Bojungikgi-tang for Chemotherapy-induced Leukopenia: A Systematic Review and Meta-Analysis

Abstract

Chemotherapy-induced leukopenia is a common side effect of cytotoxic anticancer drugs. It can deprive patients of treatment opportunities, resulting in the delay, reduction, or discontinuation of chemotherapy or other anticancer drug administration. Two researchers searched English, Chinese, Japanese, and Korean electronic databases, without limiting the time period and language, using search terms such as “Bojungikgi,” “WBC,” “leuko,” and “neutrop.” Among the human randomized controlled studies in which Bojungikgi-tang was administered to patients who underwent chemotherapy, studies reporting leukopenia-related outcomes were selected, and data extraction, bias risk assessment, and meta-analysis were performed on the selected papers. Ten studies were selected, and a systematic review with meta-analysis was conducted. Nine papers were published in China and the total number of participants was 715. As a result of administering Bojungikgi-tang to these patients, the number of patients with chemotherapy-induced leukopenia significantly decreased (OR: 0.41, 95% CI: 0.27-0.61, P = .0001, I² = 35%). Further, white blood cell counts were compared with that of the control group, and it showed an effect on prevention (MD: 0.64, 95% CI: 0.46-0.83, P < .00001, I² = 90%). A pronounced effect was observed, especially when administered after a diagnosis based on the pattern identification, such as Qi deficiency. (OR: 0.32, 95% CI: 0.18-0.58, P = .0002, I² = 0%). However, all studies had a high risk of bias due to non-blinding, and most studies had a high or uncertain risk of bias in creating random assignment orders and concealing them. Bojungikgi-tang has an effect on the prevention and treatment of chemotherapy-induced leukopenia. The effect rate can be increased when administered after proper diagnosis, and the possibility of adverse reactions and side effects is lower than that of Granulocyte-Colony Stimulating Factor (G-CSF) injection. Bojungikgi-tang appears to be useful in the treatment and prevention of leukopenia caused by cytotoxic anticancer drugs. However, it is necessary to conduct high-quality clinical studies in the future, considering the possibility of local and language bias, heterogeneity of carcinoma and intervention, and the risk of bias.

Registration: PROSPERO CRD4202341054.

Introduction

Cytotoxic anticancer drugs can cause toxicity in rapidly dividing normal cells, such as bone marrow. The effects of this cytotoxicity appear as symptoms such as leukopenia or neutropenia, anemia, thrombocytopenia. ¹

Among these, leukopenia or neutropenia due to myelosuppression is the most common side effect after chemotherapy, along with thrombocytopenia, and is the most common cause of patients undergoing chemotherapy visiting the emergency room. ² The frequency of infection is related to the degree and duration of leukopenia or neutropenia, which becomes a factor in increasing the mortality rate of patients. Even after recovery, it causes deprivation of treatment opportunities such that patients may not receive chemotherapy at the scheduled time, or reduce or discontinue anticancer drug intake. ³ Granulocyte-colony stimulating factor (G-CSF) injections are used to prevent and treat leukopenia and neutropenia after chemotherapy. However, some patients who receive the injection experience fatal splenic rupture, acute respiratory distress syndrome, anaphylaxis, or more commonly, bone pain. ⁴

Recently, systematic reviews and meta-analyses have suggested that various herbal treatments can be helpful for treating leukopenia or neutropenia induced by chemotherapy^5,6; however, the search and data analysis were conducted without limiting the composition of herbal medicines. Therefore, the effectiveness of specific prescriptions was not mentioned. Most of these prescriptions are mainly used in China; therefore, to apply herbal medicine treatment based on evidence, it must be administered in the form of a decoction in other countries. This form reduces medication adherence and increases the patient’s financial burden. ⁷

Bojungikgi-tang (Buzhongyiqi-tang, in Chinese) was first recorded in Dongwon Lee’s 『Dongwon Ten Medical Books』 in 1232 AD. It is composed of Astragali radix, Ginseng radix, Atractylodis rhizoma alba, Glycyrrhizae radix et rhizoma, Gigantis radix, Citri unshius pericarpium, Cimicifugae rhizoma, and Bupleuri radix. This is a representative prescription that tonifies the middle energizer and strengthens the Qi and is widely used in the form of granules in East Asian countries. ⁸

Therefore, in this study, we analyzed human randomized controlled studies in which Bojungikgi-tang was administered to patients receiving chemotherapy and assessed its efficacy on chemotherapy-induced leukopenia.

