RONG-NAN CHIEN, 1 YUN-FAN LIAW, 1 AND MARK ATKINS 2 FOR THE ASIAN HEPATITIS LAMIVUDINE TRIAL GROUP

Transcription

1 Pretherapy Alanine Transaminase Level as a Determinant for Hepatitis B e Antigen Seroconversion During Lamivudine Therapy in Patients With Chronic Hepatitis B RONG-NAN CHIEN, 1 YUN-FAN LIAW, 1 AND MARK ATKINS 2 FOR THE ASIAN HEPATITIS LAMIVUDINE TRIAL GROUP In the reported Asian lamivudine trial, the rate of hepatitis B e antigen (HBeAg) seroconversion, defined as HBeAg/hepatitis B virus (HBV) DNA seroclearance and development of anti-hbe, during 52 weeks of treatment was only 13% to 16%. To evaluate whether any factors influenced HBeAg seroconversion, data from 345 patients in that trial were reanalyzed to correlate HBeAg seroconversion with variables including treatment, age, gender, body build, histology, baseline HBV-DNA levels, and alanine transaminase (ALT) levels. Exploratory analysis using stepwise modeling revealed that HBeAg seroconversion correlated highly with pretherapy ALT (P F.001) followed by lamivudine therapy (P.013), but only marginally with baseline HBV-DNA (P.071) and cirrhosis (P.066) for lamivudine 100 mg and placebo comparison. Among these four variables, only pretherapy ALT still had a highly significant (P F.001) correlation and lamivudine therapy had a borderline association (P.066) for lamivudine 25 mg and placebo comparison. Categorical analysis revealed that HBeAg seroconversion occurred earlier and the cumulative rate was significantly higher in patients with pretherapy ALT values over 2 times the upper limit of normal (ULN) as compared with treated patients with lower ALT levels or untreated control patients with the same ALT levels (P F.001, respectively). The highest HBeAg seroconversion rate was observed in 100 mg lamivudine-treated patients with ALT levels greater than 5 times the ULN (64%) compared with patients with ALT 2 to 5 times the ULN (26%, P.03); and ALT less than 2 times the ULN, (5%, P F.001). These results suggest that pretherapy ALT is the strongest determinant for HBeAg seroconversion during lamivudine therapy, and should be considered in selecting patients for treatment. (HEPATOLOGY 1999;30: ) Abbreviations: HBV, hepatitis B virus; HBeAg/anti-HBe, hepatitis B e antigen and antibody; ALT, alanine transaminase; HAI, hepatitis activity index; ULN, upper limit of normal; CI, confidence interval; cccdna, covalently closed circular DNA; YMDD, tyrosine-methionine-aspartate-aspartate. From the 1 Liver Research Unit, Chang Gung Memorial Hospital, Chang Gung University, Taipei, Taiwan; and the 2 Department of Anti-Infectives and Hepatitis, Glaxo Wellcome, Research and Development, Greenford, United Kingdom. Received February 16, 1999; accepted June 7, Supported by a grant from Glaxo-Wellcome. Address reprint requests to: Yun-Fan Liaw, M.D., Liver Research Unit, Chang Gung Memorial Hospital, 199, Tung Hwa North Road, Taipei, Taiwan liveryfl@tp.silkera.net; fax: Copyright 1999 by the American Association for the Study of Liver Diseases /99/ $3.00/0 770 Lamivudine, the (-) enantiomer of 2 -deoxy-3 -thiacytidine, is a nuleoside analogue with potent direct antiviral effects against hepatitis B virus (HBV) both in vitro 1 and in vivo. 2-3 Treatment with lamivudine may also result in enhanced HBV-specific T-cell responses. 4 Although HBV suppression during lamivudine therapy can result in reductions in hepatic inflammation even in the absence of hepatitis B e antigen (HBeAg) seroconversion, 5 HBeAg seroconversion has been accepted as an important indicator of response and has been considered a critical step to ensure sustained HBV suppression. 6 The HBeAg seroconversion rate in a large series of Asian patients treated with lamivudine 100 mg/d for 1 year was only 16%, 5 which is lower than the HBeAg loss rate of 25% to 40% after interferon therapy with or without corticosteroid priming. 