Four years of treatment with lamivudine: clinical and virological evaluations in HBe antigen-negative chronic hepatitis B


Dr S. Gaia, Molinette Hospital, C. Bramante 88, 10126 Turin, Italy.


Aim : To evaluate the clinical and virological impact of the prolonged use of lamivudine in 94 patients with HBe antigen-negative chronic hepatitis B.

Methods : Initial virological and biochemical responses were obtained in 84 (89%) and in 83 (88%) patients respectively.

Results : The virological response peaked within the first 12 months, but diminished to 39% at 48 months because of drug resistance. Overall a virological breakthrough developed in 44 patients (52.4%). After virological breakthrough, the actuarial probability of maintaining biochemical remission diminished to 15% at 24 months and 0% at 29 months. There was no response in 10.6%. Polymerase gene mutations were observed in 82.5% of virological breakthroughs but also in 75% of the non-responders. Overall 7.4% of patients developed a hepatocellular carcinoma.

Conclusion : Almost 90% of patients responded initially to lamivudine but the emergence of drug resistance progressively reduced the rate of virological remission to 39% at the fourth year of therapy. YMDD mutants explained the 75% of lamivudine resistances and were also selected very early in non-responders. Although the biochemical response is invariably lost within 29 months of the YMDD mutant's duration, the clinical outcome was benign despite severe postvirological breakthrough hepatitic flares in about 12% of cases.


Most information on the efficacy of lamivudine (Lam) in the treatment of chronic hepatitis B (CHB) has been derived from studies on patients with disease sustained by wild-type hepatitis B virus (HBV).1, 2 In this setting therapeutic remissions are marked by durable seroconversion from HBe antigen (HBeAg) to antibodies to HBeAg (anti-HBe) and by the normalization of liver enzymes and clearance of serum HBV DNA.

However, over 90% of contemporary patients with CHB in the Mediterranean area lack HBeAg and many have instead anti-HBe in serum.3–5 The disease is sustained by mutants of HBV that contain nucleotide substitutions in the core/precore regions of the viral genome; it often runs a discontinuous but relentless course, punctuated by intermittent flares of viraemia and cytolysis.5–7

Only one major therapeutic trial and many small series open-label studies have been performed to establish the efficacy of Lam in HBeAg-negative CHB (e-CHB).8–16 In this setting, the efficacy of therapy is difficult to assess, as the only parameters of response are the normalization of aminotransferase and the disappearance of serum HBV DNA; in addition the intermittent nature of the disease makes it difficult to distinguish spontaneous from therapy-induced remissions.

Not surprisingly, the indications to use of Lam in e-CHB are poorly defined.11–15 Although the indication has emerged to give the drug over the long-term in order to control reactivation of viraemia and associated necroinflammation, the efficacy, safety and ultimate clinical impact of long-term therapy with Lam in unselected populations of e-CHB is as yet unknown.

In this study, we report our experience with the prolonged routine use (up to 4 years) of Lam in 94 Italian patients affected by e-CHB.

Patients and methods


Between June 1998 and December 2001, 94 consecutive adult patients from San Giovanni Battista Hospital (Turin, Italy) with e-CHB were treated with Lam 100 mg dose daily. They were followed up until 31 January 2003 and were included in this retrospective analysis. Median duration of treatment was 33 months, between 12 and 54.

The demographic and clinical features of patients are outlined in Table 1.

Table 1.  Characteristics at baseline of 94 patients treated with lamivudine
  1. * Times upper limit of normal.

  2. † 1+, 1–10 pg/mL; 2+, 11–100 pg/mL; 3+, 101–1000 pg/mL; 4+, >1000 pg/mL.

  3. ‡ According to Ishak score.

