Administration of insulin during surgery.

infiltration by the surgeon may be required, particularly of the carotid sheath. This should be considered a supplementary part of the technique rather than a failure requiring conversion to general anaesthesia. Sir,—I was interested to read the paper by Raucoules-Aime and colleagues [1] on perioperative blood glucose control in diabetics using bolus administration of insulin in place of a continuous infusion. I do not agree with the authors' conclusion that bolus administration is comparable in safety and efficacy with an infusion. First, the control group received an infusion of insulin at a constant rate of 1.25 u.h" 1. One of the main advantages of administering insulin by infusion is that the rate can be titrated according to blood glucose concentration: constant rate administration defeats this purpose. In addition, this group received bolus administration of insulin, as required for hyperglycaemic episodes, which detracts from their value as a control group as they were not devoid of the metabolic instability caused by bolus administration of insulin. One patient in the bolus only group suffered severe hypo-glycaemia after a bolus dose of insulin. While the authors state that this was not significantly different from the incidence of hypoglycaemia in the infusion group (zero), the small numbers of patients involved suggest that complacency about the safety of bolus administration would be misplaced. These patients were undergoing blood glucose estimations every 15 min, which is far more frequently than can often be carried out under normal circumstances, especially during a major case with only one anaesthetist; such frequent glucose estimations probably negate some of the "simplicity" of bolus administration and, as seen, would still not rule out potentially fatal hypoglycaemia. As insulin has a short half-life, it is likely that 2-hourly bolus administrations would lead to periods when circulating insulin concentrations are very low, especially in insulin-dependent diabetics with negligible endogenous secretion. Although the authors found no evidence of ketosis up to the time of tracheal extubation, it is possible that ketosis, acidosis, catabolism, electrolyte abnormalities, glycosuria and dehydration would occur intermittently in the ensuing hours in patients who had undergone major procedures and were receiving no fluids orally (and hence their perioperative diabetic regimen). The risks of hypoglycaemia after the bolus would also be greater in the ward, where the request for 15-min blood glucose estimations would be highly optimistic. Inhibition of endogenous insulin secretion in the bolus group, as detected by reduction in …