Secondary Logo

Journal Logo

Animal Models

Vacuum-Assisted Venous Drainage during Fetal Cardiopulmonary Bypass

Lubbers, Walter C.; Baker, R Scott; Sedgwick, Joseph A.; Lam, Christopher T.; Schenbeck, Jennifer L.*; McNamara, Jerri L.*; Ferguson, Robert E.*; Lombardi, John*; Gardner, Aimee*; Clark, Kenneth E.; Eghtesady, Pirooz

Author Information
doi: 10.1097/01.mat.0000178966.79876.3d
  • Free


Fetal cardiopulmonary bypass (CPB) has been a goal since it was proven feasible in the early 1980s,1 yet the eccentricities and frailties of the fetal physiology have posed several as of yet insurmountable challenges to researchers and clinicians. The largest obstacle in development of a successful protocol for fetal CPB has been the failure not of the fetal heart but of the placenta. In animal models, in most cases a massive rise in placental vascular resistance follows the introduction of CPB within 30–90 minutes, resulting in fetal acidosis, hypoxia, and finally fetal demise.1–5

This increase in placental dysfunction is often at least partly assigned to inadequate perfusion of the placenta.5–7 The need for high flow rates to preserve placental perfusion has become one of the principles of fetal CPB; however, high flow rates in the fetus are difficult to achieve. Miniature fetal blood vessels limit the size of applicable cannulae (especially venous), which in turn limit the achievable perfusion rates. The fetus's small size also requires reducing pump prime volume to prevent fetal blood hemodilution and to minimize blood surface contact area. This calls for the closeness of the perfusion pump to the surgical field, which further decreases gravity drainage.

One variation to conventional CPB that has been effective at many centers in solving both of these problems is the use of vacuum-assisted venous drainage (VAVD) while on CPB. First used in the late 1990s, this technique improves venous drainage by applying a vacuum to a sealed venous reservoir.8 Many have shown VAVD to be efficacious in many specialized applications of bypass, including neonatal and minimally invasive cardiac surgery.9,10 Use of VAVD allows for reduced circuit size and prime volume, preventing the excessive hemodilution often faced in neonatal patients.8,10 It also improves venous drainage and flow rates through the smaller cannulae used in neonates and in peripheral cannulation for minimally invasive cardiac surgery.8,9,11 These characteristics make VAVD ideal in fulfilling the technical requirements of fetal CPB, but concerns still exist about the safety of this technique. Some have noted VAVD increases the likelihood of gaseous embolism in the arterial line during CPB, especially with air entrained into the venous line12–15; however, reports of the severity to which VAVD aggravates gaseous embolism conflict.16

With the preceding in mind, we hypothesized that applying VAVD to our model of fetal CPB would increase venous return. In turn, this would allow us to use higher CPB flow rates, which would decrease placental dysfunction at initiation of bypass. This is the first report of application of VAVD technology to CPB support in the fetus.

Materials and Methods

Animal Care

All animals received humane care according to the principles drafted by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences (NIH publication 85-23, revised 1985). The Committee for Animal Care at Cincinnati Children's Hospital Medical Center also approved the experimental protocol. Fourteen time-dated Suffolk-cross pregnant ewes (mean gestational age 101 days, range 90–105 days) were used for this study. They were housed communally, had free access to food (sheep pellets) and water, and were allowed free movement about the pen. We kept the animals on a 12-hour light/dark cycle.

Fetal Surgical and CPB Procedures

Ewes were anesthetized with ketamine and diazepam and maintained with isoflurane. After placement of arterial and venous catheters in the ewe, we made a midline laparotomy and a small hysterotomy through which we could expose a fetal forelimb. After placement of fetal arterial and venous lines, we performed a standard sternotomy and pericardiotomy to expose the heart and great vessels. After administering heparin (300 units/kg, fetal weight estimated based on fetal gestational normograms) and confirming a satisfactory activated clotting time, we cannulated the fetal vessels. For arterial cannulation, we introduced a 6Fr arterial cannula into the pulmonary artery and for venous cannulation, we placed an angled venous cannulae (8–12 Fr) into the right atrium through the superior vena cava (SVC). The size of the venous cannulae was dependent on the size of the fetal SVC.

