High Flow Rates During Modified Ultrafiltration Decrease Cerebral Blood Flow Velocity and Venous Oxygen Saturation in Infants

Transcription

1 High Flow Rates During Modified Ultrafiltration Decrease Cerebral Blood Flow Velocity and Venous Oxygen Saturation in Infants Rosendo A. Rodriguez, MD, PhD, Marc Ruel, MD, MPH, Lothar Broecker, CPC, and Garry Cornel, MB, BS Departments of Anesthesiology, Cardiac Division, Surgery, Division of Cardiac Surgery, and Epidemiology, University of Ottawa Heart Institute and Children s Hospital of Eastern Ontario, Ottawa, Ontario, Canada Background. The intracranial hemodynamic effects of modified ultrafiltration in children are unknown. We investigated the effects of different blood flow rates during modified ultrafiltration on the cerebral hemodynamics of children with weights above and below 10 kg. Methods. Thirty-one children (weights: < 10 kg, n 21; > 10 kg, n 10) undergoing cardiopulmonary bypass were studied. Middle-cerebral artery blood flow velocities and cerebral mixed venous oxygen saturations were measured before, five minutes from the beginning, and at the end of ultrafiltration. Patients were classified according to their blood flow rates during ultrafiltration in three groups: high (> 20 ml/kg/min), moderate (10 19 ml/kg/ min), and low flow rates (< 10 ml/kg/min). Results. During modified ultrafiltration, blood pressures and hematocrit increased (p < 0.001), but cerebral blood flow velocities and mixed venous oxygen saturations decreased (p < 0.001). A significant correlation was found between blood flow rates of ultrafiltration and the decline in mean cerebral blood flow velocity (r 0.48; p 0.005) and cerebral oxygen saturation (r 0.49; p 0.005) or hematocrit increase (r 0.59; p 0.001). Infants exposed to high flow rates had greater reduction of cerebral blood flow velocity and regional mixed venous saturation and higher hematocrit at the end of ultrafiltration compared with those subjected to moderate and low flow rates (p < 0.04). No significant difference was found between moderate and low flow groups. The flow rate of ultrafiltration was the only independent predictor of the changes in cerebral mixed venous oxygen saturation (p 0.033). Conclusions. High blood flow rates through the ultrafilter during modified ultrafiltration transiently decrease the cerebral circulation in young infants compared with lower blood flow rates. These effects may be related to an increased diastolic runoff from the aorta into the ultrafiltration circuit that leads to a stealing effect from the intracranial circulation, which may be important in infants with dysfunctional cerebral autoregulation. (Ann Thorac Surg 2005;80:22 8) 2005 by The Society of Thoracic Surgeons Modified ultrafiltration (MUF) has previously been shown to be beneficial in pediatric cardiac surgery [1 3]. Young infants and neonates in particular, who often experience long cardiopulmonary bypass (CPB) times, hypothermia, and hemodilution, appear to show the greatest benefit from MUF because of their excessive accumulation of water and exaggerated inflammatory response [3 5]. During MUF, the volume driven through the ultrafilter is taken from the patient by withdrawing blood from the aortic cannula and arterial line. In an effort to decrease the duration of MUF, high blood flow rates through the ultrafilter are often used [2, 4]. However, the resultant rapid withdrawal of blood from the aortic cannula at high flow rates, particularly in small infants who weigh less than 10 kg, may cause diastolic runoff from the aorta and steal flow from the carotid circulation [2, 6]. In addition, the rapid intravascular Accepted for publication Jan 10, Address reprint requests to Dr Rodriguez, University of Ottawa Heart Institute, Room H-341, 40 Ruskin Street, Ottawa, Ontario K1Y 4W7, Canada; rrodriguez@ottawaheart.ca. changes that result from temporary imbalance between extracted and replaced volumes may exacerbate these effects [6, 7]. Previous studies [7, 8] have indicated that carotid stealing may be associated with reduced cerebral perfusion. This phenomenon may be particularly critical in newborn infants with dysfunctional cerebral autoregulation [9, 10]. However, despite the above data and the extensive use of MUF in clinical practice, the effects of using high blood flow rates on the intracranial cerebral hemodynamics of young children are still unknown. The spectral characteristics of the middle cerebral artery (MCA) blood flow velocity profile can be recorded with transcranial Doppler (TCD) and provide information about the functional status of the intracranial cerebral hemodynamics in children [11]. The velocity changes of major basal arteries such as the MCA show a good correlation with changes in cerebral blood flow [12, 13]. In addition, noninvasive cerebral oximetry using nearinfrared spectrophotometry (NIRS) measures changes in the regional cerebral mixed venous oxygen saturation 2005 by The Society of Thoracic Surgeons /05/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg RODRIGUEZ ET AL 2005;80:22 8 MUF AND CEREBRAL PERFUSION IN INFANTS (rcvos) [14]. The NIRS has been used to describe the relationship between the cerebral oxygen supply and demand in critically ill neonates, infants, children, and adults [15, 16]. In this study, we used TCD and NIRS to evaluate the effects of different MUF flow rates on the cerebral