I remember the malware campaigns the Kaspersky post refers to, long time ago... yes, back then there were big email campaigns where malware authors started to use password-protected archives and wrote the password in the email body. Kaspersky added code to extract the email body and try its words as the archive password. So... the malware authors started to send the passwords as images. Kaspersky added a simple OCR, trying to get the password even out of the image. Distorted images appeared after that I believe...
In between, Avast (and most likely not just Avast) detected those attachments without looking for the passwords, without trying to unpack them - because the files were quite specific and could be detected as such.
In very specific cases, Avast internally unpacks some password-protected archives, but no, it doesn't try to extract the passwords from emails. While it could be done, in my opinion it's too easy to bypass (obfuscating on HTML level or including images, making the user see something else than the extraction code does), plus for bigger archives it could cause unwanted performance issues. Asking the user for a password (if that's what Bitdefender is doing)... could be quite a hassle as well (windows asking for a password popping up during a scan - or even out of the blue because an application on background is currently downloading its update in form of a known password-protected archive).
Unless something special happens, I personally don't think it's worth the effort and associated troubles - considering the actual content of the archive has to be unpacked anyway before it could do any harm (and then the antivirus should detect it).