Metabolism

Between 2% and 10% of ingested ethanol is excreted intact by the kidneys and lungs, but the major fraction is metabolized by hepatic alcohol dehydrogenase (ADH) to acetaldehyde by the following reaction²:

At high blood ethanol levels, a particular isoform of the hepatic microsomal cytochrome P450 enzyme (CYP2E1) provides an additional, albeit normally minor, oxidative pathway for ethanol metabolism:

This alternative pathway is inducible with chronic ethanol exposure. Minor amounts of ethanol can also be metabolized by peroxisomal catalase:

Acetaldehyde produced by any of the preceding reactions is converted by hepatic acetaldehyde dehydrogenase (ALDH) to acetate:

Acetate can then enter the tricarboxylic acid cycle and ultimately be metabolized to carbon dioxide (CO₂) and water. Polymorphisms in the dehydrogenase enzymes can result in increased production rates or diminished metabolic clearance of acetaldehyde. As a consequence, some individuals experience marked vasodilation, facial flushing, tachycardia, and other unpleasant symptoms after ethanol consumption because of the effects of excessive acetaldehyde accumulation. Alleles leading to this reaction are particularly prevalent in persons of Chinese or Japanese descent but are uncommon in Caucasians.²

Metabolic conversion of ethanol to acetaldehyde and acetate by dehydrogenases raises the ratio of reduced nicotinamide adenine dinucleotide (NADH) relative to its oxidized form (NAD⁺). This change in intracellular redox state favors conversion of pyruvate to lactate by lactate dehydrogenase (LDH) and can thereby raise the blood lactate concentration:

The resulting increase in blood lactate level is usually small, however, and the presence of lactic acidosis should prompt consideration of an alternative cause such as circulatory shock or seizures.³

Ethanol elimination generally follows zero-order kinetics, with elimination rates of 5 to 10 g/h in nonhabituated subjects, approximately corresponding to a fall in blood ethanol concentration of 10 to 25 mg/dL/h. This rate can more than double in individuals who are chronically habituated to high doses of ethanol.

Clinical Manifestations

Excessive chronic ingestion of ethanol plays a causative role in a number of important diseases such as cirrhosis, hepatitis, pancreatitis, cardiomyopathy, and malignancies. Ethanol use can result in gastrointestinal hemorrhage by several mechanisms including gastritis, ulcers, esophageal varices, and Mallory-Weiss tears.

Acute intoxication can induce cardiac dysrhythmias, particularly atrial fibrillation. As denoted by the descriptive sobriquet, “holiday heart syndrome,” this phenomenon frequently occurs during an alcoholic binge. A variety of neurologic abnormalities are associated with chronic alcoholism, including Wernicke-Korsakoff syndrome, chronic cerebellar ataxia, Marchiafava-Bignami syndrome, and central pontine myelinolysis.⁴ Wernicke encephalopathy can manifest as lethargy, confusion, truncal ataxia, nystagmus, and ophthalmoplegia, whereas Korsakoff dementia manifests as retentive memory impairment, confabulation, and learning deficits.⁵

Acutely, ethanol has well-known, dose-dependent inebriating and sedating effects (Table 171-1), although remarkable variability in this relationship is observed in some individuals.⁴ These central nervous system (CNS) effects appear to be at least partly caused by interference with N-methyl-D-aspartate receptor and perhaps γ-aminobutyric acid receptor function.^4,6,7 The cognitive, behavioral, perceptual, and psychomotor effects of ethanol intoxication play a causative role in a substantial proportion of deaths and injuries involving motor vehicle–related trauma, accidental drownings, residential fires, homicides, and suicides. The legal driving threshold for blood ethanol concentration is 80 mg/dL in the United States for operators aged 21 years or older. Tachycardia, mydriasis, diaphoresis, hypotension, and hypothermia can occur in cases of marked intoxication. Blood ethanol concentrations of approximately 350 mg/dL have been associated with fatal outcomes, although many patients have survived much higher levels, including one subject who reportedly survived a level of 1500 mg/dL.⁸

TABLE 171-1 Relationship Between Blood Ethanol Concentration and Clinical Manifestations*

* This information serves only as an imperfect guide, because considerable variability and overlap is possible, and individuals with chronic heavy ethanol exposure often develop learned tolerance.

Laboratory Manifestations

Blood ethanol concentration correlates at least approximately with the manifestations of intoxication (see Table 171-1). In chronic alcoholic subjects, a blood ethanol concentration below 250 mg/dL is an unlikely explanation for alterations in consciousness and should prompt a search for an alternative cause.⁸ Numerous other blood test abnormalities can be seen in intoxicated subjects, particularly in patients with chronic ethanol abuse: hyponatremia, hypokalemia, hypomagnesemia, hypophosphatemia, hypoglycemia, thrombocytopenia, and coagulopathy. Elevated activities of various circulating enzymes including amylase, lipase, creatine phosphokinase, transaminases, and γ-glutamyl transpeptidase, can occur as a reflection of alcohol-induced pancreatitis, rhabdomyolysis, hepatitis, or cirrhosis. The latter can also result in hyperbilirubinemia and hypoalbuminemia.

Treatment

In the absence of associated illness or injury (Table 171-2), mild to moderate intoxication requires no special treatment other than abstinence and a period of observation. Regardless of the degree of intoxication, withdrawal precautions are recommended for chronic imbibers, particularly those with a history of heavy chronic use or alcohol withdrawal manifestations. The treatment of severe ethanol intoxication is largely supportive. As with any patient who presents to the hospital in an unconscious state, initial empirical treatment should include IV thiamine, dextrose, and naloxone, once adequate airway, ventilation, and perfusion are ensured. Gastric lavage and activated charcoal administration are of dubious value for hastening removal of ethanol from the body.^9–12

TABLE 171-2 Concomitant or Complicating Disorders Associated with Alcohol Intoxication or Withdrawal

The unconscious, stuporous, or delirious patient with ethanol intoxication can present a diagnostic challenge. Historical information is often lacking or inadequate, and the physical examination can be compromised by lack of cooperation. A central concern is that another disorder may be present in lieu of or in addition to ethanol intoxication. The other disorder may be chiefly responsible for the alteration in consciousness or may require specific urgent treatment. For example, inebriated subjects are at high risk for trauma (e.g., battery, falls, motor vehicle accidents) and therefore should be evaluated for physical injuries. Subdural hematoma is a particular concern, and any findings or suspicion of head injury should prompt cranial imaging by computed tomography. Chronic ethanol abuse also predisposes to infection, particularly aspiration pneumonia. Pneumococcal or Listeria meningitis, although not as common, is a consideration in the intoxicated patient with an altered sensorium, fever, and other compatible findings. Additional potentially confounding problems include concomitant toxic ingestions or drug overdoses, psychiatric disorders, alcohol withdrawal, and in patients with advanced cirrhosis, hepatic encephalopathy or spontaneous bacterial peritonitis.

