This item is only available as the following downloads:
1 CORTISOL DYSREGULATION AND MOOD DISTURBANCE IN WOMEN UNDERGOING SURGERY FOR SUSPECTED ENDOMETRIAL CANCER By TIMOTHY S. SANNES A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFIL LMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Timothy S. Sannes
3 To my father, whose patience, support and mentorship have made shaped me into the individual I am today a nd for whom I credit with instilling the drive in me to pursue an advance degree.
4 ACKNOWLEDGMENTS First and foremost, I would like to acknowledge my mentor, Deidre Pereira. Without her constant guidance, attention to detail and supportive mentorship, t his project would not have been possible. She has always been there throughout my graduate training, ensuring that my work i s the highest quality possible. She also always makes time to laugh, which I have found particularly helpful throughout this graduat e school adventure In addition, I would like to thank my co mentors, Naser Chegini and Michael Marsiske. Naser provided unrelenting positivity and enthusiasm for the scientific process. Despite rarely having space for me to sit in his office with the stac ks of scientific papers adorning his furniture, he always made time to share his scientific thinking and honesty about his opinions surrounding academia. Michael has nurtured my growth as a scientist since being a student in courses and serving as his teac hing assistant. Without his considerate teaching style and patience, I would not have learned the analytic techniques applied throughout this project. I would also like to briefly acknowledge a number of student colleagues, each of who have encouraged me and pushed me to excel in my graduate work and beyond. This includes, Daniel Kay, Joe Dzierzewski, Stephanie Garey Smith Burel Goodin and Elizabeth Spehalski. Furthermore, I express my sincerest gratitude to the members of the Gynecologic Oncology team at the University of Florida that helped to make this research possible; specifically, Linda Morgan and Insley Baldwin for their upbeat spirit and constant support. I would also like to thank all of the women who volunteered for this study without them; th is work would not have been possible.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 Epidemiology and Treatment of Endometrial Cancer ................................ ............. 14 Psychological Functioning among Women with Endometri al Cancer ..................... 15 Psychoneuroimmunology in Cancer ................................ ................................ ....... 16 Cortisol: Statistical Approaches and Calculating Variability ................................ .... 18 Biobehavioral Relationships with Cortisol ................................ ............................... 21 Coping as a Potential Buffer against Effects of Stress/Mood Disturbances on Cortisol ................................ ................................ ................................ ................ 23 Comparing Methods of Operationalizing Cortisol Dysregulation ............................. 25 Purpose of the Current Study ................................ ................................ ................. 27 Specific Aims and Hypotheses ................................ ................................ ............... 28 Aim 1: ................................ ................................ ................................ ............... 28 Aim 2: ................................ ................................ ................................ ............... 28 Aim 3: ................................ ................................ ................................ ............... 28 Secondary/Exploratory Aim 4: ................................ ................................ .......... 29 2 METHODS ................................ ................................ ................................ .............. 30 Participants ................................ ................................ ................................ ............. 30 Procedures ................................ ................................ ................................ ............. 31 Psychosocial Assessment ................................ ................................ ...................... 31 Beck Sc ale for Suicide Ideation (BSS; Appendix A; Beck & Steer, 1991) ........ 31 Psychotic Screening Module of the Structured Clinical Interview for DSM IV for Non clinical Populations (SCID NP; Appendix B; First, Spitzer, Williams, & Gibbon, 1995) ................................ ................................ ............. 32 Brief COPE (Appendix C; Carver, 1997) ................................ .......................... 33 Recent Health Behaviors (RHB; Appendix D; Pereira, unpublished) Questionnaire ................................ ................................ ................................ 34 Structured Interview Guide for the Hamilton Anxiety and Depression Scale (SIGH AD; Appendix E; Williams, 1988) ................................ ...................... 34 Biomedical Measures ................................ ................................ .............................. 35
6 Biobehavioral Control Variables ................................ ................................ ....... 35 Prescription medications ................................ ................................ ............ 35 Disease status and tumor subtype ................................ ............................. 36 Charlson Comorbidity Index (Charlson, Pompei, Ales, & MacKenzie, 1987) ................................ ................................ ................................ ...... 36 Salivary Cortisol Collection ................................ ................................ ............... 37 Statistical Procedures ................................ ................................ ............................. 38 Data Preparation ................................ ................................ .............................. 38 Cortisol Slope ................................ ................................ ............................. 38 Area under the Curve with Respect to Increase (AUCi) ............................. 38 Cortisol variability via Multi level Modeling ................................ ................. 39 Morning to Evening Ratio ................................ ................................ .......... 39 Descriptive Statistics ................................ ................................ .................. 40 Specific Aims and Hypotheses ................................ ................................ ......... 40 Aim 1: ................................ ................................ ................................ ......... 40 Aim 2: ................................ ................................ ................................ ......... 42 Aim 3: ................................ ................................ ................................ ......... 43 Secondary/Exploratory Aim 4: ................................ ................................ .... 45 Power and Sample Size ................................ ................................ .......................... 45 3 RESULTS ................................ ................................ ................................ ............... 47 Participant Characteristics ................................ ................................ ...................... 47 Depressive and Anxious Symptomatology ................................ .............................. 47 Cortisol values ................................ ................................ ................................ ........ 48 Results: AIM 1: ................................ ................................ ................................ ....... 48 Results: Aim 2: ................................ ................................ ................................ ........ 51 Linear Time Trend (Cortisol Slope) ................................ ................................ .. 51 Depressive/Anxious Symptoms and Cortisol Slope ................................ .......... 52 Resul ts: Aim 3: ................................ ................................ ................................ ........ 53 Descriptive Statistics for Acceptance and Positive Reframing Items ................ 53 Relations among Acceptance, Positive Reframi ng and Cortisol Variability ...... 53 Positive Reframing and Acceptance as Moderators of Anxiety/Depressive ..... 53 Symptoms and Cortisol Va riability ................................ ................................ .... 53 Results: Secondary/Exploratory Aim 4: ................................ ................................ ... 55 Cortisol Slope ................................ ................................ ................................ ... 55 Area Under the Curve with Respect to Increase (AUCi) ................................ ... 56 Morning to evening Ratio of Cortisol ................................ ................................ 56 Comparison among Cortiso l Calculations ................................ ......................... 56 4 DISCUSSION ................................ ................................ ................................ ......... 73 Biobehavioral Control Variables and Intraindividual Cortisol Variability and Cortisol Slope ................................ ................................ ................................ ...... 73 Depressive Symptoms, Anxious Symptoms, and Intraindividual Cortisol Variability ................................ ................................ ................................ ............. 74 Depressive Symptoms, Anxious Symptoms, and C ortisol Slope ............................ 76
7 Examining Unique Contributions of Anxiety and Depression to Intraindividual Cortisol Variability and Cortisol Slope ................................ ................................ .. 78 Moderating Effects of Positive Reframing and Acceptance ................................ .... 79 Comparisons among Calculations of Cortisol Output ................................ ............. 82 Clinical Implications of the Current Study ................................ ............................... 84 Study Limitations ................................ ................................ ................................ .... 86 Future Directions ................................ ................................ ................................ .... 90 Conclusions ................................ ................................ ................................ ............ 92 APPENDIX A BECK SCALE FOR SUIDICAL IDEATION ................................ ............................. 94 B BRIEF COPE ................................ ................................ ................................ .......... 97 C PSYCHOTIC SCREENING MODULE OF THE STRUCTURED CLINICAL INTERVIEW FOR DSM IV FOR NON CLINICAL POPULATIONS ......................... 99 D RECENT HEALTH BEHAVIORS QUESTIONNAIRE (RHB) ................................ 105 E STRUCTURED INTERVIEW GUIDE FOR THE HAMILTON DEPRESSION AND ANXIETY SCALES ................................ ................................ ....................... 106 LIST OF REFERENCES ................................ ................................ ............................. 115 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 126
8 LIST OF TABLES Table page 3 1 Comparison of Participants Included and Excluded in Data A nalyses. ............... 57 3 2 Predicting Intraindividual Cortisol Variability from Depressive Symptoms. ......... 58 3 3 Predicting Intraindividual Cort isol Variability from Anxiety Symptoms. ............... 58 3 4 Predicting Intraindividual Cortisol Variability from Anxiety Symptoms; Controlling for Concurrent Depressive Symptoms. ................................ ............. 58 3 5 Predicting Intraindividual Cortisol Variability from Depressive Symptoms; Controlling for Concurrent Anxiety Symptoms. ................................ ................... 58 3 6 Unconditional Growth Models of Cortisol Values. ................................ ............... 59 3 7 Relationships Between Potential Covariates and Cortisol Slope using Multilevel Modeling ................................ ................................ ............................. 60 3 8 Predicting Initial Cortisol Value from Anxiety and Depressive Symptoms. ......... 61 3 9 Predicting Cortisol Slope from Anxiety and Depressive Symptoms. ................... 61 3 10 Predicting Initial Cortisol Value from Iso Anxiety and Iso Depressive Symptoms. ................................ ................................ ................................ .......... 61 3 11 Predicting Linear Cortisol Slope from Iso Anxiety and Iso Depressive Symptoms. ................................ ................................ ................................ .......... 62 3 12 Examining Positive Reframing as a Potential Moderator of a Relationship between Depressive Symptoms and Cortisol Variability. ................................ .... 63 3 13 Examining Acceptance as a Potential Moderator of a Relationship between Depressive Symptoms and Cortisol Variability. ................................ .................. 63 3 14 Examining Acceptance as a Potential Moderator of a Re lationship between Anxiety Symptoms and Cortisol Variability ................................ ........................ 63 3 15 Examining Positive Reframing as a Potential Moderator of a Relationship between Anxiety Symptoms and Cortisol Variability. ................................ .......... 64 3 16 Examining Positive Reframing as a Potential Moderator of a Relationship between Depressive Symptoms and Cortisol Slope. ................................ .......... 65 3 17 Examining Acceptance as a Potential Moderator of a Relationship between Depressive Symptoms and Cortisol Slope. ................................ ........................ 66
9 3 18 Examining Acceptance as a Potential Moderator of a Relationship between Anxiety Symptoms and Cortisol Slope. ................................ ............................... 67 3 19 Examining Positive Reframing as a Potential Moderator of a Relationship between Anxiety Symptoms and Cortisol Slope. ................................ ................ 68 3 20 Pearson Correlations among Methods of Operationalizing Cortisol Dysregulation. ................................ ................................ ................................ ..... 69
10 LIST OF FIGURES Figure page 2 1 Calculation for Area under the Curve with Respect to Increase (AUCi) .............. 46 3 1 Average Cortisol Values Across the Four Timepoints of Collection. ................... 70 3 2 Cortisol Variability among Participants with the Highest Depressive (HD) Symptom Scores. Cortisol values and corresponding regression lines are displayed. ................................ ................................ ................................ ........... 71 3 3 Cortisol Variability among Participants with the Lowest Depressive (LD) Symptom Scores. Cortisol values and corresponding regression lines are displayed.. ................................ ................................ ................................ .......... 72
11 LIST OF ABBREVIATION S ACTH Adrenocorticotropic Hormone AIC AUCi Area under the Curve with Respect to Increase ANS Autonomic Nervous System BIC CAR Cortisol Awaken i ng Response CRF Corticotrophin Releasing Factor HLM Hierarchical Linear Modeling HPA Hypo thalamic Pitu i tary Adrenal Axis ISD Intraindividual standard deviation PCA Principal Component Analysis PNI Psychoneuroimmunology SE Standard Error TAH BSO Total Abdominal Hysterectomy and Bilateral S alpingo oopherectomy 2LL Negative 2 log likelihood
12 Abs tract o f Dissertation Presented to t he Graduate School of t he University o f Florida i n Partial Fulfillment of t he Requirements for t he Degree o f Doctor o f Philosophy CORTISOL DYSREGULATION AND MOOD DISTURBANCE IN WOMEN UNDERGOING SURGERY FOR SUSPECTED ENDOMETRIAL CANCER By Timothy S Sannes August 2013 Chair: Deidre B. Pereira Major: Psychology Endometrial cancer, the most common gynecologic cancer in the United States, represents a substantial public health burden. In spite of this very little research has examined the biopsychosocial factors associated with quality of life and health outcomes in endometrial cancer. Cortisol is a potent anti inflammatory stress hormone that may promote tumorigenesis. Multiple methods of operation alizing cortisol dysregulation, including diurnal cortisol rhythm (slope), have been used in the cancer literature. However, little research has examined cortisol variability as an indicator of cortisol dysregulation in individuals with cancer. Recent pu blished research in psychiatric populations has suggested that mood disorders, such as depression and bipolar disorder, are associated with a more erratic, or highly variable, production of the stress h ormone, cortisol. This study examine d the relationshi p between mood disturbance (anxious/depressive symptomatology) and (a) cortisol variability and (b) cortisol slope in women undergoing surgery for suspected endometrial cancer. In addition, frequency of use of positive reapprais al and acceptance coping w ere examined as moderators (buffers) of the relationship between mood disturbance and cortisol dysregulation. A secondary, exploratory, aim of the current study examine d
13 how multiple methods of operationalizing cortisol dysregulation (i.e., diurnal slope, variability, total output, and ratio of evening to morning values) compare. Participants were 82 women with nonmetastatic endometrial cancer. Anxious symptoms were not associated with either cortisol slope or intraindividual variability, and depressive s ymptoms were unrelated to cortisol slope. However, after controlling for presence of poorer prognosis cancer subtypes, greater depressive symptoms (excluding symptoms possibly/definitely due to health/treatment factors) in the week preceding surgery were si gnificantly related to greater cortisol int raindividual variability ( = .21 p = .0 48 ). This relationship was not moderated by either greater use of positive reappraisal or acceptance as adaptive coping strategies Overall, t hese results suggest that d epressive symptoms prior to surgery for suspected endometrial cancer are related to greater cortisol intraindividual variability, which is suggestive of more erratic HPA axis arousal and ma y represent a unique conceptualization of HPA axis dysregulation. Future research should examine whether mood symptoms may be associated with compromised health outcomes via erratic HPA axis arousal in this population.
14 CHAPTER 1 INTRODUCTION Epidemiology and Treatment of Endometrial Cancer Endometrial c ancer is the m ost common gynecologic cancer in the United States and the fourth most common cancer among women. The American Cancer Society (ACS, 2010) estimated that 43,470 new cases were diagnosed and 7,950 women died from the disease in 2010. It is estimated that 80 % of endometrial carcinomas are of the endometrioid type. More aggressive forms of the disease include papillary serous and clear cell carcinomas, both of which are classified as high grade cancers (Amant et al., 2005) Many of the risk factors of endometr ial cancer are related to exposure to unopposed estrogen, including early age at menarche, late age at menopause and the use of unopposed estrogen replacement therapy. Other risk factors include family history of certain cancers obesity, hypertension an d diabetes (Purdie & Green, 2001) Conversely, combined oral contraceptive use, smoking, low fat diets and physical exercise decrease the risk for endometrial cancer; likely via their reduction in unopposed estrogen exposure. While many cases of endometr ioid adenocarcinomas are caught at an early stage and have a very favorable prognosis and survival rate, the 5 year survival rate for stages III and IV endometrial cancers are 58% and 17% respectively (American Cancer Society [ACS] 2010). Compared to th e mortality rates of breast and ovarian cancers, which have decreased steadily over the past several decades, the mortality rate of endometrial cancer has remained stable since 1991 (A CS 2010). In summary, these statistics underscore the fact that endome trial cancer remains a significant public health burden.
15 Most women with suspected endometrial cancer undergo total abdominal hysterectomy and bilateral salpingo oopherectomy (TAH BSO), which involves the removal of the uterus and surrounding organs, incl uding the cervix, ovaries and fallopian tubes, depending on the extent of the disease. With aggressive surgical treatment, 83% of all women diagnosed with endometrial cancer will survive at least 5 years (A CS, 2010). Once the uterus is removed, the depth o f myometrial invasion can be assessed, which is critical for determining cancer stage in women without evidence of regional/advanced disease. Radiation therapy and/or adjuvant chemotherapy are generally recommended following surgery for women with regiona l/advanced endometrioid cancer and women with non endometrioid morphologies. As the endometrioid type is the more common type of endometrial carcinoma, it is the focus of the remainder of the following review and the current study. Psychological Functionin g a mong Women with Endometrial Cancer Women faced with the diagnosis and corresponding treatment for gynecologic cancer may experience significant psychological distress and/or mood disturbance. In a review of the literature on psychiatric disorders in gyn ecologic malignancies, greater prevalence of depression, anxiety and adjustment disorder were noted compared to the general population. Similarly, in this review, mood disturbance was observed to worsen over the course of cancer treatment (Thompson & Shear 1998). More recently, i t has been estimated that the prevalence of depression in gynecologic cancer patients is 12% 23% (Massie, 2004) Other studies have demonstrated significant psychological distress in ovarian (Norton et al., 2004) and cervical cance r (Cull et al., 1993) specifically Whi le poorer health related quality of life has been observed in endometrial cancer patients compared to age matched, healthy controls (Klee & Machin, 2001),
16 recent reports have suggested that quality of life may not be affected in gynecologic cancer. Specifically, in 57 gynecologic cancer patients undergoing treatment over the one year study period anxiety and depression ratings remained low and did not change over the year of treatment ( Y avas et al., 2012). However, t here was not a control group in this report and the sample was comprised of various gynecologic cancers. In endometrial cancer specifically, recent published work suggested that while sexual dysfunction r emained a significant concern over the first five y ears following treatment, minimal levels of depression were reported ( Onujiogu et al., 2011). Indeed, the heterogeneity of reviewed studies as well as the paucity of literaure examining psychological functioning in this population was has been recognized (Thompson & Shear, 1998), and could account for a predictable inconsistency in study outcomes. Despite these apparent conflicting findings examining quality of life in endometrial cancer, no published research has examined mood disturbance in endometrial cancer at the time of diagnosis specifically. Taken together, these studies suggest that psychological distress may occur in women with gynecological cancers. This is significant to the extent that potential psychological distress may be associated w ith neuroendocrine functioning and immunity associated with tumorigenesis. This theoretical model is described below. Psychoneuroimmunology in Cancer The biobehavioral model of tumorigenesis provides a testable model for examining pathways by which stress may impact health outcomes in cancer populations (Antoni, et al., 2006) This model (also referred to as the psychoneuroimmunologic [PNI] model) integrates theoretical concepts and research findings from psychology, neuroscience, endocrinology, and immuno logy. According to this model, stressors activate two
17 physiological systems: the autonomic nervous system and the hypothalamic pituitary adrenal (HPA) axis, (for review see Antoni et al 2006). These systems are thought to correspond to the fight or flig ht and defeat withdrawal responses, respectively. The autonomic nervous system releases catecholamines, such as epinephrine and norepinephrine which exert a powerful influence on the tumor microenvironment by upregulating tumor promoting, inflammatory fa ctors such as matrix metalloproteinases (Lutgendorf et al., 2008; Lutgendorf et al., 2005) and interluekin 6 (Lutgendorf, Anderson, Sorosky, Buller, & Lubaroff, 2000). These catecholaminergic effects may be potentiated by molecules released by the HPA axi s (Nakane et al., 1990) Once this stress system is activated, the hypothalamus releases corticotrophin releasing factor (CRF), which in turn stimulates the pituitary to release adrenocorticotropic hormone (ACTH), which in turn stimulates the adrenal medu lla to release a host of hormones, including cortisol. Cortisol, a potent immunosuppressive agent, exert s downstream effects on tumor biology by inhibiting immune cells (e.g., inhibiting lymphocyte proliferation and increasing lymphocyte apoptosis) and imp airing cell mediated immunity by decreasing natural killer cell and T lymphocyte cytotoxicity (Reiche, Nunes, & Morimoto, 2004) and inhibition of apoptosis (cell death; Wu et al., 2004) These data highlight how cortisol significantly suppresses cell mediated immunity, resulting in downstream effects on the tumor microenvironment. As such, examining dysfunction of the HPA axis via the analysis of cortisol rhythms provides a model by whic h psychological factors may influence health outcomes in endometrial cancer.
