Double-blinded placebo-controlled study of Phyllanthus urinaris for the treatment of chronic hepatitis B


Professor J. Y. Y. Sung, Department of Medicine and Therapeutics, 9/F Prince of Wales Hospital, Shatin, Hong Kong.


Background : Previous studies suggested that Phyllanthus species have an anti-viral effect on hepatitis B, but methodologies have been inadequate.

Aims : This study aimed to investigate the anti-viral effect of Phyllanthus urinaris.

Methods : Chronic hepatitis B patients with positive hepatitis B e-antigen (HBeAg), hepatitis B virus (HBV) DNA > 500 000 copies/mL and elevated alanine transaminase (ALT) were recruited. Patients were randomized into groups of 12 receiving P. urinaris 1, 2 and 3 g three times daily for 6 months or placebo (six cases). The primary endpoint was HBV DNA reduction, and secondary endpoints were HBeAg seroconversion and ALT normalization.

Results : On an intention-to-treat analysis there was no difference in log10[HBV DNA] reduction of the Phyllanthus 1-g (0.18 ± 1.42), 2-g (0.33 ± 1.08) and 3-g (0.85 ± 1.30) groups vs. placebo (0.28 ± 0.85) (P = 0.90, 0.92 and 0.38, respectively) at the end of treatment. The percentage of patients among the placebo, Phyllanthus 1-g, 2-g and 3-g groups undergoing HBeAg seroconversion (0%, 9.1%, 8.3% and 16.7%, respectively) and ALT normalization (0%, 0%, 8.3% and 33.3%) were not significantly different at the end of treatment. No delayed virological or biochemical response was documented at 24 weeks after the cessation of treatment. No serious adverse event was reported.

Conclusion : P. urinaris treatment for 6 months has no demonstrable anti-viral effect in chronic hepatitis B.


The plant of the genus Phyllanthus has been widely used in traditional Chinese medicine for the treatment of chronic hepatitis B virus (HBV) infection. In vitro studies suggested that different Phyllanthus species could inhibit DNA polymerase activity and mRNA transcription and hence suppress the replication of HBV.1–4 A recent study has also demonstrated anti-inflammatory potential of Phyllanthus amarus in vitro and in vivo.5 A preliminary study in India using P. amarus claimed a remarkable benefit in HBsAg clearance,6 but this result could not be reproduced by other groups using the same Phyllanthus species.7, 8 Recently, a systemic review by the Cochrane Hepato-Biliary Group suggested that Phyllanthus species possessed activities in clearance of HBsAg, HBeAg and HBV DNA as well as normalization of liver enzymes.9 However, most of the previous studies devoided proper randomization, used a relatively small sample size and lacked an intention-to-treat analysis.

In a pilot study conducted in China, patients received Phyllanthus urinaria from Henan were found more likely to lose HBeAg than those receiving Phyllanthus niruria or P. amarus.8 Unfortunately, a lack of clearly defined patient selection criteria and randomization strategy and a lack of HBV DNA testing and side-effect profile documentation weakened the data of this study. A properly designed clinical trial would be warranted in order to assess the anti-viral efficacy of P. urinaris. We conducted a phase-II, dose-finding study to evaluate the anti-viral efficacy and safety of P. urinaris in adults with chronic HBV infection.

Patients and methods

Preparation of Phyllanthus urinaris

One lot of P. urinaria was obtained from Hainan Province of China, and the voucher specimen was kept at the Guangzhou University of Traditional Chinese Medicine. Dried whole plant was washed and cut to pieces < 1 cm in size. Plant material was mixed with five times v/w of 80% ethanol and extracted by reflux at 80 °C three times for 2, 1 and 1 h, respectively. The extracts were combined, filtered, and the total volume was reduced to 15% by removing the ethanol under reduced pressure. The concentrated extract was stored overnight at 10 °C, filtered and concentrated to a plaster. The final product was tested for micro-organism contamination, heavy metals and insecticides according to World Health Organization standards for food. Gallic acid, an ingredient of P. urinaria, was used as a chemical marker. The content of gallic acid in the final powder product was not less than 2.0 mg/g.


