A practical approach to the indirect diagnosis of Mycobacterium avium subsp. paratuberculosis infection in cattle is performed by ELISA test. Nevertheless, ELISA has been criticized as not being sufficiently sensitive because gave positive results after bacterial excretion in feces has began. A commercially available indirect ELISA test (IDEXX) allows to discriminate serum samples on the basis of three possible results: negative (net sample OD ( 15% of net positive control OD), positive (net sample OD ( 30% of net positive control OD) and doubtful (net sample OD values ranging from 15 to 30% of net positive control OD). To improve test sensitivity and to differentiate doubtful animals as positive or negative, an anamnestic ELISA was applied. At the aim, 300 animals aged 2 - 4 years, of which 100 medium-strong-negative (net sample OD ( 5% of net positive control OD), 100 weak-negative (net sample OD values ranging from 5 to 15% of net positive control OD) and 100 doubtful, as detected by ELISA, were intradermally inoculated with johnin. In addition, 300 animals aged 2 - 4 years, of which 100 medium-strong-negative, 100 weak-negative and 100 doubtful were kept as non-inoculated controls. After 20 days post-inoculation, sera were collected and subjected to ELISA (anamnestic ELISA). 54 doubtful and 20 weak-negative animals became positive after johnin, otherwise no medium-strong-negative animals seroconverted. Only 22 doubtful and 4 weak-negative controls became spontaneously positive at second sampling. Statistical analysis (chi-square) pointed out a significant difference of seroconversion rate between johnin-treated and controls for either doubtful or weak-negative animals. Therefore, anamnestic ELISA may be considered a promising tool to improve serologic detection of paratuberculosis infection in cattle.