18 Cortisol: Statistical Approaches and Calculating Variability Cortisol production corresponds to a circadian rhythm, with a peak in the early morning upon awakening (referred to a values throughout the day, and a slight increase in cortisol before bedtime ( Krieger 1979). Although there are a number of approaches to modeling this diurnal pattern in cortisol (Adam & Kumari, 2009) it r emains difficult to compare cortisol as an outcome variable across studies when different calculations are presented between published work Indeed, a n initial meeting on the study of HPA axis dysregulation identified the lack of a firm consensus on the mo st appropriate method for the analysis of cortisol (Stewart & Seeman, 2000) Recent work highlighted that up 13 different calculations of cortisol can been applied to the same dataset (Fekedulegn et al., 2007) highlighting the myriad of cortisol calculati ons presented in the published literature One commonly used approach has been to calculate cortisol slope across multiple days of cortisol collection (Kraemer et al., 2006; Sephton et al., 2000; Simonelli, Fowler, Maxwell, & Andersen, 2008) Using this m ethod, values of cortisol throughout the day are regressed on time, and the resulting regression weight (or slope coefficient; ) is assigned to each individual. With this approach, a flattened cortisol rhythm throughout the day (represented by a less nega tive weight ) corresponds to a dysregulation of the HPA axis in its ability to respond to continued endogenous cortisol production. Such dysregulation is observed in (Laudat et al., 1988) as well as under condi tions of chronic psychological stress (Gregory, Cohen & Kim, 2002) In sum, while many methods exist for quantifying the diurnal rhythm observed in cortisol output throughout the day, cortisol slope remains one of the most widely applied
19 There is a ric h literature linking dysregulated diurnal cortisol slope with disturbances in mood. In patients with major depressive disorder, a flattening or blunting of this cortisol slope is often observed, which suggests that the body becomes less responsive to the r epeated exposure of endogenous cortisol (Miller, Cohen, & Ritchey, 2002) By the body responding to a lesser degree to continued endogenous cortisol, it is hypothesized that cortisol production continues throughout the day, resulting in higher cortisol le vels at the end of the day. H igher evening cortisol, resulting in a flatter cortisol slope throughout the day, has been observed in those suffering from chronic stress (Adam et al., 2006), patients with psychotic major depression ( Belanoff Kalehzan, Sund Ficek, & Schatzberg, 2001) and depressed patients with coronary artery disease ( Bhattacharyya Molloy, & Steptoe, 2008). In addition, those who repress ed their emotions and were classified as highly anxious display ed flatter cortisol slopes (Giese Davi s et al., 2006) Taken together, these data suggest that disturbances in mood (depression and anxiety) are related to a flatter cortisol slope throughout the day. More recently, investigators have focused on capturing individual differences in the cortiso l response through multilevel modeling techniques (Adam & Kumari, 2009; Hruschka, Kohrt, & Worthman, 2005) including hierarchical linear modeling (HLM) (Singer & Willet, 2003) which allow for the estimation of individual differences in cortisol slope and overall output Moreover, these techniques remain flexible in the presence of incomplete data, such that individual trajectories are still estimated in the presence of missing data points. This is a common occurrence in cortisol sampling, as individuals may be asked to collect repeated salivary samples throughout the day and thus may miss collections (Adam & Kumari, 2009) These models will still estimate an individual
20 cortisol slope (for a day or for a person), even if one data point is missing; an adva ntage over ordinary least squares regression analyses (reviewed above in the assigning of a weight to each individual ) that may employ list wise deletion, thereby eliminating valuable data. Despite this trend in the research to increasingly apply flexi ble models such as HLM, very little work has incorporated measures of intra individual variability in examining HPA axis functioning. Using these methods, one may calculate cortisol intraindividual variability or intraindividual standard deviation (ISD ; Ma cDonald, Hultsch, & Bunce, 2006) (Peeters, Nicolson, & Berkhof, 2004) cortisol pulsatility (Young, Abelson, & Lightman, 2004) and approximate entropy in cortisol production (Posener et al., 2004) Recent reports suggest that, for men but not women older age is related to greater variability in cortisol rhythm (Almeida, Piazza, & Stawski, 2009), suggesting an interaction of age and gender in cortisol variability However, it should be noted that this investigation examined only variability around the cortisol awakening response (CAR), which describes the initial spike seen in cortisol upon waking (Fries, Dettenborn, & Kirschbaum, 2009) possibly providing a variability estimate distinct from variability estimates across an entire day. Other predictors of change in cortis ol variability suggest that variability decreases during the first years of life (Tollenaar et al., 2010). These emerging studies raise the possibility that operationalizing cortisol output as intraindividual cortisol variability may provide information about HPA axis functioning
21 that is not captured through other methods of operationalizing cortisol regulation. Furthermore, the possibility exists that psychosocial factors may have unique relationships with intraindividual cortisol variability. Consiste nt with this, research is beginning to identify how mood factors may be associated with intraindividual cortisol variability. For instance, i n the clinical mood disorder literature Peeters and colleagues found greater cortisol variability in individuals with major depressive disorder compared to controls (Peeters et al. 2004). Similarly, Posener and colleagues found the men with major depressive disorder had significantly more erratic cortisol production compared to healthy controls; however, there were no significant differences between d epressed men and healthy control s o n cortisol slope or total cortisol output (Posener et al., 2004) Other research has revealed greater cortisol variability among individuals with remitted bipolar disorder compared to controls. More specifically, there was a significant association between greater depressive episode severity and more frequent episode recurrence and greater cortisol variability (Havermans, Nicolson, Berkhof, & Devries, 2010) Altogether, these findings s uggest that mood disturbances are associated with a more erratic pattern of cortisol production, even when similar relationships fail to emerge between mood disturbances and cortisol slope or total cortisol output. Biobehavioral Relationships with Cortis ol In addition to the various methods applied to cortisol data in the extant literature a number of research studies in clinical oncology populations have linked cortisol to important clinical outcomes. cancers, including breast and ovarian cancers. The earliest of this research examined psychosocial relationships with cortisol among breast cancer patients. Porter and colleagues examined cortisol responses to follow up mammography, a stressor
22 associat ed with the threat of cancer diagnosis/recurrence among breast cancer survivors and controls They found that, i n response to follow up mammography, breast cancer survivors experience d higher l evels of cortisol a month prior to their scheduled mammograph y (baseline) and displayed a suppressed cortisol response (i.e., decreased levels of cortisol from baseline to mammography) compared to healthy controls (Porter et al., 2003) Th e s e data suggest that breast cancer patients may experience higher initial lev els of HPA axis output (cortisol) but experience blunted physiological reactions to cancer related stressors. Furthermore, Giese Davis and colleagues found that breast cancer patients who suppress emotions or display high anxiety display a flatter cortiso l rhythm (slope ) ( Giese Davis et al., 2006) compared to their counterparts highlighting that anxiety and emotional regulation strategies are related to cor t isol dysregulation. In addition, women with more advanced breast cancer had greater average cortis ol output compared to those with less advanced disease (Abercrombie et al., 2004) suggesting that cortisol may be a marker of disease status Additionally, flatter diurnal cortisol slope an abnormal diurnal rhythm, predict ed earlier mortality in breast cancer patients with metastatic disease (Sephton, Sapolsky, Kraemer, & Spiegel, 2000) implicating HPA axis in cancer survival More recent ly follow up analyses of this group of metastatic breast cancer patients uncovered that a decrease in depressive sym ptoms irrespective of receiving supportive therapy or education, was related to a significantly longer survival time (Giese Davis et al., 2011) suggesting that reductions in depressive symptoms may also be related to longer survival time in breast cancer In sum, cortisol responses to psychological stress may be disrupted in breast cancer patients, and these disruptions cortisol rhythm, in addition
23 to reductions in depressive symptoms, may be related to important clinical outcomes, such as survival. In o varian cancer, relationships have begun to emerge on the underlying pathways through which HPA axis dysregulation may impact health outcomes. For instance, greater afternoon, evening, and total cortisol output are associated with greater circulating interl eukin 6 (IL 6) a proinflammatory cytokine implicated in ovarian carcinogenesis (Lutgendorf et al., 2008 a ). Moreover, higher vegetative depression scores are related to higher evening cortisol levels (Lutgendorf et al., 2008b; Weinrib et al., 2010), as wel l as lower morning to evening cortisol ratio s in ovarian cancer (Weinrib et al., 2010) Thus dysregulated cortisol production represents one mechanism by which psychological factors may be associated with health outcomes in ovarian cancer O verall, t hese data highlight that cortisol is related to important inflammatory pathways that may exert downstream effects on tumor biology in cancers affecting women Coping as a Potential Buffer against Effects of Stress/Mood Disturbances on Corti s ol While stress an d mood disturbances may impact tumo r igenesis through dysregulated cortisol production, adaptive coping strategies may buffer this relationship. The Transactional Model of Stress and Coping (Lazarus & Folkman, 1984) posits that the use of adaptive coping st rategies may buffer the negative effects of stress on mood, behavioral efforts to manage specific external and/or internal demands that are appraised as taxing or exceedin (Lazarus & Folkman, 1984 p. 141 ) When viewed from this definition, the demands of a given situation are evaluated, in addition to the evaluation of the given resources to handle that demand. If
24 the resources are deemed inad equate, then the demands are often considered a given event is likely considered manageable (e.g., not stressful). A large body of work has examined coping in women w ith cancer suggesting that maladaptive coping may hinder psychological adjustment For example, avoidant coping is significantly related to poorer well being and a greater distressed mood in extensively treated gynecologic cancer patients (Costanzo, Lutge ndorf, Rothrock, & Anderson, 2006) suggesting that avoidance may represent an ineffective method of coping. Negative spiritual coping (defined as those reporting a less secure relationship with a higher power) also confers a risk to depression following a diagnosis of gynecologic cancer (Boscaglia, Clarke, Jobling, & Quinn, 2005) A lower reported meaning in life has been shown to partially mediate the relationship between physical sequelae and depressive symptoms in gynecologic cancer survivors (Simonelli et al., 2008) suggesting that feelings of meaning and purpose may be important components of successful coping strategies. Overall, these results suggest that the use of maladaptive coping strategies ( e.g., avoidance) are related to mood disturbance in wo men diagnosed with, and treated for, cancer. However, use of adaptive coping strategies may facilitate positive adjustment and well being among women with cancer. Lutgendorf and colleagues found that women reporting greater positive reframing and acceptan ce coping skills report better functional and emotional well being a year following a diagnosis of gynecologic cancer, even after controlling for the effects of medical treatments on well being (Lutgendorf et al., 2002)
25 relationship between uncertainty and the appraisal of danger and opportunity in gynecological cancer (Mishel & Sorenson, 1991) suggesting that remaining active in one s emotional response to cancer may be ben eficial. Furthermore, the use of a cceptance religio us coping, and distraction were rated by patients as very helpful in the adjustment to a diagnosis of breast or gynecologic cancers (Lauver, Connolly Nelson, & Vang, 2007) further elucidating what patien strategies. C hanges in appraisals, such as increased benefit finding and perceived ability to relax have been shown to mediate the effects of a stress reduction program on levels of serum cortisol among women with breast cancer (Antoni et al., 2009; Phillips et al., 2008) Taken together, these results suggest that adaptive coping may moderate (exacerbate or buffer) the relationship between perceived stress/distress and a variety of outcomes in women with cancer. Compari ng Methods of Operationalizing Cortisol Dysregulation consensus conference examining the differ ent measures of cortisol dysregulation held in 1999 by the MacArthur Research Network on SES and Health, area under the curve a promising measure most notably in establishing a link between cortisol levels and psychological functio ning dysregulation However, none of the methods of operationalizing cortisol dysregulation was determined to be superior to the others (Stewart & Seeman, 2000) Furthermore, very few investigators have examined the relations hip between psychosocial factors and multiple methods of operationalizing cortisol in a single study, making it difficult to determine whether some psychosocial factors may be more closely associated with
26 certain methods of operationalizing cortisol dysreg ulation than other methods. Moreover it is unclear the extent to which various methods of operationalizing cortisol dysregulation may be associated. One report ( Vedhara, Tuinstra, Miles, Sanderman, & Ranchor, 2006 ) identified significant correlations am ong cortisol slope, cortisol area under the curve with respect to increase (AUCi), area under the curve with respect to ground (AUCg) and early morning cortisol peak; however, these relationships were different for the breast cancer patients and the contro l group studied. Both breast cancer patients and control subjects demonstrated a significant negative relationship between AUCi and cortisol slope, suggesting that higher cortisol reactivity was related to a flatter cortisol slope, regardless of health st atus. However, only the breast cancer patients demonstrated a significant positive relationship between morning cortisol peak and AUCg (higher morning cortisol values were associated with greater cortisol output), while only control group subjects demonst rated a significant negative relationship between AUCg and AUCi (greater cortisol output was associated with lower cortisol reactivity; Vedhara et al., 2006) Thus, different methods of operationalizing cortisol dysregulation may contain unique informatio n about HPA axis functioning, and potentially, may offer singular information about how psychosocial factors may be associated with clinical outcomes in cancer. However, n o other published research has examined the extent to which these multiple methods ma y be related to one another in a sample of individuals with a highly prevalent cancer. The examination of this research question in a sample of women with endometrial cancer may advance the field by establishing whether individual methods of operationaliz ing cortisol dysregulation
27 represent unique aspects of cortisol functioning, suggesting that each method is worthy of investigation in its own right. In summary, there remains little consensus on the most appropriate way to characterize diurnal cortisol r hythms in stress research although modeling slope via HLM is becoming increasing popular. HLM allows for the estimation of individual differences and for the calculation of intraindividual cortisol variability. While recent work has examined cortisol v ariability in healthy and psychiatric populations, no research to date has extended this approach to cancer populations. Therefore, applying an estimate of cortisol intraindividual variability to women undergoing surgery for endometrial cancer 1) is timely as an extension of the current literature on various methods of quantifying cortisol production and 2) fills a gap in the literature on stress and cortisol functioning in a population that has been largely ignored from a PNI framework. Purpose of the Curr ent Study Virtually no published research has examined the relationship between psychosocial factors and cortisol in endometrial cancer. Furthermore, no published research has examined relationship between psychosocial factors and cortisol variability in any cancers affecting wom en. This study add s to the current body of research on PNI relationship between cortisol variability and anxious and depressive symptoms in endometrial cancer, 2) to examine the relationship between cortisol slope and anxious and depressive symptoms in endometrial cancer, and 3) to examine adaptive coping strategies as moderators of the relationship between mood and cortisol variability and slope in endom etrial cancer. Across the literature examining psychosocial factors and
28 cortisol dysregulation, there is little consensus on the most appropriate way to statistically operationalize cortisol. Therefore, a secondary, exploratory aim was to examine how the four different operationalizations of cortisol dysfunction compare to one another. Specific Aims and Hypotheses Aim 1: To examine the relationship between mood disturbance (depressive symptomatology, anxious symptomatology) and intraindividual cortisol variability among women with suspected endometrial cancer prior to surgical resection. Hypothesis 1a: Women with greater depressive symptomatology will have greater cortisol variability than those with less depressive symptomatology. Hypothesis 1b: Wome n with greater anxious symptomatology will have greater cortisol variability than those with less anxious symptomatology. Aim 2: To examine the relationship between mood disturbance (depressive symptomatology, anxious symptomatology) and cortisol slope a mong women with suspected endometrial cancer prior to surgical resection. Hypothesis 2a: Women with greater depressive symptomatology will have flatter cortisol slopes than those with less depressive symptomatology. Hypothesis 2b : Women with greater anxiou s symptomatology will have flatter cortisol slopes than those with less anxious symptomatology. Aim 3: To examine moderating effects of adaptive coping (positive reframing, acceptance) on relationships between mood disturbance (depressive symptomatology, anxious symptomatology) and cortisol (intraindividual variability and slope) in women with suspected endometrial cancer prior to surgical resection
29 Hypothesis 3a: Greater use of adaptive coping skills will moderate the relationship between moo d disturba nce and cortisol variability, such that women with mood disturbance and more frequent use of adaptive coping will have less cortisol variability than those with mood disturbance and less frequent use of adaptive coping. Hypothesis 3b: Greater use of adapt ive coping skills will moderate the relationship between mood disturbance and cortisol slope, such that women with mood disturbance and more frequent use of adaptive coping will have steeper cortisol slopes than those with mood disturbance and less frequen t use of adaptive coping. Secondary/Exploratory Aim 4: To compare four operationalizations of diurnal cortisol rhythm/output (i.e., intraindividual variability, slope, AUCi, and morning to evening ratio) as indicators of c ortisol d ysregulation. Hypothes is 4: Greater morning to evening ratio cortisol will be significantly related to cortisol slope, given that the two calculations appear to contain redundant information. However, d ue to the exploratory nature of this aim, other than this hypothesized relat ionship, no specific hypothes e s are offered about which operationalizations of cortisol will be significantly related to one another.
30 CHAPTER 2 METHODS The current study pursued the above specific aims in a sample of women who underwent surgical consult ation at UF & Shands for suspected endometrial cancer between 2003 and 2009. The sample was comprised of women recruited and enrolled into a larger, longitudinal study examining psychoneuroimmunologic relations in women with endometrial cancer immediatel y prior to surgery and six to eight weeks following surgery (ACS Grant #: IRG 01 188 01 and NIH Grant #: R03 CA 117480). Specifically, this study examined the relations between mood disturbance and cortisol dysregulation at the presurgical timepoint only; therefore, the study design was nonexperimental and cross sectional. Mood data were operative visit, typically the day immediately prior to surgery. Cortisol data were collected via saliva sampling a t four times points per day for the three days preceding their pre operative study visit. All procedures were conducted according to the rules and regulations of the Institutional Review Board (IRB) of the University of Florida. This study is IRB approved (approval number 69 2004). Participants Participants were women 18 years of age and older who were scheduled to undergo surgery (TAH BSO) for suspected endometrial cancer All participants were fluent in spoken English. Exclusion criteria included recurr ent endometrial cancer, cancer originating from a site other than the endometrium, neoadjuvant chemotherapy, pre surgical radiation therapy, or documented severe psychiatric illness
31 Procedures Recruitment for the study took place in the Gynecologic Oncolo gy clinic at UF&Shands hospital. The medical team at this clinic included an attending physician, residents and one nurse practitioner. Potentially eligible participants were approached by a study staff member during their initial consultation visit with Gyncologic Oncology for an abnormal endometrial biopsy. If they were interested in study participation, they reviewed and signed a University of Florida Institutional Review Board approved informed consent form Once consented, participants underwent a br ief suicidal and psychosis screening and those screening negative for these mental health issues were scheduled to return to the clinic one week later for their p re operative study visit. They were also provided with a packet of psychosocial questionnair es, and a salivary cortisol collection kit. At this pre surgical appointment, participants returned these materials and completed a psychosocial interview to assess depressive and anxious symptomatology. Participants received $20 compensation for study p articipation. Saliva samples were immediately transported to the laboratory for processing and charts and abstracted history of comorbid medical conditions and tumor ce llular classification, stage, and grade data. Psychosocial Assessment The following measures were administered at the initial consultation visit to determine study eligibility Beck S cale for Suicide Ideation (BSS ; Appendix A; Beck & Steer, 1991) The BSS is a 21 item, self report measure of the presence and severity of suicidal ideation. The BSS has 5 initial screening items. If these items are rated as zero by the
32 participant, they do not co mplete the rest of the measure as suicidal ideation is not pres ent. All item s are rated on a 3 point scale from 0 to 2, and all 21 i tems contribute to the total BSS score with a range of 0 to 42 points The BSS has been shown to have adequate test retest reliability and a Cronbach of .96 in psychiatric populations (Pinninti, Steer, Rissmiller, Nelson, & Beck, 2002) Any participant endors ing suicidal ideation on the BSS underwent more comprehensive psychological evaluation under the licensed supervision of the Principal Investigator (DBP) and was deemed ineligible for further study participation. Psychotic Screening Module of the Structured Clinical Interview for DSM IV for Non clinical Populations (SCID NP ; Appendix B ; First, Spitzer, Williams, & Gibbon, 1995) The SCID NP is a semi structured interview for the diagnosis of DSM IV Axis I disorders in non psychiatric populations. The Psychotic Screening Module screens for the presence of hallucinations and delusions that may herald an underlying psychotic disorder. This screening measure has been used with other patients with major medical illness, such as HIV (Penedo et al., 2003) and demonstrates excellent interrate r reliability (Maffei et al., 1997) As with the BSS, a ny participant endors ing current psychotic symptoms und erwent more comprehensive psychological evaluation under the licensed supervision of the Principal Investigator (DBP) Women with current psychotic symptoms were not eligible for participation in this study. The following measures were completed by parti cipants between their initial consultation and their pre operative study visit.
33 Brief COPE ( Appendix C; Carver, 1997) The Brief COPE was developed to measure various ways individuals cope with stressful events. Originally, the COPE measure was tested in a college student sample and contained 60 items with 15 scales contained within. In this form, the COPE demonstrat ed an acceptable mean Cronbach value of .71 across the various subscales (Carver, Scheier, & Weintraub, 1989) However, it in an effort to reduce participant burden, a shortened version was developed containing 28 items This version was then tested on a population of breast cancer patients (Carver, 1997) and has been applied to other medical populations, such as those living with HIV (Sing h, 2004) and those with severe psychopathology, such as schizophrenia, major depressive disorder and schizophrenia (Meyer, 2001) Reliability estimates are good, ranging from 0.72 to 0.84 for the various subscales (Cooper, Katona, & Livingston, 2008) Whil e these studies compiled various items together to create these subscales, the (Carver, 2007) distinguishes the 14 scales based on pairs of items. These include: self distraction, a cti ve coping, denial, substance use, u se of em otional/ i ns trumental support, behavioral disengagement, venting, positive reframing, planning, humor, acceptance, religion and self blame items. For the current study, the acceptance and positive reframing items ( highlighted in Appendix C) were selected based on t he relevant literature previously reviewed. The combination of these items demonstrated adequate reliability in our sample acceptance and positive reframing items, respectively.
