In an announcement released May 15, the FDA called attention to 20 case reports it had received between March 2013 and June 6, 2014 of 20 persons, most with type 2 diabetes, who had been treated with Sodium-glucose Cotransporter-2 (SGLT2) Inhibitors such as canagliflozin (Invokana), dapagliflozin (Farxiga, Forxiga), and empagliflozin (Jardiance), for an average of 2 weeks, but in some for as much as six months and who had diabetic ketoacidosis. Diabetic ketoacidosis is typically a condition of uncontrolled diabetes seen in type 1 diabetes, or in adolescents with severe type 2 diabetes, with elevated blood glucose and evidence of ketoacidosis with elevated blood or urine ketones and acidosis with a high blood “anion gap,” reflecting the presence of substances called organic anions in the bloodstream. Another unusual feature noted in the FDA release was that blood glucose levels were typically only mildly elevated, below 200 mg/dl.
What is ketoacidosis?
The body uses insulin not only to move glucose into cells, but also as a signal to increase fat and protein synthesis in fat cells and other body tissues. When a persons who does not have diabetes goes a long period without eating, ketones act as an important source of energy. Their insulin concentrations in blood fall, acting as a negative signal to cause the body to break down fat into fatty acids and protein into amino acids. A subsequent step, also signaled by low levels of insulin, is the further breakdown of fatty acids and the removal of amino groups from certain amino acids to form ketone bodies, particularly an organic acid called beta-hydroxy butyrate, as well as acetoacetic acid, which is less important in ketoacidosis but which is measured in urine ketone test strips. For a diabetic person who has true insulin deficiency, as usually seen in type 1 diabetes, when untreated, or markedly undertreated, sometimes even for just a several hour period, the body shifts from storage of nutrients such as glucose, fat, and proteins to their release and to formation of ketones. This particularly can occur under stress, with infection, injury, or illness, and diabetic ketoacidosis develops with a typical pattern of high glucose levels, acidosis, often with a dangerous degree of dehydration.
What might occur with the SGLT2 inhibitors?
These new glucose-lowering medicines, developed for type 2 diabetes, take advantage of the fact that the kidneys filter around 180 grams of glucose per day. In people who do not have diabetes, very little appears in the urine, so it all must be reabsorbed, an energy-requiring process that uses transport molecules, the most important of which is SGLT2. To put this in perspective, the normal blood glucose is less than 100 mg per 100 ml, or about 1 g per liter. For an adult with 4-5 liters of blood, then, there are only about 4-5 g of glucose. Extracellular body water is approximately 20 liters, so at any time this space might hold about 20 g of glucose. More glucose is stored in tissues, particularly in the liver, but it can easily be understood that inhibitors of SGLT2 can be potent glucose lowering drugs, even if they only cause loss of 30-50 g of glucose per day – an amount that is approximately as much as produced by the liver per day. These drugs lower glucose levels in a fashion not depending on insulin, and so, when given to diabetic patients who do not produce sufficient insulin, one could imagine that ketone production can take place even when the blood glucose is being lowered.
This seems to have occurred in two diabetic patients who developed ketoacidosis described in a recent article from the United Kingdom (Hine J, Paterson H, Abrol E, Russell-Jones D, Herring R. SGLT inhibition and euglycaemic diabetic ketoacidosis. Lancet Diabetes Endocrinol 2015 Published Online May 27, 2015). Although not having type 1 diabetes, one had developed diabetes because of chronic pancreatitis, and the other had type 2 diabetes but subsequently had removal of part of the pancreas for a tumor. Both were treated with dapagliflozin, but were under stress, had ketoacidosis, but had blood glucose under 200 mg/dl. The authors commented that “by masking the hyperglycaemia typically associated with insulin deficiency, SGLT2 inhibition could delay diagnosis in these patients with the attendant risk of progression to life-threatening metabolic derangement.” It is interesting that two careful studies of the effect of SGLT2 inhibitors showed that they do not lower blood glucose as much as would be expected, as there appears to be an increase in glucose production by the liver with these agents. This may be caused by an increase in production of the hormone glucagon, which could also be a stress factor that might increase the risk of ketoacidosis developing.
What should be done?
First, we need to recognize that only a tiny number of cases of diabetic ketoacidosis have been reported with SGLT2 inhibitor treatment, so this is unlikely to be a major issue for these drugs, which are very useful in controlling blood glucose in a number of ways. We should, though, be careful in using SGLT2 inhibitors in people who have relatively marked insulin deficiency, both with type 1 diabetes, for whom the drugs are being investigated but are not currently approved, and for certain people with more severe forms of type 2 diabetes. For these patients, insulin or other medicines which increase insulin production so that there is little risk of developing real deficiency will be needed. And we do need to remember that diabetes is not a condition for which a certain approach to treatment is always appropriate for a given person. Anyone with diabetes who feels they might be becoming ill should be in touch with their physician, and, for my patients who have greater degrees of insulin deficiency, when I use SGLT2 inhibitors I am now beginning to recommend keeping urine ketone test strips to use in case of any symptoms of fever, gastrointestinal upset, etc.