Allen knew that carbohydrates are the component of the diet that most affects blood glucose. Many practitioners of the time, and earlier, prescribed an “animal diet” or other low carbohydrate variation to clear glycosuria. Critics of the Allen-Joslin therapy claimed as much success, without near-total food privation, using diets low on carbohydrates but allowing more fat calories [5, 6]. Why then did Allen – and Joslin following him – virtually starve their patients?
Allen’s Early Work
Frederick Madison Allen (1879–1964) graduated from the University of California at Berkeley and completed his M.D. there in 1907. He spent the years 1909–1912 as a teaching fellow at Harvard Medical School, conducting experiments on cats and dogs, largely at his own expense. An austere man, he describes himself in an unpublished memoir (private collection of Alfred Henderson, Bethesda, MD, USA) as living like a hermit, continually working 7 days a week.
Allen produced diabetes in his animals by partial pancreatectomy, the degree of disease depending on how much pancreas was removed. According to his friend and biographer Alfred Henderson, “There resulted from this experimental period a mass of manuscript material, all written in longhand which appeared capable of considerable improvement. No publisher could be found to wade through his crude manuscript” [7]. With a subsidy from Allen’s father, Harvard University Press published the manuscript as Studies Concerning Glycosuria and Diabetes, which ran to 1,179 pages. In this book, Allen gives no prescription for calorie deprivation as a therapy for human diabetes, nor does he describe any felicitous effect of starvation on his animal subjects. To the contrary, Allen occasionally reaches conclusions that seem counter to his later position, for example, “Fat feeding is not to be feared in diabetics” [8].
Promoting Starvation
In 1913, Allen left Harvard and was appointed a nonresident assistant physician to work on diabetes in the newly established Hospital of the Rockefeller Institute in New York. In his memoir, written years later, Allen recalled:
I quickly followed up the first clue from my Harvard work, proving that diabetes in partially depancreatized dogs, which was too severe to be controlled on a customary protein-fat diet or on any diet while the animals were fat, could be controlled and kept controlled by starving them and then dieting so as to keep them thin... Within a few months I was able to ask for human patients.
His longest experiment to date was on a dog kept free from glycosuria for 6 months and still alive. Allen cautioned, “If an attempt is made to increase the weight of such an animal, glycosuria soon appears and must be checked by renewed fasting. Such dogs, though very thin, are vigorous and lively” [9].
Allen presented his results anecdotally, without quantitative data, not unusual in presentations of that era, but he does not even report the number of animals used. It would have been impossible to even intuitively assess the reliability of his results or the implied differences between experimental and control animals, or if indeed there were control animals. He regards dogs as adequate models for human diabetics, and he functionally equates the dog’s surgically reduced pancreas with human diabetes.
Allen adds that he has treated a “limited” number of patients by prolonged fasting and calorie restriction. “The results obtained indicate that the same method employed in rendering the diabetic dog free of glycosuria and prolonging its life is efficacious in eliminating glycosuria and acidosis in the human patient” [9].
By fall of 1914, Allen had treated 8 patients, and the hospital director thought his results sufficient to justify undertaking this work on a larger scale [10]. Allen envisioned diabetes occurring when a predisposed pancreas is pushed over the edge by excessive diet. The “overstrained” pancreas could no longer process glucose adequately, causing sugar to appear in the urine. Conversely, Allen thought an overstrained pancreas is rested by diminished diet, and that this respite may gradually strengthen its functioning. “The attempt to put on weight, according to the time-honored traditions of diabetic treatment, is one of the surest ways of bringing back all the symptoms and sending the patient downhill” [11].
A year later, addressing a medical group in May 1915, Allen is more emphatic. Reporting no new experiments but referring generally to surgically produced diabetes in animals, he noted that the measures ordinarily used in human diabetics, namely brief fasting and carbohydrate restriction, may be insufficient to keep dogs sugar free. In severe cases, “the initial fast must sometimes be measured in weeks rather than in days. The subsequent diet must be such as to keep the animal at a low level of weight... If glycosuria is prevented, the animals may remain lively and strong though thin... The treatment of (human) diabetes at the Rockefeller Hospital has been based upon these animal experiments” [12]. Allen had by this time treated 44 patients at Rockefeller, claiming to have cleared up glycosuria with an initial fast, sometimes lasting as long as 10 days.
It is difficult to tell the fate of Allen’s 44 patients from his closing statement in the May 1915 talk: “Among the patients treated thus far, during a variable number of months in the hospital and at home, spontaneous downward progress has not yet been observed.” My tabulation of these cases, which Allen later published [13], shows that of 41 patients whose condition was known at the end of 1917, 46% were dead or, in 1 case, “nearly dead.”
Allen was deeply committed to his new regimen, often interpreting comas or deaths to lapses by patients from their prescribed diets. He claimed repeatedly that his therapy was based on his animal experimentation, without publicly reporting these experiments in detail. I am also unaware of anyone replicating his results. An article in The New York Times of February 13, 1916, says Allen’s new therapy was based on animal experiments performed at the Rockefeller Institute by I. Kleiner and S. Meltzer, but this was not true [14].
Historian Michael Bliss characterizes Allen as a “stern, cold, tireless scientist, utterly convinced of the validity of his approach.” His determination to apply his methods ruthlessly led to a decision by the Rockefeller Institute to take away his control of the diabetes clinic [15]. In 1917, as the USA entered World War I, the head of the Institute, Simon Flexner, informed Allen that he would be a captain in charge of a diabetic service at Lakewood, New Jersey, effectively terminating his employment. Given his checkered career to this point and the dubious scientific basis for his treatment, one may wonder why his therapy would soon become so popular and Allen so prominent.
Joslin’s Enlistment
Elliott Proctor Joslin (1869–1962) was born in the small town of Oxford, Massachusetts, into a prosperous family of religious Congregationalists. He received his M.D. from Harvard Medical School. After further training, he opened a private practice in Boston in 1898, became the first doctor in the USA to specialize in diabetes, and was the founder of today’s Joslin Diabetes Center. His book, The Treatment of Diabetes Mellitus [16], went through many editions, indicative of Joslin’s prominence.