Long-term intersession variability for single-breath diffusing capacity

Background: Characterizing long-term diffusing capacity (DL(CO)) variability is important in assessing quality control for DL(CO) equipment and patient management. Long-term DL(CO) variability has not been reported.

Objectives: It was the aim of this study to characterize long-term variability of DL(CO) in a cohort of biocontrols and to compare different methods of selecting a target value.

Methods: Longitudinal DL(CO) monitoring of biocontrols was performed as part of the inhaled insulin development program; 288 biocontrols were tested twice monthly for up to 5 years using a standardized technique. Variability, expressed either as percent change or DL(CO) units, was assessed using three different target values.

Results: The 90th percentile for mean intersession change in DL(CO) was between 10.9 and 15.8% (2.6-4.1 units) depending on the target value. Variability was lowest when the mean of all DL(CO) tests was used as the target value and highest when the baseline DL(CO) was used. The average of the first six DL(CO) tests provided an accurate estimate of the mean DL(CO) value. Using this target, the 90th percentile for mean intersession change was 12.3% and 3.0 units. Variability was stable over time and there were no meaningful associations between variability and demographic factors.

Conclusions: DL(CO) biocontrol deviations >12% or >3.0 units, from the average of the first six tests, indicate that the instrument is not within quality control limits and should be carefully evaluated before further patient testing.