xpath check if document contains text

I want to get all XML nodes that contains specified text. For example, when searching for text "aa" in document:

<book>
 <content>
    aaaaaaa
    <markup/>
    bbbbbb
 </content>
 <author>
 aabbcc
 </author>
</book>

I should get nodes content and author.

I was trying to use XPath "//*[contains(text(), 'aa')]" but it fails when node has some subnodes that split text (like markup which splits content in above example).

Do you know how to solve that problem?