Endocrine-Related Cancer (1998) 5 111-129
Tumours producing hypoglycaemia
V Marks and J D Teale
European Institute of Health and Medical Sciences, University of Surrey, Guildford GU2 5RF, UK and Department of Clinical
Biochemistry and Nutrition, Royal Surrey County Hospital, Guildford GU2 5XX, UK
(Requests for offprints should be addressed to V Marks)
Introduction
Hypoglycaemia is one of the commonest of medical
emergencies, mainly due to its high frequency as a
complication of insulin therapy of diabetes. It sometimes
arises, however, ‘spontaneously’ as a manifestation of
other diseases. The present review deals exclusively with
hypoglycaemia caused by neoplastic disease.
For a very brief period, between 1927 and 1930,
tumours capable of producing hypoglycaemia were
thought always to secrete insulin and arise exclusively
within the pancreas. In the intervening years it has become
increasingly apparent that almost any type of neoplasm
can cause hypoglycaemia through a number of different
mechanisms (Marks & Rose 1964, Laurent et al. 1971,
Daughaday 1989, Fajans & Vinik 1989).
Symptomatology
Hypoglycaemia may be the presenting symptom of a
tumour, a terminal epiphenomenon or anything in
between, but is always rare. It is usually episodic and
provoked by fasting rather than reactive to the ingestion of
food. The one notable exception is hypoglycaemia due to
pheochromocytoma which may be of either variety
(Gorden et al. 1981, Hiramatsu et al. 1987, Paineau et al.
1988, Akiba et al. 1990).
Symptoms of acute neuroglycopenia, character-
istically associated with insulin overdose, are unusual in
tumour-induced hypoglycaemia. Instead there is a gradual
loss of cognitive function associated with diminished
motor activity, lethargy, the onset of somnolence and a
gradual drift into stupor or coma. Recovery may occur
spontaneously at any time but is hastened by the ingestion
of carbohydrate by mouth or the injection of glucose or
glucagon.
Biochemically, tumour-induced hypoglycaemia is
usually associated with hypoketonaemia (<300 µmol/l)
unlike that of most other types of ‘spontaneous’ hypo-
glycaemia apart from insulinoma (Teale et al. 1987). Only
rarely are patients desperately ill in between hypo-
glycaemic episodes except when it is a terminal event in
patients with liver metastases.
Endocrine-Related Cancer (1998) 5 111-129 © 1998 Society for Endocrinology Printed in Great Britain
1351-0088/98/005-111 $08.00/0
Classification of hypoglycaemia-
producing tumours
Tumours producing hypoglycaemia can, broadly
speaking, be divided - on the basis of the mechanism by
which they most commonly produce hypoglycaemia -
into: (1) insulin-secreting tumours; (2) non-islet cell
(insulin-like growth factor-II (IGF-II)-secreting) tumours;
(3) myeloma, lymphoma and leukaemia; and (4) meta-
static neoplasia.
Hypoglycaemia produced by tumours other than
insulinomas is usually referred to as ‘non-islet cell tumour
hypoglycaemia’ (NICTH) and has more than one cause.
Pancreatic endocrine tumours
Insulin-secreting tumours are the commonest hormone-
producing neoplasms of the gastrointestinal tract. They
occur with an incidence of between 0.5 and 1 per million
of the population per year (Kavlie & White 1972,
Atkinson et al. 1981, Cullen & Ong 1987, Watson et al.
1989). More than 80% are solitary benign adenomas,
composed mainly or exclusively of morphologically
normal B-cells. About 10% are metastatic, i.e. malignant,
whilst a further 10% are multiple but behave as benign.
Women outnumber men in the ratio of 6:4 for benign
but not for malignant insulinomas. The highest incidence
is in middle aged adults, though both benign and
malignant insulinomas have been observed very rarely in
children and in the very elderly.
Insulin-secreting tumours
Clinically significant insulinomas are generally between
10 and 20 mm in diameter at the time of diagnosis,
although tumours as large as 150 mm or as small as 5 mm
in diameter have been found during surgery. Occasionally
very small tumours are overlooked at operation and found
only after very thorough histopathological examination of
the resected pancreas, or not at all.
In fewer than 5% of cases of true hyperinsulinaemic
hypoglycaemia can no tumour be found. Whilst some of
these may be due to a small tumour being overlooked
either at operation or at autopsy others are due to a
functional abnormality of the B-cells similar to that
111
Marks and Teale: Tumours producing hypoglycaemia
causing persistent hyperinsulinaemic hypoglycaemia of
infants (PHHI) which may, or may not, manifest itself
histologically as islet hyperplasia (nesidioblastosis). This
condition must be distinguished from the nosologically
distinct condition of microadenomatosis which can occur
either as an isolated histopathological finding or as part of
multiple endocrine neoplasia type I (MEN-I) (Diarmuid et
al. 1994).
The natural history of insulinomas (Marks 1991)
suggests that apart from their propensity for causing
hypoglycaemia - from which death is fortunately very
rare - they produce no adverse effects. Specifically they
appear not to cause obstruction to the hepatic or pancreatic
ducts or to undergo malignant transformation. There are
many reports, especially in the older literature, of
symptoms attributable to hypoglycaemia due to
insulinoma extending over a period of 25 years or more
before its true cause was identified and removed (Fonseca
et al. 1989).
The average time between onset of symptoms and
diagnosis of its cause as due to insulinoma is now about
1 year. Delay, when it does occur, is almost always the
consequence of reluctance by the patient to seek help, or
failure of the practitioner to suspect hypoglycaemia. Once
suspected, diagnosis of insulinoma seldom presents any
real difficulty (Marks & Rose 1981, Marks 1991, Marks &
Teale 1996, Marks 1997, Nauck et al. 1997).
Insulinomas are distributed evenly throughout the
pancreas and only very rarely ectopically (2%). In this
respect, as in the relative infrequency of their malignancy,
insulinomas differ from other hormone-producing
tumours of the gastrointestinal tract.
The commonest ectopic sites are the duodenum and
the immediate vicinity of the pancreas, but less than 1% of
the 677 insulinomas collected from the literature by
Laurent and co-workers prior to 1971 were located outside
the pancreas. Only a tiny number of ectopic tumours have
been reported in recent years (Yoshikawa & Wakasa 1980,
Miyazaki et al. 1986).
Insulinoma histology Many insulinomas are composed
entirely of seemingly normal B-cells. More thorough
examination, especially by electron microscopy and
immunocytochemistry, usually reveals a small but
variable number of other pancreatic endocrine cell types,
of which somatostatin-containing D-cells are the
commonest.
Insulinomas show variable and unpredictable staining
with traditional as well as with immunohistological
reagents, including synaptophycin, neuron-specific
enolase and chromogranin, which have been used as
general markers of neuroendocrine tumours or tumours of
amine precursor uptake and decarboxylation cells
(APUDomas). Some stain very poorly or not at all even
with anti-insulin antisera.
112
Attempts to classify insulinomas into prognostically
useful histopathological categories have been made for 60
years or more with little success (Graeme-Cook et al.
1990). It became apparent soon after their initial
identification that classical morphological and histo-
logical criteria were of little value for distinguishing
malignant from benign tumours. Only the presence of
metastases, or their subsequent appearance, could be
relied upon to make the distinction.
Creutzfeldt and co-workers (1973) classified insulin-
omas into four types on the basis of their histological,
ultramicroscopic and immunocytological appearances.
Type 1 is the commonest and accounts for almost half
of the cases. In this type, virtually every cell contains
typical β-granules.
Type 2, the second commonest, accounts for 25% of
the cases. Most of the cells contain a number of atypical
granules - demonstrably different from those of normal A,
B, D or PP cells - in addition to typical β-granules. The
atypical granules resemble the coated β-granules that are
found in small numbers in normal B-cells and within
which proinsulin is believed to be converted into insulin
and C-peptide (Orci et al. 1988).
Type 3 tumours consist solely of cells containing
atypical β-granules. None contains typical ones. Though
not specifically identified as such, these probably
correspond to predominantly, or exclusively, proinsulin-
secreting tumours (proinulinomas).
Type 4 tumours included the only two metastatic
carcinomas in the Creutzfeldt series. None of the four
tumours so classified had any of the characteristic histo-
logical features of an insulinoma. All gave negative
reactions with classical histological and immunocyto-
chemical stains. They all showed ultramicroscopic signs
of high functional activity. Tumours of this type would
possibly, in former times, have been classified as
carcinomas of unknown origin or carcinoids rather than
‘insulinomas’, even in experienced histopathological
laboratories.
Berger and co-workers (1983) classified hypo-
glycaemia producing tumours of the pancreas on their
functional rather than histological characteristics.
Group A tumours were characterised by the almost
complete suppressibility of insulin and proinsulin
secretion by both diazoxide and somatostatin. Histo-
logically they were characterised by an abundance of
well-granulated B-cells in a trabecular arrangement.
Group B tumours resisted the suppressant effects of
diazoxide and somatostatin and their cells were arranged
in a ‘medullary pattern’ and contained fewer less well
granulated β-granules.
Only 20% of circulating total immunoreactive insulin
(IRI) was proinsulin and proinsulin-like components in
group A patients, which is not very dissimilar to that of

