Abstract
The cardiopulmonary bypass (CPB) procedure has been shown to be a possible cause of postoperative neurological morbidity for various reasons, including: large amounts of gaseous microemboli (GME) reaching the patient and hypoperfusion of the patient due to “stolen” blood flow. This study used a simulated CPB circuit identical to that in a clinical setting to examine three different hollow-fiber membrane oxygenators without intergrated arterial filters - the Capiox RX05, the Quadrox-i neonatal, and the KIDS D100 - to determine their ability to reduce the number of GME delivered to the neonatal patient and their hemodynamic properties in response to varying flow rates, normothermic vs hypothermic conditions, and open vs closed purge line. The circuit was primed with Ringer’s Lactate and then human blood with a hematocrit of 30%. Injections of 5cc bolusses of air were injected into the venous line proximal to the venous reservoir over a thirty-second interval. Six injections were done for each oxygenator at each of the eight different experimental conditions for a total of 64 experiments per oxygenator (192 total injections). A flow probe, pressure transducer, and Emboli Detection and Classification (EDAC) quantifier transducer were positioned both upstream and downstream of the oxygenator to measure differences in each parameter. Results demonstrated that the Capiox RX05 is the most effective oxygenator at reducing the number of microemboli that potentially can be delivered to the neonatal patient. In regards to the hemodynamic properties, the Quadrox-i has the most favorable results, with the lowest mean pressure drop and the best energy retention across the oxygenator.
Keywords
Introduction
The cardiopulmonary bypass procedure is known to contribute to postoperative neurological damage with multi-factorial complications. One of the major contributing factors to this damage is the accumulation of air in the closed circuit that is returned to the patient via an arterial line and can easily travel to the brain and occlude the blood vessels 1 . To prevent complications caused by these microemboli, filters are placed in the venous reservoir, oxygenator, and arterial line in the bypass circuit to trap gaseous and particulate emboli by size. However, 3-dimensional deformations of the microemboli may cause them to escape the size limitation of the filter pores. While transcranial Doppler (TCD) ultrasound is used to detect microemboli delivered to the patient, it can only detect sizes >40 µm and cannot classify them by size 2 . Previous experiments in our laboratory, using the Emboli Detection and Classification (EDAC) quantifier system (Luna Innovations, Inc., Roanoke, VA, USA), show that most of the microemboli are cleared by the oxygenator and the arterial filter and the majority (99%) of the microemboli reaching the patient are <40µm, which is not detectable by TCD 3 . Therefore, EDAC is the system of choice in our laboratory for the evaluation of different oxygenators, arterial filters, and venous reservoirs because it is not only able to detect the volume and size down to 10µm, but it can also classify these microemboli by size 3 .
In previous experiments, we found correlations between the formation of microemboli and different cardiopulmonary bypass conditions. A smaller number and smaller size of microemboli are associated with normothermic conditions (35°C) at a low flow rate while a greater number and size of microemboli are associated with an hypothermic temperature (25°C) and a higher flow rate 3 . In addition, we calculated the pressure drop and energy loss across the oxygenator, which is higher at hypothermic vs normothermic conditions. It was shown that an open arterial purge line also reduces the number of microemboli reaching the patient, but at the same time shunts the returning blood away from the patient via the purge line of the arterial filter, which can result in hypoperfusion 4 . The significance of pressure drop is magnified in neonatal patients; therefore, it is critical to monitor the mean pressures and hemodynamic energies (energy equivalent pressure (EEP), surplus hemodynamic energy (SHE), and total hemodynamic energy (THE)) across the oxygenator and arterial filter throughout the experiment 4 .
The objectives of this experiment are to evaluate: 1) the effectiveness of three different hollow-fiber membrane oxygenators in reducing the number and size of microemboli delivered to the neonatal patient and 2) the hemodynamic performance of the oxygenators during the cardiopulmonary bypass procedure in a simulated model. We tested two flow rates (400mL/min and 700mL/min), two temperatures (25°C and 35°C) and opened and closed the purge line to simulate different clinical conditions that could affect the amount and size of microemboli captured during the procedure.
