Pulse Oximetry Is Unreliable in Patients on Veno-Venous Extracorporeal Membrane Oxygenation Caused by Unrecognized Carboxyhemoglobinemia

Abstract

Continuous bedside pulse oximetry (SpO2) is universally used to monitor oxygenation for patients supported on veno-venous extracorporeal membrane oxygenation (VV-ECMO). Yet, elevated carboxyhemoglobin (COHb), a known event in VV-ECMO, diminishes the reliability of SpO2. This retrospective cohort study aims to assess the accuracy of SpO2 compared with oxyhemoglobin (SaO2) and quantify COHb levels by co-oximetry in the VV-ECMO population. Forty patients on VV_ECMO from 2012 to 2017 underwent 1,119 simultaneous SaO2 and SpO2 measurements. Most patients were male (60%) with average age of 46 years. SpO2 overestimated SaO2 values by 2.35% at time of cannulation and 0.0061% for each additional hour on VV-ECMO (p < 0.0001). Twenty-nine (72.5%) patients developed elevated COHb (>3% of hemoglobin saturation) at least once during VV-ECMO support and 602 (40.2%) arterial blood gases yielded elevated COHb levels. Mean duration for ECMO with elevated COHb was 244 hours compared with 98 hours in patients without (p < 0.0048). Patients who developed COHb were younger (mean age 40 vs. 55 years, p < 0.024) and had single-site double-lumen cannulation (odds ratio = 4.5, p = 0.23). At time of cannulation, mean COHb was 2.18% and increased by 0.0054% for each additional hour (p < 0.0001). For every 1% increase in COHb, SaO2 decreased by 1.1% (p < 0.0001). During VV-ECMO, SpO2 often overestimates SaO2 by substantial margins. This is attributable to rising COHb levels proportional to duration on VV-ECMO. In this population where adequate oxygen delivery is often marginal, clinicians should be wary of the reliability of continuous pulse oximetry to assess oxygenation.