MATERIALS AND METHODS

Data sources: PubMed, Scopus, Web of Science and Cochrane Central Register of Controlled Trials were searched for studies evaluating efficacy and/or tolerability of herbal medicines in any types of IBD. Data were collected from 1966-2013 (up to Feb). The search terms were: “herb” or “plant” or “herbal” and “inflammatory bowel disease”. There was no language restriction. The reference list from retrieved articles was also reviewed for additional applicable studies.

Study selection: Controlled trials evaluating the efficacy and/or tolerability of herbal medicines in patients with any types of IBD were considered. “clinical response”, “remission”, “any adverse events” and “serious adverse events” were the key outcomes of interest. All published studies as well as abstracts presented at meetings were evaluated. Two reviewers independently examined the title and abstract of each article to eliminate duplicates, reviews, case studies and uncontrolled trials.

The reviewers independently extracted data on patients' characteristics, therapeutic regimens, dosage, trial duration and outcome measures. There was no disagreement between reviewers.

Assessment of trial quality: Jadad score which indicates the quality of the studies based on their description of randomization, blinding and dropouts (withdrawals) was used to assess the methodological quality of trials (Jadad and Enkin, 2007). The quality scale ranged from 0-5 points with a low quality report of score 2 or less and a high quality report of score at least 3.

Statistical analysis: Data from selected studies were extracted in the form of 2×2 tables by study characteristics. Included studies were weighted and pooled. Data were analyzed using StatsDirect software version 2.7.9. Relative Risk (RR) and 95% confidence intervals (95% CI) were calculated using Mantel-Haenszel, Rothman-Boice (for fixed effects) or Der Simonian-Laird (for random effects) methods. The Cochran Q test was used to test heterogeneity and p<0.05 considered significant. In case of heterogeneity or few included studies, the random effects model was used. Funnel plot was used as publication bias indicator.

RESULTS

The electronic searches yielded 1224 items; 698 from PubMed, 5 from Cochrane Central, 35 from Web of Science and 355 from Scopus. From these studies, 41 were scrutinized in full text.

Thirty four reports were considered ineligible. Thus, 8 trials were included in the analysis represented 812 patients (Fig. 1) (Tang et al., 2011; Gupta et al., 1997; 2001; Fernandez-Banares et al., 1999; Gong et al., 2012; Tong et al., 2010; Ling et al., 2010; Chen et al., 1994).

Fig. 1:	Flow diagram of the study selection process

From these 8 studies, 3 obtained Jadad score of 3 or more (Tang et al., 2011; Fernandez-Banares et al., 1999; Gong et al., 2012), 4 gained score of 2 or less (Gupta et al., 1997; 2001; Tong et al., 2010; Ling et al., 2010) and one with undetermined score because of its language was Japanese (Chen et al., 1994) (Table 1).

Table 1:	Characteristics of studies included in the meta-analysis

Among studies included, 5 investigated the efficacy and/or tolerability of herbal medicines in UC (Tang et al., 2011, Gupta et al., 1997; Fernandez-Banares et al., 1999; Gong et al., 2012; Tong et al., 2010), 1 in chronic colitis (Gupta et al., 2001), 1 in intractable colitis (Chen et al., 1994) and 1 in any type of IBD (Ling et al., 2010). Four herbal products were used in included studies: Andrographis paniculata in one (Tang et al., 2011), Boswellia serrata in 2 (Gupta et al., 1997; 2001), Plantago ovata in one (Fernandez-Banares et al., 1999) and traditional Chinese medicine in 4 studies (Gong et al., 2012; Tong et al., 2010; Ling et al., 2010; Chen et al., 1994). Induction of treatment was investigated in seven studies and duration of these studies is 4-8 weeks (Tang et al., 2011; Gupta et al., 1997; 2001; Gong et al., 2012; Tong et al., 2010; Ling et al., 2010; Chen et al., 1994). Maintenance of remission was evaluated in two studies and duration of these studies was 1 year (Fernandez-Banares et al., 1999). The 5-ASAs used in included studies were sulfasalazine or mesalazine (Table 2).

