|Home | About | Journals | Submit | Contact Us | Français|
Following treatment of hepatitis B virus (HBV) monoinfection, HBV-specific T-cell responses increase significantly; however, little is known about the recovery of HBV-specific T-cell responses following HBV-active highly active antiretroviral therapy (HAART) in HIV-HBV coinfected patients. HIV-HBV coinfected patients who were treatment naïve and initiating HBV-active HAART were recruited as part of a prospective cohort study in Thailand and followed for 48 weeks (n = 24). Production of gamma interferon (IFN-γ) and tumor necrosis factor α (TNF-α) in both HBV- and HIV-specific CD8+ T cells was quantified using intracellular cytokine staining on whole blood. Following HBV-active HAART, the median (interquartile range) log decline from week 0 to week 48 for HBV DNA was 5.8 log (range, 3.4 to 6.7) IU/ml, and for HIV RNA it was 3.1 (range, 2.9 to 3.5) log copies/ml (P < 0.001 for both). The frequency of HIV Gag-specific CD8+ T-cell responses significantly decreased (IFN-γ, P < 0.001; TNF-α, P = 0.05). In contrast, there was no significant change in the frequency (IFN-γ, P = 0.21; TNF-α, P = 0.61; and IFN-γ and TNF-α, P = 0.11) or magnitude (IFN-γ, P = 0.13; TNF-α, P = 0.13; and IFN-γ and TNF-α, P = 0.13) of HBV-specific CD8+ T-cell responses over 48 weeks of HBV-active HAART. Of the 14 individuals who were HBV e antigen (HBeAg) positive, 5/14 (36%) lost HBeAg during the 48 weeks of follow-up. HBV-specific CD8+ T cells were detected in 4/5 (80%) of patients prior to HBeAg loss. Results from this study show no sustained change in the HBV-specific CD8+ T-cell response following HBV-active HAART. These findings may have implications for the duration of treatment of HBV in HIV-HBV coinfected patients, particularly in HBeAg-positive disease.
Individuals infected with human immunodeficiency virus (HIV) and hepatitis B virus (HBV) are at increased risk of liver disease progression and liver-related mortality (35). Despite the introduction of effective highly active antiretroviral therapy (HAART), liver disease remains a major cause of non-AIDS-related deaths in HIV-1-infected patients (31). Current guidelines recommend the early consideration of HBV-active HAART in the majority of coinfected individuals (28), and treatment of both HBV and HIV is generally lifelong. This is in contrast to HBV-monoinfected patients, where HBV treatment ceases following production of antibody to HBV e antigen (HBeAg) or HBV surface antigen (HBsAg) (23). HBeAg and HBsAg seroconversions are considered important endpoints of treatment as they are associated with HBV DNA clearance, normalization of alanine aminotransferase (ALT), and a reduction in the risk of liver disease (12).
Little is known about the immune events precipitating HBeAg or HBsAg seroconversion. However, a reduction in antigen burden following anti-HBV treatment may reduce T-cell tolerance and exhaustion, allowing for a more efficient HBV-specific T-cell and B-cell immune response against either HBeAg and/or HBsAg (11, 13, 21). Circulating HBV-specific CD4+ and CD8+ T cells are rarely detected in untreated chronic HBV infection (5, 24). Following treatment of HBV monoinfection with nucleos(t)ide analogues such as lamivudine (LMV), there is an increase in functional HBV-specific CD4+ and CD8+ T cells both in the peripheral blood (5, 18) and within the liver (32). However, recovery of HBV-specific T cells appears to be transient and has been shown to decline following long-term therapy (5, 14, 20).
We have previously shown that the HBV-specific T-cell response is impaired in HIV-HBV coinfection (7, 9). In one small observational study (n = 5), HBV-active HAART was associated with the recovery of CD8+ HBV-specific T cells (19); however, in this study, two patients had received prior HAART, and the HBV-specific T-cell responses were examined only during the first 24 weeks of treatment (19). In addition, HBeAg status was not defined, and HBV-specific T-cell responses were measured only by IFN-γ production following stimulation with HLA-A2-restricted epitopes (19).
In the present study, we used an overlapping peptide library covering the complete HBV genome to assess change in HBV-specific CD8+ T cells following the introduction of HBV-active HAART in treatment-naïve HIV-HBV-coinfected patients in Thailand. Overall, we show that there was no sustained change in the magnitude, frequency, or quality of HBV-specific T-cell responses following initiation of effective HBV-active HAART.
