Perfusion

Perfusion

http://prf.sagepub.com/

Techniques of paediatric modified ultrafiltration: 1996 survey results Edward Darling, Kathy Nanry, Ian Shearer, David Kaemmer and Scott Lawson Perfusion 1998 13: 93 DOI: 10.1177/026765919801300204 The online version of this article can be found at: http://prf.sagepub.com/content/13/2/93

Published by: http://www.sagepublications.com

Additional services and information for Perfusion can be found at: Email Alerts: http://prf.sagepub.com/cgi/alerts Subscriptions: http://prf.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://prf.sagepub.com/content/13/2/93.refs.html

>> Version of Record - Mar 1, 1998 What is This?

Downloaded from prf.sagepub.com at UNIV OF CO HEALTH SCIENCE CTR on April 11, 2013

Perfusion 1998; 13: 93–103

Techniques of paediatric modified ultrafiltration: 1996 survey results Edward Darling, Kathy Nanry, Ian Shearer, David Kaemmer and Scott Lawson Perfusion Services, Duke University Medical Center, Durham, North Carolina

In September 1996, perfusionists from 50 paediatric open-heart surgery programmes were contacted to identify centres that are currently using the technique of modified ultrafiltration (MUF). Of the 50 centres contacted, 22 (44%) were utilizing the technique. These centres were surveyed on the following: neonatal circuit description, patient entry criteria, MUF circuit description, conduct of MUF, use of extracorporeal safety devices and/or modifications, and technical complications. All 22 centres used roller pumps and membrane oxygenators. In 19 centres, MUF was utilized exclusively in the arteriovenous mode (86%), while two centres (9%) used the venovenous mode and one centre (5%) used both methods. Most (82%) of the 22 MUF centres used a blood cardioplegia system for myocardial preservation. After cardiopulmonary bypass (CPB), these blood cardioplegia systems were often converted for use as MUF circuits in a variety of ways. Other methods of accessing the CPB circuit for MUF included utilizing either a recirculation line or a dedicated port added to the circuit specifically for MUF. Blood flow rates during MUF, pump strategies, haemoconcentrator vacuum levels and endpoints were variable from centre to centre. Technical complications related to MUF were reported by 82% of the surveyed MUF centres. The most common complication, air cavitating into the circuit, was reported by 15 centres. From these data, we propose recommendations on the integration of MUF into CPB circuits, the conduct of perfusion during MUF, and appropriate safety considerations to minimize technical complications.

Address for correspondence: E Darling, Perfusion Services, Duke University Medical Center, Box 3082, Durham, NC 27710, USA.

Presented at the Annual Seminar of the American Academy of Cardiovascular Perfusion, San Diego, California, 31 January– 3 February 1997.

© Arnold 1998

0267–6591(98)PF199OA


94

E Darling et al.

Introduction Modified ultrafiltration (MUF) was introduced in 1991 by Elliott et al. at the Hospital for Sick Children in London as a valid technique in combating the rise in total body water and tissue oedema that often accompanies a paediatric cardiopulmonary bypass (CPB).1,2 Subsequent literature supports the concept that paediatric MUF reverses haemodilution, reduces total body water, improves haemodynamics, decreases the need for blood transfusions, minimizes myocardial oedema, improves cerebral metabolic recovery after circulatory arrest and improves respiratory mechanics in patients who have undergone CPB.3–8 Some of these benefits are thought to be derived from reducing serum levels of various inflammatory mediators that are generated during CPB by ultrafiltration.9–15 The reported effectiveness of MUF has led to widespread interest in the technique. In a comprehensive North American paediatric perfusion survey in 1994, Groom and associates report that 44% of North American paediatric cardiac centres use the MUF technique.16 Yet, despite two publications that addressed the technical integration of a MUF circuit with a conventional CPB circuit17,18 other aspects of MUF, including patient entry criteria, perfusion management, safety considerations and end-points have not been clearly addressed in the literature. In this paper, we report the survey results of 22 North American paediatric open-heart centres that are currently using the MUF technique. This survey places emphasis on the technical elements of MUF, resulting in general recommendations on safe, effective conduct of MUF.

