Why Were “Starvation Diets” Promoted for Diabetes in the Pre-Insulin Period?

In this 2011 paper, Allan Mazur reviews Frederick Allen’s early support of “starvation diets” to treat diabetes. Mazur concludes Allen’s extreme dietary interventions may have caused as much harm as good; low-carbohydrate diets, as used by other researchers during the same period, may have been just as effective with fewer risks.

Prior to the discovery of insulin in 1922, there was no effective treatment for diabetes. The modern distinction between Type 1 and Type 2 diabetes had not been established, but it was known that diabetics who presented in childhood (those we now know likely had Type 1 diabetes) had a particularly poor prognosis, with few surviving more than a few months after diagnosis (2). Allen, and later Elliott Joslin — both prominent diabetes specialists — converged on the use of extremely low-calorie diets to treat diabetes.

Allen graduated from the University of California, Berkeley, in 1907. For the next six years, he served in minor academic and teaching roles at Johns Hopkins University and Harvard Medical School. During this time, seeking a research project of interest, he self-funded animal experiments. His primary focus was diabetes.

Allen surgically induced diabetes in animal models through a complete or partial pancreatectomy (i.e., removal of the pancreas). Partial pancreatectomy would create a pattern of symptoms similar to those seen in human diabetics, with glycosuria (glucose in the urine) the primary diagnostic symptom. By feeding animals a variety of diets, he found the most severe cases of diabetes could only be resolved through a highly restrictive diet involving severe caloric restriction most days and complete fasting multiple days per week; if caloric intake increased, glycosuria would return. This early animal research was published in the 1,179-page manuscript Studies Concerning Glycosuria and Diabetes, which he published with his father’s financial support.

By 1913, now working at Rockefeller Institute, Allen began to test his highly restrictive diets in humans. By 1915, he had treated 44 diabetic patients, each with an initial fast of as long as 10 days (until sugar was no longer present in urine) followed by severe calorie restriction to prevent any return of glycosuria.

Patient compliance was a constant challenge. As an example, Allen describes one blind 12-year-old diabetic boy who ate birdseed, toothpaste, and meals hidden in his room by his mother. The patient died four months after treatment began.

Allen and his colleagues provided records of these early patients. He did not compare patients to a control group or use statistics to assess the reliability, consistency, or magnitude of his findings; instead, in line with the standards of the time, he inspected individual cases and sought to understand why the treatment had or had not succeeded for a given patient. Of 41 patients treated between 1914 and 1917 for whom data was gathered, 46% were deceased by the end of the observation period. Among those under 20 at age of admission, however, 78% died over the same period.

Allen remained an adamant supporter of the use of fasting in diabetic patients, arguing that any intervention leading to weight regain will also lead to restoration of glycosuria and patient deterioration. As World War I approached, however, he was gradually removed from his clinical responsibilities, in part due to the extreme demands he placed both on patients and his fellow clinical staff.

Joslin, who had been an assistant professor at Harvard Medical School while Allen was a teaching fellow, opened a private practice dedicated to the treatment of diabetes in Boston in 1898, the first of its kind in the United States. He quickly rose to national recognition, with his textbook The Treatment of Diabetes Mellitus published in multiple editions. Allen and Joslin had, at minimum, discussed diabetic patient care while both were at Harvard; they also were members of the same undergraduate fraternity, and Allen’s work had been supported by a bequest from Joslin’s aunt. By 1915, Joslin wrote of Allen’s work in glowing terms, arguing his starvation diets provided the first tool to effectively treat the disease.

As head of the largest diabetes clinic in the country, Joslin had 62 patients under age 15 die in his care between 1898 and 1916, all to coma. Joslin adopted Allen’s treatment in 1916, successively reducing the fat, then protein, then carbohydrate content of patients’ diets until glycosuria resolved. Joslin treated 48 cases without a diabetes-related death in the first year, noting fasting increased expected survival of these patients from months to years. During the same period, three patients starved to death; Joslin blamed himself for these deaths, viewing them as the result of clinical mismanagement.

Joslin continued to promote Allen’s work during the following years, and by the time World War I was over, Allen was similarly well-known within the diabetes community. In 1920, he converted an abandoned estate in Morristown, New Jersey, into the Physiatric Institute, which in addition to treating patients served as a research center and hosted seminars evangelizing the benefits of severe dietary restriction in diabetic patients.

Around this time, Allen treated his most famous patient, Elizabeth Hughes, daughter of prominent politician Charles Evan Hughes. Diagnosed as diabetic in 1919 at age 11, Allen placed Hughes on a 500-calorie-per-day diet. This diet allowed her to survive, though her condition rapidly deteriorated during the winter of 1921/1922. However, Hughes recovered and survived to age 73, not because of Allen’s diet but because she became one of the first patients given then-experimental insulin therapy. By 1922, insulin eliminated the need for Allen’s extreme diets, and he would later contribute to research guiding the safe and effective use of exogenous insulin.

It is worth noting Allen and Joslin’s methods encountered criticism from other prominent experts. Two of their major critics were Louis “Harry” Newburgh and Phil Marsh, who treated diabetics at the University of Michigan using diets that were only moderately calorically restricted but severely restricted in carbohydrates. While Newburgh and Marsh also published only preliminary evidence, this evidence suggests their diets were similarly effective at rapidly resolving glycosuria in diabetics but with fewer side effects. Existing evidence on their low-carb intervention is consistent with modern research on the effectiveness of ketogenic diets to treat Type 1 and Type 2 diabetes.

Overall, the data we have suggests Allen and Joslin’s fasting diet was far from a cure for (Type 1) diabetes, though it certainly allowed some patients who would have otherwise died to survive until the discovery of insulin. Modern evidence has shown calorie restriction and carbohydrate restriction are powerful tools to mitigate the severity and symptoms of both Type 1 and Type 2 diabetes. In this sense, Allen, Joslin, Newburgh, and Marsh deserve credit as early proponents of non-pharmacological treatments for this chronic disease.

---

References

1. The Treatment of Diabetes Mellitus (E. Joslin)
2. The use of a high fat diet in the treatment of diabetes mellitus
3. Studies Concerning Glycosuria (F. Allen)
4. Studies Concerning Diabetes
5. Investigative and scientific phases of the diabetic question
6. Total dietary regulation in the treatment of diabetes
7. Prolonged fasting in diabetes (F. Allen)
8. Bittersweet: Diabetes, insulin and the transformation of illness
9. Present-day treatment and prognosis in diabetes (E. Joslin)
10. Radical new method of treating diabetes (New York Times)
11. Observations on the progressiveness of diabetes (F. Allen)