A fraction with a major band of 14kDa was obtained from crude cyst fluid of Taenia solium cysticerci by 2-step chromatography. A first fraction isolated by gel filtration (Sephacryl S-300 high resolution) was purified using an anion exchange column (Mono Q HR 5/5) on high performance liquid chromatography. Evaluation of the analytic sensitivity of this fraction (F3) was carried out in an antibody enzyme linked immunosorbent assay (Ab-ELISA-F3) using serum samples from pigs experimentally infected with different doses of T. solium eggs. The cross-reactivity of F3 was evaluated with serum samples from pigs that were naturally or experimentally infected with Taenia hydatigena, Taenia saginata asiatica, Fasciola hepatica, Trichinella spiralis, Metastrongylus apri, Trypanosoma congolense and Sarcoptes scabiei, and with serum samples of rabbits hyper-immunised with metacestode cyst fluid of T. hydatigena and T. solium. Antibody titres of lightly or heavily infected pigs differed in their kinetics. However, the increase in F3-specific antibodies could not be related to the infection level. Analysis of the specificity of the F3 showed that serum samples of pigs infected with other parasites did not recognise this antigen. Cross-reaction with T. hydatigena occurred in ELISA using cyst fluid as antigen, but the F3 antigen fraction was not recognized by rabbit hyper-immune serum samples to T. hydatigena. Evaluation of the diagnostic sensitivity and specificity of the Ab-ELISA-F3 was done by a non-parametric receiver operating characteristic (ROC) analysis using 66 serum samples from Zambian village pigs. The total number of cysticerci of these pigs was determined by dissection (28 pigs harboured T. solium cysticerci and 38 were negative at dissection). In addition, 58 serum samples from Cameroonian pigs (28 pigs from cysticercosis-free farms and 30 pigs with cysticerci at tongue inspection) were used in a separate ROC analysis. The results from the ROC analysis yielded a low diagnostic value (area under ROC curve=0.48) with the sera from the Zambian pigs while a relatively high diagnostic value was obtained with the sera from Cameroonian pigs (area under ROC curve=0.78). The main factor contributing to a low diagnostic value based on the Zambian serum samples seemed to be the false-positive reactions that were likely caused by the occurrence of transient antibodies in the non-infected animals.