CDH1 Gene Promoter Polymorphism and the Risk of Gastric Cancer: A Meta-Analysis

Abstract

Objective

CDH1 polymorphism has been suggested to be associated with susceptibility to gastric cancer (GC). However, relevant findings remain inconsistent across studies. This meta‑analysis was performed to evaluate the relationship between CDH1 −160C/A and −347G/GA polymorphisms and GC risk.

Methods

Relevant studies were searched in PubMed, Web of Science, Google Scholar, Cochrane Library, Embase, CNKI, and Wanfang databases up to December 20, 2025. Pooled OR and 95% (95% CI) were calculated using fixed‑ or random‑effects models.

Results

A total of 25 articles were included, involving 28 studies (4,836 cases and 6,187 controls) for the −160C/A polymorphism and 8 studies (1,201 cases and 1,611 controls) for the −347G/GA polymorphism. Overall, significant associations were identified between CDH1 polymorphisms and GC risk: −160C/A (recessive model: OR=1.28, 95% CI: 1.03-1.58; AA vs. CC: OR=1.28, 95% CI: 1.01-1.61); -347G/GA (G/GA vs. GG: OR=1.23, 95% CI: 1.04-1.44). In stratified analyses by histological type, the CDH1 −347G/GA polymorphism was associated with an elevated risk of diffuse‑type GC (G/GA vs. GG: OR=1.21, 95% CI: 1.01-1.62).

Conclusions

This meta-analysis indicates that the two CDH1 promoter polymorphisms may contribute to GC susceptibility. Further large-scale and well-designed studies are needed to verify these results.

INTRODUCTION

Gastric cancer (GC) is one of the most prevalent malignant tumors of the digestive system worldwide. In 2022, there were more than 968,000 new GC cases and approximately 660,000 deaths globally, making it the fifth most common cancer and the fifth leading cause of cancer-related mortality¹. The exact etiology and pathogenesis of GC have not been fully elucidated. Epidemiological evidence shows that GC development involves multiple risk factors and complex interactions between genetic predisposition and environmental exposures².

E-cadherin, a calcium-dependent transmembrane glycoprotein encoded by the CDH1 gene located at chromosome 16q22.1, is critical for preserving epithelial cell adhesion and tissue architecture^{3, 4}. It acts as a classical tumor suppressor. Partial or complete loss of E-cadherin expression is frequently observed in various human carcinomas, often resulting from transcriptional repression⁵.

Two common polymorphisms in the CDH1 promoter region, -160C>A (rs16260) and -347G/GA (rs5030625), can directly regulate transcriptional activity and may influence individual susceptibility to multiple cancers, including GC^{6, 7}. Although numerous studies have explored the links between these CDH1 promoter variants and GC risk^{8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37}, the results remain controversial. To clarify these associations, we conducted a comprehensive meta-analysis to quantitatively assess the relationship between CDH1 -160C/A and -347G/GA polymorphisms and GC susceptibility.

MATERIALS and METHODS

Study Registration and Guidelines

This systematic review and meta-analysis was registered in the NIHR PROSPERO database (Registration no. CRD420261283279). The study was performed and reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines³⁸.

Search Strategy

A systematic literature search was conducted in PubMed, Web of Science, Google Scholar, theCochrane Library, Embase, CNKI, and Wanfang databases for articles published up to December 20, 2025. Combinations of keywords and subject terms were used: “E-cadherin” OR “CDH1” AND “genetic polymorphism” OR “polymorphism” OR “variant” AND “gastric cancer” OR “GC” OR “stomach cancer” OR “gastric tumor” OR “gastric neoplasm”. Only human studies were included, with no language restrictions.

Reference lists of retrieved articles and reviews were manually screened to identify additional eligible studies. Grey literature, including unpublished data, theses, dissertations, and conference abstracts, was also searched manually.

