J. R. Neal et al.: J Extra Corpor Technol 2025, 57, 147--152 149
Table 1. Comparison of cardiac index( CI) for a one-hour clamp flash trial with a 3.5-minute clamping period versus a continuous low flow trial for a hypothetical adult and pediatric patient. Note: Percent decrease in total support-- L / min and time is a comparison between continuous low flow and clamp flash trials.
Weight( kg)
BSA( m 2)
Trial type
CI during flow( L / mim / m 2)
Flow rate( L / min)
Total support( L / h)
% Decrease in total support( L / h)
Supported time( min)
% Decrease in support minutes( L / h)
Adult |
80 |
2 |
Continuous low flow |
0.5 |
1 |
60 |
-- |
60 |
-- |
|
80 |
2 |
Clamp flash |
0.3 |
0.6 |
9 |
85 |
7.5 |
88 |
Pediatric |
4 |
0.48 |
Continuous low flow |
0.25 |
0.12 |
7.2 |
-- |
60 |
-- |
|
4 |
0.48 |
Clamp flash |
0.29 |
0.14 |
1 |
86 |
7.5 |
88 |
clamp periods, but not in the arterial and venous lines to the patient. During our flash periods, our center aims to only have enough flow to the patient to change over the 60 mL present in each of the arterial and venous lines. The patient flow during the flash is set between 0.12 and 0.14 LPM for 30 s. The venous line is reclamped after the flash period. During the clamp period, the tubing clamp is placed only on the patient side of the venous line before the manifold re-entry. Similar to adult patients, sweep is only provided to the oxygenator during the period from just before until just after the flash. This prevents the circuit from becoming hypocapnic and hyperoxemic. Careful control of the sweep and fractional delivered oxygen( fdO2) is performed by the perfusionist or ECMO specialist. If needed, a circuit arterial blood gas( ABG) is performed during the clamp period to assess the gas settings during the flash. For both adult and pediatric patients, ABGs and lactates are drawn every 30 minutes during the trials.
Pediatric patients can have a more varied heart physiology than adults [ 11 ]. Children with congenital cyanotic heart lesions need special attention in controlling the paO2 and paCO2. During the flash, the goal is not only to prevent excessive hemodynamic support but also to prevent excessive respiratory support. Given this, after the second or third clamp, a blood gas off the manifold line is used to assess the circuit’ s paO2 and paCO2, which is provided during the flash. For patients with cyanotic defects, the goal is not to elevate the SaO2 more than 2 to 3 mmHg above the patient’ s baseline during the period of clamping. This prevents the flash from over-supporting the patient and misrepresenting the child’ s respiratory ability during the clamp periods.
An additional consideration during the trial is the location of the clamp placement on the venous line during flow cessation. The area that was clamped is varied during each subsequent clamping to prevent excessive wear and / or future thrombin / thrombus buildup at that location of the clamping( Fig. 1). This can sometimes happen after an unsuccessful wean. While minor and insignificant on the venous side of the circuit, this is a concern on the arterial limb due to the potential embolic risk to the patient’ s systemic arterial side. Before implementing the purposeful relocation of the clamping location, this buildup of thrombotic deposition was noted on the venous line, which has been dramatically reduced following integration of this relatively minor, but important, protocol adjustment. The arterial line should be free of any red cell incorporated thrombus laydown and assessed for this before starting a clamp / flash trial, as this thrombus laydown could become mobile during the stagnant period. Then, during the flash, be sent into the arterial cannula or the patient.
The use of heparin as the only means of anticoagulation for ECMO has dramatically decreased over the last 14 years. Our center, in addition to many others, has increased the use of direct thrombin inhibitors( DTIs) to offset the need for antithrombin supplementation [ 10, 12 ]. Heparin is advantageous when protecting blood in states of stasis and from thrombosis, as compared to using DTIs only. Additionally, an issue with using DTIs is that there is a prolongation of the ACT that is artificial [ 13, 14 ]. Normally, this is not a concern since APTTs are used more often in ECMO. However, during a clamp and flash trial, the ACT is the critical lab value due to its quick result. This can lead to a situation where using artificially elevated ACT levels, whose absolute value was thought to be safe, is inadequate, yielding an elevated and difficult to quantify risk for thrombosis. Given the inaccuracy of the ACT in the context of concurrent bivalirudin administration, a solution was deemed necessary. Our center looked at previous research showing a direct, limited correlation between ACT and APTT values with heparin [ 15 ]. The study showed that low-level heparin administration is reflected in the values of these two tests. Consequently, our protocol was modified to apply these findings and compare the APTT and ACT values with a bivalirudin drip only versus the values on ECMO patients with a heparin drip. The difference in ACT values at similar APTTs led us to believe that the bivalirudin skews the ACT approximately 20--30 s higher. Normally, adult patients on a therapeutic bivalirudin infusion should have an ACT of roughly 230--250 s to start the clamp flash trial after a heparin bolus. In the pediatric setting with patients on a bivalirudin drip, the goal ACT after the heparin bolus is between 210 and 225 s, since only the arterial and venous lines are left stagnant during the clamp period( Table 2). Previously published institutional experience with bivalirudin, as well as numerous meta-analyses of retrospective studies, supported the viability and potential superiority of bivalirudin as compared to heparin-based systemic anticoagulation in ECMO [ 16 ]. Despite this reassuring data, it is essential to recall important limitations of bivalirudin therapy. Of these, the most important with regard to the application of clamps on ECMO circuitry is the organ-independent proteolysis of the parent compound, resulting in loss of regional anticoagulation in areas of relative or absolute blood stagnation. In effect, in areas of stasis, the drug is metabolized without requiring circulation through an effector organ, and as such, fresh delivery of bivalirudin is required to maintain systemic anticoagulation.