Distress, Neuroimmune Dysregulation, and Clinical Outcomes in Women Undergoing Total Abdominal Hysterectomy and Bilatera...

MISSING IMAGE

Material Information

Title:
Distress, Neuroimmune Dysregulation, and Clinical Outcomes in Women Undergoing Total Abdominal Hysterectomy and Bilateral Salpingo Oophorectomy for Suspected Endometrial Cancer
Physical Description:
1 online resource (84 p.)
Language:
english
Creator:
Patidar, Seema M
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology, Clinical and Health Psychology
Committee Chair:
Pereira, Deidre B
Committee Members:
Mccrae, Christina Smith
Price, Catherine Elizabeth
Schumacher, Jessica Rachael

Subjects

Subjects / Keywords:
pni -- tah-bso
Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre:
Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Endometrial cancer is the most common gynecologic cancer in the United States (American Cancer Society, 2010). Psychoneuroimmunology literature has identified relationships between psychosocial factors, such as stress and depression, and clinical outcomes in various cancer populations (Antoni et al., 2006 and Rodrigue etal., 1999). Specific components of distress, including mood disturbance (Sephton et al., 2000), pain (Thornton et al., 2010), and sleep disturbance (Rich et al., 2005) have been associated with cortisol and cytokine production patterns that portend poorer clinical outcomes in cancer. However, no published research has examined relations among these psychosocial factors (mood disturbance, pain and sleep disturbance), cortisol dysregulation, vascular endothelial growth factor (VEGF), and clinical outcomes in women with endometrial cancer.  The current study included 113 women (Mage= 61.38, SD=9.09) who completed psychosocial and biological assessments prior to undergoing surgery for suspected endometrial cancer. Structural equation modeling was utilized to examine the direct andindirect relationships of Distress and Neuroimmune Dysregulation on Negative Clinical outcomes among this sample of women. All models demonstrated that sleep disturbance, pain and mood disturbance significantly contributed toDistress (Figures 3.1, 3.2 and 3.3). Additionally, abnormal cortisol production(lower AUCi) was significantly associated with advanced disease stage (pathcoefficient=-0.06, p= 0.03) and greater VEGF levels were significantly associated with greater post-operative complications (path coefficient=0.09, p=0.05) in this sample. Furthermore, an exploratory model evaluated the relationship between Distress and Neuroimmune function and included IL-6 as another neuroimmune factor. This model demonstrated the best overall fit (c2=7.95, AIC=45.95), but did not find asignificant relationship between Distress and Neuroimmune dysregulation (pathcoefficient=0.03, p=0.71).  Overall, these results support the development of interventions targeting sleep, pain and mood disturbance to improve quality of life in women undergoing surgery for suspected endometrial cancer. While the models could not fully evaluatemediation, model fit encouraged future studies to further evaluate the potential relationships among these factors. Additional research is necessary to determine if these symptoms of Distress are associated with clinical outcomes in women with endometrial cancer.
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 Seema M Patidar.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Pereira, Deidre B.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045175:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 DISTRESS, NEUROIMMUN E DYSREGULATION, AND CLINICAL OUTCOMES IN WOMEN UNDERGOING TOT AL ABDOMINAL HYSTERE CTOMY AND BILATERAL SALPINGO OOPHORECTOM Y FOR SUSPECTED ENDO METRIAL CANCER B y SEEMA M PATIDAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 201 3

PAGE 2

2 2013 Seema M Patidar

PAGE 3

3 Dedicated to my family

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank Deidre Pereira and my lab mates for their support and a researcher and I greatly appreciate her eagerness to share her knowledge and experiences wit h me. I also thank the members of my supervisory committee, Dr. Christina McCrae, Dr. Catherine Price, and Dr. Jessica Schumacher. Additionally, I would like to thank my family and friends for their continued encouragement and support through this process Finally, I greatly appreciate the participation of all the patients in this study and wish them the best.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 11 INTRODUCTION ................................ ................................ ................................ ........... 13 Epidemiology and Etiology of Endometrial Cancer ................................ ................. 13 Prognosis and Treatment of Endometrial Cancer ................................ ................... 14 Outcomes and Complications Following Endometrial Cancer Surgery ................... 15 Biobehavioral (Psychoneuroimmunologic) Model of Tumorigenesis & Cancer Outcomes ................................ ................................ ................................ ............ 16 Neuroimmunity and Clinical Outcomes in Cancer ................................ ................... 17 Distress, Neuroimmunity, and Clinical Outcomes in Can cer ................................ ... 18 Mood Disturbance ................................ ................................ ............................ 19 Pain ................................ ................................ ................................ .................. 20 Sleep Disturbance ................................ ................................ ............................ 21 Significance of Present Study ................................ ................................ ................. 22 Specific Aims ................................ ................................ ................................ .......... 23 METHODS ................................ ................................ ................................ .................... 26 Study Design ................................ ................................ ................................ .......... 26 Participants ................................ ................................ ................................ ............. 27 Procedures ................................ ................................ ................................ ............. 27 Assessment of Presurgical Psychosocial Factors ................................ ................... 28 Demographics ................................ ................................ ................................ .. 28 Mood Disturbance ................................ ................................ ............................ 28 Sleep Disturbance ................................ ................................ ............................ 29 Pain ................................ ................................ ................................ .................. 30 Assessment of Neuroimmunity ................................ ................................ ............... 31 Cortisol ................................ ................................ ................................ ............. 31 VEGF ................................ ................................ ................................ ................ 32 Assessmen t of Postsurgical Clinical Outcomes ................................ ...................... 32 FIGO Disease Stage ................................ ................................ ........................ 32 Postoperative Complications ................................ ................................ ............ 33 Assessment of Potential C ontrol Variables ................................ ............................. 34 Statistical Analyses of Study Aims ................................ ................................ .......... 34 Sample Size and Model Specification ................................ ................................ ..... 35

PAGE 6

6 Power Analysis ................................ ................................ ................................ 35 Model Specifications ................................ ................................ ........................ 36 RESULTS ................................ ................................ ................................ ...................... 41 Sample Characteristics ................................ ................................ ........................... 41 Descriptive Statistics for Measured Variables ................................ ......................... 42 Mood, Pain, and Sleep ................................ ................................ ..................... 42 VEGF and Cortisol AUCi ................................ ................................ .................. 42 Post operative Complications ................................ ................................ ........... 43 Biobehavioral Control Variables ................................ ................................ ....... 43 Analyses of Primary A ims ................................ ................................ ....................... 44 Associations Among Control Variables and Clinical Outcomes ........................ 44 Associations Among Measured Variables ................................ ........................ 44 Proposed Model and Subsequent Model Adjustments ................................ ..... 45 Model 1: Distress, Neuroimmune Markers and Postoperative Complications .. 46 Model 2: Distress, Neuroimmune Markers and Disease Stage ........................ 47 Exploratory Analyses ................................ ................................ .............................. 48 Model 3: Distress and Neuroimmune Dysregulation ................................ ........ 48 DISCU SSION ................................ ................................ ................................ ................ 56 Primary Aims ................................ ................................ ................................ .......... 56 Exploratory Analysis ................................ ................................ ............................... 64 Clinical Significance ................................ ................................ ................................ 66 Strengths and Limitations ................................ ................................ ....................... 68 ................................ ................................ ... 75 LIST OF REFERENCES ................................ ................................ ............................... 76 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 84

PAGE 7

7 LIST OF TABLES Table P age 3 1 FIGO Disease Stage ................................ ................................ .......................... 50 3 2 Postoperative Complication Categories ................................ .............................. 50 3 3 Zero order Correlations Among Measured Variables and Controls .................... 50 3 4 Partial Correlations (Controls: Age, BMI, Comorbidity) ................................ ....... 51 3 5 Model Summary ................................ ................................ ................................ 51 3 6 Fit Statistics ................................ ................................ ................................ ........ 51 3 7 Standardized Regression Weights ................................ ................................ ..... 51 3 8 Unstandardized Regression Weights ................................ ................................ 52

PAGE 8

8 LIST OF FIGURES Figure P age 1 1 Structural regression model to be examined in the proposed study ................... 25 2 1 Study Design ................................ ................................ ................................ ...... 40 3 1 Postoperative Complications Model with Standardized Regressions ................. 53 3 2 Disease Stage Model with Standardized Regressions ................................ ....... 54 3 3 Distress and Neuroimmune Dysregulation Model with Standardized Regressions ................................ ................................ ................................ ........ 55

PAGE 9

9 LIST OF ABBREVIATION S VEGF Vascular Endothelial Growth Factor AUCi Area Under the Curve with respect to Increase IL 6 Interleukin 6 ACS American Cancer Society BMI Body Mass Index TAH BSO Total Abdominal Hysterectomy & Bilateral Salpingo Oophorectomy PNI Psychoneuroimmunology HPA Hypothalamic Pituitary Adrenal ANS Autonomic Nervous System Th T Helper SEM Structural Equation Modeling NCI National Cancer Institute IRB Investiga tional Regulatory Board EDTA Ethylenediaminetetraacetic Acid SIGH AD Structured Interview Guide for the Hamilton Anxiety & Depression Scales PSQI Pittsburgh Sleep Quality Index MASQ MacArthur Sociodemographic Questionnaire PCOQ Patient Centered Outcomes Questionnaire ELISA Enzyme Linked Immunosorbent Assay CTCAE Common Terminology Criteria for Adverse Events AMOS Analysis of Moment Structures FIML Full Information Maximum Likelihood CBT Cognitive Behavioral Therapy IFN Interferon

PAGE 10

10 TNF Tumor Necrosis Factor SAM Sympatho Adrenal Medullary Hct Hematocrit Hgb Hemoglobin WBC White Blood Cells RBC Red Blood Cells AIC Akaike Information Cr i terion df Degrees of Freedom CFI Component Fit Index TLI Tucker Lewis Index RMSEA Root Mean Square Error of Approximation CR Critical Ratio SE Standard Error

PAGE 11

11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DISTRESS, NEUROIMMUN E DYSREGULATION, AND C LINICAL OUTCOMES IN WOMEN UNDERGOING TOT AL ABDOMINAL HYSTERE CTOMY AND BILATERAL SALPINGO OOPHORECTOM Y FOR SUSPECTED ENDO METRIAL CANCER By Seema M Patidar August 2013 Chair: Deidre B Pereira Major: Psychology Endometrial cancer is the most common gynecologic cancer in the United States (American Cancer Society, 2010). Psychoneuroimmunology literature has identified relationships between psychosocial factors, such as stress and depression, and clinical outcomes in various cancer populations (Antoni et al., 2006 and Rodrigue et al., 1999). Specific components of distress, including mood disturbance (Sephton et al., 2000), pain (Thornton et al., 2010), and sleep disturbance (Rich et al., 2005) have been associated with cortisol and cytokine production patterns that portend poorer clinical outcomes in cancer. However, no published research has examined relations among these psychosocial factors (mood disturbance, pain and sleep disturbance), cortisol dysregulation, v ascular endothelial growth factor (VEGF), and clinical outcomes in women with endometrial cancer. The current study included 113 women ( M age= 61.38, SD =9.09) who completed psychosocial and biological assessments prior to undergoing surgery for suspected endometrial cancer. Structural equation modeling was utilized to examine the direct and indirect relationships of Distress and Neuroimmune Dysregulation on Negative Clinical

PAGE 12

12 outcomes among this sample of women. All models demonstrated that sleep disturbanc e, pain and mood disturbance significantly contributed to Distress (Figures 3.1, 3.2 and 3.3). Additionally, abnormal cortisol production (lower AUCi) was significantly associated with advanced disease stage (path coefficient= 0.06, p= 0.03) and greater VE GF levels were significantly associated with greater post operative complications (path coefficient=0.09, p= 0.05) in this sample. Furthermore, an exploratory model evaluated the relationship between Distress and Neuroimmune function and included IL 6 as an other neuroimmune factor. This model demonstrated the best overall fit ( 2=7.95, AIC=45.95), but did not find a significant relationship between Distress and Neuroimmune dysregulation (path coefficient=0.03, p= 0.71). Overall, these results support the de velopment of interventions targeting sleep, pain and mood disturbance to improve quality of life in women undergoing surgery for suspected endometrial cancer. While the models could not fully evaluate mediation, model fit encouraged future studies to furth er evaluate the potential relationships among these factors. Additional research is necessary to determine if these symptoms of Distress are associated with clinical outcomes in women with endometrial cancer.