34 R ecent H ealth B ehaviors (RHB; Appendi x D; Pereira, unpublished) Questionnaire To collect information on health and/or eating behaviors during the day s of salivary cortisol collection a short questionnaire was administered Based on the relevant literature ( Ranganathan, et al., 2009; Lovallo et al., 2005) eac h of the following was identified as impacting cortisol output caffeine, bananas, plantains and any illicit drug use Thus, the RHB asked about consumption of these during the three days of saliva collection The measure was developed by the Principal Investigator (DBP) for the purposes of this study Two scores were created from this measure for the current study: a dichotomous variable for whether or not participants endorsed smoking during the days of collection and a sum score of the remaining variables o n the measure. Smoking was examined separately as its impact on cortisol output is well documented (Steptoe & Ussher, 2006) The following was administered during the pre operative study visit. Structured Interview Gui de for the Hamilton Anxiety and Depression Scale (SIGH AD; Appendix E; Williams, 1988) Mood di sturbance ( a nxious and d epressive s ymptoms) during the week prior to surgery was measured using the SIGH AD. This measure has been used widely in medical popula tions, demonstrating adequate reliability and validity in breast cancer populations (Cruess, Antoni, Kumar, & Schneiderman, 2000) and patients living with HIV (Brown, et al., 1992) for the anxiety and the depression scales were .84 and .75, respectively (Prado, et al., 2002) Doctoral students in clinical health psychology administered SIGH AD interviews after being trained to administer the interview by the PI, a licensed psychologist. An abbreviated (24 item) version of the S IGH AD was used in order to reduce patient
35 burden and to remove items confounded with endometrial cancer symptomatology, including genitourinary symptoms or weight loss. This abbreviated version includes 15 depression items and 9 anxiety items to be score d by the interviewer. Following the interview, symptom severity ratings are summed to yield total scores for depressive (possible range of 0 44) and anxious symptomatology (possible range of 0 29): both of which are scored in the direction of greater score s signifying greater symptomatology. Any of these ratings judged by the interviewer as possibly or definitely due to health/treatment factors w ere then subtracted from the total depressive symptomatology score to yield total score not confounded by organic factors For the purposes of the current study, the anxious and depressive symptomatology scales w ere examined separately with both scales demonstrating excellent reliability (Cronbach .83 and .79, respectively ). Further, the depressive scale in t he current study demonstrated excellent concurrent validity, as it was significantly correlated with Beck Depression Inventory II (BDI II) scores (Beck, Steer & Brown, 1996) when organic symptoms were included r (54) = .71, p < .001, as well as when organic symptoms were excluded, r (54) = .67, p < .001. Biomedical Measures Biobehavioral Control Variables Prescription m edications Medical record review and abstraction of current medications was performed. In particular, use of medications known to affect H PA axis functioning (Granger, Hibel, Fortunato, & Kapelewski, 2009) was collected via medical record review and abstraction. This included p rogesterone NE dopamine reuptake inhibitors, anti hypertensives, and adrenocortical steroids The total number of HPA axis modifying
36 medications currently prescribed was examined as a potential biobehavioral control variable in subsequent analyses Disease s tatus and t umor s ubtype More advanced d isease status has been associated with dysregulated cortisol rhythm in breast cancer patients (Abercrombie et al., 2004) and therefore disease status was abstracted via medical record review Disease status was classified using International Federation of Gynecology and Obstetrics (FIGO) cancer stage (I, I I, III, IV), and FIGO cancer grade (low grade, intermediate grade, high grade). In addition, to control for the potential relationship between adenocarcinoma subtype (Type I versus Type II; Bokhman, 1983 ) and cortisol a dichotomous variable was created Specifically, p articipants with preinvasive endometrial disease (complex with Type II adenocarcinomas (clear cell, mucinous, or uterine papillary serous) were Charlson C omorbidity I ndex ( Charlson, Pompei, Ales, & MacKenzie, 1987 ) Given that greater medical comorbidity is associated with more dysregulated cortisol output (Rotman Pikielny, et al., 2006) medical comorbidity was assessed using the Charlson C omorbi dity I ndex This index contains a list of 17 medical conditions (e.g., dementia, peptic ulcer disease, cancer). Each condition is assigned a weight, with more serious conditions (e.g., metastatic solid tumor) b eing assigned a higher weight (6 points) tha n less serious conditions (e.g., peptic ulcer disease; 1 point) Medical record review and abstraction were used to assess for and document the presence of these comorbid medical conditions, and each participant was assigned an index score comprised of th e sum of all weighted items.
37 Salivary Cortisol Collection Participants collected saliva samples at four time points across three consecutive days immediately prior to their surgery The timepoints included 8:00, 12:00, 17:00 and 21:00 hours and were sele cted based upon prior literature linking cortisol with mortality in cancer populations (Sephton et al., 2000) Participants were provided with a collection kit comprised of 13 Salivettes and instructions within a small, insulated co o ler. Instructions wer e provided for participants not to brush their teeth, eat or drink, or smoke for a half hour prior to collecting salivary samples. P articipants were also asked to record the exact time of samples collectio n and store/ refrigerate their samples in the insula ted cooler provided until their return to study staff. Once study personnel received the saliva samples, saliva samples were frozen at 80 degrees Celsius until they were shipped to Salimetrics Inc. (State College, PA) for assaying. Samples were assayed f or salivary cortisol in singlet using a highly sensitive enzyme immunoassay (ELISA; Salimetrics, State College, PA). In brief, ELISA is a used by Salimetrics applies anti bodies to a particular substance of interest (e.g., cortisol), and then a secondary antibody is then applied which includes an enzyme known to react with the substance of interest. This enzyme is then mixed with a specific substrate, creating a change in c olor that signifies the amount of the substance present; in this case, cortisol. The change of color can be measured with a spectrometer and the amount of absorbance then predicts the precise amount of cortisol through a previously established standard cu rve calculation The test used 25 l of saliva per determination, has a lower limit of sensitivity of 0.003 g/dl, standard curve range from 0.012 g/dL to 3.0 g/dL, an average intra assay coefficient of variation of 3.5% and an
38 average inter assay coeffi cient of variation of 5.1%. Method accuracy determined by spike and recovery averaged 100% and linearity determined by serial dilution averaged 91.7% (Curran 2010 ) Statistical Procedures Data Preparation Four calculations were employed to estimate cor tisol output: 1) cortisol regressed upon the time of collection to create a beta weight value (Sephton et al., 2000) 2) AUCi (Pruessner, Kirschbaum, Meinlschmid, & Hellhammer, 2003) 3) intraindividual cortisol variability via multi level modeling techniq ues (Hruschka et al., 2005) and 4) cortisol morning to evening ratio (Jehn et al., 2006; Weinrib et al., 2010) The s pecific calculation of these four techniques is described in greater detail below. Cortisol Slope Cortisol slope was calculated by applyin g ordinary least squares regression Sephton et al, 2000) which regresses the four cortisol values collected on the 4 time points of collection (8:00AM, 12:00PM, 5:00PM and 9:00PM) Given that there are three days of collection, the cor tisol value of each t ime point wa s averaged for each per son in this calculation. By estimating the linea r slope for each individual, a is assigned, which then can be used as a separate variable for subsequent analyses. Area under the Curve with Respect to Increase (AUCi) Co rtisol AUCi was calculated from a previously published formula (Pruessner, et al., 2003) The two formulas presented by Pruessner et al. (2003 ) are thought to measure different processes and include AUC with respect to ground (AUCg) and AUC with respect to increase (AUCi). The former calculation measures total cortisol output and is a proxy for total hormonal output over the day, whereas the latter formula is a
39 proxy for HPA axis reactivity, as the basal level of cortisol for a given individual is not inclu ded in the formula (Pruessner, et al., 2003) The current study was interested in stress reactivity, given the previously reviewed literature on HPA axis reactivity alterations in women with cancer (Porter et al., 2003 ) only AUCi was used in the current an alyses. The calcula tion for AUCi is presented in Figure 2 1 Greater ( less negative) values signify more total cortisol output and greater HPA axis reactivity. Cortisol variability via Multi level Modeling In order to examine variability in the context of data with signifi cant time trends, the data were first potential confounds (Hultsch & MacDonald, 2004) To this end, HLM was applied to model accurately the diurnal time trend of c or tisol, and the residuals were saved. In the was then saved as a separate variable, creating a variable of cortisol intraindividual variability. Morning to Evening Ratio The formula for the calculation of the m orning to e vening r atio in this study was calculated based on published literature (Jehn, et al., 2006) as follows: ((morning cortisol nocturnal cortisol)/morning cortisol) x 100). Lower numbers signify a more abno rmal cortisol rhythm due to either (a) a smaller positive difference between morning and evening cortisol levels (i.e., creating a flattened, but negative ) or (b) a greater negative difference between morning and evening cortisol levels (i.e., creating a truly positive slope).
40 D escriptive Statistics All variables (including the aforementioned cortisol calculations ) were examined for normality to ensure the appropriate use of parametric statistics Additionally, outliers were examined, and if data remaine d nonnormal, values above/below three standard deviations were eliminated from subsequent data analysis. If appropriate, a transformation was applied to ensure normality of all data Specific Aims and Hypotheses Aim 1: To examine the relationship between mood disturbance (depressive symptomatology, anxious symptomatology ) and intraindividual cortisol variability among women with suspected endometrial cancer prior to surgical resection. To examine support for hypotheses 1a and 1b, which assert that women r eporting greater depressive and anxious symptoms will have greater cortisol variability, multi level modeling (MLM) was used Specifically, MLM was used to create an intraindividual cortisol variability estimate, and then b ivariate Pearson c orrelation ana lyses were conducted to examine the relationship between cortisol intraindividual variability and the psychological variables of interest (depressive and anxious symptoms). If found to be significant, significant psychological variable cortisol variabili ty correlations were explored further using hierarchical regression (ordinary least squares regression in which predictors are entered in theoretically ordered blocks), while controlling for variables associated with cortisol variability The following con trol variables, reviewed above, were examined and included in the analyses if they were significantly related to cortisol variability: number of HPA modifying medications
41 currently prescribed, current cigarette smoking, FIGO cancer stage, FIGO tumor grade and Charlson C omorbidity Index score. As exploratory an alyses within this Aim, the unique contribution s of anxiety and depression to cortisol variability were examined To do so, hierarch ical regression analysis was used to orthogonalize depression and anxiety scores (Little, Boviard & Widaman, 2006) By regressing out the depression and anxiety score s from the total score, two variables were created : depression excluding any anxiety effects, and anxiety excluding any depression effects. These variables will be referred to hereafter work using similar statistical approaches ( Kubzansky, Cole, Kawachi, Vokonas, & Sparrow, 2006) Based upon the majority of published work in cortisol variability focusing on major depressive disorder ( Peeters et al., 2004; Posener et al., 200 4 ), in conjunction with the paucity of literature examining anxiety cortisol variability relations, it was hypothesize d that anxiety would not be s ignificantly related to cortisol variability once the contribution of depression was regressed out of the relati onship. Conversely, it was hypothesized a significant relationship would emerge between cortisol variability and depressive symptomatology, once the contribution of anxiety was regressed out of the relationship All analyses for Aim 1 were conducted using SPSS version 20.0 ( SPSS Inc., Chicago IL ) MLM modeling of the cortisol trend was used to create the intraindividual cortisol variability esti mate with the MIXED feature within SPSS
42 Aim 2: To examine the relationship between mood disturbance (depressive symptomatology, anxious symptomatology ) and cortisol slope among women with suspected endometrial cancer prior to surgical resection. To examin e support for hypotheses 2a and 2b, which assert that women who report greater depressive symptoms will have f latter cortisol slopes, MLM was also used. As previously reviewed, MLM has been increasingly used to model the diurnal time trends observed in co rtisol (Adam & Gunnar, 2001; Almeida, Piazza, & Stawski, 2009; Hruschka, Kohrt, & Worthman, 2005). Following Singer and Willett (2003), only the k 2 polynomial trends can be identified; therefore we test ed for linear and quadratic time tre nds. The quadrat ic trend was orthogonalized to remove multicollinearity with the linear trend. To make estimates of cortisol more interpretabl e, saliva collection times were rescaled to have the starting value equal zero while keeping the original scale between points of time to preserve the original distance between points of collection. The sample average time at which each of the four cortisol measurements was tak en (over the three days) was used as the time of day predic tor. Data in this study w ere represented by 3 le vels of data collection: time of cortisol collection within day of collection (Level 1), time of cortisol collection across all 3 days of collection (Level 2) and between persons differences within day (Level 3). The possible effect of day of cortisol col lection was also estimated (Level 2). The final analyses include d two levels of analysis: time within day and between person estimations of cortisol within day, as we did not anticipate a significant cortisol time trend across days of collection. To exam ine between person predictors of inter est, two separate models were constructed: one including SIGH AD Depression as a continuous predict or (Hypothesis 2a) and one
43 including SIGH AD Anxiety as a continuous predictor (Hypothesis 2b). These were entered as Level 2 predictors to explain individual differences in linear cortisol slope. Both depression and anxie ty scores from the SIGH AD were centered to ensure that parameter estimates are more interpretable (Blackwell, de Leon, & Miller, 2006). As in Aim 1, t he fol lowing control variables were examined and included in the analyses if they significantly predict the cortisol time trend or if they significantly improve d the model fit based on comparing 2 log likelihood ratios in maximum likelihood estimation: nu mber of HPA modifying medications currently prescribed, current cigarette smoking, FIGO cancer stage, FIGO tumor grade, and Charlson C omorbidity Index score As explorator y analyses, these models were re run excluding the 8AM timepoint which has been well documented in the research literature as being highly reactive to psychosocial stressors ( e.g., Chida & Steptoe, 2009) The exclusion of this timep oint allowed for the examination of the relationship between depressive/anxious symptoms and afternoon to e vening cortisol slope, as this relationship may differ once 8AM cortisol has been eliminated. Similar to the approach used to model cortisol trend in specific a im 1, this a im use d the MIXED feature within SPSS Aim 3: To examine moderating effects of adaptive coping (positive reframing, acceptance) on relationships between mood disturbance (depressive symptomatology, anxious symptomatology) and cortisol (intraindividual variability and slope) in women with suspected endometrial cancer prior to surgical resection. To examine support for hypotheses 3a and 3b, which asserts that the greater use of adaptive coping skills will moderate the relationship between mood disturbance and cortisol variability and cortisol slope, HLM (Hypothesis 3a) and ordinary lea st squares
44 regression analyses (Hypothesis 3b) were used. To test Hypo thesis 3a, coping scores were entered as a Level 2, between person predictors, in addition to depressive and anxious symptom scores. The two coping scores (Acceptance and Positive Refra ming) were tested as moderators of the predicted anxiety/depression and cortisol relationships, following the criteria put forth by Aiken & West (1991). In summary four separate MLM models were conducted including: 1) d epressive symptoms and p ositive re framing items, 2) d epressive symptoms and a cceptance coping items, 3) a nxious symptoms and a cceptance coping items, and 4) a nxious symptom and p ositive reframing scores all as predictors of the w ithin day linear cortisol trend Additionally, to examine the potential moderating effects of coping on cortisol variability (Hypothesis 3b), ordinary least squares regression analyses were conducted, with cortisol intraindividual variability as the dependent variabl e. As above, four models were tested to predict corti sol intraindividual variability, controlling for variables that were significantly related to cortisol variability (as listed in specific aims 1 and 2) in the first block of the regression equation. Moderation was examined using the methods of Aiken & West (1991) Specifically, cortisol variability was regressed on the presence of high risk adenocarcinomas, the main effects of each mood variable, the main effects of each moderator variable, and the cross product of each mood/moderator variable. T o control for the potential impact of multicollinearity inherent with the use of interaction terms, the unique effect of each interaction was computed by orthogonalizing the product term Specifically, residuals are saved from a linear regression in which i nteraction term is regressed on the predictor and moderator variables, and these residuals are used as
45 the interaction term in subsequent regressions ( Little et al., 2006). Similarly, interaction terms were entered into the multilevel model to predict the within day linear trend. For these analyses, each predictor (anxious/depressive symptoms and acceptance/positive reframing) was orthogonalized with the time variable, to control for the potential confound of multicollinearity between predictors. Secondary /Exploratory Aim 4: To compare four operationalizations of diurnal cortisol rhythm/output (i.e., intraindividual variability, slope, AUCi, and morning to evening ratio) as indicators of c ortisol d ysregulation. As described above, f our separate methods of operationalizing cortisol output were calculated. Relations among these estimates were examined using Pearson correlation analysis. Power and Sample Size Prior to data analyses, a priori power analyses were conducted to determine the sample size that would be necessary to detect specific effect sizes with power = .80 and = .05 These specific effect sizes were drawn from prior published literature Studies by Havermans et al (2010) and Peeters et al. (2004) suggest that correlations between psycho logical factors, such as depressed mood, and cortisol variability range from r = .28 to r = size convention (Cohen, 1992) Using multiple linear regression with 2 predictors (e.g ., a control variable and a predictor of interest) and = .05, it was determined that a sample size of 113 would be necessary to detect an f 2 = .088 (equivalent to r = .285) with power = .80. At the time of data analyses, it was determined that 85 participants contributed complete data on the variables of in terest. With a sample size of 85, the
46 achieved power to detect an f 2 = .088 with 2 predictors and = .05 was .68, while the achieved power to detect this effect size with 1 predictor and = .05 was .79. Thus, in order to maintain adequate power, effort s were made to limit the number of predictors used. Snijders (2005) suggests performing power analyses for MLM procedures in a manner similar to those for multiple regression analyses Giese Davis et al. (2004) found the effect sizes between psychosocia l factors and diurnal cortisol slope in cancer patients ranged from d = .75 to d = .85. Using multiple linear regression with 2 predictors and = .05, a sample size of 64 would be necessary to detect an f 2 = .16 (equivalent to d = .80) with power = .80. Given that the sample size for the curr ent study was 85 the study appeared to be is adequately powered to detect a true relationship between p sychosocial factors and diurnal cortisol slope with MLM Figure 2 1 Calculation for Area under the Curve with Respect to Increase (AUCi)
47 CHAPTER 3 RESULTS Participant Characteristics One hundred thirty four women were enrolled into the study. Of the se participants, 26 were excluded, as they did not contribute any psychosocial or cortisol data due to systematic data collection problems (e.g., participant was discharged from clinic without the knowledge of the study researchers and before any data coul d be collected). Of the 108 remaining participants, 85 contributed complete psychosocial data and more than 1 saliva sample. Analyses were conducted on these 85 participants. Comparison of these 85 participants to the 23 who provided only partial psycho social and/or cortisol data revealed that there were no statistically significant differences on major demographic, psychosocial, or cancer related variables between the final sample of 85 and the 23 providing only partial data (see Table 3 1). The sampl e on which analyses were conducted was primarily Caucasian (93%) and older ( M = 61.89, SD = 8.96) Mean BMI was 35.95 kg/m 2 ( SD = 10.92 kg/m 2 ) with 56 (66%) of the participants within the obese range ( BMI above 30 kg/m 2 ) and 14 participants (17%) within the normal and overweight ranges. The majority of participants had surgically diagnosed Stage I disease ( n = 55, 65% of the sample analyzed) and well differentiated FIGO grades ( n = 45, 54% of the sample analyzed; Table 3 1 ). Further, the majority of parti cipants had tumors classified as Type I ( n = 73, 89%), with a small percentage having tumors classified as Type II ( n = 9, 11%). Depressive and Anxious Symptomatology Depression scores ranged from 0 to 30 with a mean of 7.81 ( SD = 5.59). Excluding those symptoms attributable to organic causes (e.g., medications, direct or
48 indirect effects of the tumor ), depressi on scores ranged from 0 to 30 ( M = 6.75 SD = 5.64) A nxiety scores ranged from 0 to 21, with a mean of 5.27 ( SD = 4.09). Depressive symptomat ology scores were non normally distributed, and therefore a square root transformation was applied resulting in adequate characteristics of normality (Shapiro p = .17 ) Cortisol values A total of 884 cortisol samples were collected across t he 85 participants. were determined to be 3 standard deviations from the mean ( M = .19 g/dL SD = .48 g/dL ) and were eliminated from further data analyses. After the removal of these outliers, cortisol data displayed adequate characteristics of normality (Shapiro test, p = 11 ). The mean raw cortisol values for each time of coll ection are presented in Figure 3 1 Results: AIM 1: To examine the relationship betwe en mood disturbance (depressive symptomatology, anxious symptomatology) and intraindividual cortisol variability among women with suspected endometrial cancer prior to surgical resection. Cortisol variability data were examined for normality. A square root transformation was applied to this data. After applying this transformation, the data displayed adequate characteristics of normality (Shapiro p = .34 ). Intraindividual cortisol variability values ranged from .16 to .58 to ( M = .29, SD = .09 1 ). Co rtisol variability was not related to BMI, r (84) = .0 6, p = .62 d = .12; age, r (85) = .1 7, p = .13 d = .35; any smoking across saliva collection days t (60) =.58, p =.57 d = 1.42; or Charlson C omorbidity index scores r (84) = .10 p =.36 d = .21 Ho wever, the presence of high risk
49 endometrial cancer subtype with cortisol variability; specifically, participants with Type II adenocarcinomas had significantly lower cortisol variability than those with Type I adenocarcinomas, t (81) = 2.33, p = .023 d = .97 Depressive/anxious symptomatology and cortisol variability Using Pearson correlations, cortisol variability was not related to anxiety symptoms, r (85) = .14 p = .19, d = .28. However, cortisol variability was marginally significantly associated wit h depressive symptomatology, r (85) = .21 p = 0 5 1 d = .43 suggesting that those with greater depressive symptoms had higher intraindividual cortisol variability. Given the significant difference in cortisol variability between Type I and Type II adeno carcinomas a hierarchical regression (ordinary least squares regression in which predictors are entered in theoretically ordered blocks) was conducted on cortisol variability cont rolling for presence of high risk subtype After cont rolling for the presence of high risk endometrial cancer subtype greater depressive symptomatology was significantly associated with greater cortisol variability ( = .2 1 p = .048, d = .43 ) Depressive symptoms explain ed 4.6 % of the variance in intraind ividual cortisol variability above a nd beyond presence of high risk subtype (Table 3 2). However, a fter controlling for high risk endometrial cancer subtype anxiety remained unrelated to cortisol variability ( = .11 p = .30 d = .22 ) in regression anal yses (Table 3 3) To further illustrate the relationship between depressive symptomatology and cortisol variability, raw cortisol values were plotted for the five participants reporting the greatest depressive sympt oms prior to surgery (Figure 3 2 ) and th e five participants with the lowest depressive symptoms (i.e., no depressive sympto ms) prior to surgery (Figure
50 3 3 ). As suggested by these Figures, compared to participants with no depressive symptoms, participants reporting the greatest depressive sympt oms demonstrated greater dispersion of raw cortisol values from their individual regression slopes. As exploratory analyses within AIM 1, the anxiety and depressive symptom scores were orthogonalized to isolate the unique contributions of anxiety without depressive symptoms and depressi on without anxiety symptoms To create these orthogonalized variables, linear regression was applied with anxiety as predictor of the total SIGH AD score, controlling for depressive symptoms The residuals were then saved f rom this regression equation, creating a variable representing anxiety symptoms without the contribution of depressive symptoms (iso anxiety) Cortisol variability was then regressed on t he se iso a nxiety and iso depressive symptom scores. Results revealed that n either iso anxiety symptoms nor iso depressive symptoms were related to cortisol variability ( r (85) = .064, p = .56 d = .13 and r (85) = .16, p = .16 d = .32, respectively). G iven the significant relationship between the presence of high risk endo metrial cancer subtype and cortisol variability, a hierarchical regression was conducted to predict cortisol variability with iso anxiety and iso depression as separate predictors controlling for the pre sence of high risk subtype. Iso anxiety symptoms di d not significantly predict cortisol variability ( = .12 p = .29, d = .24 ) (Table 3 4). However, greater iso depressive symptoms approached significance in its association with greater cortisol variability after controlling for the pr esence of high risk subtypes ( = .19 p = 08 0 d = .39 ) (Table 3 5 ) Iso depression explained 3.6 % of the variance in cortisol variability above and beyond the presence of high risk endometrial cancer subtype.
