<%BANNER%>

Determinants of Cerebral Oxygenation and Response to Hypotension Treatment in Very Low Birth Weight Neonates

Permanent Link: http://ufdc.ufl.edu/UFE0042974/00001

Material Information

Title: Determinants of Cerebral Oxygenation and Response to Hypotension Treatment in Very Low Birth Weight Neonates
Physical Description: 1 online resource (40 p.)
Language: english
Creator: GARNER,RACHEL
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: HYPOTENSION -- NEONATE -- NIRS -- PREMATURE -- VLBW
Clinical Investigation (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The normal physiologic range for blood pressure to assure adequate cerebral blood flow and oxygen delivery is unknown in very preterm infants. Some studies suggest that under a mean arterial blood pressure of 30 mmHg, cerebral blood flow is pressure-passive, but above that measure, cerebral blood flow is maintained. Ideally, one would want to monitor cerebral blood flow and target therapy at cerebral blood flow improvement. Near infrared spectroscopy (NIRS) can be used to continuously estimate regional tissue oxygenation, the primary function of cerebral blood flow, thus by-passing the intermediate step of measuring the blood flow directly. We used NIRS to continuously monitor cerebral oxygenation in 50 very low birth weight (VLBW, less than 1500 grams) neonates during the first three days of life, a time at which they are most likely to receive treatment for hypotension. Using regression analysis, we created a model for cerebral oxygenation (rSO2) including several variables; however, our analysis found that blood pressure did not fit the model. Instead, our cerebral oxygenation model included: hematocrit (Hct), systemic oxygen saturations (SpO2) and the partial pressure of carbon dioxide (pCO2). After generating the rSO2 model for all 50 patients, we generated a model specifically for the hypotensive study patients. Significant determinants of rSO2 using only hypotensive subjects in the model were Hct, birth weight and race. Based on the results from these models, we found blood pressure to be a poor surrogate for cerebral oxygenation. Additionally, we evaluated the impact of two conventional treatments, normal saline and dopamine, for neonatal hypotension on blood pressure and cerebral oxygenation. We found that while normal saline and dopamine increased blood pressure, they did not change cerebral oxygenation. Our hypotension treatment data suggest that further work should explore different therapies, targeting rSO2 improvement rather than blood pressure improvement. Based on this study, we recommend that future experiments be conducted to understand the role of factors that we identified as important to rSO2. Considering that hematocrit was the one consistent variable of significance in both rSO2 models, it may be worthwhile exploring the effects of blood transfusions on cerebral oxygenation in neonates with low hematocrit and low rSO2. Furthermore, a randomized trial evaluating the effects of red blood cell transfusions along with other therapies on rSO2 is warranted. Finally, our study was limited to the neonates? stay in the neonatal intensive care unit. Additional investigation of long term neurodevelopmental outcomes is necessary in future studies.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by RACHEL GARNER.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Limacher, Marian C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-10-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042974:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042974/00001

Material Information

Title: Determinants of Cerebral Oxygenation and Response to Hypotension Treatment in Very Low Birth Weight Neonates
Physical Description: 1 online resource (40 p.)
Language: english
Creator: GARNER,RACHEL
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: HYPOTENSION -- NEONATE -- NIRS -- PREMATURE -- VLBW
Clinical Investigation (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The normal physiologic range for blood pressure to assure adequate cerebral blood flow and oxygen delivery is unknown in very preterm infants. Some studies suggest that under a mean arterial blood pressure of 30 mmHg, cerebral blood flow is pressure-passive, but above that measure, cerebral blood flow is maintained. Ideally, one would want to monitor cerebral blood flow and target therapy at cerebral blood flow improvement. Near infrared spectroscopy (NIRS) can be used to continuously estimate regional tissue oxygenation, the primary function of cerebral blood flow, thus by-passing the intermediate step of measuring the blood flow directly. We used NIRS to continuously monitor cerebral oxygenation in 50 very low birth weight (VLBW, less than 1500 grams) neonates during the first three days of life, a time at which they are most likely to receive treatment for hypotension. Using regression analysis, we created a model for cerebral oxygenation (rSO2) including several variables; however, our analysis found that blood pressure did not fit the model. Instead, our cerebral oxygenation model included: hematocrit (Hct), systemic oxygen saturations (SpO2) and the partial pressure of carbon dioxide (pCO2). After generating the rSO2 model for all 50 patients, we generated a model specifically for the hypotensive study patients. Significant determinants of rSO2 using only hypotensive subjects in the model were Hct, birth weight and race. Based on the results from these models, we found blood pressure to be a poor surrogate for cerebral oxygenation. Additionally, we evaluated the impact of two conventional treatments, normal saline and dopamine, for neonatal hypotension on blood pressure and cerebral oxygenation. We found that while normal saline and dopamine increased blood pressure, they did not change cerebral oxygenation. Our hypotension treatment data suggest that further work should explore different therapies, targeting rSO2 improvement rather than blood pressure improvement. Based on this study, we recommend that future experiments be conducted to understand the role of factors that we identified as important to rSO2. Considering that hematocrit was the one consistent variable of significance in both rSO2 models, it may be worthwhile exploring the effects of blood transfusions on cerebral oxygenation in neonates with low hematocrit and low rSO2. Furthermore, a randomized trial evaluating the effects of red blood cell transfusions along with other therapies on rSO2 is warranted. Finally, our study was limited to the neonates? stay in the neonatal intensive care unit. Additional investigation of long term neurodevelopmental outcomes is necessary in future studies.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by RACHEL GARNER.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Limacher, Marian C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-10-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042974:00001


This item has the following downloads:


Full Text

PAGE 1

1 DETERMINANTS OF CEREBRAL OXYGENATION AND RESPONSE TO HYPOTENSION TREATMENT IN VERY LOW BIRTH WEIGHT NEONATES By RACHEL SUE GARNER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PART IAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Rachel Sue Garner

PAGE 3

3 To m y f amily Mom, Dad and Jonnie for all their love and support and to Grandma Gardner, who I wish could have l ived to see my dreams fulfilled

