Role of Stress in the Etiology and Treatment of Diabetes Mellitus

RICHARD S. SURWIT, PH.D., AND MARK S. SCHNEIDER, PH.D.

Stress has long been suspected as having major effects on metabolic activity. The effects of stress on glucose
metabolism are mediated by a variety of "counter-regulatory" hormones that are released in response to
stress and that result in elevated blood glucose levels and decreased insulin action. This energy mobilizing
effect is of adaptive importance in a healthy organism. However, in diabetes, because of a relative or absolute
lack of insulin, stress-induced increases in blood glucose cannot be adequately metabolized. Thus, stress is
a potential contributor to chronic hyperglycemia in diabetes, although its exact role is unclear. Although
there is some suggestion from retrospective human studies that stress can precipitate type I diabetes, animal
studies are contradictory with different stressors either having facilatory or inhibitory effects upon the
development of the disease. Human investigations in patients with established diabetes are equally confusing
with some showing that stress can stimulate hyperglycemia, hypoglycemia or have no effect at all on
glycemic status. There is more consistent evidence supporting the role of stress in animal models of type II
diabetes. However, human studies on the role of stress on the course of established type II diabetes are few.
Intervention studies suggest that behavioral or pharmacologic intervention to manage stress may contribute
significantly to diabetes treatment, but more long-term research is needed. It is concluded that further
research is needed to establish the importance of behavioral factors in the etiology and management of
diabetes, and several areas of methodologic improvement are suggested.
Key words: Diabetes mellitus, stress, hyperglycemia, hypoglycemia, catecholamines, sympathetic nervous
system.

Diabetes mellitus is a group of disorders that share
a common defect in the control of carbohydrate
metabolism. The term diabetes mellitus refers to the
sweet tasting urine of patients with the disease, a
symptom first described by Areteus over 2000 years
ago. By the end of the 19th century, physicians were
beginning to refer to at least two types of diabetes
mellitus. William Osier (1) described one type of
diabetes, usually appearing in youth and always
fatal, and another, which he termed diabetes of
obesity which was treatable by diet. These two basic
distinctions are still used today and the conditions
are referred to as type I or insulin dependent dia-
betes, and type II or noninsulin dependent diabetes,
respectively. It is estimated that over 14 million
Americans have diabetes with 13 million cases of
type II diabetes and between 500,000 to 1 million
cases of type I diabetes (2). Because type II diabetes
is associated with aging, this form of the disease is
becoming more and more prevalent. Approximately
15% of all American adults over 65 are thought to
have the disease (2).
Type I diabetes is now known to be caused by an
From the Department of Psychiatry, Duke University Medical
Center, Durham, North Carolina.
Address reprint requests to: Richard S. Surwit, Ph.D., Box 3842,
Duke University Medical Center, Durham, NC 27710.
Received for publication June 16, 1992; revision received Oc-
tober 28, 1992
autoimmune process in which the beta, or insulin-
producing cells in the Islets of Langerhans, are tar-
geted and destroyed by the body's own immune
system (3). Without insulin, glucose cannot be uti-
lized by most somatic tissue (the brain is one organ
that does not require insulin to metabolize glucose)
and protein and fat are metabolized instead. Fat
metabolism produces large amounts of ketoacids,
which if left untreated can lead to death. The tend-
ency to develop this immune response is inherited,
but as only 50% of identical twins are concordant
for type I diabetes, other factors such as environment
must play a role (3).
Type II diabetes is probably a heterogenous group
of conditions in which insulin secretion is not ab-
solutely compromised. In some patients, insulin se-
cretion can be greater than normal, the metabolic
defect being that the body is insulin resistant and
does not respond to its own insulin. In other type II
patients, insulin resistance may be minimal, the
primary problem being that the pancreas does not
secrete adequate quantities of insulin in response to
glucose stimulation (4). Both varieties of type II dia-
betes can usually be treated by a controlled diet,
weight reduction (which reduces insulin resistance),
and/or oral medications that stimulate additional
insulin secretion and reduce insulin resistance.
However, insulin therapy can be used for these
patients as well in certain cases.
Conventional theories of the pathophysiology of
type II diabetes suggest that the disease is due to

380 Psychosomatic Medicine 55:380-393 (1993)

0033-3174/93/55O4-O38O$O3.OO/O
Copyright © 1993 by the American Psychosorr
STRESS AND DIABETES

either a primary defect in the beta cell, making it physiology of the disease. In the following pages, we
less responsive to glucose stimulation, or to the will attempt to provide a current overview of the
severe insulin resistance that eventually exhausts research on how stress may contribute to the patho-
beta cell function (5). However, attempts to find a physiology of type I and type II diabetes as well as a
defect in the beta cell or a mechanism for insulin review of the studies that have assessed the utility
resistance in somatic cells have been frustrating. of behavioral and pharmacologic anxiolytic therapy
Examination of genes that code for insulin produc- in the treatment of these diseases. Our purpose is 3-
tion or insulin receptor expression have failed to fold. First, we educate the reader as to the status of
find defects in these functions (6). Over the last 10 our knowledge in this area. Second, we hope to
years, there has been speculation that the autonomic constructively point out the weaknesses in previous
nervous system is involved in the pathophysiology studies that need to be addressed in the future.
of type II diabetes (7-11). This possibility was fore- Above all, we hope to stimulate interest among those
seen by Claude Bernard who found that hypergly- involved in psychosomatic research in what we feel
cemia could be produced in normal rabbits by le- is a most fascinating problem.
sioning the area of the hypothalamus. More recently,
several researchers have shown that hyperglycemia
can be produced by chemical stimulation of the
brain with morphine and by a variety of endogenous EFFECTS OF STRESS ON GLUCOSE
neuropeptides and can be abolished by bilateral METABOLISM
adrenalectomy (9, 12). Hyperglycemia has also been
found to occur from a slow intravenous infusion of Stress hormones are referred to as "counter-regu-
epinephrine (13) and from the type of stress that latory" by endocrinologists because they generally
results in prolonged sympathetic discharge (14). have a hyperglycemic effect. These effects are sum-
Autonomic activity leading to metabolic decompen- marized in Table 1. Adrenergic stimulation of pan-
sation could be stimulated by stress (11, 15), or by creatic islet cells can lead to either facilitation or
the effects of dietary fat and carbohydrate on sym- inhibition of insulin secretion. Beta adrenergic stim-
pathetic outflow (e.g., 16, 17). ulation is facilitory to insulin output while alpha-2
adrenergic stimulation is inhibitory. Beta adrenergic
With proper medical treatment, the consequences stimulation also stimulates glucagon release from
of severe hyperglycemia can be avoided and life can the pancreatic alpha cells. Glucagon, in turn, stim-
be prolonged. However, moderately elevated blood ulates glucose production in the liver. Beta adrener-
glucose levels over a long period may be related to gic stimulation also promotes the conversion of gly-
the eventual appearance of the "long-term" compli- cogen to glucose in the liver as well as fat to free
cations of diabetes. It is widely believed that hyper- fatty acids in adipose tissue (22-24). Branches of the
glycemia is the key factor in the development of parasympathetic vagus nerve innervate the pan-
neuropathy, nephropathy, and probably retinopathy creatic islets as well, and stimulation of the right
(5). It is associated with atherosclerotic disease as vagus causes increased insulin secretion (23). In this
well. Furthermore, intensive treatment regimes de- fashion, both sympathetic and parasympathetic in-
signed to reduce hyperglycemia have had favorable nervation of the pancreas modulate the normal reg-
effects on risk factors such as lipid abnormalities ulation of carbohydrate metabolism.
which may be important in the development of Stressful stimuli can also result in elevated blood
atherosclerosis (18). More recently, investigators
have been studying the relationship between hyper-
tension and diabetes. Between 30 to 50% of patients TABLE 1. Effects of stress hormones on glucose metabolism
with newly diagnosed type II diabetes have hyper- Hepatic _. „, ,
, ,. , Glucose . , . Blood
tension (19). Hormone Insulin glucose
,
... . Lypolysis
. utilization ' r '
,
glucose
As we have previously noted, both type I and type production
II diabetes provide excellent disease paradigms Acetylchohne T T — — T
within which the role of psychosocial stimuli can Norepinephrine I T I T t
Epinephrine T I T T
be evaluated (20, 21). Of all chronic diseases, perhaps Cortisol n T i T T
none demands more compliance on the part of the Beta-Endorphin — — — t
patient than diabetes, and few diseases impact the Growth hormone — I — T
family of the patient as significantly. One of the —
At low levels epinephrine acts as a beta agonist that facilitates
major questions regarding the role of behavior in insulin secretion, but at high levels it is primarily an alpha-2 agonist
diabetes mellitus is how stress may effect the patho- that inhibits insulin secretion.

