To investigate if the core population hypothesis is applicable to patients with genital chlamydia infections. Retrospective cross sectional study. Two genitourinary medicine (GUM) clinics in the city of Birmingham and eight adjacent clinics. All patients with chlamydia (n = 665) or gonorrhoea (n = 584) attending between 1 October 1995 and 30 September 1996 with a postcode within the Birmingham health district. Controls were 727 patients seen in the same period with no infection. Postcodes were used to calculate population prevalence rates per 100,000 aged 15-65 in the 39 wards of the city and to estimate the socioeconomic status using the Super Profile (SP). Ethnic specific rates were also calculated. Data were obtained on gonorrhoea and chlamydia isolation from all the major laboratories of the city over the same time period. GUM clinic attenders accounted for 67.6% and 82.5% of all chlamydia and gonorrhoea isolates reported by the laboratories and that were available for our epidemiological analysis. Both infections were more common in men and in black ethnic groups. However, patients with gonorrhoea only infection were more likely to be of black ethnicity than those with chlamydia only infection (p = 0.0001) and to have different SP distribution (p = 0.0001). On logistic regression age < 20 years, male sex, black ethnicity, and living in neighbourhoods with SP J ("have nots") were predictive of both infections compared with controls. Overall chlamydia and gonorrhoea prevalence rates were 129 and 98.4 per 10(5) respectively. Corresponding rates for whites was 64.7 and 37.2 and for black ethnic groups 1105 and 1183 per 10(5) of each ethnic group. Eight adjacent wards accounted for 41% of the chlamydia and 66.5% of the gonorrhoea. In a large urban setting patients attending GUM clinics with chlamydia belong to core population groups with similar, but not identical, sociodemographic characteristics to patients with gonorrhoea infection.