51 Results: Aim 2: To examine the relationship between mood disturbance (depressive symptomatology, anxious sympt omatology) and cortisol slope among women with suspected endometrial cancer prior to surgical resection. Linear Time Trend (Cortisol Slope) As previously stated, multilevel modeling was applied to model the cortisol data and account for the diurnal time tr end typically observed in cortisol rhythms. As expected, a significant negative linear time trend was observed in cortisol ( = 0.014, SE = .001 p < .001) as well as a significant positive quadratic trend at the end of the day ( = .002 SE = .0003, p < .001). As seen in Table 3 6 each model represented an increasingly better fit to the data, represented by decreasing 2 log likelihood statistics. Further, when the nested model tests w ere compared, each iteration remained highly significant ( p values < .001 ; data not shown ). Also as expected, there w ere significant random effects, in both the first measurement of cortisol at 8AM ( p < .001), linear time trend within the day ( p < .001) and the quadratic trend observed at the end of the day (p < .0 01 ; data not shown ). There was not a significant trend of cortisol across days ( = .002, SE = .005, p = .67 ; data not shown ). Further, inclusion of this nonsignificant day trend resulted in convergence problems and it was therefore excluded in subsequent analys es. Control variables were entered into the multilevel model prior to testing for the predictor variables of interest (depressive/anxious symptoms) and retained if a) the model improved based on a significant chi square nested model tests or b) the coeffi cient estimate significantly contributed to the model. Out of the possible control variables investigated ( smoking across saliva collection days, cancer stage, tumor
52 grade, adenocarcinoma subtype, medi cation use), only Charlson C omorbidity Index score was significantly related ( = .004, SE = .00 2 p = .006) to cortisol slope in the multile vel model; therefore, Charlson C omorbidity Index score was retained as a control variable in the depression and anxiety conditional growth model s (Table 3 7). Depressive/A nxious Symptoms and Cortisol Slope After controlling for Charlson C omorbidity Index score, neither depressive nor anxious symptoms predicted the initial value of cortisol ( = .013, SE = .012, p = 30 and = .0008, SE = .003 p = .8 1 ; respectively) (Table 3 8) Similarly, and contrary to hypotheses, neither depressive nor anxious symptoms predicted linear cortisol slope across the three days preceding surgery ( = .001, SE = .0009, p = .2 6 and = .000 3 SE = .000 3 p = 29 ; respectively) (Table 3 9) Furthermore, neither depressive nor anxious symptoms were related to the positive quadratic trend observed in the cortisol rhythm within day (data not shown). As in analyses in AIM 1, iso depression and iso anxiety were entered in the multilevel model to predict linear cortisol slope. Iso a nxiety remained unrelated to the .00 8 SE = .00 6 p = .1 8 ) or linear cortisol slope throughout the day ( = 0.000 5 SE = 0.000 5, p = .32 ) (Tables 3 10 and 3 11, respectively ) However, greater iso depressive symptoms w ere m arginally associated with a greater i nitial value of cortisol ( = 0.007, SE = 0.004, p = .0 78 ) G reater iso depressive symptoms approached significance in its relationship with a steeper cortisol slope ( = 0.0006 SE = 0.0003, p = .095) (Tables 3 10 a nd 3 11, respectively). As exploratory analyses within this aim, these models were re run excluding the 8AM time point; however, the elimination of this data point led to repeated convergence problems, and therefore these models are not reported.
53 Results: Aim 3: To examine moderating effects of adaptive coping (positive reframing, acceptance) on relationships between mood disturbance (depressive symptomatology, anxious symptomatology) and cortisol (intraindividual variability and slope) in women with suspe cted endometrial cancer prior to surgical resection. Descriptive Statistics for Acceptance and Positive Reframing Items Of the 85 participants completing other variables of interest (SIGH AD and providing cortisol samples), only 72 participants completed t he COPE. Positive reframing responses ranged from 2 to 8 ( M = 5.26 SD = 1.85). Responses on the a cceptance item on the COPE ranged from 3 to 8 ( M = 6.63 SD = 1.33). Both items displayed adequate characteristics of normality and therefore the raw values w ere used in subsequent analyses. Relations among Acceptance, Positive Reframing and Cortisol Variability There were no significant correlations between p ositive r eframing scores and cortisol variability ( r (70) = .10 p =.39 d = .20 ). Similarly, there was no significant correlation between a cceptance scores and cortisol variability ( r (72) = .18 p = .12 d = .37 ). Positive Reframing and Acceptance as Moderators of Anxiety/Depressive Symptoms and Cortisol Variability To conduct moderation analyses, the ste ps of Aiken & West (1991) were applied, controlling for the main effects of each predictor variable in the hierarchical regression equation. Further, to control for the potential impact of multicollinearity inherent with the use of interaction terms, the u nique effect of each interaction was computed by orthogonalizing the product term (e.g., residuals are saved from a separate hierarchical
54 regression in which each variable is entered as independent variables and the product term is the dependent variable ; Little et al., 2006). As seen in Table 3 12, there was no main effect of positive reframing on cortisol variability ( = .11, p = .35, d = .22 ) ; further, positive reframing did not moderate the relationship between depression and cortisol variability ( = .15, p = .21, d = .30; Table 3 12 ) Similarly, as show n in Table 3 13, there was no main effect of acceptance on cortisol variability ( = .19, p = .11, d = .38 ) and acceptance did not moderate the relationship between depression and cortisol variabil ity ( = .04 2 p = .72, d = .08 4 ) However, as presented in Table 3 14, there was a marginally significant relationship between acceptance and cortisol variability ( = .22, p = .07 2 d = .45 ) such that greater use of acceptance as a coping strategy was m arginally related to greater intraindividual cortisol variability. In this model, acceptance did not moderate the relationship between anxiety and cortisol variability ( = 01 0 p = .93, d = .02 4 ; Table 3 14 ) As s een in Table 3 15, there was no main ef fect of positive reframing on cortisol variability ( = .11, p = .38, d = .22 ) ; furthermore, positive reframing did not moderating the relationship between anxiety and cortisol variability ( = .03 4 p = .79, d = .06 2 ) As outlined in specific aim 3, e ac h term was entered into separate multilevel models (in the same order of the ordinary least squares regression models just presented). As presented in Table 3 16, there was no main effect of positive reframing on cortisol slope ( = 0.0002, SE = 0.006, p = .80 ); further more positive reframing did not moderate the relationship between depression and cortisol slope ( = 0.008, SE = 0.006, p = .14 ). Similarly, as seen in Table 3 17, there was not a main effect of acceptance on cortisol slope ( = 0.001, SE = 0.001, p = .17 ), nor did acceptance
55 moderate the relationship between depression and cortisol slope ( = 0.003, SE = 0.005, p = .64 ). As demonstrated in Table 3 18, there was no relationship between acceptance and cortisol slope ( = 0.001, SE = 0.001 p = .12 ), and acceptance did not moderate the relationship between anxiety and cortisol slope ( = 0.00002, SE = 0.003, p = .10 ). As presented in Table 3 19, there was no main effect of positive reframing on the relationship between anxiety and cortisol slope ( = 0.00008, SE = 0.0006, p = .89 ), and positive reframing did not moderate the relationship between anxiety and cortisol slope ( = 0.003, SE = 0.002, p = .13 ). Results: Secondary/Exploratory Aim 4: To compare four operationalizations of diurnal c ortisol rhythm/output (i.e., intraindividual variability, slope, area under the curve with respect to increase [AUCi], and morning to evening ratio) as indicators of Cortisol Dysregulation. Cortisol Slope Cortisol slope was calculated by regressing the fou r cortisol values collected on the time of day of collection and assigning a value to each individual as an estimate of the linear slope. Consistent with prior work examining cortisol slope in cancer populations (Sephton et al., 2000), a natural log tran sformation was applied to raw cortisol values prior to the calculation of cortisol slope. Following this transformation, cortisol slope values displayed adequate characteristics of normality (Shapiro Wilks test p = .05 4 ). The mean of cortisol slope was .096 ( SD = .054), ranging from .25 to .039. For these data, a greater number is associated with a flatter diurnal pattern of linear cortisol slope.
56 Area U nder the Curve with Respect to Increase (AUCi) AUCi v alues ranged from 46.09 to 6.88 ( M = 24.93 SD = 7.13), displaying adequate characteristics of normality (Shapiro Wilks test p = 77 ). In these data, a greater ( less negative ) value corresponds to greater cortisol output throughout the day, after controlling for the baseline cortisol, thus signify ing greater HPA axis reactivity. Morning to evening Ratio of C ortisol The raw morning evening ratio values ranged from 72.46 to 94.35 ( M = 63.81 SD = 31.61). In these data, a lower value signifies a more abnormal slope. These values remained non normal and therefore a B lom transformation was applied, resulting in a normal distribution (Shapiro Wilks test p = 10 ). Comparison a mong Cortisol Calculation s Pearson correlations were first calculated to examine the relationship among the four methods of ope rationalizing cortisol output. Cortisol intraindividual variability was not associated with either morning to evening cortisol ratio, r (82) = .002, p = .99, or cortisol slope, r (82) = .016, p = .89; furthermore, cortisol slope was not correlated with AUCi, r (83) = .14 p = .20 However, greater AUCi ( greater cortisol output ) was significantly associated with greater cortisol variability, r (83) = .62 p < .001 and marginally associated with lower morning to evening cortisol ratio, r (82) = .21 p = .06 s uggesting that greater cortisol reactivity (AUCi) was associated with more erratic cortisol output and lower difference between morning and evening cortisol. A highly significant relationship also emerged between cortisol slope and morning to evening cor tisol ratio, r (82) = .93 p < .001, such that a flatter cortisol slope was associated with a lower difference between morning and evening cortisol (Table 3 20).
57 Table 3 1. Comparison of Participants Included and Excluded in Data Analyse s Variable Includ ed in Analyses ( n = 85) Excluded from Analyses ( n =23) Test Statistic Effect Size M ( SD ) n M ( SD ) n t value X 2 p value d V SIGH AD Depression 7.81 (5.59) 7.39 (4.57) .33 .74 .082 SIGH AD Depression (with items attributable to health/trea tment factors excluded) 6.75 (5.64) 5.91 (4.02) .67 .51 .17 SIGH AD Anxiety 5.27 (4.09) 4.78 (3.55) .52 .61 .13 Age 61.89 (8.96) 59.39 (9.56) 1.17 .24 .27 BMI 35.95 (10.92) 38.09 (12.12) .81 .42 .19 Cancer Grade a 3.15 .37 .17 Benign 6 2 Well differentiated 45 14 Moderately differentiated 24 2 Poorly differentiated 8 2 Cancer Stage 1.13 .77 .10 Benign 6 2 Stage I 55 17 Stage II 13 2 Stage III 11 2 Race 3.64 .16 .18 Caucasian 79 18 African American 5 4 Mixed 1 0 Ethnicity .80 .37 .09 Not H ispanic /Latino 75 17 Hispanic/Latino 4 2
58 Table 3 2. Pre dicting Intraindividual Cortisol Variability from Depressive Symptoms Step Number Predictor Variable R 2 R 2 F R 2 p 1 Presence of High Risk Endometrial Cancer Subtype .06 4 .2 5 .06 4 5.45 .022 2 SIGH AD Depression Score .11 .21 .0 46 4.05 .048 n = 82. Significance of Model, F (2, 79) = 4.85, p = .01 0 Table 3 3 Predicting Intraindividual Cortiso l Variability from Anxiety Symptoms. Step Number Predictor Variable R 2 R 2 F R 2 p 1 Presence of High Risk Endometrial Cancer Subtype .06 4 .25 .06 4 5.45 .022 2 SIGH AD Anxiety Score .0 77 .11 .01 3 1.11 .30 n = 82. Significance of Model, F (2, 7 9) = 3.29, p = .04 3 Table 3 4. Predicting Intraindividual Cortisol Variability from Anxiety Symptoms; Controlling for Concurrent Depressive Symptoms. Step Number Predictor Variable R 2 R 2 F R 2 p 1 Presence of High Risk Endometrial Cancer Subty pe .06 4 .25 .06 4 5.45 .022 2 SIGH AD Iso Anxiety Score 077 .12 .01 3 1.15 .29 n = 82. Significance of Model, F (2, 79) = 3.30, p = .04 2 Table 3 5. Predicting Intraindividual C ortisol Variability from Depressive Symptoms; Controlling for Con current Anxiety Symptoms. Step Number Predictor Variable R 2 R 2 F R 2 p 1 Presence of High Risk Endometrial Cancer Subtype .06 4 .2 5 .06 4 5.45 .022 2 SIGH AD Iso Depression Score .10 .19 .0 36 3.14 .080 n = 82. Significance of Model, F (2, 79) = 4.37, p = .01 6
59 Table 3 6 Unconditional Growth Models of C ortisol Values. Fixed effects 2LL AIC BIC Residual Intercept Coefficient SE T p Unconditional Means Model 640.63 636.63 627.06 0.028 0.16 n/a n/a 28.73 .000 Random Intercept 759.45 753.45 739.10 0.022 0.007 0.16 0.011 15.47 .000 Unconditional Growth Model (Linear) 997.81 989.81 970.68 0.016 0.008 0.014 0.001 16.67 .000 Unconditional Growth Model (Random Effects of Linear) 1004.52 994.52 970.60 0.015 0.010 0.015 0.001 14.22 .000 Unconditional Growth Model (Li near and Quadratic Time ; including Random Effects) 1065.55 1051.55 1018.06 0.012 0.012 0.002 0.00 03 6.14 .000 Random Effects Variance Wald Statistic p Random Intercept .007 4.89 .000 Random Linear .00003 2.010 .044 Random Linear and Quadratic .000002 2.20 .028 Criteria (all displayed in smaller is better forms); SE = Standard Error ; n/a = not applicable. *All multi level models controlled for Charlson comorbidity score, as this was significantly related to linear cortisol slope.
60 Table 3 7 Relationships Between Potential Covariates and Cortisol Slope using Multilevel Model ing Covariate Tested* 2LL AIC BIC Residual Intercept Coefficient SE T p Presence of High Risk Endometrial Cancer Subtype 1075. 20 1059. 20 1021.1 7 0.012 0.012 0.085 0.043 1.9 8 .052 Total Number of HPA axis Modifying Medication s Currently Used 1067.9 3 1051.9 3 1013.65 0.012 0.012 0.064 0.041 1.55 .1 3 Cancer Stage 1069.0 1 1053.0 1 1014.73 0.012 0.012 0.030 0.016 1.8 8 .065 Tumor Grade 1030.1 3 1014.1 3 976.0 2 0.013 0.012 0.017 0.018 0.9 6 .34 Any Cigarette Smoking Across Saliva Collection Days 900.6 3 884.6 3 848. 97 0.010 0.010 0.003 0.039 0.085 .93 Charlson Comorbidity Index Score 1073.9 3 1057.9 3 1019.9 3 0.012 0.012 0.004 0.002 2.7 8 .006 played in smaller is better forms); SE = Standard Error. *All models were run separately and fit statistics were compared to the Unconditional Growth Model (which included fixed and random effects of linear and quadratic time trends; 2 LL = 1065.55, AIC = 1051.55, BIC = 1018.06; Table 3 6, above.)
61 Table 3 8 Predicting Initial Cortisol Value from Anxiety and Depressive Sympto ms. 2LL AIC BIC Residual Intercept Coefficient SE T p SIGH AD Depression Score 1075.04 1057.04 1014.29 0.012 0.012 0.013 0.012 1.06 30 SIGH AD Anxiety Score 1073.9 9 1055.9 9 1013.2 4 0.012 0.012 0.00 08 0.003 0.2 5 .8 1 2LL = Negative 2 log likelih Bayesian Criteria (all displayed in smaller is better forms); SE = Standard Error. *All multilevel models controlled for Charlson C omorbidity Index score, as this was significantly related to linear cortisol slope. Table 3 9. Predicting Cortisol Slope from Anxiety and Depressive Symptoms. 2LL AIC BIC Residual Intercept Coefficient SE T p SIGH AD Depression Score 1075.2 4 1057.2 4 1014.4 9 0.012 0.012 0.001 0.00 09 1.1 5 .2 6 S IGH AD Anxiety S core 1075.0 4 1057.0 4 1014.2 9 0.012 0.012 0.000 3 0.000 3 1.0 6 .29 Bayesian Criteria (all displayed in smaller is better forms); SE = Standard Error. *All multilevel models controlled for Charlson C omorbidity Index score, as this was significantly related to linear cortisol slope. Table 3 10 Predicting Initial Cortisol Value from Iso Anxiety and Iso Depressive Symptoms. 2LL AIC BIC Residual Intercept Coefficient SE T p SIGH AD Iso Depression Score 1073.42 1055.42 1012.67 0.012 0.012 0.007 0.004 1.79 .078 SIGH AD Iso Anxiety Score 1072.0 6 1054.06 1011.31 0.012 0.012 0.008 0.006 1.3 5 .18 Bayesian Criteria (all displayed in smaller is better forms); SE = Standard Error. *All multilevel models controlled for Charlson C omorbidity Index score, as this was significantly related to linear cortisol slope.
62 Table 3 11 Pred icting Linear Cortisol Slope from Iso Anxiety and Iso Depressive Symptoms. 2LL AIC BIC Residual Intercept Coefficient SE T p SIGH AD Iso Depression Score 1073.0 9 1055.0 9 1012.3 4 0.012 0.012 0.00 06 0.000 3 1.68 0.095 SIGH AD Iso Anxiety Score 1071. 2 7 1053.2 7 1010.5 2 0.012 0.012 0.000 5 0.000 5 1.001 0.3 2 Bayesian Criteria (all displayed in smaller is better forms); SE = Standard Error. *All multilevel models co ntrolled for Charlson comorbidity Index score, as this was significantly related to linear cortisol slop e
63 Table 3 12. Examining Positive Reframing as a Potential Moderator of a Relationship b etween Depressive Symptoms and Cortisol Variability. Step Numb er Predictor Variable R 2 p R 2 F R 2 1 Presence of High Risk Endometrial Cancer Subtype .04 1 .21 .08 1 .04 1 2.91 2 SIGH AD Depressive Symptoms .12 .21 .08 2 .07 7 1.85 Positive Reframing .11 .35 Positive R eframing X Depressive Sympt oms* .15 .21 n = 68. Significance of Model, F (4, 63) = 2.20, p = .0 79 *Represents the residualized interaction term Table 3 13. Examining Acceptance as a Potential Moderator of a Relationship b etween Depressi ve Symptoms and Cortisol Variability Step Number Predictor Variable R 2 p R 2 F R 2 1 Presence of High Risk Endometrial Cancer Subtype .04 1 .20 .09 3 .04 1 2.37 2 SIGH AD Depressive Symptoms .13 .23 .0 56 .09 0 2.24 Acceptance .19 .11 Acceptance X Depressive Symptoms* .04 2 .72 n = 70. Significanc e of Model, F (4, 65) = 2.44, p = .0 55 *Represents the residualized interaction term. Table 3 14. Examining Acceptance as a Potential M oderator of a Relationship b etween Anxiety Symptoms and Cortisol Variability Step Number Predictor Variable R 2 p R 2 F R 2 1 Presence of High Risk Endometrial Cancer Subtype .04 1 .20 .09 3 .04 1 2.91 2 SIGH AD Anxiety symptoms .11 .18 .14 .0 65 1.57 Acceptance .22 .07 2 Acceptance X Anxiety Symptoms* .01 0 .93 n = 70. Significance of Model, F (4, 6 5) = 1.92, p = .12 *Represents the residualized interaction term.
