M Mahmood1, J Wilkinson2, J Manning3, A L Herrick1. 1. a Centre for Musculoskeletal Research, Salford Royal NHS Foundation Trust , The University of Manchester, Manchester Academic Health Science Centre , Manchester , UK. 2. b Centre for Biostatistics, Institute of Population Health , The University of Manchester, Manchester Academic Health Science Centre , Manchester , UK. 3. c Rheumatology Directorate , Salford Royal NHS Foundation Trust , Salford , UK.
Abstract
OBJECTIVES: Systemic sclerosis (SSc)-related calcinosis is often painful and disabling. Our aim was to examine its clinical and serological associates, and whether it is possible to build a model to predict the presence of calcinosis. METHOD: This was a retrospective cross-sectional study. Clinical and demographic variables examined were: age, gender, disease subtype, SSc duration, previous intravenous prostanoid infusions, surgical debridement and/or amputation, autoantibody status (anticentromere and antitopoisomerase), pulmonary fibrosis, and pulmonary hypertension. Univariable logistic regression was used to investigate associations between demographic and clinical factors and the odds of clinical calcinosis, then multivariable regression to obtain adjusted odds ratios (ORs) and confidence intervals (CIs). RESULTS: A total of 317 patients (86% female, median age 60 years) were included: 94 (30%) had clinically apparent calcinosis. Age and gender were similar in those with and without calcinosis. Only surgical debridement (OR 3.55, 95% CI 1.71-7.53), anticentromere status (OR 2.32, 95% CI 1.27-4.32), and disease duration (OR 1.08, 95% CI 1.04-1.11) remained significant predictors after adjusting for other variables. In combination, the selected variables explained approximately 18% of the variation in the outcome but did not grant sufficient predictive power to discriminate between those with and without calcinosis [Nagelkerke's R(2) = 0.18, area under the receiver operating characteristic curve (AuROC) = 0.51, both adjusted for optimism]. CONCLUSIONS: History of surgical debridement, positive anticentromere antibody, and disease duration were independently associated with calcinosis. However, the low explanatory and discriminatory power of a multiple logistic regression model suggests there are other important predictors of calcinosis unaccounted for in this analysis.
OBJECTIVES:Systemic sclerosis (SSc)-related calcinosis is often painful and disabling. Our aim was to examine its clinical and serological associates, and whether it is possible to build a model to predict the presence of calcinosis. METHOD: This was a retrospective cross-sectional study. Clinical and demographic variables examined were: age, gender, disease subtype, SSc duration, previous intravenous prostanoid infusions, surgical debridement and/or amputation, autoantibody status (anticentromere and antitopoisomerase), pulmonary fibrosis, and pulmonary hypertension. Univariable logistic regression was used to investigate associations between demographic and clinical factors and the odds of clinical calcinosis, then multivariable regression to obtain adjusted odds ratios (ORs) and confidence intervals (CIs). RESULTS: A total of 317 patients (86% female, median age 60 years) were included: 94 (30%) had clinically apparent calcinosis. Age and gender were similar in those with and without calcinosis. Only surgical debridement (OR 3.55, 95% CI 1.71-7.53), anticentromere status (OR 2.32, 95% CI 1.27-4.32), and disease duration (OR 1.08, 95% CI 1.04-1.11) remained significant predictors after adjusting for other variables. In combination, the selected variables explained approximately 18% of the variation in the outcome but did not grant sufficient predictive power to discriminate between those with and without calcinosis [Nagelkerke's R(2) = 0.18, area under the receiver operating characteristic curve (AuROC) = 0.51, both adjusted for optimism]. CONCLUSIONS: History of surgical debridement, positive anticentromere antibody, and disease duration were independently associated with calcinosis. However, the low explanatory and discriminatory power of a multiple logistic regression model suggests there are other important predictors of calcinosis unaccounted for in this analysis.