PAGE 13

13 CHAPTER 1 INTRODUCTION Epidemiology and Etiology of Endometrial Cancer Endometrial cancer is defined as cancer of the uterine lining, or endometrium. Endometrial cancer is the most common gynecologic cancer and the second leading cause of gynecologic cance r death in the United States. Approximately 43,470 new cases of endometrial cancer and 7,950 deaths from the disease were expected to occur in 2010 (ACS, 2010). There are several categories of endometrial cancer, including epithelial carcinoma and stromal/ mesenchymal tumors. The epithelial cancers include adenocarcinomas, of which the vast majority are endometrioid (80%; ACS, 2011). Less common types of endometrial adenocarcinomas include papillary serous and clear cell, which are histologically similar to those found in the ovary and fallopian tube and have a poorer prognosis than endometrioid cancers. Adenocarcinomas originate from abnormal glandular cells of the endometrium (ACS, 2011). Risk factors for developing endometrial adenocarcinomas include olde r age, estrogen use without progesterone, early menarche and/or late menopause, obesity, diabetes, infertility or never having children, Lynch syndrome, tamoxifen use, and polycystic ovary syndrome (ACS, 2011). Specifically, these risk factors have been im plicated in the pathophysiology of endometrial adenocarcinomas due to alterations in estrogen and progesterone levels in women, as well as promotion of inflammation (Wallace et al., 2010). Most endometrial cancer cells contain estrogen and progesterone rec eptors. Circulating hormones may bind to these receptors, leading to cellular proliferation of the endometrium (ACS, 2010). Although post menopausal women no longer have monthly ovarian production of these two hormones, they still

PAGE 14

14 experience production of estrogen from fatty tissue in the body. Post menopausal women with obesity have particularly elevated estrogen levels, and these elevated estrogen levels may lead to proliferation of the endometrium and increased risk of endometrial cancer (ACS, 2011). Obesity is also associated with chronic, systemic inflammation, and this has been suggested as a plausible biological mechanism for the association between obesity and endometrial cancer. A recent study found that elevations of specific cytokines, such a s Interleukin (IL) 6, significantly mediated the relationship between higher body mass index (BMI) and epithelial endometrial cancer risk (Dossus et al., 2010). This builds upon other research that discovered higher levels of proinflammatory cytokines in w omen with endometrial cancer (Slater et al., 2006). Prognosis and Treatment of Endometrial Cancer Most cases of endometrial cancer (69%) are diagnosed at an early stage. Post menopausal vaginal bleeding occurs in approximately 90% of women with endometri al cancer and is often the symptom that prompts early diagnostic evaluation (ACS, 2010). Survival rates in endometrial adenocarcinomas are higher if the initial diagnostic stage indicates local disease (current staging guidelines are provided in Appendix A ). Five year survival rates are relatively high in early stage disease (stage I: 75 88%, stage II: 69%), while more advanced stages (regional/distant disease) are associated with lower survival rates (stage III: 47 58%, stage IV: 15 17%). Myometrial invasi on of disease in earlier stages may serve as an indicator of survival, with greater invasion associated with poorer prognosis (ACS, 2011). Research indicates that disease stage at the time of surgery is one of the most important prognostic factors of endom etrial cancer (Sorosky, 2008).

PAGE 15

15 National Comprehensive Cancer Network (NCCN) guidelines (2011) recommend women with suspected diagnosis of endometrial cancer undergo surgical staging, which then guides additional treatment decisions (ACS, 2011). Surgical staging for endometrial cancer usually includes total hysterectomy with bilateral salpingo oophorectomy (TAH BSO) and pelvic washings, as well as pelvic and para aortic lymphadenectomy. Additional surgical procedures may include omentectomy, peritoneal was hings and/or debulking/cytoreduction. Women with advanced stage disease and women who are poor surgical candidates are likely to receive radiation and/or chemotherapy, as well (Sorosky, 2008). Outcomes and Complications Following Endometrial Cancer Surger y Although survival rates following TAH BSO for endometrial cancer are quite high, morbidity during the acute postsurgical period is prevalent. As described above, TAH BSO for endometrial cancer is invasive, and hospitalization following surgery can last from three to seven days even in the absence of any acute complications. Acute recovery (e.g., medical clearance to drive and lift > 10 lbs, etc.) can take four to six weeks (ACS, 2011), with full surgical recovery often requiring at least six months. R ates of acute complications following TAH BSO for endometrial cancer range from 15 23% for minor complications (e.g., urinary tract infection, anemia) and 12 19% for major, life threatening complications (e.g., arrhythmias, pulmonary embolism) (Janda et al ., 2010; Mourits et al., 2010). Factors associated with increased risk for acute postsurgical complications following TAH BSO for endometrial cancer include the presence of medical comorbidity (Tozzi et al., 2005; Mourits et al. 2010; Vaknin et al., 2009) and stage III/IV disease (Lambrou et al., 2004). Although psychosocial factors have been implicated in the development of negative clinical outcomes in other common

PAGE 16

16 cancers, such as breast cancer (Hjerl, et al. 2003 and Satin, et al. 2009), few psychosoci al factors have been evaluated as risk factors for poor outcomes following surgery for endometrial cancer. Biobehavioral (Psychoneuroimmunologic) Model of Tumorigenesis & Cancer Outcomes Given the rates of acute medical morbidity following TAH BSO for endometrial cancer, there is a need to identify modifiable presurgical risk factors associated with acute postoperative complications and cancer outcomes in this setting. Presurgical psychos ocial factors may emerge as potentially modifiable risk factors. Specifically, according to the biobehavioral (psychoneuroimmunologic [PNI]) model of tumorigenesis (Antoni et al., 2006), psychosocial factors may influence the tumor microenvironment and ca ncer outcomes by activating the hypothalamic pituitary adrenal (HPA) axis and autonomic nervous system (ANS). Activation of these stress systems then results in a cascade of events, including the release of stress hormones (e.g., glucocorticoids and catec holamines), which in turn may promote inflammatory, immunosuppressive, and angiogenic immune responses that (a) favor tumorigenesis (Antoni et al., 2006), and (b) promote neuroimmune mediated postsurgical complications (Menger & Vollmar, 2004). Presurgica l psychosocial factors that may be particularly relevant to examine within this model include those that are highly prevalent, can be validly assessed, and are potentially modifiable within the perioperative setting. Mood disturbance, sleep disturbance, an d pain are three psychosocial/behavioral factors that may be particularly important to target in the perioperative setting. Clinically significant anxiety (21%), depression (21 46%), sleep disturbance (37 60%), and pain (42 63%) are highly prevalent in wo men with gynecologic cancer (Thompson & Shear, 1998; Bradley et al.,

PAGE 17

17 2006; Lutgendorf et al., 2001; Portenoy et al., 1994; Rummans et al., 1998). Although these psychosocial experiences are highly prevalent in women with gynecologic cancers, there is evid ence that they are responsive to psychosocial interventions. Specifically, there is emerging evidence that cognitive behavioral interventions are effective at reducing pain, sleep disturbance, and mood disturbance, possibly in tandem, among cancer patients (Theobald, 2004; Dalton et al., 2004; Davidson et al., 2001; Antoni et al., 2009; Kwekkeboom et al., 2010). Furthermore, there is emerging evidence that these cognitive behavioral interventions, which target psychosocial factors, also influence neuroimmu nity by improving cortisol patterns (i.e., lowering evening serum cortisol levels) and cell mediated immunity (i.e., increasing production of interferon (IFN) gamma) in women with breast cancer (Antoni et al., 2009; Cruess, 2000; Savard et al., 2005), sugg esting that cognitive behavioral approaches may buffer neuroimmune responses associated with tumorigenesis and poor clinical outcomes in cancer. However, research has yet to examine how these common psychosocial experiences are associated with neuroimmunit y and clinical outcomes in women with gynecologic cancer. A comprehensive review of the literature linking these psychosocial factors (distress) with neuroimmunity and clinical outcomes is presented below. Neuroimmunity and Clinical Outcomes in Cancer Spe cific glucocorticoids (cortisol) and proinflammatory cytokines (VEGF) may be associated with clinical outcomes in women with cancer. For example, cortisol dysregulation (i.e. flattened diurnal slope) has been associated with increased mortality in women w ith breast cancer (Sephton et al., 2000). Additionally, cortisol may work synergistically with catecholamines to upregulate expression of VEGF (Antoni et al.,

PAGE 18

18 2006). This is important, as VEGF is a proangiogenic and proinflammatory cytokine, promoting gr owth of tumor vasculature and cell mediated inflammation in cancer (Reinders et al., 2003). Specifically in ovarian cancer, greater VEGF in serum and cytosol has been associated with tumor growth and angiogenesis, as well as the presence of metastatic dise ase and poorer survival (Lutgendorf et al., 2003; Thaker et al., 2007). Elevated VEGF in serum has also been found in women with endometrial cancer, as compared to women with non malignant conditions (Shaarawy and El Sharkawy, 2001). This finding is signif icant due to the association between elevations of VEGF in cytosol (intracellular fluid) and shorter disease free survival in this population (Chen et al., 2001). In addition, research on pro inflammatory markers (cytokines) in patients with heart disease found that the presence of greater preoperative proinflammatory cytokines was associated with more severe postoperative complications in patients undergoing surgery for heart disease (Kaireviciute et al., 2010). Taken together, these findings suggest that both cortisol and VEGF may be important prognostic markers for disease progression and negative clinical outcomes in women with endometrial cancer. Distress, Neuroimmunity, and Clinical Outcomes in Cancer Distress can impact patient quality of life and i s a notable clinical concern in cancer patients. The experience of distress in cancer often encompasses a range of physical, social and emotional factors (Vitek, Rosenzweig & Stollings, 2007). Common components of distress among cancer patients include de pression, anxiety, pain and sleep disturbance. Of note, research suggests that mood disturbance (depression and anxiety), pain and sleep disturbance co occur and maintain each other, thus impacting

PAGE 19

19 the quality of life in cancer patients (Shapiro et al., 2 003; Theobald 2004; Beck et al., 2005). As alluded to previously, the biobehavioral model of tumorigenesis posits that distress may impact cancer outcomes, such as postoperative outcomes and disease progression, by influencing cortisol production, impairi ng cell mediated immunity, and upregulating proinflammatory and proangiogenic cytokines (Antoni et al., 2006). However, research has yet to examine the aggregate effects of distress (mood disturbance, pain and sleep disturbance) and neuroimmune function o n clinical outcomes during the perioperative period in women with suspected endometrial cancer. Mood Disturbance Depression can act as a chronic stressor, altering HPA function and immunity, subsequently leaving cancer patients susceptible to infection af ter surgery (Spiegel & Giese Davis, 2003; Tjemsland et al., 1997). In patients undergoing hematopoietic stem cell transplant for hematologic cancers, mood disturbance (depression and anxiety) has been identified as a predictor of negative clinical outcomes such as length of survival after transplantation (Rodrigue et al., 1999). The association between depression and elevated VEGF in women with ovarian cancer suggests that mood may contribute indirectly to micrometastatic spread of disease (Lutgendorf et a l., 2008). While limited, methodologically rigorous research has been conducted with the a priori aim of examining the impact of depression on clinical outcomes in gynecologic cancer, there is a relatively rich literature on depression and clinical outcome s in heart disease. Depression has been implicated in the progression and mortality of patients with heart disease (Blumenthal et al., 2003). Additionally, autonomic nervous system (ANS) dysregulation may be the mechanism by which depression is associated with negative

