Review Article

World Journal for Pediatric and

Congenital Heart Surgery
Ultrafiltration in Pediatric Cardiac 2019, Vol. 10(6) 778-788
ª The Author(s) 2019
Article reuse guidelines:
Surgery Review sagepub.com/journals-permissions
DOI: 10.1177/2150135119870176
journals.sagepub.com/home/pch

Joel Bierer, MD1, Roger Stanzel, PhD, CPC2,

Mark Henderson, CPC, CCP2, Suvro Sett, MD1,
and David Horne, MD1

Abstract
Introduction: The use of cardiopulmonary bypass in pediatric cardiac surgery is associated with significant inflammation, fluid
overload, and end-organ dysfunction yielding morbidity and mortality. For decades, various intraoperative ultrafiltration tech-
niques such as conventional ultrafiltration, modified ultrafiltration (MUF), zero-balance ultrafiltration (ZBUF), and combination
techniques (ZBUF-MUF) have been used to mitigate these toxicities and promote improved postoperative outcomes. However,
there is currently no consensus on the ultrafiltration technique or strategy that yields the most benefit for infants and children
undergoing open heart surgery. Methods: A librarian-conducted PubMed literature search from 1990 to 2018 yielded 90 clinical
studies or publications on the various forms of ultrafiltration and the impact on physiologic markers and clinical outcomes. All
publications were reviewed, summarized, and conclusions synthesized. The data sets were not combined for systematic or meta-
analysis due to significant heterogeneity in study protocols and patient populations. Results: Modified ultrafiltration significantly
promotes improved myocardial function, reduction in fluid overload, and reduced bleeding and transfusion complications. Fur-
thermore, ZBUF has shown a consistent reduction in inflammatory cytokines and improved pulmonary function and compliance.
There is conflicting evidence that MUF, ZBUF, and ZBUF-MUF culminate in reduced ventilation time and intensive care unit stay.
Conclusion: Various modes of ultrafiltration have been shown to be associated with improved physiologic function or clinical
outcomes in pediatric cardiac surgery. There are some inconsistent trial results that can be explained by heterogeneity in
ultrafiltration, clinical staff preferences, and institution protocols. Ultrafiltration has some essential benefit as it is ubiquitously
used at pediatric heart centers; however, the optimal protocol could be yet identified.

Keywords
congenital, pediatric, cardiac, open-heart, cardiovascular surgery, modified, conventional, zero-balance, ultrafiltration,
cardiopulmonary, bypass, inflammation

Submitted May 08, 2019; Accepted July 18, 2019.

Introduction hypothesized to be a result of MUF’s superior ability to

remove fluid, concentrate coagulation factors, and extract
Children undergoing heart surgery requiring cardiopulmonary
inflammatory cytokines. This translated into some observa-
bypass (CPB) are subject to significant hemodilution and sys-
tions of improved clinical outcomes as cardiopulmonary func-
temic inflammatory response from circuit exposure.1,2 These
tion was optimized by reduced tissue edema.5 Furthermore,
insults culminate in capillary leak syndrome, associated tissue
reduced blood loss and transfusion requirements were thought
edema, and organ dysfunction leading to increased morbidity
to be related to coagulation factor and fibrinogen
and mortality.2 In the 1980s, conventional ultrafiltration (CUF)
was developed to remove excess fluid during the rewarming
phase of CPB. Conventional ultrafiltration, postoperative diur- 1
Division of Cardiac Surgery, Department of Surgery, Dalhousie University,
esis, and even peritoneal dialysis did not significantly resolve Halifax, Nova Scotia, Canada
2
the fluid overload state, morbidity, or mortality. In 1991, Naik Department of Clinical Perfusion, Nova Scotia Health Authority, Halifax,
et al developed modified ultrafiltration (MUF) facilitating large Nova Scotia, Canada
volumes of filtrate to be removed in an efficient manner, which
Corresponding Author:
seemed to improve postoperative clinical outcomes.3,4 David Horne, IWK Health Centre, 2nd Floor Children’s Site, PO Box 9700,
Improved cardiopulmonary function, reduced postopera- Halifax, Nova Scotia, Canada B3K 6R8.
tive bleeding, and diminished inflammatory syndromes were Email: david.horne@iwk.nshealth.ca
Bierer et al 779

