Decrease of serum immunoglobulin M levels in some diabetic patients

Contradictory results have been published on serum immunoglobulin levels in diabetes. Our study population consisted of 26 "juvenile" IDDM patients (males/females 14/12, mean age 15.8 +/- 2.4 yrs), 42 "adult-onset" IDDM patients (25/17, 45.1 +/- 15.2 yrs), 62 NIDDM patients (27/35, 59.8 +/- 7.7 yrs), 128 controls. IgM has been measured by a highly standardized endpoint radial immunodiffusion. Since age and sex significantly influence serum IgM levels, we calculated Z values using the formula: log (Xobs:Xexp): SDexp, where Xobs is the measured IgM in any individual and Xexp and SDexp are the expected (geometric) mean and standard deviation of the log IgM values for each year of age in both sexes, as previously calculated by orthogonal polynomials from a large population of "laboratory controls" (n = 755; 10-70 yrs). These Z values (+/- SD) in juvenile IDDM (-0.442 +/- 0.988) and NIDDM (-0.559 +/- 1.215) were significantly lower (P = 0.035 and P less than 10(-3)) than in "laboratory controls" (0 +/- 1). Values in adult IDDM (-0.0225 +/- 1.213; P = 0.24) and in controls (-0.018 +/- 1.04) did not differ significantly from the "laboratory controls". The prevalence of diabetes within the four quartiles of the IgM distribution differed significantly from the one expected according to the null hypothesis (chi-square = 42.2; 3df; P less than 10(-4]. This is also applied to juvenile IDDM (P less than 0.05) and NIDDM (P less than 10(-4], but not to adult-onset IDDM. These results suggest that the low IgM levels may partly contribute to the poorly explained increase in susceptibility to infections in some diabetics.