Prognostic Accuracy of Three COPD Classification Systems in Relation to Long-Term Mortality of COPD Patients: A Prospective Multicenter Study.
Marek PlutinskyKristian BratMichal SvobodaJaromir ZatloukalPatrice PopelkovaVladimir KoblizekPublished in: Lung (2019)
Recent research showed group B patients express higher mortality compared to group C patients when GOLD A-D grouping is used. We aimed to compare the prognostic accuracy of three GOLD classification systems, I-IV ("pre-2011"), A-D ("2011-2016") and A-D ("2017-present") in relation to mortality, exacerbation risk, quality of life (QoL) assessment and specific treatments use in a real-life COPD cohort. We used the data of 720 patients from the Czech Multicenter Research Database of COPD. Four-year mortality and time-to-exacerbation using the GOLD "pre-2011", "2011-2016" and "2017-present" classification schemes were assessed. Moreover, distribution of specific treatments use and QoL measures were analyzed. The GOLD I-IV classification system showed gradual increase in 4-year mortality across the stages (GOLD II 18.8%, III 28.5%, IV 38.7%) (p = 0.001). Using the A-D "2011-2016" classification scheme, group C patients had lower mortality (16.7%) than group B (18.7%) (p = 0.009). The A-D "2017-present" classification showed higher mortality in group B (25.5%) compared to group C (20%) (p = 0.05). For additional outcomes, the GOLD I-IV scheme showed highest match between the calculated 4-year exacerbation risk and QoL measures and GOLD stage/grouping. In terms of specific treatment distributions, various patterns for each GOLD classification system were observed with best match of GOLD "2017-present" system to the layout of GOLD groups and categories. We conclude the GOLD I-IV classification system had the highest accuracy related to mortality, QoL measures and exacerbation risk prediction, while the A-D "2017-present" scheme was most accurate within severity of symptoms prediction reflected also by more frequent specific treatments use.
Keyphrases
- end stage renal disease
- chronic obstructive pulmonary disease
- newly diagnosed
- ejection fraction
- cardiovascular events
- chronic kidney disease
- deep learning
- machine learning
- peritoneal dialysis
- risk factors
- prognostic factors
- emergency department
- silver nanoparticles
- cardiovascular disease
- skeletal muscle
- intensive care unit
- mass spectrometry
- metabolic syndrome
- depressive symptoms
- artificial intelligence
- cross sectional
- mechanical ventilation
- extracorporeal membrane oxygenation
- patient reported
- big data
- data analysis