Materials and methods A group of 50 paediatric open-heart centres in North America were arbitrarily selected using a list obtained from the Congenital Heart Surgeon’s Society. Perfusionists from each of these centres were contacted by telephone and asked if they used the MUF technique. Centres answering in the affirmative were then interviewed on a variety of questions relating to MUF. These questions covered:

• • • • • • • •

Annual paediatric case load. MUF entry criteria. MUF circuit description. Safety devices specifically for MUF. Years of MUF experience. Neonatal circuit description. MUF conduct and pump strategies. MUF technical complications.

To narrow the focus, questions on perfusion equipment were limited to the institution’s smallest ‘newborn’ circuitry. This survey, therefore, represents MUF approaches in smaller children.

Results Of the 50 paediatric open-heart centres contacted, 22 programmes (44%) reported using the MUF technique following paediatric CPB. There were five centres (10%) who had previously tried MUF, but have since discontinued its use. Modified ultrafiltration was performed in an arteriovenous (AV) configuration by 86% of the centres. Two centres used a venovenous (VV) configuration and one centre used both methods. The average annual paediatric caseload and average years of MUF experience are shown in Table 1. Perfusion equipment All 22 centres used membrane oxygenators and arterial roller pumps. Half of the centres used a hardshell, ‘open system’ venous reservoir, while the other half used a ‘closed system’ collapsible venous bag reservoir. Arterial-line filters were used in 86% of the centres, while 9% used a bubble trap and 5% (one centre) used neither. Most centres (82%) used blood cardioplegia for myocardial protection. Tubing size (for the neonatal patients) of the AV loop consisted of 1/4″ × 1/4″ internal diameter (ID) polyvinyl chloride (PVC) tubing in 77% of the centres. Three centres were using 3/16″ ID for the arterial line. Priming volumes ranged from Table 1 Annual case load and MUF experience of survey participants (n = 22)

Annual paediatric cases Years of MUF experience


Average

Range

247 ± 153 1.9 ± 1.1

50–550 0.1–4.5

Techniques of paediatric modified ultrafiltration: 1996 survey results 95 Circuit The programme responded to three primary MUF circuit considerations: 1) integration of MUF into the CPB circuit; 2) type of haemoconcentrator; and 3) method of re-infusion.

Figure 1 Pie chart showing patient selection criteria for MUF

400 to 1000 ml with a 22-centre average of 632 ± 166 ml. Entry criteria The criteria used to determine if MUF would be used on patients was variable among the surveyed centres (Figure 1). Nine centres (41%) used MUF on every open-heart paediatric case. A weight-based criterion was used by ten (45%) centres and one centre (5%) was using surgeon’s preference as the primary criterion. Two centres (9%) based MUF use on other criteria such as operative procedure and patient age. The weight-based criterion was distributed across a wide spectrum of sizes (Table 2). Two centres avoided using MUF on patients weighing under 4 and 5 kg, respectively, citing possible retrograde arterial-line flow problems associated with using an 8 French aortic cannula during MUF.

Table 2

MUF selection criteria based on patient weight

Number of centres

MUF weight criteria

3 2 1 1 1 1 1

All patients < 10 kg All patients < 15 kg All patients < 20 kg All patients < 25 kg All patients < 30 kg All patients > 4 kg Patients between 5–25 kg

Integration of MUF with the CPB circuit. Of the 18 centres which used a blood cardioplegia system, 11 (61%) directly converted the system for MUF use post-CPB (Figure 2). Another two programmes used the blood cardioplegia pump and heat exchanger for MUF, but accessed the arterial line at a luer site on the aortic cannula. Haemoconcentrators were predominately positioned between the cardioplegia pump and the cardioplegia system’s heat exchanger. The use of a dedicated MUF site on the arterial line was used in five (23%) institutions. These sites were usually proximal to the arterial filter via either a ‘Y’ connector or a luer site on an existing connector (Figure 3). Two centres used the oxygenator recirculation line as a point of access. Haemoconcentrators used for MUF. The Hemocor HPH400 haemoconcentrator (Minntech Corp., Minneapolis, MN, USA) was used by 73% of the centres for MUF. Three centres used the paediatric Cobe (Cobe Cardiovascular, Inc., Arvada, CO, USA) haemoconcentrator. The remaining three programmes used other haemoconcentrators such as the Gambro AB (Lund, Sweden) or Dideco (Mirandola, Italy) brands. MUF infusion line. After passing through the haemoconcentrator, concentrated blood must be returned to the patient. Of the 18 centres using blood cardioplegia systems for MUF, 15 (83%) used the cardioplegia solution table line to deliver the concentrated blood back into the right atrium (RA). A dedicated MUF infusion line was used by seven centres, including one centre which used a warming transfusion line, Hot Line (Level 1, Smiths Industries, Rockland, Maine, ME, USA), as the MUF infusion tubing. Fifteen centres (68%) used the venous cannula as the MUF infusion site. Of these centres, 13 attached the MUF infusion line to a luer on the venous cannula. The remaining two centres Y-jointed the MUF return line into the venous line with a connector.


