Retrospective analysis of eliminating modified ultrafiltration after pediatric cardiopulmonary bypass

Abstract

Modified ultrafiltration (MUF) is a technique which is commonly used immediately post-cardiopulmonary bypass (CPB) for open heart surgery in children. There are many advantages of MUF, but there are also a number of less reported disadvantages. At our institution, after considering all of the available data, a decision was made to no longer perform MUF. The primary motivation being the simplified and miniaturized CPB circuit would reduce hemodilution, decrease our likelihood of reaching our transfusion trigger during CPB and, potentially, improve safety. This study reports the before and after data from this practice change. A total of 160 patients less than 8kg were studied over 38 months and divided into neonatal and pediatric cohorts. Parameters reported in this study include: demographics, hematocrit, blood product transfusion, hemostasis, hemodynamics and outcomes. Although retrospective, our analysis supports an advantage of preventing hemodilution (via circuit miniaturization) versus reversing hemodilution (via MUF) at our institution with the patient population we examined.

Keywords

ultrafiltration pediatric cardiopulmonary bypass modified ultrafiltration inflammatory mediators transfusion vasoactive inotrope score conventional ultrafiltration cardiac surgery

Introduction

Modified ultrafiltration (MUF) is an ultrafiltration (UF) technique used immediately post-cardiopulmonary bypass (CPB) that was first described in 1991 by Naik et al.¹ The main goals of MUF are: removal of extravascular water, hemoconcentration, re-infusion of residual blood in the CPB circuit and the removal of inflammatory mediators. Reported benefits of MUF include: increased hematocrit (Hct),^1-6 platelets (Plt)^3,4,7 and plasma proteins;^3,6 decreased blood product transfusion,^3-5,8-10 ventilator time (VT),^4,9 chest tube output (CTO),^4,5,8-11 total body water post-CPB,^1,7,8 intensive care unit (ICU)^2,4,7,9 and hospital¹⁰ lengths of stay (LOS) and improved hemodynamic^2,4,8,12 and respiratory function.^2,13,14 In a 2011 survey by Harvey et al., 71% of pediatric centers report using MUF.¹⁵

Potential disadvantages of MUF are rarely described in the literature, but can be significant. These include increased complexity, prime volume and foreign surface area of the CPB circuit possibly increasing hemodilution and inflammation. Other disadvantages include; hemodynamic instability, cerebral steal,¹⁶ aortic obstruction caused by the aortic cannula, decreased core temperature, prolongation of exposure to non-endothelial surfaces and longer operating time with additional costs. Adverse events associated with MUF are rare, but can be serious. They include CPB circuit rupture, accidental cardioplegia delivery resulting in fibrillation, accidental hypothermia (potentially causing fibrillation), air entrainment into the CPB circuit and air embolism to the patient.¹⁷

Recently there have been significant improvements in CPB equipment and techniques as well as an increased knowledge of CPB-related pathology and physiology. These improvements have coincided with improvements in surgical, anesthesia, blood banking and critical care practices. Today, patients are exposed to less hemodilution and foreign surfaces, have less inflammation and edema post-CPB and have improved outcomes. These changes have stimulated a re-evaluation of many CPB-related techniques/practices, including MUF.

At our institution, we routinely performed MUF on all patients <25 kg and on larger patients on a case by case basis. The advantage of no longer performing MUF was to simplify and miniaturize our CPB circuit. Our motivations to change practice were that this simplified and miniaturized CPB circuit could reduce hemodilution, decrease our likelihood of reaching our transfusion trigger during CPB and improved safety. This is a retrospective analysis of the impact of this practice change, comparing MUF patients to non-MUF patients.

Methods

Patients

We conducted a retrospective study of two cohorts of patients before and after eliminating MUF from our practice. Approval was obtained from the Children’s Hospital Colorado, Organizational Research Risk & Quality Improvement Review Panel. One hundred and sixty patients weighing less than 8 kg who underwent CPB for surgical repair of congenital heart defects at Children’s Hospital Colorado from December 2010 to January 2014 were evaluated. The surgical team included three surgeons, four perfusionists and eight anesthesiologists. One anesthesiologist joined the team in 2011. There were no other significant changes to surgical, anesthesia or perfusion personnel or techniques during this period. We stopped performing MUF in April 2013. The cut-off of 8 kg was chosen to ensure homogeneity since all patients <8 kg receive a blood prime at our institution. The 160 patients were divided into pediatric and neonatal cohorts as shown in Figure 1. Prior to the study, we determined that 50 patients in each limb of the pediatric cohort and 30 patients in each limb of the neonatal cohort would both give the study enough power to determine differences and allow us to avoid going back too far historically (MUF groups) or to include patients prospectively (No-MUF groups - since we already had performed >30 neonatal and >50 pediatric No-MUF cases). In the MUF groups, we started including MUF patients consecutively in reverse chronological order from our last MUF patient who met inclusion criteria until we had enough patients in each MUF group. We did the same in chronological order from the time we stopped performing MUF in the No-MUF groups. Exclusion criteria included; patients undergoing a heart transplant, ventricular assist device placement or extracorporeal membrane oxygenation; patients requiring pre-CPB intra-op allogeneic packed red blood cells (RBCs) and patients who required unplanned re-operation. There were no deaths in any group during the operative or immediate post-operative (post-op) period.

Cardiopulmonary bypass

Protocols for CPB in both groups were identical. Not performing MUF in the No-MUF groups was the only change in clinical practice during the study period. CPB circuits containing either the FX05 (Terumo Cardiovascular, Ann Arbor, MI) or VKMO 11000 (Maquet Cardiovascular, Wayne, NJ) oxygenators with integrated arterial filters and SMAR_xT (LivaNova, Arvada, CO) coated tubing with 3/16-inch arterial, ¼-inch venous lines and a ¼-inch pump boot were used for all groups. The cardioplegia (CPG) circuit consisted of 1:4 tubing set with an integrated stopcock manifold (for MUF), a Vanguard BCD (LivaNova) CPG device and a single large roller pump. In the MUF groups, a DHF0.2 (LivaNova) UF device was linked to the CPG circuit and the main CPB circuit via a stopcock manifold and tubing. This setup allowed conventional ultrafiltration (CUF) and CPG delivery during CPB (Figure 2a) and MUF via the CPG circuit post-CPB (Figure 2b) without having to move the UF device. In the No-MUF groups, either a Hemocor HPH® Jr. pediatric hemoconcentrator (Medivators Inc., Minneapolis, MN) or the DHF0.2 was incorporated into the CPB circuit via a simple arterio-venous shunt (Figure 2c). Suction (-150 mmHg) was applied to the effluent side of the UF device for both CUF and MUF.

