The very first example in Cox and Hinkley of sufficiency concerns the Poisson distribution on p.19, which concludes that the sum, or equally the mean, is a sufficient statistic for the Poisson. The thread cited concludes that the median cannot be a sufficient statistic, by showing that such an assumption leads to a contradiction. I am reluctant to reply myself for what (paraphrasing Littlewood in another context) are delicate but sufficient grounds (pun intended).