Large-scale international validation of the ADO index in subjects with COPD: an individual subject data analysis of 10 cohorts

Abstract

Background: Little evidence on the validity of simple and widely applicable tools to predict mortality in patients with chronic obstructive pulmonary disease (COPD) exists.

Objective: To conduct a large international study to validate the ADO index that uses age, dyspnoea and FEV(1) to predict 3-year mortality and to update it in order to make prediction of mortality in COPD patients as generalisable as possible.

Design: Individual subject data analysis of 10 European and American cohorts (n=13 914).

Setting: Population-based, primary, secondary and tertiary care.

Patients: COPD GOLD stages I-IV.

Measurements: We validated the original ADO index. We then obtained an updated ADO index in half of our cohorts to improve its predictive accuracy, which in turn was validated comprehensively in the remaining cohorts using discrimination, calibration and decision curve analysis and a number of sensitivity analyses.

Results: 1350 (9.7%) of all subjects with COPD (60% male, mean age 61 years, mean FEV(1) 66% predicted) had died at 3 years. The original ADO index showed high discrimination but poor calibration (p<0.001 for difference between predicted and observed risk). The updated ADO index (scores from 0 to 14) preserved excellent discrimination (area under curve 0.81, 95% CI 0.80 to 0.82) but showed much improved calibration with predicted 3-year risks from 0.7% (95% CI 0.6% to 0.9%, score of 0) to 64.5% (61.2% to 67.7%, score of 14). The ADO index showed higher net benefit in subjects at low-to-moderate risk of 3-year mortality than FEV(1) alone.

Interpretation: The updated 15-point ADO index accurately predicts 3-year mortality across the COPD severity spectrum and can be used to inform patients about their prognosis, clinical trial study design or benefit harm assessment of medical interventions.

Figures

Figure 1

Update and validation of the ADO index in 13 914 subjects with chronic obstructive pulmonary disease (COPD). The upper part of the figure shows the predictive performance of the updated ADO index in 10 221 subjects with COPD from the Copenhagen City Heart Study, Lung Health Study, National Emphysema Treatment Trial, PLATINO and the Phenotype and Course of COPD Study. The calibration plot shows the predicted and observed risks for 10 equally sized group with increasing risk of 3-year mortality. The discrimination plot shows the area under the curve. The lower part of the figure shows the predictive performance of the updated ADO index in the validation cohort with 3693 subjects from the Cardiovascular Health Study, Basque COPD study, Jackson Heart Study, Barmelweid Study and the Quality of Life of Chronic Obstructive Pulmonary Disease Study (SEPOC).

Figure 2

Accuracy of four strategies to classify subjects into risk categories. The upper part (A) of the figure shows the net benefit of six strategies to classify subjects with chronic obstructive pulmonary disease. The higher the values for net benefit the more patients are correctly classified. Two strategies do not use any predictors but assume that all patients would be above or below a risk threshold. The four other strategies use the ADO index, age, dyspnoea or FEV1 and associated risks of 3-year mortality to classify patients. Net benefit is defined as the difference between the proportion of correctly classified subjects and the proportion of subjects classified incorrectly to be at or above a risk threshold (eg, 5% risk). The line for considering all patients to be above a risk threshold crosses that line for considering all patients to be below a risk threshold at the death rate observed (9.7%). The lower part of the graph (B) shows that, for example at a threshold of 5% mortality risk, using the ADO index would result in a net benefit similar to the reduction of 33 incorrectly classified patients per 100 subjects compared to considering all patients to be above a 5% mortality risk. Using age, dyspnoea or FEV1 only would reduce it by only 24, 10 and 18 per 100 subjects, respectively. The graph is restricted to subjects at low to moderate risk for 3-year mortality (<20%) where most uncertainty about the balance between benefits and harms from treatment may exist.