hemodynamics in children with weights above and below 10 kg. We tested the hypothesis that high MUF flow rates, particularly in small infants ( 10 kg), would result in periods of low MCA blood flow velocities and reduced rcvos. Patients and Methods Patients and Procedures This study included pediatric patients who required surgery for congenital heart disease and in whom MUF was used. The study was approved by the Institutional Research Ethics Board. The clinical decision to perform MUF in these patients was based on the patient s hemodynamic conditions, weight, hemodilution, and length of CPB. The MUF circuit consisted of a blood cardioplegia device with a heat exchanger and an ultrafilter. Blood was drawn from the aortic cannula through the arterial line, passed through the ultrafilter, and returned to the right atrium through a venous cannula. The vacuum line of the ultrafilter was set between 100 and 120 mm Hg. According to our standard institutional practice, MUF flow rates between 70 and 100 ml/min were used in children with weights equal or less than 10 kg. Flow rates of 150 ml/min were applied for those above 10 kg. Data Analysis In order to control for interindividual variations, we expressed our measurements in absolute changes (measurement minus baseline) from the baseline (eg, hematocrit and rcvos) or in percentages (measurement/ baseline) relative to the baseline (eg, blood pressure, flow velocities, heart rate). Patients were classified according to their MUF flow rates in three groups: high ( 20 ml/kg/min), moderate (10 to 19 ml/kg/min), and low flow rates ( 10 ml/kg/min). Nonparametric statistics were used. Spearman rank correlation analysis examined the associations between MUF flow rates (ml/kg/min) and continuous clinical variables. Nonparametric Mann-Whitney tests for independent samples estimated the differences between the flow groups with respect to demographics and hemodynamic parameters. Pairwise comparisons with the Wilcoxon sign rank or the Kruskall-Wallis tests assessed for differences between measurements. All tests used p 0.05 as the threshold value for statistical significance. When multiple comparisons were performed, the alpha value was adjusted using Bonferroni correction. A multivariable linear regression analysis was performed on the cerebral indicators by incorporating significant univariate descriptors and by forcing into the model several physiologic and intraoperative characteristics. Analyses were performed using SPSS version 10.0 (SPSS, Chicago, IL) and Intercooled Stata Version 8 (Stata, College Station, TX). Continuous demographic factors and physiologic variables are presented as median (25th, 75th quartiles). 23 CARDIOVASCULAR Brain Monitoring Prior to surgical incision, a 2-MHz pulsed wave Doppler flat probe connected to a TCD system (Medasonics, Plus, Nicolet, Madison, WS) was secured with an elastic band on the child s right temporal window for monitoring the cerebral blood flow velocities (CBFV) in the MCA. Established acoustic and visual guidelines were followed to identify the M-1 segment of the MCA [17]. In addition, a NIRS patch (INVOS 3100, Somanetics, Troy, MI) was applied on the right forehead (receivers, 30 and 40 mm) to monitor the rcvos throughout the operation. The CBFV and rcvos were assessed as follows: (a) immediately before the initiation of MUF (pre-muf baseline); (b) 5 minutes after the beginning; and (c) at the end of the ultrafiltration procedure. During MUF, changes in the anesthetic and hemodynamic management were minimized and the inspired fraction of oxygen (Fio 2 ) and end-tidal CO 2 (ETco 2) were maintained constant. Hematocrit, arterial ph, systemic arterial oxygen saturation (Sao 2 ), blood pressure, central venous pressure, and heart rate were documented immediately before MUF and at the end of the ultrafiltration procedure. All patients were followed after surgery; the first postoperative hematocrit at arrival in the intensive care unit, length of stay, and the presence of cardiac and neurologic complications at one-month follow-up were all documented. Results Demographics We studied 31 pediatric patients. Table 1 summarizes associated comorbidity, surgical information, and postoperative complications according to MUF flow groups. All patients survived and were discharged from hospital in stable condition. Twenty-one of the studied patients had weights less than or equal to 10 kg. Nine of these patients had high, 11 moderate, and one low flow rates during MUF. In contrast, 8 of the 10 children who had weights greater than 10 kg were exposed to low flows and only 2 to moderate flows. The median length of stay was 6 days [range, 2 to 35 days]. There were no differences in the length of stay (high: 8 days [6, 19 days]; moderate: 7 days [4, 24 days]; low: 5 days [2, 10 days]; p 0.90], extracted volume of the ultrafiltrate (high: 550 ml [462, 637 ml]; moderate: 500 ml [425, 675 ml]; low: 550 ml [550, 650 ml]; p 0.46), or postoperative hematocrit (high: 0.37 [0.30, 0.47]; moderate: 0.30 [0.23, 0.38]; low: 0.28 [0.25, 0.34]; p 0.13) among the three groups. In addition, the durations of aortic cross-clamping (high: 79 min [44, 100 min]; moderate: 55 min [13, 82 min]; low: 48 min [24, 140 min]; p 0.94) or modified ultrafiltration (high: 12 min [10, 15 min]; moderate: 11 min [10, 15 min]; low: 11 min [9, 14 min], p 0.19] were not different among groups.

