P7S Medical Review: Spring 1997, Vol.4, No.1
Diabetic Acidosis Revisited

Abbie Knowlton, M.D.

Columbia University College of Physicians and Surgeons

In 1933 the Journal of Clinical Investigation published the paper "On Diabetic Acidosis" written by three physicians in the Department of Medicine at CPMC--Dana Atchley, Robert Loeb and Dickinson Richards--with the laboratory assistance of Ethel Benedict and Mary Driscoll. This paper remains to this day, some 63 years later, the most comprehensive study of the physiological changes which take place in the development of this potentially life threatening illness. It is unique of its kind.

The study was carried out long before any laws existed mandating the obtaining of informed consent from patients undergoing experimental studies. The authors were ahead of their time in their sensitivity to their patients as the comment at the bottom of the initial page indicates. Both patients were fully informed as to any possible risks involved.

Dr. Dana Atchley

Dr. Robert Loeb

The data obtained involved meticulous and laborious work, done long before the existence of the autoanalyser, the chem 20. Each value was determined individually, many took days to complete and required weighing the sample on an analytical balance. Determination of potassium, for example, involved 14 separate steps, 15 if the sample required an initial ashing. The study entailed daily measurements of urine, aliquots of dietary intake with each batch of food which was freshly prepared every several days, and measurements of stool specimens pooled in 3-4 day collections. The analyses included sodium and magnesium, potassium, calcium, chloride, phosphorus, total nitrogen, inorganic sulfate, creatinine, glucose, ketones, organic acids, and fluid. Analyses of blood were done on five or six occasions and included, in addition to the above, total protein, oxygen, hematocrit, and pH.

Whether by virtue of the authors' medical prescience or by luck, the two patients selected for the study would by current criteria be labeled a type I diabetic (T.M.) and the other (W.O'C.) a type II. Both patients were young men, both taking up to 90 units of insulin daily and each patient remained on a measured diet throughout the period of study. After several days in which each was accustomed to the dietary regimen, each was observed for a control foreperiod (10 days for T.M., 16 days for W.O'C.), then in both patients insulin was discontinued, the regimen in all other respects remaining unchanged.

T.M., who would now be labeled a type I diabetic, complained within a day of thirst, lost an average of a kilo per day in weight, and spilled 100-150 grams of sugar per day in his urine. By the end of his 4th day he was obviously ill, with profound weakness, nausea and vomiting, and was in acidosis with a serum pH of 7.21 and bicarbonate of 14.6. During the insulin withdrawal period his daily excretion of "Na and Mg" and, most strikingly, that of K⁺ soared (see Graph I of the data prepared by Dr. Vincent Butler). With the reinstitution of fluid and insulin he was symptom free within 18 hours and back in balance within 10 days (see Tables 4 and 5).

W.O'C., who would now be labeled a type II diabetic, had a slightly longer control foreperiod due to an intercurrent URI. When, after 16 days his daily insulin was withdrawn, mild thirst on day 2 was his only complaint. Similarly to T.M. he had polyuria and spilled over 120 grams of glucose daily in his urine. In the first two days he had an impressive loss of "Na plus Mg" in his urine but only a modest loss of K⁺. By the end of 12 days he had lost 3.7 kg but had no fall in bicarbonate, no acidosis and spilled no significant amounts of ketone bodies in his urine (see Tables 8 and 9, Graph II).

The authors summarized their findings under the conclusions section at the end of the paper. It might be reasonable to add to these conclusions and emphasize the strikingly different response of the patient T.M., who rapidly developed acidosis following insulin withdrawal, with that of W.O'C. who did not. Both patients spilled equivalent amounts of sugar, both lost weight and water and "Na plus Mg" to a similar extent during the experimental period (the one in 4 days the other in 12) establishing that this was independent of acidosis and dependent on the glycosuria. However, T.M. lost overall more than twice the amount of K⁺ in 4 days than W.O'C. did in 12 days. This, I believe, emphasizes the severity of potassium depletion which marks acidosis and mandates its replenishment as well as sodium, fluids and insulin in the appropriate therapy of such a patient, assuming adequate renal function is established.