Efficacy and Safety of Brucea javanica Oil Emulsion Injection in the Treatment of Gastric Cancer: A Systematic Review and Meta-Analysis

Background: Gastric cancer (GC) is one of the most common digestive tract cancers and ranks fifth in the incidence of malignant tumors worldwide. Brucea javanica oil emulsion injection (BJOEI), a Chinese patent medicine extracted from Brucea javanica (Yadanzi in Chinese Pinyin), is widely used as an adjuvant treatment for GC in China. This systematic review and meta-analysis aimed to evaluate the available data on the efficacy and safety of BJOEI in the treatment of GC and assess the quality of the synthesized evidence. Methods: A comprehensive search was performed on PubMed, EMBASE, CENTRAL, Web of Science, Chinese Biomedical Literature Database (CBM), China National Knowledge Infrastructure (CNKI), Wanfang database and Chinese Scientific Journals Database (VIP database), and other potential resources, such as the Chinese Clinical Trial Registry (ChiCTR) and ClinicalTrials.gov from their inception to July 31, 2021. Randomized controlled trials (RCTs) comparing the therapeutic effects of BJOEI combined with conventional therapy to those of conventional therapy alone were included. We used RevMan 5.3 for data analysis and quality evaluation of the included studies and assessed the evidence quality based on the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) criteria. Results: Eighteen RCTs involving 1,210 patients were included, and the meta-analysis results demonstrated that compared with the control group (conventional therapy), the experimental group (BJOEI combined with conventional therapy) showed a significantly improved overall response rate (ORR) (risk ratio [RR] = 1.52, 95% CI: 1.36–1.69, P < 0.00001), clinical benefit rate (CBR) (RR = 1.17, 95% CI: 1.11–1.23, P < 0.00001), performance status (RR = 1.72, 95% CI: 1.46–2.01, P < 0.00001), and reduced incidence of the following adverse drug reactions (ADRs): neutropenia, leukopenia, nausea and vomiting, diarrhea, liver damage, hand-foot syndrome, and peripheral sensory nerve toxicity. Subgroup analysis showed that the BJOEI intervention could significantly improve the ORR and CBR in patients with GC when combined with FOLFOX4, XELOX, and other chemotherapeutics. Conclusion: The evidence presented in this study supports the fact that BJOEI combined with conventional chemotherapy provides a statistically significant and clinically important effect in the improvement of ORR, CBR, performance status, and ADR reduction in patients with GC. To further support this conclusion, more rigorously designed, large-scale, and multicenter RCTs are needed in the future.


INTRODUCTION
Gastric cancer (GC) is one of the most common malignant tumors of the digestive tract and ranks fifth in incidence worldwide (1). There were ∼1.089 million new cases of GC worldwide, of which 43.9% were reported in China, in 2020 (2). Due to the lack of specific symptoms in early GC, the diagnosis is often made at an advanced disease stage, and the mortality rate is high (3). At present, radical resection is still the main GC treatment, but most patients experience recurrence within 3 years after surgery. The postoperative recurrence rate of patients with locally advanced GC is as high as 50-80%. Once patients experience recurrence and metastasis after the operation, even if palliative chemotherapy is administered again, the 5-year survival rate remains low (4)(5)(6)(7). Moreover, molecularly targeted therapy and immunotherapy of GC lag behind those of many other tumor types, and better survival benefits are still being explored (8).
Traditional Chinese Medicine (TCM) has a long historical tradition and currently attracts extensive attention because of its potential treatment benefits in the field of oncology. Our team has been committed to investigating the preventive and therapeutic values of TCM for many years (9)(10)(11). Brucea javanica oil emulsion injection (BJOEI) is a Chinese patent medicine extracted from Brucea javanica (Yadanzi in Chinese Pinyin). Its main active component is quassinoid sand fatty acids, which exert anticancer effects through multiple mechanisms (12). Studies have shown the synergistic effects of BJOEI combined with chemoradiotherapy on tumor attenuation, such as reversal of chemotherapy resistance, reduction of the recurrence and metastasis rates, and improvement of the quality of life (13)(14)(15)(16). Although several existing systematic reviews have been conducted to evaluate the clinical efficacy of BJOEI in GC, none of them assessed the quality of the synthesized evidence and arrived at definitive conclusions (13,(17)(18)(19). The most recent one was reported by Wu et al. in 2018, in which the retrieval deadline was January 2017 (17). With the growing number of studies on the value of BJOEI in GC treatment, more randomized controlled trials (RCTs) have been published in recent years (20)(21)(22)(23). Therefore, we conducted a systematic review to evaluate all available evidence of the efficacy and safety of BJOEI in the treatment of GC and assessed the quality of the synthesized evidence.