normal subjects. In group B patients on the other hand they
accounted for 50% or more of total IRI. The difference
reflects the smaller proportion of coated to uncoated
β-granules in the group A patients.
Despite their sophistication compared with previous
classifications, neither the Creutzfeldt nor Berger classi-
fications enables a prognostically important distinction to
be made between malignant and non-malignant tumours.
Nor, for that matter (Graeme-Cooke et al. 1990), does the
observation that malignant, but not benign, insulinomas
express - and sometimes secrete - the subunits of human
chorionic gonadotrophin which had originally appeared to
offer this possibility.
Confusion between malignant insulinomas mas-
querading as metastatic carcinoids may, and sometimes
does, occur. This is less likely to occur if immuno-
staining with insulin antisera is carried out routinely on all
pancreatic carcinoids since symptoms from hypogly-
caemia are not inevitable until quite late in the course of
the disease (Marks 1995).
Amyloid A substance with the histological appearance of
amyloid was recognised as being deposited in varying
amounts in insulinomas whether benign or metastatic long
before it was isolated and characterised chemically
(McIntyre 1989, Porte & Kahn 1989, O’Brien et al. 1994).
Islet amyloid polypeptide (also known as amylin) is a 37
amino acid polypeptide which is co-secreted with insulin
from both normal and diseased B-cells. It may be co-
deposited with the pentomeric amyloid protein (pentraxin)
common to all varieties of amyloid in the islets of patients
with non-insulin-dependent diabetes mellitus (NIDDM)
as well as in patients with pancreatic endocrine tumours
(Crowley et al. 1996), though it is rare in all except
insulinomas.
It has been suggested that amylin might be of
pathogenic importance in the aetiology of NIDDM. It is
now known, however, to be co-produced and co-secreted
with insulin by normal as well as by insulinoma B-cells
and those of patients with NIDDM. It is still unknown,
however, why islet amyloid polypeptide should be
deposited in large amounts in some insulinomas but not in
others, nor whether it has any prognostic significance. It
does not occur in children with (familial) PHHI (Rother et
al. 1995), which, in its nosology at least, resembles
insulinoma.
Mixed cell tumours and carcinoids
While most benign insulinomas contain isolated D and/or
A cells, their presence in greater numbers is usually
associated with malignancy, i.e. metastasising tumours.
Metastatic islet cell tumours often resemble carcinoids
histologically and may be reported as such unless special
staining techniques are used. They may be associated with
no evidence of endocrinological disturbance until quite
Endocrine-Related Cancer (1998) 5 111-129
late in life and long after the initial histological diagnosis
has been made.
Functional transformation from a pancreatic endocrine
tumour manifesting features of a ‘glucagonoma’ into an
‘insulinoma’ and vice versa has been described. So too
have manifestations of ectopic hormone production which
are more often seen with other types of pancreatic
endocrine tumour most of which are malignant. Among
the most common syndromes associated with ectopic
hormone production by insulin-secreting carcinomas of
the pancreas are Cushing’s syndrome, Zollinger-Ellison
syndrome, hypersomatostatinaemia, acromegaly, carcin-
oidosis, and goitre and hypercalcitoninaemia (Laurent et
al. 1971, Wright et al. 1980, Marks & Rose 1981, Bugalho
et al. 1994).
Metastatic insulinomas and mixed islet cell tumours
vary in their aggressiveness (Ericksson et al. 1989). Pro-
gression, from the first appearance of symptoms to death
from inexorable hypoglycaemia can occur in 6 months or
less unless effective antihypoglycaemic treatment with
diazoxide plus chlorothiazide, surgical ablation (de-
bulking), chemotherapy, hepatic artery embolisation, or
cryotherapy either alone or in combination, can be
achieved (Valette & Souquet 1989, Wells et al. 1990,
Krentz et al. 1996).
Providing hypoglycaemia can be prevented, good
quality-of-life with survival (for 10 years or more) can
occur since these tumours, though malignant, are often
very slow growing and do not produce the typical features
of metastatic disease such as anorexia, nausea and loss of
body weight until late in the course of the disease.
Aetiology of pancreatic endocrine tumours
Most insulinomas occur sporadically, a further 5-10% as
part of the MEN-I syndrome (Diarmuid et al. 1994,
Thakker 1997). An association with NIDDM has been
suggested on the basis of a small number of family studies
but a link has not been established (Service et al. 1976).
A relative insensitivity to exogenous insulin is often
observed in patients harbouring insulinomas and generally
attributed to, though far from established as due to, ‘down
regulation’ of insulin receptors (Nankervis et al. 1985,
McGee et al. 1987, de Kreutzenberg et al. 1994). It has
long been known to persist after successful removal of the
tumour and abolition of hypoglycaemia, and in a minority
of cases insulin manifests itself as impaired glucose
tolerance or even frank diabetes. While this may be
secondary to damage done to the pancreas during
resection of the tumour (Dunn 1971) it may also be due to
persistence of a primary defect in peripheral insulin
sensitivity (de Kreutzenberg et al. 1994) similar to that
seen in patients with NIDDM.
It is probably no coincidence that no example of a
benign insulinoma developing in a patient with type I
113
Marks and Teale: Tumours producing hypoglycaemia
diabetes (insulin-dependent diabetes mellitus (IDDM))
has ever been reported, in contrast to patients with type II
diabetes (NIDDM) in whom it occurs with possibly an
even greater frequency than in the general population
(Heik et al. 1988, Grunberger 1993). A contrary view has
been expressed by Kane et al. (1993) who found pre-
existing diabetes in only 1 of their series of 313 patients
with insulinoma, making it a very rare occurrence indeed.
A case of malignant insulinoma has, however, recently
been reported in a patient with incontrovertible type I
diabetes (Svartberg et al. 1996).
MEN-I
Insulinomas are the commonest manifestation of MEN-I
after parathyroid and pituitary adenomas. They do not
occur in the other varieties of multiple endocrine adeno-
matosis, i.e. MEN-IIa and -IIb. The familial occurrence of
seemingly solitary insulinomas (Maioli et al. 1992) has
been reported on several occasions, but whether this
represents a forme-fruste of MEN-I (Diarmuid et al. 1994)
or a specific abnormality awaits gene analysis of affected
individuals (Larsson et al. 1988).
Hyperinsulinaemic hypoglycaemia in children and
young adults is particularly likely to be due to either
MEN-I or to the persistent functional hyperinsulinism of
infancy, formerly referred to as nesidioblastosis (Mathew
et al. 1988, Losada et al. 1995). In children below the
age of 4 years the histopathological features of solitary
insulinomas removed at operation resemble those of
‘nesidioblastosis’ more closely than those of solitary
benign insulinomas in adults.
Chemical pathology
Insulin Spontaneous hypoglycaemia was believed from
the time of its first description as a clinical entity, until the
introduction of radioimmunoassays for insulin in blood, to
result from overproduction of insulin by an insulin-
secreting tumour similar to that of thyroxine in thyro-
toxicosis rather than, as is now quite apparent, from its
inappropriate rather than excessive secretion (Samols &
Marks 1963).
Normal glucose homeostasis depends, in large part,
upon the ability of the very modest fall in blood glucose
concentration produced by fasting to below a pre-
determined ‘set’ to suppress endogenous insulin secretion.
The suppression is not ordinarily complete, however,
(unlike in IDDM) so that although peripheral glucose-
dependent tissues no longer abstract glucose from the
glucose pool, lipolysis in adipose tissue and glycogen-
olysis in the liver is not sufficiently disinhibited as to
permit rampant ketosis and hyperglycaemia. Instead
plasma ketones (and non-esterified fatty acid (NEFA)
levels) rise only modestly and net glucose production by
the liver (appearance) is constrained to match utilisation
114
by the brain and other obligatory glucose using tissues
(Marks & Rose 1981).
It is the inability of insulin-secreting tumours of the
pancreas to shut down their insulin secretion in response
to hypoglycaemia, whether induced by fasting or exo-
genous insulin, that is their biochemical hallmark. This
anomaly occurs only very rarely in other disorders, most
notably PHHI or ‘nesidioblastosis’ and sulphonylurea
overdose (Teale et al. 1989).
Though insulinoma B-cells may be functionally
abnormal in other ways, none is as uniformly present as
failure to suppress insulin secretion in response to hypo-
glycaemia. Most respond poorly, if at all, to the insulino-
trophic effects of hyperglycaemia per se (produced by
intravenous glucose for example) and do not exhibit a first
phase insulin response to intravenous glucose.
They do, on the other hand, generally respond
normally or excessively to glucagon and other insulino-
trophic enteric hormones (gastric inhibitory peptide and
glucagon-like peptide-I (GLP-I)) that are released into the
blood following ingestion of a meal. This may account for
the rare examples of insulinoma that present with a history
of reactive rather than of fasting hypoglycaemia.
The insulinotrophic amino acid L-arginine generally
produces a normal insulinaemic response in insulinoma
patients. This is in contrast to L-leucine which - whether
given orally or intravenously - often provokes excessive
insulin secretion and precipitates hypoglycaemia. A
similar response is observed only in patients with the
leucine-sensitive variety of PHHI and normal subjects
pretreated with sulphonylureas.
Insulinomas may respond to hypercalcaemia by
secreting insulin in amounts sufficient to produce hypo-
glycaemia. Advantage has been taken of this fact -like that
of other insulin secretagogues, such as tolbutamide and
glucagon - in the diagnosis of insulin-secreting tumours.
All such tests are of limited diagnostic use because of
their low predictive value. They are most useful
(Doppman et al. 1995) when employed in conjunction
with percutaneous trans-hepatic catheterisation for pre-
operative localisation, should this be considered
necessary, of insulinomas already diagnosed by more
reliable methods.
Proinsulin Proinsulin ordinarily constitutes a higher
proportion of circulating total IRI in patients with
insulinomas than in healthy people, especially in patients
harbouring malignant tumours (Samols 1965, Hayashi et
al. 1977, Cohen et al. 1986, Cohen & Camus 1988,
Hampton et al. 1988, Hale et al. 1991, Marks et al. 1992,
Tasaka et al. 1993, Gorden et al. 1995). It is still unclear
whether this is a consequence of the neoplastic process or
a factor in its aetiology, as it might be if the propensity to
develop an insulinoma is greater in people with a
predisposition to NIDDM.

Rarely, benign tumours have been found that secrete
proinsulin as their main (Alsever et al. 1975) or even sole
insulin-like product, leading to the suggestion that they
should be called ‘proinsulinomas’ (Cohen & Camus
1988).
In one such case, recently studied by us, all of the very
small amount of circulating insulin appeared to derive
from the normal islets and it responded to insulinotrophic
stimulation with glucose, glucagon etc. and to suppression
by octreotide, whereas the proinsulinoma itself did not.
Non-pancreatic tumours
In 1929 Nadler and Wolfer published a case of hypo-
glycaemia associated with primary carcinoma of the liver.
One year later Doege described a patient with episodes of
altered consciousness from whom he removed a large
mediastinal tumour. The significance of the symptoms
was not appreciated, however, until some 3 years later
when they recurred and were shown to be of hypo-
glycaemic origin (Marks & Rose 1964, Laurent et al.
1971).
More cases of hypoglycaemia occurring in association
with hepatic and other types of carcinoma, or with
mesenchymal tumours, were described over the years, in
which it was the presenting or dominant feature of the
patient’s illness. In others it occurred only as a terminal
event.
Hypoglycaemia of both types is referred to, through-
out the rest of this article, as NICTH and has many causes.
Types of tumour
The association that has attracted most interest has been
that between hypoglycaemia and large fibrosarcomas,
usually of low or moderate malignancy, arising in the
thorax or retroperitoneal space. Hypoglycaemia is, how-
ever, rare even in them. In one large series of cases
(England et al. 1989) only 3% with benign and 11% with
malignant fibromatous tumours developed hypo-
glycaemia at any time in their evolution. Apart from
hepatomas the proportion of epithelial tumours mani-
festing hypoglycaemia is even smaller, probably less than
0.1%.
The true incidence of tumour-induced hypoglycaemia
is, however, difficult to ascertain and no population study
comparable with those for insulinoma has been published.
In our own experience of cases of hypoglycaemia referred
for diagnosis from all over Britain, NICTH is about one-
quarter (25%) as common as insulinoma but this is
probably an underestimate. Many cases undoubtedly go
unrecognised especially in patients with metastatic
disease.
The cause of NICTH was, from the time of its
discovery until comparatively recently, the subject of
intense speculation (Field et al. 1963, Laurent et al. 1971,
Endocrine-Related Cancer (1998) 5 111-129
Marks & Rose 1981). It is now recognised that in most
cases it is a consequence of overproduction of an aberrant
form of IGF-II (Megyesi et al. 1974, 1975, Gorden et al.
1981, Axelrod & Ron 1988, Ron et al. 1989, Lowe et al.
1989, Shapiro et al. 1990, Teale & Marks 1990,
Daughaday & Trivedi 1992). Other, even rarer causes, are
sometimes encountered, however, and these will be
discussed later.
Histological classification
Though fibromas and fibrosarcomas are probably the best
known example of tumour-producing NICTH, no histo-
logical type of tumour is exempt. Hypoglycaemia, once
manifest, is usually both profound and relentless. It may,
however, respond to surgical ablation or debulking of the
tumour - at least temporarily. Surprisingly hypoglycaemia
does not always reappear even if the tumour recurs and
grows to its former size.
In patients who present with NICTH the anamnesis is
similar to that of insulinoma. Differentiation is, however,
easily made on the basis of plasma hormone assays (Marks
& Teale 1996). Unlike insulinoma, in which hypo-
glycaemia is always the presenting symptom, in NICTH
the diagnosis of neoplasia may already have been made by
the time hypoglycaemia appears.
NICTH is not necessarily an indicator of size or
invasiveness of the tumour, though fibromas and fibro-
sarcomas are invariably large or very large by the time
they manifest it. Only very rarely do such tumours weigh
less than 500 g and most weigh between 2 and 4 kg. The
largest so far recorded weighed over 20 kg. Most are in the
thorax or in the retroperitoneal space and may, despite
their large size, have been clinically silent.
Other tumours of a special interest because of their
association with hypoglycaemia are hepatomas, haem-
angiopericytomas, pheochromocytomas, adrenocortical
adenomas and carcinomas, myelomas, lymphomas and
leukaemias. Even tumours of the meninges have been
described as causing NICTH (Ferguson & Flinn 1995,
Phuphanich et al. 1995) and it would appear reasonable to
assume that no type of tumour is exempt (Table 1).
Craniopharyngiomas, pituitary adenomas and tumours
of the brain that have been treated by irradiation may cause
hypoglycaemia secondary to hypopituitarism - usually as
a late phenomenon but occasionally as the presenting
symptom. They will not be considered here.
Chemical pathology of NICTH
Despite a few early reports of dubious authenticity in
which insulin was claimed to be the causative agent
(Pavelic & Popovic 1981, Baltic et al. 1985) NICTH is
now known not to be caused by insulin. Indeed, doubt has
been expressed as to whether ectopic insulin production
has ever been demonstrated in a tumour that was not of
115
Marks and Teale: Tumours producing hypoglycaemia

