Kuratani et al Congenital Heart Disease

Modified versus conventional ultrafiltration in pediatric cardiac

surgery: A meta-analysis of randomized controlled trials comparing
clinical outcome parameters
Norifumi Kuratani, MD, PhD,a Piyaporn Bunsangjaroen, MD,b Thanaphon Srimueang, MD,b
Eiji Masaki, MD, PhD,c Takaaki Suzuki, MD, PhD,d and Toshiyuki Katogi, MD, PhDd

Objective: Although previous studies have demonstrated that modified ultrafiltration improves laboratory pa-
rameters in pediatric cardiac surgery, the clinical outcome data have been inconsistent. We performed
a meta-analysis of randomized controlled trials comparing modified versus conventional ultrafiltration.
Methods: We conducted a comprehensive search of the literature to identify clinical trials that met our inclusion
criteria. To be included, studies had to be prospective randomized trials that compared modified ultrafiltration
and conventional ultrafiltration in pediatric cardiac surgery using cardiopulmonary bypass. We focused on the
following outcome variables: hematocrit and mean arterial blood pressure after cardiopulmonary bypass,
amount of chest tube drainage after surgery, time to extubation, and length of stay in the intensive care unit.
The random effects model was used to determine the pooled effect estimates. The estimators of treatment effects
were expressed as the weighted mean difference with 95% confidence intervals. The heterogeneity of collected
data was also evaluated.
Results: We screened 54 studies, 8 of which satisfied our inclusion criteria. Combined analysis revealed that
modified ultrafiltration resulted in significantly higher postbypass hematocrit and higher mean arterial blood
pressure. Benefits in postoperative blood loss, ventilator time, and intensive care unit stay were not apparent.
There was significant heterogeneity among the studies surveyed.
Conclusions: The advantage of modified ultrafiltration over conventional ultrafiltration consists of significant
improvement of clinical conditions in the immediate postbypass period. The postoperative outcome parameters
were not significantly influenced. We should also take into account possible clinical or methodologic variations
in the currently available ultrafiltration studies. (J Thorac Cardiovasc Surg 2011;142:861-7)

CHD
modiluted, and cooled or those who have experienced
Earn CME credits at lengthy CPB times, are at increased risk for the develop-
http://cme.ctsnetjournals.org ment of many of the complications associated with CPB.
Because hemodilution by CPB priming is significant in pe-
diatric patients, ultrafiltration to remove excess water can
counteract the deleterious effects of CPB. Because conven-
Although cardiopulmonary bypass (CPB) is essential in pe- tional ultrafiltration (CUF), which is carried out while CPB
diatric cardiac surgery, it is widely known that CPB itself is running, offers limited filtration efficiency, modified ul-
can contribute to the development of significant morbidity trafiltration (MUF) was introduced1 and has gained popu-
after surgery. Children, especially those who are small, he- larity over the past 2 decades. Because MUF is performed
immediately after the termination of CPB, MUF removes
From the Division of Pediatric Anesthesia,a International University of Health and excess fluid with greater efficiency than CUF. The expected
Welfare Hospital, Tochigi, Japan; Department of Anesthesiology,b Khon Khan benefits of MUF include a reduction of total body water and
University, Khon Khan, Thailand; Division of Dento-oral Anesthesiology,c Tohoku
University Graduate School of Dentistry, Miyagi, Japan; and Department of Pedi-
the removal of inflammatory mediators, both of which con-
atric Cardiac Surgery,d Saimata International Medical Center, Saitama, Japan. tribute to restoring normal organ function and improving
This work was funded solely from the departmental resources of the International outcome.
University of Health and Welfare and Saitama Medical University.
Disclosures: Authors have nothing to disclose with regard to commercial support.
The potential benefits of MUF have been supported by
Received for publication Aug 3, 2010; revisions received Feb 16, 2011; accepted for numerous laboratory and clinical studies. Although MUF
publication April 4, 2011; available ahead of print May 6, 2011. is decidedly beneficial in terms of certain laboratory param-
Address for reprints: Norifumi Kuratani, MD, PhD, Division of Pediatric Anesthesia,
International University of Health and Welfare Hospital, 537-3 Iguchi, Nasushio-
eters, the available data on clinical outcomes are insufficient
bara, Tochigi 329-2763 Japan, Tel: þ81-287-37-2221, Fax: þ81-287-39-3001 and inconsistent. It is worth considering that MUF has some
(E-mail: [email protected]). potential disadvantages, including possible technical com-
0022-5223/$36.00
Copyright Ó 2011 by The American Association for Thoracic Surgery
plications. The question has therefore been raised as to
doi:10.1016/j.jtcvs.2011.04.001 whether MUF should be performed on all pediatric patients

