A 66-year-old man presented to hospital following two weeks of diarrhoea and worsening confusion, unsteady gait and muscle twitching. His oral intake and urine output were markedly reduced. The medical history included bipolar disorder, class III obesity, type 2 diabetes, hypertension and dyslipidaemia. He was being treated with venlafaxine, mirtazapine, lithium, aripiprazole, metformin, empagliflozin, olmesartan and rosuvastatin, which he continued while unwell. He had been on lithium for nearly 20 years with no recent change of dose. Serum trough concentrations of lithium were maintained around 0.6 mmol/L (target range 0.6–0.8 mmol/L). Renal function was normal when measured three months earlier.
On examination, the man was lethargic and not orientated to time or place. His blood pressure was 100/65 mmHg with otherwise normal vital signs. He had a coarse tremor, myoclonic jerks and generalised hyperreflexia, but no focal neurological deficit.
Investigations revealed normal serum sodium, potassium and glucose concentrations, but urea was 25 mmol/L and creatinine was 1130 micromol/L. Serum lithium had increased to 2.7 mmol/L. There was a moderately severe anion gap metabolic acidosis but a normal lactate. An ECG showed a normal sinus rhythm. CT of the abdomen excluded structural abnormalities that may have accounted for impaired kidney function or diarrhoea.
The patient was diagnosed with severe acute kidney injury precipitated by hypovolaemia and subsequent neurotoxicity from lithium accumulation. All drugs were temporarily ceased, intravenous fluids were given and he was admitted to intensive care. Continuous renal replacement therapy was provided under vasopressor support until serum lithium concentrations approached 1 mmol/L and renal function improved. The diarrhoea settled, but the man’s neurological recovery was slow and complicated by hypernatraemia, ileus and hospital-acquired pneumonia.
After four weeks, the patient was transferred to a rehabilitation facility and lithium recommenced at half the usual dose. His daily urine output remained greater than 3 L which was consistent with nephrogenic diabetes insipidus.