Efficiency of Protective Interventions on Irinotecan-Induced Diarrhea: A Systematic Review and Meta-Analysis

Abstract

Background:

Irinotecan is widely used in the treatment of various solid tumors, but the adverse effects from it, especially diarrhea, limit its use. Several clinical trials of prophylactic treatment of irinotecan-induced diarrhea (IID) have been ongoing, and some of the data are controversial. This encouraged us to conduct a meta-analysis of the effects of interventions on preventing IID.

Method:

This systematic review was conducted based on the PRISMA statement. We performed literature searches from PubMed, Web of Science, Embase, and Cochrane Library. The number registered in PROSPERO is CRD42022368633. After searching 1034 articles in the database and references, 8 studies were included in this meta-analysis.

Result:

The RR of high-grade diarrhea and all-grade diarrhea were 0.31 (I² = 51%, 95% CI: 0.14-0.69; P = .004) and .76 (I² = 65%, 95% CI: 0.62-0.93; P < .008) respectively, thus the use of intervention measures for preventing IID is effective, and the risk reduction of high-grade diarrhea was more significant. Subgroup analysis revealed that the monotherapy group (RR: 0.48, 95% CI: 0.21-1.13, I² = 0%) and combination therapy group (RR: 0.14, 95% CI: 0.06-0.32, I² = 0%) in the risk of high-grade diarrhea had no significant heterogeneity within the groups, and traditional herbal medicines (Kampo medicine Hangeshashin-to, PHY906 and hot ironing with Moxa Salt Packet on Tianshu and Shangjuxu) were effective preventive measures (RR:0.20, 95% CI: 0.07-0.60, I² = 0%). The Jadad scores for traditional herbal medicines studies were 3, and the follow-up duration was only 2 to 6 weeks.

Conclusion:

This systematic review and meta-analysis suggest that preventive treatments significantly reduced the risk of high-grade and all-grade diarrhea, confirming the efficacy in the incidence and severity of IID, among which traditional herbal medicines (baicalin-containing) provided a protective effect in reducing the severity of IID. However, the traditional herbal medicines studies were of low quality. Combined irinotecan therapy can obtain better preventive effects than monotherapy of IID. These would be helpful for the prevention of IID in clinical practice.

Introduction

Irinotecan, a semisynthetic derivative of camptothecin, is an anticancer drug that inhibits nucleic acid synthesis by targeting topoisomerase I.^1,2 It has been widely used in treating various solid tumors in adults and children,^3,4 including colorectal cancer most commonly,^5,6 as well as lung cancer, ⁷ breast cancer, ⁸ sarcoma,^3,4 esophageal cancer, ⁹ pancreatic cancer, ¹⁰ cholangiocarcinoma, ¹¹ gastric cancer, ¹² urothelial cell carcinoma, ¹³ etc. Experimental and clinical studies reveal that irinotecan not only can be used in monotherapy, ¹⁴ but also can be used in combination with cytotoxic agents (eg, 5-fluorouracil and oxaliplatin), monoclonal antibodies (eg, cetuximab and bevacizumab), kinase inhibitors (eg, fruquintinib, apatinib, dasatinib, regorafenib, and sunitinib), and cell-cycle checkpoint inhibitors. ¹⁵ Irinotecan is related to dose-limiting adverse events, primarily diarrhea, neutropenia, and asthenia. Due to its complex activation and deactivation mechanisms, irinotecan causes severe late diarrhea in 9% to 31% of patients undergoing irinotecan treatment. Such toxicity leads to premature termination of the drug or reduced dose intensity, leading to poor quality of life of cancer patients, expensive health care costs and low efficacy of the drug in approximately 40% of patients. Sometimes, it even leads to death.^16,17

