Hyperviscosity secondary to high hemoglobin is rarely reported as a limiting factor for extracorporeal circulation. This case report describes an uncommon cause of limitation of efficiency in a venovenous extracorporeal membrane oxygenation (VV ECMO) device related to polycythemia and its novel management. The patient provided informed consent for scientific publication.
A laparoscopic sleeve gastrectomy was performed for a 47 year old obese (170 kg, 178 cm, body mass index 53.7 kg/m2, and body surface area [BSA] 2.90 m2), hypertensive, and sleep apnoeic patient. The preoperative laboratory evaluation showed a high hemoglobin 19.7 g/dl and hematocrit 51%, without any other abnormality. Bariatric surgery was performed under general anesthesia (propofol and remifentanil) with the following respiratory parameters: tidal volume (Vt) 10 ml/kg ideal body weight (IBW), respiratory rate 12 cpm, and positive end-expiratory pressure (PEEP) 10 cm H2O. No complications were reported during surgery and the patient underwent a successful extubation.
An acute respiratory failure, associated with bilateral radiologic infiltrates, occurred within 12 h. An empirical antibiotic therapy was initiated, and high-flow nasal oxygen therapy was initiated. The patient was assessed with a transthoracic cardiac ultrasonography: the ventricles morphology and contractility were normal and the cardiac output was estimated at about 9 L/min according to the pulsed wave Doppler technique, the cardiac index being in the normal range (3.1 L/min/m2). The ratio of the velocities of the mitral early diastolic wave and diastolic mitral valve annulus velocity (E/E′) was 5.8 indicating normal filling pressure and ruling out any cardiac failure. The diagnosis of severe acute respiratory distress syndrome (ARDS) was set according to the Berlin criteria (Arterial oxygen pressure [PaO2]/Oxygen inspired fraction [FiO2]: 61). Invasive and protective mechanical ventilation (6 ml/kg IBW) was applied. The FiO2 was set to 1 and PEEP to 20 cm H2O according to the PEEP tables of the ARDS network.1 Despite a neuromuscular blocking therapy (cisatracurium) and a deep sedation, the patient’s oxygenation worsened down to a PaO2/FiO2 of 51. Because of the patient’s extreme instability, a VV ECMO support was initiated (Stockert SCPC Centrifugal Pump; SORIN Group, Munchen, Germany and EOS ECMO PMP, 1.2 m2; LivaNova Inc. formerly Sorin, Arvada, CO). A 25F “lighthouse tip” drainage cannula (Bio-Medicus 25F × 50 cm; Medtronic, Minneapolis, MN) was placed in the right femoral vein, and a 21F “lighthouse tip” cannula (Bio-Medicus 21 × 18 cm) was placed in the right jugular vein (return cannula) under real-time ultrasonography surveillance. The position of the drainage cannula was verified with ultrasound to be at the entrance of the right atrium. We aimed to reach an arterial oxygen saturation above 85% and an ECMO flow of 60% of the measured cardiac output. Continuous monitoring of right radial artery blood pressure and oxygen arterial saturation was associated with intermittent arterial preoxygenator and postoxygenator blood gases as needed. Oxygenation improved dramatically, suggesting a low recirculation fraction, as observed with the sharp visual contrast between the dark desaturated drainage blood and the bright red blood in the return cannula. No indirect signs of hemolysis, such as hematuria or decrease in hemoglobin level, were observed. Immediately after the initiation of ECMO, we achieved a maximum ECMO flow of about 4.5 L/min despite a pump velocity of 3,500 revolutions per minute (RPM), which is almost its highest capacity. The pressure before and after the oxygenator was elevated without any obstruction of the return cannula. We decided to decrease hemoglobin levels to prevent hemolysis and ECMO pump failure. A total blood volume of 1300 ml was withdrawn while the same volume of 0.9% saline was reinjected to ensure hemodynamic stability. Hemoglobin and hematocrit decreased accordingly: 13.2 g/dl and 40%, respectively. The pressure before and after the oxygenator decreased and the ECMO blood flow increased more than 1 L/min reaching more than 60% of the assessed cardiac output, allowing pump velocity reduction to 3,154 RPM without any norepinephrine increase (Table 1). Continuous heparin administration was initiated in order to achieve an activated clotting time of 1.5 to 2 times of normal, according to local protocol.
An ultraprotective ventilation (Vt = 3 ml/kg IBW; PEEP 20 cm H2O, FiO2 0.3) was implemented. A radiologic and clinical improvement of the patient allowed the weaning from mechanical ventilation on postoperative day (POD) 5 and from VV ECMO on POD7.
Clear liquids and enteral nutrition were uneventfully introduced on POD3 and POD5, respectively. The patient was discharged from Intensive Care Unit on POD12 and from hospital on POD14. Follow-up at 1 month was unremarkable.
We review the case of an obese, polycythemic, critically ill patient developing severe ARDS and requiring VV ECMO support. It is well known that low levels of hemoglobin may affect the capacity of ECMO to increase arterial blood oxygenation. Nevertheless, the effects of high hemoglobin levels on ECMO function have not been studied thoroughly. In this case, the patient had very high hemoglobin levels, presumably secondary to obesity hypoventilation syndrome. In a previous retrospective study, a hemoglobin level higher than 13 g/L was an independent predictive factor for hemolysis.2 Owing to the non-Newtonian fluid shear thinning properties of blood, the dynamic viscosity decreases with increased flow or shear, the latter improving rheology in the capillaries for example. However, high pump velocity for long periods of time can precipitate hemolysis by way of increasing the mechanical stress on red blood cells.3 We did not observe any hemolysis, possibly because patient’s hemoglobin and pump velocity were decreased early after ECMO initiation.