The CPB prime of all groups consisted of Plasmalyte-A (Baxter Healthcare Corporation, Deerfield, IL), one unit of RBCs and one unit of fresh frozen plasma (FFP). Heparin, sodium bicarbonate, mannitol, calcium chloride and aminocaproic acid were added to the prime, based on patient body weight. The RBCs and FFP underwent pre-CPB UF with 1000 ml of Plasmalyte-A for the removal of excess potassium, lactic acid, dextrose and citrate. Approximately 150 ml of this ultrafiltered prime blood was sequestered and administered during the rewarming phase of CPB. The estimated total prime volume of the CPB circuit was 300 ml in the MUF groups and 210 ml in the No-MUF groups. This difference in prime volume is due to the lower prime volume of the UF circuits between the groups (44 ml vs. 15 ml). In addition, for the MUF groups, the entire CPG circuit gets primed with patient blood for MUF versus only one-fifth of the CPG circuit in the No-MUF groups for CPG delivery (86 ml vs. 27 ml) (vs. 0 ml for non-CPG cases).

On CPB, the Hct was maintained at or above 30% during CPB and 35% at weaning, flow rates were maintained at 125–200 ml/kg/min and the mean arterial blood pressure was kept above 35 mmHg. Continuous mixed venous oxygen saturation, inline arterial blood gas monitoring (CDI 500, Terumo Cardiovascular) and cerebral near-infrared spectroscopy (NIRS) (INVOS^TM, Medtronic, Minneapolis, MN) were standard on all cases to assure adequacy of perfusion. Activated clotting times were maintained greater than 480 seconds. Vacuum-assisted venous drainage was used on every case. Hypothermic delNido cardioplegia (1:4, blood to CPG ratio) was given initially at 30 ml/kg and subsequent doses of 15 ml/kg at 60 to 90-minute intervals. A cell saver (CS) device (Fresenius C.A.T.S., Bad Homburg, Germany) was standard on all cases. Residual circuit volume post CPB/MUF was processed via the CS and re-infused by the anesthesiologist or bedside nurse as needed.

Ultrafiltration

All groups received conventional CUF during CPB to remove excess volume in the venous reservoir and to maintain the Hct at or above 30%. In the No-MUF groups, the UF device was de-primed into the venous reservoir just prior to weaning and this volume was used to fill the patient during CPB weaning. In the MUF groups, it was often necessary to add Plasmalyte-A to the venous reservoir to prime the CPG circuit with patient blood and to wean the patient from CPB.

Arterial–venous MUF was performed on the MUF groups immediately post CPB. The venous line was detached from the venous cannula and this volume was drained into the CPB circuit. Blood was removed from the patient via the arterial cannula, pumped through the UF device, then through the CPG device (warming the blood) and back to the patient via the CPG line which was attached to the venous cannula in the right atrium. Residual blood volume in the venous reservoir and oxygenator was pumped into this circuit via the arterial pump to replace the volume removed via MUF and to maintain adequate patient filling pressures. MUF flow rates ranged from 10 to 30 ml/kg/min and the technique was terminated when all residual venous reservoir and oxygenator volume had been re-infused. Pressure servo-regulation was used to detect negative circuit pressures which could lead to air entrainment into the CPB circuit. There were no known adverse events related to MUF in any patient. Upon completion of CPB (No-MUF groups) or MUF (MUF groups), heparin was reversed with protamine and blood products were administered as necessary to achieve adequate hemostasis.

Statistical analysis

Peri-operative (peri-op) data were collected retrospectively. The ICU was blinded to our change in practice. Data were obtained from an audit of transfusion, CPB, anesthesia and ICU records. Statistical analysis was performed using SAS 9.4 software (SAS Institute, Cary, NC). Outcome variables were compared between two groups using a two sample t-test and the Chi-square test as appropriate. Linear mixed effects model was used to examine the change in Hct across different time points.

Results

Demographics

Table 1 shows the demographic information for both cohorts. In the pediatric cohort, the Aristotle Comprehensive Complexity score was significantly higher in the MUF group versus the No-MUF group (7.5 ± 2 vs. 6.8 ± 1.3, p = 0.03). In the neonatal cohort, a higher percentage of patients in the MUF group had single ventricle anatomy (50% vs. 30%, p = 0.11) and received deep hypothermic circulatory arrest (DHCA) (73% vs. 53%, p = 0.108) compared to the No-MUF group.

Table 1.

Demographics.

	MUF	No-MUF	p-Value
	Pediatric Cohort
N	50	50
Age (days)	138 ± 70	152 ± 79	0.34
Weight (Kg)	5.4 ± 1.3	5.4 ± 1.3	1.0
BSA (m²)	0.3 ± 0.1	0.3 ± 0.1	0.86
SV anatomy (N)	5 (10%)	7 (14%)	0.53
Redo (N)	8 (16%)	6 (12%)	0.33
DHCA (N)	7 (14%)	5 (10%)	0.53
Aristotle Score	7.5 ± 2	6.8 ± 1.3	0.03
CPB time (min)	108 ± 43	102.5 ± 38	0.47
X-CL time (min)	61.5 ± 30	61.4 ± 27	0.98
	Neonatal Cohort
N	30	30
Age (days)	7.1 ± 5.3	7.0 ± 5.4	0.94
Weight (Kg)	3.1 ± 0.6	3.1 ± 0.4	0.90
BSA (m²)	0.2 ± 0	0.2 ± 0	0.83
SV anatomy (N)	15 (50%)	9 (30%)	0.11
Redo (N)	0	0
DHCA (N)	22 (73%)	16 (53%)	0.108
Aristotle Score	11.0 ± 3.1	10.6 ± 2.6	0.58
CPB time (min)	133.7 ± 47.3	127.2 ± 42.7	0.57
X-CL time (min)	64.9 ± 31.7	65.9 ± 31.2	0.91

BSA: body surface area; SV: single ventricle; DHCA: deep hypothermic circulatory arrest; CPB: cardiopulmonary bypass; X-CL: cross clamp.