3 24 RODRIGUEZ ET AL Ann Thorac Surg MUF AND CEREBRAL PERFUSION IN INFANTS 2005;80:22 8 Table 1. Clinical History, Surgical Characteristics, and Postoperative Complications Characteristic Low Flow a Moderate Flow a High Flow a Preoperative comorbidity: Down s syndrome Prematurity Intraventricular hemorrhage Intraoperative procedure: Total circulatory arrest b Bidirectional Glenn or TCPC Repair of atrial and/or ventricular septal defects Correction of tetralogy of Fallot Valve repairs Postoperative complications: Choreoathetosis c Seizures (no electroencephalographic changes) Neurodevelopmental delay d Sepsis Cardiac rhythm disturbances e a Values represent number of patients; b correction of TAPVR or transposition of Great Vessels; c presence of choreoathetosis was observed in the patient with preoperative history of intraventricular hemorrhage; d developmental delay was found in patients with Down s syndrome; e electrocardiographic evidence of junctional rhythm, third-degree AV-block or episodes of ventricular tachycardia. AV atrioventricular; TAPVR total anomalous pulmonary venous return; TCPC total cavopulmonary connection. Cerebral Effects of MUF Flow Rates During MUF, arterial blood pressures and hematocrit increased (p 0.001), but cerebral blood flow velocities and rcvos decreased (p 0.001). Often, the end of ultrafiltration was associated with the lowest values in cerebral blood flow velocities and rcvos. At the end of MUF, the rcvos had declined by 4% (7%, 1%) and the peak, mean, and diastolic MCA-CBFV were reduced by 23% (35%, 17%), 28% (43%, 19%), and 18% (56%, 0%), respectively, from baseline. In contrast, systolic, mean, and diastolic arterial blood pressures increased by 12% (3%, 26%), 22% (2%, 35%), and 28% (10%, 45%) while the hematocrit improved by 7% (5%, 12%) from pre-muf values. Before MUF, 21 children had rcvos values greater than 60%, 8 between 52% and 59%, and only 2 below 52%. At the end of MUF, the distribution of patients within these rcvos ranges changed to 18, 8, and 5, respectively. A significant correlation was found between MUF blood flow rates and the decline in mean (r 0.487, p 0.005), peak (r 0.42, p 0.02), and diastolic (r 0.34, p 0.06) MCA blood flow velocities or rcvos (r 0.493, p 0.005). In this relationship, higher blood flow rates through the ultrafilter were associated with a greater decline in CBFV or rcvos. In addition, MUF flow rates were associated with improvement in hematocrit Fig 1. Relationship between mean cerebral blood flow velocities (Mean CBFV) and blood flow rates during modified ultrafiltration (MUF) in 31 children (Œ) after cardiopulmonary bypass. Flow velocities in the middle cerebral artery are expressed in percentages (%) relative to the pre-muf baseline. Higher blood flow rates of ultrafiltration during MUF were associated with a significant reduction in the middle cerebral artery flow velocities (r 0.49; p 0.005). The horizontal dotted line indicates the baseline (100%) for all patients. (r 0.587, p 0.001); however, there was no correlation between changes in hematocrit and the alterations in mean CBFV (r 0.32, p 0.15) or rcvos (r 0.018, p 0.93). Figures 1 and 2 show the associations between MUF flow rates and the changes in mean CBFV and rcvos, respectively. MUF Flow Groups Patients in the high flow group were younger (p 0.001) and had a lower weight (p 0.001) than those in the moderate and low flow groups. Table 2 shows the demographic and baseline physiologic characteristics according to flow group. There were no significant differences in the baseline characteristics, CVP, ETco 2,Fio 2,Sao 2, core- Fig 2. Relationship between the absolute changes ( ) in transcranial mixed venous oxygen saturation (rcvos) and blood flow rates during modified ultrafiltration (MUF) in 31 children (Œ) after cardiopulmonary bypass. The rcvos are expressed as the difference between the values of the rcvos measured before MUF and those recorded at the end of MUF. The use of high blood flow rates during MUF were associated with a significant reduction in the rcvos as indicated by more negative values on the rcvos (r 0.49; p 0.005). The horizontal dotted line indicates no absolute change ( rcvos 0%).

