The general issue is that text objects may use embedded fonts with specific glyphs assigned to specific letters. I.e. if you have a text object with some text like "abcdef" then the embedded font may contain glyphs for these ("abcdef" letters) only but not for other letters. So if you replace "abcdef" with "xyz" then the PDF will not display these "xyz" as no glyphs are available for these letters to be displayed.
So I would consider the following workflow:
- Iterate through all the text objects;
- Add new text objects created from scratch on top of PDF file and set the same properties (font, position, etc) but with a different text; This step could require you to have the same fonts installed on your as were used in the original PDF but you may check for installed fonts and use another font for a new text object. This way iTextSharp or another PDF tool will embed a new font object for a new text object.
- Remove original text object once you have created a duplicated text object;
- Process every text object with the workflow described above;
- Save the modified PDF document into a new file.
