Presented in part: 7th Conference on Retroviruses and Opportunistic Infections, 30 January–2 February 2000, San Francisco (abstract 286).
Individuals coinfected with human immunodeficiency virus 1 (HIV-1) and hepatitis B virus (HBV) often receive treatment with an antiretroviral regimen including lamivudine. Lamivudine monotherapy for HBV may lead to drug-resistant mutations in a significant number of patients. The virological and biochemical responses of 8 patients coinfected with HBV/HIV-1 treated with both lamivudine and famciclovir were studied. Patients exhibiting HBV viral rebound at 1 year were analyzed for the emergence of HBV polymerase mutations. Only 1 patient had no prior exposure to lamivudine. Addition of famciclovir to the treatment regimen resulted in a median fall in HBV DNA level of 0.33 log10 at 3 months and an overall rise in HBV DNA level of 3 log10 at 12 months. The only patient in whom durable viral suppression and HBV e antigen seroconversion were noted began receiving lamivudine and famciclovir simultaneously. HBV polymerase gene sequencing identified resistance-associated mutations in 6 of 7 patients with viral rebound. Sequential nucleoside analogue therapy is unlikely to be successful in achieving long-term suppression of HBV replication, and combination therapy should be considered at treatment initiation.
Chronic hepatitis B virus (HBV) infection occurs in up to 20% of an HIV-positive population [1]. Many of these patients are positive for HBV e antigen (HBeAg), with a high level of circulating HBV DNA [2], and are at increased likelihood of developing cirrhosis and hepatocellular carcinoma. The introduction of highly active antiretroviral therapy (HAART) has led to a reduction in HIV-related morbidity and mortality [3], and more individuals are consequently at risk of long-term HBV-related complications.
Traditional therapies for hepatitis B such as IFN-α are poorly tolerated by patients infected with HIV and have low efficacy [4]. Lamivudine (3TC), the (-) enantiomer of 2′, 3′-dideoxy-3′-thiacytidine, is a potent inhibitor of both HIV [5] and HBV [6] and is well tolerated in all patient groups. In immunocompetent patients, administration of lamivudine results in HBeAg seroconversion rates of between 14% and 16% at 1 year [7, 8] and 27% at 2 years [9], and these rates continue to rise with increasing duration of therapy. Several studies have demonstrated the drug's efficacy in the HIV-infected population [4, 10, 11], with reports of similar seroconversion rates [11, 12]. In those who do not seroconvert to HBeAg antibody, viral rebound invariably occurs on cessation of therapy. Unfortunately, lamivudine monotherapy has been reported to cause the emergence of drug-resistant HBV after 1 year in 15%–32% of patients with chronic hepatitis B [7, 8, 13]. This incidence is lower than that of the emergence of lamivudine-resistant HIV, but the resistance similarly involves mutations within the homologous YMDD motif around the catalytic site of reverse transcriptase [14]. A recent virological analysis of patients coinfected with HIV-1/HBV from the CAESAR study reported resistance developing in 14% of these patients receiving lamivudine monotherapy for HBV at 1 year [15]. This may limit the usefulness of lamivudine as a long-term oral agent for the control of HBV, which is of particular concern for patients coinfected with HIV, for whom lamivudine is often prescribed as part of HAART regimens.
Another nucleoside analogue that has potent anti-HBV activity is famciclovir, the oral prodrug of penciclovir. Famciclovir produces dose-dependent falls in HBV levels in immunocompetent patients [16] and has been administered to patients undergoing orthotopic liver transplantation for decompensated cirrhosis related to hepatitis B, in order to decrease viral replication and limit graft reinfection [17]. An in vitro study of the lamivudine/penciclovir combination has shown its synergy in inhibiting HBV replication, a finding suggesting that dual nucleoside therapy may be superior to monotherapy [18].
We therefore analyzed the virological and biochemical responses to combination lamivudine/famciclovir therapy in 8 patients coinfected with HIV-1/HBV within our clinic, seven of whom had previously received lamivudine as the sole anti-HBV agent.