7,8 After the 1-year results of the Asian lamivudine trial had been reported, 5 we reviewed the study records at our site and found that the HBeAg seroconversion rate of our 35 patients treated with lamivudine 100 mg/d was 34% (Liaw and Chien, unpublished data, September, 1998). This much higher HBeAg seroconversion rate, compared with that of the whole Asian study, suggests that some factors might have influenced the response to lamivudine therapy. Actually, the Asian lamivudine study did show that patients with HBeAg seroconversion had higher median pretherapy alanine transaminase (ALT) levels. However, similar to earlier studies showing the predictive value of pretherapy ALT on response to interferon, 7,9 no further categorical analysis in ALT/HBeAg response was conducted in the reported Asian study. 5 Because the hepatic injuries in patients with chronic HBV infection are the results of cytotoxic T-cell mediated immune hepatocytolysis, 9 higher ALT levels reflect stronger immune responses of the host against HBV. It is of interest to know whether and to what extent such antiviral defenses of the host influence the response to lamivudine therapy. We therefore conducted an exploratory retrospective analysis on the data from the previously reported Asian lamivudine study 5 to examine the HBeAg seroconversion response in relation to clinical variables, with special emphasis on baseline ALT levels, in patients with chronic hepatitis B. PATIENTS AND METHODS The patients and results of a 1-year Asian lamivudine study have been described elsewhere. 5 None of the patients received specific HBV therapy, including interferon, within 6 months before entry. The primary end-point for that study was improvement in liver histology and a detailed categorical analysis of the HBeAg seroconversion data has not been performed previously. Briefly, a dose of 100 mg of lamivudine resulted in a significant improvement in liver

2 HEPATOLOGY Vol. 30, No. 3, 1999 CHIEN ET AL. 771 necroinflammatory activity (52% vs. 25%; P.001) and ALT normalization (72% vs. 24%; P.001) compared with a placebo. The rate of HBeAg seroconversion, defined as HBeAg/HBV DNA seroclearance and development of anti-hbe, in patients treated with 25 mg lamivudine, 100 mg lamivudine, and a placebo was 13% (17 of 135 patients), 16% (22 of 140 patients), and 4% (3 of 70 patients), respectively. The HBeAg seroconversion rates in this study were reanalyzed to examine correlations with variables including treatment, age, gender, weight, body mass index, Knodell hepatitis activity index (HAI), presence of cirrhosis (HAI fibrosis score of 4), baseline HBV-DNA (pg/ml) levels, and ALT levels (times ULN). Because ALT is the most easily measured clinical indicator, special emphasis was paid to the baseline ALT levels of the patients in respective groups to determine whether the pretherapy ALT value was a determinant for HBeAg seroconversion during lamivudine therapy. Based on the findings that patients with ALT levels less than and greater than 5 times ULN have significantly different spontaneous HBeAg seroconversion rates, 8 an ALT level over 5 times the ULN was included in this analysis. Serum HBV DNA was quantified with a solution hybridization assay (Abbott Laboratories, North Chicago, IL) that has a lower limit of detection of 1.6 pg/ml. HBeAg and anti-hbe were detected by a qualitative HBeAg immunoassay (Abbott). Statistical Analysis. 2 test or Fishers Exact test, where appropriate, was used to investigate the relationship between pretherapy ALT and HBeAg seroconversion. All statistical testing was two tailed at the 5% level. An exploratory analysis using stepwise modeling was performed to determine a prognostic model and to estimate treatment efficacy in the presence of the following prognostic variables: baseline ALT levels (single value at the start of therapy expressed as times ULN), HBV DNA (pg/ml), HAI (overall Knodell HAI score), weight, body mass index, gender, age, and cirrhosis (indicated by a score of 4 on the fibrosis component of the Knodell HAI score). ALT, HBV DNA, HAI, weight, age, and body mass index were fitted as continuous variables. The stepwise logistic model was used with generous P values for entry (P.2) and exit (P.25) criteria so that the study was controlled for characteristics with a marginal association. Patients with missing values were not included in the logistic regression analysis. Missing data for HBeAg and anti-hbe were handled using the last observation carried forward. Patients who withdrew from the study before endpoint analysis and without achieving HBeAg