Eligible patients included males and females aged 18–65 with positive hepatitis B surface antigen (HBsAg) for at least 6 months. These patients must also have positive (HBeAg), serum HBV DNA (as determined by DNA crossing-linking assay; NAXCOR XLnt™, NAXCOR, Menlo Park, CA) of at least 0.5 × 106 copies/mL and alanine transaminase (ALT) 1.5–5 times upper limit of normal at the screening visit. Patients were excluded if they had decompensated liver disease (defined by serum albumin≤36 g/dL, bilirubin ≥ 15 g/dL, prothrombin time ≥ 2 s prolonged, or a history of ascites, variceal haemorrhage or hepatic encephalopathy), pancytopenia (defined as haemoglobin < 11 g/dL, white cell count < 4000/mm3 or platelets < 105/mm3). Patients with a history of using interferon or anti-viral agents were also excluded. Women of child-bearing age must have shown a negative urine pregnancy test at screening and before randomization. Other exclusion criteria included co-infection by hepatitis C virus, a history of hepatocellular carcinoma, serious medical illness, active substance or alcohol abuse and concurrent use of corticosteroids or immunosuppressive agents. The study was approved by the ethics committee of the Chinese University of Hong Kong, and all patients gave witnessed written informed consent before enrolment.

Study design

This study was a phase II, double-blinded, dose-finding, randomized trial of three different doses of P. urinaris (1, 2 and 3 g three times daily; abbreviated as Phyllanthus 1, Phyllanthus 2 and Phyllanthus 3 groups, respectively) vs. placebo in a ratio of 2 : 2 : 2 : 1 for 24 weeks. The study medication was prepared as Phyllanthus 500 mg or placebo capsule, and each patient was required to take six capsules (either study drug, placebo or combination) three times daily to make up the tested dosing. Randomization and treatment assignment were performed within 4 weeks after the screening procedures had been satisfactorily completed. Randomization was performed by random numbers generated through a computer by a third party who was not involved in patient management. The random numbers were placed in concealed envelopes. Unblinding was only allowed in cases of medical emergency in which a knowledge of the investigational drug was critical to patient management.

After the first clinic visit at baseline, patients were asked to return at weeks 1, 2 and 4, and then every 4 weeks thereafter until the end of 24 weeks. After the cessation of the investigational drug, patients were followed up every 4 weeks until the end of week 48. At each clinic visit, laboratory tests were performed to assess liver biochemistry, safety of treatment and adverse events. Serum was assayed for HBV DNA (at baseline, weeks 24 and 48), HBeAg and antibody to HBeAg (anti-HBe) (at baseline, weeks 12, 24, 36 and 48), and HBsAg and antibody to HBsAg (anti-HBs) (at baseline, weeks 24 and 48). Serum samples in which HBV DNA was undetectable by DNA cross-linking assay (NAXCOR) were tested by real-time polymerase chain reaction (PCR), which had a detection limit of 102 viral copies/mL.

Laboratory assays

Serological assays.

HBsAg and antibody to hepatitis C virus were tested by commercially available enzyme-linked immunosorbent assay kits (Abbott GmBH Diagnostika, Wiesbaden-Delkenheim, Germany). HBeAg and anti-HBe were measured by enzyme-linked immunosorbent assay (Sanofi Diagnostics, Pasteur, France). Serum HBV DNA was quantified using the DNA cross-linking assay (NAXCOR XLnt™, NAXCOR, Menlo Park, CA) that had a detection limit of 0.5 × 106 copies/mL as described previously.10 Serum samples that had undetectable HBV DNA by the cross-linking assay were tested by the TaqMan real-time PCR system (PE Biosystems, Foster City, CA) as described previously.11, 12 In brief, PCR primers flanking HBV genome from nucleotides 1549–1653 (sense primer 5′-CCGTCTGTGCCTTCTCATCTG and antisense primer 5′-AGTCCAAGAGTYCTCTTATGYAAGACCTT) and a fluorescent probe (5′-CCGTGTGCACTTCGCTTCACCTCTGC) were used. HBV DNA was extracted by a QIAGEN QIAamp DNA Mini Kit (Qiagen Inc., Chatsworth, CA) according to the manufacturer's instructions. The reaction mixture was subjected to 45 cycles of PCR, and the intensities of the fluorescent dyes in each reaction were read automatically during PCR cycling in a PE-Applied Biosystem Detector 7700 machine. The data were analysed by sequence detector software (version 1.7a; Perkin Elmer, Inc., Foster City, California). A standard curve was generated by serial 10-fold dilution of EUROHEP genotype D HBV standard (from Dr KH Heerman, University of Goettingen, Goettingen, Germany) that contained 2.7 × 109 viral copies/mL. The range of HBV DNA detection was from 102 to 109 copies/mL with a correlation coefficient of the standard curve routinely greater than 0.990. The standard curve was used to calculate the precise quantities of HBV DNA molecules for the unknown samples.