Hematocrit

Table 2 shows Hct data for both cohorts. Hct data was evaluated at five time points: baseline, lowest on CPB (lowest-Hct), last on CPB (last-CPB), immediately post CPB/MUF (post-CPB/MUF), first in the ICU (ICU-Hct) and is reported as the mean ± 2 SE. We also evaluated the change in Hct from baseline to post-CPB/MUF and from baseline to ICU-Hct (ΔHcts).

Table 2.

Hematocrit data.

	MUF	No-MUF	p-Value
	Pediatric Cohort
Baseline (%)	34.6 ± 6.1	34.1 ± 5.2	0.63
Lowest CPB (%)	31.1 ± 3.9	32.8 ± 3	0.017
Last CPB (%)	34.4 ± 4.3	36.0 ± 4.4	0.07
Post CPB/MUF (%)	38.1 ± 4.2	35.9 ± 4.4	0.016
ICU (%)	42.6 ± 5.2	41.4 ± 4.9	0.25
Δ Hct Baseline to post CPB/MUF	+3.5 ± 6.6	+1.9 ± 4.7	0.17
Δ Hct Baseline to ICU	+8.0 ± 7.7	+7.3 ± 6.5	0.65
	Neonatal Cohort
Baseline (%)	40.5 ± 6.5	42.4 ± 5.9	0.25
Lowest CPB (%)	32.9 ± 3.5	31.7 ± 4.2	0.22
Last CPB (%)	36.8 ± 4.2	40.9 ± 4.9	0.0009
Post CPB/MUF (%)	38.6 ± 6.6	38.5 ± 5.2	0.93
ICU (%)	43.3 ± 5.6	42.2 ± 6.8	0.52
Δ Hct Baseline to post CPB/MUF	−1.9 ± 8.3	−3.9 ± 7.3	0.32
Δ Hct Baseline to ICU	+2.8 ± 7.3	−0.1 ± 8.9	0.17

CPB: cardiopulmonary bypass; MUF: modified ultrafiltration; ICU: intensive care unit; Hct: hematocrit.

In the pediatric cohort, the lowest Hct was significantly higher in the No-MUF group versus the MUF group (32.8% ± 3 vs. 31.1% ± 3.9, p = 0.017). The post-CPB/MUF Hct was significantly higher in the MUF group versus the No-MUF group (38.1% ± 4.2 vs. 35.9% ± 4.4, p = 0.016). Figure 3 is a graph of the mean Hct ± 2 SE at each time point.

In the neonatal cohort, the last CPB Hct was significantly higher in the No-MUF group versus the MUF group (40.9% ± 4.9 vs. 36.8% ± 4.2, p = 0.0009). Figure 4 shows the mean Hct ± 2 SE at each time point.

Red blood cell transfusion

Table 3 shows the RBC transfusion data for both cohorts and includes the CPB, intra-operative (intra-op), ICU and peri-op time periods. ICU data pertains to the first 24 hours post-op. Peri-op data includes the intra-op period plus the first 24 hours post-op. RBC volumes are recorded as the mean volume (ml/kg) ± 2 SE. Second donor exposure rates are recorded for all four time periods. All patients in both cohorts received one unit of RBCs in the CPB prime (1^st donor exposure).

Table 3.

PRBC transfusion data.

	MUF	No-MUF	p-Value
	Pediatric Cohort
CPB Prime (ml/kg)	49.7 ± 13.4	49.9 ± 14.4	0.94
Total CPB (ml/kg)	54.4 ± 18.7	49.9 ± 14.4	0.18
Intra-op (ml/kg)	55.6 ± 20.5	50.9 ± 15.8	0.19
Peri-op (ml/kg)	56.7 ± 20.9	51.6 ± 15.2	0.16
2^nd exposure CPB (N)	6 (12%)	0	0.011
2^nd exposure intra-op (N)	7 (14%)	1 (2%)	0.02
PRBCs transfused ICU (N)	4 (8%)	3 (6%)	0.69
2^nd exposure ICU (N)	3 (6%)	3 (6%)	1.0
2^nd exposure peri-op (N)	10 (20%)	4 (8%)	0.08
Donor exposures per patient (units)	1.22 ± 0.46	1.08 ± 0.27	0.07
	Neonatal Cohort
CPB Prime (ml/kg)	84.4 ± 21.8	82.8 ± 15.4	0.74
Total CPB (ml/kg)	91.8 ± 25.9	82.8 ± 15.4	0.10
Intra-op (ml/kg)	94.9 ± 28.8	84.8 ± 17.2	0.10
Peri-op (ml/kg)	99.1 ± 28.9	86.4 ± 18.3	0.046
2^nd exposure CPB (N)	5 (16.7%)	0	0.019
2^nd exposure intra-op (N)	7 (23.3%)	1 (3.3%)	0.022
PRBCs transfused ICU (N)	11 (36.7%)	4 (13.3%)	0.036
2^nd exposure ICU (N)	8 (26.7%)	4 (13.3%)	0.19
2^nd exposure peri-op (N)	15 (50%)	5 (16.7%)	0.006
Donor exposures per patient (units)	1.5 ± 0.51	1.17 ± 0.38	0.006

CPB: cardiopulmonary bypass; PRBCs: packed red blood cells; ICU: intensive care unit.

In the pediatric cohort, the MUF group received a 2^nd RBC donor exposure significantly more often than the No-MUF group during the CPB (12% vs. 0%, p = 0.011), intra-op (14% vs. 2%, p = 0.02) and peri-op (20% vs. 8%, p = 0.08) periods. There was no difference in the percentage of patients receiving RBCs in the ICU or receiving a 2^nd RBC donor exposure in the ICU. Overall, this translated into an average of 1.22 RBC donor exposures per patient during the peri-op period in the MUF group versus 1.08 RBC donor exposures per patient in the No-MUF group (p = 0.07). Figure 5 shows the combined pediatric and neonatal cohort hematocrit and Figure 6 shows the total peri-op PRBC volume.

