The efficacy of Jianpi Yiqi therapy for chronic atrophic gastritis: A systematic review and meta-analysis

Jianpi Yiqi therapy (JYT) is a classical therapy in treating chronic atrophic gastritis (CAG), but the clinical effects of it are still contentious. The purpose of this article is to evaluate the efficacy and safety of JYT for CAG. Seven electronic databases including PubMed, Embase, Springer Link, CNKI (China National Knowledge Infrastructure), VIP (Chinese Scientific Journals Database), Wan-fang database, and CBM (Chinese Biomedicine Database) were searched from their inception to November 1, 2016. 13 randomized controlled trials (RCTs) with a total of 1119 participants were identified for analysis. Meta-analyses demonstrated that both JYT (RR 1.41; 95% CI 1.27, 1.57; P < 0.00001) and JYT + western medicine (RR 1.27; 95% CI 1.17, 1.38; P < 0.00001) were more efficacious than only western medicine. Furthermore, JYT had potential improvement on traditional Chinese medicine (TCM) symptoms scores such as stomachache, stomach distention, belching, fatigue, et al. In addition, no serious adverse events were reported in the selected trials. The Cochrane Collaboration’s risk of bias tool was evaluated for the weaknesses of methodological quality, while the quality level of Grades of Recommendations Assessment Development and Evaluation (GRADE) evidence classification indicated “Very low”. This meta-analysis indicates that JYT may have potential effects on the treatment of patients with CAG. However, due to limitations of methodological quality and small sample size of the included studies, further standardized research of rigorous design should be needed.


Introduction
Chronic atrophic gastritis (CAG), characterized by a loss of normal glandular structures and a collapse of the reticular skeleton of the gastric mucosa (GM), is a well-established premalignant lesion of gastric cancer (GC) in the absence of specific clinical manifestations [1][2][3]. Recently, with the increasing of incidence and prevalence of CAG in China, the risk of GC has been growing, causing significant reductions in patients' quality of life and increasing substantial healthcare costs [4].
Although great progress has been made in elaborating the pathogenesis of CAG, most western medicines, including Helicobacter pylori (Hp) eradication, acid suppression, and nonsteroidal anti-inflammatory drugs, remain unsatisfied [5]. Due to chronicity and recurrence of this disease, many sufferers have put their concentrations on alternative treatments such as traditional Chinese medicine (TCM). Invigorating spleen and reinforcing qi (Chinese name in pinyin "Jianpi Yiqi") is a classical Chinese therapy in treating CAG [6]. However, the current state of evidence of Jianpi Yiqi therapy (JYT) for CAG has been unknown. Therefore, a systematic review and meta-analysis of randomized, conventional western medicine controlled trials was conducted to evaluate its efficacy and safety.

Search strategy
We comprehensively searched for publications in the following electronic databases from their inception through November 1, 2016: PubMed, Embase, Springer Link, CNKI (China National Knowledge Infrastructure), VIP (Chinese Scientific Journals Database), Wan-fang database, CBM (Chinese Biomedicine Database). The following general wording of the search terms were individually used or in combination: "traditional Chinese medicine", "Chinese herbal medicine", "herbal formula", "herbs", "alternative medicine", "Jianpi", "Yiqi", "chronic atrophic gastritis", "atrophic gastritis", "precancerous lesions of gastric cancer", "randomized controlled trial". No restriction for publication was placed on language. Electronic searches of omissive relevant studies were supplemented by the manual searches. As for the grey literature, we had retrieved them through trying our best to contact with manufacturers and pharmacists who produced herbal formulae based on Jianpi Yiqi Therapy.

Selection criteria
Studies meeting all of the following criteria were conducted in this meta-analysis. (1) Patients were definitely diagnosed CAG by gastroscopy and pathology. (2) The age of all participants was above 18 years old. (3) Experiment groups used Chinese herbal medicine of JYT, while control groups used conventional western medicine. (4) Treatment course was not less than 1 month. (5) The Jadad score was not less than 1.

