Development and validation a nomogram to predict long-term mortality risks of PRISm and mild-to-moderate COPD based on NHANES 2007-2012.

Chronic Obstructive Pulmonary Disease (COPD) can be prevented in the pre-clinical and early stages. However, very limited prediction models of COPD focus on Preserved Ratio Impaired spirometry (PRISm) and early stages. To fill this gap, this study aimed to develop and validate a nomogram to predict long-term mortality risks of PRISm and early COPD. We obtained data of participants in the US National Health and Nutrition Examination Surveys 2007-2012 and the available mortality follow-up data from the date of survey participation to Dec 31, 2019. The study population (n = 1043) was randomly divided into training and validation datasets at a ratio of 7:3. The cox proportional hazards model was applied to select significant prognostic risk factors of COPD in the training dataset. Besides, the predictive power and clinical usage value were assessed by the area under time dependent receiver operating characteristic curve (time-dependent AUROC), calibration curves and decision curve analysis (DCA). Moreover, directed acyclic graph (DAG) was utilized to plot causal associations between risk factors and mortality. We developed an accurate and easy to use nomogram using six predictors (age, passive smoking, alkaline phosphotase, gamma glutamyl transferase, lactate dehydrogenase, potassium). The nomogram had satisfactory predictive performance, as the time-dependent AUROC with 95% confidence interval (CI) at 7.5 years was 0.78 (0.69-0.84) and 0.80 (0.67, 0.87) in the training and validation datasets, respectively. The calibration curves and DCA also showed that the nomogram had good clinical usage value. Compared with the low-risk groups, the Hazard Ratio in the high-risk group was 2.25 (95% CI 1.29-3.94) in the validation datasets, respectively. DAG shown that there had directly associations of passive smoking and lactate dehydrogenase with all-cause mortality. The nomogram has the potential to identify high-risk populations in the pre-clinical and early stages of COPD.