Methods

The data search, reporting, and discussion were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) 2020 statement. ⁹

Criteria for Selection and Exclusion

Type of study

We included studies that conducted human randomized controlled trials (RCTs). All the other studies were excluded.

Participants

The included studies involved patients with cancer using anticancer drugs without restrictions on sex, race, nationality, age, or type of cancer. Patients with leukopenia not caused by anticancer drugs or with unclear cases were excluded.

Treatment group

Bojungikgi-tang and modified Bojungikgi-tang were used as interventions in the treatment group without restrictions on formulations. Studies in which the composition of Bojungikgi-tang could not be confirmed or in which additional interventions were performed together with Bojungikgi-tang were excluded. However, cases in which additional interventions were identical to those of the control group were included.

Control group

Studies in which no treatment was performed in the control group, placebo herbal medicine or additional interventions that were identical to those in the treatment group, or general treatment for leukopenia was administered were all included.

Outcome indicators

Primary outcome indicators included white blood cell (WBC) count in the peripheral blood before and after chemotherapy, the effective rate of WBC count recovery, and the number of patients with leukopenia.

Search Sources and Search Strategies

Representative academic information from the electronic databases of each language region was used as the search source. Academic Search Ultimate (ASU), e-journals, and MEDLINE via EBSCO; The Cochrane Library; China National Knowledge Infrastructure (CNKI); Wanfang Data; Japanese databases Citation Information by NII (CiNii); Japan Science Technology Information Aggregator, Electronic (J-STAGE); KoreaMed; Korean Medical Database (KMBASE); Research Information Sharing Service (RISS), and Science ON were used.

“Bojungikgi,” “WBC,” “leuko,” “leuco,” and “neutrop” were used as search terms. Although chemotherapy was considered a search term, those related to Bojungikgi-tang as an intervention and leukocytes or neutrophils as outcome measures were adopted to include all cases in which only the specific name of the anticancer drug was described in the thesis title or abstract. In addition, to include both leukocyte and neutrophil count-related and leukopenia- or neutropenia-related outcome measures, search terms excluding suffixes (-cyte, -hil, and-penia) were used.

Considering that the English name Bojungikgi-tang has been used in various Romanizations in Korean publications ¹⁰ and that Chinese and Japanese languages, which have the same Chinese characters, have different pronunciations, various search terms have been added. In addition, Chinese, Japanese, and Korean search terms were added to the databases for each language region and searched. If a search was possible without errors, even if the database was not in the corresponding language region, words from all languages were added and searched. The search date was March 14, 2023, and two researchers (LA and SP) searched with no publication year and language restrictions, and then merged the bibliographic information retrieved by each researcher. Supplemental File 1 shows the search strategies of all databases.

Study Selection and Data Extraction

After removing duplicate studies, the original texts of the selected studies were downloaded after excluding studies that did not meet the criteria. After going through the above process, the main text of the studies was reviewed to confirm additional information, and the papers to be analyzed were finally selected.

Two researchers independently selected or excluded studies according to the selection and exclusion criteria, and a consensus was reached through discussion if there was a disagreement between the researchers.

For the final selected papers, the 2 researchers independently extracted the last name of the first author; year of publication; country and language of publication; number, sex, and age of participants; type of cancer; type of anticancer drug; intervention of treatment and control groups; course of treatment; major outcome measures; pattern identification; and formulation and composition of Bojungikgi-tang.

Risk of Bias Assessment

To assess the risk of bias of the final selected studies, Cochrane’s Risk of Bias (RoB) 2.0 ¹¹ was applied using Review Manager 5.4 (Cochrane, London, UK). If there were more than 10 studies included in the meta-analysis, publication bias was assessed by using a funnel plot. Two researchers performed a risk of bias assessment using the above program, and a consensus was reached through discussion in case of disagreement between researchers.

Data Analysis

When 2 or more studies used the same outcome measure, a meta-analysis was conducted using Review Manager version 5.4.

