SATHEESH NAIR AND ROBERT P.PERRILLO

Transcription

1 Serum Alanine Aminotransferase Flares During Interferon Treatment of Chronic Hepatitis B: Is Sustained Clearance of HBV DNA Dependent on Levels of Pretreatment Viremia? SATHEESH NAIR AND ROBERT P.PERRILLO During interferon treatment of chronic hepatitis B, an alanine aminotransferase (ALT) flare may herald a sustained loss of viral replication, but the relationship between virologic response, the extent of a flare, and pretreatment hepatitis B virus (HBV) DNA level has not been defined. We retrospectively examined the impact of an ALT flare on sustained virologic response in 121 interferon-treated patients and 42 untreated controls with either low-level (<100 pg/ ml) or high-level (>100 pg/ml) viremia. The degree of ALT flare was classified as mild (increase in ALT of IU/L above baseline), moderate (increase of 172 to 343 IU/L above baseline), and severe (increase of >344 IU/L above baseline). Undetectable serum HBV DNA and hepatitis B e antigen (HBeAg) loss were significantly more likely at the end of follow-up in patients having a flare (P.0001 and.001, respectively). In the high viremia group, a proportionate increase in virologic response was observed as the degree of flare increased. By multivariate analysis, high baseline HBV DNA, high pretreatment ALT, and both moderate and severe ALT flare were independently predictive of a virologic response with severe flare being the most powerful predictor for a sustained loss of serum HBV DNA (odds ratio, 5.3; P.004). Severe flare was predictive of a virologic response in the high but not low viremia group. We conclude that a virologic response in patients with highlevel viremia is dependent on the degree of ALT flare. Induction of a robust flare may enhance virologic response when high-level viremia is detected. (HEPATOLOGY 2001;34: ) Interferon and lamivudine are the only drugs that have been specifically licensed as treatment for chronic hepatitis B at the current time. While the virologic, biochemical, and serologic responses to treatment are similar with these agents, only interferon has been proven to have a direct immunostimulatory effect. Most studies have shown that the response rate to interferon is greater among patients with low-level viremia Abbreviations: ALT, alanine aminotransferase; HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; ULN, upper limit of normal; CI, confidence interval. From the Section of Gastroenterology and Hepatology, Ochsner Clinic, New Orleans, LA. Received June 6, 2001; accepted August 6, Address reprint requests to: Robert P. Perrillo, M.D., Ochner Clinic, 1514 Jefferson Highway, New Orleans, LA rperrillo@ochsner.org; fax: Copyright 2001 by the American Association for the Study of Liver Diseases /01/ $35.00/0 doi: /jhep and high levels of alanine aminotransferase (ALT) pretreatment, which reaffirms the importance of the immune response in reducing the level of hepatitis B virus (HBV) replication. 1,2 Flares occur in 25% to 40% of patients who are treated with interferon and are considered to be reflective of an augmented cell-mediated immune response to HBV. 3,4 Although ALT flares are considered to be a favorable indicator that a response is likely, they have not been shown to be independently predictive of a response, and the degree of flare has not been specifically studied in relationship to the degree of pretreatment viremia. These are important considerations because a high level of viremia may require a brisk immune response to HBV as reflected by a greater degree of ALT increase during treatment. By contrast, the response rate in patients with low-level viremia could be independent of such an effect. To test this hypothesis further, we analyzed the relationship between ALT flare and pretreatment viremia in a large cohort of patients with chronic hepatitis B who were treated with interferon between 1986 and Particular attention was given to the correlation between the degree of flare and the frequency of a virologic