For irinotecan-induced diarrhea (IID), antidiarrheal and symptomatic treatments are widely used in clinics. While experts recommend using loperamide and octreotide, only limited effectiveness has been observed. Loperamide has a good antidiarrheal effect, but excessive exposure may lead to arrhythmia and intestinal obstruction.^18,19 Likewise, octreotide is effective but shows the risk of aggravating neutropenia. ²⁰ Thus, focusing on the prevention of the occurrence of IID, rather than the treatment, would be valuable. In the current reports on prevention measures of IID, probiotic, ²¹ activated charcoal, ²² and neomycin ²³ are often discussed. Herbal medicines have also attracted great interest from doctors owing to the effects against the adverse effects of cancer treatment. Huangqin decoction, PHY906, Hanch-Shashin-to, Sairei-to, Shengjiang Xiexin decoction, Banxia Xiexin decoction, and Hypericum Perforatum have been studied in animal models and clinical trials and explore their application to the treatment and prevention of diarrhea caused by irinotecan. ²⁴ We have great interest in whether preventive interventions before courses of chemotherapy will be effective in alleviating IID. In the current study, a meta-analysis of 8 published randomized, controlled trials (RCTs) was performed to better understand the efficacy of various interventions to prevent IID, hoping to guide the preventive treatments of IID. For the safety of the patients, it should be noted that most of the included studies used antidiarrheal drugs such as loperamide after the onset of diarrhea. Although prophylaxis was used on the day of irinotecan chemotherapy or before the chemotherapy, and loperamide was used after the onset of diarrhea, the use of antidiarrheal drugs inevitably may affect the efficacy.

Methods

This systematic review and meta-analysis were reported according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 ²⁵ and registered on the International Prospective Register of Systematic Reviews (PROSPERO: CRD42022368633).

Literature Sources and Search

The search time started with establishing each database and ended on September 14, 2022. With the help of librarians (BL) and statisticians (RC), the search terms were determined, and 2 researchers independently conducted comprehensive literature searches on the 4 major electronic databases (PubMed, Web of Science, Embase and Cochrane Library). The search did not limit the language but was limited to RCTs. The strategy combined or separated Medical Subject Heading (MeSH) terms and all fields of keywords. We searched databases using combinations of the following keywords: “Irinotecan,” “diarrhea,” “prevention and control,” and “Randomized Controlled Trial.”

Inclusion Criteria

Articles that met the following criteria were included: (1) Participants of any race, age or gender and who were diagnosed with cancer at any stage, under chemotherapy with irinotecan; (2) Interventions for irinotecan-induced diarrhea are given on the day of irinotecan or before the chemotherapy; (3) The exposure factor was irinotecan-induced diarrhea, and diarrhea was defined by the Institute’s Common Terminology Criteria for Adverse Event; (4) randomized, controlled trials; and (5) data for events or event rate and sample size were available. All RCTs were included without a time restriction.

Exclusion Criteria

Articles that met any of the following conditions were excluded: (1) unavailable original full text; (2) duplicating published literature; (3) incomplete or missing research data; and (4) conference abstracts, case reports, guidelines, meta-analysis, reviews Letters, non-RCTs, non-related topic, animal experiments, or no data acquisition.

Data Extraction

The 2 researchers independently extracted the data included in the study into standardized forms. If there was a disagreement, it was discussed with the third researcher until an agreement was reached. We extracted the following data from the included studies: study first author, published year, country, research style, sample size, age, sex ratio (male/female), tumor type, chemotherapy regimens, follow-up time of diarrhea, preventive treatments, and diarrhea criteria. Grade 1 to grade 5 diarrhea was defined as all-grade diarrhea, and grade 3 to grade 5 diarrhea was defined as high-grade diarrhea.

Quality Assessment

Two researchers independently conducted quality evaluations based on the Cochrane Collaboration tool and the modified Jadad scale. Disagreements were resolved through discussion with a third researcher. The Cochrane risk assessment tool makes high-risk, low-risk, or unclear-risk judgments on random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias. ²⁶ The modified Jadad scale was scored from 4 aspects: randomization (0: Not randomized or inappropriate method of randomization; 1: The study was described as randomized; 2: The method of randomization was told and it was appropriate), concealment of allocation (0: The process of allocation concealment was not described; 1: The study was defined using the allocation concealment method; 2: The method of allocation concealment was described appropriately), double blinding (0: No blind or inappropriate form of blinding; 1: The study was defined as double-blind; 2. The process of double blinding was told and it was appropriate), and withdrawals and dropouts (0: The follow-up was not described; 1: Withdrawals and dropouts were described). ²⁷ Scores of 1-3 and 4-7 are considered low quality and high quality, respectively.

