The advantage of CTOD is that it provides a K1 value at slow loading rates. A K1 value can be used for determining allowable flaw sizes where as charpy energy can do so only poorly. If one also looks for a minimal amount of CTOD stretch zone formation, say 0.15mm, then it is likley the piece has good toughness against both slow and rapid fracture.
The charpy test is a better (as well as faster and less expensive) test when the upper shelf material toughness is less than 100J. (Some might put this in the 100J to 200J range) Below 100J, the charpy test is a high strain test and indicates toughness pretty well. Particularly as one should also have pre-bend charpy data to compare to.
When pre-bend toughness is over 100J, charpy energy may be miss leading if too much of the hammer energy is being consumed by crack initation. In that case, the test is not measuring toughness against high strain rate and may result in a miss-leading high toughness value. Even so, a high charpy energy value is still pretty good and acceptable for many applications. Charpy is still the most commonly used standard.
In my opinion, CTOD should be reserved for critical applications where technology is being stretched in one way or another. Standard applications that have been done many times before should stick with charpy.