11

VetBooks.ir

CANINE DIABETES
MELLITUS

CLASSIFICATION AND ETIOLOGY
PATHOPHYSIOLOGY . . . . . . . . . .
SIGNALMENT . . . . . . . . . . . . . . . .
ANAMNESIS . . . . . . . . . . . . . . . . .
PHYSICAL EXAMINATION . . . . . . .
ESTABLISHING THE DIAGNOSIS
OF DIABETES MELLITUS . . . . . .
CLINICAL PATHOLOGIC
ABNORMALITIES . . . . . . . . . . . .
TREATMENT OF NONKETOTIC
DIABETES MELLITUS . . . . . . . . .
Insulin Therapy . . . . . . . . . . . . . .
Dietary Therapy . . . . . . . . . . . . .
Exercise . . . . . . . . . . . . . . . . . . .
Oral Hypoglycemic Drugs . . . . . .
Identification and Control of
Concurrent Problems . . . . . . .
Initial Adjustments in Insulin
Therapy . . . . . . . . . . . . . . . . . .
TECHNIQUES FOR MONITORING
DIABETIC CONTROL . . . . . . . . .
. 486 History and Physical Examination .
. 491 Serum Fructosamine
. 492 Concentration . . . . . . . . . . . . . .
. 492 Blood Glycated Hemoglobin
. 492 Concentration . . . . . . . . . . . . . .
Urine Glucose Monitoring . . . . . .
. 493 Serial Blood Glucose Curve . . . . . .
Role of Serum Fructosamine in
. 493 Aggressive, Excitable, or
Stressed Dogs . . . . . . . . . . . . . .
. 496 INSULIN THERAPY DURING
. 496 SURGERY . . . . . . . . . . . . . . . . . . .
. 499 COMPLICATIONS OF INSULIN
. 503 THERAPY . . . . . . . . . . . . . . . . . . .
. 504 Symptomatic and Asymptomatic
Hypoglycemia . . . . . . . . . . . . . .
. 506 Recurrence or Persistence of
Clinical Signs . . . . . . . . . . . . . .
. 507 Insulin Underdosage . . . . . . . . . . .
Insulin Overdosing and Glucose
. 508 Counterregulation (Somogyi
508 Phenomenon) . . . . . . . . . . . . . . 519
Short Duration of Insulin Effect . . . 520
509 Prolonged Duration of Insulin
Effect . . . . . . . . . . . . . . . . . . . . 521
510 Inadequate Insulin Absorption . . . . 522
511 Circulating Anti-Insulin
511 Antibodies . . . . . . . . . . . . . . . . 522
Allergic Reactions to Insulin . . . . . 524
Concurrent Disorders Causing
516 Insulin Resistance . . . . . . . . . . . 524
CHRONIC COMPLICATIONS OF
516 DIABETES MELLITUS . . . . . . . . . . 530
Cataracts . . . . . . . . . . . . . . . . . . . 531
517 Lens-Induced Uveitis . . . . . . . . . . 531
Diabetic Retinopathy . . . . . . . . . . 532
517 Diabetic Neuropathy . . . . . . . . . . 532
Diabetic Nephropathy . . . . . . . . . 532
518 Systemic Hypertension . . . . . . . . . 533
519 PANCREATIC ISLET
TRANSPLANTATION . . . . . . . . . . 533
PROGNOSIS . . . . . . . . . . . . . . . . . . 535

The endocrine pancreas is composed of the islets of
Langerhans, which are dispersed as “small islands” in
a “sea” of exocrine-secreting acinar cells. Four distinct
cell types have been identified within these islets on
the basis of staining properties and morphology—
alpha cells, which secrete glucagon; beta cells, which
secrete insulin; delta cells, which secrete somatostatin;
and F cells, which secrete pancreatic polypeptide. Dys-
function involving any of these cell lines ultimately
results in either an excess or a deficiency of the respec-
tive hormone in the circulation. In the dog and cat, the
most common disorder of the endocrine pancreas is
diabetes mellitus, which results from an absolute or
relative insulin deficiency due to deficient insulin
secretion by the beta cells. The incidence of diabetes
mellitus is similar for the dog and cat, with a reported
frequency varying from 1 in 100 to 1 in 500 (Panciera
et al, 1990).
CLASSIFICATION AND ETIOLOGY
Diabetes mellitus is classified according to the disease
in humans, that is, as type 1 and type 2 based on the
pathophysiologic mechanisms and pathogenic alter-
ations affecting the beta cells.
TYPE 1 DIABETES MELLITUS. Type 1 diabetes mellitus
is characterized by a combination of genetic suscepti-
bility and immunologic destruction of beta cells with
progressive and eventually complete insulin insuffi-
ciency (Eisenbarth, 1986; Palmer and McCulloch, 1991).
Autoantibodies are directed against several islet
components, including insulin, beta cell, and glutamic
acid decarboxylase (GAD) (Srikanta et al, 1985; Verge
et al, 1996; Gorus et al, 1997). Serum anti-insulin,
anti-beta cell, and/or anti-GAD autoantibodies are
commonly identified at the time diabetes is diagnosed
and are also detected early in the development of
type 1 diabetes, prior to the onset of hyperglycemia or
clinical signs (Verge et al, 1996; Gorus et al, 1997).
Screening individuals for serum anti-insulin, anti-beta
cell, and anti-GAD autoantibodies is used to identify
individuals at risk for development of type 1 diabetes.
Identification of anti-GAD autoantibodies or positive
results for all three autoantibodies has the highest
predictive value for development of type 1 diabetes
(Tuomilehto et al, 1994; Schatz et al, 1994; Hagopian
et al, 1995).
Immune-mediated destruction of the islets has been
conceptually divided into six stages in humans, begin-
ning with genetic susceptibility (Table 11-1) (Eisenbarth,
1986). Stage 2 involves a triggering event that leads

486
 CANINE DIABETES MELLITUS / 487

TABLE 11-1 STAGES IN THE DEVELOPMENT OF but normal insulin secretion is maintained. During stage
IMMUNE-MEDIATED INSULIN-DEPENDENT 4, immunologic abnormalities persist, but glucose-
VetBooks.ir

DIABETES MELLITUS IN HUMANS stimulated insulin secretion is progressively lost, despite

maintenance of euglycemia. Overt (clinical) diabetes
Stage 1 Genetic susceptibility
Stage 2 Triggering event develops in stage 5, although some residual insulin
Stage 3 Active autoimmunity and normal insulin secretion secretion remains. Stage 6 is characterized by complete
Stage 4 Active autoimmunity and impaired insulin secretion beta-cell destruction. As a result, individuals with type 1
Stage 5 Overt diabetes with residual insulin secretion diabetes are dependent on insulin to control glycemia,
Stage 6 Overt diabetes with complete beta-cell destruction
avoid ketoacidosis, and survive. Based on our current
understanding of the disorder, it seems likely that the
pathogenesis of diabetes mellitus in dogs progresses
to beta-cell autoimmunity. Environmental factors that through similar stages. Unfortunately, we do not
trigger the development of beta-cell immunity are poorly typically identify diabetes in dogs until stage 5 or 6, at
defined but probably include drugs and infectious which time insulin therapy is mandatory (Fig. 11-1).
agents. One interesting environmental factor that has TYPE 2 DIABETES MELLITUS. Type 2 diabetes mellitus
been proposed in humans involves consumption of is characterized by insulin resistance and “dysfunc-
cow’s milk during infancy. Some humans with newly tional” beta cells (Reaven, 1988; Leahy, 1990), defects
diagnosed insulin-dependent diabetes mellitus (IDDM) that are believed to be genetic in origin, evident for a
have antibodies to bovine serum albumin and a 17- decade or longer before hyperglycemia and clinical
amino acid bovine serum albumin peptide. It has been signs of diabetes develop, and with deleterious effects
hypothesized that the ingestion of cow’s milk that can be accentuated by environmental factors such
in early life can initiate beta-cell destruction through as obesity (Warram et al, 1990; Martin et al, 1992).
molecular mimicry between these proteins in cow’s milk Debate still continues among human diabetologists
and a beta-cell protein (Karjalainen et al, 1992; Savilahti et about which defect—insulin resistance or impaired
al, 1993; Gerstein, 1994), although there is skepticism insulin secretion—initiates the cascade of events
concerning an etiologic role of these antibodies (Atkinson leading to overt diabetes mellitus. Hepatic resistance
et al, 1993). Stage 3 is the period of active autoimmunity, to insulin, potentially induced by increased free fatty

C
2.4

S
Plasma C-peptide concentration (pM/ml)

S A 2.0
S S
S S
B

C Peptide
1.4
A Insulin B
S S A
S S 1.0
S S 0.8
B

0.6

0.4

0.2

Pre 5 10 20 30
Time (mins)

FIGURE 11-1. A, Schematic of the conversion of proinsulin (top) to insulin. Proteolytic cleavage of
proinsulin forms equimolar concentrations of C peptide and insulin, which are stored in secretory
granules of beta cells. B, Mean plasma C-peptide concentration before and after IV administration of
1 mg glucagon in 24 healthy dogs (broken line-open circles), 35 dogs with diabetes mellitus and low
baseline C-peptide concentration (broken line-triangles), 7 dogs with diabetes mellitus and increased
baseline C-peptide concentration (solid line-Xs), and 8 dogs with naturally acquired hyper-
adrenocorticism (solid line-solid circles). * = significantly (P<0.05) different from baseline value;
+ = significantly (P<0.05) different from corresponding time in healthy dogs. (B from Nelson RW:
Diabetes mellitus. In Ettinger SJ, Feldman EC (eds): Textbook of Veterinary Internal Medicine, 4th ed.
Philadelphia, WB Saunders Co, 1995, p 1511.)
 488 / CANINE DIABETES MELLITUS
acids in the portal circulation, results in excessive basal
hepatic glucose production and fasting hyperglycemia
VetBooks.ir
(DeFronzo et al, 1989; Groop et al, 1989; Bergman and
Mittleman, 1998; Massillon et al, 1997). Muscle insulin
resistance impairs the ability of endogenously secreted
insulin to augment muscle glucose uptake in response
to a meal, resulting in an excessive postprandial
increase in the blood glucose concentration (Mitrakou
et al, 1990). Defects in insulin receptor function, insulin
receptor–signal transduction pathway, glucose trans-
port and phosphorylation, glycogen synthesis, and
glucose oxidation contribute to muscle insulin resis-
tance (DeFronzo, 1997). Impaired insulin secretion also
plays a major role in the pathogenesis of glucose
intolerance in humans with type 2 diabetes (Polonsky,
1995). Total amounts of insulin secreted during glucose
tolerance testing may be increased, decreased, or
normal compared with the normal fasting animal. How-
ever, relative to the severity of insulin resist-
ance and prevailing hyperglycemia, even elevated
serum insulin concentrations are deficient (Saad et al,
1989). As the fasting blood glucose concentration
increases, insulin secretion decreases progressively, and
by the time the fasting blood glucose concentration is
greater than 180 to 200 mg/dl, the insulin response to
glucose tolerance testing is deficient in absolute terms.
Humans with type 2 diabetes are typically not
dependent on insulin to control the disease. Control
of the diabetic state is usually possible through diet,
exercise, and oral hypoglycemic drugs. However,
insulin treatment may be necessary in some type 2
diabetics if insulin resistance and beta cell dysfunction
are severe. As such, humans with type 2 diabetes can
have IDDM or non–insulin-dependent diabetes mellitus
(NIDDM). Insulin resistance and impaired insulin
secretion have been identified in dogs and cats with
obesity (Nelson et al, 1990; Appleton et al, 2001), and
obesity is accepted as a contributing factor for the
development of diabetes mellitus in dogs and cats
(Panciera et al, 1990; Scarlett and Donoghue, 1998).
Low insulin sensitivity has also been identified in a
group of lean cats subsequently identified to be at
greater risk for developing impaired glucose tolerance
with obesity than lean cats with higher insulin sensi-
tivity that subsequently developed obesity (Appleton
et al, 2001). The authors speculated that these cats
may be more at risk for progressing to overt diabetes
mellitus. The role of genetics, if any, for variations in
insulin sensitivity and risk for developing diabetes in
cats remains to be determined.
INSULIN-DEPENDENT AND NON–INSULIN-DEPENDENT
DIABETES MELLITUS. The classification of human dia-
betics as type 1 or type 2 is based on familial history,
clinical presentation, and results of immunologic (e.g.,
identification of serum anti–beta cell and anti-insulin
autoantibodies) and insulin secretagogue tests (Unger
and Foster, 1998). In contrast, classifying diabetes as
IDDM and NIDDM is based on the need for insulin
therapy to control glycemia, avoid ketoacidosis, and
survive. Individuals with IDDM must receive insulin
to avoid ketoacidosis, whereas control of glycemia and
avoidance of ketoacidosis can be accomplished through
diet, exercise, and oral hypoglycemic drugs in indi-
viduals with NIDDM. Approximately 10% of diabetic
humans in the United States have type 1 diabetes,
which is insulin-dependent. Approximately 90% of
diabetic humans have type 2 diabetes, which may be
insulin-dependent or non–insulin-dependent, depend-
ing on the severity of insulin resistance and beta cell
dysfunction.
It is more clinically relevant and perhaps more
accurate to classify diabetes in dogs and cats as IDDM
or NIDDM rather than as type 1 or type 2. Familial
history is rarely available in diabetic dogs and cats; the
clinical presentation is usually not helpful in differen-
tiating type 1 and type 2 diabetes, especially in cats
(Nelson et al, 1993); insulin secretagogue tests are not
routinely performed, and their results may be mis-
leading (Kirk et al, 1993); and autoantibody tests for
type 1 diabetes are not readily available. Therefore, as
clinicians, we usually classify diabetic dogs and cats
as either IDDM or NIDDM based on their need for
insulin treatment. This can be confusing because some
diabetics, especially cats, can initially appear to have
NIDDM progressing to IDDM, or flip back and forth
between IDDM and NIDDM as severity of insulin
resistance and impairment of beta-cell function waxes
and wanes. Apparent changes in the diabetic state (i.e.,
IDDM and NIDDM) are understandable when one
realizes that islet pathology may be mild to severe and
progressive or static; that the ability of the pancreas
to secrete insulin depends on the severity of islet
pathology and can decrease with time; that respon-
siveness of tissues to insulin varies, often in con-
junction with the presence or absence of concurrent
inflammatory, infectious, neoplastic, or hormonal
disorders; and that all these variables affect the
animal’s need for insulin, insulin dosage, and ease of
diabetic regulation. Use of serum insulin concen-
trations to differentiate between IDDM and NIDDM is
discussed on page 495.
INSULIN-DEPENDENT DIABETES IN DOGS. In our
experience, virtually all dogs have IDDM at the time
diabetes mellitus is diagnosed. IDDM is characterized
by hypoinsulinemia, essentially no increase in endo-
genous serum insulin concentration following adminis-
tration of an insulin secretagogue (e.g., glucose or
glucagon) at any time following diagnosis of the
disease, failure to establish glycemic control with diet
and/or oral hypoglycemic drugs, and an absolute
necessity for exogenous insulin to maintain glycemic
control. The cause of IDDM has been poorly charac-
terized in dogs but is undoubtedly multifactorial (Table
11-2). Genetic predispositions have been suggested by
familial associations in dogs and by pedigree analysis
of Keeshonds (see Table 11-3) (Guptill et al, 1999; Hess
et al, 2000a). The most common pathologic lesions
in dogs with diabetes mellitus are a reduction in the
number and size of pancreatic islets, a decrease in the
number of beta cells within islets, and hydropic
ballooning degeneration of beta cells (see Fig. 12-1,
page 541). In some dogs, an extreme form of the disease
 CANINE DIABETES MELLITUS / 489

may occur, represented by a congenital absolute TABLE 11-2 POTENTIAL FACTORS INVOLVED
deficiency of beta cells and pancreatic islet hypoplasia IN THE ETIOPATHOGENESIS OF DIABETES
VetBooks.ir

or aplasia (Alejandro et al, 1988; Atkins et al, 1988). MELLITUS IN DOGS AND CATS
Less severe changes of pancreatic islets and beta cells
Dog Cat
may predispose the adult dog to diabetes mellitus
after it has been exposed to environmental factors, Genetics Islet amyloidosis
Immune-mediated insulitis Obesity
such as infectious agents, insulin-antagonistic diseases Pancreatitis Pancreatitis
and drugs, obesity, and pancreatitis. Environmental Obesity Concurrent hormonal disease
factors may induce beta cell degeneration secondary to Concurrent hormonal disease Hyperadrenocorticism
chronic insulin resistance or may cause release of beta Hyperadrenocorticism Acromegaly
cell proteins that induce immune-mediated destruction Diestrus-induced excess of Hyperthyroidism
growth hormone Drugs
of the islets (Nerup, 1994). Studies evaluating for anti- Hypothyroidism Megestrol acetate
beta cell autoantibodies in diabetic dogs have been Drugs Glucocorticoids
conflicting, having been identified in newly diagnosed Glucocorticoids Infection
diabetic dogs with IDDM in one study (Hoenig and Infection Concurrent illness
Concurrent illness Renal insufficiency
Dawe, 1992) but not in another (Haines, 1986). Immune- Renal insufficiency Cardiac disease
mediated insulitis has also been described in diabetic Cardiac disease Hyperlipidemia (?)
dogs (Alejandro et al, 1988). Seemingly, autoimmune Hyperlipidemia Genetics (?)
mechanisms, in conjunction with environmental factors, Islet amyloidosis (?) Immune-mediated insulitis (?)
may play a role in the initiation and progression of
diabetes in dogs.
Clinically, pancreatitis is often seen in dogs with TABLE 11-3 BREED RISKS FOR DEVELOPING
diabetes mellitus and has been suggested as a cause of DIABETES MELLITUS DERIVED FROM ANALYSIS OF
diabetes after destruction of the islets. However, the THE VETERINARY MEDICAL DATABASE (VMDB)
incidence of histologically identifiable pancreatitis in FROM 1970 TO 1993*†
diabetic dogs is only 30% to 40%. Although destruc-
Odds
tion of beta cells secondary to pancreatitis is an obvious Breed Cases Control Ratio
explanation for the development of hypoinsulinemic
diabetes mellitus, other perhaps more complex factors Australian Terrier 33 3 9.39
Standard Schnauzer 96 14 5.85
are involved in the development of diabetes mellitus Miniature Schnauzer 526 88 5.10
in dogs without obvious exocrine pancreatic lesions. Bichon Frise 39 11 3.03
NON–INSULIN-DEPENDENT DIABETES IN DOGS. Spitz 34 10 2.90
Obesity-induced carbohydrate intolerance has been Fox Terrier 88 28 2.68
Miniature Poodle 712 244 2.49
documented in dogs (Mattheeuws et al, 1984), and Samoyed 159 56 2.42
minute amounts of amyloid have been identified in Cairn Terrier 61 23 2.26
the islets of some dogs with diabetes mellitus. Despite Keeshond 47 18 2.23
these findings, clinical recognition of NIDDM is very Maltese 42 20 1.79
uncommon in the dog. A juvenile form of canine dia- Toy Poodle 186 90 1.76
Lhasa Apso 85 47 1.54
betes mellitus that closely resembles human maturity- Yorkshire Terrier 96 57 1.44
onset diabetes of the young, a subclassification of Mixed Breed 1755 1498 1.00 (Reference group)
NIDDM, has been described. Measurement of plasma English Springer 62 77 0.69
C-peptide concentrations during insulin response Spaniel
Irish Setter 66 84 0.67
testing also suggests the presence of some continuing Beagle 70 94 0.64
beta-cell function in a small percentage of diabetic English Setter 29 41 0.60
dogs (see Fig. 11-1). C-peptide is the connecting peptide Basset Hound 28 43 0.56
found in the proinsulin molecule and is secreted into Rottweiler 37 62 0.51
the circulation in equimolar concentrations as insulin. Boston Terrier 27 45 0.56
Doberman Pinscher 103 180 0.49
Increased plasma C-peptide concentrations in these Labrador Retriever 199 375 0.45
latter dogs suggest either a severe form of type 2 Australian Shepherd 32 62 0.44
diabetes mellitus or residual beta-cell function in dogs Cocker Spaniel 77 186 0.35
with type 1 diabetes mellitus. Unfortunately, clinical Golden Retriever 91 274 0.28
Shetland Sheepdog 27 109 0.21
characteristics of juvenile canine NIDDM and the Collie 25 104 0.21
presence of some C-peptide secretory capabilities German Shepherd 68 317 0.18
resemble IDDM, in that dogs with these conditions are *From Guptill L, et al: Is canine diabetes on the increase? In Recent Advances
treated with insulin to manage hyperglycemia. in Clinical Management of Diabetes Mellitus, Iams Company, Dayton, Ohio,
TRANSIENT DIABETES IN DOGS. Transient or reversible 1999, p 24.
†
The VMDB comprises medical records of 24 veterinary schools in the
diabetes is extremely uncommon in dogs and usually United States and Canada. VMDB case records analyzed included those
occurs in dogs with subclinical diabetes treated with from first hospital visits of 6078 dogs with a diagnosis of diabetes mellitus
insulin antagonistic drugs (e.g., glucocorticoids) or in and 5922 randomly selected dogs with first hospital visits for any diagnosis
other than diabetes mellitus seen at the same veterinary schools in the same
the very early stages of an insulin-antagonistic disorder year. Only breeds with more than 25 cases of diabetes mellitus are included.
(e.g., diestrus in the bitch, hyperadrenocorticism). Such
 490 / CANINE DIABETES MELLITUS

dogs have a reduced but adequate mass of functional glycemic control using dosages of insulin considerably
beta cells to maintain carbohydrate tolerance when less than what would be expected (i.e., <0.2 U/kg per
VetBooks.ir

insulin resistance is not present, but they are unable
to secrete an adequate amount of insulin to maintain
euglycemia in the presence of insulin antagonism.
Early recognition and correction of the insulin antag-
onism may reestablish euglycemia without the long-
term need for insulin therapy. Failure to quickly correct
the insulin antagonism will result in progressive loss
of beta cells, eventual development of IDDM, and the
life-long requirement for insulin treatment to control
the hyperglycemia. Additional causes of transient or
reversible IDDM include NIDDM and the “honey-
moon period” of IDDM. A transient increase in insulin
secretion and reduced insulin dosage requirements
may occur during the initial weeks to months after the
diagnosis of IDDM in humans; this is called the honey-
moon period (Rossetti et al, 1990). A syndrome similar
to the honeymoon period occurs in some newly diag-
nosed diabetic dogs and is characterized by excellent
injection) (Fig. 11-2). Presumably, the existence of
residual beta-cell function when diabetes is diagnosed
(see Fig. 11-1) and correction of glucose toxicity (see
page 544) after initiation of insulin therapy account for
the initial ease of treating the diabetic state. Continuing
progressive destruction of residual functioning beta
cells results in worsening loss of endogenous insulin
secretory capacity and a greater need for exogenous
insulin to control the diabetes. As a result, glycemic
control becomes more difficult to maintain, and insulin
dosages increase to more commonly required amounts
(0.5 to 1.0 U/kg per injection). This increase in insulin
requirements usually occurs within the first 6 months
of treatment.
SECONDARY DIABETES IN DOGS. Secondary diabetes
mellitus is carbohydrate intolerance secondary to con-
current insulin antagonistic disease or medications.
Examples include the bitch in diestrus and the cat

500

400
Blood glucose concentration (mg/dl)

300

200

100

0
8 am Noon 4 pm 8 pm

FIGURE 11-2. Blood glucose curve in a 32-kg male dog receiving 0.3 U/kg SC beef/pork source NPH
insulin (solid line-solid circles). The blood glucose curve was obtained shortly after initiating insulin
therapy. Five months later, glycemic control deteriorated and clinical signs recurred despite increasing
the insulin dosage to 0.6 U/kg (broken line-solid circles). The dog was referred for possible insulin
resistance. Insulin underdosage was pursued initially and glycemic control was reestablished at an
insulin dosage of 1.0 U/kg (solid line-open circles). ↑ = insulin injection and food.

