Sure, if you can afford longer shutter it's much better for noise than pushing ISO. I wonder if this remains the same for very long exposure times (minutes) where the heating of the sensor makes termal noise dominant, but I think even in that case longer exposure time is better than internal amplification (ISO).
Yesterday I compared a shot at ISO100 and T=30s with ISO800 and T=4min (240s), and the first one was better in noise:

BTW, I switched the in-camera noise reduction (dark frame) OFF and A LOT of hot pixels like that one in the picture near the lamp appeared. Those hot pixels don't appear at any exposure time if the noise reduction is ON, and I thought noise reduction only acts for longer than 25 or 30s exposure times.
Does this mean these pixels need such long exposure times to reveal themselves? or they could be mapped from the camera manufacturing and are only mapped when noise reduction is activated (even if dark is not substracted for short exposure times)? There are a lot of them, about 15 or 20 in the whole image area.