partially resolved but still account for some of the

Table 1 Histological classification of 68 consecutive difficulties experienced by authors using poorly validated
tumours producing NICTH through overproduction of assay techniques.
IGF-II investigated in Guildford. The commonest cause of NICTH is undoubtedly the
overproduction of an aberrant form of IGF-II (Daughaday
Histology Number
1997). Most, if not all of the other biochemical effects that
Carcinoma 31 are observed, such as suppression of insulin, glucagon,
Lung 7 growth hormone (GH) secretion and depression of
Pancreas 6 circulating NEFA and ketone levels, stem from this.
Stomach 5 There are, however, other rare causes of NICTH such
Undifferentiated 3 as autoantibodies to the insulin receptor, overwhelming
Adrenal 2 metastatic destruction of the liver and the production of
Kidney 2
tumour necrosis factor (TNF) and other hypoglycaemic
Oesophagus 2
Ovary 1 cytokines, and these are considered later.
Prostate 1 IGF-I and IGF-II Description of the IGFs, their isolation,
Breast 1 nomenclature and full range of biological properties is
Larynx 1 beyond the scope of this review (Froesch et al. 1963, Jacob
et al. 1968, Froesch & Zapf 1985, Clemmons 1989). So
Sarcoma/fibroma 23
too is discussion of the IGF-binding proteins (IGFBPs)
Fibroma 6
Mesothelioma 5
(Baxter & Martin 1989, Baxter et al. 1995, Baxter 1996)
Fibrosarcoma 4 of which at least six types have been identified in plasma
Haemangiopericytoma 4 and possibly a seventh.
Neurofibroma 1 The IGFs were identified by their ability to mimic the
Leiomyosarcoma 2 effects of insulin upon isolated tissues in vitro in the
Sarcoma of kidney 1 presence of neutralising anti-insulin antibodies. Together,
the two IGFs account for most of the biological activity
Hepatoma 4
described in the older literature as non-suppressible
Carcinoid/neuroendocrine 1
insulin-like activity.
Both IGF-I and IGF-II are structurally similar to
Unknown 9 proinsulin. Immunologically they differ both from it and
from one another. Immunoassays have been developed for
each of the IGFs with little or no cross-reactivity between
‘pancreatic’ origin. The few cases of ectopic insulin them and none at all with proinsulin. Clear separation of
production extant in the literature have usually either the three chemical species was, however, not possible with
been metastatic APUDomas (Sasaki et al. 1980) possibly earlier bio- and radioreceptor assays and this may explain
of pancreatic origin, or are insufficiently well documented some of the confusion found in the older literature
to substantiate the claim that they were insulin or pro- surrounding investigations into the aetiology of NICTH.
insulin producing. A very recent case of cervical cancer Both IGF-I and IGF-II circulate almost completely
investigated by us is a possible exception since it produced (>90%) bound to specific IGFBPs. This is in complete
insulin, C-peptide and proinsulin in prodigious amounts. contrast to insulin and proinsulin for which there are no
In an otherwise typical case of NICTH described by Lyall specific binding proteins.
et al. (1975) plasma insulin levels were both absolutely The two IGFs bind with more or less equal avidity to
and inappropriately high during hypoglycaemia prior to each of the IGFBPs. The complex formed associates with
removal of a 680 g thoracic neurofibrosarcoma but an acid-labile subunit (ALS) to make a ternary complex.
returned to normal post-operatively with restoration of This has a molecular mass in the region of 150 kDa, which
normoglycaemia. The tumour contained no secretory virtually confines it to the intravascular compartment. In
granules and extracts of it contained no insulin. The normal healthy subjects the total concentration of IGF is,
possibility that the tumour secreted an insulinotrophic in molar terms, equal to or slightly less than that of the
substance cannot be dismissed. combined IGFBPs.
The controversy as to whether NICTH is due to Production of the IGFBPs and their release into the
overproduction of one or other of the two IGFs (IGF-I and circulation is determined by a variety of factors.
IGF-II) or to some other substance with insulin-like Production of the most plentiful (IGFBP-3) is linked, in
properties was largely a consequence of difficulties with some poorly understood way, to the production of IGF-I
assay technology (Baxter 1990). These have now been itself. This, in turn, is linked to GH action on the liver.

116

Though many, if not all, tissues, are capable of producing
it, IGF-I in the circulation derives almost entirely from the
liver. The liver is also the main, if not sole, source of the
IGFBPs. The origin of circulating IGF-II is seemingly not
exclusively the liver.
The normal molar ratio of IGF-II:IGF-I in plasma
varies but is usually about 3:1. Their concentration,
whether expressed in molar or mass units, is many
hundreds of times that of insulin (Table 2). They do not
ordinarily exert an insulin-like effect, however, either
on blood glucose or NEFA levels because they are
virtually excluded from contact with insulin and IGF
receptors on tissues due to their incorporation into the
IGFBP/ALS/IGF ternary complex. This may explain why,
despite their normal concentration in patients with
uncontrolled IDDM, IGFs exert no hypoglycaemic or
antilipolytic effect.
Big IGF-II The demonstration of increased amounts of
the mRNA for IGF-II in tumour tissue taken from patients
with hypoglycaemia-producing tumours established a role
for it in the pathogenesis of NICTH (Lowe et al. 1989,
Shapiro et al. 1990). It was, however, only after radio-
immunoassays for circulating IGF-I, IGF-II, and latterly,
the E-domain of the ‘large’ or aberrant form of IGF-II,
became available (Perdue et al. 1991) that the key role of
IGF-II in the pathogenesis of NICTH has been proved.
Indeed the presence, in plasma, of increased concentration
of IGF-II, mostly of the ‘big’ but also the normal variety
(Daughaday et al. 1988, 1990), coupled with a gross
reduction in the amount of IGF-I is so characteristic of
NICTH that it is diagnostically useful (Teale & Marks
1990).
It has still to be determined whether the reduced
affinity of ‘big’ IGF-II for IGFBPs (Daughaday &
Kapadia 1989, Daughaday 1997) or the reduction in
Table 2 Diagnostic parameters in 50 consecutive cases of NICTH.
NICTH patients
Age (years) 16 ±15 (30-91)
Glucose (mmol/l) 2.4±2.5
Immunoreactive insulin (pmol/l) <25
C-peptide (pmol/l) <75
β-hydroxybutyrate (µmol/l) 143±151
Growth hormone (mU/l) 4.6±6.2
Total IGF-I (nmol/l) 5±2
Total IGF-II (nmol/l) 102±35
IGF-II:IGF-I molar ratio 21.8±10.7
‘Big’ IGF-II (nmol/l) 20.1±6.0
* Including hyperinsulinism: iatrogenic and spontaneous.
** Excluding hypoglycaemia caused by hyperinsulinism: iatrogenic or spontaneous.
Endocrine-Related Cancer (1998) 5 111-129
circulating amounts of IGFBP-3 and ALS is the main
reason that these tumours produce hypoglycaemia. They
do so by increasing the availability of ‘free’ IGF-II to bind
to IGF receptors - and possibly to insulin receptors - in the
tissues.
Immunoassay Immunoassays for plasma IGF-I have
been used clinically for many years in the assessment of
malnutrition, for diagnosing and monitoring the response
to treatment in acromegaly and to human growth
hormone (hGH) in hypopituitarism (Teale & Marks 1986).
Clinical assays for IGF-II are, however, comparatively
new and still poorly validated (Baxter 1990). Most
antisera fail to distinguish between regular IGF-II and the
larger molecular forms possessing the E-domain. This is
ordinarily cleaved from proIGF-II during intracellular
processing but in NICTH there is a defect which leads to
the appearance in the plasma of a disproportionately large
amount of the uncleaved peptide. This is analogous to the
disproportionate ratio of proinsulin to insulin in the blood
of most patients with insulinomas.
IGF-II:IGF-I ratio Teale & Marks (1990) were the first
to draw attention to the distortion of the IGF-II:IGF-I ratio
in patients with NICTH and suggest that it might prove
useful in its diagnosis, especially in cases in which IGF-II
levels themselves were ‘normal’ or only slightly elevated.
Their early conclusions have been confirmed repeatedly
and a molar ratio of total plasma IGF-II:IGF-I of more
than 10 is virtually pathognomonic of NICTH. Apart from
NICTH, abnormally high IGF-II:IGF-I ratios are found
only in very rare cases of hypoglycaemia due to sepsis or
cachexia, but in them the absolute amounts of IGF-II are
generally subnormal as well as those of IGF-I.
Results in 50 consecutive patients with NICTH whose
plasma was assayed in Guildford are shown in Table 2.
Twenty-eight patients were women and 22 men. Seventy-
Control
hypoglycaemic subjects*
<3.0
Variable, depending on cause
Variable, depending on cause
>300**
>10*
18±14
68±12
3.8±1.5
9.8±2.0
117
Marks and Teale: Tumours producing hypoglycaemia
six percent were aged 60 or more, the youngest was 28
years old, the oldest 91. All had blood glucose levels
below 2.2 mmol/l at some time in the course of their
illness, although only 44 patients were hypoglycaemic at
the time blood was collected for hormone assay. Plasma
β-hydroxybutyrate levels were below 300 µmol/l in 37 of
the 42 patients in whom they were measured whilst
hypoglycaemic. In none did they exceed 600 µmol/l.
More than 50% of the tumours in which the histology
was known were of epithelial origin. Carcinomas of
the lung and of the stomach were the most common
(Table 1).
Other hormones Abnormalities of many other hormones
concerned with glucose homeostasis are characteristic of
NICTH (Marks 1976). Plasma insulin, proinsulin and
C-peptide levels are invariably low during hypoglycaemia
and do not respond normally to hyperglycaemia and other
insulinotrophic stimuli (Marks & Samols 1966). Plasma
GH levels are also low or very low during spontaneous
hypoglycaemia and this is thought to be due to a direct
suppressant action of IGF-II on GH secretion. Patients
with insulinomas may also have low plasma GH levels
during spontaneous hypoglycaemia, but in them it is
thought to be secondary to hypothalamic adaptation to
hypoglycaemia itself. Plasma glucagon levels are
invariably low in patients with NICTH and do not respond
normally to glucagonotrophic stimuli. This is unlike
adrenaline, which responds normally to hypoglycaemia
(Eastman et al. 1992), but too few patients have yet been
investigated to validate this conclusion.
Pathophysiology of NICTH
The initiating factor in NICTH appears to be primary
overproduction of predominantly ‘big’, but also of regular
IGF-II, by the tumour. Thereon the mechanism is
speculative.
Our own view is that primary overproduction of
IGF-II by the tumour leads to an increase in the
concentration of ‘free’ IGF-II in the plasma and a
secondary reduction in GH secretion. One consequence of
this is reduced production and secretion by the liver of
IGFBP - especially IGFBP-3, which is far and away the
most plentiful and ‘GH-dependent’ - as well as of
IGF-I itself. There is also a marked reduction in circulating
ALS.
The overall reduction in total IGFBPs - despite a
notable and probably ‘compensatory’ increase in IGFBP-2
- permits more of the IGF-II to circulate in the ‘free’ or
unbound form and consequently to a greater percentage of
it becoming available for binding to IGF receptors - as
well as to insulin receptors with which it cross-reacts
(Fradkin et al. 1989) - and which are present on liver,
muscle and all other peripheral tissues.
118
That this can be accomplished without a significant
increase in total plasma IGF-II was recently demonstrated
by Moller et al. (1996) in a case of NICTH in which ‘free’
IGF-II was increased 30-fold and ‘free’ IGF-I more than
4-fold despite an almost insignificant increase in total
plasma IGF-II and an 85% reduction in IGF-I.
Hypoglycaemia results from a combination of
decreased release of glucose into the blood by the liver and
grossly accelerated glucose utilisation by muscle and
other tissues. In NICTH this can reach prodigious
proportions, much greater, for example, than is ordinarily
encountered in insulinoma in which insulin resistance is a
common feature.
The 3- to 5-fold increase in insulin receptor number in
a patient with NICTH due to colon cancer described by
Stuart et al. (1986) was probably a result of ‘up-
regulation’ of insulin receptors in response to suppressed
plasma insulin levels that are so characteristic of NICTH.
This is, however, unlikely to be the explanation for the
increased insulin sensitivity that was held to account for
the occurrence of (insulin-induced) hypoglycaemia in an
insulin-dependent diabetic patient after he developed a
primary hepatoma (Barzilai et al. 1991) since in this
patient the sole source of insulin had been from a bottle.
Similarly, in a case described by Sturrock et al. (1997), in
which the IGF-II:IGF-I ratio was high as a result of a
pancreatic carcinoma, spontaneous hypoglycaemia
developed late in the course of her illness necessitating
discontinuation of insulin therapy.
Non-insulin-dependent glucose uptake by the tumour
itself, reflected in the hyperlactacidaemia often found in
these cases, may also make a small contribution to the
hypoglycaemia but it is not, as was once thought, a key
factor in its pathogenesis.
The role of depressed counterregulatory hormone
secretion presumably by IGF-II in NICTH is difficult to
assess. The hyperglycaemic response to exogenous
glucagon is often exaggerated unlike the insulinaemic
response, which is suppressed. Likewise exogenous GH
may alleviate hypoglycaemia in NICTH without affecting
plasma IGF-II levels (Teale & Marks 1998).
Unlike the counterregulatory polypeptide hormones,
both adrenaline and nor-adrenaline secretory mechanisms
are said to remain intact in patients with NICTH (Eastman
et al. 1992).
Associated conditions
Laurent et al. (1971) drew attention to the high
incidence (10%) of thyroid enlargement, with or without
thyrotoxicosis, in patients with NICTH and to its
association with acromegaly. Skin tags, excessive oiliness
of the skin and rhinophyma have also been observed
(Trivedi et al. 1995), but whether these features, which are