treated with megestrol acetate (progesterone). Hyper-
insulinemia may be initially present in these animals.
VetBooks.ir
However, with persistence of the insulin antagonistic
disorder, beta-cell function becomes impaired, and
permanent diabetes mellitus, typically IDDM, may
develop. If the insulin antagonistic disorder resolves
while some beta-cell function is still present, permanent
overt diabetes mellitus may not develop. However,
animals that become euglycemic after treatment of the
insulin antagonistic disorder or discontinuation of the
insulin antagonistic drug remain candidates to become
“subclinical” diabetics, and an attempt should be made
to avoid insulin antagonistic drugs and disorders to
prevent overt diabetes mellitus from developing.
Older bitches occasionally develop diabetes during
diestrus or pregnancy, presumably as a result of the
insulin antagonistic actions of progesterone (see
page 526). The diabetic state may resolve once the
progesterone concentration declines to anestrual levels
or may persist and require life-long insulin therapy.
Development of diabetes during diestrus or pregnancy
resembles gestational diabetes in humans. In humans,
gestational diabetes is restricted to pregnant women in
whom the onset or recognition of diabetes or impaired
glucose tolerance occurs during pregnancy (Unger and
Foster, 1998). After pregnancy termination, the woman
may remain clinically diabetic, may revert to a sub-
clinical diabetic state with impaired glucose tolerance,
or may revert to a normal glucose-tolerant state. For
the latter groups, the risk for developing overt diabetes
later in life ranges from 10% to 40%, depending on
severity of obesity. Bitches with transient diabetes
caused by diestrus have a high likelihood of develop-
ing permanent IDDM during the next estrus. For this
reason, all bitches with “gestational” diabetes should be
spayed as soon as possible after diabetes is diagnosed.
Once the diagnosis of diabetes is established, dogs
should be considered to have IDDM, and treatment
with insulin should be initiated, even if secondary
diabetes mellitus is suspected. Measurement of base-
line serum insulin concentration or evaluation of serum
insulin concentrations following administration of
an insulin secretagogue (e.g., glucose, glucagon; see
Table 12-1, page 547) are generally not recommended
in newly diagnosed diabetic dogs because of the high
prevalence of IDDM. The vast majority of newly diag-
nosed diabetic dogs will have either undetectable
serum insulin concentrations or serum insulin concen-
trations in the lower half of the normal reference range
(i.e., less than 12 μU/ml); findings consistent with
IDDM and hypoinsulinemia. Measurement of serum
insulin concentration may provide prognostic infor-
mation in newly diagnosed diabetic dogs in which there
is a strong suspicion of secondary diabetes mellitus
(e.g., intact bitch in diestrus; severe acute pancreatitis).
Serum insulin concentrations greater than one SD above
the reference mean (>18 μU/ml in our laboratory)
suggest the existence of functional beta cells and the
possibility for secondary diabetes mellitus and the
potential for reversion to a non-insulin-requiring state
if the underlying cause of the insulin antagonism can
CANINE DIABETES MELLITUS / 491
be identified and corrected. Insulin treatment is still
recommended in these situations to correct hyper-
glycemia and decrease the “stress” placed on the beta
cells while waiting for the insulin antagonism to
resolve. In theory, glucose toxicity may interfere with
accurate assessment of serum insulin concentrations in
diabetic dogs in a manner similar to that in diabetic
cats (see page 544) (Nelson et al, 1998a). However, the
syndrome of glucose toxicity has not yet been clearly
established in diabetic dogs, in part because most dogs
are truly hypoinsulinemic at the time diabetes mellitus
is diagnosed.
PATHOPHYSIOLOGY
Diabetes mellitus results from a relative or absolute
deficiency of insulin secretion by the beta cells. Insulin
deficiency, in turn, causes decreased tissue utilization
of glucose, amino acids, and fatty acids, accelerated
hepatic glycogenolysis and gluconeogenesis, and
accumulation of glucose in the circulation, causing
hyperglycemia. Glucose obtained from the diet also
accumulates in the circulation. As the blood glucose
concentration increases, the ability of the renal tubular
cells to resorb glucose from the glomerular ultrafiltrate
is exceeded, resulting in glycosuria. In dogs, this
typically occurs whenever the blood glucose concen-
tration exceeds 180 to 220 mg/dl. The threshold for
glucose resorption appears more variable in cats,
ranging from 200 to 280 mg/dl. Glycosuria creates an
osmotic diuresis, causing polyuria. Compensatory poly-
dipsia prevents dehydration. The diminished peripheral
tissue utilization of ingested glucose results in weight
loss as the body attempts to compensate for perceived
“starvation.”
The interaction of the “satiety center” in the ventro-
medial region of the hypothalamus with the “feeding
center” in the lateral region of the hypothalamus is
responsible for controlling the amount of food ingested
(Ganong, 1991). The feeding center, responsible for
evoking eating behavior, is chronically functioning but
can be transiently inhibited by the satiety center after
food ingestion. The amount of glucose entering the
cells in the satiety center directly affects the feeling of
hunger; the more glucose that enters these cells, the
less the feeling of hunger and vice versa (Ganong,
1991). The ability of glucose to enter the cells in the
satiety center is mediated by insulin. In diabetics with
a relative or absolute lack of insulin, glucose does not
enter satiety center cells, resulting in failure to inhibit
the feeding center. Thus these individuals become
polyphagic despite hyperglycemia.
The four classic signs of diabetes mellitus are
polyuria, polydipsia, polyphagia, and weight loss. The
severity of these signs is directly related to the severity
of hyperglycemia. As these signs become obvious to
the owner, the pet is brought to the veterinarian for
care. Unfortunately, some dogs and cats are not
identified by their owners as having signs of disease,
and these untreated animals may ultimately develop
 492 / CANINE DIABETES MELLITUS
diabetic ketoacidosis (DKA). See Chapter 13 for a
detailed discussion of the pathophysiology of DKA.
VetBooks.ir
SIGNALMENT
Most dogs are 4 to 14 years old at the time diabetes
mellitus is diagnosed, with a peak prevalence at 7 to
10 years of age. Juvenile-onset diabetes occurs in dogs
less than 1 year of age and is uncommon. Female dogs
are affected about twice as frequently as male dogs.
Genetic predispositions for or against the develop-
ment of diabetes have been suggested by familial asso-
ciations in dogs and by pedigree analysis of Keeshonds
(see Table 11-3; Guptill et al, 1999; Hess et al, 2000a).
Breed popularity can also affect perceptions of predis-
position. For example, epidemiologic studies suggest
that the Labrador retriever is at decreased risk for
development of diabetes mellitus; yet this is one of the
most frequently affected breeds in our practice, pre-
sumably because of their popularity in our region.
ANAMNESIS
The history in virtually all diabetic dogs includes the
classic signs of polydipsia, polyuria, polyphagia, and
weight loss. Polyuria and polydipsia do not develop
until hyperglycemia results in glycosuria. Occasionally
an owner brings in a dog because of sudden blindness
caused by cataract formation (Fig. 11-3). The classic
signs of diabetes mellitus may have gone unnoticed or
been considered irrelevant by the owner. If the clinical
signs associated with uncomplicated diabetes are not
observed by the owner and impaired vision caused
by cataracts does not develop, a diabetic dog is at risk
for the development of systemic signs of illness as
progressive ketonemia and metabolic acidosis develop
FIGURE 11-3. A, Bilateral cataracts causing blindness in a diabetic dog. B, Mature cataract with suture
lines in a diabetic Collie.
(see page 584). The time sequence from the onset of
initial clinical signs to the development of DKA is
unpredictable, ranging from days to weeks.
A complete history is extremely important even in
the “obvious” diabetic dog to explore for concurrent
disorders, which are almost always present at the time
diabetes mellitus is diagnosed. The clinician should
always ask, “Why has the dog developed clinical signs
of diabetes now?” In many dogs the insulin antagnism
caused by concurrent disorders such as pancreatitis,
bacterial infections, recent estrus, congestive heart
failure, or hyperadrenocorticism is the final insult
leading to overt diabetes. Identification and treatment
of concurrent disorders plays an integral role in the
successful management of the diabetic dog, and a
thorough history is the first step toward identification
of these disorders.
PHYSICAL EXAMINATION
Performance of a thorough physical examination is
imperative in any dog suspected to have diabetes
mellitus, in part because of the high prevalence of con-
current disorders that can affect response to treatment.
The physical examination findings in a dog with newly
diagnosed diabetes depend on whether DKA is present
and its severity, on the duration of diabetes prior to its
diagnosis, and on the nature of any other concurrent
disorder. The nonketotic diabetic dog has no classic
physical examination findings. Many diabetic dogs are
obese but are otherwise in good physical condition.
Dogs with prolonged untreated diabetes may have
lost weight but are rarely emaciated unless concurrent
disease (e.g., pancreatic exocrine insufficiency) is
present. Lethargy may be evident; the haircoat may be
sparse, with dry, brittle, and lusterless hair; and scales
from hyperkeratosis may be present. Diabetes-induced

hepatic lipidosis may cause hepatomegaly. Lenticular
changes consistent with cataract formation are another
VetBooks.ir
common clinical finding in diabetic dogs. Additional
abnormalities may be identified in the dog with
diabetic ketoacidosis (see page 584).
ESTABLISHING THE DIAGNOSIS OF
DIABETES MELLITUS
A diagnosis of diabetes mellitus requires the presence
of appropriate clinical signs (i.e., polyuria, polydipsia,
polyphagia, weight loss) and documentation of per-
sistent fasting hyperglycemia and glycosuria. Measure-
ment of the blood glucose concentration using a
portable blood glucose monitoring device (see page 512)
and testing for the presence of glycosuria using urine
reagent test strips (e.g., KetoDiastix; Ames Division,
Miles Laboratories Inc., Elkhart, Ind.) allows the rapid
confirmation of diabetes mellitus. The concurrent
documentation of ketonuria establishes a diagnosis of
diabetic ketosis or ketoacidosis.
It is important to document both persistent hyper-
glycemia and glycosuria to establish a diagnosis of
diabetes mellitus, because hyperglycemia differentiates
diabetes mellitus from primary renal glycosuria,
whereas glycosuria differentiates diabetes mellitus
from other causes of hyperglycemia (Table 11-4).
Stress-induced hyperglycemia is a common problem
in cats and occasionally occurs in dogs, especially
those that are very excited, hyperactive, or aggressive.
Refer to page 567 for more information on stress-
induced hyperglycemia.
Mild hyperglycemia (i.e., 130 to 180 mg/dl) is
clinically silent and is usually an unexpected and
unsuspected finding. If the dog with mild hyper-
glycemia is examined for polyuria and polydipsia, a
disorder other than clinical diabetes mellitus should
be sought, most notably hyperadrenocorticism. Mild
TABLE 11-4 CAUSES OF HYPERGLYCEMIA IN
DOGS AND CATS
Diabetes mellitus*
“Stress” (cat)*
Postprandial (diets containing monosaccharides, disaccharides,
and propylene glycol)
Hyperadrenocorticism*
Acromegaly (cat)
Diestrus (bitch)
Pheochromocytoma (dog)
Pancreatitis
Exocrine pancreatic neoplasia
Renal insufficiency
Drug therapy*
Glucocorticoids
Progestagens
Megestrol acetate
Thiazide diuretics
Dextrose-containing fluids*
Parenteral nutrition*
Head trauma
*Common cause
CANINE DIABETES MELLITUS / 493
hyperglycemia can occur up to 2 hours postprandially
in some dogs following consumption of soft moist
foods (Holste et al, 1989), in “stressed” dogs, in early
diabetes mellitus, and with disorders causing insulin
resistance (see page 524). A diagnostic evaluation for
disorders causing insulin resistance is indicated if mild
hyperglycemia persists in the fasted, unstressed dog.
Insulin therapy is not indicated in these animals,
because clinical diabetes mellitus is not present.
CLINICAL PATHOLOGIC ABNORMALITIES
OVERVIEW OF PATIENT EVALUATION. A thorough
evaluation of the dog’s overall health is recommended
once the diagnosis of diabetes mellitus has been
established to identify any disease that may be causing
or contributing to the carbohydrate intolerance (e.g.,
hyperadrenocorticism), that may result from the carbo-
hydrate intolerance (e.g., bacterial cystitis), or that may
mandate a modification of therapy (e.g., pancreatitis).
The minimum laboratory evaluation in any “healthy”
nonketotic diabetic dog should include a CBC, serum
biochemical panel, and urinalysis with bacterial
culture. Serum progesterone concentration should be
determined if diabetes mellitus is diagnosed in an
intact bitch, regardless of her cycling history. If avail-
able, abdominal ultrasound is indicated to assess for
pancreatitis, adrenomegaly, pyometritis in an intact
bitch, and abnormalities affecting the liver and urinary
tract (e.g., changes consistent with pyelonephritis or
cystitis). Because of the relatively high prevalence of
pancreatitis in diabetic dogs, measurement of serum
lipase or serum trypsin–like immunoreactivity (TLI)
should be considered if abdominal ultrasound is not
available. Measurement of the baseline serum insulin
concentration or an insulin response test is not routinely
done. Additional tests may be warranted after obtain-
ing the history, performing the physical examination,
or identifying ketoacidosis. The laboratory evaluation
of dogs with glycosuria and ketonuria is discussed in
detail in Chapter 13, page 586. Potential clinical
pathologic abnormalities are listed in Table 11-5.
COMPLETE BLOOD COUNT. Results of a CBC are
usually normal in the uncomplicated diabetic pet. A
mild polycythemia may be present if the dog is dehy-
drated. An elevation of the white blood cell count may
be caused by either an infectious process or severe
inflammation, especially if an underlying pancreatitis
is present. The presence of toxic or degenerative
neutrophils or a significant shift toward immaturity of
the cells supports the presence of an infectious process
as the cause of the leukocytosis.
SERUM BIOCHEMICAL PANEL. The prevalence and
severity of abnormalities identified in the serum
biochemistry panel are dependent on the duration of
untreated diabetes and the presence of concurrent
disease, most notably pancreatitis. The serum bio-
chemical panel is often unremarkable in “healthy”
diabetic dogs without significant concurrent disease,
aside from hyperglycemia and hypercholesterolemia.
 494 / CANINE DIABETES MELLITUS
TABLE 11-5 CLINICOPATHOLOGIC
ABNORMALITIES COMMONLY FOUND IN
VetBooks.ir
DOGS AND CATS WITH UNCOMPLICATED
DIABETES MELLITUS
Complete Blood Count
Typically normal
Neutrophilic leukocytosis, toxic neutrophils if pancreatitis or
infection present
Biochemistry Panel
Hyperglycemia
Hypercholesterolemia
Hypertriglyceridemia (lipemia)
Increased alanine aminotransferase activity (typically <500 IU/L)
Increased alkaline phosphatase activity (typically <500 IU/L)
Urinalysis
Urine specific gravity typically >1.025
Glycosuria
Variable ketonuria
Proteinuria
Bacteriuria
Ancillary Tests
Hyperlipasemia if pancreatitis present
Hyperamylasemia if pancreatitis present
Serum trypsinlike immunoreactivity usually normal
Low with pancreatic exocrine insufficiency
High with acute pancreatitis
Normal to high with chronic pancreatitis
Variable serum baseline insulin concentration
IDDM: low, normal
NIDDM: low, normal, increased
Insulin resistance induced: low, normal, increased
IDDM, Insulin-dependent diabetes mellitus; NIDDM, non-insulin-dependent
diabetes mellitus.
The most common abnormalities are an increase in
serum alanine transaminase and alkaline phosphatase
activities and hypercholesterolemia. The increase in
liver enzyme activities is usually mild (less than 500
IU/L) and a result of hepatic lipidosis. Serum alkaline
phosphatase activities in excess of 500 IU/L should
raise suspicion for concurrent hyperadrenocorticism,
especially if other abnormalities consistent with
hyperadrenocorticism are identified in the laboratory
data (see Chapter 6). Serum alanine transaminase
activities in excess of 500 IU/L should raise suspicion
for hepatopathy other than hepatic lipidosis, especially
if additional abnormalities in endogenous liver func-
tion tests (e.g., low urea nitrogen, hypoalbuminemia,
increased serum bile acids) are identified. An increase
in the serum total bilirubin concentration should raise
suspicion for extrahepatic biliary obstruction caused by
concurrent pancreatitis. When appropriate, abdominal
ultrasound and histologic evaluation of a liver biopsy
specimen may be indicated to establish concurrent
liver disease.
The BUN and serum creatinine concentrations are
usually normal in the uncomplicated diabetic. An
elevation in these parameters may be due to either
primary renal failure or prerenal uremia secondary
to dehydration. Primary renal failure as a result of
glomerulosclerosis, damage specifically related to
hyperglycemia, is a well-recognized complication in
humans but is uncommon in dogs (see Diabetic
Nephropathy, page 532). Evaluation of urine specific
gravity should help differentiate primary renal failure
from prerenal uremia.
Alterations in serum electrolytes and acid-base
parameters are common in pets with DKA and are
discussed in the section dealing with therapy for the
severe ketoacidotic diabetic (see page 596).
URINALYSIS. Abnormalities identified in the uri-
nalysis that are consistent with diabetes mellitus include
glycosuria, ketonuria, proteinuria, and bacteriuria with
or without associated pyuria and hematuria. The dog
with uncomplicated diabetes usually has glycosuria
without ketonuria. However, a relatively healthy dia-
betic may also have trace to small amounts of ketones
in the urine. If large amounts of ketones are present in
the urine, especially in an animal with systemic signs
of illness (e.g., lethargy, vomiting, diarrhea, or dehy-
dration), a diagnosis of DKA should be made and the
animal treated appropriately.
The presence and severity of glycosuria should be
considered when interpreting the urine specific gravity.
Despite polyuria and polydipsia, urine specific gravities
typically range from 1.025 to 1.035 in untreated diabetic
dogs, in part because of the large amount of glucose
in the urine. As a general rule of thumb, 2% or 4+
glycosuria as measured on urine reagent test strips will
increase the urine specific gravity 0.008 to 0.010 when
urine specific gravity is measured by refractometry. As
such, identification of a urine specific gravity less than
1.020 in combination with 2% glycosuria suggests a
concurrent polyuric/polydipsic disorder, most notably
hyperadrenocorticism or renal insufficiency.
Proteinuria may be the result of urinary tract infec-
tion or glomerular damage secondary to disruption of
the basement membrane (Struble et al, 1998). Because
of the high incidence of infection, the urine sediment
should be carefully inspected for changes consistent
with infection, including white blood cells, red blood
cells, and bacteria. Failure to identify pyuria and
hematuria does not rule out urinary tract infection
(McGuire et al, 2002). Because of the relatively high
prevalence of concurrent urinary tract infections in dia-
betic dogs, urine obtained by antepubic cystocentesis
using aseptic technique should be submitted for
bacterial culture and sensitivity testing in all dogs with
newly diagnosed diabetes mellitus, regardless of the
findings on urinalysis (Hess et al, 2000b).
SERUM CHOLESTEROL AND TRIGLYCERIDE CONCEN-
TRATIONS. Hyperlipidemia and obvious lipemia are
common in the untreated diabetic. Uncontrolled
diabetes is accompanied by an increase in the blood
concentration of triglycerides, cholesterol, lipoproteins,
chylomicrons, and free fatty acids (DeBowes, 1987).
Hypertriglyceridemia is responsible for the lipemia,
which can be seen in a peripheral blood sample.
Hypertriglyceridemia with increases in chylomicrons
and very low density lipoprotein (VLDL) triglycerides
results from insulin deficiency and the associated

curtailment in lipoprotein lipase activity (Eckel, 1989).
The enzyme lipoprotein lipase aids in the metabolism
VetBooks.ir
of the triglyceride-rich VLDLs and chylomicrons.
Increased concentrations of VLDL triglyceride also
result from excess hepatic production (induced by
increased circulating free fatty acids), obesity, and a
high caloric intake (Massillon et al, 1997; Unger and
Foster, 1998).
Overall, blood cholesterol concentrations are
increased in diabetics, but to a much lesser degree than
is hypertriglyceridemia. In humans, several factors
contribute to an increase in LDL cholesterol concen-
trations, including increased LDL synthesis from
VLDLs; reduced activity of LDL receptors, impairing
LDL clearance; and excess consumption of saturated
fatty acids (Howard et al, 1987; Unger and Foster, 1998).
In humans, high density lipoprotein (HDL) cholesterol
levels are often low, presumably as a result of accel-
erated catabolism (Witzum et al, 1982). The combination
of high LDL and low HDL cholesterol concentrations
may play a role in the accelerated development of
atherosclerotic vascular disease and coronary heart
disease, which is the major long-term complication of
diabetes in humans (Garg and Grundy, 1990). Similar
vascular complications have been infrequently docu-
mented in diabetic dogs and cats (Hess et al, 2002).
Fortunately, most lipid derangements can be improved
with insulin and dietary therapy.
PANCREATIC ENZYMES. Blood tests to assess for the
presence of pancreatitis should always be considered
in the newly diagnosed diabetic dog, especially if
abdominal ultrasound is not available. Measurement
of serum lipase concentration and serum TLI are most
commonly recommended. In theory, dogs with con-
comitant active pancreatitis should have an increase
in serum lipase concentration and serum TLI. Unfor-
tunately, serum lipase concentrations and serum TLI
do not always correlate accurately with the presence
or absence of pancreatitis (Hess et al, 1998). Pancreatic
enzyme concentrations can be increased in dogs with
a histologically confirmed normal pancreas and normal
in dogs with histologically confirmed inflammation of
the pancreas, especially when the inflammatory process
is chronic and mild. For these reasons, interpretation
of serum lipase and TLI results should always be done
in context with the history, physical examination
findings, and additional findings on the laboratory
tests. In our experience, abdominal ultrasound is the
single best diagnostic test for identifying pancreatitis
in the dog and should be considered if pancreatitis
is suspected after evaluation of the history, physical
examination, and laboratory test results. The concomi-
tant presence of pancreatitis may necessitate the insti-
gation of intensive fluid therapy and a highly digestible,
low-fat diet, which may otherwise not have been done.
Identification of chronic pancreatitis also has important
prognostic implications regarding success of establish-
ing and maintaining control of glycemia and long-
term survival (see page 528).
Measurement of serum TLI is also used to diagnose
exocrine pancreatic insufficiency; an uncommon com-
CANINE DIABETES MELLITUS / 495
plication of diabetes mellitus that presumably develops
as a sequela of chronic pancreatitis (Wiberg et al, 1999;
Wiberg and Westermarck, 2002). Exocrine pancreatic
insufficiency should be suspected in diabetic dogs that
are difficult to regulate with insulin and are thin or
emaciated despite polyphagia (see page 528).
SERUM THYROXINE CONCENTRATION. The veterinarian
may periodically have to interpret a serum thyroxine
(T4) concentration in a diabetic dog, either because
serum T4 is a routine part of the serum biochemistry
panel or because hypothyroidism is suspected after
a review of the history, clinical signs, and physical
examination findings. Interpretation of serum T4 results
must be done cautiously, especially in a dog with newly
diagnosed diabetes mellitus and concurrent illness
such as pancreatitis or infection. “Healthy” diabetic
dogs without concurrent illness usually have normal
serum T4 concentrations. However, the more severe
the diabetic state and the more severe the concurrent
illness, the more likely serum T4 concentrations will be
decreased because of the euthyroid sick syndrome
rather than because of hypothyroidism (see Fig. 3-25,
page 121). As a general rule, in a newly diagnosed
diabetic dog with a concurrent low serum T4 concen-
tration, we treat the diabetes and reevaluate the serum
T4 concentration once control of glycemia has been
established. If hypothyroidism is strongly suspected at
the time diabetes is diagnosed, we evaluate serum
free T4 and thyrotropin (TSH) concentrations before
initiating sodium levothyroxine treatment. See
Chapter 3 for a more detailed discussion of the effects
of concurrent illness and drug therapy on serum
thyroid hormone concentrations and the tests used to
diagnose hypothyroidism in dogs.
SERUM INSULIN CONCENTRATION. Measurement of
serum insulin concentration, either baseline or after
the administration of an insulin secretagogue (see
Table 12-1, page 547), is not a routine part of our diag-
nostic evaluation of the newly diagnosed diabetic dog.
In theory, identifying increased endogenous serum
insulin concentrations (i.e., >18 μU/ml) in a newly
diagnosed diabetic dog would suggest the early stages
of secondary diabetes, especially if an underlying
insulin antagonistic disorder can be identified and
treated (see page 490). Unfortunately, the suppressive
effects of hyperglycemia on beta-cell function (i.e., glu-
cose toxicity) may interfere with accurate interpretation
of serum insulin results (Nelson et al, 1998a). Although
increased serum insulin concentrations suggest the
existence of functional beta cells, finding low serum
insulin concentrations (i.e., <12 μU/ml) may not rule
out the existence of functional beta cells, a problem
commonly encountered in diabetic cats (see page 544).
However, because the vast majority of dogs with newly
diagnosed diabetes have IDDM and serum insulin
concentration is typically in the lower half of normal
or undetectable, routine measurement of serum
insulin concentration is not a cost-effective diagnostic
procedure. The exception are dogs suspected to be in
the early stages of secondary diabetes, most notably
intact bitches in diestrus.
 496 / CANINE DIABETES MELLITUS

TABLE 11-6 DIFFERENCES IN THE AMINO ACID SEQUENCE OF THE INSULIN MOLECULE IN THE DOG AND
CAT VERSUS SOURCES OF COMMERCIAL INSULIN
VetBooks.ir

COMPARISON WITH DOG COMPARISON WITH CAT

AMINO ACID POSITION AMINO ACID POSITION

A8 A10 A18 B30 A8 A10 A18 B30
Dog Thr Ile Asn Ala Cat Ala Val His Ala
Pig Thr Ile Asn Ala Cow Ala Val Asn Ala
Human Thr Ile Asn Thr Pig Thr Ile Asn Ala
Cow Ala Val Asn Ala Human Thr Ile Asn Thr
Differences between the dog or cat and commercial insulins are italicized.