The Journal of Thoracic and Cardiovascular Surgery c Volume 142, Number 4 861
 Congenital Heart Disease Kuratani et al

the final decision was made by the referee author (N.K.). Data abstraction
was also performed independently by 2 authors (P.B., T.S.) using standard-
Abbreviations and Acronyms
ized data collection forms. Inasmuch as the proposed clinical advantages of
CPB ¼ cardiopulmonary bypass MUF include hemoconcentration, reduced blood loss, and improvement of
CI ¼ confidence interval cardiovascular and respiratory function,3 we focused on the following out-
CUF ¼ conventional ultrafiltration come variables: hematocrit and blood pressure after CPB, amount of chest
ICU ¼ intensive care unit tube drainage within 48 hours after surgery, time to extubation, and length
of stay in the intensive care unit (ICU). Morbidities attributed to MUF and
MUF ¼ modified ultrafiltration
mortalities as a result of any cause were also collected. The clinical data,
WMD ¼ weighted mean difference expressed as mean  standard deviation, were extracted from each article.
When the standard error was reported, we determined the standard devia-
tion as the standard error multiplied by the square root of the number of
subjects. Variables that were not reported numerically were estimated by
who have undergone CPB. The MUF circuit has an artificial extrapolating data from the published figures. When the median data
surface that can elicit additional inflammatory responses. were reported, the mean and standard deviation were estimated by assum-
MUF requires additional time (typically, 15-20 minutes) af- ing that the mean was equivalent to the median and that the standard devi-
ation was half of the median value.
ter the termination of CPB, and it incurs additional costs. To All statistical analyses were performed using RevMan 4.2.10 (The Co-
support the rational application of MUF to pediatric cardiac chrane Collaboration, Oxford, United Kingdom). The random effects
patients, we need unbiased data regarding its clinical pa- model was used to determine the pooled effect estimates. The estimators
rameters and outcomes. of treatment effects were expressed as the weighted mean difference
Meta-analysis is a statistical tool that can be used to eval- (WMD) with 95% confidence intervals (CIs). Because eligible studies
showed clinical and methodologic diversity, the heterogeneity of collected
uate published data in both qualitative and quantitative data was assessed using a homogeneity test based on the c2 test and I2. The
ways, accounting for variations in characteristics that can I2 statistic was used to assess the impact of heterogeneity on the results.4
influence the overall estimate of outcomes of interest. The This statistic indicates the percentage of variability in effect estimates
statistical aggregation of randomized trials through meta- that is due to heterogeneity rather than sampling error.4 Owing to the
analysis allows for increased statistical power in detecting low power of this test, especially when trials have a small sample size or
are few in number, we determined a minimum cutoff P value of .10 and
potential differences in clinical outcomes. In this report, I2 value of 50% as a threshold of homogeneity to avoid false negative
we present a meta-analysis of randomized controlled trials results; P <.10 and I2 > 50% indicated heterogeneity and the combined
intended to clarify the clinical benefit of MUF in pediatric results should therefore be interpreted with caution.
cardiac surgery.
CHD

