Association between GDF5 rs143383 genetic polymorphism and musculoskeletal degenerative diseases susceptibility: a meta-analysis

A computerized literature search was conducted in the PubMed, Scopus, Web of Science (WOS), Embase, and the Cochrane Library databases up to April 20, 2018. The search method of our study followed the terms such as: (“IDD” or “LDD” or “LDH” or “LBP” or “Intervertebral Disc Degeneration” or “OA” or “osteoarthritis”) and (“GDF5” or “rs143383” or “GDF5 + 104 T/C”) and (“polymorphisms” or “variants” or “variation” or “SNP”). Eligible articles that matched the inclusion criteria were included. Moreover, the references of articles were examined one by one to avoid missing any eligible studies. When the important data were not available, we tried to contact researchers of some articles.

A study that is eligible for inclusion must meet the following criteria: (1) case–control study or cohort study including both case and control groups, (2) detection of GDF5 polymorphisms and IDD or OA risk, (3) having an accessible genotype frequency for calculating an odds ratio (OR) or hazard ratio (HR) with 95% confidence interval (95% CI), (4) genotype frequencies in controls must conform to Hardy-Weinberg equilibrium (HWE). Whereas, reviews, case reports or serious, or similar works were all eliminated. We also eliminated the studies with genotype frequencies not in HWE [24].

Two investigators (Xin Huang, Weiyue Zhang) were assigned to assess the eligibility of all studies. And the relevant data for analysis were extracted on their own. Moreover, a third investigator (Zengwu Shao) resolved the disagreements when necessary. The important data were collected as follows: name of first author, year, countries, ethnicity, sample size, disease, sex, age, genotyping methods, and allele frequencies of GDF5 rs143383. The study quality was assessed in accordance with the Newcastle-Ottawa Scale (NOS). The study was considered high quality with the scores were ≥ 7.

The statistical data was analyzed by Stata version 14.0. Outcomes were calculated by odds ratios (ORs) and 95% confidence intervals (CIs). Genotype frequencies of GDF5 rs143383 polymorphism for HWE were calculated using the chi-square test, and P < 0.05 was regarded as significant disequilibrium. The chi-square test and the I² statistic were utilized to assess the between-study heterogeneity. If an I² value of < 50%, it was considered that no significant heterogeneity existed [25]. A random effects model was utilized when there was a significant heterogeneity. On the contrary, the fixed effects model was utilized. Moreover, we further made subgroup analyses to evaluate the source of heterogeneity. Begg’s and Egger’s methods were mainly utilized to assess publication bias. And sensitivity analyses were to evaluate the stability of major outcomes and possible source of heterogeneity.

The study search is shown in the flow diagram (Fig. 1). 108 relevant articles were collected during the databases search. Furthermore, 75 were eliminated during abstract review, and 33 for further review. During the full-text review, 18 articles were eliminated for the following reasons: seven were neither case–control study or cohort study, five were not associated with IDD or OA, four were not GDF5 polymorphisms on IDD or OA risk, two was not consistent with HWE. To sum up, 15 studies with 5915 cases and 12,252 controls were included in the present study.

The main features of each eligible study are summarized in details (Table 1). Among these eligible studies, three articles investigated value of GDF5 rs143383 in IDD risk in seven independent populations. In addition, twelve articles examined effects of GDF5 rs143383 on the risk of OA in fifteen independent populations. Ten in the included studies had been made among Asian populations, and twelve were in Caucasian populations. The years for publication ranged from 2007 to 2017. In all these articles, genotype frequencies in controls conformed to HWE (Table 2). According to NOS, the quality scores of all eligible articles ranged from 7 to 8, which indicated a good quality (Additional file 1: Table S1). Furthermore, the definitions of diseases, inclusions, and exclusions of patients in each study are also shown (Additional file 1: Table S2).

A total of 5915 patients and 12,252 controls were included in our study on GDF5 rs143383 polymorphism. There was a significant relationship between GDF5 rs143383 polymorphism and increased risk of musculoskeletal degenerative diseases under each genetic model (allele model: OR = 1.32, 95% CI 1.19–1.48, P = 0.000; homozygote model: OR = 1.80, 95%CI 1.49–2.16, P = 0.000; heterozygote model: OR = 1.37, 95%CI 1.21–1.55, P = 0.000; dominant model: OR = 1.56, 95%CI 1.39–1.75, P = 0.000; recessive model: OR = 1.39, 95%CI 1.20–1.60, P = 0.000) (Fig. 2 and Table 3). The heterogeneity of studies on this polymorphism was< 50%, under homozygote, heterozygote and dominant models.

Because of heterogeneity, we conducted stratified analyses based on different disease types and ethnicity. Stratified analyses based on disease type revealed a significant relationship between the GDF5 rs143383 polymorphism and increased risk of IDD under all genetic models studied (allele model: OR = 1.45, 95% CI 1.13–1.86, P = 0.004; homozygote model: OR = 2.33, 95%CI 1.55–3.51, P = 0.000; heterozygote model: OR = 1.84, 95%CI 1.34–2.52, P = 0.000; dominant model: OR = 2.11, 95%CI 1.57–2.86, P = 0.000; recessive model: OR = 1.44, 95%CI 1.02–2.04, P = 0.037). Additionally, rs143383 polymorphism was related with increased OA risk in all genetic models (allele model: OR = 1.28, 95% CI 1.28 1.14–1.44, P = 0.000; homozygote model: OR = 1.67, 95%CI 1.36–2.05, P = 0.000; heterozygote model: OR = 1.29, 95%CI 1.13–1.48, P = 0.000; dominant model: OR = 1.46, 95%CI 1.29–1.66, P = 0.000; recessive model: OR = 1.35, 95%CI 1.16–1.59, P = 0.000) (Table 3).

When stratified with ethnicity, the outcomes revealed that this polymorphism was statistically related with increased risk of musculoskeletal degenerative diseases in Asian populations under all genetic models studied (allele model: OR = 1.49, 95% CI 1.25–1.77, P = 0.000; homozygote model: OR = 2.28, 95%CI 1.71–3.05, P = 0.000; heterozygote model: OR = 1.49, 95%CI 1.25–1.78, P = 0.000; dominant model: OR = 1.82, 95%CI 1.54–2.16, P = 0.000; recessive model: OR = 1.55, 95%CI 1.25–1.92, P = 0.000). In the Caucasian subgroup, a significant relationship between rs143383 polymorphism and increased risk of musculoskeletal degenerative diseases under all genetic models was also observed in our study (allele model: OR = 1.18, 95% CI 1.09–1.28, P = 0.000; homozygote model: OR = 1.46, 95%CI 1.23–1.74, P = 0.000; heterozygote model: OR = 1.27, 95%CI 1.07–1.51, P = 0.006; dominant model: OR = 1.35 95%CI 1.15–1.58, P = 0.000; recessive model: OR = 1.21, 95%CI 1.03–1.42, P = 0.019) (Table 3).

No obvious publication bias was shown in the funnel plot. In addition, there was no obvious publication bias according to Begg’s test (P = 0.338) and Egger’s test (P = 0.246). Therefore, we could exclude the possibility of publication bias. The sensitivity analysis revealed that the main outcomes of our study did not alter greatly when deleting studies one by one (Additional file 2).