Inclusion and Exclusion Criteria

Studies were included if they met the following criteria: 1) Used clear diagnostic criteria for GC; 2) Investigated the association between GC and CDH1 -160C/A or -347G/GA polymorphism; 3) Adopted a case-control design; 4) Provided sufficient genotype frequency data for cases and controls. Studies were excluded if they: 1) were case-only studies, case reports, or reviews; 2) lacked raw genotype data; or 3) were duplicate publications.

Data Extraction

Two investigators (Wang J and Ye CC) independently extracted the following information: first author, publication year, country, ethnicity, control source, genotyping method, sample sizes, and Hardy-Weinberg equilibrium (HWE) P value in controls. Discrepancies were resolved by discussion until consensus was reached.

Quality Assessment

Two reviewers independently assessed methodological quality using the Newcastle-Ottawa Scale (NOS). This scale awards a maximum of 9 stars across three domains: selection of study groups, comparability of groups, and ascertainment of exposure or outcome. Studies with a score ≥6 stars were regarded as high-quality.

Statistical Analysis

Pooled ORs and 95% CIs were calculated under four genetic models: dominant, recessive, heterozygote comparison and homozygote comparison. Heterogeneity was tested using the χ²‑based Q test and quantified by the I² statistic³⁹. A random‑effects model was used if significant heterogeneity existed (P ≤ 0.05 or I² > 50%)⁴⁰; otherwise, a fixed‑effects model was applied⁴¹.

Meta‑regression was performed to explore sources of heterogeneity. HWE was tested using the χ² test. Subgroup analyses were conducted by ethnicity, control source, genotyping method, and tumor location. Sensitivity analysis was performed to assess the stability of the results. Publication bias was evaluated using Begg’s funnel plot and Egger’s test^{42, 43}. All statistical analyses were performed using RevMan 5.4 and Stata 12.0.

RESULTS

Study Characteristics

The initial database search retrieved 319 potentially relevant articles. Scanning the titles and abstracts resulted in the exclusion of 289 articles. Subsequently, we excluded two studies^{33, 34, 35, 36} because they did not present detailed genotyping information. We also excluded one study³⁷ because it included the overlapped data with those included in the analysis³². Finally, 25 full-text articles were included in this meta-analysis. Because two articles included different GC groups, we treated them separately in this meta-analysis. The detailed screening process is shown in Figure 1. 28 studies from 25 articles were included^{8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32}. For the -160C/A polymorphism, 28 studies with 4836 cases and 6187 controls were analyzed. For the -347G/GA polymorphism, 8 studies with 1201 cases and 1611 controls were included. The characteristics of the eligible studies are listed in Table 1. All included studies were considered high quality with NOS scores ≥6 stars. Genotype distributions in control groups were consistent with HWE in most studies, except for two^{11, 23} and three studies^{8, 18, 32} seperately.

Figure 1.

Figure 1. Flow diagram of literature screening and study selection process.

