Duration and frequency of spirometry needed to accurately reflect longitudinal change of FEV1 in COPD

Abstract

<b>Background:</b> The slope of FEV1 decline over time is commonly used to reflect the rate of disease progression in COPD. The frequency and duration of spirometric testing needed to accurately report the true slope is unknown. <b>Objective:</b> To define the minimum frequency and follow-up duration needed to accurately describe the rate of FEV1 decline among patients with moderate to very severe COPD enrolled in the 4-year UPLIFT trial. <b>Method:</b> Using a linear mixed effects model, we performed a post-hoc analysis of the rate of FEV1 decline measured in 4412 subjects to estimate slopes for different iterations of biannual or annual testing over variable duration up to 4 years. Slope was assessed for both pre- and post-bronchodilator (BD) values. All models were compared to a reference of biannual spirometry for the 4-year trial. <b>Results:</b> Overall rate of post-BD FEV1 decline measured biannually for 4 years (44.6 ml; 95% CI 42.5-46.6) did not differ significantly from annual spirometry over the same time (43.7 ml; 95% CI 41.3-46.1) or biannual spirometry over the first 2 years (44.3 ml; 95% CI 41.1-47.5). Accuracy was preserved for at least three values each 18 months apart (41.9 ml; 95% CI 39.4-44.4). Models based on less than three values or duration less than 2 years were characterized by relative underestimation of the slope. Findings were consistent for pre- and post-BD values. <b>Conclusion:</b> Accurate measurement of annualized decline in FEV1 is achieved by at least three measurements as far as 18 months apart, with no further accuracy obtained by annual or biannual measurements beyond two years. This has important implications for clinical monitoring and therapeutic trial design in COPD.