Among all HIV-1-positive patients attending the Chelsea & Westminster Hospital (London), 8 patients with HBV coinfection were identified who had been treated with famciclovir (500 mg b.i.d., in addition to HAART including lamivudine) as part of their clinical care. All were HBeAg-positive, and no patients were coinfected with hepatitis C virus or δ virus. Seven of the patients had previously received lamivudine, for a median of 12 months (range, 6–24 months). Six patients were receiving lamivudine at the time of commencement of therapy with famciclovir, and 1 (patient 7) had received it as part of a previous antiretroviral regimen. This patient had recently taken a “drug holiday” but recommenced HAART with lamivudine when famciclovir was added to the regimen. One patient (number 8) started receiving both drugs simultaneously and had no prior exposure to either; lamivudine was substituted for stavudine in his HAART regimen (table 1).
Liver function during therapy with lamivudine (3TC) and famciclovir: A, Median for 8 patients; B, Individual values for the eight patients. Error bars represent interquartile range for each point. ALT, alanine aminotransferase value.
A, Four measures of serum lactic acid levels (mmol/L) in 5 patients: first available, highest, before beginning second regimen, and after beginning second regimen. B, Four measures of weight (kg) in 5 patients: beginning first regimen, at time of discontinuing treatment, beginning second regimen, and during second regimen. See table 1 for descriptions of the regimens.
Patients characteristics and their previous highly active antiretroviral therapy (HAART), at time of initiation of famciclovir therapy.
Patients were routinely assessed monthly, and blood specimens were obtained (1) prior to initiation of lamivudine therapy, (2) when dual therapy with lamivudine and famciclovir was commenced, and (3) at 3, 6, and 12 months after this for measurement of serum alanine aminotransferase (ALT) levels (normal range, 0–37 IU/mL). HBV DNA levels, HBeAg, and antibodies to HBeAg were also measured at these time points. Serum samples were prospectively stored at -20̊C and were then retrospectively collected for further virological studies. HBV quantification was undertaken (Roche Amplicor PCR monitor [Roche Diagnostic Laboratories]; detection range, 400–4,000,000,000 copies/mL (2.6 log10 to 9.6 log10). CD4 cell counts were determined with standard flow cytometry techniques, and plasma HIV-1 RNA levels by assay (Chiron branched DNA [bDNA] QUANTIPLEX HIV RNA assay, version 2.0 [Chiron Behring]; lower limit of detection, 500 copies/mL).
Genotype sequencing was performed prior to the addition of famciclovir to the treatment regimen and for all patients with detectable viremia during therapy with lamivudine-famciclovir at 12 months. For HBV PCR and sequencing, HBV DNA was extracted from serum by addition of 10 µL of NaOH (0.2-mL concentration) to 10 µL of serum, incubating at 40̊C for 1 h, and neutralizing with 30 µL of Tris-HCL (0.2M; pH, 7.5). The HBV polymerase gene was amplified between nucleotides 466 and 843. Population sequencing was carried out with use of an ABI 377 sequencer with Big Dye terminators (ABI) [19]. The nomenclature for amino acid changes in HBV polymerase is as described elsewhere [20].
All patients were male, white, and had sex between men as their risk factor for HIV transmission. A total of 3 of the 8 patients had a prior AIDS-defining illness. One patient undergoing HAART had an HIV virus load of 6968 copies/mL (patient 6), and one (patient 7) was receiving no antiretrovirals and had an HIV virus load of 135,617 copies/mL. This patient's HIV virus load subsequently became undetectable on resumption of HAART together with famciclovir treatment. The remaining patients were all undergoing HAART and had undetectable HIV virus loads (<500 copies/mL) for at least 6 months, suggesting good HIV control. The median CD4 cell count was 342 cells/mm3 (range, 47–450 cells/mm3).