of volume removal, or requirements of additional fluid to the

Abbreviations and Acronyms CPB circuit which can be counterproductive.
A-V arterial-venous
CPB cardiopulmonary bypass
CUF conventional ultrafiltration
CVP central venous pressure Modified Ultrafiltration
ET1 endothelin-1 Modified ultrafiltration is used after the patient is weaned from
ICU intensive cares unit
CPB, negating the limitation of maintaining venous reservoir
IL interleukin
LOS length of stay minimum volume as in CUF.1 Modified ultrafiltration removes
LV left ventricular blood directly from the patient, flowing through the ultrafiltra-
MAP mean arterial pressure tor, then directly returns the concentrated volume to the patient.
MUF modified ultrafiltration It was first described and still commonly used in an A-V (aorta/
PVR pulmonary vascular resistance arterial to right atrium/venous) configuration (see Figure 2A).
RBCs red blood cells Due to concerns about air embolism, and coronary and cerebral
TNF tumor necrosis factor
steal syndromes with this configuration, V-A MUF was devel-
V-A venous-arterial
ZBUF zero-balance ultrafiltration oped to mitigate these effects (see Figure 2B). Modified ultra-
filtration target end points include an approximate 15-minute
time point, hematocrit of approximately 40%, or depletion of
concentration. 5 However, conflicting evidence emerged the venous reservoir. The ultrafiltration capacity can be altered
regarding the impact of MUF on inflammation and even by manipulating the flow or vacuum pressure on the filtrate
improvements in clinical outcomes. 5 To this end, zero- collection. Because both the A-V or V-A MUF system can be
balance ultrafiltration (ZBUF) was developed in the late isolated from the CPB circuit, blood volume can be indepen-
1990s allowing for larger absolute blood volumes to be fil- dently transfused from the CPB circuit in the usual fashion to
trated, with more consistent evidence showing cytokine maintain satisfactory hemodynamics after weaning from CPB.
removal.6,7 These ultrafiltration techniques and current sup-
porting evidence for different physiological effects and clinical
outcomes are the focus of this review.
Zero-Balance Ultrafiltration
Zero-balance ultrafiltration has a similar configuration to CUF.
Methods While on CPB, blood is shunted postoxygenator through an
ultrafiltrator membrane returning concentrated blood to the
A librarian-conducted literature search on PubMed and venous reservoir and thus again not isolated from the bypass
PubMed Clinical Queries from 1990 to 2018 yielded 90 clinical circuit. However, unlike CUF, ZBUF replaces the volume of
studies or publications investigating the various forms of ultra- ultrafiltrate removed with the same amount of crystalloid back
filtration and the impact on physiologic markers and clinical into the venous reservoir. Therefore, significantly larger
outcomes. All publications were reviewed and summarized, volumes of blood can be ultrafiltered without running into
including 43 clinical trials as well as basic science investiga- circuit volume contraction.6,9 The higher absolute volume of
tions and previous reviews. The data sets were not combined blood filtered correlated with increased mass of cytokine and
for systematic or meta-analysis due to significant heterogeneity noxious substance removal.6,9,10 The optimal rate of fluid turn-
in the study protocol and patient populations. This review is a over in the ZBUF method is unclear.
narrative describing intraoperative ultrafiltration techniques
and perceived benefits or pitfalls and identifies important areas
for future research.
Combination Techniques
To benefit from each technique’s strength, they can be com-
Results bined within the same operation. Zero-balance ultrafiltration–
MUF has ZBUF within the CPB circuit as previously
Conventional Ultrafiltration described, and MUF in its dedicated circuit isolated during
Conventional ultrafiltration is traditionally done during the CPB and opened after weaning as previously described. Simi-
rewarming phase while on CPB.1 Postoxygenator, some blood larly, CUF-MUF can be combined using two independent cir-
is shunted through an ultrafiltration membrane extracting dis- cuits. However, original combination models required physical
cardable filtrate by hydrostatic pressure. A common ultrafilter manipulation and alteration of the circuits, which added com-
membrane pore size is 65 kDa.8 Concentrated blood is then plexity and risk of an air embolism to the procedure. To this
returned to the venous reservoir (see Figure 1). Continuous end, the Halifax method or simple modified ultrafiltration pub-
volume removal with CUF will lead to volume depletion in lished in 2000 popularised a V-A MUF configuration which
the venous reservoir which is incompatible with CPB. This also allowed for CUF or ZBUF without circuit alteration (see
limitation necessitates intermittent use of CUF, slower rates Figure 3).1,11
780 World Journal for Pediatric and Congenital Heart Surgery 10(6)

Figure 1. Conventional ultrafiltration (CUF) schematic. Tubing clamp denotes position to exclude CUF from the bypass circuit. Dotted line
denotes deoxygenated blood.

Coagulation Unfortunately, there are relatively few studies looking at ultra-

filtration impact on coagulation factors relative to the many
Normal homeostasis and the avoidance of thrombosis or bleed-
studies looking at cytokine and inflammatory mediator
ing relies on the balance of procoagulant and anticoagulant
concentration.
factors.12 Exposure to CPB both dilutes clotting factors and
Compared to control, the MUF concentration of platelets,
stimulates complement-activated inflammation, endothelial
fibrinogen, and some coagulation factors also translates to sig-
dysfunction, and the extrinsic coagulation pathway.12 These nificantly reduced RBC transfusions of between 50% and 80%
cause CPB-associated coagulopathy. At birth, healthy neonates (see Table 1).1,17,24,28–30 However, there is disagreement
have roughly 50% of the concentrations of factor II, VII, IX, X, regarding other clinical outcomes including postoperative
XI, XII, antithrombin III, protein C, protein S, and plasmino- bleeding or chest tube output, as there are four studies3,24,28,30
gen.12 Because of the relatively small intravascular volume of a that suggest reduced bleeding with MUF over controls and four
neonate or low-weight infant, there is significant hemodilution studies17,29,31,32 that indicate no difference (see Table 1). It
of platelets by 70% and coagulation factors by 50% with should be noted that studies which demonstrate decreased
exposure to CPB prime volume.13 This dilution disrupts the bleeding with MUF are driven by results from patients who
coagulation balance. Small patients with prematurity, ductal- are smaller than 10 kg, subjected to hypothermia, or required
dependent circulation, reduced cardiac output, hepatic dysfunc- periods of low and intermediate CPB flow.1,28 There seems to
tion, and sepsis are at even higher risk of hemorrhagic or be limited benefit with MUF, with regard to reduced bleeding
thrombotic complications after open-heart surgery with complications, in larger children. While there are very few
CPB.12,14 studies looking at platelet or plasma transfusions, Bando et al
A common MUF end point target is hemoconcentration by demonstrated that ZBUF-MUF required less platelet and
achieving a hematocrit of 40%. There is a substantial body of plasma transfusions than CUF.33 However, Williams et al
evidence that MUF can hemoconcentrate to this effect.1,14–24 found no difference in a similar comparison.34 A recent retro-
Furthermore, there is evidence that higher flow or longer dura- spective study, which argued miniaturized circuits circumvent
tion of MUF significantly improves the hemoconcentration the utility of MUF, showed that CUF-MUF had significantly
capabilities.25,26 Red blood cells (RBCs), platelets, prothrom- more chest tube output, cryoprecipitate or fresh frozen plasma
bin (70 kDa), fibrinogen (340 kDa), and albumin (65 kDa) are use, and platelet transfusion. Furthermore, another recent ran-
larger than common ultrafiltrator membrane pore sizes of 60 to domized trial showed CUF-MUF had significantly less RBC
65 kDa and have been shown to be concentrated by MUF; transfusions than CUF, but at the expense of more platelet
however, smaller coagulation factors such as factor IX (55 transfusions.27,35
kDa) and X (38 kDa) were not.8,16 Two other studies also Since its inception, one of the most significant benefits of
supported significant fibrinogen concentration with MUF.14,27 MUF includes its ability to hemoconcentrate and preserve
Bierer et al 781

Figure 2. Modified ultrafiltration (MUF) schematic. (A) Arterial-venous (A-V) MUF and (B) venous-arterial (V-A) MUF. Tubing clamps are
engaged to exclude the bypass circuit after the patient is weaned to allow for MUF. Dotted line denotes deoxygenated blood.

coagulation factors to reduce bleeding and transfusion require- in children undergoing open-heart surgery. Specifically, small
ments. 5,36 Although MUF is well studied in this regard, coagulation factors such as thrombin (39 kDa), IX (55 kDa), X
ZBUF-MUF has much less literature or evidence with regard (38 kDa) might be susceptible to depletion with ZBUF strate-
to recovery of hemodilution coagulopathy brought on by CPB gies. However, this seems to be a theoretical risk as no study
782 World Journal for Pediatric and Congenital Heart Surgery 10(6)

Figure 3. Simplified modified ultrafiltration (SMUF) schematic. Tubing clamp is engaged to exclude the bypass circuit after the patient is weaned
to allow for MUF. During bypass, the venous line would be left open to allow for cardiopulmonary bypass as well as ultrafiltration. Dotted line
denotes deoxygenated blood.