64 Table 3 15. Examining Positive Reframing as a Potential Moderator of a Relationship between Anxiety Symptoms and Cortisol Variability. Step Number Predictor Variable R 2 p R 2 F R 2 1 Presence of High Risk Endometrial Cancer Subtype .0 45 .21 .081 .0 45 3.14 2 SIGH AD Anxiety symptoms .07 2 .12 .37 .0 26 .60 Positive reframing .11 .38 Positive reframing X Anxiety Symptoms* .03 4 .79 n = 68. Significance of Model, F (4, 63) = 1.22, p = .31 *Represents the residualized interaction term.
65 Table 3 16 Examining Positive Reframing as a Potential Moderator of a Relationship b etween Depressi ve Symptoms and Cortisol Slope. 2LL AIC BIC Residual Intercept Coefficient SE T p 890.2 3 868.2 3 818.1 2 0.012 0.011 SIGH AD Depression Score 0.002 0.001 2.12 .037 Positive Reframing 0.000 2 0.00 6 0.25 .80 SIGH A D Depression Score X Positive Reframing 0.008 0.006 1.49 .14 *Represents the resid ualized interaction term. Bayesian Criteria (all displayed in smaller is better forms); SE = Standard Error. *All multilevel models controlled for Charlson comorbidity score, as this was significantly related to linear cortisol slope.
66 Table 3 17 Examining Acceptance as a Potential Moderator of a Relationship between Depressive Symptoms and Cortisol Slope. 2LL AIC BIC Residual Intercept Coefficient SE T p 939. 02 917.02 866.54 0.011 0.012 SIGH AD Depression Score 0.002 0.001 2.289 .024 Acceptance 0.001 0.001 1.380 .17 SIGH AD Depression Score X Acceptance* 0.003 0.005 0.476 .6 4 *Represents the residualized interaction term. 2LL = Bayesian Criteria (all displayed in smaller is better forms); SE = Standard Error. All multilevel models controlled for Charlson C omorbidity score, as this was significantly related to linear cortisol slope.
67 Table 3 18 Examining Acceptance as a Potential Moderator of a Relationship b etween Anxiety Symptoms and Cortisol Slope. 2LL AIC BIC Residual Intercept Coefficient SE T p 935.8 1 913.4 4 863.33 0.011 0.012 SI GH AD Anxiety Score 0.000 4 0.000 3 1.463 .1 5 Acceptance 0.001 0.001 1.570 .12 SIGH AD Anxiety Score X Acceptance* 0.000 02 0.003 0.006 .99 *Represents the residualized interaction term. 2LL = Negative 2 log likelihood; AIC = Akai Bayesian Criteria (all displayed in smaller is better forms); SE = Standard Error. All multilevel models controlled for Charlson C omorbidity score, as this was significantly related to linear cortisol slope.
68 T able 3 19 Examining Positive Reframing as a Potential Moderator of a Relationship b etween An xiety Symptoms and Cortisol Slope. 2LL AIC BIC Residual Intercept Coefficient SE T p 887.0 1 865.0 1 81 5.00 0.012 0.012 SIGH AD Anxiety Score 0.0 00 3 0.000 3 1.053 0. 30 Positive Reframing 0.000 08 0.00 06 0.134 0.89 SIGH AD Anxiety Score X Positive Reframing* 0.003 0.002 1.548 0.1 3 *Represents the residualized interaction term. displayed in smaller is better forms); SE = Standard Error. All multilevel models controlled for Charlson C omorbidity score, as this was significantly related to linear cortisol slope.
69 Table 3 20 Pears on Correlations among Methods of Operationalizing Cortisol Dysregulation. 1. 2. 3. 4. 1. Intraindividual Cortisol Variability 2. Morning to evening Cortisol Ratio .002 3. Cortisol Slope .0 16 .93*** 4. Area Under the Curve (with respe ct to increase) .6 2 *** .21* .14 p = .0 56 ; *** p < .001
70 Figure 3 1. Average Cortisol Values Across the Four Timepoints of Collection
71 Figure 3 2. Cortisol V ariability among Participants w ith the H ighest D epressive (HD) S ymptom S cores. These par ticipants demonstrate h igh cortisol variability as their actual cortisol values are dispersed from their average pattern
72 Figure 3 3. Cortisol V ariability among P articipants with the Lowest Depressive (LD) Symptom Scores These select 5 participants reported no depres sive symptoms prior to surgery and demon strate low cortisol variability
73 CHAPTER 4 DISCUSSION The current study sought to examine relationships between HPA axis dysregulation and anxious and depressive symptoms in a group of women unde rgoing surgery for suspected endometrial cancer. While multiple operationalizations of HPA axis dysregulation (e.g., analysis of diurnal cortisol rhythm) have been put forth in the extant literature, there is little consensus on the most appropriate metho d. T he current study sought to examine intraindividual cortisol variability a relatively new and novel indicator of how erratic or variable cortisol output is throughout the day, and its relation to anxious and depressive symptomatology Based upon publi shed literature on cortisol variability and mood disorders (Havermans et al., 2010 ; Peeters et al., 2004; Posener et al., 2004) i t was hypothesized that greater intraindividual cortisol variability would be related to greater anxious and depressive sympto ms prior to surgery. In addition, linear cortisol slope was examined (a widely applied analytic technique) and it was hypothesized that a flatter slope would be related to great anxious and depressive symptomatology preceding surgery for endometrial cancer Finally, acceptance and positive reframing coping strategies were examined as moderators of any mood cortisol relationships. E xploratory analyse s examined relations among cortisol variability, diurnal cortisol slope, cortisol AUCi, and morning to evening ratio of cort isol throughout the day. Biobehavioral Control Variables and Intraindividual Cortisol Variability and Cortisol Slope Of the potential biobehavioral control variables examined, women with high risk adenocarcinoma subtype s demonstrated higher intraindividual cortisol variability than those with low risk adenocarcinoma subtypes, suggesting that more aggressive disease may be related to HPA axis dysregulation in our study sample. The correlational design
74 of the study prevents the ability to dete rmine the direction of the relationship between aggressive endometrial cancer variants and intraindividual cortisol variability; nonetheless, these data may have implications for understanding the pathophysiology of endometrial carcinogenesis. The curren t study also found a significant correlation between greater number of medical comorbidities and flatter diurnal cortisol slope. This is in concert with prior literature (Rotman Pikielny, et al., 2006) consistent with hypotheses predicting that a flatten ed cortisol slope is negative health consequences. The utility of a flattened cortisol slope in the prediction of mortality has been observed in cancer populations (Sephton et al., 2000) and, more recently, in nonclinical popula tions (Kumari et al., 2011). Future studies should remain mindful of potential covariates affecting HPA axis outcomes such as intraindividual cortisol variability and slope t hat are unique to cancer populations Depressive Symptoms, Anxious Symptoms, an d Intraindividual Cortisol Variability Consistent with hypotheses, greater depressive symptomatology in the week prior to surgery was significantly associated with greater intraindividual cortisol variability after controlling for the presence of high risk endometrial cancer subtypes. These findings are in accord with prior work demonstrating higher cortisol variability in major depressive disorder (Peeters, et al., 2004) and episode severity in remitted bipolar disorder (Havermans, et al., 2010 ). It is po ssible that greater intraindividual cortisol variability may be indicative of greater circadian rhythm disruption. For instance, it is well established that individuals with chronic and severe disruptions in the sleep wake cycle (i.e., a circadian rhythm) demonstrate greater variability in subjective and
75 objectives measures of sleep efficiency than nonimpaired controls (Buysse, et al., 2010). Thus, it may be worthy in future investigations to compare cortisol variability with quantitative measurements of the sleep wake cycle, such as sleep efficiency, sleep quality or subjective sleep complaints. A recent investigation suggested that, in a population of 57 breast cancer patients, those who reported more intrusive thoughts had lower autocorrelation coeffi cients or more variable rest/activity rhythms as measured by actigraphy. In contrast, less variability in rest / activity rhythms was significantly correlated with a steeper decline in the diurnal cortisol slope throughout the day, both suggestive of a m ore orderly circadian rhythm (Dedert et al. 2012). Greater intraindividual cortisol variability may follow a similar pattern as these findings with actigraphy data indicating a disrupted circadian rhythm. Extrapolating from this research, it is possible that high levels of intraindividual cortisol variability are indicative of HPA axis functioning that is erratic, unpredictable, and inappropriately under and over responsive to the actual demands of the host and his/her environment. While it is presently unknown whether high intraindividual cortisol variability has negative long term health implications in cancer, it is noteworthy that there is a growing body of research suggesting that circadian rhythm disruption is directly associated with carcinogenesi s (Sephton & Spiegel, 2003; Touitou, Bogdan, Levi, Benavid es, & Auzeby, 1996 ). Future research should examine intraindividual cortisol variability, as well as other indicators of circadian disruption such as sleep disturbances, longitudinally to establish whether they are associated with cancer outcomes, such as disease free survival.
76 Contrary to hypotheses anxiety was not significantly associated with intraindividual cortisol variability in the present study. Notably, no published research to date h as examined the relationship between anxiety symptoms and intraindividual cortisol variability and as such, there is a paucity of research on which to generate hypotheses. The research that has been conducted on anxiety and cortisol is limited to examini ng slope, total diurnal output, and/ or cortisol awakening response as outcome variables and the results of these investigations have yielded inconsistent findings (see section below for a detailed summary of this literature ). Depressive Symptoms, Anx ious Symptoms, and Cortisol Slope Results revealed that, in contrast to hypotheses and related prior research (Bhattacharyya et al., 2008) depressive symptomatology was not related to cortisol slope. There are a number of possible explanations for this re sult. First, the SIGH AD queried only about the incidence and severity of mood symptoms in the week prior to depression or an assessment of the chronicity of mood symptoms It is possible that the physiological resistance to glucorticoid secretion that would yield a flattened cortisol slope may only be observed among those with major depressive disorder or prolonged, moderate to severe depressive symptomatology (Miller, C o hen, & Ritchey, 2002). Thus the nonsignificant results of the present study may not be comparable to those obtained among patients with severe or long standing depressive symptoms. In long standing depression, the negative feedback loop that attenuate s cortisol production during perceived stress i s less robust exogenous (e.g., dexamethasone suppression test) cortisol (Mil ler et al., 2002)
77 Second, participants in the present sample reported relatively low levels of depressive symptomatology, and the nonsignificant relationship between depression and cortisol slope may be due to the low amount of variance in depressive sym ptomatology. Also contrary to hypotheses, anxiety was not significantly associated with cortisol slope in the present study. This is in contrast with some published research among breast cancer patients. Giese Davis and colleagues (2004) found that, am ong women with metastatic breast cancer, flatter diurnal cortisol slopes were found among women categorized as (a) high repressors, and (b) high anxiety high repressors compared to those categorized as self assured. However, women categorized as high anxi ety, only, did not have flatter cortisol slopes than those categorized as high repressors or self assured. These findings suggest that there may not be a main effect of anxiety on cortisol slope. Moreover, published research on morning cortisol output a mong individuals with clinical anxiety disorders has yielded inconsistent findings. For example, Vreeburg and colleagues (2012) found that individuals with anxiety disorders display ed an exaggerated cortisol response upon awakening; while leagues (201 0 ) found that clinically anxious individuals had lower levels of morning cortisol compared to non anxious individuals 0 ). While these results with awakening cortisol may not be directly comparable to the current results ( e.g., morning cortisol levels may be independent of intraindividual cortisol variability), they underscore the inconsistent relationships between anxiety and HPA axis functioning in the published literature
78 Finally, Antoni and colleagues (2009) recently found that 10 week group based cognitive b ehavioral s tress m anagement (CBSM) intervention improved mood symptoms, lowered cortisol levels, and promoted adaptive cytokine regulation in women with breast cancer. However, reductions in anxiety did not mediate the relationship between CBSM group assignment and reductions in cortisol suggesting that reductions in cortisol may not always be driven by reductions in anxiety. Future investigations should remain mindful to include potential mediators/moderators of a nxiety cortisol relationships. Examining Unique Contributions of Anxiety and Depression to Intraindividual Cortisol Variability and Cortisol Slope To examine the unique contributions of anxiety and depression to the outcomes of interest (cortisol slope and variability), iso anxiety and iso depression scores were created in exploratory analyses. This approach originates from an underlying statistical framework acknowledging that anxiety and depression are consistently correlated (Burns & Eidelson, 1998). In the present study the unique effect of anxiety was unrelated to intraindividual cortisol variability and cortisol slope. However, when the unique effects of depression were examined, greater iso depression was only marginally associated with greater intr aindividual cortisol variability Compared to the significant relationship between depression and cortisol variability a smaller effect size was observed between iso depression and cortisol variability ( d = .43 versus d = .39 ; respectively ) This is in ac cord with an anticipated reduced variance when residualized approaches control for other variables that are collinear with the predictor (in this instance, concurrent anxiety) of interest ( e.g., depression) T hese findings are in parallel with the signific ant findings between depressive symptoms scores and cortisol
79 variability; however, they suggest that it may be helpful to consider iso depressive and iso anxious symptoms in future studies S imilar approaches have associated the unique effects of anxiety and anger with important health outcomes such as incidence of coronary artery disease (Kubzansky et al., 2006). Examining the unique effect of depressive and anxious symptoms (e.g., iso depression and iso anxiety ) may provide information not captured by th e sum totals of anxiety and depression scores. Moderating Effects of Positive Reframing and Acceptance Neither acceptance nor positive reframing significantly moderated any of the mood cortisol relationships tested. One possible explanation for this is the lack of variance in both of these items taken from the Brief COPE (Carver, 1997). Positive reframing responses ranged from 2 to 8 ( M = 5.26 SD = 1.85), while responses on the Acceptance item on the brief COPE ranged from 3 to 8 ( M = 6.63 SD = 1.33). The possible scores on these two items have a maximum possible score of 8; therefore while this maximum total score from the two items selected was achieved in the current sample, the lower end of possible responses (a r limiti ng the variance and lessening the chances of detecting a moderating effect. This reduced variance is inherent in selecting two items from the larger 32 item Brief COPE measure; however, the author of this instrument dissuades researchers from combining it from which to choose (Carver, 1997). Other work examining the coping process may shed light on combining items or creating appropriate subscales. For instance, in recen t confirmatory factor analyses, it is suggested that the 3 most common distinctions of coping be used; including problem vs. emotion focused coping, approach vs. avoidanc e and cognitive vs. behavioral (Skinner, Edge, Altman, & Sherwood, 2003).
80 These empir responses in addition to extending the current understanding of the coping process. Furthermore, i t may be worthwhile to combine more items of the Brief COPE to increase the over all variance in responses as others have recently done in breast cancer populations (Dedert et al. 2012). If future research were able to incorporate more comprehensive measures of coping, it may provide support for a moderating effect of coping on mood co rtisol relations. Although acceptance did not moderate the relationship between depressed mood and cortisol variability, a marginally significant relationship emerged between acceptance as a main effect and cortisol variability However, t he direction of this relationship suggested that a greater use of acceptance as a coping strategy was related to greater intraindividual cortisol variability ( = .22, p = .072, d = .45 ) which is contrary to hypotheses and recent published literature. For instance, Turner Cobb and colleagues (2010) found that greater acceptance coping was associated with lower cortisol output (AUCg) among caregivers of individ uals with acquired brain injuries at admission, 6 weeks follow up and 6 month follow up These results suggest that greater use of acceptance coping may buffer the effects of caregiver stress on high cortisol output. In the current study sample, acceptance may not buffer depression diagnosis prior to surgery may constitute a maladaptive form of coping. Engaging in a prescribe d medical regimen may be more adaptive in the face of cancer diagnosis. Therefore, while acceptance may buffer relationships between distress and cortisol in caregivers (Turner Cobb et al., 2010),
81 newly diagnosed cancer patients may not benefit from applyi ng acceptance as a method of coping. Alternatively, i t is possible that the original hypotheses of positive reframing and acceptance as moderating any depressive symptom cortisol relationships may not have been warranted, if the definition moderation inclu ded additional criteria put forth by Chmura Kraemer et al., (2008). Baron and Kenny (1986) originally proposed moderation as a conceptual definition under what conditions a moderator (in this case, coping) influences a relationship between a predictor (an xious/depressive symptoms) and an outcome (cortisol variability/slope). However, more recently, the definition of moderation has been expounded upon (Kraemer, Stice, Kazdin, Offord, & Kupfer, 2001 ) with eligibility and analytic criteria. Specifically, elig ibility refers to the temporal precedence (e.g., a moderator precedes the predictor variable), while analytic criteria refers to a statistical relationship between the predictor variable and the moderator (in this case, a relationship between depressive/an xious symptoms and coping). First, in the present study, this would suggest that acceptance or positive reframing would precede the onset of anxious or depr essive symptoms in women underg oing surgery W hen viewed from this framework, coping may be a trai t level or personality variable. However, it may also suggest that there are other, more appropriate candidate variables for moderation, such as social support, spirituality or adaptive health behaviors (e.g., exercise, healthy eating habits etc.) Second the MacArthur definition suggests that a n analytic criteria be established between the predictor and the moderator; namely, that the predictor and moderator be orthogonal (uncorrelated). However, o ur analytic approach of testing moderation, which residua lizes out the unique effect of the
82 interaction term likely controls for this possibility. Thus, this criterion was fulfilled in this present study. However, these criteria highlight the fact that care should be taken to ensure that additional candidate m oderator variables are orthogonal to the predictor variables in future research. Comparisons among Calculations of Cortisol Output Exploratory analyses in this study examined two additional operationalizations of cortisol dysregulation and their relations hip to intraindividual cortisol variability and cortisol slope. Such a comparison has not been conducted in the literature examining HPA axis dysregulation in cancer population, thus examining each of these relationships in detail may be informative. Speci fically, the intraindividual variability estimate generated via the multilevel modeling procedures in the current study while similar to other cortisol variability calculations (Peeters et al., 2004; Young et al., 2004; Posener et al., 2004) has not be en compared to other widely published calculations of cortisol rhythms. In the current study, there was a highly significant relationship between cortisol variability and AUCi r (83) = .62 p < .001. AUCi is thought to represent HPA axis reactivity via cor tisol production throughout the day and it has been shown to be positively correlate d with perceived stress (Pruessner et al., 2003). Thus, the current results suggest that cortisol variability may be related to other, better established cortisol calcula tions that also are related to psychological variables of interest (e.g., perceived stress). Interestingly, there was no significant relationship between intraindividual cortisol variability and the morning to evening ratio of cortisol r (82) = .002 p = .9 9. This is and suggests that it is a marker of cortisol dysregulation in gynecologic cancer Their
