Treatment of chronic hepatitis C in HIV/HCV-coinfection with interferon α-2b+ full-course vs. 16-week delayed ribavirin

Authors


Abstract

Human immunodeficiency virus (HIV)-infected patients increasingly experience the consequences of chronic hepatitis C virus (HCV) coinfection. This trial randomized 107 patients coinfected with HIV and HCV to receive 48 weeks of interferon alfa-2b (IFN) 3 million units three times weekly plus either a full course of ribavirin (RBV) at 800 mg/day (group A; n = 53) or 16 weeks of placebo, followed by RBV (group B; n = 54). The primary endpoint of sustained viral response (SVR) rate (undetectable HCV RNA at posttreatment week 24) was not different between groups A (11.3%) and B (5.6%; P = .32). Within group A, the SVR rate was lower in genotype 1 (2.5%) than in genotypes 2 through 4 (41.7%; P = .002). Fifty-five patients discontinued therapy prematurely, mostly because of adverse events or patient decisions. At treatment week 12, the percentage of CD4+ cells rose in group A (+4.1%; P < .001), but not in group B (−0.3%). A significant proportion (22%) of patients who were HIV viremic at baseline had undetectable HIV RNA at week 12. By week 16, the hemoglobin level decreased more in group A (−2,52 g/dL) than in group B (−1.02 g/dL; P < .001). In group A, the hemoglobin decline was steeper in patients receiving zidovudine (azidothymidine [AZT], −3.64 g/dL vs. no AZT, −2.08 g/dL), and patients receiving zidovudine had more anemia-related RBV dose reductions (AZT, 60% vs. no AZT, 16%). In conclusion, HCV therapy with IFN plus RBV is relatively safe in patients coinfected with HIV and HCV, but frequent treatment discontinuations and anemia-related RBV dose reductions contribute to a poor SVR rate. Control of HIV infection improves rather than worsens during therapy. (HEPATOLOGY 2004;39:989–998.)

Mortality resulting from human immunodeficiency virus (HIV) has declined dramatically since the introduction of highly active antiretroviral therapy in 1996. As HIV-infected persons live longer, they increasingly have health problems from comorbidities not related to HIV, chief among them coinfection with the hepatitis C virus (HCV). Persons coinfected with HIV and HCV have a higher mortality rate and a higher rate of hospitalization than patients infected with HIV alone.1 With declining overall mortality among patients with HIV or acquired immunodeficiency syndrome, relative mortality related to liver disease (especially from HCV infection) is increasing, reaching up to 50% in some centers.2, 3 The prevalence of HCV coinfection in HIV-infected persons depends largely on the method of transmission of HIV itself. It is high with parenteral transmission, such as injection drug use (91%) or transfusion of blood or blood products (71%), but low in sexually transmitted HIV (7.3%).4 Accordingly, one large HIV clinical trial cohort in the United States (18% injection drug use or transfusion) reported a prevalence of coinfection with HIV and HCV of 16%, but this study group does not claim to be representative of other HIV-infected populations or of all HIV-infected persons in the United States.5 Several studies from the era before highly active antiretroviral therapy have shown that HCV-related liver disease progresses faster in persons coinfected with HIV and HCV than in persons infected with HCV alone, particularly in patients with advanced immunodeficiency (CD4+ cells <200/mm3), but the mechanism is not well understood.6–8

Two randomized controlled studies in patients infected with HCV alone have shown that combination therapy with interferon alfa-2b (IFN) and ribavirin (RBV) has a superior sustained viral response rate than interferon alone.9, 10 The present study was conducted to evaluate safety and efficacy of IFN and RBV combination therapy in patients infected with both HIV and HCV.

Abbreviations

HIV, human immunodeficiency virus; HCV, hepatitis C virus; IFN, interferon alfa-2b; RBV, ribavirin; ALT, alanine aminotransferase; SVR, sustained viralogic response; AZT, azidothymidine; ddI, didanosine.

Patients and Methods

Patient Selection.

Adult patients were eligible if they were positive for HIV antibody and for plasma HCV RNA and had never received IFN or RBV. Liver biopsy results had to show necroinflammation (any degree), or where liver biopsy was contraindicated (n = 4), an elevated alanine aminotransferase (ALT) level was required. Patients were receiving a stable regimen of approved antiretroviral medication or no HIV treatment at all. Patients were excluded if they had decompensated liver disease; were seropositive for hepatitis B surface antigen; had active ischemic heart disease; had unstable psychiatric conditions, including active alcohol and substance use; had active HIV-related opportunistic infections; or were pregnant or breastfeeding. Because of concern for increased hepatotoxicity, patients were not allowed to be on concurrent therapy with ganciclovir, hydroxyurea, isoniazid, pyrazinamide, rifampin, or rifabutin. The study was approved by the institutional review board of each participating institution, and all patients signed an informed consent.

Study Design.