Materials and methods
Circuit design
The circuit design for this experiment simulated cardiopulmonary bypass of a neonatal patient. The circuit included a Jostra HL-20 heart-lung machine (Maquet, Wayne, NJ, USA), a Jostra-30 heat-cooler system (Maquet, Wayne, NJ, USA), 5 feet of arterial tubing, and 6 feet of venous tubing from a COBE Heart and Lung Perfusion Pack with ¼ inch diameter (COBE Cardiovascular Inc., Arvada, CO, USA). Three different hollow-fiber membrane oxygenators were used and included the Capiox BabyRX05 (Terumo Corporation, Tokyo, Japan), the Quadrox-i Neonatal (Maquet) and the Kids D100 (Sorin Group, Milan, Italy). The specifications of the oxygenators can be found in Table 1. These oxygenators were used with their respective venous reservoirs and the specifications for these can be found in Table 2. The arterial filters used were the Capiox AF02 pediatric arterial filter (Terumo Corporation, Tokyo, Japan) with the Capiox BabyRX05 and the Quadrox-i Neonatal and the KIDS D130 Neonatal arterial filter (Sorin Group, Milan, Italy) with the KIDS D100. The purge line setup was composed of 48-inch long tubing(1/8-1/32) without a one way valve and a COBE 5-port manifold (Sorin Infant Perfusion Pack, Sorin Cardiovascular Inc.,) that connected the arterial filter to the venous reservoir 4 . The mean arterial pressure was controlled using a Hoffman clamp proximal to the pseudo patient and was maintained at 100 mmHg. The circuit is identical to the clinical circuit used in the operating room at our institution and was de-aired according to clinical standards. The complete circuit setup can be found in Figure 1.
Oxygenator specifications
Venous reservoir specifications
The experimental setup
Three disposable pressure transducers, model 041500503A (Argon Medical Inc., Ithaca, NY, USA) were used to record pressure waveforms at both the pre-oxygenator and post-oxygenator sites. A dual Transonic ultrasound flow sensor model 6XL (Transonic Systems Inc., Ithaca, NY, USA) connected to the Transonic flow-meter model TS410 was used to record pump flow. A custom-made data acquisition system with a sampling rate of 1000 samples/sec was used to collect data with a custom LabView software program (National Instruments, Austin, TX, USA). Flow and pressure data were collected two minutes after the injection of a 5cc air bolus in each trial. The pressure drop (ΔP) was calculated by subtracting the post-oxygenator mean pressure from the pre-oxygenator mean pressure.
The circuit was first primed with lactated Ringer’s solution and then human blood was added as some lactated Ringer’s was removed. The circuit temperature varied, based on the trial, and was either kept at 35°C or 25°C and the blood hematocrit level was adjusted to 30% for all experiments. The total priming volume was 500 mL and the pseudo-neonatal patient was maintained at 200mL.
Experimental design
The mock circuit was subjected to two neonatal flow rates (400 ml/min and 700 ml/min), two temperatures (25°C and 35°C) and two purge line conditions (open and closed). Five cubic centiliter bolusses of air were injected into the venous line proximal to the venous reservoir over a thirty-second interval. There were a total of eight different experimental conditions and six injections were done for each oxygenator at each experimental condition for a total of 64 experiments per oxygenator (192 total injections). The EDAC quantification system was used to collect data. Three transducers were connected to the circuit at three positions: pre-oxygenator, post-oxygenator, and post-arterial filter. Data was collected for 5 minutes after each injection of air to allow for the emboli to clear and for the EDAC reading to return to baseline.
The EDAC recorded gaseous microemboli counts simultaneously at all three locations. The EDAC provided the total emboli volume over each 5-minute interval at each transducer as well as the total emboli count split into three categories; emboli between 10 and 20 microns, emboli between 20 and 40 microns, and emboli over 40 microns and up to as large as the connector diameter.
The EDAC quantifier uses ultrasound technology to detect microemboli flowing through the tubing of the circuit. Each of the three EDAC transducers was placed adjacent to the CPB tubing and secured with a clamp. Ultrasound-coupling gel was used to insure connection integrity. The EDAC system was connected to a Windows PC via a USB for data analysis. According to the manufacturer specifications, the EDAC system is capable of counting emboli at rates up to 1,000 emboli per second with diameters of 10 microns and greater and at flow rates between 200 ml/min and 6 L/min 5 .