characteristically associated with over- rather than under-
production of GH, are related to, and caused by, over-
expression of IGF-II activity is unknown.
Coincidence of fibromas and insulinomas The co-
incident presence of a large mesenchymal tumour capable
of producing NICTH in patients harbouring an
insulinoma has been reported on several occasions both
before and after the advent of immunoassays capable of
differentiating between them (Bruno & Ober 1962,
Coskey & Tranquada 1964). Since neurofibromatosis has
also been described in association with other types of
APUDoma the occurrence may be more than coincidental
(Fung & Lam1995).
Hepatomas
Hypoglycaemia due to hepatoma was described soon after
that due to insulinoma and earlier than that from any other
type of non-islet cell tumour. Many of the large series of
NICTH due to primary hepatoma have emanated from the
far east where hepatomas are common (McFadzean &
Yeung 1956, 1969) but it is also disproportionately
common in the USA (Wu et al. 1988, Shapiro et al. 1990)
as well as in the UK.
McFadzean & Yeung (1969) distinguished two types
of hepatoma causing hypoglycaemia: type A tumours,
which are far the commonest, are rapidly growing, poorly
differentiated tumours that are associated with rapid
wasting and muscular weakness; type B tumours, which
are slow growing, are histologically well differentiated
and occur in patients who appear to be well nourished.
In patients with type B tumours the requirement for
glucose to prevent hypoglycaemia can be prodigious,
some patients requiring up to 1500 g glucose
intravenously per day. Type A patients, on the other hand,
require only very modest amounts of glucose to prevent
hypoglycaemia, which is usually a terminal event
associated with cachexia. These patients correspond to
those described later in this review as suffering from
metastatic cancer.
A somewhat similar distinction between two types of
hepatoma to that suggested by McFadzean & Yeung
(1969) was drawn by Wu et al. (1988) who noted that
patients with hypoglycaemia tended to have higher plasma
IGF-II and lower plasma IGF-I levels than those without
it.
In the hypoglycaemic patients a larger proportion of
the IGF-II was in the form of partially processed or ‘big
IGF-II’, their plasma IGF-II:IGF-I ratios were higher and
GH levels lower than in normal subjects or hepatoma
patients who did not become hypoglycaemic. Treatment
with GH reduced their glucose requirements but did not
completely relieve their hypoglycaemia.
Endocrine-Related Cancer (1998) 5 111-129
Adrenocortical tumours and pheocromocytomas
Steroid-secreting tumours
Adrenocortical tumours are a rare but well-recognised
cause of NICTH (Marks & Rose 1964, Laurent et al.
1971). The hypoglycaemia was, at one time, thought to be
a consequence of the secretion of abnormal steroids with
specific hypoglycaemic properties. This now seems
unlikely and, in two recent cases of our own, the chemical
pathology was no different from that of other IGF-II-
secreting tumours. Nevertheless it seems reasonable, until
proved otherwise, to assume that some of the unusual
steroids these tumours produce may, in some way,
contribute to the production of hypoglycaemia - possibly
by interfering with normal glucocorticoid production or
action.
Catecholamine-secreting tumours
Both reactive and fasting hypoglycaemia have been
described in association with malignant pheochromo-
cytoma (Gorden et al. 1981). The cause is unlikely to be
the same in each of these situations.
Severe reactive hypoglycaemia in patients with a
pheochromocytoma with remission following its surgical
removal has been described on a number of occasions. It
has also been reported as occurring for the first time
immediately following tumour excision (Hiramatsu et al.
1987, Paineau et al. 1988, Akiba et al. 1990, Chiang et al.
1993). The mechanism in such cases is unknown, but
seemingly connected to the presence of excessive
catecholamine action either stimulating insulin secretion
or through a post-receptor effect.
Episodic, probably stimulative, hyperinsulinism has
been demonstrated in at least one patient with reactive
hypoglycaemia due to pheochromocytoma and in which
fasting hypoglycaemia was not a feature (Hiramatsu et al.
1987).
Very high plasma insulin levels were found in two
out of ten patients (12 480 pmol/l and 1144 pmol/l
respectively) who developed profound hypoglycaemia
immediately after excision of a pheochromocytoma.
Plasma insulin levels were also abnormally high, relative
to their prevailing blood glucose levels, in the remaining
eight patients (Akiba et al. 1990).
In two patients with fasting hypoglycaemia due to
pheochromocytoma studied by Gorden et al. (1981)
plasma levels of ‘IGF-like materials’ were described as
modestly raised. It seems probable, therefore, that
pheochromocytomas, like most other tumours, can
produce fasting hypoglycaemia by secreting ‘big’ IGF-II.
Lymphomas, myeloma and leukaemia
Hypoglycaemia occurs as an unusual phenomenon in
patients with various types of lymphoma (Wanebo et al.
119
Marks and Teale: Tumours producing hypoglycaemia
1966) usually as a terminal event. Very occasionally it is
the presenting feature (Braund et al. 1987, Walters et al.
1987). It is also a well-recognised complication of
leukaemia, especially in animals, but also in rare human
cases (Hilali et al. 1984, Makino et al. 1985). Sometimes
the hypoglycaemia in leukaemia is artefactual due to post-
collection disappearance of glucose from the blood
through glycolysis (Astles et al. 1995). But more often it
is real. Increased plasma lactate levels are sometimes
observed in patients with hypoglycaemia caused by
leukaemia (Makino et al. 1985) just as in patients with
hypoglycaemia and lymphoma or fibrosarcoma (Laurent
et al. 1971, Marks & Rose 1981, Durig et al. 1996).
Insulin-receptor antibodies
Evidence for the existence and pathogenic role of
stimulatory insulin-receptor antibodies in hypoglycaemia
caused by Hodgkin’s disease has been produced by two
groups of authors (Braund et al. 1987, Walters et al. 1987).
IGF-II was not measured in either of the cases reported but
in one of them IGF-I, measured by bioassay, was said to
be just below the reference range. Plasma GH measured
during spontaneous hypoglycaemia was 12.5 U/l, which is
higher than is customarily found in patients with IGF-II-
secreting tumours and consistent with the view that some
mechanism other than excessive IGF-II production was
involved.
Despite the passage of more than 10 years since these
two reports were published, no other cases of lymphoma-
induced hypoglycaemia in which antibody receptors were
implicated have, to our knowledge, been described.
In this context it is of interest that in the seven or so
patients with hypoglycaemia and lymphoma investigated
by Gorden et al. (1981) the concentration of plasma ‘IGF-
like material’, measured by a radioreceptor binding assay
and equating largely with multiplication-stimulating
activity (now known to be identical with IGF-II) was
either low, or very low compared both with a normal
reference population and patients suffering from other
types of tumour-producing NICTH. In a further case of
hypoglycaemia due to Hodgkin’s disease investigated by
us, plasma insulin, C-peptide and proinsulin levels were
all suppressed, and IGF-I, IGF-II and IGF-II:IGF-I ratio
were normal. Neither insulin autoantibodies nor insulin-
receptor antibodies were measured.
The role of stimulatory, insulin-receptor antibodies in
the pathogenesis of NICTH by neoplasms other than
lymphomas has been established by various workers
(Rochet et al. 1989, Taylor et al. 1989). It is reasonable to
assume, therefore, that they are involved in at least some
of the rare cases of hypoglycaemia secondary to
lymphoma.
120
Myeloma and primary monoclonal gammopathy
An association between hypoglycaemia and myeloma was
first reported 25 years ago and that with monoclonal
gammopathy more recently (Russell-Jones et al. 1990,
Arnqvist et al. 1993). The cause of the hypoglycaemia is
not, however, the same in all cases.
In the case of myeloma reported by Snowden et al.
(1994) an increased proportion of the IGF-II present in the
plasma was of the ‘big’ variety and the IGF-II:IGF-I
ratio was abnormally high. Onset of spontaneous hypo-
glycaemia was marked by reversal of the patient’s long-
standing need for insulin to control her NIDDM which
began 15 years earlier. Anti-insulin antibodies were not
detected in her serum.
In the cases of monoclonal gammopathy described by
Wasada et al. (1989) and by Redmon et al. (1992) hypo-
glycaemia appears to have been insulin mediated and due
to its delayed release from insulin-specific mono-clonal
autoantibodies in a manner resembling that seen in insulin
autoimmune syndrome (IAS). This was true also for the
case of myeloma described by Merlo et al. (1992) whose
monoclonal (pro)insulin binding immunoglobulin was
produced by the neoplastic cells.
In this form of NICTH, endogenous insulin - released
in response to ingestion of a meal but rendered temporarily
ineffective by binding to the autoantibody - is slowly
released during the post-absorptive period producing an
apparently inappropriate hyperinsulinaemia. But whereas
IAS is self-limiting, and part of a more general auto-
immune disorder, monoclonal gammopathy appears to be
a neoplastic disorder in which the para-protein produced
has a specific affinity for insulin or proinsulin.
Hypoglycaemia and metastatic cancer
Hypoglycaemia may occur as a terminal or agonal event
in patients dying from metastatic disease (Younus et al.
1977, Marks & Rose 1981). It is assumed to have a
different cause from hypoglycaemia occurring as a
dominant and comparatively ‘early’ feature (Marks et al.
1965, MacDougall et al. 1986) in rare cancer patients
(especially those with primary tumours of the lung,
stomach, pancreas, caecum and colon) in whom plasma
insulin and GH levels are depressed and the IGF-II:IGF-I
ratio unusually high.
Hypoglycaemia occurring in patients with over-
whelming metastatic disease was formerly attributed to
‘liver failure’ secondary to its destruction by the tumour.
An equally plausible explanation, as yet unconfirmed
(Fitzpatrick et al. 1995), is that it is mediated by various
cytokines including TNF-α, and interleukin-1 and -6
which are known to produce profound hypoglycaemia in
animals (Mahony & Tisdale 1990, Clark et al. 1992,
Battelino et al. 1996) and probably also in human beings.