RESULTS
METHODS Screening Process and Study Selection
We conducted a systematic review according to the Quality of Reporting Using electronic databases and manual search, we ini-
of Meta-analyses recommendations developed to improve the quality of
tially identified 54 articles for review. Of those, 14 studies
meta-analyses.2
A comprehensive search of the literature was performed using MED- were excluded in the primary screening inasmuch as they
LINE, the American College of Physicians Journal Club database, the were unrelated studies or review articles and 1 was a case
Cochrane Database of Systematic Reviews, the Cochrane Central Register series report. The other 40 articles were thoroughly checked
of Controlled Trials, and the Database of Abstracts of Reviews of Effects. to ensure that they met our inclusion criteria and 25 studies
The following text searches and search headings were used individually
were excluded because they did not. The clinical outcome
and in combination: ‘‘modified ultrafiltration,’’ ‘‘cardiac surgery,’’
‘‘child,’’ ‘‘infant,’’ and ‘‘cardiopulmonary bypass.’’ A manual search of data that we focused on were not available in 6 studies.
references listed in reports and reviews was also performed. Only articles One published article that was retracted later by the authors
written in English were included. The date of the most recent search was was excluded from the final analysis. Thus, 8 studies5-12
May 23, 2010. Every effort was made to find studies that reported the were identified through our defined search strategy that ful-
clinical outcome data comparing MUF versus CUF in pediatric cardiac
filled the inclusion criteria in that they contained the neces-
surgery using CPB. To be included in our analysis, studies had to be
prospective randomized trials comparing MUF and CUF in pediatric car- sary data for the planned comparison. The process of
diac surgery using CPB. Studies that compared MUF and control patients identifying eligible studies is illustrated in Figure 1.
without any ultrafiltration technique were not included, because the bene-
fits of ultrafiltration in pediatric CPB are widely acknowledged and CPB
Description of Studies
management without any ultrafiltration would not reflect actual clinical
practice. Two authors (P.B., T.S.) independently assessed each article to The details of selected trials are summarized in Table 1.
ensure that it met the aforementioned inclusion criteria. Disagreements In total, 438 patients were studied, including 232 CUF pa-
were resolved by consensus and the final decision was made by the referee tients and 206 MUF patients. In 5 studies,5,6,8,10,12 MUF
author (N.K.). group patients also underwent CUF or diluted ultrafiltration
Unmasked quality assessments on the selected published studies were during rewarming periods of CPB. Three studies5,6,12 con-
carried out by 2 investigators (P.B., T.S.) on composite aspects of study
quality (5 aspects in total, each scored as 0 or 1: randomization, compara-
cluded that MUF offered favorable clinical outcomes
bility, standardized CPB management protocol, standardized postoperative whereas the other 5 studies7-11 reported no clinically signif-
care, withdrawals). Differences in opinion were settled by consensus and icant difference between MUF and CUF. Of the 8 studies,

862 The Journal of Thoracic and Cardiovascular Surgery c October 2011

Kuratani et al Congenital Heart Disease

2 different ultrafilters, we combined the outcome data of

the 2 ultrafilters together for the meta-analysis.

Mortality, Morbidity, and Technical Complications

Only 1 study8 reported technical problems related to
MUF. Williams and coworkers8 reported 2 cases of early
termination of MUF owing to significant hypotension.
Bando and colleagues5 declared no complications related
to MUF in their study. Other studies did not mention the
technical issues related to MUF. Four studies5,7,8,11) re-
ported the overall mortality and morbidity of study patients.
Bando and associates5 reported 1 postoperative death in the
MUF group. The patient died of low cardiac output after an
arterial switch repair that was followed by 5 days of extra-
corporeal circulatory support. Wang and coworkers7
reported that 1 patient in the MUF group and 1 in the
FIGURE 1. Meta-analysis flow chart. RCT, Randomized controlled trial. CUF group died of cardiac failure and could not be weaned
from CPB. Two late deaths in the MUF group were reported
in the study by Williams and associates.8 Berdat and col-
1 study5 used venovenous MUF and the others used arterio- leagues11 reported 1 death in the CUF group and several
venous MUF.6-12 In 3 studies,9-11 the fluid volume of ultra- complications in both groups. In the studies we selected
filtration was described in the study protocol. The amount of for analysis, no mortalities attributed to MUF were
ultrafiltrate was reported in 6 studies,5,7,9-12 and 4 stud- reported.
ies5,7,10,12 reported that greater amounts of filtrate were ob-
tained in MUF group patients. The median study quality of Combined Analysis
the selected trials was 3.5 (range, 2-5). Williams and asso- Postbypass hematocrit (Figure 2, A). The hematocrit
ciates8 compared CUF and MUF with and without CUF. data for the postbypass period were reported in 5 stud-
The clinical outcome data of MUF with CUF were adopted ies,5-7,9,12 which included a total of 317 patients. All studies,