Data Analysis

The extracted efficacy data were entered in Review Manager version 5.4.1 for data synthesis and meta-analysis. The risk ratio (RR) was calculated for dichotomous data for outcomes. We used the Mantel-Haenszel method to calculate the 95% confidence interval (CI) and RRs. The 95% CI was calculated for all types of data. Heterogeneity was quantified by calculating the I² statistic. A fixed-effects model was used when no significant heterogeneity was found; otherwise, a random-effects model was used. Q statistics and I², according to the suggestion of the Cochrane Collaboration, were used to assess the heterogeneity of the included studies. P < .05 with I² > 50% was considered an indication of significant heterogeneity.^28,29 The Mantel-Haenszel method was employed to assess the significance of subgroup analyses. Trial-specific RRs of all-grade and high-grade diarrhea were combined to compute summary estimates using a random-effects model, which considers both within-study and between-study variations. Subgroup analysis was conducted to examine whether RRs of all-grade and high-grade diarrhea varied for chemotherapy regimens (monotherapy vs combination therapy), preventive treatments (traditional herbal medicine vs modern drugs), follow-up duration of diarrhea (≤4 weeks vs >4 weeks), and published year (before 2010 vs after 2010). To investigate the robustness of the findings, we conducted sensitivity analyses. In the case of clear outliers, we conducted a sensitivity analysis by removing the outlier trial (calculated by STATA 17.0) and comparing the new meta-analytic result to the result before removal. ³⁰

Results

The initial search yielded 1034 potentially relevant trials. Due to duplication, we excluded 196 studies. Seven hundred seventy-one more studies were excluded for at least one of the following reasons: Conference abstracts, Case reports, Guidelines, Meta-analysis, Reviews, Letters, Non-related topic, Animal experiments, and non-English. The remaining 67 studies were carefully screened, and 59 were excluded for Reviews, Non-related topic, No data acquisition, Not RCT, and others. Eight studies of 8 randomized, controlled trials were included in this review (see Figure 1).

Figure 1.

PRISMA diagram presenting the literature search and selection showing numbers of articles at each stage.

Assessment of Risk of Bias

The research quality evaluation based on the Jadad scale and the Cochrane risk of bias assessment manual is shown in Table 1 and Figure 2. According to the Jadad scale, 3 of the studies included in the analysis were of high quality, and the remaining 5 were of low quality.

Table 1.

Characteristics of the Included Studies.

Study	Country	research style	Sample size (n)	Age (years)	Tumor type	chemotherapy regimens	Precautionary measures	Follow-up Duration	Diarrhea criteria	Jadad score
Kee et al 31	USA	Multicenter RCT	100	57 (20-83)	Metastatic colorectal cancer	Irinotecan (140-740 mg)	Calcium aluminosilicate clay (1000 mg, po, q6h, 6 weeks, beginning with the CHE)	6 weeks	NCI-CTCAE (3.0 or 4.0)	5
de Man et al 32	Netherlands	Single RCT	19	≥18	Colorectal, esophagus and other cancers	Irinotecan (600 mg, homozygous UGT1A1*28 420 mg)	Protein and calorie restriction (restricted diet, 5 days, beginning 4 days before the CHE)	7-10 days	NCI-CTCAE (4.03)	3
de Jong et al 23	Netherlands	Multicenter RCT	62	58 (36-80)	Gastric, colorectal and other cancers	Irinotecan (350 mg/m2)	Neomycin (660 mg, po, tid, 3 days, starting 2 days before CHE)	3 weeks	NCI-CTCAE (2.0)	7
Sergio et al 22	Mexico	RCT	22	1-18	Rhabdomyosarcoma, PNET and other	Irinotecan + cisplatinum/doxorubicin/carboplatin (12 mg/m2/day days 1-5 and 8-12)	Activated Charcoal (250 mg, po, tid, until the end of the cycle, beginning 1 days before the CHE)	12 days	NCI-CTCAE	3
Mori et al 33	Japan	Single RCT	41	≤75	Advanced non-small-cell lung cancer	Irinotecan + cisplatin+ rG-CSF (160mg/m2)	Kampo medicine Hangeshashin-to (7.5 g, po, tid, at least 24 days, beginning 3 days before the CHE)	4 weeks	NCI-CTCAE (2.0)	3
Mego et al 21	Slovakia	Multicenter RCT	46	/	Colon carcinoma and rectal carcinoma	Irinotecan + cetuximab	Probiotics (10 × 109CFU of bacteria, po, tid, 12 weeks, beginning with the CHE)	12 weeks	NCI-CTCAE (4.1)	7
Kummar et al 34	USA	RCT	17	60 (18-75)	Advanced colorectal cancer	IFL regimen (irinotecan + 5-FU+ leucovorinat, 125 mg/m2)	PHY906 (using the Fibonacci scheme at successive dose levels of 1.2, 2.4, and 3.6 g, respectively, po, beginning with the CHE)	6 weeks	NCI-CTCAE (2.0)	3
Lai and Wang 35	China	Single RCT	120	25-75	Colorectal cancer	FOLFIRI regimen (irinotecan + calcium folinate+ fluorouracil, 180mg/m2)	Hot Ironing with Moxa Salt Packet on Tianshu and Shangjuxu (homemade moxa salt packs, qd, 5days, beginning 1 days before the CHE)	2 weeks	NCI-CTCAE (3.0)	3