This randomized placebo-controlled multicenter study was carried out by the American Foundation for AIDS Research (amfAR) DCRI 010 Study Group at 21 sites in the United States from May 1998 through May 2002. Patients were randomized 1 to 1 to receive IFN at 3 million units (MU) three times weekly and either RBV at 800 mg/day (group A) or matching placebo (group B). Randomization was performed via telephone to the central randomization desk. Treatment allocation was stratified by site, and participants were assigned the next available slot using a random variable-block design. Patients in the RBV group continued their regimen for 48 weeks. Patients in the placebo group who were still HCV RNA positive at treatment week 12 were switched at week 16 to open-label RBV at 800 mg/day until week 48. This measure was designed as a salvage therapy for nonresponders to IFN monotherapy. Patients were followed up for 24 weeks after completion of the treatment period. Patients who discontinued study treatment prematurely were followed up throughout the treatment and follow-up period. Dose reductions of IFN from 3.0 to 1.5 MU three times weekly (or discontinuations) were required for white blood cell counts of less than 1,500/mm3 (<1,000 mm3), absolute neutrophil counts of less than 750/mm3 (<500/mm3), and platelet counts of less than 50,000 (<25,000/mm3). Similarly, doses of RBV or placebo were reduced from 800 to 600 mg/day (discontinued) for a level of hemoglobin of less than 9.0 g/dL (<7.5 g/dL) or a drop of 2.0 g/dL or more from baseline.

The primary endpoint was the rate of sustained viral response (SVR) with undetectable plasma HCV RNA (<100 copies/mL) at study week 72 (posttreatment week 24). Viral response was measured on an intention-to-treat basis, that is, lacking HCV RNA data were treated as viral nonresponse. Secondary efficacy endpoints were rates of viral response at weeks 12 and 48 (end-of-treatment response). Safety endpoints were incidence of clinical adverse events; rate of dose reduction or discontinuation of study drugs; changes in hemoglobin level white blood cell, neutrophil, or platelet counts; and changes in HIV viral load and in CD4+ cell count during the first 16 weeks (placebo-controlled phase).

Quantitative plasma HCV RNA testing and HCV genotyping were performed centrally by National Genetics Institute (Los Angeles, CA) using the SuperQuant™ polymerase chain reaction (PCR) assay with a dynamic range of 100 to 100 million HCV RNA copies/mL (National Genetics Institute) and the Inno-Lipa Probe™ assay (Bayer Diagnostics, Tarrytown, NY), respectively. All other laboratory tests and liver histologic interpretations (without quantitative scoring) were performed locally at each site.

A data monitoring safety board reviewed unblinded efficacy and safety results every 6 months for any trend that would require early termination of the study.

Statistical Analysis.

The sample size of the study was calculated to show a difference in the SVR rate between the full-course and the delayed RBV group of at least 20 percentage points (30% vs. 10%) with a two-tailed α of 0.05 and a β of 0.20 (power, 80%). Under these conditions, 124 patients were required. Assuming a dropout rate of 33%, the study size was set at 200 patients. Categorical variables were compared using χ2 analysis or Fisher exact test as appropriate, and continuous variables were compared by Student's t test. Univariate and multivariate logistic regression analyses were performed to determine association with week 12 viral response and with SVR of the following 12 baseline factors: treatment group (A vs. B), HCV genotype (1 vs. 2 through 4), HCV RNA level (>2.00 million vs. <2.00 million copies/mL), presence of cirrhosis, black race, normal ALT level, age more than 40 years, gender, antiretroviral therapy with zidovudine, therapy with didanosine or stavudine, detectable HIV RNA level, and CD4+ cells less than 200/mm3.

Results

Patient Characteristics.

In November 2000, the data monitoring safety board halted enrollment at 110 patients because of slow patient accrual and a very low likelihood that further enrollment would significantly alter the primary endpoint analysis. Among the 110 patients, 53 were randomized to RBV and 57 to placebo. Patients were well matched between the two groups, as indicated in Tables 1, 2, and 3, except that the placebo group had a slightly higher white blood cell count than the IFN group, but this was not considered to have altered the outcome measures. All major ethnic groups were well represented, with white, Latino, and black patients each representing approximately one third of patients. Most patients had stable HIV disease at baseline, 56% had undetectable plasma HIV RNA, and 90% had CD4+ cell counts of 200/mm3 or more, with a median of 494/mm3. The median baseline level of HCV RNA was 4.55 million copies/mL, and the rate of HCV genotype 1 was 77%, levels similar to those reported from previous HCV monoinfection trials and for the U.S. population.10, 11 Cirrhosis was present on liver histologic analysis in 16% of participants, which is a higher prevalence than was reported in two large trials of IFN and RBV in HCV monoinfection, where it was 4% to 5%.9, 10

Table 1. Baseline Patient Demographics
 IFN Plus Full-Course RBV (n = 53)IFN Plus 16-Week Delayed RBV (n = 53)*All Patients (n = 106)*
  • *

    Demographic and baseline data were missing for one patient in the delayed RBV group.

