STRESS AND METABOLIC CONTROL IN DIABETES MELLITUS: METHODOLOGICAL ISSUES AND AN ILLUSTRATIVE ANALYSIS 1,2 John R. Kramer, Ph.D., Johannes Ledolter, Ph.D., George N. Manos, Ph.D., and Meg L. Bayless, R.N. University of Iowa ABSTRACT Rationale: The purpose of this article was twofold: a) to review studies of stress and glycemic control in diabetes, and b) to present a data analysis that illustrates the complexities of investi- gating stress in relation to blood glucose. The literature review emphasized human studies and the strengths and weaknesses of alternative designs. Special consideration was given to longitudi- nal investigations, and an analysis of data from the Diabetes Control and Complications Trial (DCCT) was presented to exem- plify this approach. Nine individuals with Type 1 diabetes who participated in this project at the University of lowa were studied over a period of 2 years. Stress was multiply measured (Life Experiences Survey, Hassles Scale, Perceived Stress Scale) as was blood glucose control (daily reflectance meter readings; monthly HbAlc). Within-subject time-series analyses and a combined longitudinal~cross-sectional model were used to analyze data. Two of the nine subjects manifested significant correlations between stress and HbAle, and six subjects exhibited significant associa- tions between stress and daily level or variability of glucose readings. The latter correlations varied in sign and appeared to cluster around specific individuals rather than a particular mea- sure of stress or blood glucose. Conclusion: While the subjects may not represent the full spectrum of individuals with Type i diabetes, results were consistent with earlier longitudinal research in suggesting that the strength and direction of the relationship between stress and blood glucose control varies considerably between individuals. (Ann Behav Med 2000, 22 (1): 17-28) INTRODUCTION Diabetes mellitus, a metabolic disorder characterized by elevated blood glucose levels, can result in microvascular and macrovascular complications, increasing both morbidity and mor- tality and decreasing the quality of life. Type 1 diabetes, previously referred to as IDDM (Insulin Dependent Diabetes Mellitus) or Juvenile Onset Diabetes, results from a lack of beta-cell insulin l Preparation of this manuscript was supported in part by grant DK-30659 from the National Institutes of Health. 2 The authors wish to thank the DCCT Ancillary Studies and DCCT Publications and Presentations Committees which supplied guidance throughout the formulation and execution of this project. We also are grateful to Bonnie Tindal, R.N., Carrie Otepka, and Julie Schwartzendru- ber for their considerable help with data entry. John Tsimikas, Ph.D. is acknowledged for his help with the preliminary analysis of the data. Finally, we thank Kathleen Bucholz, Ph.D. who provided helpful com- ments on an earlier draft of this manuscript. Reprint Address: J. R. Kramer, Ph.D., University of Iowa College of Medicine, Psychiatry Research 1-117 MEB, Iowa City, IA 52242-1000. 2000 by The Society of Behavioral Medicine. production from a cell-mediated, autoimmune destructive process (1). It comprises approximately 10% of all cases of diabetes, affecting predominately children, adolescents, and young adults. Individuals with Type 1 diabetes require exogenous insulin to sustain life, and current clinical practice guidelines recommend frequent monitoring of glycemic status by both the patient and health care provider, nutrition therapy, and at least quarterly evaluations to assess complications (2). Type 2 diabetes, previ- ously referred to as NIDDM (Non-Insulin-Dependent Diabetes Mellitus) or Adult Onset Diabetes, is characterized by insulin resistance and/or deficiency (1). It is by far the most common form of diabetes and affects predominately adults in mid or later life, often in the context of obesity. Individuals with Type 2 diabetes may or may not require exogenous insulin use, which usually is instituted only after modification of diet and a trial of one or more oral hypoglycemic medications. Over the past four decades, researchers have compiled evidence that stress may have an adverse influence on human blood glucose regulation among individuals with both types of diabetes. The potential implications of this phenomenon are considerable. Because chronically high levels of blood glucose have been shown to substantially increase the risk of complications in Type 1 (3) and possibly Type 2 diabetes (2,4), learning to cope with stressful events might improve long-term outcome. This article will address the relationship between stress and glucose control in two ways. First, human studies will be reviewed, with an emphasis on the strengths and weaknesses of different methodologies. A special area of consideration will be longitudinal field investigations. Second, a 2-year time-series analysis of variation in stress and blood glucose among nine individuals with Type 1 diabetes will be performed in order to illustrate the specific issues and complexities involved with this type of research. In this article, the underlying conceptualization of stress is that of an event or condition which is perceived as threatening to one's well-being and which adversely affects thoughts, emotions, behavior, and/or physiological functioning. Although particular researchers may restrict their focus to emphasize either stressors, subject character- istics, or associated physiological processes, we share the widely- held position that stress is a joint function of all these domains (see 5,6). LITERATURE REVIEW Evidence for a link between stress and blood glucose in human diabetes primarily comes from four types of investigations. The body's physiological responses to stressors initially described by Selye and Cannon in the 1930s and 1940s (7,8) now are known to include the activation of hormones such as catecholamines, cortisol, and endogenous opiates. These hormones increase glu- cose production, inhibit the release of insulin, and/or increase insulin resistance; thus, they have the potential to elevate circulat- ing blood glucose levels (9-11). In human infusion studies, investigators have obtained the expected increases in blood 17