Deficient production of interleukin-2 has been reported in Type I diabetes, but its cause has not been elucidated. We therefore measured interleukin-2 production in 27 patients with Type I diabetes, 20 patients with Type II diabetes (6 requiring insulin), 5 monozygotic twin pairs discordant for Type I diabetes, and 10 nondiabetic persons with islet-cell antibodies. Interleukin-2 production was decreased in patients with Type I diabetes as compared with controls (35.8±2.5 vs. 61.6±4.6 percent, P<0.001). Interleukin-2 production did not differ between patients with Type II diabetes and controls, regardless of whether the patients used insulin. Twins with Type I diabetes had decreased interleukin-2 production as compared with normal controls (33.2±5.4 vs. 61.6±4.6 percent, P<0.001) and with their nondiabetic twins (33.2±5.4 vs. 54.5±3.4 percent, P<0.005). lnterleukin-2 production in nondiabetic twins and in nondiabetic persons with islet-cell antibodies was normal. There was no correlation between glycosylated hemoglobin levels and interleukin-2 production in any diabetic group.

We conclude that patients with Type I diabetes have an acquired defect in interleukin-2 production, whereas patients with Type II diabetes do not, and that this defect is not correlated with an ongoing autoimmune process, with hyperglycemia, or with insulin administration or oral hypoglycemic therapy. Thus, the defect appears to be related to marked beta-cell destruction, although not to the metabolic consequences thereof or the responsible autoimmune process. (N Engl J Med 1986; 315: 920–4.)