How to handle missing values for categorical principal component analysis

I want to carry out a pca analysis on a set of postive and negative variables where each variable is assigned a binary value of 0 or 1. However, i am faced with the issue of missing values to an extent that i wouldn't consider negligible. I would appreciate meaningful contributions on how to address this issue