Article Text

Download PDFPDF

Tenofovir disoproxil fumarate rescue therapy following failure of both lamivudine and adefovir dipivoxil in chronic hepatitis B
  1. S J Patterson1,
  2. J George2,
  3. S I Strasser3,
  4. A U Lee4,
  5. W Sievert5,
  6. A J Nicoll6,
  7. P V Desmond7,
  8. S K Roberts8,
  9. S Locarnini9,
  10. S Bowden9,
  11. P W Angus1
  1. 1Austin Health, Victoria, Australia
  2. 2Westmead Hospital, NSW, Australia
  3. 3Royal Prince Alfred Hospital, NSW, Australia
  4. 4Concord Repatriation General Hospital, NSW, Australia
  5. 5Monash Medical Centre, Victoria, Australia
  6. 6Royal Melbourne Hospital, Victoria, Australia
  7. 7St Vincent's Hospital, Victoria, Australia
  8. 8The Alfred Hospital, Victoria, Australia
  9. 9Victorian Infectious Diseases Reference Laboratory, Victoria, Australia
  1. Correspondence to Dr Scott Patterson, Liver Transplant Unit, Austin Health, Studley Road, Heidelberg, VIC 3084, Australia; scott.patterson{at}


Objective To determine the efficacy of tenofovir disoproxil fumarate (TDF) in adults with chronic hepatitis B virus (HBV) infection who had previously failed lamivudine (LAM) and had significant viral replication (HBV DNA >105 copies/ml if HBeAg positive, >104 copies/ml if HBeAg negative) despite at least 24 weeks of treatment with adefovir dipivoxil (ADV).

Design A prospective open-label study of TDF 300 mg daily. Patients receiving combination ADV/LAM prior to baseline were switched to TDF/LAM.

Setting Multiple tertiary referral centres.

Methods Sixty patients were enrolled. The median age was 48.5 years (range 21–80), 46 (77%) were male and 40 (67%) were HBeAg positive. Thirty-eight patients (63%) were switched from ADV to TDF, the remainder from ADV/LAM to TDF/LAM. At baseline, substitutions conferring resistance to LAM or ADV were present in 20 patients (33%) and 17 patients (28%), respectively. The median baseline viral load was 5.33 log10 IU/ml (range 2.81–8.04). Patients initially treated with TDF monotherapy with persistent viral replication at or after 24 weeks were switched to TDF/LAM. The main outcome measures were change in HBV viral load from baseline and percentage of patients achieving an undetectable viral load (<15 IU/ml).

Results Results are reported at 96 weeks of treatment. One patient discontinued TDF at 10 days due to rash. The time-weighted change in viral load from baseline to week 12 was −2.19 log10 IU/ml overall. The median change in HBV DNA from baseline to weeks 12, 24, 48 and 96 was −2.86, −3.23, −3.75 and −4.03 log10 IU/ml, respectively. At 48 and 96 weeks, 27/59 (46%) and 38/59 (64%) patients achieved a HBV DNA <15 IU/ml. The response was independent of baseline LAM therapy or mutations conferring ADV resistance.

Conclusions In heavily pretreated patients with a high rate of genotypic resistance, TDF retains significant activity against HBV although this appears diminished in comparison with studies of naïve patients.

  • Antiviral
  • resistance
  • rtN236T
  • rtA181T/V
  • antiviral therapy
  • hepatitis B
View Full Text

Statistics from

Significance of the study

What is already know about this subject?

  • Long-term anti-HBV nucleos(t)ide therapy is effective in preventing progressive HBV disease but may be complicated by resistance.

  • Prior to the availability of tenofovir disoproxil fumarate, there was no ideal agent for use in patients with previous failure or resistance to both lamivudine and adefovir dipivoxil.

  • Tenofovir disoproxil fumarate has been shown to be effective in adefovir-resistant HBV; however, prospective studies are lacking.

What are the new findings?

  • In this prospective study of 60 heavily pretreated patients, all with prior failure of lamivudine and suboptimal response to at least 24 weeks of adefovir and a high rate of genotypic resistance, tenofovir retained significant activity against HBV.

  • 64% of patients achieved an undetectable viral load with 96 weeks of therapy. This is inferior to that observed in studies of naïve patients and in previous studies of tenofovir rescue therapy.