The median ALT value when HAART including lamivudine alone was originally initiated was 104 IU/mL (interquartile range, 70–157 IU/mL) (figure 1). This had fallen to 63 IU/mL (interquartile range, 47–103 IU/mL) before the addition of famciclovir. Four of the eight patients exhibited a transient rise in ALT at 3 months, but overall there were no significant changes in transaminase values between initiation of therapy and 12 months later in this small group of patients.
Median baseline HBV virus load prior to the commencement of treatment with any nucleosides with anti-HBV activity was 9.6 log10 (equivalent to ç4 × 109 copies/mL; figure 2). During lamivudine therapy this fell significantly to a median of 4.75 log10 (P = .0079), and 1 patient's HBV virus load became undetectable (<400 copies/mL; patient 3; Wilcoxon 2-sample test). Addition of famciclovir at time 0 resulted in a further yet small (nonsignificant) reduction in HBV virus load of 0.33 log10 at 3 months (to 4.4 log10) and of 0.2 log10 at 6 months (to 4.55 log10). After 12 months' therapy with combination lamivudine-famciclovir, the median HBV virus load had increased with respect to baseline (i.e., time of famciclovir addition), to 7.8 log10 from 4.75 log10 (P = .084; Wilcoxon 2-sample test [significance limited by small patient group]).
Durable viral suppression with an undetectable HBV virus load at 12 months was achieved only in the patient whose anti-HBV therapy commenced with administration of lamivudine and famciclovir together. This patient's ALT normalized and seroconversion to HBeAg antibodies occurred after 13 months of therapy, although he remains HBsAg-positive.
Patient 3, in whom HBV suppression was achieved with lamivudine alone, subsequently experienced marked viral rebound after 12 months of dual therapy, with a rise in HBV to 8.2 log10 (1.4 × 108 copies/mL). This was associated with a flare in ALT to 433 IU/mL but no signs of clinical hepatic decompensation. Therapy was continued and transaminase values subsequently fell, although they remain elevated at approximately twice the upper limit. (They were previously normal.)
All patients tolerated both medications without serious side effects, and no new AIDS-defining illness or serious HIV-related problems occurred throughout therapy. No patient developed decompensated liver disease.
Sequencing of the HBV polymerase gene was performed for all patients at baseline (i.e., during administration of lamivudine but prior to the addition of famciclovir) and for all patients with detectable HBV viremia at 12 months of dual therapy to identify and characterize resistance mutations. The results are shown in table 2. The most common mutation sequence found at 12 months was the combination of L526M with M550V, observed in 4 patients (numbers 1, 2, 4, and 7), although this had already emerged with lamivudine alone in 1 individual (patient 4). Patient 5 had an L526M mutation with M550I, and patient 3 had no mutation at the 526 codon and an M550I substitution. Patient 6 rebounded with wild-type virus.
Resistance mutations occurring during nucleoside analogue therapy for chronic Hepatitis B are well recognized and limit their potential for use as long-term single agents. Trials of lamivudine therapy have shown resistance occurring in immunocompetent patients [8, 21] and those who are immunosuppressed following orthotopic liver transplantation [13, 22]. The majority of these mutations involve substitution of methionine by valine or isoleucine at position 550 in the highly conserved YMDD locus in domain C of the polymerase gene (analogous to the M184V mutation in lamivudine-resistant HIV) [14]. Although virus with this mutation typically emerges more slowly than in HIV, it decreases lamivudine susceptibility by at least 20–100-fold and is associated with a rapid rise of HBV DNA in serum. The M550V mutation, in particular, has been linked with the substitution of leucine for methionine at position 526 in the B domain [21], whereas the M550I mutation generally has not [14].
These patterns of mutations have been termed “group 1” and “group 2,” respectively, although more recent studies suggest that more complex patterns of resistance may occur [19, 23]. Two studies have shown the development of lamivudine-resistant HBV in HIV-infected patients despite the use of higher doses (300 mg q.d. as compared with 100 mg q.d.) [15, 24]. In both studies this occurred after 12–17 months and was always associated with mutations in the YMDD region, usually accompanied by an L526M change in the B domain.
