Background: The establishment of an accurate diagnostic protocol for canine visceral leishmaniosis (CanL) is a significant laboratory challenge and the lack of a reliable reference standard is one of the major problems. The aim of this study was to compare in situ hybridization (ISH), immunohistochemistry (IHC) and parasitological culture (PC) for detection of L. infantum in skin, spleen, lymph node and bone marrow of clinically healthy and sick seropositive dogs. Findings: The study included 65 dogs positive with both DPP® and ELISA for anti-Leishmania antibodies. In situ hybridization of spleen or lymph node had the highest positivity rates of L. infantum detection. The total positivity rates for IHC, ISH and PC were 70 %, 68.1 % and 65.8 %, respectively. When combining techniques, the positivity rates were 81.5 % in the spleen, 79.0 % in lymph nodes, 59.0 % in bone marrow and 52.3 % in the skin. The highest percentage of infected dogs (87.7 %) was detected by using lymph node samples. When examining only skin, positivity was significantly higher in sick dogs than in the clinically healthy dogs. Infection with L. infantum was confirmed in 95.8 % of sick dogs and in 82.4 % of healthy dogs. Conclusions: Considering the advantages of accurately diagnosing different Leishmania species and of being more sensitive than PC, ISH should be considered as reference standard test for the diagnosis of CanL. Spleen and lymph node are the most suitable tissues to confirm infection with L. infantum in seropositive dogs. The testing of only skin from clinically healthy dogs may result in a high percentage of false negative results.