It's not their interpretation that surprised us, it's their results. They found that MORE insulin secreted after a meal or glucose intake meant LESS weight gain in this group of Pima Indians. But from what I could gather, high fasting insulin levels were positively associated with body fat ('Fasting plasma IRI levels were highly correlated to initial fat mass (r = 0.70; P < 0.001)', IRI = immunoreactive insulin). So maybe high insulin is bad only when it's high all the time.