Leishmania donovani causes visceral leishmaniasis (VL), which is typically fatal without treatment. There is substantial variation between individuals in rates of disease progression, response to treatment and incidence of post-treatment sequelae, specifically post-kala-azar dermal leishmaniasis (PKDL). Nevertheless, the majority of infected people are asymptomatic carriers. Hamsters and mice are commonly used as models of fatal and non-fatal VL, respectively. Host and parasite genetics are likely to be important factors, but in general the reasons for heterogeneous disease presentation in humans and animal models are poorly understood. Host microbiota has become established as a factor in cutaneous forms of leishmaniasis but this has not been studied in VL. We induced intestinal dysbiosis in mice and hamsters by long-term treatment with broad-spectrum antibiotics in their drinking water. There were no significant differences in disease presentation in dysbiotic mice. In contrast, dysbiotic hamsters infected with L. donovani had delayed onset and progression of weight loss. Half of control hamsters had a rapid progression phenotype compared with none of the ABX-treated animals and the nine-month survival rate was significantly improved compared to untreated controls (40% vs. 10%). Antibiotic-treated hamsters also had significantly less severe hepatosplenomegaly, which was accompanied by a distinct cytokine gene expression profile. The protective effect was not explained by differences in parasite loads or haematological profiles. We further found evidence that the gut-liver axis is a key aspect of fatal VL progression in hamsters, including intestinal parasitism, bacterial translocation to the liver, malakoplakia and iron sequestration, none of which occurred in non-progressing murine VL. Diverse bacterial genera were cultured from VL affected livers, of which Rodentibacter was specifically absent from ABX-treated hamsters, indicating this pathobiont may play a role in promoting disease progression. The results provide experimental support for antibiotic prophylaxis against secondary bacterial infections as an adjunct therapy in human VL patients.