In the neonatal cohort, the MUF group received significantly more RBC volume than the No-MUF group during the peri-op period (99.1 ± 28.9 ml/kg vs. 86.4 ± 18.3 ml/kg, p = 0.046) (Figure 7). The MUF group also received a 2^nd RBC donor exposure significantly more often than No-MUF patients during the CPB (16.37% vs. 0%, p = 0.019), intra-op (23.3% vs. 3.3%, p = 0.022) and peri-op (50% vs. 16.7%, p = 0.006) periods. In the ICU, MUF patients received RBCs significantly more frequently than No-MUF patients (36.7% vs. 13.3%, p = 0.036) and this was a new 2^nd RBC exposure for 26.7% of MUF patients versus 13.3% of No-MUF patients (p = 0.19). Overall, this translated into an average of 1.5 RBC donor exposures per patient during the peri-op period in the MUF group versus 1.17 RBC donor exposures per patient in the No-MUF group (p = 0.006).

Non-RBC transfusion, post-op labs, chest tube output, time out of the operating room

Table 4 shows the transfusion data for FFP, cryopreci-pitate (Cryo), and Plt for the entire peri-op period; post-op labs (platelet count, fibrinogen, international normalized ratio (INR)), chest tube output (CTO) and time in minutes from weaning off CPB to leaving the operating room (T1). Post-op labs were drawn within the first 15 minutes upon arrival in the ICU. CTO (ml/kg) includes the first 24 hours in the ICU. Volume data are presented as the mean volume (ml/kg) ± 2 SE. All patients in both cohorts received one unit of FFP in the CPB prime.

Table 4.

Peri-op fresh frozen plasma (FFP), cryoprecipitate (Cryo) and platelet (Plt) transfusion data, post-op labs, chest tube output (CTO) on leaving the operating room (T1).

	MUF	No-MUF	p-Value
	Pediatric Cohort
Prime FFP (ml/kg)	44.5 ± 14.7	43.4 ± 12.6	0.69
Peri-op FFP (ml/kg)	44.9 ± 14.7	44.2 ± 12.2	0.80
FFP 2^nd exposure (N)	1 (2%)	4 (8%)	0.168
Cryo (ml/kg)	6.6 ± 4.5	4.6 ± 3.4	0.014
Platelets (ml/kg)	14.6 ± 9	13.2 ± 10.3	0.46
Baseline platelet count (x10³/uL)	354 ± 92	347 ± 85	0.69
ICU platelet count (x10³/uL )	219.7 ± 75	221.3 ± 77	0.91
Fibrinogen (mg/dl)	301 ± 70	273 ± 60	0.04
INR	1.18 ± 0.1	1.17 ± 0.1	0.5
Chest tube output (ml/kg)	18.3 ± 11	12.9 ± 6.7	0.004
T1 (mins)	77.2 ± 18.5	66.7 ± 18.4	0.005
	Neonatal Cohort
Prime FFP (ml/kg)	74.8 ± 17.5	73.7 ± 13.7	0.78
Peri-op FFP (ml/kg)	77.9 ± 20.9	74.9 ± 14.5	0.52
FFP 2^nd exposure (N)	6 (20%)	2 (6.6%)	0.13
Cryo (ml/kg)	11.6 ± 4.9	8.8 ± 4.6	0.028
Platelets (ml/kg)	27.5 ± 17.6	21.1 ± 13.3	0.11
Baseline platelet count (x10³/uL )	277 ± 78	247 ± 118	0.25
ICU platelet count (x10³/uL )	237.7 ± 107	286.2 ± 143	0.14
Fibrinogen (mg/dl)	330.7 ± 90	312.1 ± 72	0.39
INR	1.24 ± 0.12	1.24 ± 0.18	0.96
Chest tube output (ml/kg)	17.7 ± 9.5	18.0 ± 9.9	0.92
T1 (mins)	89.7 ± 17.4	82.0 ± 22	0.13

INR: international normalized ratio.

In the pediatric cohort, the MUF group received significantly more Cryo than the No-MUF group (6.6 ± 4.5 ml/kg vs. 4.6 ± 3.4 ml/kg, p = 0.014); T1 was significantly longer in the MUF group versus the No-MUF group (77.2 ± 18.5 mins vs. 66.7 ± 18.4 mins, p = 0.005). Fibrinogen (Fbg) was significantly higher in the MUF group versus the No-MUF group (301 ± 70 mg/dl vs. 273 ± 60 mg/dl, p = 0.04) and the CTO was significantly higher in the MUF group versus the No-MUF group (18.3 ± 11 ml/kg vs. 12.9 ± 6.7 ml/kg, p = 0.004).

In the neonatal cohort, the MUF group received significantly more Cryo than the No-MUF group (11.6 ± 4.9 ml/kg vs. 8.8 ± 4.6 ml/kg, p = 0.028).

Ultrafiltration (UF) and cell saver (CS) volumes, ventilator time (VT), ICU and hospital lengths of stay

Table 5 shows UF and CS volumes, ventilator time (VT) and ICU and hospital lengths of stay (LOS). UF and CS volumes are reported as ml/kg. VT was calculated in hours from post-op arrival in the ICU to extubation. ICU LOS was calculated in days, from arrival in the ICU to transfer to a non-intensive care unit. Hospital LOS was calculated in days, from arrival in the ICU to hospital discharge.

Table 5.

Ultrafiltration and cell saver volume, ventilator time, intensive care unit and hospital length of stay.

	MUF	No-MUF	p-Value
	Pediatric Cohort
UF volume (ml/kg)	148.8 ± 89.3	98.6 ± 47.9	0.001
Cell saver product (ml/kg)	22.7 ± 10.5	27.0 ± 10	0.06
Ventilator time (hours)	23.0 ± 33.4	17.8 ± 26.7	0.39
ICU LOS (days)	3.8 ± 3	3.3 ± 2.2	0.34
Hospital LOS (days)	9.1 ± 10.8	6.9 ± 4.5	0.19
	Neonatal Cohort
UF volume (ml/kg)	247.5 ± 142.5	151.8 ± 60.2	0.003
Cell saver product (ml/kg)	48.3 ± 31.9	51.0 ± 16.9	0.71
Ventilator time (hours)	69.0 ± 51.8	65.5 ± 49.8	0.78
ICU LOS (days)	12.2 ± 13	9.8 ± 8.8	0.40
Hospital LOS (days)	21.5 ± 14.6	26.3 ± 18.4	0.26

UF: ultrafiltration; ICU: intensive care unit; LOS: length of stay.

