The association between HLA antigens and ageing is not clear. Ageing in women was associated with B40 and DR5 in a recent study, but other studies yielded conflicting results. In none of the studies, however, did the young and elderly samples originate from the same homogeneous population. Homogeneity is dependent on geographic origin. The aim of this study was to investigate whether differences in geographic origin between age groups could explain the age-associated differences in the frequencies of B40 and DR5. The authors used the new design of a 'birth-place-restricted comparison' in which the origin of all subjects was ascertained. The total study population comprised 1010 young women aged 25-40 years and 660 elderly women aged 85 years and older. The 'birth-place-restricted comparison' included 66 young and 285 elderly women from one geographic area (Leiden, the Netherlands). Men were not included because ageing in men was not associated with HLA antigens in a recent study. In the total population, the frequency of B40 in young women of different origin varied between 16 and 28%, and the frequency of DR5 between 11 and 23%. Similar differences were observed in the elderly women. In the 'birth-place-restricted comparison', the frequency of B40 was 15% in the young women and 11% in the elderly women (difference 4%, 95% confidence interval, -5 to 13%). The frequency of DR5 was 20% in the young women, and 28% in the elderly women (difference 8%, 95% confidence interval, -4 to 19%). Thus, marked differences in HLA antigen frequency were found between populations of various geographic origins. Definition and ascertainment of the target population are therefore necessary in genetic studies of ageing. In such a 'birth-place-restricted comparison', the authors confirmed that ageing in women was negatively associated with HLA-B40 and positively associated with HLA-DR5.