A thorough evaluation for common associated illnesses and injuries should include physical and laboratory examinations for evidence of head, neck, and somatic trauma, rhabdomyolysis, pancreatitis, hepatic dysfunction, coagulopathy, blood dyscrasias, and fluid and electrolyte derangements. Accordingly, routine laboratory testing should include a complete blood count, prothrombin and partial thromboplastin times, serum assays for electrolytes (including sodium, potassium, chloride, total CO₂ content, magnesium, and phosphorus), glucose, liver and kidney function tests, and amylase, lipase, transaminases, and creatine phosphokinase activities. Screening for alternative or concomitant intoxications or overdoses is occasionally fruitful.¹³ Identification of metabolic acidosis should prompt investigation for alcoholic ketoacidosis, lactic acidosis, renal failure, and relevant toxic ingestions, particularly methanol and ethylene glycol. Microbiological cultures are indicated if there are signs of serious infection.

Intravenous thiamine and a multivitamin preparation containing folate are routinely administered to hospitalized patients with alcohol intoxication or withdrawal. Parenteral thiamine (50 or 100 mg) is given during the initial phase of management, regardless of the level of sensorium, to prevent or treat Wernicke-Korsakoff syndrome.⁵

Hydration is necessary in some intoxicated patients. Dextrose-containing saline solutions are usually the fluid of choice to correct dehydration and prevent hypoglycemia. Dextrose administration is traditionally preceded by thiamine dosing. Patients with hypoglycemia require rapid IV injection of dextrose followed by a continuous dextrose infusion titrated to the results of frequent serial blood glucose tests. Hypokalemia, hypomagnesemia, and hypophosphatemia should be corrected with the use of appropriate oral or parenteral supplementation. Patients with anemia may require further investigation for gastrointestinal hemorrhage. Patients requiring admission to an intensive care unit (ICU) should have a chest radiograph as well as electrocardiographic evaluation.

Oxygenation may be assessed either by pulse oximetry or by arterial blood gas analysis, and supplemental oxygen should be provided as necessary. Administration of vitamin K, fresh frozen plasma, or platelet transfusions may be necessary if there is gastrointestinal or other hemorrhage and coagulopathy or severe thrombocytopenia. The level of consciousness should be monitored periodically. Hemodialysis has been employed and is effective at removing ethanol from the body, but in general, this modality poses greater risks than simply providing supportive care and allowing physiologic ethanol elimination. Its use might be warranted in rare cases of profound life-threatening ethanol intoxication, or if there are other reasons for dialysis.^14,15

Metabolism

Although the precise metabolic mechanisms that lead to the development of AKA are incompletely understood, several mechanisms appear to be operative. Abnormal insulin and counterregulatory hormone levels occur,¹⁶ but the disorder is distinct from simple starvation and diabetes mellitus. Ethanol results in inhibition of gluconeogenesis and depletion of glycogen stores, leading to low glucose availability, particularly when coupled with fasting. Hypoglycemia causes release of epinephrine, cortisol, and growth hormone, as well as decreased insulin production; these are all factors that favor ketone synthesis. Ethanol metabolism results in a surfeit of acetate and NADH, which promotes lactate and ketone production. Marked ketonemia results in acidosis and ketonuria. The latter causes osmotic diuresis, intravascular volume depletion, and electrolyte losses. Thus, starvation, dehydration, excessive acetate production, an altered redox state, hormonal imbalances, and perhaps genetic predisposition are all potentially involved.¹⁷

The so-called ketone bodies that accumulate in all forms of endogenous ketoacidosis are acetone, β-hydroxybutyrate, and acetoacetate. Acetone is only a minor product produced by decarboxylation of acetoacetate, either spontaneously or catalyzed by acetoacetate decarboxylase (AAD):

Acetone is excreted in the breath and urine, where it may be detected by physical examination or urinalysis, respectively. β-Hydroxybutyrate and acetoacetate are interconvertible by the enzyme β-hydroxybutyrate dehydrogenase (βHD), and the two compounds normally exist in equilibrium:

In both AKA and diabetic ketoacidosis (DKA), β-hydroxybutyrate is quantitatively the more important molecule. However, the ratio of β-hydroxybutyrate to acetoacetate tends to be higher in AKA (typically 5 : 1 but sometimes exceeding 10 : 1),¹⁸ compared with DKA (typically 3 : 1).

Clinical Manifestations

AKA characteristically develops 24 to 72 hours after an alcoholic debauch as the blood ethanol concentration is declining, during which time the subject ceases ethanol consumption and has little or no caloric intake. Gastrointestinal symptoms predominate and include anorexia, nausea, epigastric pain, and vomiting.^19,20 The subject usually has a temporary aversion to food and alcoholic beverages and complains of malaise. On physical examination, there is a clear sensorium in most cases. The odor of acetone may be detectable on the subject’s breath. Tachypnea or Kussmaul respirations may be evident if there is marked acidemia. Tachycardia and other signs of volume depletion may be apparent. In some cases, manifestations of underlying cirrhosis (e.g., jaundice, ascites, ecchymoses, hemorrhoids) or other disorders commonly associated with chronic alcohol abuse (see Table 171-2) may be present.

Laboratory Manifestations

The key laboratory findings in AKA are metabolic acidosis, ketonemia, and ketonuria in the presence of a normal, low, or mildly elevated blood glucose concentration. Ethanol may be detectable in the blood, but it is not a requirement for the diagnosis and is frequently not detectable by the time the patient presents to the hospital. If the acidosis is clinically significant, elevation of the serum anion gap is expected. Other causes of metabolic acidosis must be excluded. Simple starvation can cause mild ketoacidosis, but with simple starvation the serum total CO₂ content or bicarbonate concentration generally remains above 18 mmol/L. DKA and renal failure are readily excluded by routine blood glucose and creatinine measurements. Lactic acidosis may be suggested by the associated clinical setting (e.g., seizures, hypotension), but it should be excluded by direct assay. Mild degrees of hyperlactatemia can occur in AKA, but concentrations greater than 3 mmol/L should prompt consideration of occult hypoperfusion, seizures, or another cause. Occult toxic ingestions also require exclusion, particularly ingestions of methanol, ethylene glycol, and salicylate intoxication.^15,21–24 Ingestion of exogenous acetone or isopropanol can cause marked ketosis due to acetonemia, but in isolation these intoxications are not associated with anion gap elevation or metabolic acidosis unless the poisoning is severe enough to cause seizures or circulatory shock, thereby resulting in lactic acidosis.

The high ratio of β-hydroxybutyrate to acetoacetate seen in AKA has clinical relevance when interpreting laboratory tests. A common assay for ketone bodies uses the semiquantitative nitroprusside reaction. Nitroprusside reacts colorimetrically with acetone and acetoacetate but not with β-hydroxybutyrate. As a result, and in comparison with DKA, the degree of ketonemia detectable in AKA is often disproportionately low relative to the degree of metabolic acidosis present. Therefore, severe metabolic acidosis due to DKA is typically associated with marked levels of ketosis, whereas severe acidemia in AKA may appear to be associated with only mild to moderate ketosis by nitroprusside-based testing. In milder cases of AKA, those associated with a mild degree of metabolic acidosis in which the acidosis is due mostly to elevation of β-hydroxybutyrate, the acetoacetate and acetone levels may not be sufficiently elevated to yield detectable ketosis by the nitroprusside test.

