Sphericity and robust ANOVA

Hi all,

Can anyone tell me if performing a robust factorial ANOVA with trimmed means is sufficient to account to a lack of sphericity in my data, or should I be using the Greenhouse-Geisser correction?
Just to be clear, my intention is to use a robust test anyway because of other idiosyncrasies in the data - I'm just not sure if I need to do more to account for the lack of sphericity.

Thanks in advance.