Yes, correlation does not imply causation, but it's a darned good place to start.

Assuming all classes were equal and balanced in PvP, we can assume that all classes would have the same representation. Thus, combining the two tourneys (NA + EU), we would see 6 of each class in the top ten; if all classes were equally good and played equally well, we would expect a near-random distribution of classes in the top ten (or any grouping.) Here's the actual breakdown:

Death Knights: 8 + 5 = 13
Druids: 0 + 0 = 0
Hunters: 3 + 7 = 10
Mages: 1 + 4 = 5
Paladins: 8 + 6 = 14
Priests: 0 + 3 = 3
Rogue: 0 + 3 = 3
Shaman: 2 + 1 = 3
Warlock: 8 + 1 = 9
Warrior: 0 + 0 = 0

The only class that comes close to average representation is Mage. All other classes are way above or way below the expected zero. Granted this isn't an ideal sample size, but it shows DKs, hunters, pallies, and locks are preferred far more than would be expected if all classes were equal. Does this imply they're OP? Not sure. Probably more to the point, it shows that top-end arena players prefer those classes and thus those classes are better-suited to PvP than the other classes. If pressed, I might say that DKs and pallies are OP with druids and warriors UP.

More data needed. This was just a one-minute, back of the napkin-type examination.

Reply