The first observations on associations between blood groups and infectious diseases were made in the 1950s, but the underlying mechanisms were not elucidated for many years. This could have been due to limited explanations for the epidemiological findings or to conflicting reports of associations between different blood groups with the same disease. An example of the latter is the large numbers of papers on Helicobacter pylori and ABO or Lewis blood groups/secretor status during the past few years which have reported inconsistent or conflicting results. Because determination of blood groups is a relatively simple and inexpensive procedure, many investigators have used it for quick “simple” studies without consideration of possible confounding factors. For all studies on blood groups and infection, the following points (gained with the experience of hindsight) need to be considered in planning or assessment of surveys:
The disease or organism under investigation needs to be clearly defined. Severity of the symptoms should be also be considered, e.g., differentiation of cases of Escherichia coli O157 infection between patients with uncomplicated diarrhoeal disease and those that develop haemolytic uraemic syndrome (HUS) (Blackwell et al., 2002).
It should be made clear that the investigation examined an outbreak or defined epidemic due to a particular strain in contrast to sporadic cases which could be due to strains with different antigenic characteristics or virulence factors.
Different populations express different quantities of antigens such as H, Lewisa, or Lewisb.