  • Although no association between response and baseline adefovir resistance mutations (rtA181T/V and rtN236T) was observed, such mutations became detectable over the course of the study in four patients.

How might it impact on clinical practice in the future?

  • This study confirms that tenofovir disoproxil fumarate should be the preferred oral anti-HBV agent (of those currently available) in patients with prior failure of or resistance to lamivudine and adefovir.

The development of safe effective oral antiviral agents has revolutionised the management of chronic hepatitis B. Although suppression of viral replication with these drugs has been shown to alter the natural history of the disease and diminish the incidence of complications, long-term efficacy is compromised by the emergence of drug resistance.1–3 This is a major problem with lamivudine, with up to 70% of patients developing resistance after 4 years of treatment.4

Adefovir dipivoxil has been widely used in the treatment of patients with lamivudine-resistant hepatitis B virus (HBV) infection.5 However, up to 25% of patients treated with adefovir for lamivudine-resistant HBV infection fail to achieve a satisfactory virological response and 30% of naïve patients develop adefovir resistance at 5 years.6 Until recently, the treatment options for patients who fail sequential lamivudine then adefovir monotherapy have been limited. One option is entecavir, a drug that has an excellent resistance profile in treatment-naïve patients even with long-term therapy.7 This profile is compromised in the presence of lamivudine resistance and, in this setting, genotypic entecavir resistance is observed in >50% of patients at 5 years. Another option is tenofovir disoproxil fumarate (DF), an acyclic nucleotide analogue and structural congener of adefovir that has potent antiviral activity against wild-type HBV.8 In treatment-naïve patients, tenofovir DF durably suppresses HBV and drug resistance has not been encountered with up to 3 years of continuous therapy.9 Despite its structural similarity to adefovir, tenofovir DF partially inhibits adefovir-resistant HBV and it is also highly effective against lamivudine-resistant virus, suggesting that this drug may be an effective treatment for patients who have previously failed treatment with lamivudine and adefovir.10–12

We therefore conducted a prospective trial of tenofovir rescue therapy in patients with lamivudine-resistant HBV who failed to achieve an adequate response to rescue therapy with adefovir. The study included both patients who had received adefovir as monotherapy and those who had received combination therapy with lamivudine.


This was a prospective open-label multicentre trial that recruited patients from eight tertiary referral hospitals in Victoria and New South Wales, Australia from September 2006 to December 2007. The relevant institutional ethics review committees approved the trial at each site and all subjects gave written informed consent prior to screening for the study.

Adult patients were considered for the study if they satisfied all of the following criteria: chronic hepatitis B (defined as detectable hepatitis B surface antigen (HBsAg) for at least 6 months); previous failure of lamivudine therapy due to the emergence of drug resistance; current treatment with adefovir dipivoxil (with or without lamivudine) for at least 24 consecutive weeks leading up to baseline; and the presence of significant persistent or breakthrough viraemia. The threshold used to define significant viraemia was 5 log10 copies/ml in hepatitis B e antigen (HBeAg) positive patients and 4 log10 copies/ml in those who were HBeAg negative, equivalent to ∼4.24 log10 IU/ml and ∼3.24 log10 IU/ml, respectively. Patients remained eligible if they were receiving lamivudine and adefovir concurrently at screening, but only if this regime had been in place and unchanged for at least 24 weeks. In addition, patients were required to have a creatinine clearance of >50 ml/min (estimated by Cockcroft–Gault calculation), haemoglobin >8 g/dl, neutrophil count >1000/mm3 and alanine transaminase (ALT) not elevated to more than 10 times the upper limit of normal.

Patients were excluded if they had previously used tenofovir DF; had recently (<6 months) used any other agent with anti-HBV activity other than adefovir or lamivudine; had coexisting hepatitis C, hepatitis D or HIV infection; had evidence of hepatocellular carcinoma; had decompensated liver disease or significant renal, cardiovascular, respiratory or neurological comorbidity; or were pregnant, breastfeeding or unwilling to use contraception during the study. Patients potentially satisfying the inclusion and exclusion criteria were offered participation in the study and then screened. If they continued to satisfy the inclusion criteria, a baseline visit was scheduled within 30 days of screening.

