Résumé : In experimental murine trypanosomiasis, resistance is often scored as the capacity to control peak parasitemia levels, which results in prolonged survival. Infection-induced pathology has not systematically been used as a resistance criterion. Because this parameter could be the most relevant for comparative analysis of natural and experimental infections, as well as for understanding of pathology-associated immune alterations, we analyzed Trypanosoma brucei infections in 4 different established conventional mouse models, as well as in tumor necrosis factor (TNF)-deficient and TNF-receptor-deficient mice. Results indicate the following: (1) there is no correlation between peak parasitemia control or survival and the induction of infection-associated anemia, loss of body weight, liver pathology, reduced locomotor activity, and general morbidity; (2) serum levels of TNF, interferon- gamma, and interleukin-10, which are known to affect survival, do not correlate with induction of pathology; and (3) infection-induced occurrence of lipopolysaccharide hypersensitivity does not correlate with survival. However, one parameter that was found to correlate with the inhibition of trypanosomiasis-associated pathology in all models was the shedding of soluble p75 TNF-receptor during peak parasitemia stages. These results are important for future cytokine and trypanosomiasis pathology studies, because the interplay between TNF and the soluble receptors it sheds has not been considered in either human clinical sleeping sickness studies or in veterinary trypanosomiasis research.