Abbreviations: RCT, randomized controlled trial; 5-FU, 5-fluorouracil; po, take orally; q6h, every 6 hours; tid, 3 times a day; qd, once a day; NCI-CTCAE, National Cancer Institute Common Terminology Criteria for Adverse Events; CHE, chemotherapy.

Figure 2.

(A) Risk of bias graph: review authors’ judgments about each risk of bias item presented as percentages across all included studies. (B) Risk of bias summary: review authors’ judgments about each risk of bias item for each included study.

Characteristics of the Included Studies

The characteristics of the 8 included studies are summarized in Table 1. All analyses were randomized controlled trials, including 3 multicenter studies. Three studies were published before 2010, and 5 were published after 2010. The most common tumors in these studies were colon and rectal cancers, followed by gastric, esophagus, cholangitis, adenocarcinoma of unknown primary, rhabdomyosarcoma, osteosarcoma, PNET, retinoblastoma, neuroblastoma, hepatoblastoma, medulloblastoma, myofibroblastic tumor, suprarenal carcinoma, advanced non-small-cell lung cancer, and others. Two studies used irinotecan monotherapy, 5 used irinotecan in combination with other drugs, and the remaining one included monotherapy and combination therapy. Preventive treatments were classified as traditional herbal medicines, including Kampo medicine Hangeshashin-to (TJ-14), PHY906, hot ironing with Moxa Salt Packet on Tianshu and Shangjuxu, and modern medicines, including calcium aluminosilicate clay, neomycin, activated Charcoal, and probiotics. Protein and calorie restriction, involving the drinking of a synthetic powder brewed in 200 to 300 mL water to achieve 30% caloric restriction and 70% protein restriction, was not included in either group. The Jadad scores of the included studies ranged from 3 to 7.

Primary Outcome

Risk ratio of high-grade diarrhea

High-grade diarrhea occurred in 19 of 215 patients (8.8%) in the intervention arms and 55 of 241 (22.8%) in the control arms. A meta-analysis of the RR of high-grade diarrhea was performed in 8 RCTs (see Figure 3A). The RR of high-grade diarrhea was 0.31 (95% CI, 0.14-0.69; P = .004). Thus, the use of interventions for the prevention of irinotecan-induced high-grade diarrhea is effective. Significant heterogeneity was detected among the studies (I² = 51%, P = .04), which might be due to differences in the published year, chemotherapy regimens, preventive treatments, and follow-up duration of diarrhea. Therefore, subgroup analyses were performed according to these differences.

Figure 3.

(A) Forest plot of comparison: Intervention versus control, outcome: Risk Ratios (RRs) of High-Grade Diarrhea. (B) Forest plot of comparison: Intervention versus control, outcome: Risk Ratios (RRs) of All-Grade Diarrhea.

Secondary Outcome

Risk ratio (RR) of all-grade diarrhea

All-grade diarrhea occurred in 126 of 256 patients (49.2%) in the intervention arms and 169 of 241 (70.1%) in the control arms. A meta-analysis of the RR of all-grade diarrhea was performed for 8 RCTs (see Figure 3B). The RR of all-grade diarrhea was 0.76 (95% CI: 0.62-0.93; P < .008). Thus, the use of interventions for the prevention of irinotecan-induced all-grade diarrhea is effective. Significant heterogeneity was also detected among the studies (I² = 65%, P = .006). Therefore, subgroup analyses were performed according to the published year, chemotherapy regimens, preventive treatments, and follow-up duration of diarrhea.