83 research demonstrated vegetative symptoms of depression and greater functional disability. If the morning to evening ratio is examined closely, it becomes apparent that the calculation may be a proxy for cortisol slope as the ratio is derived from factoring the dif ference between the morning and evening values of cortisol. Indeed, in the current study, a highly significant relationship was uncovered between a flatter cortisol slope a nd a lower difference in morning evening ratio of cortisol r (82) = .93 p < .001. T hese findings call for future investigators to clarify axis outcome variables, despite recent published work continuing to use this term to describe within day morning to evening cortisol calculat ions (Jain et al. 2012). Even with these inconsistent descriptions in the published literature it may be fruitful to conduct a head to head comparison of psychological factors and these various calculations of cortisol variability in gynecologic cancer po pulations. Moreover, it is noteworthy that the current results demonstrated a significant relationship between intraindividual cortisol variability and AUCi, but no significant relationship between intraindividual cortisol variability and cortisol slope o r the morning to evening ratio, suggesting that the intraindividual variability examined in the current study may be more similar to stress reactivity (e.g., AUCi) than an impaired negative feedback loop of the HPA axis (e.g., cortisol slope). These resul ts underscore the need for investigators to meticulously explain the conceptual and mathematical definitions applied to statistically examining cortisol data. In line with comparing cortisol calculations within the same datasets, Fekedulegn and colleagues (2007) examined various mathematical calculations of cortisol
84 awakening response (CAR). By applying principal component analysis (PCA), the study was able to define two parameters across the calculations of CAR: total hormonal secretion and parameters ass ociated sensitivity or with the time course of cortisol weighting in the PCA analysis. Peak cortisol and area under the curve with respec t to ground had the first and second highe st loadings, respectively on the total hormonal sec retion parameter. Further, area under the curve with respect to increase (AUCi) and measurement (in this study, 45 minutes following wake time) maintained the highest loadings on the sensitivity or time course of cortisol measurement parameter. When comparing these results to our smaller sample and less statistically complex analyses it is interesting that AUCi and intrain dividual cortisol variability were correlated in our sample ( r (83) = .62, p < .001), while AUCi and cortisol slope were associated in the Fekedulegn et al. (2007) study The relationship between AUCi and the morning to evening cortisol ratio approached sig nificance in our sample ( r (82) = .21, p = .06; however, cortisol slope and AUCi were unrelated in our data ( r (83) = .14, p = .20. While the results of Fekedulegn et al. (2007) and our results are not directly comparable, when the current data are interpre ted in light of the results from this recent PCA analysis, it begs the question of what relationships might emerge if component analyses were applied to cortisol calculations in endometrial cancer populations. Clinical Implications of the Current Study The potential clinical implications of the current study are highlighted by the importance of e xamining relations between psychological factors and biological stress systems during the perioperati ve period of surgery for cancer, which has recently been
85 identi fied as a high priority research area (Garssen, Boomsma, & Beelen, 2010). For instance, surgery can promote micrometastases that evade tumor surveillance, which may be promoted by adrenergic excitation and immunosupression during surgery (for review see Be n Eliyahu, 2003). Analysis of monocyte subpopulations also suggests a trend toward immunosupression prior to surgery (Bartal et al., 2010). Moreover, preoperative stress significantly impairs important inflammatory processes at the surgical wound site and can slow time to recovery (Broadbent, Petrie, Alley, & Booth, 2003). Psychological variables, such as depressive/anxious symptoms measured in the current study, may potentiate suppression of such immunological parameters. Similarly, it has been suggested that the perioperative period may represent a time in which psychological interventions may exert a potent effect on dampening endocrine and immune responses (Garssen et al., 2010) and recent reports suggest that a cognitive behavi oral stress management p rogram reduces cortisol and promotes adaptive cytokine regulation in cancer (Antoni, et al., 2009). Thus psychological interventions targeting depression and anxiety may be particularly important during the perioperative period in attenuating possible dec rements in the immune response and thereby influencing surgical outcomes. variability are unclear at this time, the overarching concept of variability may be shared by othe r physiological systems. For instance, circadian rhythmicity plays a central role in functioning of the autonomic nervous system (ANS), and therefore HPA axis and ANS functioning are increasingly studied concurrently (Licht et al., 2010; Stalder, Evans, Hu cklebridge, & Clow, 2011). To this end, investigators compared the cortisol
86 awakening response (CAR) to heart rate and heart rate variability in a group of 42 healthy participants. Results revealed that more elevated CAR responses (thought to represent an over reactivity of the HPA axis) were related to higher heart rate and lower levels of heart rate variability (Stalder et al., 2011). Thus, the concept of variability, examined in the current study by analysis of cortisol output, is seen in other physiolog ical systems, such as heart rate variability. This is particularly salient when considering clinical outcomes, as there is a rich literature linking lower heart rate variability with mortality (Villareal, Liu, & Massumi, 2002). Future research should exam ine cortisol variability longitudinally, as well as other indicators of circadian disruption such as heart rate variability, to examine the association between cortisol variability and cancer outcomes, such as disease free survival. Study Limitations The findings of the present study should be interpreted in light of several limitations. First, the study design was nonexperimental, and as such, causal interpretations cannot be made. Specifically, while it was assumed that the mood symptoms examined i n this study would produce downstream effects on the HPA axis (cortisol slope and variability), the bidrectional relationship between endocrine/immune functioning and mood states is well established (Reiche et al. 2004). As such, the possibility remains th at higher intraindividual cortisol variability may have preceded increases in anxious/depressive symptoms. By only measuring cortisol output across the three days prior to surgery, assessing changes in cortisol longitudinally was not possible. In addition, prior studies have approached intraindividual cortisol variability and mood symptoms from case control approaches (Peeters et al., 2004; Havermans et al., 2010), allowing for group comparisons between clinically significant disturbances in
87 mood (major dep ressive disorder and bipolar disorder) and healthy control groups. While not truly experimental designs, these approaches allow for group comparisons and extend the clinical implications of intraindividual cortisol variability by examining across clinicall y impaired populations. In addition, future research should employ experimental designs to assess causality. Second, although the study design captures three of data, only limited cortisol variance or fluctuation is captured across the 12 time points. Indeed, other studies investigating intraindividual cortisol variability captured plasma cortisol every hour for 24 hours, ensuring a robust estimation of within person variability (Posener et al., 2004). Comprehen sive sampling of cortisol in a design such as this, allows for the examination of the entire diurnal trend in cortisol output; including the cortisol awakening response (CAR), cortisol slope and intraindividual cortisol variability. Cortisol varies, not o nly within person, but also displays dramatic fluctuations within days as well as across days. It has been suggested that diurnal cortisol slope can demonstrate a 100% change within 30 minutes (Kirschbaum & Hellhammer, 1994) further emphasizing the import ance of repeated sampling of cortisol output In summary, a larger sample with more cortisol measurements may yield more robust findings, or possibly, different results. In addition to the limitation of limited variance of the intraindividual cortisol v ariability estimate in the current study, a third limitation is that the mixed model analysis was only able to accurately model a portion of the within day cortisol trend. Therefore, the residuals used to calculate the intraindividual variability may not a ccurately control for the entire time trend seen in the diurnal cortisol variation within the
88 day. With only four time points collected within each day, only the linear and quadratic time trends were modeled (following the k 2 polynomial convention; Singer & Willet, 2003), therefore modeling the diurnal cortisol rhythm may remain incomplete. Indeed, recent work has applied more complex mathematical modeling approaches (e.g., ) observed in diurnal cortisol throughout the day. To collect such data, cortisol is collected continuously every 30 minutes for a 24 hour period. Such robust datasets may allow for stronger conclusions or different results; however, our results are in con cert with these rapid 2004) suggesting a more erratic pattern being related to depressed mood. In addition, applying the residuals to additional mixed model approaches may be war ranted, as others have suggested (Almeida et al., 2009). For example, we were unable to test the mixed models in our data after excluding the 8AM time point due to convergence problems that are likely due to a small sample size. In addition to boosting the overall sample size and number of cortisol samples collected in future investigations, it may also be informative to include different error structures and varying autocorrelations between time points of cortisol collection in calculating intraindividual cortisol variability (Peeters et al., 2004). Fourth, the study design lacked a benign disease only comparison group. Such a comparison group would allow for the ability to control for the aforementioned influence of disease status on HPA axis functioning In the current study, three participants were included in the current data analysis deemed to have benign disease at surgical staging; however, our analyses were re run excluding these participants and the results
89 remained unchanged. A larger comparison group of benign patients would allow for conclusions to be drawn regarding the disease process on HPA axis outcomes. Additionally, inclusion of a non surgical comparison group might also be informative and would allow the investigator to isolate the uniqu e psychological aspects of anticipating associated treatment. Finally, the study sample had a low percentage of racial and ethnic minority women (7% identified themse lves as non Caucasian). Given that African American women with endometrial cancer have poorer survival rates than Caucasian women (Yap indings to women who carry the highest disease burden. Moreover, with such a small percentage of the sample identifying themselves as minorities, it did not allow for us to test the hypothesis that the outcome variables of interest (cortisol variability a nd slope) may be impacted by racial differences. Inde ed, there is a rich literature highlighting cortisol differences between different racial groups. For example, African American and Hispanic youth display flatter cortisol slopes than their Caucasian cou nterparts (DeSantis et al., 2007), and, further, such relati onships may be independent of socioeconomic status (Cohen et al., 2006 ) In Mexican American adolescents, perceived discrimination was related to greater cortisol output (AUC) and a more pronounce d cortisol awakening response (Zeiders, Doane & Roosa, 2012). Extrapolating from this research, these data suggest one possible pathophysiological pathway for racial di sparities in health outcomes. Future research should oversample racial/ethnic minority women affected by gynecologic cancers in
90 order to maximize external validity minority status as a potential covariate in models examining cortisol. Future Directions These limitations wa rrant further d iscussion of potential areas of future research. First, c omparing intraindividual cortisol variability in cancer populations to healthy controls may be an area worthy of future investigation, as cancer populations display HPA axis dysregulation when compar ed to healthy controls (Abercrombie et al., 2004); potentially confounding psychological factors stress system relationships in cancer populations. The challenge of adequately including relevant control variables in research in psychoneuroimmunology has been highlighted by others (Segerstrom, 2009), and cancer severity may be one such confound. For instance, a more advanced stage of disease is related to flatter cortisol rhythm (Abercrombie et al., 2004) and ovarian cancer patients demonstrate a lower mor ning to evening ratio of cortisol (suggestive of a flattened slope) compared to benign cancer patients and healthy controls (Weinrib et al., 2010). As highlighted above, in the data for the current study, the presence of high risk endometrial cancer subty pe was associated with intravindividual cortisol variability, further highlighting that disease status may be related to HPA axis functioning. The inclusion of a noncancer control group in future research would allow fo r such factors to be examined in grou p comparisons. Secondly, f uture investigations may benefit from measuring intraindividual cortisol variability longitudinally, in addition to including measure of depressive symptoms across the perio perative period. For instance, Y avas and colleagues (2012 ) reported that anxiety and depression ratings did not change over the course of a one year treatment period for cervical and endometrial cancer patients ; however, it remains
91 unclear if depressive and anxious symptoms change during the perioperative period in endometrial cancer patients. If depression remained unchanged from the pre to post operative timepoints in the current sample does intraindividual cortisol variability remain unchanged across these two time points as well? Similarly, would the signifi cant relationship between depression and intraindividual cortisol variability uncovered in th e present study remain following surgery? Given the relatively recent application of intraindividual cortisol variability calculations, there is a paucity of liter ature from which to draw hypotheses on cortisol variability change over time. However, Peeters et al. (2004) suggested that HPA axis variability might represent the early stages of HPA axis dysregulation a hypothesis that could only be tested through lon gitudinal study designs. Finally, l ongitudinal study designs may also benefit from comprehensive measurement of stress and/or depressive symptoms throughout the day. One interpretation of the current findings with depression and cortisol variability is tha t a more chaotic HPA axis output is not commensurate with environmental stressors (Yehuda et al., 1996) and thus within day measurements of mood would allow for the This is particularly relevant, as ecological momentary assessment of positive affect has recently been shown to be related to survival in a large cohort ( N = 3,853 ) of older adults, suggesting the utility of continuously measuring not only negative mood symptoms (e.g., depression) but also positive aspects of mood (Steptoe & Wardle, 2011). Further, in a smaller sample of men ( N = 72) positive affect was inversely related to cortisol ( early in the day and a greater increase after waking ; Steptoe, Hamer, Gibson
92 & Wa rdle, 2007), highlighting how measuring psychological variables across the day may contribute to our understanding of HPA axis dysregulation. How such relationships might extrapolate to HPA axis dysregulation and cortisol variability is unknown at the curr ent time; however, such robust collection of within day changes in mood and/or affect may help connect the results of current study to important clinical outcomes, such as survival. Conclusions This study is the first to examine intraindividual cortisol v ariability and mood disturbance in a cancer population. More commonly applied operationalizations of cortisol were also examined with their relation to mood disturbance. Additionally, the study extends relevant work on coping as a potential moderator betwe en mood disturbance and HPA axis dysregulation. Finally, the present study is among the first to conduct head to head comparisons of various calculations of cortisol output in cancer populations In summary, this investigation found that greater depressiv e symptoms were significantly related to greater intraindividual cortisol variability but unrelated to linear cortisol slope in a group of women undergoing surgery for suspected endometrial cancer. Anxious symptoms were not significantly related to either outcome variable of HPA axis dysregulation (e.g., cortisol slope or variability). Further, acceptance and positi ve reframing aspects of coping did not moderate the relationship between depressive symptoms and intraindividual cortisol variability or corti sol slope. Future research should examine whether depressive symptoms may be associated with meaningful negative clinical outcomes in women undergoing surgery for suspected
93 endometrial cancer through its effects on cortisol variability, which may represent a novel index of HPA axis dysregulatio n.
94 APPENDIX A B ECK SCALE FOR SUIDIC AL IDEATION 1) __ 0 = I have a moderate to strong wish to live. 1 = I have a weak wish to live. 2 = I have no wish to live 2) __ 0 = I have no wish to die. 1 = I hav e a weak wish to die 2 = I have a moderate to strong wish to die. 3) __ 0 = My reasons for living outweigh my reasons for dying. 1 = My reasons for living or dying are about equal. 2 = My reasons for dying outweigh my reasons for living. 4) __ 0 = I have no desire to kill myself. 1 = I have a weak desire to kill myself. 2 = I have a moderate to strong desire to kill myself. 5) __ 0 = I would try to save my life if I found myself in a life threatening situation. 1 = I would take a chance on life or death if I found myself in a life threatening situation. 2 = I would not take the steps necessary to avoid death if I found myself in a life threatening situation. Subtotal Part 1_____________________ 6) __ 0 = I have brief periods of thinking about killing myself which pass quickly. 1 = I have periods of thinking about killing myself which last for moderate amounts of time. 2 = I have long periods of thinking about killing myself. 7) __ 0 = I rarely or only occasionally think about killing myself. 1 = I have frequent thoughts about killing myself. 2 = I continuously think about killing myself. 8) __ 0 = I do not accept the idea of killing myself. 1 = I neither accept nor reject the idea of killing myself. Directions: I am going to read you various groups of statements. Please tell me which statement in each group best describes how you have been feeling for the past week, including today
95 2 = I accept the idea of killi ng myself. 9) __ 0 = I can keep myself from committing suicide. 1 = I am unsure that I can keep myself from committing suicide. 2 = I cannot keep myself from committing suicide. 10) __0 = I would not kill myself because of my family, friends, religion possible injury from an unsuccessful attempt, etc. 1 = I am somewhat concerned about killing myself because of my family, friends, religion, possible injury from an unsuccessful attempt, etc. 2 = I am not or only a little concerned about killing myself because of my family, friends, religion, possible injury from an unsuccessful attempt, etc. 11) __0 = My reasons for wanting to commit suicide are primarily aimed at influencing other people, such as getting even with people, making people happier, makin g people pay attention to me, etc. 1 = My reasons for wanting to commit suicide are not only aimed at influencing other people, but also represent a way of solving my problems. 2 = My reasons for wanting to commit suicide are primarily based upon escaping from my problems. 12) __ 0 = I have no specific plan about how to kill myself. 1 = I have considered ways of killing myself, but have not worked out the details. 2 = I have a specific plan for killing myself. 13) __ 0 = I do not have access to a meth od for an opportunity to kill myself. 1 = The method that I would use for committing suicide takes time, and I really do not have a good opportunity to use this method. 2 = I have access or anticipate having access to the method that I would choose for kil ling myself and also have or shall have the opportunity to use it. 14) __ 0 = I do not have the courage or the ability to commit suicide. 1 = I am unsure that I have the courage or the ability to commit suicide. 2 = I have the courage and the ability t o commit suicide. 15) __ 0 = I do not expect to make a suicide attempt. 1 = I am unsure that I shall make a suicide attempt. 2 = I am sure that I shall make a suicide attempt. 16) __ 0 = I have made no preparations for committing suicide. 1 = I have made some preparations for committing suicide. 2 = I have almost finished or completed my preparations for committing suicide. 17) __ 0 = I have not written a suicide note. 1 = I have thought about writing a suicide note or have started to write one, b ut have not completed it.
96 2 = I have completed a suicide note. 18) __0 = I have made no arrangements for what will happen after I have committed suicide. 1 = I have thought about making some arrangements for what will happen after I have committed suici de. 2 = I have made definite arrangements for what will happen after I have committed suicide. 19) __ 0 = I have not hidden my desire to kill myself from people. 1 = I have held back telling people about wanting to kill myself. 2 = I have attempted to hide, conceal, or lie about wanting to commit suicide. 20) __ 0 = I have never attempted suicide. 1 = I have attempted suicide once. 2 = I have attempted suicide two or more times. 21) __ 0 = My wish to die during the last suicide attempt was lo w. 1 = My wish to die during the last suicide attempt was moderate. 2 = My wish to die during the last suicide attempt was high. Subtotal Part 2________________ Total Score__________________ If the patient has previously attempted suicide, please continue with the next question.
