All ECTRs will remove poisons with a molecular weight (MW) below 1000 Da, but diffusion and convection will remove them more efficiently. Solutes larger than 1000 Da will be better removed by convection, adsorption, or centrifugation, although modern intermittent hemodialysis (IHD) can clear poisons up to 10,000 Da ( Table 66.4 ). Convection from hemodiafiltration (HDF) and continuous renal replacement therapy (CRRT) can remove poisons with a MW up to 50,000 Da. Adsorption or centrifugation can remove poisons with a MW exceeding 100,000 Da.

The degree of protein binding impacts removal. Hemofiltration and hemodialysis only remove unbound poison, as the poison-protein complex exceeds the pore size of the hemofilter/dialyzer. Diffusion and convection remove poisons with a protein binding up to 80%, with a few exceptions. Hemoperfusion can remove poisons with a protein binding up to 90% to 95%, as binding to the adsorbent (activated carbon or, less commonly, a resin) competes with binding to plasma proteins. The clearance depends on the affinity of the poison for the adsorbent.

The degree of protein binding is influenced by the protein concentration, poison concentration, and pathologic states (e.g., uremia). In poisonings, saturation of protein binding sites (e.g., valproic acid and salicylate) may occur, increasing the free (unbound) fraction, which promotes removal by ECTR. Phenytoin quickly dissociates from albumin, so the unbound pool is constantly being removed by ECTR.

Extracorporeal removal will be useless if the endogenous clearance of a poison far outweighs its ECTR clearance. This explains why hemodialysis is futile for cocaine or toluene, which have high endogenous clearances (>4 mL/minute/kg). Additionally, significant impairment in endogenous metabolic and/or elimination pathways may increase the efficacy of ECTRs; for example, the renal clearance of metformin decreases from 500 mL/min to 0 mL/min in anuria, rendering ECTR clearance clinically significant.

The V _D is the apparent volume where a poison is distributed at equilibrium before clearance occurs (e.g., with rapid intravenous administration) and is calculated by dividing the total poison in the body by its concentration. This relationship assumes a single homogenous compartment of water in which the poison is distributed.

Poisons that distribute extensively in tissues (e.g., tricyclic antidepressants) have a high V _D ; conversely, poisons that distribute in total body water, such as methanol, have a low V _D (<1 L/Kg). Because ECTRs only remove poisons in the intravascular space, they will not significantly eliminate poisons with a high V _D . For example, digoxin easily crosses the dialyzer; however, because of its high V _D (7 L/kg), <5% of the total body burden will be removed during a 6-hour IHD. Many publications erroneously conclude that poisons with a high V _D are amenable to ECTR removal on the basis of a high extracorporeal clearance or reduction of serum concentrations. ECTR may occasionally be considered for poisonings with a high V _D (e.g., methotrexate and thallium) early after exposure and before distribution to tissues, when there is still a high burden of poison in the blood.

During IHD, the poison diffuses from the blood compartment to a dialysate flowing in a countercurrent direction, separated by a semipermeable dialysis membrane. The principles that dictate solute removal in IHD also apply to poison elimination (see Chapter 62 ). The characteristics of a xenobiotic that favor efficient removal by IHD are low endogenous clearance (<4 mL/min/kg), low V _D (<1–2 L/kg), low MW (<5000 _– 10,000 Da), and low protein binding (<80%).

The type of dialysis membrane, its surface area, pore size, and the blood and dialysate flow rates will influence poison clearance ( Table 66.5 ). New high-flux membranes and catheters allow the removal of larger-sized poisons considered nondialyzable 30 years ago. Increasing the dialysate flow rate and, more importantly, the blood flow rate will promote diffusion and elimination of the poison. Filters with large surfaces and maximal blood and dialysate flow rates should be used unless contraindicated (e.g., concern for disequilibrium syndrome).

IHD has several advantages over other ECTRs in poisonings because it rapidly removes poisons and corrects volume overload, acid-base, and electrolyte abnormalities. IHD is also more available, less expensive, and quicker to implement.

The most common acute complication of IHD is hypotension, which is seen in patients with acute kidney injury (AKI) or end-stage kidney failure (ESKF) requiring fluid removal. Because poisoned patients rarely require net ultrafiltration, dialytic hypotension more likely reflects the effect of the poison. While hypotension is often a marker of poisoning severity and can be an indication for ECTR, it may limit the ECTR’s performance.

During hemoperfusion, blood circulates through an extracorporeal circuit equipped with a charcoal or resin cartridge onto which the poison is adsorbed. Compared with diffusion, adsorption is not as limited regarding the MW, lipid solubility, or protein binding of the poison. Alcohols and most metals, however, are poorly adsorbed. Although certain exchange resins (XAD-4) can remove organic solutes and nonpolar (lipid soluble) poisons, they are unavailable in some countries (e.g., the United States).

Current hemoperfusion devices have improved biocompatibility. Coating of the sorbent material minimizes direct contact between the adsorbent and blood constituents, reducing the risk of embolization without impairing its adsorptive capacity. The circuit requires more anticoagulation than IHD and blood flows are limited to 350 mL/min because of hemolysis.

Hemoperfusion carries a higher risk of complications than IHD, namely the nonselective adsorption of certain cells and molecules. ^, Thrombocytopenia, hypoglycemia, and hypocalcemia can occur within hours. Cartridge costs are several-fold higher than for dialyzers and need to be replaced every 3 to 4 hours due to saturation. With technologic advancements in IHD, clearances of theophylline and phenobarbital by hemoperfusion or IHD are now comparable. ^{, ,} Hemoperfusion does not correct electrolyte and acid-base disturbances or fluid overload and is less available. Consequently, IHD is preferred in most settings where hemoperfusion is indicated.

