I think it was probably just a matter of using some variable that was easy to control and test and get clear experimental results on. Given the presenter's whole philosophy on the matter, I think he'd agree that there's much more in play than just glucose response.