Statistical analysis
Analysis of variance (ANOVA) models were fit to the continuous outcomes (e.g., change in mean arterial pressure (MAP), change in THE, arterial line THE, and shunt flow) to compare the oxygenators (Baby RX05, Quadrox-i, and KIDS D100)) and purge line status (open and closed) at a given flow rate (400 and 750 ml/min) and temperature (25 and 35 degrees). A linear mixed-effects model was fit to the continuous outcomes (e.g., MAP, THE, and total emboli count) to compare the oxygenators and purge line status at a given location (pre- or post-oxygenator or arterial line), flow rate, and temperature. The linear mixed-effects model is an extension of linear regression that accounts for the within-subject variability inherent in repeated measures designs 6 . In this study, the repeated factor is the location. For each outcome, p-values were adjusted for multiple comparisons testing, using the Tukey-Kramer procedure. All hypotheses tests were two-sided and all analyses were performed using version 9.2 of the SAS System for Windows (SAS Institute Inc., Cary, NC).
Results
Gaseous microemboli detection
Increasing the flow rate from 400ml/min to 700 ml/min at both normothermic (37°C) and hypothermic conditions (25°C) resulted in an increase in the gaseous microemboli (GME) in the pre-oxygenator, post-oxygenator, and arterial line regardless of the oxygenator used in the circuit. It is also noted that, when the temperature is decreased to hypothermic conditions while maintaining the same flow rate, the total gaseous emboli also increase. Both the RX05 and KIDS D100 show a general trend that, when the purge line is open, the total GME decrease. However, the Quadrox-i shows an opposite trend of more GME detected at the pre-oxygenator and post- oxygenator, but not necessarily at the arterial line when the purge line is opened (p<0.01).
As is shown in Table 3, at a flow rate of 400 ml/min and a temperature of 35°C, all oxygenators essentially clear all microemboli before they enter the arterial line. As the flow rate is increased to 700 ml/min (Table 4), the RX05 has the least amount of GME in all lines, whether the purge line is closed or open (p<0.01). The KIDS D100 oxygenator has the most GME when the purge line is closed. However, when the purge line is open, the Quadrox-i has more GME in the pre-oxygenator line, but not the post-oxygenator or arterial lines.
Gaseous microemboli count at 35°C and 400 ml/min
No statistically significant difference was found in comparisons between oxygenators, detection sites as well as different purge line status.
Gaseous microemboli count at 35°C and 700 ml/min
When the purge line is closed, p<0.01, KIDS vs. the other two oxygenators at pre-oxygenator and post-oxygenator sites; RX05 vs. Quadrox-i at pre-oxygenator site;
When the purge line is open, p<0.01, RX05 vs. Quadrox-i & KIDS at the pre-oxygenator site; Quadrox-i vs. KIDS at the pre-oxygenator site.
When switching the purge line status, p<0.01, KIDS with open purge line vs. closed purge line at the pre- and post-oxygenator sites, Quadrox-i with open purge line vs. closed purge line at the pre-oxygenator site.
Table 5 shows that, at hypothermic conditions and a flow rate of 400 ml/min all oxygenators clear the GME before they enter the arterial line. As the flow rate is increased to 700 ml/min (Table 6), the RX05 clears the greatest amount of GME with the purge line both open and closed compared to the other two oxygenators. Also, the Quadrox-i has less GME than the KIDS in the post-oxygenator line, but the KIDS has less GME than the Quadrox-i in the arterial line.
Gaseous microemboli count at 25°C and 400 ml/min
p<0.001, Quadrox-i Neonatal vs. RX05 and KIDS D100 with an open purge line and Quadrox-i Neonatal with an open purge line vs. a closed purge line.
Gaseous microemboli count at 25°C and 700 ml/min
When the purge line is closed, p<0.01, Quadrox-i vs. other two oxygenators at all sites; RX05 vs. KIDS at the pre-oxygenator site.
When the purge line is open, p<0.01, RX05 vs. Quadrox-i at all sites; Quadrox-i vs. KIDS at the pre-oxygenator site; RX05 vs. KIDS at the post-oxygenator site and the arterial line.
When switching the purge line status, p<0.01, RX05 with open purge line vs. closed purge line at the pre-oxygenator site; KIDS with open purge line vs. closed purge line at all sites; QUadrox-i with open purge line vs. closed purge line at the pre-oxygenator site.
Therefore, at the lower flow rate, regardless of temperature, all oxygenators clear almost all GME, but, at a higher flow rate (700 ml/min), the RX05 clears the greatest amount of GME, regardless of temperature or purge line status.
Stolen blood flow
In general, the amount of blood “stolen” is larger at normothermic conditions than hypothermic conditions, but it does not change much at different flow rates. As shown in Tables 7 and 8, though statistically significant (p < 0.001), there is very little difference among all of the oxygenators and the amount of blood stolen from the pseudo-patient.
