Evaluation of Patients With Local Reactions to Insulin With Skin Tests and In Vitro Techniques

Abstract

To better understand the part played by IgE and IgG antibody in the production of dermal reactions to insulin and the usefulness of skin tests in the evaluation of these reactions, we studied 21 diabetic patients referred for evaluation of large local insulin reactions, 46 diabetic patients without local insulin reactions, and 22 healthy nondiabetic controls. Study subjects were skin tested with 15 different insulins, and the results were evaluated over 48 h. All control subjects and 41 of 46 diabetic patients without local reactions were skin-test negative to insulin. The 11% of diabetic patients who reacted had positive wheal-and-flare reactions at 20 min to animal-species insulin but negative skin tests to human insulin. Study revealed two subgroups of patients with histories of local reactions. Ten (48%) of these patients had negative skin tests to insulin. The 11% of diabetic patients who reacted had positive wheal-and-flare reactions at 20 min to animal-species insulin but negative skin tests to human insulin. Study revealed two subgroups of patients with histories of local reactions. Ten (48%) of these patients had negative skin tests to insulin. Five of this subgroup remained skin-test negative to quantities of ≥8 U insulin/skin test. Eleven (52%) of the patients formed a subgroup with positive insulin skin tests; most of these patients were skin-test positive to human insulin and to beef, pork, or both insulins as well. Although the group mean insulin-specific IgE values of this latter subgroup were significantly higher than those of any other study group, overlap of these individual IgE values did not allow separation of specific individuals with positive skin tests from those of patients on insulin without dermal reactions. IgE-to-IgG insulin-specific antibody ratios in the subgroup with histories of reactions but negative skin tests were lower (0.72—0.86) than those of the patient subgroup with positive skin tests (1.98—2.36) or diabetics without reactions (0.98–1.61). We hypothesize that patients in the subgroup with positive skin tests had a primary IgE response to insulin, whereas patients in the subgroup with negative skin tests were undergoing a change from a predominant IgE response to a predominant IgG response to insulin. We conclude that skin testing with 1 U insulin/test is a simple, sensitive, and cost-effective method to detect clinically meaningful levels of insulin-specific IgE associated with large local reactions. Furthermore, up to half of the patients referred for specialized evaluation of large local reactions will have unreproducible reactions and negative skin tests that appear to reflect a process of spontaneous desensitization.