PAGE 4

4 ACKNOWLEDGMENTS Professionally, I thank Dr. David J. Burchfield and the other members of my research oversight and thesis committees, Drs. Marian Limacher, Daniel Driscoll, and Charles Wood, for their mentorship during this project in this early part of my academic career I thank the American Heart Association and the Department of Pediatrics at the University of Florida for financial support. I thank the parents of the participants and the participants without whom this work would not have been possible. I thank Cindy Mi ller who assisted in enrolling eligible study participants. On a personal note, I thank my parents and brother for their encouragement throughout the years in support of my medical career.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 6 LIST OF FIGURES .......................................................................................................... 7 ABSTRACT ..................................................................................................................... 8 CHAPTER 1 INTRODUCTION .................................................................................................... 10 Clinical Relevance .................................................................................................. 10 Near Infrared Spectroscopy .................................................................................... 13 Summary ................................................................................................................ 15 2 MATERIALS AND METHODS ................................................................................ 16 Overall Study Design .............................................................................................. 16 Overview .......................................................................................................... 16 Patient Selection .............................................................................................. 16 Clinical Protocol ................................................................................................ 16 Statistical Considerations ........................................................................................ 17 3 RESULTS ............................................................................................................... 21 Patient Demographics ............................................................................................. 21 rSO2 Model ............................................................................................................. 21 rSO2 Model in Hypotensive Patients ...................................................................... 21 Hypotension Treatment ........................................................................................... 22 Discharge and Death Outcomes ............................................................................. 22 4 DISCUSSION ......................................................................................................... 32 Summary and Significance of Results .................................................................... 32 Study Limitations .................................................................................................... 35 Future Directions .................................................................................................... 36 REFERENCES .............................................................................................................. 37 BIOGRAPHICAL SKETCH ............................................................................................ 40

PAGE 6

6 LIST OF TABLES Table page 2 1 Sample size and power calculation. ................................................................... 20 3 1 Descriptive statistics of variables in rSO2 model. ............................................... 23 3 2 Preliminary regression output for rSO2 model. ................................................... 23 3 3 Final regression output for rSO2 model. ............................................................. 23 3 4 Complete patient data for all variables used in rSO2 regression model. ............ 24 3 5 Descriptive statistics of variables in rSO2 model of hypotensive patients. ......... 25 3 6 Initial regression output for rSO2 model in hypotensive patients. ....................... 25 3 7 Final regression output for rSO2 model in hypotensive patients. ........................ 26 3 8 Complete patient data for all variables used in rSO2 regression model of hypotensive patients. .......................................................................................... 26 3 9 Correlation between time to discharge and variables. ........................................ 27 3 10 Correlation between corrected gestational age at discharge and variables. ....... 27 3 11 Cause of death for study subjects. ..................................................................... 27 3 12 Correlation between variables and dea th. .......................................................... 27

PAGE 7

7 LIST OF FIGURES Figure page 3 1 Changes in blood pressure after normal saline bolus and dopamine infusion. ... 28 3 2 Changes in cerebral oxygenation with normal saline bolus and dopamine infusion. .............................................................................................................. 29 3 3 Systemic oxygen saturation and corrected gestational age at discharge correlation. .......................................................................................................... 30 3 4 Birth weight and corrected gestation al age at discharge correlation ................... 31

PAGE 8

8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree Master of Science DETERMINANTS OF CEREBRAL OXYGENATION AND RESPONSE TO HYPOTENSION TREATMENT IN VERY LOW BIRTH WEIGHT N EONATES By Rachel Sue Garner May 2011 Chair: Marian C. Limacher Major: Medical Sciences Clinical and Translational Science The normal physiologic range for blood pressure to assure adequate cerebral blood flow and oxygen delivery is unknown in very preterm infants. Some studies suggest that under a mean arterial blood pressure of 30 mmHg, cerebral blood flow is pressurepassi ve, but above that measure, cerebral blood flow is maintained. Ideally, one would want to monitor cerebral blood flow and target therapy at cerebral blood flow improvement. Near infrared spectroscopy (NIRS) can be used to continuously estimate regional t issue oxygenation, the primary function of cerebral blood flow, thus by passing the intermediate step of measuring the blood flow directly. We used NIRS to continuously monitor cerebral oxygenation in 50 very low birth weight (VLBW, less than 1500 grams) neonates during the first three days of life, a time at which they are most likely to receive treatment for hypotension. Using regression analysis, we created a model for cerebral oxygenation (rSO2) including several variables; however, our analysis found that blood pressure did not fit the model. Instead, our cerebral oxygenation model included: hematocrit (Hct), systemic oxygen saturations (SpO2) and the partial pressure of carbon dioxide (pCO2).

PAGE 9

9 After generating the rSO2 model for all 50 patients, w e generated a model specifically for the hypotensive study patients. Significant determinants of rSO2 using only hypotensive subjects in the model were Hct, birth weight and race. Based on the results from these models, we found blood pressure to be a poor surrogate for cerebral oxygenation. Additionally, we evaluated the impact of two conventional treatments, normal saline and dopamine, for neonatal hypotension on blood pressure and cerebral oxygenation. We found that while normal saline and dopamine increased blood pressure, they did not change cerebral oxygenation. Our hypotension treatment data suggest that further work should explore different therapies, targeting rSO2 improvement rather than blood pressure improvement. Based on this study, we recommend that future experiments be conducted to understand the role of factors that we identified as important to rSO2. Considering that hematocrit was the one consistent variable of significance in both rSO2 models, it may be worthwhile exploring the effects of blood transfusions on cerebral oxygenation in neonates with low hematocrit and low rSO2. Furthermore, a randomized trial evaluating the effects of red blood cell transfusions along with other therapies on rSO2 is warranted. Finally, our study was limited to the neonates stay in the neonatal intensive care unit. Additional investigation of long term neurodevelopmental outcomes is necessary in future studies.

PAGE 10

10 CHAPTER 1 INTRODUCTION Clinical Relevance Although neonatology has flourished as a ped iatric specialty since the 1960s, the fundamental question of what is an adequate blood pressure in very low birth weight neonates has not been answered. Several investigators have described the normal distribution of blood pressures over the f irst several days of life in low birth weight infants .1 5 However, delivery of adequate oxygen many be more important than having blood pressure within a stat istical distribution of normal, which may not be adequate in the sick, very low birth weight (VLBW, <1500 grams) baby. The circulatory system maintains oxygen delivery to the organs of the body. Oxygen delivery, in turn, depends upon the oxygen carrying ca pacity of the blood, the oxygen content in the blood, and the volume of blood delivered to the tissues.6 Oxygen content is easy to measure but measurement of blood flow in VLBW neonates is much more challenging, so clinicians often use blood pressure as a surrogate. For many years, neonatal circulatory support has been based on an assumed proportionality between blood pressure and systemic blood flow, particularly within the cerebral circulation. However, having a blood pressure within a statistical distribution of normal, especially in the sick, very low birth weight (VLBW, <1500 grams) baby, may not be adequate to assure oxygen delivery to the tissues. Limited evidence in human neonates suggests that the neonatal cerebral circulation i s not pressurepassive, but rather autoregulated.7 8 Autoregulation implies that within a certain range of blood pressures, cerebral circulation is relatively c onstant,