Psychosomatic Medicine 55:380-393 (1993) 381


glucose levels via several different hypothalamic-
pituitary pathways. Cortisol causes enhanced glu-
cose production by the liver and diminished cellular
glucose uptake. Bjorntorp (15) has hypothesized that
this mechanism may lead to both central obesity
and a predisposition to diabetes if individuals are
exposed to chronic stress. Stress-induced release of
growth hormone and beta-endorphin can also de-
crease glucose uptake, suppress insulin secretion,
and elevate glucose levels (9). There are also vagal
afferents form the liver to the CNS as well as recep-
tors in the brain for numerous peptide hormones
originally found in the gut (25). The adaptive benefit
of the mechanism of stress-induced energy mobili-
zation in healthy individuals is obvious, but in dia-
betic individuals, where glucose metabolism is com-
promised, these stress effects can be problematic.
STRESS AND TYPE I DIABETES
Because diabetes is really a group of very different
diseases, it is important to examine the question of
how stress may affect diabetes onset and control in
relation to specific diagnostic categories. Accord-
ingly, we will review studies on type I and type II
diabetes separately. Because animal models of both
types of diabetes have been studied, and stress ef-
fects have been examined in human clinical studies,
we will review both types of studies for each type of
diabetes.
Animal Models
Animal research has provided evidence to suggest
that stress effects the "onset" of diabetes. Half a
century ago, Cannon (26) described stress-induced
hyperglycemia in normal cats. In this early investi-
gation 12 cats were confined in a holder, for varying
lengths of time, which were dependent on the ani-
mals' reaction to this novel situation. The cats were
given a large quantity of water by stomach tube and
urine was drained promptly. In all of the cases, sugar
was absent from the urine before the animal became
excited, conversely, the stress intervention invaria-
bly resulted in glycosuria. Cannon (26) observed an
apparent relationship between the animals' emo-
tional state and the onset of the hyperglycemia.
Specifically, those animals that seemed to be fright-
ened or enraged developed glycosuria more quickly
than animals that responded to the confinement in
a calm manner. Although Cannon described stress-
induced hyperglycemia in normal animals, there
382
R. S. SURWIT AND M. S. SCHNEIDER
have only been a limited number of studies on the
effects of stress on diabetic animals.
Surgically partially pancreatectomized animals
have been shown to develop either transient or
permanent diabetes following restraint stress (27).
Although these animals did not develop diabetes
spontaneously, restraint stress produced transient or
permanent diabetes in a significant percentage of
pancreatectomized animals. Some animals who
were not surgically altered became hyperglycemic
following the stressor, but none developed diabetes.
More recently, chemical pancreatectomy with B-cell
cytotoxins such as alloxan and streptozotocin has
been used instead of surgical procedures. Through
control of the dosage, partial or complete destruction
of pancreatic beta cell mass can be produced, mim-
icking the clinical picture of type II and type I
diabetes, respectively.
One group of investigators (28) reported that light-
shock stimulation could inhibit the development of
streptozotocin-induced diabetes in young mice re-
ceiving a single dose of streptozotocin, but the mech-
anism of this effect was not defined. Because the
administration of exogenous steroids can inhibit the
development of another streptozotocin-induced
model of diabetes (29), it is possible that the protec-
tive effect of stress on the development of diabetes
observed in animals treated with a single dose of
streptozotocin may have been mediated through the
adrenal corticotropic effects.
There is also a genetic model for type I diabetes,
the diabetes-prone BB Wistar rat (30). These animals
are prone to spontaneously develop an autoimmune
insulitis resulting in diabetes by the time they are 5
months old. Carter et al. (31) found that a combina-
tion of behavioral stressors such as restraint and
crowding could lower the age of onset of diabetes as
well as to increase the percentage of animals who
ultimately developed diabetes. However, BB rats
possess other endocrine and immune abnormalities
that limit the generalizability of these studies to
humans.
Bellush and Rowland (32) found that rats made
diabetic with streptozotocin had reliably higher nor-
epinephrine than controls both before and after foot-
shock. Epinephrine was found to increase in diabetic
animals after stress and decrease in controls. The
authors conclude that the elevated norepinephrine
levels in the diabetic rats suggests increased sym-
pathetic activity in those animals. Lee, Konorska,
and McGarty (33) found similar results in a subgroup
of animals they studied. These studies highlight the
complex nature of the relationship between envi-
Psychosomatic Medicine 55:380-393 (1993)
STRESS AND DIABETES
ronmental manipulations and diabetes in animals
who are susceptible to the disease.
Human Studies
Role of stress in the onset of type 1 diabetes. Stress
has been suspected to play a role in the onset of
Type I diabetes in humans as well. Fifty percent of
identical twins are concordant for type I diabetes.
Because twins usually show evidence of autoim-
mune abnormalities, it has therefore been postu-
lated that some environmental stimulus is necessary
for the overt expression of the disease. Stress may
therefore affect the onset of type I diabetes by trig-
gering this autoimmune abnormality. However, con-
clusive evidence for this mechanism is lacking.
Studies have shown that people with diabetes are
more likely to suffer a major family loss before
diagnosis (34-36). However, these studies tend to be
poorly controlled and/or they rely on recall of spe-
cific life events. One of the better studies that did
compare diabetic patients with similar age nondi-
abetic siblings and a matched neighborhood control
group found that diabetic patients have significantly
more severe life events within the 3 years before
diagnosis than either control group (34).
Stress and gJucose control in type I diabetes. The
investigation of the effects of psychological stress on
glucose metabolism in diabetic patients was first
undertaken by Hinkle and his co-workers (37-39).
Their studies demonstrated increases in blood glu-
cose and ketones following stressful psychiatric in-
terviews in diabetic patients. However, their work
was not tightly controlled and is difficult to inter-
pret. Vandenbergh, Sussman, and Titus (40) exam-
ined the impact of hypnotically induced emotion on
predominantly type I subjects in a controlled fash-
ion. Coincident with nonsignificant increases in
plasma-free fatty acids, they observed blood glucose
decreases. These findings were replicated in a sub-
sequent study (41). However, these investigators did
not fully document whether or not their subjects
had endogenous insulin reserves.
Not all studies have demonstrated that laboratory
stress leads to blood glucose changes in type I dia-
betes patients. In one study, patients with good gly-
cemic control were compared with individuals in
an acute state of insulin insufficiency. Following
mental arithmetic and public speaking stressors, nei-
ther group showed significant changes in blood glu-
cose levels (42). Since artificially induced insulin
deficiency may not be comparable to chronic poor
glycemic control, Gilbert et al. (43) compared chil-
Psychosomatic Medicine 55:380-393 (1993)
dren with good glycemic control with those who had
a history of poor control as measured by glycosolated
hemoglobin. Neither group showed a significant
change in blood glucose after a public speaking
stressor although the poorly controlled patients had
higher poststress urine glucose levels.
Some studies have raised the possibility of idio-
syncratic changes in blood glucose after laboratory
stressors. Bradley (44) found that noise stress in-
creased blood glucose in initially hyperglycemic di-
abetic subjects and decreased blood glucose in ini-
tially hypoglycemic diabetic subjects, although
Carter et al. (45) found that mental arithmetic can
produce both increases or decreases in blood glucose
in patients with type I diabetes and that the direction
of blood glucose change is idiosyncratic. This group
later replicated this finding and demonstrated that
about 50% of patients had clinically and statistically
significant increases or decreases in blood glucose
following an active stressor, mental arithmetic (46).
There were no significant overall changes and no
individual data was reported for the passive stressor,
viewing a gory film. A recent study has suggested
that these idiosyncratic differences in glycemic re-
sponse to stress may be due to differential changes
in injection site blood flow in response to stress (47).
These changes are thought to be mediated by cate-
cholamine activity on local blood flow.
Personality variables have been related to gly-
cemic responsivity to stress in diabetic patients.
Stabler et al. (48) found that type A children showed
elevated blood glucose levels after playing a stressful
videogame while Type B children showed decreased
blood glucose levels. Type A children had higher
glycohemoglobin levels than Type B children. Gly-
cohemoglobin is a glycosylated protein that can be
used as an index of average blood glucose. A subse-
quent study failed to replicate the glycohemoglobin
differences between types A and B children al-
though glucose reactivity to stress was not assessed
(49).
Several physical stressors such as illness or trauma
have been shown to cause hyperglycemia and even-
tual ketoacidosis in patients with type I diabetes
(50). McLesky, Lewis, and Woodruff (51) found a
clear hyperglycemic response in patients with both
types I and II diabetes during a surgical stress. Other
types of negative life events have also been shown
to have an influence on diabetic control. Chase and
Jackson (52) found that in children, life stress is
correlated with both glycosolated hemoglobin and
blood glucose suggesting that it has an impact on
both long- and short-term indices of glycemic con-
trol. Brand, Johnson, and Johnson (53) on the other
383