PAGE 20

20 postoperative outcomes (i.e., extended hospitalizations) in patients undergoing major heart surgery (Dao et al., 2010). Anxiety has also been associated with neuroendocrine and immune dysfunction, and this relationship may be partic ularly robust among women with cancer. Studies of women with breast cancer have demonstrated that higher stress and anxiety may modulate patterns of diurnal cortisol production that promote unfavorable clinical outcomes (Antoni et al., 2009; Vedhara et al ., 2003). In addition, greater presurgical anxiety, pain and stress may lead to negative postsurgical outcomes, such as delayed wound healing and risk of infection (Broadbent et al., 2003; Ben Eliyahu, 2003; Pearson et al., 2005). Overall, these results s uggest that mood states, such as depression and anxiety, may promote negative clinical outcomes among women with cancer via alterations of both cortisol production and VEGF. However, no published research to date has examined these plausible relations. Pa in Pain has been identified as a critical, yet undertreated, symptom in cancer patients that may become chronic in more advanced stages. Severe, chronic pain may function as a stressor, leading to HPA axis dysregulation in the form of either overstimulatio n (greater cortisol output) (Tennant & Hermann, 2002) or understimulation (lower cortisol pain was associated with greater proinflammatory cytokine production after surgery. While limited data indicates that presurgical pain predicts greater incidence of postoperative complications in surgical populations, research has suggested that

PAGE 21

21 presurgi cal pain may be associated with longer hospitalizations after TAH BSO (Mourits et al., 2010). Women undergoing gastric bypass surgery also demonstrated relationships between presurgical pain and extended recovery times, as well as the presence of wound hea ling complications after surgery (McGuire et al., 2006). Taken together, these data suggest that greater presurgical pain may be a predictor of more negative post surgical outcomes in cancer. Given that chronic pain may operate as a significant stressor that is capable to modulating cancer related neuroimmunity (Antoni et al., 2006), the possibility exists that more severe presurgical pain may be associated with poorer surgical outcomes among women with endometrial cancer. Sleep Disturbance Sleep is of ten evaluated as a subjective construct and factors of sleep disturbance can include quality, duration, latency, and disruption (Buysse et al., 1989). Various characteristics of poor sleep, such as greater number of awakenings, lack of restful sleep, and p oor sleep quality, have been related to lower cortisol awakening (Backhaus, Junghanns, & Hohagen, 2004). Specifically, experiencing a lack of sleep may lead to patterns of c ortisol dysregulation that reflect impaired hippocampal regulation of negative feedback of the HPA axis (Spiegel et al., 1999; Irwin et al., 2003). Furthermore, chronic insomnia has been associated with impaired neuroendocrine function and cancer progres sion (Vgontzas & Chrousos, 2002). Sephton and Spiegel (2003) suggested that frequent awakenings are a prominent behavioral disturbance in cancer patients, particularly those with advanced disease and those receiving treatment in the hospital. They further posit that poor sleep marked by irregular sleep wake cycles

PAGE 22

22 may be associated with circadian disruption that in turn may influence disease progression (Sephton and Spiegel, 2003). Specifically, one of the functions of nocturnal sleep is to promote cell me diated immune responses (T helper cell type 1 [Th1] immunity), including anti tumor immunity, and suppress humoral immune responses (T helper cell type 2 [Th2] immunity), which are less effective at fighting tumors and characteristic of advanced stage dise ase (Dimitrov et al., 2004). Cortisol dysregulation due to sleep disturbance may promote a Th2 dominant immune response, ineffective anti tumor immunity, and disease progression. Also, sleep disturbance has been associated with higher levels of VEGF prio r to treatments for cancer, indicating that patients with preoperative sleep difficulty may also be experiencing immunologic changes promoting angiogenesis and/or inflammation (Mills et al., 2005 and Guess et al., 2009). Taken together, these findings su ggest that it is plausible that sleep disturbance is related to patterns of neuroimmunity that are associated with poorer outcomes among individuals with cancer. Given (a) the comorbidity among sleep disturbance, mood disturbance, and pain in individuals with cancer, and (b) the associations between each of these factors and neuroimmunity/poorer clinical outcomes in cancer, a logical next step is to examine the aggregate effect of these components on neuroimmunity/clinical outcomes in cancer. Significanc e of Present Study The current study aimed to examine the direct and indirect relationships among presurgical distress, presurgical neuroimmune dysregulation and post surgical clinical outcomes among women undergoing TAH BSO for suspected endometrial cance r. Presurgical distress was comprised of mood disturbance (anxiety and depression), pain

PAGE 23

23 levels, and sleep disturbance prior to surgery. Presurgical neuroimmune dysregulation was evaluated through measurements of salivary cortisol throughout the day and V EGF prior to surgery. Post surgical clinical outcomes included surgically staged disease (disease stage) and incidence/severity of postoperative complications. Structural equation modeling (SEM), specifically structural regression modeling, was utilized to examine a theoretical model of the shared variance explained by these factors. Specific to the current study, comorbid pain, sleep and mood disturbance were expected to be present in women prior to surgery and associated with neuroimmune function. As prev iously described, these symptoms often occur simultaneously and may be modified through cognitive behavioral interventions. Examination of the relationships among distress, neuroimmune function and important clinical outcomes may encourage more holistic ev aluation and treatment of women with suspected endometrial cancer. Specific Aims The overarching aims of the present study were two fold: (1) to derive three underlying biopsychosocial constructs (Distress, Neuroimmune Dysregulation, and Negative Clinic al Outcomes) from a specified set of observed variables (Distress pain, sleep disturbance, and mood disturbance; Neuroimmune Dysregulation vascular endothelial growth factor [VEGF] and diurnal cortisol area under the curve with respect to increase [AUC i]; and Negative Clinical Outcomes disease stage and postoperative complications), and (2) to model the relationships among these resultant theoretical constructs (Distress, Neuroimmune Dysregulation, and Negative Clinical Outcomes) in a sample of women undergoing surgery for suspected end ometrial cancer (Figure 1 1). Structural equation modeling (SEM) (i.e., structural regression modeling)

PAGE 24

24 was used to assess the adequacy of the model depicted in Figure 1 1. Specifically, the use of structural regressio n modeling allowed the following specific aims to be pursued: A im 1 To examine the total effect of presurgical distress (pain, sleep disturbance, and mood disturbance) on clinical outcomes (surgically staged disease, number of postoperative complications) in women undergoing TAH BSO for suspected endometrial cancer. Hypothesis 1 Greater presurgical distress was associated with worse clinical outcomes. A im 2 To examine the direct effect of presurgical distress (pain, sleep disturbance, and mood disturbance) on presurgical neuroimmune dysregulation (vascular endothelial growth factor [VEGF], diurnal cortisol area under the curve with respect to increase [AUCi]). Hypothesis 2 Greater presurgical distress was associated with greater pre surgical neuroimmune dysregulation. Aim 3 To examine the direct effect of presurgical neuroimmune dysregulation (VEGF, diurnal cortisol AUCi) on clinical outcomes (surgically staged disease, number of postoperative complications) in women undergoing TA H BSO for suspected endometrial cancer. Hypothesis 3 Greater presurgical neuroimmune dysregulation was associated with worse clinical outcomes. A im 4 To examine the indirect effect of presurgical distress (pain, sleep disturbance, and mood disturbance) on clinical outcomes (surgically staged disease, number of postoperative complications) in women undergoing TAH BSO for suspected endometrial cancer.

PAGE 25

25 Hypothesis 4 Greater presurgical neuroimmune dysregulation would mediate the relationsh ip between greater presurgical distress and worse clinical outcomes. Figure 1 1. Structural regression model to be examined in the proposed study

PAGE 26

26 CHAPTER 2 METHODS Study Design This study used data drawn from a larger, parent study examining psychoneurimmunologic relations in endometrial cancer immediately prior to surgery and four to six weeks following surgery (American Cancer Society [IRG 01 188 01] and National Cancer Institu te [R03 CA 117480], PI: Deidre Pereira). The design of the parent study was observational/non experimental and longitudinal, including biopsychosocial assessment at two timepoints (pre surgery and post surgery). This observational/non experimental stud y design is outlined in Figure 2 1. All participants were recruited from the Shands Gynecologic Oncology clinic in Gainesville, Florida as part of a study (PI: Deidre Pereira) funded by the American Cancer Society (IRG 01 188 01) and National Cancer Instit ute (R03 CA 117480). They were consented prior to surgery (TAH BSO) for suspected endometrial cancer. They completed a psychosocial assessment and collected biological samples, including saliva, during the three days prior to surgery. These assessments wer records. The study was conducted according to the regulations of the University of Florida Institutional Review Board (IRB). This study uti lized psychosocial and salivary cortisol data collected at the preoperative visit and negative clinical outcomes data obtained through medical record abstraction following the post surgical visit. Variables that may confound true relationships among pred ictors, mediators, and outcomes were assessed through presurgical self report measures and post surgical medical abstraction.

PAGE 27

27 Participants Participants from the parent study were recruited between 2003 and 2009. Inclusion criteria for the parent study wer e (a) women undergoing TAH BSO for suspected endometrial cancer, (b) fluency in spoken English, and (c) at least 18 years old. Exclusion criteria were (a) recurrent endometrial cancer, (b) metastasis to the uterine corpus from another site, (c) presurgical chemotherapy or radiotherapy, (d) current psychotic disorder, and (e) current suicidal ideation/plan. Participants for the present study included those with at least partial psychosocial, cortisol, and VEGF data at presurgery for whom clinical outcome dat a could be abstracted from medical records during the immediate postsurgery period. Procedures During their initial consultation visits, preliminarily eligible women were identified by their physician and were approached by study personnel. Study personn el introduced the study and reviewed the IRB approved informed consent with interested patients. After signing the consent form, patients completed a brief assessment screening for suicidality and psychosis. If a patient was still eligible, they received s tudy materials, which included psychosocial questionnaires and a saliva kit. The study saliva kit included 12 Salivettes (Sarstedt, Inc., Newton, NC), a cryomarker, and soft sided cooler for storage. Participants completed questionnaires and collected sal iva samples during the three days prior to their preoperative visit. Saliva samples were collected at four specific intervals (8am, 12pm, 5pm, and 9pm) across three consecutive days. When participants returned to the clinic for their preoperative visit, th ey returned questionnaires and saliva samples. Also, study personnel conducted a psychosocial interview, escorted patients

PAGE 28

28 to their blood draw, and compensated them for participation in this portion of the study. Blood samples were collected in ethylenedia minetetraacetic acid (EDTA) tubes and Biology at the University of Florida for processing. Assessment of Presurgical Psychosocial Factors Study personnel administered the Stru ctured Interview Guide for the Hamilton Anxiety and Depression Scale (SIGH AD) (Williams, 1988) and the Pittsburgh Sleep Quality Inventory (PSQI) (Buysse et al., 1989) in interview format during the preoperative visit and participants completed a set of qu estionnaires prior to undergoing surgery. These questionnaires included the MacArthur Sociodemographic Questionnaire (MASQ) (Adler et al., 2000) and the Patient Centered Outcomes Questionnaire (PCOQ) (Robinson et al., 2005). All measures used in the curren t study were included as appendices (Appendix B). Demographics Demographic factors were assessed by the MASQ (Adler et al., 2000). This questionnaire was developed by the MacArthur foundation to assess both subjective and objective social status. The MASQ assessed subjective social status by asking participants to indicate their perceived position in society. Objective social status was assessed by family structure, education level, employment status, and financial status. Mood Disturbance Mood disturbance was operationalized as total anxious and depressive symptomatology. Anxiety and depressive symptoms were assessed using the Structured Interview Guide for the Hamilton Anxiety and Depression Scale (SIGH AD) (Williams, 1988). The SIGH AD was developed as a semi structured interview, and in