Subgroup Analysis

Subgroup analysis according to published year

Many relevant studies have been completed in the past 20 years, and we defined the subgroup analysis based on those published before 2010 and after 2010. Three studies^22,23,33 were published before 2010. Compared with the control group, preventive treatments decreased the risk of high-grade diarrhea caused by irinotecan (8.8% vs 41.0%, RR: 0.21, 95% CI: 0.05-0.79, I² = 64%, P = .02), but there was no statistically significant change in the risk of all-grade diarrhea (58.2% vs 89.7%, RR: 0.75, 95% CI: 0.53-1.08, I² = 84%, P = .12). For the 5 studies^{21,31,32,34,35} published after 2010, preventative treatments did not significantly change the risk of high-grade diarrhea caused by irinotecan (6.7% vs 14.1%, RR: 0.42, 95% CI: 0.14-1.25, I² = 42%, P = .14), but decreased the risk of all-grade diarrhea (44.2% vs 60.7%, RR: 0.76, 95% CI: 0.60-0.97, I² = 32%, P = .03) (see Tables 2 and 3, Supplemental Figure 1, and Supplemental Figure 2)

Table 2.

Summary Risk Ratios (RRs) of High-Grade Diarrhea in the Subgroup Analysis.

High-grade diarrhea	Trials, n	Total events, n (%)		RR (95% CI)	P	P for group difference
		Interventions	Control
Year
Before 2010	3	8/91 (8.8%)	32/78 (41.0%)	0.21 (0.05,0.79)	.02	.43
After 2010	5	11/165 (6.7%)	23/163 (14.1%)	0.42 (0.14,1.25)	.12
Chemotherapy regimens
Monotherapy	2	6/47 (12.8%)	15/53 (28.3%)	0.48 (0.21,1.13)	.1	.04
Combination therapy	5	5/160 (3.1%)	35/142 (24.6%)	0.14 (0.06,0.32)	<.00001
Precautionary measures
Traditional herbal medicines	3	2/92 (3.3%)	20/98 (20.4%)	0.20 (0.07,0.60)	.004	.48
Modern drugs	4	15/145 (10.3%)	31/124 (25.0%)	0.37 (0.10,1.40)	.14
Follow-up duration
≤4 weeks	5	11/170 (6.5%)	44/157 (28.0%)	0.24 (0.11,0.52)	.0003	.48
>4 weeks	3	8/86 (7.4%)	11/84 (13.1%)	0.50 (0.08,3.11)	.46

Table 3.

Summary Risk Ratios (RRs) of All-Grade Diarrhea in the Subgroup Analysis.

All-grade diarrhea	Trials, n	Total events, n (%)		RR (95% CI)	P	P for group difference
		Interventions	Control
Year
Before 2010	3	53/91 (58.2%)	70/78 (89.7%)	0.75 (0.53,1.08)	.12	.97
After 2010	5	73/165 (44.2%)	99/163 (60.7%)	0.76 (0.60,0.97)	.03
Chemotherapy regimens
Monotherapy	2	38/47 (80.9%)	48/53 (90.6%)	0.90 (0.77,1.05)	.19	.13
Combination therapy	5	56/160 (45.4%)	87/142 (61.2%)	0.61 (0.37,0.99)	.04
Precautionary measures
Traditional herbal medicines	3	34/92 (37.0%)	58/98 (59.2%)	0.68 (0.35,1.33)	.26	.85
Modern drugs	4	78/145 (53.8%)	95/124 (76.6%)	0.73 (0.53,1.01)	.06
Follow-up duration
≤4 weeks	5	79/170 (46.5%)	112/157 (71.3%)	0.72 (0.53,0.98)	.04	.5
>4 weeks	3	47/86 (54.7%)	57/84 (67.9%)	0.83 (0.66,1.04)	.11

Subgroup analysis according to chemotherapy regimens

Seven studies^{21 -23,32 -35} reported on chemotherapy regimens. Subgroup analysis based on chemotherapy regimens showed monotherapy did not significantly change the risk of all-grade diarrhea (80.9% vs 90.6%, RR: 0.90, 95% CI: 0.77-1.05, I² = 0%, P = .19) and the risk of high-grade diarrhea caused by irinotecan (12.8% vs 28.3%, RR: 0.48, 95% CI: 0.21-1.13, I² = 0%, P = .10). The combination therapy group had a decrease in both the risk of high-grade diarrhea (3.1% vs 24.6%, RR: 0.14, 95% CI: 0.06-0.32, I² = 0%, P < .00001) and the risk of all-grade diarrhea caused by irinotecan (45.4% vs 61.2%, RR:0.61, 95% CI: 0.37-0.99, I² = 80%, P = .04). There was no heterogeneity in high-grade diarrhea with monotherapy (I² = 0%) or combination therapy (I² = 0%) (see Tables 2 and 3, Supplemental Figure 3, Supplemental Figure 4).