For dichotomous variables, odds ratios (OR) and two-sided 95% confidence intervals (CIs) were presented, and for continuous variables, the mean difference (MD) and 95% CIs were presented.

Statistical heterogeneity was evaluated using the I² value; if it exceeded 50%, a distinct heterogeneity was considered and a random-effects model was used; otherwise, a fixed-effects model was used. However, while the random-effects model is often the appropriate model, there are cases where it cannot be implemented properly, because there are too few studies to obtain an accurate estimate of the between-studies variance. If the number of studies included in the meta-analysis was less than 10, the estimate of the difference between studies was considered inaccurate; therefore, a fixed-effects model was used regardless of the I² value. ¹²

Certainty of the Evidence

The Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) ¹³ approach was used to assess the certainty of the estimates and to produce evidence profiles, with the certainty of evidence graded as high, moderate, low, or very low (GRADEpro GDT software; McMaster University and Evidence Prime Inc, Hamilton, Ontario, Canada).

Results

Search Results and Study Selection

A total of 699 papers were searched in 13 academic information electronic databases; Among 699 studies, 481 studies, excluding duplicates, were reviewed. The flow diagram of the study based on the PRISMA 2020 has been shown in Figure 1.

Figure 1.

Flowchart of the study selection.

Selected Studies Analysis

General characteristics of included studies

A summary of selected studies has been presented in Table 1. Nine studies^{14 -22} were published in Chinese involving research in China, and one study ²³ in English involved research in Japan. Two studies^20,21 in Chinese were Master’s theses.

Table 1.

General Characteristics of the Included Studies.

Study ID	Sample size (Male/Female)		Intervention		Pattern identification	Main outcomes
	Age (y)
	Experimental	Control	Experimental	Control
Chen et al 14	30 (24/6)	30 (25/5)	BJT decoction modified by symptoms a 200 ml bid	Xuebao 5cap tid	NA	WBC RecoveryRate, WBC counts
	44.76 ± 14.98	42.36 ± 15.11
Qin 15	27 (19/8)	26 (20/6)	Modified BJT decoction bid	None	Spleen-Stomach Qi Deficiency	RECIST, TCM Symptom Score, KPS, WHO Toxicity Grades b
	59.04 ± 8.716	57.73 ± 10.724
Wang and Sang 16	40 (23/17)	40 (27/13)	BJT decoction modified by symptoms a bid	Xuebao 5cap tid	NA	WBC RecoveryRate, WBC counts
	55.7 ± 3.1	57.4 ± 1.3
Li et al 17	40 (25/15)	40 (26/14)	Modified BJT decoction modified by symptoms a bid	None	Qi Deficiency	CD4+, CD8+, CD4+/CD8+, TCM Symptom Score, Incidence of Toxic and Side effects, b EORTC QLQ-C30
	55.4 ± 9.6	55.2 ± 9.8
Lang 18	25 (17/8)	25 (16/9)	Modified BJT decoction modified by symptoms a bid	None	NA	CD4+, CD8+, CD4+/CD8+, Incidence of Toxic and Side Effects b
	51.5 ± 20.5	53.1 ± 20.4
Okabe et al 23	55 (37/18)	55 (37/18)	TJ-41(Hochu-ekki-to) granules 2.5 g tid	None	NA	Completion Rate of Chemotherapy for 1 y, Incidence of Adverse Events by CTCAE, b Relative Dose Intensity, RFS, OS
	49-76	35-77
Zhang and Yin 19	40 (22/18)	40 (21/19)	Buzhong Yiqi Granules 3 g tid	None	NA	WBC, RBC, Hb, PLT, TCM Symptom Score, Incidence of Adverse Events
	62.37 ± 7.45	63.97 ± 6.49
Xu 20	31 (18/13)	31 (16/15)	Modified BJT decoction modified by symptoms a 250-300 ml bid	None	Lung-Spleen Qi deficiency	KPS, TCM Symptom Score, RECIST, PFS, WHO Toxicity Grades b
	62.94 ± 5.81	61.03 ± 4.98
Liu 21	40 (23/17)	40 (24/16)	BJT decoction 100 ml tid + DSP 4tab tid	DSP 4tab tid	NA	WBC, Neu, PLT, Hb, WHO Toxicity Grades, b Fullness Symptom Grade
	56.15 ± 9.77	56.15 ± 8.89
Ying et al 22	30 (19/11)	30 (16/14)	BJT decoction 300 ml bid	None	Qi-Yin deficiency	TCM Symptom Score, IgA, IgM, IgG, C3, C4, CD3+, CD4+/CD8+, CD4+, CD8+, WBC, PLT, Incidence of Adverse Events b
	64.61 ± 11.53	66.2 ± 12.1

