Associations Between Affect and Glycemia: A Two-Way Street?

F or over three decades, researchers have recognized a link between such processes as stress, coping, and affective disturbances with indices of glycemic control in both type 1 and type 2 diabetes. ¹ Concomitantly, the burgeoning research in psychoneuroimmunology has provided many elegant demonstrations of the complex interplay of psychological experience and neuroendocrine regulation of the immune system. ² Two articles in this issue of Diabetes Technology & Therapeutics (Penckofer et al. ³ and Kruyt et al. ⁴ ) report quite different approaches to clarifying these complex processes.

The basic question seems simple enough: how, and under what conditions, are the type and degree of affect experienced by a patient with diabetes related to that patient's glycemic control? Yet, giving this a bit more thought, the question becomes exceptionally complex, and a complete understanding of these phenomena will probably require correspondingly complex and sophisticated research methodologies and analytic approaches. For example, negative affect may induce hyperglycemia directly through effects on counterregulatory hormones and insulin resistance, as suggested by Kruyt et al. ⁴ But, negative affect may also have indirect glycemic effects by inhibiting effective self-regulation of diabetes. The patient who consistently fails to achieve control of hypoglycemic episodes or of postprandial glucose excursions may become discouraged about the merits of meticulous self-care and anxious about the probability of recurrence of these episodes. Conversely, suboptimal diabetes self-management may cause some patients to become more anxious, depressed, and overwhelmed by their condition. And, it is likely that each of these processes is in play simultaneously in many, if not most, patients' circumstances. The two articles in this issue of DT&T have taken different perspectives of the possible associations between psychological processes and glucose metabolism.

Penckofer et al. ³ examined the associations between scores obtained by 23 women with type 2 diabetes on well-validated questionnaire measures of depression, anxiety, anger, and quality of life with a variety of measures of glycemic variability derived primarily from 24-h continuous glucose monitor profiles. Some associations between certain measures of quality of life and mood and certain measures of glycemic variability were statistically significant. The authors recognized that an assertion that glycemic variability causes affective changes is no more supportable by their data than is the assertion that certain affective states induce greater glycemic variability. But, this observation begs the question: how could the causal direction of these associations, if any, be disentangled? Continuous glucose monitoring, as illustrated by this study, provides a reasonable method for characterizing patients' momentary glucose fluctuations. Studies that concurrently measure patients' momentary affective fluctuations over an interval during which a continuous glucose monitor is being used could be a valuable step forward for this type of research. ⁵ For example, patients could be prompted by a smartphone application to characterize their current affect at specific frequent intervals using an instrument such as the Profile of Mood States ⁶ repeatedly during a period of continuous glucose monitoring. Sophisticated analytic techniques using autocorrelative or time-series methods could be applied to such data.

Kruyt et al. ⁴ reported a study of the acute glycemic effects of stress in the form of bungee jumping on glucose levels, pancreatic β-cell function, insulin resistance, and markers of stress and immune function in 20 healthy, experienced bungee jumpers who were randomized to either receive or not receive 3 days of pretreatment with propanolol. The results showed that this type of acute stress induced acute disturbances of glucose metabolism that were independent of systemic effects on immune function. Other studies of the acute glycemic effects of laboratory-based stressors have tended to show that glycemic effects of stress are often highly idiosyncratic with substantial variation across individuals with diabetes. ¹ Given that observation, one possibility suggested by this study might be to categorize individuals in a controlled setting in terms of their glycemic reactivity to stress and then to follow them in a naturalistic study as above, seeking to quantify the frequency and intensity of daily stressful events and to determine the extent to which those events induce predictable glycemic changes.

Articles such as those discussed here illustrate a challenging set of questions for diabetes researchers. It is hoped that this will lead to more studies that use modern technologies toward a more complete understanding of the complex interactions between psychological and physiological processes in patients with diabetes.