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

Study	Year	Country	Ethnicity	Source of control	Genotyping method	Genotype (case/control)			P HWE
						WT Ho	Ht	VR Ho
-160C/A	-	-	-	-	-	CC	CA	AA	-
Akçakaya et al.8	2021	Turkey	Caucasian	HB	PCR-RFLP	13/15	50/87	15/11	0
Al-Moundhr et al.9	2010	Oman	Asian	PB	Sequencing	93/93	60/65	21/8	0.428
Aslan et al.10	2024	Turkey	Caucasian	HB	TaqMan	53/54	43/42	9/9	0.838
Borges et al.11	2010	Brazil	Mixed	HB	Sequencing	27/32	20/14	11/5	0.09
Bustos-Carpinteyro et al.12	2016	Mexico	Mixed	HB	PCR	27/21	14/24	4/3	0.257
Cattaneo et al.13	2007	Italy	Caucasian	PB	PCR-RFLP	50/139	51/89	6/18	0.476
Chu et al.14	2014	China	Asian	HB	Sequencing	48/84	44/44	15/6	0.938
Corso et al.15	2009	Italy	Caucasian	PB	PCR-RFLP	206/185	163/185	43/38	0.395
Humar et al.16	2002	Italy	Caucasian	HB	PCR-RFLP	17/40	26/27	10/3	0.555
Jenab et al.17	2008	Mixed	Caucasian	HB	TaqMan	119/451	101/408	25/90	0.869
Kuraoka et al.18	2003	Japan	Asian	HB	PCR-RFLP	61/32	34/52	11/6	0.013
Lu et al.19	2005	China	Asian	PB	PCR-RFLP	119/152	75/91	12/18	0.391
Medina-Franco et al.20	2007	Mexico	Mixed	HB	SSCP	15/44	16/30	8/4	0.699
Park et al.21	2003	Korea	Asian	HB	SSCP	186/85	92/55	14/6	0.429
Pharoah-C et al.22	2002	Canada	Caucasian	HB	PCR-RFLP	58/43	76/44	14/6	0.231
Pharoah-G et al.22	2002	Germany	Caucasian	HB	PCR-RFLP/TaqMan	61/22	58/15	13/5	0.345
Pharoah-P et al.22	2002	Portugal	Caucasian	HB	SSCP	62/153	80/151	11/27	0.223
Shekarriz et al.23	2021	Iran	Aisan	HB	PCR-RFLP	49/51	39/39	9/5	0.479
Shin et al.24	2004	Korea	Asian	HB	PCR-RFLP	21/110	6/31	1/1	0.454
Song et al.25	2005	China	Asian	PB	PCR-DHPLC	58/55	38/41	6/5	0.448
Wu et al.26	2002	China	Asian	HB	PCR-RFLP	95/83	102/94	4/19	0.302
Sun et al.27	2025	China	Asian	HB	Sequecom	284/279	171/170	29/22	0.545
Vishteh et al.28	2022	Iran	Asian	HB	MS-PCR	19/30	22/9	7/2	0.256
Yamada et al.29	2007	Japan	Asian	HB	PCR-RFLP	93/187	51/93	4/12	0.919
Zhan et al.30	2012	China	Asian	HB	PCR-LDR	219/196	116/137	26/21	0.647
Zhang et al.31	2008	China	Asian	PB	PCR-RFLP	53/123	38/65	5/8	0.872
Zhang B et al.31	2008	China	Asian	HB	PCR-RFLP	365/403	173/194	34/28	0.453
Zhang XF et al.32	2008	China	Asian	HB	PCR-RFLP	170/228	62/96	7/19	0.042
-347G/GA	-	-	-	-	-	GG	G/GA	GA/GA	-
Akçakaya et al.8	2021	Turkey	Caucasian	HB	PCR-RFLP	60/89	16/24	2/0	0.207
Borges et al.11	2010	Brazil	Mxied	HB	Sequencing	29/37	11/11	6/5	0.011
Bustos-Carpinteyro et al.12	2016	Mexico	Mixed	HB	PCR	20/22	21/19	4/3	0.683
Shekarriz et al.23	2021	Iran	Asian	HB	PCR-RFLP	50/58	39/17	8/20	0
Shin et al.24	2004	Korea	Asian	HB	PCR-RFLP	16/103	11/39	1/0	0.058
Zhang et al.31	2008	China	Asian	PB	PCR-RFLP	60/103	30/78	6/15	0.965
Zhang B et al.31	2008	China	Asian	HB	PCR-RFLP	296/340	246/237	30/48	0.456
Zhang XF et al.32	2008	China	Asian	HB	PCR-RFLP	117/194	109/132	13/17	0.364

HWE, Hardy-Weinberg equilibrium; P_HWE was calculated by goodness-of fit χ²-test, P_HWE < 0.05 was considered statistically significant; PB: population-based; HB: hospital-based; PCR-RFLP: polymerase chain reaction-restriction fragment length polymorphism; SSCP, single-strand conformation polymorphism; PCR-LDR, polymerase chain reaction-ligation detection reaction; DHPLC, Denaturing high-performance liquid chromatography; Ht, heterozygote; VR Ho, variant homozygote; WT Ho, wide-type homozygote.