Abbreviations: BJT, Bojungikgi-tang; bid, twice a day; cap, capsule; tid, 3 times a day; WBC, white blood cell; RECIST, response evaluation criteria tn solid tumors; TCM, Traditional Chinese Medicine; KPS, Karnofsky performance scale; WHO, World Health Organization; EORTC QLQ-C30, The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; CTCAE, Common Terminology Criteria for Adverse Events; RFS, relapse-free survival; OS, overall survival; RBC, red blood cell; Hb, hemoglobin; PLT, platelet; PFS, progression-free survival; DSP, Diyushengbai tablet; tab, Tablet; Neu, Neutrophil.

^aAdded herbs depending on symptoms such as nausea, vomiting and loss of appetite.

^bContain results for the number of patients with leukopenia.

The total number of participants was 715, and each study included 50 to 110 participants; the average number of participants was 71.5. Among the participants, there were 455 males (64%), and there were more males than females in each group in all studies. In most studies, the average age of the participants was 50s or 60s (years).

Cancer types included lung cancer (276 patients, 43%), stomach cancer (194 patients, 31%), colorectal cancer (103 patients, 16%), breast cancer (30 patients, 5%), esophageal cancer (26 patients, 4%), and malignant lymphoma (6 patients, 1%). In Wang and Sang’s ¹⁶ study of 80 participants, the type(s) of cancer was not recorded. Anticancer drugs used were cisplatin, cyclophosphamide, docetaxel, epirubicin hydrochloride, fluorouracil, gemcitabine, oxaliplatin, paclitaxel, prednisone, S-1 (Tegafur, 5-chloro-2,4-dihydroxypyridine and potassium oxonate), and vincristine. All anticancer drugs except prednisone were cytotoxic. Combination chemotherapy using 2 or more anticancer agents was performed in all studies except for the one performed by Wang and Sang, ¹⁶ which did not record the anticancer agents administered to the participants.

Intervention analysis of the treatment and control groups

Seven^{15,17 -20,22,23} out of a total of 10 studies were designed to compare the Bojungikki-tang administration group with no treatment group, and 2 studies^14,16 compared the Bojungikki-tang administration group with the Xuebao capsule (XBJ) administration group. In the one study, ²¹ Bojungikgi-tang and Diyushengbai tablets (DSP) were administered to the treatment group, and only DSP was administered to the control group before comparison.

Most studies were conducted within 2 to 20 weeks, but Okabe et al’s ²³ study was conducted for approximately a year. Bojungikgi-tang in the form of a decoction was administered to the treatment group in 8 studies,^{14 -18,20 -22} of which seven^{14 -18,20,22} administered the decoction twice per day, and only one study, ²¹ in combination with DSP, administered the decoction thrice per day. In these 2 studies,^19,23 Bojungikgi-tang granules (TJ-41) produced by Tsumura & Co. in Japan and Bojungikki-tang granules produced by Beijing Handian Pharmaceutical Co., Ltd. were administered 3 times a day.

Four studies^15,17,20,22 included pattern identification as the criterion for selecting the study participants, and each patient was diagnosed with Qi deficiency, Spleen-Stomach Qi deficiency, Lung-Spleen Qi deficiency, or Qi-Yin deficiency. In all 4 studies, Bojungikgi-tang was administered to patients with Qi deficiency.

The composition of the decoctions and granules has been presented in Table 2. Due to the nature of decoctions to which herbs could be added or subtracted according to individual diagnosis and symptoms, the composition of Bojungikgi-tang used in each study was not the same. The herbs commonly included in all studies were Astragali radix, Ginseng radix or Codonopsis pilosulae radix, Atractylodis rhizoma alba, and Glycyrrhizae radix et rhizoma. Angelicae Gigantis radix and Citri unshius pericarpium were used in 9 studies, and Cimicifugae Rhizoma and Bupleuri Radix were used in 8. In 5 of the studies^{14,16 -18,20} using decoctions, additional herbs were added for additional symptoms such as nausea, vomiting, and anorexia.