Because vomiting and dehydration are frequent manifestations in AKA, metabolic alkalosis can complicate the acid-base derangement. The combination of metabolic acidosis (from ketoacidosis) and metabolic alkalosis (from vomiting and volume contraction) can result in arterial pH and blood gas values that underestimate the severity of one or both of these metabolic disturbances. For example, mild metabolic alkalosis can be obscured by the presence of moderate or severe metabolic acidosis. Rarely, both metabolic processes are present and of approximately equal severity. In this situation, blood pH and bicarbonate concentration can be within normal limits despite the acid-base disturbances.²³ Or, the metabolic alkalosis can predominate and obscure the acidosis. The serum anion gap can aid in detecting these situations. An abnormally high anion gap suggests metabolic acidosis even if no acid-base disorder is evident by arterial blood gas analysis. In the face of a wide serum anion gap, the quotient of the delta anion gap (i.e., the subject’s anion gap minus the average normal anion gap) divided by the delta bicarbonate (i.e., the average normal bicarbonate concentration minus the subject’s blood bicarbonate concentration) should equal unity in organic metabolic acidoses if there is no metabolic alkalosis.²⁵ A quotient well above unity (e.g., >1.2) is evidence of concomitant metabolic alkalosis.

Treatment

Alternative explanations for the metabolic acidosis should be promptly excluded.²⁴ As in acute alcohol intoxication, the initial assessment should focus on identifying relevant alternative, underlying, or complicating illnesses or injuries that may require specific urgent therapy. Although patients with AKA sometimes have severe metabolic acidemia, the acid-base disturbance usually responds rapidly to IV hydration and ample dextrose administration.¹⁷ Rapid infusion of 50 mL of 50% dextrose is indicated if hypoglycemia is identified. Five percent dextrose in normal saline is infused IV, at a high rate initially, to correct any hypovolemia or hypoglycemia and provide substrate for metabolic correction of the ketoacidosis. Thereafter, dextrose-containing normal or half-normal saline can be substituted at a high maintenance infusion rate, titrated to ongoing fluid losses. Ample dextrose administration is key to reversing the metabolic acidosis. The blood glucose concentration should be monitored frequently to allow detection of recurrent hypoglycemia or any intolerance to the provided glucose load.

In addition to specific tests related to acid-base imbalances, the same screening laboratory studies listed for acute alcohol intoxication should be evaluated. Serial acid-base and serum electrolyte testing is performed to monitor the response of the acidosis to treatment and to monitor for specific electrolyte abnormalities. Sodium bicarbonate and insulin are rarely if ever necessary. Potassium, magnesium, or phosphorus supplementation is provided if a deficiency is found by blood testing. Thiamine and multivitamins are indicated routinely. Because vomiting is common, the patient should be given nothing by mouth initially. Gastric intubation may be indicated if there is recent or ongoing vomiting, evidence of pancreatitis, or suspicion of gastrointestinal hemorrhage. Ethanol withdrawal precautions are observed.

Clinical Manifestations

The syndrome is traditionally classified into four stages, although the stages do not always follow the indicated sequence, and not every patient develops every stage.²⁹ The time of development of each stage is also quite variable, and overlaps can occur. A typical temporal sequence is described.

The first stage occurs 6 to 24 hours or more after the last drink or after a somewhat longer period of markedly decreased ethanol intake. Manifestations include anxiety, restlessness, decreased attention, tremulousness, insomnia, and craving for alcoholic beverages. Stage 2, which occurs about 24 hours after the onset of abstinence, is characterized by hallucinations, misperceptions, irritability, and vivid dreams.³⁰ Hallucinations may be auditory, but more often they are visual or tactile. Formication, the delusional sensation of insects crawling on the skin, and vivid or threatening visual hallucinations are particularly common. During this stage, the patient may appear otherwise lucid or somewhat confused, hypervigilant, and easily startled or misled. In stage 3, which commonly occurs 7 to 48 hours after cessation of drinking, seizures occur, usually of the grand mal variety.⁴ The seizures classically manifest as a cluster of brief tonic-clonic convulsions, at one time referred to as “rum fits.” They are more likely to occur in subjects with a history of repeated withdrawal episodes.³² A relatively lucid interval ranging from hours to 2 or 3 days is sometimes seen between stages 3 and 4. Stage 4 manifests 2 to 6 days or more after initiation of abstinence and consists of a global confusional state associated with signs of neuronal excitation and severe autonomic hyperactivity. Vernacular usage notwithstanding, the term delirium tremens specifically refers to stage 4 of withdrawal. Only a small minority of individuals with alcohol withdrawal develop delirium tremens. Tremors, hallucinations, and seizures are common during this stage. As is characteristic of delirium in general, the degree of confusion and disorientation can wax and wane. Hyperadrenergic manifestations may include diaphoresis, flushing, mydriasis, tachycardia, hypertension, and low-grade fever.⁴

Laboratory Manifestations

There are no specific laboratory manifestations of ethanol withdrawal. Laboratory abnormalities are a reflection of any concomitant or underlying disorders such as cirrhosis, coagulopathy, gastrointestinal bleeding, infection, pancreatitis, or aspiration pneumonitis. Electrolyte disorders are common, particularly hypokalemia, hypomagnesemia, and hypophosphatemia. Serum creatine phosphokinase activity should be evaluated because rhabdomyolysis is a common complicating problem and if severe can lead to renal failure or compartment syndrome.

Treatment

Early-stage withdrawal with mild symptoms does not generally require treatment in an ICU setting. Full-blown delirium tremens, on the other hand, often requires more vigilant monitoring than can be provided on many general medical or surgical units. Comorbid conditions that should prompt special consideration for ICU admission include acute coronary syndromes, congestive heart failure, severe sepsis, acute gastrointestinal bleeding, pancreatitis, hepatic failure, spontaneous bacterial peritonitis, hypothermia, and hyperthermia. Other factors to consider include advanced age, renal failure, severe electrolyte deficiencies, marked rhabdomyolysis, symptomatic hypoglycemia, recurrent or prolonged seizures, cardiac dysrhythmias, hypotension, and respiratory or airway compromise.

Initial steps in management include ensuring that a patent airway is present and that ventilation, oxygenation, and perfusion are adequate; establishing IV access; and excluding serious coexisting or complicating disorders. Subsequent treatment focuses mainly on judiciously titrated sedation and vigilant monitoring for progression of the syndrome or development of complications. All patients with alcohol withdrawal are given prophylactic multivitamin supplements including parenteral thiamine and folate, and fluid deficits and electrolyte deficiencies are corrected.³³ Routine administration of magnesium sulfate in the absence of hypomagnesemia has not been shown to be beneficial.^34,35 Prophylaxis against deep vein thrombosis is recommended.

A calm, nonthreatening, protective environment with frequent verbal orientation and reassurance is provided to allay anxiety and fear and to minimize agitation. This approach may suffice in milder cases, but more advanced withdrawal necessitates pharmacologic intervention. The principle underlying this pharmacotherapy is that administration of a cross-tolerant agent to achieve light to moderate sedation will ameliorate the severe manifestations of withdrawal (including autonomic and psychomotor hyperactivity), provide subjective relief, protect the patient from self-harm, and allow specific therapeutic interventions until spontaneous recovery occurs.

The agent of choice is a benzodiazepine given orally in milder cases or IV in more severe withdrawal states.^{30,33,36–38} Limited evidence suggests that symptom-triggered dosing is superior to fixed-schedule benzodiazepine dosing.³⁹ Individualized dosing requires the expert judgment of an experienced clinician, but practicality often necessitates substitution of protocol-driven dosing schemes. These typically use a quantitative assessment scale such as the Revised Clinical Institute Withdrawal Assessment Scale for Alcohol to score the degree of withdrawal manifestations.^40,41 Lorazepam can be administered IV in incremental doses, starting with 1 or 2 mg, followed by intermittent (e.g., every 2-6 hours) IV dosing or a continuous IV infusion (e.g., initiated at 1 mg/h and titrated to effect).^29,42 Alternatively, midazolam can be employed, beginning with 2 to 4 mg by IV injection, followed by 2 mg/h by continuous IV infusion, which may be titrated to effect. Diazepam is another option, given initially in titrated doses of 5 to 10 mg at intervals as frequent as every 10 minutes if necessary until a calm but awake level of consciousness is achieved. Subsequent dosing at 5 to 20 mg every 4 to 6 hours is typically required with this agent. Prolonged administration of diazepam can lead to prolonged duration of sedation due to accumulation of the parent drug and an active metabolite, both of which have long half-lives. This effect is less likely to occur with lorazepam.

Oral benzodiazepines have been employed commonly in mild cases of withdrawal that do not require IV sedation.^30,31 These agents also can be used in more serious cases after the severe manifestations have abated and parenteral benzodiazepines are no longer required. Typical oral chlordiazepoxide dosage is 25 to 100 mg every 6 to 12 hours. Intramuscular administration is sometimes employed, but it entails a less predictable dose-response due to erratic absorption, and there is the potential for a depot effect.

Other sedative-hypnotic drugs can be effective but are not considered first-line therapeutic agents.^33,36 Barbiturates have a long history of successful use. The most commonly used agent is phenobarbital, which can be difficult to titrate because of its long duration of action. The shorter-acting barbiturate, pentobarbital, also has been employed. Oral ethanol and, in the past, paraldehyde have been used but have been discouraged, in part because of the risks of aspiration and gastric irritation, but also because their use can be interpreted as reinforcing the acceptability of using alcoholic beverages, either in general or for treatment of withdrawal symptoms. The latter criticism has also been directed at the use of ethanol administered IV for this purpose. A randomized trial examining IV ethanol administration for alcohol withdrawal prophylaxis in trauma ICU patients found no advantage compared to benzodiazepine management.⁴³ Propofol is effective, but it is not a first-line agent and is not recommended unless an endotracheal tube is in place and mechanical ventilation is used.²⁹ Regardless of the specific sedative agent employed, appropriate dose titration is crucial. The goal is to ameliorate the manifestations of withdrawal without causing excessive sedation. Sedation should be titrated with the use of an objective sedation scale such as the Ramsay Sedation Scale,⁴⁴ the Riker Sedation-Agitation Scale,⁴⁵ or the Richmond Agitation-Sedation Scale.⁴⁶ The goal should be to achieve a calm awake state or, if that is not feasible, a state of light somnolence from which the patient can easily be aroused and is able to respond verbally.

Clonidine may be administered if hyperautonomic symptoms are prominent.^47–49 Typical oral dosing is 0.1 to 0.2 mg every 6 to 12 hours. β-Adrenergic receptor blockers are not recommended for routine use, but barring contraindications, they may be considered in selected cases as adjunctive agents for controlling severe hyperadrenergic manifestations. Haloperidol and other neuroleptic agents are not routinely used, because they can lower the threshold for seizures. In selected cases, haloperidol may be used in conjunction with benzodiazepines for marked agitation or hallucinations, but this agent or similar drugs should probably not be used as monotherapy.³⁶

Seizure precautions should be instituted for all patients in withdrawal. Withdrawal seizures are managed primarily with benzodiazepines, which usually are effective at the doses used for sedation.⁴² In refractory cases, higher doses may be necessary but may necessitate endotracheal intubation and mechanical ventilation. Concomitant use of other anticonvulsants also can be considered. Barbiturates may be used for this purpose, but phenytoin is usually ineffective unless the seizures are due to a specific cause other than alcohol withdrawal, such as underlying epilepsy or a complicating acute disorder of the CNS (e.g., meningitis, head trauma).^33,50,51 In such cases, phenytoin is usually the anticonvulsant of choice. A variety of other anticonvulsant and sedative drugs have been studied for potential use in treating alcohol withdrawal, including valproic acid, baclofen, γ-hydroxybutyrate, gabapentin, oxcarbazepine, and carbamazepine. However, data on safety and efficacy are limited, particularly for hospitalized patients and those with comorbid illness.^52–59

Once severe manifestations have been controlled with parenteral sedation for a period of at least 24 hours, tapering of the dose can be attempted. If tapering of sedation is tolerated, further gradual tapering is attempted, with the goal of substituting oral for parenteral benzodiazepine administration. This process typically takes up to several days, but there is substantial variability.

Metabolism

Other than its inebriant and mucosal irritant effects, methanol per se is nontoxic. However, it is metabolized slowly to formaldehyde:

and then rapidly to formic acid, depicted here as its dissociation products, formate and a hydrogen ion⁷⁰:

Formic acid production can result in metabolic acidosis. Independent of the acidosis, formic acid inhibits cytochrome oxidase and has direct neurotoxic effects, particularly affecting the retina and optic nerves.^68,71–76 Small amounts of methanol are present as congeners in fermented alcoholic beverages.⁷⁷ Small amounts are also formed during the metabolism of certain fruits and vegetables and by metabolism of the artificial sweetener, aspartame.^78,79 However, the quantity of methanol available or formed from these sources is small, and there are enzyme systems present in the body that can convert these small amounts of formate to harmless CO₂ (Figure 171-1). The large amounts of formate produced in serious cases of methanol intoxication overwhelm these enzymes, resulting in toxic accumulation of formate. Certain nonhuman mammalian species have enzymes with much higher activity for metabolism of formate; even large quantities of methanol are nontoxic to these species. Methanol ingested by these species is still converted to formaldehyde and formate, but these toxins are rapidly metabolized to CO₂ so that significant formate accumulation does not occur. The enzymes that convert formate to CO₂ require folinic acid, the activated form of folic acid, as an obligate cofactor.⁷⁰

Figure 171-1 Metabolic pathways involved in methanol metabolism, showing the role of folate derivatives as enzymatic cofactors operative in the elimination of formic acid. ADP, adenosine diphosphate; ATP, adenosine triphosphate; NAD⁺ and NADH, oxidized and reduced forms of nicotinamide adenine dinucleotide, respectively; NADP⁺ and NADPH, oxidized and reduced forms of nicotinamide adenine dinucleotide phosphate, respectively; THF, tetrahydrofolate.

(Adapted from Kruse JA. Methanol poisoning. Intensive Care Med 1992;18:391–7, with permission.)

Clinical Manifestations

Like ethanol, methanol has dose-dependent sedating and inebriating effects that manifest shortly after ingestion, but methanol is less potent in this regard. Both alcohols also have similar gastrointestinal irritant effects that can provoke nausea, vomiting, abdominal pain, gastritis, hematemesis, and pancreatitis, although methanol may be more potent in this regard. Methanol ingestion can lead to additional CNS manifestations that are not observed with ethanol intoxication, which can sometimes provide helpful clinical clues in cases of occult methanol intoxication.^66,68 These more specific manifestations are caused by formate, the end product of methanol metabolism. There is a characteristic delay, usually 12 to 24 hours, between ingestion and development of these manifestations, and this delay is attributable to the relatively slow conversion of methanol to formaldehyde. Delayed CNS manifestations can include cerebral edema, seizures, signs of meningeal irritation, and cerebral infarction (particularly infarction of basal ganglia).⁸⁰ However, the most specific clinical findings are ocular and range from mildly blurred vision to visual field defects or tunnel vision to complete and sometimes permanent blindness.^66,68 Other possible ocular manifestations include scotomata, scintillations, papilledema, and loss of pupillary light reflexes. Most survivors recover visual function, but permanent visual deficits occur in as many as a third of patients with serious intoxication. If the metabolic acidosis is severe, it can result in Kussmaul respirations and dyspnea. In the most severe cases of poisoning, profound acidosis, respiratory failure, and circulatory shock intervene. Severe global brain injury and brain death can also occur.

If the patient offers historical information detailing an obvious toxic ingestion, the diagnosis of methanol intoxication is straightforward. Confirmatory diagnostic studies can be obtained and treatment initiated. In other cases, the diagnosis is not straightforward. Some poisoned patients may be unable to provide any history because of stupor or coma. Alert patients may be unaware that the alcoholic beverages they were provided were adulterated. Others may be alert and aware that they ingested a toxic substance but unwilling to provide the necessary history because of fear of social stigmatization or legal recrimination or as a manifestation of irrational or sociopathic behavior.

The presence of methanol may be detectable on the intoxicated patient’s breath, but the agent’s subtle odor can be difficult to appreciate and may be confused with ethanol. As a corollary, if a patient who appears inebriated has no breath odor of any type of alcohol, suspicion should be raised of methanol or a related toxic ingestion. In some cases, a faint odor reminiscent of formalin may be noticeable on the patient’s breath. The obvious presence of ethanol on the breath does not exclude the possibility of methanol ingestion; co-ingestions involving these two alcohols are frequent.

Laboratory Manifestations

The clinical laboratory can be helpful by providing clues to the diagnosis in cases of occult intoxication and by corroborating cases with a clear history of methanol ingestion. The serum total CO₂ content may be abnormally low as a consequence of metabolic acidosis due to formic acid production. The dissociation product of formic acid, formate, is negatively charged, and can widen the serum anion gap. Arterial blood gas analysis can corroborate the presence of metabolic acidosis. Metabolic acidosis associated with a wide serum anion gap has a limited number of causes, the most common of which are lactic acidosis, ketoacidosis, and renal failure.^23,24 These other causes of wide-gap metabolic acidosis are easily excluded by measuring the concentrations in blood of lactate, ketones, glucose, and creatinine. Certain toxins (e.g., propylene glycol) can result in lactic acidosis by direct metabolic conversion of the parent compound to lactate. More commonly, lactic acidosis can occur in association with any toxic exposure or drug overdose that causes seizures or circulatory shock (e.g., iron, isoniazid). The metabolic acidosis seen in methanol intoxication is mainly due to formic acid formation but can also be due in part to lactic acidosis secondary to these other mechanisms. Analogous to ethanol metabolism, conversion of methanol to formaldehyde and formic acid leads to a reducing environment in cells, which tends to increase lactate concentration. By inhibiting cytochromes, formate also may interfere with normal aerobic metabolism and lead to an increase in anaerobic glycolysis with resulting lactic acidosis. Therefore, hyperlactatemia does not exclude methanol poisoning. A few other toxic agents besides methanol, notably ethylene glycol and salicylates, can directly cause a wide anion gap metabolic acidosis. Although measurement of plasma formate concentration would seem to be a rational method to confirm the diagnosis of methanol poisoning, this assay is rarely available in hospital laboratories.⁸¹

Life-threatening methanol poisoning can result in profound metabolic acidosis which sometimes is refractory to large doses of sodium bicarbonate. However, even with severe methanol exposure, metabolic acidosis may be absent if testing is performed within a few hours after the ingestion.⁸¹ In these cases, the plasma methanol level may be very high, but the slow rate of its metabolism has not allowed for appreciable conversion to formic acid. Therefore, in the presence of a compatible history for toxic alcohol ingestion, the absence of a wide anion gap or hypobicarbonatemia should not be regarded as excluding the possibility of methanol poisoning.

A potentially useful screening test for recognition of methanol exposure early in its course is the serum osmolality gap. Serum osmolality is determined by the concentration of osmotically active solutes, or osmoles.^82,83 Osmotic activity is directly proportional to the osmole concentration of a solution, which is directly proportional to the mass concentration of the solute and inversely proportional to the solute’s molecular weight. Therefore, to have an appreciable effect on osmolality, a solute must be present at relatively high mass concentration and have a relatively low molecular weight. For example, albumin is present at relatively large mass concentrations in serum, normally averaging about 4000 mg/dL, in comparison with urea, which normally averages only about 10 mg/dL. However, albumin has a far higher molecular weight (approximately 69,000 daltons, compared with 60 daltons for urea), making its osmolar concentration less than 1 mOsm/L. The elemental ions sodium and chloride are present in appreciable mass concentration, and their atomic weight is comparatively low (23 and 35 daltons, respectively), making them quantitatively important serum osmoles. Therefore, total serum osmolality normally comprises sodium, low atomic or molecular weight anions, plus urea and glucose; although many other osmoles are present in serum, their collective contribution is comparatively small. Based on these principles, serum osmolality may be estimated by the following formula⁸³:

where the serum sodium concentration (Na) is given in mmol/L, and the serum urea nitrogen (SUN) and serum glucose concentrations are in mg/dL. The divisors, 2.8 and 18, are necessary to convert the conventional units of mg/dL to mmol/L. They are based on the molecular weights of the respective compounds.

Because ethanol is osmotically active and may be present in the blood in relatively high concentrations, the formula may be expanded to include a term for ethanol:

Here, the units for ethanol are mg/dL, and the divisor is based on the molecular weight of ethanol, 46 daltons. These millimolar concentration units technically provide an estimate of serum osmolarity; however, for practical purposes, they can be equated to millimolal units and designated osmolality (i.e., milliosmoles per kilogram of water). Just as ethanol can appreciably affect serum osmolality, so too can methanol.^84,85 The serum osmolality may be estimated from the formula shown and compared with a more direct measurement of serum osmolality; the difference between the two results affords a method for detecting and crudely quantifying the concentration of exogenous osmoles such as methanol. This is accomplished by means of the following formula:

Measured serum osmolality is determined in most clinical chemistry laboratories by analysis of the freezing point of the sample. Freezing point represents a colligative property of solutions that is depressed in proportion to osmolality, regardless of the chemical nature of the osmoles. This method, therefore, allows an empirical assessment of osmolality. The normal serum osmole gap is typically less than 10 mOsm/kg H₂O with this formula. Appreciable elevation of the osmole gap suggests the presence of an exogenous osmole (e.g., methanol). The only toxins that can appreciably affect the osmole gap are those that have a low molecular weight and can accumulate in relatively high concentration in the blood. A number of other exogenous compounds besides methanol meet these criteria, including ethylene glycol, acetone, isopropanol, propylene glycol, and acetonitrile, all of which have been reported to increase osmolality and the osmole gap.^15,23,82,83

The constellation of laboratory findings that includes metabolic acidosis along with abnormal widening of both the serum anion gap and the serum osmole gap provides presumptive or corroborative evidence of methanol (or ethylene glycol) poisoning in compatible clinical settings. However, the serum osmole gap is not foolproof, and it has important limitations. False-positive results have been described in cases of circulatory shock, DKA or AKA, the hyperglycemic nonketotic dehydration syndrome, chronic renal failure, and multiple organ system failure.⁸³ False-negative results can occur if the ingestion involved a small but still potentially lethal volume of methanol. When assessing the serum osmole gap, it is important to ensure that all relevant measurements are made from the same serum specimen to minimize variability due to temporal changes in individual analyte concentrations. Some clinical chemistry laboratories assay serum osmolality by the dew point or vapor pressure method. For technical reasons, this method yields spuriously low osmolality readings in the presence of ethanol, methanol, and other volatile alcohols, and therefore it should not be used to assess the osmole gap.⁸³

Methanol assays are available in many clinical chemistry laboratories and provide a direct assessment of methanol concentration in serum samples. This test is not definitive, because patients who present late after methanol intake may have metabolized much or all of the ingested alcohol, although the toxic byproducts may be present in appreciable concentration.⁸⁶ The delay between ingestion and presentation represents another factor that may explain the wide range of blood methanol concentrations reportedly associated with fatal outcome.⁸⁷ Methanol assay results should be interpreted in conjunction with assessments of acid-base status, serum anion gap, and serum osmole gap, as well as the history and clinical findings.

Treatment

As with any toxic ingestion, the patient’s airway and ventilation must be immediately assessed and, if necessary, adequate support provided. Circulatory shock is treated with fluid resuscitation, inotropic support, and vasopressor agents, as appropriate. Whether the patient is initially unstable or not, close monitoring of vital signs, cardiopulmonary status, and neurologic status is indicated. Vomiting should not be induced because of the risk of aspiration and the lack of demonstrable benefit. Gastric lavage is unlikely to be of value unless the patient presents within 1 hour after ingestion. Activated charcoal is also of dubious benefit unless there is a concomitant toxic ingestant.^{9,10,12,88–93} However, co-intoxication with another drug or toxin should be considered routinely. Accordingly, naloxone should be administered if the subject is unconscious. Blood and urine samples should be obtained for toxicologic screening. As in acute ethanol intoxication, complicating and occult underlying comorbid disorders must be considered (see Table 171-2).

Specimens should also be obtained for diagnostic laboratory tests. However, because specific toxicologic identification is not available on site at all hospital laboratories, antidotal therapy should not be delayed if there is an obvious history of methanol ingestion.^8,70,93,94 Even if “stat” testing is available, methanol intoxication may not be considered in occult cases until routine laboratory test results are obtained and reveal unexplained metabolic acidosis. In such cases, the preliminary laboratory test results in conjunction with a compatible setting and perhaps physical findings may allow a presumptive diagnosis to be made and antidotal treatment to be initiated. Treatment predicated on the presumptive diagnosis can be stopped if further studies convincingly argue against methanol intoxication. Symptomatic poisoned patients require ICU admission for frequent monitoring of vital signs and level of consciousness and to provide specific antidotal treatment, which consists of ethanol or fomepizole administration, hemodialysis, and folate administration.

Ethanol has been the conventional form of antidotal pharmacotherapy for methanol intoxication. The principle is that the enzymes, alcohol dehydrogenase and aldehyde dehydrogenase, have higher affinity for ethanol than for methanol, and ethanol thereby serves as an effective competitive inhibitor.^95–98 As a result, conversion of methanol to formaldehyde and formate is significantly slowed in the presence of ethanol, allowing methanol to be excreted by the kidneys and lungs, and by hemodialysis if that modality is employed. If inhibition is incomplete, the body may be able to safely eliminate the much smaller amounts of formaldehyde and formate that are metabolically produced from the methanol. Indications for ethanol therapy include a serum methanol concentration greater than 20 mg/dL or a history or strong clinical suspicion of methanol ingestion in conjunction with either an elevated osmole gap or evidence of metabolic acidosis (e.g., arterial blood pH < 7.30 and bicarbonate < 20 mmol/L).

Ethanol can be given orally, by gastric instillation, or by vein. Oral dosing can be considered in mild cases if the patient is completely alert and is accustomed to drinking liquor. The solution is usually prepared from commercially available liquor, available in many hospital formularies, and diluted to a final concentration of 20% ethanol. Even with dilution, subjects who are uninitiated to drinking this quantity of ethanol over a short interval are unlikely to avoid vomiting. Vomiting increases the risk of aspiration, particularly when coupled with the sedating effects of the administered ethanol and the potential CNS effects of the ingested methanol. For these reasons, IV ethanol administration is usually preferred over the oral route. Intravenous administration of ethanol can be accomplished with the use of a sterile solution of either 5% or 10% (volume/volume) ethanol in 5% (weight/volume) dextrose. These solutions are markedly hyperosmolar (approximately 2000 mOsm/kg H₂O for 10% ethanol in 5% dextrose and water), and therefore must be administered through a central venous catheter.

A loading dose is given so as to rapidly effect maximal enzyme inhibition. The goal is to achieve a serum ethanol level of 100 to 150 mg/dL. Based on the volume of distribution of ethanol (0.6-0.7 L/kg in men, slightly less in women and elderly subjects, and less in obese subjects) and a target serum ethanol concentration of 100 mg/dL, the necessary loading dose is theoretically 600 mg/kg in terms of absolute ethanol. Given the specific gravity of absolute ethanol (0.79), this is equivalent to a dose of 0.76 mL/kg in terms of absolute ethanol. Absolute (i.e., 100%) ethanol is unlikely to be available in a hospital formulary. Oral loading can be accomplished using 100 proof liquor, which is 50% ethanol by volume (equivalent to 40 g/dL), at a dose of 1.5 mL/kg. Alternatively, IV loading using a 5% (volume/volume) solution of ethanol in dextrose and water (i.e., an ethanol concentration of 4 g/dL by weight/volume) would require 15 mL/kg, typically administered over 1 hour. The dosing calculations described frequently underestimate the ethanol dose necessary to achieve the target level. Loading doses of 700 mg/kg given IV, or even higher doses if given orally, are more likely to achieve the goal initially.¹ If the patient’s current ethanol concentration is already at or above the targeted level due to co-ingestion of ethanol, no ethanol loading dose is required. A proportionately lower loading dose is used in patients with a preexisting subtherapeutic blood ethanol concentration.

Maintenance dosing is required to maintain the targeted blood ethanol concentration. For patients with little or no history of ethanol exposure, the average required maintenance dose has been estimated to be about 70 mg/kg/h, in terms of absolute ethanol. For oral dosing with 100 proof liquor, this is equivalent to 0.18 mL/kg/h. Hourly oral maintenance doses are necessary and should be diluted to 20% to minimize epigastric pain and emesis. Using IV maintenance dosing, a continuous infusion of 5% ethanol solution is administered at 1.8 mL/kg/h. Subjects who consume ethanol chronically on a regular basis metabolize ethanol at considerably higher rates and therefore require higher maintenance doses. The necessary maintenance dose in such cases depends on the individual’s exposure history, but it can be 2 to 3 times higher than the cited dose, or even higher in some cases. If hemodialysis is used during ethanol therapy, it will effectively remove ethanol from the body. Therefore, the maintenance ethanol dose has to be increased, often doubled or tripled, during dialysis. Accurate prediction of the required ethanol dosing in individual cases is not possible, and maintenance of the desired therapeutic ethanol concentration can be challenging. Serial serum ethanol levels are obtained every 1 to 2 hours to allow the ethanol dosing to be titrated, striving to maintain a serum ethanol level between 100 and 150 mg/dL. Lower serum ethanol concentrations risk incomplete inhibition and toxicity from the products of methanol metabolism. The risks with higher levels are sedation, inebriation, and impairment of protective airway reflexes. Treatment is continued until serum methanol levels are less than 20 mg/dL. Intravenous ethanol therapy can entail large fluid volumes to achieve and maintain the desired blood ethanol level; for this reason, attention to fluid balance is important.

