
N. Iqbal, A. Byard, B. M. Singh, H. N. Buch
Diabetes and Endocrine CentreNew Cross Hospital Wolverhampton WV10 0QP UK
Correspondence to: Dr.N Iqbal, SpR Diabetes and Endocrinology, Diabetes Centre, Kings Mill Hospital, Sutton in Ashfield NG17 4JL
E-mail: nosheen26@hotmail.com
SMJ 2006 52(1): 55
Abstract
A
75 year-old man diagnosed to have a glucagonoma and as he was deemed unfit for
curative surgery, was managed conservatively. Over the next 7 years as a result
of increase in tumour size and development of hepatic metastases he became
progressively more symptomatic and chemotherapy was administered for palliation.
Six months later, he experienced several episodes of symptomatic hypoglycaemia,
which persisted despite withdrawal of insulin therapy. Biochemical assessment
confirmed spontaneous endogenous hyperinsulinaemia. Hypoglycaemic symptoms
responded partially to somatostatin analogues although over the next few months
he continued to deteriorate and 8 years after the initial diagnosis died of
tumour-related cachexia. This patient presents an unusual scenario where a
neuroendocrine tumour initially presented as a glucagonoma with subsequent
change in the predominant hormone secretion to insulin. In view of its temporal
association, we postulate that following chemotherapy; one cell type was more
successfully reduced as compared to others, resulting in predominance of insulin
secreting cells.
Introduction
Neuroendocrine
tumours are uncommon and often present difficult therapeutic challenges. They
may be composed of different cell types and may be multihormonal. We present a
case that illustrates several interesting features of neuro-endocrine tumours.
Case
Report
In
1995, a 68-year old man presented with a 12-month history of watery diarrhoea,
weight loss, flushing, upper abdominal pain but no skin rash. He was on lifelong
warfarin for 2 previous episodes of deep vein thrombosis. Full blood count,
renal function and oesophago-duodenoscopy were normal. Abdominal CT scan showed
a 5cm pancreatic mass and lymphadenopathy around coeliac axis. The patient had
an exploratory laparotomy although only partial excision could be performed as
the mass was densely adherent to the coeliac axis and patient was deemed unfit
for a Whipple’s procedure necessary for total excision. Histological
appearance was consistent with that of a neuroendocrine tumour with presence of
poorly defined islands of small, regular non-mitotic cells staining positive for
agyrophil granules. Plasma glucagon was elevated on two occasions (94 and 116
pmol/l (normal <50 pmol/l)). Remaining gastrointestinal hormones, urinary
5-H1AA and catecholamines were within the reference ranges. Soon after initial
presentation the patient developed hyperglycaemia, which was well controlled on
small doses of insulin (HbA1C 5.3%).
A
diagnosis of glucagonoma was made. Over
the next 7 years, he was managed symptomatically in conjunction with an
oncologist. In 2002, in view of worsening symptoms he underwent a repeat CT scan
that showed an increase in tumour size to 8.5 cm and an octreoscan showed the
presence of at least three areas of liver metastases, which were not seen on CT
scan. He received 5 cycles of
Carboplatin and Etoposide with a view to reducing tumour load. Six
months after completion of chemotherapy he developed episodes of spontaneous
hypoglycaemia, which persisted despite discontinuation of insulin. During one
such episode biochemical assessment showed plasma glucose of 2.0 mmol/l, plasma
insulin 166 pmol/L, C-peptide 1285 pmol/L, proinsulin 196 pmol/l, a normal IGF-1
and a negative sulphonylurea screen. This
confirmed spontaneous hyperinsulinaemic hypoglycaemia and the dominant tumour
product appeared to have changed from glucagon to insulin. Hypoglycaemia was
partially controlled with long acting somatostatin analogue. However over the
next few months he continued to deteriorate and died of tumour-related cachexia
8 years after first diagnosis.
Discussion
We
have presented a patient with metastatic neuroendocrine tumour in whom there was
a change in the predominant secretory product from glucagon to insulin. This was
accompanied by a corresponding change in the clinical picture from that of
insulin requiring hyperglycaemia to that of episodic spontaneous hypoglycaemia.
A change in the secretory product is a recognised phenomenon with neuroendocrine
tumours although a change from glucagon to insulin is unusual. Reports have
shown that in a series of 24 patients with pancreatic endocrine tumours, 11 had
glucagonomas, who then went on to develop elevated levels of other hormones
during follow up.1 In common with our patient, most of the patients
in this series had metastatic disease at initial diagnosis. The median time to
elevation of a second hormone was 19 months from the original diagnosis, and was
associated with new clinical symptoms and deterioration in health. There are
several hypotheses for the underlying mechanism for a change in secretory
product of a neuroendocrine tumour. One of these suggests a spontaneous
de-differentiation of tumour cells to pluripotent cells with subsequent
differentiation to cells with capability of producing a different hormone.