Table 1. Bleeding and Transfusion by MUF and Control.a

Control MUF RBC Control RBC

Study Design Patients MUF Bleeding Bleeding P Transfusion Transfusion P

Naik et al, 19911 RCT 50 12.5 mL/kg/24 h 19.0 mL/kg/24 h <.0002 3.0 mL/kg/24 h 15.5 mL/kg/24 h <.0002
Ad et al, 199628 RCT 80 11.0 mL/kg/24 h 26.8 mL/kg/24 h <.05 7.4 mL/kg/24 h 20.2 mL/kg/24 h <.05
Wang et al, 199817 RCT 40 NR NR NS 7.8 mL/kg 21.5 mL/kg <.05
Chew et al, 200231 RCT 18 0.67 mL/kg/24 h 1.09 mL/kg/24 h NS 6.5 mL/kg/24 h 5.9 mL/kg/24 h NS
Draaisma et al, 199724 Retro 198 20.1 mL/kg 29.1 mL/kg <.05 24.0 mL/kg 27.7 mL/kg <.05
Koutlas et al, 199730 Retro 120 8.4 mL/kg/24 h 16.0 mL/kg/24 h <.01 29 mL/kg/24 h 63 mL/kg/24 h <.001
Kameyama et al, 200029 Retro 100 19.8 mL/kg/24 h 16.0 mL/kg/24 h NS 47 mL/kg/24 h 113 mL/kg/24 h <.01
Kotani et al, 200832 Retro 36 28 mL/kg/24 h 44 mL/kg/24 h NS NR NR NR
Abbreviations: MUF, modified ultrafiltration; NR, not recorded; NS, nonsignificant result, no P value given; RBC, red blood cell; RCT, randomized controlled trial;
Retro, retrospective trial.
a
Controls used no ultrafiltration.

has shown an increased risk of bleeding complications with A large retrospective study of 1,540 patients supported this
ZBUF protocol. finding, as children with positive fluid balance following open
heart surgery experienced a significantly increased risk of
mortality (odds ratio ¼ 1.73).37 It was most significant in
Fluid Balance neonates and small infants <6 kg; in fact, those with positive
Excess fluid volume acquired during open-heart surgery with balance were twice as likely to die as those with zero or
CPB and subsequent morbidity and mortality were the indica- negative balance.37 However, it should be noted that extreme
tions to utilize ultrafiltration in this patient population. Sources negative balance more than 40 mL/kg had significantly
of fluid accumulation include fluid overload by preoperative higher mortality than those with balances between 40 mL/
heart failure, perioperative intravenous fluids, CPB prime, car- kg and 0 mL/kg.37 It is unclear whether there is a causal
dioplegia, operative field irrigation fluid, intravenous medica- relationship between positive fluid balance and mortality or
tion administration, and blood product resuscitation. Fluid whether the positive fluid balance is due to resuscitation for
balance is calculated by recording all the fluid inputs and sub- poor perioperative cardiopulmonary function related to the
tracting outputs such as urine volume and surgical blood loss. primary cardiac pathology. All randomized studies on various
Bierer et al 783

Table 2. Ultrafiltrate Mean Volumes by Technique. inflammatory syndrome of cardiopulmonary dysfunction,