seroconversion were considered as treatment failures, i.e., no HBeAg seroconversion. Missing values for HBV DNA were considered as failures, i.e., HBV DNA positive. The analysis software used was Statistical Analysis System (SAS; SAS Institute, NC) version 6.08 running under Virtual Memory System (VMS; Digital Equipment Corporation, MA). RESULTS Of the 358 patients enrolled in the study, 345 (96%) patients completed 1 year of lamivudine or placebo treatment. There were no differences among the three treatment groups in terms of pretherapy ALT and HBV-DNA levels. A stepwise logistic regression analysis revealed that HBeAg seroconversion was highly correlated with pretherapy ALT levels. For the comparison between the placebo and 100-mg dose of lamivudine, the stepwise modeling revealed that baseline ALT levels (odds ratio: 1.9; 95% confidence interval (CI): ; P.001) was the strongest predictor of HBeAg seroconversion followed by treatment (odds ratio: 7.5; 95% CI: ; P.013). Cirrhosis (odds ratio: 0.16; 95% CI: ; P.066) and baseline HBV-DNA levels (odds ratio: 0.99; 95% CI: ; P.071) showed a borderline association. Thus, the estimated odds ratio of response associated with lamivudine 100 mg comparing placebo, and after adjusting for the effects of baseline ALT, HBV DNA, and cirrhosis, is 7.5 (95% CI: ; P.013). That is, after controlling for imbalances in important prognostic factors, those patients receiving lamivudine 100 mg have seven times the odds of responding to therapy as those patients receiving a placebo, with a 95% CI ranging from 2 to 37. For comparison between the placebo and the 25-mg dose of lamivudine, stepwise modeling revealed that baseline ALT levels (odds ratio: 1.6; 95% CI: ; P.001) was the strongest predictor of response followed by body mass index (odds ratio: 0.6; 95% CI: ; P.011). Treatment (odds ratio: 4.0; 95% CI: ; P.066), baseline HAI (odds ratio: 1.1; 95% CI: ; P.067), and weight (odds ratio: 1.1; 95% CI: ; P.072) showed a borderline association. Baseline ALT was the strongest predictor of HBeAg seroconversion for both lamivudine doses (Table 1). The estimated odds ratio of response associated with baseline ALT levels (times ULN) after adjusting the effects of other factors is 1.9 and 1.6 for patients treated with lamivudine 100 mg/d and 25 mg/d, respectively (P.001). That is, patients with ALT levels greater than 5 times the ULN have 9.5 (5 1.9) and 8.0 (5 1.6) times the odds of response, respectively. The results of the stepwise modeling should be considered as exploratory owing to the relatively small number of patients who seroconverted. The 95% CI for the treatment effects are wide, indicating that a larger sample size would be required for a more accurate estimate; this study was not powered to compare HBeAg seroconversion rates. The HBeAg seroconversion rate of each group, in relation to pretherapy ALT categories, are shown in Table 2. The HBeAg seroconversion rates increased gradually along with increasing ALT levels ( 2 for trend: P.001). The same trend is apparent for most centers but the numbers of patients participating is small. The patients with pretherapy ALT levels greater than 2 times the ULN showed a significantly higher HBeAg seroconversion rate than the spontaneous HBeAg seroconversion rate in their counterparts receiving a placebo (34% in the 100-mg lamivudine group and 32% in the 25-mg lamivudine group vs. 7%; P.01 and.02, respectively), and higher than treated patients with ALT levels less than 2 times the ULN (P.001). The highest HBeAg seroconversion rates were in patients with pretherapy ALT levels greater than 5 times the ULN, and was significantly higher than that in patients with pretherapy ALT levels between 2 and 5 times the ULN or less than 2 times the ULN TABLE 1. Stepwise Logistic Regression Analysis on Baseline Variables and HBeAg Seroconversion During 1 Year of Lamivudine Therapy Baseline Variables 100 mg/d vs. Placebo 25 mg/d vs. Placebo OR (CI) P OR (CI) P Treatment (yes/no) 7.5 ( ) ( ).066 ALT ( ULN) 1.94 ( ) ( ).001 HBV DNA (pg/ml) 0.99 ( ) HAI score ( ).067 Cirrhosis (yes/no) 0.16 ( ) BMI ( ).011 Weight (kg) ( ).072 Age (yr) Gender (M/F) Abbreviations: OR (CI), odds ratio (95% confidence interval); BMI, body mass index.