Quantitative Synthesis

For the -160C/A polymorphism, significant associations with GC risk were found in the recessive model (OR=1.28, 95 % CI: 1.03-1.58) and in the AA vs. CC comparison (OR=1.28, 95% CI: 1.01-1.61) (Table 2 and Figure 2).

Table 2. Summary of ORs of the CDH1 -160C/A polymorphism and GC risk.

SNP	N	Dominant model			Recessive model			CA vs. CC			AA vs. CC
-	-	OR (95% CI)	Pa	I2	OR (95% CI)	Pa	I2	OR (95% CI)	Pa	I2	OR (95% CI)	Pa	I2
Total	28	1.07 (0.94,1.21)	0.0004	54%	1.28 (1.03,1.58)	0.03	37%	1.02(0.90,1.16)	0.003	47%	1.28 (1.01,1.61)	0.009	43%
Ethnicity	-	-	-	-	-	-	-	-	-	-	-	-	-
Asian	16	1.01 (0.86,1.18)	0.003	56%	1.22 (0.89,1.68)	0.02	47%	0.97 (0.83,1.13)	0.01	50%	1.20 (0.86,1.69)	0.008	52%
Caucasian	9	1.14 (0.93,1.41)	0.07	45%	1.18 (0.90,1.55)	0.32	14%	1.12 (0.90,1.39)	0.07	44%	1.17 (0.88,1.54)	0.35	10%
Mixed	3	1.29 (0.54,3.06)	0.03	72%	2.59 (1.23,5.44)	0.47	0	1.07 (0.47,2.45)	0.06	64%	2.71 (1.10,6.68)	0.26	25%
Source of control	-	-	-	-	-	-	-		-	-	-	-	-
PB	6	1.05 (0.87,1.27)	0.27	22%	1.18 (0.84,1.65)	0.36	8%	1.03 (0.83,1.29)	0.16	37%	1.15 (0.84,1.58)	0.45	0
HB	22	1.07 (0.92,1.26)	0.0002	60%	1.32 (1.01,1.71)	0.02	44%	1.02 (0.88,1.19)	0.003	51%	1.32 (0.99,1.77)	0.004	51%
Genotyping method	-	-	-	-	-	-	-	-	-	-	-	-	-
PCR-RFLP	14	1.01 (0.85,1.20)	0.01	52%	1.12 (0.79,1.58)	0.02	50%	1.00 (0.84,1.19)	0.02	51%	1.08 (0.75,1.54)	0.02	49%
SSCP	3	1.17 (0.73,1.88)	0.07	63%	1.51 (0.61,3.73)	0.07	62%	1.09 (0.71,1.67)	0.12	52%	1.63 (0.62,4.24)	0.07	63%
Sequencing	3	1.57 (1.01,2,43)	0.15	47%	2.78 (1.59,4.85)	0.82	0	1.32 (0.83,2.09)	0.16	45%	3.09 (1.74,5.48)	0.71	0
Others	8	1.00 (0.80,1.27)	0.04	52%	1.18 (0.90,1.55)	0.93	0	0.97 (0.76,1.23)	0.05	50%	1.16 (0.88,1.53)	0.79	0
Location	-	-	-	-	-	-	-	-	-	-	-	-	-
Cardia	4	0.85 (0.61,1.17)	0.32	14%	1.13 (0.63,2.04)	0.42	0	0.90(0.62,1.29)	0.35	4%	1.07(0.58,1.98)	0.04	0
Non-cardia	4	0.89 (0.73,1.07)	0.92	0	0.77 (0.34,1.78)	0.02	76%	0.87(0.70,1.07)	0.78	0	0.73(0.33,1.64)	0.03	73%
Histotype	-	-	-	-	-	-	-	-	-	-	-	-	-
Intestinal	11	0.97 (0.84,1.11)	0.72	0	1.27 (0.96,1.68)	0.45	0	0.93(0.80,1.08)	0.62	0	1.17(0.88,1.54)	0.51	0
Diffused	12	1.00 (0.75,1.33)	0.0002	69%	1.12 (0.54,2.34)	<0.00001	77%	1.04(0.78,1.39)	0.002	63%	1.01(0.74,1.39)	0.06	44%
Gender	-	-	-	-	-	-	-	-	-	-	-	-	-
Male	3	0.85 (0.69,1.06)	0.6	0	1.46 (0.94,2.26)	0.8	0	0.78(0.61,1.01)	0.3	8%	1.29(0.82,2.04)	0.63	0
Female	3	0.90 (0.70,1.14)	0.18	41%	0.70 (0.47,1.06)	0.92	0	0.92(0.70,1.21)	0.68	0	0.65(0.42,1.01)	0.83	0
Lymphatic metastasis	-	-	-	-	-	-	-	-	-	-	-	-	-
Yes	4	0.60 (0.35,1.03)	0.02	71%	-	-	-	-	-	-	-	-	-
No	4	0.79 (0.55,1.12)	0.93	0	-	-	-	-	-	-	-	-	-
Smoking	-	-	-	-	-	-	-	-	-	-	-	-	-
Yes	2	1.13 (0.78,1.64)	0.56	0	-	-	-	-	-	-	-	-	-
No	2	0.75 (0.49,1.14)	0.91	0	-	-	-	-	-	-	-	-	-