PAGE 11

11 so that cerebral blood flow rises or falls only when the blood pressure is outside that range. Moreover, some evidence also exists that during illness, autoregulation fails9 10 and this failure has been associated with neur ological injuries in VLBW neonates .11 The normal physiologic range for blood pressure to assure adequate cerebral blood flow and oxygen delivery is unknown in very preterm infants. Despite this, up to 40% of VLBW neonates are treated for hypotension in the first 3 days of life.12 Definitions and manageme nt of hypotension among VLBW neonates vary across neonatal intensive care units around the country. The lack of standardization may be due to a combination of reasons including the fact that blood pressure tends to increase hourly over the first day of lif e.1 5 The clinical relevance of hypotension in the VLBW preterm neonate is con troversial. Concern for hypotension is raised because it has been recognized as a risk for poor neurodevelopmental outcomes .13 15 Several studies document central nervous system morbidities associated with low blood pressure.1, 7 One particular study published in 1987 found that in preterm neonates, a mean blood pressure remaining below 30 mmHg for an hour was associated with severe hemorrhage, ischemic cerebral lesions, or death within 48 hours, while no severe lesions developed in patients with a mean blood pressure .16 Similarly, Munro et al .7, using near infrared spectroscopy ( NIRS ) to measure cerebral blood flow in 17 extremely low birth weight infants, showed a breakpoint at approximately 30 mmHg in the cerebral blood flow mean arterial pressure autoregulation curve. These s tudies would suggest that under a mean blood pressure of 30 mmHg, cerebral blood flow is pressurepassive, but above that measure, cerebral blood flow is maintained. Pressurepassivity occurs when the

PAGE 12

12 vascular bed is maximally dilated and blood flow decreases passively in response to further reductions in blood pressure. Additional implications from these studies1,7,16 suggest that maintaining mean arterial blood pressure above 30 mmHg would be important for optimal neurodevelopmental outcome in VLBW infants, and this level has been incorporated as a practice standard in a major neonatal textbook.17 Contrasting this notion that blood pressure below a certain level leads to CNS morbidities, other s have suggest ed that blood pressure alone is unreliabl e in predicting cerebral perfusion .18,19 Furthermore, other reports show no relationship between blood pressure and short term neurological injury .8 Currently, treatment of hypotension in V LBW neonates includes volume expansion, inotropic agents (dopamine and dobutamine) and steroids, using the patients mean arterial pressure as a marker for improvement. However, the goal of treating hypotension should be restoration of organ perfusion, par ticularly cerebral perfusion, and in that light, effects of various treatments differ. Using 133Xe clearance to measure cerebral blood flow, Lundstr o m et al .20 found that although dopamine led to a higher blood pressure response com pared to albumin infusion, albumin led to higher cerebral blood flow. Likewise, Osborn et al .21 demonstrated that treatment of hypotensive neonates with dopamine led to an increase in systemic blood pressure without an increase in cerebral blood flow, as estimated by superior vena cava flow Yet, dobutamine improved cerebral blood flow by approximately 25% with no appreciable change in blood pressure.20 T reatments for hypotension clearly need to be studied more rigorously, with improvement in cerebral perfusion and cerebral oxygenation as new endpoints.

PAGE 13

13 Near Infrared Spectroscopy Non invasive methods for determining cerebral blood flow have been developed but are not clinically practical Such techniques such as superior vena cava blood flow2 2 and Doppler23 are cumbersome, technically challenging and cannot be performed continuously. Likewise, functional echocardiography has been used to measure cardiac output, but its use is limited by the requirement of a trained operator and an inability to record information continuously. In contrast, NIRS utili zes the near infrared region of the electromagnetic spectrum and can be used to continuously estimate regional tissue oxygenation2426, the primary function of cerebral blood flow, thus by passing the need to invasively measure the blood flow directly. When using NIRS, light photons are aimed into the skin over the forehead. After being scattered about inside the scalp, skull, and brain, some fraction of the entering photons return and exit the skin ("reflectance"). By measuring the quantity of returning photons as a function of wavelength, one can infer the spectral absorption of the underlying tissue. Human tissues (skin, fat, bone) are translucent to NIR photons of wavelengths between 650 and 1100 nm. How ever, at 730 nm and 810 nm wavelengths, red colored hemoglobin molecules within red blood cells have the highest light absorption. The oxygenated state of the hemoglobin will absorb more light at 810 nm than hemoglobin does at 730 nm, and using this princi ple, one can measure oxygenated to total hemoglobin ratios, or the percentage of hemoglobin that is carrying oxygen. Since tissue oxygen delivery is the ultimate goal of circulatory support, measurement of tissue oxygenation should be useful in assessment of adequacy of blood pressure and obviate the need for measuring flow. This principle was proven in a study using newborn lambs which found a correlation between cerebral tissue

PAGE 14

14 oxygenation and changes in cerebral blood flow .27 Somanetics (Troy, MI) has developed the INVOS Cerebral Oximeter, a NIRS tool which measures cerebral oxygen saturation, thus providing a noninvasive means of continuously monitoring cerebral oxygenation and ultimately, cerebral perfusion. Data regarding the impact of low cerebral oxygenation on brain injury has been limited to date, with most studies performed in animal models or adults. In animals, histological evidence for cerebral injury occurs with cerebral saturations of <40% for 30 minutes .28 Similar findings have been reported in adult patients. Edmonds et al .29 reported that cerebral saturations below 40% and declines of more than 25% from baseline are associated with neurologic dysfunction and other adverse outcomes and that declines in cerebral rSO2 below 50%, or more than 20% from baseline, have shown cause for concern and intervention. Currently, minimal data link low regional cerebral oxygenation in premature, VLBW infants to brain injury. A recent case control study in preterm neonates found that in the first two weeks of life, those with germinal matrix hemorrhages or intraventricular hemorrhages had lower cerebral oxygen saturations than those who did not develop a hemorrhage.30 C ontemporary monitoring in neonatal intensive care units utilizes continuous blood pressure and noninvasive arterial pulse oximetry, not the more costly NIRS technology, and infers adequacy of tissue perfusion and oxygenation based on blood pressure and arterial oxygenation alone. Therefore, it would be clinically important to determine if the use of blood pressure and arterial oxygen saturation can predict adequacy of cerebral oxygenation.