We started bypass using gravity siphon drainage, and then increased flow rates to the maximum achievable level. After 15 minutes of bypass, we introduced VAVD at vacuum pressures of either –20 mm Hg or –40 mm Hg, and then again adjusted flows rates to the maximum achievable level, i.e., the maximum flow rate achievable while maintaining fetal mean arterial pressure at ∼40 mm Hg and maintaining a constant volume of blood in the cardiotomy reservoir. The fetal mean arterial pressure of 40 mm Hg was chosen based on previous studies indicating collapse and loss of placental microvascular compliance at pressures below this value.5,17

We terminated bypass after 60 minutes. We gave protamine to the fetus and removed the cannulae. After closing the fetal chest, we returned the fetus to the uterus. Fetal arterial and venous lines were exteriorized through the maternal flank for continued administration of drugs and measurement of fetal hemodynamics and blood gasses. We continued to observe the ewe and fetus until it was clear that fetal death was imminent (bradycardia and severe acidosis), at which point we euthanized the ewe and fetus using high-potassium pentobarbital.

Circuit Design

Our CPB circuit (Figure 1) consisted of a roller pump (SARNS 7000, Ann Arbor, MI), hard-shell venous reservoir (Terumo RxO5, Ann Arbor, MI), hemoconcentrator (PAN-06, Asahi Medical Co., Ltd., Tokyo, Japan), and heat exchanger (CSC-14, Cobe Cardiovascular, Arvada, CO). The placenta functioned as the oxygenator in these experiments. Vacuum was applied to the reservoir using house suction and a vacuum regulator. We recorded applied vacuum pressure by a pressure transducer placed at the venous line manifold on the cardiotomy.

Figure 1.
Figure 1.:
Schematic diagram of our CPB circuit with VAVD capability.

The circuit was primed using Plasmalyte-A (Baxter Health care Corp, Deerfield, IL) and adult sheep blood containing heparin (500 ml); we adjusted prime calcium concentration and hematocrit (∼35) to match the fetal values.

Hemodynamic and Blood Gas Surveying

We monitored maternal and fetal hemodynamics (CVP, HR, SBP, DBP, MAP), maternal pCo2, Spo2 (oxygen saturation), temperature, electrocardiogram and fetal Spo2, temperature and echocardiogram continuously using Surgivet Vital Signs Monitor (Surgivet Inc., Waukesha, WI; catalog # V1886). We also confirmed maternal arterial pressure noninvasively (cuffed forearm).

We sampled fetal and maternal arterial blood gas (pCo2, pO2), bicarbonate (HCO3), SaO2, and pH values on a pH/blood gas analyzer, Rapidlab 800 machine (Bayer Diagnostics, Tarrytown, NY) or i-STAT (Abbot, East Windsor, NJ). We measured maternal blood gas values every 30 minutes to ensure satisfactory ventilation and electrolyte levels. After placement of fetal catheters, we measured fetal blood gas values every 15 minutes until completion of the study.

Other Recorded Data

Besides the preceding measurements, we recorded both maternal and fetal electrolytes, ionized calcium, hemoglobin and hematocrit values using the Rapidlab 800 System (Bayer Diagnostics) or i-STAT machine (Abbot). We performed these measurements at the same intervals as for blood gas measurements. Fetal weights were estimated based on gestational age normograms at the beginning of the experiments and confirmed after euthanasia.

Statistical Analysis

We performed our statistical analysis using Microsoft Excel XP (Microsoft, Redmond, WA). We analyzed difference between maximum achieved flows using gravity siphon drainage and VAVD using Student's t test. Differences between groups were assessed by general factorial analysis of variance. Differences were considered significant when p < .05. All values are expressed as the mean with standard deviations.


We conducted bypass for 60 minutes in 14 fetal lambs (90–100 days gestation). All fetuses were alive at termination of bypass. For conduct of CPB, venous cannulae were selected based on the size of the SVC. In three fetuses, 8Fr venous cannulae were used; in two fetuses, 10Fr venous cannulae were used; and in nine fetuses, 12Fr venous cannulae were used (Table 1). When we began our studies, we initially aimed to test VAVD at vacuum pressures used clinically, i.e., ∼ –40 mmHg. However, after experiencing atrial injuries at these pressures (see below), we elected to also test lower vacuum pressures (–20 mm Hg) as well. For this reason, in the initial two fetuses, which were the only ones in which we used 10Fr cannulae, we were only able to test VAVD at –40 mm Hg.