Fomepizole (4-methylpyrazole) is a newer therapeutic alternative to ethanol.^99–102 The indications for fomepizole use are the same as for ethanol therapy. Like ethanol, fomepizole inhibits alcohol dehydrogenase, but it is considerably more costly than ethanol. Nevertheless, fomepizole has supplanted ethanol at many centers, owing to its advantage of being easier to dose and titrate and because it has no sedative effects. Frequent serial blood ethanol assays are avoided. Compared with oral dosing of ethanol, there is no risk of nausea, vomiting, gastritis, or abdominal pain with fomepizole. Compared with IV ethanol administration, there is less risk of overhydration.

Fomepizole is given IV as a loading dose of 15 mg/kg, followed by 10 mg/kg every 12 hours for 4 doses and then 15 mg/kg every 12 hours until the serum methanol concentration is less than 20 mg/dL. Each dose is infused over 30 minutes. Dosing is altered if hemodialysis is employed. If dialysis is initiated, the next slated dose is given immediately if 6 hours or longer has elapsed since the last dose, but no dose is given if it has been less than 6 hours since the last dose. During ongoing hemodialysis, fomepizole is dosed at intervals of 4 hours. At termination of hemodialysis, if less than 1 hour has elapsed since the last dose, no fomepizole is administered; if 1 to 3 hours has elapsed between the last dose and the end of dialysis, half of the next scheduled fomepizole dose is administered; and if more than 3 hours has elapsed, the next scheduled fomepizole dose is given at the end of hemodialysis. Subsequent fomepizole dosing after hemodialysis is every 12 hours.

Severe methanol poisoning can be associated with profound metabolic acidosis in some cases. Traditionally, sodium bicarbonate was a staple part of the treatment for most causes of metabolic acidosis, but lack of demonstrable efficacy has tempered its routine use, particularly in the treatment of lactic acidosis and DKA. There are laboratory animal data and anecdotal clinical reports ascribing benefit to bicarbonate administration in cases of alcohol or glycol poisoning. Specifically, administration of bicarbonate is claimed to be capable of reversing ocular manifestations and lowering mortality, but controlled clinical trials are lacking. There also is evidence that undissociated formic acid is more toxic than the dissociation product, formate; increasing the extracellular fluid pH favors conversion of formic acid to formate.¹⁰³ Given the potential severity of the acidosis and the likely benefit of alkali therapy, sodium bicarbonate is recommended for subjects with an arterial pH less than 7.30, although intentional alkalemia is not advocated.

Ethanol and fomepizole minimize conversion of methanol to its toxic metabolites, but these forms of pharmacotherapy do not hasten elimination of methanol from the body. Methanol is excreted by the kidneys and lungs, but only slowly. Hemodialysis can effectively and more rapidly remove methanol and its toxic metabolites from the body. Charcoal or resin hemoperfusion techniques are not effective, and peritoneal dialysis is recommended only if hemodialysis is not available. Hemodialysis is recommended as a supplement to ethanol or fomepizole in patients with serious degrees of methanol intoxication. Serious intoxication is defined by the presence of metabolic acidosis, a serum methanol level above 50 mg/dL, any type of subjective or objective ocular findings, or other findings that indicate severe poisoning. Hemodialysis also is recommended if there is renal impairment. As previously noted, fomepizole and ethanol dosing must be altered during hemodialysis. Methods have been described to incorporate ethanol into the dialysate to facilitate maintaining therapeutic ethanol levels during hemodialysis.¹⁰⁴ The endpoint for dialysis is a serum methanol level less than 20 mg/dL and normalization of the anion gap, indicating clearance of formate. Direct measurement of plasma formate would be a logical method of monitoring if rapid assays were available.

In humans and certain nonhuman primates, formate is only slowly metabolized, allowing the development of acidosis and ocular pathology if substantial amounts of methanol are ingested. Monkeys given large doses of folinic or folic acid before or after methanol administration had lower formate levels and less toxicity than control animals.¹⁰⁵ Based on these and other experimental data, large doses of folic or folinic acid are recommended in clinical methanol poisoning. Typical recommendations are to administer 50 mg of folinic or folic acid IV every 4 to 6 hours. Folic acid must be reduced to tetrahydrofolate before it can serve as a cofactor for metabolizing formate. Folinic acid does not require reduction and therefore is the preferred form of the vitamin when available.

Metabolism

The metabolism of ethylene glycol is more complicated than that of methanol.^106,107 As with methanol, the parent compound is only minimally toxic, but its metabolites are very toxic. Also, in common with methanol, the initial step in metabolism is catalyzed by alcohol dehydrogenase (Figure 171-2). The action of alcohol dehydrogenases converts ethylene glycol to glycoaldehyde, which can be converted further to glyoxal. Both glycoaldehyde and glyoxal are metabolized first to glycolic acid, then more slowly to glyoxylic acid, and finally to oxalic acid. Glycoaldehyde and glyoxylate have demonstrable nephrotoxicity in isolated rodent renal tubular segments, whereas glycolate, oxalate, and ethylene glycol do not.¹⁰⁸ Glycolate and probably some of the other metabolites are also neurotoxic. Oxalic acid can precipitate as calcium oxalate crystals within various tissues, including notably the renal parenchyma and tubules.

Figure 171-2 Metabolic pathways involved in ethylene glycol metabolism, with schematic morphologies of representative urinary crystals. LDH, lactate dehydrogenase; THF, tetrahydrofolate.

(Adapted from Kruse JA. Ethylene glycol intoxication. J Intensive Care Med 1992;7:234–43, with permission.)

Clinical Manifestations

The initial effects involve the CNS and typically manifest within 30 minutes to 12 hours after ingestion.¹⁵ The CNS manifestations of ethylene glycol poisoning can range from effects that are similar to those seen with acute ethanol intoxication, such as excitement, confusion, disorientation, and ataxia, to signs of CNS depression, such as lethargy, stupor, or coma. Nausea, vomiting, myoclonus, and seizures also can occur. Cranial nerve deficits including nystagmus, ophthalmoplegia, facial palsy, dysarthria, and dysphagia have been reported. There are also rare case reports of pupillary abnormalities and changes in visual acuity, but these findings are not characteristic; if they do occur, they may be the result of co-ingestion of methanol. Classically, the second phase manifests 12 to 24 hours after ingestion and consists of cardiorespiratory effects which may include dyspnea and a Kussmaul respiratory pattern secondary to metabolic acidosis or pulmonary edema. The latter can result in frank respiratory failure necessitating endotracheal intubation and mechanical ventilation. Tachycardia, hypotension, frank circulatory shock, coma, and death also can occur during this phase. The third phase, which usually takes 1 to 3 days to manifest, consists of renal failure, either oliguric or nonoliguric, due to acute tubular necrosis. Flank pain also can occur. The time course of each phase of intoxication is variable, and overlap is frequent.