response. PATIENTS AND METHODS The data were extracted from the patient records of a randomized, placebo-controlled multicenter trial conducted from 1986 to 1990, which compared the relative efficacy of different doses of interferon to a combination of interferon and prednisone in the treatment of chronic hepatitis B. 1 This trial involved 3 treatment arms and 1 untreated control arm. The active treatment arms were as follows: (A) 6 weeks of prednisone in decreasing doses of 60 mg, 40 mg, and 20 mg each for 2 weeks followed in turn by a 2-week interval off treatment and 16 weeks of recombinant interferon alfa-2b given in a dose of 5 MU/d; (B) a 6-week course of a matching oral placebo followed by a 2-week interval off treatment and interferon in a dose of 5 MU/d for 16 weeks; or (C) a 6-week course of placebo followed by a 2-week interval off treatment and 16 weeks of interferon in a dose of 1 MU/d. One hundred sixty-nine patients were enrolled in the study. All patients met the following criteria for inclusion: age of 18 years or older, presence of hepatitis B surface antigen (HBsAg) for at least 6 months, positive tests for hepatitis B e antigen (HBeAg) and HBV DNA documented on 3 or more occasions at least 1 month apart during 6 months before entry, elevated serum levels of ALT on at least 3 occasions before entry with an average value 1.3 times the upper limit of normal (ULN), compensated liver disease (prothrombin time less than 3 seconds prolonged, serum albumin 30 g/l, bilirubin levels 2 mg/dl [34 mol/l], and evidence of chronic hepatitis B on liver biopsy). All patients were seen at 2-week intervals during the initial phase of prednisone or placebo treatment; after 1, 2, 4, 8, 12, and 16 weeks of interferon therapy and again 1, 3, and 6 months after treatment. At each visit, serum samples were obtained for measurements of serum ALT, HBV DNA (solution hybridization

2 1022 NAIR AND PERRILLO HEPATOLOGY November 2001 TABLE 1. Comparison Between Patients With Different Levels of Pretreatment Viremia HBV DNA (<100 pg/ml) (n 78) HBV DNA (>100 pg/ml) (n 85) P Median HBV DNA, pg/ml (range) 54 ( ) 171 ( ).0001 Sustained undetectable HBV DNA 50% 22%.0001 Loss of HBeAg/HBV DNA 39% 17%.001 Pretreatment ALT, IU/L (range) 154 (60-605) 132 (54-468) NS Pretreatment ALT ( 4 ULN, %)* 44% 32% NS Peak ALT, IU/L (range) 260 ( ) 227 ( ) NS Flare 48% 47% NS Treatment Untreated control (n 42) Low dose (n 41) NS High dose (n 38) NS Prednisone interferon (n 42) NS NOTE. All values are expressed as median. Abbreviation: NS, not significant. *Pretreatment ALT level of 172 IU/L. Flare defined as any increase in ALT 86 IU/L above baseline value. Compared with untreated controls. assay; Abbott Laboratories, North Chicago, IL), and HBeAg by enzyme-linked immunosorbent assay (ELISA; Abbott Laboratories). Hepatitis B surface antigen was evaluated at the end of treatment as well as at the end of a 6-month posttreatment follow-up period (ELISA, Abbott Laboratories). In the present study, we divided the patients into 2 groups based on mean pretreatment HBV DNA levels as follows: patients with low viremia (HBV DNA 100 pg/ml) and those with high viremia (HBV DNA 100 pg/ml). The decision to use a 100 pg/ml cutoff was based on considerations of sample size and the observation that serum HBV DNA levels above this value have been shown to correlate with a diminished response to interferon alfa. 1 A pretreatment ALT value was calculated as the mean of multiple observations during the 6-month pre-enrollment period (ULN 43 IU/L). Similarly, the pretreatment HBV DNA was a mean of HBV DNA levels obtained during the pre-enrollment period. Peak ALT was defined as the highest ALT level at any time during the course of treatment. To qualify as a flare, ALT values had to increase 86 IU/L above baseline at any time during treatment. The degree of flare was arbitrarily graded as mild (increase of IU/L above baseline which is equal to an increase of approximately 2-4 ULN), moderate ( IU/L above baseline, equal to