Studies have shown that in animal studies using rats that cultured clonal cell
lines from insulinomas dedifferentiated, and were then able to express the genes
for insulin, glucagon, somatostatin and angiotensin.2, 3 Another
postulates the co-existence of two tumours with one of them manifesting later
than the other. The literature has two case reports of patients with co-existing
tumours and one of the patients had a coexisting glucagonoma and insulinoma.
However we believe, that in our patient, the tumour was co-secreting both
glucagon and insulin. It is known that up to 69% of patients have polyfunctional
endocrine tumours that secrete more than one hormone, but only present with a
clinical syndrome due to the predominant one.4, 5, 6 Insulin is the
most common hormone that is co-secreted with glucagon. Chemotherapeutic agents
could have selectively influenced glucagon-producing cells while sparing
insulin-secreting cells leading to unopposed action of insulin. The temporal
association with administration of chemotherapy and the relatively “mild”
diabetes mellitus support this hypothesis. 92% of patients in the above series
had received chemotherapy prior to the onset of secondary syndrome.
The
diagnosis of glucagonoma was based on typical symptoms, presence of diabetes,
history of deep vein thrombosis and a glucagon level greater than twice the
upper limit of the reference range (116pmol/l, normal <50) although typical
skin rash had not yet developed over 8 years in our patient, which is a
recognised feature. Plasma glucagon was only mildly raised, although in cases
reported in literature glucagon level have ranged from 82 to 14,300 pg/ml
(<61 pg/ml) in patients with confirmed glucagonoma. It is well recognised for
the clinical symptoms to be disproportionate to plasma glucagon levels as was
seen in our patient.6
Another
interesting feature was the failure of CT scan to detect the presence of hepatic
metastases that were demonstrated on the octreoscan performed around the same
time. This interesting radiological paradox of the lower sensitivity of CT and
MRI in demonstrating metastases from a neuroendocrine tumour has been shown
previously.7
Our patient survived for 9 years after he was
diagnosed to have an inoperable pancreatic neuroendocrine tumour. Prolonged
survival in inoperable and metastatic pancreatic endocrine tumours is
recognised.8 In one series,9 only 9 out of 21 patients
with glucagonoma had tumour related deaths while another10 noted a
10-year survival of 28% in another series of patients with malignant
glucagonomas.
Conclusion
We
report a case of a patient with an inoperable and metastatic neuroendocrine
tumour, with a prolonged survival after the initial diagnosis. The dominant
tumour secretion changed from glucagon to insulin after 7 years and following 5
cycles of chemotherapy. The reason for this transformation is postulated to be
due to selective reduction in one type of tumour cells as compared to another
following administration of chemotherapy.
Learning
Points
§
Varying effects of
neuroendocrine tumours can be seen in the same patient following treatment
§
Clinical presentations in
the acute setting may indeed change from hyperglycaemia to hypoglycaemia or
vice-versa
§
Neuroendocrine tumours can
have a slow progression over time though metastatic features may be present
References
Wynick
D, Williams SJ, Bloom SR. Symptomatic secondary hormone syndromes in
patients with established malignant pancreatic endocrine tumours. N Engl J
Med 1988; 319(10): 605-07.
Lips
CJM, van der Sluys Veer J, van der Donk JA, van Dam RH. Common precursor
molecule as origin for the ectopic hormone producing tumour syndrome. Lancet
1978; 1: 16-8.
Baylin
SB, Mendelsohn G. Ectopic (inappropriate) hormone production by tumours:
mechanisms involved and the biological and clinical implications. Endocr Rev
1980; 1: 45-77.
Ohneda
A, Otsuki M, Fujiya H, Yaginuma N et al. A malignant insulinoma transformed
into a glucagonoma syndrome. Diabetes 1979; 28(11): 962-69.
Leichter
SB. Clinical and metabolic aspects of glucagonoma. Medicine 1980; 59(2):
100-13.
Wermers
RA, Fatourechi V, Kvols LK. Clinical Spectrum of Hyperglucagonemia
associated with malignant neuroendocrine tumours. Mayo Clinic Proceedings
1996; 71(11): 1030-38.
Shi
W, Johnston CF, Buchanan KD, Larkin C et al. Localization of neuroendocrine
tumours with [111In] DTPA-octreotide scintigraphy (Octreoscan): a
comparative study with CT and MR imaging QJM 1998; 91(4): 295-301.
D’Arcangues
CM, Awoke S, Lawrence GD. Metastatic insulinoma with long survival and
glucagonoma syndrome. Ann Intern Med 1984; 100(2): 233-35.
Wermers
RA, Fatourechi V, Wynne AG, Kvols LK, Lloyd RV. The glucagonoma syndrome:
clinical and pathologic features in 21 patients. Medicine 1996; 75(2):
53-63.
Grama D, Eriksson B, Matersson H, Cedermark B et al. Clinical characteristics, treatment and survival in patients with pancreatic tumours causing hormonal syndromes. World J Surg 1992; 16(4): 632-9.