Questions and Answers: Refeeding Patients with Concomitant Type 1 Diabetes and Anorexia Nervosa

Q. Are there any new approaches to managing patients with anorexia nervosa (AN) complicated by Type 1 diabetes mellitus (DM)? (GP, Austin, TX)

A. These patients can pose serious treatment challenges, whether in long-term management or during acute refeeding. When eating disorders and Type I DM co-occur, risks for microvascular diabetic complications such as retinopathy, nephropathy, and neuropathy are known to be increased (Rydell, et al., N Engl J Med. 1997; 332: 1849). Contemporary approaches to diabetic management, such as sliding-scale insulin protocols or basal-bolus protocols, try to achieve tight glucose control (glucose levels <140 mg/dL) in hopes of avoiding retinopathy and nephropathy, as well as neuropathy (N Engl J Med. 1993; 329:977).

In a recent article in the International Journal of Eating Disorders, Drs. Carrie Brown and Philip S. Mehler argue the case for using an alternative approach of “permissive hyperglycemia” for such patients (Int J Eat Disord. 2014; 47:671). The idea of a “permissive” metabolic derangement has gained acceptance in other fields. Perhaps the best example is permissive hypotension or low blood pressure in patients with bleeding emergencies. The rationale is that severe comorbid AN places a patient at increased risk for dangerous levels of hypoglycemia due to a lack of normal glycogen (sugar) stores in the liver, which would typically help to prevent bouts of hyperglycemia.

The authors report an intriguing and challenging case of a 23-year-old woman with a history of severe restricting subtype AN and Type 1 DM dating back to childhood. The young woman had a long history of multiple admissions for treatment of diabetic ketoacidosis. When the critically ill woman was admitted for treatment, her body mass index was only 9.1 kg/m². There were laboratory signs of starvation hepatitis and her HbA1C on admission was 9.8%, indicating recent poor glucose control.

Despite a dietary plan of 1400 kcal per day, the patient developed signs of the refeeding syndrome, and required intravenous replenishment of potassium, magnesium, calcium, and phosphorus. Her serum blood glucose levels dropped as low as 24 mg/dL without any insulin administration shortly after starting the dietary plan. The authors theorize that this was due to her profound malnourished state and the residual effects of intermediate-insulin doses she had self-administered just before being admitted to the hospital. Thus, on day 2 of hospitalization, intravenous 5% dextrose in normal saline (50 mL/hr) was needed to prevent hypoglycemia.

The patient was transferred to the intensive care unit, but remained susceptible to severe hyperglycemia but was also prone to hypoglycemia. After receiving IV dextrose, her blood sugar level quickly rose above 200 mg/dL. An insulin drip was also started but eventually she transitioned to subcutaneous long-acting glargine insulin, at 1 unit every 24 hours. As treatment continued, the patient had wide fluctuations in serum glucose levels over 200 mg/dL, but only rare episodes of hypoglycemia.

As the authors note, diabetic patients with severe AN, especially early in refeeding, have insulin needs that are markedly diminished from what one might expect for a patient who by definition has no inherent ability to produce and secrete any endogenous insulin. This is likely explained by low weight, poor liver function, and low fat stores. The authors also pointed out that while it is clear that excessive hyperglycemia is associated with microvascular complications, these concerns are only relevant over the course of the lifetime of a patient with Type 1 diabetes and are unlikely to be clinically significant for a period of only weeks during a structured refeeding program for a diabetic patient with severe AN—as long as the level of hyperglycemia is not excessive (defined as <250 mg/dL). Thus allowing “permissive hyperglycemia” is certainly more conducive to building the therapeutic trust so critical during the refeeding program of a patient with AN.

The authors recommend that this approach be followed during the early stages of refeeding, as opposed to a nutritional rehabilitation program, where tight glucose control is sought. The latter approach would increase the risk of hypoglycemia, necessitating rescue sugar, and causing more distress to the patient.

As in the case of their patient, the authors suggest that permissive hyperglycemia be included in the refeeding guideline for Type 1 diabetes patients with AN. A blood glucose level in the range of 180 to 240 mg/dL is in harmony with the recently published Standards of Medical Care in Diabetes, for persons with diabetes and a significant chronic medical comorbidity that places them at risk for hypoglycemia.

— SC