Pulse Oximetry Carboxyhemoglobin Saturation Overestimates Carbon Monoxide Rebreathing Derived Measures Of Total Hemoglobin Mass In Healthy Adults

PURPOSE: Carbon monoxide (CO) rebreathing is commonly used to measure total hemoglobin mass (Hbmass). This test requires measurement of carboxyhemoglobin saturation (COHb) from blood samples collected before and serially after the CO rebreathing procedures. However, this method is invasive and expensive. Pulse CO-oximetry may be a portable, cost-effective and non-invasive tool for estimating carboxyhemoglobin saturation (SpCO). Two studies have compared SpCO derived measurement of total Hbmass against COHb-based analyses, with one study showing acceptable and the other showing poor agreement. This study tested the hypothesis that SpCO derived measurement of total Hbmass has poor agreement with COHb-based analyses. METHODS: Following 20 min of supine rest, fourteen healthy adults (2 women, age: 28 ± 5 y, body surface area: 2.0 ± 0.3 m2) completed a modified 2 min CO rebreathing test that involved the delivery of 1.0 mL CO per kg body mass into a closed circuit spirometer connected to a 5 L anesthesia bag filled with 100% O2 and a CO2 scrubber. Capillary blood was taken pre- and 6 and 8 min post-CO rebreathing commencement and analyzed for COHb (OSM3, Radiometer, sensitive to the nearest tenth). SpCO was recorded for 1 min before and every 30 s during and following rebreathing using pulse CO-oximetry (Rad-57, Masimo Corp., sensitive to the nearest whole-number). Data are presented as mean ± SD. RESULTS: Following CO rebreathing the peak increase in COHb (6.0 ± 1.2%) was greater than the peak increase in SpCO (5 ± 1%, P = 0.008), which was consistent when COHb was rounded to the nearest whole-number (P = 0.004). Total Hbmass was lower when calculated via COHb (951 ± 301 g) versus SpCO (1145 ± 320 g, P = 0.041). Total Hbmass calculated via COHb was moderately, positively correlated with SpCO derived values (r = 0.470, P = 0.090). Bland-Altman plot analyses indicated wide 95% limits of agreement (-821 to 434 g) and a bias towards overestimating total Hbmass with SpCO (Bias: +194 ± 320 g), although 13 of 14 (93%) individual differences fell within the limits of agreement. CONCLUSIONS: Pulse oximetry derived SpCO has poor agreement and overestimates CO rebreathing measures of total Hbmass in healthy adults compared to when COHb is measured using blood gas analyses. Supported by Office of Naval Research award N00014-20-1-2593.