Some critical remarks on the problem of HLA association with disease.

The results concerning the association of HLA antigens with ITP, pernicious anaemia, hemangiomas and Alzheimer's disease are compared with those of other authors and the causes for differences are discussed. The different number of examined patients does not seem to be the main cause of those differences. Selecting of the correct method for statistical evaluation and the differences in the criteria for diagnosing the disease are more important. The unsubstantial differences in the homogeneity of the patients group and controls are negligible.