3 772 CHIEN ET AL. HEPATOLOGY September 1999 TABLE 2. HBeAg Seroconversion in Relation to Pretherapy ALT Level During 1 Year of Lamivudine Therapy Baseline ALT Placebo n 70 Lamivudine (25 mg) n 135 Lamivudine (100 mg) n ULN 1/42 (2%) 2/88 (2%) a 4/87 (5%) b 2 ULN 2/28 (7%) c 15/47 (32%) d 18/53 (34%) e 2 5 ULN 1/21 (5%) 8/30 (27%) f 11/42 (26%) g 5 ULN 1/7 (14%) 7/17 (41%) h 7/11 (64%) k Total 3/70 (4%) 17/135 (13%) 22/140 (16%) NOTE: Data expressed as No. with seroconversion/no. of patients (%). a vs. d, a vs. f, a vs. h, b vs. e, b vs. g, and b vs. k: P.001; c vs. d: P.02; c vs. e: P.01; g vs. k: P.03; a vs. f vs. h, b vs. g vs. k: P.001 ( 2 for trend). who received lamivudine 100 mg daily (64% vs. 26% and 5%; P.03 and P.001, respectively). In patients with pretherapy ALT levels greater than 5 times the ULN, there was no significant difference in baseline HBV-DNA levels between patients with and without HBeAg seroconversion. In contrast, the HBeAg seroconversion rate in patients with pretherapy ALT levels less than 2 times the ULN was not higher than the spontaneous HBeAg seroconversion rate in the placebo group. In addition, HBeAg seroconversion occurred earlier in patients with baseline ALT levels greater than 2 times the ULN, particularly those greater than 5 times the ULN, as shown in Fig. 1. HBeAg seroconversion may occur as early as 4 weeks after the first dose of lamivudine (Fig. 2). DISCUSSION Previous studies with interferon have shown that there is a significant correlation between pretherapy ALT levels and HBeAg seroconversion. 7,9 A 3-month lamivudine study and a famciclovir study also showed that sustained HBeAg seroconversion and/or HBV-DNA suppression were more likely in patients with initially high ALT levels. 2,6 Although the present analysis should be considered as exploratory owing to the relatively small number of patients who seroconverted, the results add support to earlier observations and further indicate that HBeAg seroconversion during lamivudine therapy increased significantly along with increasing pretherapy ALT. FIG. 1. Time to HBeAg seroconversion during lamivudine therapy in relation to pretherapy ALT level. Note that HBeAg seroconversion occurs earlier and the cumulative rate is higher in patients with baseline ALT values greater than 2 times the ULN. This trend is most obvious in patients with a pretherapy ALT level greater than 5 times the ULN, particularly those patients treated with 100 mg of lamivudine w as compared with those treated with 25 mg of lamivudine daily or placebo. FIG. 2. Lamivudine therapy in a patient with a pretherapy ALT level greater than 5 times the ULN. Note that shortly after starting lamivudine (100 mg/d) therapy, there was a precipitous decline of HBV-DNA level followed by ALT normalization and HBeAg seroconversion to anti-hbe that was sustained after the end of therapy. The hatched area at the bottom of the figure represents the normal range of ALT ALT (U/L), HBV-DNA (pg/ml). It is possible that with a larger sample size other predictors such as HBV DNA and baseline HAI score would become significant. Of note is that patients with a pretherapy ALT level greater than 5 times the ULN had an HBeAg seroconversion rate up to 64%, which is the highest response rate ever reported in the treatment of chronic HBV infection. 8 In addition, the high HBeAg seroconversion rate in such patients was not significantly associated with their baseline HBV-DNA level, which was assessed to have a borderline association with the overall HBeAg seroconversion in the whole study population. Because the number of patients with ALT levels greater than 5 times the ULN was relatively small, a larger sample size would be required for a more accurate estimate. We have since treated 98 patients with a baseline ALT level greater than 5 times the ULN, including 66 patients with ALT levels greater than 10 times the ULN, and found that the cumulative HBeAg seroconversion rate at the end of 6 months of lamivudine therapy was 67% for patients with ALT levels between 5 and 10 times the ULN and 85% for those with ALT levels greater than 10 times the ULN (Chien and Liaw, unpublished data, March 1999) in contrast to a historical control of 20% and 30%, respectively. 