PAGE 29

29 the current study, trained study personnel conducted the interview and rated symptomatology scores. Ratings associated with symptoms judged by the interviewer to be possibly/definitely due to the direct physiological effects of a medical condition, medication, treatment, or substance, were then subtracted from the total scores in order to yield depressive and anxious symptomatology scores unconfound ed by medical factors. The SIGH AD has previously been modified and utilized in research with patients with HIV (Brown et al., 1992, Mitrani et al., 2011). The anxiety and depression scales have well s alpha: SIGH A= 0.83, SIGH D: 0.83 (Mitrani et al., 2011). In the parent study, the original SIGH AD was modified to minimize participant burden and to exclude symptoms that can be reliably attributed to the direct physiological effects of endometrial cancer, such as abdominal pain/discomfort. The resulting abbreviated version was comprised of 24 items, including 15 depression items and 9 anxiety items. This modified score of the SIGH AD demonstrated strong internal Total scores range from 0 to 74, with higher scores indicating greater mood disturbance. The current study used the sum of the depressive symptomatology score (excluding any ratings possibly/definitely associated with physiological factors) and anxious sy mptomatology score (excluding any ratings possibly/definitely associated with physiological factors) to operationalize total mood disturbance. Sleep Disturbance Sleep disturbance was assessed using the PSQI (Buysse et al., 1989). The PSQI generated seven subscale scores (subjective sleep quality, sleep latency, sleep

PAGE 30

30 duration, sleep efficiency, sleep disturbances, use of sleep medication and daytime dysfunction), as well as a total (global) score. The PSQI has been used to evaluate sleep disturbance in ca ncer patients, including patients with gynecologic cancer, and construct validity (Beck et al., 2004; Davidson et al., 2002). This measure was administered in interview fo rmat, but still demonstrated similar internal consistency disturbance, with higher total PSQI scores indicating greater sleep disturbance. Pain Pain was assessed using items from a self report measure, the Patient Centered pain, (ii) fatigue, (iii) emotional distress and (iv) interference with daily activities, (b) perceptions of successful and exp ected treatment outcomes across these four domains, and (c) importance of improvement across these four domains (Robinson, et al., 2005). The PCOQ has demonstrated reliability ( r (21)=0.84 to 0.90, p <0.001) and validity ( r (21)=0.52, p <0.001) compared to oth er standardized measures, and has been used in research with various chronic pain populations (Brown, et al., 2008). The PCOQ 10 0 numeric rating scale (0=none, 10 0 =worst pain imaginabl e) with higher ratings indicating higher Inventory ( r (42)=0.78, p <0.001) in the current sample.

PAGE 31

31 Assessment of Neuroimmunity Cortisol Cortisol was measured in saliva, which has been considered a reliable measure of free circulating cortisol in the body (Kirschbaum & Hellhammer, 1994). Participants collected saliva samples at four specific times (8am, 12pm, 5pm, and 9pm) across three consecutive days prior to their preoperati ve clinic visit. Participants were asked to record the accurate time of their sampling if it differed from the expected times. Saliva samples were stored at 80 C until they could be shipped to Salimetrics (State College, PA) for analysis using Enzyme Link ed Immunosorbent Assay (ELISA). As described by Salimetrics (State College, PA), ELISA procedures for cortisol require adding an unknown amount of antigen (test sample) to a surface covered by cortisol antibodies. The antigen bound to the antibodies that w ere present and the magnitude of fluorescence emitted was assessed to determine cortisol concentration. All samples was assayed in duplicate and the test requires 25 L of saliva per determination. The test had a lower limi t of sensitivity of 0.0003 g/dL standard curve r ange of 0.012 g/dL to 3.0 g/dL average intra assay coefficient of variation of 3.5%, and average inter assay coefficient of 5.1%. The accuracy (100.8%) of this method was determined by spike and recovery and linearity (91.7%) was determined by serial dilution. Values were assessed from matched serum and saliva samples and demonstrated a strong linear relationship ( r (47) = 0.91, p < 0.0001). Various methods have been used to quantify cortisol output in psychoneuroimmunology litera ture and each method demonstrates a unique characteristic of cortisol production. Specifically, cortisol has commonly been evaluated using cortisol awakening response, diurnal cortisol slope, area under the curve with

PAGE 32

32 respect to ground and area under the c urve with respect to increase (Vedhara et al., 2006). In the current study, cortisol area under the curve with respect to increase (AUCi) was chosen as a measure of cortisol production. Pruessner and colleagues (2003) provided a trapezoidal formula (below) for computation of cortisol AUCi, which were used in the current study. AUCi was chosen because this measure represents systemic reactivity and dysregulation of cortisol production. Cortisol AUCi was calculated for each day of saliva collection, and th e study utilized an average value for each participant. VEGF Concentration of plasma VEGF was assessed in blood samples collected during the preoperative assessment visit. The blood samples were processed and plasma samples were stored for future assaying of immune factors, such as VEGF, by Dr. ally available ELISA kits were used to assay VEGF concentrations (pg/mL) in samples (VEGF Quantikine Kit, R&D Diagnostics, Minneapolis, MN). Briefly, each plate contains fluorescent beads linked with cytokine specific antibodies, and application of the sam ples to this plate allows cytokines to bind with these antibodies. The liquid sample is then assessed for VEGF concentration. Assessment of Postsurgical Clinical Outcomes FIGO Disease Stage Disease stage was determined surgically and was abstracted via me dical record review and classified according to International Federation of Gynecology and

PAGE 33

33 Obstetrics anatomic/prognostic group (Appendix A) (FIGO, 2008). Disease stage was assigned a numerical value based on the following scale: 0= benign/no evidence of c ancer, 1=Stage I, 2=Stage II, 3 = Stage III, and 4 = Stage IV. Postoperative Complications Incidence and severity of postoperative complications was abstracted via medical al medical records were cross referenced with adverse events listed in the Common Terminology Criteria for Adverse Events 4.03 (CTCAE; National Cancer Institute, 2009). The guide (included in Appendix B) was intended to foster reporting of adverse events a ssociated with cancer interventions. An adverse event was defined as an unfavorable and unexpected sign, symptom or disease that is temporally associated with a medical procedure, like TAH BSO, that may be directly related to the procedure. All adverse eve nts, or postoperative complications, were categorized by anatomical system, etiology, or purpose within an organ system. Each adverse event is also assigned a grade (1=mild, 2=moderate, 3=severe, 4=life threatening, 5=death), and the criteria for assignin g each grade are described in detail. For the present study, the entirety of the reviewed by two independent raters to assess for the occurrence of any events listed w ithin the CTCAE. The two raters demonstrated consistent post operative complication infection, and anemia. For each event recorded in the medical record, the severity of the event was graded using the criteria listed in the CTCAE. The two raters achieved consensus regarding any methodological questions or any post operative complications that required additional clarification according to the CTCAE (Table 2 2). For example,

PAGE 34

34 b lood loss is a notable category in the CTCAE, but the medical record of each participant has various ways to record the amount of blood loss. Therefore, clear numerical rating guidelines were determined based on what would be an expected amount of blood lo ss and what should be rated according to the CTCAE guidelines for blood loss (including hematocrit, hemoglobin and red blood cell counts). The outcome variable of interest was the sum of the severity scores for each postoperative complication (adverse even published PNI research has used the CTCAE to assess the incidence and severity of adverse events, the CTCAE has been utilized to assess long term follow up of postoperative complications in women undergoing TAH for endometrial cancer (Janda et al., 2010 and Mourits et al., 2010). Assessment of Potential Control Variables report measures. The MASQ was utilized to evalu ate potential objective and subjective control variables, such as age, marital status, and race/ethnicity. Additionally, medical record abstraction provided data on comorbid conditions and BMI, as these factors have been linked to worse disease outcome in cancer (Reeves et al., 2007, Satariano and Ragland, 1994, and Everett et al., 2003). Medical comorbidity was assessed using the Charlson Comorbidity index, with higher scores indicating a larger number of pre existing medical conditions (Charlson et al., 1987). Statistical Analyses of Study Aims The present study sought to examine the direct and indirect relationships among distress, neuroimmune dysregulation and negative clinical outcomes through structural regression modeling, a subtype of structural eq uation modeling (SEM) (Figure 1 1). A

PAGE 35

35 structural regression model is the combination of a path analysis model and measurement model (Kline, 2005). Simply stated, this approach allowed evaluation of relationships among observed variables and latent variable s within a model. SEM methods better accounted for measurement error variance than other common statistical approaches. Analysis of Moment Structures (AMOS) software (Arbuckle, 2008) used full information maximum likelihood (FIML) estimates to account for missing data (Carter, 2006). FIML was suggested when using SEM in small to medium sample sizes. Since the assumption of normally distributed continuous data plays a significant role in SEM, the measured data were evaluated for normality prior to running s tatistical analyses. Measured variables were statistically transformed, as needed, to attain normal distributions. Specific latent variables, with multiple indicators (at least 2), were examined in the current model (Table 2 1). The model fit was evalu ated recursively, and measured variables were expected to load onto their respective latent factors. Modifications to this model were guided by theoretical and statistical assumptions of SEM. However, if an appropriate model could not be fit to the data, a simpler path analysis model was utilized to examine the primary aims. Given that this was the case, Aims 1, 2, and 3 were examined by using linear regression analyses. A total of 3 equations (1 per aim) were tested. If the null hypotheses were rejecte d for each Aim, then Aim 4 (mediation) was tested using the methods of Preacher and Hayes (2004). Sample Size and Model Specification Power Analysis Few studies have evaluated psychosocial predictors of cortisol and VEGF and clinical outcomes among canc er populations. As such, there are few studies from which

PAGE 36

36 to draw effect sizes for adequate sample size estimation. However, a primary study by Vedhara and colleagues (2006) examining the effects of psychosocial predictors of cortisol and clinical outcom es among cancer populations demonstrated a moderate effect size ( r Another study conducted by Lutgendorf and colleagues (2002) investigated psychosocial predictors of VEGF and clinical outc omes in women with ovarian cancer and found large effect sizes ( f 2 =0.53). A recent study using an SEM model to evaluate relationships among psychosocial factors and neuroimmune factors found a moderate effect size for the relationship between psychosocial factors (pain, depression and fatigue) and cortisol ( r =0.35; Thornton et al., 2010). Therefore, the effect sizes among psychosocial factors, neuroimmunity, and clinical outcomes were expected to be in the moderate effect size range. Kline (2005) reviews power analysis approaches for SEM and suggested evaluation of power for multiple linear regressions as an estimation of power for SEM deemed sufficient to attain moderat e effect sizes ( f 2 =0.3). The sample size for the present study was estimated a priori to include data from 100 participants, as use of SEM and FIML approaches will allow inclusion of participants with some missing data. Thus, the a priori sample size esti mate was deemed to be adequate to detect these effect sizes. Model Specifications Additionally, few studies have utilized SEM approaches in psycho oncology, in part due to small sample sizes and high attrition (or death) in oncology samples (Schnoll et al. 2004). However, for the current study, a sample size of 100 participants was

PAGE 37

37 deemed to be sufficient a prior for accurate parameter estimates and appropriate error estimates. A similar sample size was utilized to evaluate an SEM model of similar complexi ty in a recent psycho oncology study of breast cancer patients (Thornton et al., 2010). In order to support a robust model with this sample size, data were examined for normality and efforts were made to transform all non normally distributed data (Lei & L omax, 2005). Another method to determine adequate sample size was the ratio of participants to observed (measured) variables. The current study utilized 7 observable variables (refer to Figure 1 1), so the suggested ratio of 10 cases per observed variable, was exceeded, signifying a robust model (Kline, 2005). measured variables is greater than the number of estimated parameters (Thompson, 2000). Keeping a positive value of degrees of freedom was a primary assumption in maximum number of parameters that can be estimated in a specific model. Degrees of freedom were determined by the following formula { df = [ p *( p +1)]/2}, where p = the number of observed variables (Thompson, 2000). Using this formula { df =28=[7*(8)]/2}, the current model could not estimate more than 27 parameters. Parameters in this model included the error associated with each measured variable ( 7), the associations between the indicators and respective latent variables (7), the correlations among measured variables (3) and the correlations among latent variables (3); a total of only 20 parameters. Additional parameters were estimated in recursive versions of this model, to fully examine all theoretical parameters. As previously mentioned, if the data were significantly non normal or sample size limited power of subsequent models, a

PAGE 38

38 simpler path analysis model was planned for examination of the stu dy aims among the observed variables.

