A point I made in yesterday's post is worth expanding on - my scepticism about the advent of the Magic Pizza Pill. It's not just hard to develop an obesity drug. It's hard to fight almost any of the biological pathways, even when they're doing harm. Evolution has whittled some of them down to lean, mean, biochemical machines, and the ones that haven't had that treatment are often shaggy, baroque palaces of backup redundant redundancy.

Obesity (and feeding behavior in general) is a good example of the latter. Every year or two, another feeding/satiety signaling pathway is elucidated, and everyone gets excited that this might be the key. The peptide ghrelin is the current example. I hope it works, but the previous star players (leptin, neuropeptide Y, galanin, and so on and so on) have all fallen on their faces. It's often for the same reason: too many backup systems, too strong a thumb on the homeostatic scale. The body really resists delicate tinkering with something as important as eating.