8 In contrast, lamivudine therapy in patients with lower or normal pretherapy ALT values did not show a higher HBeAg seroconversion rate than the placebo-treated control group. These findings may explain the relatively low first year HBeAg seroconversion rate in the Asian study in which a third of patients had normal ALT levels, approximately two thirds of patients had an ALT level less than 2 times the ULN, and the median ALT level was 1.5 times the ULN before therapy. 5 Additionally, previous interferon studies often report HBeAg loss, 7 whereas this study reports full HBeAg seroconversion (undetectable HBV DNA and HBeAg and detectable anti-hbe), which will also result in a lower response rate. Because ALT elevations in patients with chronic hepatitis B are the results of T-cell mediated hepatocytolysis, 10 the level of ALT elevation reflects the level of T-cell immune response of the patients to HBV. The results of the present analysis, therefore, suggest that antiviral agents like lamivudine are more effective, in terms of HBeAg

4 HEPATOLOGY Vol. 30, No. 3, 1999 CHIEN ET AL. 773 seroconversion, in patients who have mounted an ongoing immune response to HBV. It has been shown that chronic infection of the hepatocyte with hepadnavirus is maintained by the presence of viral covalently closed circular DNA (cccdna) in the nucleus. This viral DNA species is the template for viral transcription and its production is regulated and amplified by an intracellular pathway in which newly synthesized genomic DNA is transported to the nucleus. 11 Although potent HBV polymerase inhibitor can block new cccdna synthesis and preexisting cccdna may be degraded when hepatocyte death occurs, 12 residual amounts of stable cccdna remain. 13 Thus, complete viral clearance by antiviral agents alone may be difficult in a host lacking adequate specific antiviral immune defenses. 13 In view of these basic mechanisms, the high HBeAg seroconversion rate in patients with high baseline ALT levels seems to be the result of a concerted effort of (1) the immune-mediated killing of the hepatocytes harboring cccdna by the antiviral defenses of the host, and (2) the potent direct antiviral effect of lamivudine and the enhanced CD-4 responses resulting from lamivudine therapy. 4 This may suggest that alternative treatment strategies need to be defined for patients in the immune tolerance phase or with a low anti-hbv immune response. The emergence of mutations in the tyrosine-methionineaspartate-aspartate (YMDD) locus of the HBV-RNA dependent DNA polymerase may reduce the likelihood of response and sometimes may induce severe hepatitis during prolonged lamivudine therapy. 15 However, these HBV variants do not arise before at least the completion of 24 weeks of treatment in immunocompetent patients. 5 Hence, a treatment strategy that shortens the duration of lamivudine therapy required for viral clearance and HBeAg seroconversion may conceivably reduce the risk of the emergence of these variants. The present analysis revealed that HBeAg seroconversion occurred earlier and the cumulative response rate was higher in patients with higher baseline ALT values than those with lower ALT levels. It seems reasonable to suggest that chronic hepatitis B patients with higher ALT levels, particularly those greater than 5 times the ULN, can be HBeAg seroconverted with lamivudine alone. In addition, lamivudine-induced HBeAg seroconversions appear to be durable after treatment is stopped, like those induced by interferon, 14 and therefore can be used as an indicator to consider discontinuing lamivudine therapy. This is a very important practical point to clinicians because serum ALT is the most easily measured test and is more widely available than HBV-DNA testing or liver biopsy, and they may select patients accordingly for the most cost-effective treatment. In contrast, patients with normal or low ALT levels but active HBV replication have little chance of HBeAg seroconversion during 1 year of lamivudine therapy, and prolonged therapy is therefore required for sustained HBV suppression. Despite the finding that HBV suppression in the absence of HBeAg seroconversion can result in substantial histological improvement, the incidence of YMDD mutation and its impact may increase along with increased duration of therapy. 