Potential conflict of interest: Nothing to report.
Four hundred and eighty Chinese subjects with hepatitis B e antigen (HBeAg)–positive chronic hepatitis B (CHB) were enrolled in a multicenter, double-blind, randomized, placebo-controlled study of adefovir dipivoxil (ADV) 10 mg once daily. There was a significant difference in reduction of serum hepatitis B virus (HBV) DNA after 12 weeks between subjects who received ADV and those who received the placebo (3.4 and 0.1 log10 copies/mL, respectively, P < .001). Further reductions in serum HBV DNA and increases in the proportion of subjects with an HBV DNA level of at most 105 copies/mL, with HBV DNA undetectable, and with ALT normalization were observed in ADV-treated subjects at week 52 (median HBV DNA reduction of 4.5 log10 copies/mL, 67% with HBV DNA ≤ 105 copies/mL, 28% with HBV DNA undetectable, and 79% with ALT normalization). Subjects who initially received ADV lost some treatment benefit after being rerandomized to the placebo in week 40. Subjects with YMDD mutant HBV at baseline had virological, biochemical, and serological responses to treatment that were similar to those of subjects with wild-type HBV. The incidence of clinically adverse events was similar in nature and severity between the treatment groups, and there was no evidence of renal toxicity. No adefovir-related HBV mutations were identified. In conclusion, treatment with ADV 10 mg daily over 52 weeks was safe and effective in Chinese subjects with HBeAg-positive CHB and did not lead to the emergence of drug resistance. The study is continuing for an additional 4 years with all subjects on open-label ADV 10 mg daily. (HEPATOLOGY 2006;44:108–116.)
Despite the availability of effective vaccines for hepatitis B virus (HBV) since the 1980s, it is estimated that at least 350 million people worldwide remain chronically infected with this virus.1 The incidence of chronic infection is higher in Asia, where more than 8% of the population of China, or more than 100 million people, is chronically infected.1 Although subjects may remain relatively healthy for years, chronic hepatitis B (CHB) is a progressive disease that can result in cirrhosis or hepatocellular carcinoma.
The aim of treatment is to achieve sustained suppression of HBV replication in order to arrest or reverse the progression of hepatic injury.2, 3 Lamivudine and interferon alpha are both licensed for the treatment of CHB in China. Lamivudine is a well-tolerated oral agent that suppresses viral replication, enhances hepatitis B e antigen (HBeAg) seroconversion, improves liver histology, and normalizes serum alanine aminotransferase (ALT).4, 5 Long-term treatment delays clinical progression to hepatic decompensation or occurrence of hepatocellular carcinoma in subjects with advanced liver disease.6 However, the long-term benefits are limited by the emergence of drug resistance in up to 69% of subjects after 5 years of treatment.7–9 Treatment with interferon-α requires parenteral administration and is associated with significant treatment-limiting side effects.10 Therefore, there is a continued need for well-tolerated efficacious agents that have a high barrier to the development of resistance.
Adefovir dipivoxil (ADV) is a synthetic nucleotide analogue of adenosine monophosphate. Multinational placebo-controlled clinical trials in HBeAg-positive and -negative subjects showed biochemical, virological, and histological improvement and a safety profile at the 10-mg dose similar to those of the placebo.11, 12 Long-term treatment with ADV of those with HBeAg-negative CHB for up to 4 years resulted in improved responses, was well tolerated, and the emergence of resistance was delayed and infrequent.13–15 ADV is effective against both wild-type and lamivudine-resistant forms of HBV both in vitro and in vivo.16–18 Although no adefovir-associated resistance mutations were identified during treatment for up to 48 weeks, adefovir-resistant strains, which are susceptible to lamivudine therapy, have been identified in less than 2% (4/238) of subjects treated up to 96 weeks and in 18% of a cohort of HBeAg-negative subjects treated for up to 4 years.19–23
This placebo-controlled study tested the antiviral efficacy and safety of ADV 10 mg once daily in Chinese subjects with HBeAg-positive CHB and also investigated the impact of cessation of therapy under double-blind conditions. The 52-week results are reported here. The study is ongoing and is designed to evaluate treatment for up to 5 years.
HBeAg, hepatitis B e antigen; CHB, chronic hepatitis B; ADV, adefovir dipivoxil; HBV, hepatitis B virus; DNA, deoxyribonucleic acid; ALT, alanine aminotransferase; ULN, upper limit of normal; YMDD, tyrosine- methionine-aspartate-aspartate mutation.