In the pediatric and neonatal cohorts, the UF volume was significantly greater in the MUF group versus the No-MUF group (148.8 ± 89.3 ml/kg vs. 98.6 ± 47.9 ml/kg, p = 0.001) and (247.5 ±142.5 ml/kg vs. 151.8 ± 60.2 ml/kg, p = 0.003), respectively. There was no difference in CS volume, VT or ICU and hospital LOS.

Vasoactive inotrope score

Table 6 shows the vasoactive inotrope score (VIS) data. VIS was recorded at three time points: weaning from CPB (VIS-CPB), leaving the OR (VIS-OR) and 24hrs post-op (VIS-24hr). VIS was calculated using the formula described by Gaies et al.¹⁸

Table 6.

Vasoactive inotrope scores.

	MUF	No-MUF	p-Value
	Pediatric Cohort
VIS-CPB	9.8 ± 3.6	8.6 ± 3.2	0.07
VIS-OR	8.5 ± 4	7.0 ± 3.2	0.031
VIS-24hr	4.3 ± 3.8	3.5 ± 5.0	0.38
	Neonatal Cohort
VIS-CPB	12.8 ± 4.1	11.4 ± 4.6	0.21
VIS-OR	13.1 ± 5.6	10.3 ± 5.6	0.06
VIS-24hr	12.0 ± 7.6	9.9 ± 6.3	0.25

VIS = Dopamine dose (µg/kg/min) + dobutamine dose (µg/kg/min) + 100 x epinephrine dose (µg/kg/min) + 10 x milrinone dose (µg/kg/min) + 10,000 x vasopressin dose (U/kg/min) + 100 x norepinephrine dose (µg/kg/min)

In the pediatric cohort, VIS was lower in the No-MUF group compared to the MUF group at all three time periods. VIS-OR was significantly lower in the No-MUF group versus the MUF group (7.0 ± 3.2 vs. 8.5 ± 4, p = 0.031).

In the neonatal cohort, VIS was lower in the No-MUF group compared to the MUF group at all 3 time periods, nearly reaching significance for VIS-OR (p = 0.06).

Discussion

While our results may seem unexpected, there is evidence from the literature that patients who undergo MUF may encounter more hemodilution and, potentially, more edema compared to No-MUF control groups. In multiple studies, control groups who didn’t receive MUF had a higher Hct after weaning from CPB than the MUF groups to whom they were compared.^{2,3,5,7,9,12,13} In a study by Kotani et al., the Hct was significantly higher in the control group compared to the MUF group.² Draaisma et al. reported a higher prime volume, attributed to priming of the MUF circuit (+102 ml), and lower mean values of hemoglobin and Hct during CPB in the MUF group.⁵ Thompson et al. reported that the Hct immediately after CPB was significantly lower than at baseline in MUF patients, but, in patients only undergoing conventional ultrafiltration (CUF), the Hct immediately after CPB did not differ significantly from baseline.¹⁹ Davies et al. reported that MUF improved left ventricular (LV) function. However, the LV function after CPB in the MUF patients was lower than the No-MUF control groups and the improvement in LV function after MUF was up to a level similar to control patients.¹² In a randomized control trial by Hennein et al., the MUF group had a higher prime volume and received more intra-op blood than the CUF group and the control group (no UF) to which it was compared. In addition, the last Hct on CPB was much higher in the CUF group than in the MUF group.⁷

In a report on minimizing hemodilution during CPB, by Charette et al., they do not perform MUF on neonates and small infants, stating, “It is our experience that, when the CPB circuit prime volume is around 200 ml and intraoperative hemoconcentration has been employed, it may be counterproductive to prime a MUF circuit (necessitating the addition of volume) or to remove volume from the patient post bypass (they are already dry).”²⁰ In a study by Merkle et al. on using a non-blood prime in neonates and infants, they did not perform MUF on five out of nine patients because of the volume required to prime the MUF circuit.²¹ Williams et al. report no longer performing MUF on infants since they had similar results between their MUF and CUF groups, but the incidence of UF-related complications, although uncommon, was greater for MUF.²²

Many early pro-MUF studies involved significant hemodilution and compared MUF groups to control groups that received no UF.^3,5,8,13 Today, UF during CPB to reverse hemodilution is so widely accepted and the negative effects of severe hemodilution are so well known that it would be unlikely to get approval for a study involving significant patient hemodilution and patients receiving no UF. In a meta-analysis by Kuratani et al., they excluded studies like this stating, “Studies that compared MUF and control patients without any ultrafiltration technique were not included, because the benefits of ultrafiltration in pediatric CPB are widely acknowledged and CPB management without any ultrafiltration would not reflect actual clinical practice.” Evaluating only studies with control groups that received UF during CPB, they found that post-operative outcome parameters were not significantly influenced by MUF.²³ Since hemodilution and transfusion both have injurious effects and the exact point at which one becomes more harmful than the other will likely never be known (due to multiple variables that would effect this for every individual patient), it is important to adopt strategies that both reduce hemodilution and help avoid transfusion. By eliminating MUF, we hoped to achieve both.

By no longer performing MUF at the conclusion of CPB, we were able to minimize our CPB circuit and simplify how we perform CPB. This was accomplished by eliminating tubing that linked the UF device to the CPG and the main portions of the CPB circuit; using a lower prime, lower surface area UF device for neonates and smaller pediatric patients (since we no longer needed to rapidly ultrafilter during MUF); and no longer needing to flush out the CPG solution and prime the CPG circuit with patient blood. The latter was a major disadvantage of our MUF technique, which resulted in shifting as much as 100 ml of blood from the venous reservoir to the CPG circuit, often necessitating the addition of extra crystalloid volume into the reservoir. This resulted in a drop in patient Hct just prior to weaning, which often led to the transfusion of RBCs to wean with an acceptable Hct and appropriate hemodynamics. Now, since we frequently only give a single dose of 1:4 cardioplegia (1 part blood), the significance of the prime volume and surface area of the CPG circuit is minimized as it pertains to patient hemodilution and loss of reservoir volume. This difference is even more significant for operations in which CPG is not required - we completely avoid priming the CPG circuit with patient blood and exposing the patient to this portion of the CPB circuit. In addition, since we no longer need the UF device post-CPB for MUF, we can de-prime this portion of the circuit into the venous reservoir prior to weaning and use this additional volume to fill the patient. This combination of less tubing, a smaller UF device, not priming the CPG circuit with blood and de-priming the UF circuit equals a shift in volume prior to weaning (approximately +150 ml). The extra volume available allows additional UF to be carried out, resulting in weaning at an Hct that is consistently equal to or greater than our target post-MUF Hct. The smaller the patient the more significant these changes are likely to be. As a result, we are always able to wean at a higher Hct when not performing MUF compared to if we were to perform MUF. Essentially, the difference between the techniques is performing UF just prior to weaning (No-MUF) versus performing UF just after weaning (MUF), with the former allowing a greater volume of UF during CPB and a higher Hct during weaning, the latter allowing less UF during CPB, but, generally, a greater volume of UF overall. The major advantage is that the former is safer, simpler and potentially cheaper, allows reduction in our CPB circuit foreign surface area and time of exposure and comes with a decreased chance of reaching our RBC transfusion trigger.