N: number of studies; ^aTest for heterogeneity. ^# Random-effect model was used when the P for heterogeneity test was ≤0.05, otherwise the fixed-effect model was used. CI: confidence interval, OR: odds ratio.

Figure 2.

Figure 2. Forest plots of meta-analysis for the association between CDH1 -160C/A polymorphism and gastric cancer susceptibility. (A)Recessive model, (B) AA vs CC comparison.

Subgroup analyses stratified by ethnicity showed comparable outcomes in mixed-ethnicity populations. For the recessive model, the pooled OR was 2.59 (95% CI: 1.23-5.44), and for the AA vs. CC comparison, the OR was 2.71 (95%CI: 1.10-6.68). When stratified by control source, a statistically significant association was identified in the hospital-based (HB) subgroup under the recessive model (OR=1.32, 95% CI: 1.01-1.71) (Table 2).

Following stratification by genotyping methodology, significant associations were observed in studies that adopted the sequencing technique. The OR values were 1.57 (95% CI: 1.01-2.43) for the dominant model, 2.78 (95% CI: 1.59-4.85) for the recessive model, and 3.09 (95% CI: 1.74-5.48) for the AA vs. CC contrast (Table 2).

Furthermore, we evaluated potential links between the −160C/A polymorphism and various clinical features. No significant associations were observed for any of the assessed parameters, including sex, smoking status, histological type, tumor location, and lymph node metastasis (Table 2).

For the -347G/GA polymorphism, a significant association with GC risk was identified in the G/GA vs GG model (OR=1.23, 95% CI: 1.04-1.44) (Figure 3), but not in other genetic models (dominant model: OR=1.16, 95% CI: 1.00-1.35; recessive model: OR=0.79, 95 %CI: 0.58-1.07; GA/GA vs GG model: OR=0.87, 95 % CI: 0.63-1.19). In stratified analysis by histological subtype of gastric cancer, a significant association was observed between the CDH1 -347G/GA polymorphism and the risk of diffuse-type GC in the G/GA vs GG comparison (OR=1.21, 95% CI: 1.01-1.62) (Figure 4, Table 3).

Figure 3.

Figure 3. Forest plot of meta-analysis for the association between CDH1 -347G/GA polymorphism and gastric cancer risk (G/GA vs GG model).

Figure 4.

Figure 4. Subgroup analysis by histological type for the association between CDH1 -347G/GA polymorphism and diffuse-type gastric cancer risk (G/GA vs GG model).

Table 3. Summary of ORs for CDH1 -347G/GA polymorphism and GC risk.