It is imperative that the radioimmunoassay (RIA) a serious and potentially fatal complication of therapy.
for insulin be validated for the species in which it is Hypoglycemia is most apt to occur as the result of
being used and that normal reference ranges for the overzealous insulin therapy. The veterinarian must
species in question have been established. Commercial balance the benefits of tight glucose control obtainable
endocrine laboratories use RIAs designed for use in with aggressive insulin therapy against the risk for
humans. Fortunately, the amino acid sequence of human hypoglycemia.
and dog insulin are almost identical (Table 11-6), so
RIAs designed for measurement of serum insulin con-
centrations in humans work well in dogs. The same is Insulin Therapy
not true for cats.
OVERVIEW OF INSULIN TYPES. Commercial insulin is
categorized by promptness, duration, and intensity of
TREATMENT OF NONKETOTIC DIABETES action after subcutaneous administration (Table 11-8).
MELLITUS Commonly used insulins for the long-term manage-
ment of diabetics include isophane (NPH), lente, ultra-
GOALS OF THERAPY. The primary goal of therapy is lente, and protamine zinc (PZI) insulins. NPH and PZI
elimination of the owner-observed signs occurring insulin preparations contain the fish protein protamine
secondary to hyperglycemia and glycosuria. A persis- and zinc to delay insulin absorption and prolong the
tence of clinical signs and the development of chronic duration of insulin effect (Davidson et al, 1991). The
complications (Table 11-7) are directly correlated with lente family of insulins rely on alterations in zinc con-
the severity and duration of hyperglycemia. Limiting tent and the size of zinc-insulin crystals to alter the rate
blood glucose concentration fluctuations and main- of absorption from the subcutaneous site of deposition.
taining near-normal glycemia will help minimize the The larger the crystals, the slower the rate of absorption
severity of clinical signs and prevent the complications and the longer the duration of effect. The lente insulins
of poorly controlled diabetes. In the diabetic dog, this contain no foreign protein (i.e., protamine). Ultralente
can be accomplished through proper insulin therapy, is a microcrystalline, long-acting insulin preparation.
diet, exercise, and the prevention or control of con- Lente insulin is a mixture of three parts of short-acting,
current inflammatory, infectious, neoplastic, and amorphous insulin and seven parts of long-acting,
hormonal disorders. microcrystalline insulin. Lente insulin is considered an
Although it is worthwhile attempting to normalize intermediate-acting insulin, although plasma insulin
the blood glucose concentration, the veterinarian must concentrations may remain increased for longer than
also guard against the development of hypoglycemia, 14 hours following subcutaneous administration in
some dogs (Graham et al, 1997).
Insulin Mixtures. Mixtures of short- and long-acting
insulin have been developed in an attempt to mimic
TABLE 11-7 COMPLICATIONS OF DIABETES the increase in portal insulin concentrations during
MELLITUS IN DOGS AND CATS and immediately following consumption of a meal,
Common Uncommon thereby minimizing postprandial hyperglycemia.
Iatrogenic hypoglycemia Peripheral neuropathy (dog)
NPH insulin can be mixed with regular crystalline
Persistent polyuria, polydipsia, Glomerulonephropathy, insulin, and if injected immediately, the regular insulin
weight loss glomerulosclerosis remains rapid-acting. Stable premixed 70% NPH/30%
Cataracts (dog) Retinopathy regular and 50% NPH/50% regular preparations are
Bacterial infections, especially Exocrine pancreatic available (e.g., Humulin 70/30, Eli Lilly and Co.,
in the urinary tract insufficiency
Pancreatitis Gastric paresis Indianapolis, IN; Mixtard HM 70/30, Nordisk-USA,
Ketoacidosis Diabetic diarrhea Princeton, NJ). In our experience, these premixed
Hepatic lipidosis Diabetic dermatopathy (dog) preparations are quite potent, causing a rapid decrease
Peripheral neuropathy (cat) (i.e., superficial necrolytic in blood glucose concentration within 60 to 90 minutes
dermatitis)
of subcutaneous administration (Fig. 11-4). In addition,
 CANINE DIABETES MELLITUS / 497

TABLE 11-8 PROPERTIES OF RECOMBINANT HUMAN INSULIN PREPARATIONS USED IN DIABETIC DOGS
AND CATS* AND BEEF/PORK PZI INSULIN USED IN DIABETIC CATS
VetBooks.ir

TIME OF MAXIMUM EFFECT (HR) DURATION OF EFFECT (HR)

Type of Insulin Route of Administration Onset of Effect Dog Cat Dog Cat
1 1
Regular crystalline IV Immediate ⁄2 –2 ⁄2 –2 1 –40 1 –40
IM 10–30 min 1 –4 1 –4 3 –80 3 –80
SC 10–30 min 1 –5 1 –5 4 –10 4 –10
NPH (isophane)† SC 1
⁄2 –2 hr 2 –10 2 –8 6 –18 4 –12
Lente†‡ SC 1
⁄2 –2 hr 2 –10 2 –10 8 –20 6 –18
1
Ultralente SC ⁄2 –8 hr 4 –16 4 –16 8 –24 6 –24
PZI‡ SC 1
⁄2 –4 hr – 4 –14 – 6 –20
*Purified pork insulin has similar properties; beef/pork insulin mixtures are less potent and may have a longer duration of action than recombinant human
insulins.
†
Initial insulins of choice for the diabetic dog.
‡
Initial insulins of choice for the diabetic cat.

the duration of effect has usually been short (<8 hours). 500
We generally use these insulin mixtures only as a last

Blood glucose concentration (mg/dl)

resort when more conventional insulin preparations
have been ineffective in establishing control of glycemia. 400
Although regular insulin remains fast acting when
added to NPH, when added to lente insulin, regular 300
insulin binds to excess zinc in the lente, blunting
regular insulin’s quick effect (Galloway, 1988). A
Insulin Analogs. Recently, recombinant DNA 200
technology has been applied for the production of
insulin analogs with faster and slower absorption 100
characteristics than human insulin. Rapid-acting
insulin analogs include insulin lispro (Humalog, Eli
Lilly, Indianapolis, IN) and insulin aspart (Novolog, 0
8 AM NOON 4 PM 8 PM
NovoNordisk, Princeton, NJ). The rate-limiting step in
the absorption of human insulin is a hexamer formation 500
of insulin molecules that occurs at high concentrations
Blood glucose concentration (mg/dl)

of insulin such as those obtained in the injectable fluid

(Brange et al, 1988). The hexamers of insulin molecules 400
slowly dissociate before absorption into the circulation
occurs. By replacing certain amino acids in the insulin 300
molecule, the tendency to self-associate can be
reduced without affecting the insulin-receptor B
kinetics. Insulin lispro is produced by inverting the 200
natural amino acid sequence of the B-chain at B28
(proline) and B29 (lysine), and insulin aspart is 100
produced by substituting aspartic acid for proline in
position B28 (Maarten et al, 1997; Lindholm et al,
2002). As a consequence of these alterations, insulin 0
lispro and insulin aspart exhibit monomeric behavior 8 AM NOON 4 PM 8 PM
in solution and display a rapid absorption, faster
pharmacodynamic action, and shorter duration of FIGURE 11-4. A, Blood glucose curve in a miniature poodle receiving
effect than short-acting regular crystalline insulin recombinant human lente insulin, 6 U/kg body weight (▲) and
(Howey et al, 1994; Home et al, 1997; Lindholm et al, recombinant human 70/30 NPH/regular insulin, 3 U/kg body
1999). Insulin lispro and insulin aspart are the current weight (●) SC. B, Blood glucose curve in an 8-kg cat receiving 4 U
prandial insulins (i.e., insulin administered before recombinant human ultralente insulin (●) and 4 U of recombinant
human 70/30 NPH/regular insulin (▲) SC. ↑ = insulin injection and
each meal) of choice for control of postprandial blood food. (From Nelson RW: Diabetes mellitus. In Ettinger SJ, Feldman
glucose concentrations in human diabetics and are EC (eds): Textbook of Veterinary Internal Medicine, 4th ed.
typically administered three times a day before each of Philadelphia, WB Saunders Co, 1995, p 1528.)
the three main meals (breakfast, lunch, dinner). The
role, if any, of these insulins for the treatment of
diabetes in dogs or cats remains to be determined.
Because of their extremely short duration of effect,
insulin lispro and insulin aspart would have to be
 498 / CANINE DIABETES MELLITUS
used in conjunction with a longer-acting insulin prep-
aration to maintain control of glycemia.
VetBooks.ir
Insulin glargine (Lantus, Aventis Pharmaceuticals,
Bridgewater, NJ) is a long-acting insulin analog that
differs from human insulin by the replacement of
asparagine with glycine at position A21 on the A-chain
and the addition of two arginines to the C-terminus of
the B-chain of the insulin molecule (Pieber et al, 2000).
These modifications result in a shift of the isoelectric
point from a pH of 5.4 toward a neutral pH, which
makes insulin glargine more soluble at a slightly acidic
pH and less soluble at a physiological pH than native
human insulin. As a consequence, insulin glargine
forms microprecipitates in the subcutaneous tissue
at the site of injection from which small amounts of
insulin glargine are slowly released. In humans, the
slow sustained release of insulin glargine from these
microprecipitates results in a relatively constant con-
centration/time profile over a 24-hour period with
no pronounced peak in serum insulin. The glucose-
lowering effect of insulin glargine is similar to that
of human insulin, the onset of action following sub-
cutaneous administration is slower than NPH insulin,
and the duration of effect is prolonged, compared with
NPH insulin (Owens et al, 2000). Insulin glargine is
currently recommended as a basal insulin (i.e.,
sustained long-acting insulin used to inhibit hepatic
glucose production) administered once a day at
bedtime and used in conjunction with either rapid-
acting insulin analogs or oral hypoglycemic drugs in
human diabetics (Rosenstock et al, 2000; Rosenstock
et al, 2001).
In a preliminary study evaluating the pharmaco-
kinetics and pharmacodynamics of insulin glargine,
PZI, and porcine lente insulin in nine healthy cats,
most of the parameters evaluated (i.e., onset of action,
glucose nadir, time for blood glucose concentration to
return to baseline, mean daily blood glucose concen-
tration, and area under the 24-hour blood glucose
curve) were similar for insulin glargine and PZI
(Marshall and Rand, 2002). Similar studies performed
in diabetic cats have yet to be reported. In our expe-
rience, insulin glargine has a duration of effect ranging
from 10 to 16 hours in most diabetic dogs and cats
in which it was used. We have not yet encountered
problems with inadequate absorption of insulin
glargine, as described with ultralente insulin (see
page 571), although it seems likely that this problem
will be encountered as we gain more experience with
insulin glargine. Currently, we consider using insulin
glargine in diabetic dogs and cats with problems of
short duration of effect of NPH, lente, and in cats, PZI
insulin (see pages 520 and 570). Based on our current
experiences, we do not consider insulin glargine a first
choice insulin for the treatment of diabetes in dogs
or cats.
SPECIES OF INSULIN. For each type of insulin (i.e.,
NPH, lente, ultralente), the clinician must also choose
the species of insulin to be administered. Historically,
insulin preparations were available as beef/pork
combinations (e.g., Iletin I, Eli Lilly Co, Indianapolis,
IN), purified beef or pork (e.g., Iletin II, Eli Lilly and
Co), and recombinant human insulin (e.g., Humulin,
Eli Lilly and Co). Approximately 90% of beef/pork
insulin preparation is beef insulin. Currently, more
than 95% of human diabetics requiring insulin injec-
tions are treated with recombinant human insulin
preparations and the majority of the rest are treated
with purified pork insulin preparations. Commercial
production of beef/pork and purified beef insulin for
use by diabetic humans has been severely curtailed in
the United States. Because the human market essen-
tially dictates the insulin preparations available for use
in diabetic dogs and cats, almost all diabetic dogs and
most diabetic cats are treated with recombinant human
insulin preparations. Fortunately, the development of
anti-insulin antibodies following chronic administration
of recombinant human or porcine insulin to diabetic
dogs appears uncommon (see page 522) (Feldman et
al, 1983; Harb-Hauser et al, 1998). In contrast, bovine
insulin is antigenic in diabetic dogs and stimulates
formation of anti-insulin antibodies in 40% to 65% of
diabetic dogs in which it is used (Feldman et al, 1983;
Haines, 1986; Harb-Hauser et al, 1998; Davison et
al, 2003).
Presumably, the structure and amino acid sequence
of the injected insulin relative to the native endogenous
insulin influences the development of anti-insulin anti-
bodies. Although differences exist in the amino acid
sequence of human, bovine, and canine insulin (see
Table 11-6) (Hallden et al, 1986; Ganong, 1991), studies
suggest that conformational insulin epitopes are more
important than linear subunits of the insulin molecule.
Anti-insulin antibodies induced by exogenous insulin
in humans cross-react with homologous insulin and
require intact conformation of the insulin molecule for
binding (Thomas et al, 1985; Thomas et al, 1988; Nell
and Thomas, 1989). In a recent study evaluating anti-
insulin antibody formation in diabetic dogs treated
with bovine insulin, the greatest anti-insulin antibody
reactivity was directed against the whole insulin
protein rather than the A- or B-chain (Davison et al,
2003). Cross-reactivity with homologous insulin was
also identified. Surprisingly, the insulin B-chain rather
than the A-chain was the more reactive component
of the insulin molecule. The reason for differences in
A-chain versus B-chain antigenicity is not clear,
especially considering that the amino acid sequence of
the A-chain, not the B-chain, differs between the cow
and dog.
Anti-insulin antibodies may affect the pharmaco-
kinetics of the exogenously administered insulin by
several mechanisms, which may either enhance or
reduce the pharmacodynamic response. Antibodies
may enhance and prolong the pharmacodynamic
action by serving as a carrier, or they may reduce
insulin action by neutralization (Bolli et al, 1984;
Marshall et al, 1988; Lahtela et al, 1997). Antibodies
may also have no apparent clinical effect on insulin
dosage or status of glycemic control (Lindholm et al,
2002). In our experience, anti-insulin antibody produc-
tion in diabetic dogs can alter the duration of insulin

activity and prolong its duration of action or have a
deleterious impact on insulin effectiveness, reduce the
VetBooks.ir
ability to maintain control of glycemia, and in extreme
cases, cause severe insulin resistance (see page 522).
Short duration of insulin effect is a more prevalent
problem with recombinant human insulin than with
beef-containing insulin preparations, presumably
because of lack of anti-insulin antibody production.
However, the deleterious impact of anti-insulin anti-
bodies on control of glycemia associated with beef-
containing insulin far outweighs any potential benefits
related to duration of insulin effect.
Recently, beef/pork PZI insulin became commer-
cially available for use in diabetic cats (see page 550).
This insulin preparation is not routinely recommended
for use in diabetic dogs because of concerns for anti-
insulin antibody production directed at the beef insulin
component of the insulin preparation, as discussed
above. Purified porcine lente insulin (Caninsulin,
Intervet, Ontario, Canada) is marketed for use in
diabetic dogs in Canada and Europe but is not yet
approved by the FDA for use in the United States. The
amino acid sequences of canine and porcine insulin
are identical (see Table 11-6) (Hallden et al, 1986;
Ganong, 1991). The immunogenicity and duration of
effect of porcine insulin preparations is similar to
recombinant human insulin preparations (Feldman
et al, 1983; Harb-Hauser et al, 1998). In theory, the
effectiveness of purified porcine lente insulin and
recombinant human lente insulin should be similar.
INITIAL INSULIN TREATMENT RECOMMENDATIONS.
Intermediate-acting insulin (i.e., lente, NPH) is the
initial insulin of choice for establishing control of
glycemia in diabetic dogs (see Table 11-8). Recombinant
human-source insulin should be used to avoid
problems with insulin effectiveness presumably
caused by circulating insulin antibodies (see page 522)
Insulin therapy is begun with lente or NPH insulin of
recombinant human origin at an approximate dosage
of 0.25 U/kg twice a day. Dietary therapy is initiated
concurrently (see below). Because greater than 90% of
diabetic dogs require recombinant human lente or NPH
insulin twice daily, we prefer to start with twice-a-day
insulin therapy (Hess and Ward, 2000). Establishing
control of glycemia is easier and problems with hypo-
glycemia and the Somogyi effect (see page 519) are less
likely when twice daily insulin therapy is initiated
while the insulin dose is low, that is, at the time insulin
treatment is initiated. Glycemic regulation is more
problematic and development of hypoglycemia and
the Somogyi effect more likely when poorly controlled
diabetic dogs receiving a high dose of insulin once
daily are switched to insulin twice a day but the dose
per injection is arbitrarily chosen.
Dietary Therapy
Dietary therapy plays an important role in the successful
management of the diabetic dog. Adjustments in diet
and feeding practices should be directed at correcting
CANINE DIABETES MELLITUS / 499
TABLE 11-9 RECOMMENDATIONS FOR DIETARY
TREATMENT OF DIABETES MELLITUS IN DOGS
I. Dietary composition
Increased fiber content (see Table 11-10)
Digestible carbohydrate content >45% of metabolizable energy
Fat content <30% of metabolizable energy
Protein content <30% of metabolizable energy
II. Feed canned and/or dry kibble foods; avoid diets containing
monosaccharides, disaccharides, and propylene glycol
III. Caloric intake and obesity
Average daily caloric intake in geriatric pet: 40-60 kcal/kg
Adjust daily caloric intake on individual basis
Eliminate obesity, if present, by:
Increasing daily exercise
Decreasing daily caloric intake
Feeding low-calorie-dense, low-fat, high-fiber (preferred in
diabetics) or low-calorie-dense, low-fat, low-fiber diet
designed for weight loss
IV. Feeding schedule
Maintain consistent caloric content of the meals
Maintain consistent timing of feeding
Feed within time frame of insulin action
Feed one-half the total daily caloric intake at time of each
insulin injection
Let “nibbler” dogs continue to nibble throughout day and
night
or preventing obesity, maintaining consistency in the
timing and caloric content of the meals, and furnishing
a diet that helps minimize the postprandial increase in
blood glucose concentration (Table 11-9; Nelson and
Lewis, 1990). Correction of obesity and increasing the
fiber content of the diet are perhaps the two most
beneficial steps that can be taken to improve control of
glycemia.
DIETARY FIBER. Diets containing increased fiber con-
tent are beneficial for treating obesity and improving
control of glycemia in diabetics (Nelson et al, 1991;
Nelson et al, 1998b). Several mechanisms have been
proposed to explain the fiber-induced slowing of
intestinal glucose absorption and the corresponding
improvement in glycemic control of the diabetic dog.
These include a delay in the gastric emptying of
nutrients; a delay in the intestinal absorption of
nutrients, mostly likely resulting from an effect on the
diffusion of glucose toward the brush border of the
intestine; and a fiber-induced effect on the release of
regulatory gastrointestinal tract hormones into the
circulation (Meyer et al, 1988; Anderson et al, 1991;
Nuttall, 1993). The ability of the food fiber to form a
viscous gel and thus impair convective transfer of
glucose and water to the absorptive surface of the
intestine appears to be of greatest importance (Nuttall,
1993). The more viscous soluble fibers (e.g., gums,
pectins) slow glucose diffusion to a greater degree
than do the less viscous insoluble fibers (e.g., lignin,
cellulose) and, as such, are believed to be of greater
benefit in improving control of glycemia in diabetic
human beings (Anderson and Akanji, 1991; Nuttall,
1993). Others have found both fiber types to be
beneficial in diabetic human beings (Villaume et al,
1984; Vaaler, 1986). Studies in diabetic dogs have
documented glycemic improvement in response to the
 500 / CANINE DIABETES MELLITUS

consumption of diets containing increased amounts of diet can be added and the quantity of the soluble fiber
soluble and insoluble fiber (Fig. 11-5) (Nelson et al, diet decreased.
VetBooks.ir

1991; Graham et al, 1994; Nelson et al, 1998b). Most
commercial high-fiber diets predominantly contain
insoluble fiber, although diets containing mixtures of
soluble and insoluble fiber are becoming available
(Table 11-10). The amount of fiber varies considerably
among products, ranging from 3% to 25% of dry matter
(normal diets contain less than 2% fiber on a dry
matter basis). In general, diets containing 12% or more
insoluble fiber or 8% or more of a mixture of soluble
and insoluble fiber are most likely to be effective in
improving glycemic control in diabetic dogs.
The dog’s susceptibility to the complications of
high-fiber diets, its body weight and condition, and
the presence of a concurrent disease (e.g., pancreatitis,
renal failure) in which diet is an important aspect of
therapy ultimately dictate which, if any, fiber diet is
fed. Common clinical complications of high insoluble
fiber diets include excessive frequency of defecation;
constipation and obstipation; hypoglycemia 1 to 2 weeks
after the increase in fiber content of the diet; and
refusal to eat the diet (Table 11-11). Complications of
soluble fiber-containing diets include soft to watery
stools; excessive flatulence; hypoglycemia 1 to 2 weeks
after the increase in fiber content of the diet; and refusal
to eat the diet. If firm stools or constipation become a
problem with high insoluble fiber diets, a mixture of
insoluble and soluble fiber diets can be fed or soluble
fiber (e.g., sugar-free Metamucil, canned pumpkin) can
be added to the diet to soften the stool. Alternatively, if
soft or watery diarrhea or flatulence become a problem
with soluble fiber-containing diets, an insoluble fiber
Refusal to consume diets containing increased
amounts of fiber may occur initially or develop after
several months of eating the diet. If palatability is a
problem initially, the dog can be gradually switched
from its regular diet to a diet containing small
amounts of fiber, followed by a gradual switch to diets
containing more fiber. Refusal to consume high-fiber
diets months after their initiation is usually a result of
boredom with the food. Periodic changes in the types
of high-fiber diets and mixtures of diets have been
helpful in alleviating this problem. Diets containing an
increased amount of fiber should always be considered
in the list of differential diagnoses for inappetence in a
diabetic dog.
Diets containing an increased amount of fiber
should not be fed to thin or emaciated diabetic dogs,
because high-fiber diets have a low caloric density,
which can interfere with weight gain and may result in
further weight loss. For thin diabetic dogs to gain
weight, usually glycemic control must be reestablished
through insulin therapy and the feeding of a higher-
calorie–dense, lower-fiber diet designed for mainte-
nance. Once a normal body weight has been attained,
a diet containing more fiber can be gradually sub-
stituted for the previous diet.
DIETARY PROTEIN. The protein content of the diet
remains controversial in humans with diabetes because
of the potential for both restricted and liberal protein
intakes to exert beneficial and adverse effects (Wylie-
Rosett, 1988; Malik and Jaspan, 1989). Although protein
is a much less potent insulin secretagogue than glucose,

FIGURE 11-5. Mean (± standard error of the mean) serum concentrations of glucose in 11 dogs with
naturally occurring diabetes mellitus fed high-insoluble fiber (i.e., cellulose; solid line) and low-fiber
(broken line) diet. ↑ = Insulin administration and consumption of half of daily caloric intake; * = p<0.05,
compared with low fiber diet. (From Nelson RW et al: Effect of dietary insoluble fiber on glycemic
control in dogs with naturally occurring diabetes mellitus, JAVMA 212:280, 1998.)
 CANINE DIABETES MELLITUS / 501

TABLE 11-10 APPROXIMATE NUTRIENT CONTENT OF SOME COMMERCIALLY AVAILABLE HIGH-FIBER

DOG FOODS
VetBooks.ir

Crude Fiber* Carbohydrates† Fat† Protein† Calories (Can/Cup)‡

Prescription Diet r/d

Canned 26 45 22 33 249
Dry 24 49 21 30 205

Prescription Diet w/d

Canned 14 54 29 16 390
Dry 17 62 19 19 226

Purina OM
Canned 19 17 28 55 189
Dry 11 47 16 37 276

Science Diet Light

Canned 10 58 24 18 364
Dry 14 62 18 20 295

Purina DCO
Dry 8 45 32 24 320

Waltham Glucomodulation Control Diet

Canned 10 55 11 34 339
Dry 5 62 10 28 316

Iams Eukanuba Optimum Weight Control

Dry 3 53 21 26 253
*Expressed as a percent of the diet dry matter.
†
Expressed as percent metabolizable energy derived from carbohydrate, fat, or protein.
‡
Expressed as kilocalories of metabolizable energy per can (15 oz.) or for the dry diets per standard (8 oz. volume) measuring cup full.