Patients with HIV-1 and chronic HBV were recruited from King Chulalongkorn Memorial Hospital and HIV-NAT, the Thai Red Cross AIDS Research Centre, Thailand (n = 32), as a substudy of two prospective randomized clinical trials for initiation of HBV-active HAART (Tenofovir in HIV/HBV Coinfection [TICO] study) and HIV-NAT 023, both funded by Gilead Sciences, San Francisco, CA. Participation was with the approval of the hospital ethics committee, and signed consent was obtained. Inclusion criteria were HIV-1 infection documented by enzyme-linked immunosorbent assay (ELISA), age ≥18 years, HBV DNA of >2 × 103 IU/ml, HBsAg positive for >6 months, and hepatitis C virus (HCV) antibody negative. All patients were naïve to treatment for either HIV or HBV. HBeAg serology was performed using a standard commercial assay (Abbott HBe EIA; Abbott Laboratories, Abbott Park, IL). Clinical details on the patients included in the TICO study have been published previously (25). To be included in this immunology substudy, an evaluable whole-blood specimen was required prior to the initiation of HAART (n = 24). This was defined as a specimen that would provide at least 8,000 CD8+ T cells by flow cytometry and had a background response to dimethyl sulfoxide (DMSO) of <0.4% cytokine-positive T cells.
All patients were treated with an HBV-active efavirenz-based HAART regimen defined as including LMV/emtricitabine (FTC) and/or tenofovir (TDF).
Intracellular cytokine staining (ICS) was performed as previously described (7). Briefly, 200 μl of fresh whole blood was incubated for 6 h at 37°C with an overlapping HBV peptide library that covered the entire HBV genome and was designed for the genotype A consensus sequence and included additional peptides for genotypes B, C, and D in areas of high variability (8, 9) or with HIV-1 Gag peptides (obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health) and either DMSO as a negative control or pokeweed mitogen (PWM; Sigma) and staphylococcal enterotoxin B (SEB; St. Louis, MO) as positive controls (Fig. (Fig.1).1). Surface staining was performed for 30 min at room temperature (RT) with anti-CD8-phycoerythrin (PE) (BD Biosciences, San Jose, CA), and intracellular staining was performed, following erythrocyte lysis and permeabilization, for 1 h (at RT) with anti-gamma interferon-fluorescein isothiocyanate (IFN-γ-FITC) and anti-tumor necrosis factor alpha-allophycocyanin (TNF-α-APC; BD Biosciences). Cells were acquired on a FACSCalibur instrument (BD Biosciences) and analyzed using WEASEL, version 2 (Walter and Eliza Hall Institute, Parkville, Australia).
If a patient did not respond to mitogen, did not have at least 8,000 CD8+ cells collected, or had a DMSO (background) response of >0.4% cytokine-positive cells, the HBV-specific response was not considered to be evaluable and was not included in the analysis. Cytokine responses that were less than 2-fold and less than 0.05% above background were considered undetectable and included as zero (Fig. (Fig.1).1). The magnitude of patient responses was determined as the sum of the percentage of CD8+ cytokine-positive cells for each peptide pool. The frequency of response was determined as the number of patients who had a response above zero to at least one peptide pool, after subtraction of background. The number of patients with a detectable response was expressed as a percentage of the number of patients who had evaluable specimens at each time point. To determine the specificity of the response, we calculated the mean response to each gene product as a proportion of the total proteome response for all patients. If a patient did not have a detectable response to a particular gene product, then the response was included as zero.
HBV DNA testing was performed as previously described (7) at the Victorian Infectious Diseases Laboratory (VIDRL), Melbourne. Briefly, HBV DNA was measured using the Versant HBV DNA 3.0 bDNA assay (lower limit of detection, 3.6 × 102 IU/ml; Bayer HealthCare, Tarrytown, NY) or, for samples below the lower limit of detection, the COBAS TaqMan HBV Test (lower limit of detection, 3 × 101 IU/ml; Roche Diagnostics, Branchberg, NJ). HIV RNA was quantified using the Bayer Versant HIV bDNA assay as per the manufacturer's instructions (lower limit of detection, 50 copies/ml; Bayer HealthCare-Diagnostics).
All liver biopsy specimens were examined by the same two pathologists (St. Vincent's Hospital, Melbourne, Australia), and the Metavir histological score for activity and fibrosis was used (3). Liver biopsy was optional.
