That was a very informative reply Blacksheep 
Thanks
Another point which occurred to me is how were the withdrawal times actually calculated? Did the research company use in vitro tissue culture or was the work done on live and dead whole animals? Were sheep or cattle used to establish withdrawal times? What was the size of the sample? Were measurements taken every day, or at larger intervals. Was the work done on just one or two breeds of sheep? I could imagine that metabolism would be quicker in a fast growing commercial lamb than in say a slow finishing Primitive.
My pharmaceutical expertise is much more at the other end of the process
Pharnorth knows this far better and has added to the discussion already. But potentially any and all of what you say above. As an illustration though, metabolism isn't one fixed thing where all processes run at the same rate and are all faster/slower - it's massively complex and hard to predict in some cases. A slow finishing animal may well have a faster metabolism too - burning more energy and thus gaining less weight from the food consumed. But that doesn't necessarily mean it breaks down drugs faster.
We humans have a range of generic variations in things like liver enzyme activity - a good example is in relation to codeine. Some humans lack the enzyme necessary to break it down and since it is only really when our bodies break it down to morphine that it provides pain relief they find the drug completely ineffective. Most of us have the enzyme and it works but again a small proportion are what are termed ultra-fast metabolisers and they turn the codeine into morphine so quickly that the levels generated spike high and fast, leading to side effects and sometimes toxicity. I have no doubt that livestock will have these sorts of variations too, the effects of which could have been magnified in the purified breeds we have developed, or within some flocks - and there'll be no real way of knowing.
Is the manufacturer hedging their bets and going for the worst case scenario, or averaging out all the results? How was the statistical analysis of the data carried out?
Hard to know without looking at the data in its entirety (which probably isn't available to us without detailed discussions with the manufacturer). However simply on the grounds of what is economically feasible I can't see that they will have performed formal testing in representative samples of every sheep breed, for example. So yes there will certainly be averaging and the addition of some degree of safety margin to allow for the variation described above.
I think what trish.farm is looking for is to be able to work out what would be the risk for her eating the meat and what might the consequences be if she goes ahead.
Yep, and we do estimations like this in humans regularly, for example trying to estimate what foetal exposure there may have been to a drug the mother used before she knew she was pregnant. But it is really difficult to do and impossible to be completely accurate about, even when the specialist knowledge is brought to bear.