PAGE 39

39 Table 2 1 Latent and measured v ariables Latent Variable Measured Variables Distress Pain (PCOQ), Sleep Disturbance (PSQI), and Mood Disturbance (SIGH AD: depression and anxiety) Neuroimmune Dysregulation VEGF and Cortisol AUCi Negative Clinical Outcomes Disease Stage and Postoperative Complications Table 2 2 CTCAE Consensus Items Discrepancy: Details Regarding Coding Decisions: Coding Blood Loss Hemoglobin values are the best indicator of anemia (used in the CTCAE) but the medical record often reports a variety lab values (red blood cells, hematocrit, or specific anemia subtype values). If hemoglobin was unavailable, a combination lab values was r equired to attain a severity rating for blood loss. Detailed values determined for each level of severity: 0= Nothing noted or only reported hematocrit (>26) 1= Mild: Asymptomatic Hgb 10 12, Hct 25 35, RBC 3 4 2= Moderate: Symptomatic Hgb 8 10, Hct 2 0 25, RBC 2 3 3= Severe: Transfusion indicated Hgb <8, Hct <25, RBC 2 Pain Ratings Normal postoperative pain includes routine medication: Abnormal pain levels were identified by a plan to monitor pain ratings (1), patient report of uncontrolled pain (1) and/or changes in pain medication (2) Prophylactic Antibiotics Did not code as a complication unless infection was present. Ileus vs. obstruction Different treatment modalities (TPN vs. elective interventions) led to different categorization of com plications. Abnormal Lab Values Any other irregular lab value being monitored is rated here if it does not fit a CTCAE category: White Blood Cells: High values: 1= Mild: A s y mptomatic WBC 10 17 2= Moderate: Symptomatic 17 100 3= Severe: Intervention >100 =leukocytosis (corresponds with CTCAE) Low values: 1= Mild: WBC<1000 2= Moderate: WBC<500 3=Severe: WBC not reg istered Platelet abnormalities: When monitoring minor changes, coded as 1 Respiratory Concerns Identification of complications and differentiation of categories: 1) When only treatment methods are recorded code as a 2 ex. incentive spirometer used for treatment of atelectasis 2) Atelectasis and pleural effusion should be coded separately. While they are likely to co occur, they are separate CTCAE categories which can result from different pathophysiology and have varying consequences. 3) Hypoxia is coded separately from other comp lications.

PAGE 40

40 Figure 2 1 Study d esign T0: Review Informed Consent & Screening Measures T1: Pre op Psychosocial Interview, Questionnaires, Blood Draw & Saliva Collection Surgery: TAH BSO T2: Post op Psychosocial Interview, Questionnaires, Blood Draw & Saliva Collection ~ 1 week ~ 4 weeks

PAGE 41

41 CHAPTER 3 RESULTS Sample Characteristics A total of 134 women who underwent surgery for suspected endometrial cancer were enrolled in the original study. From this sample, 21 women withdrew prior to surgery, did not complete preoperative measures, and did not have clinical outcome data available. Their data was a source of systematic missing data, which should not be included in SEM analyses. Therefore, the final sample was comprised of 113 women, who may have incomplete data at random. Since the women who withdrew from the study did not complete assessments prior to surgery, no comparative analyses could be conducted between the included and excluded participants. Descriptive statistics were utilized to evaluate normality of the data all biological variables, non normal psychosocial data and outco me data were natural log transformed prior to running statistical analyses. The final sample of women (N=113) were primarily Caucasian (91.10%), ranging in age from 35 84 ( M= 61.38, SD = 9.09), and 79.2% of the sample had at least a high school education. Re garding their current marital status, 57%.10% of women were currently married, 19.10% were divorced/separated, 17.10% were widowed, and 76.70% were never married. Overall, these women (N=80) reported an average family income of $25,000 $34,999, although a number of women did not complete this item of the MASQ.

PAGE 42

42 Descriptive Statistics for Measured Variables Mood, Pain, and S leep The mean composite SIGH AD score was 12.89 ( SD =9.02, range 0 51). PCO Q ratings indicated that while 36.4% of women were experiencing no pain, over 50% of women indicated a usual pain level of 10 or greater; usual pain ratings ranged from 0 98 ( M =21.25, SD =27.15). Furthermore, most women acknowledged some difficulty sleepin g. PSQI global scores ranging from 1 18 ( M =6.94, SD =4.30). Using a modified cut off score of 8, as recommended by Carpenter & Andrykowski (1998), in the current study approximately 30% of women in this sample had poor sleep quality, which corresponded to ( M =1.01, SD =0.79) on a scale from 1 (very poor) to 5 (excellent). Additionally, about 50% of women reported sleep efficiencies lower than 85%, which has been considered the upper limit for s leep efficiency in someone described as a poor sleeper (Espie et al., 2003). Additionally, within this sample of women 665.7% reported getting less than or equal to 7 hours of sleep each night, 96.2% experienced sleep disturbances (subscale score>0) and 65 .4% reported sleep onset latency greater than 15 minutes (subscale score>0). Despite the overall sleep difficulty these women were experiencing prior to surgery, only 310.8% of women reported use of sleep medications and 15.4% acknowledged frequent daytime dysfunction (subscale score>1). VEGF and Cortisol AUCi cortisol AUCi levels. Mean VEGF concentration was 265.91 ( SD =340.09), while mean natural log transformed cortisol AUCi level was 1.05 ( SD =0.74). These values were transformed for parametric analyses.

PAGE 43

43 FIGO Disease S tage While 9 women (8%.0%) were found to have benign disease (Stage 0), most women were found to have a diagnosis of cancer ( N =105, 92.0%). Surgical staging c ommonly revealed Stage Ia ( N =210, 197.7%) and Ib ( N =46, 410.7%) endometrial cancer, and no women in this sample had metastasis at the time of surgery (stage IV, Table 3 1). Of the women who received a diagnosis of cancer, 88 (78%) had a diagnosis of endome trial adenocarcinoma, endometrioid type. Post operative C omplications Thirty six women (321.86%) experienced no postoperative complications. Of the remaining 77 women who experienced postoperative complications, the mean overall sum score (accounting f or severity of each complication) was 3.77 ( SD =5.60, range =1 35) The most common com plications included respiratory thoracic and mediastinal disorders ( N =43) or blood and lymphatic system disorders ( N =52). There were a total of 210 different complication s identified, which were categorized as mild (37%), moderate (30%), and severe (33%) complications according to the CTCAE. Biobehavioral Control V ariables M =36.45 kg/m2 obesity class II, SD = 11.06 ) and medical comorbidity at the time of surgery (Charlson comorbidity score: M =2.50, SD =1.04). Specifically, this sample of women included women with 0 5 comorbid conditions, with 61 women having one documented comorbid condition and 67 women having two o r more documented comorbid conditions. Comorbid conditions included diabetes ( N =38), coronary artery disease ( N =10), connective tissue disorder ( N =8), chronic pulmonary disease ( N =7), cerebrovascular

PAGE 44

44 disease ( N =5), congestive heart failure ( N =2), peripheral vascular disease ( N =2), and moderate to severe renal disease ( N =1). Analyses of Primary Aims Associations Among Control Variables and Clinical Outcomes Correlations were computed to examine the relationship between potential control variables and dependent variables/clinical outcomes (Table 3 3). Age, Body Mass Index (BMI), and Charlson Comorbidity Score were all significantly related to neuroimmune markers and/or clinical outcomes in the study. Older age was significantly associated with great er cortisol AUCi levels, ( r( 81)=0.24, p< 0.05). With regard to clinical outcomes, low er BMI, ( r (109)= 0.24, p< 0.05,) and higher Charlson comorbidity scores, ( r (110)=0.24, p< 0.05) were significantly associated with more advanced disease stage. Age, BMI, a nd Charlson Comorbidity scores were controlled for in the correlation Subsequent Model Adjustments control variables were added to the models. As such, these control variables were removed from the SEM models. Associations Among Measured Variables Zero order correlations were calculated among the me asured variables (Table 3 4). PSQI global sleep ratings and SIGH AD ratings were hi ghly correlated, suggesting that participants who experience greater sleep disturbance may also experience greater psychological distress, ( r (94)=0.66, p< 0.001). PCO Q pain ratings were significantly correlated with PSQI global ratings, indicating that par ticipants with higher levels of usual pain may also experience greater sleep disturbance ( r (69)=0.34, p< 0.001). PCO Q ratings were highly, positively correlated with SIGH AD ratings, ( r (73)=0.30, p< 0.001).

PAGE 45

45 Correlation analysis did not reveal relationships a mong Distress factors were not significantly associated with any neuroimmune markers or clinical outcomes (Table 3 3). There was a marginal, positive correlation b etween VEGF and Cortisol AUCi, ( r (64)=0.21, p= 0.09). However, post operative complication s and dise ase stage were not associated, ( r (109)=0.07, p= 0.49). Cortisol AUCi was negatively correlated with FIGO disease stage, ( r (81)= 0.22, p< 0.05), but not postoperative complication scores, ( r (80)= 0.05, p= 0.67. VEGF values were not associated with disease stage, ( r (81)=0.05, p= 0.64,) or postoperative complication scores, ( r (79)=0.10, p= 0.37). Correlation analyses among measured variables while controlling for BMI, Age and Charlson Comorbidity Scores revealed that PSQI global sleep ratings and PCO Q pain ratings were still significantly associated with SIGH AD ratings (PSQI: r (43)=0.52, p< 0.001; PCO Q : r (43)=0.39, p< 0.015) (Table 3 5). However, the relationship between higher levels of pain and greater sleep difficulty bec ame only marginally signifi cant ( r (43)=0.25, p= 0.08). VEGF and cortisol AUCi were not significantly associated, ( r (43)=0.19, p= 0.22). Higher VEGF levels were marginally no longer significantly associated with greater post operative complications, ( r (43)=0.13, p= 0.4206), or diseas e stage, r (43)=0.18, p= 0.25., Additionally, while lower cortisol AUCi was not significantly associated with more advanced disease stage, ( r (43)= 0.23, p< 0.05=0.13), or postoperative complications, r (43)= 0.05, p= 0.72. The symptoms of distress were not rel ated to the neuroimmune markers or negative clinical outcomes (Table 3 4). Proposed Model and Subsequent Model Adjustments The proposed model intended to examine a mediational relationship among three underlying biopsychosocial constructs (Distress, Neuro immune Dysregulation, and Negative Clinical Outcomes) from a specified set of observed variables (Distress