Subgroup analysis according to preventive treatments

We explored the impact of traditional herbal medicine and modern drugs on the RR of diarrhea; RRs were calculated according to the type of preventive treatments. Subgroup analysis based on preventive treatments showed that the traditional herbal medicine group had significantly decreased risk of high-grade diarrhea caused by irinotecan (3.3% vs 20.4%, RR: 0.20, 95% CI: 0.07-0.60, I² = 0%, P = .004), while the risk of all-grade diarrhea caused by irinotecan (37.0% vs 59.2%, RR: 0.68, 95% CI: 0.35-1.33, I² = 82%, P = .26) did not have statistically significant differences. The modern drugs group decreased the risk of high-grade diarrhea caused by irinotecan (10.3% vs 25.0%, RR: 0.37, 95% CI: 0.10-1.40, I² = 74%, P = .009), but the risk of all-grade diarrhea (53.8% vs 76.6%, RR: 0.73, 95% CI: 0.53-1.01, I² = 73%, P = .06) was not significantly different (see Tables 2 and 3, Supplemental Figure 5, and Supplemental Figure 6).

Subgroup analysis according to follow-up duration of diarrhea

Studies were further stratified according to the Follow-up duration of diarrhea (≤4 weeks vs >4 weeks). Subgroup analysis based on Follow-up duration of diarrhea showed that the ≤4 weeks group had decreased risk of high-grade diarrhea (6.5% vs 28.0%, RR: 0.24, 95% CI: 0.11-0.52, I² = 28%, P = .0003) and risk of all-grade diarrhea caused by irinotecan (46.5% vs 71.3%, RR: 0.72, 95% CI: 0.53-0.98, I² = 82%, P = .04). The group with a follow-up duration of diarrhea ≥4 weeks showed no statistically significant difference in the risk of high-grade diarrhea (7.4% vs 13.1%, RR: 0.50, 95% CI: 0.08-3.11, I² = 52%, P = .46) and the risk of all-grade diarrhea caused by irinotecan (54.7% vs 67.9%, RR: 0.83, 95% CI: 0.66-1.04, I² = 0%, P = .11) (see Tables 2 and 3, Supplemental Figure 7, Supplemental Figure 8).

Sensitivity analysis

Sensitivity analysis by STATA indicated that the combined effect size of Gallegos-Castorena Sergio 2008²² was significantly different from the total combined effect size in all-grade diarrhea and high-grade diarrhea, thus we consider it as an outlier trial. After excluding the outlier trial, the risk of all-grade diarrhea (7.4% vs 13.1%, RR: 0.84, 95% CI: 0.72-0.97, I² = 37%, P = .14) and high-grade diarrhea (7.4% vs 13.1%, RR: 0.42, 95% CI: 0.20-0.89, I² = 34%, P = .17) were all reduced after preventive treatments. The heterogeneity was significantly decreased from the previous. Therefore, the robustness of the systematic review is considered to be poor. This study will be discussed carefully (see Figure 4, Supplemental Figure 9, and Supplemental Figure 10).

Figure 4.

(A) Sensitivity analysis of high-grade diarrhea. (B) Sensitivity analysis of all-grade diarrhea.

Discussion

As reported in previous studies, while irinotecan remains the primary antitumor agent for treating colorectal cancer, irinotecan-treated patients are likely to suffer adverse events. Although many promising agents are being investigated, no approved prophylactic measures are available against IID. ³⁶ To the best of our knowledge, this is the first meta-analysis to evaluate the efficacy of preventive treatments for IID. The analysis of 8 randomized controlled studies showed a significant reduction in the risk of all-grade and high-grade diarrhea after preventive treatment. Compared with the intervention group, the risk of all-grade and high-grade diarrhea in the control group was 1.42 times and 2.59 times, respectively, and the risk reduction of high-grade diarrhea was more significant. The results showed that preventive IID treatments effectively prevented high-grade diarrhea. Several subgroup and sensitivity analyses were carried out to identify sources of statistical heterogeneity among trials.

In the subgroup analysis, we found P for group difference is .04 for the chemotherapy regimens in high-grade diarrhea and no heterogeneity in high-grade diarrhea with monotherapy (I² = 0%) or combination therapy (I² = 0%), suggesting that a significant difference between the chemotherapy regimens and that the chemotherapy regimens are the source of heterogeneity in high-grade diarrhea. Therefore, we carefully discuss the chemotherapy regimens in the following paragraphs. The other subgroups (P for Group Differences >.05) had no statistical significance. However, we found that the traditional herbal medicines group had decreased risk of high-grade diarrhea caused by irinotecan (RR: 0.20), I² = 0%, P = .004, suggesting that there was a significant difference between the experimental and the control group within the traditional herbal medicines group and the heterogeneity was not significant. At present, the field of traditional herbal medicines is developing rapidly and there is still a lot of space to explore, so we are discussing more possibilities.