4 Ann Thorac Surg RODRIGUEZ ET AL 2005;80:22 8 MUF AND CEREBRAL PERFUSION IN INFANTS Table 2. Demographic and Baseline Physiologic Characteristics (Median [25th, 75th Quartiles]) Characteristic Low Flow Moderate Flow High Flow Number of patients 10 kg Number of patients 10 kg Weight (kg) 19 (16,22) 7 (6,8) c 4 (3,5) a,b Age (months) 73 (60,98) 9 (5,15) c 2 (0.4,3) a,b CPB duration (min) 93 (75,211) 122 (87,170) 140 (95,207) Hematocrit 0.19 (0.17,0.25) 0.23 (0.21,0.26) 0.23 (0.21,0.29) Peak CBFV (cm/s) 83 (75,94) 71 (58,89) 50 (44,73) a,b Mean CBFV (cm/s) 47 (50,52) 39 (30,52) 24 (19,27) a,b Diastolic CBFV (cm/s) 28 (16,31) 25 (14,36) 5 (2,9) a,b Cerebral venous O 2 saturation 65 (60,79) 65 (54,72) 61 (54,75) Systolic blood pressure (mm Hg) 78 (74,89) 77 (70,88) 64 (51,85) Mean blood pressure (mm Hg) 55 (54,59) 52 (48,64) 43 (36,57) Core (rectal) temperature ( C) 35.5 (35.1,35.7) 36.0 (35.0,36.4) 35.0 (34.5,35.4) Central venous pressure (mm Hg) 9 (9,11) 9 (9,11) 10 (4,12) Heart rate (beats/min) 108 (99,120) 140 (114,147) c 141 (117,156) a,b ETco 2 (mm Hg) 29 (25,32) 33 (27,37) 25 (22,30) Inspired fraction of O 2 (%) 100 (100,100) 100 (97,100) 100 (100,100) Systemic arterial O 2 saturation (%) CARDIOVASCULAR Mann-Whitney (Bonferroni correction) [p 0.05]: a High vs low flow; b high vs moderate flow; c moderate vs low flow. CBFV cerebral blood flow velocities; CPB cardiopulmonary bypass. temperature, aortic cross-clamp time, CPB duration, or volume of the ultrafiltrate among flow groups (see Table 2). In contrast, the heart rate was significantly higher for patients in the high (p 0.03) and moderate (p 0.027) flow groups compared to those in the low flow group. In addition, the pre-muf baseline mean, peak, and diastolic flow velocities had lower values in the high flow group compared with the other two groups, but there were no differences in the baseline rcvos (see Table 2). The effects of MUF on CBFV, rcvos, and hematocrit were greater for children exposed to high flow rates as opposed to those in the other two groups, but there were no significant differences regarding changes in heart rate, CVP, Fio 2,ETco 2, and MUF durations among all three groups. Table 3 summarizes the cerebral hemodynamic changes for all flow groups. The decline in the mean CBFV (p 0.042) and rcvos (p 0.033) or increases in hematocrit (p 0.006) was greater for infants with high blood flow rates compared with children under low flow rates; however, their differences in systolic (p 0.22), mean (p 0.072), and diastolic (p 0.09) blood pressures were only marginal. A reduction in the baseline rcvos Table 3. Changes in Hemodynamic Parameters (Median [25th, 75th Quartiles]) Parameter Low Flow Moderate Flow High Flow Post-MUF hematocrit 0.25 (0.22,0.29) 0.30 (0.26,0.34) 0.36 (0.30,0.47) a Hematocrit (%) 6 (3,6) 7 (5,9) 13 (10,19) a rcvos (%) 1 ( 2, 1) 2 ( 5, 1) 8 ( 15, 3) a Mean CBFV (%) 19 ( 20, 3) 28 ( 47, 23) 36 ( 43, 29) a Peak CBFV (%) 18 ( 24, 15) 27 ( 38, 14) 28 ( 42, 19) Diastolic CBFV (%) 3 ( 16, 14) 48 ( 69, 32) c 33 ( 94,0) Systolic blood pressure (%) 7 (4,23) 7 ( 1,2) 24 (12,48) b Mean blood pressure (%) 12 (4,20) 5 ( 3,24) 40 (29,50) b Diastolic blood pressure (%) 20 (11,33) 15 (3,29) 47 (38,60) b Heart rate (%) 1 ( 5,2) 2 ( 12,3) 2 ( 8,9) Central venous pressure (mm Hg) 7 (5,13) 10 (6,12) 9 (4,11) Systemic arterial O 2 saturation 100 (100,100) 100 (98,100) 100 (96,100) Inspired fraction of oxygen 82 (40,94) 41 (27,51) 38 (32,69) Mann-Whitney (Bonferroni correction) [p 0.05]: a High vs low flow; b high vs moderate flow; c moderate vs low flow. BP blood pressure; CBFV cerebral blood flow velocity; MUF modified ultrafiltration; rcvos regional cerebral mixed venous oxygen saturation; change, expressed either as absolute change (rcvos and hematocrit) or percentage (CBFV, blood pressure, heart rate) from baseline.