vasodilation, and increased total oxygen consumption.50,51
Ultrafiltrate Mean
Ultrafiltration Type Studies Volume Range (mL/kg)
Neonates, small infants under 1.8kg, ductal-dependent
defects, cyanotic defects, reduced ventricular function and
CUF 7 20–373 pre-existing multi-organ dysfunction are all risk factors for
MUF 20 7.5–143.0 increased CPB-associated morbidity.52
ZBUF 3 74–196 There have been many clinical studies which monitor vari-
CUF-MUF 5 34–352
ous cytokines and inflammatory markers relative to ultrafiltra-
ZBUF-MUF 4 150–261
tion techniques. No investigation was able to demonstrate a
Abbreviations: CUF, conventional ultrafiltration; MUF, modified ultrafiltration; significant reduction in endotoxin, IL-1, IL-6, IL-8, or TNF-a
ZBUF, zero-balance ultrafiltration. when MUF was compared to control.17,31,45 However, Wang
et al demonstrated that high flow and long duration MUF was
ultrafiltration techniques show that a wide range of volume able to remove significantly more IL-6, but not TNF-a, than
can successfully be taken off, but only a few consider the less intensive MUF protocol.18 Wang et al additionally showed
ultrafiltrate volume in relation to overall operative fluid bal- MUF was not able to remove IL-6, IL-8, TNF-a, or elastase
ance (see Table 2).4,6,10,17-20,27,28,33,34,38-47 when compared to CUF.4 However, Sever et al demonstrated
It is unclear whether these ultrafiltrate volumes have any that CUF-MUF significantly reduced IL-8 when compared to
clinical implications. Only five studies used ultrafiltration end CUF alone.53 There is a larger body of evidence regarding
points targeted to net fluid balances: ultrafiltrate volume target cytokine removal by ZBUF techniques. Two studies showed
30% to 130% of total fluid input,3,6 ultrafiltrate volume target ZBUF was able to remove significant amounts of IL-6, IL-8,
80% to100% of total prime volume,38,48 and ultrafiltrate vol- and TNF-a when compared to control.21,45 Further studies
ume target equal to cardioplegia volume plus 40 to 70 mL/kg.49 observed additional benefit of significantly reduced IL-1 and
Modified ultrafiltration end points almost always are a time C3a, decreased ventilation, and even intensive care unit (ICU)
point between 10 and 20 minutes, preset hematocrit goal, or length of stay (LOS) compared to control.6,10 Some studies
depletion of the venous reservoir; there doesn’t seem to be a found that complement C3a and C5a rise and fall with exposure
consideration of net fluid balance. Conventional ultrafiltration to CPB, but, although isolated in the ultrafiltrate, MUF did not
end point is when rewarming is complete. Therefore, the sub- significantly alter the kinetics of these factors when compared
stantial lack of ultrafiltrate volume target, which seeks to to control.45,54 Yndgaard et al found a significant reduction in
achieve a net zero or negative operative fluid balance, could endotoxin with MUF.55
be one reason why there is such heterogeneity in the outcomes Although MUF does not seem useful in removing cytokines,
of ultrafiltration studies. ZBUF appears to have more favorable evidence in the removal
of IL-1, IL-6, IL-8, TNF-a, and C3a. Interestingly, the two
studies that demonstrated significant cytokine augmentation
Inflammation also showed decreased ventilation and even ICU LOS com-
Cardiopulmonary bypass is associated with a systemic pared to control.6,10 The conclusion was that ZBUF is more
inflammatory response and patients typically develop signif- effective, relative to other ultrafiltration techniques or controls,
icant hypotension with a distributive pattern, various degrees at removing these inflammatory mediators and dampening the
of vasoplegia, and pulmonary dysfunction which is exagger- severity of CPB-associated acute pulmonary injury.10,56 Most
ated over the reaction expected from surgical trauma.2 There ultrafilters used have size cutoffs around 60 to 65 kDa.8 Tumor
appears to be three pro-inflammatory stimuli unique to open- necrosis factor a, in its soluble form, is 17 kDa, with a 51 kDa
heart surgery and CPB. First, the nonendothelialized trimeric form; IL-1a and IL-1b are 17 kDa, and finally, IL-8, as
synthetic circuit activates proteases, and subsequently, the well as, both C3a and C5a are 10 kDa.57–60 Therefore, these
alternate complement pathway producing C3a and C5a.2,50 factors should theoretically all be removed by ultrafiltration
Secondly, hypoxemia and tissue ischemia experienced during and theoretically control the CPB-induced inflammatory
CPB leads to cellular death and release of pro-inflammatory syndrome.
signals.2,50,51 Thirdly, endotoxemia, from Gram-negative
bacteremia that opportunistically translocate the intestinal
lining during relative gut ischemia associated with CPB, con-
Myocardial Function
tributes significantly to the systemic inflammatory Prevention and reversal of myocardial edema from cross-clamp
response.2,50,51 These processes stimulate action of nuclear and arrest ischemia, CPB-induced inflammation and fluid over-
factor-kappa B (NF-kB), particularly in leukocytes, and load are hypothesized to be one of the dominant benefits of
thereby production of pro-inflammatory cytokines interleukin ultrafiltration. Compared to control, patients randomized to
1 (IL-1), IL-6, IL-8, and tumor necrosis factor a (TNF-a) in MUF had significantly improved systolic performance,15 left
a positive feedback loop.50 The end product is leukocyte ventricular (LV) mean ejection pressure,46 preload-recruitable
extravasation, increased endothelial permeability, and inter- stroke work index,46 contractility,61 LV diastolic compliance,46
stitial edema, which clinically presents as the CPB-associated and LV wall thickness. 46,61 Nonrandomized studies also
784 World Journal for Pediatric and Congenital Heart Surgery 10(6)

demonstrated increases in cardiac index, 43 stroke volume In addition to removing volume by MUF, ZBUF-MUF stra-
index,43 myocardial performance index,22 and fractional short- tegies appear to be correlated with improved pulmonary func-
ening. 22 However, some randomized studies detected no tion by the removal of endothelin-1 (ET1) and other
improvement in cardiac index and LV stroke index45 as well inflammatory cytokines.40,49 Endothelin-1 is a small, 21 amino
as no change in diastolic performance.15 Again, there is hetero- acid polypeptide with pulmonary vasoconstrictive properties,
geneity in patient populations and ultrafiltration protocols, which is well-documented in many disorders involving pul-
which is perhaps why there is a discrepancy in results. Further- monary hypertension.49 Bando et al showed that ET1 levels
more, the two negative studies that did not show improved were significantly reduced by over 50% with ZBUF-MUF
cardiac function with MUF over control were both small and compared to control and then temporally correlated this with
could be underpowered. decreased pulmonary pressure to systemic pressure ratios
Honjo et al used an own control study design where patients which were sustained into the postoperative period.49 Simi-
had echocardiography done before and after MUF.22 They larly, Hiramatsu et al showed that ZBUF-MUF could signifi-
found improved global ventricular function as measured by cantly reduce the ET1 levels which correlated with
myocardial performance index as well as fractional shortening considerably reduced pulmonary vascular resistance (PVR),
with MUF.22 Interestingly, the magnitude of benefit was nega- which again was sustained into the postoperative period.40 This
tively correlated with length of CPB and aortic cross-clamp group also detected a positive correlation between PVR and
time.22 Overall, this study, in addition to the study by Chatur- ET1. There seems to be convincing evidence that pulmonary
vedi et al, showed no change in any diastolic function para- physiology can be most enhanced by the use of ZBUF-MUF to
meter which adds to the polarity in the literature as Davies et al remove both excess fluid and ET1 and other inflammatory
found improved diastolic performance.15,22,46 Improvement in mediators. Unfortunately, although there are many studies sug-
global ventricular function by MUF was thought to be related gesting improvement in pulmonary physiologic parameters
to reduced myocardial edema as a decrease in ventricular wall with MUF or ZBUF-MUF, they often fail to demonstrate con-
thickness has been observed.46,61 Additional benefits of MUF sistent clinical relevance between studies (see
with direct relation to cardiac function include the potential Table 3).3,4,6,9,10,15,17,21,27,29,30–34,37–39,42,45–47,52,61
filtration of unidentified myocardial depressant factors, such Measuring clinical outcomes such as ventilation time and
as TNF-a and IL-1b, and increased oxygen delivery by ICU LOS can be confounded by factors such as surgeon, per-
hemoconcentration.22,50 fusionist, or intensivist preference on therapy choices and insti-
As described, MUF along with other forms of ultrafiltration tutional ICU protocols. The lack of ultrafiltration
significantly improve hemodynamics in the postoperative standardization can explain the variation in clinical results. A
period. Although there is some ambiguity, there is evidence heuristic review between studies that had significant and non-
that these observations can be partially attributed to improved significant ventilation or ICU LOS reductions revealed that
ventricular function. To the authors’ knowledge, there is no longer total ultrafiltration time and larger volume removed,
current literature relating the quantity of ultrafiltrate removed, as seen with ZBUF-MUF, are most likely to have substantial
and therefore, the relationship to overall fluid balance and the benefit compared to control. It should be noted that although
magnitude of myocardial enhancement. many studies demonstrated no difference between various
forms or combinations of ultrafiltration, the comparators usu-
ally have similar volumes of ultrafiltrate, which might also
explain the nonsignificant findings seen in many studies. It
Pulmonary Function appears that ZBUF is particularly useful in promoting
Acute pulmonary injury is a common manifestation of CPB- improved pulmonary function as filtration occurs from the
mediated inflammatory response, third spacing, and fluid over- beginning to the end of the inflammatory insult of CPB.
load leading to increased ventilatory support and ICU LOS.
Modified ultrafiltration is postulated to attenuate this response
by removing volume load and thereby reducing hydrostatic Hemodynamics
pressure and pulmonary edema. Many studies have consis- Improved hemodynamic parameters with ultrafiltration has
tently shown that MUF, with or without ZBUF, increases static been one of the main and well-recognized benefits since its
pulmonary compliance21,26,38 and dynamic pulmonary compli- inception. The removal of excess fluid and inflammatory
ance,21,26,38 while decreasing the A-a gradient,6,31,38 pulmon- cytokines have long since been the hypothesized rationale
ary to systemic arterial pressure ratio,40,49 respiratory index,29 of improved cardiopulmonary function with prevention of
and airway resistance.9 In comparison, there is only a single the inflammatory distributive and vasoplegic syndrome
study which shows no improvement in pulmonary compliance, related to CPB. When MUF is used, compared to control,
pulmonary resistance, or A-a gradient when ZBUF-MUF was studies show significantly improved systolic blood pres-
compared to ZBUF or MUF individually.34 There appears to be sure,3,18–20,22,25,26,32,43,61 diastolic blood pressure,3,25,32,43,61,62
agreement from randomized studies indicating that ZBUF and and mean arterial pressure (MAP).23,25,31,43,45 No study offers
MUF together or separately have a positive impact on pulmon- evidence which contradicts these findings. Disappointingly,
ary physiologic parameters. only one of these studies comparing MUF to control revealed
 Table 3. Ventilation Time, ICU Length of Stay, and Inotrope Requirements Using Different Ultrafiltration Techniques.a
Ventilation Time (hours) ICU LOS (days) Inotrope Requirement