Endpoints and statistical analysis.

The primary end-point was a reduction of mean HBV DNA by at least 2 log10 units at the end of treatment (week 24). Assuming that the standard deviation of mean HBV DNA reduction was 1 log10 unit, and allowing a patient dropout rate of 20%, we estimated that 12 patients in each treatment group and six patients in the control group would give a power of 85% to detect a 2 log10 difference in mean HBV DNA levels between any of the treatment groups vs. the placebo group at an α-level of 0.05.

Secondary endpoints included HBeAg seroconversion, HBsAg clearance and ALT normalization at the end of treatment and the end of follow-up (week 48), as well as reduction of mean HBV DNA levels at the end of follow-up. The intention-to-treat analysis included all randomized patients who were HBeAg-positive at baseline and received at least one dose of the study medication.

Statistical tests were performed by SPSS software (version 11.0). HBV DNA was logarithmically transformed to a normal distribution, and log10[HBV DNA] reductions among different treatment groups were compared by Student's t-test. The reduction of HBV DNA on escalating the dose of Phyllanthus was assessed by repeated measures analysis of variance. The proportions of patients reaching different virological and biochemical endpoints as well as the adverse events were compared by the chi-square (χ2-test) test and Fisher's exact test as deemed appropriate. Statistical significance was taken as P < 0.05. Bonferroni's correction was used for multiple comparisons. All statistical tests were two-sided.


Study population

The study was conducted between July 2000 and December 2002. Eighty Chinese patients were screened and 42 patients were enrolled. Twelve patients were randomly assigned to receive 1, 2 or 3 g of P. urinaris and six patients received placebo. Two patients, one in the placebo group and another in the Phyllanthus 1 group, were excluded because of negative HBeAg at baseline visit. Forty patients were included in the intention-to-treat analysis.

The baseline characteristics of the study patients are summarized in Table 1. There was no significant difference in the age, gender ratio, body weight, ALT levels and log10HBV DNA between any of the Phyllanthus-treated groups and the placebo group.

Table 1.  Baseline characteristics of the study patients
(n = 5)
(n = 11)
(n = 12)
(n = 12)
  1. P1, P2 and P3 are the P-values for comparisons of Phyllanthus-1, Phyllanthus-2 and Phyllanthus-3 groups vs. placebo, respectively; statistical significant P-value ≤0.0125.

Age (year)
 Mean (s.d.)33 (8)38 (8)33 (11)36 (9)0.230.990.48
 Median (range)33 (25–44)39 (19–55)33 (19–48)37 (21–51)   
Male sex (%)204558670.590.290.13
Body weight (kg)
 Mean (s.d.)49 (10)61 (13)68 (16)65 (12)0.110.0230.018
 Median (range)48 (42–65)57 (45–88)65 (49–95)67 (42–84)   
ALT (IU/l)
 Mean (s.d.)114 (33)130 (59)139 (65)124 (47)0.580.430.69
 Median (range)107 (91–170)125 (58–276)122 56–283)129 (50–222)   
Log10[HBV DNA}
 Mean (s.d.)8.15 (1.39)7.58 (0.87)7.94 (0.93)7.83 (1.14)0.320.720.62
 Median (range)8.42 (6.67–9.57)8.00 (6.12–8.70)7.51 (6.71–9.15)7.56 (6.38–9.47)   

Virological response

The reduction in HBV DNA is shown in Table 2. Six and 11 of 40 samples at the end of treatment and end of follow-up, respectively, had undetectable HBV DNA by cross-linking assay requiring real-time PCR for quantification. The mean log10[HBV DNA] reduction in all treatment groups was less than 1 at the end of treatment (week 24) and at the end of follow-up (week 48). There was no difference in the HBV DNA reduction by any treatment group as compared to placebo group. By repeated measurement of analysis of variance, interaction between time and treatment was not significant (P = 0.81), and there was no significant difference (P = 0.67) in HBV DNA levels among different groups (Figure 1).