96

E Darling et al.


Techniques of paediatric modified ultrafiltration: 1996 survey results


97

98

E Darling et al.

Seven centres (32%) preferred to remove the venous cannula during MUF. For MUF infusion in these cases, two centres used a 10 French vent catheter (DLP Inc., Grand Rapids, MI, USA) inserted into the RA after the venous cannula is removed. Three centres simply cut off the luer end of the cardioplegia table line and inserted this tubing into the RA. Two centres pulled back their retrograde cardioplegia cannula into the RA for MUF infusion. Venovenous MUF. An alternative to the more common AV MUF, is the VV approach MUF. This technique was used in three (14%) centres. Unlike AV MUF, VV MUF avoids the use of the arterial line of the CPB circuit for postoperative ultrafiltration. Blood is drawn from the venous cannula, pumped through a haemoconcentrator, and then returned back to the RA. As filling pressures drop during ultrafiltration, circuit volume can be infused via the aortic cannula to maintain appropriate patient volume status (Figure 4). Alternative MUF approach. One programme used the Blood Monitor Pump bm 11 (Baxter Healthcare Corp., Deerfield, IL, USA) as the MUF pump instead of one of the roller pumps on the heart– lung console. This device was developed for pumpassisted continuous venovenous haemofiltration or haemodiafiltration. The unit can be primed independently and then integrated with the CPB circuit when MUF is to be used. This device is equipped with safety devices, such as an air bubble detector, an arterial and venous servo-regulated pressure monitor and a blood leak detector.

Table 3

Conduct of MUF Determination of MUF pump flow rate. Responses on MUF pump flow rates could be categorized into four groups. Eleven centres (50%) used fixed flow rates (independent of patient size). Six programmes (27%) indexed MUF pump flow rates to the patient’s weight. Four programmes (18%) used arbitrary MUF flow rates based upon an individual patient’s haemodynamic stability. One centre (5%) based MUF flow upon a calculation of 10% of the patient’s normal cardiac output. Individual programme responses to flow rates are shown in Table 3. Pump strategies. All the AV MUF centres utilized a similar pump control strategy. Once the MUF pump flow is established, this pump is usually maintained at a constant flow rate. The arterial pump flow is varied to maintain an optimal filling pressure for the patient. When patient filling pressures are low, the arterial pump flow may be increased, resulting in less volume removed from the patient and greater titration from the CPB circuit into the MUF circuit. If filling pressures become excessive, the arterial pump flow may be slowed, thus removing more volume from the patient for ultrafiltration. Heightened communication between anaesthesiologist, perfusionist and surgeon concerning optimal filling pressures during MUF were reported by all centres. Most centres limit arterial pump speed to equal, but not exceeding, the MUF pump rate. This prevents antegrade/retrograde flow oscillations in the arterial line, therefore, decreasing the chance of pumping any entrained air into the patient’s arterial system.

Individual centre responses to calculated MUF flow rates

Centre

Established flow rate

Centre

Flow indexed to weight

1 2 3 4 5 6 7 8 9 10 11 Average

100–200 ml/min 150–300 ml/min 100 ml/min 70–80 ml/min 100–200 ml/min 200–250 ml/min 25 200 ml/min 200 ml/min 300 ml/min 100–125 ml/min 100 ml/min 120–196 ml/min

12 13 14 15 16 17

15–30 ml/kg/min 15–20 ml/kg/min 10–15 ml/kg/min 10 ml/kg/min 10–15 ml/kg/min 10–30 ml/kg/min

Average

12–22 ml/kg/min


Techniques of paediatric modified ultrafiltration: 1996 survey results


99

100

E Darling et al.