Median age (range), yrs42 (27–67)41 (29–62)42 (27–67)
Gender, female8 (15%)12 (23%)20 (19%)
Race or ethnicity   
 White19 (36%)19 (36%)38 (36%)
 Latino or Hispanic16 (30%)20 (38%)36 (34%)
 Black15 (28%)12 (23%)27 (25%)
 Other3 (6%)2 (4%)5 (4%)
HIV risk factors (single or multiple)   
 Sexual contact32 (60%)37 (70%)69 (65%)
 Injection drug use29 (55%)24 (45%)53 (50%)
 Blood transfusion5 (9%)3 (6%)8 (8%)
 Unknown2 (4%)1 (2%)3 (3%)
Table 2. Baseline Characteristics of HCV-Infected Patients
 IFN Plus Full-Course RBV (n = 53)IFN Plus 16-Week Delayed RBV (n = 53)*All Patients (n = 106)*
  • NOTE. All parameters were well matched between the randomized groups.

  • *

    Demographic and baseline data were missing for one patient in the delayed RBV group.

  • Abbreviation: ULN, upper limit of normal.

HCV RNA, copies/mL(n = 52)(n = 52)(n = 104)
 Median (range)3.95 M4.85 M4.55 M (7,700–58 M)
 >2.00 M copies/mL, n (%)37 (71%)45 (87%)82 (79%)
HCV genotype(n = 52)(n = 53)(n = 105)
 Type 140 (77%)41 (77%)81 (77%)
 Type 22 (4%)7 (13%)9 (9%)
 Type 310 (19%)4 (8%)14 (13%)
 Type 40 (0%)1 (2%)1 (1%)
ALT(n = 49)(n = 53)(n = 102)
 Value (U/L), mean (range)10595100 (18–362)
 Multiple of ULN, mean (range)1.941.841.89 (0.40–7.54)
 ALT normal (%)12 (24%)9 (17%)21 (21%)
Liver pathologic features   
 No fibrosis8/48 (17%)6/53 (11%)14/101 (14%)
 Cirrhosis7/49 (14%)9/54 (17%)16/103 (16%)
Table 3. Baseline Characteristics of HIV-Infected Patients
 IFN Plus Full-Course RBV (n = 53)IFN Plus 16-Week Delayed RBV (n = 53)*All Patients (n = 106)*
  • *

    Demographic and baseline data were missing for one patient in the delayed RBV group.

  • P = .013 for group A vs. group B. All other parameters were well matched between the randomized groups.

HIV-1 RNA(n = 48)(n = 47)(n = 95)
 Copies/mL, median (maximum)<400<400<400 (514,000)
 Undetectable (%)26 (54%)27 (57%)53 (56%)
CD4+ cells(n = 51)(n = 51)(n = 102)
 Absolute count per mm3, mean (range)458520489 (12–1,195)
 CD4+ cells < 200/mm3, number (%)7 (14%)3 (6%)10 (10%)
 Percent CD4+ cells, mean (range)26%25%25% (3–53%)
 Percent CD4+ < 14%, number (%)9 (18%)3 (6%)12 (12%)
Prior acquired immunodeficiency syndrome-defining illness, number (%)12 (23%)17 (32%)29 (27%)
Hemoglobin g/dL, mean (range)14.714.514.6 (10.8–18.0)
White blood cell count, per mm34,8005,5305,181 (2,100–10,100)
Neutrophil count, per mm32,5002,5562,530 (780–5,187)
Platelet count, 1000 per mm3204190197 (61–456)
Antiretroviral (ARV) therapy(n = 53)(n = 53)(n = 106)
 No ARV therapy9 (17%)8 (15%)17 (16%)
 Nucleoside analog (NA)44 (83%)44 (83%)88 (83%)
  Zidovudine (AZT)15 (28%)13 (25%)28 (26%)
  Stavudine (d4T)26 (49%)31 (58%)57 (54%)
  Didanosine (ddI)7 (13%)4 (8%)11 (10%)
 Protease inhibitor (PI)32 (60%)29 (54%)61 (58%)
 Nonnucleoside reverse transcriptase inhibitor (NNRTI)14 (26%)20 (38%)34 (32%)

Among the patients in group B, all but 3 of 57 randomized subjects required switching to open-label RBV at week 16 because of detectable HCV RNA at week 12. As a result, three different RBV treatment schedules emerged: full-course RBV (group A; n = 53), 16-week delayed RBV (group B-RBV; n = 54), and no RBV (group B-placebo; n = 3). To allow comparison between two groups who each received the same therapy, it was decided to exclude the three patients who were continued on placebo from the efficacy analysis (except for week 12) and from the safety analysis (except during weeks 0 through 16). All other efficacy and safety analyses were performed between patients receiving full-course RBV (group A; n = 53) and those receiving 16-week delayed RBV (group B; n = 54), with a total sample size of 107 subjects. The reduction of the sample size from the originally required 124 to 107 slightly reduced the statistical power to detect a difference in SVR rate, from 80% to 72%.

Viral Response.

