Bias in association studies of systemic lupus erythematosus susceptibility due to geographical variation in the frequency of a programmed cell death 1 polymorphism across Europe

We obtained eight collections of DNA samples from ethnically matched systemic lupus erythematosus (SLE) patients and controls from five European countries totaling 783 patients and 1210 controls. A highly significant cline in the frequency of the PD1.3 A allele was found among controls but not among SLE patients. The frequency of the PD1.3 A allele increased from the Northeast to the Southwest of Europe. The cline was clearly apparent (P=1.2 × 10^-6) when data from controls of other five SLE susceptibility studies were included in the analysis. This variation has severely biased SLE association studies owing to the lack of parallel changes in SLE patients. As a consequence, the PD1.3 A allele was more common in SLE patients than in controls in the Northeast and Center of Europe, similar to controls in Southeast Europe, and less frequent than in the controls in the Southwest of the Continent. This dissociation in allele frequencies between SLE patients and controls in different subpopulations indicated that programmed cell death 1 variation and disease susceptibility are not independent but the type of relationship is currently unclear. As allele frequency clines are common in other polymorphisms their impact in genetic epidemiology studies should be carefully considered.