Table 2:	Investigated outcomes in studies included in the meta-analysis of the efficacy and safety of herbal remedies in any type of IBD compared to 5-aminosalisylates
AE: Adverse events, H: Herbal remedy, IBD: Inflammatory bowel disease, M: Mesalazine, S: Sulfasalazine, TCM: Traditional Chinese medicine, UC: Ulcerative colitis

Scientific name of plant(s) used in herbal medicine, study design, inclusion and exclusion criteria, interventions, concomitant medications, patients’ characteristics, duration of study and definition of outcomes investigated in each included study have been shown in Table 1. Results of investigated outcomes for each included study have been demonstrated in Table 2.

EFFICACY

Clinical remission: The summary for Relative Risk (RR) of clinical remission in IBD patients for seven included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gupta et al., 1997; 2001; Gong et al., 2012; Tong et al., 2010; Ling et al., 2010; Chen et al., 1994) was 1.28 with 95% CI = 1.07-1.54 (p = 0.008, Fig. 2a). The Cochrane Q test for heterogeneity indicated that the studies are not heterogeneous (p = 0.38, Fig. 2b) and could be combined, thus the fixed effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for clinical remission in IBD patients was 2.42 (95% CI =-0.52 to 5.35, p = 0.09) and Kendall's tau = 0.43, p = 0.24 (Fig. 2c).

The summary for relative risk (RR) of clinical remission in UC patients for four included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gupta et al., 1997; Gong et al., 2012; Tong et al., 2010) was 1.06 with 95% CI = 0.85-1.33 (p = 0.61, Fig. 3a). The Cochrane Q test for heterogeneity indicated that the studies are not heterogeneous (p = 0.65, Fig. 3b) and could be combined, thus the fixed effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for clinical remission in UC patients was 1.72 (95% CI =-1.7 to 5.15, p = 0.16 and Kendall's tau = 0.67, p = 0.33 (Fig. 3c).

Fig. 2(a-c):

(a) Individual and pooled relative risk for the outcome of “clinical remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients, (b) Heterogeneity indicators for the outcome “clinical remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients and (c) Publication bias indicators for the outcome of “clinical remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients

Fig. 3(a-c):

(a) Individual and pooled relative risk for the outcome of “clinical remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients, (b) Heterogeneity indicators for the outcome of “clinical remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients and (c) Publication bias indicators for the outcome of “clinical remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients

Based on plant type, RR of clinical remission was significant for TCM (1.29; 95% CI = 1.04-1.59, p = 0.02) and non-significant for Andrographis paniculata and Boswellia serrata (Table 3).

Clinical response: The summary for RR of clinical response in IBD patients for five included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gong et al., 2012; Tong et al., 2010; Ling et al., 2010; Chen et al., 1994) was 1.19 with 95% CI = 1.01-1.39 (p = 0.04, Fig. 4a). The Cochrane Q test for heterogeneity indicated that the studies are heterogeneous (p = 0.03, Fig. 4b) and could not be combined, thus the random effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for clinical response in IBD patients was 6.72 (95% CI =-6.57 to 20, p = 0.21) and Kendall's tau = 0.6, p = 0.23 (Fig. 4c).

The summary for RR of clinical response in UC patients for three included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gong et al., 2012; Tong et al., 2010) was 1.1 with 95% CI = 0.91-1.33 (p = 0.32, Fig. 5a). The Cochrane Q test for heterogeneity indicated that the studies are heterogeneous (p = 0.06, Fig. 5b) and could be combined but because of few included studies the random effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for clinical response in UC patients could not be calculated because of too few strata.

Table 3:	Results obtained from sub-analysis of included studies based on plant type
AE: Adverse events, IBD: Inflammatory bowel disease, TCM: Traditional Chinese medicine, UC: Ulcerative colitis

Fig. 4(a-c):

(a) Individual and pooled relative risk for the outcome of “clinical response” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients (b) Heterogeneity indicators for the outcome of “clinical response” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients and (c) Publication bias indicators for the outcome of “clinical response” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients

Based on plant type, RR of clinical response was significant for TCM (1.25; 95% CI = 1.11-1.4, p = 0.0001) and non-significant for Andrographis paniculata (Table 3).

