The Efficacy of Chinese Herbal Medicine as an Adjunctive Therapy for Advanced Non-small Cell Lung Cancer: A Systematic Review and Meta-analysis

Many published studies reflect the growing application of complementary and alternative medicine, particularly Chinese herbal medicine (CHM) use in combination with conventional cancer therapy for advanced non-small cell lung cancer (NSCLC), but its efficacy remains largely unexplored. The purpose of this study is to evaluate the efficacy of CHM combined with conventional chemotherapy (CT) in the treatment of advanced NSCLC. Publications in 11 electronic databases were extensively searched, and 24 trials were included for analysis. A sum of 2,109 patients was enrolled in these studies, at which 1,064 patients participated in CT combined CHM and 1,039 in CT (six patients dropped out and were not reported the group enrolled). Compared to using CT alone, CHM combined with CT significantly increase one-year survival rate (RR = 1.36, 95% CI = 1.15–1.60, p = 0.0003). Besides, the combined therapy significantly increased immediate tumor response (RR = 1.36, 95% CI = 1.19–1.56, p<1.0E−5) and improved Karnofsky performance score (KPS) (RR = 2.90, 95% CI = 1.62–5.18, p = 0.0003). Combined therapy remarkably reduced the nausea and vomiting at toxicity grade of III–IV (RR = 0.24, 95% CI = 0.12–0.50, p = 0.0001) and prevented the decline of hemoglobin and platelet in patients under CT at toxicity grade of I–IV (RR = 0.64, 95% CI = 0.51–0.80, p<0.0001). Moreover, the herbs that are frequently used in NSCLC patients were identified. This systematic review suggests that CHM as an adjuvant therapy can reduce CT toxicity, prolong survival rate, enhance immediate tumor response, and improve KPS in advanced NSCLC patients. However, due to the lack of large-scale randomized clinical trials in the included studies, further larger scale trials are needed.


Introduction
Lung cancer is the most common malignancy worldwide and a leading cause of cancer-related deaths. In 2012, it is estimated that 160,300 deaths (87,700 in men, 72,600 in women) from lung cancer would occur in the United States [1].
The non-small cell lung cancer (NSCLC) is the most common form of lung cancer, which accounts for approximately 85% of all lung cancer cases. Nowadays the standard treatment for patient with advanced NSCLC who has a good performance status, platinum-based chemotherapy (CT) is the first-line regimen [2,3]. However, platinum-based CT has the potential for severe adverse events, and the International Adjuvant Lung Cancer Trial suggests that more deaths in the CT group and the benefit of CT decreased over time [4,5]. In spite of the development of new CT regimens use in the treatment of NSCLC, the prognosis of the patients remains poor. Its five-year survival rate is as low as 15.9% [6,7]. Thus, there is an increasing awareness to maximize tumor control, prolong overall survival, minimize CT side-effects and improve quality of life (QoL).
In complementary and alternative medicine (CAM), Chinese herbal medicine (CHM) has become increasingly popular for the patients with advanced NSCLC. Recent studies have reported some CHMs in associated with platinum-based CT have definite superiority in relieving the symptoms of lung cancer patients, reducing the severe adverse effects (AEs) of standard cancer therapy, enhancing short-term efficacy and improving patients QoL [8]. There are a variety of herbs being used in different combinations and forms, such as oral administration and intravenous injection, to treat advanced NSCLC combined with platinum-based CT. However, albeit a number of studies being published in Chinese, the evidence on efficacy of CHM as adjuvant therapy to conventional CT is not well demonstrated in the Western world. Thus, the aim of this study is to carry out a comprehensive systematic review about the efficacy of CHM as an adjunctive therapy for advanced NSCLC.

Methods
Clinical trials were retrieved from 11 databases as well as from conference papers and theses. The studies were reviewed independently by SGL and HYC. The data from included studies were extracted by the first reviewer and verified by the second reviewer. Discrepancies were rectified referring to the original articles. Only the studies which satisfied the criteria were included in the meta-analysis.

Search Strategy
The terms retrieved in databases were as the following: (

Inclusion Criteria
Studies included in the meta-analysis had to meet all of the following criteria: (1) Participants: NSCLC patients had to be diagnosed by pathological sections and were treated by the CT. (2) Type of studies: only clinical randomized controlled trials (RCTs) were eligible. (3) Type of intervention: studies reported CT combined with or without CHM. For studies using other agents as the third arm, only the two arms using CHM and/or CT will be included for analysis. (4) Type of outcome measurements: overall survival rate and tumor response were the main outcome measurements; other outcome measurement included reduction in AEs of CT, improvement in clinical symptoms and blood disorders were also considered.

Exclusion Criteria
Clinical trials were excluded if they did not meet the above criteria. In addition, studies with the followings were also excluded: (1) CHM were used in both of the intervention group and control group; (2) Non-original research (e.g. review article, letter to the editor); (3) Duplicated publications.

