Influence of carbon monoxide leaks on the measurement error of total haemoglobin mass

Total haemoglobin mass (Hb(mass)) can be assessed with low measurement error using carbon monoxide (CO) rebreathing. However, variability in measurement error of Hb(mass) has been reported across laboratories and it has previously been suggested that CO leaks contribute to this variability. As a result of employing a standardized leak monitoring procedure using two CO detectors, we were able to retrospectively examine the impact of CO leaks on Hb(mass) values from past test-retest studies in our laboratory using the optimized CO rebreathing method. Test-retest data were collected to determine measurement error, with subjects tested twice within 5 days. Test-retest data were placed into separate categories based on magnitude and duration of CO leak observed during one of the two tests. The No Leak category contained test-retest data in which no leak occurred during either test. The Minor Leak category contained test-retest data in which one of the tests had a CO leak of magnitude less than 30 ppm and less than 5 seconds duration, whereas the Major Leak category included test-retest data in which a leak greater than this magnitude or duration occurred. Measurement error was lowest in the No Leak category (1.9%; 95%CI: 1.6-2.3%; n = 56), approximately doubled in the Minor Leaks category (3.6%; 95%CI: 2.6-6.1%; n = 13), and dramatically increased in the Major Leaks category (9.3%; 95%CI: 6.3-17.6%; n = 10). We recommend careful monitoring of potential CO leaks using multiple detectors. To minimize measurement error, tests in which any CO leak is detected should be excluded.