At the baseline visit, patients commenced treatment with 300 mg tenofovir daily (Gilead Sciences, East Melbourne, Victoria, Australia). After the baseline visit the dosage of tenofovir was adjusted according to estimated creatinine clearance in accordance with the manufacturer's guidelines. Those taking lamivudine and adefovir prior to baseline continued on lamivudine treatment in addition to tenofovir DF throughout the study. The remainder received tenofovir DF monotherapy with the option of commencing lamivudine add-on therapy if they had a suboptimal response as defined by a viral load >351 IU/ml (or 2.55 log10 IU/ml, the lower limit of detection by the screening assay VERSANT HBV DNA 3.0 bDNA, Siemens Healthcare Diagnostics, Tarrytown, New Jersey, USA) at or beyond 24 weeks.

Patients were reviewed (physician assessment, physical examination, urine pregnancy test for women of childbearing potential, study medication accounting) at 4, 8 and 12 weeks and then at 12-weekly intervals. At each study visit blood was taken for serum biochemistry (particularly ALT, creatinine, total bilirubin and albumin), HBsAg and hepatitis B surface antibody (anti-HBs), HBeAg and anti-HBe and measurement of HBV DNA. In addition, baseline serum was tested for HBV genotype and resistance sequence analysis was performed on the HBV polymerase gene. Resistance sequencing was also performed every 48 weeks during the study or upon virological breakthrough (confirmed >1 log10 IU/ml increase in viral load from nadir).

The primary end point of the study was change in HBV load expressed as both median fall in load from baseline and as a time-weighted change from baseline to week 12 (DAVG12). Secondary end points of the study included the proportion of patients achieving a viral load <15 IU/ml by the Abbott RealTime HBV assay (Abbott Molecular, Des Plaines, Illinois, USA), the incidence of virological breakthrough or genotypic resistance, the incidence of HBeAg seroconversion and safety, including effects on renal function. This analysis of study outcomes was performed after 96 weeks of follow-up in all patients.

HBV assays

Viral markers (HBsAg, anti-HBs, HBeAg and anti-HBe) were measured by standard commercial immunoassays. HBV loads at screening, baseline and over the course of the study were assessed by either the VERSANT HBV DNA 3.0 bDNA assay or the Abbott RealTime HBV PCR assay according to the local reference laboratory. The lower limits of detection (LLOD) for these assays are 351 IU/ml and 15 IU/ml, respectively. Viral loads initially assayed by the bDNA assay with a result below the LLOD were repeated using the Abbott HBV DNA assay on stored serum. All results reported in copies/ml were converted to IU/ml using the manufacturer's recommended conversion factor.

HBV genotyping and sequencing of the polymerase/envelope region was performed in a single reference laboratory as previously described.13 Successful HBV genotyping by this method requires a viral load of approximately log10 2.55 IU/ml (2000 copies/ml). The identification of a quasi-species population with a polymerase sequence variant requires the presence of a similar quantity of virus.

Statistical methods

Continuous variables are expressed as median (range) unless specified otherwise. Categorical data are presented as number (%). Time-weighted average change in viral load (DAVGt) was calculated by dividing the area under the response–time curve from the first post-baseline measurement to the measurement at time t by the time between those two measurements, and then subtracting the baseline viral load. DAVG is a more robust measure of change in viral load than a simple difference in medians. p Values <0.05 were considered statistically significant; the statistical test used is specified with the corresponding result.


Between September 2006 and December 2007, 67 patients were screened for the study in eight sites. Six patients were ineligible on screening (viral load below inclusion criteria) and one was lost to follow-up; the 60 remaining patients were enrolled in the study and received tenofovir DF. The baseline (first day of study medication) characteristics of the study population are shown in table 1. When subgrouped by HBeAg status or treatment with lamivudine at baseline, the subgroups were well matched with the exception that there were fewer patients of European ethnicity in the HBeAg negative subgroup than in the HBeAg positive subgroup (3/40 vs 6/20, p=0.048, Fisher's exact test). The baseline viral load did not differ significantly between these subgroups.

Table 1

Baseline patient demographics and characteristics

All patients had been exposed to lamivudine and adefovir. The median duration of drug exposure was 37.5 months (range 8.2–102.5) for lamivudine and 21.9 months (range 6–85.4) for adefovir. In addition, six patients had been previously treated with standard or pegylated interferon for a median of 6.1 months (range 4–12) and three patients had received entecavir for a median of 9.1 months (range 8.1–30.4).