-	N	Dominant model			Recessive model		-	G/GA vs. GG		-	GA/GA vs. GG		-
-	-	OR (95% CI)	Pa	I2	OR (95% CI)	Pa	I2	OR (95% CI)	Pa	I2	OR (95% CI)	Pa	I2
Total	8	1.16 (1.00,1.35)	0.31	-	0.79 (0.58,1.079	0.12	-	1,23 (1.04,1.44)	0.11	-	0.87 (0.63,1.19)	0.18	-
Ethnicity	-	-	-	-	-	-	-	-	-	-	-	-	-
Asian	5	1.17 (0.89,1.54)	0.09	-	0.74 (0.43,1.27)	0.10	-	1.30 (0.92,1.84)	0.02	-	0.78 (0.56,1.10)	0.15	-
Others	3	1.22 (0.78,1.91)	0.94	-	1.63 (0.64,4.15)	0.59	-	1.12 (0.69,1.819	0.90	-	1.86 (0.73,4.71)	0.62	-
Histotype	-	-	-	-	-	-	-	-	-	-	-	-	-
intestinal	5	1.01 (0.80,1.27)	0.29	-	0.65 (0.41,1.03)	0.48	-	1.10 (0.86,1.41)	0.11	-	0.68 (0.42,1.10)	0.61	-
diffused	5	1.16 (0.93,1.45)	0.62	-	0.66 (0.43,1.02)	0.48	-	1.21 (1.01,1.62)	0.16	-	0.75 (0.48,1.18)	0.68	-
Gender	-	-	-	-	-	-	-	-	-	-	-	-	-
Male	2	0.78 (0.32,1.92)	0.007	-	0.83 (0.50,1.41)	0.72	-	0.77 (0.28,2.13)	0.004	-	0.86 (0.51,1.47)	0.83	-
Female	2	1.03 (0.71,1.48)	0.53	-	0.50 (0.22,1.12)	0.36	-	1.14 (0.78,1.67)	0.63	-	0.53 (0.23,1.21)	0.32	-
Lymphatic metastasis	-	-	-	-
Yes	3	1.21 (0.98,1.49)	0.57
No	3	1.15 (0.86, 1.55)	0.68
Smoking	-	-	-
Yes	2	0.93 (0.43, 2.05)	0.07
No	2	0.30 (0.02, 3.87)	0.03

N: number of studies, ^a Test for heterogeneity, CI: confidence interval, OR: odds ratio.

In subgroup analyses stratified by ethnicity, no significant association was detected among Asian, Caucasian, or mixed-ethnicity populations under any genetic model. Additionally, no significant relationship was found between the -347G/GA polymorphism and the evaluated clinical characteristics, including sex, smoking status, and lymph node metastasis (Table 3).

Meta-regression and Sensitivity Analysis

Marked between-study heterogeneity was detected for the -160C/A polymorphism in the overall analysis across all genetic models (dominant model: I²=54%, P=0.004; recessive model: I²=37%, p=0.03; CA vs. CC: I²=47%, p=0.003; AA vs. CC: I²=43%, p=0.009). Meta-regression was conducted to identify potential sources of heterogeneity. The findings indicated that ethnicity and control source were the main factors contributing to heterogeneity (p<0.05). A leave-one-out sensitivity analysis was performed to further test the robustness of the pooled results. The significance and direction of the estimates remained unchanged after omitting any single study, thereby verifying the stability and reliability of the results.

Publication Bias

The risk of publication bias was evaluated using Begg’s funnel plot and Egger’s regression test. Visual inspection of the funnel plots showed no obvious asymmetry (Figure 5). Consistently, Egger’s test revealed no significant publication bias for either polymorphism. For the -160C/A polymorphism, the results were: dominant model (p-value =0.064), recessive model (p=0.233), CA vs. CC (p=0.121), and AA vs. CC (p=0.122). For the -347G/GA polymorphism, p-values were: dominant model (p=0.622); recessive model (p=0.056); G/GA vs. GG (p=0.733); and GA/GA vs. GG (p=0.05).