Outcome Measures
Survival rate, tumor response of CHM on the number of patients with complete response (CR) or partial response (PR), as well as those with progressive disease (PD) based on the WHO scale were examined. The improved or stable performance status of patients were investigated based on the Karnofsky performance score (KPS), in which 100 refers to a normal subject with no complaints, 70 refers to a patient unable to carry on normal activity, 50 refers to a patient who requires considerable assistance, 40 refers to a disabled patient and 30 refers to a hospitalizationrecommended patient. The efficacy of CHM on relieving the AEs of CT including nausea and vomiting, hemoglobin (HB), platelet (PLT) were studied by grading the acute and subacute AEs of cancer treatment.

Quality Assessment
Methodological quality of RCTs was assessed using the fivepoint Jadad scale [9]. All trials were reviewed by at least two reviewers and any disagreement was resolved by third reviewer consensus. In addition, the risk of bias for the included studies was also assessed.

Data Analysis
The Review Manager 5.1 software (Nordic Cochrane Centre, Copenhagen, Denmark) was employed for data analysis. The effect data is expressed as relative risk (RR) with 95% confidence interval (CI). If the heterogeneity exists in pooled studies (I 2 .50%), a random model was applied, otherwise the fix model was applied. Statistic significant difference was considered as p,0.05.

Characteristics of the Included Studies
In the study, 2,998 articles were retrieved. 30 studies were finally included. Among these studies, six studies were not pooled for analysis as four studies had Jadad score 2 [10][11][12][13], and other two studies did not report tumor-node-metastasis (TNM) staging information [14,15]. Therefore, 24 eligible studies were included for meta-analysis. The study selection process details were described in Figure 1. A sum of 2,109 patients was enrolled in these studies, at which 1,064 patients participated in CT combined CHM (CTC) and 1,039 in CT (six patients dropped out and were not reported the group enrolled). A total 78 patients withdraw or dropped out, 36 patients in CTC, 36 in CT and six patients in groups not specified.
All the patients recruited in the 24 studies were at stage III to IV of NSCLC TNM, and all of the studies were graded at least Jadad score 3. The risk of bias of all studies was shown in Table S1. The course of treatment varied from 4 to 16 weeks in the included studies. A list of therapeutic approaches and outcome assessment in each study, and the quality of studies assessed by five-point Jadad scale was listed in Table 1. All studies had claimed the baseline were comparable except one study [32] including age, gender, histopathology or TNM stage as shown in Table S2.

Performance Status
Two types of KPS data were reported in the studies, the improvement of KPS (ten-point cutoff) and the value of KPS in pre-and post-treatment. Six studies of the 24 studies, with evaluation of 526 patients were being analyzed. 35.1% and 10.9% of patients reported improved (the increase of KPS$10) in CTC (n = 270) and in CT (n = 256), respectively. Significant findings with improvement were shown in the CTC (RR = 2.90, 95% CI = 1.62-5.18, p = 0.0003, six studies, 526 patients) (Figure 4) [21,26,29,30,35,36]. There was no significant heterogeneity among these studies (I 2 = 51%). Dropping one of any studies did not alter the result that favor of CTC. The value of KPS was reported with pre-treatment in seven studies [19,[24][25][26]30,34,37] and post-treatment in four studies [24,25,34,37]. The pooled studies showed that the KPS of pre-treatment had no significant difference in CTC and CT (SMD = 20.04, 95% CI = 20.20-0.12, p = 0.64, I 2 = 0%). However, the pooled studies indicated the heterogeneity in the four studies of post-treatment (SMD = 1.03, 95% CI = 20.09-2.14, p = 0.07, I 2 = 95%). Interestingly, these four studies all claimed significant improvement in CTC compared to CT.

Reduction in CT Toxicity
Nausea and vomiting are common AEs of CT. A significant reduction of nausea and vomiting at toxicity grade of III-IV in CTC compared to CT therapy was found (RR = 0.24, 95% CI = 0.12-0.50, p = 1.0E24, five studies, 350 patients) [19,24,25,30,31] (Figure 5A). However, there was significant heterogeneity in the studies with reduction of nausea and vomiting at toxicity grade of I-IV (data not shown). One study reported a significant reduction of nausea and vomiting in CTC compared to CT at the 10 th day of second treatment, based on the questionnaire EORTC QLQ-LC43 (combination of EORTC QLQ-C30 and QLQ-LC13) [38]. The study was not pooled due to the different data types.

Herbs Frequently used in NSCLC
19 studies have reported herbs and decoctions. Among them, Radix Astragalus, Radix Adenophorae and Radix Ophiopogonis are the most commonly used herbs for NSCLC (Table 2).