Endoscopic remission: The summary for Relative Risk (RR) of endoscopic remission in IBD patients for five included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gupta et al., 1997; 2001; Gong et al., 2012; Tong et al., 2010) was 0.85 with 95% CI = 0.34-2.12 (p = 0.73, Fig. 6a). The Cochrane Q test for heterogeneity indicated that the studies are heterogeneous (p<0.0001, Fig. 6b) and could not be combined, thus the random effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for endoscopic remission in IBD patients was 1.79 (95% CI =-12.16 to 15.74, p = 0.71) and Kendall's tau = 0, p = 0.82 (Fig. 6c).

The summary for RR of endoscopic remission in UC patients for four included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gupta et al., 1997; Gong et al., 2012; Tong et al., 2010) was 0.81 with 95% CI = 0.31 to 2.13 (p = 0.67, Fig. 7a). The Cochrane Q test for heterogeneity indicated that the studies are heterogeneous (p<0.0001, Fig. 7b) and could not be combined, thus the random effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for endoscopic remission in UC patients was 9.78 (95% CI =-48.2 to 67.77, p = 0.54 and Kendall's tau = 0.33, p = 0.75 (Fig. 7c).

Fig. 5(a-b):	(a) Individual and pooled relative risk for the outcome of “clinical response” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients and (b) Heterogeneity indicators for the outcome of “clinical response” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients

Based on plant type, RR of endoscopic remission was non-significant for Andrographis paniculata, Boswellia serrata and TCM (Table 3).

Endoscopic response: The summary for RR of endoscopic response in IBD patients for four included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gupta et al., 1997; 2001; Tong et al., 2010) was 1.14 with 95% CI = 0.99-1.3 (p = 0.07, Fig. 8a). The Cochrane Q test for heterogeneity indicated that the studies are not heterogeneous (p = 0.51, Fig. 8b) and could be combined, thus the fixed effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for endoscopic response in IBD patients was-0.61 (95% CI =-7.26 to 6.04, p = 0.73) and Kendall's tau =-0.33, p = 0.33 (Fig. 8c).

The summary for RR of endoscopic response in UC patients for three included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gupta et al., 1997; Tong et al., 2010) was 1.14 with 95% CI = 0.98-1.32 (p = 0.08, Fig. 9a). The Cochrane Q test for heterogeneity indicated that the studies are not heterogeneous (p = 0.35, Fig. 9b) and could be combined but because of few included studies the random effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for endoscopic response in UC patients could not be calculated because of too few strata.

Based on plant type, RR of endoscopic response was non-significant for Andrographis paniculata, Boswellia serrata and TCM (Table 3).

Histological remission: The Relative Risk (RR) of histological remission in IBD (UC) patients for comparison of herbal medicines to 5-ASAs (Tang et al., 2011) was 0.8 with 95% CI = 0.05-13.72, a non-significant RR.

Fig. 6(a-c):

a) Individual and pooled relative risk for the outcome of “endoscopic remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients, (b) Heterogeneity indicators for the outcome “endoscopic remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients and (c) Publication bias indicators for the outcome of “endoscopic remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients

Histological response: The Relative Risk (RR) of histological response in IBD (UC) patients for comparison of herbal medicines to 5-ASAs (Tang et al., 2011) was 1.32 with 95% CI = 0.64 to 2.9, a non-significant RR.

Relapse: The summary for RR of relapse in IBD patients for two included trials comparing herbal medicines with 5-ASAs (Gupta et al., 2001; Fernandez-Banares et al., 1999) was 1.05 with 95% CI = 0.6-1.83 (p = 0.87, Fig. 10a). The Cochrane Q test for heterogeneity indicated that the studies are not heterogeneous (p = 0.94, Fig. 10b) and could be combined but because of few included studies the random effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for relapse in IBD patients could not be calculated because of too few strata.

Based on plant type, RR of relapse was non-significant for Boswellia serrata and Plantago ovata (Table 3).

TOLERABILITY

Any adverse events: The summary for RR of any adverse events in IBD patients for five included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gupta et al., 2001; Fernandez-Banares et al., 1999; Tong et al., 2010) was 1.31 with 95% CI = 0.8-2.14 (p = 0.28, Fig. 11a).

