Brenner and Rector's The Kidney, 8th ed.

CHAPTER 62. Extracorporeal Treatment of Poisoning

James P. Smith   Ingrid J. Chang



Principles and Techniques for Enhanced Renal Elimination of Toxins, 2081



Forced Diuresis and Manipulation of Urinary pH, 2081



Principles Governing Drug Removal by Extracorporeal Techniques, 2082



Drug-Related Factors, 2083



Dialysis-Related Factors, 2083



Measurements of Drug Removal Efficiency: Dialysate Clearance, Extraction Ratio, and Sieving Coefficient, 2084



Extracorporeal Techniques for Drug Removal, 2084



Indications for the Use of Extracorporeal Detoxification, 2084



Intermittent Hemodialysis, 2084



Hemoperfusion, 2085



Hemodialysis-Hemoperfusion, 2085



Hemofiltration, 2085



Continuous Renal Replacement Therapy, 2086



Sustained Low-Efficiency Dialysis, 2086



Peritoneal Dialysis, 2086



Plasma Exchange and Plasmapheresis, 2086



Intoxications Responsive to Extracorporeal Therapy, 2086



The Alcohols—Ethylene Glycol, Methanol, Isopropanol, 2086



Lithium, 2092



Salicylates, 2094



Theophylline, 2096



Valproic Acid, 2097

Poisoning may result from an unintentional toxic exposure, an intentional misuse of a substance with or without self-destructive intent, therapeutic error, environmental or occupational exposure, or malicious activity. According to the Toxic Exposure Surveillance System data compiled by the American Association of Poison Control Centers, over 2.4 million human exposure cases were reported by 62 participating centers in 2004, an average of one poison exposure every 13 seconds.[1] Although approximately 84% of reported exposures were unintentional, suicidal intention was suspected in 8% of reported toxin exposures. Therapeutic errors, including drug-drug interactions, were responsible for 222,644 reported cases of intoxication in 2004.[1]

Although the majority of reported exposures occurred in the pediatric population, poisoned adults comprised the majority of the fatalities. Of the 1183 deaths caused by toxic ingestion reported in 2004, 88% of victims were older than 19 years. The most common classes of drugs involved as the primary toxic substance in fatal cases were analgesics, antidepressants, stimulants and recreational drugs, sedative/hypnotics/antipsychotics, and cardiovascular agents.[1]

Although the minority of reported exposures require treatment in a health care facility (22.4% in 2004), nearly 15% of those that seek treatment require critical care.[1] All toxic exposures should initially be considered life-threatening. The general approach to the intoxicated patient necessitates prompt resuscitation and stabilization, clinical and laboratory evaluation, and gastrointestinal decontamination (if appropriate), and is extensively reviewed elsewhere. [2] [3] In certain cases, toxin elimination can be enhanced using various decontamination techniques, urinary alkalinization, and/or extracorporeal therapy. The frequency and types of interventions utilized to enhance toxin elimination in reported human exposure cases in 2004 are outlined in Table 62-1 . Due to the multitude of confounding variables inherent to toxic exposures, evidence-based treatment recommendations derived from prospective, randomized, controlled trials are rare. Therefore, in addition to reviewing recommendations based on animal studies and human case reports, the treating physician must have a working knowledge of clinical pharmacokinetics and the relationship between a toxin's characteristics and the ability to enhance its elimination with extracorporeal therapy. The purpose of this chapter is to review fundamental concepts of the extracorporeal treatment of acute poisonings.

TABLE 62-1   -- Therapy Provided in Human Exposure Cases in 2004


Number of Cases




 Activated charcoal (single-dose)




 Gastric lavage


 Ipecac syrup


 Other emetic


 Whole bowel irrigation


Measures to enhance elimination

 Activated charcoal (multiple-dose)






 Other extracorporeal procedure


Specific antidotes


Other interventions







Adapted from Watson WA, Litovitz TL, Rodgers GC Jr, et al: 2004 Annual Report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med 23(5):589–666, 2005.





Forced Diuresis and Manipulation of Urinary pH

The physiologic mechanisms involved in the elimination of drugs by the kidney include (1) glomerular filtration of the drug, (2) active proximal tubule secretion of the drug into the urine, and (3) passive reabsorption of the drug in the distal tubule. Plasma containing the toxin is initially filtered by the glomerulus, and glomerular clearance of the drug is primarily dependent on its degree of protein binding because only unbound drug is readily filtered by the glomerulus. As plasma perfuses the tubules, drug excretion may also occur through active proximal tubule secretion mediated by organic anion transporters located in the straight segment of the proximal tubule. The final physiological process that affects drug elimination by the kidney is passive distal tubular reabsorption, an energy independent process that occurs more commonly with the elimination of lipid soluble, non-ionized drugs.

Historically, forced diuresis through volume expansion using isotonic fluids (0.9% NaCl or lactated Ringer solution) with or without concomitant diuretics has been used to enhance renal elimination of toxins. Early studies suggested that despite maximal proximal tubular secretion of a solute, urine volumes of 200 to 300 mL/hr inhibited solute tubular reabsorption because the dilute urine prevented a favorable concentration gradient for passive reabsorption in the distal tubule.[4] It was therefore hypothesized that achieving a urine flow rate of 3 to 6 mL/kg/hr would limit tubular reabsorption of the drug and minimize toxin accumulation. When passive tubular reabsorption of a toxin was pH sensitive, though, increased urine volume through forced diuresis did not significantly enhance drug elimination when combined with urinary alkalinization.[3] In addition, common complications associated with forced diuresis include volume overload, pulmonary edema, cerebral edema, electrolyte disorders (hyponatremia, hypokalemia), and metabolic alkalosis. With the lack of clinical evidence supporting the efficacy of forced diuresis in the management of overdoses and the significant risks associated with this technique, forced diuresis is no longer recommended in the management of any acute poisonings.

Manipulation of urinary pH has also been used to enhance renal elimination of a toxin in the management of acute poisonings. The underlying mechanism is based on the concept of ion trapping. Because cell membranes are generally more permeable to lipid soluble non-polar, non-ionized forms, the goal of manipulating urinary pH is to favor the formation of the ionized form in the tubular lumen. By shifting the drug to its ionized form it becomes lipid insoluble, and passive reabsorption of the specific drug is inhibited. The ionized drug becomes “trapped” in the tubular lumen and is excreted in the urine. The dissociation of a weak acid or base into its ionized state is determined by its dissociation constant (pKa). The pKa represents the pH at which a drug is 50% ionized and 50% non-ionized. For example, the pKa of salicylic acid is 3.0. When the urinary pH is 3.0, salicylate exists in a 1:1 ratio of ionized to non-ionized form. Alkalinization of the urine to a pH of 7.4 increases the ratio to 20,000:1 with predominance of the ionized form. The shift to the ionized form results in decreased passive reabsorption and increased renal elimination of the drug.

The clinical efficacy of urine alkalinization is dependent on the relative contribution of renal clearance to the total body clearance of active drug. Theoretically, each unit of change in urine pH results in a 10-fold change in renal clearance; however, if only 1% of the ingested drug is excreted unchanged in the urine, even a 20-fold increase in renal elimination will have no clinically significant effect on total drug clearance from the body.[5] Criteria that determine whether a drug is amenable to urinary alkalinization are (1) it is eliminated unchanged by the kidney, (2) it is distributed primarily in the extracellular fluid compartment, (3) it possesses low protein binding, and (4) it is weakly acidic with a dissociation constant (pKa) of 3.0 to 7.5. Urine alkalinization enhances renal excretion of weak acids such as salicylates, phenobarbital, chlorpropamide, 2,4-dichlorophenoxyacetic acid and other chlorophenoxy herbicides, mecocrop, diflunisal, fluoride, and methotrexate.[5]

Urine alkalinization can be achieved with administration of sodium bicarbonate (1 to 2 mEq/kg every 3 to 4 hours) using an initial 1 ampule bolus of sodium bicarbonate followed by intravenous D5W with 2 to 3 ampules of sodium bicarbonate (50 mEq NaHCO3 per ampule) added to each liter at 250 mL/hr although the rate of infusion should be initially based on the volume status of the patient. Following volume expansion, the rate of fluid replacement should match the urine output with a goal of 2 to 3 mL/kg/hr. During urine alkalinization serum electrolytes and urinary pH must be closely monitored every 2 to 3 hours with the target urine pH range between pH 7.5 to 8.5. Complications include hypokalemia, hypernatremia, fluid overload, pulmonary edema, cerebral edema, and alkalemia. Relative contraindications are congestive heart failure, renal failure, pulmonary edema, and cerebral edema. The degree of hypokalemia may be profound because not only does alkaline diuresis promote increased urinary potassium excretion but alkalemia also results in an intracellular potassium shift; however, it is essential to correct hypokalemia in order for urine alkalinization to be effective. In the setting of hypokalemia, potassium reabsorption at the distal tubule occurs in exchange for hydrogen ion, thereby inhibiting alkaline urine excretion. Therefore if hypokalemia remains uncorrected during urine alkalinization, not only will the nephron be unable to produce an alkaline urine, but the patient will also be at higher risk for the development of alkalemia.[3] Carbonic anhydrase inhibitors (e.g., acetazolamide) have also been used to alkalinize urine in the past but are no longer recommended because they may exacerbate systemic metabolic acidosis.

Previously, urinary acidification (goal urinary pH 5.5 to 6.5) was administered to enhance renal elimination of weak bases such as amphetamines, fenfluramine, phencyclidine, and quinine. In order to achieve an acid diuresis, arginine (10 g) or lysine hydrochloride (10 g) was administered in D5NS over 30 minutes followed by oral ammonium chloride (4 g) every 2 hours. However, the lack of evidence supporting its clinical efficacy and its associated serious sequelae (i.e., rhabdomyolysis, acute renal failure, and exacerbation of systemic metabolic acidosis) has limited its utility. Urine acidification no longer has any role in the management of acute poisonings.


The elimination of a drug or toxin by extracorporeal therapy depends on both the properties of the drug and the extracorporeal technique chosen ( Table 62-2 ). Drug pharmacokinetic and pharmacodynamic characteristics that affect toxin removal by extracorporeal techniques include molecular weight, degree of protein binding, volume of distribution, and degree of redistribution (“rebound”) after therapy. Therapy-specific factors to consider include the extracorporeal modality used (e.g., hemodialysis, hemofiltration, hemoperfusion), dialysis membrane characteristics (dialyzer surface area, membrane pore structure), hemoperfusion cartridge characteristics, blood flow rate, and dialysate flow rate.[6]

TABLE 62-2   -- Factors That Enhance Drug Removal by Extracorporeal Therapy

Drug-Related Factors

Therapy-Related Factors

Low molecular weight (<500 D)

Large surface area of dialysis membrane

Low protein binding (<80%)

High-flux dialyzer

Low volume of distribution (<1 L/kg)

High blood and dialysate flow

Water soluble

Increased ultrafiltration rate

Fast redistribution from peripheral compartments into the blood

Increased time on dialysis




Drug-Related Factors

Molecular Weight

Toxin removal during hemodialysis occurs primarily by diffusion into the toxin-free dialysate. As molecular weight increases, the rate of diffusion across a given membrane progressively decreases. Low molecular weight solutes (<500 daltons) diffuse easily through the pores of even low-flux dialysis membranes such that their clearance is primarily dependent on membrane surface area, blood-to-dialysate concentration gradient, and blood and dialysate flow rates. Larger solutes (>1000 daltons in the case of low-flux membranes) have limited membrane permeability and are removed through convective clearance (“solute drag” from ultrafiltration) rather than diffusion; therefore, dialytic time and membrane surface area determine their rate of removal.

Protein Binding

The quantity of toxin that is bound to plasma or tissue proteins is another determinant of drug removal by hemodialysis and hemofiltration. Drugs that are highly bound have a low proportion of drug available for removal by hemodialysis or hemofiltration. The diffusive removal of a drug by hemodialysis requires unbound drug to cross the dialysis membrane because the larger size of the drug-protein complex limits its dialysis membrane permeability. Convective removal by hemofiltration also does not occur to any significant degree because the drug is bound to non-ultrafilterable plasma proteins. Hemoperfusion, however, may be more effective in these cases as the adsorbent (e.g., activated carbon) competes with plasma proteins for toxin binding and is dependent purely on the adsorbent's affinity for the toxin.

The degree of protein binding can be influenced by many factors, including alterations in plasma protein concentration and the presence of different pathologic states.[7] For example, hypoalbuminemia results in less protein available for binding, and uremic organic acid accumulation leads to a reduction in binding sites for acidic drugs (e.g., salicylates, warfarin, phenytoin). These alterations increase the fraction of unbound, biologically active drug relative to the total drug concentration. As a result, it is possible that therapeutic efficacy (or toxicity) may occur at a lower total drug concentration than observed in the normal state.[7] However, because hepatic enzymes metabolize unbound drug, it is possible that these alterations in protein binding may actually enhance drug clearance.[6] Drug levels become invaluable when dealing with such complex pharmacokinetic situations.

Highly protein-bound drugs include arsenic, calcium channel blockers, diazepam, phenytoin, salicylate and other nonsteroidal anti-inflammatory drugs (NSAIDs), thyroxine, and tricyclic antidepressants. Compounds with minimal protein binding include the alcohols (methanol, ethylene glycol, isopropanol, ethanol), aminoglycosides, and lithium.[8] Alterations in protein binding are especially clinically relevant for drugs that exhibit a narrow therapeutic index (e.g., lithium, digoxin).

Volume of Distribution

A drug's volume of distribution (Vd) is the apparent volume into which the drug is distributed at equilibrium and before metabolic clearance begins. Vd may be calculated by dividing the total drug in the body by its serum concentration:

This mathematical relationship assumes that the body is a single compartment of homogeneous water into which the drug is distributed. Therefore, drugs that are extensively bound to tissue (e.g., digoxin) have a large Vd; conversely, drugs that are primarily restricted to the intravascular space (e.g., phenytoin) have a small Vd.

Because highly bound drugs and toxins are not available to the dialysis membrane, dialytic removal will be insignificant for drugs with a large Vd. Even if the extraction ratio of the drug across the dialyzer is 1.0 (see later), dialytic removal will be clinically insignificant if the amount of drug in the blood is small relative to the total amount in the body. For example, digoxin is readily dialyzable given its low protein binding (20%–25%) and relatively low molecular weight (780 daltons). However, because of its large Vd (4–7 L/kg), both hemodialysis and hemofiltration make an inconsequential impact on the total body burden. Therefore, these extracorporeal therapies are not indicated for digoxin overdose. Other drugs with a large Vd (>2 L/kg) include calcium channel blockers, β-blockers, chloroquine, colchicine, quinidine, phenothiazines, quinine, strychnine, metoclopramide, and tricyclic antidepressants.