Table 4 Reported vacuum levels applied to haemoconcentrators during MUF Vacuum level

Number of centres

Maximum Work up to (−200) mmHg (−80) mmHg (−100) mmHg (−150) mmHg (−200) mmHg

6 (27%) 5 (22%) 1 (5%) 3 (14%) 3 (14%) 4 (18%)

Vacuum levels. The reported amount of vacuum applied to the haemoconcentrator ranged from conservative to aggressive. Five centres (22%) started at no vacuum or low vacuum and gradually increased to 200 mmHg vacuum. The remaining 17 centres (78%) applied a fixed vacuum to the haemoconcentrator during ultrafiltration (Table 4). MUF end point. Reported parameters for the termination of MUF were variable. Ten centres (45%) stopped when the circuit contents were completely salvaged. Five centres (22.5%) used a time-based end point of 10–20-min duration. One programme (4.5%) continued as long as the surgeon’s patience permitted. One centre (4.5%) used an haematocritbased end point. Four centres (18%) used a combination of the above parameters. The remaining centre (4.5%) calculated the patient’s circulating blood volume and continued MUF until twice this amount in ultrafiltrate was removed.

Technical complications Technical complications that were directly related to the MUF technique have occurred in 82% of the surveyed centres. The most common complication was air cavitating out of solution in the arterial line and it was experienced at least once by 68% of the programmes (Table 5). Patient cooling during MUF was a factor in centres not using a blood cardioplegia system and was described by 18% of surveyed centres. Other complications included MUF circuit over-pressurization and disruption due to a clamped line, accidental cardioplegia solution infusion when using a blood cardioplegia system, a clotted MUF circuit and transient patient exsanguination due to an unclamped oxygenator recirculation line during MUF. While these complications often led to the immediate

discontinuation of MUF, there were no reported incidence of patient morbidity or mortality associated with MUF. Safety devices and practices The MUF centres were asked if they incorporated any additional devices, modifications and/or practices to minimize potential complications relating to the MUF technique. These results are shown in Table 6. Several strategies were reported to reduce the potential of air cavitation due to a negative pressure developing in the arterial line during MUF. One such modification was the use of pressure servo-regulation to stop the MUF pump if a negative pressure develops in the arterial line. Many centres without servo-regulation capability reported ‘extravigilant’ monitoring of the arterial line pressure to enable an immediate response (stopping the MUF pump) in a negative pressure situation. Some centres even required an additional perfusionist during MUF to exclusively watch for a drop in arterial line pressure. Other strategies include having the surgeon maintain optimal aortic position and bypassing the arterial-line filter during MUF to lower resistance to retrograde flow. For additional protection against air emboli, the following were reported: inverting the haemoconcentrator with blood entering the top, the use of a bubble trap in the MUF circuit and appropriate positioning of the bubble detector. Significant patient cooling during MUF was reported exclusively by centres not using a blood cardioplegia system with its integral heat exchanger. To prevent patient cooling, the following was reported: pre-warming the haemoconcentrator by circulating warm blood through it before MUF use, incorporating a dedicated MUF circuit heat exchanger, using a heated transfusion line and warming the operating room. Table 5

Reported technical complications during MUF

Complication

Number of centres

Air cavitation Patient cooling Circuit disruption CPS* infusion Clotted MUF circuit Patient exsangination

15 (68%) 4 (18%) 2 1 1 1

*CPS, cardioplegia solution.


Techniques of paediatric modified ultrafiltration: 1996 survey results 101 Table 6

Reported responses on safety devices or modifications for MUF

Safety devices and modifications

Number of centres

Negative pressure servo-regulation Additional perfusionist in room Extra-vigilant arterial line pressure monitoring Bypass arterial line filter Invert haemoconcentrator Add bubble trap to MUF circuit Change position of bubble detector Pre-warm haemoconcentrator Add heat exchanger to MUF circuit Use a heating transfusion line for MUF infusion Adjust OR temperature to very warm room Positive pressure servo-regulation MUF pump Use ultrasonic flowmeter Placement of in-line HCT* sensor

3 2 10 4 4 1 1 2 1 1 1 4 2 1

*HCT, haematocrit.

To avoid accidental MUF circuit over-pressurization and disruption, positive pressure servoregulation of the MUF pump was also cited. Other modifications included the use of an ultrasonic flowmeter to determine actual flow rate and direction down the arterial line during MUF and appropriate placement of a haematocrit sensor to enable monitoring of haematocrit during MUF.

