This study sought to determine whether a specific polymerase chain reaction (PCR) for enterotoxigenic Escherichia coli (ETEC) toxins after chaotropic extraction of DNA from stool would increase the detection of ETEC over that of conventional oligonucleotide probe hybridization of 5 E. coli colonies per stool sample (a standard method). By DNA hybridization, 29 (21%) of 140 patients were positive for ETEC, and 59 (42%) of 140 were positive for ETEC when PCR was used. Sensitivity of the PCR assay was confirmed through spiked stool experiments to be approximately 100-1000 ETEC colonies per sample. Specificity of the assay was determined by showing an absence of ETEC by the PCR technique in a subgroup of 48 subjects and by confirming the presence of ETEC DNA of positive samples by dot blot procedure. PCR technique detected significantly more ETEC infections in these subjects than did the hybridization method (P<.0001).