PAGE 15

15 Summary By continuously monitoring very preterm neonates during the first 3 days of life, a time at which they are likely to receive hypotension treatment12, we can determine whether there is a relationship between systemic blood pressure, systemic oxygen saturation and cerebral oxygenation (using NIRS). We can also determine if there is a critical blood pressure required to maintain adequate cerebral perfusi on. Finally, we can analyze the data surrounding any treatment periods for hypotension to determine both blood pressures and cerebral saturations response to that particular treatment.

PAGE 16

16 CHAPTER 2 MATERIALS AND METHOD S Overall Study Design Overview We performed an observational study of blood pressure and cerebral oxygenation in fifty very low birth weight (VLBW) neonates. This study was approved by the University of Florida Institutional Review Board. Enrollment began in November 2008 and concluded in April 2010 after accrual of the fiftieth study subject. Informed consent was obtained from a parent prior to enrollment. Patient S election Infants less than 30 weeks gestational age ( GA ) and less than 1500 grams requiring arterial access were enrolled i n the study. Infants were excluded from the study if they had cyanotic congenital heart disease, limited viability or major congenital malformations. Clinical P rotocol As part of routine care, systemic oxygen saturation (SpO2) and arterial blood pressur e (blood pressure) were monitored using either a Agilent (Santa Clara, CA) monitor or a Philips MP30 (Amsterdam, The Netherlands) monitor. As part of the study, cerebral oxygenation (rSO2) was monitored using Near Infrared Spectroscopy (NIRS) using a Soma netics (Troy, MI) INVOS Cerebral Oximeter. A neonatal NIRS sensor was placed centrally across the neonates forehead. Both the vital sign monitor and the INVOS were attached to a Vital SyncTM (Somanetics, Troy, MI) computer which received transmitted d ata points every 3060 seconds. Data collection began after

PAGE 17

17 arterial access was established and continued for 72 hours. The data points were downloaded into a spreadsheet at the completion of the 72 hour monitoring period. In addition to vital signs, d emographic information including gestational age, birth weight, race, and gender were recorded. Laboratory values including hematocrit (Hct) and partial pressure of carbon dioxide in arterial blood samples (pCO2) were also recorded. During the study perio d, if the primary clinical care team determined that a study patient required treatment for hypotension, they determined the method, all dosages and duration of treatment. While our NICU has no standard protocol in place which dictates hypotension treatment in VLBW neonates, the clinical staff typically employs an approach that begins with volume expansion with a 1020 ml/kg normal saline bolus I f the neonate remains hypotensive, they initiate a dopamine infusion at 5 mcg/kg/min, which is then titrated u ntil either the desired blood pressure is achieved or it is maximized at 20 mcg/kg/min. If the neonate remains hypotensive once he/she has received the maximal dopamine infusion, the clinical staff typically begins supraphysiologic steroid replacement wi th hydrocortisone at 20 to 30 mg/m2 per day IV in 2 or 3 doses. Statistical Considerations To generate a model for determinants of cerebral oxygenation, we performed a regression analysis of GA, birth weight ( BW ) race, gender, Hct pCO2, mean arterial blood pressure (MAP), and systemic arterial oxygen saturation (SpO2). The pCO2 was obtained from the study subjects arterial blood gas and hematocrit from the subjects complete blood cell count (CBC), obtained as part of routine care in the NICU. The other variables in the model were obtained from data collected at the same time. All

PAGE 18

18 independent variables were included in the original model and the least significant excluded until only significant variables at p the model. The regression analysis was performed on all 50 study subjects. An additional analysis was performed on 20 subjects who were hypotensive (defined as a MAP less than 30 mmHg) at the time of pCO2 attainment. To determine the blood pressure an d cerebral oxygenation changes in response to treatment of hypotension, in the 20 patients treated for hypotension, we determined the average cerebral oxygenation and mean arterial blood pressure (MAP) during the 30 minutes prior to treatment and during t he 30 minutes following completion of normal saline (NS) boluses. When patients were treated with dopamine, we determined their average cerebral oxygenation and MAP after 30 minutes of dopamine infusion. We used a paired t test to assess changes in cerebral oxygenation and MAP after treatment. To determine if there were associations between the variables analyzed in the cerebral oxygenation model and time to discharge, corrected gestational age at discharge, and death, we performed a bivariate correlation analysis to generate a Pearsons correlation coefficient. SPSS version 18.0 (IBM, Somers, NY) was used for all statistical analysis. We considered a p value < 0.05 as statistically significant. Assuming a standard deviation of 10 or 20% of the measur e, we anticipated detecting a 20% change in blood pressure, assuming alpha of 0.05 and 2tailed design (Table 21). Given that hypotensive patients all had starting blood pressures <30 mmHg, this narrowed the standard deviation of measurement considerably from nonhypotensive patients. In order to achieve a power of 0.8, we would require 1520

PAGE 19

19 hypotensive patients. Previous studies have demonstrated that up to 40% of VLBW neonates are treated for hypotension12. Thus, we required an overall sample size of 50 patients in order to accrue 1520 hypotensive patients.

PAGE 20

20 Table 21. Sample size and power calculation N Percent c hange BP Std Dev 15 0.05 20% 3 1.0 15 0.05 20% 6 0.76 20 0.05 20% 3 1.0 20 0.05 20% 6 0.89

PAGE 21

21 CHAPTER 3 RESULTS Patient Demographics The average GA of patients enrolled was 26.5 weeks with a standard deviation of 1.8 weeks. Enrolled patients had an average BW of 929 grams with a standard deviation of 229 grams. Racial breakdown of study subjects included 46% African American, 38% Cau casian and 16% other (including Hispanic, Asian and multiracial). Among the study subjects, 54% were female and 46% were male. rSO2 Model We created a model for rSO2 incorporating variables which potentially impact cerebral oxygenation and cerebral perf usion ( Table 31 ). The initial regression model we generated (Table 3 2 ), accounting for all variables, indicated that several variables were not statistically significant. After eliminating the least significant variables in a step wise fashion until only statistically significant variables remained, we generated a model that included SpO2, pCO2 and Hct (Table 3 3 ). Our final regression model after evaluating all variables (Table 34), 93.516 + 1.208 SpO2 + 0.836 Hct + 0.359 pCO2, yielded an R2 of 0.296 for goodness of fit. rSO2 Model in Hypotensive Patients We additionally created a model for rSO2 during times of hypotension (see Table 3 5 for descriptive statistics of variables). The initial regression model we generated (Table 3 6 ), accounting for all variables, indicated that several variables were not statistically significant. After eliminating the least significant variable in a stepwise fashion until only statistically significant variables remained, we generated a model that i n cluded Hct, BW and Race (Table 37 ). Our final regression model after evaluating all