Hematocrit/RBC transfusion

When evaluating the Hct data in the pediatric cohort, it is important to consider that patients in the MUF group received a greater volume of RBCs during the CPB, intra-op and peri-op periods and received a 2^nd donor exposure significantly more often during these same periods. This could be the reason that patients in the MUF group had a significantly higher post-CPB/MUF Hct compared to No-MUF patients. No-MUF patients having a significantly higher lowest-Hct, a higher last-CPB Hct and a similar ICU-Hct despite receiving less RBCs may be evidence of an advantage of CPB circuit miniaturization by eliminating MUF.

Similar to the pediatric data, the neonatal cohort Hct data is in the setting of a MUF group which received more RBCs during the CPB, intra-op and peri-op time periods. The neonatal MUF group also had significantly higher 2^nd RBC donor exposure rates during those same time periods. Despite this, the last-CPB Hct was significantly higher in the No-MUF group and there was no difference in the post-CPB/MUF and ICU Hcts.

When evaluating the Hct data, it is also important to consider that, in the MUF groups, the lab sample for the last-CPB Hct was usually drawn prior to priming of the MUF circuit and, thus, did not reflect the hemodilution which often occurred just prior to weaning. Likewise, in the No-MUF groups, this sample didn’t reflect the hemoconcentration that often occurred just prior to weaning. Despite this, in the neonatal cohort, the No-MUF group did have a significantly higher last-CPB Hct compared to the MUF group. Had this been a randomized controlled trial (RCT), we could have timed this sample to more accurately assess the patient Hct immediately prior to weaning from CPB. Once we stopped performing MUF, we began recording the weaning Hct (via the inline venous sat/Hct monitor or via lab result if it was collected just prior to weaning). In the pediatric No-MUF group, the average weaning Hct was 39.7% (range: 32-46%) and 29 of 50 patients (58%) had a weaning Hct ⩾40%. In the neonatal No-MUF group, the average weaning Hct was 44.4% (range: 38-50%) and 29 of 30 patients (96.6%) had a weaning Hct ⩾40%. If we were to compare this weaning-Hct data to the last-CPB Hct in the MUF groups or if we had a more accurate weaning Hct in the MUF groups, it would likely show a significantly higher weaning Hct in No-MUF patients despite a lower RBC transfusion rate in these groups. This supports the significance of the change in fluid balance when eliminating MUF, particularly in neonatal patients.

Since performing MUF often contributed to our need to transfuse additional RBCs during CPB, this transfusion may have been avoided had we decided not to perform MUF. On the contrary, in No-MUF cases in which additional RBCs were required during CPB, had we decided to perform MUF, we would always have required those additional RBCs and, potentially, would have required a greater volume, since the Hct would have been even lower after priming the MUF circuit (and due to the larger CPB circuit we would be using). In every scenario, the Hct weaning from CPB should be higher than if we had performed MUF. While there are cases in which the post-CPB Hct will be higher after performing MUF, this is at the expense of a higher transfusion risk, the potential for MUF-related adverse events and a more complex, higher prime volume, higher foreign surface area circuit.

Non-RBC transfusion/hemostasis (FFP, Cryo, Plt, CTO)

Similar to Hct data, when evaluating post-op coagulation labs, it is important to consider transfusion rates of Plt, Cryo and FFP between groups (Table 4). The MUF groups in both cohorts received significantly more Cryo and more Plt and FFP (neonatal cohort) than their respective No-MUF groups. This, not greater hemoconcentration in MUF patients, might be why fibrinogen levels were significantly higher in the pediatric MUF group compared to the No-MUF group. In the neonatal cohort, the platelet count was higher in the No-MUF group despite the MUF group receiving more platelets and having a higher baseline platelet count.

Many studies have reported decreased CTO in MUF patients.^4,5,8-11 In our analysis, the CTO in the neonatal cohort was similar among the groups and was significantly less in the pediatric No-MUF group (p = 0.004). Thompson et al. also had less CTO in their control group.¹⁹ The same or less CTO in our analysis supports a similar or improved coagulopathic state in our No-MUF groups, particularly considering our No-MUF patients received less blood products to achieve hemostasis.

Time from CPB weaning to leaving OR (T1)

Some MUF studies report performing MUF for as long as 20 minutes post-CPB.^10,13,24 This prolongs the patients’ exposure to the non-endothelial surfaces of the CPB circuit and, potentially, the OR time. Our average MUF time in both MUF groups was only 9 minutes. In No-MUF cases, we felt immediate protamine administration following CPB along with an optimized coagulopathic state resulted in frequently achieving rapid hemostasis. As a result, we began recording the chest closure time (we did not routinely collect this data for MUF patients). The average time from CPB weaning to the chest being closed in the No-MUF pediatric cohort was 28 minutes (range: 11-64) and, in the No-MUF neonatal cohort, it was 39 minutes (range: 20-110). Since we weren’t able to compare this data to the MUF groups, we decided to compare the difference in time from CPB weaning to leaving the operating room (T1). While there are many variables that contribute to this time, we evaluated this as a rough measure of the speed of hemostasis and patient stability. T1 in the pediatric No-MUF group was significantly shorter than in the MUF group. This is possible evidence of more rapid hemostasis and improved stability in No-MUF patients. However, it may simply be the result of eliminating a procedure that typically took 9 minutes after weaning from CPB. Either way, eliminating MUF did result in a quicker time from CPB weaning to leaving the OR and shortened the patients’ exposure to non-endothelial surfaces. At centers that perform MUF for a longer time period on average than we do, the time benefit of eliminating MUF could be more significant.