Age (years, mean ± s.d.)46 ± 9.8
Gender (male/female)75/19
Chronic hepatitis (total number/biopsies)‡72/50
Cirrhosis (total number/biopsies)‡22/9
ALT (×ULN*) median (range)3 (1.5–20)
HBV DNA hybridization, median (range)1+ (1+ to 4+)†
Naïve/previous treatment59/35

Criteria for inclusion in the analysis were: (i) HBsAg-positive for at least 6 months, HBeAg-negative and anti-HBe-positive; (ii) serum alanine aminotransferase (ALT) values >1.5 times the upper limit of normal (ULN) in at least two determinations during the 6 months preceding therapy; (iii) positive serum HBV DNA, tested by hybridization assay, non-amplified technique (>10E5 copies/mL); (iv) CH or compensated liver cirrhosis. We did not include patients co-infected by the HCV, HDV and HIV.

Chronic liver disease was confirmed by a liver biopsy taken within 18 months before therapy in 59 patients: 50 had chronic active hepatitis, nine had cirrhosis. Among the 35 patients for whom no liver biopsy was available, 13 were considered to have cirrhosis (nine patients Child-Pugh A and four Child-Pugh B grade), diagnosed on the basis of signs of portal hypertension by upper endoscopy (six patients had oesophagus varices) or ultrasound. The remaining 22 patients had persistently abnormal ALT levels without clinical signs of cirrhosis. They refused liver biopsy before therapy.

Thirty-one patients had received prior treatment with α-interferon (IFN): eight had relapsed during therapy after an initial response and 23 had not responded. Three patients had received prior treatment with Famciclovir and one with Entecavir. All previous treatments were discontinued at least 6 months before Lam therapy. None patient was previously treated with Lam.


Follow-up.  The patients were followed-up at 3-month intervals. Before therapy and at each clinic visit routine haematological, biochemical and virological tests [HBsAg, HBsAb, HBeAg, HBeAb, IgM-antiHBc and HBV DNA by polymerase chain reaction (PCR)] were performed. The patients who had a virological breakthrough (VBK) and those who did not respond to therapy were tested for HBV polymerase mutants by Inno-Lipa HBV DR assay. Seventeen patients with different therapy outcomes were analysed retrospectively for sequence variations in the core-promoter, precore and polymerase rt-domain region B/C at baseline, and during therapy (at the appearance of VBK and at the end of follow-up).

All patients underwent ultrasound of the upper abdomen every 6 months.

Paired liver biopsies taken at baseline and during Lam therapy were obtained in 20 patients.

Therapy outcome definitions

The outcome was defined as (i) virological response (VR): negative HBV DNA by PCR for at least two consecutive determinations, (ii) biochemical response (BR): decrease in serum ALT to within the normal range during therapy, (iii) VBK:8 HBV DNA-positive by PCR (>1000 copies/mL) for at least two consecutive determinations after an initial VR, (iv) no response (NR): persistence of HBV DNA viraemia detectable by PCR on-therapy.

Serology.  Serum HBV markers (HBsAg, HBsAb, HBeAg, HBeAb, IgM anti-HBcore), anti-HCV, anti-HDV and anti-HIV were tested with immunoenzymatic assays (EIA; Abbott Diagnostic Inc., Chicago, IL, USA). HBV DNA load was performed by Dygene Hybrid-Capture I (sensitivity 700 000 copies/mL) at baseline, and during therapy by an amplified technique, PCR (Amplicor HBV, Monitor; Roche, Branchburg, NJ, USA; sensitivity of 1000 copies/mL).17

HBV polymerase mutants assay.  Polymerase mutants were determined using a commercially available kit (Inno-LiPa HBV DR, Innogenetics N.V., Belgium) based on reverse hybridization of PCR amplified HBV DNA against specific probes immobilized on paper strips.18, 19

Genotyping.  The HBV genotype was determined using a PCR-based method with primers specific for the six major genotypes (A–F).20

Statistical analysis

Patients who were lost at follow-up were considered as failure of treatment. All the data were analysed by stats direct software, version 1.21. The Kaplan–Meier’ method21 was used to estimate the cumulative rate of biochemical and virological remission during the study period. Statistical significance analysis of the histological response was performed with the Wilcoxon rank sum test. A P-value of <0.05 was considered statistically significant.



