148 J. R. Neal et al.: J Extra Corpor Technol 2025, 57, 147--152
to the decreased velocity, leading to an increased length of time the blood stays in the circuit. Turning down the flow serves as a crucial step in ECMO weaning by making the heart produce more of the function of generating cardiac output. As the ECMO flow is titrated down, failure of the heart to tolerate the physiologic load, inclusive of both alterations in preload and afterload along with the requirement to generate additional native cardiac output may yield exacerbations in acidosis driven by increased lactate production, systemic hypotension, and / or disadvantageous changes in pulmonary vascular mechanics [ 2, 7 ]. The act of VA ECMO support also mechanically helps unload the right ventricle with continuous flow, regardless of how low the ECMO flow is titrated. It is important to recognize that any continuous forward flow will serve to temper the physiologic impacts of these changes while also providing an unloading function to varying degrees as defined by native reserve and amount of ongoing ECMO flow. Due to the multitude of variables contributing to this phase of care that complicate the determination of viability of separation, along with the potential risks associated with weaning as noted above, the timely, efficient, and effective separation is critical in reducing the morbidity and mortality profile of ECMO support.
Description
From 2009 to 2011, our center had several ECMO patients who were expected to be successfully removed following weaning at a low continuous blood flow of about 0.75--1 LPM of blood flow with reasonable inotropic support, acceptable hemodynamic parameters, lactate levels, and arterial blood gases( ABG). Upon arrival in the operating room( OR) for decannulation, hypotension ensued during attempted decannulation, necessitating reinstitution of ECMO support. Right ventricular( RV) failure was postulated to be causative in many cases, which is similar to reported findings in other papers [ 9 ]. The need for reinsertion of cannulas is not without risk, given the difficulties associated with multiple re-cannulation of vessels, whether with a femoral or central approach [ 10 ].
These experiences yielded a reconceptualization of our approach to weaning and adaptation of a process similar to weaning from cardiopulmonary bypass. More specifically, this novel approach is contingent upon the intentional cessation of circuit flow while concurrently leaving cannulas indwelling and in continuity with the mechanical circuitry. To utilize this method, additional anticoagulation beyond standard levels on ECMO is required to prevent thrombus from forming in areas of relative or absolute blood stasis, including the cannulas. If managed like cardiopulmonary bypass( CPB), where heparin levels are maintained to produce an activated clotting time( ACT) over 480 s, it would be feasible for the ECMO circuit flow to be stopped for an extended period. This would be similar to CPB circuits being able to have flow cessation in them for over 1 h. However, this dramatic increase in anticoagulation would be fraught with bleeding risks. At our institution, a decision was made to perform a more“ middle of the road” approach with a slight increase of the ACT allowing intermittent clamping(“ clamp”) of the ECMO circuit between 3 and 4 min followed by a period of blood flow, which we call a
“ flash,” lasting for 30 s. By intermittently pausing and having periods of flow cessation, the unloading of the heart by the ECMO circuit is stopped, thereby unmasking physiologic reserve and hemodynamic tolerance of the patient.
The safe time of cessation was determined through consideration of the ACT. If it takes three and a half minutes( 210 s) for an activated sample to thrombose, temporary cessation for that duration of time was deemed feasible without undue risk of circuit or cannula thrombosis. ACTs are drawn from the patient every 30 min during the trial, once a goal ACT is reached. Following this logic, the blood that had been stagnant is flushed back into the patient to allow new blood that had not been stagnant to come to rest in the circuit for another period of 3--4 min. By repeating this clamping and flashing, we could maintain circuit and cannula patency and protect the patient from either thrombosis or bleeding. This technique has now been utilized as a standard approach for the past 13 years. While we have no official number of patients to report, a conservative estimate is that many hundreds of adult-sized patients, from teenagers to eighty-year-olds and over 65 neonatal and smaller pediatric patients, have had this trial implemented during their ECMO course. Importantly, there have been no occurrences of loss of circuit integrity or patient thrombosis during these clamp flash trials when the protocols have been followed.
During the 30-second flash period, the blood flow of the ECMO circuit is set so that the circuit volume that was stagnant is completely flushed into the patient. At our center, the adult Cardiohelp( Maquet Rasatatt, Germany) is utilized with a circuit prime volume of approximately 600 mL( including cannulas). Thus, the circuit flow during the flash is set to 1.2 LPM for 30 s. The RPMs that provide the patient with 1.2 LPM of flow during the flash period are maintained during the clamp period. Our adult circuit does not contain shunts, bridges, or manifolds. While clamped, there is complete cessation of blood flow in the circuit. In addition, during the clamp period, the sweep is shut off to prevent the blood in the oxygenator from becoming hypocapnic and hyperoxemic. Only just before and just after the flash, is there a sweep going through the circuit. The sweep is adjusted to maintain appropriate pO 2 and pCO2 levels coming out of the oxygenator during the flash. By having these intermittent periods of flow over an hour, the patients’ RV is unsupported by ECMO for the large majority of the time( Table 1).
With the use of the clamp flash trial in our pediatric patients, several modifications are required. Our primary ECMO circuits have consisted of, depending on patient size, either a Pedimag or Centrimag pump head( Abbott, Abbott Park, USA) and either a pediatric QuadroxiD( Maquet, Rastatt, Germany) or, more recently, an AMG PMP oxygenator( Euro- Set, Medolla, Italy). Our center also includes a recirculating manifold that is either 1 / 8” or 3 / 16” diameter tubing with a flow between 0.12 and 0.6 LPM. The overall circuit volume is approximately 230 ml, including the arterial and venous lines of approximately 60 ml each. In larger pediatric patients, our center employs a Cardiohelp 5.0( Maquet Rasatatt, Germany) with a modified 1 / 4 or 3 / 8 inch venous line and a modified 1 / 4 arterial line with a 3 / 16 inch recirculating manifold. In our pediatric population, given that we have a manifold, flow is maintained in the circuit, pump, and oxygenator during the