Using subgroup analysis to identify potential risk factors, we found that the chemotherapy regimen is the source of heterogeneity in high-grade diarrhea. Combined therapy can obtain better preventive effects of IID. There were significant subgroup differences in different chemotherapy regimens for all-grade and high-grade diarrhea, and there was no heterogeneity in high-grade diarrhea with monotherapy (I² = 0%) or combination therapy (I² = 0%). This indicates that the chemotherapy regimen is a crucial factor in the heterogeneity of high-grade diarrhea. Studies have shown that diarrhea with irinotecan may become severe when sequential administration is performed. ³⁷ Therefore, irinotecan is often used in combination with other antitumor drugs in clinical practice. ³⁸ With the development of research, irinotecan chemotherapy regimens have been changed from combination of 5-fluorouracil (5-FU) + leucovorin ³⁹ to ternary treatment (FOLFIRI), FOLFOXIRI and FOLFIRINOX,^40,41 and FOLFIRI and FOLFOXIRI regimens with added monoclonal antibodies bevacizumab, ⁴² cetuximab, ⁴³ and panitumumab. ⁴⁴ These chemotherapy regimens are increasingly effective in treating cancer. The above studies have shown that the current therapeutic effect of combination therapy is not only significantly better in the treatment of tumors, but also in the preventive treatment of high-grade diarrhea than that of irinotecan monotherapy, suggesting that using a combination of drugs with irinotecan could be a preferable selection.

The subgroup analysis result suggests that traditional herbal medicines were significantly effective against high-grade diarrhea, as well as the completion of tumor chemotherapy. Traditional herbal medicines mentioned in this review include Kampo medicine Hangeshashin-to (TJ-14), PHY906 and hot ironing with Moxa Salt Packet on Tianshu and Shangjuxu. PHY906 and TJ-14 are traditional herb medicines from China and Japan, respectively, and their major component is baicalin (BG).^34,45 BG is a natural glucuronide conjugate with an inhibitory effect against β-glucuronidase. ⁴⁶ It is conceivable that BG-containing herb medicines are orally administered to ameliorate gastrointestinal toxicity via the inhibition of bacterial β-glucuronidase and suppressing the formation of toxic SN-38 from SN-38G. Another preventive intervention of traditional herbal medicines is hot ironing with Moxa Salt Packet on Tianshu and Shangjuxu, in which the moxa salt pack may stimulate the meridian Qi through the stimulation of acupuncture points, mobilizing the meridian function and creating a lasting warming effect.^35,47

It is well known that SN-38 is the primary active metabolite of irinotecan. However, BG has been found to reduce the gastrointestinal exposure of SN-38 via inhibition of β-glucuronidase. Therefore, it is necessary to study the effect of BG on irinotecan’s antitumor performance. In one of our included studies, using a murine Colon 38 xenograft model, it was observed that BG-containing herbs did not diminish the antitumor impact of irinotecan and even significantly enhanced the antitumor activity of irinotecan. It also attenuated irinotecan-induced toxicity by reducing body weight loss and mortality. In the analysis of drug metabolism in clinical patients, the mean and median of key pharmacokinetic parameters (eg, irinotecan with C_max ranging from 476 to 2390 ng/mL and AUC values ranging from 2574 to 11 228 hours·ng/mL, SN-38 with C_max ranging from 3.4 to 37.1 ng/mL and AUC values ranging from 47.3 to 452.8 hours·ng/mL) of irinotecan and SN-38 were comparable no matter for the patients who taking baicalin-containing herbs or a placebo. ³⁴ This indicated that BG-containing herbs do not alter the pharmacokinetic parameters of irinotecan as well as the metabolism of SN-38, the active metabolite of irinotecan. In addition, another study has confirmed that BG-containing herbs could enhance the effect of irinotecan on inhibiting the proliferation and inducing apoptosis of HCT116 cells in vitro and enhance the antitumor effect of irinotecan in HCT116 xenograft tumor model in vivo. A reduction in gastrointestinal toxicity caused by irinotecan was observed. ⁴⁸ In conclusion, BG could play an anti-proliferation and anti-apoptosis role in colon cancer cells in vitro and in vivo. BG in preliminary studies can alleviate IID without reducing the anticancer effect of irinotecan, but whether β-glucuronidase inhibitors can have the same effect requires rigorous and more research.