In four patients (three in group A and one in group B), HCV RNA was negative at week 56 or 72, but no HCV RNA data were available at week 48 (end of treatment). Because a spontaneous reversion from end-of-treatment HCV viremia to posttreatment week 8 or week 24 undetectable HCV RNA level has never been reported, these four patients were assumed to have had a week-48 viral response. All other viral response rates were calculated on an intention-to-treat basis as mentioned above. At treatment week 12, significantly more patients in the full-course RBV group had an early viral response than those in the delayed RBV group (23% [12/53] vs. 5% [3/57]; P = .011). This difference was significant for genotypes 2 through 4 (P = .014), but not for genotype 1 (P = .20). By contrast, the end-of-treatment response rate was similar between group A (18.9% [10/53]) and group B (7.4% [4/54]; P = .093), and so was the rate of SVR (group A, 11.3% [6/53] vs. group B, 5.6% [3/54]; P = .32). In the full-course RBV group, patients infected with HCV genotype 1 had significantly lower viral response rates at all time points than patients with genotypes 2 through 4: week 12, genotype 1 of 10.0% versus genotypes 2 through 4 of 66.7% (P < .001); week 48 (end-of-treatment response), genotype 1 of 10.0% versus genotypes 2 through 4 of 50.0% (P = .006); and week 72 (SVR), genotype 1 of 2.5% versus genotypes 2 through 4 of 41.7% (P = .002). The efficacy results are displayed Fig. 1. The three patients in group B with a week 12 viral response who were continued on placebo are described here separately. The first patient, a 31-year-old man (genotype 3) had a viral breakthrough on treatment and his HCV RNA levels at weeks 48 and 72 were 2.50 M and 4.20 M copies/mL, respectively. The second patient, a 38-year-old woman (genotype 2), had a week-48 viral response, but she did not come for follow-up HCV RNA testing, and her outcome was considered response relapse. The third patient, a 36-year-old man (genotype 1), dropped out of the study after week 12, and no further HCV RNA levels were available. He was considered a viral nonresponder.

Figure 1.

Viral response rates at weeks 12, 48, and 72. Three patients with a viral response at treatment week 12 who were continued on placebo and not switched to RBV were excluded from the efficacy analysis of weeks 48 and 72. end Rx = end of treatment; end F/U = end of follow-up.

Univariate analysis identified 3 of the 12 examined factors to be associated with a week 12 viral response: HCV genotype 1 versus 2 through 4 (odds ratio [OR], 0.1; 95% CI, 0.03– 0.4; P < .001), treatment group A versus group B (OR, 5.3; 95% CI, 1.4–20.0; P = .011), and black versus nonblack race (0% vs. 18%; OR, 0.0; P = .020). However, in multivariable logistic regression analysis, only treatment group (A vs. B; OR, 0.09; 95% CI, 0.02–0.4; P = .003) and genotype (1 vs. 2 through 4; OR, 15.9; 95% CI, 2.5–100; P = .003) were independently associated with week 12 viral response. In the univariate analysis of SVR, only HCV genotype was significantly associated with an SVR (1 vs. 2 through 4; OR, 0.07; 95% CI, 0.01–0.35; P < .001). A multivariable logistic regression analysis was precluded by the small number of subjects (n = 9) with a primary endpoint of SVR.

Safety.

Fifty-five patients (51%) discontinued the study drugs before completing 48 weeks of therapy, of whom 23 (21%) discontinued because of adverse events, mainly depression, anxiety, anemia, and fatigue. Other reasons for early treatment discontinuation were a decision by the investigator to stop study treatment prematurely because of perceived insufficient viral response during treatment in 11%, patient choice or lack of follow-up in 12%, and relapse to substance use in 4%. Dose reductions, including temporary dose discontinuations, of IFN occurred in 23 patients (22%), mainly the result of neutropenia, neuropsychiatric symptoms, or fatigue. The dose of RBV was reduced in 36 patients (34%), most commonly because of anemia (19%). The incidence of anemia-related RBV dose reductions was higher in group A (28%) than in group B (11%; P = .032). In group A (but not group B), coadministration of zidovudine (azidothymidine [AZT]) led to higher rates of RBV dose reductions, both for all causes (AZT, 67% vs. no AZT, 24%; P = .003) and for anemia (AZT, 60% vs. no AZT, 16%; P = .001). The results are displayed in Table 4.

Table 4. Safety Analysis: Early Treatment Discontinuations and Dose Reductions (Including Temporary Dose Discontinuations) With Reasons With and Without Coadministration of Zidovudine
 IFN Plus Full-Course RBV (n = 53)IFN Plus 16-Week Delayed RBV (n = 54)P ValueAll Patients (n = 107)
  • *

    Patients may have had more than one adverse event associated with a dose reduction.

  • P < .01.