Recently, CytoSorb, a hemoadsorption device containing polymer beads integrated into CRRT prefilter, has been developed to remove cytokines in septic shock. The device can adsorb xenobiotics with a MW up to 60 kDa and must be changed after 4 to 8 hours. Although there is experience with poisonings, it is uncertain if CytoSorb provides any advantage over IHD, given that most poisons are small. There are reports of successful treatment of amitriptyline, quetiapine, venlafaxine, digitoxin, and flecainide poisonings. However, given the large V _D of these poisons, the impact of CytoSorb on clinical outcomes is uncertain.

During hemofiltration (HF), solute and solvent are removed by solvent drag (convection) and replaced by a physiologic solution. Clearance is dependent on the sieving coefficient (SC) of the membrane, which is the ratio of a solute concentration in the ultrafiltrate to its concentration in the plasma. A solute that freely crosses the membrane has an SC of 1, while a SC of 0 means that the solute cannot cross the membrane. Poison elimination by HF depends on factors similar to those regulating diffusion. However, convective transport also allows the removal of larger poisons, up to 50,000 Da, which has limited relevance in toxicology because most poisons have a MW below 2000 Da.

Diffusion and convection are increasingly combined (hemodiafiltration). Adsorption (hemoperfusion) and diffusion (IHD) are sometimes used concurrently, in series, to maximize poison clearance. However, the clearance of this combination is not additive and increases complications. Clearances are additive if the therapies are implemented in parallel, with two distinct circuits.

Most ECTRs can be offered intermittently or continuously. CRRT is popular in the intensive care unit (ICU) to manage AKI, and available in various forms (see Chapter 64): continuous venovenous hemodialysis (CVVHD), continuous venovenous hemofiltration (CVVH), and continuous venovenous hemodiafiltration. Their role in poisoning remains uncertain. The solute clearance with CRRT is lower than with IHD because of lower blood and effluent flow rates. In AKI, CRRT is often chosen due to better tolerability for fluid removal in shock. However, in poisoned patients, net ultrafiltration is rarely required and IHD therefore preferred given its superior poison clearance. CRRT may be used after IHD to avoid a rebound in poison concentration; however, the benefit of this practice is uncertain (see later).

Sustained low-efficiency dialysis (SLED) is a hybrid technique between IHD and CRRT, ^, which provides a lower poison clearance than IHD over the same period.

PD clears substances by diffusion but has a limited role in acute poisoning, as the maximum clearance is 15 mL/min, representing <10% of that achieved during IHD. However, PD does not require an extracorporeal circuit and may be easier to perform in neonates and the only technique available is resource-poor settings.

Plasmapheresis involves the withdrawal of venous blood and separation of plasma from blood cells by filtration or centrifugation and reinfusion of cells with autologous plasma or another replacement solution (Chapter 65). In TPE, the removed plasma is discarded and replaced by 5% albumin or fresh-frozen plasma. Clearance is limited by the plasma removal rate (≈50 mL/min). ^, TPE is considered for highly protein-bound (>95%) or large poisons (>50,000 Da), such as cisplatin , L-thyroxine , and vincristine. TPE has been used to treat near-fatal infusion reactions to rituximab. Complications include those associated with catheter installation, bleeding, hypocalcemia, and hypersensitivity reactions to the replacement solution. ^,

During exchange transfusion (ET), whole blood or red blood cells are removed by apheresis and replaced with transfused blood products. ET may be considered in poisons causing massive hemolysis (e.g., sodium chlorate or nitrite poisoning in a patient with G6PD deficiency) or in infants, as it is technically simpler than IHD.

Albumin dialysis is historically used to replace liver function in fulminant hepatitis or severe cirrhosis. There are three types of “albumin dialysis.” Single-pass albumin dialysis (SPAD) is similar to CRRT with albumin added to the dialysate. Molecular Adsorbents Recirculation System (MARS) is identical to SPAD, but the albumin-enhanced dialysate (with the adsorbed poison) is recycled after going through a dialysis filter with resin and charcoal cartridges. Prometheus combines albumin adsorption with IHD after selective filtration of the albumin fraction through a polysulfone filter.

These devices can theoretically remove albumin-bound xenobiotics, bile acids, and bilirubin better than IHD or CRRT and have been used to support liver function in hepatoxicity caused by acetaminophen and amatoxin-containing mushroom poisonings. However, they provide inferior clearances for theophylline, valproic acid, and phenytoin. Their role remains unclear due to their limited clearance and availability, as well as high cost and complication rates.

EG, methanol, and isopropanol are small unbound molecules with a small V _D (0.6 L/Kg). Peak concentrations occur rapidly after ingestion. Approximately 30% of EL is excreted unmetabolized by the kidneys; the remainder is oxidized to glycolaldehyde by alcohol dehydrogenase (ALDH) in the liver and converted to glycolic acid by aldehyde dehydrogenase, followed by the conversion of glycolic acid to glyoxylic acid ( Fig. 66.4 , A). Metabolic acidosis results mainly from the accumulation of glycolic acid, while oxalate precipitates in tissues including kidneys, causing AKI. Glycolate cross-reacts with some lactate assays on point-of-care blood analyzers, which falsely elevates serum lactate levels. Any EL ingestion over 0.1 mL/kg requires treatment. Deaths have been reported with ingestions as low as 30 mL.

Metabolism of toxic alcohols.