TABLE 11-11 COMMON COMPLICATIONS
ASSOCIATED WITH FEEDING DIETS CONTAINING
INCREASED QUANTITIES OF FIBER
Inappetence caused by poor palatability or boredom with food
Increased frequency of defecation
Constipation and obstipation (insoluble fiber)
Soft stools and diarrhea (soluble fiber)
Increased flatulence (soluble fiber)
Weight loss
Hypoglycemia
variation in dietary protein may influence metabolic
control of diabetes by altering gluconeogenic substrate
availability and counterregulatory hormone secretion
(Spillar et al, 1987; Krezowski et al, 1986; Henry, 1994).
Prolonged consumption of excessive dietary protein,
especially in conjunction with excessive phosphorus
and sodium, may contribute to the progression of
diabetic nephropathy in humans (Hostetter et al, 1982;
Brenner et al, 1982), and consumption of low-protein
diets may impede or delay the rate of development of
diabetic renal disease (Friedman, 1982). The impact, if
any, of liberal dietary protein consumption on renal
function is controversial in dogs and cats, although
dietary protein restriction is recommended once renal
insufficiency is identified (Harte et al, 1994; Devaux
et al, 1996). Because diabetes mellitus and renal insuffi-
ciency commonly occur together, it seems prudent to
recommend a dietary protein intake that meets daily
requirements but is not excessive (i.e., less than 30%
protein on a metabolizable energy basis in dogs).
Lower protein intake is indicated when evidence of
renal insufficiency exists.
DIETARY FAT. Derangements in fat metabolism are
common in diabetic dogs and include increased serum
concentrations of cholesterol, triglycerides, lipoproteins,
chylomicrons, and free fatty acids, hepatic lipidosis,
atherosclerosis, and a predisposition for development
of pancreatitis (DeBowes et al, 1987; Hess et al, 2002).
Feeding high-fat diets may also cause insulin resistance,
promote hepatic glucose production, and in healthy
dogs, suppress β-cell function (Massillon et al, 1997;
Kaiyala et al, 1999). These findings strongly support
feeding diets that are relatively low in fat content, that
is, less than 30% fat on a metabolizable energy basis.
Feeding lower-fat diets will help minimize the risk for
pancreatitis, control some aspects of hyperlipidemia,
and reduce overall caloric intake to favor weight loss
or maintenance (Remillard, 1999). A higher fat content
may be needed for weight gain in thin or emaciated
diabetic dogs.
CALORIC INTAKE AND OBESITY. Obesity can cause
impaired glucose tolerance in dogs (Mattheeuws et al,
1984) and may be an important factor accounting for
variations in response to insulin therapy in diabetic
dogs. Weight reduction improves glucose tolerance in
obese dogs, presumably via improvement in obesity-
 502 / CANINE DIABETES MELLITUS
induced insulin resistance (Wolfsheimer et al, 1993).
Successful weight reduction usually requires a com-
VetBooks.ir
bination of restriction of caloric intake, feeding low-
calorie dense diets (i.e., high fiber content), and
increasing caloric expenditure through exercise.
The dog’s current body weight should be recorded,
and the final ideal body weight of the dog should be
calculated. The ideal body weight can be estimated
either by reviewing the medical record for the body
weight when the dog was in an ideal body condition,
or by using breed-specific body weight charts. It is
very important to set realistic and obtainable goals for
weight loss in order to maintain client compliance.
If the ideal body weight is below 15% of the current
body weight, it is crucial to use a stepwise process to
gradually achieve ideal body weight. The pet’s initial
goal should be set at 15% body weight loss. Once this
goal has been achieved, a new target body weight can
be selected until the dog has reached an ideal body
weight. To achieve 15% body weight loss, dogs can
be fed 55 × [initial body weight (kg) 0.75] kcal per day
(Elliott, 2003). When fed at this level, dogs will achieve
15% body weight loss in approximately 12 weeks. The
calculated amount of calories to achieve 15% body
weight loss should be compared with the current daily
caloric intake obtained from the dietary history. Most
pets will be consuming more calories than required for
15% body weight loss. However, if the number of
calories to achieve weight loss is actually less than the
current daily caloric intake, then the dietary history
should be reevaluated to search for additional calories.
If no additional daily calories are identified, then the
daily caloric intake of the pet should be reduced by
15% to 20%. Although there are several diets specifically
formulated for weight reduction in dogs, diets that
utilize fiber are recommended in obese diabetic dogs
for reasons previously discussed. The quantity of food
to be fed is determined by dividing the daily caloric
requirement by the number of calories per can or
cupful of the diet. The amount fed per meal and the
timing of the meals are dictated, in part by the pet’s
insulin treatment regimen and by the owner’s schedule
(see Feeding Schedule, right column). In addition to
reducing the daily caloric intake, every effort should
be made to increase the daily energy expenditure by
encouraging exercise.
Dogs on weight reduction programs should be
reevaluated every 2 weeks. Body weight should be
recorded and the dietary history reviewed. Ideally,
dogs should achieve 1% to 2% body weight loss per
week. If the rate of weight loss exceeds 2% body
weight loss per week, the number of calories fed to the
pet should be increased by 10% to 15%. If there has not
been any weight loss, the dietary history should be
reevaluated for additional calories. If none are found,
the daily caloric intake should be further reduced by
10% to 15%. Once the ideal body weight of the dog has
been achieved, the daily caloric intake can be adjusted
to maintain optimal body weight and a diet intended
for maintenance of the diabetic dog initiated (e.g.,
Prescription Diet W/D, Hill’s Pet Products, Topeka, KS;
Iams Eukanuba Optimum Weight Control, Iams Co.,
Dayton, Ohio).
FEEDING SCHEDULE. The feeding schedule should
be designed to enhance the actions of insulin and
minimize postprandial hyperglycemia. The develop-
ment of postprandial hyperglycemia depends, in part,
on the amount of food consumed per meal, the rate at
which glucose and other nutrients are absorbed from
the intestine, and the effectiveness of exogenous and
endogenous insulin during this time. The daily caloric
intake should be ingested when insulin is still present
in the circulation and is capable of disposing of
glucose absorbed from the meal. If the meals are
consumed while exogenous insulin is still meta-
bolically active, the postprandial increase in blood
glucose concentration is minimal or absent. In
contrast, feeding the diabetic dog after insulin action
has waned results in increasing blood glucose con-
centration beginning 1 to 2 hours postprandially
(Fig. 11-6). If this occurs, either the type of insulin,
frequency of insulin administration, or timing of the
meals in relationship to the insulin injection should be
adjusted.
Typically, dogs receiving exogenous insulin twice a
day are fed equal-sized meals at the time of each insulin
injection. If the dog is receiving exogenous insulin
once a day, one half of the daily caloric intake is fed at
the time of the insulin injection and the remaining half
approximately 8 to 10 hours later. Unfortunately, the
eating behaviors of dogs vary considerably, from finicky
eaters that nibble on food periodically throughout the
day to gluttonous dogs that quickly consume every-
thing placed in their food dish. Gluttonous dogs are
fed as discussed above. Finicky dogs generally resist
attempts by owners to convert them to a “gluttonous”
type of eating behavior, which can be frustrating to the
owner instructed to have their pet eat all of its food
at the time of the insulin injection. However, if one
adheres to the principle that feeding multiple small
meals rather than one large meal within the time frame
of insulin action helps minimize the hyperglycemic
effect of each meal, then allowing a finicky eater to eat
whenever it wants should help control fluctuations in
blood glucose. For this reason, dogs that are finicky
and nibble throughout the day should be allowed to
continue their pattern of eating. For these dogs, the
food should remain available beginning at the time of
each insulin injection and the dog allowed to choose
when and how much to eat. The process is repeated at
the time of the next insulin injection.
MODIFICATIONS IN DIETARY THERAPY. Diabetes mellitus
often occurs in conjunction with other diseases (e.g.,
renal failure, heart failure). Dietary therapy is used
in the management of various diseases. Whenever
possible, dietary therapy for all disorders should be
“blended”; however, if this is not possible, dietary
therapy for the most serious disorder should take
priority. For example, dietary therapy for chronic renal
failure, heart failure, or recurring pancreatitis is a
higher priority than dietary therapy for diabetes
mellitus. Dietary therapy for diabetes mellitus should
 CANINE DIABETES MELLITUS / 503

600 60
VetBooks.ir

500 50
Blood glucose concentration (mg/dl)

Serum insulin concentration (␮U/ml)

400 40

300 30

200 20

100 10

8 AM Noon 8 PM Mid 8 AM

FIGURE 11-6. Mean blood glucose (solid line) and serum insulin (broken line) concentrations in eight
dogs with diabetes mellitus treated with beef/pork source NPH insulin SC once daily. The duration of
NPH effect is too short, resulting in prolonged periods of hyperglycemia beginning shortly after
feeding the evening meal. ↑ = insulin injection; * = equal-sized meals consumed. (From Nelson RW:
Diabetes mellitus. In Ettinger SJ, Feldman EC (eds): Textbook of Veterinary Internal Medicine, 4th ed.
Philadelphia, WB Saunders Co, 1995, p 1525.)

be considered adjunctive; glycemic control can be effect is unknown. Although we have successfully fed
maintained with insulin, regardless of the diet fed. diets containing increased amounts of fiber to several
Acute and chronic pancreatitis and exocrine pan- diabetic dogs and cats with EPI, studies are needed to
creatic insufficiency (EPI) are closely associated with critically assess the effects of dietary fiber on this com-
diabetes mellitus in the dog. Fortunately, many of the bination of disorders.
dietary principles for diabetes, pancreatitis, and EPI
are similar (Table 11-12). Many diabetic dogs with
concurrent pancreatitis tolerate high-fiber diets once Exercise
pancreatitis has subsided with the feeding of low-fat,
highly digestible diets. Such diets are also recommended Exercise plays an important role in maintaining
for dogs with EPI (Lewis et al, 1987). In vitro studies glycemic control in the diabetic dog by helping
indicate that some fiber sources impair pancreatic promote weight loss and by eliminating the insulin
enzyme activity, although cellulose, the fiber most resistance induced by obesity. Exercise also has a
commonly used in commercial high-fiber dog foods, glucose-lowering effect by increasing the mobilization
did not affect pancreatic enzyme activity (Dutta and of insulin from its injection site, presumably resulting
Hlasko, 1985). Clinical trials documenting deleterious from increased blood and lymph flow, by increasing
effects of dietary fiber in vivo have not been reported blood flow (and therefore insulin delivery) to exercising
in diabetic dogs or cats with EPI. The ability of oral muscles, by stimulating translocation (i.e., upregulation)
pancreatic enzyme supplements to counter any such of glucose transporters (primarily GLUT-4) in muscle

TABLE 11-12 COMPARISON OF GENERAL GUIDELINES FOR DIETARY TREATMENT OF DOGS AND CATS
WITH DIABETES MELLITUS, PANCREATITIS, AND EXOCRINE PANCREATIC INSUFFICIENCY
Dietary Factor Diabetes Mellitus Pancreatitis Exocrine Pancreatic Insufficiency
Digestibility Normal High High
Fat content Low Low Low
Carbohydrate content High Normal to low Normal to low
Protein content Normal to moderate restriction Moderate restriction Normal
Feeding schedule Twice a day with insulin injection Small meals often Small meals often
Caloric intake Correct and avoid obesity Correct and avoid obesity Correct and avoid weight loss
 504 / CANINE DIABETES MELLITUS
cells, and by increasing glucose effectiveness (i.e.,
ability of hyperglycemia to promote glucose disposal
VetBooks.ir
at basal insulin concentrations) (Fernqvist et al, 1986;
Galante et al, 1995; Phillips et al, 1996; Nishida et al,
2001). The daily routine for diabetic dogs should
include exercise, preferably at the same time each day.
Strenuous and sporadic exercise can cause severe
hypoglycemia and should be avoided. The insulin
dose should be decreased in dogs subjected to sporadic
strenuous exercise (e.g., hunting dogs during hunting
season) on those days of anticipated increased exer-
cise. The reduction in insulin dose required to prevent
hypoglycemia is variable and determined by trial and
error. We recommend reducing the insulin dose by
50% initially and making further adjustments based
on the occurrence of symptomatic hypoglycemia and
the severity of polyuria and polydipsia that develops
during the ensuing 24 to 48 hours. In addition, owners
must be aware of the signs of hypoglycemia and have
a source of glucose (e.g., Karo syrup, candy, food)
readily available to give their dog should any of these
signs develop.
Oral Hypoglycemic Drugs
OVERVIEW. In the United States, five classes of oral
hypoglycemic drugs are approved for the treatment of
NIDDM in human beings: sulfonylureas, meglitinides,
biguanides, thiazolidinediones, and α-glucosidase
inhibitors. These drugs work by stimulating pancreatic
insulin secretion, enhancing tissue sensitivity to insulin,
or slowing postprandial intestinal glucose absorption
(DeFronzo, 1999). In addition, chromium and vanadium
are trace minerals that may also enhance tissue sensi-
tivity to insulin. Oral hypoglycemic drugs are not
routinely used to treat diabetes in dogs, in part because
dogs develop IDDM and most dogs are hypo-
insulinemic at the time diabetes is diagnosed. Drugs
that stimulate pancreatic insulin secretion are ineffec-
tive because the beta cells have been destroyed. Insulin-
sensitizing drugs require the presence of circulating
insulin for efficacy and, as such, are ineffective as the
primary mode of treatment in hypoinsulinemic diabetic
dogs. Drugs that slow postprandial intestinal glucose
absorption improve control of glycemia in diabetic
dogs but are ineffective as the primary mode of treat-
ment, are expensive, and adverse reactions are common
(see Acarbose below). To date, only the α-glucosidase
inhibitors and chromium have been critically evaluated
in diabetic dogs. Refer to Chapter 12, page 555 for more
information on oral hypoglycemic drugs.
ALPHA-GLUCOSIDASE INHIBITORS. Acarbose (Precose,
Bayer, West Haven, CT) and miglitol (Glyset, Pharmacia
& Upjohn Inc, Peapack, NJ) are complex oligo-
saccharides of microbial origin that competitively
inhibit α-glucosidases (glucoamylase, sucrase, maltase,
and isomaltase) in the brush border of the small
intestinal mucosa (Lembcke et al, 1985; Balfour and
McTavish, 1993). Inhibition of these enzymes delays
digestion of complex carbohydrates and disaccharides
to monosaccharides. This inhibition causes increased
carbohydrate digestion within the ileum and, to a lesser
extent, colon. Most important, it delays absorption of
glucose from the intestinal tract and decreases post-
prandial blood glucose and insulin concentrations.
Controlled clinical studies involving acarbose-treated
human beings with NIDDM and IDDM have docu-
mented significant improvement in glycemic control
(Rios, 1994; Coniff et al, 1995; Kelley et al, 1998).
Improvements were characterized by significant
reductions in postprandial blood glucose concentra-
tions, blood glycated protein concentrations, and daily
insulin requirements. Placebo-controlled clinical studies
recently completed in healthy and diabetic dogs
documented a decrease in postprandial total glucose
absorption and total insulin secretion when healthy
dogs were treated with acarbose, compared with
placebo, and a decrease in daily insulin dose, mean
blood glucose concentration during an 8-hour blood
sampling period, and blood glycated protein concen-
trations in diabetic dogs treated with acarbose, com-
pared with placebo (Fig. 11-7; Robertson et al, 1999;
Nelson et al, 2000).
Results of these studies suggest that acarbose may
be beneficial in improving control of glycemia in some
dogs with IDDM. However, diarrhea and weight loss
as a result of carbohydrate malassimilation were
common adverse effects, occurring in approximately
35% of dogs (Robertson et al, 1999; Nelson et al, 2000).
Diarrhea was more prevalent at higher doses of acarbose
(i.e., 100 and 200 mg/dog) and typically resolved within
2 to 3 days of discontinuing the medication. Because of
cost and prevalence of adverse effects, acarbose should
probably be reserved for treating poorly controlled
diabetic dogs in which the cause for poor control of
glycemia cannot be identified and insulin treatment,
by itself, is ineffective in preventing clinical signs of
diabetes. The initial acarbose dosage is low (i.e., 12.5 to
25 mg at each meal) regardless of the body weight of
the dog. The benefit of this drug is dependent on its
interaction with the meal; it should only be given at
the time of feeding. A stepwise increase to 50 mg/dog
and, in large dogs (>25 kg) a further increase to
100 mg/dog, can be considered in dogs that fail to show
improvement in control of glycemia after 2 weeks
using the 12.5 to 25 mg/meal dosage regimen. How-
ever, adverse reactions (especially diarrhea) are more
likely to occur at these higher doses.
CHROMIUM. Chromium is a ubiquitous trace ele-
ment that exerts insulin-like effects in vitro. The exact
mechanism of action is not known but the overall
effect of chromium is to increase insulin sensitivity,
presumably through a post-receptor mechanism of
action (Anderson, 1992; Striffler et al, 1995). Chromium
does not increase serum insulin concentrations.
Chromium is an essential cofactor for insulin function
and chromium deficiency results in insulin resistance.
The effects of chromium treatment on glucose tolerance
and control of glycemia in humans with diabetes is
controversial. Some studies have identified improved
control of glycemia when humans with NIDDM were
 CANINE DIABETES MELLITUS / 505

Total AUI6h (␮U/ml/6 hr +SD)

3000
VetBooks.ir

2000
*
*
1000 A

0
0 25 50 100 200

Acarbose dose (mg/dog)

FBG MBG8h Gly Hgb

(mg/dl) (mg/dl) (%)

400 400 8

300 300 6 *

200 200 4 B
*
100 100 2

0 0 0

Placebo Acarbose

FIGURE 11-7. A, Mean total insulin secretion for 5 healthy dogs during the first 6 hours after
consumption of a meal and a placebo or 25, 50, 100, or 200 mg of acarbose. Error bars represent
standard error of the mean. * = p<0.05, compared with value obtained after treatment with the placebo.
(From, Robertson J, et al: Effects of the α-glucosidase inhibitor acarbose on postprandial serum glucose
and insulin concentrations in healthy dogs. Am J Vet Res 60:541, 1999.) B, Fasting blood glucose (FBG),
mean blood glucose over an 8-hour time period (MGB8h), and blood total glycosylated hemoglobin
(Gly Hgb) in 5 dogs with insulin-dependent diabetes mellitus treated with insulin and placebo (solid
bars) and insulin and acarbose (hatched bars) for 2 months each in a randomly assigned treatment
sequence. Error bars represent standard deviation. *p = <0.05, compared with placebo value.

treated with chromium picolinate (Mossop, 1983;
Evans, 1989; Anderson et al, 1997); other studies have
failed to identify an effect in human diabetics
(Rabinowitz et al, 1983; Abraham et al, 1992). In one
study, dietary chromium picolinate supplementation
improved results of an intravenous glucose tolerance
test in healthy dogs, compared with healthy dogs not
treated with chromium (Spears et al, 1998). Other
studies failed to identify an effect of dietary chromium
picolinate supplementation on glucose tolerance in
obese dogs during weight reduction (Gross et al, 2000)
or in obese and non-obese cats after 6 weeks of chro-
mium picolinate supplementation (Cohn et al, 1999).
The effect of oral chromium picolinate on control of
glycemia in insulin-treated diabetic dogs was recently
studied at our hospital (Schachter et al, 2001). Glycemic
control was evaluated monthly for 6 months; chro-
mium picolinate (200 to 400 μg per os twice daily) was
administered during the last 3 months. Chromium
picolinate did not affect the results of the complete
blood count, serum biochemistry panel or urinalysis,
nor did it improve control of glycemia in the diabetic
dogs (Fig. 11-8). Similar studies have not been reported
in diabetic cats.
Chromium picolinate is considered a nutraceutical
in the United States and can be purchased in health
food and drug stores. It is inexpensive, and there are
no known toxic effects associated with its ingestion.
Although preliminary studies failed to identify an effect
on control of glycemia, chromium picolinate treatment
may be considered in poorly controlled diabetic dogs
when the cause for poor control of glycemia cannot
be identified and when insulin treatment, by itself, is
ineffective in preventing clinical signs of diabetes. One
commercial diet (Eukanuba Optimum Weight Control,
Iams Co, Dayton, OH) currently marketed for the treat-
ment of diabetes in dogs in the United States contains
chromium picolinate.
HERBS, SUPPLEMENTS, AND VITAMINS. An increasing
number of humans, primarily with type 2 diabetes, are
trying alternative therapies that include herbs, supple-
ments, and vitamins in conjunction with or in lieu of
 506 / CANINE DIABETES MELLITUS

Blood glucose (mg/dl)

500
VetBooks.ir

400

300

200

100

0
0 2 hrs 4 hrs 6 hrs 8 hrs 10 hrs

FIGURE 11-8. Mean (± standard deviation) blood glucose concentrations obtained from 13 dogs with
insulin-dependent diabetes mellitus before insulin and feeding (Time 0) and for 10 hours after insulin
administration only (broken line) or after insulin and chromium tripicolinate administration (solid line).
Dogs were treated for 3 months with insulin and then 3 months with insulin and chromium
tripicolinate. Mean blood glucose concentrations are the mean of all corresponding blood glucose
values obtained during the 10-hour blood sample collection period for each dog at 1, 2, and
3 months of each treatment period. Arrow indicates time of insulin or insulin and chromium picolinate.
(From, Schachter S, et al: Oral chromium picolinate and control of glycemia in insulin-treated diabetic
dogs. J Vet Intern Med 15:379, 2001.)

the more conventional treatment options for diabetes
discussed above. The goals for using herbs, supple-
ments, and vitamins are primarily centered around
decreasing blood glucose, triglyceride, and cholesterol
concentrations, delaying the onset of long-term compli-
cations of diabetes (e.g., coronary artery disease, retin-
opathy), and improving the overall well-being of the
patient (Table 11-13; Roszler, 2001). Proposed beneficial
effects vary with the herb, supplement, or vitamin
used and include delaying nutrient absorption from
the gastrointestinal tract, stimulating insulin secretion,
improving insulin sensitivity, altering lipid metabolism,
improving circulation, and benefits attributed to
antioxidant properties. Some herbs, supplements, and
vitamins, such as ginseng, chromium, fish oils, and
psyllium, have been critically evaluated for efficacy
whereas others are recommended based primarily on
folk lore and testimonials (Pastors et al, 1991; Striffler
et al, 1995; Vuksan et al, 2001). To date, chromium and
vanadium are the only two supplements that have
been critically evaluated in diabetic dogs and cats (see
page 504). Critical studies assessing the effects of
herbs, supplements, and vitamins on diabetic control
and complications are needed before these alternative
therapies can be recommended. Intuitively, it seems
doubtful that the herbs, supplements, and vitamins
listed in Table 11-13 could have much of an impact
in diabetic dogs, in part because these therapies are
primarily used for treating type 2 diabetes and
delaying chronic diabetic complications, both of which
are uncommon in dogs. Although type 2 diabetes and
peripheral neuropathy do occur in cats, the chronic
administration of herbs, supplements, and vitamins to
a carnivore whose metabolism of the active ingredients
in these supplements may differ tremendously from
omnivores is worrisome without prior studies evalu-
ating their efficacy and safety.
Identification and Control of Concurrent
Problems
Concurrent disease and administration of insulin-
antagonistic drugs are commonly identified in the dog
with newly diagnosed diabetes mellitus (Hess et al,
2000b; Peikes et al, 2001). Concurrent disease and
insulin-antagonistic drugs can interfere with tissue
responsiveness to insulin. Tissue responsiveness to
insulin may be impaired as a result of decreased
number of insulin receptors at the surface of the cell
membrane, alterations in insulin receptor binding
affinity, or impairment in one of several postreceptor
steps responsible for activation of glucose transport
systems (Unger and Foster, 1998). Loss of tissue
responsiveness results in insulin resistance and the
severity of insulin resistance is dependent, in part, on
the underlying etiology (see page 524). Insulin resis-
tance may be mild and easily overcome by increasing
the dosage of insulin or may be severe, causing sus-
tained and marked hyperglycemia regardless of the
type and dosage of insulin administered. Some causes
of insulin resistance are readily apparent at the time
diabetes is diagnosed, such as obesity and the admin-
istration of insulin-antagonistic drugs (e.g., gluco-
corticoids). Other causes of insulin resistance are not
readily apparent and require an extensive diagnostic
evaluation to be identified. In general, any concurrent
inflammatory, infectious, hormonal, or neoplastic
disorder can cause insulin resistance and interfere
with the effectiveness of insulin therapy. Identification
 CANINE DIABETES MELLITUS / 507

TABLE 11-13 HERBS, SUPPLEMENTS, AND therapy. The objective during this first visit is not to
VITAMINS THAT HAVE BEEN USED TO TREAT establish perfect glycemic control before sending the
VetBooks.ir

DIABETES MELLITUS IN HUMANS. DOSAGES LISTED dog home. Rather, the objective is to begin to reverse
ARE RECOMMENDATIONS FOR HUMAN DIABETICS the metabolic derangements induced by the disease,
allow the dog to equilibrate to the insulin and change
Improve Hyperglycemia Improve Hyperlipidemia in diet, teach the owner how to administer insulin, and
Alpha-lipoic acid (100-600 mg Vitamin E give the owner a few days to become accustomed to
daily) Evening primrose oil (200 to
Vitamin C (250-1000 mg daily) 500 mg daily)
treating the diabetic dog at home. Adjustments in
Vitamin E (800 IU daily) Fish oils (e.g., omega-3 fatty insulin therapy are made on subsequent evaluations,
Chromium (400 μg daily) acids) once the owner and pet have become accustomed to
Vanadium (5 to 25 mg daily) Selenium (400 μg daily) the treatment regimen.
Fenugreek seeds (5 to 30 g daily) Diabetic dogs are typically evaluated once weekly
American and Asian ginseng Prevent Cataracts
(200 mg daily) until an effective insulin treatment protocol is
Alpha-lipoic acid identified. We inform the owner at the time insulin
Gymnema sylvestre (200 mg
Vitamin C
twice daily) therapy is initiated that it will take approximately one
Psyllium seeds month to establish a satisfactory insulin treatment
Cinnamon (1⁄4 to 1 tsp daily) Prevent Retinopathy
Vitamin E protocol, assuming unidentified insulin antagonistic
Prevent Coronary Artery Pycnogenol (pine bark extract) disease is not present. The goals of therapy are also
Disease explained to the owner. During this month, changes in
Antioxidant insulin dosage, type, and frequency of insulin
Vitamin C
Vitamin E Alpha-lipoic acid administration are common and should be anticipated
Quercetin (100 mg three times Vitamin A (100 to 400 IU by the owner. At each evaluation, the owner’s sub-
daily) daily)
Vitamin C
jective opinion of water intake, urine output, and
Vitamin E overall health of their pet is discussed; a complete
Improve Circulation Pycnogenol physical examination is performed; change in body
Gingko biloba Quercetin weight is noted; and serial blood glucose measure-
Pycnogenol Selenium ments between 7–9 AM and 4–6 PM are assessed.
Prevent/Control Pain from Adjustments in insulin therapy are based on this
Neuropathy information, the pet is sent home, and an appointment
Alpha-lipoic acid is scheduled for the next week to reevaluate the
Vitamin B6 (<100 mg daily) response to any change in therapy. Glycemic control is
Capsaicin (cayenne pepper) attained when clinical signs of diabetes have resolved,
(Topical ointment) the pet is healthy and interactive in the home, its body
Evening primrose oil
weight is stable, the owner is satisfied with the
From Roszler J: Herbs, supplements and vitamins: What to try, what to buy. progress of therapy, and if possible, the blood glucose
Diabetes Interviews August, 2001, p. 45.
concentrations range between 100 and 250 mg/dl
throughout the day.
and treatment of concurrent disease plays an integral Many factors affect the dog’s glycemic control from
role in the successful management of the diabetic dog. day to day, including variations in insulin adminis-
A thorough history, physical examination, and com- tration and absorption, dietary indiscretions and caloric
plete diagnostic evaluation is imperative in the newly intake, amount of exercise, and variables that affect
diagnosed diabetic dog (see Clinical Pathologic insulin responsiveness (e.g., stress, concurrent inflam-
Abnormalities, page 493). mation, infection). As a consequence, the insulin dosage
required to maintain glycemic control typically changes
(increase or decrease) with time (Fig. 11-9). Neverthe-
Initial Adjustments in Insulin Therapy less, a fixed dosage of insulin is administered at home
during the first few months of therapy and changes in
Diabetic dogs require several days to equilibrate to insulin dosage are made only after the owner consults
changes in insulin dosage or preparation. Therefore with the veterinarian.
newly diagnosed diabetic dogs are typically hospitalized Adjustments in insulin dosage are common, and
for no more than 24 to 48 hours to finish the diagnostic eventually a range of “safe” insulin dosages effective
evaluation of the dog and to begin insulin therapy. in maintaining glycemic control are established.
During hospitalization, blood glucose concentrations Insulin dosages outside of the “safe” range usually
are typically determined at the time insulin is cause signs of hypoglycemia or hyperglycemia. As the
administered and at 11 AM, 2 PM, and 5 PM. The intent insulin dosage range becomes apparent and as con-
is to identify hypoglycemia (blood glucose <80 mg/dl) fidence is gained in the owner’s ability to recognize
in those dogs that are unusually sensitive to the actions signs of hypoglycemia and hyperglycemia, the owner
of insulin. If hypoglycemia occurs, the insulin dosage is eventually allowed to make slight adjustments in the
is decreased before sending the dog home. The insulin insulin dosage at home based on clinical observations
dosage is not adjusted in those dogs that remain of their pet’s well-being. However, the owner is
hyperglycemic during these first few days of insulin instructed to stay within the agreed-upon insulin
 508 / CANINE DIABETES MELLITUS

500
VetBooks.ir

400

Blood glucose concentration (mg/dl)

300

200

100

0
8 AM Noon 4 PM 8 PM

FIGURE 11-9. Blood glucose curves in a 7-kg spayed cat receiving 1.2 U/kg beef/pork source lente
insulin (solid line-solid circles) 1 month after initiating insulin therapy, 0.6 U/kg beef/pork source lente
insulin (broken line-solid circles) 4 months later, and 0.8 U/kg beef/pork source lente insulin (broken line-
open circles) 10 months later. Note the fluctuation in insulin dosage required to maintain good glycemic
control. ↑ = SC insulin injection and food.

dosage range. If the insulin dosage is at the upper or
lower end of the established range and the pet is still
symptomatic, the owner is instructed to call us before
making further adjustments in the insulin dosage.
normal blood glucose concentrations is not necessary
in diabetic dogs. Most owners are happy, and most
dogs are healthy and relatively asymptomatic if
most blood glucose concentrations are kept between
100 mg/dl and 250 mg/dl.