The median decline in HBV DNA and HIV RNA viral loads was assessed and compared between HBeAg-positive and HBeAg-negative groups with a Wilcoxon rank sign test. Changes in CD4 counts over time were assessed by Kruskal-Wallis, with Dunn's posttest. Differences between patient groups were assessed by a Mann-Whitney U test. To compare the number of HBeAg-positive and HBeAg-negative patients who had undetectable levels of HBV DNA after 48 weeks, we used a test of proportions (chi-square). Changes in magnitude and frequency of T-cell responses over time were assessed by a generalized estimating equation (GEE) and chi-square test for trend in proportions, respectively. The specificity of the HBV peptide pool responses was analyzed by two-way analysis of variance (ANOVA) and a Bonferroni posttest. Data were analyzed using STATA (StataCorp, College Station, TX) and GraphPad Prism, version 5 (GraphPad Software, Inc., San Diego, CA). Statistical significance was set at a P value of ≤ 0.05.
Patient demographics prior to initiation of HBV-active HAART (n = 24) and following 48 weeks of treatment (n = 22; two patients, one HBeAg positive and one HBeAg negative, did not complete 48 weeks of follow-up) are summarized in Table Table1.1. The patients had advanced HIV infection with a median (interquartile range [IQR]) CD4 count of 60 (range, 29 to 200) cells/μl and median HIV RNA of 4.8 (range, 4.6 to 5.2) log copies/ml. Baseline HBV DNA was 8.5 log (range, 7.6 to 8.8 log) IU/ml baseline, and ALT was 39 (range, 34 to 78) U/liter. Over half of the cohort was HBeAg positive (14/24, or 58%). Patients were infected only with genotype C (8/23, or 78%; one patient did not have HBV genotype data available at baseline) and genotype B (5/23, or 22%). Patients had minimal liver disease identified on liver biopsy specimens (n = 19; median A and F scores were both 1).
As expected, following initiation of HAART, there was a significant decline in HIV RNA (median [IQR] log decline from week 0 to week 48 was 3.1 [range, 2.9 to 3.5] log copies/ml) with no significant difference in HIV RNA declines between HBeAg-positive and HBeAg-negative patients (P = 0.32, Wilcoxon rank sign test). All patients had an undetectable HIV RNA (<50 copies/ml) at week 48. CD4+ T-cell counts increased significantly in all patients, with a median increase of 166 (range, 97 to 237) cells/μl (P = 0.002, Kruskal-Wallis test) (Table (Table11).
The median (IQR) decline in HBV DNA from week 0 to week 48 was 5.8 log (range, 3.4 to 6.7) IU/ml. The median log decline in HBV DNA for HBeAg-positive and HBeAg-negative patients was 5.4 (range, 3.6 to 6.8) log IU/ml and 6.6 (range, 2.3 to 6.8) log IU/ml, respectively (P = 0.02, Wilcoxon rank sign test) (Table (Table1).1). Twelve of 22 (55%) patients had an undetectable HBV DNA level (<30 IU/ml) at week 48. At baseline, HBV DNA was lower in HBeAg-negative individuals than in HBeAg-positive individuals but the difference did not reach statistical significance (P = 0.07, Mann-Whitney U test). More HBeAg-negative patients than HBeAg-positive patients had undetectable HBV DNA at 48 weeks (7/9 and 5/13, respectively; P = 0.07, chi-squared test).
ALT levels did not significantly change over time (data not shown). Two patients had a hepatic flare, defined as an increase in ALT of >200 U/liter from baseline or more than five times the upper limit of normal of ALT (see Fig. Fig.3B).3B). One of these patients died at week 12 secondary to hepatic flare.
