Lamivudine and 24 weeks of lamivudine/interferon combination therapy for hepatitis B e antigen-positive chronic hepatitis B in interferon nonresponders
Journal of Hepatology
Volume 38, Issue 6, June 2003, Pages 818-826
Eugene R. Schiff1, Jules L. Dienstag2, Selim Karayalcin3, Ian S. Grimm4, Robert P. Perrillo5, Petr Husa6, R. A. de Man7, Zachary Goodman8, Lynn D. Condreay9, Lynn M. Crowther9, Mary A. Woessner9, Penny J. McPhillips10, Nathaniel A. Brown and the International Lamivudine Investigator Group.
1 Division of Hepatology, University of Miami, Jackson Medical Towers, 1500 N.W. 12th Avenue, Suite 1101, Miami, FL 33136, USA. 2 Massachusetts General Hospital/Harvard Medical School, Boston, MA, USA. 3 Ankara University Medical School, Ankara, Turkey. 4 University of North Carolina, Chapel Hill, NC, USA. 5 Ochsner Clinic, New Orleans, LA, USA. 6 Clinic of Infectious Diseases, Brno-, Bohunice, Czech Republic. 7 Academisch Ziekenhuis Dijkzigt Rotterdam, Rotterdam, Netherlands. 8 Armed Forces Institute of Pathology, Washington, DC, USA. 9 GlaxoSmithKline, Research Triangle Park, NC, USA
10 GlaxoSmithKline, Greenford, UK
Lamivudine is effective in treatment-naive patients with chronic hepatitis B, but its role in interferon nonresponders has not been described. We assessed lamivudine treatment, with or without added interferon, in patients with hepatitis B e antigen (HBeAg)-positive chronic hepatitis B who had failed interferon therapy previously.
Patients were randomized to lamivudine (100 mg) or placebo for 52 weeks or to a 24-week regimen of lamivudine plus interferon. Primary treatment comparisons were at week 52, with a 16-week posttreatment follow-up period. Measurements included histology (primary endpoint), HBeAg response, normalization of alanine aminotransferase, reduction of hepatitis B virus (HBV) DNA, and safety.
Among 238 patients, histologic response was significantly more common in patients treated with lamivudine (52 versus placebo 25%, P=0.002) or the combination regimen (32%, P=0.01). HBeAg loss was also more common with
lamivudine (33 versus 13 versus 21%), as were virologic and alanine aminotransferase responses. Among 28 subjects with HBeAg loss/seroconversion, 71% had durable responses 16 weeks posttreatment.
The authors conclude that Lamivudine for 52 weeks is as effective in interferon nonresponders as in previously reported treatment-naive patients; however, a combination of lamivudine for 24 weeks and interferon for 16 weeks was not effective in this population.
Chronic hepatitis B is among the leading causes of death worldwide, and progressive liver damage in patients with hepatitis B is associated with high-level hepatitis B virus (HBV) replication. The goal of antiviral therapy, then, is viral eradication or long-term suppression of HBV replication.
Interferon-, the first approved treatment for hepatitis B, is administered by injection, difficult to tolerate, but results in a 20-30% loss of hepatitis B e antigen (HBeAg), a serologic response associated with durable benefit. Because patients with high HBV DNA, low alanine aminotransferase (ALT), and early-life acquisition of HBV infection do not to respond to interferon (IFN), and because most patients fail to respond to IFN, additional treatments were developed.
Lamivudine was shown to be effective in three randomized-controlled trials involving previously untreated (`IFN-naive') patients with chronic hepatitis B. Such clinical trials and the trial described here led to regulatory approval of
lamivudine as therapy for chronic hepatitis B. The present trial was designed to assess the efficacy of lamivudine in HBeAg-seropositive IFN nonresponders, i.e. patients with more treatment-refractory hepatitis B for whom no effective treatment was available. We also incorporated into this trial an assessment of a combination of lamivudine plus the then-approved treatment for hepatitis B, 16 weeks of IFN.
Eligible patients, >16 years, had HBsAg>6 months, HBeAg, hybridization-assay-detectable HBV DNA, ALT1.3 x the upper limit of normal (ULN), histologic chronic hepatitis, and previous treatment with 240 million units (MU) of IFN (half the full course recommended in the United States product label for IFN -2b); IFN must have been completed 6 months earlier, and patients must have failed IFN for lack of efficacy, not intolerance. Other inclusion and exclusion criteria were identical to those in previous western lamivudine trials.
