Abbreviations
RBC
red blood cell count
AST
aspartate aminotransferase
CK
creatine kinase
NEF
non‐esterified fatty acids
BHB
ß‐hydroxybutyrate
DAB
diaminobenzidine
A 5‐year‐old Dexter cow in the fifth month of pregnancy was referred to the Clinic
for Ruminants and Swine, Department of Veterinary Medicine, Freie Universität Berlin,
Germany, by a local practitioner due to general weakness and ataxia that did not respond
to treatment. Approximately 3 months before hospitalization the cow had suffered from
two bouts of acute catarrhal mastitis, which were successfully treated by frequent
milking and parenteral administration of tylosin1 at a dosage of 10 mg/kg bodyweight
once daily on ten consecutive days. Approximately 6 weeks after the second bout of
mastitis (2 weeks before hospitalization), the cow had weakness and ataxia. There
was severe hypoglycemia2 (glucose concentration 21.6 mg/dL) while complete blood cell
counts and differentials, bilirubin, blood urea, creatinine, magnesium, calcium, phosphate
and iron concentrations remained within the reference range.3
At examination, the cow's posture was characterized by a wide‐based stance and reluctance
to walk. The cow had symmetrical ataxia and hypermetria of all limbs, intention tremors
of the head and a deficit in its menace responses. The ataxia was graded as 3−4 out
of 5.1 The cow demonstrated bilateral mydriasis and delayed pupillary reflexes (both
direct and indirect). Other cranial nerve function and spinal reflexes were normal
and there were no indications of head tilt, tail weakness, bladder atony, and perineal
hypalgesia. The cow showed no signs of circling. Given the symmetrical ataxia and
hypermetria of all limbs, the intention tremors and the deficit in its menace responses
the tentative neurologic diagnosis was cerebellar ataxia. The body condition score
was assessed 2 of 5.2 Respiratory rate (36/min), heart rate (72/min), and body temperature
(38.7°C) were within the reference range. Upon examination of the digestive tract,
dysphagia was observed while the animal was ruminating. Rumen fluid was dripping from
the oral cavity and a discharge containing rumen fluid was draining from the nostrils
intermittently. Urinalysis yielded normal results. The color of urine was yellow to
light amber and the specific gravity was normal (1.030; reference range 1.025–1.045).
Neither glucose nor ketone bodies were detected in the urine.
CBC and serum biochemistry revealed a low reticulocyte count (5.63 × 106 μL; reference
range 6–8 × 106 μL) and a slight left shift (band neutrophils 0.31 × 103 μL; reference
range 0–0.3 × 103 μL, segmented neutrophils 1.26 × 103 μL; reference range 1.3–4.5
× 103 μL), increased AST (86 U/L; reference range 0–50 U/L) and CK activities (860
U/L; reference range 0–150 U/L). Venous blood gas analysis identified a slightly increased
Base Excess (6 mmol/L; reference range −3–+3 mmol/L), while the blood pH was within
the reference range (7.39; reference range 7.35–7.45). NEFA4 (0.6 mmol/L, reference
range ≤0.4 mmol/L) were slightly increased, but beta‐hydroxybuyrate5 (0.2 mmol/L,
reference range ≤1.0 mmol/L) was normal. The plasma glucose concentration revealed
insulinoma as a tentative diagnosis based on the normal BHB concentration a negative
energy balance was considered unlikely. The increase in the activity of AST and CK
was probably because of transportation. To exclude cerebrocortical necrosis, serum
total thiamine concentration was determined by HPLC technique.3 Total thiamine was
41 μg/L (reference range of >50 μg/L).
No abnormalities were detected on endoscopy of the upper gastrointestinal and respiratory
tracts or ultrasonographic examination of the liver, gallbladder, spleen, intestine,
rumen, reticulum and kidney. On transabdominal ultrasonographic examination, pregnancy
in an advanced stage was ascertained and fetal viability confirmed by measuring fetal
heart rate (110 beats/min; normal fetal heart rate 90–125 beats/min).