They are currently thought to be involved in the
production of hypoglycaemia in septicaemia and over-
whelming infections with malaria parasites (Krishna et al.
1994, Rockett et al. 1994) possibly by stimulating insulin
release (Elased et al. 1996). The situation is, however, still
far from clear.
Diagnosis
The diagnosis of NICTH commences with its recognition
as the cause of the patient’s neuroglycopenic symptoms.
This rarely poses any problem providing the possibility
is considered since fasting hypoglycaemia, once it has
developed, is relentless and easily confirmed by
measurement of the blood glucose concentration. It
proceeds through differential diagnosis (Marks & Teale
1996) to elucidation of the cause upon which rational and
specific treatment depends. Only the differential diagnosis
of fasting hypoglycaemia will be considered here.
Insulinoma
Diagnosis depends upon establishing inappropriate insulin
(or rarely proinsulin) secretion during hypoglycaemia
provoked by fasting or rigorous exercise. Paradoxically
the measurement of C-peptide in plasma is a more
sensitive indicator of inappropriate secretion than the
measurement of insulin itself (Hattori et al. 1994) as the
concentration of C-peptide may be raised in peripheral
blood even when that of insulin is suppressed. This
anomaly, which has also been reported in children with
PHHI (nesidioblastosis) is probably due to a greater than
normal fractional extraction of insulin by the liver. This
leaves the C-peptide unaffected and consequently free to
appear in peripheral blood at an inappropriately high level.
Inappropriate C-peptidaemia (i.e. >100 pmol/l) and
hyperinsulinaemia (i.e. >30 pmol/l) in the presence of
fasting hypoglycaemia (blood glucose <2.2mmol/l) is
virtually pathognomonic for insulinoma or PHHI except
when there is accidental or deliberate overdosing with
sulphonylurea or other insulinotrophic drugs. Hyper
C-peptidaemia may also, very rarely, appear to occur in
those patients with IAS in whom the autoantibodies (to
endogenous proinsulin) bind C-peptide as well as insulin
(Russell-Jones et al. 1990, Merlo et al. 1992). The high
titre of autoantibodies that are invariably present in AIS
makes distinction from insulinoma comparatively easy
providing the possibility is considered and excluded.
Hyperproinsulinaemia (>20 pmol/l) is not diagnostic
of any single condition. Nevertheless the presence in
plasma of proinsulin at a concentration >50 pmol/l in a
patient with a history suggestive of spontaneous fasting
hypoglycaemia is presumptive evidence of an insulin-
oma, whilst its absence, or presence at concentration
<20 pmol/l, renders the diagnosis unlikely though not
Endocrine-Related Cancer (1998) 5 111-129
impossible (Cohen et al. 1986, Hampton et al. 1988,
Nauck et al. 1990, Tasaka et al. 1993, Gorden et al. 1995).
The ‘gold standard’ for the diagnosis of hyper-
insulinism remains Whipple’s triad (which consists of
symptoms due to hypoglycaemia demonstrated during
fasting and which are relieved exclusively by intravenous
glucose) plus the presence of inappropriately high
peripheral plasma IRI (>30 pmol/l), C-peptide
(>300 pmol/l) and proinsulin (>20 pmol/l) levels in the
presence of fasting (or exercise-induced) hypoglycaemia.
When these conditions are met the diagnosis of
insulinoma can be made preoperatively with as much
certainty as is realistically possible. If they are not, some
other explanation for the patient’s hypoglycaemia should
be sought.
Although it is fashionable to do so, calculation of a
‘glucose:insulin ratio’ does not determine the
appropriateness of insulinaemia (C-peptidaemia) during
hypoglycaemic episodes. The ratio is not only method-
dependent but also scientifically spurious and should
never be used (Hattori et al. 1994). Instead a reference
range for plasma insulin, proinsulin and C-peptide
concentrations in the presence of hypoglycaemia
(<2.2 mmol/l in children and young adults; <3.0 mmol/l in
subjects over 60 years of age) should be established
against which those obtained in patients with hypo-
glycaemia can be assessed (Marks & Teale 1996).
C-peptide suppression tests, in which plasma
C-peptide levels act as a surrogate for endogenous insulin
production, have enjoyed considerable popularity ever
since the discovery of C-peptide and assays for measuring
it became available. They are, however, no more than
screening tests capable of reducing to a manageable size
the number of patients in whom the rigorous exercise test
or admission to hospital for prolonged fasting are in order
since the incidence of both false positives and false
negatives is high. Their main value is for distinguishing
recurrence of hyperinsulinism following successful
removal of one insulinoma when others were also present
but missed at operation from a recurrence of symptoms
from psychogenic origin (Sata et al. 1995).
None of the many stimulation tests, e.g. tolbutamide,
glucagon, L-leucine, intravenous or oral glucose tolerance
tests, that have been advocated over the past half-century
adds materially to the diagnosis (Marks & Rose 1981,
Marks 1989). Their use should be discontinued except
where facilities for C-peptide, insulin and proinsulin
measurement are not available.
Localisation
Almost every localisation technique has been
advocated as a preoperative aid to surgical removal of
insulinomas but none withstands critical survey (Table 3).
This includes somatostatin receptor scintigraphy (not
121
Marks and Teale: Tumours producing hypoglycaemia

specifically shown in Table 3) the latest fashionable
localising technique and which, though very good for
localising gastrinomas (86% positive), is no better than
any other for localising insulinomas preoperatively (Pitre
et al. 1996, Zimmer et al. 1996).
It is quite clear that no preoperative localising
technique - including endoscopic ultrasonography,
undoubtedly the most sensitive currently available - is
sufficiently reliable to make a diagnosis of insulinoma if
the biochemical findings are negative nor to justify delay
in operating if they fail to localise a tumour proved
biochemically to be present (Daggett et al. 1981, Owens
et al. 1995, Marks & Teale 1996).
In contrast to preoperative localisation, intraoperative
ultrasonography has proved to be extremely useful by
enabling even experienced surgeons to localise tumours
that their unaided fingers have missed.
NICTH
The diagnosis of NICTH is suggested by the association
of severe fasting hypoglycaemia with low plasma NEFA
and β-hydroxybutyrate levels and undetectable plasma
insulin, C-peptide and proinsulin levels, and low GH
levels. It is established by the presence of an abnormally
high IGF-II:IGF-I molar ratio (>10:1; normal ratio circa
3:1).
An increased concentration of the E-domain of IGF-II
is also typically found but is not, in our experience, as
predictive as the IGF-II:IGF-I ratio in arriving at a
diagnosis (Table 2). Measurements of IGF-II alone should
not be relied upon for diagnosis and may be misleading.
Despite earlier optimism, measurements of IGFBPs
have not proved clinically useful as yet but have helped to
throw light on the mechanism by which tumours produce
NICTH. We expect, however, that measurements of ‘free’
IGF-II and IGF-I will be so when they become more
readily available and have been properly evaluated
(Chung & Henry 1996, Moller et al. 1996).
Localisation of the IGF-II-secreting tumour producing
NICTH rarely gives rise to difficulty once the diagnosis
has been established. Occasionally, however, when the
tumour is unusually small or obscured by the pelvis, for
example, it may come to light only after very thorough
investigation using modern imaging techniques. Since

Table 3 Accuracy of localisation of insulinomas for various imaging techniques: literature survey.

Number positive

Preoperative Intraoperative Endoscopic

Authors Angiography CAT MRI Catheter ultrasound ultrasound ultrasonograph

Katz et al. (1986) 6/15 1/5 11/13 1/6

Galiber et al. (1988) 14/26 7/23 17/28 21/28
Bottger et al. (1990) 20/30 3/14 10/13 12/16
Mårtensson et al. (1990) 6/15 2/5 11/11 1/6
Doherty et al. (1991) 9/26 4/23 2/8 17/22 6/23
Vinik et al. (1991) 13/32 4/25 0/4 29/29
Lightdale et al. (1991) 0/4 0/5 4/4
Liessi et al. (1992) 2/7 6/7 5/7 6/7 6/7
Pasieka et al. (1992) 18/41 9/35 34/36
Rosch et al. (1992) 32/39*
Thompson et al. (1993) 8/15 2/13 0/1 6/13 9/11
Menegaux et al. (1993) 2/2
Thompson et al. (1994) 7/10
Chang et al. (1994) 5/10 3/8 4/4 1/8 1/1 1/1
Knobel et al. (1995) 0/1 0/1 1/1 0/1 1/1
Moore et al. (1995) 1/3 0/3 3/4 0/3 4/4
Pitre et al. (1996) 18/18 10/11
Schumacher et al. (1996) 8/14
Zimmer et al. (1996) 3/14 1/14 1/14 13/14

Total number positive 97/224 43/181 11/39 117/129 38/101 71/85 65/82

Percent positive 46 24 28 91 38 84 79

*Multicentre trial in endocrine pancreatic tumours: cases include, but not exclusively, insulinomas.
CAT = computerised axial tomography; MRI = magnetic resonance imaging.

122

some of these tumours contain somatostatin receptors
which can be visualised using 111In-octreotide imaging
(Perros et al. 1996) the possibility of using this technique
to localise a hidden tumour or see whether it has
metastasised should be borne in mind in case of difficulty.
Treatment
Insulinoma
Treatment of tumour-induced hypoglycaemia, as in all
types of hypoglycaemia, can be divided into two phases;
palliation of an acute episode with intravenous glucose or
intramuscular glucagon and definitive, specific treatment
directed at its ultimate cause. Intramuscular glucagon will
invariably relieve fasting hypoglycaemia in patients with
insulinoma, but, because of its propensity to stimulate
endogenous insulin secretion, may cause profound
rebound hypoglycaemia unless food is given as soon as the
swallowing reflex is re-established.
Benign insulin-secreting tumours are best treated
surgically unless there are overwhelming clinical reasons
for not doing so. Removal of an adenoma is extremely
effective although some long-term follow-up series, but
not all, suggest that there may be an increased incidence
of neuropsychiatric aberration, peptic ulcer disease and
NIDDM in those who survived the operation (Dunn 1971,
Galbut & Markowitz 1980). Operative mortality is now
down to less than 2% in the hands of experienced
pancreatic surgeons and virtually confined to those over
70 years of age. In them palliative therapy with
diazoxide/chlorothiazide may be worth using since it is
often remarkably effective and produces few if any
unpleasant side-effects.
Malignant insulin-secreting tumours are often very
slow growing and providing hypoglycaemia can be
avoided many years of good life may follow initiation of
antihypoglycaemic therapy and/or ablation of the primary
tumour and as many of the metastases as possible (Marks
& Rose 1981, Krentz et al. 1996).
Palliative drug therapy with a combination of
diazoxide and a diuretic thiazide is often effective in
relieving the hypoglycaemia if this persists after surgery.
Long-acting and synthetic somatostatin preparations have
been used with varying success in palliation. They may
either improve hypoglycaemia, by inhibiting insulin
secretion by the tumour, or worsen it by inhibiting
secretion of the counterregulatory hormones - glucagon
and somatotropin - without affecting insulin production
itself (Alberts & Falkson 1988, Stehouwer et al. 1989,
Gama et al. 1995, Meeking et al. 1997). Since improve-
ment is commoner than the reverse a therapeutic trial is
probably worthwhile.
Tumoricidal therapy with streptozotocin and other
antineoplastic agents of which 5-fluorouracil and
Endocrine-Related Cancer (1998) 5 111-129
adriamycin are the most valuable is sometimes beneficial.
Favourable results have also been achieved with hepatic
artery thromboembolisation, which may produce re-
missions lasting a year or more, and cryogenic destruction
of metastases. These manoeuvres can be repeated if and
when relapse occurs (Krentz et al. 1996).
NICTH
Patients with NICTH may undergo complete remission
following surgical removal of a benign or locally
invasive tumour. Even partial removal often reduces or
abolishes hypoglycaemia although it usually recurs if and
when the tumour re-establishes itself.
Complete removal of the tumour abolishes any
tendency to fasting hypoglycaemia. It also produces a rise
in IGF-I and restores the blood glucose and plasma insulin
responses to food to normal. Immunoreactive IGF-II
levels fall - even if they were not abnormally high (in
absolute terms) preoperatively - and the IGF-II:IGF-I ratio
returns to normal (i.e. <10). The plasma GH and glucagon
responses to hypoglycaemia and other appropriate stimuli
are restored and the plasma IGFBP-3 concentration rises
to a level which, in molar terms, exceeds the combined
concentration of IGF-I and IGF-II (Perros et al. 1996,
Teale & Marks 1998).
Contrary to expectations, diazoxide-chlorothiazide
treatment often alleviates NICTH - though less
predictably than with insulinomas. Both hGH and
prednisolone have been used sometimes with great and
seemingly specific effect (Mitchell et al. 1965, Khaleeli et
al. 1991, Teale et al. 1992, Hunter et al. 1994, Agus et al.
1995, Katz et al. 1996, Perros et al. 1996). In the case of
hGH the effect was thought to be mediated by increasing
the production of IGFBP-3 but this now seems unlikely
(Teale & Marks 1998). In the past the failure to observe a
sustained beneficial effect of exogenous hGH after an
early favourable response was reported and may have
been due to the relatively low doses used or to the
development of GH-specific antibodies (Samaan et al.
1990).
Large doses of prednisolone may bring about a
remarkable improvement in both glucose homeostasis and
other biochemical markers of NICTH (Baxter et al. 1995,
Perros et al. 1996). In particular there is a rise in plasma
ALS and IGFBP-3 levels (Teale & Marks 1998) and a fall
in ‘big IGF-II’ (Baxter et al. 1995). The insulinaemic
response to oral glucose is restored and there may be
restoration of the endogenous GH response to appropriate
stimuli including hypoglycaemia.
Treatment with other anti-inflammatory and immuno-
suppressive agents is worthy of trial in cases where insulin
antibodies are implicated. It may produce prolonged
123
Marks and Teale: Tumours producing hypoglycaemia
remissions (Taylor et al. 1989) as may specific antitumour
chemotherapy.
References
Agus MSD, Levit Katz LE, Satin-Smith M, Meadows AT, Hintz
RL & Cohen P 1995 Non-islet-cell tumor associated with
hypoglycemia in a child: successful long-term therapy with
growth hormone. Journal of Pediatrics 127 403-407.
Akiba M, Kodama T, Ito Y, Obara T & Fujimoto Y 1990
Hypoglycemia induced by excessive rebound secretion of
insulin after removal of pheochromocytoma. World Journal of
Surgery 14 316-324.
Alberts AS & Falkson G 1988 Rapid reversal of life-threatening
hypoglycaemia with a somatostatin analogue (octreotide): a
case report. Journal of South African Medicine 74 75-76.
Alsever RN, Roberts JP, Gerber JG, Mako ME & Rubenstein AH
1975 Insulinoma with low circulating insulin levels: the
diagnostic value of proinsulin measurements. Annals of
Internal Medicine 82 347-350.
Arnqvist HJ, Halban PA, Mathiesen UL, Zahnd G & von Schenck
H 1993 Hypoglycaemia caused by atypical insulin antibodies
in a patient with benign monoclonal gammopathy. Journal of
Internal Medicine 234 421-427.
Astles JR, Petros WP, Peters WP & Sedor FA 1995 Artifactual
hypoglycemia associated with hematopoietic cytokines.
Archives of Pathological Laboratory Medicine 119 713-716.
Atkinson AB, Hadden DR, Kennedy TL, Montgomery DAD,
McIlrath E & Weaver JA 1981 Insulinomas in Northern
Ireland between 1960 and 1980. Ulster Medical Journal 50
21-32.
Axelrod L & Ron D 1988 Insulin-like growth factor II and the
riddle of tumor-induced hypoglycemia. New England Journal
of Medicine 319 1477-1479.
Baltic V, Levanat S, Petek M, Bratic-Mikes V, Pavelic K & Vuk-
Pavlovic S 1985 Elevated levels of substances
immunologically cross-reactive with insulin in blood of
patients with malignant and nonmalignant pulmonary tissue
proliferation. Oncology 42 174-178.
Barzilai N, Cohen P, Barillan R, McIntyre N & Karnieli E 1991
Case report - increased insulin sensitivity in tumor
hypoglycemia in a diabetic patient - glucose metabolism in
tumor hypoglycemia. American Journal of the Medical
Sciences 302 229-234.
Battelino T, Goto M & Zeller WP 1996 Dexamethasone
attenuates hypoglycemia in ten day old rats treated with TNF
alpha. Research Communications in Molecular Pathology and
Pharmacology 92 149-154.
Baxter RC 1990 Radioimmunoassay for insulin-like growth
factor (IGF)-II: interference by pure IGF-binding proteins.
Journal of Immunoassay 11 445-458.
Baxter RC 1996 The role of insulin-like growth factors and their
binding proteins in tumor hypoglycemia. Hormone Research
46 195-201.
Baxter RC & Martin JL 1989 Binding proteins for the insulin-like
growth factors: structure, regulation and function. Progress in
Growth Factor Research 1 49-68.
124
Baxter RC, Holman SR, Corbould A, Stranks S, Ho PJ & Braund
W 1995 Regulation of the insulin-like growth factors and their
binding proteins by glucocorticoid and growth hormone in
nonislet cell tumor hypoglycemia. Journal of Clinical
Endocrinology and Metabolism 80 2700-2708.
Berger M, Bordi C, Cuppers H-J, Berchthold P & Gries FA 1983
Functional and morphologic characterization of human
insulinomas. Diabetes 32 921-931.
Bottger TC, Weber W, Beyer J & Junginger T 1990 Value of
tumor localization in patients with insulinoma. World Journal
of Surgery 14 107-112.
Braund WJ, Naylor BA, Williamson DH, Bailey ID, Clark A,
Chapel HM & Turner RC 1987 Autoimmunity to insulin
receptor and hypoglycaemia in patients with Hodgkin’s
disease. Lancet i 237-240.
Bruno MS & Ober WB 1962 Hypoglycemia and intrathoracic
mass. New York State Journal of Medicine 62 2005-2008.
Bugalho MJ, Roque L, Sobrinho LG, Hoog A, Nunes JF,
Almeida JM, Leitao CN & Santos JR 1994 Calcitonin-
producing insulinoma: clinical, immunocytochemical and
cytogenetical study. Clinical Endocrinology 41 257-260.
Chang HY, Huang HS, Lin JD, Huang BY, Huang MJ & Jeng LB
1994 Insulinoma - clinical experience in ten cases. Chang
Keng I Hsueh 17 28-38.
Chiang WL, Tsai MH, Lieu CS, Yang SF & Lin PC 1993
Postoperative hypoglycemia after pheochromocytoma
resection (Chinese). Ma-Tsui-Hsueh-Tsa-Chi 31 273-276.
Chung J & Henry RR 1996 Mechanisms of tumor-induced
hypoglycemia with intraabdominal hemangiopericytoma.
Journal of Clinical Endocrinology and Metabolism 81 919-
925.
Clark IA, Gray KM, Rockett EJ, Cowden WB, Rockett KA,
Ferrante A & Aggarwal BB 1992 Increased lymphotoxin in
human malarial serum, and the ability of this cytokine to
increase plasma interleukin-6 and cause hypoglycaemia in
mice: implications for malarial pathology. Transactions of the
Royal Society of Tropical Medicine and Hygiene 86 602-607.
Clemmons DR 1989 Structural and functional analysis of insulin-
like growth factors. British Medical Bulletin 45 465-480.
Cohen RM & Camus F 1988 Update on insulinomas or the case
of the missing (pro)insulinoma. Diabetes Care 11 506-507.
Cohen RM, Given BD, Licinio-Paixao J, Provow SA, Rue PA,
Frank BH, Root MA, Polonsky KS, Tager HS & Rubenstein
AH 1986 Proinsulin radioimmunoassay in the evaluation of
insulinomas and familial hyperproinsulinemia. Metabolism 35
1137-1146.
Coskey RL & Tranquada RE 1964 Insulinoma and multiple
neurofibromatosis. Metabolism 13 312-324.
Creutzfeldt W, Arnold R, Creutzfeldt C, Deuticke V, Frerichs H
& Track NS 1973 Biochemical and morphological
investigations of 30 human insulinomas. Diabetologia 9 217-
231.
Crowley PF, Slavin JL & Rode-J 1996 Massive amyloid
deposition in pancreatic vipoma: a case report. Pathology 28
377-379.
Cullen RM & Ong CE 1987 Insulinoma in Auckland 1970-1985.
New Zealand Medical Journal 100 560-562.