PAGE 22

22 variables (Table 38), 9.74 + 1.085 Hct + 0.033 BW 6.446 Race yielded an R2 of 0.562 for goodness of fit. Hypotension Treatment We evaluated the effects of normal saline boluses and dopamine infusions on the blood pressure and cerebral oxygenation in very low birth weight neonates. While NS and dopamine increased study subjects blood pressures (Figure 31), they did not increase cerebral oxygenation (Figure 32). Discharge and Death Outcomes The average time to discharge was 86 days, with an average corrected gestational age at discharge of 39 weeks with a standard deviation of 3 weeks. Discharge data w ere unavailable for two patients who were transferred to community hospitals for continued care. Several variables were associated with longer time to discharge including: lower birth weight, lower gestational age, lower rSO2, lower MAP and lower SpO2 (Table 3 9 ). The only two variables associated with an older corrected gestational age at discharge included lower birt h weight and lower SpO2 (Table 310 Figures 33 and 34). There were eight deaths among the study subjects which were due to respiratory, infectious and gastrointesti nal complications (see Table 311 for etiologies). Upon analysis, the only variables associated with death included lower birth weight and lower MAP (Table 312 ).

PAGE 23

23 Table 31. Descriptive statistics of variables in rSO2 model Mean Std. d eviation N rSO2 (%) 73.02000 11.431232 50 MAP (mmHg) 31.70000 6.178468 50 SpO2 (%) 95.40000 4.035556 50 Hct (%) 43.01400 5.526597 50 pCO2 (mmHg) 42.76000 10.860075 50 GA (weeks) 26.52000 1.787028 50 BW (grams) 929.08000 228.917046 50 Table 32. Preliminary regression output for rSO2 model Coefficients a Unstandardized coefficients Standardized coefficients t Sig. B Std. error Beta (Constant) 106.091 52.211 2.032 .049 MAP .432 .318 .233 1.356 .182 SpO2 1.056 .475 .373 2.223 .032 Hct .895 .279 .433 3.211 .003 pCO2 .260 .173 .247 1.503 .140 GA 1.724 1.240 .269 1.390 .172 BW .001 .009 .015 .087 .931 Gender .756 3.013 .033 .251 .803 Race 1.609 2.254 .100 .714 .479 a Dependent variable: rSO2. Table 33. Final regression output for rSO2 model Coefficients a Unstandardized coefficients Standardized coefficients t Sig. B Std. error Beta (Constant) 93.516 45.194 2.069 .044 SpO2 1.208 .413 .426 2.924 .005 Hct .836 .257 .404 3.253 .002 pCO2 .359 .153 .341 2.343 .024 a Dependent variable: rSO2

PAGE 24

24 Table 34. Complete patient data for all variables used in rSO2 regression model Subject MAP (mmHg) SpO2 (%) rSO2 (%) Hct (%) pCO2 (mmHg) GA (weeks) BW (grams) Gender Race 1 30 94 72 41 45 27 940 1 2 2 45 97 66 37.5 30 28 1180 0 1 3 29 100 76 39.7 23 29 1166 0 3 4 25 94 46 42.3 40 23 597 0 2 5 43 94 78 58.6 45 28 762 0 1 6 25 94 90 43.8 42 25 705 0 1 7 27 84 78 50.8 59 28 1175 1 3 8 41 100 57 38.4 20 28 970 0 2 9 29 98 76 50.1 37 24 810 1 3 10 32 93 55 41.9 41 28 1040 0 1 11 32 100 72 43.6 37 28 975 1 3 12 38 96 80 50.2 40 28 975 0 1 13 28 100 81 39.4 46 27 1074 0 1 14 37 92 68 44.7 78 28 1102 1 1 15 40 95 64 29.4 54 28 1045 1 1 16 35 95 80 49.2 46 29 786 0 1 17 34 100 83 42.8 38 27 875 0 2 18 31 100 74 48.5 39 28 1333 0 2 19 30 99 62 42.7 44 26 1045 0 2 20 32 91 79 46.3 51 27 982 0 1 21 23 90 55 33.6 33 23 607 0 2 22 28 89 75 38.7 62 28 1207 0 3 23 27 89 71 44 57 27 761 1 3 24 27 98 81 41.5 40 25 811 1 2 25 22 95 72 44.1 41 25 840 1 2 26 35 93 73 45.1 50 25 762 0 2 27 31 100 88 41.9 38 26 815 1 1 28 31 98 95 44.7 42 27 949 0 2 29 26 100 91 34.6 60 28 1441 1 1 30 16 93 53 38.3 51 26 865 1 2 31 28 94 51 37.2 41 25 690 0 2 32 32 95 77 44.6 40 24 595 0 2 33 32 100 94 42.8 32 29 749 1 3 34 30 93 66 37 57 25 691 0 2 35 29 95 59 30.6 43 24 855 1 2 36 43 100 61 41.1 23 29 1437 1 2 37 40 96 77 41.9 36 28 719 1 2 38 42 100 94 50.3 53 29 1270 0 1 39 37 94 68 50.9 35 25 700 1 2 40 33 97 73 42.7 34 26 880 1 1

PAGE 25

25 Table 34. Continued. Subject MAP (mmHg) SpO2 (%) rSO2 (%) Hct (%) pCO2 (mmHg) GA (weeks) BW (grams) Gender Race 41 24 92 77 41.3 42 26 962 0 1 42 33 95 88 47.9 51 28 1313 1 3 43 39 100 82 48.1 37 25 1050 1 1 44 31 93 72 51.1 41 26 1109 1 2 45 35 96 68 44.2 40 26 1082 0 2 46 38 98 63 44.1 32 28 642 0 2 47 23 93 70 38.4 40 23 590 0 2 48 28 83 79 46.9 53 24 704 0 1 49 25 96 68 44 53 24 630 1 1 50 34 99 73 38.2 26 28 1191 1 1 Gender variable: females coded as 0 and males coded as 1. Race variable: Caucasians coded as 1, African Americans as 2 and Other as 3. Table 35. Descriptive statistics of variables in rSO2 model of hypotensive patients Mean Std. Deviation N rSO2 (%) 70.75000 12.481038 20 MAP (mmHg) 25.80000 3.138890 20 SpO2 (%) 93.45000 4.795557 20 Hct (%) 41.22000 5.175306 20 pCO2 (mmHg) 45.45000 9.832679 20 GA (weeks) 25.45000 1.848897 20 BW (grams) 862.60000 237.075471 20 Table 36. Initial regression output for rSO2 model in hypotensive patients Coefficients a Unstandardized coefficients Standardized coefficients t Sig. B Std. error Beta (Constant) 21.339 95.312 .224 .827 Hct 1.000 .522 .415 1.915 .082 BW .024 .023 .460 1.036 .322 Race 8.158 4.069 .496 2.005 .070 MAP .900 .838 .226 1.074 .306 SpO2 .025 .758 .009 .032 .975 pCO2 .164 .387 .130 .425 .679 GA 1.225 2.957 .182 .414 .687 Gender 2.746 6.255 .112 .439 .669 a Dependent variable: rSO2.