97 APPENDIX B BRIEF COPE 2 Not at all A little bit A medium amount A lot take my mind off things. 1 2 3 4 concentrating my efforts on doing 1 2 3 4 1 2 3 4 myself feel better. 1 2 3 4 t from my husband/partner. 1 2 3 4 friends. 1 2 3 4 1 2 3 4 situation better. 1 2 3 4 happened. 1 2 3 4 feelings escape. 1 2 3 4 drugs to help me get through it. 1 2 3 4 with, my spouse/partner to make me feel better. 1 2 3 4 friends to make me feel better. 1 2 3 4 14. light, to make it seem more positive. 1 2 3 4 n trying to come up with a strategy, or plan, about what to do. 1 2 3 4
98 Not at all A little bit A medium amount A lot from my husband/partner. 1 2 3 4 from my friends. 1 2 3 4 the attempt to cope. 1 2 3 4 19. 1 2 3 4 1 2 3 4 less like going to movies, watching TV, reading, daydream ing, sleeping, or shopping 1 2 3 4 fact that this as happened 1 2 3 4 1 2 3 4 spiritual beliefs. 1 2 3 4 25. 1 2 3 4 take. 1 2 3 4 1 2 3 4 1 2 3 4
99 APPENDIX C PSYCHOTIC SCREENING MODUL E OF THE STRUCTURED CLINICAL INTERVIEW FOR DSM IV FOR N ON CLINICAL POPULATIONS Psychotic Symptoms B/C PSYCHOTIC SCREENING MODULE (FOR SCID I/NP OR P W/PSYCHOTIC SCREEN) THIS MODULE IS FOR CODING PSYCHOTIC AND ASSOCIATED SXS THAT HAVE BEEN PRES ENT RESEARCH SETTINGS WHERE THOSE WITH A HISTORY OF PSYCHOTIC SXS THAT ARE NOT DUE TO SUBSTANCE USE OR A GENERAL MEDICAL CONDITION OR THAT OCCUR OUTSIDE THE CONTEXT OF A MOOD DISORDE R ARE TO BE EXCLUDED. INDICATE THE PERIOD OF TIME DURING WHICH THE SYMPTOM WAS PRESENT. WHETHER THE SYMPTOM IS DEFINI A POSSIBLE OR DEFINITE ETIOLOGIC SUBSTANCE (INCLUDING MEDICATIONS) OR GENERAL MEDICAL CONDITION. THE FOLLOWING QUESTIONS MAY BE USEFUL IF THE OVERVIEW HAS NOT ALREADY PROVIDED THE INFORMATION: Just before (PSYCHOTI C SXS) began, were you using drugs? ... on any medications? ...did you drink much more than usual or stop drinking after you had been drinking a lot for a while? ...were you physically ill? IF YES TO ANY: Has there been a time when you had (PSYCHO TIC SXS) and were not (USING DRUGS/TAKING MEDICATION/CHANGING YOUR DRINKING HABITS/ILL)? Now I am going to ask you about DELUSIONS unusual experiences that people False personal beliefs based on incorrect inference about sometimes have. external re ality and firmly sustained in spite of what almost everyone else believes and in spite of what constitutes incontrovertible and obvious proof or evidence to the contrary. The belief is not one ordinarily accepted by other members of the s culture or subculture. Code overvalued ideas (unreasonable
100 and sustained beliefs that are maintained with less than delusional Has it ever seemed like people Delusion of reference, i.e., events, ? 1 2 3 BC1 were talking ab out you or taking objects, or other people in the special notice of you? environment have a particular or 1 3 BC2 IF YES: Were you convinced unusual significance. they were talking about you or POSS/DEF PRI did you think it might have been DESCRIBE: SUBST/GMC MARY your imagination? What about receiving special messages from the TV, radio, or newspaper, or from the way things were arranged around you? Psychotic Symptoms B/C. 2 What about anyone going out of Persecutory delusion, i.e., the ? 1 2 3 BC3 their way to give you a hard time, individual (or his or her group) is or trying to hurt you? being attacked, harassed, cheated, persecuted, or conspir ed against. 1 3 BC4 DESCRIBE: POSS/DEF PRI SUBST/GMC MARY Did you ever feel that you were Grandiose delusion, i.e., content ? 1 2 3 BC5 especially important in some involves exaggerated power, way, or that you had special knowledge or importance, or a powers to do things that other special relationship to a deity or 1 3 BC6 famous person. POSS/DEF PRI
101 DESCRIBE: SUBST/GMC MARY Did you ever feel that something Somatic delusion, i.e., content ? 1 2 3 BC7 was very wrong with you involves change or disturbance in physically even though your body appearance or functioning. doctor said nothing was wrong... 1 3 BC8 like you had cancer or some other DESCRIBE: terrible disease? POSS/DEF PRI SUBST/GMC MARY Have you ever been convinced that something was very wrong with the way a part or parts of your body looked? (Did you ever feel that something strange was happening to parts of your body?) (Did you ever have any unusual Other delusions ? 1 2 3 BC9 religious experiences?) Check if: (Did you ever feel that you had ____ religious delusions 1 3 BC10 committed a crime or done ____ delusions of guilt BC11 something terrible for which you ____ jealous delusions POSS/DEF PRI BC12 should be punished?) ____ erotomanic delusions SUBST/GMC MARY BC13 BC14 DESCRIBE: Psychotic Symptoms B/C. 3 HALLUCINATIONS (PSYCHOTIC) A sensory perception that has the compelling sense of reality of a true perception but occurs without external stimulation of the relev ant HALLUCINATIONS THAT ARE SO TRANSIENT AS TO BE WITHOUT DIAGNOSTIC
102 SIGNIFICANCE) Did you ever hear things that other Auditory hallucinations when fully ? 1 2 3 BC15 s noises, awake, heard either inside or outside or the voices of people whispering of head or talking? (Were you awake at the time?) DESCRIBE: IF YES: What did you hear? How often did you hear it? 1 3 BC16 POS S/DEF PRI SUBST/GMC MARY IF VOICES: Did they comment A voice keeping up a running commentary ? 1 2 3 BC17 on what you were doing or thinking? they occur How many vo ices did you hear? Two or more voices conversing with each ? 1 2 3 BC18 Were they talking to each other? other Did you ever have visions or Visual hallucinations ? 1 2 3 BC19 see things that other people DE SCRIBE: awake at the time?) 1 3 BC20 NOTE: DISTINGUISH FROM AN POSS/DEF PRI ILLUSION, I.E., A MISPERCEPTION SUBST/GMC MARY OF A REAL EXTERNAL STIMULUS. Psychotic Symptoms B/C. 4
103 W hat about strange sensations in Tactile hallucinations, e.g., ? 1 2 3 BC21 your body or on your skin? electricity DESCRIBE: 1 3 BC22 POSS/DEF PRI SUBST/GMC MARY (What about smelling or ta sting Other hallucinations, e.g., ? 1 2 3 BC23 gustatory, olfactory smell or taste?) Check if: 1 3 BC24 ____ gustatory BC25 ____ olfactory POSS/DEF PRI BC2 6 SUBST/GMC MARY DESCRIBE: ? 1 3 BC27 GO TO A PRI NEXT MARY MODULE PSYCHO TIC SX HAS BEEN PRESENT IF A MAJOR DEPRESSIVE OR Psychotic symptoms occur at times ? 1 3 BC28 MANIC EPISODE HAS EVER other than during mood syndromes BEEN PRESENT: Has there ever
104 been a time when you had PSYCHOTIC PSY (PSYCHOTIC SXS) and you SYNDROMES OR PSYCHOTIC MOOD DIS CHOTIC were not (DEPRESSED/ S XS W/O MOOD EPISODES. ORDER. IF DISOR MANIC)? ALLOWED BY DER SYMPTOMS OCCUR STUDY, GO LIKELY EXCLUSIVELY DURING TO NEXT UNEQUIVOCAL MOOD MODULE. SYNDROMES. EXPLORE DETAIL S AND DESCRIBE DIAGNOSTIC SIGNIFICANCE:
105 APPENDIX D RECENT HEALTH BEHAVI ORS QUESTIONNAIRE (RHB) Did you use any of the following during the days of your saliva collection? Substance Usage During Saliva Collection Caffeine (coffee, tea, colas, etc.) 0 = No 1 = Yes Chocolate 0 = No 1 = Yes Figs, bananas, or plantains 0 = No 1 = Yes Alcohol _______________ (# of drinks) Nicotine _____________ (# of cigarettes) Recreational drugs Circle all that apply: Cannabis C ocaine/Amphetamines Hallucinogens Opioids Sedatives/Hypnotics/Anxiolitics Inhalants PCP Antihistamines 0 = No 1 = Yes
106 APPENDIX E S TRUCTURED INTERVIEW GUIDE FOR THE H AMILTON DEPRESSION A ND ANXIETY SCALES questions about the past week. How have you been feeling since last (DAY OF WEEK)? IF OUTPATIENT: Have you been working? IF NOT: Why not? week? Have you been feeling down or depressed? IF YES: Have you been feeling worse in the morning? Sad? Hopeless? In the last week, how often have you felt (OWN EQUIVALENT)? Every day? All day? Have you been crying at all? IF SCORED 1 4 ABOVE, ASK: How long have you been feeling this way? DEPRESSED MOOD (sadness, hopeless, helpless, worthless): 0 = absent 1 = indicated only on questioning 2 = spontaneously reported verbally 3 = communicated non verbally, i.e. facial expression, posture, voice, tendency to weep 4 = VIRTUALLY ONLY these feeling states reported in spontaneous verbal and non verbal communication 1 =POSS/DEF ORG 3 = NOT ORG
107 How have you been spending your time this past week (when not at work)? Have you felt interested in doing (THOSE THINGS), or do you feel you have to push yourself to do them? Have you stopped doing anything you used to do? Is there anything you look forward to? (AT FOLLOW UP : Has your interest been back to normal?) In the last week, have you had trouble concentrating or trouble remembering things? (How much?) How has your appetite been this past week? (What about compared to your usual appetite?) Have you had to force yourself to eat? Have other people had to urge you to eat? WORK AND ACTIVITIES: 0 = No difficulty 1 = thoughts and feeling of incapacity, fatigue, or weakness related to activities, work or hobbies 2 = loss of interest in activity, hobbies, or work by direct report of the patient or indirect in listlessness, indecisi on and vacillation (feels he has to push self to work or activities) 3 = decrease in actual time spent in activities or decrease in productivity. In hosp., pt. spends less than 3 hrs/day in activities (hospital job or hobbies) exclusive of ward chores 4 = stopped working bec. Of present illness. In hospital, no activities except ward chores, or fails to perform ward chores unassisted 1 = POSS/DEF ORG 3 = NOT ORG INTELLECTUAL (difficulty in concentrating, poor memory) 0 = not present 1 = mild 2 = modera te 3 severe 4 = very severe SOMATIC SYMPTOMS GASTROINTESTINAL: 0 = none 1 = loss of appetite but eating without encouragement 2 = difficulty eating without urging 1 = POSS/DEF ORG 3 = NOT ORG
108 How have you been sleeping over the last week? Have you had any trouble falling asleep at the beginning of the night? (Right after you go to bed, how long has it been taking you to fall a sleep?) How many nights this week have you had trouble falling asleep? During the past week, have you been waking up in the middle of the night? IF YES: Do you get out of bed? What do you do? (Only go to the bathroom?) When you get back in bed, are you able to fall right back asleep? Have you felt your sleeping has been restless or disturbed some nights? What time have you been waking up in the morning for the last time, this past week? IF EARLY, Is that with an alarm clock, or do you just wake up yourself? What time do you usually wake up (that is, before you got depressed)? INSOMNIA EARLY: 0 = no difficulty falling asleep 1 = co mplains of occasional difficulty falling asleep i.e., more than hour 2 = complains of nightly difficulty falling asleep 1 = POSS/DEF ORG 3 = NOT ORG INSOMNIA MIDDLE: 0 = no difficulty 1 = complains of being restless and disturbed during the night 2 = waking during the night any getting out of bed (except to void) 1 = POSS/DEF ORG 3 = NOT ORG INSOMNIA LATE: 0 = no difficulty 1 = waking in early hours of morning but goes back to sleep 2 = unable to fall asleep again if gets out of bed 1 = POSS/D EF ORG 3 = NOT ORG
109 In the last week, have you had broken sleep, dreams, or nightmares? Have you felt tired when you wake up? (How bad has that been?) How has your energy be en this past week? Have you been tired all the time? This week, have you had any backaches, headaches, or muscle aches? This week, have you felt any heaviness in your limbs, back or head? In the past week, have you lost interest in things, or no longer enjoyed your hobbies? Have you felt worse in the morning? INSOMNIA (difficulty in falling asleep, broken sleep, unsatisfying sleep and fatigue on waking, dreams, nightmares, night terrors): 0 = not present 1 = mild 2 = moderate 3 = severe 4 = very severe SOMATIC SYMPTOMS GENERAL: 0 = none 1 = heaviness in limbs, back or head. Backaches, headache, muscle aches. Loss of energy and fatigability. 2 = any clear cut symptoms 1 = POSS/DEF ORG 3 = NOT ORG DEPRESSED MOOD (loss of interest, lack of pleasure in hobbies, depression, early waking, d iurnal swing): 0 = not present 1 = mild 2 = moderate 3 = severe 4 = very severe
110 Have you been especially critical of done things wrong, or let others down? IF YES: What have your thoughts been? Have you been feeling guilty ab out (THIS DEPRESSION) on yourself in some way? being sick? This past week, have you had any thoughts that life is not worth living, or about having thoughts of hurting or even killing yourself? IF YES: What have you thought about? Have you actually done anything to hurt yourself? In the last week, how much have you been worrying (not just about ev eryday concerns)? How much have you been worrying about the worst that can happen, or Have you been feeling especially irritable this past week? FEELINGS OF GUILT: 0 = absent 1 = self reproach, feels he has let people down 2 = ideas of guilt or rumination over past errors or sinful deeds 3 = present illness is a puni shment. Delusions of guilt 4 = hears accusatory or denunciatory voices and/or experiences threatening visual hallucinations SUICIDE: 0 = absent 1 = feels life is not worth living 2 = wishes he were dead or any thoughts of possible death to self 3 = suici dal ideas or gesture 4 = attempts at suicide ANXIOUS MOOD (worries, anticipation of the worst, fearful anticipation, irritability) 0 = not present 1 = mild 2 = moderate 3 = severe 4 = very severe
111 Have you been feeling especially tense this past week? Have you b een worrying a lot about little ordinarily worry about? IF YES: Like what, for example? In the past week, how much have you had any of these things: being startled easily, crying easily, trembling, feeling restle ss because of nervousness, not being able to relax? This past week, have you been afraid of the dark, of strangers, of being left alone, of animals, of traffic, or of crowds? IF YES: How afraid? In this past week, have you had any of these ph ysical symptoms (READ LIST PAUSING AFTER EACH SX REPLY)? How much have these things been bothering you this past week? (How bad have they gotten? How much of the time, or how often, have you had them?) TO MEDICATION (E.G., DRY MOUTH AND IMIPRAMINE) ANXIETY PSYCHIC: 0 = no difficulty 1 = subjective tensio n and irritability 2 = worrying about minor matters 3 = apprehensive attitude apparent in face or speech 4 = fears expressed without questioning TENSION (feelings of tension, fatigability, startle response, moved to tears easily, trembling, feelings of re stlessness, inability to relax): 0 = not present 1 = mild 2 = moderate 3 = severe 4 = very severe FEARS (of dark, of strangers, of being left alone, of animals, of traffic or crowds?): 0 = not present 1 = mild 2 = moderate 3 = severe 4 = very severe ANXIETY SOMATIC (physiologic concomitants of anxiety such as C V heart palpitations, headaches RES hyperventilating, sighing 0 = absent 1 = mild 2 = moderate 3 = severe 4 = very severe
112 In the last week, how much have your thoughts been focused on your physical health or how your body is working (compared to your normal thinking)? Do you complain much about how you feel physically? Have you found yourself a sking for help with things you could really do yourself? IF YES: like what for example? How often has that happened? RATING BASED ON OBSERVATION RATING BASED ON OBSERVATION DURING INTERVIEW HYPOCHONDRIASIS: 0 = not present 1 = self absorption (bodily) 2 = preoccupation with health 3 = frequent complaints, requests for help, etc. 4 = hypochondriacal delusions INSIGHT: 0 = acknowledges being depressed and ill OR not currently depressed 1 = acknowledges illness but attributes cause to bad food, climate, o verwork, virus, need for rest, etc 2 = denies being ill at all AGITATION: 0 = none 1 = fidgetiness 2 = playing with bands, hair, etc. 4 = hand wringing nail biting, hair pulling biting of lips
113 Evidence of anxiety during the interview, such as fi dgeting, restlessness or pacing, tremor of hands, furrowed brow, strained face, sighing or rapid respiration, facial pallor, swallowing, etc. RATING BASED ON OBSERVATION DURING INTERVIEW BEHAVIOR AT INTERVIEW (f idgeting, restlessness or pacing tremor of hands, furrowed brow, strained face, sighing or rapid respiration, facial pallor, swallowing, etc.): 0 = not present 1 = mild 2 = moderate 3 = severe 4 = very severe RETARDATION (slowness of thought and speech, impaired ability to concentrate, decreased motor activity): 0 = normal speech and thought 1 = slight retardation at interview 2 = obvious retardation at interview 3 = interview difficult 4 = complete stupor 1 = POSS/DEF ORG 3 = NOT ORG
114 TOTAL HAMILTON DEPRESSION SCORE __________ TO TAL HAMILTON DEPRESSION SCORE IGNORING ORGANIC RATINGS ______ TOTAL HAMILTON ANXIETY SCORE __________
115 LIST OF REFERENCES Abercrombie, H. C., Giese Davis, J., Sephton, S., Epel, E. S., Turner Cobb, J. M., & Spiegel, D. (2004). Flattened cortisol rhythms in metastatic breast cancer patie nts. Psychoneuroendocrinology, 29 (8), 1082 1092. doi: doi:10.1016/j.psyneuen.2003.11.003 Adam, E. K., & Kumari, M. (2009). Assessing salivary cortisol in large scale, epidemiological research. Psychoneuroendocrinology, 34 (10), 1423 1436. doi:10.1016/j.psy neuen.2009.06.011 Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, London, Sage. Amant, F., Moerman, P., Neven, P., Timmerman, D., Van Limbergen, E., & Vergote, I. 2005). Endometrial cancer. Lan cet, 366 (9484), 491 505. doi:10.1016/S0140 6736(05)67063 8 American Cancer Society (2010. March). What are the key statistics about endometrial cancer? Retrieved on April 14 th 2011 from: http://www.cancer.org/cancer/endometrialcancer/detailedguide/endometrial uterine cancer key statistics Antoni, M. H., Lechner, S., Diaz, A., Vargas, S., Holley, H., Phillips, K., et al. (2009). Cognitive behavioral stress management effects on psychosocial and physiological adaptation in women undergoing treatment for breast cancer. Brain Behavior and Immunity, 23 (5), 580 591. doi:10.1016/j.bbi.2008.09.005 Antoni, M. H., Lutgendorf, S. K., Cole, S. W., D habhar, 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 (3), 240 248. doi:10.1038/nrc1820 Bartal, I., Melamed, R., Greenfeld, K., Atzil, S., Glasner, A., Domankevich, V., et al. (2010). Immune perturbations in patients along the perioperative period: alterations in cell surface markers and leukocyte subtypes before and after surgery. Brain Behav ior and Immun ity 24 (3), 376 386. doi: 10.1016/j.b bi.2009.02.010, Beck, A. T., & Steer, R. A. (1991). Manual for the Beck Scale for Suicide Ideation San Antonio, TX: The Psychological Corporation. Beck AT Steer RA and Brown GK (1996) Manual for the Beck Depression Inventory II San Antonio, TX: Psychol ogical Corporation
116 Belanoff, J. K., Kalehzan, M., Sund, B., Fleming Ficek, S. K., & Schatzberg, A. F. (2001). Cortisol activity and cognitive changes in psychotic major depression. American Journal of Psychiatry 158(10), 1612 1616. doi: 10.1176/appi.ajp.1 58.10.1612 Ben Eliyahu, S. (2003). The promotion of tumor metastasis by surgery and stress: Immunological basis and implications for psychoneuroimmunology. Brain Behavior and Immunity, 17 (1S), 27 36. D oi : 10.1016/S0889 1591(02)00063 6 Bhattacharyya, M., Mol loy, G., & Steptoe, A. (2008). Depression is associated with flatter cortisol rhythms in patients with coronary artery disease. Journal of Psychosomatic R esearch 65(2), 107 114. doi: 10.1016/j.jpsychores.2008.03.012 Blackhart, G. C ., L. A. Eckel, et al. (2 007). Salivary cortisol in response to acute social rej ection and acceptance by peers. Biological Psychology 75(3): 267 276. doi :10.3109/02699051003789211 Bokhman, J. V. (1983). Two pathogenetic types of endometrial carcinoma. Gynecologic Oncology 15(1) 10 17. doi: 10.1016/0090 8258(83)90111 7. PMID 6822361. Boscaglia, N., Clarke, D. M., Jobling, T. W., & Quinn, M. A. (2005). The contribution of spirituality and spiritual coping to anxiety and depression in women with a recent diagnosis of gynecological cancer. International Journal of Gynecologic Cancer, 15 (5), 755 761. doi:10.1111/j.1525 1438.2005.00248.x Broadbent, E., Petrie, K. J., Alley, P. G., & Booth, R. J. (2003). Psychological stress impairs early wound repair following surgery. Psychosom atic Me d icine 65 (5), 865 869. D oi : 10.1097/ 01.PSY.0000088589.92699.30 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 (5), 588 601. doi: 35400003176337.0060 Burns, D.D.; Eidelson, R.J. Why are depression and anxiety correlated? A test of the tripartite model. Journal of Consulting and Clinical Psychology, Vol 66(3), Jun 1998, 461 473. doi:10.1037/0022 006X.66.3.461 Carv er C.S. (2007) Brief COPE. Retrieved on March 19 th 2012 from : http://www.psy.miami.edu/faculty/ccarver/sclBrCOPE.html Carver, C. S. (1997). You want to measure coping but your proto col's too long: consider the brief COPE. International Journal of Behavioral Medicine, 4 (1), 92 100. doi: 10.1207/s15327558ijbm0401_6 Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach. J ournal of Personality and Social Psychology 56, 267 283. doi:10.1037/0022 35188.8.131.527
117 Gery, S. and H. P. Koeffler (2009). Circadian rhythms and cancer Cell Cycle 9 (6): 1097 103. doi: 10.4161/cc.9.6.11046 Charlson, M. E., Pompei, P., Ales, K. L., & MacKe nzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of Chronic Disease, 40 (5), 373 383. doi:10.1016/0021 9681(87)90171 8 Chmura Kraemer, H., M Kiernan, M., Marilyn, E. & Kupf er, D.J. (2008). How and why criteria defining moderators and mediators differ between the Baron & Kenny and MacArthur approaches. Health Psychology 27(2S): S101 doi:10.1037/0278 6133.27.2(Suppl.).S101 Cohen, J. (1992). A power primer. Psychological Bul letin, 112 (1), 155 159. doi:10.1037/0033 2909.112.1.155 Cohen, S., J. E. Schwartz, J.E., Epel, E. Kirschbaum, C. Sidney, S. & Seema n, T. Socioeconomic status, race, and diurnal cortisol decline in the Coronary Artery Risk Development in Young Adults (CAR DIA) Study." Psychosomatic Medicine 68(1): 41 50 doi: 10.1097/01.psy.0000195967.51768 Cooper, C., Katona, C., & Livingston, G. (2008). Validity and reliability of the brief COPE in carers of people with dementia: the LASER AD Study. Journal of Nervous & M ental Disease, 196 (11), 838 843. d oi : 10.1097/NMD.0b013e31818b504c Costanzo, E. S., Lutgendorf, S. K., Rothrock, N. E., & Anderson, B. (2006). Coping and quality of life among women extensively treated for gynecologic cancer. Psychooncology, 15 (2), 132 142. doi :10.1002/pon.930 Cruess, D. G., Antoni, M. H., Kumar, M., & Schneiderman, N. (2000). Reductions in salivary cortisol are associated with mood improvement during relaxation training among HIV seropositive men. Journal of Behavioral Medicine, 23 (2), 107 122. doi: 10.1023/A:100541991702 Cull, A., Cowie, V. J., Farquharson, D. I., Livingstone, J. R., Smart, G. E., & Elton, R. A. (1993). Early stage cervical cancer: psychosocial and sexual outcomes of treatment. British Journal of Cancer, 68 (6), 1216 1220. do i:10.1038/bjc.1993.507 Curran, M. Personal communication on behalf of Salimetric s inc. January 12 th 2011. Dedert, E. E. Lush, E.K et al. (2012). Stress, coping, and circadian d isruption a mong Women Awaiting Breast Cancer Surgery. Annals of Behavioral Medi cine: 1 11. doi: 10.1007/s12160 012 9352 y DeSantis, A. S., Adam E.K. et al. (2007). Racial/ethnic differences in cortisol diurnal rhythms in a community sample of adolescents J ournal of Adolesc ent Health 41(1): 3 13. doi: 1 0.1016/j.jadohealth.2007.03.006
118 Eismann, E. A ., Lush, E. et al. (2010). Circadian effects in cancer relevant psychoneuroendocrine and immune pathways Psychoneuroendocrinology 35(7): 963 76. doi : 10.1016/j.psyneuen.2009.12.011 Fekedulegn D. B., M. E. Andrew, M.E. et al. (2007). Area und er the curve and other summary indicators of repeated waking cortisol measurements. Psychosom atic Med icine 69(7): 651 9. doi:10.1097/ PSY.0b013e31814c405c First, Michael B., Spitzer, Robert L, Gibbon Miriam, and Williams, Janet B.W. (November, 2002) : Struc tured Clinical Interview for DSM IV TR Axis I Disorders, Research Version, Non patient Edition. (SCID I/NP) New York: Biometrics Research, New York State Psychiatric Institut e Fries, E., Dettenborn, L., & Kirschbaum, C. (2009). The cortisol awakening respo nse (CAR): facts and future directions. International Journal of Psychophysiology, 72 (1), 67 73. doi:10.1016/j.ijpsycho.2008.03.014 Garssen, B., Boomsma, M., & Beelen, R. H., 2010. Psychological factors in immunomodulation induced by cancer surgery: A rev iew. Biological Psychology 85(1), 1 13. doi: 10.1016/j.biopsycho.2010.05.006, Giese Davis, J., Wilhelm, F. H., Conrad, A., Abercrombie, H. C., Sephton, S., Yutsis, M., et al. (2006). Depression and stress reactivity in metastatic breast cancer. Psychosomat ic Medicine, 68 (5), 675 683. doi: 10.1097/ 01.psy.0000238216.88515.e5 Granger, D. A., Hibel, L. C., Fortunato, C. K., & Kapelewski, C. H. (2009). Medication effects on salivary cortisol: tactics and strategy to minimize impact in behavioral and developmental science. Psychoneuroendocrinology 34(1 0), 1437 1448. doi: /10.1016/j.psyneuen.2009.06.017 Havermans, R., Nicolson, N. A., Berkhof, J., & Devries, M. W. (2010). Patterns of salivary cortisol secretion and responses to daily events in patients with remitted bipolar disorder. Psychoneuroendocrinol ogy, 36 (2), 258 265. doi:10.1016/j.psyneuen.2010.07.016 Hruschka, D. J., Kohrt, B. A., & Worthman, C. M. (2005). Estimating between and within individual variationin cortisol levels using multilevel models. Psychoneuroendocrinology, 30 (7), 698 714. doi: 10.1016/j.psyneuen.2005.03.002 Hultsch, D., & MacDonald, S. (2004). Intraindividual variability in performance as a theoretical window onto cognitive aging. New frontiers in cognitive aging 65 88. Oxford University Press. Innominato, P. F., O. Palesh, et al. (2010). Regulation of circadian rhythms and hypothalamic pituitary adrenal axis: an ove rlooked interaction in cancer. Lancet Oncol ogy 11(9): 816 7. doi:10.1016/S1470 2045(10)70171 6
119 Jain, S., D. Pavlik, et al. (2012). Complementary medicine for fatigue and cortisol variability in breast cancer survivors: a randomized controlled trial. Cancer 118(3): 777 87. doi :10.1002/cncr.26345 Jehn, C. F., D. Kuehnhardt, Bartholomae, A., Pfeiffer, S., Krebs, M., Regierer, A. C., et al. (2006). Biomarkers of depressio n in cancer patients. Cancer, 107 (11), 2723 2729. doi: 10.1002/cncr.22294 Klee, M., & Machin, D. (2001). Health related quality of life of patients with endometrial cancer who are disease free following external irradiation. Acta Oncologica, 40 (7), 816 824. doi:10.1080/02841860152703436 Kraemer, H. C., Giese Davis, J., Yutsis, M., O'Hara, R., Neri, E., Gallagher Thompson, D., et al. (2006). Design decisions to optimize reliability of daytime cortisol slopes in an older population. American Journal of Geriatr ic Psychiatry, 14 (4), 325 333. doi:10.1097/01.JGP.0000201816.26786.5b Kraemer, H. C., Stice, E., Kazdin, A., Offord, D., & Kupfer, D. (2001). How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. Am erican Journal of Psychiatry, 158 (6), 848 856. doi: 10.1176/appi.ajp.158.6.848 Krieger, D. T. (1979). Rhythms in CR H ACTH and corticosteroids. Endocrine Reviews 123 142. Kubzansky, L. D., Cole, S. R., Kawachi, I., Vokonas, P., & Sparrow, D. (2006). Shared and unique contributions of anger, anxiety, and depression to coronary heart disease: a prospective study in the normative aging study. Annals of Behavioral Medicine 31(1), 21 29. doi :10.1207/s15324796abm3101_5 Kumari, M., M. Shipley, et al. (2011). Asso ciation of diurnal patterns in salivary cortisol with all cause and cardiovascular mortality: findings from the Whitehall II study. J ournal of Clin ical Endocrinol ogy and Metab olism 96(5): 1478 85. doi:10.1210/jc.2010 2137 Laudat, M. H., Cerdas, S., Fourni er, C., Guiban, D., Guilhaume, B., & Luton, J. P. (1988). Salivary cortisol measurement: a practical approach to assess pituitary adrenal function. Journal of Clinical Endocrinology and Metabolism 66(2), 343 348. doi: 10.1210/jcem 66 2 343 Lauver, D. R., C onnolly Nelson, K., & Vang, P. (2007). Stressors and coping strategies among female cancer survivors after treatments. Cancer Nursing, 30 (2), 101. doi:10.1097/01.NCC.0000265003.56817.2c Lazarus, R., & Folkman, S. (1984). Stress, appraisal, and coping : Spri nger Publishing Company. New York, N.Y.