Genotyping and resistance sequencing was performed on baseline serum from 59 patients (one patient exited the study after 10 days (see below) and therefore stored baseline serum was not assayed). Sequence mutations conferring resistance to adefovir and lamivudine were detected in 17/59 (28.8%) and 20/59 (33.9%) patients, respectively. The presence of baseline polymerase sequence mutations conferring antiviral resistance is detailed in table 2. The most frequent genotype was C in 31 patients (52.5%), followed by D (25.4%), B (13.6%) and A (8.5%).

Table 2

Presence of baseline polymerase sequence mutations conferring nucleos(t)ide resistance

Median viral load at baseline did not differ significantly according to the presence of rt181T/V or rtN236T (either separately or in combination when compared with the remaining patients in the study). The presence of the rtM204I/V mutation both alone and in combination with rtL180M was associated with a lower median baseline viral load compared with the remaining patients in the study (median 4.97 vs 6.09 log10 IU/ml, p=0.011 and median 4.94 vs 6.07 log10 IU/ml, p=0.019, respectively; table 3). The median baseline viral load showed no association with HBV genotype.

Table 3

Median change in HBV viral load over study

Patient survival, withdrawal and follow-up

All 60 patients enrolled in the study were alive after 96 weeks. One patient developed a rash within 10 days of baseline and was withdrawn from the study and switched to alternative anti-HBV therapy. The following results do not include this patient; the minimum follow-up in the remaining patients is 96 weeks.

Effects of treatment on HBV DNA

The changes in viral load during the study, both overall and for the various subgroups, is shown in table 3 and in figures 1, 2 and 3A–C. The median change in viral load from baseline to week 12 was −2.86 log10 IU/ml. The median DAVG12 was −2.19 log10 IU/ml. All patients had a primary response to treatment defined as a viral load reduction of >1 log10 IU/ml by 12 weeks.14 With the exception of one subject, all had a fall of >2 log10 IU/ml in viral load by 24 weeks of therapy (this was achieved at 36 weeks in the remaining patient who had received tenofovir DF and lamivudine from baseline). The median viral load continued to fall throughout the course of the study reaching −4.03 log10 IU/ml at 96 weeks. Through 96 weeks of therapy there were no significant differences in the median change in viral load between HBeAg positive and HBeAg negative patients; patients receiving tenofovir DF versus those receiving tenofovir DF/lamivudine combination; patients with HBV genotype A, B, C or D versus the remainder; or according to the presence or absence of baseline sequence mutations (specifically rtM204I, combination rtM204I and rtL180M, rtA181T/V, rtN236T or combination rtN236T and rtA181T/V). Change in viral load from baseline was positively related to baseline viral load—that is, the half of the cohort with a greater than median baseline viral load had a greater median decline in viral load (p=0.039 at 24 weeks, p<0.001 at all study visits from 36 to 96 weeks, Mann–Whitney test).

Figure 1

Change in viral load and proportion of patients achieving <15 IU/ml in all patients. Dotted line denotes viral load median and 25th/75th percentiles. Shadowed area denotes proportion of patients achieving <15 IU/ml.

Figure 2

Individual viral load profiles and proportion of patients achieving <15 IU/ml for patients with adefovir resistance mutations. Individual patients are graphed (patients with combination substitutions rtA181T and rtN236T are graphed below for clarity). Shadowed area denotes proportion of patients achieving <15 IU/ml. (A) Presence of rtA181T/V alone. (B) Presence of rtN236T alone or in combination with rtA181T/V. LLOD, lower limit of detection (15 IU/ml).

Figure 3

Change in viral load and proportion of patients achieving <15 IU/ml by treatment. Dotted lines denote viral load medians and 25th/75th percentiles. Shadowed areas denote proportion of patients achieving <15 IU/ml. (A) Tenofovir (TDF) from baseline. (B) Tenofovir/lamivudine combination (TDF and LAM) from baseline. (C) Baseline tenofovir switched to tenofovir/lamivudine combination due to persistent viral replication at 24 weeks (subset of those shown in A). LLOD, lower limit of detection (15 IU/ml).