Table 2.

The Composition of Bojungikgi-tang of the Included Studies.

Study ID	Formulation	Composition
Chen et al 14	Decoction	Astragali Radix 30 g, Codonopsis Pilosulae Radix 18 g, Angelicae Gigantis Radix 10 g, Atractylodis Rhizoma Alba 12 g, Cimicifugae Rhizoma 6 g, Bupleuri Radix 6 g, Citri Unshius Pericarpium 6 g, Glycyrrhizae Radix et Rhizoma 6 g
Qin 15	Decoction	Astragali Radix 30 g, Codonopsis Pilosulae Radix 20 g, Atractylodis Rhizoma Alba 10 g, Roasted Glycyrrhizae Radix et Rhizoma 15 g, Angelicae Gigantis Radix 10 g, Cnidii Rhizoma 10 g, Citri Unshius Pericarpium 6 g, Curcumae Rhizoma 6 g, Hedyotidis Herba 30 g, Coicis Semen 20 g
Wang and Sang 16	Decoction	Astragali Radix 30 g, Codonopsis Pilosulae Radix 18 g, Angelicae Gigantis Radix 10 g, Atractylodis Rhizoma Alba 12 g, Cimicifugae Rhizoma 6 g, Bupleuri Radix 6 g, Citri Unshius Pericarpium 6 g, Glycyrrhizae Radix et Rhizoma 6 g
Li et al 17	Decoction	Astragali Radix 60 g, Setaria Italica Germinatus 25 g, Hordei Fructus Germinatus 25 g, Atractylodis Rhizoma Alba 15 g, Ginseng Radix 15 g, Citri Unshius Pericarpium 10 g, Bupleuri Radix 10 g, Ligustri Fructus 10 g, Roasted Glycyrrhizae Radix et Rhizoma 10 g, Cimicifugae Rhizoma 10 g, Ecliptae Herba 10 g, Angelicae Gigantis Radix 10 g, Amomi Fructus 6 g
Lang 18	Decoction	Astragali Radix 60 g, Setaria Italica Germinatus 25 g, Hordei Fructus Germinatus 25 g, Atractylodis Rhizoma Alba 15 g, Ginseng Radix 15 g, Citri Unshius Pericarpium 10 g, Bupleuri Radix 10 g, Ligustri Fructus 10 g, Roasted Glycyrrhizae Radix et Rhizoma 10 g, Cimicifugae Rhizoma 10 g, Ecliptae Herba 10 g, Angelicae Gigantis Radix 10 g, Amomi Fructus 6 g
Okabe et al 23	Granules	Astragali Radix 4 g, Atractylodis Rhizoma Alba 4 g, Ginseng Radix 4 g, Angelicae Gigantis Radix 3 g, Bupleuri Radix 2 g, Zizyphi Fructus 2 g, Citri Unshius Pericarpium 2 g, Glycyrrhizae Radix et Rhizoma 1.5 g, Cimicifugae Rhizoma 1 g, Zingiberis Rhizoma Recens 0.5 g (per 5 g of a dried extract)
Zhang and Yin 19	Granules	Roasted Astragali Radix, Codonopsis Pilosulae Radix, Roasted Glycyrrhizae Radix et Rhizoma, Angelicae Gigantis Radix, Parched Atractylodis Rhizoma Alba, Cimicifugae Rhizoma, Bupleuri Radix, Citri Unshius Pericarpium, Zingiberis Rhizoma Recens, Zizyphi Fructus
Xu 20	Decoction	Astragali Radix 15 g, Codonopsis Pilosulae Radix 10 g, Poria Sclerotium 10 g, Atractylodis Rhizoma Alba 10 g, Adenophorae Radix 10 g, Polygonati Rhizoma 10 g, Arisaematis Rhizoma 10 g, Ostreae Testa 25 g, Prunellae Spica 15 g, Roasted Glycyrrhizae Radix et Rhizoma 6 g
Liu 21	Decoction	Astragali Radix 25 g, Parched Atractylodis Rhizoma Alba 15 g, Cimicifugae Rhizoma 15 g, Roasted Glycyrrhizae Radix et Rhizoma 8 g, Angelicae Gigantis Radix 15 g, Rehmanniae Radix Preparata 15 g, Paeoniae Radix 15 g, Cnidii Rhizoma 15 g, Poria Sclerotium 15 g, Citri Unshius Pericarpium 15 g, Ginseng Radix 15 g, Bupleuri Radix 15 g
Ying et al 22	Decoction	Astragali Radix 20 g, Ginseng Radix 10 g, Atractylodis Rhizoma Alba 10 g, Roasted Glycyrrhizae Radix et Rhizoma 10 g, Angelicae Gigantis Radix 5 g, Citri Unshius Pericarpium 5 g, Cimicifugae Rhizoma 5 g, Bupleuri Radix 5 g, Zingiberis Rhizoma Recens 10 g, Zizyphi Fructus 6 pieces