Fig. 7(a-c):

(a) Individual and pooled relative risk for the outcome of “endoscopic remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients, (b) Heterogeneity indicators for the outcome “endoscopic remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients and (c) Publication bias indicators for the outcome of “endoscopic remission” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients

The Cochrane Q test for heterogeneity indicated that the studies are not heterogeneous (p = 0.99, Fig. 11b) and could be combined, thus the fixed effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for any adverse events in IBD patients was -0.19 (95% CI = -1.47 to 1.08, p = 0.59) and Kendall's tau = -0.33, p = 0.33 (Fig. 11c).

Serious adverse events: The summary for Relative Risk (RR) of serious adverse events in IBD patients for two included trials comparing herbal medicines with 5-ASAs (Tang et al., 2011; Gong et al., 2012) was 1.8 with 95% CI = 0.13 to 24.5 (p = 0.66, Fig. 12a). The Cochrane Q test for heterogeneity indicated that the studies are not heterogeneous (p = 0.28, Fig. 12b) and could be combined but because of few included studies the random effects for individual and summary of RR was applied. Regression of normalized effect vs. precision for all included studies for serious adverse events in IBD patients could not be calculated because of too few strata.

DISCUSSION

The 5-ASAs are usually considered as the first line treatment for IBD. Since, the use of herbal remedies for the treatment of chronic gastrointestinal disorders like IBD is increasing (Farzaei et al., 2013; Rahimi and Abdollahi, 2013, Rahimi and Abdollahi, 2012; Rahimi et al., 2009a, 2013b), a meta-analysis was conducted to compare the efficacy of herbal remedies with 5-ASAs.

Fig. 8(a-c):

(a) Individual and pooled relative risk for the outcome of “endoscopic response” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients, (b) Heterogeneity indicators for the outcome “endoscopic response” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients and (c) Publication bias indicators for the outcome of “endoscopic response” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients

Fig. 9(a-b)	(a) Individual and pooled relative risk for the outcome of “endoscopic response” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients and (b) Heterogeneity indicators for the outcome “endoscopic response” in the studies considering herbal medicines comparing to 5-ASAs therapy in UC patients

Fig. 10(a-b):	(a) Individual and pooled relative risk for the outcome of “relapse” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients and (b) Heterogeneity indicators for the outcome of “relapse” in the studies considering herbal medicines comparing with 5-ASAs therapy in IBD patients

The results of this meta-analysis showed that induction of clinical response and remission by herbal remedies is significant when compared with 5-ASAs. Regarding other outcomes endoscopic efficacy, histological efficacy, relapse, any adverse related to efficacy including events and serious adverse events, no significant difference was seen between herbal medicines and 5-ASAs.

The present meta-analysis may have been inevitably limited by small sample sizes of studies and heterogeneity. Because the included trials involved herbal medicines containing different plants administered to patients with various subtypes of IBD, the trials were disaggregated.

Fig. 11(a-c):

(a) Individual and pooled relative risk for the outcome of “any adverse events” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients, (b) Heterogeneity indicators for the outcome “any adverse events” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients and (c) Publication bias indicators for the outcome of “any adverse events” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients

Fig. 12(a-b):	(a) Individual and pooled relative risk for the outcome of “serious adverse events” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients and (b) Heterogeneity indicators for the outcome “serious adverse events” in the studies considering herbal medicines comparing to 5-ASAs therapy in IBD patients

Thus, sub-analyses based on type of IBD and plant type were performed. The results of sub-analyses based on IBD type showed that there is no significant difference between herbal medicines and 5-ASAs in inducing or maintaining efficacy in UC and chronic colitis. But inducing clinical response and remission by herbal medicines in patients with intractable colitis was significant compared to that of 5-ASAs. The sub-analysis based on type of herbal medicine conducted on 4 groups: Andrographis paniculata, Boswellia serrata, Plantago ovata and Traditional Chinese Medicine (TCM). TCM showed significant effect in induction of clinical response, clinical remission and endoscopic response in comparison with 5-ASAs. Other three groups did not show significant difference compared to 5-ASAs in any investigated outcomes.

CONCLUSION

Overall, it seems that efficacy and tolerability of herbal medicines in IBD is comparable to 5-ASAs, but the evidence is too limited to make any confident conclusions. Further high quality, large controlled trials are warranted to better conclusion.

ACKNOWLEDGMENTS

This study is the outcome of an in-house financially non-supported study. Authors wish to National Elite Foundation (NEF) and the Iran National Science Foundation (INSF).