5-Oxoprolinuria as a cause of high anion gap metabolic acidosis

We report a case of high anion gap metabolic acidosis (HAGMA) caused by 5-oxoprolinuria resulting from chronic intermittent paracetamol therapy, malnutrition and concomitant moderate renal/hepatic dysfunction.

A 36-year-old Caucasian woman was brought to our intensive care unit after intubation at another hospital. She initially presented with drowsiness and tachypnoea. The initial working diagnosis was bibasilar bronchopneumonia, and she received a dose of ceftriaxone. Arterial blood gas (prior to intubation) showed pH 6.99, partial pressure of CO₂ 14 mmHg, partial pressure of O₂ 80 mmHg and 92% saturation on oxygen 2 l min⁻¹ via nasal cannulae. She had past medical history of migraine, seizures, depression, chronic obstructive pulmonary disease, malnutrition, anaemia and deranged hepatic transaminases for 4 years following a prior suicide attempt with paracetamol. She had a history of marijuana, methamphetamine, dextromethorphan cold tablet and ethanol abuse, in addition to 40 pack-year history of smoking. Her prescription medications included divalproex sodium, levetiracetam, olanzapine, promethazine, mirtazapine, fenofibrate, lactulose, omeprazole and paracetamol as needed. On examination, she was malnourished (body mass index 17.9 kg m⁻²), pale, tachypnoeic (40 breaths min⁻¹) and tachycardic (hear rate 102 min⁻¹) with a blood pressure of 105/60 mmHg and weak peripheral pulses.

Further laboratory studies showed an anion gap of 20, with bicarbonate 5 mequiv l⁻¹ and creatinine 1.5 mg dl⁻¹ (132.6 µmol l⁻¹). The Δanion gap/Δbicarbonate ratio = 0.42 suggested a mixed anion and non-anion gap acidosis. The latter was presumed to be from renal dysfunction. Other pertinent laboratory values were as follows: haemoglobin 11.9 g dl⁻¹ (119 g l⁻¹), white blood cell count 40 000 mm⁻³ (40 × 10⁹ l⁻¹) and platelets 854 000 mm⁻³ (854 × 10⁹ l⁻¹). Total bilirubin was normal, aspartate aminotransferase 322 U l⁻¹ (normal 5–40 U l⁻¹), alanine aminotransferase 101 U l⁻¹ (normal 7–56 U l⁻¹), alkaline phosphatase 296 U l⁻¹ (normal 38–126 U l⁻¹), international normalized ratio 2.5, ammonia 78 µg dl⁻¹ (45.7 µmol l⁻¹; normal 17–60 µg dl⁻¹), lipase 411 U l⁻¹ (normal 7–60 U l⁻¹), amylase 68 U l⁻¹, activated partial thromboplastin time 50 s and valproate concentration 24 mg l⁻¹ (subtherapeutic). Urinalysis and drug of abuse screen were unremarkable.

The following causes of metabolic acidosis were negative: lactic acid 1.0 mmol l⁻¹, ethanol <0.01 mg dl⁻¹, methanol <5 mg dl⁻¹, isopropranolol <5 mg dl⁻¹, ethylene glycol <10 mg dl⁻¹, propylene glycol <0.04 mg dl⁻¹, salicylate 4.3 mg dl⁻¹, paracetamol <1.2 µg ml⁻¹ and β-hydroxybutyrate 0.2 mmol l⁻¹; and d-lactic acid, acetaldehyde, paraldehyde and acetone were not detected.

The patient's acidosis did not resolve over the next 2 days, so she was started on sodium bicarbonate infusion and later switched to continuous veno-venous haemodialysis, which proved very effective in correcting her metabolic derangement.

To explore other aetiologies of HAGMA, urine organic acid screen was sent on the initial urine sample, which revealed a markedly elevated excretion of 5-oxoproline. A diagnosis of HAGMA secondary to 5-oxoprolinuria was made. The patient regained her normal neurological state with no acid–base disturbance, and repeat urine organic acid screen on discharge 7 days later was negative, indicating a transient 5-oxoprolinuria. The transient 5-oxoprolinuria and the absence of haemolytic anaemia makes the concomitant existence of a genetic deficiency of glutathione synthetase in this patient unlikely.

Excess 5-oxoproline (also known as pyroglutamic acid) production is a rare cause of HAGMA. It is an intermediary in the γ-glutamyl cycle, which facilitates the transport of the tripeptide glutathione (glutamyl-cystinyl-glycine) and its constituent amino acids across cellular membranes and regenerates glutathione intracellularly. Reduced glutathione is required for detoxification and minimization of free-radical-induced oxidative stress. As shown in Figure 1, excess 5-oxoproline is generated via the γ-glutamyl cyclotransferase enzyme when glutathione synthetase is deficient [1]. Glutathione depletion exerts a negative feedback on γ-glutamyl cysteine synthetase. This negative feedback is decreased in situations of glutathione depletion, thus increasing production of γ-glutamyl cysteine, which acts as a substrate for γ-glutamyl cyclotransferase enzyme to generate 5-oxoproline. Glutathione depletion is seen in liver disease, paracetamol use, alcohol abuse, fad diets (e.g. low-protein diet), glycine deficiency, malnutrition and severe sepsis [1–4]. 5-Oxoproline is cleared renally and thus accumulates in renal dysfunction. It is oxidized by the enzyme 5-oxoprolinase to l-glutamate; however, certain drugs (flucloxacillin, netilmicin and vigabatrin) can inhibit 5-oxoprolinase, hence preventing its degradation [2, 5]. 5-Oxoproline concentrations are also increased in patients with burns and those on total parenteral nutrition [1].

Figure 1.

γ-Glutamyl cycle showing generation of 5-oxoproline (pyroglutamic acid). AKI, acute kidney injury

Our patient was a malnourished woman with history of chronic paracetamol therapy, which in combination probably led to glutathione depletion. Importantly, supratherapeutic/toxic paracetamol concentrations do not appear to be required to cause 5-oxoprolinuria, as was seen in our patient and as reported by others [2, 6, 7]. The management of metabolic acidosis from 5-oxoprolinuria involves discontinuing the offending agents (antibiotics or antiepileptics) and providing supportive care. Some authors suggest using N-acetylcysteine to regenerate glutathione or using intravenous bicarbonate if pH is <7.0 [5, 8, 9].

Physicians should consider screening urine for organic acids when other common causes of HAGMA have been excluded, especially in patients who have been taking paracetamol.

Competing Interests

There are no competing interests to declare.