I would think that the main issue you would run into would be engine stoichiometry rather than aero... lower density would negatively affect the engine performance obviously.
But we are talking air density fluctuations somewhere in the region of <10% variance, ignoring other factors such as relative humidity and particulate content etc...
In terms of the theoretical equations that apply, both the drag and downforce equations vary proportionally with density. There is one way that you would swing the equations to imply a lower density is better though... There is a lot of confusion regarding whether to use planform or frontal area for force calculations. In aerospace, planform is used for lift whilst in racing what is typically done, is to use frontal area for both. In some racing teams (including F1 from my experience) uses a "fixed" area constant value to calculate them; this is due to the fact that with aerofoil pitch and rake angle changes, along with others, the actual frontal area changes and it is simpler to say that "this tweak gives x-many points of additional downforce" and ignore the area all together than to recalculate everything.
By this logic, if you were to swap to using frontal area for drag and planform area for lift, given that planform is usually much larger than the frontal area, you could argue that a lower density air would be better for aerodynamic efficiency (L/D) than a higher one, excluding all other factors. Whether this solution actually occurs in reality, I can't say off the top of my head... Within CFD, forces are typically calculated by summation of all of the pressure vectors on a surface and then the components in line with the specified coordinate system correspond to the lift and drag figures... i.e. the two figures for lift and drag are generated from one "magnitude" of pressure which is calculated using either coupled or uncoupled equations for velocity, pressure and density (or energy) and "not" area... specifically