Patients and Methods
Subjects were identified by the 12 investigators across seven cities in China. Eligible subjects included men and women 18 years and older who, at screening (up to 28 days prior to randomization and, for HBsAg, ≥6 months prior to study enrollment), had detectable hepatitis B surface antigen (HBsAg); detectable hepatitis B e antigen (HBeAg); serum HBV DNA ≥ 106 copies/mL (Roche COBAS Amplicor Monitor PCR assay, lower limit of detection 200 copies/mL, revised to 300 copies/mL during the course of the study); serum ALT level more than 1 times the upper limit of normal, ULN (and >2× ULN sometime within the previous 6 months).
Criteria for exclusion included evidence of hepatocellular carcinoma; clinical signs of liver decompensation; serum creatinine greater than 1.5 mg/dL (≥130 μmol/L); ALT more than 10 times the ULN; seropositivity for hepatitis C or D virus or HIV; lamivudine therapy within 3 months prior to screening; and ADV therapy or any other anti-HBV therapy within the previous 6 months. Subjects were not permitted to receive systemic antiviral agents, immunomodulators, immunosuppressive therapy, Chinese traditional medicines, or agents known to lower ALT levels during the study.
The study was approved by the local ethics committees and conducted in accordance with the guidelines of the Declaration of Helsinki (1996).
The primary study objective was to assess efficacy and safety after 52 weeks of treatment and was designed as a registration study in close consultation with the Chinese regulatory agency.
Subjects returned to the clinic within 4 weeks of screening for a baseline assessment. They were then randomly assigned to receive either ADV 10 mg once daily (QD) or matching placebo tablets (both manufactured by Gilead Science Inc, CA) in a 3:1 ratio for 12 weeks (Fig. 1) to evaluate the initial antiviral response. A central telephone randomization system, RAMOS (Randomisation and Medication Ordering Systems), was used for clinical supplies management and treatment assignment via random-number generation in blocks of eight. Double-blind study drug and open-label supplies were packaged, labeled, and numerically coded for each treatment according to the computer-generated random code.
At week 12, all subjects started open-label ADV 10 mg QD for 28 weeks. After this, subjects who received ADV in the first 12 weeks were rerandomized to receive either ADV or placebo in a 2:1 ratio for 12 weeks to evaluate the effect of discontinuing therapy in a subgroup of subjects. During this period, placebo recipients who experienced progression of liver disease were eligible to receive open-label ADV. Subjects who initially received placebo continued to receive open-label ADV.
Enrolled subjects returned for study visits in weeks 4, 8, 12, 16, 28, 40, 44, 48, and 52 for specified assessments. Subjects were evaluated for the presence of YMDD mutations at baseline and for ADV mutations at week 52.
Clinical data were collected, monitored, and entered into a database by GlaxoSmithKline. Hematology and biochemical analyses were conducted at each site. Serology testing was conducted by MDS Pharma Services (Beijing, China). Genotypic resistance analyses were performed by QUEST Diagnostics Ltd. (Van Nuys, CA). The sponsor conducted the statistical analyses.
Efficacy End Points.
The primary and secondary efficacy end points in this study were the log10 reduction from baseline of serum HBV DNA (Roche COBAS Amplicor Monitor™ PCR assay) in week 12 and the proportion of subjects with ALT normalization in week 12, respectively. These early end points were measured at the request of the local regulatory agency. Other efficacy end points to week 52 included the log10 reduction in HBV DNA; the proportion of subjects with an HBV DNA concentration of less than 105 copies/mL, HBV DNA undetectability, ALT normalization, HBeAg loss and seroconversion, and time to disease progression for subjects who discontinued ADV after week 40. A severe flare was defined as a confirmed elevation of ALT to more than 5 times the ULN with a serum bilirubin more than twice the ULN. A preplanned analysis of efficacy measures was performed according to the presence or absence of YMDD mutation at baseline.
Testing for Genotypic Resistance to Adefovir and Lamivudine.