Study Design Patients MUF Control P MUF Control P MUF Control P

Naik et al, 19913 RCT 50 NR NR NS NR NR NS

Saatvedt et al, 199645 RCT 18 6 6 NS 3 4 NS
Wang et al, 199817 RCT 40 NR NR NS NR NR NS
Davies et al, 199746 RCT 21 24 26 NS 4 4 NS 156 mg/kg/24 h 865 mg/kg/24 h .03
Keenan et al, 200021 RCT 38 140 90 .08 10.0 6.3 .1 Epinephrine 53% Epinephrine 47% .07
Dobutamine 18% Dobutamine 13% .07
Amrinone 47% Amrinone 27% .02
Chew et al, 200231 RCT 18 NR NR NS
Chaturvedi et al, 199915 Pro 22 NR NR NS
Kotani et al, 200832 Pro 36 52.3 54.3 NS 4.2 5.8 <.05 80 hours 95 hours NS
Koutlas et al, 199730 Retro 120 21 22 NS 4.6 6.1 NS
Kameyama et al, 200029 Retro 100 23.8 50.6 <.05 8.6 mg/kg/min 6.7 mg/kg/min NS

MUF CUF MUF CUF MUF CUF

Wang et al, 19964 RCT 50 NR NR NS NR NR NS NR NR NS

Thompson et al, 200148 RCT 110 60.0 62.4 .82 5.0 4.6 .21
Takabayashi et al, 200763 Retro 30 86% patients 66% patients NS

CUF-MUF CUF CUF-MUF CUF CUF-MUF CUF

39
Maluf et al, 2003 RCT 41 95.7 94.8 .98 6.5 7.0 .80
Sever et al, 200453 RCT 27 19.8 50.6 .017 1.6 3.4 .014
Mahmoud et al, 200538 RCT 40 12 13 .4 1.7 1.6 .9 NR NR NS
Ziyaeifard et al, 201647 RCT 46 12.3 34.3 .004 4.4 7.4 .007 Milrinone 36.6 hours Milrinone 64.1 hours .04
Epinephrine 72.7 hours Epinephrine 131.6 hours .001
Dopamine 62.4 hours Dopamine 81.6 hours .239
Dobutamine 26.0 hours Dobutamine 86.4 hours .002
McRobb et al, 201727 Retro 160 Pediatric 23.0 Pediatric 17.8 .39 Pediatric 3.8 Pediatric 3.3 .34 Pediatric VIS 4.3 Pediatric VIS 3.5 .38
Neonatal 69.0 Neonatal 65.0 .79 Neonatal 12.2 Neonatal 9.8 .40 Neonatal VIS 12.0 Neonatal VIS 9.9 .25

ZBUF Control ZBUF Control ZBUF Control

10
Liu et al, 2007 RCT 30 7.8 9.6 <.05 1.2 1.3 NS

ZBUF-MUF Control ZBUF-MUF Control ZBUF-MUF Control

9
Huang et al, 2003 RCT 30 9.3 8.7 >.05 2.0 2.5 <.05

ZBUF-MUF MUF ZBUF-MUF MUF ZBUF-MUF MUF

Journois et al, 19966 RCT 20 10.8 28.2 .02

ZBUF-MUF CUF ZBUF-MUF CUF ZBUF-MUF CUF

49
Bando et al, 1998 RCT 24 68 178 <.05 NR NR NS

ZBUF-MUF MUF or Control ZBUF-MUF MUF or Control ZBUF-MUF MUF or Control

42
Ming et al, 2001 RCT 80 NR NR NS NR NR NS

Abbreviations: CUF, conventional ultrafiltration; ICU, intensive cares unit; LOS, length of stay; MUF, modified ultrafiltration; NR, not recorded; NS, nonsignificant result and no P value given; Pro, prospective
trial; RCT, randomized controlled trial; Retro, retrospective trial; VIS, vasoactive–inotropic score with increasing score indicating severe illness; ZBUF, zero-balance ultrafiltration.
a