This was a multicenter, randomized, partially blinded trial of three treatment regimens. All three groups underwent study observation for 68 weeks, which included, for the majority of subjects, a primary treatment period of 52 weeks and a 16-week follow-up period. The primary study goal was to compare 1 year of lamivudine to 1 year of placebo in IFN nonresponders. Histologic improvement, defined as a 2-point reduction in Histologic Activity Index (HAI), was the primary efficacy endpoint, as in other lamivudine registration trials. Although recently, IFN retreatment was reported to be effective in a proportion of prior IFN nonresponders, at the time this study was designed, IFN nonresponders were considered unlikely to respond to retreatment; therefore, an IFN monotherapy arm was not included. A
combination regimen (lamivudine plus IFN) was incorporated into the protocol; however, we were constrained by regulatory considerations to limit IFN to its approved 16-week course. Because previous IFN trials had shown enhanced responsiveness in subjects with low-level HBV DNA, we postulated that HBV DNA reduction with lamivudine prior to introducing IFN would improve the likelihood of a response. Therefore, the combination arm was limited to six months consisting of eight weeks of lamivudine pretreatment followed by the regulatory-approved 16 weeks of IFN, with continued lamivudine. The design of this IFN arm paralleled those in previous IFN trials in having a 2-month lead-in period and the principal efficacy evaluations 6 months posttreatment, i.e. at week 52. Finally, a follow-up period (weeks 52-68) was incorporated to provide a preliminary controlled assessment of the safety and efficacy of stopping versus continuing lamivudine after 1 year of treatment.
Patients' treatment was blinded until week 8, when a randomization envelope revealed whether the patient would receive open-label IFN. Treatment assignments in LAM and PLA remained blinded until end-of-study week 68. Patients' week-48 anti-HBe status (but not treatment assignment) was unblinded at week 52, and those who had seroconverted discontinued
study drug at week 52 but continued posttreatment monitoring to week 68 to determine the short-term durability of HbeAg seroconversion. To assess stopping versus continuing lamivudine in nonseroconverters at week 52, we subjected LAM patients to a double-blinded secondary randomization (1:1) to continue lamivudine (LAM/lAM) or to switch to placebo (LAM/PLA) for an additional 16 weeks. To maintain the study blind in the monotherapy groups during the week 52-68 period, patients initially randomized to PLA continued to receive placebo for the full 68 weeks.
Liver biopsies were performed within 12 months of baseline and at week 52. Biopsy slides - blinded to patient, treatment assignment, study center, date, and sequence (baseline versus week 52) - were read by hepatopathologist Z.G. and assigned an HAI score. In addition, the pathologist performed blinded ranking assessments of necroinflammatory and fibrosis features for paired biopsy slides for each patient, as described.
Representing 63 centers in 11 countries, 238 patients were assigned randomly (2:1:1) to three well-matched groups, LAM (119), PLA (56), or LAM+IFN (63). Study participants were somewhat older and had a higher (19%) prevalence of cirrhosis at baseline than those in lamivudine studies of interferon-naive patients; 209 (88%) patients completed the study, 110/119 (92%) LAM, 46/56 (82%) PLA, and 53/63 (84%) LAM+IFN.
Histologic response occurred in 62/119 (52%) LAM, 14/56 (25%) PLA (P=0.002), and 20/63 (32%) LAM+IFN (P=0.01). Worsening of HAI (by 2 points) occurred less commonly in LAM (8/119, 7%) than PLA (9/56, 16%) and LAM+IFN (13/63, 21%). The 26% with missing biopsy data (13/56 (23%) PLA; 30/119 (25%) LAM; 19/63 (30%) LAM+IFN) - nonsignificantly different among treatment groups and similar to the 16-33% rate reported in previous multicenter hepatitis trials - were counted as nonresponders. Median decrease in HAI was 4 points in LAM, 0 in PLA (P=0.002), and 1 in LAM+IFN (P=0.01).
