Well, still can't find myostatin half life in the blood, but follistatin (see "Release of Activin and Follistatin during Cardiovascular Procedures Is Largely due to Heparin Administration," Phillips et al, 2000) and myostatin propeptide both seem to be around 2 hours. Sooo... If myostatin is similar, that could be why you need so much compared to the ng/ml concentration in the blood: If half of it turns over every 2 hours, or in other words your entire supply of myostatin is replaced 6 times a day, for a two week dose of antibody you would need to bind 6 myostatin/day x 14 days = 84 times the amount in your blood at any one time. If you multiply the blood concentration (e.g., 20ng/ml) x 5000 ml blood in average adult male x 4 for an excess of antibody x 100 (let's just use 100 instead of 84, for round numbers), you get a two week dose of 40mg, which is starting to look very much like the lowest numbers that had an effect in the human study of post-menopausal women.
A lot of speculation in there. Mainly, I don't know for sure that myostatin's half life isn't much longer. And, I do know that the chimeric antibody has about a 7 day half life. What does that mean, practically? Not sure. Does it bind up some mystatin, keep it permanently and then it's out of commission, or can myostatin be degraded while bound to the antibody, allowing the antibody to be recyled, thus binding more antibody (a lot more) before it is used up? That's a pretty important question given that it changes the dose by a factor of 84 (under my assumed half life). I suspect that the antibody *is* reused or they never would have found a 7 day half life for it. So, this is all mainly to say that I can see how the 1+mg/kg doses *could* be justified, but I can also see ug doses working.