5,15,16 Therefore, strategies using the manipulation of the host immune responses against HBV to shorten the duration of and to increase the response to lamivudine therapy need to be evaluated in patients with lower or normal ALT levels. In conclusion, patients with higher pretherapy ALT values ( 2 times the ULN) have significantly enhanced HBeAg seroconversion during 1 year of lamivudine therapy compared with patients with lower ALT levels or placebocontrolled patients. Patients with an ALT level greater than 5 times the ULN have the highest seroconversion rate and, therefore, can be HBeAg seroconverted with lamivudine alone. A combination of lamivudine and immunomodulatory agents should be evaluated in controlled trials to see if HBeAg seroconversion rates can be further enhanced in patients who are in the immune tolerance phase or who have a low anti-hbv immune response. Acknowledgment: The authors thank the excellent technical assistance of C. Y. Chen and the secretarial assistance of S. C. Chen and S. C. Chu. APPENDIX In addition to the authors, the following investigators have participated in the Asia Hepatitis Lamivudine Therapy Group: Clinical Group: Ching-Lung Lai, Man-Fung Yuen, Wai-Man Wong, Department of Medicine, Queen Mary Hospital, Hong Kong; Nancy WY Leung, Department of Medicine, Prince of Wales Hospital, Hong Kong; Ting-Tsung Chang, Department of Internal Medicine, National Cheng Kung University Hospital, Taiwan; Richard Guan, Ivy Yap, Department of Medicine, National University Hospital, Singapore; Dar-In Tai, Chang Gung Memorial Hospital, Kaohsiung, Taiwan; Keng-Yeen Ng, Han-Seong Ng, Wang-Cheng Chow, Singapore General Hospital, Singapore. Histopathologist: Pui-Chee Wu, Department of Pathology, Queen Mary Hospital, Hong Kong. Glaxo Wellcome: D Fraser Gray, Julie C Dent, Judy Barber, Lynn Condreay, Sally Stephenson. REFERENCES 1. Doong SL, Tsai CH, Schinazi RF, Liotta DC, Cheng YC. Inhibition of the replication of hepatitis B virus in vitro by 2, 3 -dideoxy-3 thiacytidine and related analogues. Proc Natl Acad Sci U S A 1991;88: Dienstag JL, Perrillo RP, Schiff ER, Bartholomew M, Vicary C, Rubin M. A preliminary trial of lamivudine for chronic hepatitis B infection. N Engl J Med 1995;333: Nevens F, Main J, Honkoop P, Tyrrell DL, Barber J, Sullivan MT, Fevery J, et al. Lamivudine therapy for chronic hepatitis B: a six-month randomized dose-ranging study. Gastroenterology 1997;113: 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: Lai CL, Chien RN, Leung NWY, Chang TT, Guan R, Tai DI, Ng KY, et al. A one year trial of lamivudine for chronic hepatitis B. N Engl J Med 1998;339: Zoulim F, Trepo C. Drug therapy for chronic hepatitis B: antiviral efficacy and influence of hepatitis B virus polymerase mutations on the outcome of therapy. J Hepatol 1998;29: Wong DKH, Cheung AM, O Rourke K, Naylor CD, Detsky AS, Heathcote J. Effect of alpha-interferon treatment in patients with hepatitis B e antigen-positive chronic hepatitis B; a meta-analysis. Ann Intern Med 1993;119: Liaw YF. Current trends in therapy for chronic viral hepatitis. J Gastroenterol Hepatol 1997;12:S346-S Lok ASF, Chang MG, Watson G, Ayola B. Predictive value of aminotransferase and hepatitis B virus DNA levels on response to interferon therapy for chronic hepatitis B. J Viral Hepat 1998;5: Liaw YF, Tsai SL. Pathogenesis and clinical significance of acute exacerbations and remissions in patients with chronic hepatitis B virus infection. Viral Hep Rev 1997;3: Wu TT, Coates L, Aldrich CE, Summers J, Mason WS. In hepatocytes

5 774 CHIEN ET AL. HEPATOLOGY September 1999 infected with duck hepatitis B virus, the template for viral RNA synthesis is amplified by an intracellular pathway. Virology 1990;175: Fourel I, Cullen J, Saputelli J, Aldrich C, Schaffer P, Averett D, Pugh J, et al. Evidence that hepatocyte turnover is required for rapid clearance of duck hepatitis B virus during antiviral therapy of chronically infected ducks. J Virol 1994;68: Mason WS, Cullen J, Moraleda G, Saputelli J, Aldrich CE, Miller DS, Tennant B, et al. Lamivudine therapy of WHV-infected woodchucks. Virology 1998;245; Schiff E, Cianciara J, Kowdley K, Norkrans G, Perillo R, Tong M, Crowther L, et al. Durability of HbeAg seroconversion after lamivudine monotherapy in controlled phase II and III trials [Abstr]. HEPATOLOGY 1998;28(Suppl):163A. 15. Liaw YF, Chien RN, Yeh CT, Tsai SL, Chu CM. Acute exacerbation and hepatitis B virus clearance after emergence of YMDD motif mutation during lamivudine therapy. HEPATOLOGY 1999; 30 (in press). 16. Liaw YF, Lai CL, Leung NWY, Chang TT, Guan R, Tai DI, Ng KY, et al. Two year lamivudine therapy in chronic hepatitis B infection: results of a placebo controlled multicenter study in Asia [Abstract]. Gastroenterology 1998;114:A1289.