The Authors examined adrenal insufficiency The incidence of nonlethal CO poisoning is not established. CO is a colorless, odorless, nonirritating gas. Its specific gravity is 0.97 relative to air, so it does not stratify. CO has a high affinity for hemoglobin. It reversibly displaces oxygen from hemoglobin to produce carboxyhemoglobin (COHb), resulting in tissue hypoxia.
The symptoms of CO poisoning are protean and vague, resulting in a high rate of misdiagnosis. Correct treatment requires aggressive oxygen therapy and the select use of hyperbaric oxygen (HBO).

SYSTEMIC MANIFESTATIONS

Cardiac

Although CO poisoning affects every organ system (see Table -2), the brain and heart, having the highest metabolic requirements, are most susceptible to CO toxicity. When coronary artery disease is present, a low serum COHb level can produce myocardial ischemia and infarction. In fact, ischemia and infarction can occur in the absence of coronary artery disease. In a study of burn victims, 5 of 18 patients (28 percent) with COHb levels greater than 10 percent sustained myocardial infarctions. Decreased exercise tolerance also results from CO poisoning.
Myocardial ischemia is most prominent in the subendocardial and subepicardial regions of the ventricles. This may produce papillary muscle dysfunction, abnormal ventricular wall motion, and mitral valve prolapse. The electrocardiogram often reflects ischemia, showing ST-segment and T-wave abnormalities. Atrial flutter, atrial fibrillation, premature ventricular tachycardia, and conduction system disturbances are also seen. The ventricular fibrillation threshold is lowered in cases of severe CO poisoning. However, in patients with existing cardiac arrhythmias, mild elevations of COHb (up to 5 percent) are not reported to increase the frequency of single or multiple ventricular ectopic beats during rest or exercise.

Ophthalmologic

Visual disturbances are frequent and correlate with the duration of exposure. Funduscopic examination may reveal flame-shaped retinal hemorrhages. They are caused by hypoxia and may be unilateral. The hemorrhages often resolve without visual deficit. A sensitive indicator of CO toxicity is the presence of bright red retinal veins. These may be seen before the mucous membranes become erythematous and appear at lower COHb levels. Blindness is infrequent and is usually temporary.

Dermatologic

The classic cherry-red skin of CO poisoning is rarely seen; more often, victims are pale or cyanotic. Dermal necrosis with bullae formation may occur anywhere but especially at pressure point areas in the comatose patient and over areas of myonecrosis. Myonecrosis from pressure or prolonged tissue hypoxia may cause rhabdomyolysis, a potential cause of renal failure. Other dermal changes include sweat gland necrosis, alopecia, edema, and erythematous patches.

Other

Other physiologic effects of CO poisoning include noncardiogenic pulmonary edema, bowel ischemia, hepatic failure, vestibular dysfunction, hearing loss, disseminated intravascular coagulation, and thrombotic thrombocytopenic purpura. Even low COHb levels can significantly diminish exercise tolerance in patients with chronic destructive pulmonary disease.

Fetal

Up to 10 percent of patients receiving HBO are pregnant. In pregnancy, the fetus is particularly susceptible, as CO readily crosses the placenta. During exposure, fetal COHb levels lag behind the maternal level, achieving equilibration after 14 to 24 h. After equilibration with maternal blood, fetal COHb levels are 10 to 15 percent higher than maternal levels. In addition, because the fetal oxygen-hemoglobin dissociation curve is shifted to the left, small amounts of COHb can markedly decrease the fetal oxygen tension and content. Fetal elimination of CO is prolonged, resulting in a fetal CO elimination half-life that is three and a half times longer than that of the mother. Pregnant women should receive more aggressive and prolonged oxygen therapy. Longo suggests that the mother should receive oxygen for a period five times longer than the time period needed to complete just the maternal course of therapy. For example, if 4 h is required to completely treat the mother, an additional 20 h of oxygen treatment is required. Fetal heart tones should be obtained, and cardiotachodynomanometry instituted at 20 weeks’ gestation or greater. Up to 60 percent of children born to women surviving CO poisoning have neurologic sequelae. In addition, CO is teratogenic and causes lower-birth-weight infants. Fetal death has occurred in cases of nonlethal maternal poisoning. Fetal demise may be seen immediately or be delayed. Fetal outcome roughly correlates to maternal symptoms at the time of CO exposure; however, there may be no relationship between fetal death and maternal COHb levels.