This case reveals the difficulty to achieve an appropriate ECMO output with high hemoglobin levels and stresses the importance of blood viscosity. Importantly, a high ECMO output was required in this patient because of the ultraprotective ventilation and the patient’s cardiac output (> 9 L/min). One may argue that inadequate VV ECMO flow may be because of inappropriate set-up or size of cannulas and expect an obese patient to have wider central veins. However, the size of the drainage cannula was determined after ultrasonographic deep vein diameter evaluation as to obtain the highest output, and greater diameter cannulas were considered at risk. The length of the cannula does increase resistance and an alternative configuration (drainage from the superior vena cava with a short multistaged cannula and return flow via inferior vena cava) would result in less recirculation and more effective ECMO flow.4 Although the elevated native cardiac output may hide recirculation, it was not considered to be the core issue as our main concern was rather the elevated pressures than a lack of oxygenation. Therefore, we assumed that the low output was neither related to the length or design of the cannulas, nor the set-up used. Poiseuille’s law, a commonly used simplification of Navier–Stokes equation describes the laminar flow of an incompressible fluid: the output is inversely related to blood viscosity (Figure 1), and importantly, the relation between blood viscosity and hematocrit is not linear but exponential.3 Accordingly, moderate changes in hematocrit may be associated with conspicuous changes in blood viscosity, sufficient to impact on the cardiopulmonary bypass (CPB) pressure gradient. Moreover, nonphysiologic turbulent flow that occurs in ECMO results in viscous shear stresses, loss of energy and an increased risk for hemolysis.5 We observed a significant improvement in the ECMO pump’s performance after the hematocrit reduction. Given that the ECMO’s output was sufficient and no hemolysis was observed, the hematocrit was not decreased any further.
High preoxygenator pressures and CPB dysfunction have already been reported in polycythemic patients undergoing elective cardiac surgery. In a coronary artery bypass graft case,6 the hematocrit reached 51% and the oxygenator had to be changed despite the administration of acetylsalicylic acid. Interestingly enough, in an aortic valve replacement and coronary artery bypass graft case,7 the patient’s hematocrit at the time of CPB initiation was only 38%. The authors administrated platelet antiaggregants medications in addition to further hemodilution, allowing surgery to proceed, without oxygenator change. Our case differs first because ECMO oxygenators are different from those used in cardiac surgery and second because the major reason suspected is high blood viscosity rather than thrombosis. Importantly, despite the high-pressure levels before and after the oxygenator, neither the pressure gradient nor the flow through the oxygenator were higher than recommended by the company (EOS ECMO). Immediately after the hemodilution, the oxygenator pressure gradient substantially decreased and no oxygenator change was required over the entire ECMO support period, even though heparin was not administered at therapeutic doses. This suggests that a thrombus formation within the hemofilter upon ECMO initiation was unlikely and that the major reason for increasing pressure in the circuit was the high viscosity of the blood.
Hematocrit and hemoglobin level reduction used to be achieved for centuries by phlebotomy and some indications such as polycythemia vera, hemochromatosis, and other iron overload remain up-to-date.3 In this particular case, an isovolumic blood withdrawal was preferred to avoid any hemodynamic compromise. This intervention can be compared to acute normovolemic hemodilution (ANH), described in cardiac surgery as a blood conservation technique, although its benefits are debated.8 ANH associates phlebotomy along with crystalloid or colloid infusion, before the surgery; the blood withdrawn being administered if required. Erythrapheresis was another option that may be more effective, although at higher cost; however, the association of erythrapheresis with ECMO may be time-consuming as well as a technical challenge.
Venovenous extracorporeal oxygenation device may be a life-saving treatment in severe obese patients with ARDS, but hyperviscosity may be a limiting factor to achieve optimal ECMO output. Acute normovolemic hemodilution may be a safe, inexpensive and efficient technique to reduce hemoglobin or hematocrit levels. Further evaluation should be performed to determine the optimal hematocrit for patients supported with VV ECMO.
1. Brower RG, Matthay MA, Morris A, et al. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome
. N Engl J Med 2000.342: 1301–1308.
2. Jenks CL, Zia A, Venkataraman R, Raman L. High hemoglobin is an independent risk factor for the development of hemolysis during pediatric extracorporeal life support. J Intensive Care Med 2017.34.
3. Kwaan HC, Wang J. Hyperviscosity
vera and other red cell abnormalities. Semin Thromb Hemost 2003.29: 451–458.
4. Frenckner B, Broman M, Broomé M. Position of draining venous cannula in extracorporeal membrane oxygenation
for respiratory and respiratory/circulatory support in adult patients. Crit Care 2018.22: 163
5. Yen JH, Chen SF, Chern MK, Lu PC. The effect of turbulent viscous shear stress on red blood cell hemolysis. J Artif Organs 2014.17: 178–185.
6. Lehot JJ, Was B, Dendeleu L, Jegaden O. [Oxygenator thrombosis without heparin resistance in polycythemia
vera]. Ann Fr Anesth Reanim 2012.31(suppl 1): S14–S17.
7. Fieldwalker MA, Jackson SC, Seal D. High transoxygenator pressure gradient in a patient with polycythemia
vera. J Cardiothorac Vasc Anesth 2010.24: 104–108.
8. Barile L, Fominskiy E, Di Tomasso N, et al. Acute normovolemic hemodilution
reduces allogeneic red blood cell transfusion in cardiac surgery: A systematic review and meta-analysis of randomized trials. Anesth Analg 2017.124: 743–752.