Do you have other datasets that have the same problem?
Yes, indeed, last autumn, I was animating a Therion course for the French federation, and I had to use 5.2.2, because the course examples had the same problem. And Xavier also reported the same problem, some time earlier in 2009, on the Therion list.
I have progressively stripped your sample file down to 4 simple sequential legs, removed the layouts from the export extended elevation to xvi, made various other simplifications to try to make it similar to mine, and still I get the xvi 450000 issue with ONLY your sample dataset and not with any of my datasets. There are more changes I can make, but it seems the blind trial and error approach might take a while... Maybe an author might have some insight as to what the cause might be? Bruce