5 26 RODRIGUEZ ET AL Ann Thorac Surg MUF AND CEREBRAL PERFUSION IN INFANTS 2005;80:22 8 Table 4. Linear Regression Prediction Model a of the Changes in Regional Cerebral Mixed Venous Oxygen Saturation During Modified Ultrafiltration Variable Coefficient 95% CI p Significant independent predictors MUF flow rate (ml/kg/min) , Other variables included in the model Pre-MUF hematocrit (%) , Pre-MUF mean arterial pressure (mm Hg) , Weight (kg) , Age (days) , Type of surgery b , a Prediction model significant at p (adjusted R ); b grouped in the model as: I: complex procedures such as correction of TAPVR, TGV, bidirectional Glenn procedures, TCPC, TOF; II: repair of atrial and/or ventricular septal defects; and III: valve repairs. MUF modified ultrafiltration; TAPVR total anomalous pulmonary venous return; TCPC total cavopulmonary connection; TGV transposition of Great Vessels; TOF tetralogy of Fallot. greater than 13% during MUF was documented in 3 infants who were exposed to high MUF flow rates (rcvos change: 17%, 15%, 14%) and only one in the moderate flow group (rcvos change: 16%). A decline in pre-muf rcvos between 5% and 8% from pre-muf values was observed in 5 infants of the high flow group but in only 3 of the moderate flow group. None of the patients exposed to low MUF flow rates had reductions in rcvos greater than 5%. Patients in the high flow group had significantly lower rcvos (p 0.043) and comparable reductions in CBFV at the end of MUF relative to those of the moderate flow group (see Table 3), but their improvement in hematocrit was similar (p 0.063). This was in contrast to the changes in systolic (p 0.051), mean (p 0.003), and diastolic (p 0.021) blood pressures that were higher in the high flow group. In addition, the effects on the peak (p 0.29) and mean (p 0.09) flow velocities between the moderate and low flow groups were nonsignificant. This was in contrast to the diastolic flow velocities that remained lower in patients with moderate flow rates compared with those under low flows (p 0.033). No significant differences were found in rcvos (p 0.43), hematocrit (p 0.16) or systolic (p 0.36), mean (p 0.60), and diastolic (p 0.60) blood pressures between these two groups. Multivariate Predictors of Cerebral Venous Oxygen Saturation Table 4 summarizes the results of the multivariable regression analysis. The MUF flow rate was the only independent predictor of the changes in rcvos during MUF, with an average decline from baseline rcvos of 0.4% (95% confidence interval: 0.8%, 0.04%; p 0.033) for each ml kg min increase in MUF flow rate. The type of surgery, weight, age, baseline hematocrit, or mean arterial pressure had no significant effects on the changes in rcvos. Comment Cardiopulmonary bypass in pediatric cardiac surgery exposes young infants and neonates to hemodilution, hypothermia, and systemic inflammatory response [4]. The resultant increase in total body water, particularly in the interstitial tissue, can have adverse effects on the function of the lungs, heart, and brain [3]. Modified ultrafiltration has been developed as a method to reduce these adverse effects in pediatric cardiac patients [2]. However, the possibility of rapid intercompartmental fluid shifts leading to hemodynamic instability and the risk of stealing flow from the carotid circulation in small infants are some of the concerns associated with this procedure. The effects of these hemodynamic changes on the intracranial cerebral circulation have not been studied. Our findings indicate that middle cerebral artery blood flow velocities and cerebral mixed venous oxygen saturations decrease during MUF despite mild increases in systemic blood pressures or hematocrit. These cerebral effects were greater when high blood flow rates through the MUF circuit were used. In addition, we found that the change in MUF flow rate was the only independent predictor of the changes in regional mixed venous oxygen saturation during MUF. Small infants, particularly those under 10 kg, who were exposed to the highest blood flow rates in our study ( 20 ml/kg/min), showed greater decline in both physiologic indicators compared with older children receiving lower blood flow rates. The reduction on cerebral blood flow velocities and mixed venous oxygen saturation during high MUF flow rates in small infants may be indicative of transiently reduced cerebral perfusion. These hemodynamic changes may be in part associated with an increased diastolic runoff from the aorta as a consequence of extracting higher blood volumes from the child s aorta into the MUF circuit [2, 6]. This phenomenon eventually leads to a stealing effect from the head and neck vessels [6, 8]. These effects might be amplified in young infants because of their smaller circulating blood volume and the short anatomical distance between the head vessels and the aorta. Cerebral stealing is a phenomenon that has been previously described associated with arteriovenous shunts [8, 18, 19] and subclavian artery stenosis [20, 21].