At week 52, HBeAg seroconversion had occurred in an indistinguishable 19/108 (18%) LAM, 7/53 (13%) PLA, and 7/57 (12%) LAM+IFN (Fig. 3). Still insignificant but trending higher were week-68 HBeAg seroconversions in LAM/lAM (14/58, 24%) versus LAM/PLA (7/50, 14%), PLA (7/53, 13%), or LAM+IFN (5/57, 9%).
By week 52, HBV DNA responses had occurred in 102/110 (93%) LAM, 23/54 (43%) PLA, and 56/57 (98%) LAM+IFN. Sustained HBV DNA responses to week 52 were significantly more common in LAM (60/110, 55%) than PLA (9/54, 17%; P<0.001) and LAM+IFN (13/57, 23%; P=0.002). Retrospectively, because serum was tested for YMDD-variant HBV (see below) at end-of-therapy time points, we were able to compare the proportion of patients per group in whom HBV DNA became undetectable by PCR: 26/48 (54%) LAM+IFN (week 24), 34/99 (34%) LAM (week 52), and 8/47 (17%) PLA (week 52).
By week 52, HBsAg loss had occurred in 2/119 (2%) LAM and 4/63 (6%) LAM+IFN and, at week 68, in 1/60 (2%) LAM/lAM, 2/59 (3%) LAM/PLA, and 3/63 (5%) LAM+IFN.
Among patients with baseline ALT elevation (seven normalized between screening and baseline), significantly more LAM (51/115, 44%) experienced sustained ALT responses through week 52 than PLA (8/54, 15%; P<0.001) and LAM+IFN (11/62, 18%; P=0.005). Sustained ALT responses to the end of study (week 68) were higher in LAM/lAM (22/58, 38%) than LAM/PLA (11/57, 19%), PLA (5/54, 9%), and LAM+IFN (11/62, 18%).
At week 52, 27/99 (27%) of available serum from LAM patients, but none in PLA or LAM+IFN, had detectable YMDD-variant virus. Median HBV DNA and ALT levels at week 52 remained below baseline for patients with YMDD-variant HBV (HBV DNA 32 versus 161 pg/ml; ALT 71 versus 112 U/l, Fig. 6). In addition, 2/27 (7%) patients with YMDD variants achieved HBeAg seroconversion by week 52; at end-of-study week 68, one remained HBeAg-negative and both HBV DNA-negative.
The safety profile of lamivudine was similar to that of placebo during initial and extended lamivudine treatment beyond week 52. The proportion of patients experiencing adverse events was greatest in LAM+IFN, reflecting interferon side effect: malaise/fatigue, fever, headache, nausea/vomiting, hais loss, muscle pain, were the most common side effects for patients taking LAM/IFN.
Forty-five serious adverse events were reported in 24 patients; 32/45, 71% were judged by investigators as unrelated/unlikely-to-be-related to study treatment. Among events considered attributable/possibly attributable to study medication, no difference in pattern occurred between lamivudine and placebo recipients.
The frequency of laboratory abnormalities was similar among the three treatment groups. Grade 3-4 creatine phosphokinase abnormalities, but not clinical musculoskeletal pain, were slightly more common in LAM than PLA. Safety observations in patients who acquired YMDD variants were similar to those in patients who had not. Patients with YMDD variants were somewhat more likely to have a Grade 3-4 ALT elevation during treatment but less likely to experience posttreatment ALT elevation.
Elevations of ALT were asymptomatic, and no patients in any group experienced hepatic decompensation during or after treatment. During 52 weeks of study, patients in LAM+IFN were more likely to fulfill one of the first three ALT-elevation criteria, but in no treatment group did patients demonstrate ALT elevation plus bilirubin elevation. During the week 52-68 follow-up period, more patients in LAM/PLA exhibited ALT elevations, but none was clinically severe; one patient in LAM/lAM had an ALT elevation associated with transient bilirubin elevation that resolved before study completion.
Discussion by authors
In IFN nonresponders with chronic hepatitis B, 1 year of lamivudine improved liver histology and serum ALT levels and suppressed HBV replication, with good tolerability; thus, lamivudine-associated responses are not limited to patients who respond to IFN.
