The etiology of poisoning may or may not be disclosed by the patient history. Even when a history is available, its accuracy and reliability must be assessed. The identity of the toxin involved is incorrectly reported by up to 50% of patients with intentional ingestions [33]. The amount reportedly taken is also unreliable. Hence, in such patients, the history should be approached with caution. Layperson misidentification of acetaminophen as aspirin and vice versa is also relatively common. To avoid missing the correct diagnosis, the presence or absence of both drugs should be confirmed by laboratory analysis when an overdose of any kind is suspected.

Drugs in pill form can often be identified by the imprint code, the alphabetical and numeric markings on tablets and capsules. A listing of imprint codes with the corresponding trade name and ingredient(s) can be found in the Identidex portion of Poisindex [27], which is available at virtually all poison centers in the United States. It also provides the identities of street drugs based on their slang names. Prescription drugs may be identified by contacting the dispensing pharmacy. Drug samples can sometimes be identified by direct chemical analysis (Toxicology Screening section). Police and government toxicology laboratories may be of assistance when illicit drug use is involved.

By US law [34], the ingredients of potentially hazardous commercial products used in and around the home must be stated on their label. This information, however, is not necessarily present or accurate, and labels may be missing or unreadable. In such cases, the ingredients may be identified by consulting Poisindex [27] or a regional poison center. Alternatively, the manufacturer or distributor can be called to obtain information on drugs or products that they produce or distribute. This action may be particularly helpful if the product is an outdated formulation or a recently reformulated or released one. Most large companies maintain 24-hour emergency telephone numbers for such purposes, and many employ medical consultants who can also provide management advice. Although industrial products do not have the same labeling requirements as household ones, right-to-know legislation requires that companies make information regarding the ingredients and potential toxicity of products they make, distribute, or use available to workers and health care providers [35]. Such information can be obtained by requesting a Material Data Safety Sheet (MSDS).

Information on drugs and chemical products manufactured or obtained outside the United States can be found in Poisindex [27] and Martindale: The Complete Drug Reference [21], or obtained from a domestic or foreign poison center. Information on drugs undergoing clinical trials in the United States may also be found in Martindale, since such drugs are often already available in other countries. Most foreign poison centers have English-speaking staff or translators available.

Plants (including fungi or mushrooms), along with their active parts and chemical constituents, can be identified by consulting Poisindex [27] if either their common or botanical name is known. If the name is not known but a sample is available, a representative from a local nursery, horticultural or mycologic society, or university botany department may be of assistance in identifying it. Similarly, pet stores, zoos, veterinarians, amateur or academic entomologists, herpetologists, zoologists, and field guides can be helpful in identifying potentially venomous insects, reptiles, snakes, and other animals. Poison centers usually maintain lists of local experts who are willing to help with such identifications.

A toxidrome is a clinical syndrome that involves multiple physiologic systems and facilitates bedside identification of the culprit toxin [36]. The physiologic state of the patient can usually be characterized as excited (i.e., central nervous system [CNS] excitation with increased blood pressure, pulse, respirations, and temperature), depressed (i.e., decreased level of consciousness and decreased vital signs), discordant (i.e., inconsistent, mixed, or opposing CNS and vital sign abnormalities), or normal. The differential diagnosis can then be narrowed to the common or characteristic causes of these physiologic states (Table 117.3).

The excited state is primarily caused by sympathomimetics (agents that directly or indirectly stimulate α- and β-adrenergic receptors), anticholinergics (agents that block parasympathetic muscarinic receptors), hallucinogens, and withdrawal syndromes. The depressed state is primarily caused by sympatholytics (agents that block adrenergic receptors or depress cardiovascular activity), cholinergics (agents that directly or indirectly stimulate muscarinic receptors), opioids, or sedative hypnotics (which enhance the effect of the inhibitory CNS neurotransmitter gamma-aminobutyric acid [GABA] or depress neuronal membrane excitability). The discordant state is primarily due to asphyxiants (agents that decrease the availability, absorption, transport, or use of oxygen), membrane active agents (those that block sodium channels or otherwise alter the activity of excitable cell membranes), and agents that cause a variety of CNS syndromes due to interference with dopamine, GABA, glycine, or the synthesis, metabolism, or function of serotonin. A normal physiologic state may be due to a nontoxic exposure (Table 117.4), psychogenic illness, or presentation during the preclinical phase of poisoning. Agents that have a long preclinical phase (i.e., delayed onset of toxicity) are known as toxic “time bombs.” Delayed onset of toxicity may result from slow absorption or distribution, metabolic activation, or a mechanism of action
that involves the disruption of metabolic or synthetic pathways. Psychogenic illness should be considered when symptoms are inconsistent with the reported exposure and cannot be substantiated by objective physical findings, laboratory abnormalities, and toxicologic testing and other etiologies have been excluded [37].

An excited or depressed state may be mischaracterized as a discordant one when the activity of a stimulant or depressant is selective for a receptor subtype or results in a compensatory or opposing autonomic response. For example, hypotension caused by an alpha-blocker, β₂-agonist, or vasodilator may be
accompanied by tachycardia, and hypertension due to a selective α-agonist (e.g., phenylpropanolamine) may be accompanied by bradycardia. Severe stimulant or depressant poisoning can also cause what appears to be a discordant state (Table 117.5). For example, prolonged seizures and extreme hyperthermia caused by sympathomimetics can culminate in cardiovascular collapse as a consequence of anaerobic metabolism, acidosis, or depletion of neurotransmitters. Similarly, marked hypotension and hypoventilation caused by physiologic depressants can precipitate seizures and tachyarrhythmias as a result of ischemia, anoxia, and acidosis. In addition, paradoxic excitation can result from the preferential inhibition of cortical function that normally controls social activity by low doses of CNS depressants, most notably alcohol and other sedative hypnotics. In such cases, the physiologic state and its cause can often be correctly identified by the overall clinical picture and course of events.

The severity of mental status changes and the nature of associated autonomic findings can be used to narrow the differential diagnosis of physiological stimulation and depression to one of four subcategories (see Table 117.3). In the excited patient, marked vital sign abnormalities (e.g., severe hypertension with end-organ ischemia, tachyarrhythmias, hyperthermia, cardiovascular collapse) with minor mental status changes suggest an agent with peripheral sympathomimetic activity as the cause. Conversely, marked mental status abnormalities with nearly normal vital signs suggest a centrally acting hallucinogen. Anticholinergic poisoning (Table 117.6) can be differentiated from sympathomimetic (Table 117.7), hallucinogen, and withdrawal syndromes by the presence of dry, flushed, and hot skin; decreased or absent bowel sounds; and urinary retention. Other causes of excitation are usually accompanied by pallor, diaphoresis, and increased bowel or bladder activity.

In the patient with physiological depression, marked cardiovascular abnormalities (e.g., hypotension and bradycardia) with relatively clear sensorium suggest a peripherally acting sympatholytic, whereas marked CNS and respiratory depression with minimal pulse and blood pressure abnormalities suggest a centrally acting agent (opioid or sedative hypnotic). Cholinergic poisoning (Table 117.8) can be distinguished from other causes of physiologic depression by the presence of characteristic autonomic findings: Salivation, lacrimation, urination, defecation, GI cramps, and emesis (SLUDGE syndrome). In addition, cholinergic poisoning causes pallor and diaphoresis, whereas the skin is usually warm and dry with opioid and sedative–hypnotic poisoning.

Other findings can sometimes help narrow the differential diagnosis further. Only the most common and diagnostically useful ones are noted here. Because of limited specificity and sensitivity, the presence or absence of a particular sign or symptom cannot be used to confirm or exclude a given etiology.

Ocular findings can sometimes help narrow the diagnostic possibilities. Although mydriasis can be caused by any agent or condition that results in physiologic excitation (see Table 117.3), it is most pronounced in anticholinergic poisoning, in which it is associated with minimal pupil response to light and accommodation. Similarly, although miosis is a nonspecific manifestation of physiologic depression, it is usually most pronounced in opioid poisoning. Notable miosis can, however, also be caused by cholinergic agents and sympatholytics with alpha-blocking effects (e.g., phenothiazines). Visual disturbances suggest anticholinergic, cholinergic, digitalis, hallucinogen, methanol, and quinine poisoning. Horizontal nystagmus and disconjugate gaze are nonspecific manifestations
of sedative–hypnotic poisoning. Although vertical and rotary nystagmus can be seen in patients with lithium and phenytoin poisoning, they are most suggestive of phencyclidine intoxication. These etiologies should be readily distinguishable by assessing the physiologic state. Rapidly alternating lateral “ping-pong” gaze has been described in monoamine oxidase inhibitor poisoning. Except for abnormalities due to topical chemical exposure, both eyes are equally affected. Although failure to respond to topical miotics has been said to be diagnostic of drug-induced pupillary dilatation, this is only true for topical exposures. Hence, unilateral pupillary abnormalities should generally prompt evaluation for a central, structural lesion.

Dermatologic abnormalities may also be helpful. Flushed skin can be caused by anticholinergics, boric acid, a disulfiram-ethanol reaction, monosodium glutamate, niacin, scombroid (fish poisoning), and rapid infusion of vancomycin (red man syndrome). The skin is hot and dry in anticholinergic poisoning but normal or moist with other etiologies. Flushing should not be confused with the orange skin discoloration caused by rifampin. Pallor and diaphoresis may be due to cholinergics, hallucinogens, hypoglycemics, sympathomimetics, and drug withdrawal (see Table 117.3). As noted previously, manifestations of the SLUDGE syndrome distinguish cholinergic poisoning from other etiologies. Cyanosis may be due to agents that cause cardiovascular or respiratory depression, methemoglobinemia, pneumonitis, or simple asphyxia. In methemoglobinemia, it may have a chocolate-brown or slate-gray hue and is unaffected by oxygen administration. Cyanosis should not be confused with the blue discoloration of the skin caused by amiodarone or by topical exposure to blue dyes. The latter condition can be diagnosed by wiping the skin with acetone or alcohol. Hair loss, mucosal pigmentation, and nail abnormalities are suggestive of heavy metal poisoning (e.g., arsenic, lead, mercury, thallium).

Finally, the presence of neuromuscular abnormalities may suggest certain etiologies. Seizures and tremors can be caused by cholinergics, hypoglycemic agents, lithium, membrane-active agents, some narcotics (e.g., meperidine, propoxyphene), and most physiologic stimulants [38] (see Table 117.3). They can also occur in patients poisoned by agents that cause asphyxia, low lactate increased AGMA (see later), and cerebral hypoperfusion or hypoventilation (e.g., physiologic depressants; see Table 117.3). The most common causes of seizures due to poisoning are tricyclic antidepressants, sympathomimetics, antihistamines (primarily diphenhydramine), theophylline, and isoniazid. Although carbon monoxide, hypoglycemics, lithium, and theophylline can cause focal seizures, seizures due to poisoning are usually generalized. Because hypertensive and traumatic CNS hemorrhages are known complications of poisoning, the possibility of a structural lesion should be considered if focal signs and symptoms are present. Myoclonus suggests anticholinergic or sympathomimetic poisoning. Fasciculations are typical of cholinergic insecticide poisoning but can also be caused by sympathomimetics. Rigidity may be seen in phencyclidine and sympathomimetic poisoning and in those with CNS syndromes (see Table 117.3). Dystonic posturing is most often caused by antipsychotic agents. It is also a characteristic feature of strychnine poisoning.

Acid–base status, anion gap, serum osmolality, ketone, electrolyte, glucose, and organ function abnormalities identified by routine laboratory tests can be extremely helpful in the differential diagnosis of poisoning. As with clinical manifestations, the diagnostic sensitivity and specificity of a single finding is not sufficiently high for its presence or absence to confirm or exclude a specific etiology. The use of anion and osmolar gaps and serum ketone and lactate levels in the diagnosis of poisoning of unknown etiology is summarized in Figure 117.1.

Assessing acid–base status and calculating the anion gap is particularly important because an increased AGMA may be due to advanced ethylene glycol, methanol, and salicylate poisoning. In such cases, prompt initiation of specific therapies is essential to prevent progressive, irreversible, or fatal poisoning [39,40]. The normal anion gap is 13 ± 4 mEq per L in unselected acutely hospitalized patients. In ethylene glycol and methanol poisoning, AGMA is primarily due to the accumulation of acid metabolites. In salicylate poisoning, it is caused by the accumulation of a variety of endogenous organic acids resulting from salicylate’s interference with intermediary metabolism. Agents that cause hypoxemia, cellular asphyxia, seizures, shock, or extensive tissue necrosis can also cause an AGMA, but in these instances, the accumulation of lactic acid generated by anaerobic metabolism is responsible for the AGMA. When the underlying cause is unclear, measuring the serum lactate level may be helpful. The lactate concentration is usually low (< 5 mEq per L) or significantly less than the anion gap in ethylene glycol, methanol, and salicylate poisoning, but high (> 5 mEq per L) or nearly equal to the anion gap in conditions associated with anaerobic metabolism.

Other common toxicologic causes of a low-lactate AGMA include ethanol, which can cause ketoacidosis by disrupting intermediary metabolism in susceptible alcoholics, and toluene, which can cause renal tubular acidosis with bicarbonate wasting. Rarely, this metabolic picture occurs in poisoning by formaldehyde (which is metabolized to formic acid), paraldehyde (presumably as a result of its metabolism to acetic acid), phosphate [41], and sulfur (and possibly sulfates) [42]. It can also be seen with large overdoses of ibuprofen (and probably all nonsteroidal anti-inflammatory agents) and valproic acid (due to high levels of these acidic drugs and their metabolites) [43,44]. Metformin and nucleoside reverse transcriptase inhibitor antiretroviral agents (e.g., zidovudine or azidothymidine) can interfere with normal-lactate metabolism and cause a high-lactate AGMA at therapeutic as well as excessive doses [45,46]. A high-lactate AGMA can rarely occur soon after massive acetaminophen ingestion [47].

An abnormally low anion gap may be seen in severe bromide, calcium, iodine, lithium, magnesium, and nitrate intoxication [39,48,49]. In bromide, iodine, and lithium intoxication, the low anion gap results from spuriously elevated chloride levels, and with nitrate poisoning, it is due to falsely elevated bicarbonate levels.

Serum osmolality can help differentiate the toxic causes of a low-lactate AGMA. An increased osmole gap may be seen early in the course of ethylene glycol and methanol (when high serum levels of the parent compounds are present) but not salicylate poisoning. Although not strictly accurate from a physical chemistry perspective [50], the osmole gap is typically defined
as the difference between the measured serum osmolality and the calculated serum osmolality.

where normal serum osmolality is 290 ± 10 mOsm per kg of H₂O normal osmole gap is 5 ± 7 mOsm per kg (in unselected acutely hospitalized patients [40]).

Normal osmole gap is 5 ± 7 mOsm per kg (in unselected hospitalized patients [40])

This formula assumes that all concentrations are measured in millimoles per liter. If the glucose and BUN concentrations are measured in milligrams per deciliter, dividing them by 18 and 3, respectively, gives their approximate concentrations in millimoles.

Additional causes of an increased osmolar gap include other low-molecular-weight solutes, such as acetone, ethanol, isopropyl alcohol, magnesium, mannitol, and propylene glycol [51]. The approximate concentration of these substances that will increase the serum osmolality by 1 mOsm per kg of H₂O, calculated on the basis of their molecular weights, is shown in Table 117.9. When direct measurements are not readily available, the serum concentration of these agents can be estimated by multiplying this amount by the osmolar gap. Serum osmolality must be measured by freezing point depression rather than the headspace or vapor pressure method to detect the presence of volatile agents such as acetone and toxic alcohols. An increased osmolar gap has also been reported in alcoholic ketoacidosis and conditions causing lactic acidosis [52].

Serum ketones can also help to differentiate the toxic causes of a low-lactate AGMA. Ketosis, as defined by a positive nitroprusside reaction, is relatively common in salicylate poisoning but unusual in ethylene glycol and methanol poisoning. Ketosis is also seen in alcoholic ketoacidosis and in acetone and isopropyl alcohol poisoning.

The urinalysis, serum calcium concentration, and the overall clinical picture can also be helpful in differentiating the toxic causes of a low-lactate AGMA. Crystalluria, hypocalcemia, and back pain or flank tenderness suggest ethylene glycol; visual symptoms implicate methanol; and tinnitus or impaired
hearing point to salicylates. Crystalluria can also be caused by acyclovir [53], felbamate [54], indinavir [55], oxalate [56], primidone [57], and sulfa drugs [58]. Hypocalcemia is also seen in fluoride and oxalate [56] intoxication.

Serum potassium and glucose abnormalities may also provide clues to the etiology of poisoning [18,59,60]. Toxicologic causes of hypokalemia include barium, β₂-adrenergic agonists, calcium channel blockers, chloroquine, diuretics, insulin, licorice, methylxanthines, and toluene. Hyperkalemia can be caused by α-adrenergic agonists, angiotensin-converting enzyme inhibitors, beta-blockers, digitalis, fluoride, potassium-sparing diuretics, and trimethoprim. Common toxicologic causes of hypoglycemia are ethanol, beta-blockers, hypoglycemics, quinine, and salicylate. Common causes of hyperglycemia include acetone, β-agonists, calcium channel blockers, iron, and methylxanthines.

Common toxicologic causes of acute liver dysfunction are acetaminophen, ethanol, halogenated hydrocarbons (e.g., carbon tetrachloride), heavy metals, and mushrooms (e.g., Amanita phalloides and related species) [61]. Acute renal toxicity is most often due to ethylene glycol, halogenated hydrocarbons, heavy metals, nonsteroidal anti-inflammatory drugs, toluene, envenomations, and agents that cause hemolysis or rhabdomyolysis [62]. An elevated creatinine with a normal BUN can be seen in acetone and isopropyl alcohol poisoning because acetone interferes with colorimetric assays for creatinine, resulting in falsely high results. Acute hemolysis (in the absence of glucose-6-phosphate dehydrogenase deficiency) can result from poisoning by arsine gas, naphthalene, and inducers of methemoglobinemia. Rhabdomyolysis is associated with toluene abuse, CNS syndromes (see Table 117.3), and severe physiologic dysfunction (e.g., extreme agitation, deep or prolonged coma, hyperthermia, seizures) of any etiology [63]. The most common agents involved are sympathomimetics, ethanol, heroin, and phencyclidine.

The ECG may provide clues to the cause of poisoning [18]. Ventricular tachyarrhythmias that occur in patients with normal QRS and QT intervals suggest myocardial irritation (i.e., increased automaticity) as the underlying mechanism. Sympathomimetics, digitalis, and cardiac-sensitizing agents such as chloral hydrate and aliphatic or halogenated hydrocarbons, which potentiate the action of endogenous catecholamines, are common causes [64]. In contrast, ventricular tachyarrhythmias that occur in the setting of depolarization and repolarization abnormalities, reflected by QRS and QT interval prolongation, respectively, suggest a reentrant mechanism. Causes include electrolyte abnormalities, organophosphate insecticides, and other membrane active agents (see Table 117.1) [65,66]. Torsades de pointes (polymorphous) ventricular tachycardia strongly implicates an agent that prolongs the QT interval.

Atrioventricular conduction abnormalities (atrioventricular block) and bradyarrhythmias can be caused by beta-blockers, calcium channel blockers, digitalis, membrane-active psychotherapeutic agents, organophosphate insecticides, and α-agonists such as phenylpropanolamine. With α-agonists, they are a reflex (i.e., homeostatic) response to hypertension, but with other causes, they are associated with generalized cardiovascular depression and hypotension.

Ingested chemicals can sometimes be visualized within the GI tract by abdominal radiographic imaging, and such imaging can occasionally be helpful in suggesting the etiology or amount of an unknown ingestion. Although a large variety of chemicals can be detected by routine radiography in vitro, relatively few are visible in vivo [67]. Agents most likely to be visible on plain films are indicated by the mnemonic CHIPES (Table 117.10) [18].

Ingested drug packets may appear as uniform, ovoid or round, marble-sized densities scattered along the GI tract [29,30]. Ingested hydrocarbons may sometimes appear as a double gastric fluid level or “double bubble” because of the air–fluid and fluid–fluid interface lines created when less dense hydrocarbons layer on top of gastric fluids. Computed tomography may be superior to plain films in detecting ingested drug packets but the optimal test in this setting remains unclear [68,69]. Whether contrast should be used or not remains controversial. Abdominal ultrasound can detect ingested pills, particularly enteric-coated and sustained-released formulations [70]. Such imaging may be useful in confirming or refuting some recent specific (CHIPES) ingestions. Because the volume of pills can be determined, plain radiography may be used to guide GI decontamination.

Abnormal findings on chest radiography can be caused by a wide variety of chemicals [18,71]. Diffuse or patchy infiltrates (i.e., pneumonitis or acute lung injury) can be due to the inhalation of irritant gases (e.g., ammonia, chlorine, hydrogen sulfide, nitrogen oxides, phosgene, smoke, sulfur dioxide), fumes (e.g., beryllium, metal oxides, polymers), and vapors (e.g., acids, aldehydes, hydrocarbons, isocyanates, mercury). They can also be seen in patients who have ingested or injected cholinergic agents (e.g., carbamate and organophosphate insecticides), metabolic poisons (e.g., cyanide, carbon monoxide, heavy metals, hydrogen sulfide), paraquat, phencyclidine, salicylates, thiazide diuretics, and tocolytics and in patients with envenomations. Aspiration pneumonitis is quite common and can occur in patients with coma or seizures of any etiology [72]. Acute lung injury can also develop in any patient with prolonged or pronounced anoxia, hyperthermia, or hypotension (e.g., those with severe opioids or sedative–hypnotic or sympathomimetic poisoning). Chronic chemical exposure can cause pulmonary fibrosis, granulomas, or pleural plaques.

The use of antidotes for diagnostic purposes has largely fallen from favor. The availability of point of care blood glucose measurement negates the need for empiric intravenous dextrose in altered patients. Many antidotes may be harmful if used inappropriately, including flumazenil physostigmine, glucagon, nitrites, and chelators. Naloxone remains a reasonably safe therapy in a patient with clinical signs of opiate intoxication. Clinicians should be prepared to manage acute withdrawal and its sequelae.

Analysis of a sample of the toxin itself, or patient urine, blood, gastric contents, or hair [71,73] can sometimes be helpful in identifying the cause of poisoning. Urine is generally the best specimen to analyze because large quantities can be obtained
for extraction procedures and many chemicals are normally concentrated in urine. However, toxicology testing can detect only a small fraction of all chemicals (primarily drugs) and is not always reliable [74]. Immunoassay screens are inexpensive and provide results within minutes, but they are only capable of detecting a few agents. They suffer from many false positives and false negatives. Patients may be misdiagnosed and potentially harmed by clinicians acting solely on the results of immunoassay screens [75]. Comprehensive screens are expensive and require 2 to 6 hours for completion (excluding transportation times). Although results may increase diagnostic certainty or specificity, they rarely change disposition or treatment in patients who are asymptomatic or who have signs and symptoms consistent with the reported exposure [75,76,77,78,79,80]. Noteworthy exceptions are acetaminophen and salicylate, which are widely available, commonly ingested, sometimes misidentified, require specific treatment, and cause few or nonspecific early signs and symptoms. Hence, in most overdose patients, quantitative acetaminophen and salicylate levels are the only toxicology tests likely to be clinically useful.