Pediatric

Children make up 37 percent of patients receiving HBO therapy; children less than 5 years old account for 25 percent of victims receiving HBO. Among acutely exposed children, as many as 17 percent die and 48 percent require cardiopulmonary resuscitation (CPR). Most childhood deaths from CO poisoning are fire-related. Newborns are more susceptible to the effects of CO poisoning due to a higher fetal hemoglobin concentration which shifts the oxygen-hemoglobin curve to the left.
In children, in whom nausea is a common complaint, CO toxicity is frequently misdiagnosed as acute gastroenteritis; in young infants it can be confused with colic. Carbon monoxide toxicity has been implicated as a cause for some cases of sudden infant death syndrome.
Children are frequently affected by riding in vehicles with faulty exhaust systems, especially in the back seats of cars or in the enclosed backs of pickup trucks. In one study of 68 children requiring HBO therapy, 20 (29 percent) were exposed from riding in pickup trucks. Children riding in the same vehicle can have different symptoms with the same COHb levels or have markedly different COHb levels from the same exposure.
The treatment of CO toxicity for children is similar to that for adults; however, myringotomy is more commonly performed in children undergoing HBO therapy.

Neurologic

Since the brain has a high oxygen requirement, most acute symptoms are related to the central nervous system (CNS). In a mass exposure of 184 victims, the most common complaints were headache (90 percent dizziness (82 percent), and weakness (53 percent). Carbon monoxide poisoning has been suggested as an occult cause of syncope. Of CO-poisoned patients, 35 percent complained of spells which mimicked near-syncope. The prevalence of CO poisoning among patients presenting with seizures may be 5 to 7 percent.
The most significant pathologic changes of the brain seen with CO poisoning are white matter lesions. These include white matter demyelination, edema, focal and laminar necrosis, and petechiae. Lesions of the hippocampus are reported in half of the cases. Gray matter lesions also occur in the watershed area between the anterior and middle cerebral arteries. The characteristic pathologic injury of CO poisoning is bilateral lesions of the globus pallidus. These lesions are seen with other hypoxic-ischemic insults and may be unilateral or asymmetric. The globus pallidus lesions are most likely the result of hypoxia with concomitant ischemia. This area has relatively low oxygen requirements, protecting it from pure hypoxia, but it has a poor blood supply, making it vulnerable to hypoperfusion. Low-density lesions of the globus pallidus demonstrated on CT scan or magnetic resonance imaging (MRI) are associated with a high incidence of neurologic sequelae.
Neurologic abnormalities are frequent in acute poisoning (Table -2). The most common neurologic complaints are headache, dizziness, agitation, stupor, seizures, and coma. Other aberrations include behavioral disorders, decreased cognitive ability, gait disturbance, memory deficits, emotional lability, parkinsonism, mutism, tic disorders, and impairment of parietal lobe functions. Long-term psychiatric and neurologic sequelae are grossly apparent in 11 percent of survivors. Memory impairment occurs in up to 43 percent. The most common personality change, affective incontinence, is characterized by emotional lability and may be a consequence of damage to the hippocampus. Resolution of the neurologic sequelae, when it occurs, may take 2 years. These deficits may be permanent.
A syndrome of delayed neurologic sequelae has been described. Delayed sequelae usually occur after coma from a prolonged exposure. The patient recovers, has a symptom-free interval, then undergoes rapid neurologic deterioration. In 10 to 30 percent of poisonings, neuropsychiatric symptoms will appear within 2 to 3 weeks after exposure. The common symptoms are mental deterioration, urinary and fecal incontinence, and gait disturbance. Complete recovery occurs in two-thirds. Older patients are at greater risk for developing delayed sequelae. Diffuse demyelinization of white matter is associated with delayed deterioration. More aggressive oxygen therapy may reduce the incidence of sequelae.

The physical examination is of limited use in establishing the diagnosis of CO poisoning. Nevertheless, once the diagnosis is established, the physical examination does help define the severity of the poisoning. Vital sign abnormalities reflect the severity of the intoxication. The presence of singed nasal hairs, carbonaceous sputum, and thermal injury to the oral mucosa suggests thermal airway injury and severe CO poisoning. Auscultation may reveal wheezing from the exacerbation of preexisting pulmonary disease or from bronchospasm that results from irritation due to inhalation injury. Rales can be heard from the development of noncardiogenic pulmonary edema.

