Gut microbiota signatures in colorectal cancer: Evidence synthesis from umbrella review and probiotic intervention studies

Abstract

BACKGROUND

Aberrant alterations in the gut microbiota have been widely reported in colorectal cancer (CRC), but taxon-level associations remain heterogeneous in direction, magnitude, and reproducibility across studies.

AIM

To conducted an umbrella review with a second-order meta-analysis to identify reproducible CRC-associated taxa and to summarize gut microbiota changes after probiotic or synbiotic interventions.

METHODS

We included eligible meta-analyses of CRC-associated gut microbiota and evaluated them using AMSTAR 2 (a measurement tool to assess systematic reviews). After standardizing effect sizes and deriving missing standard errors, we performed a second-order meta-analysis to estimate pooled effect sizes, 95% confidence intervals (CIs), P values, and heterogeneity. Candidate taxa were identified using predefined criteria, and species-level findings were mapped to genera. We also added original CRC studies on probiotics and synbiotics, assessed bias using RoB 2, and combined the data via standardized mean differences (SMDs), wherever appropriate. Studies without extractable variance estimates were included in the qualitative or percentage-based synthesis.

RESULTS

Eleven meta-analyses of 3569 CRC patients and 3722 healthy controls were included. Frequently reported genera, such as Roseburia, Prevotella, Porphyromonas, Bacteroides, and Streptococcus, showed inconsistent association directions; therefore, reporting frequency alone could not reliably identify stable CRC-associated taxa. Seventeen genera met the predefined significance and reporting-frequency criteria and 8 retained significance in leave-one-out analyses. After heterogeneity-based filtering, Granulicatella and Peptoniphilus remained the most stable genera. Six probiotic/synbiotic intervention studies of 323 participants were included. The quality of the studies was generally low, with some missing outcome data and selective reporting. Pooled analysis revealed that Lactobacillus (SMD = 1.49, 95%CI: 0.67-2.31) and Bifidobacterium (SMD = 1.81, 95%CI: 0.60-3.03) both increased significantly, mainly reflecting supplementation and other taxa remains exploratory. After intervention, single-study SMDs and percentage-based analyses suggested possible increases in Faecalibacterium, Eubacterium, Bacillus, Staphylococcus, and Parabacteroides, and possible decreases in Fusobacterium, Candida, Enterobacteriaceae, Pseudomonas, Porphyromonas, Alloprevotella, Erysipelatoclostridium, Peptostreptococcus, Bacteroides, and Clostridium.

CONCLUSION

This study identifies reproducible CRC-associated microbial candidates and highlights the need to evaluate cross-study stability beyond reporting frequency. Probiotic/synbiotic interventions may alter selected microbial taxa in CRC patients, but larger shotgun metagenomics-based trials are needed to determine whether taxonomic shifts translate into ecological or metabolic restoration.

Keywords: Colorectal cancer; Gut microbiota; Umbrella review; Meta-analysis; Probiotics; Synbiotics; Microbial biomarkers

Core Tip: This study synthesizes gut microbiota signatures in colorectal cancer (CRC) through two complementary components: An umbrella review with second-order meta-analysis of CRC-associated taxa, and a synthesis of probiotic or synbiotic intervention evidence. The key innovation is distinguishing frequently reported taxa from truly stable taxa across studies. The findings also suggest that probiotic or synbiotic interventions may partially reshape CRC-related dysbiosis, but higher-quality clinical evidence is still needed.