Overall, an initial VR was observed in 84 of the 94 patients (89%) within 3–12 months of therapy (median time 6 months). The actuarial rate of VR at month 12 was 82% and decreased linearly to 39% at month 48, for the development of Lam resistance (Figure 1).

Figure 1.

Virological and biochemical responses over 48 months of therapy in 94 patients with chronic hepatitis B (e-CHB) treated with lamividine.

The initial BR was observed in 83 patients (88%); it was achieved within 33 months from the start of therapy (within 12 months in the majority). The BR was achieved 0–27 months (median 3 months) after the VR in 76 subjects and preceded it by 3–6 months in seven patients. The actuarial rate of BR peaked to 83% at month 12 and decreased progressively to 52% at month 48 (Figure 1).

There was correspondence between the VR and the BR in all except three patients, in whom aminotransferase diminished during therapy but remained higher than 1.5 times ULN. These patients had other factors of hepatic morbidity (body mass index >26, dislipidaemia and diabetes mellitus).

Paired liver biopsies obtained at baseline and while on response were available in 13 VRs; in these patients the second biopsy was performed between 17 and 36 months after starting Lam (median time 30 months). Histology showed an improvement of the necroinflammatory score (Ishak score) in 12 of 13 (92.3%) (>3 points in 10 patients, >1 point in two subjects). The medium histological grading score at baseline was 7.2 points and it decreased to 2.2 during treatment (P < 0.002). An improvement of the fibrosis score was seen in 7 of 13 patients (53.8%) and worsening in two; fibrosis remained unchanged in four patients (P = 0.006) (Figure 2).

Figure 2.

Histological modifications in 20 chronic hepatitis B (e-CHB) patients, with paired liver biopsy during therapy with lamivudine. Number of patients in brackets. NR, no response; VBK, virological breakthrough.

During treatment one patient, responsive to therapy, lost HBsAg and developed anti-HBsAg antibodies; he stopped Lam and maintained a sustained response after discontinuing the drug.


Ten patients (10.6%) were NRs (median time of therapy 17.5 months, range 12–45). All of them had a partial decrease of viraemia during the first month of therapy and in seven (70%) the reduction of HBV DNA was associated with a temporary reduction of hepatic cytolysis. During follow-up ALT remained less than two times ULN in four patients, between three and nine times ULN in five subjects, and raised to 1126 IU/L in one patient. We looked for a hidden no-compliance with an interview to the patient to exclude false-resistance to anti-viral.

Before therapy two NRs analysed had the wild-type viral sequences in the polymerase gene (data not shown). During therapy YMDD mutants were tested in eight NRs; in six (75%) there were the usual polymerase mutations in rt-domain region B and C and in two patients (25%) the sequences were wild type (Table 2). The pattern of hepatic cytolysis was similar in patients with YMDD mutants and wild-type Lam non-responsive virus.

Table 2.  Different rt-domain mutations (polymerase gene) in non-responders (NRs) and in virological breakthroughs (VBKs) tested with Inno-Lipa HBV DR, during lamivudine therapy
Polymerase mutationrt 204M (wild) (n)rt 204V (n)rt 204I (n)rt 204I + V (n)
  1. * In one VBK patient there was no hybridization of polymerase chain reaction (PCR) product in position rt 204.

NRs (8 pts)
 rt 180L (wild)2120
 rt 180M0120
VBKs* (40 pts)
 rt 180L (wild)62133
 rt 180M1932

Virological breakthrough

In the 84 responders the efficacy of therapy diminished over time because of the emergence of drug resistance. Overall a VBK developed in 44 patients (52.4%). During therapy the VBK occurred in 8% (seven of 84 responders) within the first year of therapy and increased to 34% (25 of 74), 40% (17 of 42) and 59% (13 of 22) of the responders within the second, third and fourth year of therapy, respectively. The median time of VBK development was 15 months (4–42 months) after the initial VR.