Baseline HBV isolates were analyzed for the presence of YMDD mutations in a blinded fashion. Only week 52 isolates from subjects who had breakthrough HBV DNA on ADV therapy at any time during the study were analyzed for the presence of adefovir-associated mutations. HBV DNA breakthrough was defined as an increase in the level of HBV DNA of at least 1 log10 copies/mL from the lowest point prior to week 40 while on treatment. A modified HBV genotype assay, which amplified a 866-nt region of the pol gene, was used to detect changes at five specific sites in the reverse transcriptase region of the HBV polymerase gene (rt180, rt181, rt207, rt204, rt236) known to be associated with resistance to lamivudine or ADV. In-house sequencing rather than a line-probe assay was employed.
The 36-item Short-Form Health Survey (SF-36) questionnaire assesses eight dimensions of health: physical function, social function, role limitations due to physical problems, mental health, role limitations due to emotional problems, vitality, general health perceptions, and bodily pain.24 Each dimension was assessed via a score on a scale ranging from 0 (worst) to 100 (best possible). An ANOVA model was used for the baseline visit, adjusting for sex, age, and investigator. Treatment comparisons were performed in weeks 12, 40, and 52 using an ANCOVA model in which treatment group and investigator were the main effects and baseline value of the corresponding scale, ALT (× ULN) at baseline, sex, and age were the covariates.
The study was designed to enroll 480 subjects, and expected to provide greater than 95% power to detect a mean difference in the change in HBV DNA from baseline to week 12 between ADV and placebo of 1.5 log10 copies/mL. The primary population for efficacy analyses was all randomized subjects with HBeAg-positive CHB regardless of whether the study drug was taken or if the subject completed the study. The population for the safety analyses was all subjects for whom no clear evidence of failure to take study medication was available. Results for the week 12 analysis are presented for the two treatment groups, the ADV (n = 360) and the placebo (n = 120) groups. The results of subsequent analyses are presented for three treatment groups: 1) Group PAA (n = 120)—subjects randomized to receive placebo for the first 12 weeks, open label ADV for the next 28 weeks, and ADV for the remaining 12 weeks; 2) Group AAA (n = 240)—subjects randomized to receive ADV for the first 12 weeks, open label ADV for the next 28 weeks, and ADV for the remaining 12 weeks; and 3) Group AAP (n = 120)—subjects randomized to receive ADV for the first 12 weeks, open label ADV for the next 28 weeks, and rerandomized to placebo for the remaining 12 weeks.
From December 2002 to February 2003, 480 subjects were randomized to the study. Baseline demographic and disease characteristics were similar across all treatment groups (Table 1). Subjects were all Chinese, predominantly male, with the median age of the groups ranging from 30 to 31 years. A total of 303 of the 480 (63%) subjects had previously taken medication for the treatment of CHB, including 154 (32%) who had received prior treatment with lamivudine and 176 (37%) who had received treatment with traditional Chinese medicines. The incidence of YMDD mutant HBV at baseline was similar between the treatment groups, ranging from 15% to 22% (Table 1). No subject with cirrhosis was enrolled.
Table 1. Summary of Baseline Demographic and HBV Disease Characteristics*
PAA Group (N = 120)
AAA Group (N = 240)
AAP Group (N = 120)
Screening for study eligibility occurred up to 28 days prior to randomization (baseline).
Mean ± SD
32 ± 10
31 ± 9
32 ± 10
HBV DNA (log10 copies/mL)
Mean ± SD
8.6 ± 1.0
8.6 ± 1.0
8.5 ± 1.0
ALT (× ULN)
Mean ± SD
3.8 ± 4.1
3.9 ± 3.8
3.3 ± 2.6
<1 × ULN
1-2 × ULN
2-5 × ULN
>5 × ULN
Prior lamivudine treatment
HBV DNA reduction (median change, range, 25% and 75% interquartile values), the proportion of subjects with an HBV DNA level of less than 105 copies/mL, and the proportion with HBV DNA undetectable (<300 copies/mL) in weeks 12, 40, and 52 are presented in Table 2 and Fig. 2.
At week 12, the median decrease in serum HBV DNA levels was 3.4 log10 copies/mL in the ADV group and 0.1 log10 copies/mL in the placebo group (P < .001). At week 40, the median reduction in serum HBV DNA was similar across all treatment groups. In subjects rerandomized to ADV (Groups AAA and PAA) at week 40, there were further reductions in serum HBV DNA from −4.2 to −4.5 and from −4.6 to −5.0 log10 copies/mL, respectively at week 52. In contrast, in subjects rerandomized to placebo during weeks 40-52 (Group AAP), the median reduction in serum HBV DNA decreased from −4.0 to −0.2 log10 copies/mL at week 52.