Early treatment discontinuation, n (%)25 (47%)30 (56%).4455 (51%)
 Adverse event13 (25%)10 (19%).4923 (21%)
  Depression34 7 (7%)
  Anxiety30 3 (3%)
  Anemia3 (5.7%)0 3 (3%)
  Fatigue03 3 (3%)
  Others43 7 (7%)
 Physician decision: viral nonresponse4 (8%)8 (15%).3612 (11%)
 Lost to follow-up or noncompliant3 (6%)5 (9%).728 (7%)
 Patient decision2 (4%)3 (6%)1.005 (5%)
 Substance abuse relapse2 (4%)2 (4%)1.004 (4%)
 Other reasons1 (2%)2 (4%)1.003 (3%)
Dose reduction of IFN*, n (%) [AZT, no AZT]8 (15%) [20%, 13%]15 (28%) [31%, 28%].1623 (22%) [25%, 21%]
 Neutropenia4 (8%) [7%, 8%]4 (7%) [15%, 5%]1.008 (8%) [11%, 6%]
 Neuropsychiatric3 (6%)4 (7%)1.007 (7%)
 Fatigue2 (4%)4 (7%).686 (6%)
 Headache1 (2%)3 (6%)1.004 (4%)
 Anemia2 (4%)2 (4%)1.004 (4%)
Dose reduction of RBV*, n (%) [AZT; no AZT]19 (36%) [67%, 24%]17 (31%) [46%, 28%].6936 (34%) [57%, 26%]
 Anemia15 (28%) [60%, 16%]6 (11%) [23%, 8%].03221 (20%) [41%, 11%]
 Hyperbilirubinemia6 (11%)3 (6%).329 (8%)
 Fatigue2 (4%)5 (9%).447 (7%)
 Neutropenia2 (4%)3 (6%)1.005 (5%)

There were no deaths in this trial. There was one clinical serious adverse event of mild acute pancreatitis with hospitalization on treatment day 14 in a female patient who had been assigned to IFN plus placebo and also was receiving stavudine, nevirapine, and saquinavir for HIV therapy. Within 2 days, all symptoms resolved and lipase levels normalized. The study drugs were resumed without further complication. The remaining 7 cases of elevated lipase levels more than 2.5 times the upper limit of normal were asymptomatic. The most common clinical adverse events were fatigue, headache, depression, fever, and myalgia. Only a few of them were considered to be grade III (severe) or grade IV (life threatening), the most common of them depression (5/107; 5%). Table 5 lists all adverse events with a frequency of 5% or more.

Table 5. Safety Analysis: Clinical Adverse Events and Laboratory Abnormalities With ≥5% Frequency, Total, Grade III or IV, Related to Study Drugs and Those That Led to Dose Reductions or Discontinuations
Adverse EventAny Adverse Event, n (%), Total [n (%) Grade III–IV Adverse Event]Adverse Event Related to Study Drugs, n (%) Total–n (%) Dose Change
Group A (n = 53)Group B (n = 54)Group A (n = 53)Group B (n = 54)
  1. Abbreviations: Hgb, hemoglobin level; ULN, upper limit of normal.

Clinical adverse event    
 Weakness/fatigue10 (19%) [0 (0%)]15 (28%) [2 (4%)]9 (17%)–4 (8%)14 (26%)–7 (13%)
 Headache9 (17%) [1 (2%)]12 (22%) [2 (4%)]9 (17%)–4 (8%)11 (20%)–5 (9%)
 Depression7 (13%) [2 (4%)]10 (19%) [3 (6%)]6 (11%)–3 (6%)10 (19%)–6 (11%)
 Fever2 (4%) [0 (0%)]7 (13%) [1 (2%)]1 (2%)–(2%)6 (11%)–1 (2%)
 Myalgia7 (13%) [2 (4%)]3 (6%) [1 (2%)]7 (13%)–3 (6%)3 (6%)–2 (4%)
 Nervousness5 (9%) [0 (0%)]3 (6%) [0 (0%)]5 (9%)–3 (6%)3 (6%)–2 (4%)
 Diarrhea4 (8%) [0 (0%)]3 (6%) [1 (2%)]4 (8%)–1 (2%)3 (6%)–2 (4%)
 Nausea3 (6%) [0 (0%)]4 (7%) [1 (2%)]3 (6%)–1 (2%)4 (7%)–3 (6%)
 Rash2 (4%) [0 (0%)]4 (7%) [0 (0%)]1 (2%)–1 (2%)3 (6%)–1 (2%)
 Insomnia2 (4%) [0 (0%)]3 (6%) [0 (0%)]2 (4%)–2 (4%)3 (6%)–0 (0%)
 Chills2 (4%) [0 (0%)]2 (4%) [0 (0%)]2 (4%)–2 (4%)2 (4%)–1 (2%)
 Anxiety3 (6%) [0 (0%)]1 (2%) [0 (0%)]3 (6%)–3 (6%)1 (2%)–1 (2%)
Laboratory abnormalities [value for grade III–IV]    
 Anemia [Hgb < 7.0 g/dL]17 (32%) [1 (2%)]9 (17%) [1 (2%)]17 (32%)–8 (14%)8 (15%)–7 (13%)
 Neutropenia [<750/mm3]17 (32%) [7 (13%)]14 (26%) [7 (13%)]15 (28%)–4 (8%)13 (24%)–4 (7%)
 Hyperbilirubinemia [>2.5 × ULN]14 (26%) [4 (8%)]10 (19%) [3 (6%)]14 (26%)–6 (11%)8 (15%)–2 (4%)
 Hypertriglyceridemia [>750 mg/dL]10 (19%) [4 (8%)]16 (30%) [5 (9%)]6 (11%)–2 (4%)10 (19%)–3 (6%)
 Amylase increase [>2.5 × ULN]9 (17%) [5 (9%)]5 (9%) [2 (4%)]6 (11%)–1 (2%)4 (7%)–0 (0%)
 Lipase increase [>2.5 × ULN]7 (13%) [5 (9%)]6 (11%) [3 (6%)]5 (9%)–1 (2%)4 (7%)–1 (2%)
 Hypophosphatemia [<1.5 mg/dL]9 (17%) [2 (4%)]4 (7%) [2 (4%)]6 (11%)–0 (0%)3 (6%)–1 (2%)