an increase of 4 to 8 ULN), and severe ( 344 IU/L above baseline, equal to greater than 8 ULN) when compared with pretreatment values. A virologic response to treatment was defined as the disappearance of HBV DNA, either with or without HBeAg clearance, that was sustained during a 6-month posttreatment observation period. Loss of HBeAg and disappearance of HBsAg were also evaluated at the same time points. Statistical Analysis. Pretreatment ALT was considered as a continuous as well as a dichotomous variable (values above 4 ULN (i.e., 172 U/L) and values less than 4 ULN). Likewise pretreatment HBV DNA was also analyzed as a continuous variable and dichotomous variable ( 100 pg/ml and 100 pg/ml). Because the continuous variables (pretreatment HBV DNA and pretreatment ALT) were not normally distributed, nonparametric tests were used for data analysis. Continuous measurements were analyzed by using the Mann-Whitney U test, and categorical variables were compared by using Fisher s exact test. The relationship between the degree of flare and treatment response in patients with both high and low levels of viremia was assessed, as was the relationship between ALT flare and loss of HBsAg. In the univariate analysis of predictors of response, pretreatment ALT (as a dichotomous variable), pretreatment HBV DNA (also as a dichotomous variable), degree of flare, and treatment groups were compared between responders and nonresponders. The variables that were found to be significant in the univariate analysis were used in a multiple logistic regression model to determine whether they were independently predictive of a sustained virologic response that was defined as either a sustained lack of detection of serum HBV DNA alone or in combination with disappearance of HBeAg. Because the data were not normally distributed, the results were expressed as median values with minimum and maximum values. A P value of.05 was considered significant for all analyses. The statistical software used was SPSS (SPSS for windows, version 10.0; Chicago, IL). RESULTS The treatment groups were not significantly different with respect to age, sex, and duration of hepatitis. 1 Six of the original 169 patients were excluded from the present study because of insufficient follow-up or disappearance of HBV DNA before initiation of interferon. Among the remaining patients (42 in the prednisone/interferon group, 38 in the high-dose interferon group, 41 in the low-dose interferon group, and 42 patients in the control group), 58 patients (36%) had sustained disappearance of HBV DNA during therapy and the 6-month follow-up interval, and 44 patients (27%) had both disappearance of HBV DNA and HBeAg. Ten patients (6%) had undetectable HBsAg by the end of the follow-up period. Seventy-eight patients (48%) were classified as having lowlevel viremia (pretreatment HBV DNA levels of 100 pg/ml), and 85 patients (52%) had high-level viremia (HBV DNA levels 100 pg/ml) (Table 1). The rate of sustained loss of serum HBV DNA was significantly greater in the low viremia group (50% vs. 22%; P.0001). Sustained loss of both HBeAg and HBV DNA was also higher in patients with low pretreatment HBV DNA (39% vs. 17%; P.001). The mean pretreatment ALT, peak ALT, and frequency of flare did not differ between patients with low- and high-level viremia (Table 1). The treatment groups were also evenly matched with respect to lowand high-level viremia, and none of the active treatment groups significantly differed from controls with respect to levels of viremia. Seventy-eight patients (48%) had a flare (increase in ALT of 86 IU/L from the pretreatment ALT values) during the course of treatment (Table 2). The range of values for flares

