For a multidisciplinary journal or conference, I will sometimes deliberately assign a paper one reviewer who knows significantly less about the subject. The intention is to have a slightly more detached perspective who can say whether this paper is interesting and intelligible to anybody outside of its narrow sub-sub-sub-field. It's also rare to review a paper that you are perfectly knowledgeable about, since science has so many different aspects.
What you should do when you do not perfectly understand a paper:
- Be extremely clear on which parts you are confident that you understand and which you don't.
- Do not assume the authors are wrong if you don't understand. It may be one of the gaps in your knowledge.
- Do not assume the authors are right if you don't understand. They may be blowing smoke in your eyes.
- Explain what, if anything, you found of value in the paper despite your lack of knowledge.
If your review, in combination with the others, is not sufficient, it is the responsibility of the editor or chair to obtain another. It is not your job to determine how the reviewers are distributed.
That said, if you are completely and totally lost, contact the editor / chair who assigned the paper to you and check with them. Depending on their intent, they may take you off the paper, or they may tell you that this is exactly what they want you to write down.
If you need to do this do it soon---it's terrible form to screw up somebody's reviewing schedule and either create a last-minute crisis or an extra delay for the authors.