SummaryWe compared a strip test employing recombinant K39 (rK39) antigen and protein A/colloidal gold as read-out agents with the rK39 ELISA for IgM and IgG antibodies and the direct agglutination test (DAT) using 55 sera from patients with parasitologically con®rmed visceral leishmaniasis (VL). The rK39 strip test was positive in 37/55 (67%), the DAT in 50/55 (91%) at ³ 1 : 1600 cut-off value and in 47/55 (85%) at ³ 1 : 6400 cut-off value. The rK39-ELISA gave positive IgG results for all sera; those who had a positive strip test had signi®cantly higher IgG levels than those with a negative strip test (31.1 (SD 3.6) and 17.7 U/ml (SD 9.8), respectively, P < 0.0001). A total of 31/55 (56%) sera showed a positive IgM result; of these 27 (49%) had a positive strip test. We tested 115 apparently cured VL patients with the strip test during follow-up; 68 were also tested with DAT. In the strip test, 25±43% of patients had a positive result at time points 3, 6, 9 and 12 months after treatment; for DAT (cut-off ³ 1 : 1600) these results were 67±83%. In neither test did a signi®cant decrease in positivity rates occur over time (P 0.37 for the strip test, P 0.17 for the DAT). No correlation (P 0.33) was found between a positive strip test and a positive DAT result (cut-off ³ 1 : 1600), indicating that the strip test and DAT are complementary rather than interchangeable. Of 61 endemic controls two (3%) had a positive strip test result; both had a positive leishmanin skin test. The rK39 strip test has the ideal format for use in the ®eld, but its sensitivity is limited; like DAT, but to a lesser extent, it remains positive after treatment.keywords visceral leishmaniasis, diagnosis, rK39 strip test, Sudancorrespondence Professor E.