CHD
for our analysis. In the study by Berdat and colleagues,11 except that reported by Wang’s group,7 reported the hemat-
2 different brands of ultrafilters were used. Inasmuch as ocrit value immediately after the termination of CPB or
they failed to find any significant differences between the MUF. Wang and associates7 reported the hematocrit value
TABLE 1. Summary of the studies included in the meta-anlaysis
Ultrafiltration Patient age* Study
Study (no. of patients) (mo) quality Technical complication Remarks
Bando et al5 CUF (N ¼ 50) 30.1  42.2 5 No MUF-related complication ‘‘Complex’’ congenital heart
surgery only.
DUF þ (v-v) MUF (N ¼ 50) 17.7  20.7 One death in MUF group
Server et al6 CUF (N ¼ 14) 12.94  12.98 2 Not reported
CUF þ (a-v) MUF (N ¼ 13) 9.38  1.94
Wang et al7 CUF (N ¼ 26) 43.6  33 3 Not reported One patient in each group died of
(a-v) MUF (N ¼ 24) 62.16  46.44 cardiac failure.
William et al8 CUF (N ¼ 19) 2.0  2.2 5 Two cases, MUF terminated early Two late deaths in MUF group
CUF þ (a-v) MUF (N ¼ 21) 2.9  3.45 because of hypotension
Thompson et al9 CUFy (N ¼ 67) 9.0  11.3 2 Not reported
(a-v) MUFy (N ¼ 43) 12.6  14.1
Mahmoud et al10 CUF (N ¼ 20) 11.8  3.3 3 Not reported
CUF þ (a-v) MUFy (N ¼ 20) 13.1  4.1
Berdat et al11 CUFy (N ¼ 21) 23.4  16.2 5 Not reported The data of 2 types of ultrafilters
were combined for analysis.
(a-v) MUFy (N ¼ 20) 17.3  16.6 One death in CUF, other
complications in each group
Aggarwal et al12 CUF (N ¼ 15) 33.6  13.9 4 Not reported
CUF þ (a-v) MUFy (N ¼ 15) 30  20.8
CUF, Conventional ultrafiltration; DUF, dilutional ultrafiltration; MUF, modified ultrafiltration; v-v, venovenous; a-v, arteriovenous. *Values are given as mean  standard
deviation. yThe filtration volume was standardized.

The Journal of Thoracic and Cardiovascular Surgery c Volume 142, Number 4 863
 Congenital Heart Disease Kuratani et al
CHD

FIGURE 2. Meta-analysis of clinical outcome parameters in MUF compared with CUF. Effect sizes of MUF are represented by the weight mean difference,
shown as black diamonds. Horizontal lines represent the lower and upper limits of the 95% confidence intervals. The open diamond indicates the pooled
result. MUF, Modified ultrafiltration; CUF, conventional ultrafiltration; WMD, weighted mean difference; CI, confidence interval; Ref, reference.

at the end of the operation. We treated these values as post- nificant heterogeneity was also revealed (c2 ¼ 34.16; df ¼ 4
bypass hematocrit data. Combined analysis revealed that [P <.00001]; I2 ¼ 88.3%).
MUF resulted in significantly higher postbypass hematocrit Postbypass mean arterial blood pressure (Figure 2,
levels (WMD ¼ 6.27; 95% CI, 3.45-9.09; P<.0001). Sig- B). The arterial blood pressure data in the postbypass