Analysis of main outcome measures

A total of 5 studies^{14,16,19,21,22} recorded WBC count before and after treatment, and two^14,16 of them also reported WBC Recovery Rate. One ²¹ presented the number of patients by toxicity grade presented by the World Health Organization (WHO), ²⁴ and one study ²² presented the number of patients with other side effects in addition to leukopenia. Moreover, 2 studies^15,20 presented only the number of patients by WHO toxicity grade, 2 studies^17,18 presented only the number of patients with leukopenia along with the number of patients with other side effects, and one study ²³ presented the number of patients by grade of adverse events according to the common terminology criteria for adverse events (CTCAE). ²⁵

Additionally, most studies presented various outcome measures. Five studies^{15,17,19,20,22} assessed Traditional Chinese Medicine (TCM) symptom scores before and after intervention, and 3 studies measured T-cell fraction^17,18,22 and platelet count.^19,21,22

Analysis of adverse events and dropout reports

Among the 10 selected studies, three^20,21,23 mentioned adverse reactions due to Bojungikgi-tang intervention. Of these, Xu’s ²⁰ study did not mention any adverse reactions or dropouts in the results, Liu’s ²¹ study reported that there were no adverse reactions or dropouts according to the set criteria, and Okabe et al’s ²³ study reported jaundice and skin rashes in some patients, which may have occurred due to TJ-41; however, there were no serious side effects, and it was reported that there were no dropouts due to this.

Risk of Bias Assessment

The RoB assessment was performed for all 10 selected studies, and the results are shown in Figure 2.

Figure 2.

Summary of risk of bias for 10 randomized controlled trials.

Random sequence generation

Five studies^{14,18,19,21,23} mentioned a method for generating random sequences. Three^14,21,23 of these had a low risk of bias when using a random number table or a central computer system; however, 2 studies^18,19 had a rather high risk of bias when using the allocation method according to the order of admission. In the remaining 5 studies,^{15 -17,20,22} the risk of bias was uncertain as no mention of the randomization method.

Allocation concealment

Five studies^{14,18,19,21,23} mentioned a method for allocation concealment. Three^14,21,23 of these had a low risk of bias when using a random number table or a central computer system with concealed allocation; however, 2 studies^18,19 had a rather high risk of bias, because group assignment was done according to order of admission, even though the studies used sealed envelopes or other appropriate concealment methods. In the remaining 5 studies,^{14,18,19,21,23} the risk of bias was uncertain as no mention of concealment of the allocation was made.

Blinding of participants, personnel, and outcome assessment

In all 10 studies, the risk of bias was high because the participants and researchers were not blinded, and for outcome assessment, it was uncertain because there was no mention of blinding the outcome assessors.

Incomplete outcome data and selective reporting

Regarding dealing with incomplete results, only 3 studies^20,21,23 presented the criteria for participant elimination, and among them, only two^21,23 indicated whether there was a drop out; however, as a result of reviewing the results of all 10 studies, it was found that there were no participant dropouts and no missing values due to this. Therefore, the risk of bias was low. Regarding selective reporting of results, detailed protocols for reporting the results were mentioned in 2 studies,^21,23 the risk of bias was judged to be low. Other studies were considered uncertain due to lack of detailed protocols mentioned.