Overall, 27/59 (45.8%) and 38/59 (64.4%) patients achieved an undetectable viral load <15 IU/ml with 48 and 96 weeks of therapy, respectively (figure 1). As might be expected, persistent viral replication >15 IU/ml was associated with a higher baseline viral load at all time points to week 96 (p<0.003 at weeks 12–48, p<0.017 at weeks 60–84, p=0.029 at week 96, Fisher exact test, comparing patients with a baseline viral load either greater than or less than or equal to the median baseline value). Persistent viral replication was associated with baseline HBeAg positive status up to 48 weeks (p=0.021 at 24 weeks, p=0.01 at 36 and 48 weeks) but not at later time points. There was no significant relationship between persistent viral replication >15 IU/ml and the presence or absence of substitutions conferring antiviral resistance at baseline (or combinations of substitutions as described above). The change in viral load over time for individual patients with either of the adefovir resistance mutations rtA181T/V and/or rtN236T is shown in figure 2. The use of combination tenofovir DF and lamivudine therapy from baseline had no significant effect on persistence of viral replication >15 IU/ml (figure 3A,B).

A suboptimal virological response, defined at commencement of the study as a persistent viral load >351 IU/ml at 24 weeks, occurred in 25 patients, 9 of whom had received combination therapy from baseline. As per protocol, the remaining 16 patients had lamivudine added to tenofovir DF from 24 weeks. The change in median viral load for these patients (and the proportion who achieved a viral load <15 IU/ml) is shown in figure 3C. There was no significant difference in median change in viral load from baseline through to 96 weeks in these 16 patients compared with those patients receiving combination therapy from baseline (p>0.25 at all time points, Mann–Whitney test). As expected, the requirement for lamivudine add-on therapy due to a suboptimal response after 24 weeks was associated with persistent viraemia >15 IU/ml when compared with all patients receiving tenofovir/lamivudine from baseline (p<0.015 at 24–60 and at 96 weeks, p<0.05 at 72–84 weeks; figure 3B,C). When these 16 patients with a suboptimal response to tenofovir monotherapy at 24 weeks were compared with similar suboptimal responders who had received combination therapy from baseline (9 patients), there was no difference in the proportion achieving an undetectable viral load with ongoing therapy beyond 24 weeks.

Seven confirmed episodes of virological breakthrough were observed in five patients during the study to 96 weeks, three of which (two in the same patient) are thought to have occurred in the context of poor compliance. Following five of these episodes of breakthrough, the viral load declined below the previous nadir either spontaneously or with compliance to ongoing tenofovir DF. The other two episodes occurred at 84 weeks (near the end of the follow-up period). The viral load in both patients declined at 96 weeks, but not beyond the previous nadir. In all episodes, polymerase sequencing was performed and no unique polymerase resistance mutations were identified compared with baseline.

Sequence mutations on study

In 4 of the 59 patients substitutions in the HBV polymerase associated with resistance to adefovir emerged during the study. In another 4 patients mutations detected at baseline were persistently detectable at 48 weeks (and in 2 of these patients these mutations were further detected at week 96). This is shown in table 4. No other antiviral drug resistance-associated polymerase substitutions were detected.

Table 4

Sequence mutations over study (sorted by baseline viral load)

Biochemical and serological response

Median ALT fell during the course of the study from 49 IU/l at baseline (range 17–528) to 35 IU/l (14–62) at week 96. One HBeAg negative patient had an ALT >10 times the upper limit of normal at baseline (violating study inclusion criteria); however, ALT progressively fell in this patient during the study. Elevated ALT levels were present in 33/60 patients (55%) at baseline and this had normalised in 17/33 patients at 96 weeks (51.5% of those with initially elevated ALT).

Of the 39 HBeAg positive patients at baseline, 2 (5.1%) and 4 (12.8%) lost HBeAg at 48 and 96 weeks, respectively. None seroconverted to anti-HBe positive. No patient had lost HBsAg or seroconverted to anti-HBs to 96 weeks.

Safety and side effects

Serum creatinine did not change significantly over the course of the study from a median of 80 μmol/l (range 48–116) at baseline to 83 μmol/l (range 49–122) at 48 weeks and 85 μmol/l (range 64–118) at 96 weeks. No patient had an increase in serum creatinine of >0.5 mg/dl (44.2 μmol/l). The dosage interval of tenofovir DF was adjusted to alternate days in one patient between weeks 4 and 72 according to estimated creatinine clearance (latest follow-up in this patient 120 weeks).

