On the gravitational analogue of Larmor's theorem