After reports of lactic acidosis and other mitochondrial toxicity associated with concurrent use of RBV and didanosine (ddI),12, 13 a post hoc safety analysis was performed on behalf of the U.S. Food and Drug Administration. It examined hepatic and pancreatic adverse events in patients receiving RBV and either ddI (n = 16) or no ddI (n = 84). The adverse events were defined as elevated amylase and lipase levels (pancreatic events, including the one case of pancreatitis) as well as elevated levels of ALT, aspartate aminotransferase, bilirubin, and gamma-glutamyltransfrase (hepatic events). Pancreatic events occurred in 31% of patients receiving ddI and in 16% of patients not receiving ddI (P = .16). Hepatic events occurred in 63% of patients receiving ddI and in 47% of patients not receiving ddI (P = .28). Thus, no difference in hepatic or pancreatic events was seen in patients receiving RBV with or without concurrent therapy with ddI. There was no incidence of symptomatic lactic acidosis during the study. Lactic acid levels were not obtained routinely as part of the study protocol.

Hematologic toxicity was assessed within the first 16 weeks, when the double-blind phase allowed a comparison between RBV (group A) and placebo (group B) in combination with IFN. The RBV group had a significantly lower nadir of hemoglobin level than the placebo group (12.1 vs. 13.6 g/dL) and a more pronounced change of hemoglobin from baseline (−2.52 vs. −1.04 g/dL). In 20% of patients receiving RBV, a hemoglobin decrease to less than 10.0 g/dL compared with 2% of those receiving placebo, and in 35% of patients receiving RBV, the hemoglobin level dropped by 3.0 g/dL or more compared with no patients in the placebo group. Patients in the RBV group who were also taking AZT for HIV infection, compared with patients not taking AZT, had a lower hemoglobin nadir (AZT, 10.1 g/dL vs. no AZT, 13.0 g/dL; P < .001) and a more pronounced decrease in hemoglobin concentration (AZT, −3.64 g/dL vs. no AZT, −2.08 g/dL; P = .002). Other antiretroviral drugs had no effect on anemia or other hematologic toxicity. The mean nadir leukocyte count was lower in the RBV group (3,290/mm3) than in the placebo group (4,050/mm3), and this was caused by a more pronounced lymphopenia (RBV, 1,160/mm3 vs. placebo, 1,920/mm3) and not by neutropenia. Concurrent treatment with AZT was not associated with more pronounced leukopenia or neutropenia. The results are detailed in Table 6. Recombinant human erythropoietin was used for the treatment of anemia in six patients in group A and in one patient in group B, and recombinant granulocyte colony-stimulating factor was used in seven patients whose neutrophil counts at start of recombinant granulocyte colony-stimulating factor therapy ranged from 540 to 890/mm3.

Table 6. Hematology Safety Profile During Treatment Weeks 0 to 16 and HIV Disease Markers Between Treatment Weeks 0 and 12
Change Between Week 0 (Baseline) and Week 16 [Week 0 and Week 12 for HIV Markers]IFN Plus RBVIFN Plus PlaceboP Value
  • *

    P < .05.

  • P < .01.

  • P < .001.

  • §

    P < .01 from baseline value.

  • P < .001 from baseline value.

Hemoglobin level (g/dL), mean, all patients [AZT; no AZT]   
 Nadir12.1 [10.1; 13.0]13.6 [13.4; 13.7]<.001
 Change baseline to nadir−2.52 [−3.64; −2.08]−1.04 [−1.31; −0.93]<.001
 Nadir < 10.0 g/dL (%)20% [67%; 0%]2% [7%; 0%].003
 Drop baseline to nadir > 3.0 g/dL (%)35% [69%; 21%]0% [0%; 0%]<.001
Leukocyte count (per mm3)   
 Nadir, mean3,2904,050<.001
Neutrophil count (per mm3)   
 Nadir, mean1,4301,580.27
 Nadir < 750/mm3 (%)5%14%.17
Lymphocyte count (per mm3)   
 Nadir, mean1,1601,920.005
 Nadir < 1,000/mm3 (%)65%14%<.001
Platelet count (1,000 per mm3), all patients [AZT; no AZT]   
 Nadir, mean170 [200; 157]*153 [155; 152].11
 Nadir < 50,000/mm3 (%)0%2%1.00
Mean change in absolute CD4+ cells per mm3−85§−35.15
% Mean change in CD4++4.1%−0.3%.001
Change in HIV-1 viral load   
 Log10 change, mean−0.175+0.008.36
 Undetectable week 0 to detectable week 12, total 3/36 (8%)2/16 (13%)1/20 (5%).57
 Detectable week 0 to undetectable week 12, total 8/36 (22%)§4/17 (24%)4/19 (21%)1.00