785
Controls used no ultrafiltration. Blank field indicates outcome was not studied.
786 World Journal for Pediatric and Congenital Heart Surgery 10(6)

a significant reduction in inotrope use,46 while four others circuit and preventing acute pulmonary injury. Although ultra-
showed no difference.3,21,31,32 Finally, comparison of MUF filtration more often does not yield good evidence of physio-
and control looking at heart rate reductions reveals two studies logic improvement, especially improved cardiopulmonary
that show significant difference19,20 and two that show no function, there does not always seem to be a consistent
difference46,61; three studies that show significant central enhancement of postoperative outcomes in the literature. The
venous pressure (CVP) reductions22,32,45 and three that showed discrepancies could be explained by the heterogeneity of the
no difference.19,25,43 It should be noted that those studies studies regarding patient populations, ultrafiltration protocols,
which observed significant hemodynamic improvement with myocardial protection strategies, bypass protocols, ICU proto-
MUF only showed benefit lasting a few hours into the post- cols, and differences in expert opinion. The majority of publi-
operative period. cations occurred prior to 2010 as only three clinical studies
Although the evidence illustrating MUF can improve systo- have been published since that time. Specifically, one rando-
lic and diastolic pressure when compared to controls, there is mized trial that showed reduced ventilation time and ICU LOS
less information to suggest one form of ultrafiltration is super- with MUF, a second randomized trial that showed reduced
ior to another in this regard. A single study directly comparing RBC transfusion but increased platelet transfusion with MUF,
MUF and CUF demonstrated a significant increase in systolic, and one retrospective study that showed no clinical bene-
diastolic, MAP, and reduced heart rate with MUF, but no dif- fit.27,35,47 There seems to be paucity without clinical practice
ference in inotrope use.63 When CUF-MUF is compared to consensus. Despite the noted polarity in the literature, there
CUF, there appears to be increased systolic blood pressure,47 seems to be some inherent benefit as some form of ultrafiltra-
diastolic pressure, 47 MAP,47,53 reduced heart rate,53 and tion utilized in 97% of pediatric heart surgery centers assessed
CVP47,53 without contradicting studies. Despite improved by survey in 2011.64 This review can serve as a basis for future
hemodynamic profiles, only one study showed reduced ino- research, which would more precisely characterize how ultra-
trope use with CUF-MUF over CUF47 against three studies that filtration augments the inflammatory and coagulation cascades
showed no difference.27,38,39 Finally, one study compared to ultimately define an ideal protocol. Identifying and standar-
ZBUF-MUF to ZBUF or MUF alone and found significantly dizing the optimal ultrafiltration technique could lead to sig-
higher systolic and diastolic pressures with ZBUF-MUF but no nificant reductions in morbidity and mortality for infants and
difference in inotrope use.34 Again, those studies which children undergoing open-heart surgery with CPB.
showed significant improvements were transient and matched
comparator groups within a few hours into the postoperative Authors’ Note
period. Authors had full control of the study design, methods, and production
It is clear from the available literature that ultrafiltration, of written report.
and specifically MUF, offers significant hemodynamic support.
Unfortunately, postoperative clinical benefit has not been con- Declaration of Conflicting Interests
sistently demonstrated and, furthermore, it is less obvious
The author(s) declared no potential conflicts of interest with respect to
which ultrafiltration technique might yield the optimal post- the research, authorship, and/or publication of this article.
operative outcome. Moreover, the heterogeneity in ultrafiltra-
tion protocol between studies makes it difficult to compare
Funding
study results and may explain the discrepancies in clinical out-
comes. Enhancement in postoperative hemodynamics with The author(s) received no financial support for the research, author-
ultrafiltration has been hypothesized to be a result of improved ship, and/or publication of this article.
cardiopulmonary function through reduction in tissue edema,
as well as the removal of vasodilation-inducing inflammatory References
cytokines. Further characterization of the inflammatory sub- 1. Naik SK, Knight A, Elliott MJ. A successful modification of
stances stimulated by CPB and those that can be filtered as ultrafiltration for cardiopulmonary bypass in children. Perfusion.
mentioned previously could explain and prevent cardiopul- 1991;6(1): 41-50.
monary injury as well as the distributive syndrome often 2. Butler J, Rocker GM, Westaby S. Inflammatory response to car-
observed after pediatric open-heart surgery. diopulmonary bypass. Ann Thorac Surg. 1993;55(2): 552-559.
3. Naik S, Knight A. A prospective randomized study of a modified
technique of ultrafiltration during pediatric open-heart surgery.
Comments Circulation. 1991;84(suppl 5): III422-III431.
Ultrafiltration has been used for decades to dampen the CPB- 4. Wang MJ, Chiu IS, Hsu CM, et al. Efficacy of ultrafiltration in
induced toxic responses and reduce bleeding complications, removing inflammatory mediators during pediatric cardiac oper-
inotrope use, ventilation time, and ICU LOS. Modified ultra- ations. Ann Thorac Surg. 1996;61(2): 651-656.
filtration appears to have relative superiority regarding cardiac 5. Elliot M. Modified ultrafiltration and open heart surgery in chil-
function related to fluid removal and reducing bleeding com- dren. Paediatr Anaesth. 1999;9(1): 1-5.
plications. Zero-balance ultrafiltration seems to have the cap- 6. Journois D, Israel-Biet D, Pouard P, et al. High-volume, zero-
abilities of decreasing inflammatory responses to the bypass balanced hemofiltration to reduce delayed inflammatory response
Bierer et al 787