3 HEPATOLOGY Vol. 34, No. 5, 2001 NAIR AND PERRILLO 1023 TABLE 2. Comparison Between Patients With and Without a Flare During Treatment Flare (n 78) No Flare (n 85) P Pretreatment ALT, IU/L (range) 145 (60-544) 136 (54-605) NS Pretreatment ALT ( 4 ULN)* 44% 32% NS Pretreatment HBV DNA, pg/ml (range) 101 ( ) 101 (47-532) NS Sustained loss of HBV DNA (%) Loss of HBeAg/HBV DNA (%) Loss of HBsAg (%) Treatment Untreated control (n 42) 11 (26%) 31 (74%) Low dose (n 41) 15 (49%) 26 (51%).04 High dose (n 38) 14 (45%) 24 (55%) NS Prednisone interferon (n 42) 25 (71%) 17 (29%).0001 NOTE. All continuous variables are expressed as median values. Flare is defined as an increase in ALT 86 IU/L above baseline value. Abbreviation: NS, not significant. *Pretreatment ALT greater than 4 ULN (i.e., 172 IU/L). Compared with untreated controls. ranged from 86 to 2,235 IU/L. Neither pretreatment ALT nor HBV DNA levels were significantly associated with a flare. The rate of virologic response (evaluated either as sustained disappearance of HBV DNA alone or together with loss of HBeAg) was higher among patients who had a flare (50% vs. 22%; P.0001, 40% vs. 15%, P.001, respectively; Table 2). A flare was also more frequently observed in patients with HBsAg clearance (10% vs. 2% for no flare), but this was of borderline significance (P.049). Seventy-one percent of patients receiving prednisone had a flare compared with 26% of controls (P.001, Table 2). Interferon monotherapy also resulted in a higher frequency of flare although to a lesser extent than in the prednisone pretreated patients, but this was only significantly different from controls in the low-dose interferon group. The response rate was assessed according to the degree of flare in patients with high- and low-level viremia (Fig. 1A and B). In patients with high-level viremia, a proportional increase in virologic response rate was observed as the degree of flare increased from mild to severe (Fig. 1A). In this group, 50% of patients with a severe flare showed a sustained loss of HBV DNA and 34% lost HBeAg compared with less than 10% of those without a flare (P.002 and.01, respectively). In contrast to these findings, the low viremic group did not show a clear-cut association between the degree of ALT flare and virologic response, with the only exception being a higher rate of HBeAg clearance in patients having a moderate flare when compared with individuals without a flare. Univariate analysis (Table 3) showed that pretreatment ALT, HBV DNA, degree of flare, and treatments with highdose interferon, alone or in combination with prednisone, were associated with a sustained loss of HBV DNA. In a multiple logistic regression model involving all 163 patients, high pretreatment HBV DNA ( 100 pg/ml; odds ratio, 0.267; 95% confidence interval [CI], ; P.001), pretreatment ALT ( 4 ULN; odds ratio, 3.76; 95% CI, ; P.001), and degree of flare (moderate and severe) were inde- FIG. 1. (A) Virologic response in patients with high-level viremia according to the degree of flare. There is a proportional increase in response rate as the degree of flare increases. All comparisons are with no flare group. *P.03; P.04; P.002; P.01. (B) Virologic response in patients with low-level viremia according to the degree of flare. There is no statistical relationship between the degree of flare and sustained loss of serum HBV DNA. *Moderate flare is associated with a greater rate of HBeAg loss, P.01. TABLE 3. Univariate Analysis of Factors Affecting Virologic Response Responders (n 58) Nonresponders (n 105) P Pretreatment ALT ( 4 ULN, %)* 57% 27%.001 HBV DNA 100/ 100 pg/ml 50%/22% 50%/78%.001 Flare No flare (n 85) 19 (22%) 66 (78%) Mild flare (n 25) 9 (37%) 16 (65%) NS Moderate flare (n 29) 16 (55%) 13 (45%).002 Severe flare (n 24) 14 (58%) 10 (42%).002 Treatment Untreated control 10 (24%) 32 Low dose 11 (27%) 30 NS High dose 18 (47%) Prednisone 19 (46%) NOTE. Virologic response is defined as sustained lack of detection of serum HBV DNA. Abbreviation: NS, not significant. *Pretreatment ALT greater than 4 ULN (i.e., 172 IU/L). Mild flare is an increase in ALT IU/L above baseline value; moderate flare is an increase in ALT of IU/L above baseline; severe flare is increase in ALT 344 IU/L above baseline. Compared with patients with no flare. Compared with untreated controls.