Previous studies have shown that only using the Jadad scale to evaluate research quality in the systematic review is not comprehensive. In our quality assessment based on the Cochran risk assessment manual and the Jadad scale, we found that the high-risk papers evaluated by the Cochrane collaboration tool were roughly consistent with the low-quality papers considered by the Jadad scale. In addition, the observations in the traditional herbal group were mainly high-risk and low-quality (Jadad score = 3) because they did not have double blinding or had inappropriate forms of blinding. The experiments using herbal medicine can be unified as oral administration of capsules, and placebo-controlled trials should be strictly used to improve the quality of papers. The follow-up duration for traditional herbal studies is 2 to 6 weeks and may not be sufficient to develop high-grade IID. Although the traditional herbal medicines group has good efficacy, the study quality is low. It is worthy of more research to confirm the efficacy of traditional herbal medicines.

There were no significant subgroup differences between studies before and after 2010. Studies before 2010 effectively reduced irinotecan-induced high-grade diarrhea but with high heterogeneity. The follow-up duration of diarrhea ≤4 weeks significantly reduces high-grade and all-grade diarrhea. Late-onset diarrhea is a severe gastrointestinal toxicity of irinotecan, often due to the inadequacy of drugs in treating IID and preventing all severe diarrhea. ¹⁵ Diarrhea may become worse after a large number of chemotherapy cycles and is more difficult to control using any intervention. Therefore, the reason for the worse results in studies with >4 weeks interventions may be that the shorter studies did not allow enough time to see the full effect of the standard-length regimen on diarrhea.

In the sensitivity analysis, after excluding the outlier trial, ²² the heterogeneity of all-grade diarrhea (I² = 37%) and high-grade diarrhea (I² = 34%) were significantly reduced compared to before. The outlier trial is the only study in this review with a sample age range of 1 to 18 years old, which also has a small sample size and is prone to bias as compared to other studies of adults. Physicians have long used activated charcoal to treat abdominal distension and diarrhea. ⁴⁹ Before the outlier trial, a study on activated charcoal for adult IID showed that activated charcoal could effectively relieve the symptoms. However, due to the complex influencing factors, such as the patients as their own controls and the complexity of irinotecan dose reduction and supportive care during the trial, the conclusions of this trial were limited and were not included in our systematic analysis. ⁵⁰ In contrast, a previous study has shown that multiple oral doses of activated charcoal do not modulate the clearance of CPT-11 and SN-38 in rats. The implication is that activated charcoal alone may not be very effective in preventing IID. ⁵¹ Therefore, the conclusion that activated charcoal can effectively alleviate IID still needs more research, and the outlier trial affects the stability of the systematic analysis.

Limitation

We found that almost all trials used antidiarrheal agents after the onset of diarrhea, such as loperamide. However, due to the different loperamide treatment regimens, some trials did not include detailed records of antidiarrheal drugs. Therefore, it is difficult to extract the information on antidiarrheal drugs for further analysis, which may affect the efficacy of preventive measures for IID. There were several combination chemotherapy regimens in this review. Still, they could not be further divided into groups, and the impact of other combined drugs on diarrhea was hard to assess. Only the effects of monotherapy and combination therapy on the results were analyzed. Moreover, due to the small number of studies, this review analyzed the overall effects of different preventive treatments on IID. It divided them into traditional herbs and modern medicines for specific analysis, and it was hard to investigate the impact of a single preventive therapy. In the future, with a more definite chemotherapy regimen, the adverse reactions of other common clinical drugs could be analyzed. Although this review has suggested that clinicians might use traditional herbal medicine to reduce the risk of high-grade diarrhea, further studies on the effects of different methods of traditional medicine use are necessary before clinical recommendations can be made.

Conclusion

In summary, compared to irinotecan monotherapy, combination therapy has a better antitumor effect and can obtain better preventive results in IID. Preventive treatments significantly reduced the risk of high-grade and all-grade diarrhea, confirming the efficacy in the incidence and severity of IID, among which traditional herbal medicine (baicalin-containing) was seen to be quite effective in reducing the severity of IID. Researchers should give attention to the potential of traditional herbs for the prophylactic treatment of IID, extend the follow-up duration, improve the quality of studies, and achieve the purpose of preventing IID by exploring β-glucuronidase inhibitors.