The effect of HCV therapy on markers of HIV disease was assessed comparing CD4+ cells as well as HIV viral load at week 12 with baseline. The absolute CD4+ cell count dropped significantly in the RBV group (−85/mm3; P = .001), but not in the placebo group (−35/mm3; P = .15). By contrast, the percentage of CD4+ cells among all lymphocytes rose by 4.1 percentage points (P < .001) in the RBV group compared with an insignificant decrease by 0.3 percentage points (P = .75) in the placebo group (A vs. B; P = .001). Among 36 patients with undetectable HIV viral load at baseline for whom paired HIV viral loads were available, only 3 (8%; P = NS vs. baseline) had a viral breakthrough to detectable plasma HIV RNA at week 12. By contrast, a significant proportion of patients with HIV viremia at baseline became undetectable for HIV RNA by week 12 (22% [8/36]; P = .005 vs. baseline). There was no difference between the treatment groups (24% vs. 21%).

Discussion

The double-blind phase of this trial showed that in coinfection with both HIV and HCV, as in HCV monoinfection, the addition of RBV to IFN is essential to achieve a viral response. The week 12 response was only 5% with IFN alone, but 23% with combination therapy. The rate of SVR for patients coinfected with HIV and HCV who received 48 weeks of IFN plus RBV was 11%. This was higher than for patients who received delayed RBV (6%), but the difference did not reach significance. Although the SVR rate of 11% is low and the early discontinuation rate of 51% is high, the rates are comparable with other prospective HIV and HCV coinfection trials with the same regimen. A randomized United States-wide coinfection study (genotype 1 in 77%) found a similar SVR rate of 9% in 74 patients receiving IFN 3 MU three times weekly plus RBV 800 mg/day for 48 weeks. The early discontinuation rate was 86%.14 In an uncontrolled French study, Landau et al. treated 51 patients coinfected with HIV and HCV with IFN at 3 MU three times weekly and a higher RBV dose of 1,000 to 1,200 mg/day for 48 weeks. Twenty-nine percent of patients discontinued treatment prematurely, and an SVR was achieved in 9% of patients with genotype 1 and 47% of patients with genotypes 2 or 3, rates similar to the present study.15

In contrast to this HIV and HCV coinfection study, previous treatment trials of HCV monoinfection with the same IFN dose combined with a higher RBV dose of 1,000 to 1,200 mg/day had a much higher SVR rate than this study, with overall SVR rates of 38% and 43% (29% for genotype 1 and 66% for genotypes 2 or 3) and early discontinuation rates of only 21% and 19%.9, 10, 16 Several factors could explain these differences in viral response. First, in the present study, there was a high rate of treatment discontinuation, many related to anemia. Our study had a rate of treatment discontinuation as a result of anemia of 5.7% (group A) compared with 0.4%10 and 0.7%9 in HCV monoinfection. Second, RBV dose reductions for anemia were much more frequent in this study (28%) compared with the monoinfection trials (9% and 7%). The higher rate of anemia from RBV in the coinfected patients occurred despite the lower starting dose of 800 mg/day and the reduced hemoglobin threshold to reduce RBV doses (9 g/dL), and despite the fact that HIV infection was well controlled in this population with a mean baseline hemoglobin level of 14.7 g/dL. Given a lower starting dose and more RBV dose reductions, patients in the full-course RBV arm cumulatively received lower RBV doses than in the two monoinfection trials. Of interest, in Landau et al's study, where the higher RBV dose of 1,000 to 1,200 mg/day was used, the rate of RBV dose reduction as a result of anemia was even higher (35%) than in the present study.15 Higher RBV doses have been associated with better SVR rates in combination with pegylated interferon.17 A third factor that may explain the lower SVR rate in this coinfection study compared with HCV monoinfection trials is the higher rate of cirrhosis (14% vs. 4% to 5%). Cirrhosis has been associated with lower rates of SVR.9, 18, 19

Antiretroviral treatment with AZT (the only HIV drug with a myelosuppressive effect) seems to play a significant role in the development of anemia during HCV combination therapy, an issue not previously emphasized. IFN9 and AZT20 each can cause mild anemia because of their bone marrow suppressive properties, whereas RBV monotherapy21, 22 leads to mild anemia because of dose-dependent acute hemolysis. The combination of IFN and RBV causes more profound anemia because of the impaired ability of the suppressed bone marrow to compensate for red blood cell loss from RBV-induced hemolytic anemia.9, 23 The observation in our study that anemia is more pronounced during therapy with IFN plus RBV when the patient is also taking AZT suggests that there is a cumulative myelosuppressive effect of IFN plus AZT that further reduces erythropoiesis that could compensate for the acute RBV-induced hemolysis. This directly translated to more frequent RBV dose reductions for anemia in patients receiving full-course RBV who were also taking AZT. AZT did not make a difference in anemia in patients who were receiving IFN alone, where no hemolysis needed to be compensated. A cumulative myelosuppressive effect of IFN plus zidovudine was not seen for neutrophils, suggesting that white cell precursors have a lower susceptibility to this combined effect than erythroid progenitor cells.