Stolen (purge line) blood flow at 35°C
p<0.001, Quadrox-i Neonatal vs. other two oxygenators;
p<0.001, Capiox BabyRX05 vs. KIDS D100.
Stolen (purge line) blood flow at 25°C
p<0.001, Quadrox-i Neonatal vs. other two oxygenators;
p<0.001, Capiox BabyRX05 vs. KIDS D100.
Pre-oxygenator mean pressure
Figure 2 shows that the pre-oxygenator mean pressure (PreOMP) is dependent on flow rate, purge line status and type of oxygenator, while the temperature has minimal impact on it. For all oxygenators, as the flow rate increases, the PreOMP also increases. When the purge line is opened, the PreOMP decreases slightly for all oxygenators. In regards to the oxygenators, the Quadrox-i has the lowest PreOMP and the KIDS has the highest PreOMP (p<0.001).
Mean pressure drop
As shown in Figure 3, the mean pressure drop is inversely proportional to temperature and directly proportional to the flow rate for all oxygenators. The Quadrox-i Neonatal has the lowest mean pressure drop at all temperatures, flow rates, and purge line status compared to the other 2 oxygenators. On the other hand, the KIDS D100 has the highest mean pressure drop for all of the parameters stated above (p<0.001).
Total hemodynamic energy (THE)
As shown in both Figures 4 and 5, THEpre varies with the changes of flow rate and temperature. THEpre increases as flow rate increases and as temperature decreases. On the other hand, THEpost is independent of flow rate, temperature and purge line status. The KIDS D100 has both the highest THEpre and THEpost and the Quadrox-i has the lowest THEpre and THEpost at all parameters (p<0.001).
Discussion
Other studies have shown that GME reaching the patient may have adverse effects, including morbidity and mortality 1 . The results of this study show that the Capiox Baby RX05 has the lowest GME in the pre-oxygenator, post-oxygenator, and arterial line at all temperatures and flow rates tested compared to the Quadrox-i and the KIDS D100. Based on our results, the RX05 oxygenator allows the least amount of emboli to reach the patient, which will most likely result in a better outcome for the patient.
The Quadrox-i has a trend opposite to that of the other 2 oxygenators such that GME increase when the purge line is open in all cases, except that in the post-oxygenator and arterial line at normothermic conditions. This variation may be due to the difference in the cardiotomy filter used for this particular oxygenator.
The issue of “stolen” (purge line) blood flow is important in a clinical setting because it leads to hypoperfusion of the patient 4 . Our results show that, if a patient is being pefused at 400 ml/min under normothermic conditions and the purge line is open, they are losing almost 110 ml/min to the purge line, which is 27.5% of their total blood volume. Thus, their total perfusion is not the 400 ml/min, but rather 290 ml/min. This hypoperfusion can lead to morbidity in the patient because critical organs are not receiving the blood supply that they need 4 . Further research in our laboratory will include experiments evaluating oxygenators with integrated arterial filters to see if their usage decreases the amount of blood stolen from the patient.
It is ideal to have a low pressure drop to conserve energy delivered to the patient. The Quadrox-i had the lowest pre-oxygenator mean pressure as well as the lowest mean pressure drop at all parameters in the experiement, whereas the KIDS D100 had the highest pre-oxygenator mean presure and highest pressure drop, which results in loss of energy.
Having the highest possible total hemodynamic energy when the blood reaches the patient is beneficial for insuring adequate microperfusion of the patient’s organs 7 . The KIDS D100 showed a much higher THE pre-oxygenator than both the RX05 and the Quadrox-i at both temperatures and particularly at 700ml/min. However, the retention of the THE in both the Quadrox-i and RX05 was much higher, and the post-oxygenator THE values for all three oxygenators were very similar. The lower THE retention is due to the large pressure drop seen with the KIDS D100.
Limitations
The difference in venous reserviors (see Table 2) contributes to the difference in pre-oxygenator GME differences for each of the oxygenators 8 .
Conclusions
This study showed that the Capiox RX05 is the most effective oxygenator at reducing the number and size of microemboli that can, potentially, be delivered to the neonatal patient. In regards to hemodynamic properties, the Quadrox-i Neonatal has the most favorable results, with the lowest pre-oxygenator mean pressure, lowest mean pressure drop across the oxygenator, and the best energy retention across the oxygenator as shown by the THEpre and THEpost values. Future experiments will investigate all parameters used in this study while using an oxygenator with an integrated arterial filter.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of Interest Statement
None Declared.