From a guest blogger:
Consider the following thought-experiment.
A particularly sadistic person subjects another human being to horrible 
torture so severe that the victim’s sufferings are certain to result eventually in death. The victim then begs their tormentor to diminish his or her suffering by hastening the inevitable end. If euthanasia is to be considered morally acceptable then under such circumstances it would surely be all right for the sadist to terminate the victim’s life. We could certainly declare the sadist guilty of an act of horrible torture, but the guilt of murder as well would be problematic: the quality of life of the victim had been so diminished that the value of that life might either be ignored or plausibly be said to have also been reduced to a point that it was valueless.
The experiment suggests that, if euthanasia be morally acceptable, the value of life, or the attention we pay to it, is on some kind of sliding scale. If the victim did not ask that their life be ended, we might then consider that life to be worth more than if he or she did make the request. Likewise, if the request were made by one in a state of health wherein recovery seemed possible, we might rate such a life at a higher value than that of one in a terminal condition.
This state of affairs runs the risk of becoming even more complex. If the sadist had not tortured the victim but, rather, had simply responded to a voluntary request from the other, we might attribute culpability for unjustly taking the life (albeit with consent). But we could, surely, only make such an assertion by assuming the life being lived to be of a sufficiently high value for its deliberate termination to be a bad thing. What then are we to think of murderers? Is it worse to kill an able-bodied person than a disabled one, or a happy person than someone who is depressed?
Such speculation might appear ridiculous and, at present, it would certainly be so. We generally hold that human life has a constant value regardless of that life’s circumstances. Once, however, we allow euthanasia to be morally acceptable we immediately make an opening for the notion that the value of life, or the attention we pay to its value, might change or be changed. Whether this happens by way of disease or accident or another human being is rather beside the point. In fact, in accepting euthanasia as moral, we should necessarily do away with the belief that life has an unchanging value common to all. In the event of someone becoming distressed we should be led to conclude that the life in question was, in some way, worth less than that of another person not in that condition.