In contrast to treatment-naive patients, in these IFN nonresponders, HBeAg seroconversion was not statistically better in LAM than PLA. Nonetheless, HBeAg loss was more common in LAM than PLA (as in studies of treatment-naive patients), and such responses were generally durable posttreatment. Although HbeAg seroconversion had been postulated to be more durable than HBeAg loss, in this and two previous studies, posttreatment durability of these HBeAg responses was indistinguishable.
The longer-term durability of lamivudine-related HBeAg responses is being examined in an ongoing 5-year follow-up study. Preliminary results, after a median of more than 2 years, suggest that HBeAg loss/seroconversion and ALT responses remain durable posttreatment in the vast majority of HBeAg responders. Moreover, at a median of 2 years, 9/43 (21%) of enrolled HBeAg responders from lamivudine trials in the West have achieved durable HBsAg seroconversions.
Assessments during the week 52-68 follow-up treatment period in this study and after 2-3 years of continued treatment in other studies indicate that suppression of HBV replication can be maintained with continued lamivudine treatment in most patients beyond 1 year and that continued ALT improvements and incremental increases in HBeAg responses can occur. These findings support a strategy of continuing lamivudine in patients who fail to achieve HBeAg responses during the first year of therapy, to maintain HBV suppression and histologic improvement and provide a continuing chance for HbeAg responses. A strategy of treating until after an HBeAg response occurs is also supported by the results of two other studies; however, generalization of this recommendation to decompensated or immunosuppressed patients is not advisable.
Although combination LAM+IFN in this study was inferior to LAM in histologic and ALT responses, conclusions about the superiority of LAM monotherapy cannot be made with confidence. Constrained by regulatory considerations, we could not include an IFN-comparator regimen that had not been approved. Because the United States approved duration of IFN therapy for chronic hepatitis B is only 16 weeks, we had to limit the duration of IFN in the IFN arm to four months. Therefore, we selected a period of HBV DNA reduction with lamivudine for 2 months prior to beginning IFN. Certainly, continuing lamivudine to week 52 in LAM+IFN would probably have improved histologic results for combination therapy. In a Phase III study involving interferon-naive patients, HBeAg responses at week 52 were more common with a similar LAM+IFN regimen, although HBeAg seroconversion was not statistically different among the lamivudine, IFN, and combination-therapy
groups, and other efficacy responses (histologic, ALT, and virologic) were similar or better in the lamivudine monotherapy arm. A hint in our study that the combination of lamivudine plus interferon might yield added benefit derived from the higher proportion in LAM+IFN (54% at 24 weeks) than in LAM (34% at 52 weeks) of undetectable HBV DNA by PCR at the end of therapy; however, this analysis was done retrospectively, and the study was not powered to compare these two groups for HBV DNA response. Thus, the results of study cited above and the present trial should not preclude investigations of other approaches to LAM+IFN therapy, and the results of year-long treatment trials with combinations of lamivudine plus pegylated IFN are anticipated.
This study demonstrated an excellent safety profile for lamivudine in IFN nonresponders. Among the four treatment regimens we investigated, the continuous 68-week lamivudine assignment exhibited the lowest proportion of patients with clinical adverse events or Grade 3-4 laboratory abnormalities. During treatment, LAM and PLA had a similar safety profile, while LAM+IFN had more adverse events, typical of interferon therapy. Safety data from the follow-up period indicated that mild-moderate, but not severe, posttreatment ALT elevations were more common in patients in LAM/PLA than LAM/lAM and PLA/PLA, reinforcing findings among treatment-naive patients. Accordingly, posttreatment monitoring is advisable for several months after discontinuing lamivudine.
The incidence of YMDD-variant HBV in this study was higher than that reported for Asian patients but similar to that for treatment-naive Western patients. As in previous studies, patients with YMDD-variant HBV were less likely to exhibit HBeAg responses and maintain undetectable HBV DNA and normal ALT levels but, nonetheless, tended to maintain improved HBV DNA and ALT levels.
The results of previous IFN and lamivudine trials demonstrated treatment benefits for IFN-naive patients, and this trial demonstrates that IFN nonresponders have lamivudine as a treatment option. Optimal treatment regimens for lamivudine and IFN continue to be studied, and progress on additional treatment options continues to be reported. Ultimately, treatments that reduce HBV replication should have the anticipated impact on preventing complications of chronic HBV infection.