In Theorem 7.6 of http://www.numdam.org/article/AST_2002__279__271_0.pdfIllusie - An overview of the work of K. Fujiwara, K. Kato, and C. Nakayama on logarithmic étale cohomology it is stated that if $X$ is a log regular fs log scheme and $U$ is the open subschema where the log structure is trivial, then the kummerKummer-étale site of $X$ is equivalent to the curve-tame site of $U$. The proof is not written there, and it cites a misteriousmysterious paper of Fujiwara and Kato that apparently was never published (or at least I have not been able to find it). This result seems to be well known and (as far as I understand) not trivial at all, but I cannot find a reference where this is proved.