PAGE 26

26 Table 37. Final regression output for rSO2 model in hypotensive patients Coefficients a Unstandardized coefficients Standardized coefficients t Sig. B Std. error Beta (Constant) 9.740 19.040 .512 .616 Hct 1.085 .411 .450 2.640 .018 BW .033 .009 .636 3.768 .002 Race 6.446 2.811 .392 2.293 .036 a Dependent variable: rSO2 Table 38. Complete patient data for all variables used in rSO2 regression model of hypotensive patients Subject MAP (mmHg) SpO2 (%) rSO2 (%) Hct (%) pCO2 (mmHg) GA (weeks) BW (grams) Gender Race 3 29 100 76 39.7 23 29 1166 0 3 4 25 94 46 42.3 40 23 597 0 2 6 25 94 90 43.8 42 25 705 0 1 7 27 84 78 50.8 59 28 1175 1 3 9 29 98 76 50.1 37 24 810 1 3 13 28 100 81 39.4 46 27 1074 0 1 21 23 90 55 33.6 33 23 607 0 2 22 28 89 75 38.7 62 28 1207 0 3 23 27 89 71 44 57 27 761 1 3 24 27 98 81 41.5 40 25 811 1 2 25 22 95 72 44.1 41 25 840 1 2 26 27 92 66 45.1 46 25 762 0 2 29 26 100 91 34.6 60 28 1441 1 1 30 16 93 53 38.3 51 26 865 1 2 31 28 94 51 37.2 41 25 690 0 2 35 29 95 59 30.6 43 24 855 1 2 41 24 92 77 41.3 42 26 962 0 1 47 23 93 70 38.4 40 23 590 0 2 48 28 83 79 46.9 53 24 704 0 1 49 25 96 68 44 53 24 630 1 1 Gender variable: females coded as 0 and males coded as 1. Race variable: Caucasians coded as 1, African Americans as 2 and Other as 3.

PAGE 27

27 Table 3 9 Correlation between time to discharge and variables Variable Pearson correlation Significance BW 0.708 0.01 GA 0.617 0.01 rSO2 0.324 0.041 MAP 0.437 0.005 SpO2 0.36 0.023 Hct 0.179 0.269 pCO2 0.012 0.94 Table 310. Correlation between corrected gestational age at discharge and variables Variable Pearson correlation Significance BW 0.518 0.001 GA 0.218 0.177 rSO2 0.201 0.213 MAP 0.269 0.094 SpO2 0.315 0.048 Hct 0.132 0.418 pCO2 0.07 0.67 Table 311. Cause of death for study subjects Age at death (days) Cause of death 11 Necrotizing enterocolitis 33 Necrotizing enterocolitis 36 Necrotizing enterocolitis 2 Peptostreptococcus sepsis 11 Fungal sepsis 29 Pseudomonas sepsis 1 Hypoxic respiratory failure 134 Cor pulmonale Table 3 12. Correlation between variables and death Variable Pearson correlation Significance BW 0.282 0.047 GA 0.221 0.123 rSO2 0.141 0.33 MAP 0.335 0.017 SpO2 0.057 0.692 Hct 0.082 0.572 pCO2 0.066 0.651

PAGE 28

28 Figure 31. Changes in blood pressure after normal saline bolus and dopamine infusion. Box plots of the change in MAP with normal saline bolus and dopamine infusion. The horizontal line indicates median level of MAP, the box indicates 25th and 75th percentiles and the bars denote minimum and maximum values.

PAGE 29

29 Figure 32. Changes in cerebral oxygenation with normal saline bolus and dopamine infusion. Box plots of the change in rSO2 with normal saline bolus and dopamine infusion. The horizontal line indicates median level of MAP, the box indicates 25th and 75th percentiles and the bars denote minimum and maximum values. Neither treatment resulted in a significant change in rSO2.

PAGE 30

30 Figure 33. Systemic oxygen s aturation and corrected gestational age at discharge correlation. Scatter plot of corrected gestational age at discharge and systemic oxygen saturation. Best fit line included (R2 linear of 0.099).

PAGE 31

31 Figure 34. Birth weight and corrected gestational age at discharge correlation. Scatter plot of corrected gestational age at discharge and birth weight. Best fit line included (R2 linear of 0.268)

PAGE 32

32 CHAPTER 4 DISCUSSION Summary and Significance of Results Our study confirmed that blood pressure is a poor surrogate for cerebral perfusion in VLBW neonates. In neither of our cerebral oxygenation models did blood pressure play a significant determining role. Additionally, while treatment of hypotension improved blood pressure, it did not result in an improvement in cerebral oxygenation. Overall, these findings suggest that treatment should aim at improving cerebral oxygenation rather than hypotension. Our regression models explored the impact of several variables on cerebral oxygenation in VLBW neonates. The cerebral oxygenation model inclusive of all 50 study subjects found only SpO2, Hct and pCO2 to be significant variables impacting cerebral oxygenation. From a physiologic standpoint, these variables should play an important role in cerebral oxygenation and perfusion as hematocrit determines oxygen carrying capacity, systemic oxygen saturations should impact regional oxygen saturations and pCO2 influences cerebral vascular resistance. While the variables with in this model were statistically significant, the R2 value was low at only 0.296. Thus, approximately 70% of cerebral oxygenation is determined by factors unaccounted for by the model. In contrast, the cerebral oxygenation model for hypotensive patients included Hct, BW and Race as significant variables. Interestingly, hematocrit was the only consistent variable between the two models. In contrast to the model inclusive of all study subjects, the hypotensive patient model elicited race as a variable important for cerebral oxygenation, with African American and Other subjects exhibiting a lower