Table 1
Table 1:
Average Flows Achieved through Venous Cannulae of Assorted Sizes Using Gravity Siphon Drainage Alone or Gravity Siphon Drainage with VAVD

Average flow achieved using gravity siphon drainage was 132 ml/kg/min (range 48–160 ml/kg/min). After starting VAVD, we achieved significantly (p < .05) increased average flows of 285 ml/ kg/min (range 109–481 ml/ kg/min) (Figure 2). Varying vacuum pressure (–20 or –40 mm Hg) did not yield a significant difference in achieved flow rates (195.8 vs. 224.2 ml/kg/min, p = .18) (Table 1). Use of VAVD increased flow rates by an average of 42%. Initiation of VAVD was followed by an average decrease in mean arterial pressure of 10.8 mm Hg (not statistically significant).

Figure 2.
Figure 2.:
Average maximum flow achieved using gravity siphon drainage alone and gravity drainage with VAVD. Error bars indicate standard deviation. The difference between the groups was statistically significant (p < 0.05).

Vacuum-assisted venous drainage at –40 mm Hg caused tearing in the right atrium in four fetuses. This occurred when the right atrium was drawn into the tip of the venous cannula. No injury was seen when pressures of –20 mm Hg were used. We tried to repair the tear in all four fetuses and were successful in three cases. Entrainment of air into the venous return line occurred during these repairs, but we did not observe any gross evidence of deleterious air embolism as a result. No noticeable sudden change in umbilicoplacental hemodynamics or fetal gas exchange was noted. Similarly, we did not see any ischemic changes in the fetal echocardiographic recordings.

Spontaneous pulmonary hemorrhage occurred in the two fetuses in which the highest flow rates were achieved (≥400 ml/kg/min), resulting in massive blood loss and fetal exsanguinations. This bleeding occurred in the right lung. In both cases, the fetuses were not hypertensive and fetal mean arterial pressure (30–40 mm Hg) was not above physiologic levels when hemorrhage occurred.

Fetal blood gasses were preserved while on bypass. We divided flows into three groups to compare effects of flow rates to standard flows used during neonatal cardiac surgery, in which average flows are ∼200 ml/kg/min. Low flows (<175 ml/kg/min) were achieved in three fetuses, moderate flows (175–250 ml/kg/min) in eight fetuses, and high flows (>250 ml/kg/min) in five fetuses. Fetuses in which the average flow rate was similar to that used in neonatal perfusion exhibited the best preservation of pH (Figure 3A) compared to normal physiologic values for the fetus. Carbon dioxide tensions (pCO2) and oxygen tensions (pO2) were preserved for all flow rates; fetuses in which moderate flows were achieved also had pCO2 and pO2 values closest to the normal physiologic values (Figures 3B and 3C). Although we were able to preserve fetal blood gasses and hemodynamics during CPB, severe placental dysfunction and acidosis followed completion of bypass run within 90 minutes (Figure 4) in all but one case. In the exception, the fetus tolerated bypass well and survived for several hours after bypass, when it was euthanized according to our experimental protocol.

Figure 3.
Figure 3.:
3A, 3B, and 3C: Average fetal arterial blood pH (A), pCo2 (B), and pO2 (C) before initiation and after termination of CPB, grouped by average flow achieved during the bypass run. Average flows were 138, 214, and 302 ml/kg/min for the low-, moderate-, and high-flow groups, respectively. Error bars indicate standard deviation. Differences between flow rate groups were not statistically significant; differences between pre- and postbypass values were not significant.
Figure 4.
Figure 4.:
Fetal arterial pH during and after CPB. Note that CPB began at “on bypass” time 0 and was terminated 60 minutes later at “off bypass” time 0.


Fetal CPB presents several difficult challenges in perfusion, needing high flow rates using small cannulae and a low-prime circuit. Many have used VAVD to help achieve these ends in neonatal, pediatric, and adult CPB, but no group has yet evaluated its efficacy in fetal CPB.

We have shown in this report that VAVD can in fact augment gravity siphon drainage during fetal CPB and lead to higher flow rates. Using VAVD, we were able to more than double the average flows obtained using gravity drainage alone. This increase is considerably higher than that reported in in vitro studies of VAVD using cannulae of the same size, which report a 40% increase in flows when –40 mm Hg vacuum pressure is applied.18 During several procedures, flows were limited not by inadequate venous return, but by high arterial line pressure. These high systemic fetal arterial pressures reflected the increase in placental vascular resistance noted with CPB as reported previously.