Figure 5.

Begg’s funnel plot for assessment of publication bias under the recessive model of CDH1-160C/A polymorphism.

DISCUSSION

E‑cadherin plays a key role in maintaining intercellular adhesion, preserving cell polarity, participating in cell-cell signaling, and regulating tissue morphogenesis^{44, 45}. It has been reported that approximately 33% of diffuse GCs harbor CDH1 mutations, and 50%-80% exhibit reduced expression or loss of E‑cadherin⁴⁶. The C>A polymorphism at the -160 locus of the CDH1 gene can modulate its transcriptional activity. The protective A allele is associated with higher E-cadherin expression. Conversely, reduced CDH1 mRNA transcription and the consequent absence of E-cadherin on the cell membrane can disrupt cellular adhesion, thereby promoting intestinal mucosal carcinogenesis and accelerating metastasis⁴⁷.

Regarding genetic susceptibility, Jiang et al.⁴⁸ conducted a meta‑analysis to evaluate the association between CDH1 polymorphisms and GC risk. Interestingly, they did not confirm a significant link for the CDH1 –160C/A polymorphism, a finding consistent with another earlier meta‑analysis⁴⁹. In contrast, a subsequent meta‑analysis by Jiang et al. reported that CDH1 polymorphisms were significantly associated with GC risk⁵⁰.

In the current meta-analysis, we conducted a thorough literature search across multiple academic databases and included more eligible studies. This approach expanded the overall sample size and enhanced the accuracy of risk effect estimation. The combined results demonstrated significant associations between CDH1 gene polymorphisms and GC susceptibility across specific genetic comparison models. Further subgroup analyses indicated that the -160C/A polymorphism was significantly associated with an increased risk of GC in the sequencing-based genotyping and HB control subgroups. The -347G/GA polymorphism in CDH1 was associated with a increased risk of diffuse GC. These inconsistent findings across comparisons may be explained by variations in genetic backgrounds, environmental exposures, and methodological limitations—including small sample sizes or insufficient adjustment for confounding factors.

To assess potential differential effects based on clinicopathological features, further association analyses were conducted. The combined results did not indicate a significant association between CDH1 polymorphisms and clinicopathological characteristics. However, given the limited number of studies reporting relevant data, these findings should be interpreted with caution, and further research with larger sample sizes is warranted.

Heterogeneity is a common issue in meta-analyses of genetic association studies⁵¹. In the present analysis, significant heterogeneity was observed across all genetic models in the overall comparisons. Stratified analyses revealed that heterogeneity persisted within the Asian population, PCR-RFLP and SSCP genotyping subgroups, and HB populations. In contrast, heterogeneity was substantially reduced or absent in analyses restricted to Caucasian populations, sequencing-based genotyping subgroups, and population-based studies. These patterns suggest that ethnic background, control selection criteria, and genotyping methodology may contribute to the moderate heterogeneity observed. The Sensitivity analysis, performed by iteratively excluding individual studies, indicated that the pooled odds ratio remained stable, which supports the robustness of the meta-analytic findings. Furthermore, no significant publication bias was detected, which lends credibility to the results.

Study Limitations

Several limitations should be acknowledged. First, some potentially relevant studies were excluded due to incomplete data or publication restrictions. Second, the number of studies included in some subgroup analyses was relatively small, which may reduce statistical power. Third, our results were based on unadjusted effect estimates, and the influence of potential confounding factors, such as family history and environmental exposures, was not fully considered.

This meta-analysis suggests that CDH1 -160C/A and -347G/GA polymorphisms may contribute to genetic susceptibility to GC. Further large-scale, well-designed, and comprehensive studies are needed to confirm these findings.

Ethics

Ethics Committee Approval: Since this study is a meta-analysis, ethical approval is not required.

Informed Consent: Since this study is a meta-analysis study, patient consent is not required.