The diagnostic evaluation of the CO poisoning victim must be guided by the situation. For example, the fire victim should be evaluated for cyanide toxicity. When evaluating those attempting suicide, a routine drug screen and measurement of ethanol, phenobarbital, salicylate, and acetaminophen levels may be required. Asymptomatic individuals with minor exposure do not require extensive evaluations.
A COHb level should be obtained in all cases. This should not delay the administration of 100% oxygen. Levels should be obtained by a modified spectrophotometric blood gas analyzer or, less accurately, by measuring expired CO concentration with a hand-held breath analyzer. A high expired CO concentration or serum COHb level confirms CO poisoning, but a low level does not exclude the diagnosis. If a patient has been removed from the CO-containing environment and has been given 100% oxygen, the COHb level may be deceptively normal.
An arterial blood gas measurement calculates the reported oxygen tension from the amount of oxygen that is dissolved in the blood, not from the amount of oxygen that is bound to hemoglobin. The amount of oxygen that is dissolved in the blood is essentially unaffected by CO. In addition, pulse oximetry units mistakenly measure COHb as oxyhemoglobin. Therefore, the oxygen saturation as reported by most arterial blood gas measurements and pulse oximeters will be falsely elevated. The difference between the oxygen saturation that is accurately measured and the false elevated oximetry reading is called the saturation gap. This saturation gap is characteristic of CO poisoning and correlates with the COHb level. For example, if the oximeter reads 93 percent and the arterial blood gas directly measures (not calculates) the saturation to be 88 percent, the COHb is about 5 percent. Likewise, if the pulse oximeter is 95 percent and the measured COHb is 5 percent, the true oxygen saturation is only 90 percent.
An ECG should be performed on patients with cardiac symptoms, coronary artery disease, altered mental status, or a COHb level greater than 10 percent. The ECG is often abnormal. Sinus tachycardia and ST-segment changes are the most frequently seen aberrations, although almost any abnormality may be present.
Fire victims must have a chest radiograph taken. Though frequently normal acutely, it serves as a valuable baseline study. Cyanide levels can be obtained but are of limited value. Assays for other toxic inhalants, such as acrolein, phosgene, or gaseous hydrogen chloride, are unavailable. Spirometry and fiberoptic bronchoscopy may be required.
Serial creatine phosphokinase (CPK) and lactic dehydrogenase (LDH) levels assist in determining myocardial damage. The CPK is frequently elevated. Isoenzymes, especially of the CPK, differentiate habdomyolysis from cardiac necrosis or infarction. In addition, a urinalysis may reveal myoglobinuria.
Alcohol is frequently consumed with both suicidal and accidental CO poisoning, sometimes making a rapid diagnosis difficult. Therefore, a blood alcohol level should be measured. If a suicide attempt is suspected, a drug screen may be sent for analysis. In addition, quantitative aspirin, phenobarbital, and acetaminophen levels should be obtained.
Other laboratory tests assist in confirming CO poisoning. A routine complete blood count (CBC) may show a leukocytosis. Determination of arterial blood gas and electrolyte levels may demonstrate an anion gap acidosis. The acidosis results from lactic acid production. Abnormalities are commonly seen with measurements of the blood glucose, amylase, and transaminases. These findings, however, are of limited clinical significance.
Brain CT is used to define focal or otherwise unexplainable neurologic symptoms as well as to assess for cerebral edema. MRI is a more sensitive neuroimaging modality for detecting CO-induced CNS lesions. MRI can demonstrate cerebral findings associated with CO poisoning, which include necrosis of the globus pallidus, demyelinated lesions of the white matter, and necrotic lesions of the hippocampus. MRI has also defined lesions in the anterior thalami in a CO-poisoned child.
Psychometric testing is performed on CO-poisoned patients at some hospitals. The Carbon Monoxide Screening Battery has been developed to detect patients with subtle cognitive deficit resulting from CO exposure. The battery includes: the general orientation, digit spans, trail-making, digit symbols, aphasia screening, and block design. This battery is used to assess the need for HBO treatment. After HBO treatment, the patient is tested again to determine the effectiveness of therapy.

Initial Approach

In treating CO-poisoned patients, one should remove the victims from the CO-containing atmosphere, keep them at rest to minimize oxygen requirements, immediately administer 100% oxygen, and institute cardiac monitoring. Pulse oximetry will not adequately measure oxygen status. A simple mask which allows for mixture with room air is inadequate. A tight-fitting mask with a bag reservoir is required to deliver a high FIO2. The serum elimination half-life of COHb when breathing room air is 240 to 320 min, compared to 50 to 80 min when breathing 100% oxygen. Oxygen therapy should not be discontinued until the patient is asymptomatic and the COHb level is less than 10 percent. In patients with angina, 100% oxygen should be applied until the COHb is below 2 percent or the patient is pain free.

Hyperbaric Oxygen Therapy

HBO use is empiric but is recommended for severe cases of CO toxicity. HBO is delivered in monoplace or multiplace chambers. The multiplace chamber permits treatment of multiple patients simultaneously and provides room for staff and accessory equipment, e.g., ventilators. Most treatment regimens have been empirically developed; they usually involve 100% oxygen at pressures of 2 to 3 atm for 46 to 120 min. A second treatment is given in 6 to 8 h if symptoms persist. With HBO therapy, the elimination half-life of CO is reduced to 23 min. When HBO is used to treat methylene chloride induced CO toxicity, the CO half-life changes from 13.0 h at room air to 5.8 h.
With HBO therapy at 3 atm pressure, the amount of oxygen dissolved in the blood is increased to 6.6 vol percent. This amount is sufficient to meet the demands of cerebral metabolism independent of the COHb concentration. HBO therapy may also reduce the amount of cerebral edema. HBO therapy has been shown to prevent CO-mediated lipid peroxidation in the brain.
HBO use remains somewhat empiric since no well-designed prospective study has adequately defined the benefits of HBO. Nevertheless, the growing weight of scientific evidence suggests an advantage of HBO in severe poisoning or in cases with protracted symptoms. The prompt application of HBO may reduce mortality. In one study, patients treated within 6 h of CO exposure had a mortality rate of 13.5 percent; of those treated after 6 h, 30.1 percent died. Reports have demonstrated a dramatic reduction in the incidence of neuropsychiatric sequelae when HBO is used.
As many as 10 percent of patients requiring HBO therapy are pregnant. Animal studies have raised concerns regarding the potential adverse effects of high partial pressures of oxygen on the fetus, but the pressures and durations used were greater than those used in humans. The Russians have the largest experience with HBO in pregnancy. They found a decrease in perinatal complications and mortality. A Canadian study involving pregnant women exposed to CO had five cases of severe poisoning. Of the three patients treated with high-flow (non-HBO) oxygen, two had stillbirths; the child of the third patient had cerebral palsy. The two patients treated with HBO had normal outcomes. The greatest advantage of HBO over 100% normobaric oxygen is that it reduces the time required to decrease fetal COHb. Given the safety of HBO and the danger of CO toxicity to the fetus, many recommend liberal use of HBO in pregnant women.
When HBO therapy is required, it should be started promptly. Nevertheless, delays of 7 to 20 h have still resulted in favorable outcomes. Treatment guidelines have been developed to assist in defining those patients needing HBO therapy (Table -3). However, assigning patients into different treatment plans based on the clinical presentation and the COHb level is imprecise. In addition, HBO should be used more readily in cases of pregnancy, age extremes, cardiovascular or neurologic impairment, and metabolic acidosis.
The potential benefits of HBO are based upon sound physiologic principles; the risks of treatment are minimal. During HBO treatment patients may develop otalgia and sinus discomfort. Myringotomy can be performed to ease otic pressure. Tympanic membrane rupture and epistaxis are rare. Of patients with severe CO poisoning who undergo HBO, 5 percent will have a seizure and 6 percent will vomit. The cause of these adverse reactions (CO toxicity versus HBO treatment) is unclear.
Patients with significant neurologic abnormalities, cardiovascular abnormalities, or symptomatic pregnant patients require prompt HBO treatment. Such patients should be stabilized and prepared for transfer to the nearest HBO chamber. The indications for transfer and HBO therapy may be unclear, or the patient's condition may remain unstable. Therefore, the patient's transfer and interim treatment should be discussed with the HBO chamber physician. The role of nitrites with mixed cyanide and CO exposures is not clearly defined and is theoretically harmful. The nitrites given to treat cyanide toxicity cause methemoglobinemia. This shifts the oxygen-hemoglobin dissociation curve further to the left, exacerbating tissue hypoxia. If the patient has been removed from the source and presents to the emergency department alive, sodium thiosulfate, 12.5 g intravenously, should detoxify the cyanide.
Cerebral edema should be treated with mannitol and by elevation of the head of the bed. Hyperventilation should be used in severe cases. The efficacy of glucocorticoids is debated.
Correcting the metabolic acidosis may be harmful; a mild acidosis is often a normal compensatory response. A low pH shifts the oxygen-hemoglobin dissociation curve to the right, increasing oxygen unloading to the tissue. Alkalinization will shift the oxygen-hemoglobin curve to the left, further reducing tissue availability of oxygen and exacerbating the toxic effect of CO. Nevertheless, if the pH is less than 7.20, cardiovascular performance is compromised, and judicious alkalinization is required.