(A) Ethylene glycol: The broken arrow points to inhibitors of alcohol dehydrogenase; the asterisk denotes the rate-limiting step. In the presence of the electron acceptor, nicotinamide adenine dinucleotide (NAD ⁺ ), ethylene glycol is oxidized to glycolaldehyde by alcohol dehydrogenase. Aldehyde dehydrogenase then rapidly converts glycolaldehyde to glycolic acid, which is followed by the slow conversion of glycolic acid to glyoxylic acid (the rate-limiting step). Final end products include oxalic acid, glycine, and oxalomalic acid, which are all effectively removed by intermittent hemodialysis (IHD). (B) Methanol: The broken arrow points to inhibitors of alcohol dehydrogenase; the asterisk denotes the rate-limiting step. (C) Isopropanol. NADH + H ⁺ , reduced form of nicotinamide adenine dinucleotide; TCA, tricarboxylic acid.

Methanol is metabolized principally by ALDH (85%) to formaldehyde and then oxidized by formaldehyde dehydrogenase to formic acid (see Fig. 66.4 , B). In a folate-dependent step, formic acid is transformed to water and CO ₂ . Formic acid inhibits cytochrome c oxidase in cells’ mitochondria and is responsible for toxic symptoms. The lethal dose of methanol is 10 mL.

Approximately 80% of isopropanol is metabolized to acetone via ALDH, and the remainder is eliminated unchanged in urine ( Fig. 66.4 , C). Isopropanol displays first-order elimination kinetics with an elimination half-life of 3 to 8 hours, ^, while the half-life of acetone is 10 hours. ^, Most of the CNS depressant effects are attributed to isopropanol. The lethal dose is 150 to 250 mL.

All toxic alcohols initially cause inebriation, similar to that seen with ethanol. As EL is metabolized, metabolic acidemia appears after a latent period of approximately 3 to 6 hours after ingestion, with progressive neurotoxicity (coma, seizures, cerebral edema, cranial nerve defects, radiculopathies, brainstem injury); cardiotoxicity (hypotension); and respiratory distress. AKI occurs in approximately 50% of patients due to the precipitation of calcium oxalate crystals in the kidneys and direct tubular injury from glycolate. If left untreated, EL poisoning leads to multiorgan failure and death in up to 20% of patients.

Methanol poisoning results in neurologic, visual, and gastrointestinal (GI) symptoms. CNS manifestations may include inebriation, headaches, dizziness, nausea, seizures, and cerebral edema. Methanol can produce a Parkinson-like syndrome through damage to the putamen and subcortical white matter of the basal ganglia. ^, Metabolic acidosis, caused by accumulation of formate and lactate, can be severe. Vision deficits (blurred vision and decreased visual acuity) are the hallmark of methanol poisoning and usually occur 6 to 30 hours after exposure, depending on how much ethanol is coingested. Vision deficits are permanent in 25% of cases. Death is usually caused by cardiovascular shock and respiratory arrest.

Isopropanol overdose is characterized by altered sensorium, which may range from mild inebriation, lethargy, stupor, respiratory depression, and/or coma. Isopropanol irritates the GI tract causing nausea, vomiting, gastritis, and abdominal pain. Isopropanol is toxic to myocytes and can induce severe reversible hypotension. Other findings include hypoglycemia from impaired gluconeogenesis, hypothermia, and hemolytic anemia.

Toxic alcohol assays are not widely available. EG and methanol poisoning should be considered in any patient presenting with a high anion gap metabolic acidosis and/or a high osmol gap. In early presentations, the osmol gap is more prominent, whereas in later presentations, a high anion gap predominates. As toxic alcohols are osmotically active, the osmol gap, calculated as the difference between measured osmolality (by freezing point depression) and calculated osmolarity, approximates the toxic alcohol concentration (in mmol/L) and can be converted to mg/dL ( Table 66.7 ). The osmol gap can be monitored serially during admission and dialysis when serum concentrations are unavailable. ^, The calculated osmolarity is based on the concentrations of sodium, glucose, blood urea nitrogen, and ethanol:

The normal osmol gap is between–5 and 10. In isopropanol poisoning, isopropanol and acetone contribute to the osmol gap. A “normal” osmol gap does not exclude the diagnosis of toxic alcohol poisoning: If a specific patient has a baseline gap of 0, ingestion of 25.6 mg/dL or 8 mmol/L of methanol would still lead to a “normal” osmol gap. Also, the osmol gap remains normal if a patient presents late after ingestion, once the parent alcohol has undergone complete oxidation since glycolic acid and formic acid do not contribute to the osmol gap.

Similarly, the anion gap can be used to estimate the glycolic acid and formic acid concentrations. The anion gap is the difference between measured cations (Na ⁺ and K ⁺ ) and anions (Cl ^– and HCO ₃ ^– ), representing the difference between unmeasured anions and cations, in mmol/L. Lactate can also contribute to the anion gap and needs to be factored in the equation:

Studies have shown a good correlation between the concentration of metabolites and the anion gap. ^, The absence of an anion gap suggests early presentation after EG or methanol exposure before metabolism, coingestion of ethanol, or another alcohol (isopropanol and propylene glycol).

Calcium oxalate crystals are found in the urine sediment of 50% of patients with EG exposure, 4 to 8 hours after ingestion, ^, but can also be seen with large ingestions of vitamin C or high oxalate-containing food. Urine that fluoresces under Wood lamp illumination is a unique feature of EG poisoning attributed to many types of antifreeze, which can be detected for up to 6 hours after ingestion. ^, Other laboratory abnormalities are hypocalcemia and leukocytosis.

Characteristic findings of isopropanol exposure include increased osmol gap, absent or low-grade metabolic acidosis with a normal anion gap, fruity acetone breath, acetonemia, and acetonuria without hyperglycemia. Acetonemia or acetonuria can be suspected by a positive sodium nitroprusside reaction in plasma or urine. A low concentration of serum ketones 2 hours after isopropanol ingestion (in absence of ALDH inhibition) generally excludes substantial ingestion.

