TY - JOUR
T1 - Discriminative and predictive properties of multidimensional prognostic indices of chronic obstructive pulmonary disease
T2 - A validation study in Taiwanese patients
AU - Ou, Chih Ying
AU - Chen, Chiung Zuei
AU - Yu, Chun Hsiang
AU - Shiu, Chih Hui
AU - Hsiue, Tzuen Ren
PY - 2014/7
Y1 - 2014/7
N2 - Background and objective Chronic obstructive pulmonary disease (COPD) is a pulmonary disease with systemic involvement. Several multidimensional indices have been developed to predict long-term outcomes. However, these indices have not been compared and validated in Taiwanese patients with COPD. Methods A prospective, observational, hospital-based study was designed, and a total of 621 patients were recruited from May 2006 to December 2011. Patients followed at least 1 year were enrolled and 594 patients eligible for inclusion. Three prognostic indices - the ADO (age, dyspnoea and airflow obstruction), BODEx (body mass index, airflow obstruction, dyspnoea and exacerbations), and CPI (the COPD Prognostic Index) - were validated and the predictive power of each was analysed. Results The median follow-up of the 594 patients was 33 months (range 1-72 months), and the mortality rate was 19.2% (114 deaths). All indices were significantly predictive for all-cause mortality in our validation cohort. Furthermore, the C statistics of the three indices, indicating their predictive accuracy, were all >0.7 (area under the curve of the CPI 0.718, P < 0.001, ADO 0.702, P < 0.001, BODEx 0.702, P < 0.001). Conclusions ADO, BODEx and CPI scores are useful predictors of all-cause mortality with significantly discriminative properties in Taiwanese patients with COPD. We demonstrate that the ADO, BODEx and CPI scores could act as useful predictors of all-cause mortality with significantly discriminative properties in Taiwanese patients with COPD.
AB - Background and objective Chronic obstructive pulmonary disease (COPD) is a pulmonary disease with systemic involvement. Several multidimensional indices have been developed to predict long-term outcomes. However, these indices have not been compared and validated in Taiwanese patients with COPD. Methods A prospective, observational, hospital-based study was designed, and a total of 621 patients were recruited from May 2006 to December 2011. Patients followed at least 1 year were enrolled and 594 patients eligible for inclusion. Three prognostic indices - the ADO (age, dyspnoea and airflow obstruction), BODEx (body mass index, airflow obstruction, dyspnoea and exacerbations), and CPI (the COPD Prognostic Index) - were validated and the predictive power of each was analysed. Results The median follow-up of the 594 patients was 33 months (range 1-72 months), and the mortality rate was 19.2% (114 deaths). All indices were significantly predictive for all-cause mortality in our validation cohort. Furthermore, the C statistics of the three indices, indicating their predictive accuracy, were all >0.7 (area under the curve of the CPI 0.718, P < 0.001, ADO 0.702, P < 0.001, BODEx 0.702, P < 0.001). Conclusions ADO, BODEx and CPI scores are useful predictors of all-cause mortality with significantly discriminative properties in Taiwanese patients with COPD. We demonstrate that the ADO, BODEx and CPI scores could act as useful predictors of all-cause mortality with significantly discriminative properties in Taiwanese patients with COPD.
UR - http://www.scopus.com/inward/record.url?scp=84902971143&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84902971143&partnerID=8YFLogxK
U2 - 10.1111/resp.12313
DO - 10.1111/resp.12313
M3 - Article
C2 - 24797671
AN - SCOPUS:84902971143
SN - 1323-7799
VL - 19
SP - 694
EP - 699
JO - Respirology
JF - Respirology
IS - 5
ER -