PAGE 46

46 pain, sleep disturbance, and mood disturbance; Neuroimmune Dysregulation vascular endothelial growth factor [VEGF] and diurnal cortisol area under the curve with respect to increase [AUCi]; and Negative Clinical Outcomes disease stage and postoperative complications). However, versions of this mediation model were unidentified and could not be evaluated with the proposed structure. Consequently, th e initial model structure was adjusted to theoretically evaluate the primary aims of the study utilizing only the latent construct of Distress. This construct was composed of respective measured variables that were interrelated. Given that (a) latent facto rs constructed with less than three measured variables may not be as structurally sound and (b) both the latent factors of Neuroimmune Dysfunction and Negative Clinical Outcomes contained only two measured variables, these latent factors were dismantled. The latent factor of Neuroimmune Dysregulation was separated into two individual measured variables (multiple mediators), while the latent factor of Negative Clinical Outcomes was separated into two individual outcomes of interest (Figures 3 1 and 3 2). T hen, SEM was used to assess the adequacy of these adjusted models allowing examination of the specific aims of the study. Model 1: Distress, Neuroimmune Markers and Postoperative Complications The first model (Figure 3 1) examined the indirect effect of presurgical distress (pain, sleep disturbance, and mood disturbance) on postoperative complications in women undergoing TAH BSO for suspected endometrial cancer. Greater levels of presurgical neuroimmune markers (VEGF and cortisol AUCi), were expected to m ediate the relationship between greater presurgical distress and postoperative complications. This model demonstrated that greater Distress was significantly predicted by associated with higher PCO Q pain ratings (path coefficient= 1.50, p< 0.001), higher

PAGE 47

47 P SQI scores (path coefficient=0.67, p< 0.001), and higher SIGH AD ratings of anxiety and depression (path coefficient=0.75, p< 0.001;Table 3 8). The model also found a significant relationship between higher VEGF levels and greater postoperative complications (path coefficient=0.09, p= 0.05), but not cortisol AUCi and postoperative complications (path coefficient= 0.01, p= 0.55) However, a relationship did not emerge between Distress and VEGF (path coefficient=0.14, p= 0.72) or Cortisol AUCi (path coefficient=0.57, p= 0.65). While the model was considered to be a good overall fit to the 2 =10.65, P=0.22, CFI=0.95), mediation could not be further evaluated (Tables 3 5 and 3 6). Model 2: Distress, Neuroimmune Marker s and Disease Stage The second model (Figure 3 2) examined the indirect effect of presurgical distress (pain, sleep disturbance, and mood disturbance) on surgically staged disease in women undergoing TAH BSO for suspected endometrial cancer. Greater level s of presurgical neuroimmune markers (VEGF and cortisol AUCi) were expected to mediate the relationship between greater presurgical distress and disease stage. Contrary to hypotheses, a significant relationship between lower cortisol AUCi (indicating HPA axis understimulation, rather than overstimulation) and advanced disease stage was found in this model (path coefficient= 0.06, p< 0.05); however, no relationship was found between VEGF and disease stage (path coefficient=0.09, p= 0.37). Consistent with Model 1, greater Distress was signific antly predicted associated with higher PCO Q pain ratings (path coefficient=1.48, p< 0.001), higher PSQI scores (path coefficient =0.68, p< 0.001), and higher SIGH AD ratings of anxiety and depression (path coefficient=0.75, p< 0.001; Table 3 8). There were no significant relationships found between VEGF and Distress (path coefficient=0.18, p= 0.63) or

PAGE 48

48 Cortisol AUCi and Distress (path c oefficient=0.57, p= 0.64). While the fit statistics supp 2 =10.81, P=0.21, CFI=0.95), the model was comparable to the previous model evaluating postoperative complications, since it did not elucidate a relationship among Distre ss and neuroimmune markers and did not demonstrate an improvement in fit (Table 3 6, AIC model 1=48.65, AIC model 2= 48.81). Exploratory Analyses Model 3: Distress and Neuroimmune Dysregulation One of the primary aims was to evaluate the relationship amo ng the latent factors of Distress and Neuroimmune Dysregulation. In order to evaluate this initial aim, IL 6, another critical inflammatory marker in endometrial cancer, was added to the original model. The mean level of IL 6 in this sample was 4.20 g/mL (SD=14.92). The exploratory model (Figure 3 3) was identified after the addition of IL 6 utilizing three measured variables for each latent factor. This model demonstrated so 2 =7. 95, P=0.44, CFI=1.00, AIC=45.95 (Tables 3 5 and 3 6), but still did not reveal a significant relationship among Distress and Neuroimmune Dysregulation (path coefficient=0.03, p= 0.71; Table 3 8). Higher PCO Q pain ratings (path coefficient=1.53, p< 0.001), higher PSQI scores (path coefficient=0.65, p < 0.001), and higher SIGH AD ratings of anxiety and depression (path coefficient=0.76, p< 0.001) were all consistently associated with the Distress factor. Additionally, VEGF, IL 6 and cortisol (AUCi) did not significantly contribute to the Neuroimmune Dysre gulation (VEGF: path coefficient=3.66, p= 0.34, IL 6: path coefficient=0.99, p= 0.27 and AUCi: path coefficient=10.05, p= 0.27), but no significant relationships were found with this latent factor. When the Negative Clinical

PAGE 49

49 Outcomes latent factor was added t o this model, the model was unidentified suggesting a structural and theoretical problem when this latent factor was added to the SEM models presented in this study.

PAGE 50

50 Table 3 1 FIGO d isease s tage f % Valid % Cumulative % Stage 0 9 8.0 8.0 8.0 Stage IA 21 18.6 18.6 26.5 Stage IB 46 40.7 40.7 67.3 Stage IC 9 8.0 8.0 75.2 Stage IIA 4 3.5 3.5 78.8 Stage IIB 11 9.7 9.7 88.5 Stage IIIA 8 7.1 7.1 95.6 Stage IIIB 1 0.9 0.9 96.5 Stage IIIC 4 3.5 3.5 100.0 Table 3 2 Postoperative complication c ategories Postoperative Complication Categories Frequency General Disorders and Administration Site Conditions 14 Cardiac Disorders 10 Vascular Disorders 11 Gastrointestinal Disorders 18 Respiratory, Thoracic, and Mediastinal Disorders 43 Blood and Lymphatic System Disorders 52 Metabolism and Nutrition Disorders 8 Renal and Urinary Disorders 6 Infections and Infestations 11 Nervous System Disorders 1 Injury, Poisoning and Procedural Complications 5 Investigations 28 Musculoskeletal and Connective Tissue Disorders 2 Endocrine Disorders 1 Table 3 3 Zero order correlations among measured variables and c ontrols Age BMI Comor bidity SIGH AD PCO Q PSQI Cortisol AUCi VEGF FIGO Cancer Stage Age (N=113) BMI (N=111) 0.19 Comorbidity (N=112) 0.01 0.04 SIGH AD (N=106) 0.30 *** 0.06 0.04 PCO Q (N=77) 0.09 0.21 0.06 0.30 *** PSQI(N=96) 0.32 *** 0.04 0.07 0.66 *** 0.34 *** Cortisol AUCi (N=83) 0.24 0.02 0.02 0.04 0.03 0.05 VEGF (N=83) 0.08 0.04 0.08 0.06 0.04 0.07 0.21 FIGO Cancer Stage (N=113) 0.03 0.24 0.24 0.11 0.02 0.14 0.22 0.05 Postoperative Complications (N=111) 0.05 0.17 0.13 0.02 0.13 0.09 0.05 0.10 0.07 *Indicates p< 0.05, **Indicates p< 0.01, ***Indicates p< 0.001

PAGE 51

51 Table 3 4 Partial c orrelations (Controls: Age, BMI, Comorbidity) N=43 SIGH AD PSQI PCO Q Cortisol AUCi VEGF FIGO Cancer Stage SIGH AD PSQI 0.52*** PCO Q 0.25 0. 39* AUCi 0.13 0.11 0.12 VEGF 0.16 0.08 0.07 0.19 FIGO Cancer Stage 0.11 0. 14 0.08 0.23* 0.18 Postoperative Complications 0.08 0.15 0.16 0.05 0.13 0.01 *Indicates p< 0.05, **Indicates p< 0.01, ***Indicates p< 0.001 Table 3 5 Model s ummary Model NPAR CMIN DF P CMIN/DF 1. Post operative Complications Model 19 10.65 8 0.22 1.33 2. Disease Stage Model (FIGO) 19 10.81 8 0.21 1.35 3. Distress and Neuroimmune Dysregulation 19 7.95 8 0.44 0.99 Table 3 6 Fit s tatistics Model 2 RMSEA TLI CFI Hoelter AIC 1. Postoperative Complications Model 10.65 0.05 (0.41) 0.87 0.95 164 48.65 2. Disease Stage Model (FIGO) 10.81 0.06 (0.39) 0.86 0.95 161 48.81 3. Distress and Neuroimmune Dysregulation 7.95 0.00 (0.63) 1.00 1.00 219 45.95 Table 3 7 Standardized regression w eights 1 2 3 Pain and Distress 0.43 0.42 0.43 PSQI Global and Distress 0.85 0.85 0.84 Total SIGH AD and Distress 0.74 0.73 0.74 VEGF and Distress 0.05 0.06 Cortisol AUCi and Distress 0.06 0.06 Postoperative Complications and AUCi 0.07 Postoperative Complications and VEGF 0.21 Disease Stage and AUCi 0.23 Disease Stage and VEGF 0.10 Distress and Neuroimmune Dysregulation 0.08 VEGF and Neuroimmune Dysregulation 0.49 IL 6 and Neuroimmune Dysregulation 0.40 Cortisol AUCi and Neuroimmune Dysregulation 0.42

PAGE 52

52 Table 3 8 Unstandardized regression w eights 1. Postoperative Complications Model Estimate S.E. C.R. P VEGF and Distress 0.14 0.38 0.36 0.72 Cortisol AUCi and Distress 0.57 1.23 0.46 0.65 Pain and Distress 1.50 0.52 2.88 0 .00 PSQI Global and Distress 0.67 0.23 2.88 0.00 Total SIGH AD and Distress 0.75 0.24 3.17 0.00 Post operative Complications and Cortisol 0.01 0.01 0.60 0.55 Post operative Complications and VEGF 0.09 0.05 1.95 0.05 2. Disease Stage Model Estimate S.E. C.R. P VEGF and Distress 0.18 0.38 0.48 0.63 Cortisol AUCi and Distress 0.57 1.23 0.47 0.64 Pain and Distress 1.48 0.52 2.86 0.00 PSQI Global and Distress 0.68 0.24 2.86 0.00 Total SIGH AD and Distress 0.75 0.24 3.17 0.00 FIGO Disease Stage and Cortisol 0.06 0.03 2.18 0.03 FIGO Disease Stage and VEGF 0.09 0.10 0.90 0.37 3. Distress and Neuroimmune Dysregulation Estimate S.E. C.R. P Distress and Neuroimmune Dysregulation 0.03 0.08 0.38 0.71 Pain and Distress 1.53 0.52 2.92 0.00 PSQI Global and Distress 0.65 0.22 2.92 0.00 Total SIGH AD and Distress 0.76 0.24 3.16 0.00 VEGF and Neuroimmune Dysregulation 3.66 3.82 0.96 0.34 IL 6 and Neuroimmune Dysregulation 0.99 0.90 1.10 0.27 Cortisol AUCi and Neuroimmune Dysregulation 10.05 9.11 1.10 0.27

PAGE 53

53 Figure 3 1 Postoperative complications model with standardized r egressions

PAGE 54

54 Figure 3 2 Disease stage model with standardized r egressions

PAGE 55

55 Figure 3 3 Distress and neuroimmune dysregulation model with standardized r egressions

PAGE 56

56 CHAPTER 4 DISCUSSION Primary Aims

PAGE 57

57

PAGE 58

58

PAGE 59

59

PAGE 60

60

PAGE 61

61

PAGE 62

62

PAGE 63

63

PAGE 64

64 Exploratory Analysis

PAGE 65

65

PAGE 66

66 Clinical Significance

PAGE 67

67

PAGE 68

68 Strengths and Limitations Additionally, the women who withdrew from the study did not complete assessments prior to surgery, s o no comparative analyses could be conducted between the included and excluded participants.