Production of IFN-γ and/or TNF-α in CD8+ T cells was measured by ICS pretreatment (week 0) and at weeks 4, 12, 24, and 48 following treatment initiation. Of the 32 patients initially enrolled in the study, 75% (n = 24) had measurable IFN-γ-positive (IFN-γ+) or TNF-α-positive (TNF-α+) (single) CD8+ T-cell responses and/or IFN-γ+ and TNF-α+ (dual) CD8+ T-cell responses at baseline and at least one additional ICS measurement over the study period. Patients excluded from this analysis (n = 8) did not have an evaluable sample at baseline because either there were too few CD8+ T cells available for analysis (n = 5) or no sample was available (n = 3). Prior to initiation of HBV-active HAART, 33% (8/24; IFN-γ+) and 25% (5/20; TNF-α+) had a detectable CD8+ T-cell response to at least one HBV peptide pool (Fig. (Fig.2A).2A). Responses were detected to all peptide pools without any clearly dominant pool. There was no significant difference between the numbers of individuals who had a response to each pool or the magnitude of the responses to each pool at each time point (two-way ANOVA and Bonferroni posttest) (Fig. (Fig.2C).2C). Following initiation of HBV-active HAART, the number of patients who had a detectable IFN-γ+ cytokine response to at least one peptide pool was 22%, 36%, 20%, and 16% at week 4, 12, 24, and 48, respectively (Fig. (Fig.2A).2A). The number of patients with a detectable HBV-specific TNF-α+ response was 27%, 14%, 28%, and 17% at week 4, 12, 24, and 48, respectively (Fig. (Fig.2A).2A). There were no significant changes in the frequency of HBV-specific CD8+ T-cell responses over time (IFN-γ+, P = 0.21; TNF-α+, P = 0.61; chi-square test for frequency of proportions). In addition, there was no significant change in the frequency of HBV-specific T cells when HBeAg-positive and HBeAg-negative patients were assessed separately (Fig. (Fig.2A).2A). The findings were similar when the magnitude of the HBV-specific response was evaluated (Fig. (Fig.2A2A).
CD8+ T-cell responses to HIV Gag peptides were included as a positive control in each assay. As expected, we observed a significant decline in the frequency of HIV-specific CD8+ T-cell responses over time (IFN-γ+, P < 0.001; TNF-α+, P = 0.05; chi-square test for frequency of proportions) (Fig. (Fig.2B)2B) and magnitude of IFN-γ+ CD8+ T-cell HIV Gag-specific responses over time but not TNF-α+ responses (IFN-γ+, P = 0.013; TNF-α+, P = 0.09; GEE) (Fig. (Fig.2B2B).
Given that functional antiviral T cells are associated with production of more than one antiviral cytokine (2, 4), we also assessed the capacity of HBV-specific CD8+ T cells to produce both cytokines (i.e., both IFN-γ+ and TNF-α+) following HBV-active HAART. Prior to initiation of HBV-active HAART, 6/24 (25%) had a detectable IFN-γ+ TNF-α+ CD8+ T-cell response to at least one HBV peptide pool (Fig. (Fig.2A).2A). There were no significant changes in the frequency of HBV-specific IFN-γ+ TNF-α+ CD8+ T-cell responses over time (P = 0.11, chi-square test for frequency of proportions) or magnitude (P = 0.13, GEE). However, as expected, we observed a significant decline in the frequency of HIV-specific IFN-γ+ TNF-α+ CD8+ T-cell responses over time (P = 0.001, chi-square test for frequency of proportions) and magnitude (P = 0.022, GEE) (Fig. (Fig.2B2B).
The immunological events that precede HBeAg seroconversion are not fully understood although it is likely that the efficient production of antibody to HBeAg (HBeAb) requires reconstitution of CD4+ and/or CD8+ HBV-specific T cells. We therefore examined the change in HBV-specific CD8+ T cells in individuals who lost HBeAg over 48 weeks (5/14, or 36%) (Fig. (Fig.3A).3A). Patients 1017, 1339, and 1022 were randomized to the LMV/TDF or FTC/TDF treatment group, and patients 1015 and 1025 were randomized to the FTC-only treatment group. In two patients (patients 1017 and 1025), there was a loss of HBeAg and appearance of HBeAb, but HBeAb was not maintained, and seroreversion to HBeAg occurred. Four of the five patients (80%) who lost HBeAg had clearly detectable, but only transient, HBV-specific T cells prior to loss of HBeAg and/or seroconversion. The highest detectable response was predominantly seen with TNF-α+-producing (4/5) rather than IFN-γ+-producing (1/5) CD8+ T cells, and responses were not associated with any particular HBV peptide pool. The specificity of the response in each patient was diverse (data not shown), with maximal response to peptides from the X (patient 1017 and 1339), precore (1015 and 1025), and surface (1022) proteins. None of the five patients who lost HBeAg had a detectable HBV-specific T-cell response prior to treatment. The maximum magnitude of the HBV-specific response at any time after treatment was higher in the HBeAg-positive patients who seroconverted (n = 5) than in the rest of the cohort (n = 19), but this difference did not reach statistical significance (P = 0.07) (Fig. (Fig.3C3C).