In toxic alcohol poisonings, clinicians often must initiate treatment before confirming the exposure with specific laboratory tests. Management includes supportive care, correction of acidemia, antidotal therapy, and ECTRs. High-minute ventilation should be maintained if significant acidemia is present. Because toxic alcohols are rapidly absorbed from the GI tract, decontamination is seldom performed. Aspiration of the gastric contents may be useful if the ingestion is recent.

The cornerstone of treatment for EG or methanol poisoning is to prevent their metabolism into toxic metabolites using an antidote (i.e., ethanol or fomepizole, which competes for ALDH).

Indications for these antidotes are ^, :

• 1.

  Serum concentration of EG >20 mg/dL (3.2 mmol/L) or methanol >20 mg/dL (6.3 mmol/L)

• 2.

  Documented recent (hours) ingestion of toxic amounts of EG or methanol and osmol gap >10

• 3.

  A history or strong clinical suspicion of EG or methanol poisoning and at least two of the following: arterial pH <7.3, serum bicarbonate <20 mEq/L, osmol gap >10, or oxalate crystals in urine

ALDH inhibitors are contraindicated in isopropanol poisoning; the metabolite (acetone) has limited toxicity, and administration of ethanol and fomepizole can prolong the CNS depressant effect of isopropanol. Ethanol can be given enterally, intravenously, or via the dialysate. The intravenous formulation allows immediate bioavailability and avoids GI distress. Target serum ethanol concentrations for ALDH competition are between 100 and 150 mg/dL (21.7–32.5 mmol/L); below this concentration, ALDH inhibition would not maximally inhibit metabolite formation. Above this range, there is a high risk of respiratory depression. Fomepizole (4-methylpyrazole, Antizol) has largely replaced ethanol as the primary therapy and has multiple advantages over ethanol: 1) more potent ALDH inhibition, 2) simple dosing, 3) predictable pharmacokinetics, 4) no blood monitoring, 5) few side effects and no CNS depression, and 6) longer duration of action. Its main drawback is the cost (US$1000 per dose). Either antidote is usually continued until EG or methanol concentrations are below 20 mg/dL (EG <3.2 mmol/L or methanol <6.3 mmol/L) and the patient is asymptomatic with a normal arterial pH. Table 66.8 presents dosages of ethanol and fomepizole during toxic alcohol poisonings.

Acidemia should be corrected with the administration of IV sodium bicarbonate, which enhances deprotonation of acid metabolites, making them less likely to penetrate end-organ tissues (retina and kidney) and increasing their excretion by the kidneys. An intravenous bolus (1–2 mEq/kg), followed by an infusion, if necessary, should be given to maintain an arterial pH between 7.45 and 7.55. Asymptomatic hypocalcemia is not routinely treated in EG poisoning because it could exacerbate calcium oxalate crystal formation and deposition. Pyridoxine, thiamine, and magnesium are cofactors in the metabolism of EG, and their supplementation is recommended in malnourished patients or those with known deficits. Folic acid supplementation (50 mg IV every 4 to 6 hours for 5 doses) is recommended for methanol poisoning.

ECTRs, especially IHD, are extremely efficient at removing alcohols and their toxic metabolites, as well as correcting metabolic acidosis. With optimal dialysis parameters, clearance of alcohols and metabolites can reach 250 mL/min.

The increasing availability of fomepizole has modified the indications and pertinence of ECTR because of its great efficacy in preventing metabolite formation. Patients with EG poisoning without acidosis or kidney impairment may be treated with fomepizole alone, whatever the concentration of EG. The same applies to methanol; however, endogenous clearance of methanol with fomepizole is low. Assuming a methanol half-life of 54 hours with fomepizole, a patient with a methanol concentration of 320 mg/dL (100 mmol/L) would need 9 days to reach a safe methanol concentration (<20 mg/dL or 6.3 mmol/L). Dialysis might reduce hospitalization costs and antidote requirements for patients poisoned with EG or methanol. ^, IHD is mandatory with metabolic acidosis or an increased anion gap because this indicates toxic metabolite accumulation.

Indications for ECTR for EL and methanol poisoning include the following ^, :

• •

  Serum methanol concentration >50 mg/dL (15.6 mmol/L) or serum EG >50 mmol/L

• •

  Osmol gap >50 due to methanol or EG

• •

  Serum glycolate concentration is >8–12 mmol/L

• •

  Anion gap calculated with K ≥24 mmol/L

• •

  Metabolic acidosis (pH <7.2)

• •

  Coma or seizures

• •

  Vision deficits due to methanol

• •

  AKI or CKD

The expected duration of dialysis (hours) can be estimated using an estimated half-life of EG or methanol of 2.5h during IHD ^, provided that metabolic acidosis is corrected. Monitoring of the toxic alcohol concentration or osmol gap is recommended to confirm the estimation. IHD should be continued until the parent alcohols are <20 mg/dL (EG <3 mmol/L or methanol <6 mmol/L) and metabolic acidosis is corrected. CRRT offers lower clearance and requires longer treatment but may be considered if IHD is unavailable. ^,

IHD effectively removes isopropanol and acetone and is indicated when the isopropanol concentration is >400 mg/dL (66.6 mmol/L), or with prolonged coma, hypotension, myocardial depression, tachydysrhythmias, or AKI.

The use of heparin should preferably be avoided in methanol-poisoned patients due to the increased risk of intracerebral hemorrhage.

Salicylates are rapidly absorbed from the GI tract. Peak serum concentrations occur within an hour, except for enteric-coated products. In acute overdose, bezoar formation and pylorospasm may prolong absorption and the onset of symptoms can be delayed. ^,

With therapeutic dosing, salicylic acid undergoes first-order hepatic metabolism and <10% of salicylate is excreted unchanged by the kidneys. Its elimination half-life is 2 to 4 hours. ^, In acute overdose, protein binding decreases from 90% to 50%, the V _D increases, and metabolism becomes saturated, leading to a prolonged elimination half-life (>30 hours). Salicylic acid is a weak acid (pKa of 3.0). It exists predominantly in the nonionized form in the plasma at physiologic pH:

Nonionized particles cross cell membranes including the blood-brain barrier more readily than the ionized form. Acidemia drives the above reaction to the right, increasing CNS toxicity. Alkalinization drives the reaction to the left, reducing cell toxicity and favoring urinary elimination.

