Mind the gap! An unusual metabolic acidosis

Abstract

In May, 2010, a 50-year-old woman presented to us with hip pain and a temperature of 37·5°C. She had a history of epilepsy and was taking phenobarbital, carbamazepine, and phenytoin. Her total white cell count was 12·2×109 per L, haemoglobin 152 g/L, and C-reactive protein 87·9 mg/L. Cultures of blood and fl uid aspirated from her hip grew fl ucloxacillin-sensitive Staphylococcus aureus. Our patient was treated with intravenous fl ucloxacillin 8 g and paracetamol 4 g daily. After 12 days, she became obtunded, tachypnoeic, tachycardic, and hypotensive, and she was admitted to the intensive care unit. Arterial blood gas analysis showed a profoundly high anion-gap metabolic acidosis (pH 7·277, pCO2 1·31 kPa, pO2 11·65 kPa, base excess –19·8 mmol/L, bicarbonate 4·5 mmol/L, sodium 128 mmol/L, potassium 2·6 mmol/L, chloride 96 mmol/L, anion gap 30·1 mmol/L). Her glucose was 10·1 mmol/L and lactate 2·15 mmol/L; renal function was normal (urea 4·9 mmol/L, creatinine 99 μmol/L). Urine dipstick was positive for ketones, but starvation ketosis was excluded because the plasma ketone concentration was not high enough to account for the acidosis (3-hydroxybutyrate 0·50 mmol/L, free fatty acids 0·86 mmol/L). The antibiotic regimen was changed to meropenem on the basis of fresh wound and urine cultures. We suspected a diagnosis of pyroglutamic acidaemia because of the history of sepsis treated with fl ucloxacillin and paracetomol. The diagnosis was confi rmed by high urinary concentration of 5-oxyproline (16 500 μmol/mmol creatinine). Her condition failed to improve despite optimal management of her sepsis and repeated sodium bicarbonate administration, so paracetamol treatment was stopped and she was treated with N-acetyl cysteine. Within 3 h of this change in treatment, inotropic support was weaned, and the acidosis was completely corrected within 12 h (pH 7·484, pCO2 3·29 kPa, pO2 11·57 kPa, base excess –4·4mmol/L, bicarbonate 18·1 mmol/L, anion gap 14·2mmol/L). She was ventilated for 13 days, and was discharged home after one month, having made a full recovery. Pyroglutamic acidaemia is an unusual metabolic acidosis caused by the accumulation of 5-oxyproline, occurring as a result of hereditary defects. Although such a defect is unlikely in our patient, her condition can be explained by glutathione depletion and drug-related disruption of the γ-glutamyl cycle (fi gure). Under normal circumstances, glutathione inhibits γ-glutamylcysteine synthetase, and mild depletion leads to increased enzyme activity and restoration of glutathione stores. However, severe depletion saturates glutathione synthetase, leading to an accumulation of γ-glutamylcysteine, which is converted by γ-glutamyl cyclotransferase to 5-oxyproline. Our patient had several reasons to have glutathione depletion. She was malnourished, and glutathione synth esis is sensitive to dietary intake of cysteine. Also, sepsis depletes glutathione in animals, and potentially in people. Our patient had had 12 days’ therapy with paracetamol, a metabolite of which, N-acetylbenzo quinonimine, irre versibly binds glutathione and reduces stores. Her anti-epileptic medications are hepatic enzyme inducers, and compete with paracetamol for metabolism, further depleting glutathione. Sex diff erences in glutathione transferase activity, and lower glutathione stores in women than in men, also make women more susceptible to pyroglutamic acidaemia. Finally, treatment with fl ucloxacillin inhibits the breakdown of 5-oxyproline by 5-oxyprolinase. Pyro glutamic acidaemia should be considered in cases of high anion-gap metabolic acidosis, after the common causes have been excluded, and particularly when the risk factors of female sex, starvation, sepsis, and concurrent paracetamol and fl ucloxacillin therapy are present. It has also been associated with renal impairment and abnormal liver function. If pyroglutamic acidaemia is suspected, causative medications should be stopped immediately, and urinary organic acids measured. Replenishment of glutathione stores with N-acetylcysteine may be benefi cial.