Strange and Schafermeyer's Pediatric Emergency Medicine, Fourth Edition (Strange, Pediatric Emergency Medicine), 4th Ed.

CHAPTER 127. Carbon Monoxide, Cyanide, and Smoke Inhalation

David  Juurlink


• Smoke inhalation victims may have concomitant carbon monoxide and cyanide poisoning.

• Clinical signs and symptoms of carbon monoxide poisoning are notoriously nonspecific and correlate poorly with carboxyhemoglobin values.

• Carbon monoxide poisoning should be considered for an illness affecting more than one member of a family or group from a common environment.

• Cyanide poisoning is marked by rapid onset of central nervous system and cardiovascular dysfunction.

• Key laboratory features of cyanide poisoning include marked acidemia, striking lactate elevation, and a diminished arterial–venous O2 difference.

• Hydroxycobalamin is the antidote of choice for cyanide poisoning and early administration may be life saving.

Carbon monoxide (CO) is the leading cause of toxin-related morbidity and mortality in North America, with tens of thousands of exposures and thousands of deaths each year in the United States. While most of these deaths represent suicides, the majority of pediatric fatalities result from smoke inhalation or misadventures involving combustion of fossil fuels with inadequate ventilation. In contrast, cyanide is rarely an agent of deliberate harm and is most often encountered in smoke inhalation victims.1,2 The pathophysiology of smoke inhalation is complex. From a toxicological perspective, carbon monoxide and cyanide poisoning may coexist.1,2 Critical interventions for the management of both of these poisonings include removal from the source of exposure, meticulous supportive care, and, for patients with concomitant cyanide poisoning, timely antidote administration.


Carbon monoxide (CO) is an insidious poison. It is an imperceptible gas produced by the incomplete combustion of carbon-based compounds such as wood, charcoal, gasoline, or kerosene. In children, CO poisoning is typically unintentional and most often results from malfunctioning home heating systems or proximity to inadequately ventilated generators, charcoal grills, or motor vehicles.3 Rarely, CO poisoning results from inhalation or ingestion of methylene chloride, a hydrocarbon commonly found in paint stripping products and metabolized to CO by the liver.4


The pathophysiology of CO poisoning is complex and incompletely understood. It is rapidly absorbed through the alveoli and binds to heme iron with an affinity roughly 240 times that of oxygen, resulting in the formation of carboxyhemoglobin (COHb).5,6 This produces a functional anemia as well as a conformational change in the structure of the hemoglobin molecule that shifts the oxyhemoglobin dissociation curve to the left. Collectively, these effects reduce the oxygen carrying capacity of blood and impair the release of oxygen to tissues.

In addition, CO produces a host of other cellular effects that contribute to toxicity by interfering with oxygen utilization and promoting inflammation.5 Notably, it binds to myoglobin and other heme-containing structures including mitochondrial cytochromes, impairing cellular -respiration and ATP generation.5,7 Predictably, the most prominent effects of CO poisoning involve the myocardium and brain, tissues with little ability to tolerate cellular asphyxia.6,8The tissue-specific effects of CO are illustrated by the observation that CO-related myocardial dysfunction can occur even when oxygen delivery itself is normal.7 A major consequence of CO poisoning is peroxidation of brain lipids, which promotes inflammation and may underlie the neurologic sequelae seen in patients with severe CO poisoning.7,8

The developing fetus is particularly sensitive to CO poisoning. The fetal oxyhemoglobin dissociation curve lies to the left of the adult curve, fetal hemoglobin binds CO more avidly than maternal hemoglobin, and the half-life of COHb is prolonged in utero.


The clinical signs and symptoms of CO poisoning are notoriously nonspecific and correlate poorly with the COHb level at the scene of exposure.9 (Table 127-1) Moreover, symptoms can persist long after carboxyhemoglobin is undetectable. The diagnosis can, therefore, be -difficultwhen a history of exposure is not readily apparent, or when CO poisoning is not considered until long after the patient is removed from the source of exposure. It is important to consider CO poisoning in patients with nonspecific or vague symptoms, particularly when multiple patients present simultaneously from the same location. Misdiagnosis of carbon monoxide poisoning as a viral illness such as influenza is particularly common in this setting.9

