Diagnosis of Diabetic Ketoacidosis – What about beta-hydroxybutyrate level?

The root cause of diabetic ketoacidosis is a lack of insulin secretion in response to increased release of counter-regulatory hormones such as catecholamines, cortisol, glucagon, and growth hormone. These counter-regulatory hormones are secreted in response to a physiologic stressor such as infection. The concurrent increase in gluconeogenesis and glycogenolysis, in turn, worsens the hyperglycemia. It is important to stress that elevated glucose alone does not cause DKA. And the majority of the hyperglycemia seen in DKA is from these counter-regulatory hormones, not from oral glucose intake. (It’s OK to feed your hyperglycemic patients!)

The counter-regulatory hormones also cause breakdown of adipose tissue into free fatty acids. The fatty acids undergo beta-oxidation into the ketone bodies, acetoacetate and beta-hydroxybutyrate (BHB), or are used to increase production of VLDL and triglycerides. While acetoacetate is metabolized to acetone (causing the fruity breath of patients with DKA), BHB can cross the blood-brain barrier to be used as an energy source by the brain. Remember, the lack of insulin is causing the patient’s body to act as if it is hypoglycemic. The elevated triglycerides may cause pancreatitis, which will resolve with treatment of the underlying DKA. The breakdown of adipose tissue also causes release of prostaglandins, which along with acidemia, lead to the paradoxical vasodilation that occurs despite severe dehydration.

The worsening hyperglycemia results in osmotic diuresis and subsequent depletion of the intravascular volume, triggering the renin-angiotensin-aldosterone (RAA) system. This leads to depletion of total body potassium, although the patient’s extracellular potassium may be elevated from intracellular to extracellular shifts. In the recovery phase of DKA, chloride is retained in exchange for the ketoacid anions being excreted (see the below figure). This results in a loss of bicarbonate as well, creating a superimposed hyperchloremic normal anion gap metabolic acidosis. The point is: patients in DKA have significant electrolyte disturbances that need to be monitored closely. 

Finally, keep in mind there are innumerable triggers of DKA, and the trigger is often not identified. The most common trigger is infection, followed by medication noncompliance, myocardial infarction, and alcohol abuse. One should have a low threshold to test for and treat these etiologies.

Starvation/alcoholic ketoacidosis: There is also dysfunction with the glucose and insulin regulatory system, but the dysfunction originates from decreased glucose intake. The decrease in glucose results in a lack of insulin secretion, and fatty acid oxidation ensues. As such, measurement of the patient’s blood glucose is an important clue, and treatment involves administration of glucose in addition to thiamine to prevent development of Wernicke’s encephalopathy.

Hyperosmolar syndrome: There is a very high glucose level without (or with a minimal) anion gap, and the condition generally develops more gradually than DKA. The treatment is very similar to DKA, but should be more focused on volume repletion.

The traditional criteria according to the American Diabetes Association: glucose >250, bicarb <18, pH < 7.3, anion gap >10 and positive serum or urine ketones (1).

Image courtesy of Ogawa et al, J Diabetes Investig. 2016

But, there are some problems with these criteria. A patient on Invokana (canagliflozin) may have relatively lower fingerstick of 200 and have DKA. This is an example of euglycemic DKA. The mechanism isn’t completely understood, but the majority of euglycemic DKA patients are on SGLT-2 inhibitors.  

These drugs work by inhibiting the sodium glucose transport protein 2 in the proximal tubule of the kidney, thereby blocking the reabsorption of glucose by the kidney. At the same time, they can cause an increase in glucagon. The end result is a proportional increase in glucagon without a proportional increase in insulin – more glucose is excreted renally there, so there is less of a trigger for insulin secretion. The increase in glucagon (a counter-regulatory hormone) induces ketosis, and the continuing renal excretion of glucose helps keep the patient euglycemic (2). 

So now that we know we can’t always rely on glucose level for diagnosis, what about pH? In most cases, the pH will be less than 7.3, but DKA may involve a mixed picture – there is anion gap acidosis (ketones), hyperchloremic acidosis (from ketoanion exchange for chloride in the kidney), and a contraction metabolic alkalosis. Therefore, pH is not always reliable.

So if we can’t use pH or glucose as a diagnostic guideline, then what can we use?

BHB measurement

Let’s look back at our slide showing the generation of ketones. Fatty acids are metabolized to acetyl-coenzyme A, which can’t enter the citric acid cycle. So Acetyl coA is diverted to ketone body production – enzymatic conversion of acetyl coA to acetoacetate. Acetoacetate is then reduced (in acidemic states) to BHB and acetone.

Urinalysis detects acetoacetate and is not a precise estimate of blood ketones. There may be false positives – if patient is on drugs with sulfhydryl group like captropril (3) – and false negatives – if the urine is acidic or exposed to air.

In DKA, beta hydroxybutyrate has a concentration 3-10 times higher than acetoacetate. As you treat DKA, the beta hydroxybutyrate shifts back to acetoacetate, which can elevate blood and urine ketone levels, giving the false impression that DKA is worsening.  

In hyperglycemia of > 250 mg/dL, a BHB level of more than 1.5 mmol/L has sensitivity ranging from 98-100%, specificity of 85%% – positive likelihood ratio (LR) of 6.7 and a negative LR of 0.02 for diagnosing DKA (4). At a cutoff of 3 mmol/L the sensitivity is almost 100% and specificity of 94% (5).

Bottom line: Beta-hydroxybutyrate is better for diagnosis and monitoring treatment of DKA.

1. DKA is primarily due to a lack of insulin combined with elevated counter-regulatory hormones rather than high blood glucose
2. Patients will have significant electrolyte shifts, so monitor them closely
3. Search for the DKA trigger – particularly infection
4. Patients on SGLT2 inhibitors can develop DKA with euglycemia
5. BHB measurement is the most specific way to diagnose DKA