The actuarial probability of maintaining a biochemical remission after the VBK onset decreased during time, thus 29 months from VBK all patients showed ALT elevation (Figure 3). However, in 11 (25%) VBKs there were no evidence of ALT elevation during the interim of their individual follow-up (median time 10 months after VBK, range 0–24). In those who lost biochemical remission, the ALT elevation preceded VBK by 3 months in three patients, it was simultaneous in 10 and followed the HBV resistance by 3–29 months in the majority (20 patients). After the relapse the value of ALT ranged from mild (<3 UNL) in 25 of 32 (78%), to intermediate (3 < ALT < 10 UNL) in three of 32 (9%), to high (ALT > 10 UNL) in four of 32 (13%). No patient experienced jaundice or liver decompensation.

Figure 3.

Probability of biochemical remission over time in 44 chronic hepatitis B (e-CHB) patients who developed virological breakthrough (VBK) during lamivudine therapy (Kaplan–Meier's method).

Before therapy all the eight VBKs analysed were carriers of the wild-type viral sequences in the polymerase gene (data not shown). Polymerase mutants were screened while on-therapy in 40 VBKs by Inno-LiPa, results are reported in Table 2.

Results of liver biopsy performed after a median time of 17 months from the emergence of VBK in five patients are reported in Figure 2.

Comparison of HBV sequences between responders, VBK patients and non-responders

All 17 patients resulted genotype D, 16 were serotype ayw and one was serotype adw. The precore, core-promoter, polymerase regions at baseline and during therapy were compared in 17 patients (seven responders, two NRs and eight VBKs) (data not shown). No correlations were detected between the finding of core-promoter/precore mutations at baseline and the emergence of YMDD during therapy, or the type of response to Lam.22

Adverse effects and withdrawal

Generally Lam was well-tolerated. A mild increase in amylasaemia (less than two times ULN) was observed in 14 patients.

Seven patients (two responders and five VBKs) discontinued therapy, for severe hepatic cytolysis after VBK (four patients), pregnancy (one VBK), acute pancreatitis (one responder) spontaneous discontinuation (one responder). After Lam discontinuation, transaminases relapsed in VBKs and in one responder while remained normal in the other responder.

Hepatocellular carcinoma

All 94 patients had a normal alfa-feto protein serum level before Lam treatment. Seven patients (7.4%) developed hepatocellular carcinoma (HCC) during treatment; the tumour was diagnosed after a median time of 30 months (range: 8–39) from starting therapy. Six patients had clinical or histological evidence of cirrhosis and one had histological findings of CH. Two died of liver decompensation because of the neoplasm.

At the time of HCC development five patients were VBKs (median time of HCC development after VBK evidence: 11 months) and two were NRs (polymerase mutation unknown). None of the virological responders to Lam developed HCC (P < 0.05).


This study provides data on long-term therapy with Lam in our region, based on the routine use of the drug in patients with e-CHB collected in an area where the disease is endemic. All patients included in this study had serum HBV DNA-positive by a non-amplified method regardless of low or high replication. However, the treatment was given to patients who had histological and/or clinical evidence of severe or progressive compensated CHB. We have not considered decompensated cirrhosis; these patients were treated with different protocols primarily aimed at decreasing viraemia with the prospect of liver transplantation.23

Overall 39% of our patients were in virological remission and as many as 52% were in biochemical remission at the end of their follow-up. Like in other studies,7–10 the VR was rapid; it occurred on average after 6 months of therapy within 12 months. The VR was associated with a biochemical remission in all but three patients; these patients had additional non-viral factors of liver damage that may have caused the increase in the liver enzymes. In responders for whom paired liver biopsies were obtained histology showed that the inflammatory and fibrosis scores had significantly improved. However, the small number of paired biopsies cannot permit to have definitive evaluation, but only to observe a trend of histological improvement in patient responding to Lam.