Prognosis

Many variables, such as age, smoking habit, existing cardiopulmonary disease, severity of exposure, prior state of health, and form of therapy, influence outcome. Coma, cardiac arrest, metabolic acidosis, and high COHb levels have been associated with poor neurologic outcome. These factors, however, are inconsistent predictors of outcome and therefore are of limited value. Abnormal CT findings are associated with neurologic sequelae, which will often persist.

1. Burney RE, Wu SC, Nemiroff MJ: Mass carbon monoxide poisoning: Clinical effects and results of treatment in 184 victims. Ann Emerg Med 11:394, 1982.
2. Cobb N, Etzel RA: Unintentional carbon monoxide-related deaths in the United States, 1979 through 1988. JAMA 266:659, 1991.
3. Hyperbaric Center Advisory Committee Emergency Medical Service, City of New York: A registry for carbon monoxide poisoning in New York City. Clin Toxicol 26:419, 1988.
4. Koren G, Sharav T, Pastuszak A, et al: A multicenter, prospective study of fetal outcome following accidental carbon monoxide in pregnancy. Reprod Toxicol 5:397, 1991.
5. Peirce EC, Kaufmann H, Bensky WH, et al: A registry for carbon monoxide poisoning in New York City. Clin Toxicol 26:419, 1988.

    

6. Reisdorff EJ, Shah SM: Carbon monoxide poisoning: From crib death to pickup trucks. Emerg Med Rep 14:181, 1993.
   
   
   
7. Thom SR: Carbon monoxide–mediated lipid peroxidation in the rat. J Appl Physiol 68:997, 1990.

 

---

TABLE 1

Signs and Symptoms at Various Carboxyhemoglobin Concentrations

COHb Level, % Signs and Symptoms

TABLE 2

Reported Complications of Carbon Monoxide Poisoning

System Involved Complication

---

TABLE 3

Treatment Guidelines Based on Severity of CO Poisoning


Mild poisoning
Criteria
COHb levels < 30%
No signs or symptoms of impaired cardiovascular or neurologic function
May complain of headache, nausea, or vomiting
Treatment
Admission of patients with COHb levels > 25%
Symptomatic medication
100% oxygen by non-rebreathing mask until COHb remains < 5%
Patients with underlying heart disease should be admitted and cardiac
function appropriately monitored regardless of COHb level.
Moderate poisoning
Criteria
COHB levels from 30–40%
No signs or symptoms of impaired cardiovascular or neurologic function
Treatment
Admission
Cardiovascular status should be followed closely even in the absence of
clear cardiac effects, especially in those patients with underlying heart
disease
Determination of acid-base status (will be corrected by high-flow
oxgyen)
100% oxygen by non-rebreathing mask until COHb remains < 5%
Severe poisoning
Criteria
COHb levels > 40%
or
Cardiovascular or neurologic functional impairment at any COHb level
Treatment
Admission
Cardiovascular function monitoring
Acid-base status monitoring
100% oxygen by non-rebreathing mask
Transport to a hyperbaric oxygen facility immediately if available, or if
no improvement in cardiovascular or neurologic function is seen within
4 h