Alcoholic Ketoacidosis—Underrecognized Cause of Metabolic Acidosis in the Elderly

Mahesh Krishnamurthy, MD, FACP


The Substance Abuse & Mental Health Services Administration (SAMHSA) reported that substance abuse among adults age 60 years and older is a rapidly growing health problem. The report also stated that in 2000, 17% of Americans age 65 and older had problems with prescription drug and alcohol abuse.1 Most elderly people with alcohol abuse problems have a history of early-life alcohol abuse. However, a significant proportion start drinking later in life in response to traumatic life events such as the death of a loved one, loneliness, pain, insomnia, and retirement. This subset often experiences periods of binge drinking with little or no food intake. Alcoholic ketoacidosis (AKA) is an acute anion gap metabolic acidosis that typically occurs in people with a recent history of binge drinking and little or no nutritional intake. Some patients with AKA also have intractable vomiting and dehydration, and in these cases there is a concomitant metabolic alkalosis.

Case Presentation

An 86-year-old female, who had been a widow for the past 20 years, presented to the hospital with complaints of nausea, epigastric discomfort, and breathlessness for 2 days. She confessed to a history of alcohol abuse starting shortly after her husband’s death and to regularly consuming a pint of hard liquor each day. She had been on an alcohol binge for 4-5 days without eating any food, but she had stopped consuming alcohol because of nausea for 2 days prior to presentation. Although the patient was depressed and admitted to suicidal ideation, she denied ingestion of antifreeze, methanol-containing solvents, rubbing alcohol, or salicylates. Physical examination was remarkable for only signs of dehydration and epigastric tenderness. Fundoscopy did not reveal any retinal changes. On admission, the following levels were found: serum sodium, 142 mEq/L; potassium, 4.8 mEq/L; chloride, 101 mEq/L; bicarbonate, 6 mEq/L; anion gap, 35 mEq/L; blood urea nitrogen, 16 mg/dL; glucose, 117 mg/dL; blood ethyl alcohol, 0 mg/dL; arterial pH, 7.14; PCO2, 16 mm Hg; and PO2, 130 mm Hg. The serum osmolality was 341 mOsm/kg, as compared with a calculated value of 296.2 mOsm/kg, indicating an elevated osmolal gap of 44.8 mOsm/kg. The serum acetone concentration was 84 mg/dL (16.8 mmol/L); serum lactate, 2.3 mmol/L; amylase, 954 U/L; and lipase, 210 U/L. The urinalysis, including microscopy, was negative, except for the presence of ketones. A provisional diagnosis of AKA was made, as the patient had a characteristic clinical presentation preceded by an alcohol binge that was terminated by nausea, with the last alcohol intake being 2 days before coming to hospital. Moreover, serum and urinary ketones were strongly positive. However, there was a diagnostic dilemma, as the patient presented with high anion gap and high osmolal gap metabolic acidosis, which also raised the concern of simultaneous toxic alcohol ingestion (methyl alcohol or ethylene glycol). There were also no other clinical features such as fundus changes or oxalate crystals on urinalysis to support the diagnosis of toxic alcohol ingestion. In spite of this, it was difficult to exclude simultaneous toxic alcohol ingestion in the emergency setting, especially with an unreliable history. The patient was vigorously hydrated and promptly dialyzed in view of very high osmolal gap. The blood ethylene glycol, methyl alcohol, and isopropyl alcohol returned negative 3 days later from the reference laboratory. Therefore, the patient’s physicians retrospectively concluded that the high anion gap and osmolal gap in the patient was secondary to AKA. The patient made a good recovery and was discharged home with strict instructions to stop consuming alcohol.