Outcomes

While many studies have shown decreased VT, ICU and hospital LOS with MUF, many other studies have shown little or no benefit.^{3,13,14,19,22,24,25} We wanted to confirm that eliminating MUF didn’t have a negative impact on outcomes. There was no difference between groups in either cohort for any of the outcome measures studied. Any impact on outcomes from MUF is likely related to its impact on transfusion. Although MUF is intended to decrease transfusion, in our case, because of the way our circuit was set up, MUF was actually contributing to transfusion. None of the MUF studies which show decreased transfusion and improved outcomes report comparing a MUF group to a control group with a smaller, simpler CPB circuit.

Hemodynamics (VIS)

In our experience, weaning with a high Hct (>40%) in No-MUF patients often results in very good hemodynamics (low central venous pressure (CVP), high arterial blood pressure (ABP), low inotrope requirement). This may translate into less volume being needed to fill the patient (lower CVP) compared to weaning with a lower Hct in MUF patients. Being able to wean with a lower CVP would further enhance the volume benefit previously described (less volume needed to fill the patient for weaning = more volume available for hemoconcentration) and further improve our ability to hemoconcentrate prior to weaning. It is possible that the improvement in hemodynamics often observed during MUF (increased ABP with decreasing CVP) is not due to inflammatory mediator removal or a sudden decrease in myocardial edema. A more likely explanation is a rapid increase in blood viscosity during MUF when weaning at a relatively low Hct and quickly achieving a much higher Hct. Weaning at a low Hct and attempting to quickly increase the Hct via MUF could result in unnecessary hypotension, higher inotropic requirements and needing to re-institute CPB due to hemodynamic instability. We used VIS (a measure of the amount of inotropic support a patient requires) as an indicator of the patient’s hemodynamic state. In our study, the VIS was lower in the No-MUF groups in both cohorts during all time periods. Lower VIS could be due to increased blood viscosity; however, this would only explain the lower VIS-CPB, since the Hct was generally similar between groups except when weaning from CPB. Thus, decreased transfusion rates in the No-MUF groups may have been the biggest factor in them having lower VIS during all three time points. There is evidence in the literature supporting a positive correlation between transfusion and VIS.²⁵

Safety

By no longer performing MUF, we have eliminated the possibility of any of the MUF-related adverse events or complications. Performing MUF required significant manipulation of the CPB circuit and heart-lung machine just prior to weaning in order to convert from CUF during CPB to MUF post-CPB. Some manipulations were designed to enhance safety during MUF. These manipulations often had to be reversed if CPB needed to be re-initiated or in the case of failure to wean. This added to the complexity of the weaning and re-initiation periods and could lead to errors, particularly if CPG needed to be re-administered. In addition, re-initiation of CPB or failing to wean after setting up for MUF often resulted in RBC transfusion. While we didn’t evaluate how often re-initiation or failure to wean after priming the MUF circuit occurred, once the MUF circuit is primed with blood, this volume is essentially tied up in the CPG circuit and not easily accessible, often necessitating transfusion.

Blood product administration during MUF

Previously, we reported on the technique of giving Plt and Cryo during MUF as a means to eliminate hemodilution caused by the administration of these products post-MUF.²⁶ Also, by giving these products during MUF, they could be given quicker since the administration of the products isn’t limited by the patient’s vascular capacity. The goals were to have an increased Hct post MUF/product administration, achieve earlier hemostasis and, as a result, transfuse less blood products. However, our results did not favor this technique over standard-MUF and it was abandoned just prior to completely eliminating MUF.

There are a few potential problems with this technique which we did not appreciate until after this study. When the Cryo/Plt are given during MUF, they are given prior to protamine. Subsequently, they may be given in the presence of significant bleeding and may be ‘lost’ prior to having their effect on hemostasis. As a result, many of the blood products could potentially bleed out and get wasted, either by being suctioned directly to the cell saver or indirectly by being sent to the CPB circuit via the pump suckers and eventually ending up in the cell saver. This may result in needing a greater volume of products to eventually achieve hemostasis. Our analysis showed we often had to give additional Cryo/Plt after MUF when using this technique. The most obvious problem with giving Cryo/Plt prior to protamine is it eliminates the ability to use the clinical picture to guide the transfusion of blood products. By giving the products during MUF instead of after protamine you are: 1) assuming the patient will need them and 2) estimating how much to give (based on lab values or experience) rather than observing how much non-surgical bleeding is present and transfusing as needed. This may result in the unnecessary transfusion or over-transfusion of products. This technique warrants further evaluation.

Study limitations

This was a single-center, non-randomized, retrospective evaluation including a consecutive series of patients. As such, there were some potentially important differences between our groups. In the pediatric cohort, Aristotle Scores were significantly higher in the MUF group. In the neonatal cohort, more MUF patients had single ventricle anatomy and underwent DHCA. While this could skew our results in favor of No-MUF, we feel that, even if the groups were more similar, our MUF results would be similar to our No-MUF results, but not better. Also, being a retrospective study, we did not have strict criteria (other than our normal protocols) for transfusion or evaluating readiness to extubate, transfer, etc. In addition, since the data was collected by some of the same people involved in the decision to stop performing MUF, sample bias is a possibility. However, strict inclusion and exclusion criteria were adhered to during the patient selection process. Also, as a retrospective study, it is possible that other changes in practice during the study period may have impacted results, especially considering our MUF groups were from the historically older group. However, the study only covered a three-year period and no significant changes occurred during that time (besides eliminating MUF and simplifying our CPB circuit). It is possible, with a team consisting of three surgeons, that surgical differences could have impacted some of our results. We did not evaluate the distribution of cases between surgeons within the groups. As a result of these limitations, just evaluating our No-MUF groups’ data alone, we feel confident eliminating MUF because we have low transfusion rates, very low 2^nd RBC exposure rates, low CTO, often low inotropic requirements and continue to have good outcomes.