hand found life stress to only be correlated with
urinary ketone levels in boys. These studies used
different measures of life stress and neither investi-
gated whether the findings suggest a direct effect of
stress on glycemic control or an indirect effect based
on poorer compliance. Using a prospective, in vivo
design, Halford, Cuddihy and Mortimer (54) dem-
onstrated that life stress predicted blood glucose
levels independent of diet in 50% of patients. There
were no significant differences in blood glucose lev-
els between patients who did exhibit a relationship
between stress and blood glucose and those who did
not.
In summary, some of the studies investigating the
impact of psychological stress on blood glucose in
human diabetes report that stress has a general
hyper- or hypoglycemic effect, although others find
that the response is idiosyncratic, with some patients
showing increases and others decreases in blood
glucose in response to stressful stimuli. Some data
exist to suggest that these individual differences are
related to personality variables.
Another way to study the effects of stress in type
I diabetes is to model the effects of stress on metab-
olism by administering stress hormones and meas-
uring the metabolic effects that follow. The advan-
tage of these procedures is that they are easy to
define operationally and results from them tend to
be more replicable than those involving behavioral
stress. Consequently, it is not surprising that the
data on the role of stress hormones (catecholamines,
cortisol, growth hormone), in the development of
hyperglycemia and ketosis in type I diabetes appears
to be more consistent than that from stress studies.
However, because stress orchestrates a complex in-
teraction of many neuroendocrine variables simul-
taneously, the relevance of single neuroendocrine
agonist challenge studies to diabetes must be inter-
preted with caution. Insulin-dependent diabetic
children have been shown to demonstrate elevated
blood glucose and more rapid ketone release follow-
ing epinephrine infusion compared with normal
children (55). Small, intravenous doses of epineph-
rine were shown by Ferngvist, Gunnarsson and
Linde (56) to decrease absorption of insulin in both
healthy and diabetic subjects, even though subcu-
taneous blood flow remained stable or increased. At
levels of epinephrine infusions intended to mimic
moderate physical stress, insulin absorption may be
retarded by as much as 50%. Sherwin et al. (57)
examined the effects on blood glucose of infusions
of epinephrine, glucagon and cortisol both alone and
in combination in normal and type I diabetic sub-
jects. Although the diabetic subjects received insu-
384
R. S. SURWIT AND M. S. SCHNEIDER
lin infusions for several hours before the injection
of the hormones, the elevation of blood glucose was
significantly greater in the diabetic group compared
to normal controls. This increase was sustained in
the diabetic patients over a 5-hour period, but lasted
less than 2 hours in the controls. Cortisol infusions
produced hyperglycemia in both groups with in-
creases in hepatic glucose production observed only
in diabetic patients.
BEHAVIORAL AND PHARMACOLOGIC
ANXIOLYTIC THERAPY IN TYPE I DIABETES
Despite the lack of conclusive evidence of a rela-
tionship between stress and blood glucose, there
have been several attempts to control stress re-
sponses in patients with type I diabetes. Therapeutic
approaches aimed at reducing the stress response
and its metabolic effects in diabetic patients have
included intensive family therapy (58), and long-
term beta blockade (59). However, psychotherapeu-
tic approaches are costly, and long-term beta block-
ade may increase the risk of unheralded hypogly-
cemia and interfere with the recovery from a hy-
poglycemic event in patients with type I diabetes.
Relaxation techniques have offered a potential al-
ternative. Relaxation has been used in the treatment
of a variety of autonomically mediated illnesses (60).
Relaxation seems to decrease adrenocortical activity
(61, 62) as well as circulating levels of catechol-
amines (63, 64). Thus, relaxation therapy might
serve to moderate some of the negative effects of the
stress response on metabolic control in some diabe-
tes patients. Moreover, relaxation is not associated
with the undue adverse effects caused by adrenergic
blockade.
There is a growing literature reporting the results
of experimental applications of relaxation training
in type I diabetes. Fowler, Budzynski, and Vanden-
bergh (65) reported that intensive relaxation training
and EMG biofeedback lowered the insulin require-
ments and reduced the frequency of ketoacidosis in
one patient with type I diabetes but they failed to
provide outcome data. Seeburg and DeBoer (66) re-
ported the treatment of a similar case of type I
diabetes. Although relaxation training appeared to
produce a significant reduction in the patient's in-
sulin requirement, the patient began to experience
frequent hypoglycemia and training had to be dis-
continued. Rosenbaum (67) also reported some suc-
cess with a combined program of EMG biofeedback
and relaxation training with four patients with type
Psychosomatic Medicine 55:380-393 (1993)
STRESS AND DIABETES
I diabetes and two with type II diabetes who were
being treated with insulin.
Although these case reports provide suggestive
evidence for the utility of relaxation training in the
control of plasma glucose in diabetes, they have been
relatively uncontrolled and the patient populations
reported on have been poorly defined. There have
been three recent reports of the application of relax-
ation training to patients with clearly documented
type I diabetes. Landis et al. (68) trained six patients
with type I diabetes who were being closely followed
and intensively treated. Although absolute glucose
levels, glycohemoglobin, and insulin requirements
did not change significantly after training, patients
did appear to experience an improvement in plasma
glucose control as measured by the daily range in
blood glucose levels divided by the daily insulin
dose. Bradley et al. (69), compared conventional
insulin treatment, insulin treatment with a contin-
uous subcutaneous insulin infusion, conventional
insulin treatment plus relaxation training, and con-
ventional treatment plus biofeedback-assisted relax-
ation training. No differences among the four treat-
ment groups were found on daily measures of blood
glucose or glycohemoglobin, although individual
differences within groups were considerable.
Feinglos, Hastedt, and Surwit (70) investigated
whether a relaxation program would affect glycemic
control in 20 patients with type I diabetes who
reported stress-induced elevations of blood glucose.
Patients in the treatment group were given five
biofeedback sessions and instructions to practice
twice daily during a 1-week hospitalization period
and at home after discharge. Relaxation did not
significantly improve glucose tolerance in treated
patients compared to an attention control group.
Furthermore, there were no differences between
treatment and control groups in mean daily blood
glucose, required insulin dose, or other measures of
chronic control.
SUMMARY: STRESS AND TYPE I DIABETES
The weight of the evidence from animal and hu-
man investigations argues persuasively for a role for
stress in the course of type I diabetes. There are,
however, inconsistencies in the literature that may
be attributable to methodological problems. First,
animal models that produce artificially induced di-
abetes, create pathophysiologic conditions that are
not completely analogous to human type I diabetes.
For instance, pancreatectomized animals undergo
potentially stressful procedures that differ from the
Psychosomatic Medicine 55:380-393 (1993)
more gradual onset of type I diabetes in humans
(71). The stress of the sudden onset of the condition
in these animal models may lead to changes in
catecholamines independent of the effects of exper-
imental stressors on the condition itself (33). Second,
in many of these studies, animals often have chronic
hyperglycemia which is not being controlled by
insulin, yielding results that may not be generaliz-
able to humans with type I diabetes in which rea-
sonable glycemic control is maintained with insulin.
Although the human literature contains some
contradictory findings, there are several explana-
tions for apparently disparate results in studies on
the effects of stress in human type I diabetes. First,
the term "stress" is itself a cause of confusion. Stress
has been used to describe a variety of both experi-
mental stimuli and responses. For instance, noise
can be viewed as a stressful stimulus because it can
produce an autonomic response in a passive subject.
However, mental arithmetic is a behavior in re-
sponse to the stimulus of experimental instructions
and thus not a stimulus itself. The presentation of
arithmetic problems does not evoke a response with-
out the subject performing the arithmetic and there-
fore mental arithmetic is best viewed as a response
that is accompanied by an autonomic reaction. In
the experimental studies just cited, few investigators
used the same stress paradigm. In the naturalistic,
correlational studies, the definition of stressful life
events varied as well. Hence, the disparity of results
is not surprising.
A second problem with human studies of stress in
type I diabetes is that baseline measures are often
absent or inadequate, and there are frequently no
checks on the effectiveness of the stress manipula-
tion. Some of the procedures used to control sub-
jects' blood glucose levels at baseline, such as over-
night fasting, may be stressful to the subjects (72).
Furthermore, many of the above studies did not
carefully describe the nature of their subjects' dia-
betes. The effects of a given environmental stimulus
on a patient with some endogenous insulin could be
quite different than if applied to a patient without
endogenous insulin.
Finally, disruptions in autonomic nervous system
activity, due to diabetic neuropathy, can lead to
decreased sympathetic responses in diabetic pa-
tients (73). In patients with known autonomic neu-
ropathy, mental arithmetic stress does not produce
the typical changes in skin temperature and heart
rate that is found in patients without neuropathy
and nondiabetic controls (74). Because these sym-
pathetic nervous system defects are very common
in diabetes, Cryer (73) has argued that increased
385

