In this thesis, we have investigated the effects of the radiatively-driven microscopic diffusion of iron, carbon, nitrogen and oxygen in a typical $eta$~Cephei star. We thought that it was possible that microscopic diffusion could explain recent puzzling observations in some $eta$~Cephei stars, such as a wide range of observed frequencies ($ u$~Eri and 12~Lac), the existence of low metallicity $eta$~Cephei stars (observed in the SMC and the LMC), as well as hybrid $eta$~Cephei-SPB stars ($gamma$~Peg, $psi$~Cen), and unexplained carbon, nitrogen and oxygen abundance ratios ($delta$~Cet, $eta$~Cep, $xi^1$~CMa, V2052~Oph and to a lesser extent $ u$~Eri). In order to tackle the role of radiative forces and microscopic diffusion in $eta$~Cephei stars, we had to implement them in our stellar evolution code. In this process, we also had to add the effects of mass loss through stellar winds in order to remove surface abundance anomalies and numerical instabilities. We have shown that the radiative forces are able to sustain iron against gravity in $eta$~Cephei stars, that radiatively-driven microscopic diffusion is important in the external layers of $eta$~Cephei stars, and that it induces the accumulation of a significant amount of iron in the driving region of the pulsation modes, which is the iron convective zone at 200,000~K. This accumulation leads to an enhancement of the opacity and thus favors the $kappa$-mechanism responsible for the excitation of the pulsation modes. We have shown through parametric studies that indeed more modes become unstable. Our latest computations, involving a full evolutionary study, confirm the results of our parametric studies. This provides an explanation for the wide range of frequencies observed in some $eta$~Cephei stars. It can also explain the existence of the hybrid $eta$~Cephei-SPB pulsators, because the accumulation of iron broadens the instability strips for both the $eta$~Cephei and SPB stars. The exsitence of low metallicity $eta$~Cephei stars is also explained since microscopic diffusion can locally increase the iron in the driving region, creating at least a few unstable modes. Another important result from our work is that microscopic diffusion happens very early in the evolution of $eta$~Cephei stars, in fact as soon as the star is born. It would be interesting to check if the same is true for less massive stars, as it is usually assumed that they are homogeneous during the pre-main sequence. Our results for carbon, nitrogen and oxygen show that radiative forces could possibly explain the observed excess of nitrogen. They could offer a reasonable alternative to the usual argument of rotational mixing.