Data abstraction and quality assessment
Two researchers independently conducted data extraction, including first author, publication year, whether by Hp infection, type of syndrome, sex, sample size, age, course of disease, intervention, duration, outcome measures, follow-up and side effects. We used the Cochrane Collaboration's risk of bias tool, conducted by Jadad scale in preliminary, to make evaluation of methodological quality. We could judge quality of literature from randomization, double blinding, and withdrawal or dropout in preliminary. The guidelines for assessment about Jadad scale are as follows: 1. Randomization: A method to generate the sequence of randomization will be regarded as appropriate if it allowed each study participant to have the same chance of receiving each intervention and the investigators could not predict which treatment was next. Methods of allocation using date of birth, date of admission, hospital numbers, or alternation should be not regarded as appropriate.
2. Double blinding: A study must be regarded as double blind if the word "double blind" is used. The method will be regarded as appropriate if it is stated that neither the person doing the assessments nor the study participant could identify the intervention being assessed, or if in the absence of such a statement the use of active placebos, identical placebos, or dummies is mentioned.
3. Withdrawals and dropouts: Participants who were included in the study but did not complete the observation period or who were not included in the analysis must be described. The number and the reasons for withdrawal in each group must be stated. If there were no withdrawals, it should be stated in the article. If there is no statement on withdrawals, this item must be given no points.
However, the final results of literature quality including the risk of bias evaluation were illustrated by the Cochrane tool. Disagreements were resolved after consulting with a third investigator.

Data synthesis and analysis
Review Manager 5.3 software was used for this statistical analysis. We calculated the pooled risk ratio (RR) to assess dichotomous data, while the standardized mean difference (SMD) or mean difference (MD) was used for continuous variable data, with both adopting 95% confidence intervals (CI). Heterogeneity was statistically assessed by using the χ 2 test and inconsistency index statistic (I 2 ) [7]. A model of random effect was conducted if substantial heterogeneity existed (I 2 >50% or P<0.05). We investigated possible sources of substantial heterogeneity using sensitivity analysis, which aimed at evaluating the robustness of emerging results through omitting one trial in turn. The number needed to treat (NNT) was computed as the reciprocal of the effective rate. Funnel plot was performed to evaluate if publication bias existed. In addition, the overall quality of grading evaluation for the review of evidence was calculated using GRADEprofiler version 3.6 which includes the elements of GRADE criteria such as study design, risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Description of studies
A total of 3163 relevant studies were obtained based on the search strategy and screened records. After further reviewing, 13 randomized clinical trials (RCTs) (N = 1119) satisfied all of the criteria and were included in this meta-analysis [8][9][10][11][12][13][14][15][16][17][18][19][20]. As for the grey literature, either their data did not meet our criteria or no response had been returned. In addition, although we also searched the EMBASE database which had many latest gray literatures such as meeting abstracts or latest literatures, none of them satisfied our criteria. The flow chart of literature search process was shown in Fig 1 (Flow chart of the process for literature retrieval). In addition, 13 studies were single centre trials and published in Chinese. Sample size was between 53 [16] and 130 [19]. The ages of participants were from 21 to 75 years old. The courses of disease ranged from 1 month to 26 years. The durations were from 4 weeks [13] to 24 weeks [14]. What's more, the interventions between experiment groups and control groups included the following: JYT versus western medicines (4 trials) [8,11,16,19] and JYT + western medicines versus western medicines (7 trials) [9,10,12,13,17,18,20]. The characteristics of the included studies were described in Table 1. The constituents of herbal formulae were listed in Table 2. Frequencies of usage and distribution in TCM were observed in Table 3. In addition, Chinese herbs classification can be found in Table 4. We can conclude from Tables 3 and 4 that in the all TCM categories, the proportion of invigorating spleen and reinforcing qi (Jianpi Yiqi) was    27.0%, which was the highest frequency among ten kinds of different herbs (Fig 2. TCM category rate).

Risk of bias assessment
A description of the assessment of methodological quality of the included trials can be observed in Table 5. Nine studies used a random number table [8,[12][13][14][16][17][18][19][20]. Two studies used a method of flipping a coin [10,11], while the other studies used the word "randomization", without any explanation of the random-allocation process [9,15]. Moreover, only one trial reported "single-blind" for the patients [19]. The remaining 12 studies did not mention blinding. Furthermore, none of included trials reported the concealment of allocation. In addition, taking the integrity of outcome data into account, only five trials provided the number of dropouts [8,10,11,16,18]. However, the missing data were not conducted by intention-totreat analysis. Because of the relative lack of specific information, it cannot be determined whether implementations were conducted adequately in the process of random sequence generation, blinding or allocation concealment, thus accounting for the high risk in the validity of this review (Fig 3. (a) Risk of bias summary. (b) Risk of bias graph).

