Association between 20q12 rs13041247 polymorphism and risk of nonsyndromic cleft lip with or without cleft palate: a meta-analysis

Embase, Web of Science, PubMed, the China National Knowledge Internet (CNKI) and the China Wanfang database were searched for articles published through 25 June 2019. Search terms used were as follows: (“nonsyndromic cleft lip with or without cleft palate” or “cleft palate” or “cleft lip” or “orofacial clefts” or “oral cleft” or “CL” or “CP” or “NSCL/P”) and (“MAFB” or “20q12” or “v-maf musculoaponeurotic fibrosarcoma oncogene homolog B”) and (“polymorphism” or “allele” or “gene” or “SNP”). No restrictions were imposed upon the language in which articles were published.

In order to be included in this meta-analysis, studies had to meet the following inclusion criteria: (1) Studies were case-control studies; (2) Studies were focused on NSCL/P; (3) Polymorphisms analyzed in the study included rs13041247 at the 20q12 locus; (4) Studies provided sufficient data necessary for the calculation of odds ratios (ORs) and 95% confidence intervals (CIs); (5) All necessary data was either available or was obtained within two attempts to contact the study authors. When multiple articles included overlapping patient cohorts, only the study with the most comprehensive information was included in the present meta-analysis (i.e. the study with the largest population or the most complete dataset). In addition, studies meeting the following criteria were excluded from this analysis: (1) Animals studies; (2) Reviews, letters, or abstracts lacking original data; (3) Case-reports; (4) Studies focused only on other SNPs or which lacked a control group. No language restrictions were imposed on these studies.

Two authors (Liheng Huang and Yangzhan Ou), independently reviewed identified studies in order to exclude those which were either clearly irrelevant or which were duplicates, after which a full text review was performed to identify studies meeting the inclusion criteria for this analysis. Any discrepancies were resolved via discussion with the third author (Yunpu He).The following pieces of data were extracted from these studies: Name of the first author, year of publication, country, population ethnicities, control source, study design, samples size, control Hardy-Weinberg equilibrium (HWE) p-values, genotyping methodology, and case/control genotype distributions (Table 1) [5‐17]. In addition, the Newcastle-Ottawa scale was used by two authors (Yunpu He and Shijie Tang) to independently evaluate included study quality as recommended previously [18]. In total, 8 of the included studies reported on polymorphisms in Chinese and East Asian populations, 2 focused on Caucasian populations Germany and Brazil [15, 16], and 1 each focused on Nigerian [17], Brazilian [7], Mayan Mesoamerican [8], Indian [5] ethnic populations. For those subgroups for which more than one article was available, ethnicity-based subgroup analyses were performed.

STATA 12.0 was used to obtain crude ORs and 95% CIs for each of the included articles as a means of assessing the relationship between the rs13041247 SNP and NSCL/P risk. Chi-squared tests were used to assess the HWE in the control group, revealing it not to differ significantly from HWE (P > 0.05). Ethnicity-based subgroup analyses were also conducted. Five different genetic models were used to examine the rs13041247 and NSCL/P risk according to ORs and 95% CIs: an allele model (C vs T), a homozygote model (CC vs TT), a heterozygote model (CC vs CT), a dominant model (CT + TT vs CC), and a recessive model (CC + CT vs TT). Z-tests were used to assess the significance of pooled ORs, P < 0.05 as the significance threshold. The I² statistic was used to assess heterogeneity among studies, with I² > 50% being consistent with significant heterogeneity, leading to the use of a random effect model, and I² < 50% leading to the use of a fixed effect model. Sensitivity analyses were conducted by iteratively omitting individual studies from the overall analysis, while funnel plots were used to examine the risk of publication bias.

The study selection process for the present meta-analysis is detailed in Fig. 1, with the characteristics of included studies shown in Table 1. In total, 13 case-control studies were included in the present meta-analysis, incorporating a total of 4914 cases and 5981 controls in whom rs13041247 genotyping had been performed [5‐17]. These studies included diverse populations from a range of ethnicities, including East Asian, Caucasian, India, Mayan, Brazilian, and African cohorts. The quality of these studies was assessed using the Newcastle-Ottawa scale (NOS), revealing all of these studies to be of high quality with scores > 7 out of a possible 9 (Table 2).

Table 3 highlights the results of this meta-analysis of the relationship between rs13041247 and NSCL/P risk. Ultimately, this analysis revealed no significant association between rs13041247 and NSCL/P risk in an East Asian population (C vs T: OR = 0.847, 95% CI = 0.702–1.021; CC vs TT: OR = 0.725, 95% CI = 0.494–1.063; CC vs CT: OR = 0.837, 95% CI = 0.657–1.067; CT + TT vs CC: OR = 1.265, 95% CI = 0.951–1.684; CC + CT vs TT: OR = 0.805, 95% CI = 0.630–1.029) or in a Caucasian population (C vs T: OR = 0.936, 95% CI = 0.786–1.114; CC vs TT: OR = 0.988, 95% CI = 0.674–1.446; CC vs CT: OR = 1.197, 95% CI = 0.816–1.757; CT + TT vs CC: OR = 0.918, 95% CI = 0.639–1.318; CC + CT vs TT: OR = 0.855, 95% CI = 0.677–1.081). However, when the overall combined patient/control populations were analyzed, this analysis revealed that the C allele, the CT genotype, and the CC + CT model to be associated with significantly reduced NSCL/P risk (C vs T: OR = 0.897, 95% CI: 0.723–1.114, P = 0.048; CT vs TT: OR = 0.839, 95% CI: 0.734–0.959, P = 0.01; CC + CT vs TT: OR = 0.824, 95% CI: 0.701–0.968, P = 0.019). Significantly heterogeneity was detected in all models for both the overall and Asian analysis groups (I² > 50%), whereas the heterogeneity in the Caucasian subgroup analysis was not significant. The results of this meta-analysis of the relationship between rs13041247 and NSCL/P risk for the allele model (C vs T), the heterozygote model (CC vs CT), the homozygote model (CC vs TT), the recessive model (CC + CT vs TT), and the dominant model (CT + TT vs CC) are shown in Figs. 2, 3, and 4. No significant changes in the study outcomes were detected in a sensitivity analysis, and no evidence of publication bias was detected based upon Egger’s test (East Asian P = 0.253, Caucasian P = 0.239, and Overall population P = 0.124). Similarly, no funnel plot asymmetry was detected (Figs. 5 and 6 for East Asian and Overall, respectively).