4 1024 NAIR AND PERRILLO HEPATOLOGY November 2001 FIG. 2. (A) Analysis of independent factors associated with a sustained loss of serum HBV DNA (all patients included). *P.001; P.007; P.004. (B) Analysis of independent factors associated with a sustained loss of serum HBV DNA in patients with high-level viremia. *P.009; P.02. (C) Analysis of independent factors associated with a sustained loss of serum HBV DNA in patients with low-level viremia. *P.02; P.03; OR, odds ratio. pendent predictors of response (Fig. 2A). Among these 3 variables, however, severe flare was most predictive of response (odds ratio, 5.3; 95% CI, ; P.004). Neither interferon treatment nor prednisone in combination with interferon were predictive of a virologic response. Multivariate analysis of predictors of a sustained loss of HBeAg in combination with undetectable HBV DNA showed that pretreatment ALT ( 4 ULN; odds ratio, 2.58; 95% CI, ; P.02), high HBV DNA ( 100 pg/ml; odds ratio, 0.318; 95% CI, ; P.007), and both moderate flare (odds ratio, 4.81; 95% CI, ; P.003), and severe flare (odds ratio, 4.55; 95% CI, ; P.01) were predictive of a sustained loss of HBeAg and HBV DNA. Antiviral treatments were not predictive. Multiple logistic regression analysis was performed separately for patients with high and low levels of pretreatment viremia to assess the relationship between key variables and the sustained loss of serum HBV DNA. In patients with highlevel viremia, pretreatment ALT ( 4 ULN; odds ratio, 9.6; 95% CI, ; P.009), and severe flare (odds ratio, 7.2; 95% CI, ; P.02) were the only predictors of a sustained loss of HBV DNA (Fig. 2B). In the patients with low-level viremia, the independent predictors of a sustained loss of HBV DNA were treatment with high-dose interferon (odds ratio, 7.3; 95% CI, ; P.03) and pretreatment HBV DNA (odds ratio, 0.97; 95% CI, ; P.02) (Fig. 2C). Although the overall rate of HBsAg loss was low (6%) in the entire patient population, a severe flare was associated with a significantly higher rate of HBsAg loss (21% vs. 2.4% in patients with no flare; P.01). Patients with mild and moderate flares had a slightly higher rate of clearance of

5 HEPATOLOGY Vol. 34, No. 5, 2001 NAIR AND PERRILLO 1025 HBsAg (4% and 7%, respectively) than patients with no flare; however, the difference was not statistically significant. Flares were generally well tolerated, but 1 patient with cirrhosis in the prednisone pretreatment group died of hepatic decompensation. The details in this case have been previously described. 1 DISCUSSION Spontaneous or therapy-induced flares in serum aminotransferase levels are a frequent occurrence in chronic hepatitis B and are generally considered to originate from a more vigorous attack of the cell-mediated immune system against HBV-infected hepatocytes. 3,4 Flares that occur during interferon therapy have been associated with a higher rate of virologic response. 1-3 The current study, while limited by a relatively small number of patients in each treatment group, revealed that treatment with interferon results in flares in approximately 50% of cases. From a theoretical perspective, the limited response to interferon may not only be explainable by the relatively low rate of occurrence of flares, but by a lack of sufficient magnitude so as to permanently suppress viral replication. 3 These considerations led us to study the question of whether a higher degree of flare was associated with a higher rate of virologic response and whether this had particular meaning for patients with various levels of pretreatment viremia. In the current study, we found that the presence of an ALT flare during interferon therapy was significantly associated with a virologic response. Of note, as the degree of flare increased in patients with high-level viremia, so did the rate of sustained loss of HBV DNA and HBeAg (Fig. 1A). Although patients with low-level viremia were also more likely to lose HBeAg in the presence of a moderate flare, the association between the degree of flare and virologic response was, in general, less clear in this group of patients (Fig. 1B). Multivariate analysis revealed that moderate (i.e., an absolute ALT increase of at least 172 IU/L from baseline) to severe flares (an ALT increase of at least 344 IU/L from baseline) were predictive of a sustained loss of HBV DNA (Fig. 2A). Furthermore, multivariate analysis in patients with high-level viremia showed that high pretreatment ALT and severe flares were predictive of a sustained loss of serum HBV DNA (Fig. 2B). This was not the case for patients with low-level viremia, a group in which treatment with high-dose interferon and low baseline serum HBV DNA seemed to be more predictive (Fig. 2C). An alternate explanation for the observed relationship between the degree of flare and virologic response in our highly viremic patients is that a rapid decline in HBV DNA level led to immune reconstitution. This has been proposed as an explanation for ALT flares that occur during lamivudine treatment. 