Abstract
MPORTANCE: Comorbidities are common in COPD, but quantifying their burden is difficult. Currently there is a COPD-specific comorbidity index to predict mortality and another to predict general quality of life. We sought to develop and validate a COPD-specific comorbidity score that reflects comorbidity burden on patient-centered outcomes.
MATERIALS AND METHODS: Using the COPDGene study (GOLD II-IV COPD), we developed comorbidity scores to describe patient-centered outcomes employing three techniques: 1) simple count, 2) weighted score, and 3) weighted score based upon statistical selection procedure. We tested associations, area under the Curve (AUC) and calibration statistics to validate scores internally with outcomes of respiratory disease-specific quality of life (St. George's Respiratory Questionnaire, SGRQ), six minute walk distance (6MWD), modified Medical Research Council (mMRC) dyspnea score and exacerbation risk, ultimately choosing one score for external validation in SPIROMICS.
RESULTS: Associations between comorbidities and all outcomes were comparable across the three scores. All scores added predictive ability to models including age, gender, race, current smoking status, pack-years smoked and FEV1 (p<0.001 for all comparisons). Area under the curve (AUC) was similar between all three scores across outcomes: SGRQ (range 0·7624-0·7676), MMRC (0·7590-0·7644), 6MWD (0·7531-0·7560) and exacerbation risk (0·6831-0·6919). Because of similar performance, the comorbidity count was used for external validation. In the SPIROMICS cohort, the comorbidity count performed well to predict SGRQ (AUC 0·7891), MMRC (AUC 0·7611), 6MWD (AUC 0·7086), and exacerbation risk (AUC 0·7341).
CONCLUSIONS: Quantifying comorbidity provides a more thorough understanding of the risk for patient-centered outcomes in COPD. A comorbidity count performs well to quantify comorbidity in a diverse population with COPD.
Putcha, Nirupama