Good cumulative results are nonetheless misleading unless confronted with the time of therapy duration. The cumulative VR was 89%; it peaked to a zenith of 83% after 1 year but due to the emergence of resistance to Lam it then declined to 66.2% in the second year of therapy, to 59.6% in the third and to 39% in the fourth.

When the cumulative rate of VBK (52.4%) is split over the duration of therapy, the rate of resistance increased likewise from a nadir of 17% in the first therapy year to a zenith of 59% in the fourth year of therapy.

These data are not generally different from those collected in investigational studies but the large number of patients enrolled and the prolonged follow-up make our study suitable to provide a more precise analysis of the time restraints to the response to Lam.

This is particularly pertinent to the issue of the clinical impact of the selection of YMDD mutants. Their role is currently controversial;8, 24–28 early studies based on the limited duration of the mutants suggested that they had only a minor pathogenic significance but subsequent studies based on more prolonged YMDD duration have indicated that they became pathogenic over time.8 In our experience, in the YMDD-carriers the rate of biochemical reactivation increased linearly and progressively throughout the months of YMDD duration. Ultimately 27% of the patients with VBK lasting for over 9 months exhibited normal ALT values vs. 100% of the VBK-free long-term responders, treated for the same length of time. These data confirm that the emergence of YMDD mutants implies that therapy is failing; from available data it is reasonable to assume that the disease will relapse in all YMDD patients with mutants enduring for longer than 2 years. In our patients aminotransferase rose to a peak of variable magnitude 0–29 months after the initial VBK and then remained abnormal throughout the subsequent follow-up. However, the immediate clinical impact of the biochemical breakthroughs was minor; although the ALT flare reached hepatitic levels (>10 UNL) in four patients, liver function decompensated in none.

Of note, among the patients with VBK, YMDD mutants were selected in 82.5%, but in the remaining 17.5% resistance was not associated with the emergence of these mutants; analysis of the polymerase region in eight of the latter did not disclose any significant mutation from the wild type sequence. Conversely, the analysis of the rt-domain disclosed YMDD mutants in 75% of the minority of the patients who did not respond to Lam, with only 25% of them harbouring wild-type virus.

The sum of virological data indicates that although the emergence of YMDD mutants is responsible for the largest loss of response to Lam, in 17.5% of our patients loss of response was not related to mutations in the YMDD locus of the polymerase region; therefore, if present, they are to be sought in other viral domains.

More surprising, primary non-response to Lam appeared also to be associated in the majority of cases with the presence of YMDD mutants and only a few patients were refractory to the drug for intrinsic resistance of the wild-type virus. In this setting, YMDD mutants could be presumably selected very early under pharmacological pressure rather than pre-existent to therapy; YMDD mutants were not found in any of 17 patients in whom the rt-domain was sequenced at baseline, including two of the non-responders. Neither response nor resistance to Lam could be predicted at baseline from the virological structure of the precore/core-promoter regions,22, 29 which were not different in patients who responded, did not respond or experienced a VBK.

Seven patients (7.4%) developed HCC during therapy and two of them died of liver decompensation related to the neoplasm. All of them had Lam resistance, while none responder developed HCC. Hepatocellular carcinoma was more frequent in patients with diagnosis of cirrhosis than in those with CH, as generally reported.30, 31 In our study, six of seven patients who developed HCC had cirrhosis, and all of them had active HBV replication. Further, longer studies are awaited to better understand the occurrence of this severe complication during Lam therapy.

In conclusion, Lam suppresses viral replication leading to normalization of ALT and improvements in liver histology in most patients with e-CHB. Therapy needs to be continuous but continued therapy is associated with the emergence of Lam-resistant mutants that recapitulate disease. Continuous Lam therapy should be given to e-CHB patients who have evidence of progressive (over a few years) disease, whose prognosis is poor and in whom the benefits of therapy outweigh the risk of YMDD emergence. As Adefovir Dipivoxil has the potential to inhibit replication of YMDD mutants and hence avert the resulting disease, it may provide rescue therapy in this setting.


Financial support: no grants or other financial support was used for this work.