Daggett PR, Goodburn EA, Kurtz AB, Le Quesne LP, Morris DV,
Nabarro JDN & Raphael MJ 1981 Is preoperative localisation
of insulinomas necessary? Lancet i 483-486.
Daughaday WH 1989 Hypoglycemia in patients with non-islet
cell tumors. Endocrinology and Metabolism Clinics of North
America 18 91-98.
Daughaday WH 1997 Insulin-like growth factors and tumour
hypoglycaemia. In Clinical Endocrine Oncology, pp 496-501.
Eds R Sheaves, PJ Jenkins & JAH Wass. Oxford: Blackwell
Science.
Daughaday WH & Kapadia M 1989 Significance of abnormal
serum binding of insulin-like growth factor II in the
development of hypoglycemia in patients with non-islet-cell
tumors. Proceedings of the National Academy of Sciences of
the USA 86 6778-6782.
Daughaday WH & Trivedi B 1992 Measurement of derivatives
of proinsulin-like growth factor II in serum by a radio-
immunoassay directed against the E-domain in normal
subjects and patients with nonislet cell tumor hypoglycemia.
Journal of Clinical Endocrinology and Metabolism 75 110-
115.
Daughaday WH, Emanuele MA, Brooks MH, Barbato AL,
Kapadia M & Rotwein P 1988 Synthesis and secretion of
insulin-like growth factor II by a leiomyosarcoma with
associated hypoglycemia. New England Journal of Medicine
319 1434-1440.
Daughaday WH, Wu J-C, Lee S-D & Kapadia M 1990 Abnormal
processing of pro-IGF-II in patients with hepatoma and in
some hepatitis B virus antibody-positive asymptomatic
individuals. Journal of Laboratory and Clinical Medicine 116
555-562.
Diarmuid S, O’Riordain MB, O’Brien T, van Heerden JA,
Service FJ & Grant CS 1994 Surgical management of
insulinoma associated with multiple endocrine neoplasia
Type I. World Journal of Surgery 18 488-494.
Doherty GM, Doppman JL, Shawker TH, Miller DL, Eastman
RC, Gorden P & Norton JA 1991 Results of a prospective
strategy to diagnose, localize and resect insulinomas. Surgery
110 989-996.
Doppman JL, Chang R, Fraker DL, Norton JA, Alexander HR,
Miller DL, Collier E, Skarulis MC & Gorden P 1995
Localization of insulinomas in regions of the pancreas by
intra-arterial stimulation with calcium. Annals of Internal
Medicine 123 269-273.
Dunn DC 1971 Diabetes after removal of insulin tumours of the
pancreas: a long term follow up survey of 11 patients. British
Medical Journal 2 84-86.
Durig J, Fiedler W, de Wit M, Steffen M & Hossfeld DK 1996
Lactic acidosis and hypoglycemia in a patient with high-grade
non-Hodgkin’s lymphoma and elevated circulating TNF-
alpha. Annals of Hematology 72 97-99.
Eastman RC, Carson RE, Orloff DG, Cochran CS, Perdue JF,
Rechler MM, Lanau F, Roberts CT, Shapiro J & Roth J 1992
Glucose utilization in a patient with hepatoma and hypo-
glycaemia: assessment by a positron emission tomography.
Journal of Clinical Investigation 89 1958-1963.
Endocrine-Related Cancer (1998) 5 111-129
Elased KM, Taverne J & Playfair JH 1996 Malaria, blood
glucose, and the role of tumour necrosis factor (TNF) in mice.
Clinical and Experimental Immunology 105 443-449.
England DM, Hochholzer L & McCarthy MJ 1989 Localized
benign and malignant fibrous tumors of the pleura. A
clinicopathologic review of 223 cases. American Journal of
Surgical Pathology 13 640-658.
Eriksson B, Oberg K & Skogseid B 1989 Neuroendocrine
pancreatic tumors. Clinical findings in a prospective study of
84 patients. Acta Oncologia 28 373-377.
Fajans SS & Vinik AI 1989 Insulin-producing islet cell tumors.
Endocrinology and Metabolism Clinics of North America 18
45-74.
Ferguson JM & Flinn J 1995 Intracranial meningioma with
hepatic metastases and hypoglycaemia treated by selective
hepatic arterial chemo-embolization. Australasian Radiology
39 97-99.
Field JB, Keen H, Johnson P & Herring B 1963 Insulin-like
activity of nonpancreatic tumors associated with
hypoglycemia. Journal of Clinical Endocrinology and
Metabolism 23 1229-1236.
Fitzpatrick DR, Peroni DJ & Bielefeldt-Ohmann H 1995 The role
of growth factors and cytokines in the tumorigenesis and
immunobiology of malignant mesothelioma. American
Journal of Respiratory and Cellular Molecular Biology 12
455-460.
Fonseca V, Ames D & Ginsburg J 1989 Hypoglycaemia for 26
years due to an insulinoma. Journal of the Royal Society of
Medicine 82 437-438.
Fradkin JE, Eastman RC, Lesniak MA & Roth J 1989 Specificity
spillover at the hormone receptor - exploring its role in human
disease. New England Journal of Medicine 320 640-645.
Froesch ER & Zapf J 1985 Insulin-like growth factors and
insulin: comparative aspects. Diabetologia 28 485-493.
Froesch ER, Burgi H, Ramseier EB, Bally P & Labhart A 1963
Antibody-suppressible and nonsuppressible insulin-like
activities in human serum and their physiologic significance -
an insulin assay with adipose tissue of increased precision and
specificity. Journal of Clinical Investigation 42 1816-1834.
Fung JWK & Lam KSL 1995 Neurofibromatosis and insulinoma.
Postgraduate Medical Journal 71 485-486.
Galbut DL & Markowitz AM 1980 Insulinoma: diagnosis,
surgical management and long-term follow-up. American
Journal of Surgery 139 682-690.
Galiber AK, Reading CC, Charboneau JW, Sheedy PF, James
EM, Gorman B, Grant CS, van Heerden JA & Telander RL
1988 Localization of pancreatic insulinomas: comparison of
pre- and intra-operative US with CT and angiography.
Radiology 166 405-408.
Gama R, Marks V, Wright J & Teale JD 1995 Octreotide
exacerbated fasting hypoglycaemia in a patient with a
proinsulinoma: the glucostatic importance of pancreatic
glucagon. Clinical Endocrinology 43 117-120.
Gorden P, Hendricks CM, Kahn CR, Megyesi K & Roth J 1981
Hypoglycemia associated with non-islet-cell tumor and
insulin-like growth factors. A study of the tumor types. New
England Journal of Medicine 305 1452-1455.
125
Marks and Teale: Tumours producing hypoglycaemia
Gorden P, Skarulis MC, Roach P, Comi RJ, Fraker DL, Norton
JA, Alexander HR & Doppman JL 1995 Extensive personal
experience: plasma proinsulin-like component in insulinoma:
a 25-year experience. Journal of Clinical Endocrinology and
Metabolism 80 2884-2887.
Graeme-Cook F, Nardi G & Compton CC 1990 Immuno-
cytochemical staining for human chorionic gonadotropin
subunits does not predict malignancy in insulinomas.
American Journal of Clinical Pathology 93 273-276.
Grunberger G 1993 Insulin resistance in a case of coexisting
insulinoma and type 2 diabetes. Acta Diabetologica 30 243-
240.
Hale PJ, Djurup R, Baddeley RM & Nattrass M 1991 Insulin and
proinsulin concentrations in patients with insulinoma before
and after surgical treatment. Diabetes and Nutritional
Metabolism 4 113-116.
Hampton S, Beyzavi K, Teale D & Marks V 1988 A direct assay
for proinsulin in plasma and its application in hypoglycemia.
Clinical Endocrinology 29 9-16.
Hattori N, Ishihara T, Saiki Y, Moridera K, Hino M, Ikekubo K
& Kurahachi H 1994 Insulinoma with normal plasma insulin
concentrations and insulin/glucose ratios during hypo-
glycemic episodes. Internal Medicine 33 813-816.
Hayashi M, Floyd JC Jr, Pek S & Fajans SS 1977 Insulin,
proinsulin, glucagon and gastrin in pancreatic tumors and in
plasma of patients with organic hyperinsulinism. Journal of
Clinical Endocrinology and Metabolism 44 681-694.
Heik SCW, Kloppel G, Krone W, Iben G, Priebe K & Kuhnau J
1988 Hypoglykamie durch insulinom bei diabetes mellitus.
Deutsche Medizinische Wochenschrift 113 1714-1717.
Hilali MM (Al), Majer RV & Penney O 1984 Hypoglycaemia in
acute myelomonoblastic leukaemia: report of two cases and
review of published work. British Medical Journal 289 1443-
1444.
Hiramatsu K, Takahashi K, Kanemoto N & Arimori S 1987 A
case of pheochromocytoma with transient hyperinsulinemia
and reactive hypoglycemia. Japanese Journal of Medicine 26
88-90.
Hunter SJ, Daughaday WH, Callender ME, McKnight JA,
McIlrath EM, Teale JD & Atkinson AB 1994 A case of
hepatoma associated with hypoglycaemia and overproduction
of IGF-II (E-21): beneficial effects of treatment with growth
hormone and intrahepatic adriamycin. Clinical Endocrinology
41 397-401.
Jacob A, Hauri Ch & Froesch ER 1968 Nonsuppressible insulin-
like activity in human serum III: differentiation of two distinct
molecules with nonsuppressible ILA. Journal of Clinical
Investigation 47 2678-2688.
Kane LA, Grant CS, Nippoldt TB & Service FJ 1993 Insulinoma
in a patient with NIDDM. Diabetes Care 16 1298-1300.
Katz LB, Aufses AH Jr, Rayfield E & Mitty H 1986 Preoperative
localization and intraoperative glucose monitoring in the
management of patients with pancreatic insulinoma. Surgery
Gynecology and Obstetrics 163 509-512.
Katz LE, Liu F, Baker B, Agus MS, Nunn SE, Hintz RL & Cohen
P 1996 The effect of growth hormone treatment on the insulin-
like growth factor axis in a child with non-islet cell tumor
126
hypoglycemia. Journal of Clinical Endocrinology and
Metabolism 81 1141-1146.
Kavlie H & White TT 1972 Pancreatic islet beta-cell tumours and
hyperplasia. Annals of Surgery 175 326-330.
Khaleeli A, Perumainar M, Spedding AV, Teale JD & Marks V
1991 Treatment of tumour induced hypoglycaemia with
human growth hormone. Journal of the Royal Society of
Medicine 85 303.
Knobel B, Rosman P & Czerniak A 1995 Endoscopic
ultrasonography for the preoperative localization of
insulinoma - a new, useful modality. Israeli Journal of
Medical Science 31 365-367.
Krentz AJ, Mayer D, Ollilff SP, Bailey CJ, Libman LJ & Nattrass
M 1996 Cryotherapy of hepatic metastases and regional
perfusion with low-dose streptozotocin in the management of
metastatic malignant insulinoma. Endocrine-Related Cancer
3 341-345.
de Kreutzenberg SV, Riccio A, Dorella M, Avogaro A,
Marescotti MC, Tiengo A & del Prato S 1994 Surgical
removal of insulinoma restores glucose recovery from
hypoglycaemia but does not normalize insulin action.
European Journal of Clinical Investigation 25 360-367.
Krishna S, Waller DW, ter-Kuile F, Kwiatkowski D, Crawley J,
Craddock CF, Nosten F, Chapman D, Brewster D, Holloway
PA et al. 1994 Lactic acidosis and hypoglycaemia in children
with severe malaria: pathophysiological and prognostic
significance. Transactions of the Royal Society of Tropical
Medicine and Hygiene 88 67-73.
Larsson C, Skogseid B, Oberg K, Nakamura Y & Nordenskjold
M 1988 Multiple endocrine neoplasia type 1 gene maps to
chromosome II and is lost in insulinoma. Nature 332 85-87.
Laurent J, Debry G & Floquet J 1971 Hypoglycaemic Tumours.
Amsterdam: Excerpta Medica.
Liessi G, Pasquali C, D’Andrea A, Scandellari C & Pedrazzoli S
1992 MRI in insulinomas: preliminary findings. European
Journal of Radiology 14 46-51.
Lightdale CJ, Botet JF, Woodruff JM & Brennan MF 1991
Localization of endocrine tumors of the pancreas with
endoscopic ultrasonography. Cancer 68 1815-1820.
Losada J, Sarria R, Fernandez Val JF & Lopez Ariztegui MA
1995 DNA ploidy and PCNA index in pancreatic lesions
producing hyperinsulinemic hypoglycemia. Journal of
Surgical Oncology 59 14-17.
Lowe WL, Roberts CT, LeRoith D, Rojeski MT, Merimee TJ,
Teng Fui S, Keen H, Arnold D, Mersey J, Gluzman S, Spratt
D, Eastman RC & Roth J 1989 Insulin-like growth factor II in
nonislet cell tumors associated with hypoglycemia: increased
levels of messenger ribonucleic acid. Journal of Clinical
Endocrinology and Metabolism 69 1153-1159.
Lyall SS, Marieb NJ, Wise JK, Cornog JL, Neville EC & Felig P
1975 Hyperinsulinemic hypoglycemia associated with a
neurofibrosarcoma. Archives of Internal Medicine 135 865-
867.
MacDougall IC, Fleming S & Frier BM 1986 Hypoglycaemic
coma associated with gastric carcinoma. Postgraduate
Medical Journal 62 761-764.
McFadzean AJS & Yeung TT 1956 Hypoglycaemia in primary
carcinoma of liver. Archives of Internal Medicine 98 720-736.