PAGE 33

33 cerebral oxygenation than Caucasian subjects. Pigmentation cannot account for the racial differences in cerebral oxygenation as NIR photons penetrate through skin. Thus, some unknown racial difference other than skin pigmentation accounts for varying cerebral oxygenation in Caucasians and African Americans. The R2 value for the hypotensive study subject model was 0.562, meaning that factors unaccounted for in the model account for approximately 46% of cerebral oxygenation. While this model was a better fit than the model inclusive of all study subjects, its goodness of fit was far from ideal. The hypotension treatment data revealed that while two of our conventional therapies for neonatal hypotension increase blood pressure, they do not impact cerebral oxygenation. When used as treatment for hypotension, both normal saline and dopamine resulted in statistically significant changes in blood pressure (pvalues of 0.0041 and 0.0163 respectively). Yet, neither normal saline nor dopamine resulted in any change in cerebral oxygenation (pvalues of 0.3898 and 0.8148 respectively). When physicians treat hypotension, they aim to improve end organ perfusion, particu larly cerebral perfusion. Based on our data, it appears that normal saline and dopamine improve the surrogate measure (blood pressure), but not cerebral oxygenation. Furthermore, while neonatologists have considered blood pressure an appropriate surrogat e, using the NIRS technique we found that blood pressure is not an adequate measure of cerebral perfusion. Our study also evaluated the impact of several variables we considered in our cerebral oxygenation model on discharge and surv ival outcomes. While we found several variables to be associated with longer time to discharge, we decided that time to

PAGE 34

34 discharge is not an ideal outcome measure in this population. Neonates within the study were born at varying gestational ages. As a g eneral rule, most preterm neonates reach discharge readiness close to their mothers due date. In order to be discharge ready a preterm neonate must have reached a sufficient weight to maintain an adequate body temperature outside the isolette, must be able to feed exclusively by mouth and gain weight and must outgrow apnea and bradycardia. Thus, we decided that corrected GA at discharge might provide a more standardized discharge measure for our study subjects of varying gestational ages. We found only two variables associated with an older corrected gestational age at discharge: lower birth weight and lower SpO2. Some neonates experience intrauterine growth restriction (IUGR) which results in a birth weight much lower than their counterparts who wer e a size appropriate for gestational age (AGA). IUGR neonates must reach a similar discharge weight as their AGA counterparts, which usually results in a greater corrected GA at discharge. The correlation between lower SpO2 and increased corrected GA at discharge, does not present a straightforward explanation. Perhaps the neonates with lower SpO2 at birth had worse respiratory distress syndrome and thus went on to develop chronic lung disease, increasing their corrected age at discharge. We also exami ned the causes of death in the eight study subjects who died. The causes of death were common complications of prematurity and not unexpected. Lower birth weight and lower MAP were the only two variables significantly correlated with death. Neonates who are born at a lower birth weight have a lower chance of survival than those who are born at higher birth weight .31 Three of the eight subjects died from necrotizing enterocolitis (NEC). While the etiology of NEC remains a mystery in our

PAGE 35

35 field, some have hypothesized that mesenter ic hypoperfusion is a major pathogenic factor .32, 33 Perhaps, these neonates suffered an early hypoperfusion insult which predisposed them to later development of NEC. Thus, the correlation between low MAP and death is not necessarily one of direct causation, and most likely multifactorial. Study Limitations One of the main limitations of our study was in its observational nature. The blood pressure treatment data was most affected by the observational design. Neonates were not randomized to treatment. Prior to receiving dopamine, neonates received a normal saline bolus. Had neonates been randomized to dopamine as a first line treatment of hypotension, results may have differed. Additionally, the nonrandomized nature of our study limited neonates to the approach to hypotension treatment taken by clinicians in our unit. Other units employ a different treatment approach, using other vasopressors such as dobutamine and epinephrine and corticosteroids. The effects of these hypotension treatments were not investigated in our study as they were not used routinely by our treating physicians. Our data points in the rSO2 analysis were taken from the time at which the neonate had a n arterial blood gas including a pCO2. The treating team determined the timing of blood gas acquisition. While these blood gases were obtained within the first 24 hours of life, the exact timing was not consistent among the neonates. Ideally we would like to have acquired all data points for comparison at the same postnatal hour. However, the time to establish arterial access is impossible to standardize and thus some neonates had their first blood gas obtained earlier than others.

PAGE 36

36 Future Directions Our study yielded interesting results that should direct further study. Clearly, the two conventional treatments for hypotension we explored did not improve cerebral oxygenation. However, blood pressure was not a significant variable for rSO2 in our models. We suggest that blood pressure should no longer be used as a surrogate for cerebral perfusion. Rather, neonatologists can use rSO2 as a continuous, bedside marker of cerebral perfusion. Additionally, further exploration is warranted into treatments that can improve cerebral oxygenation and cerebral perfusion. Within neonatology our mindset has focused on blood pressure as a surrogate for cerebral perfusion. If we shift our mindset to cerebral oxygenation as a marker of cerebral perfusion we should investigate treatment of low cerebral oxygenation rather than hypotension. As hematocrit was the one consistent variable of sig nificance in both rSO2 models, it may be worthwhile exploring the effects of blood transfusions on cerebral perfusion in neonates with low hematocrit and low rSO2. Additionally, past evidence has suggested that dobutamine21 and al bumin20 may improve cerebral perfusion. A randomized trial evaluating the effects of dobutamine, albumin and red blood cell transfusions on rSO2 would be a valuable next step Another avenue warranting further exploration is the effect of low rSO2 on neurologic outcomes in VLBW neonates. Neurologic outcome measures should include neuroimaging in the form of cranial ultrasounds to evaluate for intraventricular hemorrhage and term equivalent brain MRIs to evaluate for PVL. Additi onally, long term follow up of VLBW neonates should include neurodevelopmental assessment in the form of a scoring system such as the Bayley psychomotor and mental test at 1822 months, which is the typical follow up time for this patient population.