864 The Journal of Thoracic and Cardiovascular Surgery c October 2011

Kuratani et al Congenital Heart Disease

period were available in 3 studies.6,9,12 Aggarwal and DISCUSSION

associates12 reported the systolic and diastolic pressure In this meta-analysis, evidence from currently available
data and we calculated MAP using the following equation: randomized controlled studies regarding ultrafiltration in
pediatric cardiac surgery revealed that MUF augmented
MAP ¼ diastolic pressureþ1=3 hemoconcentration and facilitated the restoration of circula-
ðsystolic pressure  diastolic pressureÞ tion, as compared with CUF. However, postoperative out-
come parameters, including chest tube drainage, ventilator
The pooled results showed a significant improvement in time, and ICU stay, were not significantly influenced by
systemic blood pressure favoring the MUF group, with MUF. These findings suggest that MUF could contribute
a WMD of 9.18 (95% CI, 2.27-16.09; P ¼ .009). The het- to improving the clinical conditions immediately after
erogeneity was statistically significant (c2 ¼ 4.79; df ¼ 2 CPB, although its impact on the overall clinical outcome
[P ¼ .09]; I2 ¼ 58.3%). might not be significant.
Chest tube drainage (Figure 2, C). The mean amount of Hematocrit levels after the termination of CPB were sig-
chest tube drainage was reported in 4 trials.5,6,9,12 Bando,5 nificantly higher in MUF patients than in CUF patients. The
Sever,6 and their associates reported the amount of blood higher hematocrit in MUF patients reflects the higher effi-
loss during the first 24 hours only. Because the amount of ciency of hemoconcentration in MUF compared with
chest tube output on the second postoperative day is usually CUF. High hematocrit levels after bypass can help reduce
small, we considered these data sufficient for inclusion in the need for transfused blood and thereby offer the signifi-
our analysis. If an article reported the amount of chest tube cant benefit of minimizing homologous blood exposure.
drainage in milliliters, the data were converted to milliliter Of the 8 studies we identified in this meta-analysis, 2 stud-
per kilogram using the mean body weight data reported. ies5,6 showed a reduction in blood transfusion in MUF pa-
Of the 4 studies reporting chest tube drainage data, 35,6,12 tients compared with those who had CUF only.
(2 of which reported statistically significant results6,12) con- Our analysis demonstrated that MUF patients showed
cluded that decreasing blood loss through the chest tube had higher systemic blood pressure after CPB. This higher sys-
a favorable effect on the overall outcome. Thompson and co- temic blood pressure reflects the augmented recovery of the
workers9 reported significantly increased blood loss in the circulatory system in MUF patients. Hypothermic CPB
MUF group but did not discuss any potential reasons for it. with cardiac arrest is an extremely unphysiologic condition
The pooled analysis of these 4 studies failed to identify a sta- for the circulatory system. Myocardial edema resulting

CHD
tistically significant difference between MUF and the con- from hemodilution and increased vascular permeability
trol group (WMD ¼ 1.78; 95% CI, 8.93 to 5.36; contributes to myocardial dysfunction after CPB. In the
P ¼ .62). Statistical heterogeneity was found between trials dysfunctional heart, myocardial thickness and decreased
(c2 ¼ 52.33; df ¼ 3 [P<.00001]; I2 ¼ 94.3%). systolic function are often observed by ultrasound examina-
Duration of mechanical ventilation (Figure 2, D). All tion in the postbypass period.13 Previous studies have illus-
studies,5,6,8-12 except that reported by Wang and associ- trated that MUF reduces the edema of the myocardium and
ates,7 reported the duration of mechanical ventilation after facilitates the restoration of normal myocardial func-
surgery. The postoperative ventilatory management and ex- tion.13,14 Another possible cause of higher blood pressure
tubation criteria were mentioned in 45,8,10,12 of 7 studies. after MUF could be decreased concentrations of anesthetics
Two articles5,6 reported that MUF significantly shortened owing to the filtration process. Hodges and colleagues15
the postoperative ventilatory time compared with CUF. measured plasma anesthetic concentration after MUF and
Other studies reported no significant difference between showed that the plasma concentration of fentanyl remained
CUF and MUF patients. Combined analysis indicated that stable throughout ultrafiltration. They concluded that the
there was no difference between MUF and the control higher blood pressure in the MUF group was not likely a re-
group in terms of ventilation time (WMD ¼ 3.24; 95% sult of the decreased plasma anesthetic level.15
CI, 11.77 to 5.28; P ¼ .46). We found significant hetero- In our analysis, MUF failed to decrease the amount of
geneity among trials (c2 ¼ 22.86; df ¼ 6 [P ¼ .0008]; chest tube drainage in the ICU. Coagulopathy and hemostatic
I2 ¼ 73.7%). difficulty are common after CPB in pediatric patients. Be-
Duration of ICU stay (Figure 2, E). Data regarding the cause the coagulation system of a neonate undergoing CPB
length of ICU stay were available for 55,6,8,10,12 of the 8 trials is known to be profoundly and globally affected by hemodi-
included in this meta-analysis. No studies described the ICU lution,16 MUF is expected to reverse the adverse effects of he-
discharge criteria. The combined results failed to show that modilution on the coagulation system. Indeed, previous
MUF shortened the ICU stay on average (WMD ¼ 0.81; reports have suggested that MUF increased the concentration
95% CI, 1.82 to 0.20; P ¼ 0.12; heterogeneity: of coagulation factors and that it attenuated the coagulopathy
c2 ¼ 15.30; df ¼ 4 [P ¼ .004]; I2 ¼ 73.9%). associated with CPB.17,18 Hemostatic difficulty after CPB