Four fetuses suffered right atrial injury at vacuum pressures of –40 mm Hg because of the right atrium being sucked into the venous cannula. Similar cardiac trauma has not been previously reported for VAVD. Our observations most likely reflect the fragile nature of fetal tissues. We were able to successfully repair these tears in three cases and perform the balance of the bypass run without incident. However, we recommend that future uses of VAVD during fetal CPB consider our observations. In our continuing experiments, we no longer use vacuum pressures greater than –20 to –25 mm Hg.

Contrary to reports that VAVD worsens air embolism, especially when gas is entrained to the venous line,14 we observed no evidence of harmful air embolism in any animal. We noted no significant change even when a significant volume of air entrained during these atrial repairs. We were unable to quantify the number or volume of emboli introduced to the fetus, as we did not include a bubble detector in our circuit. However, we did not see the rapid deterioration of the fetus during bypass that would be expected had a destructive volume of air been passed to the fetus. Because in our experiments the placenta functioned as the oxygenator, any significant embolism into the placental microvasculature might have been expected to result in sudden decrease in placental perfusion. Similarly, we did not view any evidence of fetal electrocardiographic changes suggestive of air embolism into fetal coronaries. We offer this as preliminary evidence that one of the majors concerns about the use of VAVD may not in fact pose significant danger to the patient when used in fetal CPB.

Spontaneous hemorrhage of the fetal lungs during CPB like that remarked in this study has not previously been reported; however, flow rates as high as accomplished in this study (as high as 505 ml/kg/min) had also not been previously reported. In the two affected fetuses, within 10 minutes of fixing flows above 400 ml/kg/min, severe and uncontrollable bleeding began from a then-unknown source, later determined to be the parenchyma of the fetal lungs. Although fetal mean arterial pressure remained under 40 mm Hg, we saw a dramatic increase in line pressure in the CPB circuit. We suspect this resulted from placental and peripheral systemic vasoconstriction and simultaneous pulmonary vasodilation in the fetus, a reaction seen in fetal hypoxia. The most likely culprit was evolving placental dysfunction after introduction of bypass. Previous studies have suggested that with initiation and conduct of fetal CPB, flow is preferentially diverted away from the placenta and that blood flow to the pulmonary circulation increases in turn.3,19,20 The mechanism underlying this redistribution of fetal blood flow remains unknown, but most likely has significant implications for future attempts at protecting placental dysfunction after fetal CPB.

Despite the high flow rates achieved in these experiments, we were unable to reproducibly prevent the placental dysfunction that has confounded most every study of fetal CPB. Although we were able to largely preserve hemodynamics and blood gasses while on bypass, termination of bypass was almost invariably followed by a severe drop in fetal pH and sharp increase in pCO2, leading to fetal death within 90 minutes. This leads us to believe that while increased flow rates are necessary to preserve placental perfusion during fetal CPB, the principal reason for dysfunction is not a mechanical limit in perfusion. Our data are contrary to some previous reports suggesting that high flow rates may evade the problem of placental dysfunction associated with fetal CPB.3,5

Our studies had several limits. We used the placenta as our oxygenator and used adult donor blood as the substrate for our prime solution. Although we were able to adjust pH and calcium content of the prime (to match that of the fetus), we could not regulate the carbon dioxide and oxygen gas tensions before starting bypass or during bypass. Often the starting prime blood had blood oxygen levels higher than the fetus, which may have led to some early vasoconstriction and placental dysfunction. A second set of experiments to study the effect of manipulating gas tensions in the circuit prime solution is planned. We also did not specifically look for potential gaseous emboli in our experiments. Significant microemboli may have been produced during CPB that only caused placental dysfunction in a delayed fashion. In our experiments, flow rates were adjusted based on the systemic mean arterial pressure of the fetus; pharmacological manipulations of systemic or placental vascular resistance during bypass could lead to different flow rates and may have resulted in changes in placental gas exchange.

