I was just reading on cnn that they have found a lower incidence of diabetes among coffee drinkers. I find this interesting because the common consensus seems to be that caffeine lowers insulin sensitivity. Substances which have been shown to increase glucose sensitivity (such as quinindines) are probobly not in coffee in large enough quantities to be significant. I'd be very interested on Bradley's and others opinions.
My own personal theory would be that caffeine has shown to decrease appetite and then coffee drinkers would eat less and be less fat, which increases insulin sensitivity.