or controls.

View larger version (30K): [in this window] [in a new window]   Figure 1. Mean serum alanine aminotransferase levels in 29 patients with chronic hepatitis C who were treated with ribavirin (solid line) and 29 patients with chronic hepatitis C who were treated with placebo (dashed line). Levels decreased significantly among ribavirin-treated patients but not among placebo recipients (repeated-measures analysis of variance). To convert U/L to µkat/L, multiply by 0.01667.

Serum HCV RNA levels did not change during therapy in either the patients receiving ribavirin or in the controls (Figure 2). All patients remained positive for HCV RNA by PCR. The number of patients reactive for HCV RNA by the bDNA signal amplification assay decreased by 1 in both the ribavirin-treated group (from 27 to 26 patients) and the control group (from 24 to 23 patients).

View larger version (22K): [in this window] [in a new window]   Figure 2. Mean serum hepatitis C virus RNA levels as measured by branched DNA signal amplification assay in 29 patients with chronic hepatitis C who were treated with ribavirin (solid line) and 29 patients with chronic hepatitis C who received placebo (dashed line).

Histologic examination of the initial liver biopsy specimens showed chronic hepatitis in all patients and cirrhosis in 8 patients. The average degree of histologic abnormality in the biopsy specimens taken before treatment was similar in the ribavirin-treated patients and the controls. Comparison of the initial and final biopsy specimens, however, showed improvement in the overall histology activity index among ribavirin recipients (P = 0.01) but not among controls (P = 0.74) (Table 2). The improvement seen in the ribavirin-treated patients was primarily in the degree of intralobular injury (P = 0.01). The degree of fibrosis did not change in ribavirin recipients and worsened slightly in controls (P = 0.06), but no patient developed cirrhosis. The overall histology activity index score decreased in the ribavirin-treated patients (from 11.7 to 10.0; P < 0.01), but the decrease was not statistically greater than the decrease in the controls (from 11.3 to 11.0) (P = 0.09).

Histologic examination of liver specimens showed that more improvement occurred in the 10 responders to ribavirin than in the 19 nonresponders Figure 3 (P = 0.0039). The degree of fibrosis did not change, but the current activity score (the sum of the first three components of the histology activity index score) improved in all but 1 responder, who initially had had the lowest activity score. Although the histologic scores improved, the overall histologic findings did not revert to normal in any patient.

View larger version (21K): [in this window] [in a new window]   Figure 3. Total histology activity index scores for the initial and final liver biopsy specimens from 10 patients who responded (left) and 19 patients who did not respond (right) to ribavirin therapy. In the left panel, closed circles denote complete responders and open circles denote partial responders. Among responders, the mean histology activity index score decreased by 4.3 (P = 0.0039, Wilcoxon signed-rank test). Among nonresponders, the mean histology activity index score decreased by 0.3 (P = 0.64).

The initial characteristics of patients who had complete or partial responses (responders) are compared in Table 3 with those who had no response to ribavirin therapy (nonresponders). On initial evaluation, the 10 responders tended to be older and to have lower serum aminotransferase levels than the 19 patients who did not respond. Little difference was apparent between responders and nonresponders in other characteristics, including histologic findings, HCV genotype, or HCV RNA titer in serum, characteristics that tend to predict response to interferon- therapy [17]. Compliance was minimally better among the responders than among the nonresponders (97.8% compared with 95.0% [P = 0.37]). However, the average total ribavirin dose was greater in responders, probably because none of 6 patients who had a reduction in dosage (see below) was a responder.

All patients who received ribavirin were followed for 6 to 24 months after therapy was discontinued. In all but two patients (7% [CI, 1% to 23%]), serum aminotransferase levels returned to pretreatment values within 2 to 3 months after ribavirin therapy was discontinued. Two patients continued to have normal aminotransferase values for more than 1 year (22 and 30 months, respectively) after ribavirin therapy was stopped. Both patients continued to have HCV RNA detectable in serum.

Ribavirin was generally well tolerated. The predominant side effect noted was hemolysis. Hematocrits decreased approximately 13% (from a mean of 0.45 to a mean of 0.39; P = 0.0001) during the first 1 to 2 months of ribavirin therapy and were stable thereafter. Accompanying the decreases in hematocrits were increases in reticulocyte counts (from a mean of 0.017 to a mean of 0.061; P = 0.0001) and serum bilirubin levels (from a mean of 14.9 µmol/L to a mean of 17.4 µmol/L; P = 0.001). In most patients, hematocrits decreased but remained within the normal range. Only three patients developed clear-cut symptoms of anemia (weakness, headaches, shortness of breath, and palpitations), which led to a decrease in the dosage of ribavirin to 800 mg/d, after which the hematocrits stabilized and the patients had no symptoms clearly attributable to anemia. In three other patients, the ribavirin dose was decreased because of anxiety, irritability, and fatigue unrelated to anemia. One of the three patients stopped receiving ribavirin therapy after 7 months because of fatigue and irritability that persisted despite a decrease in dose but that resolved when therapy was discontinued. The total leukocyte counts also decreased significantly during therapy (from a mean of 6.2 x 10⁹/L to a mean of 5.6 x 10⁹/L; P = 0.002); this decrease was due largely to depression of the lymphocyte counts (from a mean of 2.1 x 10⁹/L to a mean of 1.5 x 10⁹/L). Neutrophil and platelet counts were essentially unchanged.

Two other side effects of ribavirin were seen. During ribavirin therapy, five patients developed chronic or refractory upper respiratory symptoms suggestive of sinusitis or ear infections. Three other patients developed pruritus. These symptoms developed during the last 6 months of treatment and resolved when ribavirin therapy was stopped. At the time of our study, these side effects were not attributed to ribavirin, and the dosage was not modified in patients who developed these symptoms.

