Psychosocial Factors in Diabetes Management 195Stress
Since the seventeenth century, psychological stress has been
suspected to be a psychosomatic factor involved in diabetes.
In the twentieth century, clinical observation and anecdotal
evidence gave way as Walter Cannon (1941) introduced the
experimental study of the effects of stress on diabetes with
his research on stress-induced hyperglycemia in normal cats.
A detailed review of the literature on stress and diabetes is
beyond the scope of this chapter (see Evans, 1985; Surwit &
Williams, 1996); however, we include a brief review of the
research linking stress and the development and management
of type 1 and type 2 diabetes.
Stress in the Etiology of Diabetes
The underlying assumption in type 1 diabetes is that the
stress response in some way disrupts the immune system of
genetically susceptible individuals, making pancreatic beta
cells more vulnerable to autoimmune destruction (Cox &
Gonder-Frederick, 1991). Only 50% of identical twins are
concordant for type 1 diabetes, suggesting that an environ-
mental stimulus may be required for expression of the
disease, although evidence for this mechanism is lacking
(Surwit & Schneider, 1993). There are numerous reports of
the development of type 1 diabetes following major stressful
life events, particularly losses (Robinson & Fuller, 1985).
However, studies of life events have been criticized method-
ologically for lack of controls, small sample sizes, and retro-
spective recall of events (Surwit, Schneider, & Feinglos,
1992). Animal research has provided limited and mixed evi-
dence of an effect for stress in the onset of type 1 diabetes
(Surwit & Schneider, 1993). Surgically pancreatized animals
have been shown to develop either transient or permanent
diabetes after restraint stress (Capponi, Kawada, Varela, &
Vargas, 1980). Studies using another animal model of type 1
diabetes, the diabetes-prone BB Wistar rat, have shown that
the combined effects of behavioral stressors, such as restraint
and crowding, lower the age of diabetes onset (Carter,
Herman, Stokes, & Cox, 1987) and increase the percentage
of animals that became diabetic compared to nonstressed
controls (Lehman, Rodin, McEwen, & Brunton, 1991). How-
ever, because of other endocrine abnormalities in these ani-
mals, generalizability of these “ndings to humans is limited
(Surwit & Schneider, 1993).
Because type 2 diabetes has a concordance rate of almost
100% among identical twins (Sperling, 1988), there is theo-
retically less of an opportunity for stress to play an etiological
role in the incidence of this diabetes type. Retrospective case
studies suggest that stress acts as a triggering factor in the de-
velopment of type 2 diabetes (Cox & Gonder-Frederick,
1991). However, there are no controlled studies of the possi-
ble role of stress in the onset of type 2 diabetes in humans. In
the past 20 years, increasing evidence suggests that the auto-
nomic nervous system is involved in the pathophysiology
of type 2 diabetes (Surwit & Feinglos, 1988). Exaggerated
glycemic reactivity to stress appears to be characteristic of
some humans who are predisposed to developing type 2 dia-
betes, such as the Pima Indians (Spraul & Anderson, 1992;
Surwit, McCubbin, Feinglos, Esposito-Del Puente, & Lillioja,
1990), as well as some animal models of type 2 diabetes
(Mikat, Hackel, Cruz, & Lebovitz, 1972; Surwit, Feinglos,
Livingston, Kuhn, & McCubbin, 1984). The data argue that
expression of hyperglycemia in these genetic animal models
is dependent on exposure to stressful stimuli. However, there
is little evidence to suggest that stress is associated with the
onset of type 2 diabetes de novo (Wales, 1995).Stress and Glycemic ControlIt has been hypothesized that stress has both direct and
indirect effects on BG control in type 1 diabetes. A direct in-
”uence implies that the stress response results in direct
hormonal/neurological effects that can, in turn, affect BG
level. The stress hormones epinephrine, cortisol, and growth
hormones are all believed to raise BG levels (Cox & Gonder-
Frederick, 1991), and it is widely reported that patients with
type 1 diabetes believe that stress affects BG (Cox et al.,
1984). Some human studies have attempted to model the ef-
fects of stress by infusing stress hormones and measuring
glucose metabolism. The data from these studies are fairly
consistent in supporting the notion of a direct and acute con-
nection between stress and BG (Kramer, Ledolter, Manos, &
Bayless, 2000; Sherwin, Shamoon, Jacob, & Sacca, 1984).
However, the infusion paradigm only partially mimics the
complexity of bodily reactions.
Studies involving laboratory stressors with type 1 diabetes
have been less consistent in demonstrating a stress-glycemic
control relationship (e.g., Gonder-Frederick, Carter, Cox, &
Clarke, 1990; Kemmer et al., 1986). Methodological factors
may partially explain the contradictory data, including lack
of control for the prestress metabolic status of the individual
(Cox, Gonder-Frederick, Clarke, & Carter, 1988). Caution
is also warranted in the potential lack of generalizability
between relatively short-lived laboratory stressors that, in
general, induce only modest physiological changes, and
real-world stressors that may be profoundly different in terms
of magnitude, duration, and spectrum (Kemmer et al., 1986).