TECHNIQUES FOR MONITORING

DIABETIC CONTROL
The basic objective of insulin therapy is to eliminate
the clinical signs of diabetes mellitus while avoiding
the common complications associated with the
disease. Common complications in dogs include
blindness caused by cataract formation, weight loss,
hypoglycemia, recurring ketosis, and poor control of
glycemia secondary to concurrent infection, inflam-
mation, neoplasia, or hormonal disorders. The devas-
tating chronic complications of human diabetes (e.g.,
nephropathy, vasculopathy, coronary artery disease)
require several decades to develop and are uncommon
in diabetic dogs. As such, the need to establish near
History and Physical Examination
The most important initial parameters to assess when
evaluating control of glycemia are the owner’s subjec-
tive opinion of severity of clinical signs and overall
health of their pet, findings on physical examination,
and stability of body weight (Briggs et al, 2000). If the
owner is happy with results of treatment, the physical
examination is supportive of good glycemic control,
and the body weight is stable, the diabetic dog is
usually adequately controlled. We prefer to talk to the
owner and perform the physical examination of the dog
at the beginning of the day (between 7:30 and 9:00 AM)
prior to or within 1 hour of insulin administration, and

we obtain blood for determination of glucose and serum
fructosamine concentration (see below) at that time. In
VetBooks.ir
most well-regulated diabetic dogs, the blood glucose
concentration measured between 7:30 and 9:00 AM and
prior to or within 1 hour of insulin administration
will be between 150 and 250 mg/dl. An early morning
blood glucose concentration less than 150 mg/dl in a
presumably well-controlled diabetic dog raises con-
cern for development of hypoglycemia several hours
after insulin administration, which either produces
no symptoms or produces clinical signs that are not
recognized by the owner. Measurement of serum
fructosamine concentration is indicated in this situation;
identification of a low serum fructosamine concen-
tration (i.e., <350 μmol/L) suggests periods of hypo-
glycemia and a need to reduce the insulin dosage.
Poor control of glycemia should be suspected, and
additional diagnostics (i.e., serial blood glucose curve;
serum fructosamine concentration; tests for concurrent
disorders) or a change in insulin therapy considered if
the owner reports clinical signs (i.e., polyuria, poly-
dipsia, lethargy, signs of hypoglycemia), the physical
examination identifies problems consistent with poor
control of glycemia (e.g., thin or emaciated, poor hair
coat), the dog is losing weight, or the blood glucose
concentration measured in the early morning is
greater than 300 mg/dl (Briggs et al, 2000).
Documenting an increased blood glucose concentra-
tion does not by itself confirm poor control of glycemia.
Stress or excitement can cause marked hyperglycemia
that does not reflect the patient’s responsiveness to
insulin and can lead to the erroneous belief that the
diabetic dog is poorly controlled (see page 567). If a
discrepancy exists between the history, physical exami-
nation findings, and blood glucose concentration or
if the dog is fractious, aggressive, excited, or scared
and the blood glucose concentration is known to be
unreliable, measurement of serum fructosamine con-
centration should be done to further evaluate status of
glycemic control.
Serum Fructosamine Concentration
Fructosamines are glycated proteins found in blood
that are used to monitor control of glycemia in diabetic
dogs and cats (Reusch et al, 1993; Crenshaw et al, 1996;
Elliott et al, 1999). Fructosamines result from an irrever-
sible, nonenzymatic, insulin-independent binding of
glucose to serum proteins. All humans and animals
have circulating fructosamines. Serum fructosamine
concentrations are a marker of mean blood glucose
concentration during the circulating lifespan of the
protein, which varies from 1 to 3 weeks, depending
on the protein (Kawamoto et al, 1991). The extent of
glycosylation of serum proteins is directly related to
the blood glucose concentration; the higher the average
blood glucose concentration during the preceding 2 to
3 weeks, the higher the serum fructosamine concen-
tration, and vice versa. Serum fructosamine concen-
tration is not affected by acute increases in the blood
CANINE DIABETES MELLITUS / 509
glucose concentration, as occurs with stress or
excitement-induced hyperglycemia (Lutz et al, 1995;
Crenshaw et al, 1996). Serum fructosamine concen-
rations can be measured during the routine evaluation
of glycemic control performed every 3 to 6 months;
to clarify the effect of stress or excitement on blood
glucose concentrations; to clarify discrepancies between
the history, physical examination findings, and serial
blood glucose concentrations; and to assess the effec-
tiveness of changes in insulin therapy (see page 567).
Serum fructosamine concentrations increase when
glycemic control of the diabetic dog or cat worsens
and decrease when glycemic control improves (see
Fig. 12-17, page 563).
Fructosamine is measured in serum, which should
be frozen and shipped on cold packs overnight to the
laboratory. Although freezing does not cause a signifi-
cant change in results, storage of serum at room tem-
perature overnight can decrease serum fructosamine
results by 10%; storage of serum in the refrigerator can
also decrease the serum fructosamine result (Jensen,
1992). An automated colorimetric assay using nitroblue
tetrazolium chloride is used for measurement of
fructosamine concentrations in serum (Baker et al,
1985). A linear relationship between serum total
protein, albumin, and fructosamine concentration has
been identified, and hypoproteinemia (total protein
<5.5 g/dl) and hypoalbuminemia (albumin <2.5 g/dl)
can decrease the serum fructosamine concentration
below the reference range in healthy dogs and presum-
ably diabetic dogs as well (Fluckiger et al, 1987; Loste
and Marca, 1999; Reusch and Haberer, 2001). A similar
decrease in serum fructosamine results has been
identified with hyperlipidemia (cholesterol >380 mg/dl
and triglycerides >150 mg/dl) and azotemia (blood urea
nitrogen >28 mg/dl and serum creatinine >1.7 mg/dl)
(Reusch and Haberer, 2001). A significant change in
serum fructosamine results was not detected in healthy
dogs with hyperproteinemia or hyperbilirubinemia.
In our laboratory, the normal reference range for
serum fructosamine in dogs is 225 to 365 μmol/L, a
range determined in healthy dogs with persistently
normal blood glucose concentrations (Briggs et al,
2000). Serum fructosamine concentration in newly diag-
nosed diabetic dogs ranged from 320 to 850 μmol/L.
The normal serum fructosamine concentration in a
few diabetic dogs suggests that hyperglycemia severe
enough to cause clinical signs had been present for
only a short time before diagnosis.
Interpretation of serum fructosamine in a diabetic
dog must take into consideration the fact that hyper-
glycemia is common, even in well-controlled diabetic
dogs (Table 11-14). Most owners are happy with their
pet’s response to insulin treatment if serum fruc-
tosamine concentrations can be kept between 350 and
450 μmol/L. Values greater than 500 μmol/L suggest
inadequate control of the diabetic state, and values
greater than 600 μmol/L indicate serious lack of
glycemic control. Serum fructosamine concentrations
in the lower half of the normal reference range (i.e.,
<300 μmol/L) or below the normal reference range
 510 / CANINE DIABETES MELLITUS

TABLE 11-14 SAMPLE HANDLING, to adjust insulin therapy (see Serial Blood Glucose
METHODOLOGY, AND NORMAL VALUES FOR Curve, page 511).
VetBooks.ir

SERUM FRUCTOSAMINE CONCENTRATIONS

MEASURED IN OUR LABORATORY IN DOGS
Fructosamine
Blood Glycated Hemoglobin Concentration
Blood sample 1-2 ml serum Glycated hemoglobin (Gly Hb) is a glycated protein
Sample handling Freeze until assayed
Methodology Automated colorimetric assay using
that results from an irreversible, nonenzymatic,
nitroblue tetrazolium chloride insulin-independent binding of glucose to hemoglobin
Factors affecting results Hypoproteinemia and in red blood cells. Blood Gly Hb is a marker of mean
hypoalbuminemia (decreased), blood glucose concentration during the circulating
hyperlipidemia (decreased), lifespan of the red blood cell, which is approximately
azotemia (decreased), storage at
room temperature (decreased) 110 days in the dog (Jain, 1993). The extent of
Normal range 225 to 365 μmol/L glycosylation of hemoglobin is directly related to the
Interpretation in blood glucose concentration; the higher the average
diabetic dogs: blood glucose concentration during the preceding 3 to
Excellent control 350-400 μmol/L
Good control 400-450 μmol/L
4 months, the higher the blood Gly Hb, and vice versa.
Fair control 450-500 μmol/L Gly Hb rather than fructosamine is used to monitor
Poor control >500 μmol/L long-term effectiveness of treatment in human dia-
Prolonged hypoglycemia <300 μmol/L betics, in part because diabetic humans self-monitor
their blood glucose and adjust their insulin dose daily
and Gly Hb assesses a longer treatment interval than
fructosamine (i.e., 3 to 4 months versus 2 to 3 weeks,
should raise concern for significant periods of hypo- respectively). In contrast, measurement of serum
glycemia in the diabetic dog. The Somogyi phenomenon fructosamine is used more commonly to assess control
(i.e., glucose counterregulation) should be suspected of glycemia in diabetic dogs and cats, in part because
if clinical signs (i.e., polyuria, polydipsia, polyphagia, the assay is readily available commercially and is
weight loss) are present in a diabetic dog with a serum better for assessing the impact of changes in insulin
fructosamine concentration less than 400 μmol/L, therapy on control of glycemia in fractious dogs and
assuming hypoproteinemia or hypoalbuminemia is cats because concentrations of fructosamine change
not present (see Table 11-14). Increased serum fruc- more quickly than Gly Hb (i.e., 2 to 3 weeks versus 3
tosamine concentrations (i.e., >500 μmol/L) suggest to 4 months, respectively).
poor control of glycemia and a need for insulin adjust- In the dog and cat, there are three fractions of Gly
ments; however, increased serum fructosamine con- Hb: one major fraction (Gly HbA1c) that binds glucose
centrations do not identify the underlying problem and two minor fractions (Gly HbA1a and Gly HbA1b)
(Fig. 11-10). Evaluation of serial measurements of blood that do not (Hasegawa et al, 1991; 1992). Measurement
glucose concentration is required to determine how of Gly HbA1c is typically used to evaluate status of

600 600 12

500 500 10

400 400 8

300 300 6

200 200 4

100 100 2

0 0 0
MBG8h Fructosamine GLY HGB
(mg/dl + SD) (␮mol/L+ SD) (% + SD)

FIGURE 11-10. Mean blood glucose concentration determined over an 8-hour period (MBG8h), serum
fructosamine concentration, and blood total glycosylated hemoglobin (Gly Hgb) in 10 diabetic dogs
with poor control of glycemia caused by the Somogyi phenomenon (hatched bars) and 12 diabetic dogs
with poor control of glycemia caused by hyperadrenocorticism-induced insulin resistance (solid bars).
Note the similar glycated protein results in both groups of dogs. Although the average blood glucose
concentration is lower on the day hypoglycemia is identified in dogs with the Somogyi phenomenon,
high blood glucose concentrations on subsequent days result in high glycated protein concentrations.
 CANINE DIABETES MELLITUS / 511

glycemic control in human diabetics, whereas studies TABLE 11-15 SAMPLE HANDLING,
in diabetic dogs have used assays that measure all METHODOLOGY, AND NORMAL VALUES FOR
VetBooks.ir

three fractions, that is, total Gly Hg (Elliott et al, 1997)
or Gly HbA1c (Marca et al, 2000; Marca and Loste,
2001). Most techniques that measure total Gly Hg have
been shown to be clinically valid for assessing degree
of diabetic control (Mahaffey and Cornelius, 1982;
Elliott et al, 1997). Depending on the methodology,
however, acute hyperglycemia may cause an increase
in the concentration of total Gly Hb. Similar studies
using Gly HbA1c to assess status of glycemic control
have not yet been reported in diabetic dogs.
Gly Hb is measured in whole blood collected in
EDTA. Blood samples can be refrigerated up to a week
without significant change in the Gly Hb concentra-
tion. In dogs, blood Gly Hb has been measured by
affinity chromatography (Wood and Smith, 1982;
Elliott et al, 1997), colorimetric analysis (Mahaffey and
Cornelius, 1981), ion-exchange high performance
liquid chromatography (Hasegawa et al, 1991), and
immunoturbidometric assay (Marca and Loste, 2001).
We measure total Gly Hb using an affinity chroma-
tography technique whose results are not affected by
acute hyperglycemia (Elliott et al, 1997). Assays for
measuring Gly Hb are designed for use in humans. As
such, it is important that the Gly Hb assay be validated
for use in the dog and that a normal reference range is
established for the dog. In our experience, several Gly
Hb assays, especially in-house automated analyzers for
rapid measurement of Gly HbA1c in human diabetics,
have not provided valid results in dogs or cats. Any
condition that affects red cell life span may affect Gly
Hb concentration. Anemia and polycythemia can
falsely decrease and increase Gly Hb concentrations,
respectively (Elliott et al, 1997). The hematocrit should
be taken into consideration when interpreting Gly Hb
concentrations.
In our laboratory, the normal reference range for
total Gly Hb as measured by affinity chromatography
in dogs is 1.7% to 4.9%; a range determined in healthy
dogs with persistently normal blood glucose concen-
trations (Elliott et al, 1997). Blood total Gly Hb in
newly diagnosed diabetic dogs ranged from 6.0% to
15.5%. Interpretation of blood Gly Hb in a diabetic dog
must take into consideration the fact that hyper-
glycemia is common, even in well-controlled diabetic
dogs (Table 11-15). Most owners are happy with their
pet’s response to insulin treatment if blood total Gly
Hb can be kept between 4% and 6%. Values greater
than 7% suggest inadequate control of the diabetic
state and values greater than 8% indicate serious lack
of glycemic control. Blood total Gly Hb less than 4%
should raise concern for significant periods of hypo-
glycemia in the diabetic dog, assuming anemia is not
present. Increased total Gly Hb (i.e., >7%) suggests
poor control of glycemia and a need for insulin adjust-
ments; however, increased total Gly Hb does not
identify the underlying problem (see Fig. 11-10). Evalu-
ation of serial measurements of blood glucose concen-
tration is required to determine how to adjust insulin
therapy (see Serial Blood Glucose Curve, right column).
BLOOD TOTAL GLYCOSYLATED HEMOGLOBIN
CONCENTRATIONS MEASURED IN OUR
LABORATORY IN DOGS
Total Glycosylated Hemoglobin
Blood sample 1-2 ml whole blood in EDTA
Sample handling Refrigerate until assayed
Methodology Affinity chromatography and
hemolysates derived from canine
red blood cells
Factors affecting results Storage at room temperature
(decreased); storage at 4° C for
longer than 7 days (decreased);
anemia (Hct<35%) (decreased)
Normal range 1.7% to 4.9%
Interpretation in
diabetic dogs:
Excellent control 4% to 5%
Good control 5% to 6%
Fair control 6% to 7%
Poor control >7%
Prolonged hypoglycemia <4%
Urine Glucose Monitoring
Occasional monitoring of urine for glycosuria and
ketonuria is helpful in those diabetic dogs that have
problems with recurring ketosis or hypoglycemia to
determine whether ketonuria or persistent negative
glycosuria is present, respectively. We do not have the
owner adjust daily insulin dosages based on morning
urine glucose measurements in diabetic dogs, except
to decrease the insulin dose in dogs with recurring
hypoglycemia and persistent negative glycosuria. In
our experience, the vast majority of diabetic dogs
develop complications as a result of owners being
misled by morning urine glucose concentrations. On
occasion we recommend evaluation (e.g., on the week-
ends) of multiple urine samples obtained throughout
the day and early evening. The well-controlled diabetic
pet should have urine that is free of glucose for most
of each 24-hour period. Persistent glycosuria through-
out the day and night suggests inadequate control of
the diabetic state and the need for a more complete
evaluation of diabetic control using techniques
discussed in this section.
Serial Blood Glucose Curve
If an adjustment in insulin therapy is deemed neces-
sary after reviewing the history, physical examination,
changes in body weight, and serum fructosamine
concentration, then a serial blood glucose curve
should be generated to provide guidance in making
the adjustment unless blood glucose measurements
are unreliable because of stress, aggression, or excite-
ment (see page 567). The serial blood glucose curve
provides guidelines for making rational adjustments
in insulin therapy. Evaluation of a serial blood glucose
 512 / CANINE DIABETES MELLITUS
curve is mandatory during the initial regulation of the
diabetic dog, is periodically of value to assess glycemic
VetBooks.ir
control despite the fact that a dog may appear to be
doing well in the home environment, and is necessary
to reestablish glycemic control in the dog in which
clinical manifestations of hyperglycemia or hypo-
glycemia have developed. Reliance on history, physical
examination, body weight, and serum fructosamine
concentration to determine when a serial blood glucose
curve is needed help reduce the frequency of perform-
ing serial blood glucose curves, reduce the number of
venipunctures, and shorten the time the dog spends in
the hospital, thereby minimizing the dog’s aversion
(and stress) to these evaluations and improving the
chances of obtaining meaningful blood glucose results
when a serial blood glucose curve is needed.
PROTOCOL FOR GENERATING THE SERIAL BLOOD GLUCOSE
CURVE IN THE HOSPITAL. When assessing glycemic
control, the veterinarian or clinician should follow the
insulin and feeding schedule used by the owner and
blood should be obtained every 1 to 2 hours through-
out the day for glucose determination. Owners of finicky
diabetic dogs should feed their pets at their home, not
at the hospital. Inappetence can profoundly alter the
results of a serial blood glucose curve (Fig. 11-11). If
insulin is usually given within 1 hour prior to the pet’s
presentation to the clinic, insulin is given by the owner
(using their insulin and syringe) in the hospital after an
initial blood glucose is obtained. The entire insulin
administration procedure should be closely evaluated
by a veterinary technician. If the owner usually
administers insulin before 6 AM, we recommend that the
owner give the insulin at 6 AM and bring the pet to the
hospital at the first appointment of the day. It is more
important to maintain the pet’s daily routine than to risk
inaccurate blood glucose results caused by inappetence
in the hospital or insulin administration at an unusual
time. The exception are those instances where the
clinician wants to evaluate owner administration
technique using insulin rather than physiologic saline.
400
Serum glucose (mg/dl)
300
200
100
0 control of glycemia. The biggest factors inducing stress
8 AM 12 PM 4 PM 8 PM Mid 4 AM
FIGURE 11-11. Mean blood glucose concentrations in 8 diabetic
dogs following administration of NPH insulin (↑) and feeding equal-
sized meals at 8 AM and 6 PM (solid line) or feeding nothing (broken
line) during the 24 hours of blood sampling. (From Nelson RW,
Couto CG: Essentials of Small Animal Internal Medicine. St Louis,
Mosby-Year Book, 1992, p 572.)
Blood glucose concentrations are typically deter-
mined by either a point-of-care glucose analyzer or
hand-held portable blood glucose monitoring device.
Commercially available portable blood glucose moni-
toring devices provide blood glucose concentrations
reasonably close to those obtained with reference
methods (i.e., glucose oxidase and hexokinase methods),
although results may consistently overestimate or
underestimate actual glucose values (Fig. 11-12; Cohn et
al, 2000; Wess and Reusch, 2000a). In our experience,
blood glucose values determined by the majority of
portable blood glucose monitoring devices are typically
lower than actual glucose values determined by
reference methods. This may result in an incorrect diag-
nosis of hypoglycemia or the misperception that
glycemic control is better than it actually is. Failure to
consider this “error” could result in insulin underdosage
and the potential for persistence of clinical signs despite
what appears to be “acceptable” blood glucose results.
By evaluating serial blood glucose measurements
every 1 to 2 hours throughout the day, the clinician will
be able to determine whether the insulin is effective and
to identify the glucose nadir, time of peak insulin effect,
duration of insulin effect, and severity of fluctuation in
blood glucose concentrations in that particular diabetic
dog. Obtaining only 1 or 2 blood glucose concentrations
has not been reliable for evaluating the effect of a given
insulin dose (Fig. 11-13). Persistent poor control of the
diabetic state often stems from misinterpretation of the
effects of insulin that is based on assessment of only 1 or
2 blood glucose concentrations.
The ideal goal of insulin therapy in diabetic dogs is
to maintain the blood glucose concentration between
100 mg/dl and 250 mg/dl throughout the day and
night. This goal can be very difficult, and in some dia-
betic dogs, impossible to attain. The ultimate decision
on whether to adjust insulin therapy must always take
into consideration the owner’s perception of how the
pet is doing at home, findings on physical examination,
changes in body weight, serum fructosamine con-
centrations, as well as results of serial blood glucose
measurements. Many diabetic dogs do well despite
blood glucose concentrations consistently in the high
100s to low 300s.
PROTOCOL FOR GENERATING THE SERIAL BLOOD GLUCOSE
CURVE AT HOME. Hyperglycemia induced by stress,
aggression, or excitement is the single biggest problem
affecting accuracy of the serial blood glucose curve,
especially in cats. Stress can override the glucose-
lowering effect of the insulin injection, causing high
blood glucose concentrations despite the presence of
adequate amounts of insulin in the circulation and
leading to a spiraling path of insulin overdosage,
hypoglycemia, Somogyi phenomenon, and poor
8 AM
hyperglycemia are hospitalization and multiple
venipunctures. An alternative to hospital-generated
blood glucose curves is to have the owner generate the
blood glucose curve at home using the ear or lip prick
technique and a portable home glucose monitoring
device that allows the owner to touch the drop of
blood on the ear or lip with the end of the glucose-test
 CANINE DIABETES MELLITUS / 513
VetBooks.ir

FIGURE 11-12. Scatterplots of the difference between blood glucose concentration obtained with four
portable blood glucose meters and concentration obtained with a reference method versus concentration
obtained with the reference method for blood samples from 170 dogs. (From Wess G, Reusch C:
Evaluation of five portable blood glucose meters for use in dogs. JAVMA 216:203, 2000.).

strip (Wess and Reusch, 2000b). This technique is
usually reserved for diabetic dogs in which the
reliability of blood glucose results generated in the
veterinary hospital is questionable. See page 565 for
more information on monitoring blood glucose con-
centrations at home.
INTERPRETING THE SERIAL BLOOD GLUCOSE CURVE.
Results of the blood glucose curve will allow the veteri-
narian to assess the effectiveness of the administered
insulin to lower the blood glucose concentration and
determine the glucose nadir and the duration of insulin
effect (Fig. 11-14). Ideally, all blood glucose concen-
trations should range between 100 and 250 mg/dl
during the time period between insulin injections.
Typically, the highest blood glucose concentrations
occur at the time of each insulin injection, but this does
not always occur.
Insulin effectiveness, glucose nadir, and duration of
insulin effect are the critical determinations from the
serial blood glucose curve. The effectiveness of insulin
is the first parameter to assess. Is the insulin effective
in lowering the blood glucose concentration? The
insulin dosage, the highest blood glucose concen-
tration, and the difference between the highest and
lowest blood glucose concentrations (i.e., blood
glucose differential) must be considered simultane-
ously when assessing insulin effectiveness. For
example, a blood glucose differential of 50 mg/dl is
acceptable if the blood glucose ranges between 120
and 170 mg/dl but is unacceptable if the blood glucose
ranges between 350 and 400 mg/dl. Similarly, a blood
glucose differential of 100 mg/dl indicates insulin
effectiveness if the patient receives 0.4 U of insulin per
kilogram of body weight but suggests insulin
 Blood glucose (mg/dl)
500

400
VetBooks.ir

300

200

100

0
8 AM Noon 4 PM 8 PM

FIGURE 11-13. Blood glucose concentration curve in a Dachshund receiving 0.8 U of recombinant
human lente insulin per kilogram body weight twice a day (solid line), a Miniature Poodle receiving
0.6 U of recombinant human lente insulin per kilogram body weight twice a day (broken line), and a
Terrier-mix receiving 1.1 U of recombinant human lente insulin per kilogram body weight twice a day
(dotted line). Insulin and food was given at 8 AM for each dog. Interpretation of the blood glucose curves
suggest short duration of insulin effect in the Dachshund, insulin underdosage in the Miniature
Poodle, and the Somogyi effect in the Terrier-mix. Notice that the blood glucose concentrations were
similar in all dogs at 2 PM and 4 PM and the glucose results at these times do not establish the diagnosis
in any of the dogs. (From Nelson RW, Couto CG: Small Animal Internal Medicine, 3rd ed. St Louis,
Mosby Inc, 2003, p 741.)

