Detection of insulin antibodies by radioassay and ELISA: interrelation and correlation with metabolic control in type I diabetes.

1. The literature suggests that the radioassay (RA) and ELISA detect different types of insulin antibodies (IA) (Wilkin et al., 1989. Diabetes, 38: 172-181). 2. In the present study we evaluated the relationship between these two antibodies and their involvement in the metabolic control of Type I diabetic (DMI) patients. 3. IA were measured by RA and ELISA in sera obtained from 34 patients (age: 9-16 years, median = 12.5 years; clinical duration of DMI: 0.1-11.0 years, median = 1.7 years) treated with different types of insulin [purified (bovine + porcine) N = 18, and monocomponent (porcine or human) N = 16] and submitted to various degrees of metabolic control as assessed by glycosylated serum protein (GSP) levels: range, 3.4-13.5%; median = 8.7%; normal value, 0.8-2.4%. 4. Insulin antibody levels measured by RA were: 3264 +/- 300 nU/ml (mean +/- SEM, normal value < 60 nU/ml) and by ELISA: 0.74 +/- 0.11 ELISA index (EI) (normal value, < 0.53). No correlation was found between IA levels measured by RA and ELISA, or between duration of the disease or insulin daily necessity and IA by either method. GSP was positively correlated with IA determined by ELISA (rS = 0.43, P < 0.01) but not with IA determined by RA. 5. The patients on purified bovine + porcine insulin had higher titers of IA by ELISA, compared to those of patients on monocomponent (0.96 +/- 0.15 vs 0.50 +/- 0.13 EI, P < 0.03, while IA levels measured by RA did not differ between groups. 6. These data show that RA or ELISA assays provide different serum titers of IA in insulin-treated diabetics and data obtained with ELISA correlated best with the metabolic control of Type I diabetic patients.