Other potential risks of bias

Other items at a risk of bias include recruitment bias, imbalance in baseline conditions, and inaccurate statistical analysis. ²⁶ All studies used appropriate statistical analysis tools and methods, and the risk of bias was judged to be low.

Data Analysis

Among the 10 selected studies, five^{14,16,19,21,22} reported WBC count in peripheral blood before and after the use of anticancer drugs, and 7 studies^{15,17,18,20 -23} reported number of patients with leukopenia. Liu’s ²¹ study and Ying et al’s ²² study reported both outcomes. Studies reporting the same results were grouped and meta-analyzed, and the studies by Liu and Ying were included in both groups.

Considering that chemotherapy-induced leukopenia usually occurs 7 to 14 days after chemotherapy, ¹ the results were also analyzed after 2 weeks of anticancer drug administration. Studies with an effective rate of recovery of the WBC count as an outcome measure were included in the selection criteria for the search, but both studies^14,16 reported WBC count; therefore, a meta-analysis was conducted only for this.

In addition, to observe the effect of Bojungikgi-tang administration after pattern identification, a meta-analysis was conducted on the studies that applied pattern identification as the participants’ selection criteria versus studies that did not.

Analysis of the number of patients with leukopenia after chemotherapy

A total of 7 studies^{15,17,18,20 -23} were included, and the Bojungikgi-tang treatment group showed a significant reduction in the number of patients with leukopenia after using anticancer drugs compared to the control group (n = 248, OR: 0.41, 95% CI: 0.27-0.61, P < .0001), and the heterogeneity between studies was low (heterogeneity: Chi² = 9.28, P = .16, I² = 35%) (Figure 3).

Figure 3.

Forest plots; experimental versus control; patients with leukopenia after chemotherapy.

White blood cell count analysis 2 weeks after chemotherapy

A total of 5 studies^{14,16,19,21,22} were included, and WBC count after chemotherapy was significantly higher in the Bojungikgi-tang group than in the control group (n = 180, MD: 0.64, 95% CI: 0.46-0.83, P < .00001), but there was high heterogeneity between the studies (heterogeneity: Chi² = 40.56, P < .00001, I² = 90%) (Figure 4).

Figure 4.

Forest plots; experimental versus control; white blood cell count 2 weeks after chemotherapy.

Analysis of the number of patients with leukopenia according to the application of Bojungikgi-tang through pattern identification

A meta-analysis was conducted on 4 studies^15,17,20,22 administering Bojungikgi-tang with pattern identification such as Qi deficiency, Spleen-Stomach Qi deficiency, Lung-Spleen Qi deficiency, and Qi-Yin deficiency (n = 128, OR: 0.32, 95% CI: 0.18-0.58, P = .0002); Also, other 3 studies^18,21,23 without pattern identification (n = 120, OR: 0.51, 95% CI: 0.29-0.87, P = .01).

However, there was no heterogeneity between studies applying pattern identification (heterogeneity: Chi² = 0.52, P = .91, I² = 0%), but there was a high tendency of heterogeneity among studies without its application (heterogeneity: Chi² = 7.52, P = .02, I² = 73%) (Figure 5).

Figure 5.

Forest plots. (A) Pattern identification for participants. (B) No pattern identification for participants.

Certainty of the Evidence

Table 3 shows a summary of the quality of evidence regarding each outcome and diagnosis subgroup. Certainty in the evidence was moderate to high. The main reason for the evidence level downgrade was the unclear risk of bias and the high heterogeneity of results.

Table 3.

The Qualities of Evidence Regarding Each Outcome and Diagnosis Subgroup.

	Anticipated absolute effects a (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty
	Risk with control	Risk with BJT
Outcomes
Incidence of leukopenia	445 per 1000	248 per 1000	OR 0.41 (0.27-0.61)	495 (7 RCTs)	⨁⨁⨁⨁
					High
WBC count	-	MD 0.64 higher	-	360 (5 RCTs)	⨁⨁⨁◯
					Moderate
Diagnosis
Pattern identification	417 per 1000	186 per 1000	OR 0.32 (0.18-0.58)	255 (4 RCTs)	⨁⨁⨁⨁
					High
Non-pattern identification	475 per 1000	316 per 1000	OR 0.51 (0.29-0.87)	240 (3 RCTs)	⨁⨁⨁◯
					Moderate

Abbreviations: CI, confidence interval; BJT, Bojungikgi-tang; MD, mean difference; OR, odds ratio.

