A two-body system hypothetically affected by an additional radial acceleration Hv r , where v r is the radial velocity of the binary's proper orbital motion, would experience long-term temporal changes of both its semimajor axis a and the eccentricity e qualitatively different from any other standard competing effect for them. Contrary to what one might reasonably expect, the analytical expressions of such rates do not vanish in the limit M → 0, where M is the mass of the primary, being independent of it. This is a general requirement that any potentially viable physical mechanism able to provide such a putative acceleration should meet. Nonetheless, if H had the same value H 0 of the Hubble parameter at present epoch, such rates of change would have magnitude close to the present-day level of accuracy in determining planetary orbital motions in our Solar System. However, general relativity, applied to a localized gravitationally bound binary system immersed in an expanding Friedmann-Lemaître-Robertson-Walker, does not predict the existence of such a putative radial acceleration at Newtonian level. Instead, it was recently shown in literature that an acceleration of order H and directed along the velocity v of the test particle occurs at post-Newtonian level. We worked out its orbital effects finding well-behaved secular rates of change for both a and e proportional to the Schwarzschild radius r s of the primary. Their magnitude is quite small: the rate of change of a amounts to just 20 µm per century in our Solar System. Finally, we discussed certain basic criteria of viability that modified models of gravity should generally meet when their observable effects are calculated.