Plasmodium falciparum malaria, as ancient as hominid evolution itself, has provoked more change within the human genome than any other pathogen. JBS Haldane observed the overlapping distributions for thalassaemia and malaria endemicity and proposed ‘balanced polymorphisms’ as advantageous heterozygous mutant states. We now appreciate the wider range of haemoglobinopathies, membranopathies, and enzymopathies as distinct evolutionary adjustments to the erythrocyte, the very compartment which P. falciparum hijacks to sicken the host. Unlike other Plasmodium species, P. falciparum’s power over the erythrocyte consists of its limitless red cell infectivity and its capacity to render the infected red blood cell (iRBC) adhesive enough to arrest in the circulation. This latter cytoadhesivity is achieved by sticky knob proteins known as ‘Plasmodium falciparum erythrocyte membrane protein-1’ (PfEMP-1), trafficked to the red cell exterior from the parasite within. PfEMP-1 is designed to latch onto endothelial cells of the post-capillary venules (‘sequestration’), as well as onto other uninfected red blood cells and platelets (‘rosetting’). In so stalling their flow towards the spleen, the iRBC doubly harms the host by resisting the first defence of reticuloendothelial clearance and congesting the host’s microvasculature. The youngest, most malaria-naïve suffer malaria’s highest case fatality rates, revealing just how critical this innate (pre-adaptive) immune control of parasitaemia is. The biochemical means by which PfEMP-1 achieves its cytoadhesive promiscuity is in part through one particular lectin-like domain, DBL1α. This domain binds not only to heparan sulphate-like glycosaminoglycans, but to two blood group antigens expressed densely on erythrocytes: the group A carbohydrate in the ABO system and antigens (including those of the Knops system) on CR1 (CD35). If indeed these ligands are critical in the molecular pathogenesis of malaria fatalities, then we might expect to observe non-adhesive variants ascending to higher prevalence in the most malaria-endemic parts of the world. The cytoadhesivity of wildtype group A hosts is theoretically, and in vitro, demonstrably mitigated by what we now know are the mutant phenotypes which define the polymorphisms of the ABO system. These include the group O or B alleles, the weaker A types and the genetics influencing the quantity of secreted (competitive) free A antigen in group A hosts. Each of these phenotypes is observed at higher frequencies in malaria-endemic areas. Certain CR1 polymorphisms are also more frequently found in these parts of the world. The assembly of in vitro, geographic and clinical evidence weighs heavily towards ABO evolution being a highly specific response to P. falciparum. Rather than bypassing invasion or enhancing clearance, these mutations are special because they highlight the importance of escape from cytoadhesion. Forthcoming are the results of the first prospective study powered to confirm the impact of ABO on malaria mortality (NCT 00707200, http://www.clinicaltrials.gov). Should the results of emerging studies confirm a survival advantage among group O individuals, the basic strategy for transfusion support in malaria may shift to greater use of group O red cells. The clinical value of this approach, more immediately available than any new drug or vaccine development, will need to be tested in clinical trials.