to cardiopulmonary bypass in children. Anesthesiology. 1996; during surgery for children with ventricular septal defects: a pilot
85(5): 965-976. study. Perfus (United Kingdom). 2018;33(3): 203-208.
7. Tallmann RD, Dumond M, Brown D. Inflammatory mediator 24. Draaisma AM, Hazekamp MG, Frank M, Anes N, Schoof PH,
removal by zero-balance ultrafiltration during cardiopulmonary Huysmans HA. Modified ultrafiltration after cardiopulmonary
bypass. Perfusion. 2002;17(2): 111-115. bypass in pediatric cardiac surgery. Ann Thorac Surg. 1997;
8. Hemocor HPH® Hemoconcentrators. http://www.medivators. 64(2): 521-525.
com/sites/default/files/50090235F_Hemocor.pdf. Accessed 25. Rodriguez RA, Ruel M, Broecker L, Cornel G. High flow rates
March, 2019. during modified ultrafiltration decrease cerebral blood flow velo-
9. Huang H, Yao T, Wang W, et al. Continuous ultrafiltration attenu- city and venous oxygen saturation in infants. Ann Thorac Surg.
ates the pulmonary injury that follows open heart surgery with 2005;80(1): 22-28.
cardiopulmonary bypass. Ann Thorac Surg. 2003;76(1): 136-140. 26. Türköz A, Tunçay E, Balci ŞT, et al. The effect of modified
10. Liu J, Ji B, Long C, Li C, Feng Z. Comparative effectiveness of ultrafiltration duration on pulmonary functions and hemody-
methylprednisolone and zero-balance ultrafiltration on inflamma- namics in newborns and infants following arterial switch oper-
tory response after pediatric cardiopulmonary bypass. Artif ation. Pediatr Crit Care Med. 2014;15(7): 600-607.
Organs. 2007;31(7): 571-575. 27. McRobb CM, Ing RJ, Lawson DS, Jaggers J, Twite M. Retro-
11. Myers GJ, Leadon RB, Mitchell LB, Ross DB. Simple modified spective analysis of eliminating modified ultrafiltration after
ultrafiltration. Perfusion. 2000;15(5): 447-452. pediatric cardiopulmonary bypass. Perfus (United Kingdom).
12. Jaggers J, Lawson JH. Coagulopathy and inflammation in neona- 2017;32(2): 97-109.
tal heart surgery: mechanisms and strategies. Ann Thorac Surg. 28. Ad N, Snir E, Katz J, Birk E, Vidne BA. Use of the modified
2006;81(6): 2360-2366. technique of ultrafiltration in pediatric open-heart surgery: a pro-
13. Kern FH, Morana NJ, Sears JJ, Hickey PR. Coagulation defects in spective study. Isr J Med Sci. 1996;32(12): 1326-1331.
neonates during cardiopulmonary bypass. Ann Thorac Surg. 29. Kameyama T, Ando F, Okamoto F, et al. The effect of modified
1992;54: 541-546. ultrafiltration in pediatric open heart surgery. Ann Thorac Cardi-
14. Friesen RH, Campbell DN, Clarke DR, Tornabene MA. Modified ovasc Surg. 2000;6(1): 19-26.
ultrafiltration attenuates dilutional coagulopathy in pediatric open 30. Koutlas TC, Gaynor JW, Nicolson SC, Steven JM, Wernovsky G,
heart operations. Ann Thorac Surg. 1997;64(6): 1787-1789. Spray TL. Morbidity after cavopulmonary connection. Ann
15. Chaturvedi RR, Shore DF, White PA, et al. Modified ultrafiltra- Thorac Sur. 1997;4975(97): 37-43.
tion improves global left ventricular systolic function after open- 31. Chew MS, Brix-Christensen V, Ravn HB, et al. Effect of modified
heart surgery in infants and children. Eur J Cardiothorac Surg. ultrafiltration on the inflammatory response in paediatric open-
1999;15(6): 742-746. heart surgery: a prospective, randomized study. Perfusion. 2002;
16. Ootaki Y, Yamaguchi M, Oshima Y, et al. Effects of modified 17(5): 327-333.
ultrafiltration on coagulation factors in pediatric cardiac surgery. 32. Kotani Y, Honjo O, Osaki S, et al. Effect of modified ultrafiltra-
Surg Today. 2002;32(3): 203-206. tion on postoperative course in neonates with complete transposi-
17. Wang W, Huang HM, Zhu DM, Chen H, Su ZK, Ding WX. tion of the great arteries undergoing arterial switch operation. Circ
Modified ultrafiltration in pediatric cardiopulmonary bypass. Per- J. 2008;72(9): 1476-1480.
fusion. 1998;13(5): 304-310. 33. Bando K, Turrentine MW, Vijay P, et al. Effect of modified
18. Wang W, Zhu DM, Huang HM, et al. Effect of flow rate, negative ultrafiltration in high-risk patients undergoing operations for con-
pressure, and duration of modified ultrafiltration on hemody- genital heart disease. Ann Thorac Surg. 1998;66(3): 821-827;
namics and inflammatory mediators. ASAIO J. 2007;53(1): 41-45. discussion 828.
19. Schlünzen L, Pedersen J, Hjortholm K, Hansen OK, Ditlevsen E. 34. Williams GD, Ramamoorthy C, Chu L, et al. Modified and con-
Modified ultrafiltration in paediatric cardiac surgery. Perfusion. ventional ultrafiltration during pediatric cardiac surgery: clinical
1998;13(2): 105-109. outcomes compared. J Thorac Cardiovasc Surg. 2006;132(6):
20. Gaynor JW, Kuypers M, van Rossem M, et al. Haemodynamic 1291-1298.
changes during modified ultrafiltration immediately following the 35. Milovanovic V, Bisenic D, Mimic B, et al. Reevaluating the
first stage of the Norwood reconstruction. Cardiol Young. 2005; importance of modified ultrafiltration in contemporary pediatric
15(1): 4-7. cardiac surgery. J Clin Med. 2018;7(12): 1-13.
21. Keenan HT, Thiagarajan R, Stephens KE, Williams G, Rama- 36. Ungerleider RM. Effects of cardiopulmonary bypass and use of
moorthy C, Lupinetti FM. Pulmonary function after modified modified ulftrafiltration. Ann Thorac Surg. 1998;65(6 suppl): 35-39.
venovenous ultrafiltration in infants: a prospective, randomized 37. Grist G, Whittaker C, Merrigan K, et al. The correlation of fluid
trial. J Thorac Cardiovasc Surg. 2000;119(3): 501-507. balance changes during cardiopulmonary bypass to mortality in
22. Honjo O, Osaki S, Kotani Y, Akagi T, Sano S. Diagnosis-based pediatric and congenital heart surgery patients. J Extra Corpor
differences in response of global ventricular performance to mod- Technol. 2011;43(4): 215-226.
ified ultrafiltration in children. Circ J. 2010;74(1): 86-92. 38. Mahmoud AS, Burhani MS, Hannef AA, et al. Effect of modified
23. Jia Z, Teng Y, Liu Y, Wang H, Li Y, Hou X. Influence of high- ultrafiltration on pulmonary function after cardiopulmonary
flow modified ultrafiltration on brain oxygenation and perfusion bypass. Chest. 2005;128(5): 3447-3453.
788 World Journal for Pediatric and Congenital Heart Surgery 10(6)

