Published Research

Chinese Herbal Medicine and Chemotherapy in the Treatment of Hepatocellular Carcinoma: A Meta-analysis of Randomized Controlled Trials

By Xiaojuan Shu, MPH (candidate), Michael McCulloch, LAc, MPH, Hang Xiao, MPH, PhD, Michael Broffman, LAc, and Jin Gao, MD, PhD 

Click here to download this study (including tables) in PDF format.

Hepatocellular carcinoma (HCC), one of the most common malignancies worldwide, is highly resistant to standard therapy. It is unclear whether chemotherapy, arte- rial embolization, or arterial chemoembolization improve survival advantage enough to justify their high toxicity. Treatment with Chinese herbal medicine has been explored, combining herbs that stimulate host immune response with those that have cytotoxic activity against HCC cells. The au- thors sought to evaluate the effectiveness of Chinese herbal medicine combined with chemotherapy. The hypothesis was that Chinese herbal medicine added to chemotherapy for the treatment of HCC would improve survival and tumor re- sponse, when compared to treatment with chemotherapy alone. Methods: The authors searched the databases TCMLARS, PubMed, and EMBASE as well as the bibliogra- phies of studies identified in the systematic search for po- tentially relevant titles or abstracts of studies in any language. They retained those that (1) treated only HCC pa- tients, (2) were described as randomized or reported that there was no statistical difference between treatment groups, (3) gave patients either Chinese herbal medicine therapy combined with chemotherapy in the treatment group or chemotherapy alone in the control group, and (4) provided data on the number of enrolled subjects and re- sponders and nonresponders for tumor response and sur- vival. The authors used random effects meta-analysis to combine data. Results: Twenty-six studies representing 2079 patients met the inclusion criteria. Chinese herbal medicine combined with chemotherapy, compared to chemotherapy alone, improved survival at 12 months (relative risk [RR], 1.55; 95% confidence interval [CI], 1.39-1.72; P < .000), 24 months (RR, 2.15; 95% CI, 1.75-2.64; P < .000), and 36 months (RR, 2.76; 95% CI, 1.95-3.91; P < .000). Tumor re- sponse increased (RR, 1.39; 95% CI, 1.24-1.56; P < .000). Conclusions: These findings provide promising evidence that combining Chinese herbal medicine with chemotherapy may benefit patients with HCC. Because of the low quality of these studies, these findings should be confirmed through conducting high-quality, rigorously controlled trials.

Hepatocellular carcinoma (HCC) is the fifth leading cause of cancer worldwide, with increasing incidence.1 The incidence ranges from 1 to 4 cases per 100 000 population in Western Europe and North America to 50 to 150 cases per 100 000 population in parts of Asia and Africa, where HCC is responsible for a large pro- portion of cancer deaths.2 Its treatment remains a con- troversial issue, despite the progress that has been made during the past decades, as most HCC patients are diagnosed at late stages and receive only palliative treatments not intended to extend survival.1 Untreated patients with advanced disease have a 1- year survival rate of 29%, a 2-year survival rate of 16%, and a 3-year survival rate of 8%.3

Arterial embolization, chemoembolization, and arterial chemotherapy are the major procedures in chemotherapy. Arterial embolization or chemoemboli- zation is a technique combining intra-arterial chemo- therapy and selective ischemia and was developed as an alternative to conventional systemic or intra-arterial chemotherapy. It is intended to induce tumor necrosis through administration of chemotherapy and an embolizing agent directly into the tumor by way of the feeding artery. The cytotoxic effect of arterial occlu- sion can be potentiated by labeling the infusion with radioactive isotopes or by adding cytotoxic drugs. Arterial embolization or chemoembolization includes procedures such as transcatheter arterial chemo- embolization, transarterial embolization, hepatic arterial chemoembolization, and hepatic arterial embolization. Arterial chemotherapy is not aimed to achieve arterial occlusion. It includes procedures such as transarterial infusion, hepatic arterial infusion, and intravenous chemotherapy. However, these proce- dures are associated with their own potentially life- threatening toxicities and complications, such as severe postembolization syndrome, hepatic insuffi- ciency, abscess, or infarction.4 In addition, they have shown only slightly improved survival or no survival advantages in previous studies.5-7 Meta-analysis of randomized controlled trials of treatment with chemoembolization or embolization has shown only a modest advantage in 2-year survival for patients with unresectable HCC, compared with conservative man- agement.8 The same meta-analysis revealed that tamoxifen provided no advantage in 1-year survival.8 Since the prognosis with existing therapy is so poor, we sought to evaluate the clinical evidence for effective- ness of Chinese herbal medicine in combination with systemic chemotherapy for the treatment of HCC.