As hepatic flare is thought to be secondary to infiltration of both HBV-specific and non-HBV-specific T cells and/or the production of proinflammatory cytokines, we also examined the frequency of circulating HBV-specific T cells in the limited number of patients with hepatic flare (n = 2). One individual (1339) (Fig. (Fig.3B)3B) had a significant increase in ALT and, as described above, clear but transiently detected HBV-specific TNF-α+ CD8+ T cells in the weeks prior to HBeAg seroconversion. This patient had only mild liver disease (necroinflammatory score [A] of 1; fibrosis score [F] of 1) at baseline. This patient also had a detectable IFN-γ response at the time of ALT flare. The other patient (1336) with hepatic flare died from hepatic decompensation. There was a clearly detectable TNF-α response in the weeks prior to flare. This patient did not consent to a liver biopsy.
This is the first report of a prospective, comprehensive longitudinal study of HBV-specific T cells following initiation of HBV-active HAART in HIV-1-HBV-coinfected individuals. Overall there was no significant increase in HBV-specific CD8+ T-cell responses over time following initiation of HBV-active HAART. The incidence of HBeAg seroconversion was high (36%), and in nearly all cases, we could detect an HBV-specific T-cell response prior to seroconversion although this was only transient.
Recovery of HBV-specific T-cell responses has been previously described following treatment of HBV monoinfection with LMV and adefovir. In these studies, HBV-specific T-cell responses increased initially and then returned to pretreatment levels within the first 24 weeks of treatment (5, 14, 20). Other studies have shown that HBV-specific T-cell responses increase again soon after HBV DNA rebound (14, 29), suggesting that there is a role for sustained antigenic stimulation in detection of HBV-specific responses. We found that there was no change in the frequency or magnitude of HBV-specific CD8+ T cells following HBV-active HAART. One explanation may have been that we missed any transient recovery; however, we evaluated HBV-specific responses at similar frequencies to previous studies of HBV monoinfection (5, 14, 20). It is possible that reconstitution of HBV-specific T-cell responses may require a longer time in the setting of HIV-HBV coinfection than in HBV monoinfection, and therefore more prolonged follow-up of this patient cohort may be of interest. An alternative explanation is that patients in our cohort were unable to generate an efficient antigen response secondary to advanced HIV infection (median CD4 count, 60 cells/μl). However, these patients could clearly generate a response to HIV Gag peptides as well as a mitogen (data not shown). Finally, another explanation may be that HBV-specific T cells may traffic to the liver following initiation of HAART and immune reconstitution and are therefore not detected in blood. We have recently shown that untreated HIV-HBV-coinfected patients express high levels of CXCL-10 (10), the principal chemokine involved in trafficking of CXCR3+ T cells to the liver in other infections such as HCV (16).
We observed a significant decline in HIV-specific responses on HAART, as previously reported. This is generally thought to occur secondary to a decline in antigenic stimulus (1, 17). Therefore, it is possible that one might expect HBV-specific T-cell responses also to decline following effective treatment; however, we found no change in the HBV-specific T-cell response despite significant declines in HBV DNA levels. This may have several explanations. First, although the median decline in HBV DNA from week 0 to week 48 was 5.8 logs, only 55% of patients had an HBV DNA of <30 IU/ml at week 48, in contrast to 100% of patients having an HIV RNA of <50 copies/ml. However, even when we compared patients who had detectable HBV DNA to those who cleared HBV DNA by 48 weeks, we found no differences in HBV-specific T-cell responses (data not shown). Second, in contrast to HIV, following treatment of HBV infection and despite significant reductions in HBV DNA levels, there was little change in levels of circulating HBV proteins, namely, HBsAg (6, 26). We performed quantitative HBsAg (qHBsAg) analysis on a subset of patients (n = 12), and while there was a decrease in qHBsAg levels following treatment, the levels overall remained very high even at 48 weeks (data not shown). This is in agreement with previous reports of a minimal decline in HBsAg following nucleoside analogue treatment in both HBV monoinfection and HIV-HBV coinfection (6, 26, 34). The persistence of HBsAg, despite treatment, may be a consequence of limited recovery of HBV-specific T cells or, alternatively, may lead to ongoing immune exhaustion and therefore suppress an effective immune response (22, 30).
