The aim of this study was to find suitable and reliable tools for demonstrating Lawsonia intracellularis in routine clinical diagnosis. Firstly, a method to prepare tissue samples before a polymerase chain reaction (PCR) was evaluated in pigs submitted for necropsy. Secondly, seven different faecal preparation methods and four different DNA polymerases were tested in single or nested PCR, with co-amplification of a mimic molecule. Thirdly, in selected pigs submitted for necropsy, tissue and faecal samples were examined histopathologically and by PCR, and blood samples were analysed serologically. Detection of L. intracellularis in tissue preparations by PCR showed good specificity and correlated to lesions found at necropsy. The sensitivity in spiked tissue samples was 10(1)-10(2) mimic molecules per tube. In faecal samples, nested PCR on boiled lysate gave the best result with a sensitivity of 10(2)-10(3) mimic molecules per reaction tube. However, because of the time-consuming procedure and the increased risk for contamination, a commercially available kit was preferred for routine diagnoses, despite a somewhat lower detection rate in subclinically infected pigs. In a few cases, the serological results differed from those obtained by PCR and by necropsy but the reason for this is not clear. This study indicates that the best method for diagnosis of acute enteritis in growers is PCR on faecal or tissue samples. To determine the presence of the bacteria in a herd, serology or repeated faecal sampling for PCR from target animals, or both, should be used.