Mysterious episodic coma

Abstract

In June, 2006, an 84-year-old man was admitted to hospital with confusion, slowed speech and movement, frequent urination, and urinary incontinence. Within 48 h, he was in a coma. Intermittently, white blood cells were found in his urine, so antibiotics were prescribed. The patient recovered gradually, over several days, and was discharged. In the next 5 months, he had six similar presentations, on each occasion becoming comatose. The working diagnosis of urinary-tract infection, as a cause of coma, was not entirely convincing: white blood cells were not always present in the patient’s urine; blood cultures were consistently negative; on blood testing, the white-cell count and concentration of C-reactive protein were normal; the patient had no fever. However, no other cause of coma had been identifi ed. The patient had no notable medical history, other than stomach cancer, which had been treated 12 years before, by roux-en-Y gastrectomy; and ongoing paroxysmal atrial fi brillation. ESR, and concentrations of bilirubin and γ-glutamyltransferase were slightly high, but results of other blood tests— including virological, immuno logi cal, and toxicological screens, and concentration of D-lactate—were normal. Ultrasonography of the abdomen showed three gallstones, but normal bile ducts, gallbladder, and liver. Abdominal radiography, with a barium swallow, showed the gastrojejunostomy, but nothing unexpected. MRI of the head, magnetic-resonance angiography, and analysis of cerebrospinal fl uid showed nothing of note. Therefore, as the patient recovered from coma in December, 2006, we kept him in hospital for observation and further investigation. 6 days after recovery, the patient again became confused and lethargic; 24 h later, he scored only 5 on the Glasgow coma scale (GCS; fi gure). We found symmetrical rigidity and hyperrefl exia; plantar refl exes were normal. Blood-gas analysis showed metabolic alkalosis (pH 7∙51, PaO2 14∙7 kPa, PaCO2 3∙6 kPa, base excess –0∙4 mmol/L). CT of the head showed no acute changes. Electroencephalography revealed diff use slowing, consistent with encephalopathy. The next day, the patient was responsive, although confused; his serum ammonia concentration was 48 μmol/L (normal range 15–55 μmol/L). A further 2 days later, the patient was mobile and lucid; 10 days after this recovery, he became unresponsive once more. The serum ammonia concentration was 235 μmol/L. We adminis tered intravenous fl umazenil (200 μg): 1 min later, the patient opened his eyes; at 5 min, he was able to obey commands; at 15 min, he spoke in single words. The morning after, he was well enough for transfer to a tertiary centre, for hydrogen breath testing, results of which were consistent with bacterial overgrowth. After 6 months of treatment with antibiotics, the serum ammonia concentration, and results of hydrogen breath testing, were normal. The patient has had no comas since antibiotic treat ment began, except for a brief recurrence after treatment had been stopped for 2 weeks. The plasma aminoacid profi le was not diagnostic for a urea-cycle defect. When last seen, in April, 2008, the patient was well. The diff erential diagnosis of episodic coma in elderly patients includes epilepsy, neurodegenerative disease with sepsis, recurrent ischaemic cerebrovascular events, and structural brain lesions causing episodic elevation in intracranial pressure. Metabolic causes include hepatic encephalopathy, D-lactic acidosis, and hyperammonaemia. The latter can be caused by a urea-cycle defect, or colon isation of the small intestine by colonic bacteria. Risk factors for bacterial overgrowth syndrome, which is under diag nosed in elderly patients, include anatomical alter ations to the intestine, drugs that inhibit gastric acid sec retion, diabetes, and advanced age. Bacterial overgrowth syn drome often presents with diarrhoea, malabsorption, and weight loss, but can present with encephalopathy caused by release of neurotoxins, notably ammonia. Chronic exposure to ammonia is thought to inhibit astro cytic uptake of γ-aminobutyric acid—hence, perhaps, the appar ent reversal of coma by fl umazenil— and to increase neuro steroid synthesis, which contributes to encephalopathy.