The Effectiveness And Mechanism Of Tonifying Kidney And Spleen Method On Preventing And Treatment Of Myelosuppression Induced By Adjuvant Chemotherapy Of Colorectal Cancer Ⅵ

3.3 Quality evaluation of included literature

Among the included literature, the high risk of bias mainly exists in the implementation of blinding of subjects and experimental personnel, allocation concealment and incomplete data results. (1) Generation of random sequence: 24 studies mentioned the method of random sequence generation and were rated as "low risk"; the rest were "unclear risk of bias". (2) Allocation concealment: 2 studies used central randomization or envelope method for concealment and were rated as "low risk"; 7 studies used alternating allocation, medical record number allocation, etc. and were rated as "high risk"; the rest were "unclear risk of bias". (3) Blinding of subjects and researchers: All studies were blank-controlled and could not be blinded, so they were rated as "high risk". (4) Blinding of outcome assessors: None of the studies mentioned the use of blinding for outcome assessors. However, the outcome indicators of this study were all laboratory measurements, or were determined based on laboratory measurements as raw data. The outcome assessors judged that the outcomes would not be affected by the lack of blinding, so it was rated as "low risk"; (5) Incomplete outcome reporting: 8 studies reported missing data, the number of missing data between groups and the reasons were unbalanced, or inappropriate methods were used to handle missing data, so they were rated as "high risk", 24 studies had no missing data, and were rated as "low risk", and the rest were rated as "unclear risk of bias". (6) Selective reporting: None of the studies mentioned the relevant information of the study protocol registration, and were rated as "unclear risk of bias". (7) Other biases: All studies reported the comparability of baseline data, which was rated as "low risk". (Figure 1-2, Figure 1-3)

Figure 1-2 Evaluation of specific projects included in the study ROB

3.4 Meta-analysis results

3.4.1 Main outcome indicators

(1) Incidence of leukopenia

Nine studies 125-33 reported the incidence of leukopenia, 16 studies [34-49 reported the incidence of grade I/II/I/IV leukopenia, and another study [501 reported the incidence of grade I-II/III-IV leukopenia. After merging the data, it was found that the heterogeneity between the studies was not large (P=29%). The results of the meta-analysis showed that oral Chinese medicine can reduce the risk of leukopenia caused by adjuvant chemotherapy for colorectal cancer (RR=0.67, 95%CI[0.61,0.75], P<0.00001). (Figure 1-4)

Note: The forest plot is not included in this text-based table. The Risk Ratio column shows the numerical values, but the graphical representation is not possible in this format.

Figure 1-4 Meta-analysis of the incidence of leukopenia in traditional Chinese medicine + chemotherapy vs chemotherapy

The inverted funnel plot shows a symmetrical trend, which basically supports the results of the meta-analysis. (Figure 1-5)

Figure 1-5 Inverted funnel plot of meta-analysis of Chinese medicine + chemotherapy vs chemotherapy leukopenia

(2) Incidence of hemoglobin reduction

Four studies125.2728.51 reported the incidence of hemoglobin reduction, and 10 studies134.35.3740.42.4.47.48I reported the incidence of grade I/II/I/IV hemoglobin reduction. The combined data found that the heterogeneity between the studies was small (I²=0%): Meta-analysis results showed that oral Chinese medicine can reduce the risk of hemoglobin reduction caused by adjuvant chemotherapy for colorectal cancer (RR=0.75, 95%CI[0.63,0.89], P=0.0009). (Figure 1-6)

| Study or Subgroup | Experimental | | Control | | Weight | Risk Ratio | Risk Ratio |