The Journal of Thoracic and Cardiovascular Surgery c Volume 142, Number 4 865
 Congenital Heart Disease Kuratani et al

does not have a simple pathologic cause, however; multiple responses after ultrafiltration.23,24 Inasmuch as MUF has
factors are involved.19 Increased inflammatory responses, higher efficiency in terms of fluid removal, it may be capa-
platelet dysfunction, and increased fibrinolysis are other ma- ble of filtering out inflammatory mediators more efficiently
jor factors that should be considered as reasons for abnormal as well. However, it is still unknown whether reduction of
hemostasis. The effects of MUF on preserving platelet func- cytokine levels by ultrafiltration can contribute to favorable
tion and fibrinolysis have not yet been fully clarified. outcome in pediatric cardiac surgery. Further study will be
Pulmonary dysfunction after CPB is common in pediatric necessary to clarify the attenuation of the inflammatory re-
cardiac surgery and may result in significant morbidity and sponse by MUF and to determine the clinical benefits.
mortality. The reasons for CPB-induced lung injury include It is important to note some of the limitations of meta-
increased interstitial lung water owing to hemodilution, analysis. Each study has different study protocols; this
lung ischemia during aortic crossclamping, and inflamma- may be the reason for the significant heterogeneity revealed
tory reaction elicited by CPB. Because MUF can eliminate by the I2 test in our meta-analysis. The justification of com-
excess water and can ameliorate inflammatory reactions, bining the results of different protocols in the calculation of
the advantages of MUF in terms of lung function have the WMD and in drawing conclusions is debatable. Factors
been noted and are widely accepted. However, our meta- that may influence study results include the type of ultrafil-
analysis failed to show the benefit of MUF on postoperative tration during CPB, type of MUF, duration of ultrafiltration
ventilation time. As Mahmoud and associates10 have during CPB, volume of ultrafiltrate obtained, end point cho-
pointed out, the advantages of MUF on lung function might sen for termination of MUF, type of hemofilter, concomitant
be of limited duration only rather than sustained for long anti-inflammatory therapies, patient characteristics, CPB
postoperative periods. An alternative view is that the variables, and complexity of cardiac surgery. In addition,
postoperative ventilation time may not reflect the real ben- because the meta-analysis is based on published articles,
efit of MUF in terms of lung function. If we consider the there is a possibility of publication bias. In this study, the
results of previous studies20,21 that demonstrated the im- omission of the unpublished, nonindexed, or non-English
provement of various pulmonary parameters, including articles that were not included may affect our conclusions.
lung compliances and respiratory indexes, we cannot elim- Although we limited our analysis to the literature in
inate the possibility that MUF facilitates the restoration of English, the effect of excluding non-English trials on the
lung function in the immediate postbypass period. results of a meta-analysis is equivocal. Some data suggest
To counteract pathologic fluid accumulation during CPB, that the exclusion of trials not published in English may ac-
CHD