Persistent hypoglycemia was demonstrated on four consecutive days starting from the
day after admission. The Dexter cow was initially treated with glucogenic precursors
(40 g propylene glycol and 40 g glycerol twice daily), which were administered orally
to control hypoglycemia. The treatment of hypoglycemia with corticosteroids was not
performed to avoid the risks of abortion. Serum samples for determination of insulin,
estrogen, and cortisol were sent to the Endocrinology Laboratory, Clinic for Cattle,
Tierärztliche Hochschule Hannover. A radioimmunoassay was applied for the quantitation
of serum insulin concentrations in cattle.6 The details for the analysis of above‐mentioned
variables have been described by Meyerholz (2014). Estrogen (246 pg/mL; reference
range >20 pg/mL for pregnant cows) and progesterone (6.3 ng/mL; reference range >5
ng/mL for pregnant cows) were detectable and confirmed pregnancy. Cortisol was also
detectable (19.8 ng/mL; reference range >2 ng/mL) ruling out adrenal insufficiency
as a cause of persistent hypoglycemia. Plasma glucose concentrations were extremely
low (33.3, 25.2, 30.6, 27.0 mg/dL; reference range 40–59 mg/dL) and serum insulin
concentrations exceeded 300 pmol/L in three of the four samples (316, 333, 109, and
501 pmol/L, respectively; reference range Holstein Friesian heifers 54–194 pmol/L)
on 4 days during hospitalization. The latter levels were higher compared to those
reported in literature when the same assay had been applied in pregnant primiparous
cows4 and also when a similar method had been applied in pregnant multiparous cows.5
No reference values, however, are available for serum insulin concentration in Dexter
cows.
Due to persistent hypoglycemia and hyperinsulinemia in the absence of ketonemia, an
insulin‐secreting tumor was suspected. An intravenous glucose tolerance test (IVGTT)
was performed to support this differential diagnosis. A long‐term catheter7 was placed
in the left jugular vein under aseptic conditions and connected with a three‐way stopcock
fitted with an extension8 tube. The tube was fixed to the skin with two simple interrupted
sutures. The system was filled with heparinised normal saline solution (50 IU/mL)
and blood samples for determination of glucose and insulin concentrations were drawn
from the catheter into appropriate 9 mL tubes9 after disposal of the 5 mL of aspirated
fluid. Following collection of the initial blood sample, a volume of 500 mL of 5%
glucose solution10 was administered over a period of 5 minutes through the long‐term
catheter (170 mg glucose/kg BW) and serial blood samples were obtained at 30 minutes
and 1‐hour intervals following glucose administration. From an initial level of 27.7
mg/dL glucose levels crossed the lower limit of the reference range (40 mg/dL) at
30 minutes following the start of the infusion, and decreased to below baseline at
90 minutes (Fig 1). Insulin concentrations crossed a high peak concentration (2180
pmol/L) from an initial level of 54 pmol/L and decreased to a level of 322 pmol/L
at 30 minutes postinfusion and remained elevated over the baseline concentration 5
hours long postinfusion (Fig 1). Determination of the glucose/insulin ratio, which
is the standard approach for diagnosis of insulinoma in dogs6 and humans,7 was not
considered an appropriate diagnostic tool in cattle as the physiological range of
insulin secretion and concentration in Dexter cows is unknown. The results of IVGTT
reflecting persisting hyperinsulinemia after plasma glucose levels had already returned
to subnormal levels further supported suspicion of an insulin‐secreting tumor. From
the results of the IVGTT, it was concluded that persisting hyperinsulinemia associated
with hypoglycemia as observed at IVGTT in this Dexter cow was demonstrative of an
insulinoma. Hypoglycemia in cattle is a common finding related to malnutrition in
young stock8 or to negative energy balance in the transition period.9 In contrast
to the case presented here, hypoglycemia originating from deficient energy supply
or increased demands of energy due to lactation is associated with increased levels
of NEFA and BHB due to lipomobilization and ketone body production. The Dexter cow
did not respond with ketone body production or excessive lipomobilization in the face
of hypoglycemia. This finding indicated a blockade of lipolysis and ketone body formation,
most likely due to persistent hyperinsulinemia which inhibits lipolysis and subsequent
ketone body production.10 The mildly increased NEFA levels of the Dexter cow might
be due to catecholamine‐induced lipolysis, which cannot be inhibited by insulin due
to reduced antilipolytic effect of insulin in pregnancy.10 The absence of excessive
lipomobilization and ketosis as well as persistence of hyperinsulinemia observed following
glucose administration, however, do justify the assumption of insulinoma but are not
conclusive as the insulin response to intravenous glucose infusion is physiologically
high (Peak values range between 137–1728 pmol/L) in dairy cows, as demonstrated in
two previous studies.11, 12 In humans, the term insulinoma is used to describe small
tumors composed of islet cell tissue that develop in the pancreas and ectopic sites
and cause hypoglycemia by their ability to secrete insulin.13 The clinical diagnosis
of insulinoma in humans is based on the presence of Whipple's triad, consisting of
the presence of clinical signs such as tremor, sweating, tachycardia, loss of consciousness,
giddiness, and blurring of vision that occur intermittently,14 persistent hypoglycemia
during fasting and, improvement of clinical signs after infusion of glucose.15 Signs
of neurologic dysfunction were reported in 30 patients with insulinoma.16 Confusion,
coma, convulsions, and weakness were predominant findings in these patients.16 Magnetic
resonance imaging and computed tomography are alternative diagnostic tools applied
in humans and small animals. Recently, intraabdominal ultrasonography was added to
the diagnostic spectrum in humans to increase the diagnostic sensitivity of insulinomas.15
Biochemical diagnosis of insulinoma in humans relies on unequivocally measurable insulin
concentrations in the fasting state, the concurrent measurement of C‐peptide together
with quantitation of ketone bodies.7 Furthermore, evaluation of the insulin–glucose
ratio is an important diagnostic parameter with a sensitivity of 93% and specificity
of 94%.7 In the present case, the two monoclonal spikes in the alpha 1‐fraction, which
were detected by serum electrophoresis, were probably due to pregnancy or increased
levels of C‐peptide. However, a validated C‐peptide assay for use in cattle is not
available yet.
