Based on your own citation, this is untrue. The citation you have linked is a meta analysis of 31 trials, many of which specify no medication use at recruitment

Based on your own citation, this is untrue. The citation you have linked is a meta analysis of 31 trials, many of which specify no medication use at recruitment. You are however correct about population being important (which is a big reason meta analyses can be very useful).