ultrafiltration to remove excess water is now a widely ac- tually result in a more conservative estimate of the treat-
cepted practice in pediatric cardiac surgery. Theoretically, ment effect.25 This may be related in part to the presence
MUF has a much higher efficiency in terms of fluid removal of publication bias where only positive findings are pub-
than does CUF, because it is carried out after the termination lished; this occurs primarily in English-language journals.
of CPB. Indeed, previous reports10,12 have indicated that the In conclusion, meta-analysis of the currently available
ultrafiltrated fluid volume was larger in MUF. Meanwhile, randomized controlled trials that examined the clinical ben-
Thompson and associates9 conducted a prospective random- efits of MUF over CUF in pediatric cardiac surgery indi-
ized study to assess the hypothesis that MUF and CUF have cates that MUF resulted in significantly higher postbypass
similar clinical effects when a standardized volume of fluid hematocrit levels and higher mean arterial blood pressure.
is removed. They concluded that hematocrit, hemodynam- Our analysis failed to show a positive impact of MUF in
ics, ventricular function, blood product requirements, and postoperative clinical parameters, including postoperative
postoperative resources used do not differ between pediatric blood loss, ventilator time, and ICU time. These findings
patients receiving CUF and those receiving MUF.9 It re- suggest that MUF, compared with CUF, can improve clini-
mains unknown whether the benefits of MUF depend solely cal conditions in the immediate postbypass period, although
on its greater efficiency at fluid removal. the benefit of MUF on patient overall outcome might not be
Another potential advantage of ultrafiltration is cytokine significant. We must, however, take into account the possi-
removal and inflammatory response attenuation. Surgical ble clinical or methodologic variations in the currently
trauma and CPB are associated with the production of var- available evidence related to MUF.
ious kinds of cytokines and inflammatory responses. These
effects are most pronounced in pediatric patients. Such in- We thank Ryuichiro Araki, PhD (Community Health Science
flammatory responses can play a role in eliciting morbidity Center, Saitama Medical University, Saitama, Japan) for his statis-
tical assistance.
and mortality in postoperative periods. Indeed, Allan and
colleagues22 have demonstrated that postoperative interleu-
kin 6 and interleukin 8 are correlated with the length of the References
1. Naik SK, Knight A, Elliott M. A prospective randomized study of a modified
ICU stay in infant cardiac surgery. Some studies have re- technique of ultrafiltration during pediatric open-heart surgery. Circulation.
ported reduced cytokine levels and reduced inflammatory 1991;84(5 Suppl):III422-31.