sympathetic nervous system activity is not likely to
be responsible for hyperglycemia in type I diabetes.
Differences in autonomic status among diabetic sub-
jects may account for the different stress effects
found in various studies, and may also explain why
exogenously administered stress hormones produce
more reliable hyperglycemic effects. Many of these
difficulties are bypassed in studies in which stress
is modeled by the infusion of "stress hormones" and
results from these studies have been more consist-
ent.
Although stress may affect glucose control in type
I diabetes, it is not yet clear that these effects can be
attenuated with behavioral intervention. Most of the
reports of stress-management interventions have
been uncontrolled and have employed few subjects.
Controlled studies have failed to find significant
effects. Individual differences, found to be important
in the effects of stress on blood glucose levels, have
not been investigated. The role of stress-reduction
therapies in the treatment of type I diabetes remains
to be elucidated.
STRESS AND TYPE II DIABETES
Animal Models
Role of stress in the onset of type II diabetes. In-
creased glycemic reactivity to stress and stress hor-
mones is characteristic of several animal models of
type II diabetes. An early observation of stress-in-
duced hyperglycemia in an animal model of spon-
taneously occurring type II diabetes was made dur-
ing metabolic studies of the sand rat (psammomys
obesus). The sand rat is a North African rodent that
develops diabetes when fed laboratory chow (in-
stead of their native diet of succulent plants) and
allowed to become obese (75). Mikat et al. (76) main-
tained sand rats on a low calorie, low carbohydrate
diet of vegetables, and saline so that they remained
euglycemic. Glucose or saline was then adminis-
tered to rats either through an esophageal tube or
intraperitoneally by injection. Similar procedures
were carried out on a group of Sprague-Dawley rats.
Sand rats experiencing esophageal intubation
showed abnormal glucose tolerance which would
be considered typical of diabetes. In contrast, in the
Sprague-Dawley rats intubation did not alter normal
glucose tolerance values. Thus, it seems that the
stress of intubation precipitates glucose intolerance
even in lean, euglycemic sand rats genetically pre-
disposed towards developing diabetes.
Surwit, Feinglos, and colleagues have studied the
386
R. S. SURWIT AND M. S. SCHNEIDER
effects of stress on hyperglycemia in another animal
model of type II diabetes, the genetically obese
mouse (C57BL/6J, ob/ob). This animal is character-
ized by a syndrome of obesity, hyperinsulinemia,
insulin resistance, hyperglycemia, and glucose in-
tolerance (75). However, there has been some con-
troversy over the degree to which the obese mouse
is hyperglycemic. Different laboratories have re-
ported varying "basal" plasma glucose levels ranging
from 130 mg/dl to over 300 mg/dl (77-79). In one
study (80) blood samples were drawn from obese
and lean mice after either a rest period or exposure
to stress. Stress consisted of restraint in a wire-mesh
cage for 60 minutes punctuated by a 5-minute period
of shaking. Both lean and obese animals had an
increase in plasma glucose levels produced by the
stress. This effect was significantly greater in the
obese animals than in their lean, nondiabetic litter-
mates. In a later study, stress hyperglycemia was
easily classically conditioned in C57BL/6J ob/ob
mice, but not in a nondiabetic control strain (81).
Thus, the expression of hyperglycemia in this ge-
netic model of type II diabetes can be seen to be
dependent upon exposure to stressful stimuli.
To study the mechanism by which stress seems to
dysregulate glucose metabolism in obese mice, Sur-
wit et al. (80) administered epinephrine to lean and
obese animals. The results indicated that the effects
of epinephrine were analogous to those of stress.
That is, epinephrine produced an increase in plasma
glucose in all animals with obese mice showing a
greater response than their lean littermates. Like-
wise, epinephrine decreased plasma insulin only in
obese mice. In a follow-up study (82), epinephrine
administration resulted in a dose-related increase in
plasma glucose in lean animals in the dose range
from 1 to 5 ^ig/10 g body weight, while plasma
glucose responses were maximal at the lowest dose
of epinephrine tested in the obese mice, suggesting
that the dose response relationship is altered in
obese animals. Furthermore, phentolamine, an al-
pha adrenergic antagonist, produced a greater in-
crease in plasma insulin in ob/ob mice than the lean
littermates, suggesting that this increased sensitivity
to catecholamines may be largely alpha adrenergic.
Altered peripheral responses to sympathetic stimuli
appear to be important in the exaggerated glycemic
responses of ob/ob mice to stress and may be an
etiologic factor in the development of diabetes in
these animals. Exaggerated sensitivity to alpha ad-
renergic stimulation has also been shown to be char-
acteristic of the KK mouse, another animal model of
type II diabetes (83).
Most recently, Surwit and colleagues (84) have
Psychosomatic Medicine 55:380-393 (1993)
STRESS AND DIABETES
shown that altered adrenergic sensitivity might be
a biologic marker for the development of type II
diabetes in the background strain for the ob/ob
mutation. These investigators observed that lean,
nondiabetic C57BL/6J mice also show exaggerated
glucose response to epinephrine when compared to
several other nondiabetic strains. When placed on a
high-fat, high-simple carbohydrate "junk food" diet,
these mice develop type II diabetes and obesity as
well as hypertension (85). This has led this investi-
gative team to speculate that altered sensitivity to
adrenergic stimulation in the pancreas, liver, and
possibly other sites as well may be related to the
pathophysiology of type II diabetes (84) and hyper-
tension (85).
Human Studies
Role of stress in the onset of type II diabetes in
humans. Despite numerous anecdotal reports of
stress preceding the onset of type II diabetes, there
are no data to argue for or against the importance of
stress in the development of this disease in humans.
However, several recent studies have suggested that
individuals "at-risk" for type II diabetes may have
some of the same autonomic nervous system abnor-
malities that have been identified in animal models.
Pima Indian Native Americans are at extraordinar-
yly high risk for developing type II diabetes; approx-
imately 60% of Pima Indian Native Americans even-
tually develop type II diabetes in adulthood, com-
pared with 5% of the white population. Surwit et al.
(86) have shown that young, euglycemic Pima Indi-
ans showed a disturbed glycemic response to behav-
ioral stress compared with white. Pima Indians and
whites were given a mixed meal that was followed
2 hours later by a 10-minute mental arithmetic
stressor. Although all subjects showed normal glu-
cose tolerance in response to the meal, 10 of 13 Pima
Indians showed a hyperglycemic response to a men-
tal arithmetic stressor although seven of eight white
controls do not. This suggests that Pima Indians who
are "at-risk" for type II diabetes may have exagger-
ated sensitivity to adrenergic stimulation. This con-
clusion was supported by another recent study in
which sympathetic nervous system activity in
young, nondiabetic Pima and white controls (87) was
directly measured. Pima Indians showed signifi-
cantly decreased sympathetic nerve activity to mus-
cle compared with white controls that is consistent
with increased adrenergic receptor sensitivity in the
periphery. Thus, there is some evidence of a preex-
isting autonomic nervous system defect that could
Psychosomatic Medicine 55:380-393 (1993)
allow stress to play a role in mediating the onset of
the disease.
Relationship Between Stress and Glycemic
Control in Type II Diabetes
Although there have been no studies of the role
of stress on the onset of type II diabetes in humans,
a modest literature on the effects of stress on control
of type II diabetes has accumulated over the last 15
years. Grant et al. (88) looked at the role of stress in
terms of the relationship between actual life events
and fluctuations in the course of diabetes mellitus.
Subjects included 15 patients with "juvenile onset"
diabetes and 22 adult-onset diabetes patients. Their
results suggested a relationship between life events,
particularly those of a negative nature, and changes
in diabetic symptomatology. However, direct meta-
bolic effects may have been confounded with stress-
related changes in adherence to the diabetic regi-
men.
McClesky et al. (50) also investigated the effect of
a physical stressor (surgery and anesthesia) on glu-
cose and glucagon levels in both normal and diabetic
patients. The subjects included nondiabetic patients,
those with type I diabetes, and individuals with type
II diabetes who were undergoing elective surgery.
Serum glucagon and glucose levels were repeatedly
sampled during the pre-, intra-, and postoperative
periods. Throughout the time sampling period, the
diabetic patients showed a clear hyperglycemic re-
sponse to the surgical stress. Naliboff, Cohen, and
Sowers (89) studied the effects of postural, hand grip,
and mental arithmetic stressors in a sample of eight
type II patients and eight controls. Although both
physical and mental arithmetic stressors produced
reliable elevations in catecholamines, glucose and
other metabolic parameters failed to change in
either group. However, some of the diabetic subjects
seemed to have autonomic neuropathy that may
have blunted metabolic responses to stress. Finally,
Goetsch et al. (90), did find acute hyperglycemic
effects of mental arithmetic in six type II patients,
but they did not measure other neuroendocrine
parameters and no control group was used.
Although the mechanism of this stress responsiv-
ity in diabetic patients has not been directly studied,
there is evidence of altered adrenergic sensitivity
and responsivity in patients with type II diabetes as
well. Linde and Deckert (91) and Robertson, Halter,
and Porte (92) observed that alpha-adrenergic block-
ade with phentolamine increased glucose-stimu-
lated insulin secretion in patients with type II dia-
387