Table 2.  Serological and biochemical response in different treatment groups as compared to placebo group
  1. EOT, end of treatment; EOS = end of study. P1, P2 and P3 are the P-values for comparisons of Phyllanthus-1, Phyllanthus-2 and Phyllanthus-3 groups vs. placebo, respectively; statistical significant P-value ≤0.0125.

Log10[HBV DNA] reduction
 Mean ± s.d.0.28 ± 0.850.18 ± 1.420.33 ± 1.080.85 ± 1.300.900.920.38
 Median, range−0.043, –0.60 to 1.420.13, –0.31 to 1.90−0.044, −1.44 to 2.600.24, –0.39 to 3.77   
 Mean ± s.d.0.50 ± 0.810.83 ± 1.731.30 ± 1.261.06 ± 1.900.690.210.54
 Median, range0.60, –0.71 to 1.400.25, −1.17 to 4.021.39, –0.41 to 3.420.48, –0.99 to 4.82   
HBeAg seroconversion
 EOT n (%)0 (0)1 (9)1 (8)2 (16)
 EOS n (%)1 (20)4 (36)2 (16)1 (8)
ALT normalization
 EOT n (%)0 (0)0 (0)1 (8)4 (33)
 EOS n (%)3 (60)3 (27)3 (25)2 (16)
Figure 1.

Estimated marginal means of log10[HBV DNA] at baseline, end of treatment (week 24) and end of study (week 48) of different treatment groups as analysed by repeated measures analysis of variance. Solid line, placebo; broken lines, Phyllanthus (▵, Phyllanthus-1; □, Phyllanthus-2; ○, Phyllanthus-3). There was no difference in HBV DNA levels among different groups.

Serological response

There was no difference in the proportion of patients developing HBeAg seroconversion in any of the treatment groups as compared to the placebo group both at the end of treatment as well as at end of post-treatment follow-up (Table 2). There was also no difference in proportion of patients developing HBeAg seroconversion on escalating the dosage of Phyllanthus at the end of treatment (P = 0.82) or the end of post-treatment follow-up (P = 0.40). None of the patients in this study developed HBsAg clearance.

Biochemical response

There was no difference in the proportion of patients who had normalization of ALT in the treatment groups as well as the placebo group, both at the end of treatment and end of post-treatment follow-up (Table 2). There was no difference in proportion of patients with normalization of ALT on escalating the dosage of Phyllanthus at the end of treatment (P = 0.065) or at the end of post-treatment follow-up (P = 0.34).

Adverse events

Phyllanthus urinaris was well tolerated, with an incidence of adverse events in the various dosage groups similar to that for placebo. The majority of adverse events were mild to moderate. The incidence of adverse events did not reveal a significant pattern related to the drug dosage. The commonest adverse events are shown in Table 3. No patient died and none required liver transplantation in this study.

Table 3.  Adverse events in patients reported by greater than 10% of patients in any of the treatment groups or placebo group during treatment
Adverse eventsPlaceboPhyllanthus-1Phyllanthus-2Phyllanthus-3P1P2P3
  1. Counts of adverse events represent patients, not events. Therefore, a patient who experienced any symptom more than once was counted once only.

  2. URTI, upper respiratory tract infection. P1, P2 and P3 are the P-values for comparisons of Phyllanthus-1, Phyllanthus-2 and Phyllanthus-3 groups vs. placebo, respectively; statistically significant P-value ≤0.0125 after Bonferroni's correction.