The immune events precipitating HBeAg seroconversion are still poorly understood even though seroconversion is an important endpoint of treatment. Spontaneous HBeAg seroconversion occurs at greater frequency in certain HBV genotypes, namely, B and C (27), and is more common in individuals with ALT levels more than two to five times the upper limit of normal (ULN) prior to treatment (11, 21, 37). In addition, HBeAg seroconversion has been associated with a lower HBV DNA level (21) and an increase in T-cell proliferation in vitro in response to core protein and HBeAg (36) at the time of ALT flare, as well as a fall in quantitative HBeAg, at least during pegylated interferon therapy (13). Recently, it has been demonstrated that the expression of the T-cell exhaustion marker PD-1 on total CD8+ T cells significantly declined in patients who had treatment-induced HBeAg seroconversion (11). Therefore, it is feasible that a reduction in HBV-related antigenic stimuli may reduce T-cell tolerance and exhaustion and lead to a more efficient HBV-specific immune response. Together with HAART-induced recovery of CD4+ T cells, this may potentially lead to higher rates of HBeAg and/or HBsAg seroconversion. We detected HBV-specific CD8+ T cells in the majority of individuals prior to HBeAg seroconversion although detection of these cells was only transient. This study was not adequately powered to determine if the magnitude of HBV-specific T cells differed in HBeAg-positive patients who seroconverted from levels in those that did not seroconvert. Therefore, this study needs to be repeated in a larger cohort to determine if the detection of an increase in HBV-specific T cells is significantly associated with, or predictive of, durable HBeAg seroconversion. In addition, it would be important to also assess the HBV-specific CD4+ T-cell response in this setting.
This study had several limitations. First, we were unfortunately unable to measure HBV-specific CD4+ T-cell responses prior to or following treatment, given the low median CD4+ T-cell count at the time of treatment initiation in this cohort. Understanding the changes in HBV-specific CD4+ T cells may be critical to our interpretation of HBV-specific CD8+ T-cell kinetics and development of antibodies to either HBeAg or HBsAg. Second, we were able to measure the production of IFN-γ and TNF-α only from HBV-specific T cells because we had access only to a four-color flow cytometer in Thailand. We selected these two cytokines, in preference to others such as interleukin-2 (IL-2) and IL-10, because our previous study of HBV monoinfected patients identified IFN-γ and TNF-α as being produced in greater magnitude and frequency than either IL-2 or IL-10 in HBV-specific CD8+ T cells in the blood and liver (8). In addition, both IFN-γ and TNF-α have been demonstrated to be important in the cytolytic and noncytolytic clearance of HBV (15). It is possible that recovery of HBV-specific cytokines other than IFN-γ and TNF-α may occur following HBV-active HAART. For example, the production of other cytokines such as IL-17, IL-2, RANTES, and CD107a may be important and should be examined in future studies. Finally, it would also be important to analyze changes in the intrahepatic compartment following initiation of HBV-active HAART. Newer techniques such as fine-needle aspiration of the liver may allow such studies in the future (33).
In this prospective clinical trial of initiation of HBV-active HAART in HIV-HBV coinfection in an Asian population, there was little change in the frequency or magnitude of HBV-specific T cells, despite good control of HBV DNA replication and significant recovery of total CD4+ T cells after 48 weeks. These findings may have implications for the duration of treatment of HBV in HIV-HBV-coinfected individuals, given that HBeAg seroconversion is an important endpoint of treatment. Longer follow-up is required to determine if the absence of recovery of HBV-specific T cells following initiation of HBV-active HAART is associated with clinical outcomes such as liver disease progression and durability of HBeAg and HBsAg seroconversion.
This study was funded by the National Institutes of Health (NIH R21 AI055379-01) and Gilead Sciences. S.R.L. is supported by the Alfred Foundation and is a National Health and Medicine Research Council Practitioner Fellow (grant 251651).
K.R. has recently received research grants/funding, honoraria, or has been a consultant or advisor to, or received lecture sponsorships from Abbott, Boehringer-Ingelheim, Bristol-Myers-Squibb, Gilead, GlaxoSmithKline, Hoffmann-LaRoche, Janssen-Cilag, Merck Sharpe & Dolme, Tibotec and Virco. S.R.L. has received research grants/funding from Roche, Bristol-Myers-Squibb, Gilead, and Pfizer.
We thank Tim Spelman for his expert statistical advice, Pip Marks for study coordination, and the patients for their participation in this study.
Published ahead of print on 6 January 2010.