Improved survival prediction from lung function data in a large population sample

Area of Science:

• Pulmonary Medicine
• Epidemiology
• Biostatistics

Background:

• Lung function impairment is commonly assessed using percent predicted forced expiratory volume in 1s (FEV(1)PP).
• FEV(1)PP retains age, height, and sex biases, potentially limiting its accuracy in predicting mortality.
• Alternative methods for expressing FEV(1) may offer improved prognostic value.

Purpose of the Study:

• To evaluate alternative methods of expressing FEV(1) for predicting all-cause and airway-related lung disease mortality.
• To compare the predictive performance of FEV(1)/ht(2) against FEV(1)PP and FEV(1)SR.

Main Methods:

• Utilized data from the Copenhagen City Heart Study, including 13,900 subjects with 25 years of follow-up.
• Employed Cox regression models to assess predictors of survival.
• Compared FEV(1) expressed as raw data, FEV(1)/ht(n), FEV(1)PP, and FEV(1)SR.

Main Results:

• Quintiles of FEV(1)/ht(2) demonstrated superior prediction of all-cause mortality compared to FEV(1)PP and FEV(1)SR in multivariate models.
• The hazard ratio for the worst quintile of FEV(1)/ht(2) for all-cause mortality was 2.8, versus 2.0 for FEV(1)PP.
• Categorized FEV(1) impairment showed significantly higher hazard ratios for all-cause (10 vs. 5) and airway disease mortality (2044 vs. 194) with FEV(1)/ht(2) compared to FEV(1)PP.

Conclusions:

• FEV(1)/ht(2) is a more effective method than FEV(1)PP for predicting survival in a general population.
• This height-standardized expression of FEV(1) impairment better reflects subsequent mortality risk.
• The findings suggest a revised approach to assessing lung function for prognostic purposes.