betes. Thus, as in animal models of this disease,
alpha-adrenergic stimulation may be having a
greater effect on insulin release in diabetic patients
than in normals. This conclusion is supported by
clinical studies in which selective blockade of alpha-
2 receptors increased both insulin secretion and
glucose disposal rate following a mixed meal in
patients with type II diabetes (93). Finally, it has
been reported that glyburide, one of the sulfonylurea
oral agents used to treat type II diabetes, binds to
alpha-2 receptors in the pancreas suggesting that
one effect of this drug on insulin secretion may be
due to antagonism of adrenergic activity (94).
Behavioral and Pharmacologic Anxiolytic
Therapy in Type II Diabetes
The investigation of the effects of stressful stimu-
lation upon diabetes control can also be approached
by investigating the effects of relaxation techniques
that have been designed to reduce or modify an
individual's response to stressful environmental
stimuli. The use of such techniques in the treatment
of diabetes has a long history. Before the discovery
of insulin, it was common to treat diabetes of obesity
that probably corresponds to what we consider to be
type II diabetes today, with rest and opiates (1). More
recently, relaxation techniques have been applied
to the treatment of type II diabetes.
As reported earlier, Rosenbaum (67) demonstrated
some improvement in two patients with type II dia-
betes using a comprehensive stress management
program. Lammers, Naliboff and Straatmeyer (95)
used a multiple baseline single subject design to
investigate specific effect of progressive muscle re-
laxation on blood glucose levels in four insulin re-
quiring type II diabetes patients. Progressive muscle
relaxation significantly lowered blood glucose levels
in two of the four subjects.
Although these studies provide suggestive evi-
dence for the utility of relaxation training in the
control of plasma glucose in diabetes, they were
relatively uncontrolled and the patient populations
were not adequately defined. There has been one
recent study of the controlled application of relaxa-
tion training to patients with clearly documented
type II diabetes. Surwit and Feinglos (96, 97) studied
the acute effects of relaxation training on glucose
tolerance in a carefully defined population of pa-
tients with type II diabetes. Nine men and three
women with type II diabetes in poor control, who
were not using insulin, served as subjects. All pa-
tients were hospitalized for 9 days and diet, exercise,
388
R. S. SURVV1T AND M. S. SCHNEIDER
and other extraneous stimuli and behaviors where
held constant. Half of the patients were assigned to
receive instruction in a modified version of progres-
sive relaxation supplemented by EMG biofeedback.
Six control patients remained in the hospital under
conditions identical to that of the treatment group
except they did not receive relaxation training. Al-
though fasting glucose improved for all subjects with
hospitalization, glucose tolerance improved signifi-
cantly in patients who had received relaxation train-
ing as compared with controls. The improvement in
glucose tolerance was not accompanied by a change
in insulin sensitivity, insulin secretion, or plasma
glucagon. Plasma cortisol and catecholamines were
measured immediately before the first and second
glucose tolerance test. Subjects receiving relaxation
also showed a decrease in plasma cortisol between
the first and second glucose tolerance tests whereas
control subjects showed an increase over the same
period (97). Plasma catecholamines decreased in
both groups from the first to the second test. How-
ever, differences in catecholamine levels between
the groups were not significant.
Recently, Surwit et al. (98) reported the results of
a larger study that attempted to see if a similar
relaxation training program could add incremen-
tally to intensive medical therapy in a sample of 38
patients with type II diabetes. All volunteers were
assigned to a relaxation training and diabetes edu-
cation or diabetes education-alone group. They were
admitted to the hospital for an initial metabolic
assessment and then given outpatient relaxation
training for 8 weeks. Patients were then readmitted
to the hospital. Assessment admissions were re-
peated at 24 and 48 weeks. Although training failed
to significantly improve any measure of glycemic
control (glucose tolerance, fasting glucose, or glyco-
hemoglobin), there was much individual variability
of metabolic response to relaxation training. Meta-
bolic response to relaxation training was predicted
by psychological variables as well as responses to
pharmacologic challenges. Individuals who showed
high trait anxiety and neuroticism scores, and who
showed large glycemic responses to epinephrine
showed greater improvements in glucose tolerance
after relaxation training than patients who showed
the opposite characteristics.
There is also evidence that anxiolytic pharmaco-
therapy can be used to attenuate the effects of stress
on hyperglycemia in type II diabetes. Surwit et al.
(99) gave lean and obese diabetic mice injections of
either alprazolam, a triazolobenzodiazepine, or ve-
hicle alone before exposure to restraint and shaking
stress or to a no stress-control period. Blood samples
Psychosomatic Medicine 55:380-393 (1993)
STRESS AND DIABETES
were then drawn and subsequently analyzed for
glucose, insulin, and corticosterone. The results in-
dicated that during stress alprazolam significantly
lowered plasma glucose in obese but not lean mice,
and it also reduced plasma corticosterone more in
obese mice than in lean mice. The implication from
these findings is that benzodiazepines might have a
role in the modification of stress hyperglycemia in
type II diabetes. These investigators have recently
shown that alprazolam will produce small but reli-
able improvements in glucose tolerance in a sample
of patients with type II diabetes (98). Patients were
given two glucose tolerance tests, one with no other
medication and one after oral administration of 1
mg of alprazolam. Alprazolam significantly reduced
incremental glucose area compared with that ob-
served on the control day.
SUMMARY: EFFECTS OF STRESS ON TYPE II
DIABETES
In general, studies exploring the effects of stress
and stress reduction techniques in type II diabetes
have been somewhat more consistent in finding
positive results. Stress has been shown to affect
glucose acutely and chronically. Numerous animal
studies exist which suggest that stress can adversely
affect glycemic control in this form of the disease.
Furthermore, there is evidence from both animals
and humans which suggests that individuals with
type II diabetes have altered adrenergic sensitivity
in the pancreas (and perhaps other sites as well)
which could make them particularly sensitive to
stressful environmental stimulation. But, while
there is substantial data showing the importance of
stress in understanding hyperglycemia in animal
models of type II diabetes, there is no direct evidence
that stress plays a clinically significant role in the
expression or control of the human disease. Fur-
thermore, although there are some studies that show
that relaxation can attenuate hyperglycemia in pa-
tients, these effects are not dramatic and have only
been demonstrated in small numbers of patients.
Human research on the role of stress in type II
diabetes has been less impressive than animal stud-
ies. Human stress studies and stress reduction inter-
vention trials are difficult to perform and most stud-
ies have only used a small number of subjects. As
in the study of type I diabetes, methodological dif-
ferences between studies make results difficult to
compare from one study to the other. Finally, al-
though the sympathetic nervous system may play a
role in the etiology of type II diabetes, glucose tox-
Psychosomatic Medicine 55:380-393 (1993)
icity can eventually destroy the insulin secretory
capability of the pancreas in humans (100) and, as
in type I diabetes, neuropathy that results from
chronic hyperglycemia (73) may make stress and
neural factors less important in established disease
than it is in developing disease.
SUMMARY AND CONCLUSIONS
Speculation regarding the role of stress in the
onset and course of diabetes has been ongoing for
over 300 years. Although the early literature did not
clearly distinguish type I from type II diabetes, later
studies have been more careful to examine sepa-
rately the effects of stress on these two, very differ-
ent diseases. There is only limited evidence to sup-
port the notion that stress may impact the onset of
type I diabetes. The literature on the effects of stress
on the course of the disease in experimental models
and in clinical diabetes is complicated. In animal
models, the effects of stress seem to depend upon
the type of stress and the particular model studied.
Also, because animal models of type I diabetes are
rarely treated with insulin, they are metabolically
unstable. It is, therefore, not immediately apparent
how relevant these studies are to clinical problems.
Human studies are confusing with some studies
showing hyperglycemic effects, some showing no
effects and some showing idiosyncratic effects, with
patients demonstrating either hyper or hypoglyce-
mia in response to the same stressor. As in animal
studies, metabolic stability is often a confounding
variable in that the degree of blood glucose control
in the subject sample varies from study to study.
Furthermore, insulin treatment regimens often dif-
fer among subjects, both within and across studies
adding a further confounding variable. The temporal
relationship between insulin dosage, meals, and
stress must be carefully controlled in order for the
metabolic effects of a behavioral manipulation to be
interpretable. A final variable confounding most of
these clinical observations is autonomic neuropathy,
which develops over time and which would com-
promise any sympathetic nervous system response
to stress (73). There are little data on effects of stress
reduction therapies in type I diabetes, and those that
exist are inconclusive. More data are needed before
we can conclude much about the effects of stress on
type I diabetes control or how these effects would
best be treated.
Although the literature on the effects of stress on
type II diabetes is no more extensive than that of
type I, results of both human and animal studies
389