PAGE 69

69

PAGE 70

70

PAGE 71

71

PAGE 72

72

PAGE 73

73

PAGE 74

74

PAGE 75

75 APPENDIX

PAGE 76

76 LIST OF REFERENCES Adler, N.E., Epel, E.S., Castellazzo, G., & Ickovics, J.R. (2000). Relationship of subjective and objective social status with psychological and physiological functioning: preliminary data in healthy white women. Health Psychology, 19, 586 592. American Cancer Society (2010). Cancer facts and fig ures 2010. Atlanta, GA: American Cancer Society. American Cancer Society (2011). Detailed Guide: Endometrial Cancer. Atlanta, GA: American Cancer Society. Antoni, M.H., Lechner, S., Diaz, A., Vargas, S., Holley, H., Phillips, K., et al. (2009). Cognitiv e behavioral stress management effects on psychosocial and physiological adaptation in women undergoing treatment for breast cancer, Brain Behavior Immunity, 23, 159 166. doi:10.1016/j.bbi.2008.09.005. Antoni, M.H., Lutgendorf, S.K., Cole, S.W., Dhabhar, F.S., Sephton, S.E., McDonald, P.G. et al. (2006). The influence of bio behavioural factors on tumour biology: pathways and mechanisms. Nature Reviews. Cancer, 6, 240 248. Arbuckle, J. L. (2008). Amos 17.0 [computer software]. Spring House, PA: SPSS. B ackhaus, J., Junghanns, K., & Hohagen, F. (2004). Sleep disturbances are correlated with decreased morning awakening salivary cortisol. Psychoneuroendocrinology, 29, 1184 1191. doi:10.1016/j.psyneuen.2004.01.010. Beck, S.L., Dudley, W., & Barsevick, A. (2 005). Pain, Sleep Disturbance, and Fatigue in Patients With Cancer: Using a Mediation Model to Test a Symptom Cluster. Oncology Nursing Forum, 32 (3), E48 E55. doi:10.1188/05.ONF.E48 E55. Beck, S.L., Schwartz, A.L., Towsley, G., Dudley, W., & Barsevick, A (2004). Psychometric evaluation of the Pittsburgh Sleep Quality Index in cancer patients. Journal of Pain and Symptom Management, 27(2), 140 148. doi:10.1016/j.jpainsymman.2003.12.002 Bellone, S., Watts, K., Cane, S., Palmieri, M., Cannon, M.J., Burnett A., Roman, J.J., Pecorelli, S., and Santin, A.D. (2005). High serum levels of interleukin 6 in endometrial carcinoma are associated with uterine serous papillary histology, a highly aggressive and chemotherapy resistant variant of endometrial cancer. Gyn ecologic Oncology, 98:92 98. doi:10.1016/j.ygyno.2005.03.016 Ben Eliyahu, S. (2003). The promotion of tumor metastasis by surgery and stress: immunological basis and implications for psychoneuroimmunology. Brain, Behavior, and Immunity, 17, S27 S36.

PAGE 77

77 Blumenthal, J.A., Lett, H.S., Babyak, M.A., White, W., Smith, P.K., Mark, D.B., Jones, R., Matthew, J.P., & Newman, M.F. (2003). Depression as a risk factor for mortality after coronary artery bypass surgery. Lancet. 362, 604 609. Bradley, S., Rose, S., L utgendorf, S., Costanzo, E., & Anderson, B. (2006) Quality of life and mental health in cervical and endometrial cancer survivors. Gynecol Oncology. 100, 479 486. doi:10.1016/j.ygyno.2005.08.023. Broadbent, E., Petrie, K.J., Alley, P.G., Booth, R.J. (2003 ). Psychological stress impairs early wound repair following surgery. Psychosomatic Medicine, 65, 865 869. Brown, G.R., Rundell, J.R., McManis, S.E., Kendall, S.N., Zachary, R., & Temoshok, L. (1992). Prevalence of psychiatric disorders in early stages of HIV infection. Psychosomatic Medicine, 54, 588 601. Brown J.L. Edwards, P.S., Atchison, J.W., Lafayette Lucey, A., Witmer, V.T., & Robinson, M.E. (2008). Defining patient centered, multi dimensional success criteria for treatment of chronic spine pain. Pain Med; 9(7): 851 862. Buysse, D.J., Reynolds, C.F., Monk, T.H., Berman, S.R., & Kupfer, D.J. (1989). Psychiatry Research 28:193 213. doi: 10.1016/0165 1781(89)90047 4 Carpe nter, J.S. & Andrykowski, M.A. (1998). Psychometric evaluation of the Pittsburgh Sleep Quality Index. Journal of Psychosomatic Research 45(1): 5 13. doi: 10.1016/S0022 3999(97)00298 5 Carter, R.L. (2006) Solutions for missing data in structural equation modeling. Res Pract Assess, 1(1):1 6. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbid ity in longitudinal studies: development and validation. J Chronic Dis, 40(5), 373 383 Chen, C.A., Cheng, W.F., Lee, C.N., Wei, L.H., Chu, J.S., Hsieh, F.J., et al. (2001). Cytosol vascular endothelial growth factor in endometrial carcinoma: Correlation w ith disease free survival. Gynecologic Oncology, 80, 207 212. Chida, Y. & Steptoe, A. (2009). Cortisol awakening response & psychosocial factors: A systematic review & meta analysis. Biological Psychology, 80: 265 278. doi:10.1016/j.biopsycho.2008.10.004 Cruess, D.G., Antoni, M.H., Kumar, M., Schneiderman, N. (2000). Reductions in salivary cor tisol are associated with mood improvement during relaxation training among HIV seropositive men. Journal of Behavioral Medicine, 23, 107 122.

PAGE 78

78 Cohen, J. (1977). Statistical power analysis for the behavioral sciences (rev. ed.). Hillsdale, NJ England: Lawr ence Erlbaum Associates, Inc. Cozzolino F, Torcia M, Lucibello M, Morbidelli L, Ziche M, Platt J, Fabiani S, Brett J, Stern D (1993) Interferon and interleukin 2 synergistically enhance basic fibroblast growth factor synthesis and induce release, promoting endothelial cell growth. J Clin Invest 91:2504 2512 Dalton, J.A., Keefe, F.J., Carlson, J., & Youngblood, R. (2004). Tailoring cognitive behavioral treatment for cancer pain. Pain Manag Nurs, 5, 3 18. doi:10.1016/S1524 9042(03)00027 4. Dao, T.K. Youssef, N.A., Gopaldas, R.R., Chu, D., Bakaeen, F., Wear, E. & Menefee, D. (2010). Autonomic cardiovascular dysregulation as a potential mechanism underlying depression and coronary artery bypass grafting surgery outcomes. Journal of Cardiothoracic Surg ery, 5 36. doi: 10.1186/1749 8090 5 36. Davidson, J.R., MacLean, A.W., Brundage, M.D., & Schulze, K. (2002). Sleep disturbance in cancer patients. Social Science and Medicine, 54, 1309 1321. Davidson, J.R., Waisberg, J.L., Brundage, M.D., & MacLean, A.W (2001). Nonpharmacologic group treatment of insomnia: a preliminary study with cancer survivors. Psychooncology. 10, 389 97. DOI: 10.1002/pon.525. Dimitrov, S., Lange, T., Tieken, S., Fehm, H.L., Born, J. (2004). Sleep associated regulation of T helper 1/T helper 2 cytokine balance in humans. Brain, Behavior, and Immunity, 18, 341 348. Dossus, L., Rinaldi, S., Becker, S., Lukanova, A., Tjonneland, A., Olsen, A., et al. (2010). Obesity, inflammatory markers, and endometrial cancer risk: a prospective cas e control study. Endocrine related cancer, 17, 1007 19. doi: 10.1677/ERC 10 0053. Espie, C. (2002). Conceptual issues in insomnia. Annu. Rev. Psychol. (53) 215 243. Geiss, A., Rohleder, N., Kirschbaum, C., Steinbach, K., Bauer, H.W., & Anton, F. (2005). Predicting the failure of disc surgery by a hypofunctional HPA axis: evidence from a prospective study on patients undergoing disc surgery, Pain 114, 104 117. doi:10.1016/j.pain.2004.12.007 Guess, J., Burch, J.B., Ogoussan, K., Armstead, C.A., Zhang, H., Wagner, S., et al. (2009). Circadian disruption, Per3, and human cytokine secretion. Integrative Cancer Therapies, 8(4):329 36. DOI: 10.1177/1534735409352029 Heffner, K., France, C., Trost, Z., Mei Ng, H., and Pigeon, W.R. (2011). Chronic low back pain, sleep disturbance and Interleukin 6. Clinical Journal of Pain, 27(1): 35 41.

PAGE 79

79 Irwin, M., Clark, C., Kennedy, B., Christian Gillin, J., & Ziegler, M. (2003). Nocturnal catecholamines and immune function in insomniacs, depressed patients, and control subjec ts. Brain, behavior, and immunity. 17(5): 365 72. doi:10.1016/S0889 1591(03)00031 X Janda, M., Gebski, V., Brand, A., Hogg, R., Jobling, T.W., Land, R., Manolitsas, T., McCartne y, A., Nascimento, M., Neesham, D., Nicklin, J.L., et al. (2010) Quality of life after total laparoscopic hysterectomy versus total abdominal hysterectomy for stage I endometrial cancer (LACE): a randomised trial. Lancet Oncol 11(8):772 780. DOI:10.1016/S1 470 2045(10)70145 5. Kaireviciute, D., Blann, A.D., Balakrishnan, B., Lane, D.A., Patel, J.V., Uzdavinys, G., et al. (2010). Characterization and validity of inflammatory mediators in the prediction of postoperative atrial fibrillation in coronary artery disease patients. Thromb Haemost, 104, 122 7. doi:10.1160/TH09 12 0837. Kirschbaum, C. & Hellhammer, D.H. (1994). Salivary cortisol in psychoneruoendocrine research: recent developments and applications. Psychoneuroendocrinology, 19, 313 333. doi:10.1016/0306 4530(94)90013 2 Kline, R. B. (2005). Principles and practice of structural equation modeling. New York, NY: The Guilford Press. Kwekkeboom, K.L., Abbott Anderson, K., & Wanta, B. (2010). Feasibility of a Patient Controlled Cognitive Behavioral Intervention for Pain, Fatigue, and Sleep Disturbance in Cancer. Oncol Nurs Forum. 37(3): E151 E159. doi:10.1188/10.ONF.E151 E159. Lambrou, N.C., Gomez Marin, O., Mirhashemi, R., et al. (2004). Optimal surgical cytoreduction in patients with stage III and stage IV endometrial carcinoma: A study of morbidity and survival. Gynecol Oncol, 93, 653 658. doi:10.1016/j.ygyno.2004.03.015 Lei, M. & Lomax, R.G (2005). The Effect of Va rying Degrees of Nonnormality in Structural Equation Modeling. Structural Equation Modeling: A Multidisciplinary Journal. 12(1): 1 27. DOI: 10.1207/s15328007sem1201_1. Lutgendorf S K Anderson B Ullrich P Johnsen E L Buller R E Sood A K et al (2001) Quality of life and mood in women with gynecologic cancer: a 1 year prospective study. Cancer 94:131 140. DOI: 10.1002/cncr.10155 Lutgendorf, S.K., Cole, S., Costanzo, E., Bradley, S., Coffin, J., Jabbari, S., et al. (2003). Stress related m ediators stimulate vascular endothelial growth factor secretion by two ovarian cancer cell lines. Clinical Cancer Research, 9, 4514 4521.