In China, herbal medicine is frequently combined with chemotherapy in the treatment of liver cancer, usually in formulas that combine various Chinese herbs into one treatment strategy. In this meta-analysis, we sought to assess the effectiveness of this therapy for HCC by analyzing data from studies that compared treatment with Chinese herbal medicine combined with chemotherapy to treatment with chemotherapy alone.


Systematic Search
Retrieval of studies in all languages was performed through systematic searching of the databases TCMLARS (1984-August 2004;, PubMed (1966-August 2004;, and EMBASE (1974-August 2004; using the keywords liver cancer, chemotherapy, Chinese medicine, and randomized controlled trials. We also searched the bibliographies of review papers and published ran- domized controlled trials identified in our systematic search.

Criteria for Inclusion
To be included in this meta-analysis, a study had to ful- fill the following criteria: (1) only patients with HCC were included, (2) studies were described as random- ized or had the design elements of quasi-randomized studies (studies did not mention adoption of random- ization but reported there was no statistical difference between 2 groups), (3) HCC patients were given ei- ther Chinese herbal medicine therapy in combination with chemotherapy in the treatment group or chemo- therapy alone in the control group, and (4) authors provided necessary data on the number of responders and nonresponders for the end points of tumor re- sponse and survival sufficient to calculate relative risks (RRs) and 95% confidence intervals (CIs).

Criteria for Exclusion
Randomized controlled trials were excluded if they (1) were not described as randomized or quasi- randomized, (2) did not provide information con- cerning survival or tumor response, (3) included patients with cholangiocellular carcinomas or liver metastases of primary cancers other than liver, (4) did not contain an experimental arm treated with Chinese medicine combined with chemotherapy and control arm treated with chemotherapy alone, or (5) were du- plicate publications of other studies previously identi- fied in our systematic search. Decisions on whether to include or exclude trials and how to group studies for analysis by end points reported were made before conducting the meta-analysis.

Criteria for Combinability
Since there is no successful standard treatment for HCC, there are few published trials using exactly the same chemotherapy treatments and same Chinese herbal formula. As has been done in previous meta- analyses of chemotherapy for HCC, we combined studies in which patients in the treatment group were treated with the new therapy being evaluated and those in the control group were treated with standard chemotherapy alone.7-9 We also sought to separately meta-analyze any subgroups of studies that all used the same herbal combination.

Data Collection and Abstraction
Two reviewers (X.S. and M.M.) independently re- viewed the retained studies. Since prior work has shown that blinding during study selection and data extraction does not result in either a clinically or a sta- tistically significant effect on the summary effect mea- sure of a meta-analysis,10 we chose a nonblinded design for data extraction.

Analysis of Outcomes
The primary outcome of interest was the proportion of patients surviving at 6, 12, 24, and 36 months. The secondary outcome of interest was tumor response.

Survival. We calculated survival using the propor- tion of subjects surviving among the total number of  subjects, separately for each treatment group. The ratio for improvement in survival at each time point was calculated as the proportion of subjects alive in the Chinese herbal medicine combined with chemother- apy group divided by the proportion of subjects alive in the chemotherapy-only group.

Objective tumor response. We selected studies that reported tumor response using the World Health Organization scale.11 Following a method used in a previous meta-analysis of studies treating HCC patients with chemotherapy,12 we calculated tumor response as any response (complete response plus partial response) divided by the total (complete response plus partial response plus no change plus progressive disease), separately for each treatment group. The RR of tumor response was calculated as the probability of any tumor response in the herbal medicine combined with chemotherapy group divided by the probability of any tumor response in the chemotherapy-only group.

Pooled analysis. We used the Stata statistical software package (version 8.0; Stata Corp, College Station, Tex) for data management and analysis. All analyses were performed on an intention-to-treat basis. We used the random effects method to calculate pooled treatment effects.13 An RR value greater than 1.0 was considered to be consistent with a beneficial effect of Chinese herbal medicine combined with chemother- apy (vs chemotherapy alone) in the treatment of patients with HCC.

Qualitative Analysis

Study quality. To evaluate study quality, we used the Jadad scale, a validated 5-point scale developed to eval- uate the quality of reporting in studies included in a meta-analysis.14,15 The scale assigns a score of 0 or 1 for each of the following quality criteria, whether (1) the study was described as randomized, (2) the authors reported the method of randomization, (3) the use of blinding was reported, (4) the method for conceal- ment of allocation was reported, and (5) authors accounted for withdrawals and dropouts. Thus, a highest-quality study would receive a score of 5 and a lowest-quality study a score of 0.


