Since Sakel's introduction of hypoglycaemia into the therapy of psychoses, discrepancies between the level of the patient's blood sugar and the severity of disturbance of consciousness have frequently been noted. The blood sugar decreases in a relatively steep curve within the first and second hour after the injection of insulin; but only after it has remained on a low level—differing from case to case—for a considerable time does the patient go into coma.

It seems useful to define the clinical signs of coma if it is intended to correlate it with laboratory findings. It is a state of unconsciousness in which the patient cannot be roused by sensory stimuli; in coma such stimuli do not provoke local and purposeful, but global movements only, if any. Once in coma the patient remains so, independently of the blood-sugar level, which can rise and fall within a certain range without visible change in the clinical picture; it may even rise to a level at which the patient was awake or could be easily wakened before he went into coma. Subcutaneous or intravenous injections of adrenalin in coma produce a rise in blood sugar without clinical effect. On the other hand, the flooding of the blood stream with sugar by nasal tube feed of a concentrated sugar solution or by intravenous injection of 33 per cent. glucose solution terminates coma and restores consciousness within a few minutes. The latter's action is so immediate that it is considered a practical rule not to stop injecting glucose until the patient is really awake.

From these observations coma appears to be a kind of trigger mechanism comparable to instinctive reactions, such as thirst, hunger or sexual desire. Once established it can only be reversed by specific means.