I highlighted the word in pdf using the code in the answer to the following question: [Highlight words in a pdf using itextsharp, not displaying highlighted word in browser]2
Due to my <50 rep this is posted as an answer but really it is a comment.
Thanks mkl for pointing me in the right direction however the problem remains for me sometimes event when this is taken into account (am working in VB).
For some PDFs (from a lists of similar documents scanned) I too get the error and it turns out that the pFont.GetWidth call is returning ZERO for each character hence the result of GetStringWidth returns ZERO also.
I note that I can pass GetWidth the whole string or one character at a time and the results are the same - when it works it works great when not I get zero.
I downloaded the latest ITEXT (5-5-10-0) and the problem remains - since this problem occurs rarely I have temporarily resorted to checking the output of getstringwidth and if it is to become zero then I estimate a response based on an average with a cross check against it being too high versus the pSingleSpaceWidth value.
I know this is quite poor but my requirements are modest, it happens rarely and I haven't the time or requisite knowledge (of Itext/PDF) to investigate how to obtain an equivalent GetWidth result accurately.
I do hope this info helps someone and possibly if someone does have the required knowledge to improve/fix this idea I am all ears.