Figure 1
Plasma glucose and serum insulin concentrations during intravenous glucose tolerance
test using 170 mg glucose/kg BW in the Dexter cow.
All treatment attempts including providing glucogenic nutrients and precursors (oral
administration of 40 g propylene glycol and 40 g glycerol twice daily throughout hospitalization)
failed and the condition of the cow deteriorated during hospitalization. Furthermore,
she had an abortion approximately in the sixth month of pregnancy. Abnormalities were
not detected on postmortem examination of the fetus. Abortion might have been caused
by undersupply of the fetus with glucose due to hypoglycemia of the mother. In pregnant
cows, glucose crosses the uterus and placenta insulin‐independently by the primary
glucose transporters (GLUT1 and GLUT3). A maternal hypoglycemia affects the fetal
glucose uptake directly because the fetus and placenta cannot sequester glucose against
its concentration gradient and the capacity of the fetus and placenta is limited to
compensate hypoglycemia by gluconeogenesis.17 Due to the clinical condition and the
unfavorable prognosis, euthanasia was elected and necropsy was performed at the Institute
of Veterinary Pathology, Faculty of Veterinary Medicine (Freie Universität Berlin).
At necropsy, multifocal, partially encapsulated, highly infiltrative, white‐grey nodules
with a maximal diameter of 2.8 cm were present in the right lobe of the pancreas (Fig
2). White nodes in pancreaticoduodenal and mesenteric lymph nodes were seen in addition.
Additional findings were chronic enteritis and a moderate chronic ulcerative abomasitis.
The proposed cause of abomasitis was chronic stress during hospitalization. For histological
evaluation, pancreatic tissue and lymph nodes were fixed by immersion in 10% neutral‐buffered
formalin for 96 hour and paraffin‐embedded. Sections of 4 μm in thickness were routinely
stained with hematoxylin and eosin (HE). These showed a multiple infiltrative coalescing
neoplastic masses of polygonal tumor cells arranged in nests and packets, overall
showing a neuroendocrine pattern (Fig 3). Multifocal necrosis and hemorrhage were
present. Tumor cells had high amounts of intensely eosinophilic cytoplasm (Fig 3)
and mitotic rate was low. Moderate pancreatic atrophy was present in the adjacent
exocrine pancreas. The aforementioned lymph nodes were almost completely infiltrated
by the tumor cells, resulting in a replacement of original lymphoid tissue. Intravascular
tumor cells were frequently observed. For a definitive diagnosis of the neoplasm,
immunohistochemistry using anti‐human insulin antibodies,11 anti‐chromogranin A, anti‐synaptophysin
and anti‐melan A was performed following routine protocols with a 15 minutes microwave
heating step in citrate buffer pH 6.0 as retrieval method for anti‐chromogranin A,
anti‐synaptophysin and anti‐melan A. A biotinylated goat anti‐mouse antibody diluted
1 : 200 was used as secondary antibody for all antibodies except anti‐insulin (goat
anti‐guinea pig). Color development was performed using DAB and hemalaun as counterstain.
The expected staining pattern for islet cell tumors is similar in different domestic
animals. Cells stain positive for chromogranin A or B depending on the neoplastic
cell type, protein gene product 9.5 (PGP9.5, synonym Ubiquitin carboxy‐terminal hydrolase
L1), synaptophysin or neuron specific enolase (NSE). In insulinomas, tumor cells are
additionally positive for insulin.18 Immunohistochemistry confirmed the clinical diagnosis
of insulin‐producing islet cell tumor in both the pancreatic neoplasm and the lymph
nodes, with the anti‐insulin antibody yielding strong intracytoplasmic signals with
occasional membrane staining (Fig 4). In addition, the neoplastic cells were melan
A‐negative, chromogranin A‐negative, and synaptophysin‐positive (not shown). Although
tumor cells stained unexpectedly negative for chromogranin A, the positive staining
for both synaptophysin and especially insulin support the diagnosis of an insulinoma.