5 This question cannot be addressed by the current study, but we believe that this is unlikely to explain our findings because patients with low-level viremia would be anticipated to have similar or even greater rates of decline in serum HBV DNA, and interferon is modest by comparison to lamivudine in its antiviral activity. Based on the findings of the current study, it is reasonable to conclude, therefore, that the degree of flare plays an important determinant role in inducing a sustained loss of viral replication in patients with high-level viral replication. This conclusion has considerable relevance for interferon treatment because patients with high-level viremia have generally been found to have a low response rate. 1,6,7 As anticipated, multivariate analysis failed to show that treatment with interferon was an independent predictor of response in patients with high-level viremia (Fig. 2B). In highly viremic patients only, pretreatment ALT and a severe flare were predictive of a response. Taken together, these observations imply that a high basal level of immunologic activity as well as an intense treatment-induced increase in immunologic response are important in achieving a virologic response in patients with highlevel viremia. Hepatitis-like flares in serum aminotransferase levels can also be induced in some patients after a short course of corticosteroids. 8,9 This observation has led to clinical trials in which corticosteroids were used in combination with interferon to enhance the efficacy of treatment. 1,10,11 This strategy has not improved overall response rates to interferon but has been shown to be associated with a higher rate of response in patients with low baseline ALT levels ( 100 IU/L). 1,12 It is noteworthy that these patients have been found to have higher levels of serum HBV DNA before treatment. 1,3 A recent study by Liaw et al. used a short course of corticosteroids before a 9-month course of lamivudine as a means of enhancing antiviral efficacy and shortening the duration of treatment with the latter. 13 In that study, patients who showed a 5-fold elevation in ALT during treatment had a significantly higher rate of HBeAg seroconversion than those in whom a flare did not occur. Unlike the current study, however, Liaw et al. did not specifically analyze the relationship between the degree of ALT flare and baseline HBV DNA. An interferon-induced flare of hepatitis is a favorable indicator that a response to treatment will occur, but patients with cirrhosis and marginal hepatic reserve are not good candidates for treatment because of a high risk of hepatic decompensation. 3,14 Infrequently, flares may be accompanied by an increase in serum bilirubin or alteration in synthetic function in patients who otherwise appear to be acceptable candidates for treatment. In our experience, this should prompt discontinuation of treatment and careful consideration to initiation of interferon at a reduced dose after the laboratory abnormalities have subsided. In conclusion, flare is an independent predictor of response in patients with chronic hepatitis B when interferon is used to treat chronic hepatitis B. The therapeutic benefit of a flare is most evident in patients with high-level viremia, a group that has been difficult to manage with interferon therapy alone. Our data leads us to believe that a higher rate of virologic response may be achieved by the induction of a robust flare in clinically stable, noncirrhotic patients with high pretreatment HBV DNA levels. One way of accomplishing this may be to use a short course of corticosteroids before interferon or nucleoside analogue therapy with an alternate strategy being the use of therapeutic vaccines that incorporate immunodominant epitopes of HBV. 15 It has been hypothesized that increasing the priming dose and shortening the duration of corticosteroid treatment may increase the chances of such a flare. 16 Antiviral studies that specifically address the therapeutic value of corticosteroids in combination with interferon or nucleoside analogues in patients with high-level viremia appear warranted.