McFadzean AJS & Yeung TT 1969 Further observations of
hypoglycaemia in hepatocellular carcinoma. American
Journal of Medicine 47 220-226.
McGee GS, Kulaylat MM & Abumrad NN 1987 Defective
glucose utilization in patients with functioning beta cell
tumors persists after tumor excision. Annals of Surgery 206
449-457.
McIntyre I 1989 Amylinamide, bone conservation and pancreatic
β-cells. Lancet ii 1026-1027.
Mahony SM & Tisdale MJ 1990 Metabolic effects of tumour
necrosis factor alpha in NMRI mice. British Journal of Cancer
61 514-519.
Maioli M, Ciccarese M, Pacifico A, Tonolo G, Ganau A, Cossu
S, Tanda F & Realdi G 1992 Familial insulinoma: description
of two cases. Acta Diabetologica 29 38-40.
Makino H, Noda K, Inagaki Y, Horie H, Osegawa M, Kanatsuka
A & Yoshida S 1985 Lactic acidosis and hypoglycemia
associated with acute leukemia. Japanese Journal of Medicine
24 257-262.
Marks V 1976 Hypoglycaemia 2 - other causes. Clinical
Endocrinology and Metabolism 5 769-782.
Marks V 1989 Diagnosis and differential diagnosis of
hypoglycemia. Mayo Clinic Proceedings 64 1558-1561.
Marks V 1991 Hypoglycemia due to pancreatic causes. In The
Endocrine Pancreas, pp 207-227. Ed E Samols. New York:
Raven Press.
Marks V 1995 Hypoglycaemia - real and unreal, lawful and
unlawful: the 1994 Banting Lecture. Diabetic Medicine 12
850-864.
Marks V 1997 Insulinomas and hypoglycaemia. In Clinical
Endocrine Oncology, pp 383-386. Eds R Sheaves, PJ Jenkins
& JAH Wass. Oxford: Blackwell Science.
Marks V & Rose FC 1964 Hypoglycaemia. Oxford: Blackwell.
Marks V & Rose FC 1981 Hypoglycaemia. Oxford: Blackwell.
Marks V & Samols E 1966 Hypoglycaemia of non-endocrine
origin. Proceedings of the Royal Society of Medicine 59 338-
340.
Marks V & Teale JD 1996 Investigation of hypoglycaemia.
Clinical Endocrinology 44 133-136.
Marks V, Auld WHR & Barr JB 1965 Carcinoma of stomach and
other non-pancreatic lesions as causes of spontaneous
hypoglycaemia. British Journal of Surgery 52 925-928.
Marks V, Teale JD & Hampton SM 1992 Proinsulin assays in
insulinomas. In Recent Advances on Hypoglycemia, Serono
Symposia Publication 89, pp 145-159. Eds D Andreani, PJ
Lefebvre, V Marks & G Tamburrano. New York: Raven Press.
Mårtensson H, Böttcher G, Sundler F & Nobin A 1990
Localization and peptide content of endocrine pancreatic
tumors. Annals of Surgery 212 607-614.
Mathew PM, Young JM, Abu-Osba YK, Mulhern BD,
Hammoudi S, Hamdan JA & S’di AR 1988 Persistent neonatal
hyperinsulinism. Clinical Pediatrics 27 148-151.
Meeking DR, Steuer A, Cavan DA, Sönksen PH & Lowy CL
1997 A combination of subcutaneous growth hormone and
octreotide for the treatment of malignant insulinoma.
Endocrinology and Metabolism 4 265-266.
Megyesi K, Kahn CR, Roth J & Gorden P 1974 Hypoglycemia
in association with extrapancreatic tumors: demonstration of
Endocrine-Related Cancer (1998) 5 111-129
elevated plasma NSILA-s by a new radioreceptor assay.
Journal of Clinical Endocrinology and Metabolism 38 931-
934.
Megyesi K, Kahn CR, Roth J & Gorden P 1975 Circulating
NSILA-s in man - preliminary studies of stimuli in vivo and of
binding to plasma components. Journal of Clinical
Endocrinology and Metabolism 41 475-484.
Menegaux F, Schmitt G, Mercadier M & Chigot JP 1993
Pancreatic insulinomas. American Journal of Surgery 165
243-248.
Merlo C, Wimpfheimer C, Muser J & Keller U 1992
Hypoglycemia and multiple myeloma. Schweizerische
Medizinische Wochenschrift 122 1622-1626.
Mitchell ML, Ernesti M, Raben MS & Scaliter H 1965 Control
of hypoglycemia with diazoxide and hormones. Annals of the
New York Academy of Sciences 150 406-414.
Miyazaki K, Funakoshi A, Nishihara S, Wasada T, Koga A &
Ibayashi H 1986 Aberrant insulinoma in the duodenum.
Gastroenterology 90 1280-1285.
Moller N, Frystyk J, Skjaerbaek C, Pedersen HB, Christensen C,
Schmitz O, Gee A, Gjedde A, Zapf J & Orskov H 1996
Systemic and regional tumour metabolism in a patient with
non-islet cell tumour hypoglycaemia: role of increased levels
of free insulin-like growth factors. Diabetologia 39 1534-
1535.
Moore NR, Rogers CE & Britton BJ 1995 Magnetic resonance
imaging of endocrine tumours of the pancreas. British Journal
of Radiology 68 341-347.
Nankervis A, Proietto J, Aitken P & Alford F 1985
Hyperinsulinaemia and insulin insensitivity: studies in
subjects with insulinoma. Diabetologia 28 427-431.
Nauck M, Stockmann F & Creutzfeldt W 1990 Evaluation of a
euglycaemic clamp procedure as a diagnostic test in
insulinoma patients. European Journal of Clinical
Investigation 20 15-28.
Nauck MA, Baum F, Seidensticker F, Roder M, Dinesen B &
Creutzfeldt W 1997 A hyperinsulinaemic, sequentially eu-
and hypoglycaemic clamp test to characterize autonomous
insulin secretion in patients with insulinoma. European
Journal of Clinical Investigation 27 109-115.
O’Brien TD, Butler AE, Roche PC, Johnson KH & Butler PC
1994 Islet amyloid polypeptide in human insulinomas.
Diabetes 43 329-336.
Orci L, Vassalli J-D & Perrelett A 1988 The insulin factory.
Scientific American 259 50-61.
Owens LV, Huth JF & Cance WG 1995 Insulinoma: pitfalls in
preoperative localization. European Journal of Surgical
Oncology 21 326-328.
Paineau J, Blanloeil Y, Legrand D, Pineau M, Letessier E,
Charbonneau P & Visset J 1988 Postoperative hypoglycemia
after excision of pheochromocytoma: a case. Presse Medicale
17 475-478.
Pasieka JL, McLeod MK, Thompson NW & Burney RE 1992
Surgical approach to insulinomas. Archives of Surgery 127
442-447.
Pavelic K & Popovic M 1981 Insulin and glucagon secretion by
renal adenocarcinoma. Cancer 48 98-100.
127
Marks and Teale: Tumours producing hypoglycaemia
Perdue JF, Gowan LK, Hudgins WR, Scheuermann J, Foster B &
Northcutt Brown S 1991 Development of a specific
radioimmunoassay for E domain containing forms of insulin-
like growth factor II. In Molecular Biology and Physiology of
Insulin, IGFs and their Receptors, pp 45-56. Eds MK Raizada
& D LeRoith. New York: Plenum Publishing Corporation.
Perros P, Simpson J, Innes JA, Teale JD & McKnight JA 1996
Non-islet cell tumour-associated hypoglycaemia: 111In-
octreotide imaging and efficacy of octreotide, growth
hormone and glucocorticosteroids. Clinical Endocrinology 44
727-731.
Phuphanich S, Jacobs L, Poulos JE & Vesely DL 1995 Case
report: Hypoglycemia secondary to a meningioma. American
Journal of the Medical Sciences 309 317-321.
Pitre J, Soubrane O, Palazzo L & Chapuis Y 1996 Endoscopic
ultrasonography for the preoperative localization of
insulinomas. Pancreas 13 55-60.
Porte D Jr & Kahn SE 1989 Hyperproinsulinemia and amyloid in
NIDDM. Clues to etiology of islet beta-cell dysfunction.
Diabetes 38 1333-1336.
Redmon B, Pyzdrowski KL, Elson MK, Kay NE, Dalmasso AP
& Nuttall FQ 1992 Hypoglycemia due to an insulin-binding
monoclonal antibody in multiple myeloma. New England
Journal of Medicine 326 994-998.
Rochet N, Blanche S, Carel JC, Fischer A, Deist FL, Griscelli C,
van Obberghen E & Le Marchand-Brustel Y 1989
Hypoglycaemia induced by antibodies to insulin receptor
following a bone marrow transplantation in an
immunodeficient child. Diabetologia 32 167-172.
Rockett KA, Awburn MM, Rocket EJ & Clark IA 1994 Tumor
necrosis factor and interleukin-1 synergy in the context of
malaria pathology. American Journal of Tropical Medicine
and Hygiene 50 735-742.
Ron D, Powers AC, Pandian MR, Godine JE & Axelrod L 1989
Increased insulin-like growth factor II production and
consequent suppression of growth hormone secretion: a dual
mechanism for tumor-induced hypoglycemia. Journal of
Clinical Endocrinology and Metabolism 68 701-706.
Rosch T, Lightdale CJ, Boten JF, Boyce GA, Sivak MV Jr,
Yasuda K, Heyder N, Palazzo L, Dancygier H & Schusdziarra
V 1992 Localization of pancreatic endocrine tumors by
endoscopic ultrasonography. New England Journal of
Medicine 326 1721-1726.
Rother KI, Carney JA, Couce M, Charlesworth J & Butler PC
1995 Islet amyloid polypeptide in pancreatic tissue of children
with persistent hyperinsulinemic hypoglycemia caused by
primary islet hyperplasia and nesidioblastosis. Journal of
Clinical Endocrinology and Metabolism 80 1956-1959.
Russell-Jones DL, Thomas SM, Simpson RD, Bevan PC &
Bateman CJT 1990 The first reported case of hypoglycaemia
as the presenting feature of invasive plasmacytoma with a
paraprotein band and grossly elevated insulin levels. British
Journal of Haematology 76 151-153.
Samaan NA, Pham FK, Sellin RV, Fernandez JF & Benjamin RS
1990 Successful treatment of hypoglycemia using glucagon in
a patient with an extrapancreatic tumor. Annals of Internal
Medicine 113 404-406.
128
Samols E 1965 Immunochemical aspects of insulin. In On the
Nature and Treatment of Diabetes, pp 226-246. Eds RS Leibel
& GA Wrenshall. Amsterdam: Excerpta Medica.
Samols E & Marks V 1963 Insulin assay in insulinomas. British
Medical Journal i 507-510.
Sasaki H, Yamatani K, Tominaga M, Ebitani I, Yawata Y, Hara
M & Kobayashi S 1980 Insulinoma-carcinoid syndrome as a
paraneuroma. Biomedical Research 1 159-164.
Sata N, Kimura W, Kanazawa T & Muto T 1995 Malignant
insulinoma causing liver metastases 8 years after the initial
surgery: report of a case. Surgery Today 25 640-642.
Schumacher B, Lubke HJ, Frieling T, Strohmeyer G & Starke AA
1996 Prospective study on the detection of insulinomas by
endoscopic ultrasonography. Endoscopy 28 273-276.
Service FJ, Dale AJD, Elveback LR & Jiang N-S 1976
Insulinoma - clinical and diagnostic features of 60 consecutive
cases. Mayo Clinic Proceedings 51 417-429.
Shapiro ET, Bell GI, Polonsky KS, Rubenstein AH, Kew MC &
Tager HS 1990 Tumor hypoglycemia: relationship to high
molecular weight insulin-like growth factor II. Journal of
Clinical Investigation 85 1672-1679.
Snowden JA, Greaves M & Page K 1994 Reversal of diabetes
associated with escape of myeloma: evidence for
inappropriate IGF-II secretion. British Journal of
Haematology 87 202-204.
Stehouwer CD, Lems WF, Fischer HR, Hackeng WH & Naafs
MA 1989 Aggravation of hypoglycemia in insulinoma
patients by the long-acting somatostatin analogue octreotide
(Sandostatin). Acta Endocrinologica 121 34-40.
Stuart CA, Prince MJ, Peters EJ, Smith FE, Townsend CM &
Poffenbarger PL 1986 Insulin receptor proliferation: a
mechanism for tumor-associated hypoglycemia. Journal of
Clinical Endocrinology and Metabolism 63 879-885.
Sturrock NDC, Selby C & Hosking DJ 1997 Spontaneous
hypoglycaemia in a non-insulin-dependent diabetes mellitus
patient with disseminated pancreatic carcinoma. Diabetic
Medicine 14 324-326.
Svartberg J, Stridsberg M, Wilander E, Andersson DE &
Eriksson B 1996 Tumour-induced hypoglycaemia in a patient
with insulin-dependent diabetes mellitus. Journal of Internal
Medicine 239 181-185.
Tasaka Y, Nakaya F, Matsumoto H, Odagiri R, Ohgawara H &
Omori Y 1993 Tumor and serum levels of proinsulin and
insulin in insulinoma patients. Endocrine Journal 40 245-248.
Taylor SI, Barbetti F, Accili D, Roth J & Gorden P 1989
Syndromes of autoimmunity and hypoglycemia:
autoantibodies directed against insulin and its receptor.
Endocrinology and Metabolism Clinics of North America 18
123-143.
Teale JD & Marks V 1986 The measurement of insulin-like
growth factor I - clinical applications and significance. Annals
of Clinical Biochemistry 23 413-424.
Teale JD & Marks V 1990 Inappropriately elevated plasma
insulin-like growth factor II in relation to suppressed insulin-
like growth factor I in the diagnosis of non-islet cell tumour
hypoglycaemia. Clinical Endocrinology 33 87-98.
Teale JD & Marks V 1998 Glucocorticoid therapy suppresses
abnormal secretion of big IGF-II by non-islet cell tumours