In week 12, serum ALT normalization was observed in 140 of the 330 subjects (42 %) in the ADV group and in 15 of the 108 subjects (14%) in the placebo group (P < .001).
In Group AAA the proportion of subjects with serum ALT normalization increased to 79% in week 52. As for virologic parameters in those subjects rerandomized to placebo for weeks 40-52 (Group AAP), biochemical benefit was rapidly lost (Table 2 and Fig. 3).
HBeAg and HBsAg Responses.
In week 52, HBeAg loss occurred in 30 of the 233 subjects (13%) who had received continuous ADV therapy (Table 2). This figure was higher in Group PAA at week 52, with HBeAg loss observed in 24 of 118 subjects (20%). Of note is that in 6 subjects in Group PAA, HBeAg loss was recorded in week 12 after only placebo therapy; HBeAg data between baseline and week 12 were not collected, so the precise timing of HBeAg loss is not known. Sixteen of the 114 subjects in Group AAP (14%) had achieved HBeAg loss after 40 weeks of ADV therapy. After week 40, when they were rerandomized to placebo, 9 of these 16 subjects regained HBeAg at some timepoint between week 40 and week 52 There was no clear correlation in these 16 AAP subjects between the timing of first-recorded HBeAg loss and either maintenance of HBeAg loss or HBeAg reversion between weeks 40 and 52 on placebo. However, the observed HBeAg reversion supports the concept that it is continuing rather than early termination of therapy after HBeAg loss is first reported.
In week 52 HBeAg seroconversion had occurred in 21 of the 118 subjects (18%) in Group PAA, a lower percentage than in the other two treatment groups (Group AAA, 19/233, 8%; Group AAP, 8/114, 7%). The higher rate of seroconversion in Group PAA subjects was influenced by six subjects showing HBeAg loss and seroconversion in the first 12 weeks while on placebo and may represent spontaneous cases of HBeAg seroconversion occurring around study entry. In this small subset of subjects in Group PAA, baseline ALT values were higher than Group PAA (mean 7.1, median 5.3, range 1.5-17.2 × ULN and mean 3.8, median 2.5, range 0.5-28.9, respectively), which could indicate that spontaneous cases of HBeAg seroconversion did occur. This phenomenon must be considered when assessing the reported incidence of HBeAg loss or HBeAg seroconversion with any CHB treatment.
Baseline median HBV DNA and ALT characteristics in this study were similar to those of subjects in previously reported studies that demonstrated superior HBeAg loss (24%) and seroconversion (12%) after 48 weeks.11 Serological data following subsequent years of ADV therapy in this Chinese study population are awaited.
No subject experienced HBsAg loss or seroconversion up to 52 weeks.
Durability of HBeAg seroconversion was not investigated within the 52-week study period. However, in the extension study (open-label ADV therapy for all subjects up to 5 years), treatment will be stopped following confirmed HBeAg seroconversion, and durability will be assessed.
Two subjects (1.6%) from Group AAP met the criterion of a severe flare following rerandomization to placebo at week 40. Both experienced a marked increase in serum ALT levels associated with increased serum bilirubin levels. Both events resolved after the subjects commenced treatment with open-label ADV.
Testing for Genotypic Resistance to Adefovir.
A total of 45 subjects (28 from Group AAA, 11 from Group PAA, and 6 from Group AAP) had an increase in serum HBV DNA of at least 1 log10 copies/mL while on adefovir from their lowest point during treatment and therefore had isolates analyzed for the presence of ADV-associated mutations at week 52. In the majority of subjects (28/45), this increase in serum HBV DNA was less than 2 log10 copies/mL, with only two subjects recording increases of more than 4 log10 copies/mL from their lowest point during treatment. Of these 45 protocol-defined breakthroughs, only 2 were associated with an ALT flare, both in subjects in Group AAP after week 40 on rerandomization to placebo. All subjects were compliant with study medication. Testing for genotypic resistance showed that none of the subjects had the N236T or A181V mutations known to be associated with resistance to ADV.
Impact of YMDD Mutations at Baseline.
