Exercise has been advocated as a means of furthering the combustion of the carbohydrates in the system. This factor Allen likewise utilizes in his plan of treatment. He rather thinks that brisk exercises are more efficacious than the milder forms. Fatigue, however, should always be avoided.

With regard to the details of treatment, Allen gives the following rules:

If the patient is moderately emaciated, with a negative carbohydrate balance and acidosis, he is put to bed and receives no food whatsoever. If coma seems imminent, the usual emergency treatment, with purging, stimulants, alkalies, and large amounts of water should, of course, be carried out. In addition to fasting, alcohol is important in the treatment at this stage. From 50 to 250 cc. of whiskey or brandy may be given in each twenty-four hours, in small doses, - from 10 to 20 cc, every one to three hours during the twenty-four. As soon as the glycosuria stops and the acidosis diminishes, which even in severe cases may be within 48 to 96 hours, the amount of alcohol and alkali may be reduced. Fasting and moderate dosage of alcohol are continued for from twenty-four to forty-eight hours longer, however, depending on the patient's strength. The alkali is then stopped, and feeding with starch is commenced, in order to clear up the last traces of ketonuria.

Green vegetables are useful because their carbohydrate and food value is so low that they can be given in considerable bulk, and this bulk is agreeable to the patient for relieving his feeling of emptiness. Neither fat nor protein is added.

For the first day, the carbohydrate content is 10 to 40 grams, divided for four equal feedings, during the day. If glycosuria remains absent, the ration for the next day is doubled. There still being no sugar in the urine, the ration on the following day may be increased to 100 grams of carbohydrate. Should glycosuria reappear, another fast day is interposed, - from 50 to 200 cc. of whiskey being given. Even severe cases of ketonuria may by this method completely clear up.

The carbohydrate of the diet is seldom reduced below 50 grams, and is preferably kept higher. If carbohydrate must be kept low, the total diet is kept low. The diet is so chosen that glycosuria, not ketonuria, is the signal of overstrain. Fasting-alcohol days are given not merely whenever this signal appears, but also at close enough intervals to prevent it from appearing, - even every two or three days, - if necessary.

No matter how low the assimilation power is, no attempt should be made to feed in excess of it. With the improvement in the condition of the patient, the carbohydrate is increased. Increase in weight, however, should not be attempted at this time. The metabolism is kept at the lowest safe level, until the patient is taking from 100 to 150 grams of carbohydrate (green vegetables) daily, with fast days interposed often enough to prevent any trace of glycosuria from appearing. The protein is cautiously added, always being kept low. In favorable cases the weight and well-being may finally improve under gradual additions of fat.

In milder cases, the treatment may be correspondingly milder. Primary loss of weight is intentional. When there is extreme cachexia and emaciation, the difficulty is greatest. The alkali treatment is not employed unless for a brief period at the outset, while severe acidosis is being combatted. Even very severe cases are amenable to this mode of treatment. The best therapy lies, however, in the application of this principle of treatment at the earliest possible stage of diabetes.

Severe acidosis can be considerably reduced by prolonged fasting. Freedom from glycosuria seems attainable in all cases of uncomplicated human diabetes before there is danger of death from starvation. No contra-indication has been met, unless it be the appearance of nausea, vomiting, and prostration while fasting. Just as fasting is continued, not for any limited number of days but as long as necessary for sugar-freedom; - so also the diet is governed not by any theoretical standard of protein or calories, but by the amount of each food that can be given in each case while keeping the urine clear. Under this program, even weak and emaciated patients have been subjected to under-nutrition, in both protein and calories for weeks or months continuously, with ultimate benefit. Any trace of sugar is the signal for a fast day, with subsequent modifications of diet; and routine fast days are often used as frequently as once a week, even in absence of glycosuria.

Two principles are important in the management of severe cases: To keep the patient permanently below weight, and to restrict the quantity of fat in the diet. The reduction in weight is beneficial to the diabetic patient, serving to spare the weakened function and increase tolerance. Sometimes a slight reduction of weight suffices even for a severe case. Again, a well-nourished patient, easily kept free from glycosuria, had to be reduced by 20 kilograms merely because of a slight stubborn ketonuria and a persistently high blood sugar. Most patients are able to regain weight, but few severe ones are able to return fully to normal weight. Any increase that is possible without return of symptoms is permitted. Any gain that brings back glycosuria or ketonuria is checked. The overtaxing of the patient's metabolism by giving fat beyond the limit of tolerance is an additional explanation of the failure to keep certain patients free from glycosuria and ketonuria under former methods of treatment.1

1F. M. Allen: Prolonged Fasting in Diabetes." Am. Jour. Med. Sciences, 1915, p. 480. 9