Insulin autoantibodies and insulin assay.

Insulin antibodies are known to interfere with the radioimmunoassay of insulin. We tested intravenous glucose tolerance on 25 insulin autoantibody-positive (IAA+) patients and 25 IAA- controls, who were matched for sex, age, and body mass index, to establish if IAA could also interfere with insulin assay. Insulin content was measured in untreated serum, serum precipitated with polyethylene glycol (PEG, free insulin), and serum extracted with acid and precipitated with PEG (total insulin). The mean untreated first-phase insulin response (I1 + 3) for IAA+ patients was 172 +/- 67.3 mU/L, significantly higher than the mean control value of 108 +/- 47.5 (P less than .001). After PEG precipitation, mean I1 + 3 in the patient group fell significantly to 105 +/- 48.4 mU/L (P less than .001), but the control value was unchanged (104 +/- 45.5). The mean percentage fall after PEG precipitation was 36.9% (patients) and 2.9% (controls) (t = 8.3, P less than .001). There was a strong correlation between the IAA titer and the interference in the insulin assay (r = .81). After total insulin extraction of IAA samples, there was a significant fall in mean I1 + 3 to 134 +/- 55.4 mU/L (P less than .001), but the control value was unchanged. IAA can significantly falsify insulin measurement, and care must be taken in the interpretation of insulin-release tests when IAA is present.