Discussion

Top Methods Results Discussion Author & Article Info References

Several pilot studies have suggested that ribavirin has a beneficial effect in patients with chronic hepatitis C [8-10]. In these studies, ribavirin was given for 3, 6, and 9 months, and the antiviral effects of treatment were not fully evaluated. Titers of HCV RNA were assessed by the relatively inaccurate method of end-point titration using PCR [8, 10]. In addition, studies of ribavirin given for only 6 months showed that the drug produced no statistically significant improvement in histologic findings in the liver [8]. These preliminary studies did not include untreated controls.

Our study was designed to address the shortcomings of the pilot studies. Ribavirin was administered for 12 months and compared with placebo. Histologic findings in the liver were carefully scored under code, and HCV RNA titers were quantified by the recently developed method of bDNA signal amplification [12]. This 12-month trial showed again that ribavirin therapy for chronic hepatitis C was accompanied by a rapid improvement in serum aminotransferase levels, which fell into the normal or nearly normal range in 35% of patients. In addition, histologic findings in the liver appeared to improve with ribavirin therapy, but only in those patients who also manifested a significant improvement in aminotransferase levels. Histologic improvement was largely in the degree of hepatocellular necrosis and inflammation. As might be expected, the degree of hepatic fibrosis did not change appreciably. Furthermore, both controls and patients who had no or little improvement in serum aminotransferase levels had no or little improvement in histologic findings in the liver. These results are compatible with what is known about the natural history of chronic hepatitis C. Improved histologic findings in the liver are also seen with interferon- therapy, largely in patients whose aminotransferase levels fall into the normal range [5, 6].

Despite improvements in serum biochemical test results and histologic findings in liver specimens, the serum HCV RNA levels of patients treated with ribavirin did not change. Studies done on liver biopsy tissue also showed no decrease in the amount of staining for HCV antigen in liver during ribavirin therapy [18]. In this regard, our results differ from those obtained with interferon-, which leads reliably to decreased HCV RNA levels in both serum and liver [11, 12, 18]. Indeed, most patients whose aminotransferase levels fall into the normal range during interferon- therapy become nonreactive for HCV RNA. Furthermore, most long-term responders to interferon- continue to have no detectable HCV RNA in serum specimens. With ribavirin, HCV RNA levels did not decrease, even among responders, and the two patients who had long-term improvement in serum aminotransferase levels both had persistence of HCV RNA in serum at levels equivalent to or even higher than levels before treatment.

The finding of a discrepancy between virologic and serum biochemical responses in patients having ribavirin therapy was unexpected and raises questions about the mechanisms of action of ribavirin in chronic hepatitis C. One possibility is that ribavirin acts not as an antiviral but as an immunomodulatory or immunosuppressive agent. This is compatible with a decrease in hepatic inflammation and necrosis with little or no change in HCV RNA levels, implying that ribavirin might have a dampening effect on immune-mediated or cytokine-mediated liver injury related to HCV infection. Supporting this hypothesis are previous reports of the immunosuppressive actions of ribavirin [19] and our findings that absolute lymphocyte counts decreased by an average of 30% among ribavirin-treated patients. However, the decrease in lymphocyte counts did not correlate with decrease in aminotransferase levels, and lymphocyte counts in the two long-term responders returned to pretreatment levels shortly after therapy was discontinued.

More likely explanations of the lack of change in HCV RNA levels during ribavirin therapy are that these levels do not reflect the degree of viral inhibition in the liver or that the effects of ribavirin on viral replication can be overcome by compensatory viral mechanisms that may lead to a shift to a less pathogenic strain or quasi-species of HCV. Interestingly, with interferon therapy, changes in serum HCV RNA status generally closely parallel changes within the liver. These are hypothetical explanations of the effects of ribavirin, but they invite interesting approaches to elucidating the mechanism of action of this agent in chronic hepatitis C.

In our study, ribavirin was generally well tolerated. Minor side effects included irritability, fatigue, upper respiratory symptoms, and pruritus. The frequency of these side effects probably accounted for the finding that 75% of patients who were willing to guess were able to correctly guess whether they were receiving ribavirin or placebo. Hemolytic anemia was the most significant side effect, but it was usually not symptomatic or severe [20]. Hemolysis was, however, dose-limiting in some patients and would have been particularly difficult if patients had had preexisting anemia. Chronic hemolysis induced by ribavirin may also lead to the deposition of excessive amounts of iron in the liver, a side effect that could limit the long-term use of ribavirin [21].

Although ribavirin appeared to have a clear-cut effect on serum aminotransferase levels, this effect was largely transient and lasted only as long as ribavirin was administered. Only 2 of the 10 responders maintained normal aminotransferase levels after ribavirin therapy was discontinued. These 2 responders were still viremic when last tested. Thus, the long-term prognosis for continued benefit is still questionable. Ribavirin therapy was not associated with the elimination of HCV RNA and was rarely associated with sustained improvement in serum aminotransferase levels; thus, it is of limited use as a single agent given for a finite period for chronic hepatitis C. Our results suggest that continuous therapy would be needed to maintain the benefit of treatment. Continuous therapy might be practical, but further studies are needed to document its long-term benefit and lack of serious side effects.

Perhaps more attractive than ribavirin alone as therapy for chronic hepatitis C is the combination of ribavirin and interferon-. Several recent reports [22-24] have suggested that this combination leads to a higher sustained response rate than either agent alone. These preliminary results provide a rationale for a definitive randomized, controlled trial comparing this combination with interferon- therapy alone in patients with chronic hepatitis C who were and were not previously treated with interferon.