Although this study was limited in scope, it does suggest that VAVD may safely and effectively be used to augment gravity siphon drainage during fetal CPB. Applying VAVD allows increased flow rates through small cannulae while reducing the overall size and prime volume of the CPB circuit. Although VAVD is not the silver bullet that will make successful CPB in the fetus possible, it is a useful tool in overcoming several of the hurdles inherent in treating the fetal cardiac patient.


1.Hanley FL: Fetal cardiac surgery. Adv Card Surg 5: 47–74, 1994.
2.Reddy VM, Liddicoat JR, Klein JR, et al: Long-term outcome after fetal cardiac bypass: Fetal survival to full term and organ abnormalities. J Thorac Cardiovasc Surg 111: 536–544, 1996.
3.Sistino J: Foetal bypass: Concepts and controversies. Perfusion 13: 111–117, 1998.
4.Saiki Y, Rebeyka I: Fetal cardiac intervention and surgery. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 4: 256–270, 2001.
5.Hawkins JA, Clark SM, Shaddy RE, Gay WA, Jr: Fetal cardiac bypass: Improved placental function with moderately high flow rates. Ann Thorac Surg 57: 293–297, 1994.
6.Hawkins JA, Paape KL, Adkins TP, et al: Extracorporeal circulation in the fetal lamb: Effects of hypothermia and perfusion rate. J Cardiovasc Surg (Torino) 32: 295–300, 1991.
7.Fenton KN, Heinemann MK, Hanley FL: Exclusion of the placenta during fetal cardiac bypass augments systemic flow and provides important information about the mechanism of placental injury. J Thorac Cardiovasc Surg 105: 502–512, 1993.
8.Banbury MK, White JA, Blackstone EH, Cosgrove III DM: Vacuum-assisted venous return reduces blood usage. Thorac Cardiovasc Surg 126: 680–687, 2003.
9.Shin H, Yozu R, Maehara T, et al: Vacuum assisted cardiopulmonary bypass in minimally invasive cardiac surgery: Its feasibility and effects on hemolysis. Artif Organs 24: 450–453, 2000.
10.Berryessa R, Wiencek R, Jacobson J, et al: Vacuum-assisted venous return in pediatric cardiopulmonary bypass. Perfusion 15: 63–67, 2000.
11.Merkle F, Boettcher W, Schulz F, et al: Perfusion technique for nonhaemic cardiopulmonary bypass prime in neonates and infants under 6 kg body weight. Perfusion 19: 229–237, 2004.
12.Kiyama H, Imazeki T, Katayama Y, et al: Vacuum-assisted venous drainage in single-access minimally invasive cardiac surgery. Artif Organs 6: 20–24, 2003.
13.LaPietra A, Grossi E, Pua B, et al: Assisted venous drainage presents the risk of undetected air microembolism. J Thorac Cardiovasc Surg 120: 856–863, 2000.
14.Willcox TW, Mitchell SJ, Gorman DF: Venous air in the bypass circuit: a source of arterial line emboli exacerbated by vacuum-assisted drainage. Ann Thorac Surg 68: 1285–1289, 1999.
15.Jegger D, Tevaearai HT, Mueller XM, et al: Limitations using the vacuum-assist venous drainage technique during cardiopulmonary bypass procedures. J Extra Corpor Technol 35: 207–211, 2003.
16.Jones TJ, Deal DD, Vernon JC, et al: Does vacuum-assisted venous drainage increase gaseous microemboli during cardiopulmonary bypass? Ann Thorac Surg 74: 2132–2137, 2002.
17.Assad RS, Lee FY, Bergner K, Hanley FL: Extracorporeal circulation in the isolated in situ lamb placenta: hemodynamic characteristics. J Appl Physiol 72: 2176–2180, 1992.
18.Humphries K, Sistino JJ: Laboratory evaluation of the pressure flow characteristics of venous cannulas during vacuum-assisted venous drainage. J Extra Corpor Technol 34: 111–1114, 2002.
19.Bradley SM, Hanley FL, Duncan BW, et al: Fetal cardiac bypass alters regional blood flows, arterial blood gases, and hemodynamics in sheep. Am J Physiol 263: H919–928, 1992.
20.Reddy VM, McElhinney DB, Rajasinghe HA, et al: Role of the endothelium in placental dysfunction after fetal cardiac bypass. J Thorac Cardiovasc Surg 117: 343–351, 1999.
Copyright © 2005 by the American Society for Artificial Internal Organs