I am going to be REALLY unpopular here, but I'm going to say it anyway (because I am a pedant)!!!!!!!!!!! Antigens CANNOT be either heterozygous or homozygous; only genes can be heterozygous or homozygous. An antigen can be described as either showing homozygous expression, or heterozygous expression. That having been said, is a red cell sample that types as K+k- phenotypically, genotypically K/K or K/Ko, or even K/k, with a mutation within the Kell gene that prevents the k antigen being expressed and detected with all anti-k grouping reagents (just in case anyone doesn't believe me - we had one!). That's got that off my chest. Now then, there is NO doubt that there are some anti-K's around that only react with K+k- red cells (dosage), but they are fairly rare, however, how many people use antibody screening red cells that are K+k-? I doubt if there are any. Therefore, we are all ruling out anti-K using red cells with apparent K antigen heterozygous expression on every single sample that (apparently) has no atypical alloantibodies present. Am I wrong about this? It follows, therefore, that, over the years, there MUST have been occasions when a patient with a very weak anti-K (one that is only detected using red cells that are apparently showing homozygous expression) and who has been transfused with K+ blood (do the maths). As far as I know, there are no papers within the literature that report a case of either a delayed or an acute transfusion reaction as a result of this. Yes, this may cause the anti-K to become stronger (and, hence, be detectable using an apparent heterozygous red cell sample showing K+k+ expression), but then, if this happens, you give K- blood. So, my considered answer is that you can exclude using K+k+ red cells. I shall now go and lie down!!!!!!!!!!!!!