Quality control for multiple breath washout tests in multicentre bronchiectasis studies: Experiences from the BRONCH-UK clinimetrics study.

INTRODUCTION: Multiple Breath Washout (MBW) to measure Lung Clearance Index (LCI) is increasingly being used as a secondary endpoint in multicentre bronchiectasis studies. LCI data quality control or "over-reading" is resource intensive and the impact is unclear. OBJECTIVES: To assess the proportion of MBW tests deemed unacceptable with over-reading, and to assess the change in LCI (number of turnovers), LCI coefficient of variation (CV%) and tidal volume (VT) CV% results after over-reading. METHODS: Data were analysed from 250 MBW tests (from 98 adult bronchiectasis patients) collected as part of the Bronch-UK Clinimetrics study in 5 UK centres. Each MBW test was over-read centrally using pre-defined criteria. MBW tests with <2 technically valid and repeatable trials were deemed unacceptable to include in analysis. In accepted tests, values for LCI, LCI CV% and VT CV% before and after over-reading, were compared. RESULTS: Insufficient data was collected in 10/250 tests. With over-reading, 30/240 (12%) were deemed unacceptable to include in analysis. In those accepted tests, overall the change in LCI, LCI CV% and VT CV% with over-reading was not statistically significant. When MBW new sites were compared to MBW expert sites, the change in LCI with over-reading was significantly greater in MBW new sites (p = 0.047). Data suggests that over-reading could be important up to at least 12 months post initiation of MBW activity. CONCLUSION: MBW over-reading was important in this study as 12% of tests were considered unacceptable. Over-reading improved test result accuracy in sites new to MBW.