To analyze the impact of YMDD-mutant HBV at baseline on efficacy parameters, only the subjects who received uninterrupted ADV therapy (Group AAA) were investigated. In these subjects changes in median HBV DNA and ALT normalization rates up to week 52 were similar regardless of YMDD-mutant HBV status at baseline (Table 3).
Table 3. Efficacy to week 52 by YMDD-mutant HBV status at baseline (Group AAA)
Non-YMDD (N = 188)
YMDD-Mutant HBV (N = 52)
HBV DNA reduction (log10 copies/mL), median (range)
Only six subjects (1.2%) discontinued study drug prematurely during the study; for three (0.6%), this was because of an adverse event (two in Group AAA and one in Group AAP). One subject withdrew after a diagnosis of IgA nephritis on renal biopsy; this subject had proteinuria at baseline and normal renal function during treatment and follow-up. The other two adverse events leading to withdrawal were back pain and alopecia. The overall rate of clinically adverse events was similar in nature and severity in all three treatment groups at week 52. Adverse events rarely occurred at a frequency of 5% or greater in any treatment group (Table 4) Group AAP had a higher incidence of hepatitis B-related adverse events (11/120, 9%), compared with a frequency of less than 1% in the other two treatment groups. These events all occurred following the rerandomization of subjects to placebo at week 40.
Table 4. Adverse Events Reported by at Least 5% of Subjects
PAA Group (N = 120)
AAA Group (N = 240)
AAP Group (N = 120)
Reactivation of hepatitis B
Upper respiratory tract infection
One subject initially randomized to placebo (Group PAA) had a serious adverse event (nasopharyngeal cancer), as did four subjects in Group AAA (ALT elevation [2 cases], spontaneous abortion, and fracture), and eight subjects in Group AAP (exacerbation of hepatitis B [8 cases] and bronchial pneumonia). All eight cases of disease exacerbation occurred after subjects were rerandomized to placebo at week 40. No subject died during the study.
Serum ALT elevation was the most common laboratory finding of toxicity. The incidence of elevation of serum ALT to more than 5 times the ULN was 21% (51/240) in Group AAA, 55% (66/120) in Group AAP, and 38% (46/120) in Group PAA,. The 21% of subjects on continuous ADV therapy for 52 weeks (Group AAA) whose ALT increased to more than 5 times the ULN is a proportion very similar to that previously reported.25 In this same treatment group, 57 of the 61 (95%) instances of ALT increasing to more than 5 times the ULN occurred in the first 8 weeks of therapy; only two ALT elevations were recorded after 12 weeks, both 5-10 × the ULN, occurring in a single subject in weeks 48 and 52. In in Group AAP the incidence of ALT elevation during treatment also occurred predominantly in the early weeks of therapy. A total of 35 of the 43 (81%) instances of ALT elevation during treatment to more than 5 times the ULN occurred in the first 8 weeks of therapy; only 6 were recorded after 12 weeks, all 5-10 times the ULN, occurring in three subjects. In Group AAP ALT elevation while not being treated (i.e., on re-randomization to placebo at week 40) predominantly occurred in weeks 48 and 52; of 120 subjects in Group AAP, 29 (24%) and 28 (23%) recorded ALT elevations of 5-10 times the ULN and more than 10 times the ULN, respectively, similar to those previous reported after cessation of therapy.26, 27 In Group PAA, the majority of instances (60/76, 79%) of ALT elevation to more than 5 times the ULN occurred while off treatment in the first 12 weeks of the study; no elevation of ALT to more than 5 times the ULN was reported after 16 weeks on treatment. Across all treatment groups only 2 instances of ALT increasing to more than 5 times the ULN (in Group AAP subjects on rerandomization to placebo) were associated with a concurrent elevation in total bilirubin to more than twice the ULN. Both subjects were switched to open-label ADV, and the events resolved.
One subject from each treatment group had an increase in serum creatinine of more than 0.5 mg/dL from baseline, confirmed by a consecutive sample. However, in each case the peak value was within the normal range, and no action was taken regarding study medication dosing. The safety profile at week 52 for Group AAA subjects with YMDD-mutant HBV at baseline was similar to that observed for subjects with wild-type HBV.
After 52 weeks, the group receiving continuous ADV (Group AAA) had a significant improvement in quality of life in five of the eight domains assessed compared with that in the group rerandomized to placebo at week 40 (Group AAP). The five domains were role limitation due to physical problems; role limitation due to emotional problems, social function, vitality, and general health. At week 52, there was no significant difference in quality of life between Groups PAA and AAA.
