Severe hypophosphataemia and insulin resistance in diabetic ketoacidosis


We were interested to read the review on hypophosphataemia (Anaesthesia 1998; 53: 895–902) and would like to report a case in which correction of phosphate levels was of importance.

A 20-year-old female insulin-dependent diabetic was admitted to hospital with a lower respiratory tract infection and established ketoacidosis. On admission, her serum glucose was 23.8 mmol. l−1, arterial pH 6.76 and base deficit 32 mmol.l−1. Resuscitation was commenced with crystalloid infusion, and blood glucose initially controlled with human insulin at 2–6 iu.h−1. Despite 6 l of intravenous fluid and administration of sodium bicarbonate, severe metabolic acidosis persisted (pH 6.91), and deterioration in level of consciousness and worsening arterial oxygenation (PO2 7.7 kPa, FIO2 1) expedited admission into the intensive care unit.

Over the next 12 h, the patient became increasingly unwell requiring sedation, mechanical ventilation, pulmonary artery catheterisation, and cardiovascular support with adrenaline and noradrenaline infusions. This improved arterial blood gases (PO2 9.2 kPa, FIO2 0.7, pH 7.2, base deficit 13.4 mmol.l−1, lactate 0.85 mmol. l−1), but could not avert acute renal failure. In addition, blood glucose became increasingly difficult to control, measuring consistently above 20 mmol.l−1 despite increasing the insulin infusion rate to a peak of 160 iu.h−1.

Approximately 16 h after admission to ICU, biochemical investigations revealed a severe hypophosphataemia of 0.02 mmol. l−1 (reference range 0.7–1.3 mmol.l−1), and replacement therapy was commenced using an infusion of potassium hydrogen phosphate (K2HPO4). During the 2 h after starting the phosphate infusion, insulin requirements decreased from 160 iu.h−1 to zero, with glucose easily controlled between 4 and 8 mmol. l−1. Notably, from this time onwards, insulin dose requirements remained low and well within the conventionally accepted therapeutic range. Potassium hydrogen phosphate (K2HPO4) was continued until hyperkalaemia, secondary to the renal failure, necessitated a change to the sodium salt (Na2HPO4).

Peritoneal dialysis was started, and during the next 12–24 h cardiovascular, pulmonary and metabolic parameters improved. The patient made a steady improvement over the next few days, and after 5 weeks in hospital she was discharged home with no apparent long-term sequelae.

In our patient, apart from phosphate replacement, no change in therapy or patient condition occurred in the 4 h prior to change in insulin sensitivity. The degree of metabolic acidosis (pH 7.2) is unlikely to have made a significant contribution to the severe insulin resistance, inotrope levels were stable and peritoneal dialysis was not commenced until several hours later. While we appreciate the role played by other therapies, especially peritoneal dialysis, in the successful management of our patient, we feel reversal of severe insulin resistance by the correction of serum phosphate levels was a significant factor in this case.