The incidence of primary biliary cirrhosis (PBC) varies widely between regions. However, little is known about variation within regions and the degree to which this may reflect environmental risk factors. The aim of this study was to describe the spatial distribution of cases of PBC in a defined region of Northeast England over a defined period, and to assess the magnitude of any departure from random spatial distribution. Seven hundred seventy patients with established PBC were identified in a previous comprehensive case finding study. A total of 3,044 control locations were randomly selected from postcode (zip code) data weighted for number of drop off points per postcode. Geographical analysis was performed by testing both for spatial variation in risk and local clustering by using previously described point process methods. Both tests used the same null hypothesis that risk of disease does not vary spatially and cases occur independently of each other. Statistically significant spatial variations in risk were found in the whole study region (P <.001) and in the major urban area within the region (P <.004). Risk was higher in the urban area of Tyneside than in the surrounding rural area. Within the rural area, spatial variation in risk was equivocal (P =.012), but there was significant (P =.001) clustering of cases (estimated average cluster effect approximately 10 excess cases within a 7-km radius). PBC occurred to a density of 10.7 cases/km(2) in the highest risk areas. In conclusion, PBC is unevenly distributed in Northeast England. This may reflect one or more environmental risk factors in its etiology.