Sometimes characters with different meanings are used by visual writers because they feel they look better. For example, quote symbols can be just a straight line, or can curve toward the quoted text. In the latter case there is a different symbol for left and right quotes. I'm not sure if UNICODE designates a separate code for apostrophe vs single quote but if so, then there are three possibilities of what a sighted person might place in the apostrophe location. Further complicating this, is that the US English keyboard has only one key that is used for both apostrophe and quote and so often the word processor replaces the single quote with a more appropriate symbol. Microsoft Word calls this smart quotes and it can be configured in the auto correct area of settings.
Now, the libelous is translating back from UNICODE to Braille and must decide for every unicode character what to do with it.
What should you do?
1. Read by character and see how NVDA reads the individual character.
2. If NVDA says apostrophe then this is a bug in NVDA Braille translation tables that are provided by libelous.
Otherwise it might just be a rogue web designer.
If you consistently find that a particular non apostrophe character is used for the apostrophe, it might still be possible to get the Braille translator to re interpret it mid-word as an apostrophe.