Influence of human insulin on symptoms and awareness of hypoglycaemia: a randomised double blind crossover trial.

OBJECTIVE To investigate the apparent increased risk of severe hypoglycaemia associated with use of human insulin by comparing the pattern of symptoms of hypoglycaemia with human insulin and porcine insulin. DESIGN Randomised controlled double blind crossover trial of treatment with human insulin and porcine insulin, with two treatment periods of six weeks. SETTING Diabetes outpatient department of a university teaching hospital in Berne, Switzerland. PATIENTS 44 patients (25 men, 19 women) aged 14 to 60 years, with insulin dependent diabetes mellitus. All patients met the following criteria: receiving treatment with fast acting soluble insulin and long acting protamine insulin; performing multiple daily fingerstick blood glucose self measurements; and had stable glycaemic control with about one mild hypoglycaemic episode a week during the preceding two months. INTERVENTION Patients were randomised to receive either human or porcine insulin for six weeks and were then changed over to the other type of insulin for a further six weeks. MAIN OUTCOME MEASURE Questionnaire recording "autonomic" and "neuroglycopenic" symptoms that occurred during hypoglycaemic episodes confirmed by a blood glucose concentration less than or equal to 2.8 mmol/l. RESULTS Insulin doses and blood glucose, glycated haemoglobin A1c, and fructosamine concentrations were similar during the two treatment periods. 493 questionnaires on hypoglycaemia (234 during treatment with human insulin and 259 during treatment with porcine insulin) were analysed. With human insulin patients were more likely to report lack of concentration (52% v 35%, p = 0.0003) and restlessness (53% v 45%, p = 0.004) and less likely to report hunger (33% v 42%, p = 0.016) than during treatment with porcine insulin. The difference in the pattern of symptoms during the two treatments was similar to that between the 12 patients with a history of recurrent hypoglycaemic coma and the 32 patients without such a history. CONCLUSIONS The pattern of symptoms associated with human insulin could impair patients' ability to take appropriate steps to avoid severe hypoglycaemia. Caution should be exercised when transferring patients from animal insulin to human insulin, and a large scale randomised trial of the two types of insulin may be justified.