The problem in your regexp is that the word boundary isn't properly detected with those chars (just like \w
and \W
are badly handled with regards to Unicode).
I'd suggest to start with
new RegExp('(^|[\\s\\.])ČčĆ抚Đđ', "i")
and to add to [\\s\\.]
the other chars you may be needing as word boundaries.
If you can't define the expected possible word boundaries, you'd better use a library to produce "Unicode compatible" regular expressions. Some are listed in this related question.