Una mujer de 30 años ingresó en el hospital en coma profundo sin causa aparente ni duración conocida. Los estudios posteriores revelaron un hiperinsulinismo endógeno con hipoglucemia como origen del cuadro. La cateterización selectiva arterial de las ramas del tronco celíaco con inyección de calcio y recogida de muestras en vena hepática permitió localizar un insulinoma en la cabeza pancreática, no evidente en ninguna de las otras exploraciones practicadas. La cirugía subsiguiente fue curativa, pero quedaron secuelas neurológicas con lesiones groseras evidentes en la RMN.
Aunque infrecuente, la posibilidad de una hipoglucemia debida a un insulinoma debe tenerse presente en el diagnóstico diferencial de un paciente en coma sin etiología evidente. El debut de un insulinoma con un único episodio de coma puede tener un efecto devastador a nivel cerebral, con importantes secuelas neurológicas.
A young woman was brought to the hospital with a deep coma of unknown duration and origin. Posterior studies revealed hypoglycaemia as the cause, and selective intraarterial calcium injection with hepatic venous sampling allowed the localization and resection of an insulinoma.
Although uncommon, hypoglycaemia due to an insulinoma should be considered in the diagnosis of any patient with a coma of unknown origin. The first symptom of an insulinoma may have a devastating effect on the brain resulting in permanent neurological sequelae.
INTRODUCTION
There are numerous diagnostic posibilities in an unconscious patient without known previous diseases. They include mass lesions, severe head injuries, hypertensive brain haemorrhage, and alcoholic and other forms of drug intoxication.
Hypoglycaemia due to an insulinoma is uncommon and difficult to diagnose, as it mimics a great variety of neurological conditions. Regardless of the absence of previous symptoms it must be suspected if no other causes of coma are obvious. Signs and symptoms of hypoglycaemia are classified into two major groups: autonomic and neuroglucopenic. The latter includes headache, dizziness, confusion, tiredness, difficulty in thinking, convulsions and loss of consciousness, among others. These symptoms differ between but are consistent from episode to episode for each patient1, and may appear without preceding autonomic symptoms (sweating, trembling, anxiety, palpitations) as the only manifestation of a hypoglycaemic attack.
A previously silent insulinoma which caused a profound hypoglycaemic coma with evidence of permanent neurologic damage in a young female is described. We detail the diagnostic tests employed to diagnose the tumour, particularly the selective intraarterial stimulation with calcium.
CASE REPORT
A 28-year-old woman with no known medical history was taken unconscious to Emergency after having been found at home in a deep coma. It was not possible to know how long she had been comatose. Before arrival at the hospital she had received naloxone, flumacenile and intravenous dextrose. Neurological examination showed her to be deeply unconscious, with response to pain and bilateral extensor plantar responses. She presented left-sided weakness with motor strenght 2/5, spasticity and conjugated left deviation of both eyes. She had no medical history of note and was not taking medication. Her family did not know previous epidodes of neurological disturbance. Blood tests including random glucose (9 mmol/l after intravenous dextrose), thyroid function, copper and caeruloplasmin were normal. Cerebrospinal fluid (CSF) was acellular with normal glucose concentration. Emergency computed tomography brain scan (CT) was also normal. Two days after admission to the intensive care unit, autoantibody screen was negative as were all the viral serologies. An electroencephalogram (EEG) showed basal activity of very few theta waves a 5 Hz with frecuent regular delta waves at 2-2,5 Hz originating from the right subcortical region. Magnetic resonance imaging (MRI) performed after seven days, demonstrated high signal areas in the region of both the globus palidus and left external capsule. Positron emission tomography (PET) with Hmpao-Tc99 confirmed an area of hypoperfusion in the left frontoparietal region. Cardiovascular examination, including echocardiogram, Holter register and doppler study of supraaortic trunks, were all normal. In view of the EEG register, the possibility of a hypoglycaemic coma was considered and a 72-hour fast was programmed. The fast had to be suspended at 14 hours because the patient had symptoms of hypoglycaemia and a plasma glucose level of 1.9 mmol/l. Plasma insulin concentration of 10.8 mUI/l (77.5 pmol/l) and C-peptide of 1.8 ng/ml (60 nmol/l) were received one week later. As her family did not realize previous symptoms suggestive of hypoglycemia and we hadn't collected a specimen of urine, a new fast was programmed with collection of plasma and urinary specimens to rule out factitious hypoglycaemia by sulphonylureas (SU). Asymptomatic hypoglycaemia of 2 mmol/l occurred at 20 hours and a symptomatic of 1.8 mmol/l at 24 hours, with