Transient elevations in ALT to >2 times the upper limit of normal were seen in 13/59 patients (22%) during the study; however, all such elevations subsequently resolved with one exception. This patient developed an elevated ALT >10 times the upper limit of normal at 48 weeks in the context of using a herbal remedy. The viral load was undetectable (<15 IU/ml) at that time and liver biopsy histology was consistent with a drug reaction. The herbal remedy was ceased, tenofovir therapy continued and ALT fell to ∼2–3 times the upper limit of normal but has not normalised at 96 weeks.

There were no significant abnormalities in other laboratory parameters during the study and no serious adverse events considered to be related to study therapy.


In this study we prospectively demonstrate that tenofovir DF produces significant suppression of HBV replication in patients with lamivudine resistance who have failed to respond adequately to adefovir. Despite a background of extensive prior treatment and the presence of clinical resistance to lamivudine and/or adefovir in a significant proportion of patients, none experienced a primary non-response to the study drug and, after 2 years of follow-up, no patient developed persistent breakthrough or novel polymerase mutations. Importantly, the viral load continued to fall throughout treatment in the majority of patients, with the median viral load reduction being −2.86, −3.75 and −4.03 log10 IU/ml at 12, 48 and 96 weeks of therapy, respectively. However, as might be expected, the virological response to tenofovir DF in this study appears to be inferior to that observed in treatment-naïve patients. The key registration studies of tenofovir DF reported that 69% of 176 HBeAg positive and 91% of 250 HBeAg negative nucleos(t)ide-naïve patients achieved a viral load of <169 copies/ml (approximately <30 IU/ml) after 48 weeks of tenofovir.15 16 For comparison, in the current study 33.3% (13/39) of HBeAg positive and 70% (14/20) of HBeAg negative patients achieved a viral load <15 IU/ml (approximately <80 copies/ml) by 48 weeks (overall 27/59 or 45.8%).

In this study the effectiveness of tenofovir DF in achieving an undetectable viral load was also less than in previously published reports of nucleos(t)ide-experienced patients receiving tenofovir. A retrospective study of the response to tenofovir DF therapy in patients with viral breakthrough on lamivudine and persistent viral replication >104 copies/ml despite adefovir found that 19/20 patients (95%) achieved an HBV load <400 copies/ml within a median of 3.5 months.11 None of these patients had genotypic adefovir resistance. A subsequent larger retrospective report from the same group included 101 patients with persistent viral replication (20 additional patients were excluded for poor compliance or genotypic adefovir resistance) of whom 91% achieved a viral load <400 copies/ml after 48 weeks of treatment.17 In a separate report of 10 patients with genotypic adefovir resistance, only 2/10 (20%) achieved the same end point.18 Finally, another study randomised 105 patients (of whom approximately 60% had been exposed to lamivudine) with suboptimal response to no more than 96 weeks of adefovir to combination tenofovir DF/emtricitabine or tenofovir DF alone. Ten of these patients had baseline genotypic adefovir resistance.19 20 At 48 weeks, approximately 80% of patients in both treatment groups and 80% of patients with baseline genotypic adefovir resistance had achieved an HBV load <400 copies/ml. When the current study is analysed using an equivalent threshold of 400 copies/ml, 61% of patients achieved this end point by 48 weeks.

A possible explanation for the lower rate of virological response in our patients is their extensive prior exposure to lamivudine and adefovir and a high rate of genotypic resistance. The patients in the current study were heavily pretreated, all having had prior lamivudine and adefovir for a median of 3 and 1.75 years, respectively. Furthermore, 17 of the 59 patients (28.8%) had genotypic resistance to adefovir and 20 (33.9%) to lamivudine. This rate of baseline genotypic adefovir resistance is more than double that in previous reports of nucleos(t)ide-experienced patients receiving tenofovir.

