Because the statistics aren't good enough. It's right on the borderline between "definitely a real effect" and "probably not a real effect". Effects reported at this level of confidence have turned out to be non-existent before. So people are being cautious: it's significant enough to be worth considering explanations for it, but not significant enough to be sure that it's not just noise in the signal.
(I know an astrophysicist who cynically remarked that it's the sort of result you include in your paper because it's flashy and suggests new physics, as opposed to being something you're sure is there.)
Anyway, WMAP continues to collect statistics, and in a few years there will be Planck, so eventually we will know whether it's real or not.