Summary, Twenty-six Type1 diabetic patients previously treated for 10-20 months with twice daily conventional bovine isophane insulin (containing at least 1000 ppm proinsulin) were changed to highly purified (< 1 ppm proinsulin) bovine isophane for 6 months (Switch group). Insulin antibody levels fell significantly from a geometric mean of 14.9 to 9.1 l-tg/1. Thirty-two patients with newly diagnosed Type 1 diabetes were treated with the same highly purified bovine isophane insulin twice daily for 6 months (Starter group). Their insulin antibody levels rose from a geometric mean of 1.9 to 8.2 ~tg/1 in contrast to values of 1.4 rising to 16.3 ~tg/1 in an age and sex matched historical control group treated from diagnosis only with twice daily conventional bovine isophane insulin. Lipoatrophy at injection sites developed in three (9%) in the Starter group treated with highly purified bovine isophane compared to 7 (22%) of those on conventional bovine isophane. Insulin dose and diabetic control did not differ between the groups. Starter and Switch groups were subsequently treated with semi-synthetic human isophane insulin for 6 months during which insulin antibody levels fell significantly from a geometric mean of 8.5 to 4.4 ~tg/1 (p< 0.001). We conclude that bovine insulin purified to less than 1 ppm proinsulin is significantly less immunogenic than its conventional proinsulin contaminated counterpart but even at this level of purity is still more immunogenic than human insulin of equivalent purity.Key words: Immunogenicity, Bovine insulin, Lipoatrophy, Insulin antibodies, Human insulin Previous studies [1][2][3] have shown that bovine insulin purified by sequential recrystallisation to > 1000ppm proinsulin (conventional bovine insulin) is much more immunogenic that its highly purified ( < 1 ppm proinsulin) porcine counterpart. Not only are levels of insulin antibody higher but antibodies to C-peptide and other islet proteins are often present after using the conventional preparation [3][4][5][6]. We previously studied patients on conventional bovine soluble and isophane insulins who were transferred to equivalent bovine preparations purified to a proinsulin content of 20-40 ppm [7]. This significantly reduced antibodies reactive with C peptide but did not change insulin antibody levels. It left open the question of whether bovine insulin is inherently more immunogenic than pork or human because the "purified" bovine still contained significantly more proinsulin than the porcine with which it was compared. In the present study we have examined the immunogenicity of highly purified bovine isophane insulin (containing < 1 ppm proinsulin) and compared it with a conventional bovine isophane and semi-synthetic human isophane insulin.
Patients
Switch groupTwenty-six patients with Type 1 diabetes were treated from diagnosis 10-20 months earlier with twice daily conventional bovine (CB) isophane insulin only. They and the other patients reported in this paper had all been started on insulin as outpatients and had never rec...