Guidelines for interpreting serial

blood glucose curve

1. Was insulin effective in lowering blood glucose?

Yes No

2. What is the lowest blood glucose? 2. Measure serum fructosamine

<80 mg/dl 80-150 mg/dl >150 mg/dl >500 μmol/L <450 μmol/L

↓ Insulin dosage ↑ Insulin dosage

and reevaluate in and reevaluate in Consider stress
7-10 days 7-10 days hyperglycemia or
Somogyi

3. What is the duration of insulin effect? 3. What is the insulin dosage?

<10 hours 10-14 hours >14 hours <1.0 u/kg >1.0 u/kg

Change to longer- Measure serum Change to shorter- ↑ Insulin dosage Consider insulin
acting insulin q12h; fructosamine acting insulin q12h; and reevaluate underdosage, Somogyi,
change to shorter- administer current in 7-14 days and causes of insulin
acting insulin TID insulin SID ineffectiveness
(see Table 11-16)
If no improvement,
Reevaluate in <450 μmol/L >500 μmol/L Reevaluate in consider Somogyi and
7-14 days 7-14 days causes of insulin
ineffectiveness (see
No change Consider Table 11-16)
Somogyi or
insulin
underdosage

FIGURE 11-14. Algorithm for interpreting results of a serial blood glucose concentration curve. (From
Nelson RW, Couto CG: Small Animal Internal Medicine, 3rd ed. St Louis, Mosby Inc, 2003, p 742.)

resistance if the patient receives 2.2 U of insulin per
kilogram.
VetBooks.ir
If the insulin is not effective in lowering the blood
glucose concentration, the clinician should consider
insulin underdosage and the differentials for insulin
ineffectiveness and resistance (see Recurrence or
Persistence of Clinical Signs, page 518). In general,
insulin underdosage should be considered if the
insulin dosage is less than 1.0 U/kg per injection in the
diabetic dog and insulin ineffectiveness and resistance
should be considered if the insulin dosage exceeds
1.0 to 1.5 U/kg per injection. The veterinarian should
always be wary of the Somogyi phenomenon, espe-
cially in toy and miniature breeds, and the effect of
stress on the blood glucose results.
If insulin is effective in lowering the blood glucose
concentration, the next parameter to assess is the
lowest blood glucose (i.e., glucose nadir). The glucose
nadir should ideally fall between 100 and 125 mg/dl.
If the glucose nadir is greater than 150 mg/dl, the
insulin dosage may need to be increased, and if the
nadir is less than 80 mg/dl, the insulin dosage should
be decreased. For the latter, the insulin dosage is
typically decreased approximately 10% to 25%. If
the dog is receiving a large amount of insulin (e.g.,
>2.2 U/kg/injection), concurrent insulin resistance or
the Somogyi phenomenon should be considered. A
diagnostic evaluation for insulin resistance may be
warranted (see page 524), and/or glycemic regulation
started over again using the insulin dose recommended
for the initial regulation of the diabetic dog (see page
496). Control of glycemia should be reevaluated 7 to
14 days after initiating the new dose of insulin and
adjustments in the insulin dose made accordingly.
Duration of insulin effect can be assessed if the
glucose nadir is greater than 80 mg/dl and there has
not been a rapid decrease in the blood glucose con-
centration after insulin administration. Assessment of
duration of insulin effect may not be valid when the
blood glucose decreases to less than 80 mg/dl or
decreases rapidly because of the potential induction of
the Somogyi phenomenon, which can falsely shorten
the apparent duration of insulin effect (see page 519).
The duration of effect is roughly defined as the time
from the insulin injection through the lowest glucose
and until the blood glucose concentration exceeds 200
to 250 mg/dl (Fig. 11-15). Duration of effect of lente
and NPH insulin is 10 to 14 hours in approximately
90% of diabetic dogs, necessitating twice daily insulin
treatment. Dogs are usually symptomatic for diabetes
if the duration of insulin effect is less than 10 hours
(see page 520) and may develop hypoglycemia or the
Somogyi phenomenon if the duration of insulin effect
is greater than 14 hours and the insulin is being
administered twice a day (Fig. 11-16).
Alterations in the dosage of insulin are often
necessary at the same time as alterations in type of
insulin. Adjustments in insulin dosage are usually
dictated by the change in type of insulin and the
glucose nadir. Longer-acting insulin (i.e., ultralente,
PZI, glargine) is less potent than intermediate-acting
CANINE DIABETES MELLITUS / 515
400
Serum glucose (mg/dl)
300
200
A
100
17h Duration
0
8 AM 12 PM 4 PM 8 PM Mid 4 AM 8 AM
400
Serum glucose (mg/dl)
300
200
B
100
13h Duration
0
8 AM 12 PM 4 PM 8 PM Mid 4 AM 8 AM
FIGURE 11-15. A, Blood glucose curve in a Poodle mix receiving
beef/pork source NPH insulin, 1.1 U/kg body weight (solid line).
Insulin is given at 8 AM, and the dog is fed at 8 AM and 6 PM. Hyper-
glycemia in excess of 225 mg/dl redevelops around midnight. The
duration of insulin action is approximately 17 hours. Switching to
beef/pork source ultralente insulin, 1.3 U/kg once daily, improved
glycemic control (broken line) in this dog. B, Blood glucose curve in a
Terrier mix receiving beef/pork source NPH insulin, 1 U/kg body
weight (solid line). Insulin is given at 8 AM, and the dog is fed at 8 AM
and 6 PM. Notice the postprandial hyperglycemia following the
evening meal. Hyperglycemia in excess of 225 mg/dl redevelops
around 9 PM. The duration of insulin action is approximately 13 hours.
Switching to beef/pork source NPH insulin, 1 U/kg twice daily at
12-hour intervals and feeding at the time of the insulin injection
improved glycemic control (broken line) in the dog. (From Nelson RW,
Couto CG: Essentials of Small Animal Internal Medicine. St Louis,
Mosby-Year Book, 1992, p 570.)
insulin (i.e., NPH, lente). It takes more of a less potent
insulin to get a comparable decrease in the blood
glucose concentration, compared with a more potent
insulin (see Fig. 12-11, page 551). When switching from
a more potent to a less potent insulin, the dose of insulin
is usually not changed if the glucose nadir is between
80 and 120 mg/dl and is increased approximately 10%
if the glucose nadir is greater than 120 mg/dl. When
switching from a less potent to a more potent insulin,
the dose of insulin is increased approximately 10% if
the glucose nadir is greater than 150 mg/dl, is usually
not changed if the glucose nadir is between 120 and
150 mg/dl, and is decreased approximately 10% if the
glucose nadir is between 80 and 120 mg/dl. Control of
glycemia should be reevaluated 7 to 14 days after
initiating the new type of insulin and adjustments in
the insulin dose made accordingly.
REPRODUCIBILITY OF THE SERIAL BLOOD GLUCOSE
CURVE. The reproducibility of serial blood glucose
 516 / CANINE DIABETES MELLITUS
Blood glucose (mg/dl)
500
VetBooks.ir
400
300
200
100
0
8 AM Noon 4 PM 8 PM
FIGURE 11-16. Blood glucose concentration curves obtained from 3
diabetic dogs treated with recombinant human lente insulin twice a
day, illustrating a difference between dogs in the duration of insulin
effect. The insulin is effective in lowering the blood glucose concen-
tration in all dogs and the blood glucose nadir is between 100 and
175 mg/dl for the dogs. However, the duration of insulin effect is
approximately 12 hours (solid line) in one dog with good control of
glycemia (ideal duration of effect), approximately 8 hours (dotted
line) in one dog with persistently poor control of glycemia (short
duration of effect), and greater than 12 hours (broken line) in one dog
with a history of good days and bad days of glycemic control
(prolonged duration of effect); a history suggestive of the Somogyi
phenomenon. (From Nelson RW, Couto CG: Small Animal Internal
Medicine, 3rd ed. St Louis, Mosby Inc, 2003, p 743.)
curves varies from dog to dog. In some dogs, results of
serial blood glucose curves may vary dramatically
from day to day or month to month, depending, in
part, on the actual amount of insulin administered and
absorbed from the subcutaneous site of deposition, and
the interaction between insulin, diet, exercise, stress,
excitement, presence of concurrent disorders, and
counterregulatory hormone secretion (i.e., secretion of
glucagon, epinephrine, cortisol, growth hormone). In
other dogs, serial blood glucose curves are reasonably
consistent from day to day and month to month. Lack
of consistency in results of serial blood glucose curves
creates frustration for many veterinarians. However, it
is important to remember that this lack of consistency
is a direct reflection of all the variables that affect the
blood glucose concentration in diabetics. Daily self-
monitoring of blood glucose concentrations and daily
adjustments in insulin dose are recommended in
human diabetics to minimize the effect of these vari-
ables on control of glycemia. A similar approach for
diabetic dogs and cats will undoubtedly become more
common in the future, as home glucose monitoring
techniques are refined. For now, initial assessment of
control of glycemia is based on the owner’s perception
of their diabetic pet’s health combined with periodic
examinations by the veterinarian. Serial blood glucose
measurements are indicated if poor control of glycemia
is suspected. The purpose of serial blood glucose
measurements is to obtain a glimpse at the actions of
insulin in that diabetic animal and ideally identify a
reason (e.g., short duration of insulin effect) that could
explain why the diabetic dog is poorly controlled.
We prefer to adjust insulin therapy based on inter-
pretation of a single serial blood glucose curve and
typically rely on owner perceptions of clinical response
and change in serum fructosamine concentration to
initially assess the impact of the adjustment on control
of glycemia (see page 509). If problems persist, we
consider repeating the serial blood glucose curve. We
rarely perform serial blood glucose curves on multiple,
consecutive days because it promotes stress-induced
hyperglycemia. In addition, information gained from
a prior serial blood glucose curve should never be
assumed to be reproducible on subsequent curves,
especially if several weeks to months have passed or
the dog has developed recurrence of clinical signs. The
protocol for the serial blood glucose curve should be
followed each time a glucose curve is generated.
Role of Serum Fructosamine in Aggressive,
Excitable, or Stressed Dogs
Serial blood glucose curves are unreliable in aggressive,
excitable, or stressed dogs because of problems related
to stress-induced hyperglycemia. In these dogs, the
clinician must make an educated guess as to where the
problem lies (e.g., wrong type of insulin, low dose),
make an adjustment in therapy, and rely on changes in
serum fructosamine to assess the benefit of the change
in treatment. Refer to page 567 for more information
on use of serum fructosamine in diabetic pets with
stress-induced hyperglycemia.
INSULIN THERAPY DURING SURGERY
Generally, surgery should be delayed in diabetic dogs
until the animal’s clinical condition is stable and the
diabetic state is controlled with insulin. The exception
are those situations in which surgery is required to
eliminate insulin resistance (e.g., ovariohysterectomy
in a diestrus bitch) or to save the animal’s life. The
surgery itself does not pose a greater risk in a stable
diabetic animal than in a nondiabetic animal. The
concern is the interplay between insulin therapy and
the lack of food intake during the perioperative period.
The stress of anesthesia and surgery also causes the
release of diabetogenic hormones, which in turn pro-
motes ketogenesis (see Chapter 13, page 580). Insulin
should be administered during the perioperative
period to prevent severe hyperglycemia and minimize
ketone formation. To compensate for the lack of food
intake and prevent hypoglycemia, the amount of
insulin administered during the perioperative period
is decreased and IV dextrose is administered, when
needed. To correct marked hyperglycemia (i.e., blood
glucose concentration greater than 300 mg/dl), regular
crystalline insulin is administered intramuscularly or
by continuous IV infusion. Frequent blood glucose

monitoring and appropriate adjustments in therapy
are the key to avoiding hypoglycemia and severe
VetBooks.ir
hyperglycemia during the perioperative period.
We use the following protocol during the peri-
operative period in dogs and cats undergoing surgery.
The day before surgery, the animal is given its normal
dose of insulin and fed as usual. Food is withheld after
10 PM. On the morning of the procedure the blood
glucose concentration is measured before the animal is
given insulin. If the blood glucose concentration is less
than 100 mg/dl, insulin is not given and an IV infusion
of 2.5% to 5% dextrose is initiated. If the blood glucose
concentration is between 100 and 200 mg/dl, one quarter
of the animal’s usual morning dose of insulin is given
and an IV infusion of dextrose is initiated. If the blood
glucose concentration is more than 200 mg/dl, one-
half of the usual morning dose of insulin is given but
the IV dextrose infusion is withheld until the blood
glucose concentration is less than 150 mg/dl. In all
three situations the blood glucose concentration is
measured every 30 to 60 minutes during the surgical
procedure. The goal is to maintain the blood glucose
concentration between 150 and 250 mg/dl during the
perioperative period. A 2.5% to 5% dextrose infusion
is administered IV and regular crystalline insulin
administered intermittently as needed to eliminate
or prevent hypoglycemia and severe hyperglycemia,
respectively. When the blood glucose concentration
exceeds 300 mg/dl, the dextrose infusion should be
discontinued and the blood glucose concentration
evaluated 30 and 60 minutes later. If the blood glucose
concentration remains greater than 300 mg/dl, regular
crystalline insulin is administered intramuscularly at
approximately 20% of the dosage of long-acting insulin
being used at home. Subsequent doses of regular
crystalline insulin should be given no more frequently
than every 4 hours, and the dosage should be adjusted
based on the effect of the first insulin injection on the
blood glucose concentrations.
On the day after surgery the diabetic dog or cat can
usually be returned to the routine schedule of insulin
administration and feeding. An animal that is not
eating can be maintained with IV dextrose infusions
and regular crystalline insulin injections given sub-
cutaneously every 6 to 8 hours. Once the animal is
eating regularly, it can be returned to its normal insulin
and feeding schedule.
COMPLICATIONS OF INSULIN THERAPY
Symptomatic and Asymptomatic Hypoglycemia
Hypoglycemia is a common complication of insulin
therapy. Hypoglycemia may be symptomatic or
asymptomatic. Asymptomatic hypoglycemia is more
prevalent than symptomatic hypoglycemia in diabetic
dogs and cats. Symptomatic hypoglycemia is most apt
to occur following sudden large increases in the insulin
dose, with excessive overlap of insulin action in dogs
receiving insulin twice a day, during unusually stren-
CANINE DIABETES MELLITUS / 517
uous exercise, and following prolonged inappetence.
In these situations, severe hypoglycemia may occur
before glucose counterregulation (i.e., secretion of
glucagon, cortisol, epinephrine, and growth hormone)
is able to compensate for and reverse the hypo-
glycemia. Signs of hypoglycemia include lethargy,
weakness, head tilting, ataxia, seizures, and coma. The
occurrence and severity of clinical signs is dependent
on the rate of blood glucose decline and the severity of
hypoglycemia. Symptomatic hypoglycemia is treated
with glucose administered as food, sugar water, or
dextrose IV (see Chapter 14, page 638). Whenever signs
of hypoglycemia occur, the owner should be instructed
to stop insulin therapy until hyperglycemia and
glycosuria recur. Urine glucose testing by the owner
with the dog in its home environment is useful for
identifying when glycosuria recurs. The adjustment in
the subsequent insulin dosage is somewhat arbitrary;
as a general rule of thumb, the insulin dosage initially
should be decreased 25% to 50% and subsequent
adjustments in the dosage based on clinical response
and results of blood glucose measurements. Failure
of glycosuria to recur following a hypoglycemic
episode suggests reversion to a non–insulin-dependent
diabetic state or impaired glucose counterregulation
(see below).
Asymptomatic hypoglycemia is typically identified
during evaluation of a serial blood glucose curve. Serum
fructosamine concentration less than 350 μmol/L
(normal range, 225 to 365 μmol/L) is also suggestive
of hypoglycemia. However, failure to identify hypo-
glycemia during a blood glucose concentration curve or
low-normal serum fructosamine concentration does not
rule out asymptomatic hypoglycemia, in part because
of hypoglycemia-induced glucose counterregulation
(Somogyi phenomenon) and transient insulin resistance
(see page 519). Treatment of asymptomatic hypo-
glycemia involves decreasing the dose of insulin,
typically 10% to 20%, and assessing the clinical response,
change in serum fructosamine concentration, and in
nonstressed dogs, blood glucose concentrations.
If hypoglycemia remains a reoccurring problem
despite reductions in the insulin dose, excessive
overlap in insulin action or reversion to a non–insulin-
dependent diabetic state should be considered. Exces-
sive overlap of insulin action results from twice-a-day
administration of insulin with a duration of effect con-
siderably longer than 12 hours in that diabetic animal
(see page 521). Reversion from an insulin-dependent
to a non–insulin-dependent diabetic state should be
suspected if hypoglycemia remains a persistent problem
despite administration of small doses of insulin once
a day, blood glucose concentrations remain below
200 mg/dl prior to insulin administration, serum fruc-
tosamine concentrations are less than 350 μmol/L, or
urine glucose test strips are consistently negative.
These findings suggest the presence of endogenously
derived circulating insulin, partial but not complete
loss of pancreatic beta cells, and concurrent disease
causing transient insulin resistance. Resolution of an
apparent insulin-dependent diabetic state is uncommon
 518 / CANINE DIABETES MELLITUS
in dogs and usually indicative of secondary diabetes
(see page 490).
VetBooks.ir
IMPAIRED GLUCOSE COUNTERREGULATION. Secretion of
the diabetogenic hormones, most notably epinephrine
and glucagon, stimulates hepatic glucose secretion
and helps counter severe hypoglycemia. A deficient
counterregulatory response to hypoglycemia has been
identified as early as 1 year after diagnosis of IDDM in
humans (White et al, 1983). As a consequence, when
the blood glucose concentration approaches 60 mg/dl,
there is no compensatory response by the body to
increase the blood sugar, and prolonged hypoglycemia
ensues. An impaired counterregulatory response to
hypoglycemia has also been documented in dogs with
IDDM (Duesberg et al, 1995). Dogs with impaired
counterregulation had more problems with hypo-
glycemia than diabetic dogs without impaired counter-
regulation. Impaired counterregulation should be
considered in a diabetic dog exquisitely sensitive to
small doses of insulin or with problems of prolonged
hypoglycemia after administration of an acceptable
dose of insulin.
INAPPETENCE. A healthy, well-regulated diabetic
dog or cat should maintain an excellent appetite.
Occasional inappetence at mealtime is not, by itself, an
indication to stop insulin therapy. Most diabetic dogs
and cats eat within a couple of hours of the insulin
injection, as the blood glucose begins to decline. If the
inappetence persists or if other signs of gastro-
intestinal dysfunction develop (e.g., vomiting), insulin
therapy should be modified or discontinued until the
veterinarian has examined the dog or cat. Common
causes of inappetence in diabetic dogs include pan-
creatitis, ketoacidosis, bacterial infection (especially
involving the urinary system), neoplasia, finicky eaters,
and boredom with high-fiber diets. Appropriate diag-
nostic and therapeutic steps should be initiated,
depending on results of the physical examination.
Recurrence or Persistence of Clinical Signs
Recurrence or persistence of clinical signs (i.e., polyuria,
polydipsia, polyphagia, weight loss) is perhaps the
most common “complication” of insulin therapy in
diabetic dogs. Recurrence or continuing clinical signs
suggest insulin ineffectiveness or resistance. In dia-
betic dogs, insulin ineffectiveness is usually caused by
problems with biologic activity of the insulin or with
owner technique in administering insulin; problems
with insulin therapy relating to the insulin type, dose,
species, or frequency of administration; or problems
with responsiveness to insulin caused by concurrent
inflammatory, infectious, neoplastic, or hormonal dis-
orders (i.e., insulin resistance). Discrepancies in the
parameters used to assess glycemic control, resulting
in an erroneous belief that the diabetic dog is poorly
controlled, should also be considered. This is usually
caused by erroneously high blood glucose concen-
trations that suggest insulin ineffectiveness; these high
values may be stress-induced and may not reflect
the patient’s responsiveness to insulin (see page 567).
When a diabetic dog is evaluated for suspected insulin
ineffectiveness, it is important that all parameters used
to assess glycemic control be critically analyzed, most
notably the owner’s perceptions of how the pet is
doing in the home environment, findings on physical
examination, and changes in body weight (see Tech-
niques for Monitoring Diabetic Control, page 508). If the
history, physical examination, change in body weight,
and serum fructosamine concentration suggest poor
control of the diabetic state, a diagnostic evaluation to
identify the cause is warranted, beginning with evalu-
ation of the owner’s insulin administration technique
and the biologic activity of the insulin preparation.
PROBLEMS WITH OWNER ADMINISTRATION AND ACTIVITY
OF THE INSULIN PREPARATION. Administration of bio-
logically inactive insulin (e.g., outdated, overheated,
mixed by shaking), use of inappropriate insulin syringes
for the concentration of insulin (e.g., U-100 syringe with
U-40 insulin), and problems with insulin administration
technique (e.g., misunderstanding how to read the
insulin syringe, inappropriate injection technique) will
result in recurrence or persistence of clinical signs and
an increase in serum fructosamine and blood glucose
concentrations because of an inadequate dosage of
insulin. Problems with owner administration and
insulin action are identified by evaluating the owner’s
insulin administration technique and by administering
new, undiluted insulin and measuring several blood
glucose concentrations throughout the day. In addition,
the skin and subcutaneous tissues should be assessed
in the area where insulin injections are given. Some
diabetic dogs develop low-grade inflammation, edema,
and thickening of the dermis and subcutaneous tissues
in areas of chronic insulin administration, and these
changes can interfere with insulin absorption following
subcutaneous administration (see Allergic Reactions
to Insulin, page 524).
Freezing and heat inactivates insulin in the bottle.
Although “room temperature” does not inactivate
insulin, we instruct owners to store insulin in the door
of the refrigerator to maintain a consistent environ-
ment for the insulin preparation. It is common practice
for humans with diabetes to replace their insulin with
a new bottle every month to avoid problems with loss
of activity or sterility. Some veterinarians advocate
similar recommendations for owners of diabetic dogs
and cats. We have not appreciated a clinically signifi-
cant loss of insulin action with time when insulin
preparations are maintained in a constant environ-
ment (i.e., refrigerator) and handled appropriately. We
do not routinely recommend purchasing a new bottle
of insulin every month, especially if the diabetic dog
or cat is doing well. However, we do evaluate the
effectiveness of a new bottle of insulin if clinical signs
recur in a diabetic dog or cat, especially if more than
75% of the owner’s insulin preparation has been used.
Dilution of insulin is a common practice, especially
in very small dogs and cats whose insulin requirements
can be quite small. Although studies evaluating the
shelf-life of diluted insulin have not been published,