120 Licht, C. M., Vreeburg, S. A., van Reedt Dortland, A. K., Giltay, E. J., Hoogendijk, W. J., DeRijk, R. H., et al. (2010). Increased sympathetic and decreased parasympathetic activity rather than changes in hypothalam ic pituitary adrenal axis activity is associated with metabolic abnormalities. Journal of Clinical Endocrinology and Metabolism, 95 (5), 2458 2466. doi:10.1210/jc.2009 2801 Little, T. D., Bovaird, J. A., & Widaman, K. F. (2006). On the m erits of o rthogonal izing p owered and p roduct t erms: i mplications for m odeling i nteractions a mong l atent v ariables. Structural Equation Modeling: A Multidisciplinary Journal 13(4), 497 519. doi:10.1207/s15328007sem1304_1 Lovallo, W. R., Whitsett, T. L., al'Absi, M., Sung, B. H., Vincent, A. S., & Wilson, M. F. (2005). Caffeine stimulation of cortisol secretion across the waking hours in relation to caffeine intake levels. Psychosom atic Med icine 67(5), 734 739. doi:10.1097/ 01.psy.0000181270.20036.06 Lutgendorf, S., Anderson, B., Sorosky, J., Buller, R., & Lubaroff, D. (2000). Interleukin 6 and Use of Social Support in Gynecologic Cancer Patients. International Journal of Behavioral Medicine, 7 (2), 127 142. d oi : 10.1207/S15327558IJBM0702_3 Lutgendorf, S. K., Anderson, B., Ullri ch, P., Johnsen, E. L., Buller, R. E., Sood, A. K., et al. (2002). Quality of life and mood in women with gynecologic cancer: a one year prospective study. Cancer, 94 (1), 131 140. doi: 10.1002/cncr.10155 Lutgendorf, S. K., Weinrib, A. Z., Penedo, F., Russel l, D., DeGeest, K., Costanzo, E. S., et al. (2008 a ). Interleukin 6, cortisol, and depressive symptoms in ovarian cancer patients. Journal of Clinical Oncology, 26 (29), 4820 4827. doi: 10.1200/JCO.2007.14.1978 Lutgendorf, S. K., Lamkin, D. M., Jennings, N. B ., Arevalo, J. M., Penedo, F., DeGeest, K., et al. (2008 b ). Biobehavioral influences on matrix metalloproteinase expression in ovarian carcinoma. Clinical Cancer Research, 14 (21), 6839 6846. doi: 10.1158/1078 0432.CCR 08 0230 Lutgendorf, S. K., Sood, A. K. Anderson, B., McGinn, S., Maiseri, H., Dao, M., et al. (2005). Social support, psychological distress, and natural killer cell activity in ovarian cancer. Journal of Clinical Oncology, 23 (28), 7105 7113. doi: 10.1200/JCO.2005.10.015 MacDonald, S. W., Hult sch, D. F., & Bunce, D. (2006). Intraindividual variability in vigilance performance: does degrading visual stimuli mimic age related "neural noise"? Journal of Clinical Experimental Neuropsychology, 28 (5), 655 675. doi: 10.1080/13803390590954245 Maffei C, Fossati A, Agostoni I, Barraco A, Bagnato M, Deborah D, Namia C, Novella L, Petrachi M. (1997) Interrater reliability and internal consistency of the structured clinical interview for DSM IV axis II personality disorders (SCID II), version 2.0. Journal o f Personality Disorders 11(3):279 84 doi:10.1521/pedi.19184.108.40.2069
121 Massie, M. J. (2004). Prevalence of depression in patients with cancer. Journal of the National Cancer Institute Monographs (32), 57 71. doi: 10.1093/jncimonographs/lgh014 Meyer, B. (2001 ). Coping with severe mental illness: R elations of the b rief COPE with symptoms, functioning, and well being. Journal of Psychopathology and Behavioral Assessment, 23 (4), 265 277. doi: 10.1023/A:1012731520781 Miller, G. E., Chen, E., & Zhou, E. S. (2007). I f it goes up, must it come down? Chronic stress and the hypothalamic pituitary adrenocortical axis in humans. Psychological Bul l et in 133 (1), 25 45. doi:10.1037/0033 2909.133.1.25 Miller, G. E., Cohen, S., & Ritchey, A. K. (2002). Chronic psychological str ess and the regulation of pro inflammatory cytokines: a glucocorticoid resistance model. Health Psychology, 21 (6), 531 541. doi: 10.1037/0278 6220.127.116.111 Mishel, M. H., & Sorenson, D. S. (1991). Uncertainty in gynecological cancer: a test of the mediating functions of mastery and coping. Nursing Research, 40 (3), 167 171. Nakane, T., Szentendrei, T., Stern, L., Virmani, M., Seely, J., & Kunos, G. (1990). Effects of IL 1 and cortisol on beta adrenergic receptors, cell proliferation, and differentiation in cu ltured human A549 lung tumor cells. Journal of Immunology, 145 (1), 260 266. Norton, T., Manne, S., Rubin, S., Carlson, J., Hernandez, E., Edelson, M., et al. (2004). Prevalence and Predictors of Psychological Distress Among Women With Ovarian Cancer. Journ al of Clinical Oncology, 22 (5), 919. doi: 10.1200/JCO.2004.07.028 O'Brien, J. T., Ames, D., Schweitzer, I., Mastwyk, M., & Colman, P. (1996). Enhanced adrenal sensitivity to adrenocorticotrophic hormone (ACTH) is evidence of HPA axis hyperactivity in Alzhe imer's disease. Psychological Medicine, 26 (1), 7 14. doi: 10.1017/S0033291700033675 O'Donovan, A., B. M. Hughes, et al. (2010). Clinical anxiety, cortisol and interleukin 6: evidence for specificity in emotion biology relationships. Brain Behav ior and Immu n ity 24(7): 1074 7. doi: 10.1016/j.bbi.2010.03.003 Onujiogu, N., Johnson, T., Seo, S., Mijal, K., Rash, J., Seaborne, L., et al. (2011). Survivors of endometrial cancer: Who is at risk for sexual dysfunction? Gynecologic oncology 123(2), 356 359. doi:10. 1016/j.ygyno.2011.07.035 Peeters, F., Nicolson, N. A., & Berkhof, J. (2004). Levels and variability of daily life cortisol secretion in major depression. Psychiatry Research, 126 (1), 1 13. doi:10.1016/j.psychres.2003.12.010
122 Penedo, F. J., Gonzalez, J. S., Davis, C., Dahn, J., Antoni, M. H., Ironson, G., et al. (2003). Coping and psychological distress among symptomatic HIV+ men who have sex with men. Annals of Behavioral Medicine, 25 (3), 203 213. doi: 10.1207/S15324796ABM2503_06 Pereira, D.B., (200 4 ) Recent health behaviors q uestionnaire (RHB). Unpublished instrument. Phillips, K. M., Antoni, M. H., Lechner, S. C., Blomberg, B. B., Llabre, M. M., Avisar, E., et al. (2008). Stress management intervention reduces serum cortisol and increases relaxation during treatment for nonmetastatic breast cancer. Psychosomatic Medicine, 70 (9), 1044 1049. doi: 10.1097/PSY.0b013e318186fb27 Pinninti, N., Steer, R., Rissmiller, D. J., Nelson, S., & Beck, A. (2002). Use of the Beck Scale for Suicide Ideation with psychiatric inp atients diagnosed with schizophrenia, schizoaffective, or bipolar disorders. Behaviour research and therapy, 40(9), 1071 1079. doi: 10.1016/S0005 7967(02)00002 5, Porter, L. S., Mishel, M., Neelon, V., Belyea, M., Pisano, E., & Soo, M. S. (2003). Cortisol l evels and responses to mammography screening in breast cancer survivors: a pilot study. Psychosomatic Medicine, 65 (5), 842 848. doi: 10.1097/ 01.PSY.0000088595.91705.C5 Posener, J. A., Charles, D., Veldhuis, J. D., Province, M. A., Williams, G. H., & Schatzberg, A. F. (2004). Process irregularity of cortisol and adrenocorticotropin secretion in men with major depressive disorder. Psychoneuroend ocrinology, 29 (9), 1129 1137. doi:10.1016/j.psyneuen.2004.01.004 Pruessner, J., Kirschbaum, C., Meinlschmid, G., & Hellhammer, D. (2003). Two formulas for computation of the area under the curve represent measures of total hormone concentration versus tim e dependent change. Psychoneuroendocrinology, 28 (7), 916 931. doi:10.1016/S0306 4530(02)00108 7 Purdie, D. M., & Green, A. C. (2001). Epidemiology of endometrial cancer. Best Practice & Research in Clinical Obstetrics Gynaecology, 15 (3), 341 354. doi:10.1 053/beog.2000.0180 Ranganathan, M., Braley, G., Pittman, B., Cooper, T., Perry, E., Krystal, J., et al. (2009). The effects of cannabinoids on serum cortisol and prolactin in humans. Psychopharmacology 203(4), 737 744. d oi : 10.1007/s00213 008 1422 2 : Reich e, E. M., Nunes, S. O., & Morimoto, H. K. (2004). Stress, depression, the immune system, and cancer. Lancet Oncology, 5 (10), 617 625. doi:10.1016/S1470 2045(04)01597 9
123 Rotman Pikielny, P., Roash, V., Chen, O., Limor, R., Stern, N., & Gur, H. G. (2006). Ser um cortisol levels in patients admitted to the department of medicine: Prognostic correlations and effects of age, infection, and comorbidity. Am erican J ournal of the Med ical Sci ences 332(2), 61 67. Ruhe, H. G., Dekker, J. J., Peen, J., Holman, R., & de Jonghe, F. (2005). Clinical use of the Hamilton Depression Rating Scale: is increased efficiency possible? A post hoc comparison of Hamilton Depression Rating Scale, Maier and Bech subscales, Clinical Global Impression, and Symptom Checklist 90 scores. Co mpr ehensive Psychiatry, 46 (6), 417 427. doi: 10.1016/j.comppsych.2005.03.001 Sahar, S. and P. Sassone Corsi (2009). Metabolism and cancer: the circadian clock connection. Nat ure Rev iews Cancer 9(12): 886 96. doi:10.1038/nrc2747 Segerstrom, S. C. (2009). Bi obehavioral controls: threats to psychoneuroimmunology research? Brain Behav ior and Immun ity 23 (7): 885 6. doi: 10.1016/j.bbi.2009.05.053 Sephton, S. E., Sapolsky, R. M., Kraemer, H. C., & Spiegel, D. (2000). Diurnal cortisol rhythm as a predictor of breas t cancer survival. Journal of the National Cancer Institute, 92 (12), 994 1000. doi: 10.1093/jnci/92.12.994 Sephton, S., & Spiegel, D., 2003. Circadian disruption in cancer: a neuroendocrine immune pathway from stress to disease? Brain Behavior and Immunity 17(5), 321 328. doi:10.1016/S0889 1591(03)00078 3 Simonelli, L. E., Fowler, J., Maxwell, G. L., & Andersen, B. L. (2008). Physical sequelae and depressive symptoms in gynecologic cancer survivors: meaning in life as a mediator. Annals of Behavioral Medici ne, 35 (3), 275 284. doi: 10.1007/s12160 008 9029 8 Singer, J. D., & Willet, J. B. (2003). Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence : Oxford University Press, Oxford. Singh, D. (2004). Psychopathology and coping in recently dia gnosed HIV/AIDS patients -the role of gender. South African Medical Journal, 94 (7), 500. Skinner, E. A., Edge, K., Altman, J., & Sherwood, H. (2003). Searching for the structure of coping: a review and critique of category systems for classifying ways of coping. Psychol ogical Bu l l etin 129 (2), 216 269. 10.1037/0033 2909.129.2.216 Snijders, T. A. B. (2005). Power and sample size in multilevel modeling. Encyclopedia of statistics in behavioral science, 3 1570 1573. doi: 10.1002/0470013192.bsa492 SPSS 17.0 (2011 ) Command Syntax Reference 2006, SPSS Inc., Chicago Ill.
124 Stalder, T., Evans, P., Huckl ebridge, F., & Clow, A. (2011). Associations between the cortisol awakening response and heart rate variability. Psychoneuroendocrinology, 36 (4), 454 462. doi: 10.1016/j.psyneuen.2010.07.020, Steptoe, A., & Ussher, M. (2006). Smoking, cortisol and nicotine. International Journal of Psychophysiology 59(3), 228 235. doi: 10.1016/j.ijpsycho.2005.10.011 Steptoe, A., Gibson, E.L., Hamer, M. & Wardle, J. (2007). Neuroendocrine and cardiovascular correlates of positive affect measured by ecological momentary a sses sment and by questionnaire. Psychoneuroendocrinology 32(1): 56 64. doi: 10.1016/j.psyneuen.2006.10.001 Ste ptoe, A. and J. Wardle (2011). Positive affect measured using ecological momentary assessment and s urvival in older men and women. Proceedings of the National Academy of Sciences 108(45): 18244 18248. doi:10.1073/pnas.1110892108 Stewart J., S., T. (2000). John D. and Catherine T. MacArthur Research Network on Socioeconomic Status and Health: Salivary Cortisol Measurement. Retrieved on June 20 th 2011 from: http://www.macses.ucsf.edu/Research/Allostatic/salivarycort.php Thompson, D. S., & Shear, M. K. (1998). Psychiatric disorders and gynecological oncology: a review of the literature. General Hospital Psychiatry, 20 (4), 241 247. doi:10.1016/S0163 8343(98)00030 9 Tollenaar, M. S., Jansen, J., Beijers, R., Riksen Walraven, J. M., & de Weerth, C. (2010). Cortisol in the first year of life: Normative values and intra individual variability. Early Human Development 86(1): 13 6. doi:10.1016/j.earlhumdev.2009.12.003 Touitou, Y., Bogdan, A., Levi, F., Benavides, M., & Auzeby, A. (1996). Disruption of the circadian patterns of serum cortisol in breast and ovarian cancer patients: r elationships with tumour marker antigens. British Journal of Cancer 74(8), 1248 1252. doi:10.1038/bjc.1996.524 Turner Cobb, J. M., Palmer, J., Aronson, D., Russell, L., Purnell, S., Osborn, M., et al. (2010). Diurnal cortisol and coping responses in close relatives of persons with acquired brain injury: a longitudinal mixed methods study. Brain Injury 24(6), 893 903. doi:10.3109/02699051003789211 Vedhara, K., Tuinstra, J., Miles, J. N., Sanderman, R., & Ranchor, A. V. (2006). Psychosocial factors associat ed with indices of cortisol production in women with breast cancer and controls. Psychoneuroendocrinology, 31 (3), 299 311. doi:10.1016/j.psyneuen.2005.08.006 Villareal, R. P., Liu, B. C., & Massumi, A. (2002). Heart rate variability and cardiovascular mort ality. Curr ent Atheroscler osis Rep orts 4 (2), 120 127. doi :10.1007/s11883 002 0035 1
125 Vreeburg, S. A., F. G. Zitman, et al. (2012). Salivary cortisol levels in persons with and without different anxiety disorders. Psychosom atic Med icine 72(4): 340 7. doi:1 0.1097/ PSY.0b013e3181d2f0c8 Weinrib, A. Z., Sephton, S. E., Degeest, K., Penedo, F., Bender, D., Zimmerman, B., et al. (2010). Diurnal cortisol dysregulation, functional disability, and depression in women with ovarian cancer. Cancer, 116 (18), 4410 4419. doi: 10.1002/cncr.25299 Williams, J. (1988). Structured Interview Guide for the Hamilton Depression and Anxiety Scales (SIGH AD). New York State Psychiatric Institute, New York, NY Wu, W., Chaudhuri, S., Brickley, D. R., Pang, D., Karrison, T., & Conzen, S D. (2004). Microarray analysis reveals glucocorticoid regulated survival genes that are associated with inhibition of apoptosis in breast epithelial cells. Cancer Research, 64 (5), 1757 1764. doi: 10.1158/0008 5472.CAN 03 2546 Yap, O. W., & Matthews, R. P. 2006. Racial and ethnic disparities in cancers of the uterine corpus. J ournal of the Nat ional Med ical Assoc ation 98 (12), 1930 1933. Yavas, G., Dogan, N. U., Yavas, C., Benzer, N., Yuce, D., & Celik, C. ( in press ). Prospective Assessment of Quality of Li fe and Psychological Distress in Patients With Gynecologic Malignancy: A 1 Year Prospective Study. International Journal of Gynecologic Cancer doi :10.1002/cncr.20385 Yehuda, R., Teicher, M. H., Trestman, R. L., Levengood, R. A., & Siever, L. J. (1996). Co rtisol regulation in posttraumatic stress disorder and major depression: a chronobiological analysis. Biological Psychiatry 40(2), 79 88. doi: 10.1016/0006 3223(95)00451 3 Young, E. A., Abelson, J., & Lightman, S. L. (2004). Cortisol pulsatility and its ro le in stress regulation and health. Frontiers in Neuroendocrinology, 25 (2), 69 76. doi: 10.1158/0008 5472.CAN 03 2546 Zeiders, K. H., L. D. Doane, et al. (2012). "Perceived discrimination and diurnal cortisol: examining relations among Mexican American adol escents." Hormones and Behavior 61(4): 541 8. doi: 10.1016/j.yhbeh.2012.01.018 Zimmerman, M., Chelminski, I., & Posternak, M. (2004). A review of studies of the Hamilton depression rating scale in healthy controls: implications for the definition of remiss ion in treatment studies of depression. The J ournal of Nerv ous and Ment al Dis ease 192 (9), 595 601. doi: 10.1097/01.nmd.0000138226.22761.39
126 BIOGRAPHICAL SKETCH Tim began his research career working at his alma mater, UNC Chapel Hill, in the Department of P sychiatry. This led to a research fellowship studying mind body medicine with various cancer populations at the National Center for Complementary and Alternative Medicine (NCCAM) in Bethesda, MD. As a graduate student at the University of Florida Tim has focused on the study of psychoneuroimmunology to better understand the processes and mechanisms by which psychological states are related to health and disease. He has remained committed to the clinical care of cancer patients, as well; seeing patients as part of the bone marrow transplant team, multi disciplinary GI oncology team, inpatient rehabilitation and consultation liaison services at UF. Outside of his clinical and academic pursuits, Tim spends his time running, playing ultimate Frisbee and sha ring time with his basset hound, Mitch.