6 1026 NAIR AND PERRILLO HEPATOLOGY November 2001 Acknowledgment: The authors would like to gratefully acknowledge the other physician investigators and collaborators who were involved with the initial study: Drs. Eugene Schiff, Gary Davis, Henry Bodenheimer, Karen Lindsay, John Payne, Jules Dienstag, Christopher O Brien, Carlo Tamburro, Ira Jacobson, Richard Sampliner, David Feit, Jay Lefkowitch, Mary Kuhns, Carlton Meschievitz, Bharati Sanghvi, Janice Albrecht, and Alexandra Gibas. REFERENCES 1. Perrillo RP, Schiff ER, Davis GL, Bodenheimer HC Jr, Lindsay K, Payne J, Dienstag JL, et al. A randomized, controlled trial of interferon alfa-2b alone and after prednisone withdrawal for the treatment of chronic hepatitis B. The Hepatitis Interventional Therapy Group. N Engl J Med 1990; 323: Hoofnagle JH, Peters M, Mullen KD, Jones DB, Rustgi V, Di Bisceglie A, Hallahan C, et al. Randomized, controlled trial of recombinant human alpha-interferon in patients with chronic hepatitis B. Gastroenterology1988;95: Perrillo RP. Acute flares in chronic hepatitis B: the natural and unnatural history of an immunologically mediated infection. Gastroenterology 2001;120: Waters JA, O Rourke S, Schlict HJ, Thomas HC. Cytotoxic T cell responses in patients with chronic hepatitis B virus infection undergoing HBe antigen antibody seroconversion. Clin Exp Immunol 1995;102: Boni C, Bertoletti A, Penna A, Cavalli A, Pilli M, Urbani S, Scognamiglio P, et al. Lamivudine treatment can restore T cell responsiveness in chronic hepatitis B. J Clin Invest 1998;102: Thomas HC, Karayiannis P, Brook G. Treatment of hepatitis B virus infection with interferon. Factors predicting response to interferon. J Hepatol 1991;13(Suppl 1):S4-S7. 7. Dusheiko GM. Alpha-interferon treatment of chronic hepatitis B infection: predictors of responsiveness. J Gastroenterol Hepatol 1991;6(Suppl 1): Scullard GH, Smith CI, Merigan TC, Robinson WS, Gregory PB. Effects of immunosuppressive therapy on viral markers in chronic active hepatitis B. Gastroenterology 1981;81: Hoofnagle JH, Davis GL, Pappas SC, Hanson RG, Peters M, Avigan MI, Waggoner JG, et al. A short course of prednisone in chronic type B hepatitis. Report of a randomized, double blind, placebo-controlled trial. Ann Intern Med 1986;104: Perrillo RP, Regenstein FG, Peters MG, DeSchryer-Keckemeti K, Bodicky CJ, Campbell CR, Kuhns MC. Prednisone withdrawal followed by recombinant alpha interferon in the treatment of chronic type B hepatitis. A randomized, controlled trial. Ann Intern Med 1988;109: Krogsgaard K, Marcellin P, Trepo C, Berthelot P, Sanches-Tapias JM, Bassendine M, Tran A, et al. Prednisone withdrawal therapy enhances the effect of human lymphoblastoid interferon in chronic hepatitis B. Interpred Trial Group. J Hepatol 1996;25: Cohard M, Poynard T, Mathurin P, Zarski JP. Prednisone-interferon combination in the treatment of chronic hepatitis B: direct and indirect metaanalysis. HEPATOLOGY 1994;20: Liaw FN, Tsai SL, Chein RN, Yeh CT, Chu CM. Prednisolone priming enhances Th1 response and efficacy of subsequent lamivudine therapy in patients with chronic hepatitis B. HEPATOLOGY 2000;32: Perrillo R, Tamburro C, Regenstein F, Balart L, Bodenheimer H, Silva M, Schiff E, et al. Low-dose, titratable interferon alfa in decompensated liver disease caused by chronic infection with hepatitis B virus. Gastroenterology 1995;109: Heathcote J, McHutchison J, Lee S, Tong M, Benner K, Minuk G, Wright T, et al. A pilot study of the CY-1899 T-cell vaccine in subjects chronically infected with hepatitis B virus. HEPATOLOGY 1999;30: Perrillo RP, Mason AL. Therapy for hepatitis B virus infection. Gastroenterol Clin North Am 1994;23: