Article Content

Diabetes mellitus is a metabolic syndrome characterized by hyperglycemia resulting from insufficient insulin secretion, insulin action, or both. Approximately 21 million children and adults in the United States have diabetes mellitus. Unfortunately, about 6.2 million of these cases remain undiagnosed. Additionally, more than 50 million people in the United States are thought to have "prediabetes" (American Diabetes Association, n.d.). Their glucose levels are higher than normal, but insufficiently high for a diagnosis of diabetes (American Diabetes Association, n.d.). Patients with life-limiting illnesses are more likely to have diabetes than the general population due to advanced age, the high prevalence of cancer (patients with cancer are more likely to have diabetes than those without cancer), and the use of medications that may elevate blood glucose.


Diabetes exists as 2 primary "types." Type 1 diabetes is predominantly an immune-mediated disease, resulting in autoimmune destruction of insulin-producing beta-pancreatic cells. Patients with type 1 diabetes are dependent on exogenous insulin secretion. They represent approximately 4% to 5% of all patients with diabetes.


The vast majority of diabetes patients have type 2 diabetes. These patients experience insulin resistance and a relative lack of insulin production. Type 2 diabetes is managed with medical nutrition therapy, exercise, and medications if necessary to improve insulin sensitivity as well as action or to enhance insulin secretion. Many patients with type 2 diabetes also use insulin as part or their entire management plan.


Significant research exists to show that improved blood glucose control delays the onset or progression of diabetes-related complications. For example, the American Diabetes Association recommends a glycosylated hemoglobin level lower than 7% (or even 6% if possible) as a 2- to 3-month measure of blood glucose control, a preprandial capillary plasma glucose level of 90 to 130 mg/dl, a peak postprandial capillary plasma glucose level lower than 180 mg/dl, and other goals for blood pressure and cholesterol management (American Diabetes Association, 2007). Whereas "tighter" blood glucose control is important for patients not facing a life-limiting illness, this is an unrealistic goal for patients with a terminal illness, and may in fact cause unacceptable symptoms such as hypoglycemia.


All patients with medical problems should have therapeutic goals. For terminally ill patients with diabetes, the glycosylated hemoglobin (A1c) level no longer provides useful information. The most important goal for an end-stage patient with diabetes is to prevent symptoms related to hyperglycemia or hypoglycemia. If the patient, family, or providers insist on monitoring blood glucose, a "looser" range is appropriate such as blood glucose maintained at 140 to 300 mg/dl, unless the patient has symptoms of hyperglycemia.


Medical nutrition therapy and physical activity are important parts of managing diabetes, but less so with hospice patients. As a matter of fact, terminally ill patients are unlikely to be participating in any significant physical activity, and their diet can be liberalized to focus on patient likes and dislikes (short of causing troubling symptoms).


When considering the drugs used to treat diabetes, it is important for healthcare providers to know which agents cause hypoglycemia and which do not. Insulin, of course, causes hypoglycemia, as does any oral medication that causes the secretion of insulin. Such medications include the oral sulfonylurea agents (e.g., glyburide, glipizide, glimepiride) and the glinides (repaglinide and nateglinide). Other medications such as metformin, the thiazolidinediones (rosiglitazone and pioglitazone), the alpha-glucosidase inhibitors (acarbose and miglitol), and newer agents such as exenatide, sitagliptin, and pramlintide are unlikely to cause hypoglycemia.


The management of patients with type 1 diabetes should continue as usual if the patient has stable nutritional status with good quality of life. If the patient has a consistent but declining appetite, you should consider switching the patient to once- or twice-daily intermediate- or long-acting insulin. If the patient has erratic oral intake, more reliance on rapid-acting insulin (e.g., insulin lispro) administered after meals should be considered if warranted. For actively dying patients (days to 1 week before death), discontinuation of insulin therapy should be considered.