Conclusion

Eliminating MUF allowed us to simplify our CPB circuit, reduce prime volumes and decrease the likelihood of reaching our RBC transfusion trigger on CPB. This analysis confirms that eliminating MUF, when it allows such circuit miniaturization, may be beneficial. Eliminating MUF did not have a negative impact on outcomes and favored many of the parameters evaluated (transfusion, CTO, inotrope usage, time from weaning to chest closure). Centers that perform MUF similar to the way we did, with practices similar to ours, might benefit from eliminating MUF and simplifying their CPB circuit. However, our results cannot necessarily be extrapolated to other centers and other patient populations. Centers that have different MUF/CPB circuits and techniques may not benefit from eliminating MUF. Eliminating MUF may also have cost benefits, both indirectly via lower transfusion rates and directly via reduced CPB circuit costs. Our analysis supports an advantage of preventing hemodilution (via circuit miniaturization) versus reversing hemodilution (via MUF) at our institution with the patient population we examined. This advantage was particularly evident in neonates.

Footnotes

Acknowledgements

We would like to acknowledge Pan Zhaoxing for his assistance with statistical analysis.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

References

Naik

Knight

Elliott

MJ.

A successful modification of ultrafiltration for cardiopulmonary bypass in children. Perfusion 1991; 6: 41–50.

Kotani

Honjo

Osaki

. Effect of modified ultrafiltration on postoperative course in neonates with complete transposition of the great arteries undergoing arterial switch operation. Circulation 2008; 72: 1476–1480.

Wang

Huang

Zhu

. Modified ultrafiltration in paediatric cardiopulmonary bypass. Perfusion 1998; 13: 304–310.

Sever

Tansel

Basaran

. The benefits of continuous ultrafiltration in pediatric cardiac surgery. Scand Cardiovasc J 2004; 38: 307–311.

Draaisma

Hazekamp

Frank

. Modified ultrafiltration after cardiopulmonary bypass in pediatric cardiac surgery. Ann Thorac Surg 1997; 64: 521–525.

Friesen

Campbell

Clarke

. Modified ultrafiltration attenuates dilutional coagulopathy in pediatric open heart operations. Ann Thorac Surg 1997; 64: 1787–1789.

Hennein

Kiziltepe

Barst

. Venovenous modified ultrafiltration after cardiopulmonary bypass in children: a prospective randomized study. J Thorac Cardiovasc Surg 1999; 117: 496–505.

Naik

Knight

Elliott

A prospective randomized study of a modified technique of ultrafiltration during pediatric open-heart surgery. Circulation 1991; 84: III422–431.

Bando

Turrentine

Vijay

. Effect of modified ultrafiltration in high-risk patients undergoing operations for congenital heart disease. Ann Thorac Surg 1998; 66: 821–827; discussion 828.

10.

Koutlas

Gaynor

Nicolson

. Modified ultrafiltration reduces postoperative morbidity after cavopulmonary connection. Ann Thorac Surg 1997; 64: 37–42; discussion 43.

11.

Aggarwal

Das

Sharma

. Efficacy of combined modified and conventional ultrafiltration during cardiac surgery in children. Ann Card Anaesth 2007; 10: 27–33.

12.

Davies

Nguyen

Gaynor

. Modified ultrafiltration improves left ventricular systolic function in infants after cardiopulmonary bypass. J Thorac Cardiovasc Surg 1998; 115: 361–369; discussion 369–370.

13.

Keenan

Thiagarajan

Stephens

. Pulmonary function after modified venovenous ultrafiltration in infants: a prospective, randomized trial. J Thorac Cardiovasc Surg 2000; 119: 501–505; discussion 506–507.

14.

Mahmoud

Burhani

Hannef

. Effect of modified ultrafiltration on pulmonary function after cardiopulmonary bypass. Chest 2005; 128: 3447–3453.

15.

Harvey

Shann

Fitzgerald

. American Society of ExtraCorporeal Technology’s International Consortium for Evidence-Based Perfusion and Pediatric Perfusion Committee. International pediatric perfusion practice: 2011 survey results. J Extra Corpor Technol 2012; 44: 186–193.

16.

Medlin

Sistino

JJ.

Cerebral oxygen saturation changes during modified ultrafiltration. Perfusion 2006; 21: 325–328.

17.

Darling

Nanry

Shearer

. Techniques of paediatric modified ultrafiltration: 1996 survey results. Perfusion 1998; 13: 93–103.

18.

Gaies

Jeffries

Niebler

. Vasoactive-inotropic score is associated with outcome after infant cardiac surgery: an analysis from the Pediatric Cardiac Critical Care Consortium and Virtual PICU System Registries. Pediatr Crit Care Med 2014; 15: 529–537.

19.

Thompson

McElhinney

Findlay

. A prospective randomized study comparing volume-standardized modified and conventional ultrafiltration in pediatric cardiac surgery. J Thorac Cardiovasc Surg 2001; 122: 220–228.

20.

Charette

Hirata

Bograd

. 180 ml and less: cardiopulmonary bypass techniques to minimize hemodilution for neonates and small infants. Perfusion 2007; 22: 327–331.

21.

Merkle

Boettcher

Schulz

. Perfusion technique for nonhaemic cardiopulmonary bypass prime in neonates and infants under 6 kg body weight. Perfusion 2004; 19: 229–237.

22.

Williams

Ramamoorthy

Chu

. Modified and conventional ultrafiltration during pediatric cardiac surgery: clinical outcomes compared. J Thorac Cardiovasc Surg 2006; 132: 1291–1298.

23.

Kuratani

Bunsangjaroen

Srimueang

. Modified versus conventional ultrafiltration in pediatric cardiac surgery: a meta-analysis of randomized controlled trials comparing clinical outcome parameters. J Thorac Cardiovasc Surg 2011; 142: 861–867.

24.

Berdat

Eichenberger

Ebell

. Elimination of proinflammatory cytokines in pediatric cardiac surgery: analysis of ultrafiltration method and filter type. J Thorac Cardiovasc Surg 2004; 127: 1688–1696.

25.

Maluf

MA.

Modified ultrafiltration in surgical correction of congenital heart disease with cardiopulmonary bypass. Perfusion 2003; 18: 61–68.

26.

Mejak

Ing

McRobb

. Cryoprecipitate and platelet administration during modified ultrafiltration in children less than 10 kg undergoing cardiac surgery. J Extra Corpor Technol 2013; 45: 107–111.