have been somewhat more consistent. This may be,
in part, because of the fact that glucose control is
more stable in these patients, but other factors have
also been proposed. Stress and stress hormones have
been typically shown to have a hyperglycemic effect
on individuals and, in contrast to type I diabetes,
stress-induced hypoglycemia has not been reported.
Furthermore, both animal and human studies pro-
vide evidence of autonomic nervous system abnor-
malities in the etiology of type II diabetes in which
there is an exaggeration of sympathetic nervous
system effects upon glucose metabolism. This has
led to the development of several theoretical models
in which the sympathoadrenal/cortical system play
a central role in the development of the disease.
Several groups of investigators have speculated that
behavioral stimuli that excite the autonomic nerv-
ous system and adrenocortical system may impair
insulin secretion and glucose utilization as the dis-
ease develops (15, 101) and evidence is emerging
that suggests that prediabetic individuals may be
acutely susceptible to this type of stimulation. How-
ever, as in type I diabetes, the clinical impact of
stress in the course of the disease is still far from
clear. Long-term evaluations of the effects of stress
on diabetes control in large numbers of patients are
needed before the importance of stress in the course
of the disease is understood. Similarly, more well-
controlled studies evaluating the efficacy of stress-
reducing behavioral or pharmacologic interventions
as an adjunct to treatment of this disease are needed
before the clinical importance of stress management
in the treatment of either type I or type II diabetes
can be established.
Future studies to assess the effects of stress on
either type I or type II diabetes can be improved by
attention to several problems common to past re-
search. First, because of the diversity of stressful
stimuli that have been used in the studies that make
up the literature, it is difficult to generalize from
one study to the next. This problem can be solved
by some standardization of stress-induction para-
digms analogous to what has taken place in cardio-
vascular research. The addition of neuroendocrine
data to substantiate the effects of stress on counter-
regulatory hormone activity would also be helpful.
Changes in catecholamine and cortisol levels, for
instance, could be expected in subjects that have
been adequately stressed. Evaluating the negative
stress effects on glucose metabolism, without these
neuroendocrine data, is presently not possible. The
evaluation of stress-management interventions as a
treatment for diabetes requires the use of appropri-
ate nonspecific control conditions. Because it is well
390
R. S. SURYVIT AND M. S. SCHNEIDER
known that any intervention is likely to focus the
patient's attention on adherence to the treatment
regimen, relaxation, and biofeedback treatments
need to be compared with nonspecific "attention-
placebo" conditions, not to "no intervention at all."
Finally, previous data strongly suggests that gly-
cemic response to stress, as well as to stress-manage-
ment interventions, is likely an individual differ-
ence, related to measurable psychometric con-
structs. Future research aimed at susceptible
individuals should be much more fruitful than pre-
vious studies using heterogeneous groups of unse-
lected subjects. It is therefore likely that as the
methodology of future studies improves, the impor-
tance of stress in our understanding of diabetes will
come into sharper focus.
Preparation of this manuscript was supported by
Research Scientist Development Award MH. 00303,
National Institutes of Health Grants ROl DK42923
and ROl DK43106, and National Institutes of Mental
Health training Grant 5T 32MHW109-03.
REFERENCES
1. Osier W: The Principles and Practice of Medicine. New York,
Appleton, 1892
2. National Diabetes Data Group: Diabetes in America, Be-
thesda, U.S. Department of Health and Human Services.
Publication 85-1468,1985
3. Palmer JL, McCulloch DK: Prediction and prevention of
IDDM. Diabetes 40:943-947,1991
4. Banerji MA, Lebovitz HE: Insulin-sensitive and insulin-re-
sistant variants in NIDDM. Diabetes 38:784-792, 1989
6. Bell GI: Molecular defects in diabetes mellitus. Diabetes
40:413-422. 1991
7. Feldberg W, Pike DA, Stubbs WA: On the origin of nonin-
sulin-dependent diabetes. Lancet 1:1263-1264,1985
8. Giugliano D: Morphine, opioid peptides, and pancreatic islet
function. Diabetes Care 7:92-98, 1984
9. Guillemin R: Hypothalamus, hormones and physiological
regulation. In Debbs-Robin E (ed), Claude Bernard and the
Internal Environment: A Memorial Symposium. New York,
Dekker, 1978, 137-156
10. Ipp E. Dhorajiwala J, Pugh W, Moossa AR, Rubinstein AH:
Effects of an enkephalin analog on pancreatic endocrine
function and glucose homeostasis in normal and diabetic
dogs. Endocrinology 111:461-463, 1982
11. Surwit RS, Feinglos MN: Stress and autonomic nervous
system in type II diabetes: A hypothesis. Diabetes Care
11:83-85, 1988
12. Van Loon GR, Appel N: Endorphin-induced hyperglycemia
is mediated by increased central sympathetic outflow to
adrenal medulla. Brain Res 204:236-241, 1981
13. Lori CF, Lori GT, Buchwald KW: The mechanism of epi-
nephrine action vs. changes in blood sugar, lactic acid, and
blood pressure during continuous intravenous injection of
epinephrine. Am J Physiol 93:273-283, 1930
Psychosomatic Medicine 55:380-393 (1993)
STRESS AND DIABETES
14. Woods SC, Smith PH, Porte D: The role of the nervous system
in metabolic regulation and its effects on diabetes and obe-
sity. In Brownless M (ed), Handbook of Diabetes Mellitus,
Vol. 3. New York, Garland, 1981, 208-271
15. Bjorntorp P: Metabolic implications of body fat distribution.
Diabetes Care 14:1132-1143, 1991
16. Berne C, Fagius J, Niklasson F: Sympathetic response to oral
carbohydrate administration. J Clin Invest 84:1403-1409,
1989
17. Kaufman LN, Peterson MM, Smith SM: Hypertension and
sympathetic hyperactivity induced in rats by high-fat or
glucose diets. Am J Physiol 26O:E95-E1OO, 1991
18. Schade DS, Santiago JV, Skyler JS, Rizza RA: Intensive In-
sulin Therapy. Princeton, Excerpta Medica, 1983
19. Fuller JH: Epidemiology of hypertension associated with
diabetes mellitus. Hypertension 7(Suppl. II):II-3-7, 1985
20. Surwit RS, Feinglos MN, Scovern AW: Diabetes and behav-
ior: A paradigm for health psychology. Am Psychol 38:255-
262, 1983
21. Surwit RS, Scovern AW, Feinglos MN: The role of behavior
in diabetes care. Diabetes Care 26:337-342, 1982
22. Clutter WE, Bier DM, Shah SD, Cryer PE: Epinephrine
plasma metabolic clearance rates, and physiologic thresh-
olds for metabolic and hemodynamic actions in man. J Clin
Invest 66:74-101, 1980
23. Williams RH, Porte D: The pancreas. In Williams RH (ed),
Textbook of Endocrinology. 5th ed. Philadelphia, Saunders,
1974,502-626
24. Landsberg L, Young JB: Sympathoadrenal system: The reg-
ulation of metabolism. In Ingbar SH (ed), Contemporary
Endocrinology, Vol. 2. New York, Plenum, 1985, 217-246
25. Whitehead WE, Schuster MM: Gastrointestinal Disorders:
Behavioral and Physiologic Basis for Treatment. New York,
Academic Press, 1985
26. Cannon WB: Bodily Changes in Pain, Hunger, Fear and Rage.
New York, MacMillan, 1941
27. Capponi R, Kawada ME, Varela C, Vargas L: Diabetes mel-
litus by repeated stress in rats bearing chemical diabetes.
Horm Metab Res 12:411-412, 1980
28. Huang SW, Plaut SM, Taylor G, Wareheim BA: Effect of
stressful stimulation on the incidence of streptozotocin in-
duced diabetes in mice. Psychosom Med 43:431-437,1981
29. Roudier M, Portha B, Picon L: Glucocorticoid-induced re-
covery from streptozotocin diabetes in the adult rat. Diabetes
29:201-205,1980
30. Nakhooda AF, Like AA, Chappell CI, Wei CN, Marliss EB:
The spontaneously diabetic Wistar rat (the "BB" rat). Studies
prior to and during the development of the overt syndrome.
Diabetologia 14:199-207, 1978
31. Carter WR, Herman J, Stokes K, Cox DJ: Promotion of dia-
betes onset by stress in BB rat. Diabetologia 30:674-675,1987
32. Bellush LL, Rowland NE. Stress and behavior in streptozo-
tocin diabetic rats: Biochemical correlates of passive avoid-
ance learning. Behav Neurosci 103:144-150, 1989
33. Lee JH, Konorska M, McCarty R: Physiological responses to
acute stress in alloxan and streptozotocin diabetic rats. Phys-
iol Behav 45:483-489, 1989
34. Robinson N, Fuller JH: Role of life events and difficulties in
the onset of diabetes mellitus. J Psychosom Res 29:583-591,
1985
35. Slawson PF, Flynn WR, Kollar EJ: Psychological factors as-
sociated with the onset of diabetes mellitus. JAMA 185:166-
170, 1963
36. Stein SP, Charles E: Emotional factors in juvenile diabetes
Psychosomatic Medicine 55:380-393 (1993)
mellitus: A study of early life experience of adolescent
diabetics. Am J Psychiatry 128:700-704,1971
37. Hinkle LE, Evans FM, Wolf S: Studies in diabetes mellitus
III: Life history of three persons with labile diabetes, and the
relation of significant experiences in their lives to the onset
and course of their disease. Psychosom Med 13:160-183,
1951
38. Hinkle LE, Evans FM, Wolf S: Studies in diabetes mellitus
IV: Life history of three persons with relatively mild, stable
diabetes, and relation of significant experiences in their lives
to the onset and course of the disease. Psychosom Med
1:184-202, 1951
39. Hinkle LE, Wolf S: Importance of life stress in course and
management of diabetes mellitus. JAMA 148:513-520, 1952
40. Vandenbergh RL, Sussman KE, Titus CC: Effects of hypnot-
ically induced acute emotional stress on carbohydrate and
lipid metabolism in patients with diabetes mellitus. Psycho-
som Med 28:382-390,1966
41. Vandenbergh RL, Sussman KE, Vaughan GD: Effects of com-
bined physical anticipatory stress on carbohydrate-lipid me-
tabolism in patients with diabetes mellitus. Psychosomatics
8:16-19, 1967
42. Kemmer F, Bisping R, Steingruber H, Baar H, Hardtmann R,
Schlaghecke R, Berger M: Psychological stress and metabolic
control in patients with type I diabetes mellitus. N Engl J
Med 314:1078-1084, 1984
43. Gilbert BO, Johnson SB, Silverstein J, Malone J: Psychological
and physiological responses to acute laboratory stressors in
insulin-dependent diabetes mellitus adolescents and non-
diabetic controls. ] Pediatr Psychol 14:577-591, 1989
44. Bradley C: Psychophysiological aspects of the management
of diabetes mellitus. Int J Mental Health 11:117-132, 1982
45. Carter WR, Gonder-Frederick LA, Cox DJ, Clarke WL, Scott
D: Effects of stress on blood glucose in IDDM. Diabetes Care
8:411-412, 1985
46. Gonder-Frederick LA, Carter WR, Cox DJ, Clarke WL: En-
vironmental stress and blood glucose change in insulin-
dependent diabetes mellitus. Health Psychol 9:503-515,
1990
47. Greenhalgh PM, Jones JR, Jackson CA, Smith CCT, Yudkin
JS: Changes in injection-site blood flow and plasma-free
insulin concentrations in1 response to stress in type I diabetic
patients. Diabetic Med 9:20-29, 1992
48. Stabler B, Surwit RS, Lane JD, Morris MA, Litton J, Feinglos
MN: Type A behavior pattern and blood glucose control in
diabetic children. Psychosom Med 49:313-316,1987
49. Stabler B, Lane JD, Ross SL, Morris MA, Litton J, Surwit RS:
Type A behavior pattern and chronic glycemic control in
individuals with IDDM. Diabetes Care 11:361-362, 1988
50. Travis L: Stress, hyperglycemia, and ketosis. In Travis L,
Brouhard BH, Schreiner BD (eds), Diabetes Mellitus in Chil-
dren and Adolescents. Philadelphia, Saunders, 1987, 137-
146
51. McLesky CH, Lewis SB, Woodruff RE: Glucagon levels dur-
ing anesthesia and surgery in normal and diabetic patients.
Diabetes 27:492, 1978
52. Chase HP, Jackson GG: Stress and sugar control in children
with insulin-dependent diabetes mellitus. J Pediatr 98:1011-
1013,1981
53. Brand AH, Johnson JH, Johnson SB: Life stress and diabetic
control in children and adolescents with insulin-dependent
diabetes. J Pediatr Psychol 11:481-495, 1986
54. Halford WK, Cuddihy S, Mortimer RH: Psychological stress
391