METHODS
This study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guidelines, and readers can access the protocol of this systematic review in the International Prospective Register of Systematic Reviews (CRD42021265646).

Inclusion Criteria
Studies that met the following criteria were included: (1) the study design was limited to RCTs, whether it was blinding or not; (2) the studies needed to meet the diagnostic criteria for GC by biopsy or postoperative pathological examination; and (3) studies provided the experimental group with BJOEI in combination with the same interventions provided to the control group.

Exclusion Criteria
Studies were excluded if any of the following reasons were involved: (1) duplicate studies; (2) inappropriate interventions; (3) incomplete data; and (4) irrelevance to outcome indicators.

Outcome Measures
Primary outcome measures included the overall response rate (ORR) and clinical benefit rate (CBR). The secondary outcome measure was the performance status. Safety outcome measures included the occurrence of adverse drug reactions (ADRs).

Literature Search Strategy
We searched the following relevant databases from inception to July 31, 2021: PubMed, EMBASE, CENTRAL, Web of Science, the Chinese Biomedical Literature Database (CBM), the China National Knowledge Infrastructure (CNKI), Wanfang database, and Chinese Scientific Journals Database (VIP database), and other potential resources, such as the Chinese Clinical Trial Registry (ChiCTR) and ClinicalTrials.gov for more study records. The combination of MeSH terms and text words was applied to study retrieval. "Stomach Neoplasms" was regarded as the MeSH term. All the strategies were adapted from different databases. The search strategies used in PubMed were as follows:

Study Selection
The search results were imported into Excel 2003. After removing duplicates, the titles and abstracts were screened for potential studies. Then, the full articles were checked to determine whether the studies met the inclusion criteria. The study selection process was independently performed by two investigators.

Data Extraction and Quality Assessment
All data were independently extracted by two investigators, and any discrepancies between the reviewers were resolved by the intercessor (JL) until consensus was reached. Data retrieved from the publications included author name, year of publication, number of patients, average age, gender, details about dosage and course of treatment, and outcome data. When necessary and feasible, the corresponding authors of the selected studies were contacted to obtain missing or incomplete data.
In terms of bias, the articles were evaluated as low risk, high risk, and unclear risk according to the following quality items: randomization generation, allocation concealment, subject blinding, outcome assessment, incomplete outcome data, and selective outcome reporting.

Statistical Analysis
Quantitative synthesis was conducted for outcomes reported in more than one homogeneous RCT. The systematic review was performed using the RevMan 5.3 software. Random-effects or fixed-effects models were chosen based on the analysis of heterogeneity. Randomized individuals were considered as unitof-analysis issues. If a meta-analysis was not appropriate because of clinical/methodological issues or statistical heterogeneity, a narrative summary of the findings or relevant subgroup analyses were used. The RR was used to evaluate dichotomous outcomes, while the mean difference (MD) was used to assess continuous variables. Each outcome numerical value was presented with 95% CIs. Funnel plots were used to test the risk of publication bias. The heterogeneity between RCTs was analyzed using the chi-square test and estimated using I 2 . Results of P ≥ 0.1 and I 2 ≤ 50% suggested a lack of significant heterogeneity, and a fixed-effects model was used accordingly; otherwise, the random-effects model was used. When conducting the meta-analysis, several subgroup analyses were performed to identify subpopulations that might be associated with differences in efficacy. The results of the sensitivity analysis were reported.

Quality of the Synthesized Evidence
Quality assessment of the synthesized evidence was performed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach (24). This assessment of evidence quality includes the risk of bias, heterogeneity, indirectness, imprecision, and publication bias. The quality of the evidence was classified as high, moderate, low, or very low.

Literature Search Results
A total of 458 clinical studies were identified based on the retrieval strategy. After screening based on the inclusion/exclusion criteria, 18 articles were selected for further analysis (Figure 1).

Quality Evaluation of the Literature
As shown in Figure 2, in terms of random sequence generation, six RCTs (20,22,23,28,29,38) were considered to have a low bias risk by applying a random number table or random envelope. Three RCTs (30,31,36) were marked as "high risk" because they divided patients according to hospitalization period, ID, and postoperative chemotherapy, respectively. The other nine RCTs (21, 25-27, 32-35, 37) did not describe the specific randomized method and were evaluated as "uncertain risk." None of the trials reported the methods of allocation concealment and blinding procedures, which indicated that there were unclear bias risks.

