In this article, we point out the flaws in Theorem 1 of the article “Badr, Elsayed, et al. ”An integer linear programming model for solving radio mean labeling problem.“ IEEE Access 8(2020) : 162343-162349” and show that the radio mean number of cycles Cn determined in Theorem 1 is incorrect when n ≥ 16 by defining radio mean labeling with lower span. We thus derive an improved upper bound on the radio mean number of Cn, n ≥ 16.