Non-constant carboxyhemoglobin elimination kinetics during normobaric oxygen therapy in acute carbon monoxide poisoning

During normobaric oxygen (NBO) therapy for acute carbon monoxide (CO) poisoning, carboxyhemoglobin (COHb) clearance is often interpreted using a single constant half-life. However, physiological and experimental evidence suggests that CO distribution beyond hemoglobin may cause time-dependent elimination. We evaluated whether serial COHb measurements during high-flow NBO show population-level attenuation of elimination over time. We retrospectively studied patients with acute CO poisoning treated with continuous high-flow NBO. Patients with at least two COHb measurements and documented uninterrupted oxygen therapy were included. Apparent elimination rate constants and half-lives were calculated using two-point log-linear kinetics. Time-dependent changes were assessed using early (< 2 h) versus late (≥ 2 h) interval classification and piecewise mixed-effects models. A total of 159 patients were included. Two-point interval analysis showed faster COHb elimination in early than late intervals. The elimination rate constant decreased as time from rescue increased. In the primary mixed-effects model, the early-phase slope was − 0.725 h⁻¹ and the post-2-h slope change was 0.525 h⁻¹, corresponding to a late-phase slope of − 0.200 h⁻¹. The piecewise model showed better fit than the mono-exponential model, and the post-2-h attenuation remained consistent after physiological and exposure-source adjustment and across sensitivity analyses. COHb elimination during NBO showed population-level time-dependent attenuation. Although individual biphasic trajectories cannot be established from two-point measurements, COHb clearance may be inadequately described by a single constant half-life. Serial COHb values should be interpreted in relation to sampling time after rescue.