Perhaps I'm misunderstanding you, but this is not really correct. Most vulnerabilities are not regressions.
Attention is what makes security issues discoverable, and popularity is what makes exploits valuable serious enough to warrant attention. The more popular software is, the more attention it gets from the security community (both black and white hat). The more popular software is, the higher the impact of an exploit is. The more popular the software is, the more significant the response is.
That doesn't mean older software is secure, or that it can't be exploited. It just means nobody is really looking at it. Fairly often, security alerts come up for software that doesn't list older releases because they didn't bother to check their EoL releases not because they're unaffected.
Take the Print Spooler vulnerability on Windows, or the ShellShock exploit in bash, or the Apache Log4j 2.x vulnerability. These vulnerabilities are all so old that they essentially work on any version of the affected software, including those that are more than a decade old for which no fix was even planned. Like the ShellShock bug in bash was found to go back to bash 1.03 released in 1989.
As long as you have an earlier version of software that was later found to be vulnerable, you should assume that it is unless you've explicitly investigated the vulnerability and found otherwise.