A five sigma signal means the chance of the result being a statistical fluke is less than one in three million. Since July, the Higgs-like signals seen by the CMS group have fallen from around 2.8 to 2.3 sigma. Those seen by the Atlas group have dropped from around 2.8 to less than two.
We're still waiting for the combined data from CMS and ATLAS. I'm guessing that they don't want to fall victim to background issues that could be the major culprit of the "bump" seen in the CDF data that no one else is seeing.