Conversely, drugs with a low Vd (Vd < 0.5–1.0 L/kg) are primarily restricted to the intravascular or water space and are more readily eliminated by extracorporeal therapies. These drugs include alcohols, salicylates, lithium, theophylline, atenolol, paracetamol, and aminoglycosides.


Redistribution, or “rebound,” refers to the movement of a drug to the plasma space from tissue/protein binding sites or from the intracellular space. If this redistribution occurs slowly, total body clearance of the substance is limited, especially if the Vd is large. The same mechanism explains why a toxin's plasma concentration will increase following discontinuation of extracorporeal therapy—the rate of plasma drug removal exceeds the rate of redistribution during therapy. The time course and magnitude of the rebound varies dramatically by drug, membrane characteristics, Vd, and degree of recirculation. This phenomenon is observed in the elimination of drugs such as lithium,[9]methotrexate,[10] and digoxin.[11] It is important to recognize clinical scenarios in which redistribution is expected to occur because repeated dialysis or continuous therapy may be required. For these toxins, a plasma concentration obtained immediately after extracorporeal therapy should not be considered a nadir, and frequent levels should be obtained in the redistribution period.

Dialysis-Related Factors

Factors that affect drug elimination during hemodialysis include characteristics of the dialysis membrane (material, surface area, porosity), concentration gradient across the membrane, blood and dialysate flow rates, and to a lesser extent ultrafiltration rate. Drug removal by hemofiltration, however, depends on convective transport for solute removal. Therefore, the rate of ultrafiltration, which is dependent on the blood flow rate, limits drug elimination with this modality. Factors that affect drug elimination by hemoperfusion include the cartridge material (type of adsorbent), size, and saturation with toxin (i.e., duration of use).[12] These characteristics are discussed later in this chapter as the different extracorporeal modalities are further described.


Dialysate clearance (C) is defined as:

where Qb is the blood flow rate, A is the pre-filter drug concentration, and V is the post-filter drug concentration. The term (A - V)/A is also referred to as the extraction ratio for a particular substance and dialyzer. This term is equal to the fraction of the drug removed from the plasma by the dialyzer. An extraction ratio of 1.0 would imply complete elimination of a substance from the blood after a single pass through the extracorporeal circuit (i.e., V = 0). Extraction ratios can also be used to describe drug removal by hemoperfusion, but the extraction ratio will decrease over time as the adsorbent becomes saturated.

Solute removal during hemofiltration occurs by convection. In this case, the efficiency of drug removal is related to the sieving coefficient (SC), which describes the ability of a drug to cross a membrane by solvent drag. Mathematically, the sieving coefficient is the ratio of the drug concentration in the ultrafiltrate to its concentration in the plasma and is defined as:

where UF is the drug concentration in the ultrafiltrate. Because the difference between A and V is not usually clinically significant, the equation can be simplified to SC = UF/A.[13] If SC = 1, the concentration of drug in the ultrafiltrate is the same as in the plasma; whereas if SC = 0, then there is no solute removal as may be seen in a toxin with extensive protein binding or large molecular size. The sieving coefficient can be used to calculate the rate of drug clearance:

Clearance (mL/min) = SC * Ultrafiltration Rate (mL/min)

Because the SC is constant for a given drug and membrane combination, clearance varies linearly with ultrafiltration rate.[13]

As discussed previously in this chapter, the efficiency with which extracorporeal therapies remove drugs and toxins from the bloodstream does not necessarily correlate with clinical utility. For drugs with a large Vd due to extensive tissue binding, the majority of the total body burden is not available to the membrane for elimination despite a high extraction ratio or sieving coefficient, and therefore these modalities would only make a minor impact on total body drug clearance.


Indications for the Use of Extracorporeal Detoxification

The decision to initiate any form of extracorporeal therapy must take into account the patient's clinical status, drug-related factors, and dialysis-associated factors.[14] There is no absolute drug level that indicates when dialysis must be initiated; instead, the decision is based on the clinical condition of the patient and potential sequelae from the poisoning. There are several factors that influence the decision to consider extracorporeal therapy in acute poisonings:[4] [8] (1) the dialysis membrane should be readily permeable to the toxin or the adsorbent should bind the poison with high affinity, (2) a clinically significant proportion of the drug is present in the intravascular space or rapidly equilibrates with plasma so that it is readily available for removal by extracorporeal techniques, (3) drug levels in the blood are directly correlated with the pharmacologic effects of the drug, and (4) extracorporeal methods increase total body clearance of the toxin by ≥30%.

Indications for the initiation of extracorporeal therapy include: [4] [12] (1) progressive clinical deterioration despite aggressive supportive therapy, (2) impairment of normal routes for drug excretion, (3) poisoning with a toxin that causes serious metabolic and/or delayed effects for which supportive therapy is ineffective (e.g., methanol intoxication may result in blindness or theophylline poisoning may cause seizures with permanent neurologic sequelae or death), (4) presence of comorbid conditions that predispose the patient to increased susceptibility to the effects of the toxin, (5) depression of midbrain function resulting in hypoventilation, hypothermia, and hypotension, (6) extracorporeal detoxification removes the drug at a rate that exceeds endogenous or spontaneous clearance, and (7) presence of overt signs and symptoms of toxicity.

Intermittent Hemodialysis

The same principles that dictate solute removal in intermittent hemodialysis (IHD) also apply to toxin elimination (see Chapters 58 and 60 for a detailed discussion). Characteristics of the drug that favor efficient toxin removal by hemodialysis are low molecular weight (<500 D), small volume of distribution (Vd < 1 L/kg), low protein binding (<80%), non-ionized state, and high water solubility/low lipid solubility. Although both hemodialysis and hemoperfusion (discussed later) are recommended for a wide variety of drug and chemical intoxicants,[12] the most common toxins amenable to hemodialysis that share the previously listed characteristics are the alcohols, salicylate, lithium, and theophylline.

Features of the dialysis system also determine clinical efficacy of drug clearance and include membrane type, membrane surface area, and blood and dialysate flow rates. Because hemodialysis depends predominantly on diffusion for solute removal with a smaller contribution from convection, toxin elimination is limited by the pore size of the dialysis membrane as the molecular weight of the drug increases. In this setting, drug clearance is more dependent on convection and the rate of ultrafiltration to increase solute drag. High-flux, biocompatible membranes are more permeable and can remove larger molecular weight compounds. Membrane surface area also influences clearance such that membranes possessing a larger surface area will eliminate drugs more effectively. Increasing both the blood flow and dialysate flow rates establishes an optimal concentration gradient between the blood and dialysate resulting in greater diffusion and elimination of the drug; however, drug clearances tend to plateau at blood flow rates of 200-300 mL/min.[12]

There are several distinct advantages of hemodialysis compared to other extracorporeal modalities in the management of acute poisonings. Both solute and volume removal occur rapidly, and the common complications associated with fatal intoxications are corrected by hemodialysis—acute kidney injury (AKI), pulmonary edema, volume overload, acid-base abnormalities (e.g., metabolic acidosis), and electrolyte disturbances (e.g., hyperkalemia).

The most common acute complication of hemodialysis is systemic hypotension. The incidence of hypotension in hemodialysis patients ranges between 15% and 30%. Although there are several strategies that may be employed to improve hemodynamic stability during hemodialysis (e.g., sodium modeling, ultrafiltration modeling, lowering dialysate temperature to 35°C, increasing dialysate calcium concentration, use of the α1-adrenergic agonist midodrine), hypotension remains the principle drawback that limits the efficacy of drug removal during this procedure.


Hemoperfusion refers to the circulation of anticoagulated blood through an extracorporeal circuit equipped with an adsorbent-containing cartridge using either activated charcoal or a resin as the adsorbent. Drug-related characteristics that influence toxin removal with hemoperfusion differ from that of hemodialysis. In hemoperfusion, drug elimination is dependent on the affinity of the charcoal or resin to adsorb the toxin rather than diffusion. Hemoperfusion is therefore not limited by molecular weight, lipid solubility, or protein binding of the drug. Charcoal hemoperfusion results in irreversible binding, whereas resin hemoperfusion reversibly binds the toxin.[12] It is the preferred method when the toxin is lipid soluble or highly protein bound and is effective for compounds ranging in molecular weight from 113 to 40,000 daltons. If the toxin is eliminated equally well by either hemodialysis or hemoperfusion then hemodialysis is the preferred treatment modality because dialysis is associated with fewer complications, allows rapid correction of electrolyte disturbances, and is technically easier to perform. At present, the clinical use of hemoperfusion occurs most commonly with theophylline poisoning. Alcohols and lithium are poorly adsorbed by activated charcoal and acute poisonings with these agents should not be treated with hemoperfusion. Although certain exchange resins (XAD-4 polystyrene resin) are highly effective in the removal of organic solutes and nonpolar, lipid-soluble drugs (e.g., glutethimide, theophylline, and short-acting barbiturates), they are no longer available in the United States. Characteristics of the hemoperfusion devices clinically available in the United States are listed in Table 62-3 .[12]

TABLE 62-3   -- Available Hemoperfusion Devices



Sorbent Type

Amount of Sorbent

Polymer Coating



Petroleum-based spherical charcoal

170 g



Biocompatible system


50, 100, 250 cc

Heparin-hydrogel polymer




100, 300 g

Cellulose acetate



Norit extruded charcoal

260 g

Cellulose acetate

From Winchester JF: Dialysis and hemoperfusion in poisoning. Adv Ren Replace Ther 9:20–30, 2002.




The extracorporeal circuit for hemoperfusion is similar to hemodialysis aside from the disposable adsorbent cartridges. Original hemoperfusion cartridges consisted of uncoated granular carbon as the activated charcoal. These earlier cartridges caused multiple complications including adsorption of platelets by the charcoal, hypocalcemia, and complement release. Current hemoperfusion devices microencapsulate the sorbent material with a thin (0.05 mm), porous semipermeable membrane, which serves to improve biocompatibility by minimizing direct contact between the adsorbent and the blood constituents without impairing the adsorptive capacity of the sorbent. Anticoagulation is performed with heparin infusion at a rate of at least 2000 U/h to avoid clotting of the system. Requirements for resin hemoperfusion are slightly higher. Blood flow rates should be maintained at 250 to 400 mL/min for efficient drug removal.

Complications of hemoperfusion include (1) mild thrombocytopenia and platelet losses of ≤30% caused by adsorption of platelets by the activated charcoal, with recovery of the platelet count within 24 to 48 hours, (2) transient leukopenia (<10% decrease) secondary to complement activation by surface contact and margination of leukocytes, (3) reduction of fibrinogen and fibronectin caused by charcoal adsorption, (4) hypothermia from the reinfusion of a large volume of unheated extracorporeal blood, (5) hypocalcemia due to absorption by charcoal, and (6) hypoglycemia. Although these complications were more common with the earlier, uncoated charcoal devices, hemoperfusion still carries a higher risk for complications compared with hemodialysis. The decision to initiate hemoperfusion should be considered with caution. The principle disadvantage of hemoperfusion is the saturation of the adsorbent cartridge over time. Saturation of the cartridge occurs after 4 to 8 hours and results from the deposition of cellular debris and plasma proteins to the adsorbent. The cartridge therefore must be changed every 6 to 8 hours. In addition, hemoperfusion does not correct fluid overload, electrolyte abnormalities, or acid-base disturbances but may be used in conjunction with hemodialysis or hemofiltration.


The concurrent use of hemodialysis and hemoperfusion is theoretically the optimal treatment in acute poisoning because it effectively removes toxin by both adsorption and diffusion. Case reports and small studies have described efficient clearance of toxins (theophylline, [15] [16] Amanita phalloides mushroom poisoning,[17] barbiturates,[18] tricyclic antidepressants,[19] organophosphates[20]) with hemodialysis-hemoperfusion; however, data comparing clinical outcomes and efficacy of hemodialysis-hemoperfusion therapy to other extracorporeal modalities are limited. Combination therapy is useful for toxins with a small volume of distribution. For drugs with larger volumes of distribution, the duration of treatment may be extended or the procedure repeated once the toxin redistributes back into the plasma.


The process of hemofiltration removes solute by convection. Drug elimination by hemofiltration depends on the rate of ultrafiltration, the drug protein binding, and the sieving coefficient of the membrane. The sieving coefficient is the mathematical expression of the ability of a solute to cross a membrane by convection (see earlier discussion of the sieving coefficient in section on Measurements of Drug Removal Efficiency). The convective transport of hemofiltration allows high flux of plasma water and efficiently eliminates high molecular weight toxins (≤40,000 D) with small volumes of distribution. Because most poisons possess a molecular size (<1000 D), the rate of removal of these smaller solutes by convection during hemofiltration is proportional to its blood concentration and independent of its size[8]; therefore, hemofiltration does not offer much advantage over hemodialysis.

Continuous Renal Replacement Therapy

Although continuous methods of extracorporeal therapy are commonly used in the management of AKI in the ICU setting, their role in the management of acute poisonings is not well established. Solute removal in continuous renal replacement therapy (CRRT) occurs through either dialysis (diffusion-based transport) or filtration (convective transport). There are various types of CRRT treatments (see Chapters 58 and 60 ): continuous arteriovenous hemodialysis (CAVHD), conti-nuous venovenous hemodialysis (CVVHD), continuous ar-teriovenous hemofiltration (CAVH), continuous venovenous hemofiltration (CVVH), continuous arteriovenous hemodiafiltration (CAVHDF), and continuous venovenous hemodiafiltration (CVVHDF).

These continuous modalities of extracorporeal treatment are rarely used in the management of acute overdose because of their lower drug clearance relative to IHD. CRRT has a distinct advantage in hemodynamically unstable patients who are unable to tolerate the rapid solute and fluid losses of conventional IHD or hemoperfusion. It has been suggested that CRRT may also be effective for the slow, continuous removal of substances that possess avid tissue binding, large volumes of distribution, slow intercompartmental transfer and are prone to “rebound phenomenon” (e.g., lithium,[21] procainamide, and methotrexate[2]). In these cases, CRRT may be used as adjuvant therapy with IHD or hemoperfusion.

Sustained Low-Efficiency Dialysis

Sustained low-efficiency dialysis (SLED) is a new technique that is described as an alternative to intermittent hemodialysis (IHD) and CRRT. In the ICU setting, various factors may hamper effective toxin removal using these traditional therapies. IHD is often complicated by hypotension as well as inadequate solute and fluid removal [22] [23]; whereas CRRT is associated with higher complexity, increased nursing requirements, continuous anticoagulation, and higher expenses for specialized equipment and customized solutions. [22] [24] SLED combines the advantages of both IHD and CRRT by using conventional hemodialysis machines at slower blood flow rates (Qb ≈200 mL/min), reduced dialysate flow rates (Qd ≈300–350 mL/min) for prolonged periods (Td ≈8–12 hours) on a daily basis. In hemodynamically unstable, critically ill patients SLED is better tolerated, results in high solute clearance and fluid removal, and does not entail the same level of complexity or costs associated with continuous therapies. [22] [25] [26] [27] [28] Although there is mounting evidence to support the use of SLED in the ICU, it remains a novel extracorporeal treatment modality for which specific studies addressing its efficacy in acute poisonings (e.g., salicylate toxicity) remain anecdotal.[29]

Peritoneal Dialysis

There is little role for peritoneal dialysis in acute poisoning. It is an inefficient method of toxin elimination, achieving only 10 to 15 mL/min of maximum drug clearance. There are no advantages of peritoneal dialysis compared to hemodialysis in acute poisonings, and peritoneal dialysis should not be employed unless other more efficient methods of extracorporeal therapy are unavailable.

Drug-related characteristics that favor toxin elimination by peritoneal dialysis include low molecular weight, water solubility, low protein binding, and low volume of distribution. Solute removal is also dependent on the characteristics of the peritoneal membrane (transport status), dialysate composition, frequency of exchanges (hourly exchanges), and dwell time (≈30 min). The composition of dialysate can be modified in several ways. Increasing the dextrose concentration of the dialysate enhances ultrafiltration and convective transport resulting in enhanced elimination of water soluble substances. Altering dialysate pH above the pKa for a weakly acidic drug promotes a shift to its ionized form leading to improved solute clearance from the body. The addition of albumin to the dialysate has also been used to enhance the elimination of highly protein bound substances such as barbiturates.[30] In the setting of hypothermia, using preheated dialysate has been shown to be clinically effective in rapidly reversing hypothermia.[31]

Plasma Exchange and Plasmapheresis

Plasma exchange is a process in which plasma is removed from the patient and replaced with fresh frozen plasma or stored plasma. The term plasmapheresis is used when the replacement fluids include other products (e.g., albumin) rather than plasma alone. The role of plasma exchange is not well defined in acute poisoning but has been used for toxins that are highly protein bound (>80%) with low volumes of distribution (<0.2 L/kg body weight). [32] [33]Poisonings complicated by massive hemolysis (e.g., hemolytic anemia from sodium chlorate poisoning) or methemoglobinuria are other indications for plasma exchange. Plasma exchange not only removes the toxin but it also eliminates red cell fragments and free hemoglobin.[34] Adverse outcomes from plasma exchange involve complications associated with vascular access placement, bleeding, and hypersensitivity reactions to the replacement plasma proteins.


The majority of poisoning cases do not require treatment with extracorporeal therapy. In fact, the drugs or toxins that are most commonly responsible for poisoning-related fatalities are not effectively treated or amenable to extracorporeal removal (e.g., acetaminophen, tricyclic antidepressants, short-acting barbiturates, stimulants, and “street drugs”). Acute intoxications that have been shown to be clinically responsive to extracorporeal methods include the alcohols (ethylene glycol, methanol, isopropanol, ethanol), lithium, salicylate, theophylline, and valproic acid. The remainder of this chapter will focus on the clinical characteristics of these agents and the role of extracorporeal therapy in the management of acute intoxication with these drugs.

The Alcohols—Ethylene Glycol, Methanol, Isopropanol

Poisonings with the toxic alcohols ethylene glycol, methanol, or isopropanol share common clinical and biochemical characteristics and are associated with significant morbidity and mortality if left untreated. If appropriate treatment is initiated early, the prognosis is excellent. These substances and their metabolites all share several characteristics that make them ideal for removal by hemodialysis: low molecular weight, small volume of distribution, water soluble, and low protein binding.

Ethylene Glycol


Ethylene glycol is a colorless, odorless, sweet-tasting substance commonly found in antifreeze, various solvents, hydraulic brake fluid, de-icing solutions, detergents, lacquers, and polishes. It is rapidly absorbed by the gastrointestinal system and reaches a peak serum concentration 1 to 4 hours after ingestion with an elimination half-life of 3 hours. The molecular weight of ethylene glycol is 62 g/mol, and it is highly water soluble with a low volume of distribution of 0.5 to 0.8 L/kg.[35] The accepted minimum lethal dose of ethylene glycol for an adult is 1.0 to 1.5 mL/kg or 100 mL.


Ethylene glycol itself is non-toxic, and it is the accumulation of its metabolites that is responsible for its severe toxicity. Figure 62-1 illustrates the oxidative pathway of ethylene glycol metabolism. In the presence of the electron acceptor, nicotinamide adenine dinucleotide (NAD), ethylene glycol is oxidized to glycoaldehyde by alcohol dehydrogenase. Aldehyde dehydrogenase then rapidly converts glycoaldehyde to glycolic acid followed by the slow conversion of glycolic acid to glyoxylic acid (the rate limiting step). The final end products include oxalic acid, glycine, oxalomalic acid, and formic acid. Both ethanol and fomepizole interfere with the metabolism of ethylene glycol to its toxic metabolites by competitively inhibiting alcohol dehydrogenase. Although pyridoxine and thiamine are cofactors involved in the metabolism of glyoxylate to its non-toxic metabolites (glycine and α-hydroxy-b-ketoadipate, respectively), there is no clinical data to support the clinical utility of these cofactors in the treatment of acute ethylene glycol intoxication.[35]

FIGURE 62-1  Metabolism of ethylene glycol. 4-MP, fomepizole; NAD+, nicotinamide adenine dinucleotide; NADH+, reduced form of nicotinamide adenine dinucleotide; TCA, tricarboxylic acid; broken arrow, inhibitors of alcohol dehydrogenase; asterisk, rate-limiting step.



The underlying mechanisms of ethylene glycol toxicity are tissue destruction from calcium oxalate deposition and profound acidosis due to the accumulation of its metabolites (e.g., glycoaldehyde, glycolic acid, lactate). Systemic calcium oxalate deposition has been observed. In the kidney, acute tubular necrosis from calcium oxalate deposition in the proximal tubules, interstitial nephritis, focal hemorrhagic cortical necrosis, direct renal cytotoxicity, and obstruction are suggested mechanisms of renal toxicity in ethylene glycol poisoning.[35] The classic pathologic findings of ethylene glycol toxicity are the presence of calcium oxalate crystals in the kidney and acute tubular necrosis. However, because the degree of renal injury does not correlate with the amount of calcium oxalate deposition in the kidney, it has been suggested that glycolic acid or its metabolites are primarily responsible for the renal failure.[35]The metabolic acidosis results from the formation of glycolic acid and increased lactic acid production.

Clinical Presentation

The clinical course of ethylene glycol toxicity occurs in three phases. The severity and progression of each stage depends on certain factors such as the dose of ethylene glycol ingested, concurrent ingestion of ethanol, and timing of treatment.[35]

Phase 1 is the neurologic phase occurring 30 minutes to 12 hours after ingestion. Shortly after ethylene glycol ingestion, patients appear inebriated similar to that observed with ethanol intoxication but without the odor of alcohol on their breath. Gastrointestinal irritation from ethylene glycol results in nausea, vomiting, and hematemesis. As ethylene glycol undergoes oxidation to glycoaldehyde and glycolic acid (4–12 hours after ingestion), symptoms of CNS depression predominate. Altered consciousness may progress to coma and seizures in severe poisonings. Cerebral edema, nystagmus, ataxia, ophthalmoplegias, myoclonic jerks, and hyporeflexia may also occur.

Phase 2 is the cardiopulmonary phase, occurring 12 to 24 hours after ingestion. During this second phase of ethylene glycol intoxication, calcium oxalate crystals deposit in the vasculature, myocardium, and lungs. Patients may develop tachycardia, mild hypertension, congestive heart failure, acute respiratory distress syndrome, severe metabolic acidosis, and multiorgan failure. Most deaths occur during this phase. [35] [36]

Phase 3 is the renal phase, occurring 24 to 72 hours after ingestion. In this final stage, calcium oxalate precipitates in the kidney resulting in flank pain, acute tubular necrosis, hypocalcemia, microscopic hematuria, and oliguric AKI.

With early medical intervention, ethylene glycol toxicity tends to resolve completely. The removal of ethylene glycol and its toxic metabolites, as well as prevention of calcium oxalate formation, can reverse tissue destruction. Renal recovery is generally complete with proper supportive therapy although permanent renal, CNS, and cranial nerve damage have been described.[37]

Diagnosis and Laboratory Data

The diagnosis of ethylene glycol toxicity should be suspected in patients who present with the following features: high anion gap metabolic acidosis, high osmolal gap, hypocalcemia, inebriation without the smell of alcohol on their breath, altered mental status, and calcium oxalate crystals in the urine.[35]

The osmolal gap is an extremely useful diagnostic tool in the various alcohol intoxications. It represents an estimation of the unmeasured, osmotically active substances in the serum. It is the difference between the measured osmolality, as determined by freezing point depression, and the calculated osmolality:

ΔOsmgap = Osmmeasured - Osmcalculated

where ΔOsmgap is the osmolal gap (mOsm/kg), Osmmeasured is the measured osmolality (mOsm/kg), and Osmcalculated is the calculated osmolality (mOsm/kg). The calculated osmolality is determined by the following formula:

where [Na] is the concentration of sodium in mEq/L, [glucose] is the concentration of glucose in mg/dL, and [BUN] is the concentration of blood urea nitrogen in mg/dL. The measured osmolality normally ranges between 270 and 290 mOsm/kg and the normal osmolal gap is <10 to 12 mOsm/kg H2O. An elevated osmolal gap suggests the presence of ethylene glycol, methanol, ethanol, isopropanol, propylene glycol, or acetone.

Although an elevated osmolal gap is a significant clinical finding, a normal osmolal gap does not exclude the diagnosis of ethylene glycol or methanol poisoning. [38] [39] In the case of ethylene glycol toxicity, ethylene glycol is an osmotically active compound but its metabolite, glycolic acid, does not contribute to the osmolal gap. As ethylene glycol is metabolized to glycolic acid, the osmolal gap may decrease to normal. Patients who present later in the course of their ethylene glycol intoxication may actually have a normal osmolal gap.[40]

The osmolal gap is also clinically useful in estimating the serum concentration of the toxin. The following equation can be used:

In this equation DOsmgap is the osmolal gap, and MW is the molecular weight of the toxin: acetone (58 daltons), ethanol (46 daltons), ethylene glycol (62 daltons), isopropanol (60 daltons), and methanol (32 daltons). Using this calculation, serial measurements of the osmolal gap can be used to monitor changes in the ethylene glycol level during hemodialysis. This is especially useful if serum drug levels are not readily available to guide therapy during treatment. At a serum concentration of 100 mg/dL, the contribution of the toxin to the osmolal gap is as follows: acetone (18 mOsm/kg H2O), ethanol (22 mOsm/kg H2O), ethylene glycol (16 mOsm/kg H2O), isopropanol (17 mOsm/kg H2O), and methanol (31 mOsm/kg H2O).[37]

Urinalysis can often provide supporting evidence of ethylene glycol toxicity. Calcium oxalate crystals (monohydrate and dihydrate forms) may be present in the urine sediment and are birefringent when viewed under polarized light. Urinary calcium oxalate crystals appear 4 to 8 hours after ingestion and are found in approximately 50% of patients with ethylene glycol intoxication. [35] [41] The presence in urine of the monohydrate form, described as prism or dumbbell-shaped, is not specific for ethylene glycol toxicity and is normally seen in individuals who ingest large amounts of vitamin C or high oxalate-containing foods (e.g., cocoa, garlic, tea, tomatoes, spinach, rhubarb). The dihydrate form, which is octahedral or tent-shaped, is present only under high urinary calcium and oxalate conditions. The presence of the dihydrate form is more specific for ethylene glycol toxicity.

Urine that fluoresces under Wood lamp illumination is another unique feature of ethylene glycol ingestion. Many types of antifreeze contain sodium fluorescein, a fluorescent dye used to detect radiator leaks. For up to 6 hours after ingestion, sodium fluorescein can be detected in the urine. The presence of fluorescent urine by Wood lamp examination suggests ethylene glycol intoxication caused by antifreeze ingestion,[42] although there are limitations to this method of detection.[43]

In addition to the high anion gap metabolic acidosis seen in ethylene glycol toxicity, other laboratory abnormalities may be observed. Hypocalcemia, caused by the formation of calcium oxalate complexes, can result in prolonged QT on ECG. Microscopic hematuria, leukocytosis, and an elevated protein level in cerebral spinal fluid may also be found.

Because the serum level of glycolic acid, which is primarily responsible for metabolic acidosis in ethylene glycol intoxication, correlates well with the increase in the anion gap and decrease in the serum bicarbonate, serial measurements of the anion gap can be used to monitor improvement in metabolic acidosis during treatment.[4]


As with any intoxication, supportive therapy should be directed toward stabilization of the airway, breathing, and circulatory system. Intravenous glucose (50 mL of 50% glucose) should be administered in patients with altered mental status. If the patient is a suspected alcoholic or is malnourished, then thiamine (100 mg every 6 hours), multivitamin supplementation, and pyridoxine (50 mg every 6 hours) should be given. Aggressive fluid resuscitation to maintain urine output, correct dehydration, and support circulatory shock should be initiated, although this must be done with caution to avoid volume overload in the setting of renal dysfunction. Metabolic acidosis should be treated with IV sodium bicarbonate therapy, especially if the serum bicarbonate is <15 mEq/L or the arterial pH is <7.35.

Seizures, which may be caused by hypocalcemia, should be controlled with standard anticonvulsant therapy. In general, asymptomatic hypocalcemia is not routinely treated in the setting of ethylene glycol intoxication because this may potentially exacerbate calcium oxalate crystal formation and deposition. If seizures persist despite adequate anticonvulsant therapy, then 10 to 20 mL of 10% calcium gluconate (0.2 to 0.3 mL/kg) can be infused slowly.

Treatment of both ethylene glycol and methanol intoxication is focused on inhibition of toxic metabolite production or dialytic removal of the toxin and its metabolites. The antidotes used in the management of ethylene glycol overdose are ethanol and fomepizole, inhibitors of alcohol dehydrogenase that prevent the metabolism of ethylene glycol to its metabolites. Indications for use of an antidote in ethylene glycol toxicity according to the American Academy of Clinical Toxicology (AACT) guidelines are (1) plasma ethylene glycol concentration more than 20 mg/dL, (2) documented recent ingestion of toxic amounts of ethylene glycol and osmolal gap of more than 10 mOsm/L, or (3) history or strong clinical suspicion of ethylene glycol poisoning and at least two of the following: arterial pH less than 7.3, serum bicarbonate less than 20 mEq/L, osmolal gap of more than10 mOsm/L, or presence of oxalate crystals in the urine.[35]

Ethanol has been the traditional antidote for both ethylene glycol and methanol intoxication. Because alcohol dehydrogenase has a 100-fold greater affinity for ethanol than ethylene glycol, the ethanol competitively inhibits alcohol dehydrogenase and prevents the metabolism of ethylene glycol to its toxic intermediates. At an ethanol serum concentration of 50 to 100 mg/dL, the active sites on alcohol dehydrogenase are saturated.[35]

Ethanol can be given intravenously as 10% ethanol diluted in D5W (10 g of ethanol per 100 mL of solution). The loading dose of ethanol is 0.6 to 0.7 g ethanol/kg (7.6 mL 10% ethanol/kg) and the maintenance dose is 66 mg ethanol/kg/hr (0.83 mL 10% ethanol/kg/hr) for nondrinkers and 154 mg ethanol/kg/hr (1.96 mL 10% ethanol/kg/hr) for alcoholics. If hemodialysis is initiated, the ethanol dose must be adjusted to replace the ethanol removed during dialysis. The maintenance dose during dialysis can be approximated by doubling the dose or more specifically, 169 mg ethanol/kg/hr (2.13 mL 10% ethanol/kg/hr) for nondrinkers and 257 mg ethanol/kg/hr (3.26 mL 10% ethanol/kg/hr) in alcoholics.[35] Ethanol may also be added to the dialysate to give a concentration of 100 mg/dL using 95% ethanol instead.[44] Serum ethanol concentrations should be monitored every 1 to 2 hours to ensure that blood levels remain therapeutic at 100 to 150 mg ethanol/dL.

The adverse side effects of ethanol are hypoglycemia, inebriation, CNS depression, emotional lability, poor coordination, and slurred speech. The altered mental status from ethanol therapy may mask the signs and symptoms of ethylene glycol toxicity. Use of ethanol requires close monitoring of serum ethanol levels to maintain therapeutic levels because its pharmacokinetic properties are erratic, and it is not an ideal antidote.

Fomepizole (4-MP, 4-methylpyrazole) is a newer antidote that is a potent competitive inhibitor of alcohol dehydrogenase that has limited side effects. Inhibition of alcohol dehydrogenase with fomepizole effectively blocks the formation of toxic metabolites and prevents end-organ damage in both ethylene glycol and methanol ingestions. The U.S. Food and Drug Administration (FDA) approved the use of fomepizole as first-line treatment for both acute ethylene glycol intoxication (December 1997) and acute methanol toxicity (December 2000). The advantages of fomepizole compared to ethanol are (1) 500- to 1000-fold more potent inhibitor of alcohol dehydrogenase than ethanol, (2) ease of dosing, (3) predictable pharmacokinetics and no blood monitoring, (4) few adverse side effects with no CNS depression or hepatotoxicity, and (5) longer duration of action than ethanol. The primary disadvantage of fomepizole is its cost. Adverse side ef-fects include headache, nausea, dizziness, eosinophilia, rash, tachycardia or bradycardia, and mild but transient elevation of hepatic transaminases although most of these were felt to be attributable to the toxic ingestion rather than fomepizole itself.[45]

Fomepizole is administered intravenously over 30 minutes. The loading dose of fomepizole is 15 mg/kg followed by 10 mg/kg every 12 hours for 4 doses (i.e., 48 hours). After 48 hours, fomepizole is continued at 15 mg/kg every 12 hours until the ethylene glycol concentration is undetectable or less than 20 mg/dL and the patient is asymptomatic with a normal arterial pH. [35] [46] During hemodialysis, the dosing interval must be changed to every 4 hours or a constant infusion of 1.0 to 1.5 mg/kg/hr can be used to maintain adequate therapeutic levels of fomepizole.[47]

Hemodialysis is an extremely effective method in the removal of ethylene glycol and its toxic metabolite, glycolic acid. The clearance of ethylene glycol ranges from 145 to 230 mL/min and depends on the type of dialyzer and blood flow rate. The elimination half-life of ethylene glycol on hemodialysis is 2.5 to 3.5 hours. Glycolic acid is also easily eliminated by hemodialysis with a clearance rate of 105 to 170 mL/min and an elimination half-life of 2.5 hours.[35]

Indications for hemodialysis are as follows [35] [48]: deteriorating clinical status despite supportive therapy, metabolic acidosis (arterial pH <7.25 to 7.30), AKI with a serum creatinine more than 3.0 mg/dL (265 mmol/L) or rise in serum creatinine by 1.0 mg/dL (90 mmol/L), or acid-base/electrolyte abnormalities unresponsive to standard treatment. Traditionally, an ethylene glycol level of more than 50 mg/dL was an indication for hemodialysis. However, reports have described patients with normal renal function and no evidence metabolic acidosis who have been effectively treated with fomepizole despite ethylene glycol levels of more than 50 mg/dL. [49] [50] In the absence of both renal insufficiency and significant metabolic acidosis, fomepizole may be used without instituting hemodialysis although if metabolic acidosis develops, hemodialysis should be initiated.[35]

Hemodialysis should be continued until the ethylene glycol level is undetectable or an ethylene glycol level less than 20 mg/dL and the disappearance of metabolic acidosis signs of systemic toxicity.[35] Rebound distribution of ethylene glycol may occur within 12 hours of discontinuing hemodialysis; therefore, serum osmolality and electrolytes should be closely monitored every 2 to 4 hours for 12 to 24 hours after withdrawal of hemodialysis. Treatment with ethanol or fomepizole should be continued after hemodialysis for several hours to protect from possible rebound effects.



Methanol, or wood alcohol, is a clear, colorless liquid with a faint, slightly alcoholic scent that is bitter tasting and highly volatile. It is often used as a solvent, an intermediate of chemical synthesis during various manufacturing processes, or as an octane booster in gasoline. The solvents containing methanol include windshield or glass cleaning solutions, enamels, printing and duplicating solutions, stains, dyes, varnishes, thinners, paint removers, camp stove fuels, solid canned fuels, and antifreeze additives for gas. Most methanol poisonings result from ingestion, although there have been rare reported cases of methanol toxicity from inhalation or transdermal absorption.

Methanol is rapidly absorbed and distributed with the peak serum concentration occurring within 30 to 60 minutes post ingestion and an elimination half-life (untreated) of 12 to 20 hours. Methanol has a molecular weight of 32 g/mol, is water soluble, easily crosses the blood-brain barrier, and has a low volume of distribution (Vd = 0.60–0.77 L/kg) equivalent to water.[51] The estimated minimum lethal dose is 10 mL,[40] although this is highly variable.


Methanol itself has relatively low toxicity, but the oxidation of methanol to its respective metabolites is responsible for the toxicity seen in methanol poisoning. As illustrated in Figure 62-2 , the metabolic pathway of methanol begins in the liver where methanol is oxidized by alcohol dehydrogenase to generate formaldehyde in the presence of NAD. The short half-life of formaldehyde (1 to 2 minutes) prevents its accumulation in the body, and it is rapidly oxidized by formaldehyde dehydrogenase to form formic acid, the principle toxin in methanol intoxication. The rate limiting step, which depends on folic acid, involves the formation of 10-formyl tetrahydrofolate from folic acid and tetrahydrofolate. 10-Formyl tetrahydrofolate is then metabolized to the end products, carbon dioxide and water. Similar to ethylene glycol toxicity, management of methanol toxicity focuses on prevention of toxic metabolite formation (i.e., formic acid) and the elimination of the toxin and its metabolites by hemodialysis. Both ethanol and fomepizole are used as antidotes for methanol poisoning because they competitively inhibit alcohol dehydrogenase and prevent formic acid accumulation.

FIGURE 62-2  Metabolism of methanol. 4-MP, fomepizole; NAD+, nicotinamide adenine dinucleotide; NADH+, reduced form of nicotinamide adenine dinucleotide; broken arrow, inhibitors of alcohol dehydrogenase; asterisk, rate-limiting step.



The primary substance responsible for the toxicity of methanol poisoning is formic acid, which inhibits cytochrome c oxidase in the mitochondria of cells through its binding to the ferric moiety of cytochrome oxidase.[51] Cellular oxidative metabolism is interrupted resulting in cell hypoxia and cell death. This mechanism of cellular injury is similar to cyanide or carbon monoxide poisoning, although formic acid is a weaker inhibitor of cytochrome c oxidase.[52] It has been suggested that formic acid may also bind to the ferric part of hemoglobin, which may account for the rare cases of methemoglobinemia observed in severe cases of methanol intoxication.

The etiology of the metabolic acidosis in methanol poisoning is multifactorial with formic acid having both direct and indirect roles in the development of acidosis.[51] Formic acid directly causes acidosis by contributing a proton as it dissociates to formate and hydrogen ions. The indirect mechanism involves the production of lactic acid as formic acid interferes with oxidative metabolism and shifts glycolysis from aerobic to anaerobic glycolysis. The increase in both lactic acid and tissue hypoxia decreases cellular pH, which promotes additional undissociated formic acid generation resulting in exacerbation of the acidosis and cellular injury. Another factor in the formation of lactate is the shift of the NAD/NADH ratio when methanol is converted to formaldehyde. The alteration in the NAD/NADH ratio promotes the metabolism of pyruvate to lactate. Acidosis in the early stages of methanol poisoning primarily results from the accumulation of formic acid. Later in the course of methanol intoxication, tissue hypoxia and disruption of mitochondrial respiration are responsible for the generation of lactic acid and the persistence of acidosis in methanol toxicity.

The accumulation of formic acid in the eye has been implicated in the ocular toxicity found in methanol ingestion.[51] Optic nerve cells possess few mitochondria and low levels of cytochrome c oxidase, so the optic nerve is extremely susceptible to the toxic effects of formic acid. With correction of the acidosis, vision improves. With resolution of the acidosis there is an increase in the dissociated form of formic acid, which is unable to readily diffuse into the CNS. The accumulation of formic acid in the eye is therefore decreased in this setting.

Clinical Presentation

The principle sites of methanol toxicity are the CNS (especially the putamen), eyes, and GI tract, and the most serious complications are blindness and death. Co-ingestion of ethanol in methanol poisoning can delay the onset of symptoms for more than 24 hours because ethanol competitively inhibits alcohol dehydrogenase. The clinical presentation of methanol intoxication is characterized by an early stage, latent interval, and delayed stage.

The early stage of methanol overdose involves CNS depression in which the patient appears inebriated and drowsy. This phase is mild and transient and is followed by a latent interval lasting from 6 to 30 hours. This period corresponds to the metabolism of methanol and the gradual accumulation of its toxic metabolite, formic acid. Typically the sensorium is clear, and the only presenting symptom may be blurred vision. The latent interval is then followed by the delayed stage in which formic acid accumulates and systemic toxicity develops.

The visual changes associated with methanol poisonings occur rapidly and symmetrically because formic acid is directly toxic to the retina. Ocular symptoms are highly variable and may include blurred vision, central scotoma, impaired pupillary response to light, decreased visual acuity, photophobia, visual field defects, or progression to complete blindness. Early findings on the ophthalmic examination are hyperemia of the optic disc and fixed, dilated pupils. Later manifestations of ocular toxicity include retinal edema, optic disc edema with loss of physiologic cupping, and optic atrophy.[53] Most often the visual abnormalities resolve, although in 25% to 33% of patients with methanol poisoning, the visual changes are permanent.[51] Both pupillary status and retinal edema correlate with the severity of methanol poisoning.[51]

The CNS effects of mild to moderate methanol toxicity are headache, vertigo, delirium, lethargy, restlessness, and confusion. In severe cases of methanol intoxication, CNS symptoms may progress to coma and seizures caused by cerebral edema. A Parkinson-like syndrome (rigidity, bradykinesia, mild tremor, masked facies, lethargy, and dementia) has also been observed and is associated with the neurotoxicity to the putamen and subcortical white matter.[51]

Other clinical manifestations of methanol toxicity include gastrointestinal symptoms characterized by nausea, vomiting diarrhea, transient elevation of hepatic transaminases, and abdominal pain from acute pancreatitis. Myoglobinuric acute kidney injury (AKI) develops rarely, although the diagnosis may be complicated by the interference of methanol with serum creatinine determination. Another uncommon complication of methanol toxicity is severe reversible cardiac failure. Poor prognostic signs include severe metabolic acidosis, elevated formic acid level, bradycardia, cardiovascular shock, anuria, and seizures or coma at presentation. The most common causes of death in methanol poisoning are respiratory failure or sudden respiratory arrest.[51]

Diagnosis and Laboratory Data

Methanol overdose should be considered in patients with visual changes, abdominal pain, high-anion gap metabolic acidosis, and elevated osmolal gap. Laboratory tests should include a complete blood count; assessment of electrolytes, serum calcium, lipase, amylase, creatine kinase, serum osmolality, and serum methanol and ethanol levels; urinalysis; and arterial blood gas determinations.

Formic acid and lactate accumulation in the body contribute to the development of an elevated anion-gap metabolic acidosis and correlate with poor clinical outcome. Because methanol is an osmotically active compound, an osmolal gap is present during the early and latent phases of intoxication. For every 100 mg/dL of methanol in the serum, methanol contributes 31 mOsm/kg H2O to the osmolal gap. As methanol is metabolized to formic acid during the delayed stage, the osmolal gap may return to normal because formate is not osmotically active. Similar to ethylene glycol toxicity, a normal osmolal gap does not exclude the diagnosis of methanol poisoning. The specifics of the osmolal gap are described in greater detail in the previous section on Ethylene Glycol: Diagnosis and Laboratory Data.

Other laboratory abnormalities include increased serum amylase levels and mean corpuscular volumes. The serum amylase is typically elevated from inflammation of the salivary glands or acute necrotizing pancreatitis. The increased mean corpuscular volume is believed to result from the toxic effects of formaldehyde on cellular ion transport.[51]


Initial management of methanol poisoning involves supportive care: stabilization of the airway, breathing, and circulation. The most serious sequelae of methanol overdose are metabolic acidosis, ocular toxicity, seizures, and coma. Treatment focuses on preventing these complications by limiting the accumulation of the toxic metabolites, formic acid, and formaldehyde.

Metabolic acidosis (pH < 7.3) is treated aggressively with sodium bicarbonate therapy, because the severity of intoxication and clinical outcome correlate with the degree of acidosis.[51] Acidosis increases the ratio of formic acid to formate, and it is its un-dissociated form (formic acid) that inhibits cytochrome oxidase more effectively relative to its dissociated form (formate). Correction of the acidosis rectifies the acid-base balance and improves pathophysiologic factors involved in methanol toxicity such that the ratio of formic acid to formate is decreased and inhibition of mitochondrial cytochrome oxidase is ameliorated.[52]

Folic acid supplementation plays a role in the manage-ment of methanol intoxication. The rate-limiting step of methanol metabolism is mediated by 10-formyl tetrahydrofolate synthetase, which is folic acid dependent. Although data are limited, folic acid increases the metabolism of formic acid to carbon dioxide and water. The suggested dose of folic acid is 50 mg intravenously every 4 hours for 5 doses and then once daily. Symptomatic patients should receive one intravenous dose of 1 mg/kg.[40] Thiamine and multivitamin supplementation should also be administered to patients suspected of ethanol abuse. Other supportive measures include aggressive intravenous fluids to maintain adequate urine output and control of seizures with standard anticonvulsants.

Early recognition and rapid treatment are essential in the management of methanol toxicity. The same alcohol dehydrogenase inhibitors, ethanol and fomepizole, are also used in methanol intoxications. Ethanol or fomepizole should be immediately administered if methanol toxicity is suspected to prevent the formation of formic acid. Indications for the use of antidotes in methanol overdose are provided in the AACT guidelines: plasma methanol concentration of more than 20 mg/dL; documented recent ingestion of toxic amounts of methanol and osmolal gap of more than 10 mOsm/kg H2O; or a history or strong clinical suspicion of methanol poisoning and at least two of the following: arterial pH less than 7.3, a serum bicarbonate level less than 20 mEq/L (mmol/L), or an osmolal gap more than 10 mOsm/kg H2O.[51]

Ethanol binds to alcohol dehydrogenase with 10 times greater affinity than methanol, and fomepizole has an even higher affinity for alcohol dehydrogenase than ethanol by 500- to 1000-fold. [48] [51] The use of ethanol and fomepizole in the management of methanol poisoning is similar to that of ethylene glycol, and the dosing of both ethanol and fomepizole is described in detail in the previous section on Ethylene Glycol: Treatment. Antidotes are continued until the methanol concentration is undetectable or symptoms and metabolic acidosis resolve and the methanol level is less than 20 mg/dL.

Hemodialysis effectively removes methanol and formic acid, corrects metabolic acidosis, and shortens the course of hospitalization. Because methanol overdose is associated with an increased frequency of intracerebral hemorrhage, the use of heparin during hemodialysis should be minimized. The clearance rates of methanol and formate are 125 to 215 mL/min and 203 mL/min, respectively, and the rates are dependent on blood flow and the type of membrane used. Hemodialysis is indicated for the following: serum methanol levels higher than 50 mg/dL, severe metabolic acidosis (pH < 7.25 to 7.30), visual changes, dose of ingested methanol more than 30 mL, seizures, deteriorating clinical status despite intensive supportive therapy, renal failure, or electrolyte abnormalities not responsive to standard therapy. [51] [54] According to AACT guidelines, a methanol concentration of more than 50 mg/dL is no longer an absolute indication for hemodialysis because fomepizole may be used as first-line treatment for methanol poisoning if there is no evidence of ophthalmologic impairment or severe acidosis.

Hemodialysis is continued until the serum methanol concentration is undetectable or until the methanol level is less than 25 mg/dL and the anion-gap metabolic acidosis and osmolal gap are normal. The presence of visual changes is not an indication for continued dialysis because ophthalmologic abnormalities may be transient or permanent. Methanol redistributes after the discontinuation of hemodialysis and may result in rebound of the methanol levels. Close monitoring of the serum osmolality and electrolytes should be continued every 2 to 4 hours for 12 to 36 hours after hemodialysis. To protect the patient from the toxic effects of methanol rebound, fomepizole or ethanol therapy should be continued for several hours after withdrawal of hemodialysis until methanol levels are undetectable or less than 20 mg/dL with resolution of acidosis and symptoms.[51]



Isopropanol (i.e., isopropyl alcohol) is a clear, colorless, bitter liquid commonly found in “rubbing alcohol,” skin lotion, hair tonics, aftershave lotion, denatured alcohol, solvents, cements, cleaning products, de-icers, and the manufacturing process of acetone and glycerin. Intoxication may occur through either ingestion or inhalation of vapors, especially in infants.

Isopropanol is rapidly absorbed by the gastrointestinal system and reaches a peak serum concentration 15 to 30 minutes following ingestion with an elimination half-life of 3 to 7 hours. It is water soluble, has a molecular weight of 60 g/mol, and a volume of distribution equivalent to water (Vd = 0.6 L/kg).[40] Unlike the other alcohols discussed previously, the parent compound isopropanol is responsible for the toxic effects observed in isopropanol intoxication. Delaying the metabolism of isopropanol therefore is not considered a beneficial method of treatment.


The metabolism of isopropanol is illustrated in Figure 62-3 . Isopropanol is oxidized to acetone by alcohol dehydrogenase in the presence of NAD. Acetone is then excreted in the breath and urine.[40]

FIGURE 62-3  Metabolism of isopropanol. NAD+, nicotinamide adenine dinucleotide; NADH+, reduced form of nicotinamide adenine dinucleotide.


Clinical Presentation

The principle targets of isopropanol intoxication are the CNS, cardiovascular, and gastrointestinal system. Clinical symptoms tend to appear within 1 hour of ingestion. Isopropanol is a potent CNS depressant, which is twice as effective as ethanol in its CNS toxicity.[55] Early effects of isopropanol overdose are therefore similar to those of ethanol: confusion, headache, poor coordination, lethargy, and dizziness. These symptoms may progress to ataxia, stupor, coma, and respiratory arrest in severe cases of poisoning.

Gastrointestinal effects include nausea, vomiting, and abdominal pain. Because isopropanol is a gastrointestinal irritant, hematemesis from severe gastritis or hemorrhage may develop. The combined effects of CNS depression and vomiting place the patient at risk for aspiration pneumonia. In animals, hepatotoxicity and fatty liver have also been described.[37]

Another unique feature of isopropanol intoxication is myocardial depression and myocyte toxicity. Hypotension is often severe and is due to multiple factors, including cardiac depression, arrhythmias from cardiomyopathy, vasodilation, dehydration, and gastrointestinal bleeding. Hypotension is the strongest predictor of mortality in isopropanol overdose.[40]

AKI may develop because of severe hypotension or myoglobinuria. Other systemic findings include hypoglycemia from impaired gluconeogenesis, hypothermia, and hemolytic anemia.

Diagnosis and Laboratory Data

The diagnosis of isopropanol overdose should be suspected in any patient with altered sensorium, acetone on the breath that smells “fruity,” elevated osmolal gap but no elevated anion-gap, and the presence of acetonemia or acetonuria (i.e., positive sodium nitroprusside reaction in serum or urine) in the absence of hyperglycemia, glycosuria, or acidosis. Laboratory tests should include complete blood count, electrolytes, serum osmolality, serum acetone, creatine kinase, urinalysis, and arterial blood gas determinations.

Because the end product of isopropanol metabolism, acetone, is not an organic acid, the anion-gap is not elevated. However, if hypotension is severe then poor tissue perfusion and hypoxia may result in lactic acid accumulation and the development of metabolic acidosis. An osmolal gap is present because isopropanol is an osmotically active compound. For a serum level of 100 mg/dL, isopropanol contributes 17 mOsm/kg H2O to the osmolal gap, and 100 mg/dL of acetone contributes an additional 18 mOsm/kg H2O. The clinical utility of the osmolal gap is discussed in further detail in the previous section (see Ethylene Glycol: Diagnosis and Laboratory Data). Other laboratory abnormalities observed in isopropanol intoxication include hypoglycemia, elevated serum creatinine, elevated creatine kinase, and elevated protein in cerebral spinal fluid.


Treatment for isopropanol toxicity focuses on appropriate supportive therapy. Because the oxidation of isopropanol does not produce toxic metabolites, agents that inhibit alcohol dehydrogenase are not used in the management of isopropanol poisoning. Intravenous fluids and pressors should be used in the setting of hypotension. Mechanical ventilation may be indicated for respiratory distress and airway protection. Gastrointestinal lavage is effective in preventing systemic absorption of isopropanol, and up to 87% to 92% of isopropanol can be removed by activated charcoal when administered in a charcoal:alcohol ratio of 20:1.[56]

Hemodialysis effectively removes isopropanol and acetone, although it is usually unnecessary except in severe cases of isopropanol intoxications. Indications for hemodialysis include an isopropanol level more than 400 mg/dL, prolonged coma, hypotension, myocardial depression or tachyarrhythmias, and renal failure.[55]


Lithium salts have been used therapeutically for nearly 150 years, initially in the treatment of gout and “re-discovered” in the 1950s for the treatment of bipolar disorder, its present-day FDA-approved indication. Lithium therapy requires careful monitoring because of its narrow therapeutic index (0.6 to 1.5 mEq/L). Patients who are especially susceptible to lithium toxicity include the elderly, those with intrinsic kidney disease or reductions in GFR secondary to medications (e.g., ACE inhibitors, NSAIDs, diuretic-induced hypovolemia), and patients with lithium-induced diabetes insipidus from chronic lithium therapy.[57]


Lithium is a 7 dalton monovalent cation orally administered as a carbonate (capsule) or citrate (liquid) salt. It is rapidly and completely absorbed in the upper gastrointestinal tract, with peak serum levels occurring within 1 to 2 hours. Sustained-release preparations are also available, with peak levels occurring 4 to 5 hours after ingestion of therapeutic doses but can continue to rise for several days in cases of acute poisoning. [57] [58] Lithium is less than 10% protein bound, is distributed in total body water (Vd 0.7–0.9 L/kg), and demonstrates preferential uptake in certain compartments over others. For example, there is a significant delay in reaching steady-state brain concentrations compared to plasma, which explains the delay between peak blood levels and central nervous system effects after an acute overdose. [59] [60] It is distributed predominantly to the intracellular compartment by active transport, and lithium exit from cells is slow and not well understood. This disposition makes removal of the total body burden by extracorporeal therapy a slow process.

Lithium is entirely eliminated by the kidneys, where it is freely filtered at the glomerulus.[61] The fractional excretion of lithium is usually 20%, with 60% of the filtered lithium reabsorbed in the proximal tubule and 20% distally. Lithium reabsorption follows that of sodium, and sodium avid states (e.g., volume depletion, NSAID use, congestive heart failure, cirrhosis) markedly increase lithium retention. The processes that accomplish lithium reabsorption throughout the nephron are incompletely understood. Although it is known that lithium can substitute for sodium or potassium on several transport proteins (e.g., the ubiquitous Na+/H+ exchanger and the amiloride-sensitive sodium channel, ENaC), the affinity for lithium at the basolateral Na+/K+-ATPase is at least an order of magnitude less than for sodium or potassium. [57] [62]

Lithium also gains entry to other cells most likely by substituting for sodium or potassium on a number of transport proteins. Accumulation seems to primarily occur in the brain white matter and distal renal tubule epithelial cells.[57]It is thought to stabilize cell membranes, and with excessive levels, reduces neural excitation and synaptic transmission.[59]

The elimination half-life ranges from 12 to 27 hours after a single dose with normal renal function. Therefore, 12-hour drug levels should be drawn 5 to 6 days after changing the lithium dose to ensure steady state. The half-life can be prolonged in chronic users of lithium (up to nearly 60 hours after more than 1 year of use) and the elderly.[61]

Clinical Presentation

Lithium overdose may present as an acute intoxication (i.e., accidental ingestion or deliberate poisoning) or as a more gradual intoxication in patients on chronic therapy. Lithium-naive patients with acute poisoning generally present with milder symptoms and have less risk of mortality than those on chronic lithium therapy because of the shorter elimination half-life in these individuals.[60] Intoxication among chronic users generally develops due to infrequent lithium level monitoring around the time of dose adjustments or because of unanticipated states of sodium retention and/or decreased GFR.

The potential symptoms and signs of lithium intoxication can be varied and extensive ( Table 62-4 ). Severity of symptoms in acute intoxication generally correlates with the serum lithium concentration and may be categorized as mild (1.5 to 2.0 mEq/L), moderate (2.0 to 2.5 mEq/L), or severe (more than 2.5 mEq/L), but symptoms may be present even when concentrations are well within the recommended therapeutic range.[58] Symptoms of mild lithium poisoning include lethargy, drowsiness, coarse hand tremor, weakness, nausea, vomiting, and diarrhea. With moderate toxicity, neurologic symptoms predominate with confusion, nystagmus, ataxia, myoclonic jerks, and dysarthria. Plasma lithium concentrations more than 3.5 mEq/L may manifest as seizures, hyperreflexia, stupor, coma, and death. Even survivors of the most extreme cases of lithium toxicity have a significant risk of permanent neurologic sequelae.[57]

TABLE 62-4   -- Symptoms, Signs, and Laboratory Abnormalities of Lithium Intoxication



Laboratory Findings






Elevated TSH


 Diffuse muscle weakness


Mental status changes

 Slurred speech

Decreased anion gap



Acute renal failure



Abnormal ECG



 Prolonged QT


 Increased deep tendon reflexes

 ST-segment depression



 Inverted T waves



 Sinus node dysfunction







 Volume depletion






From Borkan SC: Extracorporeal therapies for acute intoxications. Crit Care Clin 18:393–420, 2002.

TSH, thyroid stimulating hormone; ECG, echocardiogram.





Potential effects of chronic lithium therapy, such as nephrogenic diabetes insipidus, hyperparathyroidism, chronic interstitial nephritis (controversial), and thyroid disorders are beyond the scope of this chapter.

Diagnosis and Treatment

The best approach to lithium toxicity is prevention. Given its narrow therapeutic index, plasma levels should be monitored any time an adjustment in dose or a change in renal function occurs and periodically thereafter. The initial management of acute lithium intoxication involves supportive care, prevention of further absorption, and enhancement of lithium elimination. A serum lithium level, electrolytes, and assessment of renal function should be obtained early. A nasogastric tube may be placed for gastric lavage although activated charcoal should not be used because it does not bind lithium ions. Whole bowel irrigation with polyethylene glycol solution has also been reported to remove lithium from the GI tract preventing further absorption. [60] [61] Another initial step in lithium intoxication is volume resuscitation using half-normal saline. Administration of hypotonic fluid is especially important for patients with underlying lithium-induced diabetes insipidus because normal saline may lead to hypernatremia.

Once the patient is hemodynamically stable, treatment should focus on enhancing lithium elimination from the body. If renal function is preserved, the clearance of lithium by the kidneys is 10 to 40 mL/min. Hemodialysis, the modality of choice when extracorporeal elimination is required, removes lithium at a rate of 70 to 170 mL/min, or approximately 1 mEq/L per 4-hour treatment with a low-flux dialyzer. [57] [61] Because lithium primarily distributes to the intracellular space, serum lithium levels often rebound after hemodialysis. Therefore, the dialysis catheter should be left in place after the initial therapy, and levels should be checked frequently. Prolonged monitoring is also important when intoxication occurs with a slow-release preparation.

Indications for dialysis include coma, seizure, respiratory failure, or worsening mental status. In addition, dialysis must be utilized if the patient develops AKI, as lithium is nearly entirely excreted by the kidney. In some instances, lithium levels will fail to decrease at an acceptable rate despite normal kidney function. Whether due to continued gastrointestinal absorption or redistribution, hemodialysis may be necessary in these cases as well. Extracorporeal elimination should also be considered for any patient with a lithium level more than 6 mEq/L, those on chronic lithium therapy with a level greater than 4 mEq/L, or for patients who develop cardiac or neurologic symptoms with a level between 2.5 and 4.0 mEq/L.[57] If endogenous renal clearance is not significantly impaired, dialysis is rarely indicated for peak serum lithium levels lower than 2.5 mEq/L.