Discussion
Recent studies showed that the use of CAM has increasingly gained recognition and usage in cancer patients [40,41]. CHM is especially popular among the CAM usage as palliative care for cancer patients, but the efficacy of the combined use of CAM and CT on advanced NSCLC cancer remains under explored due to language barrier of many studies reported in Chinese language [42]. Meta-analysis is a powerful statistical analysis of results from individual studies, which increases the precision of a treatment effect and settles controversies studies [43]. In the present study, the pooled data with advanced NSCLC has shown that combined therapy significantly improved the survival rates, immediate tumor response and performance status of advanced NSCLC patients. We also found that, when compared with CT alone, the CTC therapy significantly reduce AEs associated with chemotherapeutic interventions, including nausea and vomiting, decrease in the peripheral blood leukocytes and HB ( Figure 5 and Figure 6).
In contrast to most of the previous meta-analyses in this area, our systematic review set the inclusion criteria with Jadad score $3 to increase the study quality and included studies with stage III-IV of TNM to minimize the heterogeneity among the studies. It is also encouraging to see that the adjunctive use of CHM with CT may prolong survival in advanced stage NSCLC patients, and CT-related side effects appear to be less frequent and milder in the use of concomitant CAM treatment, which suggest CAM may enhance the compliance to CT and eventually result in improving KPS of patients. The efficacy of CHM as an adjuvant therapy for NSCLC is in line with our previous findings in colorectal cancer, hepatocellular carcinoma and nasopharyngeal carcinoma [44][45][46].
According to Chinese medicine theory, illness is caused by the disharmony of yin and yang, and Chinese medicine aims to restore the balance of yin and yang to alleviate the disease symptoms. CHMs have been commonly used in Asia for thousands of years. Interestingly, it is reported that 61 out of 74 lung cancer patients recruited by 17 Community Clinical Oncology Program (CCOP) affiliates throughout the United States have used CAM while undergoing CT or radiation therapy [47]. However, the potential for integrating CAM into conventional CT treatment remains to be evaluated [48]. Radix Astragalus, Radix Adenophorae, Radix Ophiopogonis, Radix Glycyrrhizae, Poria and Oldenlandia diffusa identified in the study have the function of tonifying qi, nourishing yin and removing blood stasis, which are in accordance to the commonest symptoms in NSCLC patients undergoing CT (i.e. blood stasis, vital energy and yin deficiency). It is in line with other findings that Radix Astragalus, Radix Ophiopogonis and Oldenlandia diffusa are commonly used for NSCLC [49]. Besides, it has also been reported that there was a significant effectiveness of adding astragalus-based herbal treatment to standard CT regimens [50]. Several experimental researches have revealed that Adenophora Polysaccharides and Oldenlandia diffusa extract could effectively inhibited the growth of cancer cell lines and induced significant increase of apoptosis. Furthermore, there was a significant inhibition of lung metastases in animal model with no noticeable AEs [51,52]. Radix Adenophorae and Radix Ophiopogonis were also shown to have anti-inflammation and immunomodulating effects [53,54]. Our previous studies have demonstrated that Radix Astragalus, the dried root of Astragalus membranaceus, has anti-tumor, immunomodulating and immunorestorative effects in vivo and in vitro [55,56]. It is in accordance with other findings that Radix Astragulus has immunologic benefits by stimulating macrophage, natural killer cell activity but inhibiting T-helper cell type 2 cytokines [57]. In addition, the combination of Radix Astragulus and Radix Angelicae increased WBC, HB and PLT in cyclophosphamide-induced anemic rat [58]. This study also indicated that these two herbs are associated with enhancing erythropoietin expression [58]. Erythropoietin is also able to prevent against cisplatin cytotoxicity in cells via several mechanisms [59,60]. These findings may give some insights on the mechanism of how CHM improved the hematological parameters in this systematic review. Furthermore, Radix Astragalus also reduced toxicity induced by cyclophosphamide [61]. Though the molecular mechanism is not fully understood, the immunostimulating effects and the reduction of chemotherapy-induced toxicity may be the two major advantages for CHM as adjuvant therapy in NSCLC treatment.
Survival rate, immediate tumor response and chemo-toxicity are three major outcomes in the studies. However, not all the studies simultaneously reported the three outcomes. For example, Zhang et al. [21] reported all three outcomes while Li and Li [39] reported the reduction of chemo-toxicity only. Nevertheless, we analyzed all available data in these reports without any subjective selection. KPS is a scale for the evaluation of cancer patients. In this study, we analyzed continuous data (average of KPS scores) and discontinuous data (patient number with an increase of KPS$10). There is significant heterogeneity when we pooled studies with continuous data, although the baseline has no heterogeneity in the meta-analysis. The variation of KPS value in the studies may be due to the differences in treatment duration,  treatment methods and the herb prescriptions in different studies. Therefore, the change of KPS in patients is more accurate to indicate the efficacy of treatments in pooled studies. The interval of ten in KPS normally shows a significant change of performance status as described by previous publications [26,57,62,63]. Hence, we adopted a ten-point increase as the cutoff for improved performance status. The findings showed a significant improvement in KPS with CTC treatment.
In conclusion, the evidence from the meta-analysis of the included studies shows that CHM as adjuvant therapy has advantages in NSCLC patients. However, due to the complex nature of CHM interventions, particular attention should be paid to apply appropriate and rigorous research methodologies to investigate CHM as a holistic system [64]. Therefore a large scale RCT integrated the Chinese Medicine methodology of pattern diagnosis and treatment is warranted for further study.