Definition of normal autoantibody levels in an apparently healthy population

Clinically asymptomatic women with laboratory abnormalities in autoantibody profiles have recently been shown to experience reproductive failure. The evaluation of autoantibody panels and the establishment of normal cutoff values in clinically apparently healthy populations has therefore achieved increasing importance. Investigation of 400 clinically asymptomatic patients (200 females and 200 males) for the presence of autoantibodies to six phospholipids, five histone subfractions, and four polynucleotides revealed a nonparametric distribution of autoantibodies primarily for antiphospholipids and antihistone antibodies. This observation suggests that widely used parametric methods for the determination of normal autoantibody levels are inadequate and will give an unreasonably high incidence of abnormal results. Consequently, rather than the widely used 95% confidence interval, we used the 99% confidence interval (based on medians) to determine the upper limit of normal for various autoantibodies. This resulted in the detection of four to 13 positive patients (out of 400) per antigen, for a positivity rate of 1-3%. A more rigid definition of normal and abnormal autoantibody levels is essential to pursue accurately diagnostic and therapeutic considerations concerning the newly evolving concept of subclinically abnormal autoimmunity.