Discussion Fifty paediatric open-heart centres in the USA and Canada were contacted, 22 of which were using the technique of MUF. An interesting aspect of this survey was the reasons of the other 28 centre for not using the MUF technique. Many non-MUF centres believed that micro-prime circuits, careful cell salvaging techniques and aggressive ultrafiltration during the case can achieve similar results to that of MUF. Some centres elected not to use the MUF technique because they believed that MUF was just a current fad and that data on its effectiveness were questionable. Still others felt that their outcomes were acceptable and were unwilling to introduce change into a system that seemed to work well. Five centres said they had tried the technique at their institutions, but have since discontinued its use. Reasons given by former MUF users for discontinuation were technical difficulties, cost ineffectiveness, surgeon’s impatience and difficulty in keeping a large staff comfortable and competent using the MUF technique.

While most of the literature on MUF has focused on the benefits that this technique may have, little has been reported on the potential technical pitfalls of MUF. A striking result of this survey was that 82% of the MUF centres experienced at least one (and often multiple) significant technical complications. Considering the average MUF experience in this survey was less than two years, many of these technical complications may be attributed to the learning curve required to master this technique. Responses on the most serious and common complication, cavitation of air into the arterial line, occurred in even the more experienced centres. While withdrawing blood from the aortic cannula, rotation of the MUF pump can produce a negative pressure in the arterial line if there is high resistance to retrograde flow (e.g. kinked or obstructed aortic cannula, clamped line, excessive MUF pump flow rate). This results in air either cavitating or being drawn across the membrane oxygenator. Several centres who have experienced cavitation believe that the aortic wall was sucked down over the end of the aortic cannula causing obstruction. This danger was first described in 1994 with recommendations to servoregulate the MUF pump to a negative arterial line pressure.17 Concern over this phenomenon was evident in this survey by the reported modifications designed to minimize this threat. Surprisingly, only three centres used negative pressure servoregulation during MUF (Table 6). Groom et al. at the Fairfax Hospital in Virginia first described the use of the blood cardioplegia system for MUF after CPB.18 This was a major


102

E Darling et al.

contribution in convenience and safety as it provided to the MUF circuit: arterial line access, a heat exchanger to prevent heat loss, a pressure monitoring site and air trapping capacity. In this survey, 18 of the 22 MUF centres used a blood cardioplegia system for myocardial protection. Many converted the blood cardioplegia system for MUF application. To minimize the dead space of the circuit, some centres accessed the arterial line at a luer site on the aortic cannula connected to the inlet of the blood cardioplegia pump (see Figure 2). This method retains the benefits of using the blood cardioplegia system while allowing maximum circuit blood salvage during MUF. Caution should be exercised, however, to ensure that the cardioplegia component is removed and residual cardioplegia solution is flushed from the circuit prior to beginning MUF. The majority of MUF centres use an AV MUF configuration. This is probably due to the available literature on AV MUF and the lack of published papers on VV effectiveness. VV MUF is appealing (especially in light of the potential of air cavitation) because it does not involve retrograde flow down the arterial line and may be less problematic in patients who are haemodynamically unstable. Criticisms of the technique include losing the potential benefits of delivering oxygenated blood to the pulmonary vasculature and the recirculation that can occur when both removing and infusing blood in the RA. A recent report of VV MUF, however, demonstrated similar benefits as AV MUF.19 Although there was marked variability in patient selection for MUF in this survey, the majority of centres using a weight-based criterion did so on patients under 20 kg. The available literature seems to support the concept that MUF is most effective on smaller patients where the ratio between patient volume and extracorporeal volume/ surface area is most disparate.2 In early MUF publications, determination of the MUF pump flow rate was described in terms of a set flow rate (usually 200–300 ml/min). In a similar manner, set MUF pump flow rates are used by 50% of the centres in our survey. Because of the wide range of patient sizes in paediatric open-heart surgery, a more appropriate method of determining MUF pump flow rate may be to normalize MUF flow rate by indexing it to body weight. This avoids the potential problems of flowing too high in small patients (increasing the likelihood of producing a

negative arterial line pressure) and flowing too slowly in larger patients (extending the time it takes to MUF effectively). Indexing MUF flow to body weight was first suggested by our group17 and is currently being used by 27% of MUF centres (Table 3). Determining a MUF end point is an issue to be considered. A nearly unanimous response from all MUF centres was the surgeon’s impatience during the procedure. Rushing through MUF, however, may not be ideal as more and more data suggest that the removal of deleterious blood factors by the haemoconcentrator triggering the inflammatory response may play a part of the technique’s effectiveness.9–16 A balanced approach of effective circuit salvaging and an adequate time frame for removal of these factors is warranted. Although this survey was not all inclusive, we believe that it accurately describes a reliable snapshot of MUF as it is currently being practised in North America. This survey shows that MUF is a technique that can be adapted to the CPB circuit in many ways. The survey also illuminates that the possibility of technical complications relating to MUF can occur. Programmes wishing to begin MUF at their institutions should consider visiting an experienced MUF centre with a similar bypass circuit for observation of practices and techniques.

