An RSS feed is no different than other web resorces: to browse it anonymously via Tor, connect to each site individually, one new circuit per site.
As it stands today, RSS feed aggregators just aggregate several feeds from a user-predefined list, retrieving it in the background and presenting it in an UI.
Taking e.g. Tiny Tiny RSS, an OSS self-hosted PHP webapp, each feed in the list is checked in the background every n minutes for updates. From a network perspective, a single IP generates a recognisable pattern of traffic, either the native IP of the server, a proxy or an Tor exit node.
One may attain some level of obfuscation by having a very large set of feeds, so that one's fedlist does not have an unique, recognizable pattern. For example, by loading up as many as opml files one finds on the net, or by having a large multiuser instance of tt-rss.
But for effectiveness, it depends on the risk model one has.