rtA181T and rtN236T confer reduced sensitivity (a <5-fold reduction) to tenofovir in vitro (and a ∼10-fold reduction in susceptibility in combination).10 We have previously reported that patients with baseline rtN236T mutation were more likely to have persistent viral replication quantifiable by the bDNA assay (LLOD >351 IU/ml) at 48 weeks compared with patients without rtN236T (p=0.042).21 When a highly sensitive PCR-based assay was employed in the current study, the proportion of patients infected with HBV with this substitution who achieved an HBV DNA <15 IU/ml was no longer significantly different from those without (p=0.43 at 48 weeks and 0.085 at 96 weeks). However, when the individual profiles were examined in patients with baseline rtN236T substitution, five of seven remained HBV DNA positive at 96 weeks and in three of these patients there was no appreciable fall in viral load beyond 36 weeks (figure 2B). Within this context, it is of considerable interest that adefovir-resistant HBV variants were selected in four patients during the course of the study, and longer term follow-up will be required to determine the clinical and virological sequelae of these changes.22 A recent retrospective analysis of tenofovir rescue therapy suggested a blunting of the response to tenofovir in the presence of adefovir resistance substitutions. In this study, patients with baseline genotypic adefovir resistance had a significantly lower rate of achieving a viral load <400 copies/ml than the remainder of the cohort after a mean of 23 months of treatment (52% vs 100%).12 Although we did not show such a difference, it should be noted that the cut-off for detecting viral replication was significantly lower in the current study (15 IU/ml).

We also observed a very low proportion of patients achieving a serological end point with only 10% achieving HBeAg loss, none of whom seroconverted to anti-HBe positive (and there was no HBsAg loss or anti-HBs seroconversion). The likely explanation is that the patients in this study represented a poorly responding group who had previously failed to achieve serological end points with previous therapy.

Of note, the study design allowed for continuation of lamivudine from baseline in patients previously receiving adefovir/lamivudine, as well as for the addition of lamivudine after 24 weeks in patients with persistently significant viral replication on tenofovir DF alone. There was no difference over the study in the proportion of patients achieving an undetectable viral load or in the absolute fall in HBV DNA between patients receiving tenofovir DF alone and those receiving combination lamivudine and tenofovir DF from baseline. Of the 37 patients initially receiving tenofovir monotherapy, 16 had persistent significant viral replication at 24 weeks and were switched to combination lamivudine and tenofovir. These 16 patients started with a higher viral load (median 6.91 vs 5.32 log10 IU/ml); however, the absolute fall in viral load in this group was similar to that observed in patients receiving combination therapy from baseline (see table 3 and figure 3). As expected, significantly more of these patients defined by a suboptimal response had persistent viral replication >15 IU/ml beyond 24 weeks. However, when compared with similarly defined suboptimal responders receiving combination therapy from baseline, there was no difference in the proportion achieving an undetectable viral load (<15 IU/ml) between these two groups. Although the design of this study precludes conclusions being drawn regarding the optimum use of combination lamivudine and tenofovir, there is certainly no clear evidence that additional lamivudine in this setting provides a significant supplemental antiviral effect. However, we suggest that appropriate combination therapy should be employed where possible for the treatment of multidrug-resistant HBV in an attempt to reduce the risk of further drug resistance.

In conclusion, tenofovir DF appears to be safe and effective in patients with prior failure of lamivudine and a suboptimal response to (or failure of) adefovir therapy. Despite the high prevalence of drug resistance mutations in the study population, no patient failed to respond to treatment or had persistent rebound/breakthrough of viral replication following the initiation of tenofovir DF treatment. Importantly, novel mutations conferring resistance to tenofovir were not observed. However, the efficacy of tenofovir in this heavily pretreated population with a high rate of baseline genotypic adefovir resistance was inferior to that reported previously and further follow-up is required to determine if this will be associated with subsequent virological breakthrough.


View Abstract


  • See Commentary, p 148

  • Funding Gilead Sciences (East Melbourne Victoria, Australia) donated tenofovir disoproxil fumarate and provided research support for this study.

  • Competing interests Associate Professor Simone Strasser: participates as Speaker, in Gilead sponsored and Supported clinical trials (tenofovir) and as a member of a Gilead advisory board (tenofovir); Associate Professor Amanda Nicoll: research support from Gilead Sciences in 2004 (not ongoing); Associate Professor Stuart Roberts: member of Gilead advisory board; Professor Stephen Locarnini: research support from Gilead Sciences and Bristol Myers Squibb; Professor Peter Angus: member of Gilead advisory board and research support (see Grant support, below).

  • Ethics approval This study was conducted with the approval of the Austin Health human research ethics committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles