Chapter 4: Drug Toxicity and Poisoning

Some chemicals may inherently be poisons, such as lead, which has no known necessary physiological role in the human body, and which is known to cause neuronal injury even at very low exposure levels. Most pharmaceuticals are threshold poisons; at therapeutic dosing the drug is used to confer a health advantage, but at higher doses the drug may produce a toxic effect. For instance, iron is a nutrient essential for heme synthesis and numerous physiological enzyme functions, but overdose of ferrous sulfate can lead to life-threatening multi-organ dysfunction.

To linearize such data, the response (death) can be converted to units of deviation from the mean, or probits (probability units). The probit designates the deviation from the median; a probit of 5 corresponds to a 50% response, and because each probit equals one standard deviation, a probit of 4 equals 16% and a probit of 6 equals 84%. A plot of the percentage of the population responding, in probit units, against log dose yields a straight line (Figure 4–1B). The LD₅₀ is determined by drawing a vertical line from the point on the line where the probit unit equals 5 (50% mortality). The slope of the dose-effect curve also is important. The LD₅₀ for both compounds depicted in Figure 4–1 is the same (~10 mg/kg); however, the slopes of the dose-response curves are quite different. At a dose equal to one-half the LD₅₀ (5 mg/kg), less than 5% of the animals exposed to compound B would die, but 30% of the animals given compound A would die.

The quantal dose-response curves described thus far represent linearization of sigmoidal dose-response curves (Figure 4–1). However, not all dose-response curves follow this shape. "U"-shaped dose-response curves can be observed for essential metals and vitamins (Figure 4–3). At low dose, adverse effects are observed since there is a deficiency of these nutrients to maintain homeostasis. As dose increases, homeostasis is achieved, and the bottom of the "U"-shaped dose-response curve is reached. As dose increases to surpass the amount required to maintain homeostasis, overdose toxicity can ensue. Thus, adverse effects are seen at both low and high dose.

Drug Absorption. Aspirin poisoning is a leading cause of overdose morbidity and mortality as reported to U.S. poison control centers (Bronstein et al., 2008). In therapeutic dosing, aspirin reaches peak plasma concentrations in ~1 hour (Chapter 34). However, after overdose, several drug-related and physiology-related factors change. Aspirin overdose may cause spasm of the pyloric valve, delaying entry of the drug into the small intestine. Aspirin, especially enteric coated forms, may coalesce into bezoars, reducing the effective surface area for absorption. Peak plasma salicylate concentrations from aspirin overdose may not be reached for 4-35 hours after ingestion (Rivera et al., 2004).

Drug Elimination. After therapeutic dosing, valproic acid has an elimination half-life (t_½) of ~14 hours (see Appendix II). Valproic acid poisoning may lead to coma. In predicting the duration of coma, it is important to consider that, after overdose, first-order metabolic processes appear to become saturated and the apparent elimination t_½ may exceed 30-45 hours (Sztajnkrycer, 2002). Such consideration has important clinical ramifications pertaining to prognosis, resource utilization, and therapy.

Table 4–1 lists some pharmaceuticals notorious for their predilection to have initial symptoms develop after a typical 4-6 hour emergency medical observation period for a suspected drug overdose.

Pharmacological Toxicity. The CNS depression produced by barbiturates is largely predictable in a dose-dependent fashion. The progression of clinical effects goes from anxiolysis to sedation to somnolence to coma. Similarly, the degree of hypotension produced by nifedipine is related to the dose of the drug administered. Tardive dyskinesia (Chapter 16), an extrapyramidal motor disorder associated with use of antipsychotic medications, seems to be dependent upon duration of exposure. Pharmacological toxicity can also occur when the correct dose is given: there is phototoxicity associated with exposure to sunlight in patients treated with tetracyclines, sulfonamides, chlorpromazine, and nalidixic acid.

Pathological Toxicity. Acetaminophen is metabolized to nontoxic glucuronide and sulfate conjugates, and to a highly reactive metabolite N-acetyl-p-benzoquinoneimine (NAPQI) via CYP isoforms. NAPQI is referred to as a biologically reactive intermediate, and such intermediates often arise due to drug metabolism. At therapeutic dosing, NAPQI binds to nucleophilic glutathione; but, in overdose, glutathione depletion may lead to the pathological finding of hepatic necrosis (Figure 4–5).

Genotoxic Effects. Ionizing radiation and many environmental chemicals are known to injure DNA, and may lead to mutagenic or carcinogenic toxicities. Many of the cancer chemotherapeutic agents (Chapters 60, 61, 62, 63) may be genotoxic. Discussions of genotoxicity can be found in Chapters 6 and 7.

Type I: Anaphylactic Reactions. Anaphylaxis is mediated by IgE antibodies. The Fc portion of IgE can bind to receptors on mast cells and basophils. If the Fab portion of the antibody molecule then binds antigen, various mediators (e.g., histamine, leukotrienes, and prostaglandins) are released and cause vasodilation, edema, and an inflammatory response. The main targets of this type of reaction are the gastrointestinal (GI) tract (food allergies), the skin (urticaria and atopic dermatitis), the respiratory system (rhinitis and asthma), and the vasculature (anaphylactic shock). These responses tend to occur quickly after challenge with an antigen to which the individual has been sensitized and are termed immediate hypersensitivity reactions.

Type II: Cytolytic Reactions. Type II allergies are mediated by both IgG and IgM antibodies and usually are attributed to their capacity to activate the complement system. The major target tissues for cytolytic reactions are the cells in the circulatory system. Examples of type II allergic responses include penicillin-induced hemolytic anemia, quinidine-induced thrombocytopenic purpura, and sulfonamide-induced granulocytopenia. Fortunately, these autoimmune reactions to drugs usually subside within several months after removal of the offending agent.

Type III: Arthrus Reactions. Type III allergic reactions are mediated predominantly by IgG; the mechanism involves the generation of antigen-antibody complexes that subsequently fix complement. The complexes are deposited in the vascular endothelium, where a destructive inflammatory response called serum sickness occurs. This phenomenon contrasts with the type II reaction, in which the inflammatory response is induced by antibodies directed against tissue antigens. The clinical symptoms of serum sickness include urticarial skin eruptions, arthralgia or arthritis, lymphadenopathy, and fever. Several drugs, including commonly used antibiotics, can induce serum sickness-like reactions. These reactions usually last 6-12 days and then subside after the offending agent is eliminated.

Type IV: Delayed Hypersensitivity Reactions. These reactions are mediated by sensitized T-lymphocytes and macrophages. When sensitized cells come in contact with antigen, an inflammatory reaction is generated by the production of lymphokines and the subsequent influx of neutrophils and macrophages. An example of type IV or delayed hypersensitivity is the contact dermatitis caused by poison ivy.

An increased incidence of peripheral neuropathy is seen in patients with inherited deficiencies in acetylation when isoniazid is used to treat tuberculosis, e.g., slow and fast acetylators exist due to polymorphisms in N-acetyl transferase. Many black males (~10%) develop a serious hemolytic anemia when they receive primaquine as an antimalarial therapy. Such individuals have a deficiency of erythrocyte glucose-6-phosphate dehydrogenase. Genetically determined resistance to the anticoagulant action of warfarin is due to an alteration in the vitamin K epoxide reductase.

Interaction of Absorption. A drug may cause either an increase or a decrease in the absorption of another drug from the intestinal lumen. Ranitidine, an antagonist of histamine H₂ receptors, raises gastrointestinal pH and may increase the absorption of basic drugs such as triazolam (O'Connor-Semmes et al., 2001). Conversely, the bile-acid sequestrant cholestyramine leads to significantly reduced serum concentrations of propranolol, and may diminish its effect for a given dosage (Hibbard et al., 1984).

Interaction of Protein Binding. Many drugs, such as aspirin, barbiturates, phenytoin, sulfonamides, valproic acid, and warfarin, are highly protein-bound in the plasma, and it is the free (unbound) drug that produces the clinical effects. These drugs may have enhanced toxicity in overdose if protein binding sites become saturated, in physiological states that lead to hypoalbuminemia, or when displaced from plasma proteins by other drugs. The anticoagulant effects of warfarin may be enhanced by displacement from plasma proteins by simultaneous valproic acid therapy (Guthrie et al., 1995).

Interaction of Metabolism. A drug can frequently influence the metabolism of one or several other drugs (Chapter 6), and this is especially notable within hepatic CYPs. Acetaminophen is partially transformed by CYP2E1 to the toxic metabolite NAPQI (Figure 4–5). Intake of ethanol, a potent inducer of the 2E1 isoenzyme, may lead to increased susceptibility to acetaminophen poisoning after overdose (Dart et al., 2006). Similarly, a number of second-generation piperidine antihistamines (terfenadine, astemizole) were removed from the market when they were noted to lead to QT interval prolongation and tachydysrhythmias when co-administered with macrolide antibiotics.

Interaction of Receptor Binding. Buprenorphine is an opioid with partial agonist and antagonist receptor activities. It can be used as an analgesic but is more commonly used to treat opioid addiction. The drug binds opioid receptors with high affinity, and can prevent euphoria from concomitant use of narcotic drugs of abuse.

Interaction of Therapeutic Action. Aspirin is an inhibitor of platelet aggregation and heparin is an anticoagulant; given together they may increase risk for bleeding. Sulfonylureas cause hypoglycemia by stimulating pancreatic insulin release, whereas biguanide drugs (metformin) lead to decreased hepatic glucose production, and these drugs can be used together to control diabetic hyperglycemia.

Chemicals are first tested for toxicity by estimation of the LD₅₀ in two animal species by two routes of administration; one of these is the expected route of exposure of human beings to the chemical being tested. The number of animals that die in a 14-day period after a single dose is recorded. The animals also are examined for signs of intoxication, lethargy, behavioral modification, and morbidity. The chemical is next tested for toxicity by repeat exposure, usually for 90 days. This study is performed most often in two species by the route of intended use or exposure with at least three doses. A number of parameters are monitored during this period, and at the end of the study, organs and tissues are examined by a pathologist.

Long-term or chronic studies are carried out in animals at the same time that clinical trials are undertaken. For drugs, the length of exposure depends somewhat on the intended clinical use. If the drug normally would be used for short periods under medical supervision, as would an antimicrobial agent, a chronic exposure of animals for 6 months might suffice. If the drug would be used in human beings for longer periods, a study of chronic use for 2 years may be required.

Studies of chronic exposure often are used to determine the carcinogenic potential of chemicals. These studies usually are performed in rats and mice for the average lifetime of the species. Other tests are designed to evaluate teratogenicity (congenital malformations), perinatal and postnatal toxicity, and effects on fertility. Teratogenicity studies usually are performed by administering drugs to pregnant rats and rabbits during the period of organogenesis. As noted in Chapter 6, in silico computational methods of chemical biology systems may soon contribute to such studies.

Once a drug is judged ready to be studied in humans, a Notice of Claimed Investigational Exemption for a New Drug (IND) must be filed with the FDA. The IND includes: 1) information on the composition and source of the drug; 2) chemical and manufacturing information; 3) all data from animal studies; 4) proposed clinical plans and protocols; 5) the names and credentials of physicians who will conduct the clinical trials; and 6) a compilation of the key data relevant to study the drug in man made available to investigators and their institutional review boards (IRBs).

It often requires 4-6 years of clinical testing to accumulate and analyze all required data. Testing in humans is begun after sufficient acute and subacute animal toxicity studies have been completed. Chronic safety testing in animals, including carcinogenicity studies, is usually done concurrently with clinical trials. In each of the three formal phases of clinical trials, volunteers or patients must be informed of the investigational status of the drug as well as the possible risks and must be allowed to decline or to consent to participate and receive the drug. These regulations are based on the ethical principles set forth in the Declaration of Helsinki. In addition to the approval of the sponsoring organization and the FDA, an interdisciplinary IRB at the facility where the clinical drug trial will be conducted must review and approve the scientific and ethical plans for testing in humans.

In phase 1, the effects of the drug as a function of dosage are established in a small number (20–100) of healthy volunteers. Although a goal is to find the maximum tolerated dose, the study is designed to prevent severe toxicity. If the drug is expected to have significant toxicity, as may be the case in cancer and AIDS therapy, volunteer patients with the disease rather than disease-free volunteers are used in phase 1. Phase 1 trials are designed to determine the probable limits of the safe clinical dosage range. These trials may be non-blind or "open"; i.e., both the investigators and the subjects know what is being given. Alternatively, they may be "blinded" and/or placebo-controlled. The choice of design depends on the drug, disease, goals of investigators, and ethical considerations. Many predictable toxicities are detected in this phase. Pharmacokinetic measurements of absorption, t_½, and metabolism are often conducted. Phase 1 studies are usually performed in research centers by clinical pharmacologists.

In phase 2, the drug is studied in patients with the target disease to determine its efficacy ("proof of concept"), and the doses to be used in any follow-on trials. A modest number of patients (100–200) are studied in detail. A single-blind design may be used, with an inert placebo medication and an established active drug (positive control) in addition to the investigational agent. Phase 2 trials are usually done in special clinical centers (e.g., university hospitals). A broader range of toxicities may be detected in this phase. Phase 2 trials have the highest rate of drug failures, and only ~25% of drug candidates move on to phase 3.

In phase 3, the drug is evaluated in much larger numbers of patients with the target disease—usually thousands—to further establish and confirm safety and efficacy. Using information gathered in phases 1 and 2, phase 3 trials are designed to minimize errors caused by placebo effects, variable course of the disease, etc. Therefore, double-blind and crossover techniques are frequently used. Phase 3 trials are usually performed in settings similar to those anticipated for the ultimate use of the drug. Phase 3 studies can be difficult to design and execute and are usually expensive because of the large numbers of patients involved and the amount of data that must be collected and analyzed. The drug is formulated as intended for the market. The investigators are usually specialists in the disease being treated. Certain toxic effects, especially those caused by immunologic processes, may first become apparent in phase 3 trials.

If phase 3 results meet expectations, application is made to the FDA for permission to market the new agent. Marketing approval requires submission of a New Drug Application (NDA) (or for biologicals, a Biological License Application [BLA]) to the FDA. Note that even at this point in the process, experience with the new drug is limited to information gathered from a few thousand patients. Thus, post-marketing surveillance is crucial in determining the actual toxicity of the drug as it is administered to a much larger population in which low-frequency adverse reactions will be noticeable.

Some toxicities of pharmaceuticals can be predicted based upon their known pharmacological mechanism; however, it is often not until the post-marketing period that the therapeutic toxicity profile of a drug becomes fully appreciated. In the U.S., the approval system for new drugs typically uses only 500 to 3000 exposed subjects. Such a system is likely to identify toxicities occurring in 1% or more of patients receiving a drug (Strom, 2004). The Adverse Event Reporting System of the FDA relies upon two signals to detect rarer adverse drug events. First, the FDA requires (Code of Federal Regulations, Title 21, Volume 5, Section 314.80) drug manufacturers to perform post-marketing surveillance of prescription drugs, and similar regulations exist for nonprescription products. Second, the FDA operates a voluntary reporting system (MedWatch, online at http://www.fda.gov/Safety/MedWatch/default.htm) available to both health professionals and consumers. Hospitals may also support adverse drug event committees to investigate potential adverse drug events, and these investigations may be made available to industry and the government. Unfortunately, any national dataset will significantly underestimate the morbidity and mortality attributable to adverse drug events because of under-reporting and because it is difficult to estimate the denominator of total patient exposures for each event reported once a drug is available on the open market.

As an example, post-marketing surveillance identified the toxicity associated with the serotonin receptor-modulating drug cisapride. This drug was known to enhance GI motility and was marketed in the U.S. as a treatment for gastroesophageal reflux. Post-marketing surveillance revealed that cisapride was associated with prolongation of the QT interval and predisposition to ventricular arrhythmia. It was withdrawn from the market, and subsequent case-control studies demonstrated increased risk of arrhythmia (Hennessy et al., 2008). Cisapride is now limited in its distribution through an investigational access program managed by the manufacturer. In another instance, selective serotonin reuptake inhibitors (SSRIs, Chapter 15) are believed to be safer than monoamine oxidase inhibitors and tricyclic antidepressants, and are popularly prescribed for treatment of depression; however, current post-marketing studies are looking to see if the use of SSRIs may predispose adolescent and young adult patients to suicidality (Barbui et al., 2009). Thus, the determination of drug toxicity carries beyond the stages of drug development.

Exact quantification of nontherapeutic poisoning occurrences remains elusive. The National Institute of Drug Abuse (NIDA) commissioned the University of Michigan to perform the "Monitoring the Future" survey of U.S. high school children regarding drug use patterns (NIDA, 2009). In 2008, 47% of surveyed high school seniors reported that they had previously used an illicit drug. Among pharmaceutical abuse, 13% reported having abused a non-heroin narcotic, 10% an amphetamine drug, 8.9% a benzodiazepine, antidepressant, or antipsychotic medicine, and 8.5% a barbiturate. NIDA also monitors emergency department (ED) visits through its Drug Abuse Warning Network (DAWN). In 2006, ~250 ED visits per 100,000 population were related to drug toxicity and 49% of these visits were related to the nonmedical use and abuse of pharmaceutical agents (U.S. DHHS). The top five drugs involved in drug-related deaths reported to DAWN-participating medical examiners' offices in 2005 are presented in Table 4–3.

The AAPCC (www.aapcc.org) offers a toll-free poisoning information phone hotline throughout the U.S., and has been collecting voluntary reports of potential poison exposures for over 25 years. It currently has over 46 million human exposure case records in its database. In 2007, nearly 2.5 million cases, including 1239 fatal poisonings, were voluntarily reported to the AAPCC's National Poison Data System (NPDS). Eighty-three percent of human poison exposures reported to the NPDS were unintentional, 13% were intentional, and 2.5% were adverse drug reactions (Bronstein et al., 2008). The substances most frequently involved in human exposures and fatalities are presented in Tables 4–4 and 4–5.

Right drug, right patient, right dose, right route, right time.

The incidence of poisoning in children has decreased dramatically over the past four decades. This favorable trend is largely due to improved safety packaging of drugs, drain cleaners, turpentine, and other household chemicals; improved medical training and care; and increased public awareness of potential poisons.

From 1958-1973, aspirin ingestion was a common cause of childhood poisoning death. In 1973, regulations were instituted requiring child-resistant packaging for aspirin consumer products, and this change in packaging was associated with a 34% reduction in the aspirin-related child mortality rate (Rodgers, 2002). In 2007, no aspirin deaths among children aged less than 6 years were reported to the AAPCC (Bronstein et al., 2008). From 1983-1990, iron was the single most frequent cause of unintentional pharmaceutical ingestion fatality in children younger than 6 years, and accounted for almost one-third of such deaths. In 1997, drug products with 30 mg or more of elemental iron per dosage unit were required to be placed into unit-dose packaging. This packaging change was associated with a decrease in mortality (odds ratio of 0.07; 95% confidence interval, 0.52–0.01) (Tenenbein, 2005).

The most typically available urine drug toxicology test is an immunoassay designed to detect common drugs of abuse such as amphetamines, barbiturates, benzodiazepines, cannabis, cocaine, and opiates. Acute poisoning with these substances can usually be determined on clinical grounds, and the results of these assays are infrequently available fast enough to guide stabilization. Additionally, detection of drugs or their metabolites on a urine immunoassay does not mean that the detected drug is responsible for the currently observed poisoning illness. When ingestion of acetaminophen or aspirin cannot clearly be excluded via the exposure history, serum quantification of these drugs is recommended. An electrocardiogram (ECG) may be useful at detecting heart blocks, Na⁺ channel blockade, or K⁺ channel blockade associated with specific medication classes (Table 4–10). Further laboratory analysis, such as use of blood gas determinations, serum chemistries, complete blood counts, and other testing, should be tailored to the individual poisoning circumstance.

Gastric emptying may be attempted by induced vomiting or by gastric lavage. Historically, pharmaceutical methods to stimulate vomiting included administration of potentially toxic doses of copper sulfate or apomorphine; more recently, syrup of ipecac took their place as the favored drug to induce emesis. Gastric emptying has been shown in volunteer studies to reduce drug absorption by ~ one-third under optimal conditions (Tenenbein, 1987). A randomized trial of gastric emptying for poisoned patients did not show improvement in clinical outcome (Pond et al., 1995), but the selection of subjects and timing of therapy may have biased the study toward the null hypothesis. Based upon review of existing evidence, the American Academy of Pediatrics no longer recommends syrup of ipecac as part of its childhood injury prevention program (AAP, 2003), and the American Academy of Clinical Toxicology dissuades routine use of gastric emptying in the poisoned patient (AACT, 2004; Manoguerra and Cobaugh, 2005). The declining role of syrup of ipecac in the treatment of poisoning is evident by its use in <0.1% of cases reported to the AAPCC in 2007, compared to ~10% in 1987 (Bronstein et al., 2008).

Syrup of Ipecac. The U.S. FDA approved syrup of ipecac for sale without a prescription in 1965. The alkaloids cephaeline and emetine within syrup of ipecac act as emetics because of both a local irritant effect on the enteric tract and a central effect on the chemoreceptor trigger zone in the area postrema of the medulla. Syrup of ipecac is available in 0.5- and 1-fluid ounce containers. Ipecac is given orally at a dose of 15 mL for children up to 12 years, and 30 mL for older children and adults. Administration of ipecac is typically followed by a drink of water, and reliably produces emesis in 15-30 minutes. Contraindications for syrup of ipecac administration include existing or impending CNS depression, ingestion of a corrosive or hydrocarbon drug (due to the emergence of chemical pneumonia), or presence of a medical condition that might be exacerbated by vomiting. Ipecac has been misused by bulimic patients; in cases of Munchausen syndrome by proxy and with chronic abuse, ipecac may cause serum electrolyte abnormalities, cardiomyopathy, ventricular arrhythmia, and death.

Gastric Lavage. The procedure for gastric lavage involves passing an orogastric tube (24-French for small children, up to 40-French for adults) into the stomach with the patient in the left-lateral decubitus position with head lower than feet. Preferably, the tube will have been designed for lavage purposes, and will have sizable side holes in the tubing. After withdrawing stomach contents, 10 to 15 mL/kg (up to 250 mL) of saline lavage fluid is administered and withdrawn. This process continues until the lavage fluid returns clear. Complications of the procedure include mechanical trauma to the stomach or esophagus, pulmonary aspiration of stomach contents, and vagus nerve stimulation.

Activated Charcoal. Charcoal is created through controlled pyrolysis of organic matter, and is activated through steam or chemical treatment that increases its internal pore structure and adsorptive surface capacity. The surface of activated charcoal contains carbon moieties, such as carbonyl and hydroxyl groups, that are capable of binding poisons. The recommended dose is typically 0.5-2 g/kg of body weight, up to a maximum tolerated dose of ~75-100 g. As a rough estimate, 10 g of activated charcoal is expected to bind ~1 g of drug. The efficacy of activated charcoal at adsorbing ingested drug diminishes over time. Alcohols, corrosives, hydrocarbons, and metals are not believed to be well adsorbed by charcoal.

Complications of activated charcoal therapy include vomiting, constipation, pulmonary aspiration, and death. Charcoal slurries are black and gritty; in a series of children offered charcoal in a pediatric emergency department, only 44% of the children <6 years accepted the agent orally (Osterhoudt et al., 2004a). Nasogastric administration of charcoal increases the incidence of vomiting (Osterhoudt et al., 2004b), and may increase the risk for pulmonary aspiration. Charcoal should not be given to patients with suspected GI perforation, or to patients who may be candidates for endoscopy.

Polyethylene glycol electrolyte solution is typically administered at a rate of 25 to 40 mL/kg/h until the rectal effluent is clear and no more drug is being passed. To achieve these high administration rates a nasogastric tube may be used. Large doses have been administered without adversely affecting serum electrolyte concentrations. WBI is contraindicated in the presence of bowel obstruction or perforation, and may be complicated by abdominal distention or pulmonary aspiration.

Cathartics. The two most common categories of simple cathartics are the magnesium salts, such as magnesium citrate and magnesium sulfate, and the nondigestible carbohydrates, such as sorbitol. The use of simple cathartics has been abandoned as a GI decontamination strategy, although sorbitol is sometimes administered with single-dose activated charcoal in an effort to add sweetness and reduce its predilection toward constipation.

Manipulating Urinary pH: Urinary Alkalinization. Drugs subject to renal clearance are excreted into the urine by glomerular filtration and active tubular secretion (Chapter 2); non-ionized compounds may be reabsorbed far more rapidly than ionized polar molecules. Weakly acidic drugs are susceptible to "ion-trapping" in the urine. Aspirin is a weak acid with a pK_a = 3.0. As the pH of the urine increases, more salicylate is in its ionized form at equilibrium, and more salicylic acid is diffused into the tubular lumen of the kidney. Urinary alkalinization is also believed to speed clearance of phenobarbital, chlorpropamide, methotrexate, and chlorphenoxy herbicides. The American Academy of Clinical Toxicologists recommends urine alkalinization as first-line treatment only for moderately severe salicylate poisoning that does not meet criteria for hemodialysis (Proudfoot et al., 2004). To achieve alkalinization of the urine, 100-150 mEq of sodium bicarbonate in 1L of D5W is infused intravenously at twice the maintenance fluid requirements and then titrated to effect. Hypokalemia should be treated since it will hamper efforts to alkalinize the urine due to H⁺-K⁺ exchange in the kidney. Urine alkalinization is contraindicated in the presence of renal failure, or when the fluid administration may worsen pulmonary edema or congestive heart failure. Acetazolamide is not used to alkalinize urine as it promotes acidemia.

Multiple-Dose Activated Charcoal. Activated charcoal adsorbs drug to its surface and promotes enteral elimination. Multiple doses of activated charcoal can speed elimination of absorbed drug by two mechanisms. Charcoal may interrupt enterohepatic circulation of hepatically metabolized drug excreted in the bile, and charcoal may create a diffusion gradient across the GI mucosa and promote movement of drug from the bloodstream onto the charcoal in the intestinal lumen. Activated charcoal may be administered in multiple doses, 12.5 g/h every 1, 2, or 4 hours (smaller doses may be used for children). Complications of therapy are similar to those listed for single-dose activated charcoal. Charcoal enhances the clearance of many drugs of low molecular weight, small volume of distribution, and long elimination t_½. In the absence of good clinical outcomes data, multiple-dose activated charcoal is believed to have the most potential utility in overdoses of carbamazepine, dapsone, phenobarbital, quinine, theophylline, and yellow oleander (AACT, 1999; de Silva et al., 2003).

Extracorporeal Drug Removal. The ideal drug amenable to removal by hemodialysis has a low molecular weight, a low volume of distribution, high solubility in water, and minimal protein binding. Hemoperfusion involves passing blood through a cartridge containing adsorbent particles. The most common poisonings for which hemodialysis is sometimes used include salicylate, methanol, ethylene glycol, lithium, carbamazepine, and valproic acid. For a more exhaustive list of drugs amenable to hemodialyis or hemoperfusion, see Winchester (2002).