Prediction of Mortality Using Different COPD Risk Assessments : A 12-Year Follow-Up

Purpose: A multidimensional approach in the risk assessment of chronic obstructive pulmonary disease (COPD) is preferable. The aim of this study is to compare the prognostic ability for mortality by different COPD assessment systems; spirometric staging, classification by GOLD 2011, GOLD 2017, the age, dyspnea, obstruction (ADO) and the dyspnea, obstruction, smoking, exacerbation (DOSE) indices. Patients and Methods: A total of 490 patients diagnosed with COPD were recruited from primary and secondary care in central Sweden in 2005. The cohort was followed until 2017. Data for categorization using the different assessment systems were obtained through questionnaire data from 2005 and medical record reviews between 2000 and 2003. Kaplan-Meier survival analyses and Cox proportional hazard models were used to assess mortality risk. Receiver operating characteristic curves estimated areas under the curve (AUC) to evaluate each assessment systems' ability to predict mortality. Results: By the end of follow-up, 49% of the patients were deceased. The mortality rate was higher for patients categorized as stage 3-4, GOLD D in both GOLD classifications and those with a DOSE score above 4 and ADO score above 8. The ADO index was most accurate for predicting mortality, AUC 0.79 (95% CI 0.75-0.83) for all-cause mortality and 0.80 (95% CI 0.75-0.85) for respiratory mortality. The AUC values for stages 1-4, GOLD 2011, GOLD 2017 and DOSE index were 0.73, 0.66, 0.63 and 0.69, respectively, for allcause mortality. Conclusion: All of the risk assessment systems predict mortality. The ADO index was in this study the best predictor and could be a helpful tool in COPD risk assessment