Selection of Trials
After an initial screening of titles and abstracts, 385 potentially relevant clinical trials of HCC were identified. We then performed a second screening of these 385 studies and retained 126 published full papers, which were randomized controlled trials assessing Chinese herbal medicine in the treatment of HCC. After fur- ther evaluation, we excluded 9 studies because of nonrandomization, 13 studies because they lacked rel- evant end points, 24 studies because other cancers were included, 51 studies because they were missing either a qualified experimental group or controlled group, and 3 studies because they were duplicate pub- lications of other studies previously identified in our systematic search. This yielded a total of 24 random- ized controlled trials and 2 quasi–randomized con- trolled trials, which were identified as meeting protocol-specified inclusion criteria, totaling 2079 patients (Tables 1 and 2).

Treatment Effectiveness
Chemotherapy combined with Chinese herbal medi- cine, compared with chemotherapy alone, signifi- cantly improved survival at 12 months (RR, 1.55; 95% CI, 1.39-1.72; P < .000; Figure 1), 24 months (RR, 2.15; 95% CI, 1.75-2.64; P < .000; Figure 2), and 36 months (RR, 2.76; 95% CI, 1.95-3.91; P < .000; Figure 3). Tu- mor response was also significantly improved (RR, 1.39; 95% CI, 1.24-1.56; P < .000; Figure 4). There were no reliable data indicating significant survival improve- ment at 6 months since there was significant between- study heterogeneity (Table 3; figure not shown).

We identified 2 studies that used the exact same herbal formula (Hua Chan Su) and conducted a sepa- rate meta-analysis.16,17 The result also showed survival benefits of Hua Chan Su at 12 months (RR, 1.67; 95% CI, 1.38-2.02; P < .000; Table 3; figure not shown). It should be noted that Hua Chan Su contains bufotoxin, a digoxin-like substance that must be pro- vided in carefully measured dosages to avoid cardiac glycoside toxicity.

A Chinese-English translation table is provided as an appendix, which identifies the original names of herbal formulas as identified in included studies and the names as translated into English.

Our meta-analysis data suggest promising evidence that Chinese herbal medicine in the treatment of HCC may have potential therapeutic value. Herbal medicine combined with chemotherapy for the treat- ment of HCC can improve survival at 12, 24, and 36 months when compared with chemotherapy alone. There is additional evidence that 6-month survival may also be improved; however, this finding is limited by the presence of between-study heterogeneity. Tu- mor response was also improved by the addition of Chinese herbal medicine to chemotherapy.

In a separate meta-analysis of the only 2 studies using the exact same herbal medicine, Hua Chan Su, we found that 12-month survival was also increased.16,17


Publication bias. We found evidence for publication bias in the data for meta-analysis of survival at 6 and 12 months for all herbal formulas combined. We also found evidence for publication bias in the data for meta-analysis of tumor response.

Research techniques used in Chinese studies. Because the studies we found were of generally poor quality, we are unable to make definite conclusions from our data. Published studies from China were found to be more highly condensed than typical articles published in the Western literature, with key details of study design omitted, especially details concerning blinding of subjects and clinicians. In addition, since most stud- ies did not describe subject withdrawals or dropouts, it is not possible to adjust the analysis for censoring, and therefore pooled estimates of survival should be inter- preted with caution. However, inadequate reporting of specific details of randomization is also found in many Western medical journals. In 2004, 8 years after publication of the CONSORT Statement (intended to improve the quality of reporting of results in random- ized controlled trials), more than 40% of trials published in Western medical journals failed to use

adequate randomization methods or failed to report the method for concealment of allocation.18 In 1994, this figure was as high as 70% to 80%.19

An additional comment on the issue of blinding in these studies is warranted. Most of the studies (20 of 26) assessed in this meta-analysis provided the treat- ment group with orally administered Chinese herbal medicine combined with intravenous chemotherapy and the control group only intravenous chemother- apy. Therefore, blinding of subjects and clinicians was not possible. In such a study design setting, blinding could be achieved only if the treatment and control groups received both oral and intravenous trial medi- cations (ie, the treatment group was given true orally administered herbal medicine and intravenous che- motherapy and the control group was given placebo orally administered herbal medicine and intravenous chemotherapy).

In future trials, we propose that Chinese investiga- tors employ relatively simple measures such as ran- dom number–generating software for use in

randomization and placebo trial drugs for use in com- paring different therapies or in evaluating new and emerging extracted active components of Chinese herbal medicines. More thorough reporting of patients’ characteristics and accounting for patient withdrawals and dropouts would also be helpful.

Standardized herbal formulas. Since there is no stan- dard Chinese formula for the treatment of HCC, it is hard to identify the active components in the treat- ments. Further investigation is needed on effects of Chinese herbal medicine on HCC, on the identity of active components of Chinese herbal medicine, and on the therapeutic mechanisms underlying possible survival benefits of combining Chinese herbal medicine with chemotherapy for HCC.

Administrative and methodological support provided by the Pine Street Foundation (San Anselmo, Calif;