If a surviving liver be perfused with blood containing β.oxybutyric acid, the latter is in part converted into aceto. acetic acid.1 Minced liver or even the aqueous extract of liver tissue will effect the same reaction.2

Fischler and Kossow3 report that the formation of aceton bodies in a phlorhizinized dog is decreased in the presence of an Eck fistula, whereas if a "reversed" Eck fistula be created by diverting the blood from the vena cava into the portal vein, the excretion of aceton bodies is increased fivefold. This points to the liver as the main source of the aceton bodies, if one may accept conclusions drawn from experimental conditions so profoundly abnormal.

The quantity of the aceton bodies in the blood is given by Marriott4 as follows:

1 Embden and Engel: "Hofmeister's Beitrage," 1908, xi, 323.

2 Wakeman and Dakin: "Journal of Biological Chemistry," 1909, vi, 373.

3 Fischler and Kossow: "Deutsches Archiv fur klin. Med.," 1913, cxi, 479.

4 Marriott: "Journal of Biological Chemistry," 1913.14, xvi, 293.

The increase in the aceton bodies in the blood is greatest in diabetes mellitus in man, is not so marked in phlorhizin glycosuria in dogs, and is least of all present in depancreatized dogs. Sassa1 states that the organs of diabetic men dying in coma may contain eight times the normal quantity of β-oxy-butyric acid, the liver showing relatively the greatest storage of the substance. In one instance (Case II) 130 milligrams of β-oxybutyric acid were found in 100 grams of the body tissue of a man weighing 70 kilograms, and the author computes the presence of 85 grams of the substance within the body. Marriott's2 highest figures for 100 c.c. of diabetic blood in man are 28 milligrams of aceto-acetic acid and 45 milligrams of β-oxybutyric acid.

The demonstration by Ringer3 that propionic acid was completely converted into glucose and that higher fatty acids with uneven numbers of carbon atoms yielded glucose in so far as they might form propionic acid by β-oxidation, presents the theoretic possibility of giving to diabetics fats containing these fatty acids, which would yield innocuous glucose instead of acid bodies as the end-products of oxidation. Practical difficulties in the preparation of such fats have alone prevented Ringer from testing the efficiency of their administration to diabetic subjects.

The result of the formation of acid bodies in the organism leads to a condition of acidosis, the alkali reserves being called upon. Not only does ammonia increase in the urine, but there may be a marked fall in the carbon dioxid content of the blood due to a diminution in the quantity of bicarbonate of soda. Magnus-Levy4 reports an extreme case in which 100 c.c. of the blood of a diabetic just before death in coma contained only 3.3 c.c. of carbon dioxid instead of 40 c.c. normally present.

1 Sassa: "Biochemische Zeitschrift," 1913-14, lix, 362.

2 Marriott: "Journal of Biological Chemistry," 1914, xviii, 507.

3 Ringer: Ibid., 1912, xii, 511.

4 Magnus-Levy: "Archiv fur experimentelle Path, und Pharm.," 1901, xlv, 389.

The reduction in the carbon dioxid combining power of the blood and the consequent lowering of the carbon dioxid tension in the alveoli do not appear in the earlier days of acidosis, provided the acids formed be neutralized with ammonia.1 The withdrawal of alkali occurs later. Rona and Wilenko2 find that, despite the acidosis, the hydrogen ion concentration of the blood may remain normal on account of the compensation brought about through the removal of carbon dioxid by the lungs and of acids through the urine. Notwithstanding this control over the blood, the authors believe it possible that there may be a local increase of the hydrogen ion concentration in certain cells and tissues.

Concrete cases of blood analyses are offered by Poulton3 (see p. 221), who reports concerning the blood of 7 diabetic patients. The first 6 possessed a normal blood reaction. One of them (E. M. S.) following the first examination fell into deep coma and twenty-two hours later showed an abnormally high hydrogen ion concentration. E. H., whose blood reaction was similar, was also in deep coma. The first two patients gave no indication of coma, but all the others were drowsy. B. died in coma eighteen hours after the examination of his blood, which had been normal in reaction.

The figures are in part as follows:

1 Munzer: "Zeitschrift fur exp. Path, und Therapie," 1914, xvi, 281. 2 Rona and Wilenko: "Biochemische Zeitschrift," 1913-14, lix, 173. 3 Poulton: "Journal of Physiology," 191s, 1, p. 1.

On the basis of work on a diabetic and comatose boy weighing 32 kg., Magnus-Levy1 makes the following computation of metabolism. He purposely assumes a high requirement of energy for a lad of this size, or 50 to 55 calories per kilogram, which calls for a total of 1600 to 1700 calories. The boy burned 90 grams of protein and perhaps 200 grams of fat:

Here we perceive an extreme case of diabetic metabolism in which half the energy contained in protein is excreted in urinary sugar and 20 per cent, of that contained in fat is eliminated in the unburned β-oxybutyric acid.

This, then, is the worst picture of the perverted metabolism in diabetes. Sugar cannot burn, fat burns only as far as /3-oxybutyric acid, and as for protein, a part of its amino-acids are converted into sugar and another part into β-oxybutyric acid, neither of which can be burned.

It is notable that the phlorhizinized cancer patient of Stanley Benedict (see p. 455) who had a D : N ratio of 3.66 excreted 37 grams of /3-oxybutyric acid and 4 grams of ammonia daily, which shows that the acidosis of diabetes is coincident with a lack of sugar oxidation. In the diabetic C. K. (see p. 478) a fall in the β-oxybutyric acid excretion preceded the break in the D : N ratio (consult p. 271).