According to S6353 \d is equivalent to [0-9]. This is (turned out, I did not know) indeed the case for JavaScirpt/ECMAScript, but for most regex engines it is not.
The latter only matches the 10 digits defined in ASCII, where the first matches all characters defined as number in Unicode. I think be stating explicitly that this is only the case due to a limitation of the ECMASCript regex engine, it will help people to understand the suggestion, whithout (in later cases) applying it blindly in other languages where it turn out to make a difference.