6 Ann Thorac Surg RODRIGUEZ ET AL 2005;80:22 8 MUF AND CEREBRAL PERFUSION IN INFANTS However, this report documents this phenomenon associated with the use of high MUF flow rates. Positron emission tomography studies [20] have documented that during cerebral stealing blood flow to the brain decreases, oxygen extraction fraction increases, and the cerebral mixed venous oxygen saturation as measured by NIRS decreases [21]. In the presence of stealing, cerebral vascular resistance falls in an attempt to diminish the severity of the runoff [18]. However, when cerebral perfusion pressure decreases, the cerebral vasculature is unable to compensate for these changes and cerebral blood flow velocities and noninvasive cerebral mixed venous oxygen saturation are consequently reduced [8]. The implications of cerebral stealing during high MUF flow rates may be even more relevant after repair of complex procedures [22, 23] because some of these critically ill newborns may have a dysfunctional cerebral autoregulation [9]. Because our study did not examine the cognitive function of our patients, the potential impact of these hemodynamic changes on the brain function of these young infants is unknown. However, our study raises concerns about the use of safe MUF flow rates in small infants, particularly in those that require prolonged periods of CPB or longer episodes of deep hypothermic circulatory arrest. This is an important issue in pediatric cardiac surgery in order to minimize the exposure of young infants to the effects of reduced cerebral perfusion. During MUF, 10 of our pediatric patients had a reduction in cerebral blood flow velocities greater than 40%, but only 4 had a decline in rcvos values greater than 13% (see Figs 1 and 2). A 10% change in the rcvos at any time or a reduction in absolute rcvos values below 50% have been associated with a decreased amplitude in the somatosensory-evoked potentials after carotid occlusion during carotid endarterectomy [24]. A 40% reduction in MCA blood flow velocity, however, has been observed during the clinical manifestations of syncope or presyncope in patients with neurocardiogenic syncope [25]. In contrast, the threshold for neurologic manifestations of regional ischemia after carotid occlusion appears to be achieved when the relative decline in rcvos is greater than 20% from the preocclusion baseline [26]. Our findings did not show an association between changes in hematocrit and the alterations in blood flow velocities or oxygen saturation. Previous studies [27, 28] have found an inverse relationship between hematocrit and cerebral blood flow velocity or cerebral blood flow. Their results indicate that these effects appear to be primarily related to changes in the arterial oxygen content and brain oxygen delivery associated with a higher hematocrit [29]. In addition, a higher hematocrit level has been found associated with increased concentrations of the oxygenated fraction of hemoglobin that results in increased oxygen saturation as measured by the NIRS technique [29]. This is in contrast to our findings that indicate a reduction as opposed to an augmentation in the cerebral mixed venous oxygenation. We speculate that our changes were more likely indicative of an increased cerebral oxygen demand and/or reduced oxygen supply associated with the hemodynamic changes observed during MUF. In summary, our study indicates that high blood flow rates through the ultrafilter during MUF in small infants ( 10 kg) decrease cerebral blood flow velocities and transcranial mixed venous oxygen saturation compared with the use of lower blood flow rates in older