It doesn't make sense to say there's an upper and lower bound on the "worst" case. If that were true, the upper bound would be the new worst case, and the lower bound would be some middle case. A best or worst case is inherently always going to be big-theta.