Secondary Outcomes Performance Status
As shown in Figure 5

Sensitivity Analysis
According to the Cochrane Handbook for Systematic Reviews of Interventions (39), I 2 values between 0 and 40% indicated that heterogeneity might not be important. Therefore, we eliminated the included studies with I 2 ≥ 40% one by one and then conducted a meta-analysis. The results showed that in the CBR of BJOEI + XELOX, after excluding Zhang et al. (38), the heterogeneity was decreased from 59 to 32% (P = 0.0002; Figure 7). After excluding Ma et al. (31), the heterogeneity was decreased from 59 to 0% (P = 0.04; Figure 8). The data suggested      that Zhang et al. (38) and Ma et al. (31) were the main reasons for the heterogeneity in the CBR of BJOEI + XELOX. In terms of ADRs, after excluding Ma et al. (31), the heterogeneity of nausea and vomiting decreased was from 47 to 36% (P = 0.14; Figure 9), and the heterogeneity of peripheral sensory nerve toxicity was decreased from 41 to 0% (P = 0.57; Figure 10). In addition, after deleting Zhang et al. (38), the heterogeneity of peripheral sensory nerve toxicity was decreased from 41% to 1% (P = 0.005; Figure 11). These findings suggest that Zhang et al. (38) and Ma et al. (31) might explain the heterogeneity in ORR and ADRs.

Analysis of Publication Bias
A funnel plot of publication bias for ORR is displayed in Figure 12, which indicates that there was no evidence of significant publication bias.

Quality of Evidence Assessment
Based on the GRADE criteria, the ORR, CBR, performance status, and ADRs were all assessed as low-quality evidence, owing to the existence of clinical heterogeneity and low participant numbers in most studies (Tables 2, 3).

DISCUSSION
Despite advances in disease screening and modern technology, GC remains one of the most common malignant tumors. Its metastasis, morbidity, and mortality rates are all on the rise, while the cure, radical resection, and 5-year postoperative survival rates of patients with advanced GC are low (40). In recent years, TCM has made great progress in anti-tumor therapy, and the manufacturing technologies of Chinese medicine compounds, Chinese patent medicine, Chinese medicine extract, and Chinese medicine monomers have developed more rapidly. BJOEI is a Chinese patent medicine that is widely used in the treatment of various cancers, such as lung (41) and several gastrointestinal cancers (42,43). Previous studies have shown that its antitumor effects might be related to the following mechanisms: 1) inhibition of DNA synthesis in tumor cells (44,45); 2) induction of tumor cell apoptosis and differentiation (46)(47)(48); 3) anti-angiogenesis (49); and 4) reversion of drug resistance (50).
In this study, we searched as many RCTs as we could and conducted a meta-analysis to evaluate the treatment efficacy and safety of BJOEI in patients with GC. All available data from the collected trials were applied without intentional selection. The results showed that BJOEI combined with chemotherapy was superior to single chemotherapy in improving ORR, CBR, and performance status. Considering that the different patient regimens might lead to high outcome heterogeneity, to obtain a more convincing conclusion, we conducted a subgroup analysis according to chemotherapeutic regimens. The results showed that for each BJOEI + FOLFOX4 and BJOEI + XELOX subgroup, the ORR and CBR were significantly improved by the addition of the BJOEI intervention. Furthermore, we have paid special attention to neutropenia, leukopenia, nausea and vomiting, diarrhea, liver damage, hand-foot syndrome, and peripheral sensory nerve toxicity, which are common symptoms of chemotherapy-associated ADRs. The meta-analysis showed that the BJOEI group had fewer symptoms related to the above ADRs. However, more RCTs are needed to further demonstrate the positive effect of BJOEI in ameliorating chemotherapyassociated toxicities.
Although we strictly conducted this meta-analysis according to the review procedure released by the Cochrane Collaboration, this study has several limitations. First, the duration of the intervention is an important factor in the evaluation of efficacy. The observation time of the included studies was mainly concentrated at 12 and 6 W, and the longest was 24 W (in only one RCT). Furthermore, the long-term effects of BJOEI in the treatment of GC remain unknown. Moreover, high-quality original studies were scarce in this study. The problems in most RCTs included unexplained randomization methods, insufficient attention to allocation concealment, low utilization rate of blinding, and unreported lost follow-up cases. Finally, recent advances have renewed the hope that immune and targeted agents can be leveraged to improve patient survival (51,52). Although chemotherapy is still the backbone of therapy against GC, studies should also investigate the efficacy of BJOEI combined with immunotherapy or targeted therapy.
Due to the limitations associated with the poor quality of pooled studies, it is difficult to draw a definitive conclusion. Nevertheless, our study suggests the positive effect of BJOEI in facilitating the management of ORR, CBR, performance status, and ADRs in patients with GC. More prospectively designed, large-sample, and multicenter RCTs are expected to offer persuasive evidence to demonstrate the efficacy and safety of BJOEI.

AUTHOR CONTRIBUTIONS
JL designed this study. XW and HW performed the online database search. LC, JW, TL, and SL contributed to the data collection, extraction, and analysis. XW, HW, and LC prepared the original draft and finished the revision of the manuscript. All authors have read and approved the final manuscript.