Alcohol-related intoxications including AKA and toxic alcohol ingestions such as methanol, ethylene glycol, diethylene glycol, and propylene glycol can present with a high anion gap metabolic acidosis and increased serum osmolal gap. Isopropanol intoxication presents with hyperosmolality alone.2 AKA has to be in the differential diagnosis of an elderly patient presenting to the hospital with altered mental status and metabolic acidosis. There is increasing evidence that, rather than being benign and self-limiting, AKA may be a significant cause of mortality in patients with alcohol dependence.3

Pathophysiology of AKA

In the typical case of AKA, the onset is preceded by prolonged and massive alcohol intake, which the patient stops because of abdominal pain and repeated vomiting (usually because of gastritis or pancreatitis). Recurrent episodes of AKA have been well described.4 A marked deficiency of fluid and caloric intake in the days prior to the admission is the rule. This leads to relative excess of glucagon and catecholamines in the absence of liver glycogen stores. Moreover, because of prolonged oxidation of alcohol to acetaldehyde, there is an increased ratio of NADH to NAD+, which switches off the gluconeogenesis. This leads to marked increase in fatty acid oxidation to ketone bodies, which explains the high anion gap.5 Serum osmolal gap reflects low-molecular-weight non-ionic substances that are not accounted for by the calculated osmolality. Traditionally, high osmolal gap (> 10 mOsm/kg) has been attributed to toxic alcohol ingestion such as methyl alcohol, ethylene glycol, or isopropyl alcohol.5 Isopropyl alcohol, however, does not cause metabolic acidosis.5

There have been a few case reports of AKA with osmolal gap as high as 36.5 mOsm/kg.4,6-8 The case patient had a very high osmolal gap of 44.8 mOsm/kg. Thus, there is no reliable upper limit for osmolal gap in AKA. Braden et al8 have investigated the cause of this high osmolal gap in cases of AKA and have attributed it to glycerol, acetone, and acetone metabolites such as acetol and 1,2-propanediol. In the case patient, acetone contributed 16.8 mOsm/kg to the osmolal gap. Serum lactate and beta-hydroxybutyric acid do not, however, contribute to an osmolal gap, as they are largely ionized at physiological pH.8

Management Considerations

Current management focuses on the administration of saline and glucose to reverse volume and glycogen depletion, and the high NADH/NAD ratio. Thiamine 100 mg parenterally is an obligatory initial measure in the treatment of malnourished chronic alcoholics. Insulin therapy is contraindicated because of the risk of hypoglycemia, even in the presence of mildly elevated serum glucose, as in the case patient. Both the volume replacement and the glucose supply stimulate insulin release, leading to the termination of free fatty acid and ketone body production. Hemodialysis is extremely important, as it is not possible to rule out simultaneous toxic alcohol ingestion during acute presentations.

Outcome of the Case Patient

Three months later, the patient again presented with similar complaints. This time, her serum bicarbonate was 3 mEq/L; anion gap, 45 mEq/L; osmolal gap, 42 mOsm/kg; and she had an elevated serum ketone. Because of her prior history, she was not dialyzed and was treated with IV saline and dextrose only. The patient made a prompt recovery within 3 days.

Conclusions and Learning Points

AKA has to be in the differential diagnosis of an elderly patient presenting to the hospital with altered mental status and metabolic acidosis. There is increasing evidence that, rather than being benign and self-limiting, AKA may be a significant cause of mortality in patients with alcohol dependence.3 It has a characteristic clinical presentation occurring some time after termination of alcohol intake, often resulting in a negative blood alcohol level. Although not very well known, high osmolal gap is an important characteristic of AKA and is attributed mainly to acetone, glycerol, and acetone metabolites such as acetol and 1,2-propanediol. Because of the high osmolal gap in AKA, it is difficult at times to exclude the simultaneous presence of toxic alcohol. Given the potentially high morbidity and mortality of these intoxications, it is important for the clinician to have a high degree of suspicion for these disorders in cases of high anion gap metabolic acidosis, acute renal failure, or unexplained neurological disease so that treatment can be initiated early.2 Therefore, if clinical history is unreliable, it is safer to start hemodialysis.

The author reports no relevant financial relationships.

Dr. Krishnamurthy is Faculty in Charge, Geriatric Medical Education, Division of Geriatrics, Clinical Assistant Professor of Medicine, Drexel University, Philadelphia, PA, and Core Faculty, Internal Medicine Residency Program, Easton Hospital (Academic Affiliate Drexel University), Easton, PA.