children. A possible explanation of these effects is related to an increased diastolic runoff from the aorta into the MUF circuit that steals flow from the cerebral circulation. The changes in the cerebral circulation associated with the use of high MUF flow rates may be important, particularly after deep hypothermic circulatory arrest and in those newborn infants with dysfunctional cerebral autoregulation. The clinical relevance of these hemodynamic changes on the postoperative neurologic outcome of young infants needs to be determined. The authors are thankful for the participation of Dr Nihal Weerasena in this study. The cooperation of the staff from Anesthesia, Cardiovascular Surgery, Perfusion, and Nursing at Children s Hospital of Eastern Ontario during this study is gratefully appreciated. References 1. Chaturvedi RR, Shore DF, White PA, et al. Modified ultrafiltration improves global left ventricular systolic function after open-heart surgery in infants and children. Eur J Cardiothorac Surg 1999;15: Maluf MA. Modified ultrafiltration in surgical correction of congenital heart disease with cardiopulmonary bypass. Perfusion 2003;18[Suppl 1]: Montenegro LM, Greley WJ. The use of modified ultrafiltration during pediatric cardiac surgery is a benefit. J Cardiothorac Vasc Anesth 1998;12: Bando K, Turrentine MW, Vijay P, et al. Effect of modified ultrafiltration in high risk patients undergoing operations for congenital heart disease. Ann Thorac Surg 1998;66: Ming ZD, Wei W, Hong C, Wei Z, Xian DW. Balanced ultrafiltration, modified ultrafiltration, and balanced ultrafiltration with modified ultrafiltration in pediatric cardiopulmonary bypass. J Extra Corpor Technol 2001;33: Raja SG. Modified ultrafiltration for paediatric cardiac surgical patients: additional benefits and concerns. Chin Med J (Engl) 2004;117: Serwer GA, Armstrong BE, Anderson PAW. Noninvasive detection of retrograde descending aortic flow in infants using continuous wave Doppler ultrasonography: implications for diagnosis of aortic run-off lesions. J Pediatr 1980;97: Rodriguez RA, Cornel G, Weerasena N, Hosking MC, Murto K, Helou J. Aortic valve insufficiency and cerebral steal during pediatric cardiopulmonary bypass. J Thorac Cardiovasc Surg 1999;117: Lou HC, Lassen NA, Friis-Hansen B. Impaired autoregulation of cerebral blood flow in the distressed newborn infant. J Pediatr 1979;94: Ramamoorthy C, Lynn AM. The use of modified ultrafiltration during pediatric cardiovascular surgery is not a benefit. J Cardiothorac Vasc Anesth 1998;12: Edmonds HL Jr, Rodriguez RA, Audenaert SM, Austin EH III, Pollock SB Jr, Ganzel BL. The role of neuromonitoring in cardiovascular surgery. J Cardiothorac Vasc Anesth 1996;10: Trivedi UH, Patel RL, Turtle MRJ, Venn GE, Chambers DJ. Relative changes in cerebral blood flow during cardiac 27 CARDIOVASCULAR

7 28 RODRIGUEZ ET AL Ann Thorac Surg MUF AND CEREBRAL PERFUSION IN INFANTS 2005;80:22 8 operations using xenon-133 clearance versus transcranial Doppler sonography. Ann Thorac Surg 1997;63: Greisen G, Johansen K, Ellison PH, Fredriksen PS, Mali J, Friis-Hansen B. Cerebral blood flow in the newborn infant: comparison of Doppler ultrasound and 133 xenon clearance. J Pediatr 1984;104: Brown R, Wright G, Royston D. A comparison of two systems for assessing cerebral venous oxyhaemoglobin saturation during cardiopulmonary bypass in humans. Anaesthesia 1993;48: Kurth CD, Steven JM, Nicolson SC. Cerebral oxygenation during pediatric cardiac surgery using deep hypothermic circulatory arrest. Anesthesiology 1995;82: Watzman HM, Kurth CD, Montenegro LM, Rome J, Steven JM, Nicolson SC. Arterial and venous contributions to nearinfrared cerebral oximetry. Anesthesiology 2000;93: Bode H. Methods. In: Bode H, ed. Pediatric applications of transcranial Doppler sonography. New York, NY: Springer- Verlag, 1988: Perlman JM, Hill A, Volpe JJ. The effect of patent ductus arteriosus on flow velocity in the anterior cerebral arteries: ductal steal in the premature newborn infant. J Pediatr 1981;99: Serwer GA, Armstrong BE, Sterba RJ, Anderson PAW. Alterations in carotid arterial velocity-time profile produced by the Blalock-Taussig