Does the venous cannulation method affect gaseous embolic load to the patient during extracorporeal circulation?

Abstract

Introduction

Cardiopulmonary bypass (CPB) is essential for cardiac surgery but poses risks, including gaseous micro emboli (GME). While the incidence of stroke—a common clinical consequence of embolism—ranges from 1-5% in cardiac surgery, prevalence of GME during CPB remain poorly understood.

Objectives

To quantitatively compare GME incidence in the arterial line between cavoatrial and bicaval cannulation during open-heart surgery. Secondary exploratory objectives include evaluating the impact of venous reservoir volume on GME, and the correlation between GME in the venous and arterial lines.

Methods

This single center randomized controlled trial was conducted at Sahlgrenska University Hospital, Sweden. Patients ≥18 years undergoing planned aortic valve repair/replacement with cavoatrial cannulation, or mitral valve repair/replacement with bicaval cannulation, with or without coronary artery revascularization, were screened for eligibility. Patients were further randomized to either venous reservoir volume of ≥300 mL (control) or 200–300mL (intervention). GME detection was performed using GAMPT BCC300 with probes positioned at multiple locations within the CPB circuit.

Results

39 patients were included. No significant differences in GME quantity in the arterial line were observed between the cannulation methods (count p=.444; volume p=967). Similarly, no significant differences were found based on venous reservoir volume (count p=.074; volume p=.166). Furthermore, no significant correlation was observed between GME in venous line entering the arterial line (count p-value=492; volume p-value=.750). The CPB circuit effectively removed 99.14% of GME, with no adverse events reported.

Conclusion

No significant differences were found in arterial GME count or volume between cavoatrial and bicaval cannulation during CPB. These findings underscore the importance of the bypass circuit's air-handling capacity, as well as the role of modern oxygenators and arterial line filters in effectively minimizing the passage of GME. Trial Registration:ClinicalTrials.gov Identifier: NCT05820828URL: https://clinicaltrials.gov/ct2/show/NCT05820828

Keywords

Bubble detector Cannulation Extracorporeal circulation Gaseous micro emboli Open heart surgery

Introduction

Cardiopulmonary bypass (CPB) used in cardiac surgery is associated with the risk of causing damage to various organs, including the brain, lungs, and kidneys.¹ Emboli within the systemic circulation is a rare, but serious complication, and can lead to severe neurological damage.^2–4 As far as we know, there is no clear data on the incidence of air embolism during cardiac surgery with cardiopulmonary bypass. However, the incidence of stroke (a common clinical outcome of embolism) after cardiac surgery is approximately 1%–5%, depending on the type of procedure and patient profile.⁵ The mean incidence of clinical stroke among coronary artery bypass grafting patients in Sweden from 2006 to 2017 was reported by Jónson et al. to be 1.2%.⁶ Ideally, the cardiopulmonary bypass circuit should prevent all air emboli. However, studies have shown that circuits do not offer complete protection, as gaseous micro emboli (GME) have been detected even after passing through the oxygenator.^1,7 The precise mechanisms contributing to this phenomenon remain unclear, as GME during open heart surgery with CPB is a multifactorial issue.^4,8

Venous cannulation

For cardiac surgery that does not involve opening of the right atrium or right ventricle, for example aortic valve replacement or coronary artery bypass grafting, a cavoatrial cannulation with a two-stage cannula is often used. The cannula requires only one incision and is placed in the right atrium and has side holes that drain from the atrial level, and the tip of the cannula drains the inferior vena cava. The alternative bicaval cannulation method is often used for mitral valve surgery or procedures that require the right atrium to be opened, where two single-stage cannulas are placed in the inferior and superior vena cavae. Sometimes occlusion is used to optimize caval decompression, to improve drainage of blood and to prevent air entering the cannulas.^2,9,10

Air entry into the circuit could potentially arise from suboptimal or improper cannula placement, including misaligned side holes or failure to secure an air-tight cannulation site.^2,8 Rodriquez et al. identified a significant correlation between the volume of air in the venous cannula and the count of emboli detected in the middle cerebral artery (MCA) at the initiation of CPB.² However, the effect of venous cannulation methods on GME generation and detection of GME in the arterial line remains uncertain.