and blood glucose regulation in type I diabetic patients.
Health Psychol 9:516-528, 1990
55. Baker L, Kaye R, Haque N: Studies on metabolic homeostasis
in juvenile diabetes mellitus II: Role of catecholamines.
Diabetes 16:504-505, 1967
56. Fernqvist E, Gunnarsson R, Linde B: Influence of circulating
epinephrine on absorption of subcutaneously injected insu-
lin. Diabetes 37:694-701,1988
57. Sherwin RS, Shamoon H, Jacob R, Sacca L: Role of counter-
regulatory hormones in the metabolic response to stress in
normal and diabetic humans. In Melchionda N, Horwitz DL,
Schade DS (eds), Recent Advances in Obesity and Diabetes
Research. 1984, 327-344
58. Minuchin S, Rosman B, Baker L: Psychosomatic Families.
Cambridge, Harvard University Press, 1978
59. Baker L, Barcai A, Kaye R, Haque N: Beta-adrenergic block-
ade in juvenile diabetes: Acute studies and long-term ther-
apeutic trial. ] Pediatr 75:19-29,1969
60. Shapiro D, Surwit RS: Biofeedback. In Pomerleau OF, Brady
JP (eds), Behavioral Medicine: Theory and Practice. Balti-
more, Williams & Wilkins, 1970, 45-73
61. Jevning R, Wilson AF, Davidson JM: Adrenocortical activity
during meditation. Horm Behav 10:54-60,1978
62. DeGood DE, Redgate ES: Interrelationship of plasma cortisol
an other activation indexes during EMG biofeedback train-
ing. J Behav Med 5:213-224,1982
63. Mathew RJ, Ho BT, Kralik P, Taylor P, Claghorn JL: Cate-
cholamines and migraine: Evidence based on biofeedback-
induced changes. Headache 20:247-252,1980
64. Mathew RJ, Ho BT, Kralik P. Taylor D, Semchuk K, Weinman
M, Claghorn JL: Catechol-O-Methythransferase and cate-
cholamines in anxiety and relaxation. Psychiatr Res 3:85-
91, 1980
65. Fowler JE, Budzynski TH, Vandenbergh RL: Effects of an
EMG biofeedback program on the control of diabetes. Bio-
feedback and Self-Regulation 1:105-112,1976
66. Seeburg KN, DeBoer KF: Effects of EMG biofeedback on
diabetes. Biofeedback Self-Regulation 5:289-293,1980
67. Rosenbaum L: Biofeedback assisted stress management for
insulin treated diabetes mellitus. Biofeedback Self-Regula-
tion 8:519-532, 1983
68. Landis B, Javonic L, Landis E, Peterson CM, Groshen S,
Johnson K, Miller NE: Stress reduction program reduces
glycemia excursions 9% to 40% from baselines without
raising insulin needs. Diabetes 33:69A, 1984
69. Bradley C, Moses JL, Gamsu DS, Knight G, Ward JD: The
effects of relaxation on metabolic control in type I diabetes:
A matched control study. Diabetes 34:17A, 1985
70. Feinglos MN, Hastedt P, Surwit RS: The effects of relaxation
therapy on patients with type I diabetes mellitus. Diabetes
Care 10:72-75,1987
71. Rowland NE, Bellush LL: Diabetes mellitus: Stress, neuro-
chemistry and behavior. Neurosci Biobehav Rev 13:199-
206, 1989
72. Goetsch VL: Stress and blood glucose in diabetes mellitus:
A review an methodological commentary. Ann Behav Med
11:102-112,1989
73. Cryer PE: Decreased sympathochromaffin activity in IDDM.
Diabetes 38:405-409, 1989
74. Locatelli A, Franzetti I, Lepore G, Maglio ML, Gaudio E,
Caviezel F, Pozza G: Mental arithmatic stress as a test for
evaluation of diabetic sympathetic autonomic neuropathy.
Diabetic Med 6:490-495,1989
75. Dulin WE, Gerritsen GC, Chang A: Experimental and spon-
392
R. S. SURWIT AND M. S. SCHNEIDER
taneous diabetes in animals. In Ellenberg M, Rifkin H (eds),
Diabetes Mellitus: Theory and Practice, 3rd Edition. New
York, Medical Examination Publishing Company, 1983, 361-
408
76. Mikat EM, Hackel DB, Cruz PT, Lebovitz HE: Lowered
glucose tolerance in the sand rat (psammonys obesus) re-
sulting from esophageal intubation. Proc Soc Exp Biol Med
139:1390-1391, 1972
77. Beloff-Chain A, Freund N, Rookledge KA: Blood glucose and
serum insulin levels in lean and genetically obese mice.
Horm Metab Res 1:374-378, 1975
78. Grundleger ML, Godbole VY, Thenen SW: Age-dependent
development of insulin resistance of soleus muscle in genet-
ically obese (ob/ob) mice. Am J Physiol 239:363-371,1980
79. Dubuc PO, Cahn PJ, Ristimaki S, Willis PL: Starvation and
age effects on glycoregulation and hormone levels of C57BL/
6J ob/ob mice. Horm Metab Res 14:532-535, 1982
80. Surwit RS, Feinglos MN, Livingston EG, Kuhn CM, Mc-
Cubbin JA: Behavioral manipulation of the diabetic pheno-
type in ob/ob mice. Diabetes 33:616-618,1984
81. Surwit RS, McCubbin JA, Livingston EG. Feinglos MN: Clas-
sically conditioned hyperglycemia in the obese mouse. Psy-
chosom Med 47:565-568, 1985
82. Kuhn CM, Cochrane C, Feinglos MN, Surwit RS: Exagger-
ated peripheral responsivity to catecholamines contributes
to stress-induced hyperglycemia in the ob/ob mouse. Physiol
Biochem Behav 26:491-495, 1987
83. Fujimoto K, Sakaguchi T, Ui M: Adrenergic mechanisms in
the hyperglycaemia and hyperinsulinaemia of diabetic KK
mice. Diabetologia 20:568-572, 1981
84. Surwit RS, Kuhn CM, Cochrane C, McCubbin JA, Feinglos
MN: Diet-induced type II diabetes in C57BL/6J mice. Dia-
betes 37:1163-1167, 1988
85. Mills E, Kuhn CM, Feinglos MN, Surwit RS: Hypertension
in the C57BL/6J mouse model of non-insulin dependent
diabetes mellitus. Am J Physiol, in press
86. Surwit RS, McCubbin JA, Feinglos MN, Esposito-Del Peunte
A, Lillioja S: Glycemic reactivity to stress: A biologic marker
for development of type 2 diabetes. Diabetes 39 (Supp 1):8A,
1990
87. Spraul M, Anderson EA: Baseline and oral glucose stimu-
lated muscle sympathetic nerve activity in Pima Indians and
Caucasians. Diabetes 41(Supp. 1):189A, 1992
88. Grant I, Kyle GC, Teichman A, Mendels J: Recent life events
and diabetes in adults. Psychosom Med 36:121-128,1974
89. Naliboff BD, Cohen MJ, Sowers JD: Physiological and meta-
bolic responses to brief stress in noninsulin dependent dia-
betic and control subjects. J Psychosom Res 367-374, 1985
90. Goetsch VL, Wiebe DJ, Veltum LG, Van Dorsten B: Stress
and blood glucose in Type II diabetes mellitus. Behav Res
Therapy 28:531-537, 1990
91. Linde J, Deckert T: Increase of insulin concentration in
maturity-onset diabetics by phentolamine (Regitine) infu-
sion. Horm Metab Res 5:391-395, 1973
92. Robertson PR, Halter JB, Porte D: A role for alpha-adrenergic
receptors in abnormal insulin secretion in diabetes mellitus.
J Clin Invest 57:791-795, 1976
93. Kashiwagi A, Harano Y, Suzuki M, Kojima H, Harada M,
Nisho Y, Shigeta Y: New alpha-2 adrenergic blocker (DG-
5128) improves insulin secretion and in vivo glucose disposal
in NIDDM patients. Diabetes 35:1085-1089,1986
94. Cherksey B, Altszuler N: Tolbutamide and glyburide differ
in effectiveness to displace alpha- and beta-adrenergic ra-
Psychosomatic Medicine 55:380-393 (1993)
STRESS AND DIABETES