Laboratory Manifestations

Laboratory findings are similar to those seen in methanol poisoning. Detection of ethylene glycol in serum provides definitive evidence of the diagnosis. However, if the patient presents late and significant metabolism of the toxic agent has occurred, the measured concentration may not represent the patient’s peak ethylene glycol level. After ingestion of a substantial quantity of ethylene glycol, metabolic acidosis due to metabolic breakdown of the parent compound occurs during the first phase of intoxication.⁸⁶ The acidosis may be severe and is principally caused by glycolic acid accumulation.^109–112 Dissociation of this acid results in the accumulation of glycolate, which leads to an increase in the serum anion gap.

Measurement of the plasma glycolate concentration is a rational method of assessment, but clinical availability of the assay is lacking.⁸¹ The blood lactate concentration may be elevated because of the reducing intracellular milieu induced by ethylene glycol metabolism or as a manifestation of complicating seizures or circulatory shock. Lactate levels also may be artifactually elevated to a substantial degree because of the cross-reactivity of glycolate with lactate in certain automated lactate analyzers.¹¹³ The serum osmole gap may be elevated due to high blood levels of ethylene glycol and its metabolites. Because the molecular weight of ethylene glycol (62 daltons) is higher that of methanol (32 daltons), the osmole gap is less affected by a given amount (by weight) of ethylene glycol ingested or by a given blood level (by weight/volume) compared with methanol.⁸⁴ Therefore, the osmole gap is more likely to yield a false-negative result after ingestion of ethylene glycol, compared with a similar mass amount of methanol.

There are two notable laboratory findings that may be seen in ethylene glycol poisoning; these are findings not observed in methanol poisoning. The first is calcium oxalate crystalluria.¹¹⁴ Oxalate produced by ethylene glycol metabolism chelates calcium, forming crystals and potentially producing hypocalcemia in the process (see Figure 171-2). Two crystalline forms of this organic salt can occur. One is calcium oxalate dihydrate, also known as weddellite. These crystals have a characteristic octahedral shape, making them relatively easy to distinguish from various nonoxalate forms of crystalluria. The second form is calcium oxalate monohydrate, also known as whewellite. These crystals can be polymorphic; they can appear as monoclinic prisms or assume a needle-like, dumbbell-shaped, ovoid, or hempseed-like appearance. Hippurate crystalluria also has been described, but it can be difficult to morphologically discriminate hippurate from some forms of whewellite by light microscopy.¹⁰⁶ The finding of oxalate crystalluria corroborates the diagnosis; however, these crystals occasionally can be seen in the urine in the absence of ethylene glycol exposure, so their presence is not proof of glycol poisoning. On the other hand, because crystalluria does not uniformly occur after ethylene glycol ingestion, its absence does not exclude the diagnosis.

The other potential finding is fluorescence of the urine on exposure to ultraviolet radiation.^115,116 This finding is present when the ingested formulation of ethylene glycol contains fluorescein, a fluorescent dye added to many automotive antifreeze solutions to facilitate identification of cooling system leaks and to mitigate accidental confusion with potable liquids. The fluorescein is excreted in the urine and fluoresces yellow-green on exposure to ultraviolet light, such as from a Wood’s lamp (commonly used in emergency departments and ophthalmology clinics to detect corneal lesions after topical application of fluorescein to the eye). False-positive results have been described due to other fluorescent substances in urine (e.g., carotene, carbamazepine, niacin, benzodiazepine metabolites) and from certain types of glass or plastic specimen containers that have a high degree of native fluorescence.^115,117 False-negative results may occur if more than 4 hours has elapsed since the ingestion—that is, sufficient time for the fluorescein to be excreted, at least by some individuals. A false-negative result is obviously expected if the ingested ethylene glycol formulation did not contain fluorescein or involved a small volume. False-negative results can occur if the urine pH is less than 4.5, but this may be circumvented by urine pH testing followed by upward titration of the specimen’s pH if necessary. Owing to interfering factors and the limited ability of untrained examiners to detect fluorescence, clinical decision making should not hinge on this test in isolation.¹¹⁸

Treatment

With a few exceptions, the treatment of ethylene glycol poisoning is the same as for methanol intoxication. Gastric lavage may have some efficacy, but only if it is performed within 1 hour after the ingestion. Activated charcoal is not effective unless there is an amenable concomitant toxic ingestion.^12,119 Ethanol^98,120,121 or fomepizole^99–102 is administered to slow the conversion of the glycol to toxic intermediates; sodium bicarbonate is given if there is significant metabolic acidosis (e.g., arterial pH < 7.30); and hemodialysis is used in cases of serious intoxication to speed elimination of the parent compound and toxic metabolites. Ethanol or fomepizole is recommended if the serum ethylene glycol concentration is above 20 mg/dL. However, ethylene glycol assays are not available at all institutions, and inhibitor treatment should be initiated while awaiting definitive identification of the glycol if there is presumptive evidence of intoxication.^8,94,106 This evidence can include a clear history of recent ethylene glycol ingestion or strong clinical suspicion of ingestion in conjunction with either an elevated osmole gap, evidence of metabolic acidosis (e.g., arterial blood pH < 7.30 and bicarbonate < 20 mmol/L), or oxalate crystals in the urine. Dosing of ethanol and fomepizole is the same as for methanol intoxication. Fomepizole can be recommended over ethanol if the sensorium is depressed. Inhibitor treatment is continued until the serum ethylene glycol level falls below 20 mg/dL.

Hemodialysis can be even more important for the treatment of cases of ethylene glycol poisoning than it is for methanol intoxication, because ethylene glycol ingestion can result in severe renal dysfunction, thereby interfering with excretion of the compound and its toxic metabolites. Dialysis is conventionally recommended for all patients with serum ethylene glycol levels over 50 mg/dL. Hemodialysis is indicated for all patients with renal dysfunction and for patients with metabolic acidosis or other toxic manifestations. The conventional endpoint for dialysis is a serum ethylene glycol concentration less than 20 mg/dL in conjunction with normalization of the anion gap, indicating clearance of toxic metabolites.

Although there is some evidence that formic acid may be produced as a minor product of ethylene glycol metabolism, it probably does not play a significant role in the pathophysiology of this form of poisoning. Therefore, folate administration has not been routinely recommended. However, there is more convincing evidence that glyoxylate may be metabolized to nontoxic products by enzyme systems that rely on other vitamin cofactors, specifically pyridoxine (vitamin B₆) and thiamine (see Figure 171-2). Providing supplements of pyridoxine (e.g., 50 mg IV every 6 hours) and thiamine (e.g., 100 mg IV every 6 hours) could hasten elimination of toxic intermediates, although evidence of efficacy is quite limited.¹⁰⁶ Given the low toxicity of these vitamins, both are recommended. Magnesium is a necessary cofactor for the enzymatic degradation of glyoxylate, and supplemental magnesium should be given if there is hypomagnesemia.

Routine IV administration of calcium salts was advocated at one time as a therapeutic means of lowering oxalate levels in body fluids in cases of ethylene glycol poisoning. However, precipitation of calcium oxalate in vital organs is probably more likely to have harmful effects. Therefore, routine therapeutic administration of calcium to correct hypocalcemia is no longer advised unless the hypocalcemia is severe enough to cause manifestations.

Key Points