Ketoacidosis is a high anion gap metabolic acidosis due to an excessive blood concentration of ketone bodies (keto-anions). Ketone bodies (acetoacetate, beta-hydroxybutyrate, acetone) are released into the blood from the liver when hepatic lipid metabolism has changed to a state of increased ketogenesis. A relative or absolute insulin deficiency is present in all cases. The major reactions starting from the production of acetoacetate from hepatic acetyl CoA are outlined in the box.
The major ketone bodies are acetoacetate and beta-hydroxybutyrate and the ratio between these two acid anions depends on the prevailing redox state (eg as assessed by the NADH/NAD+ ratio).
A mixed acid-base disorder may be present (eg lactic acidosis from peripheral circulatory failure, or metabolic alkalosis from vomiting). An associated lactic acidosis may mask the presence of the ketoacidosis. This occurs because the lactic acidosis decreases the acetoacetate : beta-hydroxybutyrate ratio (ie more beta-hydroxybutyrate produced ) because NAD+ is produced in the production of lactate. The common test used to detect ketones (eg ‘Acetest’) depends on the reaction of acetoacetate (and to a lesser extent acetone) with the nitroprusside reagent. A decreased acetoacetate level may lead to a weak or absent test reaction despite high total levels of total ketoanions (acetoacetate and beta-hydroxybutyrate combined) because the beta-hydroxybutyrate is not detected.
The three major types of ketosis are:
When hepatic glycogen stores are exhausted (eg after 12-24 hours of total fasting), the liver produces ketones to provide an energy substrate for peripheral tissues. Ketoacidosis can appear after an overnight fast but it typically requires 3 to 14 days of starvation to reach maximal severity. Typical ketoanion levels are only 1 to 2 mmol/l and this will not much alter the anion gap. The acidosis even with quite prolonged fasting is only ever of mild to moderate severity with ketoanion levels up to a maximum of 3 to 5 mmol/l and plasma pH down to 7.3. This is probably due to the insulin level, which though lower, is still enough to keep the FFA levels less than 1mM. This limits substrate delivery to the liver restraining hepatic ketogenesis. Ketone bodies also stimulate some insulin release from the islets. The anion gap will usually not be much elevated.
This typical situation leading to alcoholic ketoacidosis is a chronic alcoholic who has a binge, then stops drinking and has little or no oral food intake. Food intake may be limited because of vomiting. The two key factors are the combination of ethanol and fasting. Presentation is typically a couple of days after the drinking binge has ceased.
The poor oral intake results in decreased glycogen stores, a decrease in insulin levels and an increase in glucagon levels. Hepatic metabolism of ethanol to acetaldehyde and then to acetate both involve NAD+ as a cofactor. The NADH/NAD+ ratio rises and this:
The insulin deficiency results in increased mobilisation of free fatty acids from adipose tissue. The decreased insulin/glucagon ration results in a switch in hepatic metabolism favouring increased beta-oxidation of fatty acids. This results in an increased production of acetylCoA which forms acetoacetate (a keto-acid). (The pathophysiology of the insulin deficiency and the switch in hepatic metabolism is discussed in more detail in DKA section below.)
Other points to note:
This syndrome is rapidly reversed by administration of glucose and insulin.
This diagnosis is often overlooked. A strong suspicion should be raised in any ill chronic alcoholic with a sweet ketone breath who presents to a hospital Emergency Department. Such patients are often dishevelled, and can be noisy and generally uncooperative.
A mixed acid-base disorder may be present: high anion gap due to ketoacidosis, metabolic alkalosis due to vomiting and a respiratory alkalosis.
An absolute or relative lack of insulin leads to diabetic metabolic decompensation with hyperglycaemia and ketoacidosis. A precipitating factor (eg infection, stress) which causes an excess of stress hormones (which antagonise the actions of insulin) may be present.
Since the discovery and therapeutic use of insulin, the mortality from DKA has dropped dramatically from 100% to perhaps 2 to 5% in Western countries today. (Lebovitz, 1995)
FFA mobilisation is initiated by the effect of absolute or relative insulin deficiency on fat cells. FFA levels can be quite high (eg 2.5 to 3.5 mM). This provides the liver with plenty of substrate. These FFA levels are much less then ketone levels and contribute only a small amount to the metabolic acidosis.
The major switch in hepatic lipid metabolism occurs in response not just to insulin deficiency but additionally to the concomitant rise in levels of the stress hormones (glucagon, corticosteroids, catecholamines, growth hormone). The role of glucagon is the most clearly established. The hepatic effects of a fall in the insulin:glucagon ratio are:
The net effect is an increase in the hepatic output of both ketone bodies and glucose.
The inhibition of the enzyme acetyl CoA carboxylase is probably the key step. This enzyme is inhibited by increased FFA levels, decreased insulin levels and particularly by the rise in glucagon. All three of these factors are present in DKA. The effect is to decrease the production and level of malonyl CoA. This compound has a central role in the regulation of hepatic fatty acid metabolism as is mediates the reciprocal relationship between fatty acid synthesis and oxidation. It is the first committed intermediate in fatty acid metabolism. Malonyl CoA inhibits fatty acid oxidation by inhibiting carnitine acyltransferase I.
A fall in malonyl CoA levels removes this inhibition resulting in excessive fatty acid oxidation with excessive production of acetyl CoA and excess acetoacetate.
Two basic mechanisms underlie the pathophysiology of DKA: hyperglycaemia and ketoacidosis. The above discussion shows how both these problems follow from relative insulin deficiency coupled with stress hormone excess. The problem however is not just of hepatic over-production of glucose and ketones but also of peripheral underutilisation of both glucose and ketones.
Acetoacetic acid (pKa 3.58) and beta-hydroxybutyric acid (pKa 4.70) dissociate producing H+ which is buffered by HCO3- in the blood. For each anion produced there is a loss of one bicarbonate. The increase in the anion gap (representing the increase in the unmeasured acid anions) should approximately equal the decrease in the [HCO3-]. A ‘pure’ high anion gap metabolic acidosis results.
In some cases, a hyperchloraemic metabolic acidosis develops: this is most common during the treatment phase. Why does this occur? Acetoacetate and beta-hydroxybutyrate are moderately strong acids and even at the lowest urinary pH are significantly ionised. They are excreted with a cation (usually Na+ or K+) to maintain electroneutrality. The net effect is the loss of ‘potential bicarbonate’ equal to the level of urinary ketone body loss. The HCO3- is replaced in the blood by Cl- derived from renal reabsorption, gut absorption or (particularly) IV saline administered during treatment. The effect is to cause a rise in plasma [Cl-] and the anion gap returns towards normal despite the persistence of the metabolic acidosis. At presentation, both types of acidosis may be present and the elevation in the anion gap will be less than expected for the degree of depression in the bicarbonate level (resulting in Delta ratio < 0.8).
A predominant hyperchloraemic acidosis (defined as a DKA patient with a delta ratio < 0.4) is present in about 10% of patients on arrival at hospital and in about 70% after 8 hours of treatment. Patients who are more severely dehydrated retain more keto-anions and have a lower incidence of hyperchloraemic acidosis.
Administration of large volumes of normal saline in resuscitation of patients with acute DKA promotes continued diuresis (and continued loss of ketone bodies with Na+ as the cation) and provides plenty of chloride to replace the lost ketoanions. This hyperchloraemic acidosis is slower to resolve because the keto-anions needed for regeneration of bicarbonate have been lost. Patients who have been able to maintain fluid intake during development of their illness are more likely to have a hyperchloraemic acidosis component present on admission.
It should not just be assumed that the patient only has a diabetic ketoacidosis. Possible complicating acid-base disorders are:
Renal tubular acidosis (type 4) is present in some diabetic patients and the associated urinary acidification defect can cause a hyperchloraemic normal anion gap acidosis. This syndrome (known as hyporeninemic hypoaldosteronism) occurs in some elderly diabetics who have pre-existing moderate renal insufficiency but is not a common problem in acute DKA.