shunt. Circulation 1981;63: Mase M, Yamada K, Matsumoto T, Fujimoto S, Iida A. Cerebral blood flow and metabolism of steal syndrome evaluated by PET. Neurology 1999;52: Bar-Yosef S, Sanders EG, Grocott HP. Asymmetric cerebral near-infrared oximetric measurements during cardiac surgery. J Cardiothorac Vasc Anesth 2003;17: Hovels-Gurich HH, Seghaye MC, Schnitker R, et al. Longterm neurodevelopmental outcomes in school-aged children after neonatal arterial switch operation. J Thorac Cardiovasc Surg 2002;124: Myung RJ, Kirshbom PM, Petko M, et al. Modified ultrafiltration may not improve neurologic outcome following deep hypothermic circulatory arrest. Eur J Cardiothorac Surg 2003;24: Cho H, Nemoto EM, Yonas H, Balzer J, Sclabassi RJ. Cerebral monitoring by means of oximetry and somatosensory evoked potentials during carotid endarterectomy. J Neurosurg 1998;89: Rodriguez RA, Snider K, Cornel G, Teixeira OHP. Cerebral blood flow velocity during head-up tilt table testing for pediatric syncope. Pediatrics 1999;104: Samra SK, Dy EA, Welch K, Dorje P, Zelenock GB, Stanley JC. Evaluation of a cerebral oxymeter as a monitor of cerebral ischemia during carotid endarterectomy. Anesthesiology 2000;93: Schuurman PR, Albrecht KW. Intraoperative changes of transcranial Doppler velocity: relation to arterial oxygen content and whole-blood viscosity. Ultrasound Med Biol 1999;25: Todd MM, Wu B, Maktabi M, Hindman BJ, Warner DS. Cerebral blood flow and oxygen delivery during hypoxemia and hemodilution: role of arterial oxygen content. Am J Physiol 1994;267 (Heart Circ Physiol 36):H Liem KD, Hopman JCW, Oeseburg B, de Haan AFJ, Kollée LAA. The effect of blood transfusion and haemodilution on cerebral oxygenation and haemodynamics in newborn infants investigated by near infrared spectrophotometry. Eur J Pediatr 1997;156: INVITED COMMENTARY Cardiopulmonary bypass (CPB) in children is associated with the accumulation of water as a consequence of an inflammatory capillary leak. That increase in total body water is associated with tissue edema and subsequent organ dysfunction, particularly in the heart, lungs, and brain. Modified ultrafiltration (MUF) was designed to reduce these adverse effects of CPB. Previous studies have shown that MUF after CPB in children decreases body water, removes inflammatory mediators, and improves hemodynamics. Rodriguez and associates investigated the effects of different blood flow rates during arteriovenous modified ultrafiltration on the cerebral hemodynamics of children with weights of more than and less than 10 kg. Their study indicates that high blood flow rates through the ultra-filter during arteriovenous MUF in small infants ( 10 kg) decrease cerebral blood flow velocities and transcranial mixed venous oxygen saturations compared with the use of lower blood flow rates in older children. They indicated that a possible explanation for these effects is that the MUF circuit produces an increased diastolic runoff from the aorta that steals flow from the cerebral circulation. The authors should be congratulated for bringing this important new observation to the literature. This adverse effect may be one of the potential disadvantages of an arteriovenous MUF system. Other possible disadvantages of this system include: (1) the potential for air entrainment at the aortic cannulation site, (2) limitations of flow due to the aortic size, (3) significant arterial to venous shunts, and (4) difficulty in maintaining desired filling pressure, temperature, and blood oxygen saturation. Using a veno-venous MUF system, these potential disadvantages of arteriovenous MUF can be avoided. As the authors have correctly pointed out, further study is necessary to determine what constitutes an adequate flow based on the patient s body size. I would add that comparing the impact of veno-venous versus arteriovenous MUF on cerebral blood flow and cognitive function may also be important. Ko Bando, MD Department of Cardiovascular Surgery National Cardiovascular Center Fujishirodai Suita, Osaka, Japan kobando@hsp.ncvc.go.jp 2005 by The Society of Thoracic Surgeons /05/$30.00 Published by Elsevier Inc doi: /j.athoracsur