866 The Journal of Thoracic and Cardiovascular Surgery c October 2011

Kuratani et al Congenital Heart Disease

2. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the 14. Chaturvedi RR, Shore DF, White PA, Scallan MH, Gothard JW, Redington AN,
quality of reports of meta-analyses of randomized controlled trials: The QUOROM et al. Modified ultrafiltration improves global left ventricular systolic function af-
statement. Quality of Reporting of Meta-analyses. Lancet. 1999;354:1896-900. ter open-heart surgery in infants and children. Eur J Cardiothorac Surg. 1999;15:
3. Gaynor JW. The effect of modified ultrafiltration on the postoperative course in 742-6.
patients with congenital heart disease. Semin Thorac Cardiovasc Surg Pediatr 15. Hodges UM, Berg S, Naik SK, Bower S, Lloyd-Thomas A, Elliot M. Filtration of
Card Surg Annu. 2003;6:128-39. fentanyl is not the cause of the elevation of arterial blood pressure associated with
4. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in post-bypass ultrafiltration in children. J Cardiothorac Vasc Anesth. 1994;8:
meta-analyses. BMJ. 2003;327:557-60. 653-7.
5. Bando K, Turrentine MW, Vijay P, Sharp TG, Sekine Y, Lalone BJ, et al. Effect of 16. Kern FH, Morana NJ, Sears JJ, Hickey PR. Coagulation defects in neonates dur-
modified ultrafiltration in high-risk patients undergoing operations for congenital ing cardiopulmonary bypass. Ann Thorac Surg. 1992;54:541-6.
heart disease. Ann Thorac Surg. 1998;66:821-7. 17. Ootaki Y, Yamaguchi M, Oshima Y, Yoshimura N, Oka S. Effects of modified ul-
6. Sever K, Tansel T, Basaran M, Kafali E, Ugurlucan M, Ali Sayin O, et al. The trafiltration on coagulation factors in pediatric cardiac surgery. Surg Today. 2002;
benefits of continuous ultrafiltration in pediatric cardiac surgery. Scand Cardio- 32:203-6.
vasc J. 2004;38:307-11. 18. Friesen RH, Campbell DN, Clarke DR, Tornabene MA. Modified ultrafiltration
7. Wang MJ, Chiu IS, Hsu CM, Wang CM, Lin PL, Chang CI, et al. Efficacy of ul- attenuates dilutional coagulopathy in pediatric open heart operations. Ann
trafiltration in removing inflammatory mediators during pediatric cardiac opera- Thorac Surg. 1997;64:1787-9.
tions. Ann Thorac Surg. 1996;61:651-6. 19. McEwan A. Aspects of bleeding after cardiac surgery in children. Paediatr
8. Williams GD, Ramamoorthy C, Chu L, Hammer GB, Kamra K, Boltz MG, et al. Anaesth. 2007;17:1126-33.
Modified and conventional ultrafiltration during pediatric cardiac surgery: clin- 20. Keenan HT, Thiagarajan R, Stephens KE, Williams G, Ramamoorthy C,
ical outcomes compared. J Thorac Cardiovasc Surg. 2006;132:1291-8. Lupinetti FM. Pulmonary function after modified venovenous ultrafiltration in
9. Thompson LD, McElhinney DB, Findlay P, Miller-Hance W, Chen MJ, infants: a prospective, randomized trial. J Thorac Cardiovasc Surg. 2000;119:
Minami M, et al. A prospective randomized study comparing volume- 501-5.
standardized modified and conventional ultrafiltration in pediatric cardiac sur- 21. Onoe M, Oku H, Kitayama H, Matsumoto T, Kaneda T. Modified ultrafiltration
gery. J Thorac Cardiovasc Surg. 2001;122:220-8. may improve postoperative pulmonary function in children with a ventricular
10. Mahmoud AB, Burhani MS, Hannef AA, Jamjoom AA, Al-Githmi IS, septal defect. Surg Today. 2001;31:586-90.
Baslaim GM. Effect of modified ultrafiltration on pulmonary function after car- 22. Allan CK, Newburger JW, McGrath E, Elder J, Psoinos C, Laussen PC, et al. The
diopulmonary bypass. Chest. 2005;128:3447-53. relationship between inflammatory activation and clinical outcome after infant
11. Berdat PA, Eichenberger E, Ebell J, Pfammatter JP, Pavlovic M, Zobrist C, et al. cardiopulmonary bypass. Anesth Analg. 2010;111:1244-51.
Elimination of proinflammatory cytokines in pediatric cardiac surgery: analysis 23. Wang W, Huang HM, Zhu DM, Chen H, Su ZK, Ding WX. Modified ultrafiltra-
of ultrafiltration method and filter type. J Thorac Cardiovasc Surg. 2004;127: tion in paediatric cardiopulmonary bypass. Perfusion. 1998;13:304-10.
1688-96. 24. Journois D, Israel-Biet D, Pouard P, Rolland B, Silvester W, Vouh
e P, et al.
12. Aggarwal NK, Das SN, Sharma G, Kiran U. Efficacy of combined modified and High-volume, zero-balanced hemofiltration to reduce delayed inflammatory
conventional ultrafiltration during cardiac surgery in children. Ann Card Anaesth. response to cardiopulmonary bypass in children. Anesthesiology. 1996;85:
2007;10:27-33. 965-76.
13. Davies MJ, Nguyen K, Gaynor JW, Elliott MJ. Modified ultrafiltration improves 25. J€uni P, Holenstein F, Sterne J, Bartlett C, Egger M. Direction and impact of lan-
left ventricular systolic function in infants after cardiopulmonary bypass. guage bias in meta-analyses of controlled trials: empirical study. Int J Epidemiol.
J Thorac Cardiovasc Surg. 1998;115:361-9. 2002;31:115-23.

CHD

The Journal of Thoracic and Cardiovascular Surgery c Volume 142, Number 4 867