Attempts to associate ABO and other blood types with various disease states have resulted in a large and somewhat controversial literature.1-6 The necessary requirement of large numbers of patients and appropriate control data has been difficult to meet except for the more common disease states. The review articles of Roberts7 and of Clarke8 clearly point out the pitfalls of drawing statistical conclusions from small series of patients.
Recently Fadhli and Dominguez9 reported finding an increased incidence of blood type A among 125 white patients who had multiple primary cancers. It was suggested that the probability of a second primary malignant tumor developing in these patients with blood type A could be considered greater than it was in patients whose blood type was 0, B, or AB. These authors further commented, however, on the need of larger series to confirm their findings. Tsukada and associates10 were