References 1 Naik SK, Knight A, Elliott MJ. A successful modification of ultrafiltration for cardiopulmonary bypass in children. Perfusion 1991; 6: 41–50. 2 Naik SK, Knight A, Elliott MJ. A prospective randomized study of a modified technique of ultrafiltration during pediatric open-heart surgery. Circulation 1991; 84: 422–31. 3 Naik SK, Balaji S, Elliott MJ. Modified ultrafiltration improves hemodynamics after cardiopulmonary bypass in children. J Am Coll Cardiol 1992; 19: 37A. 4 Elliott MJ. Ultrafiltration and modified ultrafiltration in pediatric open heart operations. Ann Thorac Surg 1993; 56: 1518–22. 5 Naik SK, Elliott MJ. Ultrafiltration and modified ultra-filtration in pediatric cardiopulmonary bypass. Perfusion 1993; 8: 101–12. 6 Skaryak LA, Kirshbom PM, DiBernardo LR et al. Modified ultrafiltration improves cerebral metabolic recovery after circulatory arrest. J Thorac Cardiovasc Surg 1995; 109: 744–52. 7 Gaynor JW, Tulloh R, Owen CH et al. Modified


Techniques of paediatric modified ultrafiltration: 1996 survey results 103

8

9

10

11

12

ultrafiltration reduces myocardial edema and reverses hemodilution following cardiopulmonary bypass in children. J Am Coll Cardiol 1995; 25: 200A. Meliones JM, Gaynor JW, Wilson BG et al. Modified ultrafiltration reduces airway pressures and improves lung compliance after congenital heart surgery. J Am Coll Cardiol 1995; 25: 271A. El Habbal MH, Smith L, Strobel S, Elliott MJ. Modified ultrafiltration after cardiopulmonary bypass for repair of ventricular septal defect reduces serum IL8. Circulation 1993; 88(9) part 2: I-96. Andreasson S, Gothberg S, Berggren H, Bengtsson A, Eriksson E, Risberg B. Hemofiltration modifies complement activation after extracorporeal circulation in infants. Ann Thorac Surg 1993; 56: 1515–17. Millar AB, Armstrong L, van der Linden J et al. Cytokine production and hemofiltration in children undergoing cardiopulmonary bypass. Ann Thorac Surg 1993; 56: 1499–502. Journois D, Pouard P, Greeley W, Mauriat P, Vouhe P, Safran D. Hemofiltration during cardiopulmonary bypass in pediatric cardiac surgery. Effects on hemostasis, cytokines, and complement components. Anesthesiology 1994, 81: 1181–89.

13 Journois D, Israel-Biet D, Pouard P et al. High volume, zero-balanced hemofiltration to reduce delayed inflammatory response to cardiopulmonary bypass in children. Anesthesiology 1996; 85: 965–76. 14 Saatvedt K, Linberg H, Geiran OR et al. Ultrafiltration after cardiopulmonary bypass in children: effects on hemodynamics, cytokines and complement. Cardiovasc Res 1996; 31: 596–602. 15 Wang MJ, Chui IS, Hsu CM et al. Efficacy of ultrafiltration in removing inflammatory mediators during pediatric cardiac operations. Ann Thorac Surg 1996; 61: 651–56. 16 Groom RC, Hill AG, Kurusz M et al. Paediatric perfusion practices in North America: an update. Perfusion 1995; 10: 393–401. 17 Darling EM, Shearer IR, Nanry K et al. Modified ultrafiltration in pediatric cardiopulmonary bypass. J ExtraCorpor Technol. 1994; 26: 205–209. 18 Groom RC, Akl BF, Albus RA, Hill A, Munoz R, Lefrak EA. Alternative method of ultrafiltration after cardiopulmonary bypass. Ann Thorac Surg 1994; 58: 573–74. 19 Li CM, Alfieris GM, Walker MJ et al. Modified venovenous ultrafiltration in infant cardiac surgery. J Am Coll Card 1995; 25: 200A.
