Intermediate energy (>80 MeV/nucleon) knockout reactions are a key probe of structure in highly exotic nuclei. The longitudinal momentum distribution of the heavy reaction residue are characteristic of the angular momentum of the residue states, whilst absolute cross sections probe spectroscopic strengths. Extraction of this structure information from experiment relies on reaction models, most often based on the spectator-core approximation and eikonal (straight-trajectory) reaction dynamics. Calculations for two-nucleon knockout also make the no-recoil approximation, assuming the core of the projectile is sufficiently massive that it can be fixed at the projectile centre-of-mass. This reduces the computation time required by three orders of magnitude. However, in single-nucleon removal this approximation leads to an overestimate of the elastic breakup cross section by ~20-30%, and simple estimates have suggested the overestimate of the diffraction-stripping component in two-nucleon removal could be as much as a factor of two. Here we will discuss the impact of core-recoil on two-nucleon knockout, and present new calculations that do not use the no-recoil approximation. The net effects of including core recoil result from a reduction of breakup through nucleon diffraction, offset by an increase in breakup via core diffraction. These results give improved agreement with experimental data on the relative strengths of elastic and inelastic removal mechanisms for 28Mg(−2p), albeit within large experimental uncertainties.