Venous reservoir

The venous reservoir is the first component in the circuit designed to eliminate any accidentally introduced or generated air.¹¹ Studies have consistently shown a significant correlation between low venous reservoir volumes and GME in the arterial line.^4,12–14 Rodriguez et al. recommended maintaining a venous reservoir volume of at least ≥800 mL to mitigate the risk of emboli potentially circulating to the brain.⁴ Nielsen et al. on the other hand, noted that improvements in reservoir design, such as enhanced air filtration and bubble traps, minimized the bubble count even at lower reservoir volumes.¹³ However, the study was conducted in an in vitro setting, which may not fully reflect clinical scenarios.

Hypothesis

We hypothesize that bicaval cannulation, which requires two separate cannulation incisions, will result in a higher count and volume of GME reaching the arterial line compared to cavoatrial cannulation. Additionally, we expect that lower venous reservoir volumes will be associated with an increased risk of GME entering the arterial circulation.

Objectives

The primary objective of this study was to quantitatively compare the incidence (count and volume) of GME detected in the arterial line between two commonly used venous cannulation methods, cavoatrial and bicaval, during open heart surgery with CPB. As secondary exploratory objectives, we assessed whether the volume in the venous reservoir (control ≥ 300 mL, intervention 200–300 mL) influences the quantity of GME detected in the arterial line, and the possible correlation between GME in the venous and arterial line.

Method and materials

Ethical approval and considerations

Ethical approval for this study was granted by the Swedish Ethical Review Authority in March 2023 (protocol ID: 2023-00204-01). The study adheres to the ethical principles outlined in the Declaration of Helsinki and follows guidelines established by the International Committee of Medical Journal Editors for research participant protection. The trial is registered on ClinicalTrials.gov (NCT05820828).

Informed consent and data protection

Participants received both oral and written information about the study and had time for consideration before providing informed consent. Data management complied with the European Union’s Data Protection Directive to ensure participant confidentiality and security.

Patients

This single center randomized controlled trial was conducted at Sahlgrenska University Hospital, Gothenburg, with patient screening occurring between April and August 2023. Inclusion criteria were patients aged ≥18 years undergoing planned cavoatrial cannulation for aortic valve repair or replacement (AVR) or planned bicaval cannulation for mitral valve repair or replacement (MVR) with or without coronary artery revascularization (only left internal mammary artery (LIMA) to left anterior descending artery (LAD)), with planned normothermia (35°C–37°C). Exclusion criteria were urgent or emergent surgeries, adult congenital heart disease (ACHD) surgery, acute endocarditis, redo heart surgery, intraoperative iatrogenic adverse event such as major hemorrhage, aortic dissection and/or other severe complications.

To ensure balanced allocation of participants in the two groups for secondary endpoints, patients were randomized in blocks of four, using a 1:1 ratio via closed envelope to either control group (≥300 mL in venous reservoir) or intervention group (200–300 mL in venous reservoir).

Sample size

A power analysis was performed based on previous research examining GME in the arterial line during CPB in open heart surgery.¹⁵ We hypothesized that bicaval cannulation would result in a 30% increase of detected GME in the arterial line compared to the baseline count of 72 ± 23 GME observed with cavoatrial cannulation.¹⁵ Using an alpha level of 0.05 and a power of 80%, the analysis indicated that a sample size of 17 patients per group would be required to detect a statistically significant difference. For potential dropouts, a total of 40 patients were enrolled.

Clinical management

The circuit design was consistent across all patients regardless of group assignment. A Stöckert S5 (LivaNova, Munich, Germany) with roller pump system was utilized in combination with an Inspire 8F oxygenator, which included an integrated arterial filter (38 microns), an Inspire HVR (hardshell venous reservoir) with dual-screen filters (41 and 120 microns), and an integrated cardiotomy reservoir (41 microns) (LivaNova, Mirandola, Italy). The circuit was primed with 1300 mL of crystalloid prime solution, infused with 10.000 IU of Heparin (Fresenius, Bad Homburg, Germany), and recirculated through a pre-bypass filter before cannulation. Cannula sizes ranged from 22 to 24F for arterial cannulas and 32/40F (cavoatrial) and 26F together with 28F (bicaval) for venous cannulas, selected in accordance with the manufacturer’s specified safety pressure drop threshold restrictions. Heparin was given with a target activated clotting time of ≥480 seconds before initiating CPB. CPB pump flow was set to flow index 2.4 L/min/m². Antegrade cold blood cardioplegia, served as myocardial preservation, was administered approximately every 20 minutes with at ratio of 4:1. During surgery, vacuum assisted venous drainage (VAVD) of −20mmHg was used according to surgeons’ preference. When the left heart was open, carbon dioxide (CO₂₎ insufflation of 10 L/min was applied into the surgical field.

Anesthetic procedure

Anesthesia was conducted in agreement with our department’s clinical standard, induced by administration of fentanyl (3‒5 µg/kg), propofol (1.0–1.5 mg/kg), rocuronium (0.6 mg/kg) and followed by endotracheal intubation. Sevoflurane was used to provide anesthesia throughout the procedure (0.5%–2.5%). A transesophageal echocardiographic probe was inserted in all patients to examine heart valve structure and function, correct venous cannula position, and for detection of any residual air in the heart.

Measurements

GAMPT BCC300 (GAMPT mbH, Merseburg, Germany) clamp on probes, with pulsed ultrasonic doppler system, were used to detect and quantify GME with range between 10 and 2000 μm. GME volume was expressed in microliter (µL). The three probes were positioned on the venous line (before the venous reservoir), pre-oxygenator (after the master pump), and on the arterial line. Data collection was recorded at CPB initiation and ended at CPB completion. During CPB, specific predefined ranges, including a hematocrit level of 27%–32%, PaO₂ between 18 and 24 kPa, PaCO₂ within 5‒6 kPa, SpO₂ ranging from 97% to 100%, SvO₂ > 65%, normothermia, and a mean arterial pressure between 40 and 70 mmHg were maintained.

Venous reservoir volume

To maintain a venous reservoir volume of 200–300 mL in the intervention group throughout CPB, a standardized setup was established before initiating CPB. A quick prime tubing was connected to the post oxygenator shunt, allowing excess volume to be drained into an infusion bag during CPB. This drained volume was reinfused into the reservoir prior to wash-in. Hemofiltration, cell salvage and wash-in were made according to our clinical protocol.

Statistical analysis

Statistical analysis was performed using IBM SPSS ((version 28.0.1.1 (15), IBM Corp, Armonk, NY). Continuous variables were expressed as either median with interquartile range (median (IQR)) or mean ± standard deviation (mean ± SD), depending on the distribution of the data. Normality was assessed using the Shapiro-Wilk test. The Mann Whitney U-test was applied for non-normally distributed continuous variables to compare differences between groups, while Students t test was used for normally distributed data. Spearman’s rank correlation coefficient (ρ) was used to assess correlation. A p-value of <.05 was considered statistically significant.

Results

Between April and August 2023, 40 patients were included and randomized in this trial. One patient was excluded due to per operative finding of patent foramen ovale, and cannulation strategy had to be converted from cavoatrial to bicaval cannulation, leaving 39 patients for analysis. No adverse events occurred during the trial. CONSORT template, see Figure 1. Patient demographics and baseline characteristics are detailed in Table 1.

Table 1.

Patient baseline characteristics.

	Cavoatrial cannulation n = 19	Bicaval cannulation n = 20
Female, n (%)	5 (26)	3 (15)
Age years, median (range)	69 (55‒81)	69 (43–82)
Height, cm	177 ± 11	180 ± 8
Weight, kg	87 ± 11	82 ± 13
Body surface area, m²	2.04 ± 0.18	2.02 ± 0.17
Hemoglobin, g/L	122 ± 16	131 ± 12
Hematocrit, %	36 ± 5	39 ± 4
Creatinine, mmol/L	113 ± 129	90 ± 22
History of diabetes mellitus, n (%)	6 (32)	1 (5)
History of hypertension, n (%)	13 (68)	6 (30)
History of hyperlipidemia, n (%)	6 (32)	2 (10)
History of stroke, n (%)	0 (0)	0 (0)
Smoker, n (%)	1 (5)	0 (0)
LVEF, %	53 ± 9	55 ± 8
EuroSCORE II^a (%)	1.25 ± 0.62	1.44 ± 0.88

Data is presented as mean ± SD, median and range, numbers, and percentage.

LVEF: Left ventricular ejection fraction (Simpson method).

^aEuropean System for Cardiac Operative Risk Evaluation, calculation tool to assess risks associated with open-heart surgery, range 0%‒100% with higher scores indicating higher risk.

No significant difference in GME quantity in the arterial line was observed between the two cannulation methods (Table 2).

Table 2.

Treatment characteristics and arterial line GME data during CPB.

	Cavoatrial cannulation n = 19	Bicaval cannulation n = 20	p-value
CPB time, min	81 (37)	93 (20)	.107
Cross clamp, min	59 (32)	70 (22)	.879
Reperfusion, min	12 (4)	15 (12)	.070
Pump flow, L/min, mean ± SD	4.89 ± 0.43	4.85 ± 0.41	.792
GME arterial line, count	96 (117)	120 (560)	.444
GME arterial line, counts/min	0.88 (1.26)	1.37 (7.22)	.667
GME arterial line, volume (µL)	0.14 (0.47)	0.15 (0.39)	.967
GME arterial line, volume/min (µL/min)	0.0019 (0.0042)	0.0084 (0.0042)	1.0

Data is presented as median (IQR) or mean ± SD. Mann Whitney U-test and t test.

CPB: Cardiopulmonary bypass; GME: gaseous micro emboli; µL = microliter.

Regarding exploratory analyses, measurements indicate no difference in GME count measured between the groups. However, bicaval cannulation generated a higher GME volume, compared to cavoatrial cannulation. Additionally, an increased GME count was observed after the main pump (pre-oxygenator) in the bicaval group, compared to the cavoatrial group. However, no difference in GME volume between groups was found before the oxygenator (Table 3).

Table 3.

Pre-oxygenator GME data during CPB.

	Cavoatrial cannulation n = 19	Bicaval cannulation n = 20	p-value
GME venous line before venous reservoir, count	9029 (4689)	11695 (30027)	.084
GME venous line before venous reservoir, counts/min	93 (91)	112 (348)	.079
GME venous line before venous reservoir, volume (µL)	1223 (1330)	4903 (30736)	.001
GME venous line before venous reservoir, volume/min (µL/min)	15 (13)	54 (387)	.001
GME pre-oxygenator after master pump, count	3396 (5342)	10513 (25208)	.001
GME pre-oxygenator after master pump, counts/min	46 (6)	116 (304)	.002
GME pre-oxygenator after master pump, volume (µL)	2.7 (7.9)	2.6 (3.3)	.380
GME pre-oxygenator after master pump, volume/min (µL/min)	0.04 (0.08)	0.03 (0.04)	.270

Data is presented as median (IQR). Mann Whitney U-test.

GME: gaseous micro emboli; µL: microliter; CPB: Cardiopulmonary bypass.

Additional exploratory analyses examined differences between low and high venous reservoir volumes concerning GME entering the arterial line. No significant difference in arterial GME count was found between the control group (≥300 mL) and the intervention group (200-300 mL), Table 4. The circuit effectively eliminated 99.14% (±2.44%) of incoming GME.

Table 4.

GME quantity in the arterial line based on venous reservoir volume.

	Control group (≥300 mL) n = 19	Intervention group (200-300 mL) n = 20	p-value
GME arterial line, count	69 (137)	146 (382)	.074
GME arterial line, counts/min	0.85 (1.57)	1.54 (4.09)	.101
GME arterial line, volume (µL)	0.13 (0.27)	0.26 (0.71)	.166
GME arterial line, volume/min (µL/min)	0.0013 (0.0034)	0.0027 (0.0081)	.204

Data is presented as median (IQR). Mann Whitney U-test.

GME: gaseous micro emboli; ECC: extracorporeal circulation; µL: microliter.

The Spearman’s rank correlation coefficient (ρ) between GME count and volume in the venous and arterial lines revealed a weak, non-significant correlation between the variables. Specifically, for GME count, the p-value was .492 with a linear R² of 0.106 (Figure 2(A)). For GME volume, the p-value was .750 with a linear R² of 0.014 (Figure 2(B)).

Discussion

To the best of our knowledge, this is the first study to investigate the potential impact of venous cannulation methods on both the count and volume of GME in the arterial line during open-heart surgery with CPB. Our primary finding was that there were no significant differences in either GME count or volume between the two venous cannulation methods, considering GME entering the arterial line during cardiopulmonary bypass. These well-designed circuits efficiently clear emboli through the combined functions of key components such as the venous reservoir, oxygenator, and arterial line filter (ALF).^8,16,17 Additional contributing factors may include surgical techniques, vent positioning, CO₂ flooding in the surgical field, and meticulous de-airing protocols.

One of the most notable advancements has been the ongoing refinement of venous reservoir design over the years.^8,16,17 Benim et al. demonstrated that design enhancements, such as reducing turbulence at exit points, have significantly decreased GME passage to the arterial line. Furthermore, the absolute quantity of GME passing through the oxygenator to the arterial line in both the cavoatrial and bicaval cannulation groups was markedly lower than in previous studies.^2,18,19 Contemporary oxygenators incorporate a top shunt mechanism, directing accumulating air back into the reservoir and thereby further reducing arterial air embolization.²⁰ This underscores the efficiency of current air-handling technologies in CPB circuits, which is likely a key factor in the lower arterial GME counts and volume observed in this study.

While previous research has consistently shown that the introduction of air into the venous line results in a predictable increase in embolic detection distal to the oxygenator,^2,18,19 this could not be confirmed in our study. However, a notable finding was that while GME counts were low in the arterial line, a significantly higher number of GME were detected in the pre-oxygenator position. This is likely due to the fragmentation of larger bubbles into smaller microbubbles as they pass through the venous reservoir and roller pump before entering the oxygenator. The ability of modern CPB circuits to effectively capture and eliminate larger bubbles is well documented, as both the oxygenator and arterial line filter (ALF) contribute to air removal before blood reaches the arterial circulation.^4,13,14

Another important consideration is the methodology used for GME measurement. Our study employed the GAMPT BCC300, which differs from its predecessor, the BCC200, due to software updates aimed at improving measurement accuracy.²¹ Previous studies have shown that the BCC200 overestimated GME by a factor of approximately 2–3 and concentration by approximately 20%.²¹ While the BCC 300 has been designed to address this limitation, it has not yet been widely used in research, making direct comparisons with previous studies challenging. Although our findings indicate substantially lower arterial GME counts compared to earlier reports, the differences in measurement technology should be considered when interpreting these results. Future studies comparing different emboli detection systems are needed to validate the accuracy of the BCC300 and ensure consistency in GME quantification across studies.

As secondary objective, we examined the impact of venous reservoir volume on the GME passage through the oxygenator, comparing control (≥300 mL) and intervention (200–300 mL) groups. While fluid dynamics suggest that higher flow rates and turbulence may impair the efficacy of the ALF in capturing microbubbles,^4,13,14 our findings showed no significant difference in the arterial GME count and volume even with shorter circulating time in the reservoir. The absence of significant differences between high and low venous reservoir volumes is likely attributed to the robust air-handling capabilities of the circuits, particularly under low-volume conditions. The circuit used in our study is approved for a pump flow rate of 8 L/min with minimum operating level of 150 mL volume within in the reservoir. The fact that all cases were performed above the minimum venous reservoir threshold level, may explain why we observed only a numerical increase in GME within the intervention group, without reaching statistical significance.

Further research is essential to optimize clinical practices in order to continue minimizing GME, particularly as the use of miniature extracorporeal circuits, which are potentially less efficient at air-handling,²² becomes increasingly prevalent. Additionally, it would be valuable for future studies to explore the relationship between venous reservoir volumes and the risk of GME across a broader range of flow rates and circuit designs, including potential comparisons between hard-shell and soft-shell reservoirs. These studies should be carefully designed with appropriate power calculations to ensure sufficient statistical power and reliability of the findings.

Strengths and limitations

Key strengths of this study include its prospective, randomized design, diverse patient population, and standardized procedural protocols. The consistency of having a single individual manage patient enrollment, assignment, and data collection minimized potential biases. However, variations in the surgical technique, procedural steps and venous cannulation selection may act as confounding variables, representing a limitation in this study. Another limitation of this study is that it was conducted using only one cardiopulmonary bypass circuit, which may limit the generalizability of the findings. Also, the use of VAVD based on the surgeons’ preference reduces the homogeneity of the groups. Regarding the secondary exploratory aims the study’s relatively small sample size and limited variability in venous reservoir levels between the randomized groups may have restrict its ability to adequately assess any differences between GME and venous reservoir volume.

Conclusion

No difference in arterial GME count or volume was observed between cavoatrial and bicaval cannulation during CPB. Although a notable difference in pre-oxygenator GME quantity was detected between the groups, the oxygenator’s air-handling capacity, along with the arterial line filter (ALF), effectively neutralized this difference before the GME reached the arterial line. These findings suggest that the design of modern oxygenators and ALFs plays a crucial role in minimizing the impact of cannulation techniques on GME passage. Secondary exploratory data suggest that even at lower venous reservoir volumes (200–300 mL), the oxygenator’s air handling remained efficient.

Footnotes

Acknowledgements

We would like to express gratitude to the Cardiology Research Unit for their contribution.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical statement

ORCID iDs

Petronella Torild

Anna Corderfeldt Keiller

Tor Damén

Data Availability Statement

The datasets generated during and/or analyzed during the study are available from the corresponding author on reasonable request.*

References

Zanatta

Forti

Minniti

, et al. Brain emboli distribution and differentiation during cardiopulmonary bypass. J Cardiothorac Vasc Anesth 2013; 27: 865–875.

Rodriguez

Rubens

Belway

, et al. Residual air in the venous cannula increases cerebral embolization at the onset of cardiopulmonary bypass. Eur J Cardio Thorac Surg 2006; 29: 175–180.

Borger

Feindel

. Cerebral emboli during cardiopulmonary bypass: effect of perfusionist interventions and aortic cannulas. J Extra Corpor Technol 2002; 34: 29–33.

Rodriguez

Williams

Babaev

, et al. Effect of perfusionist technique on cerebral embolization during cardiopulmonary bypass. Perfusion 2005; 20: 3–10.

Gaudino

Rahouma

Di Mauro

, et al. Early versus delayed stroke after cardiac surgery: a systematic review and meta-analysis. J Am Heart Assoc 2019; 8: e012447.

Jonsson

Barbu

Nielsen

, et al. Perioperative stroke and survival in coronary artery bypass grafting patients: a SWEDEHEART study. Eur J Cardio Thorac Surg 2022; 62: ezac025.

Kurusz

Butler

. Bubbles and bypass: an update. Perfusion 2004; 19(Suppl 1): S49–S55.

Taylor

Borger

Weisel

, et al. Cerebral microemboli during cardiopulmonary bypass: increased emboli during perfusionist interventions. Ann Thorac Surg 1999; 68: 89–93.

Punjabi

Taylor

. The science and practice of cardiopulmonary bypass: from cross circulation to ECMO and SIRS. Glob Cardiol Sci Pract 2013; 2013: 249–260.

10.

Chikwe

Cooke

Weiss

. Cardiopulmonary bypass. OSH cardiothoracic surgery. Oxford: Oxford University Press, 2013, pp. 101–126.

11.

Liu

Newland

Tully

, et al. In vitro evaluation of gaseous microemboli handling of cardiopulmonary bypass circuits with and without integrated arterial line filters. J Extra Corpor Technol 2011; 43: 107–114.

12.

Patel

Banahan

Janus

, et al. Neurological impact of emboli during adult cardiac surgery. J Neurol Sci 2020; 416: 117006.

13.

Nielsen

Funder

Jensen

, et al. Influence of venous reservoir level on microbubbles in cardiopulmonary bypass. Perfusion 2008; 23: 347–353.

14.

Norman

Sistino

Acsell

. The effectiveness of low-prime cardiopulmonary bypass circuits at removing gaseous emboli. J Extra Corpor Technol 2004; 36: 336–342.

15.

Lynch

Wells

Akers

, et al. Monitoring microemboli during cardiopulmonary bypass with the EDAC quantifier. J Extra Corpor Technol 2010; 42: 212–218.

16.

Benim

Frank

Assmann

, et al. Computational investigation of hemodynamics in hardshell venous reservoirs: a comparative study. Artif Organs 2020; 44: 411–418.

17.

Rodriguez

Belway

. Comparison of two different extracorporeal circuits on cerebral embolization during cardiopulmonary bypass in children. Perfusion 2006; 21: 247–253.

18.

Willcox

Mitchell

Gorman

. Venous air in the bypass circuit: a source of arterial line emboli exacerbated by vacuum-assisted drainage. Ann Thorac Surg 1999; 68: 1285–1289.

19.

Jones

Deal

Vernon

, et al.

How effective are cardiopulmonary bypass circuits at removing gaseous microemboli?

J Extra Corpor Technol 2002; 34: 34–39.

20.

Stammers

Miller

Francis

, et al. Goal-directed perfusion methodology for determining oxygenator performance during clinical cardiopulmonary bypass. J Extra Corpor Technol 2017; 49: 81–92.

21.

Segers

Stehouwer

de Somer

, et al. Optical verification and in-vitro characterization of two commercially available acoustic bubble counters for cardiopulmonary bypass systems. Perfusion 2018; 33: 16–24.

22.

Nollert

Schwabenland

Maktav

, et al. Miniaturized cardiopulmonary bypass in coronary artery bypass surgery: marginal impact on inflammation and coagulation but loss of safety margins. Ann Thorac Surg 2005; 80: 2326–2332.