Effects of the interventions: Primary outcomes
Comparison of effective rate. Among the included studies, 11 reported the effective rate based on the standards of the Guiding Principles for the Clinical Research of New TCM [21]: Cure, the clinical symptom disappeared; Markedly effective, the clinical symptom markedly improved; Effective, the clinical symptom improved; Ineffective, the clinical symptom did not improve even deteriorate. The effective rate was equal to (the numbers of patients whose clinical symptom improved after intervention divide total numbers of patients) × 100%. For example, the experiment group had 32 patients whose clinical symptom improved after intervention while the control group had 23 patients whose clinical symptom improved after intervention in the Chen et al. study. Moreover, the total numbers of patients in the experiment group were 36 while those in the control group were 33. Therefore, the effective rate in the experiment group was equal to (32 Ä 36) × 100% while those in the control group was equal to (23 Ä 33) × 100%. In addition, we did not perform a sensitivity analysis for good homogeneity in primary outcomes. JYT versus western medicine. Four of the thirteen trials including 371 patients (189 in the experiment groups vs. 182 in the control groups) with CAG reported the effective rate [8,11,16,19]. Although the forms of Jianpi Yiqi Theory were decoctions, the doses and methods of preparation and administration were different. Moreover, discrepancies in interventions among control groups were existed. Therefore, a random effect model was applied to estimate pooled effect size despite good homogeneity (χ 2 = 3.05, P = 0.38, I 2 = 2%) (Fig 4. Forest plot of effective rate (random effect model)). JYT showed statistically significant differences in the effective rate compared to western medicine (RR 1.41; 95% CI 1.27, 1.57; P < 0.00001) (Fig 4. Forest plot of effective rate (random effect model)). Potential publication bias was identified by asymmetrical funnel plot in Fig 5 (Funnel plot of effective rate). JYT + western medicine versus western medicine. Seven studies also evaluated the effective rate [9,10,12,13,17,18,20]: of 528 participants, 265 were assigned to the groups of JYT + western medicine, whereas 263 were assigned to the groups of western medicine. Because of the existence of discrepancies in interventions, pooled estimates were conducted by using a model of random effect in spite of no significant heterogeneity (χ 2 = 4.59, P = 0.60, I 2 = 0%) (Fig 6. Forest plot of effective rate (random effect model)). The effective rate of the experiment groups had potentially superior to that of the control groups (RR 1.27; 95% CI 1.17, 1.38; P < 0.00001) (Fig 6. Forest plot of effective rate (random effect model)). In addition, one trial reported NNT = 5 (95% CI 2.6, 5000.0) [20]. No evidence of symmetry was observed from the funnel plot in Fig 7 (Funnel plot of effective rate).

Improvement of TCM symptoms scores.
Of all the included studies, four reported the improvement of stomachache [8,12,13,15], three reported the improvement of stomach distention and belching [8,12,15], and three reported the improvement of fatigue [8,12,20]. Moreover, all of them were analyzed by a consensus [22] or semiquantitative scoring system. Although discrepancies in scoring system were existed, every study showed that JYT or combined with conventional western medicines can significantly improve these TCM symptoms caused by CAG.

The treating improvements in endoscopic and histopathologic test results
In the included trials, three reported the treating improvement in endoscopy [9,14,15] and two reported that in histopathology [14,15]. Because of few trials reporting the treating improvements in endoscopic and histopathologic test results, the two items were only qualitatively analyzed. But in the treating improvements in endoscopic and histopathologic test results, the treatment groups had potentially superior to the control groups.

Hp eradication rate
Although seven of thirteen studies described the situation of Hp infection [8,10,11,15,16,18,20], only one reported Hp eradication rate after treatment [15]. However, the study showed the experiment group had better efficacy than the control group in Hp eradication rate [15].

Adverse events
Of all the included studies, six reported no adverse reactions during JYT treatment [8, 10-12, 16, 18]. Moreover, the adverse effects of the experiment groups were no different from those of the control groups. GRADE evidence of quality GRADEprofiler software, adopted by WHO and the Cochrane collaboration, was used for rating quality of evidence and grading strength of recommendations for this systematic review. GRADE indicated that evidence quality was "Very low", which may be associated with high risk of bias within RCTs and the relatively small sample sizes of the included studies (Fig 10. GRADE quality grading evaluation) ( Table 6).

