Résumé
The baryon density of our Universe, bh2 = 0.0224+/-0.0009, as inferred from the WMAP first year of observations is used to predict the primordial abundances of light elements produced during Big Bang Nucleosynthesis (BBN). Such a baryon density, and a gravitation described by General Relativity, lead to predictions for the abundances of 4He and D in very good agreement with the observed ones; but they lead to a significant discrepancy between the calculated and the observed 7Li abundances. Supposing that the standard non gravitational sector is not modified, we consider scalar-tensor theories of gravity, and study their impact on the 7Li abundance. It is shown that solving the lithium problem requires a non trivial behaviour of the expansion rate of the Universe at the epoch of BBN, in order not to violate the constraints from 4He and D. This kind of behaviour is obtained thanks to various models, both without and with self-interaction for the scalar field.