One component of control programmes to eliminate trachoma is the treatment of Chlamydia trachomatis infection. A diagnosis of trachoma is based on clinical grounds, but the signs of active trachoma do not always correlate with the presence of C. trachomatis. During a therapeutic trial, the level of C. trachomatis infection in children with active trachoma in Guinea and Pakistan was assessed using a qualitative commercially available PCR that targeted the C. trachomatis plasmid. The influence of the quality of specimens on the efficiency of the PCR was investigated using two quantitative real-time PCRs targeting the specific omp1 gene of C. trachomatis and human chromosomal DNA, respectively. C. trachomatis was detected in c. 23% of children (aged 1-10 years) who presented with clinically active trachoma. Controls showed that PCR-related problems did not influence this detection rate. For 14% of the positive samples, C. trachomatis was detected in only one eye, with a significantly lower mean load of bacteria. These results suggest that epidemiological and therapeutic surveys should be conducted by sampling and testing both eyes. Moreover, the high variability of the cell load observed in the conjunctival swabs suggests that the effectiveness of swabbing may be questionable.