Extracorporeal Membrane Oxygenation for Pregnant and Postpartum Patients

Michael J. Wong, MD; Shobana Bharadwaj, MBBS; Jessica L. Galey, MD; Allison S. Lankford, MD; Samuel Galvagno, DO, PhD; Bhavani Shankar Kodali, MD


Anesth Analg. 2022;135(2):277-289. 

In This Article

Ventilation and Oxygenation

Oxygen consumption increases by 20% during pregnancy[30] and ECMO flows must be adjusted accordingly. While an arterial oxygen saturation (Sao2) >80% represents adequate oxygenation during ECMO support in nonpregnant adult patients,[27] such a degree of relative hypoxemia is unacceptable during pregnancy. Assuming normal uteroplacental blood flow is maintained, a minimum maternal Pao2 of 60 mm Hg (8.0 kPa) is required to ensure adequate fetal oxygenation, as reflected by maternal Sao2 >90%.[30] In practice, ECMO flows are titrated to achieve a maternal Sao2 >95%.[18] Recurrent fetal decelerations or other fetal heart rate tracing abnormalities may indicate the need to augment maternal oxygenation via increased ECMO flow or other methods such as prone positioning.[27] As such, continuous fetal heart rate tracing serves as a "fifth vital sign" for maternal well-being, since it can herald physiologic derangement before maternal decompensation.[31]

Cardiac output increases 30% to 50% by the third trimester, which may reduce the efficiency of oxygenation during ECMO.[10] Higher cardiac output results in more venous blood bypassing the drainage cannula and proceeding to the heart without being oxygenated.[32] In adult patients, ECMO flows typically begin around 60 to 80 mL/kg/min (3–4 L/m2/min); however, during pregnancy, ECMO flows of 100 to 120 mL/kg/min (5–6 L/m2/min) are often necessary to compensate for elevated cardiac output and improve oxygenation.[4] Further increasing ECMO flow eventually predisposes to undesirable recirculation between drainage and return cannulas, as well as chatter, hemolysis, and thrombocytopenia.[32,33] Beta-adrenergic antagonists can help reduce excessive cardiac output while on ECMO,[34] and this approach is acceptable for pregnant patients.[33] Esmolol or metoprolol is used to target a maternal heart rate of 100 beats per minute or lower. Additional venous drainage cannula placement may also be considered when flow is limited by cannula size such that oxygenation goals cannot be met.[35] At our institution, we have not required extra drainage cannulas for pregnant patients on ECMO support, but for nonpregnant patients, this would usually entail insertion of a 21- to 23-Fr single lumen cannula, with the most common location for additional drainage cannula placement being the left femoral vein.

During normal pregnancy, increased minute ventilation reduces Paco2 to a range of 28 to 32 mm Hg (3.7–4.3 kPa), creating a favorable partial pressure gradient for fetal carbon dioxide elimination and oxygen uptake.[36,37] Since the fetus has limited capability to buffer for major acid-base disturbances,[38,39] it is ideal to maintain this physiologic Paco2 range though it may be difficult to accomplish initially. Fetal acidosis shifts the oxygen-hemoglobin dissociation curve rightward, reducing fetal oxygen-carrying capacity and contributing to a vicious cycle of worsening hypoxemia and acidosis. Although brief periods of hypercapnia up to 60 mm Hg (8.0 kPa) are tolerated by fetuses of healthy volunteers, a safe threshold for permissive hypercapnia in the context of ongoing maternal critical illness remains undetermined.[40,41] Therefore, hypercapnia (Paco2 > 40 to 45 mm Hg, or 5.3–6.0 kPa) should be avoided, and a target Paco2 in the range of 28 to 32 mm Hg (3.7–4.3 kPa) should be gradually pursued as it becomes increasingly feasible.[11,12,27,42] Additionally, hypocapnia (Paco2 < 15 to 25 mm Hg, or 2.0–2.7 kPa) must also be avoided, as it causes fetal acidosis via uterine artery vasoconstriction.[38,43,44]

Where pregnancy is a state of chronic, partially compensated metabolic alkalosis,[37] maternal acidosis (pH < 7.25) is generally poorly tolerated by the fetus. In our experience, progressive maternal acidosis portends the development of a nonreassuring fetal heart rate tracing unless corrective measures are taken, and low maternal pH may be more deleterious than maternal hypercarbia. Thus, while its underlying causes are identified and addressed, our practice is to temporize maternal acidosis by raising maternal pH closer to 7.4 with an infusion of sodium bicarbonate (150 mEq of sodium bicarbonate in 1 L of 5% dextrose in water). In severe refractory maternal acidosis, continuous renal replacement therapy may be considered concurrent to ECMO support (eg, continuous VV hemofiltration);[45] however, this must not delay preparation for possible emergent cesarean delivery.

Of note, fetal and maternal acid-base status change in tandem, with fetal pH trending approximately 0.1 lower than maternal pH.[46] With normal uteroplacental blood flow, fetal and maternal Paco2 are similarly correlated, where fetal Paco2 is about 10 mm Hg (1.3 kPa) greater than maternal Paco2.[47] These relationships can guide the indirect assessment of fetal well-being, in addition to fetal heart rate monitoring.

Minute ventilation increases 30% to 50% in pregnancy, largely driven by a 40% increase in spontaneous tidal volume.[30] Because of this, lung-protective ventilation results in relative hypoventilation for pregnant patients, which may contribute to fetal acidosis as well as loss of fetal physiologic reserve. Increasing the respiratory rate during mechanical ventilation may not adequately address this relative maternal hypoventilation, so it may be necessary to accept somewhat higher tidal volumes and airway pressures (up to 35 cm H2O). It is noteworthy that pregnant patients were excluded from seminal investigations demonstrating the superiority of lung-protective ventilation in ARDS.[48] The use of transpulmonary pressure monitoring has been suggested in other populations to optimize ventilation parameters;[49] however, this approach has not been studied in pregnant patients. Considering increased intra-abdominal pressure and decreased chest wall compliance observed during pregnancy, it is unclear at what threshold slightly higher tidal volumes may meaningfully affect transpulmonary pressures.[40,50] Notwithstanding, VV-ECMO facilitates minimization of ventilator support during maternal respiratory failure.

For pregnant and nonpregnant patients alike, optimal ventilation parameters during ECMO have yet to be determined.[51] At our institution, "lung rest" on VV-ECMO is generally achieved using pressure control ventilation with a peak inspiratory pressure set at 20 to 25 cm H2O, positive end-expiratory pressure (PEEP) of 10 to 15 cm H2O, respiratory rate of 10 breaths per minute, and fraction of inspired oxygen (FIO2) at 30% to 40%.[52–54] This tends to be adequate for pregnant and nonpregnant patients alike, as oxygenation and carbon dioxide elimination are accomplished via the ECMO membrane lung. There is increasing awareness that elevated driving pressure (ie, tidal volume normalized to lower respiratory system compliance or plateau pressure [Pplat] minus PEEP) is associated with increased mortality during ARDS.[55,56] Thus, it is also desirable to periodically monitor driving pressure and keep it <15 to 20 cm H2O. Tidal volumes and compliance on lung rest settings are trended over the course of days to weeks and help demonstrate whether lung recovery is occurring. In some instances of refractory hypoxemia, we use airway pressure release ventilation (APRV) with pressure during inspiration (P-high) maintained <30 cm H2O to reduce the risk of barotrauma.[57]