Salicylate poisoning can be acute or chronic. Acute ingestions >150 mg/kg lead to mild-to-moderate toxicity while >500 mg/kg is potentially lethal. The therapeutic range is 10 to 30 mg/dL (0.7–2.2 mmol/L). Toxic symptoms are usually seen at concentrations exceeding 75 mg/dL (5.4 mmol/L) and high risk of lethality >10 mg/dL (7.2 mmol/L).

Acute salicylate ingestion causes nausea, vomiting, and decreased gastric motility. Acid-base abnormalities occur, with the classical finding being mixed respiratory alkalosis and high anion gap metabolic acidosis. Salicylate stimulates the respiratory center in the brainstem, leading to a fall in PCO ₂ . Metabolic acidosis is caused by accumulation of organic acids and lactic acid production from increased minute ventilation and uncoupling of mitochondrial oxidative phosphorylation. ^,

Salicylates cause a variety of neurologic symptoms including tinnitus, central hyperthermia, vertigo, agitation, delirium, hallucination, or coma. ^, As stupor progresses, there may be blunting of the respiratory response, decreasing blood pH, and increasing salicylate entry to the CNS. ^, Tinnitus occurs at salicylate concentration of 30 mg/dL (2.2 mmol/L). Early salicylate poisoning may present with hyperglycemia due to glycogenolysis, gluconeogenesis, and decreased peripheral use. Hypoglycemia occurs later and is a marker of severe poisoning.

Chronic poisoning from regular therapeutic salicylate doses can occur, especially when kidney function is decreased. Symptoms such as confusion and pulmonary edema are more prominent than noted at the same concentration following acute ingestion.

The diagnosis of salicylate poisoning is suspected from the history, clinical findings, and metabolic abnormalities. An elevated anion gap with respiratory alkalosis should prompt salicylate measurement. Because absorption may be erratic or prolonged, serial measurements (every 2–4 hours) are required to identify the peak concentration. While salicylate concentrations reflect the risk of poisoning, treatment is also prompted by symptoms. The Done nomogram is no longer used because of its poor predictive value.

General principles of poisoning management apply to salicylates. However, patients are dependent on high-minute ventilation and high serum pH to reduce salicylate entry into the CNS. Endotracheal intubation should be performed only if necessary and by an experienced clinician. Ventilator settings should replicate the patient’s respiratory pattern before intubation, although difficult because of auto-PEEP.

Further therapy aims at decreasing absorption and increasing elimination of salicylates. MDAC decreases absorption and increases elimination ^, but is less efficacious for enhanced elimination than urine alkalinization. ^{, , ,} Serum and urine alkalinization are crucial. For salicylate poisoning, urine alkalinization reduces the elimination half-life from 19 to 5 hours. Because the ionization constant (pKa) is a logarithmic function, small changes in urine pH will have a large effect on salicylate elimination. ^, The bicarbonate infusion is titrated a urine pH of 7.5 and continued until the serum salicylate concentration is <30 mg/dL (2.2 mmol/L). Alkaline urine cannot be produced in hypokalemia since kidney reabsorption of potassium occurs via the H ⁺ /K ⁺ exchange pump in exchange for hydrogen. For this reason, and because kaliuresis complicates bicarbonate therapy, potassium should be monitored and hypokalemia aggressively corrected. Preemptive potassium supplementation is recommended if serum potassium is <4 mmol/L prealkalinization. Alkalinization may be contraindicated in AKI or pulmonary edema, when IHD is preferred. Neither MDAC nor urine alkalization replace ECTR in severe poisonings.

Salicylate is highly dialyzable due to its low V _D , low MW, and saturation of protein binding at high plasma concentrations. IHD maximizes salicylate removal and corrects acid-base status. It is noteworthy that most deaths occur before ECTR initiation.

Indications for ECTR include any of the following attributable to salicylate poisoning :

• •

  Neurologic symptoms (confusion, seizures, coma)

• •

  New hypoxemia requiring supplemental oxygen (e.g., due to pulmonary edema)

• •

  pH <7.25

• •

  Serum salicylate concentration >90 mg/dL (6.5 mmol/L)

• •

  Impaired kidney function with salicylate concentration >90 mg/dL (6.5 mmol/L)

• •

  Clinical deterioration despite appropriate treatment

Treatments should be maintained until salicylate concentrations are below 19 mg/dL (1.4 mmol/L) with clinical improvement. ^,

The absorption of lithium is rapid at therapeutic doses. Peak blood concentrations are usually delayed 8 to 12 hours post-ingestion in acute poisoning. Lithium is unbound to proteins and has a V _D of 0.7 to 0.8 L/kg. It distributes slowly into the brain, explaining the absence of initial CNS effects after an acute overdose. Lithium is eliminated by the kidneys, with 80% being reabsorbed in the tubules, so total body clearance is ≈20% of GFR. Lithium reabsorption follows that of sodium, so salt-depleted states (e.g., volume depletion) increase lithium retention. The half-life of lithium with therapeutic use is usually 18 to 24 hours but can be prolonged in the elderly, chronic users, and CKD. ^, With a normal kidney function, the elimination half-life is approximately 10 hours in acute overdose (due to distribution into extravascular tissues) versus 30 to 40 hours in chronic overdose.