Discussion
This meta-analysis revealed that JYT or JYT + western medicine showed better effective rate than only western medicine and can significantly improve TCM symptoms caused by CAG such as stomachache, stomach distention, belching, fatigue, et al. However, weaknesses were identified in most trials using the Cochrane Collaboration's risk of bias tool, while the quality level of GRADE evidence classification indicated "Very low". No serious adverse events were found in the included studies. In addition, because of the Chinese herbal medicines of invigorating spleen and reinforcing qi (Jianpi Yiqi) at high proportion in the treatment of CAG (Fig 2. TCM category rate), it can suggest that JYT was possibly a promising therapy in treating CAG and provided practitioners with important reference value on clinical syndrome differentiations. The pathogenesis of CAG remains still controversial. Numerous mechanisms indicate that the development of CAG is associated with Hp infection, inflammation, gene, and autoimmune diseases [23][24][25][26][27]. Evidence for the efficacy of JYT for CAG was identified in modern pharmacological studies. Experimental data have showed that Xiangsha Liujunzi Decoction is an effective prescription for harmonizing the spleen and stomach, whose mechanisms are possibly associated with protecting GM, promoting gastric emptying, and inhibiting small intestine peristalsis too fast [28]. Another experiment has proved that Shenxiang Yangwei Powder can evidently promote the blood flow in the GM of rabbits, abate the injury of the GM of the white rats by alcohol, and also have preventive effect on lesions and secretive function of experimental CAG model [29]. In addition, clinical trials have demonstrated that Jianpi Huayu Jiedu therapy can relieve the degree of IM and glandular atrophy of GM, possibly by influencing the expression of the Cyclin E protein in the patients with precancerous lesion of GC, thus preventing the development of premalignant lesion of GC [30]. Weishu Capsule can significantly improve the clinical and pathological changes in the precancerous lesions of CAG via inducing and promoting effect of Weishu Capsule on the differentiation and maturity of IM cells and dysplasia cells, as well as inhibiting and correcting the abnormal proliferation of cells [31]. In a word, JYT may be a multitargeting management for the treatment of CAG, deserving to be studied further in vitro and in vivo.
Several potential limitations of this meta-analysis must be acknowledged. First, the methodological quality of included trials was generally poor. Because of no description of allocation concealment and double blind double dummy, this resulted in potentially high risk of selection bias and detection bias or performance bias. Furthermore, potential publication bias possibly existed because studies with favorable results were more likely to be published. Second, only one study mentioned follow-up and its period was half a year [8]. And considering atrophic gastritis as a chronic recurrent disease, its treatment sessions and follow-up periods should be long enough to evaluate long-term clinical effect of JYT. However, the courses of treatment in the included trials were all less than six months. The durations were too short to assess medium-or long-term efficacy and safety of JYT for CAG. Third, although the forms of JYT were decoctions, the doses and methods of preparation and administration were different. Moreover, discrepancies in interventions among control groups were not separately analyzed. These limitations may result in heterogeneous. Fourth, all of the included studies were based in China, not involving foreign countries. This geographically limited distribution could also result in sampling bias in CAG diagnosis. It was hard to validate that the efficacy of JYT for CAG screening applies to different populations worldwide. Fifth, most of the trials used the effective rate as the primary outcome. This will result in inability to quantitatively evaluate the efficacy of JYT for CAG. As for secondary outcomes, we qualitatively described them because of few studies reported. Therefore, the authenticity of the results awaited further proof. Sixth, although several literatures had reported that Hp eradication could possibly reduce GC risk [32][33][34][35], only one included study mentioned Hp eradication rate after treatment. This few recorded difference may potentially bring about unreliable and unbelievable results. Seventh, all of the included trials had single centers and small sample sizes, causing unstable results and inability to truly reflect general trends. Therefore, more rigorous designed RCTs are warranted to evaluate the efficacy of JYT for CAG. Furthermore, it remains urgent to make reporting quality of future research improvement strictly based on Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) or Consolidated Standards of Reporting Trials (CONSORT) statement.