Human African trypanosomiasis (HAT) or sleeping sickness is a major public health problem in sub-Saharan Africa and is due to the kinetoplastid parasite Trypanosoma brucei gambiense in West and Central Africa. The exact role of multiple infections, the basis of clinical diversity observed in patients and the determinism that leads trypanosomes into different body fluids of the host remain opened questions to date. In this paper we investigate, in three Guinean foci, whether strains found in blood, lymph or cerebrospinal fluid (CSF) or in patients at different phase of HAT (phase 1, early phase 2 and late phase 2) are representative of the focus they belong to. Amplifications of parasites directly from body fluids led to substantial amounts of allelic drop outs, especially so for blood and CSF samples, which required data recoding of all homozygous sites into missing data. While controlling for geography, date of sampling and patient's phase of the disease, we found no effect of body fluids in the genetic structure of T. b. gambiense despite the presence of mixed infections. On the contrary, we found that the strains found in patients in different phase of the disease differed genetically, with early phase patients being more likely to be infected with more recent strains than patients at a more advanced phase of the disease. Thus, the combination of date of sampling and patient's status represents a parameter to be controlled for in population genetic structure analyses. Additional studies will also be required to explore further the phenomenon of mixed infections and its consequences.