simultaneous insulin levels of 132.7 and 110.5 pmol/l and detectable C-peptide levels. Plasma and urine detection of SU were negative. Clinical evaluation was favourable, with progressive reduction of paresis and residual lacunar amnesia along with slight bradypsichia. There was improvement in EEG registers, with persistence of delta waves in the subcortical right hemisphere. Abdominal CT and selective abdominal arteriography were normal, therefore a selective arterial calcium stimulation test was decided upon. An arterial catether was placed consecutively into the principal arteries supplying the pancreas after standard selective angiography. Insulin levels were measured in samples taken from the right hepatic vein before and 30, 60, 90, 120 and 180 seconds after the injection of 0.025 mEq/kg of calcium gluconate in each artery. Results showed a peak of 76 mUI/l thirty seconds after calcium injection in the superior mesenteric artery, more than two times the basal value (26 mU/l), with no response in the splenic nor the gastroduodenal arteries. These results suggested a cephalic localization of insulinoma. During operation an 8 mm tumour was resected in the pancreatic head, and histology revealed nests of islet cells with negative study of resected adenopathies. The patient was discharged ten days after operation with normal glycaemias. Three months later, motor function had fully recovered, although very slight deterioration of the intellect and lacunar amnesia persisted. Cerebral MRI discovered the presence of the previous lesion in the left external capsule and hyperintense signal areas in the region of protuberance and bilateral talamus in T2-weighted images.
COMMENTS
Hypoglucemia due to an insulinoma can mimic a great variety of neurological conditions, including dystonic choreoatetosis2, peripheral neuropathy3, hemiplegia, cognitive alterations (even dementia4) and other psychiatric processes, frequently without previous adrenergic symptoms5. Deep coma as the first symptom of an insulinoma is not frequently reported, and there may even be a fatal outcome.
In our patient neurological damage was evident in MRI, PET and EEG, involving principally the basal ganglia and subcortical parieto-frontal areas. It has been proposed that some areas of the brain may be more susceptible than others to hypoglycaemia. This selective regional vulnerability has been attributed to local diferences in the celular metabolism and vascular supply6. These lesions can be reversed or cause permanent damage if they are not treated promptly, as that shown in our patient by MRI three months later. Such lesions have been demonstrated in recent studies7 by electrophysiological studies using evoked potentials and peri-pheral conduction velocity.
The supervised 72-hour fast is the classic diagnostic test for hypoglycaemia. With rare exceptions, insulin mediated hypoglycaemic disorders are characterised by plasma insulin concentrations of 6 mUI/l (42.7 pmol/l) or higher, with plasma glucose below 50 mg/dl (2,77 mmol/l). Ratios of glucose to insulin have fallen into disuse following demonstration that absolute insulin value offer diagnostic sensibility at over 95%8. Measurement of sulphonylureas in the plasma at the end of the fast and detection of C-peptide are an essential component of the prolonged supervised fast, in order to allow the biochemical diagnosis of an insulinoma. Criteria using proinsulin levels (greater than 25% of total insulin9,10) and the response of plasma glucose to intravenous glucagon at the end of the fast11 are also useful.
Insulinomas are located in the pancreas in 99% of cases12 and 70% of them are due to a single benign tumor13. Preoperative localization is difficult in many cases and CT, MRI and angiograhy has been used with poor results14. Somatostatin receptor scintigraphy is very useful for diagnosing of gastrinoma but not insulinoma15, while endoscopic ultrasonography appears to detect a higher percentage of tumours than the other techniques in recent reports16. Many works have recently been published revealing excellent results using a new technique with high sensitivity in the topographic localization of insulinomas. It consists of the use of selective intraarterial calcium injection17-19 with measurement of insulin concentration in the right hepatic vein. This test allows the localization of insulin-secreting tumours in regions of the pancreas, and is based on the stimulatory effect of calcium on neoplasic ß -cell20. Preoperative topographical diagnosis makes surgery considerably easier and enables partial pancreatic resection to be performed if no tumour is identified during the operation.
In conclusion, hypoglycemia due to an insulinoma is a rare disease and frequently difficult to diagnose, as it mimics a great variety of neurologic conditions, including the possibility of a single devastating event. It must be considered in the differencial diagnosis of a spontaneous coma, and when biochemical diagnosis is made, it should be localized with selective intraarterial calcium injection.