Similarly, type 2 patients with stable nutritional status and good quality of life should continue current therapy. For a consistent declining appetite and weight loss, discontinuation of diabetes therapies should be considered, or reduction of hypoglycemic agents (e.g., insulin, sulfonylurea agents, glinides) by a minimum of 50%. For erratic meal intake, use of a glinide after a meal, based on carbohydrate intake, should be considered. Discontinuation of therapy should be considered for type 2 patients who are actively dying.


One of the most difficult challenges is to dissuade families from checking the patient's blood glucose multiple times per day and to help them be amenable to relaxed blood glucose control. It is important to stress that these interventions are not a "giving up" but rather a shifting of goals, with patient comfort as the primary goal.


Let's look at the case of Mrs. Ramirez, a 64-year-old Hispanic woman admitted to hospice with a diagnosis of breast cancer. The patient has had type 2 diabetes for approximately 15 years. She has lost approximately 50 pounds over the past 6 months. Currently, she weighs about 135 pounds (she is 5 feet 4 inches tall). At admission, Mrs. Ramirez was taking metformin 1,000 mg twice daily by mouth and glipizide 10 mg twice daily to manage her diabetes.


Mrs. Ramirez's appetite has waned, but she still eats 3 meals per day. Her preprandial blood glucose values have been dropping to less than 70 mg/dl on most occasions. She occasionally complains of palpitations, sweating, and "jitteriness." These objective and subjective data are consistent with hypoglycemia, and Mrs. Ramirez's diabetes medication regimen requires adjustment. Because she is still eating consistently, although less, it would be reasonable to cut the doses of both medications in half (i.e., metformin 500 mg and glipizide 5 mg twice daily). The family continues to monitor Mrs. Ramirez's blood glucose, and her preprandial values range from 100 to 140 mg/dl without symptoms of hypoglycemia or hyperglycemia.

Figure. No caption a... - Click to enlarge in new windowFigure. No caption available.

Mrs. Ramirez continues to hold her own for several more weeks. Then her oral intake becomes fairly erratic. On 2 occasions, she has experienced moderately severe symptoms of hypoglycemia, requiring administration of a rapid-acting carbohydrate. At this point, you recommend discontinuing the metformin and glipizide as well as the blood glucose monitoring. The family becomes quite upset, stating, "The doctor said we had to check Mamma's sugar 4 times a day to prevent complications such as her going blind or needing a kidney transplant. Are you giving up on our mother?"


What is the best way to handle these concerns, as well as Mrs. Ramirez's diabetes? First, it is important to stress to the family that you are not "giving up" on the patient. As a matter of fact, you are putting the patient's comfort first and foremost. An important discussion with the family should describe how a patient with advanced illness naturally has a declining appetite, and how continuing the diabetes medications increases the risk for toxicity, particularly hypoglycemia, which can be uncomfortable or even fatal. It is also worth pointing out that fingerstick testing is not without discomfort, and that monitoring Mrs. Ramirez for the symptoms of high or low blood sugar is sufficient. Using tact and diplomacy, you can explain to the patient and family that retinopathy and nephropathy are long-term complications of diabetes, and that at this point in the patient's disease, monitoring for short-term complications (e.g., hypoglycemia and hyperglycemia) is more important. Should Mrs. Ramirez develop symptoms of high blood glucose (increased thirst, hunger and urination), she can be restarted on a lower dose of glipizide (e.g., 2.5 mg once or twice a day).


Diabetes mellitus is a common comorbid condition among patients who are terminally ill, and appropriate management is different from the treatment for patients who do not have an advanced illness, but equally important. Communication and goal setting with the patient, family, and all providers is key to avoiding unnecessary pain and adverse effects from diabetes medications.




American Diabetes Association. (n.d.). Total prevalence of diabetes and prediabetes. Retrieved January 7, 2008 from [Context Link]


American Diabetes Association. (2007). Standards of medical care in diabetes-2007. Diabetes Care, 30(Suppl 1), S4-S41. [Context Link]