OBJECTIVE

A number of studies have reported an association of GDF5 with osteoarthritis (OA) but have produced some divergent findings and their interpretation may not be straightforward.

METHODS

We investigated the association between GDF5 and OA using meta-analytic techniques, combining all published data up to Nov 2014. 16 independent samples from 11 research teams contributed data on SNP rs143383 (located in the 5'-UTR of GDF5) and knee, hip, and hand OA. The total number of cases and controls for this marker was 7,965 and 12,747 for knee OA, 6,363 and 9,727 for hip OA, and 4,335 and 5,991 for hand OA, respectively. The ORs for each OA phenotype were synthesized using random-effects models or fixed-effects models depending on the test of between-study heterogeneity.

RESULTS

Using a random-effect model, a significant difference was identified between patients with knee OA and controls for the T-allele of rs143383 (Subtotal OR = 1.18, 95 % CI=1.10-1.27, P=1.84 × 10(-6)). For hand OA, a moderate association was also observed (Subtotal OR = 1.09, 95 % CI = 1.02-1.16, P = 0.01) for SNP rs143383 in the combined population. However, non-statistically significant summary OR of hip OA was found in both combined studies (Subtotal OR = 1.22, 95 % CI = 0.97-1.53, P = 0.09) and European studies (Subtotal OR = 1.16, 95 % CI = 0.91-1.48, P = 0.23).

CONCLUSIONS

Our results demonstrate that SNP rs143383 of GDF5 is a compelling risk factor for both knee and hand OA, and provide further support for GDF5 in the etiology of OA. Further efforts to identify functional variants of GDF5 in in vitro and in vivo will be required.