Famciclovir resistance is less well-defined [25]. Some patients exhibit primary nonresponse to famciclovir with no documented mutations in domain B or C. Others develop mutations in domains B and C associated with viral breakthrough; this has been noted in both immunocompetent [26] and immunodeficient patients [27, 28]. These mutations may involve the same L526M substitution seen in lamivudine resistance [28] and suggest possible cross-resistance between nucleoside analogues. Several reports have described patients with lamivudine resistance failing to respond to famciclovir therapy [29, 30], although, conversely, patients with resistance mutations whose famciclovir therapy fails may remain susceptible to lamivudine [26, 31]. Resistance may be accelerated if the L526 mutation is present [28].
Lamivudine is a commonly used and potent agent for the treatment of HIV disease, and many coinfected patients, including the majority of those in our study, will have experienced lamivudine as unintentional monotherapy for HBV infection. Although a small proportion of these individuals may seroconvert to anti-HBe [11], the majority will not. All but 1 of our patients had detectable HBV in serum during therapy with lamivudine alone, and 6 of 7 had ALT values outside the normal range. However, only 1 patient (number 4) had lamivudine-associated resistance mutations (M550V, L526M) prior to the addition of famciclovir.
Similar to findings in HIV-negative populations [29, 30], we found that the addition of famciclovir to these patients' regimens failed to consistently suppress HBV production, with a median fall in HBV DNA of only 0.33 log10 at 3 months and an overall median gain of 3.0 log10 at 12 months. All patients with prior exposure to lamivudine demonstrated viral rebound within 12 months of the addition of famciclovir. In 4 of the 7 patients, this was associated with the emergence of amino acid changes at both L526M and M550V/I—substitutions recognized in both lamivudine and famciclovir resistance. Patient 3 had a transient flare associated with the appearance of an M550I substitution only, whereas patient 6 still has wild-type virus. This patient also had a detectable HIV virus load during combination therapy and may well have been only intermittently adherent to his medication. The only patient with an undetectable HBV load at 12 months—and who later seroconverted for HBeAg antibodies—received both lamivudine and famciclovir as initial therapy, a circumstance supporting the theory that lamivudine and famciclovir may be synergistic in vivo.
In conclusion, our study demonstrates the ineffectiveness of intensification of therapy with famciclovir for HIV-positive patients receiving lamivudine monotherapy for HBV infection. Resistance mutations are seen to occur readily with dual therapy, despite the absence of resistance prior to the addition of famciclovir. This would suggest that famciclovir fails to suppress the development of new mutations despite providing additional anti-HBV activity. Whether famciclovir may actually hasten the emergence of resistance in this setting, as has been suggested, remains unclear.
At present there are limited nucleoside analogues available with anti-HBV activity. Trials of adefovir have been halted for HIV-infected patients because of unacceptably high levels of renal toxicity, and entecavir (a deoxyguanosine analogue that inhibits HBV polymerase) is still in phase II trials. Our findings demonstrate that the sequential use of currently available anti-HBV nucleoside analogues is unlikely to be effective as a treatment strategy, a lesson learned previously with regard to HIV infection. Since lamivudine monotherapy is unlikely to be successful for the majority of patients, the use of dual agents such as lamivudine-famciclovir at initiation of therapy, with a rationale similar to that for combination therapy for HIV infection, may be the way forward. Although this study involved only small numbers of patients, it is worth noting that the only patient in whom durable HBV suppression and anti-HBeAg seroconversion were achieved had received both drugs together at the initiation of therapy.
Thus, when prescribing antiretrovirals for the treatment of HIV infection, the potential effects of nucleoside analogues on the course of HBV infection should also be considered, and the benefits of HBV suppression should be balanced against the risk of emergence of HBV drug resistance. As new and more potent nucleoside agents become available (including those with both anti-HIV and anti-HBV activity), the need for carefully controlled clinical trials of combination therapy increases.
IDSA Members: For your free access to this journal, log in via the IDSA members area.
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy