Secondary Logo

Journal Logo

Effect of Oxygenator Size on Air Removal Characteristics: A Clinical Evaluation

Stehouwer, Marco C.; de Vroege, Roel; Kelder, Johannes C.; Hofman, Frederik N.; de Mol, Bastian A.; Bruins, Peter

doi: 10.1097/MAT.0000000000000360
Adult Circulatory Support

During cardiopulmonary bypass (CPB), gaseous microemboli (GME) are released into the patients’ arterial bloodstream. Gaseous microemboli may contribute to the adverse outcome after cardiac surgery. Recently, two oxygenator models with or without integrated arterial filter (IAF) were designed and only differ in size, leading to a change of 20% in surface area of the hollow fibers and 25% in blood velocities. The aim of this study was to assess the air removal characteristics of the inspire oxygenators with or without IAF. Sixty-eight patients were randomly assigned to four different groups: optimized adult and full adult and an additional IAF. Gaseous microemboli reduction rates were measured with a bubble counter. The number of GME reduction rates showed no differences. However, both models reduced significantly less volume of GME (optimized adult: 40.6% and full adult: 50.3%) compared with both models with IAF (88.7% and 88.5%, respectively). No significant differences of reduction rates were found between both devices without IAF and also not between both models with IAF. In conclusion, the larger inspire oxygenator tends to remove more GME. No effect from size of oxygenator device with integrated screen filter on GME reduction was observed. The inspire oxygenators with IAF may be considered as an adequate GME filter.

From the *Department of Extracorporeal Circulation, St Antonius Hospital, Nieuwegein, The Netherlands; Department of Extracorporeal Circulation, HAGA Hospital, The Hague, The Netherlands; Department of Cardiology, St Antonius Hospital, Nieuwegein, The Netherlands; §Department of Cardiothoracic Surgery, St Antonius Hospital, Nieuwegein, The Netherlands; Section Cardiovascular Biomechanics, Faculty of Biomedical Technology, Technical University Eindhoven, Eindhoven, The Netherlands; and Department of Anaesthesiology, Intensive Care and Pain Management, St Antonius Hospital, Nieuwegein, The Netherlands.

Submitted for consideration October 2015; accepted for publication in revised form February 2016.

Disclosure: The authors have no conflicts of interest to report.

Correspondence: Marco C. Stehouwer, Department of Extracorporeal Circulation, St Antonius Hospital, Koekoekslaan 1, P.O. Box 2500, 3430 EM Nieuwegein, The Netherlands. E-mail:

During cardiac surgery with cardiopulmonary bypass (CPB), gaseous microemboli (GME) originate from the extracorporeal circuit and may be released into the arterial bloodstream of the patient. Gaseous microemboli are thought to contribute to the adverse outcome after cardiac surgery,1 including brain injury.

Two components of a CPB circuit are able to actively reduce GME, i.e., the oxygenator2,3 and, explicitly developed for air removal, the arterial filter.4,5

Integration of an arterial filter into an oxygenator is a contemporary concept developed to reduce the detrimental effects of CPB by reducing both prime volume and foreign surface area.

Recently, two models of adult oxygenators with or without integrated arterial filter (IAF) were designed for use during CPB. These two oxygenators, the optimized adult (Inspire 6 M; Sorin Group, Mirandola, Italy) and the full adult (Inspire 8 M; Sorin Group), could be employed depending on the required patients’ blood flow and, although identical in design, only differ in size. This difference in size leads to a change of the surface area of the hollow-fiber membrane and of the blood velocity in the devices. The optional IAF differs also in size and in surface area of the filter screen in both models. These differences lead to a change of 20% in surface area of the hollow-fiber membrane and 25% in blood velocities.

We hypothesize that these differences may affect air removal characteristics. The aim of this study was to assess the air removal characteristics during CPB of two oxygenators that differ in size and two differently sized oxygenator devices with an integrated arterial screen filter. Besides, this study provides a detailed evaluation of the air removal characteristics of the complete new inspire oxygenator family and may support the perfusionists’ choice on oxygenator device.

Back to Top | Article Outline

Materials and Methods


A prospective randomized study was performed in a teaching hospital in Nieuwegein, The Netherlands. After approval from the local Medical Ethics Committee (LTME Z-12.07), 68 patients undergoing elective cardiac surgery with CPB were randomly assigned to be perfused with both models with IAF (optimized adult + IAF or full adult + IAF group) or without IAF (optimized adult or full adult group). The specifications of the models are shown in Table 1.

Table 1

Table 1

Included were patients undergoing first time cardiac surgery and the calculated blood flow of the patient (based on a cardiac index of 2.4 L/min/m2) did not exceed 6 L/min. Excluded were emergency surgery and patients with surgical procedures combined with deep hypothermia.

Back to Top | Article Outline

CPB System

The CPB system was controlled by a HL30 heart-lung machine (Maquet, Hirrlingen, Germany) and consisted of a tip-to-tip phosphorylcholine coated tubing system (Ph.i.s.i.o; Sorin Group, Mirandola, Italy). Venous blood drained into a soft-shell reservoir (BMR 1900, Sorin Group) by gravity. A centrifugal pump (Rotaflow, Maquet) directed the blood flow into the tested oxygenator device.

All suction blood was collected in a hard shell cardiotomy reservoir (Card 43, Sorin Group) and this blood was drained back into the venous line of the CPB system.

During extracorporeal circulation, the recirculation line on the oxygenator or on the IAF, both pre- and postfilter, were constantly open, as recommended by the manufacturer, leading via the sample manifold back into the venous line.

The CPB system was primed with approximately 1,500 ml. Priming solution consisted of a combination of 500 ml gelofusin (40 g/L modified gelatin; B. Braun, Melsungen, Germany) and 1,000 ml ionolyte priming solution containing acetate: 34 mmol/L; Na+: 137 mmol/L; K+: 4.0 mmol/L; Mg2+: 1.5 mmol/L; and Cl: 110.0 mmol/L (Fresenius Kabi, Zeist, The Netherlands).

Back to Top | Article Outline

CPB Procedure

Anesthesia was induced with midazolam (0.02–0.1 mg/kg) in combination with propofol (0.5–2 mg/kg), pancuronium (0.1 mg/kg), and fentanyl (5–10 μg/kg). Anesthesia was maintained with a continuous propofol infusion (2–8 mg/kg/hr), remifentanil (5–20 μg/kg/hr), and pancuronium, when required.

After heparinization (300 IU/kg), the aorta was cannulated with an arterial Cannula 24 Fr (DLP, Medtronic, Minneapolis, MN). The right atrial appendage was cannulated with a two stage 36/51 Fr Cannula (DLP, Medtronic). When double venous cannulation was necessary, a curved 36 Fr venous single stage cannula with long tip (DLP, Medtronic) was used for cannulation of the vena cava inferior. The vena cava superior was cannulated with a curved 28 Fr cannula (DLP, Medtronic). In the aortic root, a 7 Fr venting needle (DLP, Medtronic) was placed. When necessary, left ventricular decompression was achieved with an 18 Fr Sarns vent catheter (Terumo, Ann Arbor, MI).

Cardiopulmonary bypass was initiated when the ACT reached a minimal level of 420 s measured by the Hemochron Jr (International Technidyne Corp. [ITC], Edison, NJ).

During CPB, the nasopharyngeal temperature was maintained at 32°C.

After aortic cross-clamping, preservation of the heart was achieved by topical cooling combined with infusion of 1,000 ml cold cardioplegia based on hydroxyethyl starch (HES: 60 g/L; Fresenius AG) and low sodium chloride (25 mmol/L), containing 2 mmol/L D,L-magnesium aspartate, 4 mmol/L procaine hydrochloride, 0.5 mmol/L calcium hydrochloride, 5 mmol/L potassium chloride, 10 mmol/L glucose, 200 mmol/L mannitol, and 20 mg/L dexamethason with an osmolarity of 320 mosm/L and pH 7.4.6

During the procedure, CO2 flooding was achieved with a gas flow of 6 L/min through a diffuser (CarbonAid, Cardia Innovations, Stockholm, Sweden).

During CPB, the cardiac index was maintained at 2.4 L/min/m2 and acid-base management was regulated according to the alpha-stat protocol. Mean blood pressure was maintained between 40 and 80 mm Hg.

After the initial cardioplegia dose and during rewarming, the pO2 of blood samples were measured by a blood-gas analyzer (Bayer, Rapidlab 855, Tarrytown, NY).

Shed blood was collected and processed with an autotransfusion device (Xtra; Sorin Group, Mirandola, Italy). After CPB, the residual blood in the HLM-circuit was rinsed with 1–2 L priming solution and was directed to the autotransfusion device.

After CPB, heparin was neutralized with protamine sulphate at a 1:1 ratio.

Back to Top | Article Outline

GME Detection

GME was detected with the bubble counter BCC 200 (GAMPT mbH, Zappendorf, Germany). The measurements were conducted with two noninvasive sensor probes clamped on the 3/8 inch tubing. A probe was clamped on the inflow and the other probe on the outflow tubing of the oxygenator device. The measurement is based on a self-calibrating ultrasonic Doppler device. The BCC 200 device measures the number and size of GME with a diameter ranging from 20 to 500 μm. The device specifies the bubbles with a diameter of more than 500 μm as “over range” and when the air volume is too big or too much as “bolus.” Particulate emboli do not influence the count results. Data were collected and accumulated during the complete CPB procedure.

To avoid measurement of electronic distortion produced by diathermy, the electronic filter algorithm of the BCC200 was enabled.

Back to Top | Article Outline

Data Analysis

Quantitative variables are presented as mean ± standard deviation (SD) or as median (interquartile ranges [IQR]), when appropriate. Before analysis, the data were tested for distribution according to Kolmogorov–Smirnov goodness of fit test.

We used the Student’s t-test or Mann–Whitney tests, when appropriate for comparisons focusing on two groups. ANOVA or Kruskal–Wallis test were used to assess statistical significant differences for the four groups.

Categorical variables were compared by means of the χ2 test.

A two-tailed Spearman correlation p value was used to determine the relationship between CPB time and number or volume of GME entering the devices. Also the relationship between volume of GME (in) and volume reduction within all four groups was assessed with this Spearman correlation.

Reduction percentage was calculated for every oxygenator by the formula: % reduction = [1 − GMEout/GMEin] × 100. For obtaining reduction distribution graphs, for every oxygenator at every GME size, the % reduction was calculated. Subsequently, for every GME diameter the mean over all devices was calculated and shown in the graph.

Statistical analysis was performed using SPSS 22.0 (SPSS, Chicago, IL).

Back to Top | Article Outline


Table 2 displays the demographic and intraoperative data. These baseline variables did not differ between the groups. Also the body surface area (BSA) of the patients was similar in all groups meaning that size of oxygenator was not adapted to the size of the patient.

Table 2

Table 2

Strong to weak correlations were shown between the CPB time and the number (N) and volume (V) of GME, (data of all groups were enrolled), entering (in) and leaving (out) the devices (Nin: R = 0.570, p < 0.001; Nout: R = 0.459, p < 0.001; Vin: R = 0.411, p = 0.001; Vout: R = 0.348, p = 0.008). No correlation was detected between the CPB time and number and volume reduction (R = 0.242; p = 0.067 and R = 0.009; p = 0.945, respectively).

No association was detected within all four groups between total number or volume of GME entering the devices and the % reduction: optimized adult (number: R = 0.135, p = 0.606; volume: R = 0.176, p = 0.523), optimized adult + IAF (number: R = 0.304, p = 0.0.236; volume: R = 0.466, p = 0.06), full adult (number: R = 0.179, p = 0.492; volume: R =0.196, p = 0.451), and full adult + IAF (number: R = 0.157, p = 0.548; volume: R = 0.243, p = 0.348).

Air categorized by the bubble counter in the over range group (>500 μm) or bolus group (>500 μm and >1,000 GME/s) entering the devices was measured in 40% (over range; mean number 17 ± 47) and 16% (bolus) of the patients and both categories were reduced for 100% in all oxygenator devices.

Back to Top | Article Outline

Difference in Oxygenator Size

As presented in Table 3, in the larger full adult oxygenator group, both the absolute number and volume of GME entering and leaving the devices was significantly lower compared with the optimized adult group (all p < 0.005, except volume of GME in: p = 0.013).

Table 3

Table 3

In Figure 1, the accompanying number and volume of GME reduction rates of both different sized oxygenators are shown. The number of reduction rate of the smaller optimized adult (26.0 ± 12.9 %) did not differ significantly, when compared with the full adult oxygenator (30.5 ± 15.4 %, p = 0.547). Also the volume of GME reduction rates of the optimized adult and full adult oxygenators showed no significant differences (40.6 ± 18.4% and 50.3 ± 19.3%, respectively; p = 0.153).

Figure 1

Figure 1

Back to Top | Article Outline

Difference in Size of Oxygenator Device with IAF

Table 4 shows the number and volume of GME entering and leaving the two oxygenator devices with IAF and no significant differences between these groups were shown.

Table 4

Table 4

The number and volume of GME reduction rates of the different sized oxygenator devices with IAF are shown in Figure 2. No significant differences were shown between the number reduction rates of the optimized adult with IAF and the full adult with IAF (38.0 ± 20.2% and 37.4 ± 20.4%, respectively; p = 0.744) and also no significant differences were shown in the volume reduction rates (optimized adult + IAF: 88.7 ± 7.7%; full adult + IAF: 88.5 ± 12.4%; both with a p = 0.667).

Figure 2

Figure 2

Back to Top | Article Outline

Overall Evaluation of the Four Inspire Models

The number of GME reduction rates of all four oxygenator models (Figures 1 and 2) showed no significant differences (p = 0.158). However, in volume of GME reduction rates significant differences were identified (p < 0.05). The optimized adult reduced significantly less volume of GME (40.6 ± 18.4%) when compared with both the inspire 6 and 8 with IAF (optimized adult + IAF: 88.7 ± 7.7%; full adult + IAF: 88.5 ± 12.4%; both with a p < 0.001). Also the optimized adult reduced significantly less total volume of GME (50.3 ± 19.3%) when compared with both inspire versions with IAF (p < 0.001).

The distribution between volume reduction and diameter of GME for all devices are shown in Figure 3.

Figure 3

Figure 3

Both inspire devices without arterial filter show a visible comparable relationship. The device may more prominently reduced larger GME than smaller GME. Interestingly, both distribution graphs show a much larger variance of reduction rates when compared with the distribution curves of the oxygenators with IAF. This variation in reduction rates increases with increased GME diameter size.

Both the optimized adult + IAF and full adult + IAF show a very similar distribution between GME size and reduction rate. Both devices show a negative reduction rate with GME of approximately 20 to 75 μm, followed by a steep rise of reduction rate with increasing GME diameter.

Back to Top | Article Outline


Both the optimized adult and the full adult oxygenators (without IAF), although different in size, demonstrate similar air removal characteristics, which resulted in equal GME number and volume reduction rates in both devices. Comparable air removal characteristics also apply to both models with IAF. Our findings suggest that the size of the oxygenator or size of the integrated screen filter has no effect on GME removal.

To the knowledge of the authors, no studies are available on the effect of size of oxygenator devices on air removal.

Two models of adult oxygenators, differing in size, but identical in design, enabled us to study the effect of oxygenator size on air removal characteristics. One could expect that increasing the number of hollow fibers would ameliorate the filter properties. The oxygenator functions as a barrier of air2,7 and this may be caused by two filtration mechanisms. First, the compact design of the hollow-fiber membrane possesses depth filter properties,8 and second, air may be removed by tangential (or cross flow) filtration. The larger device has a larger radius, and consequently a thicker fiber layer, probably causing increased depth filter properties. Although the oxygenator devices do not differ in length, the blood flow path in the larger device may be longer. This is due to the manufacturer’s statement that the blood flows both radially and longitudinally. Whereas, the ratio between both flows is unknown, the larger surface area may contribute to increased cross flow filtration

We observed that an increase of the surface area of the hollow-fiber membrane with 20% (from 1.43 to 1.71 m2) in the larger device tends to be translated in an exact equal increase (20%) of the volume of GME reduction from 40% to 50% (optimized adult versus full adult). Also the role of difference in radius of both oxygenator models has to be taken into account, since air can also be removed by buoyancy. This can only be achieved when the buoyancy force is higher than the opposing forces, i.e., surface adhesion and viscous drag force, known as Stokes’ drag. The average blood velocity in the larger full adult oxygenator, 0.056 m/s, is 25% lower compared with the optimized adult, 0.075 m/s, at a blood flow rate of 4 L/min (blood velocity specifications according manufacturer), which may lead to improved air removal by buoyancy properties. Computational fluid dynamics show that buoyancy probably plays no role for small GME due to strong drag forces created by high fluid velocities as shown in arterial line filters.9 Meaning that probably most GME were drawn with the blood stream into the oxygenator device barriers (i.e., hollow fibers or integrated screen filter). This is supported by our findings, since also larger GME (>100 μm) leave the devices without IAF and fractionation was observed in the devices with IAF. In worst-case scenarios, gross air may be removed by buoyancy properties. However, in this study, we observed that air entering the devices consisted almost only of GME. Moreover, no effect of device size on air removal was observed. Suggesting that other factors of the design, i.e., rheology of tangential drag force, housing geometry, are greater contributors to air removal characteristics.

Our results show that adding an IAF ameliorates air removal. Volume reduction is increased from 40% to 50% to almost 89% in both devices with IAF. The IAF of the optimized adult and full adult have a surface area of, respectively, 0.0068 and 0.0097 m2. Despite the 30% difference in surface area of the screen filter and difference in size, no influence on air removal characteristics were observed. This finding is supported by results of Riley,4 who showed no association between prime volume and air removal in 10 arterial line filter models.

Despite the much smaller surface area of the screen filters of the IAF, these filters are major contributors of air removal capacity and air volume reduction. Besides the mechanical filter properties of the screen filter, the arterial filter is designed for velocity reduction thus promoting buoyancy of bubbles, which may contribute to the air removal capacity (Figure 4). The filter frame has specific designed bars positioned in front of the inflow opening of the oxygenator that consequently redirects the flow and makes the drag force tangential. The velocity reduction in the oxygenator is probably as low as or lower than an arterial filter. So, the air reduction performance seems to be completely dependent on the “barrier” function of the screen filter pore size of the IAF.

Figure 4

Figure 4

As 60–50% of the GME passed the hollow fibers of the oxygenator, and only 12% passed the device with IAF, the reduction capacity of the arterial filters can be calculated as circa 75–80%. A reduction rate of 75–80% may be considered as a mediocre reduction rate for arterial filters when compared with an in vitro study ranking arterial filters on GME reduction.4 Moreover, the total volume reduction rate of approximately 88% is comparable to a competitive oxygenator with IAF showing 92% volume reduction, as assessed in a clinical study performed in our institute with the same bubble counter.5

Interestingly, adding an IAF did not improve the GME number reduction, although the GME volume reduction was significantly and vigorously increased. Probably this is caused by fractionation of large bubbles into more smaller bubbles as shown in the reduction graphs where in both devices with IAF a negative reduction was observed at GME < 75 μm. We previously observed this phenomenon in another oxygenator device with IAF (Quadrox-i) and these reduction graphs show very similar results compared with the current data. Among our group, also others reported about this phenomenon explicitly for screen filters.10,11

Although the number and size of GME causing clinical impairment is unknown, we argue that the use of an IAF is justified since we show that an IAF strongly ameliorates volume reduction. We consider the IAF as a safe replacement for an additional arterial line filter.

Back to Top | Article Outline


This study has some limitations. The absolute numbers and volume of GME entering the devices varied strongly per CPB procedure and significant differences were found. This variation is probably caused by the large difference in interventions and conditions (e.g., adding volume or medication, air in the venous line). In almost all groups, some procedures (approximately four) were present wherein a large amount of these interventions and conditions occurred, resulting in very high GME activity. Notably, in the full adult group none of these circumstances occurred, causing a significantly lower GME activity. We believe this can only be explained by coincidence. Although the majority of GME are observed in the first 10 min of CPB,12 a correlation was shown between CPB time and GME entering the devices. We believe that both the unequal GME load, which also comprises the number of GME per second, as the correlation with CPB time has no influence on the primary outcome % reduction since we showed with Spearman correlation tests that reduction rates were not associated with the absolute number or volume of GME entering the devices and also not with CPB time. Besides, we never observed more GME leaving then were entering the devices. This excludes the presence of a GME saturation phenomenon. In conclusion, percentage of reduction is independent of GME load and CPB time and may be considered as a valid parameter for GME removal in a clinical setting. Therefore, this is an important parameter for judging GME-removal capacities of the oxygenator and IAF.