Extending dialysis therapy to 8 to 12 hours may prevent rebound of lithium levels. [9] [63] If shorter sessions are performed, at least two treatments are typically necessary due to redistribution. Continuous therapies (e.g., CVVHD) do not reduce lithium levels as quickly as hemodialysis and are limited by the need for anticoagulation but will clear 60 to 85 L/d. There are a few reported cases that describe gradual and com-plete removal of lithium by CRRT without rebound of serum lithium concentration. Continuous therapies may be most useful for patients with chronic poisoning who may be at higher risk for permanent neurologic morbidity due to more substantial intracellular lithium accumulation, but this has not been rigorously tested.[21]


Salicylates are widely used as analgesics and anti-inflammatory medications, and aspirin (acetylsalicylic acid) is a widely prescribed anti-platelet therapy for the prevention of cardiovascular disease. Salicylates are also found in over-the-counter cold preparations, Pepto-Bismol (bismuth subsalicylate), wart removers (salicylic acid), sunscreens, musculoskeletal pain ointments, and Oil of Wintergreen (methyl salicylate), which may contain as much as 7 grams of salicylate per teaspoon.[64]


Absorption of salicylates depends on the route and formulation of the product. Toxicity resulting from an acute ingestion of nonbuffered aspirin may be apparent within 1 hour unless pylorospasm inhibits gastric emptying or concretions of tablets form in the stomach.[65] Enteric-coated products may have delayed absorption with one reported case demonstrating the onset of symptoms 35 hours after toxic ingestion.[66] Chronic use of topical salicylates also has the potential to cause toxicity, especially in children and adults with abnormal skin.[67]

Salicylate is 90% protein bound at therapeutic levels and has a Vd of 0.2 L/kg, indicating it is primarily distributed to the vascular space. With salicylate overdose, protein binding falls to 50% when the serum level reaches 80 mg/dL, which can exacerbate toxicity and significantly impact metabolism and drug elimination.[68] In addition, acidemia leads to a greater fraction of non-ionized salicylate, which has greater ability to cross the blood-brain barrier and into other tissues.

Hepatic metabolism predominates at therapeutic doses. Aspirin and salicylate salts are rapidly hydrolyzed to salicylic acid and subsequently oxidized or conjugated to glucuronic acid or glycine. Normally, less than 10% of salicylate is excreted unchanged by the kidney. With salicylate intoxication, hepatic conjugation pathways become saturated and protein binding decreases, leading to an even larger amount of unbound circulating salicylic acid. In this situation, the slow renal excretion of un-metabolized salicylic acid becomes more important clinically.[69] The elimination half-life following a therapeutic dose ranges between 2.0 and 4.5 hours, but can be extended to 18 to 36 hours following overdose.[59]

Salicylic acid is filtered at the glomerulus, actively secreted in the proximal tubule, and reabsorbed passively in the distal tubules. Elimination is influenced by urinary pH because the pKa of salicylic acid is 3.0 and ionization in the urinary space favors excretion over uptake into the renal tubular cells. This provides the rationale for urinary alkalinization to enhance elimination in cases of salicylate overdose.

Mechanism of Toxicity

A variety of acid-base abnormalities may occur with salicylate intoxication, but the classic finding is a mixed respira-tory alkalosis and high anion-gap metabolic acidosis. In one series, approximately 56% of adults presented with this pattern, 22% presented with a pure respiratory alkalosis, and 2% had a respiratory and metabolic acidosis (many of whom co-ingested a sedative).[65] Salicylate stimulates the medullary respiratory center independent of the aortic and carotid chemoreceptors, leading to an early fall in pCO2 and respiratory alkalosis. Several factors contribute to the subsequent metabolic acidosis. The primary respiratory alkalosis itself appears to promote lactic acid production based on experimental animal studies that demonstrated that without the initial fall in pCO2, lactate production is minimal. In addition, salicylates uncouple mitochondrial oxidative phosphorylation and interrupt glucose and fatty acid metabolism in the Krebs cycle. This uncoupling of oxidative phosphorylation combined with increased skeletal muscle demand for glucose leads to an increase in the production of tissue carbon dioxide, lactic acid, and ketoacids. [59] [65] This results in an anion-gap metabolic acidosis and indirect stimulation of the respiratory center. The salicylate anion itself has only a minor effect on serum pH; for example, at a plasma concentration of 60 mg/dL, salicylate contributes 3.5 mEq/L to the anion gap.

As mentioned previously, the systemic acidosis leads to an increased concentration of the non-ionized form of salicylic acid, which can easily move across cell membranes and the blood-brain barrier. Salicylic acid can produce altered mental status directly or through the selective reduction of brain glucose concentration. Cerebral edema may also play a role, perhaps secondary to capillary leak.[65]

Increased pulmonary capillary permeability may also occur in the setting of normal capillary wedge pressure. This non-cardiogenic pulmonary edema is the most common cause of major morbidity in chronic salicylate intoxication and complicates treatment. Because volume resuscitation and the administration of sodium bicarbonate are two arms of therapy, the presence of pulmonary edema is an absolute indication for hemodialysis. This complication is most common in the elderly, chronic salicylate users, and smokers.[65]

Clinical Presentation

Two distinct syndromes of salicylate intoxication may occur, depending on whether the exposure is acute or chronic ( Table 62-5 ). Following acute ingestion, nausea and vomiting frequently occur due to drug-induced gastritis and direct stimulation of the medullary chemoreceptor trigger zone. Tinnitus, tachypnea, sweating, and lethargy follow. With severe intoxication, coma, seizures, hypoglycemia, hyperthermia, oliguric AKI thought secondary to prostaglandin synthesis inhibition, non-cardiogenic pulmonary edema, and death may occur. [59] [65] [70]

TABLE 62-5   -- Clinical Effects of Acute and Chronic Salicylate Toxicity





Ill with mild to moderate dehydration

More severely ill with moderate to severe dehydration


Tinnitus, decreased hearing

Same as acute



Tachypnea, pulmonary edema


None initially, then develops manifestations of chronic toxicity

Agitation, confusion, mental status depression, seizures, coma


Nausea, vomiting

Nausea, vomiting, gastritis



Increased transaminases


Increase PT, platelet dysfunction

Same as acute



Acute renal failure

From Yip L, Dart RC, Gabow PA: Concepts and controversies in salicylate toxicity. Emerg Med Clin North Am 12:351–364, 1994.

CNS, central nervous system; GI, gastrointestinal.





Victims of chronic intoxication are usually young children or the delirious/demented elderly that present with nonspecific symptoms including confusion, dehydration, and metabolic acidosis. Chronic salicylism has been misdiagnosed as many diseases, including myocardial ischemia, sepsis, delirium, alcohol withdrawal, and psychosis. Therefore, the astute clinician must consider salicylate toxicity in the differential diagnosis, especially because severe poisoning typically occurs at a lower plasma concentration than observed with acute intoxication, and morbidity and mortality rates are higher. Cerebral and pulmonary edema are more commonly observed in this population than with acute ingestions. [59] [65]

Diagnosis and Laboratory Data

The diagnosis of salicylate intoxication is usually suspected from the history and physical examination as described earlier. The presence of urinary salicylate is rapidly detected by the presence of a purple color when ferric chloride is added to urine.[71] However, taken alone, urinary salicylate is very nonspecific in the adult population due to the prevalence of therapeutic salicylate ingestion (e.g., daily aspirin therapy for cardiovascular protection). With modern instrumentation, a quantitative serum salicylate level may be obtained in less than 15 minutes in centers with a toxicology laboratory. If these tests are not immediately available, other supportive laboratory values that may support, but not confirm, the clinical diagnosis of salicylate intoxication include a mixed respiratory alkalosis-metabolic acidosis, hypokalemia, hypernatremia due to increased insensible water losses, glucose derangements (hyperglycemia early, hypoglycemia late), and hypouricemia.

The Done nomogram, which attempted to correlate salicylate levels with toxicity, is no longer in clinical use due to its poor predictive value.[72] Generally, however, patients with higher salicylate levels tend to present with more severe symptoms. The level should be determined on admission provided that more than 4 hours have elapsed from the time of ingestion to presentation. Measurements made in the first 4 hours are difficult to interpret due to the unknown quantity of unabsorbed salicylate in the gastrointestinal tract.[70] Ingestions of enteric-coated preparations require even longer times to reveal the true extent of intoxication. Roughly 6 hours after ingestion, mild toxicity is observed with salicylate levels of 30 to 50 mg/dL, moderate toxicity with 50 to 70 mg/dL, and severe toxicity with levels greater or equal to 750 mg/L.[70] The magnitude of the level is less important in those who present with significant symptoms since treatment will be initiated regardless. In these cases, the salicylate level is most useful to follow the effectiveness and determine the duration of therapy.


The treatment of salicylate toxicity is variable and depends on the severity of exposure. During initial stabilization, intubation should be avoided unless hypoventilation is confirmed by an arterial blood gas. Intubation for impending respiratory fatigue without objective evidence of hypoventilation has resulted in death, likely due to an increased fraction of non-ionized salicylic acid when respiratory alkalosis was not maintained. If intubation is necessary (e.g., respiratory arrest), it is essential to hyperventilate the patient in an attempt to maintain alkalemia.[73]

The use of gastric lavage is controversial but may be of benefit if the dose was taken less than 1 hour before presentation.[70] Use of multiple-dose activated charcoal (MDAC) may further enhance elimination, but the evidence supporting this is weak. (Volunteers took 25 g of MDAC at 4, 6, 8, and 10 hours following 2880 mg aspirin ingestion with only modest effect.[74]) Nevertheless, many advocate initially 50 g (or 1 g/kg in children) of MDAC followed by repeat doses of 25 to 50 g every 3 to 5 hours if repeat salicylate levels continue to trend upward. [59] [75] [76]

Careful assessment of fluid and electrolyte balance is critical, and volume deficits should be replaced. Intravenous fluids should contain glucose (empiric), because of the derangements in glucose metabolism that occur with salicylate intoxication. In those with altered mental status, some recommend administering supplemental glucose (e.g., 50 to 100 mL of 50% dextrose) even if the serum glucose is normal.[65] Because systemic and urinary alkalinization is one of the initial goals of therapy, intravenous fluids should contain sodium bicarbonate (e.g., 100 to 150 mEq sodium bicarbonate in 1 L of 5% dextrose in water). If the patient is hypokalemic, either on presentation or as a consequence of alkalinizing therapy, potassium may be added once urine output is established.

Urine flow rate has a much smaller effect on salicylate clearance than alkalinization. Salicylate elimination is directly proportional to urine flow rate but logarithmically proportional to urine pH.[5] In one small series, the percentage of the dose excreted unchanged increased from 2% under acidic conditions to 30% under alkaline urine conditions.[77] The indications for urine alkalinization are not established by controlled trials, but expert opinion suggests that plasma levels of 30 to 60 mg/dL or any signs of salicylism warrants therapy with a goal urine pH greater than 7.5. [5] [65] [70] Alkalemia is not a contraindication to sodium bicarbonate administration because patients typically have a significant base deficit despite their arterial pH.[59] Acetazolamide also effectively alkalinizes the urine, but this therapy is absolutely contraindicated because it lowers arterial pH and promotes salicylate movement into the CNS and other tissues.[68] Urinary alkalinization should be continued until the plasma salicylate concentration is in the therapeutic range and symptoms have resolved.[70]

Administration of large volumes of fluid to achieve alkalinization and maintain urinary flow rates may precipitate pulmonary edema, especially in the elderly, those with heart disease, and chronic users of salicylates. Typically, the pulmonary edema associated with salicylate toxicity is due to capillary leak, and should be managed as other cases of acute lung injury or ARDS. Because of the need to maintain hyperventilation and alkalemia, hemodialysis to control volume overload is preferred to artificial ventilation.

Hemodialysis is the best extracorporeal therapy in salicylate poisoning because it can aid in the treatment of electrolyte and acid-base abnormalities while removing salicylate. Although no controlled trial comparing hemodialysis to carefully managed urinary alkalinization has been performed, the general consensus for the initiation of hemodialysis includes clinical deterioration despite supportive care, seizures, persistent altered mental status, pulmonary edema, persistent acidemia, coagulopathy, renal or hepatic dysfunction, and extremes of age. [65] [70] Salicylate levels alone are not as helpful in deciding whether to initiate hemodialysis, but dialysis should be strongly considered if the plasma salicylate concentration is greater than 80 mg/dL in adults and 70 mg/dL in children. Urinary alkalinization should be continued during hemodialysis to maximize drug elimination. [68] [70]

Continuous venovenous hemodiafiltration has also been reported to treat salicylate overdose, demonstrating a 77% reduction in plasma concentration with one 11-hour treatment without concomitant urinary alkalinization.[78]Similarly, sustained low-efficiency dialysis (SLED) has been reported to manage salicylate overdose with a relative clearance of SLED to hemodialysis of 84% based on half-life estimates. Because SLED can be performed with many standard dialyzers and has decreased nursing requirements, this is a promising therapy that deserves more study in the treatment of intoxications.[29]


Theophylline is an oral methylxanthine bronchodilator that is FDA-approved for use in asthma and chronic obstructive pulmonary disease. Although its clinical use has declined over the past decade in favor of inhaled corticosteroids and β-adrenergic agonists, it continues to be used in some patients with severe disease or those resistant to more standard therapy.


Oral theophylline (MW 180.17, anhydrous) is rapidly absorbed from the upper gastrointestinal tract although slow-release products (e.g., Theodur®) are most commonly used. Generally 40% protein bound, protein binding may be reduced in elderly patients and variable in those with chronic obstructive pulmonary disease. Reduction of protein binding leads to an increase in Vd from its typical value of 0.45 L/kg.[79]

Theophylline is extensively metabolized by the liver by several hepatic cytochrome P450 isozymes, of which the most important seems to be CYP1A2 in adults. Metabolism may become saturated within the therapeutic range, which may result in large increases in serum concentration after comparatively small dose increases. Given that the therapeutic index is very narrow (10 to 20 mg/L) with some patients exhibiting symptoms of toxicity even within this range, following serum theophylline levels is extremely important, especially when dose adjustments are made.[79]

Because of the extensive hepatic metabolism, the adult kidney only excretes 10% of ingested theophylline. The endogenous clearance is less than 4 mL/min/kg with an elimination half-life of 6 to 12 hours in adults. Drug elimination can be substantially affected if co-administered with other medications that enhance or slow hepatic metabolism. Clearance is decreased approximately 30% in the elderly, 50% with hepatic dysfunction, and 50% with congestive heart failure. Conversely, smoking can increase clearance 50 to 80% by inducing metabolic pathways. This impact of smoking (tobacco or marijuana) on theophylline clearance is important, as even abstaining from smoking for one week may result in a 40% decrease in theophylline clearance, leading to supra-therapeutic levels. Interestingly, even passive exposure to smoke has been shown to increase theophylline clearance by up to 50%.[79]

Clinical Presentation

Theophylline intoxication may occur from acute ingestion or chronic use. Sustained-release preparations are responsible for greater morbidity and mortality.[68] Because theophylline is metabolized by the cytochrome P450 system, the clinician must be aware of potential drug-drug interactions when adding drugs to a patient's regimen that includes theophylline. For any given plasma theophylline level, chronic intoxication leads to a more severe clinical syndrome than acute intoxication.[80]

Mild theophylline intoxication is characterized by nervousness, tremors, tachycardia, abdominal pain, vomiting, and diarrhea. Patients with moderate intoxication may be lethargic, disoriented, and experience supraventricular tachycardia and frequent premature ventricular contractions. Severe toxicity may manifest as seizure activity (typically generalized although focal lip-smacking or ocular deviation may occur) associated with hyperthermia, hypotension, ventricular tachycardia, rhabdomyolysis, and acute kidney injury.[68]

The seizures and ventricular arrhythmias of theophylline intoxication generally do not occur until plasma theophylline levels exceed 90 to 100 mg/L in acute overdose. With chronic intoxication, severe toxicity may occur with levels of 40 to 60 mg/L. Seizures may first occur 12 to 16 hours following acute ingestion, in part due to delayed absorption of sustained-release preparations. Therefore, levels should be repeated every 2 to 4 hours after acute oral overdose once the diagnosis is made.[59]

Diagnosis and Laboratory Data

The diagnosis of theophylline intoxication is made by measurement of the serum theophylline concentration in the setting of suggestive clinical findings. Caffeine, also a methylxanthine, may cause a similar clinical picture in cases of acute overdose and will produce a falsely elevated theophylline concentration with most assays.[59] Although the recommended therapeutic theophylline plasma concentration is 10 to 20 mg/L, nearly one third of patients with concentrations about 15 mg/L may exhibit signs of mild toxicity. Therefore, a theophylline level in the “normal” range should not exclude theophylline toxicity from the differential diagnosis in the appropriate clinical setting. With plasma levels of 20 to 30 mg/L, approximately 67% of patients are likely to demonstrate symptoms, and with levels greater than 30 mg/L, more than 90% will be clinically intoxicated.

The most common electrolyte abnormality associated with acute theophylline toxicity is hypokalemia, perhaps secondary to transcellular potassium shifting due to increased β-adrenergic tone. Because of this mechanism, total body potassium stores may be normal despite hypokalemia on presentation. As the serum theophylline concentration is reduced, serum potassium levels return to normal. Other electrolyte/metabolic abnormalities may include hypomagnesemia, hypophosphatemia, hypercalcemia, metabolic acidosis, and hyperglycemia. These metabolic abnormalities are common with acute ingestion, but tend not to occur with chronic intoxication.[59]


The treatment of theophylline poisoning depends on the severity of intoxication. Following initial hemodynamic stabilization, multiple-dose activated charcoal should be used to decrease the serum elimination half-life to 1 to 3 hours, even if intoxication occurred by the intravenous route.[68] Activated charcoal (50 g or 1 g/kg) should be administered with a cathartic (e.g., sorbitol, magnesium citrate, magnesium sulfate) every 4 hours until the serum level is less than or equal to 20 mg/L. Alternatively, whole bowel irrigation with polyethylene glycol in a balanced electrolyte solution may be used, especially for large ingestions of sustained-release preparations.[59]

The hypotension and cardiac arrhythmias associated with theophylline are the result of excessive circulating catecholamines. Typically, the use of β-adrenergic antagonists (e.g., low-dose propranolol, metoprolol, or esmolol) slows the heart rate and controls the arrhythmia.[68] Although propranolol may be more effective in raising the blood pressure (due to antagonism of β2-mediated vasodilatation), one must use caution in the setting of bronchospasm.[59] If a patient is unable to tolerate the use of these agents, then verapamil or diltiazem may be used.

Seizures associated with theophylline toxicity are usually difficult to control. Diazepam should be used initially with phenobarbital added if necessary. Phenytoin may also be used, but appears to be least effective. For recurrent seizures unresponsive to standard therapy, general anesthesia and neuromuscular paralysis may be necessary.[68]

Extracorporeal therapy is utilized when the patient is clinically unstable with life-threatening cardiac arrhythmias or seizures. Although there is no evidence-based plasma theophylline level at which extracorporeal treatment should be initiated, a level ≥100 mg/L in acute intoxication or ≥60 mg/L in chronic poisoning is generally considered an indication for extracorporeal therapy. Elderly patients or those with cardiac or liver disease may require dialysis at lower levels (e.g., ≥40 mg/L).[68]

Due to its low molecular weight, small volume of distribution, and low endogenous clearance, theophylline elimination can be significantly enhanced by hemodialysis or hemoperfusion. Hemodialysis results in a mean extraction ratio of 36%; hemoperfusion may initially demonstrate an extraction ratio of nearly 100% due to its ability to also remove protein-bound theophylline but the extraction ratio then decreases to 60% as the cartridge becomes saturated. The disadvantages of hemoperfusion are that it does not correct the electrolyte abnormalities that often accompany theophylline poisoning, cartridges must be switched every 2 to 3 hours due to saturation, and therapy can lead to severe thrombocytopenia, hypoglycemia, and hypocalcemia. [68] [81] [82] Serial hemodialysis-hemoperfusion can circumvent these problems to some extent because the hemoperfusion cartridge will not saturate as quickly, but this modality is not available in many centers. Instead, hemodialysis with high-efficiency membranes and high blood flow rates can be used to achieve a theophylline clearance rate of more than 200 mL/min, which is equivalent to charcoal hemoperfusion. Continuous venovenous hemofiltration and hemodiafiltration have also been reported in the management of theophylline toxicity. [81] [83] Extracorporeal therapy should be continued until clinical improvement and a plasma level of less than 20 mg/L is obtained, monitoring for rebound (which may increase plasma levels 5-10 mg/L) thereafter.[79]

Valproic Acid

Valproic acid was introduced in the late 1970s as an anticonvulsant drug, and is currently FDA-approved for absence seizures, complex partial seizures, and migraine treatment although it is also used for bipolar affective disorders and schizophrenia. Acute valproic acid intoxication generally results in mild, self-limited CNS depression, but serious toxicity and death have been reported.[84]


Valproic acid (MW 144.21) is available in immediate- and sustained-release preparations that demonstrate 90% bioavailability compared to the parenteral form. Peak plasma concentrations are typically observed 1 to 13 hours following ingestion depending on the preparation.[85] Ingestion with food may slow absorption but total absorption remains the same. Therapeutic serum concentrations range from 50 to 100 mg/mL.[59]

Valproic acid is typically 90% protein-bound to albumin, but this is concentration dependent with decreased protein binding observed at high valproate concentrations. Protein binding decreases to 70% at plasma concentrations greater than 150 mg/mL and 35% when greater than 300 mg/mL.[86] In addition, patients with liver or renal disease, hyperlipidemia, HIV, or advanced age also demonstrate decreased protein binding resulting in increased free fractions of valproic acid. Concordant with the high protein binding is a small volume of distribution (Vd 0.13 to 0.22 L/kg).[59]

Valproic acid is rapidly metabolized by the liver. It undergoes glucuronic acid conjugation (70%) and beta- and omega-oxidation to nine different metabolites. While the cytochrome P450-mediated omega-oxidation is normally responsible for a small component of its metabolism, it generates metabolites that have been implicated in the adverse effects of valproic acid. Less than 3% of an administered dose is normally excreted unchanged in the urine.[86]The elimination half-life normally ranges from 5 to 20 hours (average 10.6 hours), but in overdose may be prolonged to 30 hours. [59] [85]

Clinical Presentation

Acute intoxication commonly causes gastrointestinal distress (nausea, vomiting, diarrhea), CNS abnormalities (confusion, obtundation, and coma with respiratory failure), hypotension, transaminitis, and tachycardia. Miotic pupils may lead to confusion with opiate poisoning, and increased seizure frequency may be observed in patients with a preexisting seizure disorder. Free and total valproic acid plasma concentrations poorly correlate with severity of toxicity, but most patients with levels greater than 180 mg/mL usually develop some degree of CNS depression.[59]

Hyperammonemia may also occur, with or without concomitant hepatic dysfunction. Many patients remain asymptomatic although some present with hyperammonemic encephalopathy. These patients may be confused, lethargic, vomiting, and seizing. The degree of encephalopathy does not clearly correlate with serum valproic acid levels, and may occur with levels in the “normal” range.[87] The mechanism by which valproic acid, in the absence of hepatic dysfunction, leads to hyperammonemia is incompletely understood. One theory suggests that a valproic acid metabolite indirectly inhibits carbamoyl phosphate synthetase, the first enzyme of the urea cycle. A second theory involves valproic acid-induced carnitine deficiency leading to β-oxidation impairment and subsequent urea cycle inhibition.[87]

At very high serum levels (greater than 1000 mg/mL), electrolyte/metabolic abnormalities may include anion gap metabolic acidosis, hyperosmolality, hypernatremia, and hypocalcemia. Days after ingestion, other complications may arise including pancreatitis, optic nerve atrophy, cerebral edema, non-cardiogenic pulmonary edema, bone marrow suppression, and acute kidney injury.[59] Active metabolites with long half-lives may be responsible for prolonged CNS effects despite normalization of plasma valproic acid concentrations.

Diagnosis and Treatment

Diagnosis of valproic acid intoxication is based on history of exposure, CNS depression, typical metabolic disturbances, and confirmation with a serum valproic acid level. Serial valproic acid levels should be drawn, especially after ingestion of divalproex-containing preparations (Depakote® and the extended-release Depakote ER®). As mentioned earlier, the serum level correlates poorly with outcome with death occurring at peak levels ranging from 106 to 2728 mg/mL but survival reported in a patient with a peak level of 2120 mg/mL.[59]

Because of the metabolic derangements observed with valproic acid intoxication, serum electrolytes, glucose, creatinine, ammonia, liver transaminases, bilirubin, prothrombin time, amylase, serum osmolality (valproic acid levels greater than 1500 mg/mL may raise the osmolar gap by 10 mOsm/L or more), and arterial blood gas determination should also be performed. ECG monitoring for arrhythmias and prolonged QT interval is recommended.[59]

Treatment consists of initial stabilization of respiratory and cardiovascular function. Vomiting should not be induced, but activated charcoal should be administered with repeated 25 g to 50 g quantities over the first 12 to 24 hours in cases of large ingestions reducing the half-life from a mean of 12 hours to 4.8 hours.[88] If sustained-release products such as divalproex have been ingested, whole bowel irrigation may be considered if serum levels continue to rise despite multiple-dose activated charcoal.[59]

Naloxone has been reported to reverse CNS depression in some case reports, but others do not suggest benefit. It appears to be of greatest success in patients with serum valproic acid levels of approximately 190 mg/mL, but given its low risk, the benefit/risk ratio favors therapeutic trial.[88] Those with hyperammonemia, encephalopathy, and/or hepatotoxicity may respond to L-carnitine, with one retrospective study of severe valproate-induced hepatotoxicity associating a marked survival advantage to those who received intravenous carnitine therapy.[89] Carnitine should be administered for 3 to 4 days in doses of 100 mg/kg/d to 2 g/d divided in three oral doses, or 150 to 500 mg/kg/d divided in three intravenous doses.

Because the excretion of valproic acid by the kidney is minor, forced diuresis is ineffective in enhancing drug elimination. The low molecular weight and Vd of valproic acid favor dialytic clearance of unbound drug, but hemodialysis has little impact on the elimination of valproic acid at usual plasma concentrations due to its extensive protein binding. At supratherapeutic drug levels plasma proteins become saturated, and the fraction of unbound drug increases substantially. Because the free drug and metabolites are accessible to the dialysis membrane, hemodialysis leads to a four- to ten-fold decrease in elimination half-life in overdose patients.[59] In cases of severe symptomatic hyperammonemia, hemodialysis has been shown to clear ammonia approximately ten-fold greater than any other method.[85] In a multicenter case series, patients with peak valproic acid concentrations above 850 mg/mL were more likely to develop coma, respiratory depression, or metabolic acidosis, and therefore many support the early initiation of hemodialysis in these cases.[85] Of note, case reports using high-flux hemodialysis to treat valproic acid intoxication have observed a rebound effect after 5 to 13 hours, which may necessitate re-initiation of therapy.[86] This suggests that the Vd of valproic acid may be higher than previously reported, and at overdose concentrations it may not observe single compartment kinetics.

Charcoal hemoperfusion has been reported in the treatment of valproic acid intoxication as well, but its efficacy is inconsistent among case reports.[84] Tandem hemodialysis-hemoperfusion may be more effective, but one group collected data on the two techniques separately and concluded that hemodialysis was effective in enhancing elimination of valproic acid whereas hemoperfusion was limited by adsorbent saturation and thrombocytopenia.[90]


1. Watson WA, Litovitz TL, Rodgers Jr GC, et al: 2004 Annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System.  Am J Emerg Med  2005; 23:589-666.

2. Goldfrank LR, Flomenbaum NE, Lewin NA, et al: Principles of managing the poisoned or overdosed patient.   In: Goldfrank LR, ed. Goldfrank's Toxicologic Emergencies,  7th ed. New York: McGraw-Hill; 2002.

3. Mokhlesi B, Leiken JB, Murray P, et al: Adult toxicology in critical care: Part I: General approach to the intoxicated patient.  Chest  2003; 123:577-592.

4. Shaver MJ, Marx MA, Golper TA: Extracorporeal treatment of poisoning.   In: Owen WF, Periera BJG, Sayegh MH, ed. Dialysis and Transplantation: A Companion to Brenner & Rector's The Kidney,  1st ed. Philadelphia: WB Saunders; 1999:437-453.

5. Proudfoot AT, Krenzelok EP, Vale JA: Position Paper on urine alkalinization.  J Toxicol Clin Toxicol  2004; 42:1-26.

6. Lam YW, Banerji S, Hatfield C, et al: Principles of drug administration in renal insufficiency.  Clin Pharmacokinet  1997; 32:30-57.

7. Matzke GR, Frye RF: Drug administration in patients with renal insufficiency. Minimising renal and extrarenal toxicity.  Drug Saf  1997; 16:205-231.

8. Pond SM: Extracorporeal techniques in the treatment of poisoned patients.  Med J Aust  1991; 154:617-622.

9. Hansen HE, Amdisen A: Lithium intoxication. (Report of 23 cases and review of 100 cases from the literature).  Q J Med  1978; 47:123-144.

10. Grimes DJ, Bowles MR, Buttsworth JA, et al: Survival after unexpected high serum methotrexate concentrations in a patient with osteogenic sarcoma.  Drug Saf  1990; 5:447-454.