^aThe risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

Discussion

Recently, the rapid development of targeted anticancer and immuno-anticancer drugs has led to dramatic improvements in cancer treatment, but the role of cytotoxic anticancer drugs is still significant in specific cancer types and advanced-stage anticancer therapies. In addition, research has reported that the anticancer effect of immuno-anticancer drugs can be enhanced by maximizing the immune response due to changes in the immunological environment around the tumor when combined with cytotoxic anticancer drugs, rather than using immuno-anticancer drugs alone. Therefore, cytotoxic anticancer drugs’ role is changing once more.^27,28

Leukopenia or neutropenia, which often occurs as a side effect of these cytotoxic anticancer drugs’ administration, sometimes causes severe infections such as sepsis, which increases the patient’s mortality rate, delays the anticancer treatment schedule, and leads to reduction of dose or discontinuation of anticancer drugs.^1,3 As an adjuvant therapy for this, in China, herbal medicine in the form of capsules or tablets are actively used according to pattern identification. ²⁹ In Japan, based on RCTs, Sipjeondaebo-tang(Juzentaihoto in Japanese) granules are used. ³⁰

Bojungikgi-tang is a representative prescription that tonifies the middle energizer and strengthens the Qi. Recently, the effects on chronic fatigue syndrome, ³¹ atopic dermatitis, ³² and other diseases were systematically reviewed; and clinical studies on cancer-related fatigue were conducted, and significant results were obtained. ³³

Additionally, several in vivo studies have been conducted to demonstrate the efficacy of Bojungikgi-tang administration in treating decreased immunity associated with Qi deficiency. ³⁴

In this study, a systematic literature review and meta-analysis was conducted on RCTs in which Bojungikgi-tang was administered to patients undergoing chemotherapy to assess its effectiveness against leukopenia, a typical side effect of chemotherapy with cytotoxic agents. Hereby, we tried to lay the foundation for actively using Bojungikgi-tang for the treatment and prevention of chemotherapy-induced leukopenia, and the following results were obtained.

First, there is evidence that Bojungikgi-tang administration is effective in the treatment and prevention of chemotherapy-induced leukopenia. In addition, this treatment may be more effective when administered based on pattern identification, such as Qi deficiency.

Second, the degree of adverse reactions and side effects that may occur when administering Bojungikgi-tang to patients receiving chemotherapy appears to be low. However, only 3 studies reported the presence of adverse reactions and side effects.

This study has the following limitations. First, 9 of the final 10 studies were conducted in China, and the thesis was also published in Chinese; therefore, there is a possibility of location bias. Second, the total number of participants included in the analysis and the number of studies were small, the types of cancers and cytotoxic anticancer drugs used in each study were different, and the composition of Bojungikgi-tang used in the intervention was also different. Therefore, the heterogeneity of the content of each study was high.

Third, regarding the risk of bias assessment, none of the studies blinded the participants or researchers, and it was unclear whether the outcome assessors were blinded. However, because the outcome measures were reported based on the WBC count through a peripheral blood test rather than improvement in subjective symptoms, the real risk of bias due to this is considered lower than the RoB evaluation result.

Nevertheless, considering that most studies had a high risk of bias or were unable to measure risk in the process of generating random sequences and concealing allocation, efforts are needed to ensure that more high-quality clinical studies are conducted in the future.

Fourth, one study ¹⁹ conducted in Japan showed high heterogeneity. It is estimated that the heterogeneous results were due to the longer anticancer drug administration period, which was set at 1 year, compared with other studies.

Conclusions

At present, Bojungikgi-tang can be used to treat and prevent leukopenia caused by cytotoxic anticancer drugs. In East Asian countries such as Korea, China, and Japan, easy-to-take formulations such as granules are available, which may lead to high patient compliance with medication.

However, it is necessary to conduct high-quality clinical studies in various countries that supplement these subjects in the future, considering the possibility of location bias, heterogeneity of cancer types and interventions, and other risks of bias.