I found this paper, where E85 was run in a Toyota Prius hybrid. This is interesting:
While using E-85, the check engine light came on. Upon inspection this
produced a code indicating a lean condition was detected. During this testing, it
was noted that the check engine light for the lean code would come on between
140-160 miles after clearing the code. The oxygen sensor output and the shortterm
fuel trim all indicated that the Air/Fuel ratio is stoichiometric, and that the
fuel control module was maintaining the correct amount of fuel required. What is
setting off the check engine light is that the system monitors the change in the
base fuel map, which is indicated by the long-term fuel trim. If this change is
outside of the set boundaries for an extended period of time, then a code is set to
indicate that the fuel system is making greater adjustments to the fuel quantity to
maintain the correct air/fuel mixture than is necessary under normal
circumstances. The use of ethanol caused the fuel system to adapt the long-term
fuel trim by an increase of 32.81%. ! This supports the lower average fuel
economy of ethanol in the results. See figure 4. An added advantage of using E-
85 showed gains in power output, as shown on table 1. Although ethanol has a
lower energy density than gasoline, due to its lower air/fuel ratio, and effectively
cooling the intake charge, it allows more fuel to be introduced to the combustion
chamber thereby providing the power gains seen
So the check engine light does not necessarily mean that you are ACTUALLY running lean, just that the computer can't figure out where all that extra oxygen is coming from.
If one had a Scangauge II and one could ensure that a real lean condition is not occuring, one could run ethanol levels that cause the CEL to occur.
One really needs a Scangauge II...