inducing hypoglycaemia (NICTH). Clinical Endocrinology
(In Press).
Teale JD, Pearse AG & Marks V 1987 The measurement of
insulin, C-peptide and β-hydroxybutyrate in the differential
diagnosis of spontaneous hypoglycaemia. In Hypoglycaemia,
pp 281-282. Eds D Andreani, V Marks & PJ Lefebvre. New
York: Raven Press.
Teale JD, Starkey BJ, Marks V, Hampton SM, Robinson J, Fisher
RA & Levine DF 1989 The prevalence of factitious
hypoglycaemia due to sulphonylurea abuse in the UK: a
preliminary report. Practical Diabetes 6 177-178.
Teale JD, Blum WF & Marks V 1992 Alleviation of non-islet cell
tumour associated hypoglycaemia through the restorative
effect of growth hormone therapy on plasma levels of insulin-
like growth factor binding protein-3. Annals of Clinical
Biochemistry 29 314-323.
Thakker RV 1997 Multiple endocrine neoplasia type 1. In
Clinical Endocrine Oncology, pp 433-444. Eds R Sheaves, P
Jenkins & JAH Wass. Oxford: Blackwell Science.
Thompson GB, Service FJ, van Heerden JA, Carney JA,
Charboneau JW, O’Brien PC & Grant CS 1993 Reoperative
insulinomas 1927 to 1992: an institutional experience.
Surgery 114 1196-1206.
Thompson NW, Czako PF, Fritts LL, Bude R, Bansal R, Nostrant
TT & Scheiman JM 1994 Role of endoscopic ultrasonography
in the localization of insulinomas and gastrinomas. Surgery
116 1131-1138.
Trivedi N, Mithal A, Sharma AK, Mishra SK, Pandey R, Trivedi
B & Daughaday WH 1995 Non-islet cell tumour induced
hypoglycaemia with acromegaloid facial and acral swelling.
Clinical Endocrinology 42 433-435.
Valette PJ & Souquet JC 1989 Pancreatic islet cell tumors
metastatic to the liver: treatment hy hepatic artery chemo-
embolization. Hormone Research 32 77-79.
Vinik AI, Delbridge L, Moattari R, Cho K & Thompson N 1991
Transhepatic portal vein catheterization for localization of
insulinomas: a ten-year experience. Surgery 109 1-11.
Endocrine-Related Cancer (1998) 5 111-129
Walters EG, Tavar JM, Denton RM & Walters G 1987
Hypoglycaemia due to an insulin-receptor antibody in
Hodgkin’s disease. Lancet i 241-243.
Wanebo HJ, Schessinger I & Tashima CK 1966 Severe
hypoglycemia associated with terminal lymphomas. Cancer
19 1451-1459.
Wasada T, Eguchi Y, Takayama S, Yao K, Hirata Y & Ishii S 1989
Insulin autoimmune syndrome associated with benign
monoclonal gammopathy. Evidence for monoclonal insulin
autoantibodies. Diabetes Care 12 147-150.
Watson RGP, Johnston CF, O’Hare MMT, Anderson JR, Wilson
BG, Collins JSA, Sloan JM & Buchanan KD 1989 The
frequency of gastrointestinal endocrine tumours in a well-
defined population - Northern Ireland 1970-1985. Quarterly
Journal of Medicine 72 647-657.
Wells JL, Heath DA & West RJ 1990 Hepatic artery embolization
in the treatment of intractable hypoglycaemia. Journal of the
Royal Society of Medicine 83 592-593.
Wright J, Abolfathi A, Penman E & Marks V 1980 Pancreatic
somatostatinoma presenting with hypoglycaemia. Clinical
Endocrinology 12 603-608.
Wu J-C, Daughaday WH, Lee S-D, Hsiao TS, Chou C-K, Lin H-
D, Tsai Y-T & Chiang BN 1988 Radioimmunoassay of serum
IGF-I and IGF-II in patients with chronic liver diseases and
hepatocellular carcinoma with or without hypoglycemia.
Journal of Laboratory and Clinical Medicine 112 589-594.
Yoshikawa K & Wakasa H 1980 Hypoglycemia associated with
aberrant insulinoma: a case report of 16 years follow-up.
Tohoku Journal of Experimental Medicine 132 17-29.
Younus S, Soterakis J, Sossi AJ, Chawla SK & LoPresti PA 1977
Hypoglycemia secondary to metastases to the liver: a case
report and review of the literature. Gastroenterology 72 334-
337.
Zimmer T, Stolzel U, Bader M, Koppenhagen K, Hamm B, Buhr
H, Riecken EO & Wiedenmann B 1996 Endoscopic
ultrasonography and somatostatin receptor scintigraphy in the
preoperative localisation of insulinomas and gastrinomas. Gut
39 562-568.
129