Seems that some chappies have never operated in a non-ATC (i.e. CTAF) non-radar environment.............
Neville someboby obviously has, and he is closest to the mark, however I would advise extreme caution in flying a DME arrival in reverse. Port Moresby in PNG had DME departures in days of old, and they were quite different to just being the reverse of the arrival steps. I vaguely recall something about different azimuth guidance tolerances when tracking away from a station, as opposed to tracking to a station.
Meanwhile back to the question.......
The first thing I would be looking at is the takeoff minimums on the applicabe chart. If it says night ops not permitted, then it is off to the hotel. If there is no hint of that, then the following would apply.......
First of all you need to determine whether the charts for the airport in question are designed to PANS OPS or TERPS criteria, as this will determine the radius of the circling area for your aircraft type. Then determine if there are any additional circling restrictions, such as a no circling area (this could potentially make departure impossible at night in some cases). Having determined this you can then depart, ensuring the aircraft is maintained within the circling area (and observing circling speed restrictions) until the MSA/LSALT is attained, and then proceed on course. You also need to consider your non-normal scenarios and ensure that at any point during your departure you can position yourself to either return visually or intercept the instrument approach, whilst satisfying obstacle clearance requirements.