Incubation of the slides with an irrelevant antibody as negative control did not result
in specific staining (not shown).
Figure 2
Pancreas with multifocal whitish nodules, formalin‐fixed.
Figure 3
Polygonal neoplastic cells with high amounts of eosinophilic cytoplasm arranged in
a neuroendocrine pattern of nests and packets, separated by moderate amounts of fibrovascular
stroma. HE stain.
Figure 4
(A) Anti‐insulin immunohistochemistry showing variably intense cytoplasmic staining
of neoplastic cells. (B) Higher magnification. Note that not all neoplastic cells
are positive for insulin. DAB staining (brown) with hematoxylin counterstaining (blue).
Insulinomas are endocrinologically active tumors of the pancreas derived from pancreatic
beta cells and have been reported in humans14 and a number of animal species.6, 19,
20 Insulinoma in cattle has previously been identified by immunohistochemistry and
light microscopy of suspected neoplasms encountered in the slaughterhouse at routine
meat inspection.20 Most of pancreatic tumors have been described as islet cell tumors
and most of them were classified as malignant.20 In humans, insulinomas are described
as usually solitary, benign, and encapsulated small lesions with a diameter <2 cm.13
Other pancreatic tumors that can cause hyperinsulinemia include pancreatic polypeptide‐secreting
tumors (PPomas), as described in dogs.6 Clinical signs, however, are typically absent
or go unnoticed in case of tumors that primarily produce pancreatic polypeptide in
dogs.21 Pancreatic tumors, that produce both pancreatic polypeptide and insulin, can
cause persistent hypoglycemia due to hyperinsulinemia.19, 21 Therefore, histopathological
examination of pancreas tissue is necessary in order to achieve an exact diagnosis
of insulinoma.19 A paraneoplastic hypoglycemia is an important differential diagnosis
of insulinoma and occurs due to a nonislet cell tumor that secrets incompletely processed
IGF‐II, which causes glucose consumption by interacting with IGF and insulin receptors
directly.6, 22 The determination of IGF‐I and IGF‐II levels in serum and IGF‐II:IGF‐I
ratio can help to confirm or rule out the diagnosis of paraneoplastic hypoglycemia.6,
22 Further differential diagnoses of insulinoma, which might cause a persistent hypoglycemia,
include disorders of counter regulatory hormone release (glucagon, epinephrine, growth
hormone, and cortisol), especially adrenal insufficiency.23 Adrenal insufficiency
is either congenital or acquired and characterized by an insufficient production of
steroid hormones (glucocorticoids and often mineralocorticoids). ACTH stimulation
test can establish the diagnosis and distinguish whether it is primary or secondary
adrenal insufficiency.23 Addison's disease (chronic adrenal insufficiency) can occur
as a result of a primary disorder of the adrenal gland or secondary to a deficiency
of hypothalamic and pituitary hormones such as adrenocorticotropic hormone (ACTH)
or corticotropin‐releasing hormone (CRH).23 Adrenal insufficiency has never been described
in cattle except one case24 and was ruled out in the present case by evaluation of
cortisol and potassium concentrations. Neither disorders of counter regulatory hormone
release nor nonislet cell tumors are associated with hyperinsulinemia, contrary to
insulinoma. Animals with insulinoma do not exhibit a compensatory drop of insulin
secretion in the presence of hypoglycemia.25 Therefore, IVGTT seems to be the diagnostic
method of choise for insulinoma in cattle, to characterize the regulation of insulin
release and the pancreatic insulin response to changes in glucose levels.
Surgical excision of neoplastic tissue is treatment of choice in humans14 and ferrets.19
In dogs, the long‐term medical treatment is applied by oral administration of diaxozide
possibly in combination with prednisolone.6 A partial pancreatectomy contributed to
a longer survival time in dogs with insulinoma.6 Persistent hypoglycemia in the absence
of ketosis in adult dairy cattle presenting with signs of neurologic dysfunction could
indicate the presence of an insulinoma. The present case of a cow with signs of neurologic
dysfunction associated with persistent hypoglycemia and hyperinsulinemia illustrates
a case of an insulinoma with inhibition of lipolysis and ketogenesis. When diagnosis
is reached in vivo, glucocorticoid medication in combination with partial pancreatectomy
might be attempted, but possible metastases have to be considered.