The different serological tests used for leprosy are firstly the methods for the detection of antibodies (anti-PGL1, 35kD, 36kD, LAM), and secondly, the tests to detect the PGL1 antigen from the serum or urine. The antibody detection tests have a good but insufficient specificity for the diagnosis of leprosy patients and their sensitivity is generally high for the multibacillary patients but low for the paucibacillary patients. Their positive predictive value for the diagnosis of patients in a population are very low: 2.1% for the anti-PGL1 ELISA when the prevalence is 1/1000. For the early diagnosis of patients and the follow up of high risk populations, these tests are not cost effective: the number of patients detected in these populations is 10 fold lower than in the general population and the relative risks for developing the disease are not different among seropositive and among seronegative groups. In treated multibacillary patients, the IgM anti-PGL1 level decreases in correlation with the decrease of the bacillary index. For the diagnosis of M. leprae infection in a population, there was no correlation between the anti-PGL1 seroprevalence and the prevalence of the disease. Concerning the PGL1 antigen detection tests, they are specific and sensitive for the diagnosis of multibacillary patients but they cannot be used in routine for technical reasons. In conclusion and to date, the usefulness of serological tests in a leprosy control programme is quite questionable.