A practical, sensitive and specific method for the assay of antibodies to insulin in human plasma has been developed. This method involves competition for insulin antibody between a constant, trace amount of I-125-insulin and increasing concentrations of unlabeled insulin standard. The insulin standard is prepared from the soluble form of therapeutic insulin that the patient has been receiving. The quantity of antibody is calculated arbitrarily from the amount of unlabeled insulin required to displace 25 per cent of bound I-125-insulin from the antibody. The result is expressed as microunits of insulin antibody per milliliter of plasma. Since this is not the absolute amount of antibody present but does indicate a relative, quantitative immunologic response of each patient to the antigen, insulin, the result is referred to as an antibody index. The method has a precision of 20 per cent and sensitivity of 1 μU./ml. The coefficient of variation is significantly less at higher antibody indices: 8 per cent in the 1,001 to 80,000 range and 22 per cent in the 0.08 to 40 μU./ml. range.
Therapeutic subcutaneous doses of Lente or Sulphated insulin do not interfere with the antibody assay, whereas the interference from Regular insulin continues for several hours. The antibody index declines after withdrawal of insulin therapy and has a half-life of approximately eighteen days. There is no diurnal variation in antibody index. Three newborn infants of insulin-treated diabetic mothers were found to have insulin antibodies.
