J.T. is a 72-year-old man with chronic hepatitis C and Child-Pugh grade A (clinically well-compensated) cirrhosis. He takes propranolol (propranolol 20 mg PO BID) for esophageal variceal bleeding prophylaxis. He had a blood transfusion 25 years ago. Hepatitis C was diagnosed 10 years ago, and cirrhosis was diagnosed by liver biopsy 2 years ago. He does not drink alcohol. He has never been overweight. He has no personal or family history of diabetes. Over the past year, random plasma glucose levels have ranged from 110 to 180 mg/dl. The most recent random glucose was 210 mg/dl. The patient denies polydipsia, polyuria, nocturia, or any other symptoms of hyperglycemia. He weighs 150 lb. (BMI 22 kg/m2).

Physical examination findings are normal except for mild palmar erythema, spider angiomata on the upper chest, and a palpable spleen tip. Fasting blood glucose was 136 mg/dl, and hemoglobin A1c (A1C) was 6.3%. Another fasting glucose several weeks later was 128 mg/dl.

At first glance, many clinicians might assume this patient has type 2 diabetes. The history is compatible with this diagnosis. However, the absence of classic risk factors for type 2 diabetes and the appearance of new hyperglycemia in the setting of known cirrhosis makes it more likely he has “liver diabetes,” also known as hepatogenous diabetes.1,2 Patients with cirrhosis have insulin resistance. Impaired glucose tolerance (IGT) is common, and about 20–40% have diabetes.1,3 While there is no definitive test to distinguish type 2 diabetes from diabetes caused by liver disease, liver diabetes appears to be caused by hepatic dysfunction. It should be noted that the American Diabetes Association and the World Health Organization do not recognize liver diabetes as a specific type of diabetes. Regardless of whether the diagnosis is that of liver diabetes or type 2 diabetes, decisions about when and how to treat hyperglycemia should take into account comorbid conditions such as hepatic dysfunction.

This patient has only a minimal elevation in A1C, and the value is within standard treatment goals for diabetes. However, it should be noted that A1C reference ranges assume a normal erythrocyte life span. Older erythrocytes have higher A1C levels than younger cells. Any condition that reduces erythrocyte survival, such as cirrhosis4 or hemolysis resulting from hypersplenism can cause spuriously low A1C levels. Therefore, in this patient, it would be desirable to institute home blood glucose monitoring in order to better assess the severity of his hyperglycemia. The decision about whether to start treatment for any condition is based on a comparison of the risks and benefits of that treatment. First, a review is in order of the risks of each therapeutic option that should be considered for patients with hepatic dysfunction.

Diet and exercise are usually considered a very safe first-line of therapy for patients with mild hyperglycemia. However, many patients with cirrhosis are malnourished, and dietary restriction with a goal of weight loss may exacerbate hypoalbuminuria and worsen overall prognosis. If dietary restriction results in lower vitamin K intake, then a coagulopathy may result. Every class of oral hypoglycemic medication currently available in the United States has been associated with at least a small risk of hepatotoxicity. For patients with marginal hepatic function at baseline, even mild hepatotoxicity can be fatal. Hepatic dysfunction can also cause an exaggerated response to a standard dose of medication and a higher risk of side effects if the drug is metabolized by the liver. Sulfonylureas, repaglinide, metformin, and thiazolidinediones are all extensively metabolized by the liver. It is generally advised that metformin and thiazolidinediones should not be used in patients with significant hepatic dysfunction. For these reasons, many clinicians use insulin as a first-line agent to treat diabetes in cirrhotic patients. The main risk of insulin is severe hypoglycemia. Patients with cirrhosis have reduced hepatic glycogen stores. Glucagon may stimulate less hepatic glycogenolysis in cirrhotic patients than in patients without liver disease.1 Also, many patients with severe hepatic dysfunction have hepatic encephalopathy, which may impair their ability to comply with instructions about therapy. Patients with cirrhosis and diabetes have a shorter life expectancy than do nondiabetic patients with cirrhosis, but they typically die of complications of liver disease, such as gastrointestinal hemorrhage, rather than from complications of diabetes, such as cardiovascular disease.2,3,5 This suggests that in cirrhotic patients, the development of diabetes reflects a greater degree of liver failure. No studies have been conducted to determine whether patients with cirrhosis benefit from diabetes treatment. However, there are several situations in which cirrhotic patients would be expected to benefit from glucose control. Treatment for symptomatic hyperglycemia should be used to reduce symptoms. Treatment of persistent hyperglycemia would be expected to lessen the risk of infection. Patients with A1C results ≥ 7% who are awaiting liver transplantation or whose life expectancy is expected to be several years might benefit from a lower risk of diabetes complications if their diabetes is treated.

In J.T.’s case, he was observed off therapy for about 6 months. He eventually started low-dose insulin (lispro 3 units SQ before meals and glargine 5 units SQ Q HS) for persistent hyperglycemia > 200 mg/dl, A1C > 7.5%, and patient preference. He did not have any episodes of severe hypoglycemia.


You are the nurse during his initial visit to the hospital. After completing a comprehensive health history and physical examination, you move on to provide client education on his medications and three health promotion topics appropriate to his case.

Clinical Pearls

• Severe hepatic dysfunction can cause IGT and diabetes. The clinical distinction between type 2 diabetes and liver diabetes is based on the onset of diabetes relative to the onset of cirrhosis and on whether the patient has typical risk factors for type 2 diabetes.

• A1C results may be spuriously low in patients with severe liver dysfunction.

• All currently available oral hypoglycemic agents pose some risk of hepatotoxicity. Metformin and thiazolidinediones should be avoided in patients with significant hepatic dysfunction. Many clinicians consider insulin to be the first-line agent for treating diabetes in patients with significant liver disease, although some clinicians advocate the cautious use of sulfonylureas in this situation.1

• Patients with cirrhosis are especially susceptible to hypoglycemia and may respond poorly to glucagon.

• Among patients with cirrhosis and diabetes, the main cause of death is hepatic failure rather than cardiovascular disease or other complications of diabetes.

• An individualized assessment of risks of benefits of diabetes treatment should be considered for each patient.

Marguerite McNeely, MD, MPH, is an assistant professor in the Division of General Internal Medicine at the University Of Washington School Of Medicine in Seattle.


1 Petrides AS: Liver disease and diabetes mellitus. Diabetes Revs 2:2–18, 1994 2

2Holstein A, Hinze S, Thieben E, Plaschke A, Egberts E-H: Clinical implications of hepatogenous diabetes in liver cirrhosis. J Gastroenterol Hepatol 17:677–681, 2002

3 Marchesini G, Ronchi M, Forlani G, Bugianesi E, Bianchi G, Fabbri A, Zoli M, Melchionda N: Cardiovascular disease in cirrhosis. Am J Gastroenterol 94:655–662, 1999

4 Owens D, Jones EA, Carson ER: Studies on the kinetics of unconjugated [14C] bilirubin metabolism in normal subjects and patients with compensated cirrhosis. Clin Sci Mol Med 52:555–570, 1977

5 Bianchi G, Marchesini G, Zoli M, Bugianesi E, Fabbri A, Pisi E: Prognostic significance of diabetes in patients with cirrhosis. Hepatology 20:119–125, 1994