Lithium overdose is defined as “acute” after a single exposure in a naïve patient and as “chronic” after dosing or prescribing errors or when lithium clearance becomes impaired. Patients with acute poisoning have few symptoms at concentrations of 4.0 mmol/L, while clinical toxicity can occur in chronic poisonings with near therapeutic concentrations. ^, In acute poisoning, GI symptoms (nausea, vomiting, or diarrhea) and nonspecific cardiac conduction delay are predominant, although life-threatening dysrhythmias are uncommon. In chronic poisoning, symptoms are mostly neurologic, ranging from coarse tremor or dysarthria to lethargy, seizures, hyperthermia, coma, and death. ^, A protracted neurologic course can be seen after severe poisoning. Some patients develop a syndrome of irreversible lithium-effectuated neurotoxicity (SILENT), which relates to cerebellar and cognitive deficits.

Therapy is guided by the type of poisoning (acute vs. chronic), lithium concentrations, symptoms, and kidney function. ^,

Hyperthermia, dysrhythmias, and seizures should be treated according to standard protocols. Volume contraction favors proximal lithium reabsorption and should be promptly corrected with isotonic saline. Following resuscitation, hypotonic solutions or free water may be required for significant hypernatremia resulting from lithium-induced nephrogenic diabetes insipidus.

Whole bowel irrigation may prevent absorption of sustained-release formulations. ^, Oral sodium polystyrene sulfonate (Kayexalate), a cation-exchanger used for hyperkalemia, can bind unabsorbed lithium and enhance its elimination ^{, ,} and can be considered in patients without contraindications with mild-to-moderate symptoms for whom ECTR is delayed or not offered. ^,

Lithium is readily dialyzable and IHD is the modality of choice, with clearances of 180 mL/min with modern filters. Lithium concentrations often rebound after IHD, ^, but this should not be concerning, as lithium CNS concentrations continue to decrease during this time, unless there is ongoing absorption from the gut. CRRT provides clearances of approximately 25% of that by IHD. ^,

Indications for ECTR include any of the following :

• •

  Severe neurologic features (central hyperthermia, seizures, and/or altered consciousness)

• •

  Serum Li concentration >5 mmol/L

• •

  Kidney impairment with symptoms and serum Li concentration >4 mmol/L

• •

  Life-threatening dysrhythmias

The threshold for ECTR is lower with impaired kidney function or if volume repletion is not tolerated. ^,

Valproic acid has a small MW (144.21 Da) and V _D (0.2 L/kg) and a high bioavailability. Serum concentrations peak 1 to 13 hours after ingestion, depending on the preparation and dose. Therapeutic serum concentrations range from 50 to 100 mg/L (347–693 μmol/L). Protein binding of valproic acid is saturable in overdose, being 90% at therapeutic concentrations and 35% at 300 mg/L (2079 μmol/L).

The liver rapidly metabolizes valproic acid. It undergoes glucuronic acid conjugation (70%) and β- and ω-oxidation to various metabolites, while <3% is excreted unchanged in urine. In overdose, more is metabolized by CYP450-mediated ω-oxidation to 5-OH-VPA and 4-en-VPA, which are responsible for some of the toxic effects.

Most valproic acid overdoses are well tolerated. Toxicity is more likely following ingestions >200 mg/kg. ^, Acute poisoning is manifested by GI distress (nausea, vomiting, diarrhea); CNS abnormalities (confusion, obtundation, coma with respiratory failure); hypotension; and elevated transaminase concentrations. Free and total valproic acid serum concentrations poorly correlate with the severity of poisoning, but most patients with total concentrations greater than 180 mg/L (1247 μmol/L) develop some degree of CNS depression. Hyperammonemia can occur at any concentration, and when markedly elevated, it can contribute to encephalopathy, cerebral edema, and death.

At very high serum concentrations (>1000 mg/L or 6930 μmol/L), complications include high anion gap metabolic acidosis, hypernatremia, hypotension, hypocalcemia, pancreatitis, noncardiogenic pulmonary edema, bone marrow suppression, and AKI. ^, Diagnosis is based on a history of exposure, symptoms, and serum valproate concentration.

Treatment begins with stabilization of respiratory and cardiovascular function. Activated charcoal should be administered, particularly within 1 hour of exposure, but can be useful beyond that period due to the prolonged absorption phase in overdose. Supportive data are limited regarding L-carnitine as a treatment for hyperammonemic encephalopathy. ^, If used, intravenous L-carnitine can be administered with a loading dose of 100 mg/kg, followed by infusions of 50 mg/kg every 8 hours, up to a maximum of 3 g per dose until ammonia levels are decreasing.

IHD clears ammonia, corrects metabolic acidosis, and removes valproic acid in overdose due to saturation of plasma protein binding. ^,

Indications for ECTR include any of the following attributable to valproate toxicity :

• •

  Serum valproate concentration >900 mg/L (6250 μmol/L)

• •

  Shock

• •

  Cerebral edema, coma, or respiratory depression requiring mechanical ventilation

• •

  Acute hyperammonemia

• •

  pH ≤7.10

Rebound of valproic acid concentrations occurs 5 to 13 hours after IHD, requiring additional sessions. Hemoperfusion has been successfully used in cases but is limited by early column saturation. “In-series” IHD-hemoperfusion offers marginal advantage over IHD in terms of efficacy but is offset by the added cost and potential complications. CRRT is less effective and should only be used if IHD is unavailable or if cerebral edema is present. ^, Albumin dialysis, TPE, and PD are inferior therapeutic options in valproate poisoning and are not recommended. ^,

Carbamazepine inhibits sodium channels, neuronal depolarization, and glutamate release and appears to have anticholinergic effects at high concentrations.

