Here is a link to download the full article:
http://jcem.endojournals.org/cgi/co...pe2=tf_ipsecsha

Strange results. I could have accepted no correlation between insulin levels and obesity, but an inverse correlation? Wow.