After 12 weeks of treatment with ADV, Chinese subjects with HBeAg-positive CHB showed significant virological and biochemical improvement compared with those receiving placebo in this registration study. Continuous treatment with ADV over 52 weeks was well tolerated and resulted in the virological, biochemical, and serological improvements being maintained. The median HBV DNA reduction at week 52 in those subjects on continuous ADV therapy was 4.5 log10 copies/mL, with a maximum reduction of 8.0 log10 copies/mL. Some of the subjects only achieved a modest HBV DNA response (<3 log10 copies/mL reduction in 25% of subjects in week 52). Most subjects who were initially treated with ADV lost their treatment benefit after being re-randomized to receive placebo at week 40 (Group AAP). This is consistent with the findings of other studies conducted with ADV and other anti-HBV drugs and likely a result of lack of clearance of intracellular cccDNA, which is responsible for the reestablishment of viral replication.4, 11, 12, 30 Subjects who received placebo for the first 12 weeks followed by ADV had a better serological response than the subjects who received ADV for the entire 52-week period. It is not clear why these subjects had a better serological response, as the groups were well balanced and had similar disease characteristics at the time treatment was initiated, but may have been influenced by cases of spontaneous HBeAg seroconversion early in the study. However, the serological response of the patients in the group that received treatment with ADV throughout the study was lower than that of those previously reported in other studies.11, 13 It is possible this is a chance finding, and longer-term data from this study population are awaited. The response to treatment and the safety profile of subjects who had YMDD-mutant HBV at baseline were similar to those of subjects who had wild-type HBV at baseline.
The results of this study are consistent with previously published results for subjects with HBeAg-positive and HBeAg-negative CHB treated with 10 mg of ADV for 48 weeks and for subjects with lamivudine-resistant HBV and HIV coinfection.13, 14, 28, 29 Marcellin et al. reported a median HBV DNA reduction of 3.52 log10 copies/mL, ALT normalization in 48% of subjects, histological improvement in 53% of subjects, and a HBeAg seroconversion rate of 12% in subjects with HBeAg-positive CHB treated for 48 weeks.11 Hadziyannis et al. reported a median HBV DNA reduction of 3.91 log10 copies/mL, ALT normalization in 72%, and histological improvement in 64% of subjects with HBeAg-negative CHB treated for 48 weeks.12 No adefovir-related mutations were identified, which is consistent with the findings of other studies conducted for a similar duration.11, 13, 23
As previously reported, treatment with ADV for up to 52 weeks was well tolerated, and no changes in overall renal function were measured. As with other antiviral therapies, the observed distribution of increases in ALT when on treatment and when off treatment reinforces the need for careful monitoring when subjects discontinue treatment.
As in earlier studies, no evidence of the development of genotypic resistance to adefovir was found in any subject after 52 weeks of treatment.13, 14, 21–24, 28 Because many subjects are likely to require long-term treatment, the delayed onset and lower incidence of resistance makes ADV an attractive therapy option for CHB.
In conclusion, this study has shown that treatment with 10 mg of ADV daily over 52 weeks was safe and effective in Chinese subjects with HBeAg-positive CHB and that this treatment did not lead to the emergence of drug resistance. The study is ongoing and will evaluate the long-term efficacy, safety, and resistance profile of ADV for up to 5 years of treatment.
The authors thank all the other investigators in China who contributed to the study including XiaQiu Zhou (Shanghai Ruijin Hospital), DaoZhen Xu (Beijing Ditan Hospital), JiLu Yao (Sun Yet-San 3rd Hospital, Guangzhou), WeiLun Lu (Sun Yet-San 3rd Hospital, Guangzhou), YuMing Wang (Chongqing Xinan Hospital), Hong Ren (CMU 2nd Hospital, Chongqing), YaGang Chen (ZMU 1st Hospital, Hangzhou), and JunQi Niu (Jilin Unv 1st Hospital, Changchun); as well as Susan Scott (GlaxoSmithKline, UK), Jon Dixon (GlaxoSmithKline, UK), Alison Tan (GlaxoSmithKline, China), Cathy Huang (GlaxoSmithKline, China), and John Fry (San Francisco) for their contributions to this article.