PAGE 80

80 Lutgendorf, S.K., Johnsen, E.L., Cooper, B., Anderson, B., Sorosky, J.I., Buller, R.E., et al. (2002). Vascular endoth elial growth factor and social support in patients with ovarian carcinoma. Cancer, 95, 808 815. Lutgendorf S K Weinrib A Z Penedo F et al. ( 2008 ) Interleukin 6, cortisol, and depressive symptoms in ovarian cancer patients J Clin Oncol 26 : 4820 4827 DOI: 10.1200/JCO.2007.14.1978 Maccio, A. & Madeddu, C. (2012). Inflammation and Ovarian Cancer. Cytokine, 58:133 147. doi:10.1016/j.cyto.2012.01.015 McGuire, L., Heffner, K., Glaser, R., Needleman, B., Malarkey, W. B., Dickinson, S., et al. (2006). Pain and wound healing in surgical patients. Annals of Behavioral Medicine, 3,165 172. Men ger, M.D. & Vollmar, B. (2004). Surgical trauma: hyperinflammation versus immunosuppression? Langenbecks Archives of Surgery, 389, 475 84. Doi: Mills, P.J., Parker B., Dimsdale, J.E., Sadler, G.R., Ancoli Israel, S. (2005). The relationship between fatigue and quality of life and inflammation during anthracycline based chemotherapy in breast cancer. Biol Psychology, 69, 85 96. doi:10.1016/j.biopsycho.2004.11.007. Mitrani, V.B., Feaster, D.J., Weiss Laxer, N.S. & McCabe, B.E. (2011). Malaise, motivation and motherhood: predictors of engagement in behavioral interventions from a randomized controlled trial of HIV + women in drug abuse recovery. AIDS Behav (15) 396 406. DOI: 10.1007/s10461 010 9714 z. Mourits, M.J ., Bijen, C.B., Arts, H.J., ter Brugge, H.G., van der Sijde, R., Paulsen, L., Wijma, J., Bongers, M.Y., Post, W.J., van der Zee, A.G., & de Bock, G.H. (2010). Safety of laparoscopy versus laparotomy in early stage endometrial cancer: a randomised trial. La ncet Oncology. 11(8):763 771. DOI:10.1016/S1470 2045(10)70143 1. National Cancer Institute (NCI). (2009). Common Terminology Criteria for Adverse Events (CTCAE) v4.0 http://evs.nci.nih.gov/ftp1 /CTCAE/About.html Pearson, S., Maddern, G.J., & Fitridge, R. (2005). The role of pre operative state anxiety in the determination of intra operative neuroendocrine responses and recovery. British Journal of Health Psychology, 10, 299 310. Pigeon W R P erlis M L Swan J Costescu S Matteson Rusby S Walton J et al. (2008). CBT for Insomnia is associated with decreased fatigue and diminished IL 6 levels in chr onic pain patients. Sleep, 31S: A253.

PAGE 81

81 Portenoy, R.K., Thaler, H.T., Kornblith, A.B. et al. (1994). Symptom prevalence, characteristics, and distress in a cancer population. Qual Life Res. 3,183 189. DOI: 10.1007/BF00435383. Pruessner, J.C., Kirschbaum, C., Meinlschmid, G., & Hellhammer, D.H. (2003). Two formulas for computation of the ar ea under the curve represent measures of total hormone concentration versus time dependent change. Psychoneuroendocrinology, 28, 916 931. Rich, T., Innominato, P.F., Boerner, J., Mormont, M.C., Iacobelli, S., et al., (2005). Elevated serum cytokines corr elated with altered behavior, serum cortisol rhythm, and dampened 24 hour rest activity patterns in patients with metastatic colorectal cancer. Clin. Cancer Res. 11:1757 64. doi: 10.1158/1078 0432.CCR 04 2000 Reinders, M. E., Sho, M., Izawa, A., Wang, P., Mukhopadhyay, D., et al. (2003). Proinflammatory functions of vascular endothelial growth factor in alloimmunity. J. Clin. Invest. 112, 1655 1665. doi:10.1172/JCI200317712. Rodrigue, J. R. Pearman, T. P. and Moreb, J. (1999). Morbidity and mortality fo llowing bone marrow transplantation: Predictive utility of pre HSCT affective functioning, compliance, and social support stability. International Journal of Behavioral Medicine 6 pp. 241 254. doi:10.1207/s15327558ijbm0603_3 Rummans, T.A., Frost, M., Suman, V.J., et al. (1998). Quality of life and pain in patients with recurrent breast and gynecologic cancer. Psychosomatics. 39, 437 445. Sato, N., Koeda, K., Ikeda, K., Kimura, Y., Aoki, K., Iwaya, T., Akiyama, Y., Isida,K. Saito, K., and Endo, S. (2002). Randomized Study of the Preoperative Corticosteroid Administration on the Postoperative Morbidity and Cytokine Response in Patients Undergoing Surgery for Esophageal Cancer. Annals of Surgery 236(2): 184 190. DOI: 10.1097/01.SLA.0000022025.67985.53 Savard, J., Simard S., Ivers, H., et al. (2005). Randomized study on the efficacy of cognitive behavioral therapy for insomnia secondary to breast cancer, part II: Immunologic effects. J Clin Oncology. 23, 6097 610. DOI: 10.1200/JCO.2005.12.513. Schnoll, R.A., Fang, C.Y., & Manne, S.L. (2004). The application of SEM to behavioral research in oncology: Past accomplishments and future opportunities. Structural Equation Modeling. 11, 583 614. DOI: 10.1207/s15328007sem1104_5. Sephton, S.E., Sapolsky, R.M., Kraemer, H.C., & Spiegel, D. (2000). Diurnal cortisol rhyth as a predictor of breast cancer survival. Journal of National Cancer Institute 92, 994 1000.

PAGE 82

82 Sephton, S. & Spiegel, D. (2003). Circadian disruption in cancer: A neuroendocrine immune pathway from stress to disease? Brain, Behavior, and Immunity, 17, 321 328. Shaarawy M. & El Sharkawy S A. (2001). Biomarkers of intrinsic angiogenic and anti angiogenic activi ty in patients with endometrial hyperplasia and endometrial cancer. Acta Oncol. 40(4):513 518. Shapiro, S.L., Bootzin, R.R., Figueredo, A.J., Lopez, A.M., & Schwartz, G.E. (2003). The efficacy of mindfulness based stress reduction in the treatment of slee p disturbance in women with breast cancer: An exploratory study. Journal of Psychosomatic Research, 54, 85 91. Slater, M., Cooper, M. & Murphy, C.R. (2006). Human growth hormone and interleukin 6 are upregulated in endometriosis and endometrioid adenocarc inoma, Acta Histochem.108, 13 18. doi:10.1016/j.acthis.2006.01.004 Sorosky, J.I. Endometrial cancer, Obstet. Gynecol. 111 (2008 Feb), pp. 436 447. Spiegel, D. & Giese Davis, J. (2003). Depression and cancer: mechanisms and disease progression. Biol Psych iatry 2003;54:269 82. doi:10.1016/S0006 3223(03)00566 3 Spiegel, K., Leproult, R., & Van Cauter, E. (1999). Impact of sleep debt on metabolic and endocrine function. Lancet, 23, 1435 1439. Tennant, F. & Hermann, L. (2002). Normalization of serum cortisol concentration with opioid treatment of severe chronic pain, Pain Med. 3, 132 134. DOI: 10.1046/j.1526 4637.2002.02019.x. Thaker, P. H., Lutgendorf, S. K., & Sood, A. K. (2007). The neuroendocrine impact of chronic stress on cancer. Cell Cycle., 6, 430 43 3. Theobald, D.E. (2004). Cancer pain, fatigue, distress, and insomnia in cancer patients. Clinical Cornerstone 4, 6( suppl 1D), S15 S21. doi:10.1016/S1098 3597(05)80003 1 Thom pson, B. (2000). Chapter 8: Ten commandments of structural equation modeling. In L. G. Grimm, & P. R. Yarnold (Eds.), Reading and understanding more multivariate statistics, pp. 261 283. Washington, DC: APA. Thompson, D.S. & Shear, M.K. (1998). Psychiatri c disorders and gynecological oncology: A review of the literature. Gen Hosp Psychiatry, 20, 241 247. doi:10.1016/S0163 8343(98)00030 9

PAGE 83

83 Thornton, L.M., Andersen, B.L., & Blakely, W.P. (2010). The pain, depression, and fatigue symptom cluster in advanced breast cancer: covariation with the hypothalamic pituitary adrenal axis and the sympathetic nervous system. Health Psychology. 293, 333 337. doi: 10.1037/a0018836 Tjemsland, L., Soreide, J.A., Matre, R., & Malt, U.F. (1997). Pre operative [correction of P roperative] psychological variables predict immunological status in patients with operable breast cancer. Psychooncology 6, 311 320. Tozzi, R., Malur, S., Koehler, C., & Schneider, A. (2005). Analysis of morbidity in patients with endometrial cancer: Is there a commitment to offer laparoscopy? Gynecologic Oncology 97, 4 9. Vaknin, Z., Ben Ami, I., Schneider, D., Pansky, M., & Halperin, R. (2009). A comparison of perioperative morbidity, perioperative mortality, and disease specific survival in elderly w omen (>or = 70 years) versus younger women (<70 years) with endometrioid endometrial cancer, Int J Gynecol Cancer 19, 879 883. doi: 10.1111/IGC.0b013e3181a73a12 Vedhara, K., Miles, J., Bennett, P., Plummer, S., Tallon, D., Brooks, E., et al. (2003). An i nvestigation into the relationship between salivary cortisol, stress, anxiety, and depression, Biological Psychology 62, 89 96. doi:10.1016/S0301 0511(02)00128 X. Vedhara, K., Stra, J. T., Miles, J. N., Sanderman, R., & Ranchor, A. V. (2006). Psychosocia l factors associated with indices of cortisol production in women with breast cancer and controls. Psychoneuroendocrinology 31(3): 299 311. doi:10.1016/j.psyneuen.2005.08.00 VEGF Quantkine Kit, R & D Diagnostics, Minneapolis, MN. Vgontzas, A.N., Chrouso s, G.P. (2002). Sleep, the hypothalamic pituitary adrenal axis, and cytokines: multiple interactions and disturbances in sleep disorders. Endocrinology and Metabolism Clinics of North America 31, 15 36. DOI: 10.1016/S0889 8529%2801%2900005 6 Vitek, L., R osenzweig, M.Q., & Stollings, S. (2007). Distress in Patients With Cancer:Definition, Assessment, and Suggested Interventions. Clinical Journal of Oncology Nursing. 11(3), 413 418. doi: 0.1188/07.CJON.413 418 Wallace, A.E., Gibson, D.A., Saunders, P.T. & Jabbour, H.N. (2010). Inflammatory events in endometrial adenocarcinoma. J Endocrinolology 206, 141 157. DOI: 10.1677/JOE 10 0072. Williams, J.B. (1988). A structured interview guide for the Hamilton Depression Rating Scale. Archives of General Psychiat ry, 45, 742 747.

PAGE 84

84 BIOGRAPHICAL SKETCH Seema M Patidar attended the University of North Carolina where she e arned a Bachelor of Science in psychology, with a minor in c hemistry in 2006. After graduating, Seema worked at Duke University Medical Center on a study examining the prognostic benefits of exercise and anti depressant therapy in patients with heart disease.