Carbamazepine is available in immediate and modified-release formulations and is characterized by erratic and incomplete absorption, which is exacerbated in overdose due to pharmacobezoar formation and ileus. Carbamazepine has small MW (236 D), small-to-moderate V _D (1.2 L/Kg), and a protein binding of approximately 75%, which does not decrease much in overdose. It undergoes hepatic metabolism into many metabolites, the most important being carbamazepine-10,11-epoxide, which is pharmacologically active. Carbamazepine induces its own metabolism, which increases endogenous clearance with chronic use. The therapeutic concentration range is 4 to 12 mg/L (16.9–50.8 μmol/L).

Carbamazepine toxicity frequently presents with neurologic, cardiovascular, and anticholinergic symptoms, which may be delayed because of its erratic absorption. Mild toxicity (≈30 mg/L or 127 μmol/L) presents as drowsiness, nystagmus, tachycardia, hyperreflexia, or dysmetria. In severe exposures (>40 mg/L or 169 μmol/L), lethargy, seizure, coma, QRS prolongation, hypotension, and ileus may develop; death is unusual. Agranulocytosis and SIADH are typically not seen in acute poisonings.

Diagnosis relies on the history, clinical findings, and laboratory testing. Serial measurements of serum carbamazepine concentrations are required because the time to peak may be beyond 24 hours. Concentrations should be obtained every 4 to 6 hours until a definite downward trend is seen.

Most patients can be managed with supportive care including ventilatory support, benzodiazepines for seizures, vasopressors for hypotension, and sodium bicarbonate for sodium channel blockade. GI decontamination is administered even if the patient presents late after ingestion due to the prolonged absorption phase but is contraindicated if ileus is present. MDAC can enhance carbamazepine clearance and may reduce the duration of coma and need for mechanical ventilation. ^,

Severe carbamazepine poisonings can be treated using ECTR, providing better and more predictable clearance than MDAC. Carbamazepine clearance with IHD compares with that of hemoperfusion and is therefore the ECTR of choice, given all the known advantages stated earlier. ^{, ,} There are limited data on the ECTR clearance of the toxic metabolite carbamazepine-10,11-epoxide. ^, CRRT, TPE, and albumin dialysis have been used but provide lesser clearances. ^,

Indications for ECTR include any of the following attributable to carbamazepine toxicity :

• •

  Prolonged coma

• •

  Seizures, cardiovascular instability, or other signs unresponsive to supportive care

• •

  Rising carbamazepine concentrations despite MDAC

Absorption kinetics are variable and influenced by the dose, presence of ileus, and coingested drugs. Barbiturates are small molecules with low protein binding. Their metabolism is mostly hepatic, but phenobarbital undergoes some renal excretion. Chronic barbiturate use induces its own metabolism, increasing endogenous clearance.

An acute oral dose of 1 g of most barbiturates will cause serious toxicity. For phenobarbital, death can result in ingestions >2 g or concentrations >80 mg/L (345 μmol/L), without supportive care. The following discussion focuses on phenobarbital since it is the most encountered, and recommendations also apply to primidone because it is partially metabolized to phenobarbital.

Phenobarbital has prolonged clinical effects, ^, , which progress from sedation in mild exposures to apnea, hypotension, circulatory collapse, hypothermia, and coma in large ingestions. ^, Concomitant AKI, cardiac, or pulmonary disease increases the clinical sensitivity to barbiturates. Early deaths after barbiturate ingestion are caused by respiratory and cardiovascular collapse, whereas later deaths are caused by acute lung injury, ventilator-acquired pneumonia, cerebral edema, or multiorgan failure, particularly when there is delayed presentation for medical care.

Supportive measures including rewarming, hydration, and vasopressors are usually sufficient. Patients may be profoundly comatose and require mechanical ventilation for many days. MDAC increases the clearance of phenobarbital and is the first-line treatment for enhanced elimination. Urinary alkalinization can increase phenobarbital renal clearance by twofold to threefold. ^, However, because renal clearance of phenobarbital is low (<3 mL/min), alkalinization has lesser impact. A controlled trial indicated that MDAC was superior to MDAC plus urinary alkalinization, which was superior to urinary alkalinization.

No randomized studies have evaluated the effect of ECTR in humans, although the mortality rate was lower in an uncontrolled group with hemoperfusion. Modern high-flux dialyzers provide clearances at least equal to hemoperfusion cartridges.

Indications for ECTR include any of the following (in particular for phenobarbital poisoning) ^, :

• •

  Coma with respiratory depression

• •

  Hypotension not responding to vasopressors

• •

  Inefficiency of MDAC at reducing serum concentrations

When treating barbiturate poisoning in chronic users, there is a risk of precipitating barbiturate withdrawal when concentrations fall below the therapeutic range. This can be minimized by stopping ECTR and other treatments when the phenobarbital concentration is therapeutic, after which it may be prudent to reinitiate barbiturate treatment.

Phenytoin has high protein binding (90 to 95%), which decreases to 70% in the presence of CKD or hypoalbuminemia but remains almost unchanged in overdose.

Phenytoin has erratic oral bioavailability, especially in overdose. ^, Phenytoin is mainly metabolized by the liver. The elimination half-life with oral therapeutic dosing is 14 to 22 hours but significantly longer in overdose. Therapeutic serum concentrations are 10 to 20 mg/L (40–80 μmol/L) and the potentially lethal dose in adults is 2 to 5 g.

Phenytoin toxicity is predominantly neurologic, with severity loosely correlating with serum concentrations. Between 20 and 40 mg/L (80–160 μmol/L), nystagmus, ataxia, and mild CNS depression are apparent, which progress to lethargy, confusion, hypotension, coma, and seizures >40 mg/L (160 μmol/L). Cardiovascular toxicity is unusual with oral overdoses, but short-lived atrioventricular delays and bradycardia may occur following rapid intravenous infusion. Death caused by respiratory or circulatory depression is rare.