dioligands in pancreatic islet cells and membranes. Diabetes and pharmacologic manipulation of glucose tolerance. Pro-
33:499-503,1984 ceedings of the 14th International Diabetes Federation Con-
95. Lammers CA, Naliboff BD, Straatmeyer AJ: The effects of gress, Washington DC, 1991
progressive relaxation on stress and diabetic control. Behav gg S u ^ R S M c C u b b i n JA_ K u h n C M , M c G e e D i Gerstenfeld

Res Ther 22:641-650,1984 DA, Feinglos MN: Alprazolam reduces stress hyperglycemia
96. Surwit RS, Feinglos MN: The effects of relaxation on glucose • L, . . „ . „ .
. . . i. j j j. i x iw T->- in ob/ob mice. Psychosom Med 48:278-282,1986
tolerance in nomnsuhn dependent diabetes mellitus. Dia- 10
betes Care 5:337-342,1983 °- L e a h y J L : N a t u r a l h i s t o r y o f B " ce11 d V s f t l n c t i o n i n N I D D M -
97. Surwit RS, Feinglos MN: Relaxation-induced improvements Diabetes Care 13:992-1010,1990
101
in glucose tolerance is associated with decreased plasma - Surwit RS, Feinglos MN: Stress and autonomic nervous
cortisol. Diabetes Care 7:203-204,1984 system in type D diabetes: A hypothesis. Diabetes Care
98. Surwit RS, McCasKill CC, Ross SL, Feinglos MN: Behavioral 11:83-85,1988

Psychosomatic Medicine 55:380-393 (1993) 393