we recommend replacing diluted insulin preparations
every 4 to 8 weeks. Even using these guidelines, insuffi-
VetBooks.ir
cient amounts of insulin are administered when
diluted insulin is used in some dogs, despite appro-
priate dilution and insulin administration techniques,
inadequacies that are corrected when full-strength
insulin is used. Small-volume insulin syringes (U-100,
0.3 ml) should be used with full strength U-100 insulin
for dogs receiving small amounts of insulin.
PROBLEMS WITH THE INSULIN TREATMENT REGIMEN.
The most common problems causing poor control of
glycemia in this category include insulin underdosage,
the Somogyi phenomenon, short duration of effect
of lente and NPH insulin, and once-a-day insulin
administration. The insulin treatment regimen should
be critically evaluated for possible problems in these
areas, and appropriate changes made to try to improve
insulin effectiveness, especially if the history and
physical examination do not suggest a concurrent
disorder causing insulin resistance.
Insulin Underdosage
Control of glycemia can be established in most dogs
using less than 1.0 U of insulin/kg of body weight
administered twice each day. An inadequate dose of
insulin in conjunction with once-a-day insulin therapy
is a common cause for persistence of clinical signs.
In general, insulin underdosage should be considered
if the insulin dosage is <1.0 U/kg and the animal is
receiving insulin twice a day. If insulin underdosage is
suspected, the dose of insulin should be gradually
increased by 1 to 5 units/injection (depending on the
size of the dog) per week. The effectiveness of the
change in therapy should be evaluated by client per-
ception of clinical response and measurement of serum
fructosamine or serial blood glucose concentrations.
Other causes for insulin ineffectiveness and resistance
should be considered once the insulin dosage exceeds
1.0 to 1.5 units/kg/injection, the insulin is being
administered every 12 hours, and control of glycemia
remains poor. Inducing hypoglycemia and the Somogyi
phenomenon are the primary concerns when increasing
the insulin dosage to rule out insulin underdosage as
the cause of the poor control of glycemia, especially
in toy and miniature breeds (see below). A gradual
increase in the insulin dosage and close monitoring
helps minimize the development of these problems.
Insulin Overdosing and Glucose
Counterregulation (Somogyi Phenomenon)
A high dose of insulin may cause overt hypoglycemia
or induce glucose counterregulation (Somogyi phe-
nomenon) and transient insulin resistance. The
Somogyi phenomenon results from a normal physio-
logic response to impending hypoglycemia induced
by excessive insulin. When the blood glucose concen-
tration declines to less than 65 mg/dl or when the blood
CANINE DIABETES MELLITUS / 519
800
600
Blood glucose (mg/dl)
400
200
100
50
8 AM 12 4 8 PM 12 4 8 AM
Time (hr)
FIGURE 11-17. Blood glucose concentrations in a 6.1-kg Cairn terrier
after receiving beef/pork source NPH insulin at 8 AM. The dog
was fed at 8 AM and 6 PM. Solid line = 20 units; broken line = 4 units.
↑ = insulin injection. (From Feldman EC, Nelson RW: Insulin-induced
hyperglycemia in diabetic dogs. JAVMA 180:1432, 1982.)
glucose concentration decreases rapidly regardless of
the glucose nadir, direct hypoglycemia-induced stimu-
lation of hepatic glycogenolysis and secretion of
diabetogenic hormones, most notably epinephrine and
glucagon, increase the blood glucose concentration,
minimize signs of hypoglycemia, and cause marked
hyperglycemia within 12 hours of glucose counter-
regulation, in part because the diabetic dog cannot
secrete sufficient endogenous insulin to dampen the
continuing rise in the blood glucose concentration
(Fig. 11-17; see Hypoglycemia and the Counter-
regulatory Response, page 617) (Cryer and Polonsky,
1998; Karam, 2001). By the next morning, the blood
glucose concentration can be extremely elevated (400
to 800 mg/dl), and the morning urine glucose concen-
tration is consistently 1 to 2 g/dl as measured with
urine glucose test strips. If the owners of a diabetic
dog are adjusting the daily insulin dose based on the
morning urine glucose concentration, they interpret
these readings as indicating the dog received insuffi-
cient amounts of insulin, especially if hypoglycemic
signs (i.e., weakness, ataxia, bizarre behavior, or con-
vulsions) are not observed. The owners invariably
increase the insulin dose the following morning, and a
continuous cycle of worsening insulin-induced hyper-
glycemia occurs. Unrecognized short duration of
insulin effect combined with insulin dose adjustments
based on morning urine glucose concentrations is
historically the most common cause for the Somogyi
phenomenon and a major reason why we do not
recommend insulin dose adjustments based on
morning urine glucose readings.
Clinical signs of hypoglycemia are typically mild or
not recognized by the owner; clinical signs caused by
hyperglycemia tend to dominate the clinical picture.
The insulin dose that induces the Somogyi phenom-
enon is variable and unpredictable. The Somogyi
 520 / CANINE DIABETES MELLITUS
phenomenon is often suspected in poorly controlled
diabetic dogs whose insulin dosage is approaching
VetBooks.ir
2.2 U/kg body weight/injection but can also occur at
insulin dosages less than 0.5 U/kg/injection. Toy and
miniature breeds of dogs are especially susceptible to
development of the Somogyi phenomenon with lower-
than-expected doses of insulin. The Somogyi phe-
nomenon should always be suspected in any poorly
controlled diabetic dog or cat, regardless of the amount
of insulin being administered.
The diagnosis of the Somogyi phenomenon requires
demonstration of hypoglycemia (<65 mg/dl) followed
by hyperglycemia (>300 mg/dl) following insulin
administration (Feldman and Nelson, 1982). The
Somogyi phenomenon should also be suspected when
the blood glucose concentration decreases rapidly
regardless of the glucose nadir (e.g., a drop from 400 to
100 mg/dl in 2 to 3 hours). If the duration of insulin
effect is greater than 12 hours, hypoglycemia often
occurs at night following the evening dose of insulin
and the serum glucose concentration is typically greater
than 300 mg/dl the next morning (Fig. 11-18). Serum
fructosamine concentrations are unpredictable in dogs
with the Somogyi phenomenon, ranging from 300 to
greater than 600 μmol/l (see Fig. 11-10). Values below
400 μmol/l in diabetic dogs with suspected poor control
of glycemia suggest the Somogyi phenomenon. High
serum fructosamine values identify poor control of
glycemia but are not helpful in identifying the cause.
The Somogyi phenomenon should be included in the list
of causes for high serum fructosamine concentrations.
Unfortunately, the diagnosis of the Somogyi phe-
nomenon can be elusive, in part because of the effects of
the diabetogenic hormones on blood glucose concen-
trations following an episode of glucose counter-
regulation. Secretion of diabetogenic hormones during
the Somogyi phenomenon may induce insulin resistance,
which can last 24 to 72 hours after the hypoglycemic
episode. If a serial blood glucose curve is obtained on the
day glucose counterregulation occurs, hypoglycemia
will be identified and the diagnosis established.
However, if the serial blood glucose curve is obtained on
a day when insulin resistance predominates,
hypoglycemia will not be identified and the insulin dose
may be incorrectly increased in response to the high
blood glucose values (see Fig. 12-21, page 570).
Sometimes the diagnosis is established through response
to treatment, that is, improvement in clinical signs when
the insulin dose is decreased despite the high blood
glucose values identified on the serial blood glucose
curve. See page 617 for more information on insulin
resistance induced by the Somogyi phenomenon.
Therapy involves reducing the insulin dose. The
reduction in insulin dosage depends on the amount
of insulin given at the time the Somogyi phenomenon
is diagnosed. If the diabetic dog is receiving an
“acceptable” dose of insulin (i.e., ≤1.0 U/kg/injection),
the insulin dosage should be decreased 10% to 25%. If
the insulin dose exceeds these amounts, glycemic
regulation should be reinitiated using the insulin
dosage recommended for the initial regulation of the
diabetic dog (i.e., 0.25 U/kg given twice daily).
Blood glucose (mg/dl)
500
400
300
200
100
0
8 AM Noon 4 PM 8 PM
FIGURE 11-18. Blood glucose concentration curves obtained from 3
poorly controlled diabetic dogs treated with recombinant human
lente insulin twice a day, illustrating the typical blood glucose curves
suggestive of the Somogyi phenomenon. In one dog (solid line), the
glucose nadir is less than 80 mg/dl and is followed by a rapid
increase in the blood glucose concentration. In one dog (dashed line),
a rapid decrease in the blood glucose concentration occurs within
2 hours of insulin administration and is followed by a rapid increase
in the blood glucose concentration; the rapid decrease in blood
glucose stimulates glucose counterregulation despite maintaining
the blood glucose nadir above 80 mg/dl. In one dog (dotted line), the
blood glucose curve is not suggestive of the Somogyi phenomenon,
per se. However, the insulin injection causes the blood glucose to
decrease by approximately 300 mg/dl during the day, and the blood
glucose concentration at the time of the evening insulin injection is
considerably lowering than the 8 AM blood glucose concentration.
If a similar decrease in the blood glucose occurs with the evening
insulin injection, hypoglycemia and the Somogyi phenomenon
would occur at night and would explain the high blood glucose
concentration in the morning and the poor control of the diabetic
state. (From Nelson RW, Couto CG: Small Animal Internal Medicine,
3rd ed. St Louis, Mosby Inc, 2003, p 745.)
Evaluation of glycemic control (see page 508) should
be done 1 to 2 weeks after initiating the new dose of
insulin and further adjustments in the insulin dose
made accordingly.
Short Duration of Insulin Effect
For most dogs, the duration of effect of recombinant
human lente and NPH insulin is 10 to 14 hours and
twice-a-day insulin administration is effective in con-
trolling blood glucose concentrations (see Fig. 11-6).
However, in some diabetic dogs, the duration of effect
of lente and NPH insulin is less than 10 hours, a
duration that is too short to prevent periods of hyper-
glycemia and persistence of clinical signs. Diabetic
dogs with the problem of short duration of insulin
effect have persistent morning glycosuria (>1 g/dl on
urine glucose test strips). Owners of these pets usually
mention continuing problems with evening polyuria
and polydipsia or weight loss. If owners are adjusting
the daily insulin dosage based on the morning urine
glucose concentration, they usually induce the
Somogyi phenomenon as the insulin dosage is gradually
 CANINE DIABETES MELLITUS / 521

increased in response to persistent morning glyco- Prolonged Duration of Insulin Effect

suria. Serum fructosamine concentrations are variable
VetBooks.ir

but typically greater than 500 μmol/L. The diagnosis
of short duration of insulin effect is made by demon-
strating recurrence of hyperglycemia (>250 mg/dl)
within 6 to 10 hours of the insulin injection, whereas
the lowest blood glucose concentration is maintained
above 80 mg/dl (see Fig. 11-16). The clinician must
evaluate multiple blood glucose concentrations
obtained throughout the day to establish the diagnosis
(see page 511). One or two afternoon blood glucose
determinations consistently fail to identify the problem.
They may identify normal glucose concentrations or
mild hyperglycemia, findings that are not consistent
with the worries of the owner and do not identify the
underlying problem. Alternatively, one or two after-
noon blood glucose determinations may reveal severe
hyperglycemia, findings that do not differentiate short
duration of insulin effect from the Somogyi phenom-
enon or insulin resistance. Diabetic dogs that have a
short duration of insulin effect can be diagnosed only
by determining serial blood glucose concentrations.
Treatment involves changing to a longer-acting insulin
given twice daily (e.g., switching to ultralente or insulin
glargine) (Fig. 11-19) or increasing the frequency of
insulin administration (i.e., initiating TID therapy) if
the duration of effect is less than 8 hours. PZI insulin
of beef/pork source is not recommended in dogs
because of potential problems with insulin antibodies
(see below). Evaluation of glycemic control (see
page 508) should be done 1 to 2 weeks after initiating
the new dose of insulin and further adjustments in the
insulin dose made accordingly. Refer to Interpretation
of the Serial Blood Glucose Curve, page 514, for more
information on the diagnosis and treatment of short
duration of insulin effect.
In some diabetic dogs, the duration of effect of lente
and NPH insulin is greater than 12 hours, and twice-
a-day insulin administration creates problems with
hypoglycemia and the Somogyi phenomenon. In these
dogs, the glucose nadir following the morning
administration of insulin typically occurs near the
time of the evening insulin administration, and the
morning blood glucose concentration is usually
greater than 300 mg/dl (see Fig. 11-16). Gradually
decreasing blood glucose concentrations measured at
the time of sequential insulin injections is another
indication of prolonged duration of insulin effect. The
effectiveness of insulin in lowering the blood glucose
concentration is variable from day to day, presumably
because of varying concentrations of diabetogenic
hormones whose secretion was induced by prior hypo-
glycemia. Serum fructosamine concentrations are
variable but typically greater than 500 μmol/L. An
effective treatment depends, in part, on the duration of
effect of the insulin. A 24-hour blood glucose curve
should be generated following administration of insulin
once in the morning and feeding the dog at the normal
times of the day. This will allow the clinician to estimate
the duration of effect of the insulin (see Interpreting the
Serial Blood Glucose Curve, page 514). If the duration of
effect is less than 16 hours, a shorter-acting insulin given
twice a day or a lower dose of the same insulin given in
the evening, compared with the morning insulin dose,
can be tried (see Fig. 11-19). If the duration of effect is 16
hours or longer, switching to a longer-acting insulin
administered once a day or administering NPH or lente
insulin in the morning and regular insulin at bedtime
(i.e., 16 to 18 hours after the morning insulin injection)
can be tried. When using different types of insulin in the

Less potent Insulin glargine Longest duration

of effect

Ultralente

Protamine zinc

Lente

NPH

70% NPH,
30% regular crystalline

50% NPH,
50% regular crystalline

Regular crystalline
Shortest duration
More potent Insulin lyspro, aspart of effect

FIGURE 11-19. Categorization of types of commercial insulin based on the potency and duration of
effect. Notice the inverse relationship between the potency and duration of effect. (From Nelson RW,
Couto CG: Small Animal Internal Medicine, 3rd ed. St Louis, Mosby Inc, 2003, p 747.)
 522 / CANINE DIABETES MELLITUS

same 24-hour period, the goal is to have the combined dogs treated with NPH or lente insulin but is a
duration of effect of the insulins equal 24 hours. For potential concern with ultralente and insulin glargine.
VetBooks.ir

example, if the duration of effect of lente insulin is See page 571 for information on inadequate absorption
approximately 16 hours and regular insulin is 6 to 8 of longer-acting insulins.
hours, the combination of the two insulins is
approximately 22 to 24 hours. Differences in potency
of intermediate- and long-acting insulins versus Circulating Anti-Insulin Antibodies
regular insulin often necessitates use of different
dosages for the morning and evening insulin injection; Although all species of commercial insulin are
regular insulin is more potent, and less of it is required effective in diabetic dogs, insulin immunogenicity and
to get the same glycemic effect, compared with lente, production of insulin antibodies can enhance and
NPH, and ultralente insulin. prolong the pharmacodynamic action of insulin by
serving as a carrier, or they can reduce insulin action
by neutralization (Bolli et al, 1984; Marshall et al, 1988;
Inadequate Insulin Absorption Lahtela et al, 1997). Antibodies may also have no
apparent clinical effect on insulin dosage or status of
Slow or inadequate absorption of subcutaneously glycemic control (Lindholm et al, 2002). In our expe-
deposited insulin is an uncommon problem in diabetic rience, anti-insulin antibody production in diabetic dogs

500 1000

400 800
Serum insulin concentration (␮U/ml)
Blood glucose concentration (mg/dl)

300 600

200 400

100 200

0 0
8 AM Noon 4 PM 8 PM 8 AM Noon 4 PM 8 PM

FIGURE 11-20. Blood glucose and serum insulin concentrations in a 7.6-kg, spayed dog receiving 1.1
U/kg beef/pork source lente insulin (solid line) SC. The dog had severe polyuria, polydipsia, and
weight loss. A baseline serum insulin concentration was greater than 1000 μU/ml 48 hours after
discontinuing insulin therapy. Interference from anti-insulin antibodies was suspected, and the source
of insulin was changed to recombinant human insulin. Clinical signs improved within 2 weeks and a
blood glucose curve obtained 4 weeks later, with the dog receiving 0.9 U/kg recombinant human lente
insulin (broken line), showed excellent glycemic control. Presumably, loss of anti-insulin antibody
interference with the insulin radioimmunoassay allowed a more accurate assessment of changes in
the serum insulin concentration after recombinant human insulin administration. ↑ = insulin injection
and food.

can have a deleterious impact on insulin effectiveness,
hamper the ability to maintain control of glycemia, and
VetBooks.ir
in extreme cases, cause severe insulin resistance (Fig. 11-
20). Insulin antibodies can also cause erratic fluctuations
in the blood glucose concentration, with no correlation
between the timing of insulin administration and
changes in blood glucose concentration (Fig. 11-21).
Presumably, fluctuations in blood glucose concentration
result from erratic and unpredictable changes in the
circulating free (i.e., non–antibody-bound) insulin
concentration (Bolli et al, 1984). This phenomenon
causes inappropriate and potentially life-threatening
hypoglycemia at unexpected times in humans with
diabetes. We have observed a similar syndrome in
diabetic dogs treated with beef/pork insulin.
Insulin antibodies result from repeated injections
of a foreign protein (i.e., insulin). The structure and
amino acid sequence of the injected insulin relative to
400
300
Blood glucose concentration (mg/dl)
200
100
0
8 AM Noon 4 PM 8 PM
FIGURE 11-21. Blood glucose curve in a 50-kg male dog receiving
0.7 U/kg beef/pork source lente insulin (solid line) SC. Note the
erratic fluctuations in the blood glucose concentration. The dog had
polyuria, polydipsia, and weight loss and was blind from cataract
formation. A baseline serum insulin concentration was 825 μU/ml
24 hours after discontinuing insulin therapy. Interference from anti-
insulin antibodies was suspected, and the source of insulin was
changed to recombinant human insulin. Clinical signs improved
within 1 month and a blood glucose curve obtained 8 weeks later,
with the dog receiving 0.5 U/kg recombinant human lente insulin
(broken line), showed excellent glycemic control and loss of erratic
fluctuations in the blood glucose concentration.
CANINE DIABETES MELLITUS / 523
the native endogenous insulin influences the develop-
ment of anti-insulin antibodies. Conformational insulin
epitopes are believed to be more important in the
development of anti-insulin antibodies than differences
in the linear subunits of the insulin molecule, per se
(Thomas et al, 1985; Thomas et al, 1988; Nell and
Thomas, 1989; Davison et al, 2001). Nevertheless, the
more divergent the insulin molecule being administered
from the species being treated, the greater the likeli-
hood that significant or worrisome amounts of insulin
antibodies will be formed. The amino acid sequence of
canine, porcine, and recombinant human insulin are
similar, suggesting that recombinant human insulin
should not be strongly antigenic when administered to
diabetic dogs (Feldman et al, 1983; Ganong, 1991).
Studies using an ELISA to detect insulin antibodies in
serum of insulin-treated diabetic dogs identified serum
insulin antibodies in approximately 5% of dogs treated
with recombinant human insulin (Harb-Hauser et al,
1998). In contrast, the amino acid sequences of canine
and beef insulin differ, and serum insulin antibodies
have been identified in approximately 40% and 50% of
dogs treated with beef/pork-source insulin (which is
90% beef insulin) (Haines, 1986; Harb-Hauser et al, 1998)
and 65% of dogs treated with bovine insulin (Davison et
al, 2003). In our experience, the presence of serum
insulin antibodies in these dogs was associated with
erratic and often poor control of glycemia, an inability to
maintain control of glycemia for extended periods of
time, frequent adjustments in insulin dose and
occasional development of severe insulin resistance (i.e.,
blood glucose concentrations consistently >400 mg/dl).
Dogs treated with recombinant human insulin
consistently had more stable control of glycemia for
extended periods of time. In fact, ELISA titers reverted to
negative and glycemic control improved by 4 to 6 weeks
after changing from beef/pork to human recombinant
insulin in insulin antibody–positive dogs. Recombinant
human-source insulin should be used in diabetic dogs to
avoid problems with insulin effectiveness presumably
caused by circulating insulin antibodies.
Although uncommon, insulin antibody formation
should be suspected in poorly controlled diabetic dogs
treated with recombinant human insulin in which
another cause for the poor glycemic control cannot be
identified. Documentation of serum insulin–binding
antibodies should use assays that have been validated
in diabetic dogs, assays that are currently not readily
Mid 4 AM 8 AM available commercially. However, circulating insulin
antibodies may interfere with some RIA techniques
used to measure serum insulin concentration in a
manner similar to the effects of thyroid hormone
antibodies on RIA techniques for serum T3 and T4
concentrations (see Fig. 3-14, page 112). This inter-
erence can be used to raise the clinician’s index of
suspicion for insulin-binding antibodies as a cause for
insulin resistance. A single-phase separation system
using antibody-coated tubes is used in our laboratory
to measure serum insulin concentration. The presence
of insulin antibodies in the serum sample causes
spuriously high insulin values using this RIA. The
 524 / CANINE DIABETES MELLITUS
serum insulin concentration is typically less than
50 μU/ml 24 hours after insulin administration in
VetBooks.ir
diabetic dogs and cats without antibodies causing
interference with the RIA. In contrast, serum insulin
concentrations are typically greater than 400 μU/ml
and may be greater than 1000 μU/ml 24 hours after
insulin administration when insulin antibodies interfere
with the RIA results; insulin values are erroneously
increased because of the antibodies (see Fig. 11-20). A
switch to purified porcine insulin, a switch to a purer
form of insulin (i.e., regular crystalline insulin), or both,
should be considered if insulin antibodies are identified
in a diabetic dog being treated with recombinant human
insulin. Because regular crystalline insulin has a
duration of effect of approximately 6 to 8 hours when
injected subcutaneously, it should be administered every
8 hours when used to treat diabetic dogs or cats at home.
Allergic Reactions to Insulin
Significant allergic reactions to insulin occur in up to
5% of human diabetics treated with insulin and
include erythema, pruritus, induration at the injection
site, and uncommonly, systemic manifestations charac-
terized by urticaria, angioneurotic edema, or frank
anaphylaxis (Unger and Foster, 1998). Atrophy or
hypertrophy of subcutaneous tissue (i.e., lipoatrophy
and lipodystrophy) may also occur at the insulin
injection site. Many humans with insulin allergy have
histories of sensitivity to other drugs as well. Allergic
reactions to insulin have been poorly documented in
diabetic dogs and cats. Pain on injection of insulin is
usually caused by inappropriate injection technique or
site of injection and not an adverse reaction to insulin,
per se. Chronic injection of insulin in the same area of
the body may cause inflammation and thickening of
the skin and subcutaneous tissues and may be caused by
an immune reaction to insulin or some other protein (e.g.,
protamine) in the insulin bottle. Inflammation and
thickening of the skin and subcutaneous tissues may
impair insulin absorption, resulting in recurrence of
clinical signs of diabetes. Rotation of the injection site will
help prevent this problem. Rarely, diabetic dogs and cats
will develop focal subcutaneous edema and swelling at
the site of insulin injection. Insulin allergy is suspected in
these animals. Treatment includes switching to a more
homologous insulin (purified porcine insulin for dogs,
purified beef insulin for cats) and to a more purified
insulin preparation (e.g., regular crystalline insulin) in
the hopes of minimizing a potential immune reaction to
the species of insulin or some contaminant in the insulin
preparation. We have not yet confirmed systemic
allergic reactions to insulin in dogs or cats.
Concurrent Disorders Causing Insulin Resistance
Insulin resistance is a condition in which a normal
amount of insulin produces a subnormal biologic
response. Insulin resistance may result from problems
occurring before the interaction of insulin with its
receptor (i.e., prereceptor), at the receptor, or at steps
distal to the interaction of insulin and its receptor (i.e.,
postreceptor) (Ihle and Nelson, 1991). Prereceptor
problems reduce free metabolically active insulin con-
centration and include increased insulin degradation
and insulin-binding antibodies. Receptor problems
include alterations in insulin-receptor binding affinity
and concentration and insulin-receptor antibodies.
Postreceptor problems are difficult to differentiate clini-
cally from receptor problems, and both often coexist. In
dogs and cats, receptor and postreceptor abnormalities
are usually attributable to obesity or a disorder causing
excessive secretion of an insulin-antagonistic hormone,
such as cortisol, glucagon, epinephrine, growth
hormone, progesterone, or thyroid hormone.
No insulin dose clearly defines insulin resistance.
When insulin effectiveness in a diabetic dog is being
assessed, the insulin dose relative to body weight and
adequacy of glycemic control should be evaluated
simultaneously. For most diabetic dogs, good glycemic
control (i.e., blood glucose concentration between 100
and 250 mg/dl) can be achieved with less than 1.0 U of
intermediate or long-acting insulin per kilogram of
body weight given twice daily. Insulin resistance should
be suspected if control of glycemia is poor despite an
insulin dosage in excess of 1.5 U/kg, when excessive
amounts of insulin (i.e., insulin dosage >1.5 U/kg) are
necessary to maintain the blood glucose concentration
below 300 mg/dl (Fig. 11-22), and when control of
glycemia is erratic and insulin requirements are
constantly changing every few weeks in an attempt
to maintain control of glycemia (Fig. 11-23). Failure
of the blood glucose concentration to decrease below
300 mg/dl during a serial blood glucose curve is
suggestive of, but not definitive for, the presence of
insulin resistance. An insulin resistance–type blood
glucose curve can also result from stress-induced
hyperglycemia, the Somogyi phenomenon, and other
problems with insulin therapy (Table 11-16), and a
TABLE 11-16 RECOGNIZED CAUSES OF INSULIN
INEFFECTIVENESS OR INSULIN RESISTANCE IN
DIABETIC DOGS AND CATS
Caused by Insulin Therapy Caused by Concurrent Disorder
Inactive insulin Diabetogenic drugs
Diluted insulin Hyperadrenocorticism
Improper administration Diestrus (bitch)
technique Acromegaly (cat)
Inadequate dose Infection, especially of oral cavity
Somogyi effect and urinary tract
Inadequate frequency of Hypothyroidism (dog)
insulin administration Hyperthyroidism (cat)
Impaired insulin absorption, Renal insufficiency
especially ultralente insulin Liver insufficiency
Anti-insulin antibody excess Cardiac insufficiency
Glucagonoma (dog)
Pheochromocytoma
Chronic inflammation, especially
pancreatitis
Pancreatic exocrine insufficiency
Severe obesity
Hyperlipidemia
Neoplasia
 CANINE DIABETES MELLITUS / 525

500
VetBooks.ir

400

Blood glucose concentration (mg/dl)

300

200

100

0
8 AM Noon 4 PM 8 PM

FIGURE 11-22. Blood glucose curve in a 12-kg male diabetic dog with untreated hypothyroidism
receiving 2.2 U/kg recombinant human lente insulin (solid line). The large amount of insulin required
to lower the blood glucose concentration suggests insulin resistance. Glycemic control was improved
and the insulin dosage decreased to 0.9 U/kg after sodium levothyroxine therapy was initiated (broken
line). ↑ = SC insulin injection and food.