In the double-blind phase, patients receiving RBV had a significantly lower lymphocyte count compared with placebo patients, independent of zidovudine treatment. This phenomenon had already been observed in placebo-controlled trials of RBV monotherapy for chronic hepatitis C.21, 22 Although this decrease in absolute lymphocyte count did not have any clinical consequences, it led to an (artificial) drop in absolute CD4+ cell count, which is calculated as the number of lymphocytes times the percentage of CD4+ cells within lymphocytes. Indeed, the percentage of CD4+ cells, a more stable measure of cellular immune function in HIV-infected individuals, rose by 4.1 percentage points by week 12 in the IFN plus RBV group. Why the same rise in CD4+ percentage was not observed in the IFN plus placebo group is unclear, but an immunomodulatory effect of RBV24 or an antiretroviral effect of RBV through inhibition of HIV reverse transcriptase25, 26 are possible causes. There was no significant change in HIV viral load during IFN treatment with or without RBV. However, a significant number of patients with HIV viremia at baseline (22%) had undetectable plasma HIV RNA after 12 weeks of IFN treatment, again with or without RBV. This may be the result of the anti-HIV activity of IFN.27

Concerns about antagonism between RBV and the antiretroviral drugs zidovudine and stavudine have been raised because in vitro RBV inhibits conversion of zidovudine28 and stavudine29 to their active triphosphate forms through competitive inhibition of intracellular phosphorylation enzymes. This study, however, demonstrates that treatment with RBV does not have a negative effect on therapeutic control of HIV infection and may indeed favorably influence cellular immunity by raising the CD4+ percentage. In a recent study, 30 patients coinfected with HIV and HCV receiving stavudine therapy were randomized to therapy with IFN plus RBV or no treatment. Patients receiving active IFN plus RBV treatment had no change in HIV viral load at 3 months, even though these patients did have a nonsignificant trend for decreased intracellular stavudine-triphosphate concentrations compared with placebo.30 From those findings and our study, we conclude that treatment with RBV does not interfere with antiretroviral effectiveness in HIV infection, even when zidovudine or stavudine are part of the regimen.

Clinical adverse events generally were similar to the ones that clinicians expect during treatment with IFN and RBV in HCV monoinfection, and no new types of toxicity were observed. Symptomatic lactic acidosis was not observed, and one case of mild pancreatitis involved neither RBV nor didanosine treatment.

As HCV clinicians become more experienced with the management of HCV treatment-related toxicity, they discontinue treatment less frequently and manage more side effects with dose reductions and other interventions, which in turn may raise the SVR rate. This phenomenon was observed in the comparison of two trials using IFN plus RBV for HCV monoinfection, one reported in 199810 and the other from 2001.19 The early discontinuation rate dropped from 21% in 1998 to 13% in 2001, and the SVR rate for genotype 1 rose from 28% in 1998 to 33% in 2001. In addition to better management of side effects, combination therapy of pegylated IFN plus RBV gives hope for better viral response rates. This has already been shown in patients infected with HCV only,19, 22 and several randomized, controlled trials are ongoing that test this combination in patients coinfected with HIV and HCV. We conclude that combination therapy of IFN plus RBV can be given safely to patients coinfected with HIV and HCV patients, and that higher rates of sustained viral response than the current 11% are likely to be achieved in the future. Anemia seems to be an important limiting factor for the success of combination therapy in patients coinfected with HIV and HCV.

Appendix

Study Sites and Investigators.

AIDS Healthcare Foundation, Los Angeles, CA (Dale Prokupek, MD); Arizona Clinical Research Center, Tucson, AZ (Kelly Vollmer, MD); Bentley Salick Health, New York, NY (Ramon Torres, MD); Bronx Veterans Affairs Medical Center, Bronx, NY (Norbert Bräu, MD); CARE Resource, Coral Gables, FL (Allan Stein, DO); Columbia University Liver Center, New York, NY (Robert Brown, MD); Denver Public Health, Denver, CO (William Burman, MD); Digestive Health Consultants, Burbank, CA (David Schulman, MD); Emory University, Atlanta, GA (Christine Bruno, MD); Fundacion Gastroenterologia de Diego, Santurce, Puerto Rico (Maribel Rodriguez-Torres, MD); Hershey Medical Center, Hershey, PA (M. Elaine Eyster, MD); Internal Medicine Associates, Miami, FL (Barry Brand, MD); Pacific Oaks Medical Group, Beverly Hills, CA (Dale Prokupek, MD); Philadelphia FIGHT, Philadelphia, PA (Jay Kostman, MD); Research and Education Group, Portland, OR (Thomas Ward, MD); Roger Williams Medical Center, Providence, RI (Gail Skowron, MD); St. Michael's Medical Center, Newark, NJ (Jihad Slim, MD); University of California at Davis, Sacramento, CA (Jerry Powell, MD); University of California, San Diego, CA (Tarek Hassanein, MD); University of North Carolina, Chapel Hill, NC (Michael Fried, MD); University of Southern California, Los Angeles, CA (Maurizio Bonacini, MD).

Ancillary