Caution should be taken when detailed bubble distribution is used. De Somer et al.13 showed that the BCC200 bubble counter overestimates sizes of microbubbles more than 200%. However, the limitation of their study was that in a worst-case scenario, i.e., high embolic loads only one apparatus was tested.

Back to Top | Article Outline


The larger inspire oxygenator tends to remove more GME. No effect from size of oxygenator device with integrated screen filter on GME reduction was observed. This suggests that other factors in design, i.e., rheology and housing geometry, are greater contributors to air removal characteristics.

Adding an IAF strongly contributes to air removal capacities of both oxygenator devices. The inspire oxygenators with IAF reduce GME volume with 89% and may be considered as an adequate GME filter.

Back to Top | Article Outline


1. Barak M, Katz Y: Microbubbles: Pathophysiology and clinical implications. Chest 2005.128: 2918–2932.
2. De Somer F: Impact of oxygenator characteristics on its capability to remove gaseous microemboli. J Extra Corpor Technol 2007.39: 271–273.
3. Jirschik M, Keyl C, Beyersdorf F: A clinical comparison of bubble elimination in Quadrox and Polystan oxygenators. Perfusion 2009.24: 423–427.
4. Riley JB: Arterial line filters ranked for gaseous micro-emboli separation performance: An in vitro study. J Extra Corpor Technol 2008.40: 21–26.
5. Stehouwer MC, Boers C, de Vroege R, C Kelder J, Yilmaz A, Bruins P: Clinical evaluation of the air removal characteristics of an oxygenator with integrated arterial filter in a minimized extracorporeal circuit. Int J Artif Organs 2011.34: 374–382.
6. Branger AB, Eckmann DM: Theoretical and experimental intravascular gas embolism absorption dynamics. J Appl Physiol (1985) 1999.87: 1287–1295.
7. Potger KC, McMillan D, Ambrose M: Microbubble transmission during cardiotomy infusion of a hardshell venous reservoir with integrated cardiotomy versus a softshell venous reservoir with separated cardiotomy: An in vitro comparison. J Extra Corpor Technol 2013.45: 77–85.
8. De Somer F, Dierickx P, Dujardin D, Verdonck P, Van Nooten G: Can an oxygenator design potentially contribute to air embolism in cardiopulmonary bypass? A novel method for the determination of the air removal capabilities of neonatal membrane oxygenators. Perfusion 1998.13: 157–163.
9. Fiore GB, Morbiducci U, Ponzini R, Redaelli A: Bubble tracking through computational fluid dynamics in arterial line filters for cardiopulmonary bypass. ASAIO J 2009.55: 438–444.
10. Stehouwer MC, Kelder JC, van Oeveren W, de Vroege R: In vitro air removal characteristics of two neonatal cardiopulmonary bypass systems: Filtration may lead to fractionation of bubbles. Int J Artif Organs 2014.37: 688–696.
11. Johagen D, Appelblad M, Svenmarker S: Can the oxygenator screen filter reduce gaseous microemboli? J Extra Corpor Technol 2014.46: 60–66.
12. Camboni D, Schmidt S, Philipp A, et al: Microbubble activity in miniaturized and in conventional extracorporeal circulation. ASAIO J 2009.55:58–62.
13. De Somer FM, Vetrano MR, Van Beeck JP, van Nooten GJ. Extracorporeal bubbles: A word of caution. Interact Cardiovasc Thorac Surg 2010.10:995–1001.

air removal; GME; oxygenator; integrated arterial filters; design

Copyright © 2016 by the American Society for Artificial Internal Organs