Blood glucose (mg/dl)

400

300

200

100

8 AM Noon 4 PM 8 PM

FIGURE 11-23. Blood glucose curves in a 9-year-old diabetic cat with recurring clinical signs of chronic
pancreatitis. The cat developed lethargy, inappetence, polyuria, polydipsia, and weight loss, and blood
glucose concentrations were increased (solid lines) when pancreatitis was symptomatic. Clinical signs
were absent and blood glucose concentrations were in an acceptable range (broken lines) when
pancreatitis was quiescent. Blood glucose curves were obtained when the cat was receiving 4 units
ultralente insulin (solid and broken lines - no circles) SC twice a day and 3 units lente insulin (solid and
broken lines - solid circles) SC twice a day. Note the negative impact of pancreatitis on glycemic control.
↑ = SC insulin injection and food.
 526 / CANINE DIABETES MELLITUS
decrease in the blood glucose concentration below 300
mg/dl can occur with disorders causing relatively
VetBooks.ir
mild insulin resistance. Serum fructosamine concen-
trations are typically greater than 500 μmol/L in
animals with insulin resistance and can exceed 700
μmol/L if insulin resistance is persistent and severe.
Unfortunately, an increased serum fructosamine
concentration is merely indicative of poor glycemic
control not insulin resistance, per se.
The severity of insulin resistance is dependent, in
part, on the underlying etiology. Insulin resistance may
be mild and easily overcome by increasing the dosage
of insulin or may be severe, causing marked hyper-
glycemia regardless of the type and dosage of insulin
administered. Some causes of insulin resistance are
readily apparent at the time diabetes is diagnosed, such
as obesity and the administration of insulin-antagonistic
drugs (e.g., glucocorticoids, megestrol acetate). Other
causes of insulin resistance are not readily apparent
and require an extensive diagnostic evaluation to be
identified. In general, any concurrent inflammatory,
infectious, hormonal, or neoplastic disorder can cause
insulin resistance and interfere with the effectiveness
of insulin therapy (see Table 11-16). In our experience,
the most common concurrent disorders interfering with
insulin effectiveness in dogs include diabetogenic drugs
(glucocorticoids), severe obesity, hyperadrenocorticism,
diestrus, chronic pancreatitis, renal insufficiency, oral,
skin, and urinary tract infections, hyperlipidemia, and
anti-insulin antibodies in dogs receiving beef insulin.
Obtaining a complete history and performing a
thorough physical examination is the most important
step in identifying these concurrent disorders. Is the
dog severely obese? Is the dog taking any medications
that may have insulin-antagonistic properties? Has the
female diabetic dog been spayed, and if not, when was
the last estrus observed? Has the owner noticed any
clinical signs that may suggest concurrent infection
(e.g., hematuria, vaginal discharge)? Abnormalities
identified on a thorough physical examination may
suggest a concurrent insulin-antagonistic disorder or
infectious process, which will give the clinician direc-
tion in the diagnostic evaluation of the animal. If the
history and physical examination are unremarkable, a
CBC, serum biochemical analysis, serum progesterone
concentration (intact female dog), abdominal ultra-
sound, and urinalysis with bacterial culture should
be obtained to further screen for concurrent illness.
Additional tests will be dependent on the results of the
initial screening tests (Table 11-17).
OBESITY. Obesity causes a reversible insulin resis-
tance secondary to obesity-induced downregulation of
insulin receptors, reduced insulin-receptor binding
affinity, and intracellular, postreceptor defects in
glucose metabolism (Truglia et al, 1985). Weight
reduction improves tissue responsiveness to insulin,
presumably via improvement in obesity-induced
insulin resistance. Whenever possible, correction of
obesity through dietary modifications (see page 501)
and exercise should be attempted. Glycemic control
may improve if persistent hyperglycemia is caused by
TABLE 11-17 DIAGNOSTIC TESTS TO CONSIDER
FOR THE EVALUATION OF INSULIN RESISTANCE
IN DIABETIC DOGS AND CATS
CBC, serum biochemistry panel, urinalysis
Bacterial culture of the urine
Serum lipase and amylase activities (pancreatitis)
Serum trypsin-like immunoreactivity (exocrine pancreatic
insufficiency, pancreatitis)
Adrenocortical function tests
ACTH stimulation test (spontaneous or iatrogenic
hyperadrenocorticism)
Low-dose dexamethasone suppression test (spontaneous
hyperadrenocorticism)
Thyroid function tests
Baseline serum total and free thyroxine (hypothyroidism or
hyperthyroidism)
Endogenous TSH (hypothyroidism)
Thyroid-stimulating hormone stimulation test (hypothyroidism)
Thyroid-releasing hormone stimulation test (hypothyroidism or
hyperthyroidism)
Triiodothyronine suppression test (hyperthyroidism)
Serum progesterone concentration (diestrus in intact female dog)
Plasma growth hormone or serum insulin-like growth factor I
concentration (acromegaly)
Serum insulin concentration 24 hours after discontinuation of
insulin therapy (insulin antibodies)
Serum triglyceride concentration (hyperlipidemia)
Abdominal ultrasonography (adrenomegaly, adrenal mass,
pancreatitis, pancreatic mass)
Thoracic radiography (cardiomegaly, neoplasia)
Computed tomography or magnetic resonance imaging (pituitary
mass)
a combination of obesity, inappropriate diet, and
problems with insulin therapy. However, improvement
in blood glucose concentrations is unlikely if insulin
resistance is caused by concurrent inflammatory,
infectious, neoplastic, or hormonal disorders.
PROGESTOGENS. Insulin resistance induced by
diestrus should always be suspected in any newly
diagnosed or poorly controlled intact diabetic female
dog, regardless of the history concerning estrus activity.
The diestrual increase in blood progesterone concen-
tration has a direct insulin-antagonistic effect and also
stimulates secretion of growth hormone. Progesterone
reduces insulin binding and glucose transport in
tissues (Ryan and Enns, 1988). The insulin-antagonistic
effects of growth hormone result from a decrease in
the number of insulin receptors and inhibition of glu-
cose transport, possibly through effects on the expres-
sion of glucose-transporter genes (Moller and Flier,
1991). Administration of progestogens (e.g., medroxy-
progesterone acetate) in the dog also stimulates growth
hormone secretion (Eigenmann and Rijnberk, 1981;
Eigenmann et al, 1983).
Diestrus-induced insulin resistance quickly becomes
severe as plasma growth hormone concentrations
increase in response to progesterone secretion from the
corpora lutea. Insulin quickly becomes ineffective in
lowering the blood glucose concentration, despite the
administration of insulin dosages well in excess of
2.2 U/kg/injection. Life-threatening diabetic keto-
acidosis can develop. Rapid identification of increased
serum progesterone concentration followed by ovario-
 CANINE DIABETES MELLITUS / 527

TABLE 11-18 FREQUENCY OF OCCURRENCE panel, and urinalysis. Hyperglycemia and glycosuria
OF CAUSES OF INSULIN RESISTANCE IN 65 are readily apparent in the results. Physical findings
VetBooks.ir

DIABETIC DOGS AND 54 DIABETIC CATS suggestive of hyperadrenocorticism, such as endocrine

RANDOMLY EVALUATED FROM UC DAVIS alopecia, thin skin, and abdominal enlargement, are
RECORDS BETWEEN 1990 AND 1995 usually not apparent at the time diabetes is diagnosed
but often become apparent weeks later as the severity
Cause Diabetic Dogs Diabetic Cats
of hyperadrenocorticism progresses. In addition,
Hyperadrenocorticism 38% 17% hepatomegaly, hypercholesterolemia, and hepatic
Acromegaly – 19%
Bacterial infection 16% 9%
enzyme alterations are quickly ascribed to the diabetic
Diestrus 8% – state. Early indicators for possible hyperadreno-
Renal, hepatic, cardiac 6% 15% corticism in diabetic dogs include a serum alkaline
insufficiency phosphatase activity greater than 500 IU/L, urine
Hypothyroidism 9% – specific gravity less than 1.020 (especially in the presence
Hyperthyroidism – 9%
Diabetogenic drugs 3% 9% of glycosuria), and adrenomegaly on abdominal ultra-
Chronic pancreatitis 5% 6% sound. Glycemic control in diabetic dogs with untreated
Exocrine pancreatic insufficiency 5% 4% hyperadrenocorticism is variable and dependent, in
Anti-insulin antibodies 5% 2% part, on the severity and duration of hypercortisolemia.
Glucagonoma 4% –
Pheochromocytoma 2% –
Diabetic dogs presumably in the early stages of hyper-
Miscellaneous neoplasia – 9% adrenocorticism may be responsive to insulin and
initially well-controlled with typical doses of insulin.
However, during the ensuing months, control of
glycemia becomes progressively more difficult, and
hysterectomy is strongly recommended, especially if eventually severe insulin resistance develops. Hyper-
diabetic ketoacidosis is present. For most commercial adrenocorticism should always be considered in any
endocrine laboratories, a serum progesterone concen- poorly controlled diabetic dog in which problems with
tration greater than 2 ng/ml is consistent with diestrus. the insulin treatment regimen are not readily apparent,
The stimulus for growth hormone secretion declines regardless of the findings on physical examination and
once the blood progesterone concentration returns to routine diagnostic tests.
anestrus levels, following either spontaneous regression When interpreting results of tests of the pituitary-
of the corpora lutea, ovariohysterectomy, or with- adrenocortical axis (e.g., ACTH stimulation test, low-
drawal of progestogen therapy. Improvement in insulin dose dexamethasone suppression test, urine cortisol:
effectiveness typically occurs within 7 days after ovario- creatinine ratio), the influence of poorly regulated dia-
hysterectomy or withdrawal of progestogens. Carbo- betes mellitus on adrenocortical function and results of
hydrate intolerance may resolve in bitches with these diagnostic tests should be considered. Although
adequate numbers of functional beta cells once growth one study found normal adrenal function test results
hormone returns to normal baseline concentration (see in 15 well-regulated diabetic dogs (Zerbe et al, 1988),
Secondary Diabetes in Dogs, page 490). many of our poorly regulated diabetic dogs have had
HYPERADRENOCORTICISM. Hyperadrenocorticism, one or more adrenal function test results suggestive of
whether naturally occurring or due to exogenous hyperadrenocorticism despite normal pituitary and
glucocorticoid administration, is the most common adrenal glands at necropsy. A mildly exaggerated
cause of severe insulin resistance in diabetic dogs response to ACTH administration (i.e., post-ACTH
(Table 11-18). Glucocorticoids antagonize the actions cortisol, 18 to 24 μg/dl), an increased urine cortisol:
of insulin in both hepatic and peripheral cells. Gluco- creatinine ratio, and less commonly, failure of plasma
corticoids may induce insulin resistance directly by cortisol to suppress to less than 1.5 μg/dl during low-
reducing the number or efficacy of glucose transporters dose dexamethasone suppression testing can occur in
and indirectly by increasing the circulating levels of poorly regulated (and occasionally well-regulated)
glucagon and free fatty acids (Moller and Flier, 1991). diabetic dogs. Presumably, the chronic “stress” related
The decreased cellular use of glucose induced by to poorly regulated diabetes mellitus and its associated
glucocorticoids does not involve changes at the level problems is responsible for these results. Because of
of the insulin receptor. the potential interplay between the stress of poorly
Naturally occurring hyperadrenocorticism and regulated diabetes mellitus and results of tests of the
diabetes mellitus are common concurrent diseases in pituitary-adrenocortical axis, the diagnosis of hyper-
the dog. The question is often raised: which disorder adrenocorticism should always be based on a combi-
came first? Although the answer to this question is not nation of history, findings on physical examination,
known, we suspect that hyperadrenocorticism develops clinical pathologic abnormalities, results of tests of the
initially and subclinical diabetes mellitus becomes pituitary-adrenocortical axis, ultrasound evaluation of
clinically apparent as a result of the insulin resistance adrenal size, and clinical suspicion for the disorder.
caused by the hyperadrenal state. Invariably diabetes The decision to treat should be based on the etiology
mellitus is diagnosed first, in part because the initial (pituitary- versus adrenal-dependent disease), severity
diagnostic evaluation for polyuria and polydipsia of clinical signs and insulin resistance, and status of
includes evaluation of a CBC, serum biochemistry glycemic control. Treatment may not be indicated at
 528 / CANINE DIABETES MELLITUS
the time pituitary-dependent hyperadrenocorticism is
diagnosed, especially if clinical signs are absent, insulin
VetBooks.ir
resistance is not apparent, and control of glycemia is
good. The reader is referred to Chapter 6 for a complete
discussion on the diagnosis and treatment of hyper-
adrenocorticism in dogs.
BACTERIAL INFECTIONS. The detrimental impact of
diabetes on the risk for infection-related morbidity
and mortality is well recognized in humans (Bertoni
et al, 2001). Infections are also commonly identified in
diabetic dogs and cats, suggesting that diabetes causes
a similar increased susceptibility of infection in these
species (Hess et al, 2000b; Peikes et al, 2001). Proposed
mechanisms for the increased susceptibility to infections
in diabetics include decreased blood supply secondary
to the associated microangiopathy and atherosclerosis,
resulting in decreased delivery of oxygen, phagocytes,
and antibodies to the site of infection; impaired
humoral immunity, resulting in decreased antibody
production; abnormal chemotaxis of neutrophils;
defects in the phagocytosis and intracellular killing of
organisms; and impaired cell-mediated immunity
(Possilli and Leslie, 1994; McMahon and Abistrian,
1995; Joshi et al, 1999). Many of these host defense
abnormalities improve, at least partially, with better
glycemic control.
Sustained hyperglucagonemia and insulin resistance
have been documented in diabetic humans with bac-
terial infections (Nelson, 1989). The anti-insulin effects
of glucagon are confined largely to the liver (i.e., stimu-
lation of hepatic gluconeogenesis, glycogenolysis, and
ketogenesis), although impaired glucose transport has
also been described in peripheral tissues (Moller and
Flier, 1991). The clinical significance of glucagon in
insulin resistance in the diabetic dog and cat remains
to be clarified. Clinically, correction of concurrent
severe bacterial infection (most notably affecting the
urinary tract, skin, and oral cavity) has improved
glycemic control.
Bacterial infections are usually identified during
the physical examination or following evaluation of
the CBC, urinalysis, urine culture, and rarely, blood
cultures. Because of the relatively high incidence of
bacterial infections in diabetic dogs and cats, evaluation
of the effect of trial antibiotic therapy on improving
glycemic control should be considered when a cause
for insulin resistance is not identified. Under these
circumstances, a broad-spectrum, bactericidal antibiotic
(e.g., amoxicillin-clavulanate) should be administered
for 10 to 14 days and glycemic control reevaluated
before antibiotic therapy is discontinued.
RENAL INSUFFICIENCY. Renal insufficiency and dia-
betes mellitus are common geriatric diseases and often
occur concurrently, especially in older cats. Abnormal
renal function may result from the deleterious effects
of the diabetic state (i.e., diabetic nephropathy; see
page 532) or may be an independent problem that has
developed in conjunction with diabetes in the geriatric
pet. As renal function declines, human diabetics with
concurrent nephropathy are at increased risk for severe
hypoglycemia as a result of decreased renal clearance
of insulin and decreased renal glucose production by
gluconeogenesis (see page 617) (Stumvoll et al, 1997;
Rave et al, 2001). Tissue responsiveness to insulin (i.e.,
insulin sensitivity) is also attenuated, resulting in
poorer metabolic control of the diabetic state (Eidemak
et al, 1995). Prolonged duration of insulin effect, insulin
resistance, and less commonly, hypoglycemia are
recognized problems in diabetic dogs and cats with
concurrent renal insufficiency. The interplay between
progression and severity of renal insufficiency, severity
of insulin resistance, and impairment of insulin
clearance creates unpredictable fluctuations in control
of glycemia and insulin requirements and frustration
for the owner and veterinarian. In addition, reliance
on an important indicator of diabetic control (i.e.,
severity of polyuria and polydipsia) is no longer
reliable because of the concurrent renal insufficiency.
In most cases, treatment for renal insufficiency and
failure takes priority and insulin therapy is modified,
as needed, to attain the best possible control of the
diabetic state while trying to avoid hypoglycemia,
recognizing that attainment of good control will be
difficult and polyuria and polydipsia will persist
regardless of the status of glycemic control.
CHRONIC PANCREATITIS. Chronic inflammatory dis-
orders can deleteriously affect glycemic control. The
most significant inflammatory disorder in the diabetic
dog and cat is pancreatitis. Chronic pancreatitis is
identified at necropsy in approximately 35% of diabetic
dogs and 50% of diabetic cats (Alejandro et al, 1988;
Goosens et al, 1995). Most of these animals have a
similar history, characterized by poorly controlled dia-
betes, fluctuating insulin requirements, blood glucose
concentrations often greater than 300 mg/dl, intermit-
tent lethargy and inappetence, and owner concerns
that their pet is “just not doing well” (see Fig. 11-23).
An inability to correct these problems ultimately leads
to euthanasia for many dogs and cats. Documentation
of chronic pancreatitis can be difficult and must rely
on a combination of clinical signs, physical examination
findings, serum lipase concentration, serum trypsin–
like immunoreactivity, ultrasound evaluation of the
pancreas, and clinical suspicion for the disorder (see
Pancreatic Enzymes, page 495). A successful response
to therapy, including intravenous fluids, dietary modifi-
cations, and anti-inflammatory doses of prednisone, can
be difficult to attain, and many of these dogs and cats
continue to do poorly
EXOCRINE PANCREATIC INSUFFICIENCY. Exocrine
pancreatic insufficiency (EPI) is an uncommon compli-
cation of diabetes mellitus, presumably developing as
a sequela of chronic pancreatitis. In our experience, most
diabetic dogs with EPI are difficult to regulate with
insulin; have persistent weight loss despite a ravenous
appetite, ultimately becoming emaciated; and defecate
increased amounts of soft stools, not the voluminous,
rancid stools considered classic for EPI. Mild diffuse
thickening of the small intestine may be evident
during abdominal palpation. Diagnosis of EPI in the
diabetic dog requires evaluation of serum trypsin–like
immunoreactivity (TLI), which is low (i.e., <2.5 μg/L)

(Widberg et al, 1999; Widberg and Westermarck, 2002).
Treatment with pancreatic enzyme replacement therapy,
VetBooks.ir
metronidazole, and if necessary, a highly digestible,
low-fat diet has resulted in improvement in clinical
signs and glycemic control, often with a reduction in
daily insulin requirements.
HYPOTHYROIDISM. Insulin resistance has been
documented in diabetic dogs with hypothyroidism,
which resolved following correction of the thyroid
hormone deficiency (Ford et al, 1993). The mechanisms
of carbohydrate intolerance in states of thyroid
hormone deficiency are controversial. Most studies
agree that a postreceptor defect in insulin-mediated
glucose transport and metabolism exists (Arner et al,
1984; Pedersen et al, 1988). The effect of thyroid hor-
mone deficiency on insulin secretion and on insulin
receptors is less clear. Depending on the study,
insulin secretion may be increased or decreased and
insulin receptor concentration and binding affinity
may be increased, decreased, or unaffected (Czech et
al, 1980; Arner et al, 1984; Pedersen et al, 1988). Other
factors associated with hypothyroidism and diabetes
mellitus may also contribute to insulin resistance,
most notably obesity and hyperlipidemia (see below).
Interpretation of serum thyroxine (T4) concentrations
must take into consideration the status of glycemic
control of the dog, the severity of concurrent illness,
and the effects these factors may have on thyroid
gland function (i.e., euthyroid sick syndrome; see
Serum Thyroxine Concentration, page 112). A low
serum thyroxine concentration does not, by itself,
confirm hypothyroidism in the dog. If hypothyroidism
is strongly suspected after evaluation of the history,
findings on physical examination, and results of
routine blood work and serum T4 concentration, we
evaluate serum free T4 and thyrotropin (TSH) concen-
trations before initiating levothyroxine sodium treat-
ment. Refer to Chapter 3 for a more detailed discussion
of the effects of concurrent illness and drug therapy on
serum thyroid hormone concentrations and the tests
used to diagnose hypothyroidism in dogs.
If levothyroxine sodium is administered, a concur-
rent reduction in insulin dosage should be considered,
and glycemic control monitored weekly during the
first month of administration (see Fig. 11-22). Further
reductions in insulin dosage may be required if hypo-
glycemia develops as the hypothyroid-induced insulin
resistance resolves. In one report on diabetic, hypo-
thyroid dogs, insulin requirements decreased 50% to
60% within 2 weeks after initiating levothyroxine
sodium therapy (Ford et al, 1993).
PHEOCHROMOCYTOMA, GLUCAGONOMA, AND OTHER
NEOPLASIA. Neoplasia not causing hyperadreno-
corticism was believed to contribute to insulin resis-
tance in 5% to 10% of our diabetic dogs (see Table 11-18).
Insulin resistance in diabetic dogs with glucagonoma
and pheochromocytoma could be explained by excess
secretion of the diabetogenic hormones glucagon and
catecholamines, respectively. The effects of glucagon
on glycemic control and insulin action are discussed in
the bacterial infection section above and in Chapter 15.
CANINE DIABETES MELLITUS / 529
Catecholamines have both direct and indirect effects to
stimulate hepatic glycogenolysis and hepatic and renal
gluconeogenesis; provide muscle tissue with an alter-
native source of fuel by mobilizing muscle glycogen
and stimulating lipolysis; mobilize gluconeogenic pre-
cursors (e.g., lactate, alanine, and glycerol); and inhibit
glucose use by insulin-sensitive tissues such as skeletal
muscle (Moller and Flier, 1991; Cryer, 1993; Karam,
2001).
The carbohydrate intolerance that develops in
humans with pheochromocytoma is usually mild and
does not require therapy, in part because excessive
catecholamine secretion is usually episodic, not sus-
tained. Although pheochromocytoma may interfere
with the effectiveness of insulin in diabetic dogs,
development of severe insulin resistance is uncommon,
presumably for the same reasons as in humans, that is,
excess catecholamine secretion is usually episodic, not
sustained. Identification of a pancreatic or adrenal
mass with abdominal ultrasonography should raise
suspicion for glucagonoma and pheochromocytoma,
respectively, especially if other tests do not support
hyperadrenocorticism. Additional diagnostic tests for
pheochromocytoma and glucagonoma are discussed
in Chapters 9 and 15, respectively.
Lymphoma and mast cell tumor are the most
common nonendocrine tumors identified in diabetic
dogs and cats with insulin resistance. The mechanisms
involved with poor glycemic control of diabetics with
nonendocrine tumors are poorly characterized but
undoubtedly relate to effects on diabetogenic hormone
secretion, hepatic function, lipid metabolism and/or
tissue responsiveness to insulin (Vail et al, 1990; Ogilvie
et al, 1997).
HYPERTRIGLYCERIDEMIA. Fasting lipemia is common
in dogs with diabetes mellitus. Lipemia is caused by
hypertriglyceridemia, which has been associated with
insulin resistance and carbohydrate intolerance in
humans and dogs. Hypertriglyceridemia may impair
insulin receptor–binding affinity, promote downregu-
lation of insulin receptors, and cause a postreceptor
defect in insulin action (Bieger et al, 1984; Berlinger
et al, 1984). Hyperlipidemia-induced increase in
hepatic glucose secretion, in conjunction with insulin
resistance, worsens carbohydrate intolerance. The
insulin resistance caused by hypertriglyceridemia is
most commonly appreciated in diabetic dogs that
develop hypothyroidism and in diabetic Miniature
Schnauzers with idiopathic hyperlipoproteinemia but
should be considered in any poorly controlled diabetic
dog with persistent lipemia.
Hypertriglyceridemia can be confirmed by measuring
serum triglyceride concentration, preferably from a
blood sample obtained after a 24-hour fast. In our
laboratory, the upper limit of normal for serum
triglyceride concentration is 150 mg/dl in healthy
dogs. Hypertriglyceridemia may be secondary to other
disorders or may be a primary idiopathic hyper-
lipidemia disorder (see Table 3-10, page 107).
Unfortunately, hypertriglyceridemia is common in
poorly regulated diabetic dogs, and the differentiation