PAGE 37

37 REFERENCES 1. Bada H S Korones S B Perry E H Arheart K L Ray J D Pourcyrous M et al. Mean arterial blood pressure changes in premature infants and those at risk for intraventricular hemorrhage. J Pediatr 1990;117(4):607614. 2. Cordero L Giannone P J, Rich J T. Mean arterial pressure in very low birth weight (801 to 1500 g) concordant and discordant twins during the first day of life. J Perinatol 2003;23(7):545551. 3. Cordero L Timan C J, Waters H H Sachs LA. Mean arterial pressures during the first 24 hours of life in < or = 600gram birth weight infants. J Perinatol 2002;22(5):348353. 4. Cunningham S Symon A G Elton R A Zhu C McIntosh N. Intra arterial blood pressure reference ranges, death and morbidity in very low birthweight infants during the first seven days of life. Early Hum Dev 1999;56(23):151 165. 5. Lee J Rajadurai V S Tan K W. Blood pressure standards for very low birthweight infants during the first day of life. Arch Dis Child Fetal Neonatal Ed. 1999;81(3):F168170. 6. Leach R M Treacher D F. The pulmonary physician in critical care 2: oxygen delivery and consumption in the critically ill. Thorax 2002;57(2):170177. 7. Munro M J, Walker A M Barfield C P. Hypotens ive extremely low birth weight infants have reduced cerebral blood flow. Pediatrics 2004;114(6):15911596. 8. Tyszczuk L Meek J Elwell C, Wyatt J S. Cerebral blood flow is independent of mean arterial blood pressure in preterm infants undergoing intensive care. Pediatrics 1998;102(2 Pt 1):337341. 9. Blankenberg F G Loh N N Norbash A M Craychee J A Spielman D M Person B L et al. Impaired cerebrovascular autoregulation after hypoxic ischemic injury in extremely low birth weight neonates: detection with power and pulsed wave Doppler US. Radiology 1997;205(2):563568. 10. Tsuji M Saul J P du Plessis A Eichenwald E Sobh J Crocker R et al. Cerebral intravascular oxygenation correlates with mean arterial pressure in critically ill premature infants. Pediatrics 2000;106(4):625632. 11. Pryds O Greisen G Lou H Friis Hansen B. Heterogeneity of cerebral vasoreactivity in preterm infants supported by mechanical ventilation. J Pediatr 1989;115(4):638645. 12. Al Aweel I Pursley D M Rubin LP Shah B Weisberger S Richardson D K. Variations in prevalence of hypotension, hy pertension, and vasopressor use in NICUs. J Perinatol 2001;21(5):272278.

PAGE 38

38 13. Goldstein R F Thompson R J, Jr Oehler J M Brazy J E. Influence of acidosis, hypoxemia, and hypotension on neurodevelopmental outcome in very low birth weight infants. Pediatrics 1995;95(2):238243. 14. Grether J K Nelson K B Emery E S 3rd, Cummins S K. Prenatal and perinatal factors and cerebral palsy in very low birth weight infants. J Pediatr 1996;128(3):407414. 15. Hunt R W Evans N Rieger I Kluck ow M. Low superior vena cava flow and neurodevelopment at 3 years in very preterm infants. J Pediatr 2004;145(5):588592. 16. MiallAllen V M de Vries LS Whitelaw A G. Mean arterial blood pressure and neonatal cerebral lesions. Arch Dis Child 198 7;62(10):10681069. 17. Fanaroff A A Martin R J Neonatal Perinatal Medicine: Diseases of the Fetus and Infant. St. Louis: Mosby; 1992. 18. Evans N. Which inotrope for which baby? Arch Dis Child Fetal Neonatal Ed. 2006;91(3):F213220. 19. Watkins A M West C R Cooke R W. Blood pressure and cerebral haemorrhage and ischaemia in very low birthweight infants. Early Hum Dev 1989;19(2):103110. 20. Lundstrom K Pryds O Greisen G. The haemodynamic effects of dopamine and volume expansion in sick pr eterm infants. Early Hum Dev 2000;57(2):157163. 21. Osborn D Evans N Kluckow M. Randomized trial of dobutamine versus dopamine in preterm infants with low systemic blood flow. J Pediatr 2002;140(2):183191. 22. Kluckow M Evans N. Superior vena cav a flow in newborn infants: a novel marker of systemic blood flow. Arch Dis Child Fetal Neonatal Ed. 2000;82(3):F182187. 23. Ando Y Takashima S Takeshita K. Cerebral blood flow velocity in preterm neonates. Brain Dev 1985;7(4):385391. 24. Baraka A Naufal M El Khatib M. Correlation between cerebral and mixed venous oxygen saturation during moderate versus tepid hypothermic hemodiluted cardiopulmonary bypass. J Cardiothorac Vasc Anesth. 2006;20(6):819825. 25. Guarracino F. Cerebral monitoring during cardiovascular surgery. Curr Opin Anaesthesiol 2008;21(1):5054. 26. Leyvi G Bello R, Wasnick JD Plestis K. Assessment of cerebral oxygen balance during deep hypothermic circulatory arrest by continuous jugular bulb venous saturation and near inf rared spectroscopy. J Cardiothorac Vasc Anesth. 2006;20(6):826833.

PAGE 39

39 27. Wong F Y Nakamura M Alexiou T Brodecky V Walker A M. Tissue oxygenation index measured using spatially resolved spectroscopy correlates with changes in cerebral blood flow in newborn lambs. Intensive Care Med. 2009;35(8):14641470. 28. Hou X L Ding HY Zhou C L Tang X Y Ding H S Teng Y C et al. [Correlation of brain hypoxia at different degrees with brain function and brain damage investigated using near infr ared spectroscopy]. Zhonghua Er Ke Za Zhi 2007;45(7):523528. 29. Edmonds H L Jr., Ganzel B L Austin E H 3rd. Cerebral oximetry for cardiac and vascular surgery. Semin Cardiothorac Vasc Anesth 2004;8(2):147166. 30. Verhagen E A Ter Horst H J, Keating P Martijn A Van Braeckel K N Bos A F. Cerebral oxygenation in preterm infants with germinal matrix intraventricular hemorrhages. Stroke .41(12):29012907. 31. Tyson J E Parikh N A Langer J Green C Higgins RD. Intensive care f or extreme prematurity --moving beyond gestational age. N Engl J Med. 2008;358(16):16721681. 32. Mizrahi A Barlow O Berdon W Blanc W A Silverman W A. Necrotizing Enterocolitis in Premature Infants. J Pediatr 1965;66:697705. 33. Santulli T V Schullinger J N Heird W C Gongaware R D Wigger J Barlow B et al. Acute necrotizing enterocolitis in infancy: a review of 64 cases. Pediatrics 1975;55(3):376387.

PAGE 40

40 BIOGRAPHICAL SKETCH Rachel S. Garner was born in Albuquerque, NM She grew up in the San Francisco Bay a rea and graduated from Valley Christian High School (Dublin, CA) in 1997. She received her BS degree with a major in m olecular, cell and d evelopmental b iology and minors in c lassical civilizatio ns and p olitical science from the University of California, Los Angeles in 2001. Rachel then at tended Medical School at Albert Einstein College of Medicine in Bronx, NY where she received a Medical Doctorate with distinction in research in 2005. She comp leted her pediatrics residency at the University of Florida in 2008, where she received further training as a fellow in neonatology. Her fellowship research is supported by the American Heart Association Greater Southeast Affiliate Clinical Research Grant Program. After completing her fellowship in June 2011, she will continue her career in neonatology at the University of Arizona where she has accepted a faculty position.