Genetic polymorphisms inside and outside the MHC improve prediction of AS radiographic severity in addition to clinical variables.

OBJECTIVE

The aim of this study was to analyse if single nucleotide polymorphisms (SNPs) inside and outside the MHC region might improve the prediction of radiographic severity in AS.

METHODS

A cross-sectional multi-centre study was performed including 473 Spanish AS patients previously diagnosed with AS following the Modified New York Criteria and with at least 10 years of follow-up from the first symptoms of AS. Clinical variables and 384 SNPs were analysed to predict radiographic severity [BASRI-total (BASRI-t) corrected for the duration of AS since first symptoms] using multivariate forward logistic regression. Predictive power was measured by the area under the receiver operating characteristic curve (AUC), specificity, sensitivity, positive predictive value (PPV) and negative predictive value (NPV).

RESULTS

The model with the best fit measured radiographic severity as the BASRI-t 60th percentile and combined eight variables: male gender, older age at disease onset and six SNPs at ADRB1 (rs1801253), NELL1 (rs8176785) and MHC (rs1634747, rs9270986, rs7451962 and rs241453) genes. The model predictive power was defined by AUC = 0.76 (95% CI 0.71, 0.80), being significantly better than the model with only clinical variables, AUC = 0.68 (95% CI 0.63, 0.73), P = 0.0004. Internal split-sample analysis proved the validation of the model. Patient genotype for SNPs outside the MHC region, inside the MHC region and clinical variables account for 26, 38 and 36%, respectively, of the explained variability on radiographic severity prediction.

CONCLUSION

Prediction of radiographic severity in AS based on clinical variables can be significantly improved by including SNPs both inside and outside the MHC region.