Total serum phenytoin concentrations should be obtained in all suspected overdoses. Some conditions (uremia, extremes of age, or concomitant use of agents) displace phenytoin from its albumin-binding site. The free phenytoin concentration should be measured if at all possible in such cases, and toxicity is apparent with free (unbound) phenytoin concentrations >2.1 mg/L (8.3 μmol/L).

Supportive treatment is usually sufficient since most patients have an excellent outcome, although some may have a prolonged length of stay. Benzodiazepines are preferred for seizures. Due to prolonged absorption in overdose, gastrointestinal decontamination should also be considered in delayed presentations. Multiple-dose activated charcoal significantly reduces the elimination half-life of phenytoin. ^{, ,}

Experience with ECTRs is limited. Given the high protein binding of phenytoin, HP or TPE has been historically favored and used successfully. Surprisingly, IHD can enhance the elimination of phenytoin, perhaps due to its low binding constant to albumin, which ensures a constant pool of freely diffusible unbound phenytoin. This finding is attributable to high-flux, high-efficiency dialyzers—older ECTR apparatus did not significantly increase phenytoin clearance.

Neither PD nor CRRT has a role in phenytoin poisoning, and data are uncertain for albumin dialysis. In the case of actual or anticipated toxicity, with persistently elevated phenytoin concentration, IHD or HP can be considered.

Metformin has a large V _D (3 L/Kg) and high endogenous clearance from renal tubular secretion, with clearance up to 500 mL/min in patients with normal kidney function. The toxic dose in acute ingestions is likely to exceed 100 mg/kg.

Metformin toxicity manifests as GI symptoms (e.g., abdominal pain, diarrhea, nausea, and vomiting), metabolic acidosis with hyperlactatemia, and hypotension. Hypoglycemia, hypothermia, altered mental status, and acute pancreatitis have also been reported. ^, Metformin-associated lactic acidosis (MALA) usually occurs in the presence of underlying conditions, particularly AKI or CKD. MALA is variably defined as an arterial pH ≤7.35 and an elevated lactate concentration (e.g., >45 mg/dL or 5 mmol/L) from acute or chronic metformin poisoning. The extent to which metformin causes hyperlactatemia is controversial; some sources claim that the association is coincidental, ^, while others suggest that metformin is causative. The finding of asymptomatic patients who develop toxicity shortly after acute toxic ingestion gives credence to the latter hypothesis. Some, but not all, studies have reported a correlation between metformin concentrations and hyperlactatemia ^{, ,}

Management is supportive and includes intravenous hydration, treating concomitant and exacerbating conditions, normalizing acid-base abnormalities, and reducing metformin concentrations. Activated charcoal should be considered within 1 to 2 hours of an acute large ingestion. Patients who develop metabolic acidosis with hyperlactatemia, hypoglycemia, or other signs of metformin toxicity should be admitted to the ICU.

Severely acidotic patients (pH ≤7.1) should receive intravenous sodium bicarbonate, although treatment can be limited by hypernatremia and volume overload. ECTRs correct acidosis faster than intravenous bicarbonate, clear lactate and metformin, and treat volume overload and uremia. A dramatic improvement from ECTR has been described in some acutely poisoned patients soon after IHD initiation, suggesting that the benefit might be attributable to pH correction rather than metformin removal. Studies note similar mortality rates between patients who did and did not receive ECTR, but the IHD group was sicker at baseline, which may support a clinical benefit of ECTR. ^{, , ,} Metformin distributes in a smaller V _D in AKI and is more readily removed by ECTR in these cases.

Indications for ECTR include any of the following attributable to metformin toxicity :

• •

  Lactate concentration >180 mg/dL (20 mmol/L)

• •

  Arterial pH ≤7.1

• •

  Failure of supportive therapy in severe MALA

• •

  Shock, impaired kidney function, or coma

Recent research has confirmed the poor outcomes related to these conditions. IHD is preferred over CRRT due to higher metformin and lactate clearances. ^{, ,} More lactate is removed following 6 hours of IHD than 24 hours of continuous venovenous hemodiafiltration, although both remain inferior to normal physiological processes. Although IHD provides superior metformin and lactate clearance, CRRT may be used in some settings. ^, Hemoperfusion and TPE are not recommended.

Endpoints for ECTR cessation include normalization of blood pH and lactate. Shortened ECTR sessions can result in a marked rebound in lactate concentration, which may be reduced by sequentially using CRRT after IHD.

Plasma paraquat concentrations peak by 2 hours post-ingestion, with maximal tissue distribution over the next 6 hours. ^, Paraquat is distributed to most organs, especially the lungs, kidneys, and liver. It is predominantly excreted by the kidneys, with an elimination half-life of 12 hours with normal kidney function, and >48 hours with impaired kidney function. ^,

Paraquat catalyzes the formation of reactive oxygen species (ROS) through redox cycling. The resultant oxidative stress causes extensive cell injury and necrosis and a secondary inflammatory reaction. ^{, ,}

The clinical course depends on the dose and timing of presentation to a hospital. Ingestions between 20 and 40 mg/kg will usually cause AKI and multiorgan failure with death one to several weeks post-exposure, and ingestions >40 mg/kg cause death within 3 days. Most deaths are caused by acute respiratory distress syndrome (ARDS)-like presentation, progressing to pulmonary fibrosis ^, or corrosion to mucosal membranes and esophageal perforation, hepatic injury, and shock. Paraquat does not typically cause CNS effects.

Serum paraquat concentration and validated nomograms predict mortality, although assays are rarely available. A rapid and inexpensive qualitative test can be obtained by adding sodium dithionite to urine. If the urine changes from yellow to blue, it confirms recent paraquat exposure (the more intense the blue the higher the exposure).