Link to full article:
As I discussed in great detail previously, there is no actual physiological requirement for dietary carbohydrate. Most tissues can use fatty acids, the few that utilize glucose exclusively just reuse the same amounts over and over, and the brain switches to using ketones when glucose isn't available with the body making what little is required from other sources. From the standpoint of survival, the minimum amount of carbohydrates that are required in a diet is zero grams.
Of course, when carbohydrates are restricted completely, the body has to find something to make glucose out of. That something is lactate and pyruvate (produced from glucose metabolism), glycerol (from fat metabolism) and amino acids. It's the amino acid use that can be problematic since they have to come from somewhere. Under conditions of total starvation, that somewhere is generally muscle tissue; the body will readily break down protein to scavenge the amino acids it needs to produce glucose. In doing so, the muscle released alanine and glutamine (produced in the muscle from the breakdown of leucine and the branch chained amino acids, so you know) which can be converted to glucose in the liver.
Protein losses during total starvation are extremely high to start, gradually decreasing as the brain switches over to using ketones for fuel. Even so, in complete starvation there is always some loss of body protein. Over long periods of time, this goes from harmful (because function is compromised from muscle loss) to downright fatal.
From a body recomposition point of view, it should be obvious that losing muscle protein this way is bad. Researchers found years ago that providing adequate dietary protein helped to decrease if not outright eliminate the utilization of body protein for gluconeogenesis (a big word meaning the production of new glucose). Diets providing nothing but small amounts of protein (to the tune of 1.5 g/kg lean body mass or so) helped to almost eliminate the nitrogen losses inherent to starvation.