Any adverse event4 (80%)8 (72.7%)12 (100%)10 (83.3%)
Abdominal pain2 (40%)5 (45.5%)6 (50%)2 (25%)
URTI symptoms2 (40%)5 (45.5%)8 (66.7)7 (58.3%)1.000.590.62
Malaise03 (27.3%)2 (16.7%)3 (25%)0.511.000.51
Headache04 (36.3%)3 (25%)3 (25%)0.240.510.51
Insonmia2 (40%)1 (9.1%)1 (8.3%)1 (8.3%)
Increased appetite2 (40%)001 (8.3%)0.0830.0740.19
Increased flatulence0002 (16.7%)
Myalgia1 (20%)02 (16.7%)2 (16.7%)0.311.001.00
Dizziness1 (20%)3 (27.3%)2 (16.7%)2 (16.7%)
Rash1 (20%)03 (25%)00.311.000.29
Blurred vision1 (20%)1 (9.1%)
Fever001 (8.3%)2 (16.7%)
Frequency1 (20%)01 (8.3%)00.310.510.29
Dysuria1 (20%)0000.310.290.29
Dental caries02 (18.2%)
Gingivitis002 (16.7%)

One patient in the Phyllanthus 1 group was removed from the study because of a flare-up of hepatitis. Her total bilirubin was elevated to 42 mmol/L and ALT to 459 IU/L at 21 weeks of treatment. The study drug was stopped and lamivudine 100 mg daily was given. Her bilirubin and ALT gradually returned to normal. Another patient in the Phyllanthus 1 group withdrew at week 12 for personal reasons. Otherwise there was no discontinuation of treatment in the remaining 38 patients.

One patient in the Phyllanthus 2 group had creatine kinase raised to 679 IU/L at one visit, but it resolved spontaneously on subsequent visits. No serious biochemical abnormalities were experienced by any other patients in all four study groups.


As previous studies suggested that Phyllanthus species have an anti-viral effect on HBV, the primary end-point of this study was taken as an HBV DNA level reduction by Phyllanthus vs. placebo treatment. A 2 log10 HBV DNA reduction was selected, as most anti-viral agents (including lamivudine) can achieve a mean of 3 log10 or greater reduction in HBV DNA.13, 14 In this study, there was no demonstrable effect of the herbal treatment on HBV DNA levels by all three doses of P. urinaria treatment. The magnitudes of changes were small in terms of logarithmic reduction and were not significantly different from that of the placebo group. This highlighted the importance of including a placebo-controlled study and the use of logarithmic HBV DNA reduction rather than percentage HBV DNA reduction in the study of anti-viral efficacy. Loss of HBV DNA was used as an indicator of treatment response in most studies.9 However, previous HBV DNA assays had limited sensitivity, and undetectable HBV DNA might not mean clearance of the virus.10 In this study, the highly sensitive real-time PCR HBV DNA assay was used and the magnitude of HBV DNA reduction could be more accurately measured, particularly among samples which had lower HBV DNA levels before the commencement of treatment.

The dosage of Phyllanthus species in the treatment of chronic hepatitis B is uncertain. In a study from India, P. amarus 500 mg was prepared in capsules given orally three times daily.15 Other groups used different Phyllanthus species at dosages ranging from 300 to 900 mg three times daily.8 Owing to the inconspicuous anti-viral effects in previous studies and the good safety profile as reported in the past, we evaluated three different relatively higher dosages (1000–3000 mg three times daily) of P. urinaris to test for any dose-related anti-viral effect. Most patients experienced mild to moderate adverse effects on taking the studied capsules, and similar complaints were documented in patients who received Phyllanthus and placebo treatment. The commonest adverse event was abdominal discomfort, which was likely related to the large number of capsules one has to ingest.

The high rate of HBsAg clearance (up to 59%) and HBeAg loss (up to 64%) found in previous studies was not demonstrated in this study.9 The percentage of patients undergoing HBeAg seroconversion in the treatment groups (ranging from 8 to 36%) was similar to that of the placebo group as well as the previously reported spontaneous HBeAg seroconversion rates among patients in South-East Asia.16, 17 None of the patients in this study had HBsAg clearance. The discrepant results of this study from others might be related to the patient selection criteria or the species of Phyllanthus plant under study. The inadequate methodological designs of previous trials also imposed uncertainties on the validity of their results.9

In summary, this double-blinded, placebo-controlled, dose-finding, randomized study showed no clinically significant anti-viral effect in chronic hepatitis B by the herbal medicinal plant P. urinaris. Further studies of other different Phyllanthus species using critical clinical trial designs are warranted.


We are grateful for the financial support of UGC – Area of Excellence in Chinese Medicine Research, Hong Kong. We thank Xinhua Wang and Xinbai Guo from Guangzhou University of Traditional Chinese Medicine for the preparation of P. urinaris.