11. Ujhelyi MR, Robert S: Pharmacokinetic aspects of digoxin-specific Fab therapy in the management of digitalis toxicity.  Clin Pharmacokinet  1995; 28:483-493.

12. Winchester JF: Dialysis and hemoperfusion in poisoning.  Adv Ren Replace Ther  2002; 9:26-30.

13. Jacobs C: Replacement of Renal Function by Dialysis,  4th ed. Dordrecht; Boston, Kluwer Academic, 1996.

14. Amir AR, Winchester JF: Hemodialysis and hemoperfusion in the management of drug intoxication.   In: Massry SG, Glassock RJ, ed. Massry & Glassock's Textbook of Nephrology,  4th ed. Philadelphia: Lippincott Williams and Wilkins; 2001:1729-1735.

15. Heath A, Knudsen K: Role of extracorporeal drug removal in acute theophylline poisoning. A review.  Med Toxicol Adverse Drug Exp  1987; 2:294-308.

16. Higgins RM, Hearing S, Goldsmith DJ, et al: Severe theophylline poisoning: charcoal haemoperfusion or haemodialysis?.  Postgrad Med J  1995; 71:224-226.

17. Sabeel AI, Kurkus J, Lindholm T: Intensive hemodialysis and hemoperfusion treatment of Amanita mushroom poisoning.  Mycopathologia  1995; 131:107-114.

18. De Broe ME, Verpooten GA, Christiaens MA, et al: Clinical experience with prolonged combined hemoperfusion-hemodialysis treatment of severe poisoning.  Artif Organs  1981; 5:59-66.

19. Frank RD, Kierdorf HP: Is there a role for hemoperfusion/hemodialysis as a treatment option in severe tricyclic antidepressant intoxication?.  Int J Artif Organs  2000; 23:618-623.

20. Nagler J, Braeckman RA, Willems JL, et al: Combined hemoperfusion-hemodialysis in organophosphate poisoning.  J Appl Toxicol  1981; 1:199-201.

21. Leblanc M, Raymond M, Bonnardeaux A, et al: Lithium poisoning treated by high-performance continuous arteriovenous and venovenous hemodiafiltration.  Am J Kidney Dis  1996; 27:365-372.

22. Berbece AN, Richardson RM: Sustained low-efficiency dialysis in the ICU: Cost, anticoagulation, and solute removal.  Kidney Int  2006; 70:963-968.

23. Evanson JA, Himmelfarb J, Wingard R, et al: Prescribed versus delivered dialysis in acute renal failure patients.  Am J Kidney Dis  1998; 32:731-738.

24. O'Reilly P, Tolwani A: Renal replacement therapy III: IHD, CRRT, SLED.  Crit Care Clin  2005; 21:367-378.

25. Kielstein JT, Kretschmer U, Ernst T, et al: Efficacy and cardiovascular tolerability of extended dialysis in critically ill patients: A randomized controlled study.  Am J Kidney Dis  2004; 43:342-349.

26. Kumar VA, Craig M, Depner TA, et al: Extended daily dialysis: A new approach to renal replacement for acute renal failure in the intensive care unit.  Am J Kidney Dis  2000; 36:294-300.

27. Marshall MR, Golper TA, Shaver MJ, et al: Sustained low-efficiency dialysis for critically ill patients requiring renal replacement therapy.  Kidney Int  2001; 60:777-785.

28. Marshall MR, Golper TA, Shaver MJ, et al: Urea kinetics during sustained low-efficiency dialysis in critically ill patients requiring renal replacement therapy.  Am J Kidney Dis  2002; 39:556-570.

29. Lund B, Seifert SA, Mayersohn M: Efficacy of sustained low-efficiency dialysis in the treatment of salicylate toxicity.  Nephrol Dial Transplant  2005; 20:1483-1484.

30. Exaire E, Trevino-Becerra A, Monteon F: An overview of treatment with peritoneal dialysis in drug poisoning.  Contrib Nephrol  1979; 17:39-43.

31. Winchester JF: Active methods for detoxification.   In: Haddad LM, Shannon MW, Winchester JF, ed. Clinical Management of Poisoning and Drug Overdose,  3rd ed. Philadelphia: WB Saunders; 1998:175-188.

32. Choi JH, Oh JC, Kim KH, et al: Successful treatment of cisplatin overdose with plasma exchange.  Yonsei Med J  2002; 43:128-132.

33. Jones JS, Dougherty J: Current status of plasmapheresis in toxicology.  Ann Emerg Med  1986; 15:474-482.

34. Meert KL, Ellis J, Aronow R, et al: Acute ammonium dichromate poisoning.  Ann Emerg Med  1994; 24:748-750.

35. Barceloux DG, Krenzelok EP, Olson K, et al: American Academy of Clinical Toxicology Practice Guidelines on the Treatment of Ethylene Glycol Poisoning. Ad Hoc Committee.  J Toxicol Clin Toxicol  1999; 37:537-560.

36. Litovitz TL, Klein-Schwartz W, Rodgers Jr GC, et al: 2001 Annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System.  Am J Emerg Med  2002; 20:391-452.

37. Winchester JF: Methanol, isopropyl alcohol, higher alcohols, ethylene glycol, cellosolves, acetone, and oxalate.   In: Haddad LM, Shannon MW, Winchester JF, ed. Clinical Management of Poisoning and Drug Overdose,  3rd ed. Philadelphia: WB Saunders; 1998:491-504.

38. Glaser DS: Utility of the serum osmol gap in the diagnosis of methanol or ethylene glycol ingestion.  Ann Emerg Med  1996; 27:343-346.

39. Church AS, Witting MD: Laboratory testing in ethanol, methanol, ethylene glycol, and isopropanol toxicities.  J Emerg Med  1997; 15:687-692.

40. Abramson S, Singh AK: Treatment of the alcohol intoxications: Ethylene glycol, methanol and isopropanol.  Curr Opin Nephrol Hypertens  2000; 9:695-701.

41. Saladino R, Shannon M: Accidental and intentional poisonings with ethylene glycol in infancy: Diagnostic clues and management.  Pediatr Emerg Care  1991; 7:93-96.

42. Winter ML, Ellis MD, Snodgrass WR: Urine fluorescence using a Wood's lamp to detect the antifreeze additive sodium fluorescein: A qualitative adjunctive test in suspected ethylene glycol ingestions.  Ann Emerg Med  1990; 19:663-667.

43. Wallace KL, Suchard JR, Curry SC, et al: Diagnostic use of physicians' detection of urine fluorescence in a simulated ingestion of sodium fluorescein-containing antifreeze.  Ann Emerg Med  2001; 38:49-54.

44. Nzerue CM, Harvey P, Volcy J, et al: Survival after massive ethylene glycol poisoning: Role of an ethanol enriched, bicarbonate-based dialysate.  Int J Artif Organs  1999; 22:744-746.

45. Chu J, Wang RY, Hill NS: Update in clinical toxicology.  Am J Respir Crit Care Med  2002; 166:9-15.

46. Antizol (fomepizole) injection product monograph,  Minnesota, MN, Orphan Medical, Inc., 2000.

47. Jobard E, Harry P, Turcant A, et al: 4-Methylpyrazole and hemodialysis in ethylene glycol poisoning.  J Toxicol Clin Toxicol  1996; 34:373-377.

48. Megarbane B, Borron SW, Baud FJ: Current recommendations for treatment of severe toxic alcohol poisonings.  Intensive Care Med  2005; 31:189-195.

49. Borron SW, Megarbane B, Baud FJ: Fomepizole in treatment of uncomplicated ethylene glycol poisoning.  Lancet  1999; 354:831.

50. Hantson P, Hassoun A, Mahieu P: Ethylene glycol poisoning treated by intravenous 4-methylpyrazole.  Intensive Care Med  1998; 24:736-739.

51. Barceloux DG, Bond GR, Krenzelok EP, et al: American Academy of Clinical Toxicology practice guidelines on the treatment of methanol poisoning.  J Toxicol Clin Toxicol  2002; 40:415-446.

52. Liesivuori J, Savolainen H: Methanol and formic acid toxicity: Biochemical mechanisms.  Pharmacol Toxicol  1991; 69:157-163.

53. Sullivan-Mee M, Solis K: Methanol-induced vision loss.  J Am Optom Assoc  1998; 69:57-65.

54. Wadgymar A, Wu GG: Treatment of acute methanol intoxication with hemodialysis.  Am J Kidney Dis  1998; 31:897.

55. Garella S: Extracorporeal techniques in the treatment of exogenous intoxications.  Kidney Int  1988; 33:735-754.

56. Burkhart KK, Martinez MA: The adsorption of isopropanol and acetone by activated charcoal.  J Toxicol Clin Toxicol  1992; 30:371-375.

57. Timmer RT, Sands JM: Lithium intoxication.  J Am Soc Nephrol  1999; 10:666-674.

58. Finley PR, Warner MD, Peabody CA: Clinical relevance of drug interactions with lithium.  Clin Pharmacokinet  1995; 29:172-191.

59. Olson KR: San Francisco Bay Area Regional Poison Control Center: Poisoning & drug overdose,  4th ed. New York, McGraw-Hill, 2004.

60. Ellenhorn MJ, Schonwald S, Ordog G, et al: Lithium.   In: Ellenhorn MJ, Schonwald S, Ordog G, Wasserberger J, ed. Medical Toxicology: Diagnosis and Treatment of Human Poisoning,  Baltimore: Williams and Wilkins; 1997:1579.

61. Okusa MD, Crystal LJ: Clinical manifestations and management of acute lithium intoxication.  Am J Med  1994; 97:383-389.

62. Greger R: Possible sites of lithium transport in the nephron.  Kidney Int Suppl  1990; 28:S26-S30.

63. Jacobsen D, Aasen G, Frederichsen P, et al: Lithium intoxication: Pharmacokinetics during and after terminated hemodialysis in acute intoxications.  J Toxicol Clin Toxicol  1987; 25:81-94.

64. Botma M, Colquhoun-Flannery W, Leighton S: Laryngeal oedema caused by accidental ingestion of Oil of Wintergreen.  Int J Pediatr Otorhinolaryngol  2001; 58:229-232.

65. Yip L, Dart RC, Gabow PA: Concepts and controversies in salicylate toxicity.  Emerg Med Clin North Am  1994; 12:351-364.

66. Rivera W, Kleinschmidt KC, Velez LI, et al: Delayed salicylate toxicity at 35 hours without early manifestations following a single salicylate ingestion.  Ann Pharmacother  2004; 38:1186-1188.

67. Brubacher JR, Hoffman RS: Salicylism from topical salicylates: Review of the literature.  J Toxicol Clin Toxicol  1996; 34:431-436.

68. Borkan SC: Extracorporeal therapies for acute intoxications.  Crit Care Clin  2002; 18:393-420.

69. Patel DK, Hesse A, Ogunbona A, et al: Metabolism of aspirin after therapeutic and toxic doses.  Hum Exp Toxicol  1990; 9:131-136.

70. Dargan PI, Wallace CI, Jones AL: An evidence based flowchart to guide the management of acute salicylate (aspirin) overdose.  Emerg Med J  2002; 19:206-209.

71. Hoffman RJ, Nelson LS, Hoffman RS: Use of ferric chloride to identify salicylate-containing poisons.  J Toxicol Clin Toxicol  2002; 40:547-549.

72. Dugandzic RM, Tierney MG, Dickinson GE, et al: Evaluation of the validity of the Done nomogram in the management of acute salicylate intoxication.  Ann Emerg Med  1989; 18:1186-1190.

73. Greenberg MI, Hendrickson RG, Hofman M: Deleterious effects of endotracheal intubation in salicylate poisoning.  Ann Emerg Med  2003; 41:583-584.

74. Kirshenbaum LA, Mathews SC, Sitar DS, et al: Does multiple-dose charcoal therapy enhance salicylate excretion?.  Arch Intern Med  1990; 150:1281-1283.

75. Position statement and practice guidelines on the use of multi-dose activated charcoal in the treatment of acute poisoning. American Academy of Clinical Toxicology; European Association of Poisons Centres and Clinical Toxicologists.  J Toxicol Clin Toxicol  1999; 37:731-751.

76. Barone JA, Raia JJ, Huang YC: Evaluation of the effects of multiple-dose activated charcoal on the absorption of orally administered salicylate in a simulated toxic ingestion model.  Ann Emerg Med  1988; 17:34-37.

77. Vree TB, Van Ewijk-Beneken Kolmer EW, Verwey-Van Wissen CP, et al: Effect of urinary pH on the pharmacokinetics of salicylic acid, with its glycine and glucuronide conjugates in human.  Int J Clin Pharmacol Ther  1994; 32:550-558.

78. Wrathall G, Sinclair R, Moore A, et al: Three case reports of the use of haemodiafiltration in the treatment of salicylate overdose.  Hum Exp Toxicol  2001; 20:491-495.

79. Hendeles L, Jenkins J, Temple R: Revised FDA labeling guideline for theophylline oral dosage forms.  Pharmacotherapy  1995; 15:409-427.

80. Shannon M: Predictors of major toxicity after theophylline overdose.  Ann Intern Med  1993; 119:1161-1167.

81. Okada S, Teramoto S, Matsuoka R: Recovery from theophylline toxicity by continuous hemodialysis with filtration.  Ann Intern Med  2000; 133:922.

82. Shannon MW: Comparative efficacy of hemodialysis and hemoperfusion in severe theophylline intoxication.  Acad Emerg Med  1997; 4:674-678.

83. Henderson JH, McKenzie CA, Hilton PJ, et al: Continuous venovenous haemofiltration for the treatment of theophylline toxicity.  Thorax  2001; 56:242-243.

84. Hicks LK, McFarlane PA: Valproic acid overdose and haemodialysis.  Nephrol Dial Transplant  2001; 16:1483-1486.

85. Eyer F, Felgenhauer N, Gempel K, et al: Acute valproate poisoning: Pharmacokinetics, alteration in fatty acid metabolism, and changes during therapy.  J Clin Psychopharmacol  2005; 25:376-380.

86. Kane SL, Constantiner M, Staubus AE, et al: High-flux hemodialysis without hemoperfusion is effective in acute valproic acid overdose.  Ann Pharmacother  2000; 34:1146-1151.

87. McCall M, Bourgeois JA: Valproic acid-induced hyperammonemia: A case report.  J Clin Psychopharmacol  2004; 24:521-526.

88. Goldfrank LR: Goldfrank's Toxicologic Emergencies,  6th ed. Stamford, CT, Appleton & Lange, 1998.

89. Bohan TP, Helton E, McDonald I, et al: Effect of L-carnitine treatment for valproate-induced hepatotoxicity.  Neurology  2001; 56:1405-1409.

90. Franssen EJ, van Essen GG, Portman AT, et al: Valproic acid toxicokinetics: Serial hemodialysis and hemoperfusion.  Ther Drug Monit  1999; 21:289-292.