Carboxyhemglobin and Drainage Pressure During Venovenous Extracorporeal Membrane Oxygenation.

Abstract

To the Editor: We read with interest the article1 recently published by Nisar et al. on ASAIO Journal. The authors describe the frequent development of carboxyhemoglobinemia in patients treated with venovenous (V-V) extracorporeal membrane oxygenation (ECMO). As highlighted in the article, high carboxyhemoglobin levels cause significant discrepancies between pulse oximetry and actual oxyhemoglobin levels. This may lead to erroneous interpretation of patient oxygenation when the latter is assessed noninvasively. Carboxyhemoglobinemia has been described since 19492 as a marker for hemolysis. Hemolysis during V-V ECMO occurs in up to 18% of patients.3 The principal causes of mechanical hemolysis during ECMO support include circuit components (oxygenator, pump,4 cannula size) as well as patient’s characteristics.5 Centrifugal pump could be a major source of hemolysis when rotating in unsuitable conditions. High rotation speed (e.g., >3000 rpm) may result in high negative pressures at the pump head, thus promoting hydrodynamic cavitation and a high level of hemolysis.6 We performed a retrospective analysis of prospectively collected data of our historical cohort of acute respiratory distress syndrome patients treated with V-V ECMO, to compare our findings with Nisar et al. We included adult patients admitted to our ECMO unit (Rianimazione Generale, ASST Monza, Italy) between 2014 and 2020. We excluded patients with veno-venoarterial ECMO and COVID-19 patients, who may present coagulopathy as a part of the underlying disease.7 Arterial blood gas (ABG) analysis was performed with the RAPIDPoint 500e System (Siemens Healthcare, Erlangen, Germany). Elevated carboxyhemoglobin was defined according to the authors’ threshold (i.e., >3%).1 Eighty-four patients (39% females, 53 ± 10 years) were included in our analysis. ECMO configuration was femoro-femoral (86% of patients) or femoro-jugular. Median duration of ECMO support was 20 days, ranging from 3 to 83 days. Average blow flow was 3.5 ± 0.5 (2,520 ± 287 rpm), whereas mean drainage pressure was −42 ± 14 mm Hg. Overall, mean carboxyhemoglobin level was 1.1 ± 0.8%. Carboxyhemoglobin level at ECMO connection was 0.6 ± 0.5%%, and its average increase was 0.011% for each ECMO day. Among 1,653 ABG analyses, carboxyhemoglobin was >3% in only 23 samples (1.4%). A carboxyhemoglobin higher than 3% occurred at least once in eight (9.5%) patients. In our study, we found a lower incidence of hemolysis compared with Nisar et al.1 This finding may have different explanations. First, at our institution, we are used to carefully monitor the drainage pressure and adjusting the ECMO blood flow to avoid drainage pressures lower than −60/−70 mm Hg. This may limit hemolysis due to hydrodynamic cavitation. Moreover, a minor difference due to technical reason (i.e., two different blood gas analyzers) cannot be excluded. In absence of recirculation, a BF of 3–4 L/min usually allows achieving a safe oxygenation in adult patients.8 If adequate drainage cannulae are used (i.e., 21–25 Fr), this blood flow range can be usually achieved with drainage pressure of −30 to −60 mm Hg.9,10 Therefore, we hypothesize that limiting the blood flow rate to maintain “safe” drainage pressures may help in reducing the increase of carboxyhemoglobinemia during the ECMO course.