39. Maluf MA. Modified ultrafiltration in surgical correction of con- 51. Bronicki RA, Hall M. Cardiopulmonary bypass-induced inflam-
genital heart disease with cardiopulmonary bypass. Perfusion. matory response: pathophysiology and treatment. Pediatr Crit
2003;18(suppl 1): 61-68. Care Med. 2016;17(8): S272-S278. doi:10.1097/PCC.000000
40. Hiramatsu T, Imai Y, Kurosawa H, et al. Effects of dilutional and 0000000759.
modified ultrafiltration in plasma endothelin-1 and pulmonary 52. Shen I, Giacomuzzi C, Ungerleider RM. Current strategies for
vascular resistance after the Fontan procedure. Ann Thorac Surg. optimizing the use of cardiopulmonary bypass in neonates and
2002;73(3): 862-865. infants. Ann Thorac Surg. 2003;75(2): S729-734.
41. Pearl JM, Manning PB, McNamara JL, Saucier MM, Thomas 53. Sever K, Tansel T, Busaran M, et al. The benefits of continuous
DW. Effect of modified ultrafiltration on plasma thromboxane ultrafiltration in pediatric cardiac surgery. Scand Cardiovasc J.
B2, leukotriene B4, endothelin-1 in infants undergoing cardiopul- 2004;38(5): 307-311.
monary bypass. Ann Thorac Surg. 1999;68(4): 1369-1375. 54. Andreasson S, Göthberg S, Berggren H, Bengtsson A, Eriksson E,
42. Ming ZD, Wei W, Hong C, Wei Z, Xiang DW. Balanced ultra- Risberg B. Hemofiltration modifies complement activation after
filtration, modified ultrafiltration, and balanced ultrafiltration extracorporeal circulation in infants. Ann Thorac Surg. 1993;
with modified ultrafiltration in pediatric cardiopulmonary bypass. 56(6): 1515-1517.
J Extra Corpor Technol. 2001;33(4): 223-226. 55. Yndgaard S, Andersen LW, Andersen C, Petterson G, Baek L.
43. Ricci Z, Polito A, Netto R, et al. Assessment of modified ultra- The effect of modified ultrafiltration on the amount of circulating
filtration hemodynamic impact by pressure recording analytical endotoxins in children undergoing cardiopulmonary bypass. J
method during pediatric cardiac surgery. Pediatr Crit Care Med. Cardiothorac Vasc Anesth. 2000;14(4): 399-401.
56. Chew MS. Does modified ultrafiltration reduce the systemic
2013;14(4): 390-395.
inflammatory response to cardiac surgery with cardiopulmonary
44. Ziyaeifard M, Alizadehasl A, Massoumi G. Modified ultrafiltra-
bypass? Perfusion. 2004;19: S57-S60.
tion during cardiopulmonary bypass and postoperative course of
57. Palladino MA, Bahjat FR, Theodorakis EA, Moldawer LL. Anti-
pediatric cardiac surgery. Res Cardiovasc Med. 2014;3(2):
TNF-a therapies: the next generation. Nat Rev Drug Discov.
e17830. doi:10.5812/cardiovascmed.
2003;2(9): 736-746.
45. Saatvedt K, Lindberg H, Geiran OR, et al. Ultrafiltration after
58. Watanabe N, Kobayashi Y. Selective release of a processed form
cardiopulmonary bypass in children: effects on hemodynamics,
of interleukin 1a. Cytokine. 1994;6(6): 597-601.
cytokines and complement. Cardiovasc Res. 1996;31(4):
59. Brat DJ, Bellail AC, Van Meir WG. The role of interleukin-8 and
596-602.
its receptors in gliomagenesis and tumoral angiogenesis. Neuro
46. Davies M, Nguyen K, Gaynor J, Elliott M. Modified ultrafiltration
Oncol. 2005;7(2): 122-133.
improves left ventricular systolic function in infants after cardio-
60. Zhou W. The new face of anaphylatoxins in immune regulation.
pulmonary bypass. J Thorac. 1997;115(2): 361-370.
Immunobiology. 2012;217(2): 225-234.
47. Ziyaeifard M, Alizadehasl A, Aghdaii N, et al. The effect of
61. Rivera ES, Kimball TR, Bailey WW, Witt SA, Khoury PR,
combined conventional and modified ultrafiltration on mechani- Daniels SR. Effect of veno-venous ultrafiltration on myocardial
cal ventilation and hemodynamic changes in congenital heart performance immediately after cardiac surgery in children. A
surgery. J Res Med Sci. 2016;21(8): 1-7. prospective randomized study. J Am Coll Cardiol. 1998;32(3):
48. Thompson LND, McElhinney DB, Findlay P, et al. A prospective 766-772.
randomized study comparing volume-standardized modified and 62. Gaynor JW. Use of ultrafiltration during and after cardiopulmon-
conventional ultrafiltration in pediatric cardiac surgery. J Thorac ary bypass in children. J Thorac Cardiovasc Surg. 2001;122(2):
Cardiovasc Surg. 2001;122(2): 220-228. 209-211.
49. Bando K, Vijay P, Turrentine MW, et al. Dilutional and modified 63. Takabayashi S, Shimpo H, Yokoyama K, Iwata H. Relationship
ultrafiltration reduces pulmonary hypertension after operations between increased blood pressure and hematocrit during modified
for congenital heart disease: a prospective randomized study. J ultrafiltration for pediatric open heart surgery. Gen Thorac Car-
Thorac Cardiovasc Surg. 1998;115(3): 517-527. diovasc Surg. 2007;55(1): 12-18.
50. Paparella D, Yau TM, Young E. Cardiopulmonary bypass 64. Harvey B, Shann KG, Fitzgerald D, et al. International pediatric
induced inflammation: pathophysiology and treatment. An perfusion practice: 2011 survey results. J Extra Corpor Technol.
update. Eur J Cardiothorac Surg. 2002;21(2): 232-244. 2012;44(4): 186-193.