TABLE 127-1

Clinical Manifestations of Carbon Monoxide Toxicity


Headache, nausea, dizziness, and impaired concentration are among the most common presenting symptoms in patients with CO poisoning.5 More severe exposures are characterized by confusion, altered mental status, seizures, and coma. Patients who survive CO poisoning may have long-term neurologic sequelae. Some experts classify these as either persistent neurologic sequelae (PNS) or delayed neurologic sequelae (DNS).10 The main distinction between the two is that DNS is characterized by a period of neurologic improvement (or even restoration of normalcy) prior to neurologic deterioration. The onset of DNS can be dramatic and accompanied by marked abnormalities on neuro-imaging.10,11 In infants, the only suggestion of toxicity may be irritability or difficulty feeding. Children generally exhibit symptoms similar to those of adult patients but may become symptomatic sooner because of their higher metabolic rates.

The physical examination in patients with CO poisoning is of limited utility in making the diagnosis. Vital signs are normal or minimally perturbed in many patients. Cherry red appearance of the skin is often touted as a clue to the diagnosis. While this is commonly seen in patients who have died from CO poisoning, it is rarely encountered in patients presenting to the emergency department.12 Neurologic abnormalities are sometimes subtle and may only be appreciated by a detailed neurologic examination that tests comprehension, recall, and attention. More severe cases may be characterized by delirium, coma, pulmonary edema, arrhythmias, and cardiovascular instability.


It is important for clinicians to recognize that the usual methods of assessing oxygenation not only fail to detect carboxyhemoglobin, they produce falsely reassuring measurements. Conventional pulse oximetry does not distinguish COHb from oxyhemoglobin. Because both are reddish pigments, COHb is perceived as oxyhemoglobin and will artificially elevate the estimated percentage of hemoglobin that exists as oxyhemoglobin. However, oximeters capable of measuring COHb are available13 and have been used for hospital and pre-hospital point of care diagnosis. Similarly, because routine arterial blood gas analyzers measure the partial pressure of oxygen dissolved in plasma to estimate hemoglobin saturation, they also produce falsely elevated estimates of hemoglobin saturation. They will, however, provide accurate information about acid–base status and ventilation. Most laboratories use co-oximetry to measure COHb.5 This can be performed on a venous blood sample, obviating the need for arterial puncture.

The measurement of COHb is subject to several interpretive cautions. While its presence at more than trace percentages is indicative of CO exposure, these values correlate poorly with clinical signs. Because COHb concentrations decline once exposure ceases, the value obtained in hospital may be significantly lower than peak values. Importantly, severe neurologic toxicity can be present in the absence of measurable COHb if measurement was delayed, particularly when high-flow oxygenation has been administered prior to COHb measurement.

Significant acidemia is an uncommon manifestation of isolated CO toxicity but can occur in severely poisoned patients. In victims of structural fires, elevated serum lactate should prompt consideration of concomitant cyanide toxicity (discussed below). Electrocardiography and myocardial enzymes may provide evidence of cardiac injury.14 Computed tomography and, in particular, magnetic resonance imaging may reveal characteristic low-density changes in the globus pallidus and sub-cortical white matter in cases of CO poisoning.15,16


Following removal from exposure, meticulous supportive care is the mainstay of treatment. If the patient has been extricated from a fire, clinicians should maintain a high index of suspicion for concomitant cyanide poisoning, which is often accompanied by mental status abnormalities, lactic acidosis, and hemodynamic instability. Patients with obvious evidence of inhalational injury such as singed nasal hair, soot in the oropharynx, and carbonaceous sputum should be intubated immediately.

High-flow supplemental oxygen is the standard treatment of the CO-poisoned patient (Fig. 127-1). The administration of 100% oxygen (sometimes described as normobaric oxygen or NBO) significantly hastens the elimination of carboxyhemoglobin, with a mean t1/2 of 74 minutes.17 This is a safe and low-cost intervention that many authorities recommend as standard of care. Typically, 4–6 hours of 100% oxygen will suffice although it is prudent to ensure that COHb is below 5% before supplemental oxygen is discontinued. Extended treatment for pregnant women is sometimes advocated although there is no strong evidence supporting this practice.


FIGURE 127-1. Diagnostic and treatment algorithm.

The role of hyperbaric oxygen (HBO) in the management of CO poisoning remains the subject of considerable debate.18,19 Compared to high-flow supplemental oxygen, HBO clearly hastens the elimination of COHb, with a mean elimination half-life of 24 minutes at 3 atmospheres of pressure. Other surrogate measures cited by proponents of HBO include a reduction in neuronal disruption and apoptosis from reperfusion injury, as well as a 10–20-fold increase in the amount of oxygen dissolved in blood.5,18 Several randomized trials have attempted to determine the benefits of HBO in CO poisoning. All are hampered by significant methodologic limitations, as outlined in two, recent reviews,19,20 which conclude that there is insufficient evidence to recommend the use of HBO in the treatment of acute CO poisoning, and that additional, well-designed trials should be conducted.

Nevertheless, many hyperbaric physicians and toxicologists advocate the use of HBO in selected patients, in part because HBO carries few risks.18 While there is no absolute indication for HBO therapy in patients with CO poisoning, commonly quoted indications include a very elevated COHb percentage (typically greater than 25–30% although many centers advise treatment at lower values in the setting of pregnancy), coma, syncope, seizure, or evidence of myocardial ischemia. Unstable patients should not be placed in a hyperbaric chamber or transferred to another center solely for that purpose. The decision to institute HBO therapy should be made in conjunction with a medical toxicologist or other physician with carbon monoxide poisoning expertise and should take into consideration the patient’s clinical status as well as the risks of treatment and transfer to a facility with a hyperbaric chamber.


Children who are asymptomatic can be discharged after oxygen therapy has lowered COHb below 5%, provided they will not return to an environment where ongoing exposure is likely. Parents should be advised of the potential for delayed neuropsychiatric sequelae, including persistent headaches, memory lapses, irritability, and personality changes. Such changes are occasionally sudden and dramatic, but it is difficult to predict, which patients will develop sequelae. Hospitalization solely for observation is generally not warranted. Some experts suggest psychometric testing 4–6 weeks after an episode of significant CO exposure.


Cyanide poisoning is uncommon. There are several potential sources with the commonest scenario being smoke inhalation. Hydrogen cyanide gas is formed as a combustion product of wool, silk, synthetic fabrics, and building materials.1,2 Acetonitrile (methylcyanide) is found in agents used to remove sculpted nails and has caused cyanide poisoning in children.21 Pediatric cyanide poisoning has also occurred from ingestion of cyanide-containing metal cleaning solutions imported from Asia and cyanogenic glycosides, like amygdalin, found in the seeds and pits of certain plants such as apples, apricots, and peaches.22


Hydrogen cyanide gas is rapidly absorbed from the alveoli and may cause profound toxicity within seconds. Ingested cyanide salts, such as sodium cyanide and potassium cyanide, are also rapidly absorbed across the gastric mucosa and may result in toxicity within minutes. Acetonitrile releases cyanide through oxidative metabolism by the hepatic cytochrome P450 system, with a more gradual onset of toxicity, generally within 2–6 hours from the time of ingestion. Ingestion of amygdalin and other cyanogenic glycosides requires hydrolysis to release cyanide, so toxicity may also be delayed up to several hours after ingestion.

Cyanide primarily causes tissue hypoxia by binding with ferric iron (Fe3+) in cytochrome aa3 of the mitochondrial cytochrome oxidase. This impairs oxidative phosphorylation, preventing efficient cellular oxygen use and disrupting ATP production, which results in anaerobic metabolism and severe lactic acidosis. Cyanide also shifts the oxygen–hemoglobin dissociation curve to the left, further impairing oxygen delivery to the tissues. Cyanide inhibits a wide variety of other iron- and copper- containing enzymes although their contribution to clinical toxicity is uncertain. The critical targets of cyanide are those organs most dependent on oxidative phosphorylation, particularly the brain and heart.23


While chronic toxicity from low-level cyanide exposure is described, most cases of cyanide poisoning in children involve smoke inhalation or (rarely) the deliberate or malicious ingestion of cyanide salts such as potassium cyanide. Cyanide-poisoned patients are typically critically ill, although the classic triad of cardiovascular collapse, CNS depression, and lactic acidosis is not consistently seen and may be preceded by a period of hypertension, tachycardia, and simple agitation. In the absence of an obvious exposure history (such as smoke inhalation or a suicide note in a patient with access to cyanide), the diagnosis can be difficult because the clinical features are nonspecific and have other more common causes. The characteristic features of cyanide poisoning are outlined in Table 127-2. Patients may exude a “bitter almond” odor although the ability to detect this is genetically determined, and its absence is of no value. Because cyanosis requires accumulation of deoxyhemoglobin, its presence is not expected in patients with cyanide poisoning. When present it portends a grave prognosis.

TABLE 127-2

Features of Cyanide Toxicity



Cyanide levels in blood correlate well with the severity of poisoning but are not available emergently. Thus, clinicians must rely on other tests to make the diagnosis. Key laboratory features are marked acidemia, lactate elevation (often more than 8–10 mmol/L), and a diminished arterial–venous oxygen difference. An anion gap metabolic acidosis due to lactate accumulation1 is characteristic of cyanide poisoning, and severe acidemia (pH < 7.0) can be seen. The finding of an elevated oxygen concentration in central venous blood or in a venous blood gas is consistent with the diagnosis and results from impaired peripheral oxygen extraction, leaving venous blood with a higher oxygen content than expected.24 In the setting of smoke inhalation, cyanide levels correlate well with carbon monoxide levels;1however, carbon monoxide itself is rarely measured directly. Other nonspecific findings that may accompany cyanide poisoning include leukocytosis, hyperglycemia, or elevations of troponin or creatine kinase if myocardial injury has occurred.


Because cyanide poisoning can be rapidly lethal, survival depends upon immediate diagnosis, meticulous supportive care, and rapid antidote administration, with the latter being critical for survival.25,26Available antidotes capitalize on the fact that the binding of cyanide to mitochondrial Fe3+ is reversible. There are two available choices, hydroxycobalamin and the “Taylor Kit.” Hydroxycobalamin is the antidote of choice2730 and the only option for smoke inhalation victims. The Taylor Kit should be used only if hydroxycobalamin is unavailable. Consultation with a regional Poison Center is strongly recommended.


Hydroxycobalamin is a precursor of cyanocobalamin (vitamin B12). When administered to patients with cyanide poisoning, the cobalt moiety readily binds cyanide to form cyanocobalamin, a harmless compound excreted by the kidneys, leading to restoration of mitochondrial oxidative phosphorylation. Unlike elements of the Taylor Kit, it does not induce hypotension or rely upon methemoglobin production, rendering it a safer antidote for cyanide poisoning in general, and it is the only option for smoke inhalation patients with concomitant carboxyhemoglobin accumulation. Hydroxycobalamin is generally well tolerated, with the exception of transient hypertension (which likely reflects scavenging of nitric oxide) and a reddish-orange discoloration of the skin and urine that typically lasts several days.29,30 Its presence interferes with numerous colorimetric, and co-oximetric laboratory tests in an unpredictable way.31 Some examples include bilirubin, creatinine, glucose, albumin, hemoglobin and urinary pH, glucose, protein, and ketones. Notably, it can interfere with co-oximetry measurements of carboxyhemoglobin, methemoglobin, and oxyhemoglobin, and this should be considered during treatment of smoke inhalation victims.32 Clinicians should review the product monograph for advice regarding interpretation of routine blood and urine tests in patients treated with this antidote. Some authorities recommend combination therapy with hydroxycobalamin and sodium thiosulfate in patients with suspected cyanide poisoning. The initial intravenous dose of hydroxycobalamin is 5 g in adults and 50 mg/kg in children while the initial dose of sodium thiosulfate is 12.5 g in adults and 400 mg/kg in children.


For many years, the antidote for cyanide poisoning comprised a kit consisting of three components: amyl nitrite perles for inhalation along with sodium nitrite and sodium thiosulfate for intravenous administration. This was originally referred to as the “Lilly Kit” and now as the “Taylor Kit.”

The ferric (Fe3+) iron of methemoglobin produced by the nitrite components of the Taylor Kit binds cyanide to form the relatively nontoxic cyanomethemoglobin. Sodium thiosulfate provides a sulfur donor for the rhodanase-mediated conversion of cyanomethemoglobin to methemoglobin and thiocyanate. Thiocyanate is minimally toxic and is excreted by the kidneys.

Amyl nitrite perles held near the nose are used first while establishing an intravenous line and preparing the sodium nitrite solution. After an IV is established, sodium nitrite (6 mg/kg or 0.2 mL/kg of a 3% solution, not to exceed 10 mL) is administered at a rate of 2.5 mL/min. If the patient is known to have significant anemia, a proportionately lower dose is recommended to avoid excessive methemoglobin levels. In an unstable or hypotensive patient the dose may be given more slowly. Methemoglobin levels should be monitored periodically after the infusion and should not exceed 12–15%. Higher levels advocated by older references are unnecessary and dangerous. Adverse effects of nitrite administration include hypotension, headache, blurred vision, nausea, and vomiting. Sodium thiosulfate may be administered following nitrite therapy or concurrently at a separate site. The pediatric dose is 250 mg/kg (1.0 mL/kg) of a 25% solution up to 50 mL (12.5 g). Thiosulfate itself is relatively safe, but accumulation of thiocyanate, especially in patients with impaired renal function, may be associated with nausea, vomiting, arthralgias, and psychosis.

The use of nitrates in critically ill patients creates several challenges. Nitrates cause hypotension, an undesirable adverse effect in patients with pre-existing hemodynamic instability. While the optimal degree of methemoglobinemia is unknown, sodium nitrite may induce excessive methemoglobinemia, unnecessarily compromising oxygen delivery. This is of particular concern in patients extricated from fires who have both cyanide poisoning and elevated carboxyhemoglobin levels.


Several studies suggest a correlation between elevated carboxyhemoglobin levels and cyanide levels in smoke inhalation victims.1 Thus, when an elevated carboxyhemoglobin level is found in a severely ill fire victim, concomitant cyanide poisoning may be present and requires immediate treatment. This is particularly true in a fire victim who requires intubation or has a persistent metabolic acidosis, markedly elevated lactate, abnormal mental status, or cardiovascular instability.


Patients who are asymptomatic and whose exposure has apparently been minimal are observed for 4–6 hours. Those who have ingested cyanogenic glycosides are observed for at least 6 hours for evidence of the onset of toxicity. Those ingesting acetonitrile-containing compounds are observed for 12–24 hours. Symptomatic patients require antidote therapy, care in an intensive care unit, with standard airway management including the administration of 100% oxygen, continuous cardiac monitoring, and the adjunctive use of vasopressors and anticonvulsants as required. Seizures should be treated with benzodiazepines or barbiturates rather than phenytoin. If present, concomitant carbon monoxide poisoning should be treated with 100% oxygen rather than hyperbaric oxygen. Surviving patients should be referred for assessment of neuro-cognitive function.


1. Baud FJ, Barriot P, Toffis V, et al. Elevated blood cyanide concentrations in victims of smoke inhalation. N Engl J Med. 1991;325:1761–1766.

2. Clark CJ, Campbell D, Reid WH. Blood cyanohemoglobin and cyanide levels in fire survivors. Lancet. 1981;1:1332–1335.

3. Cho CH, Chiu NC, Ho CS, Peng CC. Carbon monoxide poisoning in children. Pediatr Neonatol. 2008;49:121–125.

4. Rioux JP, Myers RA. Methylene chloride poisoning: a paradigmatic review. J Emerg Med. 1988;6:227–238.

5. Weaver LK. Clinical practice. Carbon monoxide poisoning. N Engl J Med. 2009;360:1217–1225.

6. Hardy KR, Thom SR. Pathophysiology and treatment of carbon monoxide poisoning. J Toxicol Clin Toxicol. 1994;32:613–629.

7. Lippi G, Rastelli G, Meschi T, Borghi L, Cervellin G. Pathophysiology, clinics, diagnosis and treatment of heart involvement in carbon monoxide poisoning. Clin Biochem. 2012;45:1278–1285.

8. Thom SR, Keim LW. Carbon monoxide poisoning: a review epidemiology, pathophysiology, clinical findings, and treatment options including hyperbaric oxygen therapy. J Toxicol Clin Toxicol. 1989;27: 141–156.

9. Dolan MC, Haltom TL, Barrows GH, Short CS, Ferriell KM. Carboxyhemoglobin levels in patients with flu-like symptoms. Ann Emerg Med. 1987;16:782–786.

10. Pepe G, Castelli M, Nazerian P, et al. Delayed neuropsychological sequelae after carbon monoxide poisoning: predictive risk factors in the emergency department. A retrospective study. Scand J Trauma Resusc Emerg Med. 2011;19:16.

11. Yogaratnam J, Hariram J, Lee DS, et al. Delayed neuropsychiatric sequelae and recovery following carbon monoxide poisoning. Ann Acad Med Singapore. 2011;40:516–517.

12. Gorman DF, Clayton D, Gilligan JE, Webb RK. A longitudinal study of 100 consecutive admissions for carbon monoxide poisoning to the Royal Adelaide Hospital. Anaesth Intensive Care. 1992;20:311–316.

13. Barker SJ, Curry J, Redford D, Morgan S. Measurement of carboxyhemoglobin and methemoglobin by pulse oximetry. Anesthesiology. 2006;105:892–897.

14. Fotbolcu H, Incedere O, Bakal RB, Tanalp AC, Astarcioglu MA, Dindar I. Reversible myocardial stunning due to carbon monoxide exposure. Cardiovasc J Afr. 2011;22:93–95.

15. Lee WK, Yu ZH, Lee CC. Delayed neurological sequelae after carbon monoxide poisoning. Aust N Z J Psychiatry. 2008;42:430.

16. Vion-Dury J, Jiddane M, Van BY, Rumeau C, Lavielle J. Sequelae of carbon monoxide poisoning: an MRI study of two cases. J Neuroradiol. 1987;14:60–65.

17. Weaver LK, Howe S, Hopkins R, Chan KJ. Carboxyhemoglobin half-life in carbon monoxide-poisoned patients treated with 100% oxygen at atmospheric pressure. Chest. 2000;117:801–808.

18. Hampson NB, Piantadosi CA, Thom SR, Weaver LK. Practice recommendations in the diagnosis, management, and prevention of carbon monoxide poisoning. Am J Respir Crit Care Med. 2012; 186:1095–1101.

19. Buckley NA, Juurlink DN, Isbister G, Bennett MH, Lavonas EJ. Hyperbaric oxygen for carbon monoxide poisoning. Cochrane Database Syst Rev. 2011;CD002041.

20. Buckley NA, Isbister GK, Stokes B, Juurlink DN. Hyperbaric oxygen for carbon monoxide poisoning: a systematic review and critical analysis of the evidence. Toxicol Rev. 2005;24:75–92.

21. Caravati EM, Litovitz TL. Pediatric cyanide intoxication and death from an acetonitrile-containing cosmetic. JAMA. 1988;260:3470–3473.

22. Geller RJ, Barthold C, Saiers JA, Hall AH. Pediatric cyanide poisoning: causes, manifestations, management, and unmet needs. Pediatrics. 2006;118:2146–2158.

23. Hall AH, Rumack BH. Clinical toxicology of cyanide. Ann Emerg Med. 1986;15:1067–1074.

24. Johnson RP, Mellors JW. Arteriolization of venous blood gases: a clue to the diagnosis of cyanide poisoning. J Emerg Med. 1988;6: 401–404.

25. Megarbane B, Delahaye A, Goldgran-Toledano D, Baud FJ. Antidotal treatment of cyanide poisoning. J Chin Med Assoc. 2003;66:193–203.

26. Reade MC, Davies SR, Morley PT, Dennett J, Jacobs IC. Review article: management of cyanide poisoning. Emerg Med Australas. 2012;24:225–238.

27. Baud FJ. Cyanide: critical issues in diagnosis and treatment. Hum Exp Toxicol. 2007;26:191–201.

28. Borron SW, Baud FJ. Acute cyanide poisoning: clinical spectrum, diagnosis, and treatment. Arh Hig Rada Toksikol. 1996;47:307–322.

29. Borron SW. Recognition and treatment of acute cyanide poisoning. J Emerg Nurs. 2006;32:S12–S18.

30. Borron SW, Baud FJ. Antidotes for acute cyanide poisoning. Curr Pharm Biotechnol. 2012;13:1940–1948.

31. Carlsson CJ, Hansen HE, Hilstead L, et al. An evaluation of the interference of hydroxycobalamin with chemistry and co-oximetry tests on nine commonly used instruments. Scand J Clin Lab Invwest. 2011;71:378–386.

32. Lee J, Mukai D, Kreuter K, et al. Potential interference by hydroxycobalamin on cooximetry hemoglobin measurements during cyanide and smoke inhalation treatments. Ann Emerg Med. 2007;49: 802–805.