Carbon monoxide (CO) poisoning is a major cause of morbidity and mortality but prospectively assessing the severity of poisoning is difficult. Objective criteria are needed in order direct appropriate treatment. One or more biomarkers may prove to be reliable, objective tools for establishing the diagnosis, prognosis and clinical management of CO poisoning. This article reviews current literature and highlights the need for prospective studies to identify biomarkers that accurately predict clinical outcomes.
By Dr Phi-Nga Jeannie Le and Prof. S.R. Thom
Carbon monoxide (CO) is the agent that, world-wide, causes the most injury and death by poisoning worldwide [1]. Clinical diagnosis of acute CO poisoning and treatment decisions are based on nonspecific criteria. When CO poisoning is suspected, measurement of blood carboxyhemoglobin (COHb) by spectrophotometry is the standard method to confirm exposure. The usefulness of quantifying the COHb level is limited, however, because COHb correlates poorly with symptoms and with clinical outcome. CO pathophysiology is not due simply to hypoxia and this may account for the poor correlation. A combination of CO-mediated hypoxia/ischemia, ATP depletion, excitotoxicity, oxidative stress, lipid peroxidation and immunological responses lead to cardiovascular and brain injuries. Objective measures are thus needed to establish the severity of CO poisoning and guide treatment decisions.
The standard treatment for CO poisoning is administration of supplemental oxygen. This will hasten the dissociation of CO from hemoglobin and provide enhanced tissue oxygenation, but supplemental oxygen at ambient pressure does not counter CO-induced excitotoxicity, oxidative stress or the pathological inflammatory cascade. Hyperbaric oxygen (HBO2) is the most extensively studied treatment for CO poisoning and the only treatment investigated in randomised clinical trials. HBO2 reduces the incidence of cardiovascular mortality and neurological morbidity in animal models because it can inhibit many of the CO-induced pathological pathways – including (paradoxically at first pass) CO-mediated oxidative stress such as brain lipid peroxidation. Clinical efficacy of HBO2 therapy for reducing neurological sequelae has been shown in several trials, but – in this age of meta-analysis – debate persists over its role [2, 3]. Patients suffering from loss of consciousness or with a COHb level above 25% are often referred for HBO2 treatment. If reliable laboratory markers were available to stratify risk, they could be used to assist treatment decisions and to appropriately match patients in clinical investigations [see Figure 1].
Blood/plasma markers S100 beta and enolase
Several groups have investigated whether blood levels of neuron-specific proteins can be used to assess CO poison severity and morbidity risk. The first report on this subject found that S100 beta protein, an enzyme found in astroglia, and neuron-specific enolase were not reliably correlated with severity of poisoning or clinical outcome in a series of 20 patients evaluated at time of admission and for up to 48 hours in hospital [4]. Larger patient series have found that S100 beta is elevated in unconscious patients; hence those more seriously poisoned than conscious victims [5, 6]. More recently, elevations of S100 beta were shown to exhibit a strong negative correlation with Glasgow Coma Scores; the level decreased more rapidly in patients treated with HBO2 [7].
Activated neutrophils
One method for assessing intracellular neutrophil activation is measuring by flow cytometry the surface expression of b2 integrins. A significant elevation in b2 integrin expression was observed in patients exposed to CO for over three hours and this was associated with the presence of platelet-neutrophil aggregation, an early element in biochemical and functional deficits in an animal model [8]. No significant elevations in total leukocyte or neutrophil counts were observed in this study of 50 consecutive patients, although another group reported that CO victims (a series of eight patients versus eight controls) had abnormally high neutrophil counts [9].
Myocardial proteins
Elevations in creatine kinase-MB, troponin-T, brain natriuretic peptide (BNP) or N-terminal pro-BNP often are associated with transient myocardial hypokinesis in the first hours to days after acute CO poisoning, even in patients with normal electrocardiograms [10-16]. Among adults, including those who do not exhibit myocardial infarction on ECG, there is a heightened risk of death long after symptomatic recovery from acute CO poisoning [10, 11, 17]. Therefore, there is a need for objective evaluation of cardiac injury.
Inflammatory markers
Elevations in plasma cortisol and granulocyte-colony stimulating factor were observed in a series of eight CO victims [9]. A complex pattern of plasma protein abnormalities was reported in a series of 63 samples from patients selected on the basis of a range of poison severity, and values were compared against 42 age and sex-matched controls [18]. The concentrations of 99 proteins out of 180 analysed displayed a significant difference from the controls in at least one of the CO poisoning groups and 14 proteins were significantly different from the controls in all CO patient groups [Table 1].
Conflicting results have been reported for abnormalities in inflammatory and coagulation pathway proteins associated with low-level, environmental CO exposures. Elevated C Reactive Protein was found in one trial but not in another by the same group [19, 20]. Plasma fibrinogen was increased in one study [21], decreased in one [22] and unchanged in two [19, 23]. Other studies have reported an elevation in soluble intercellular adhesion molecule-1, along with decreases in factor VII, serum albumin and prothrombin time [19, 23, 24].
Cerebrospinal fluid (CSF) abnormalities
Neurological dysfunction in animal studies is caused by an inflammatory cascade involving alterations in myelin basic protein. Elevations in CSF interleukin 6 and myelin basic protein have been found in patients who go on to develop delayed neurological sequelae [25-27].
Discussion
Despite over a century of study, CO poisoning remains a fertile field for research. Not only is more work needed on pathophysiology, but a prospective study is urgently needed where patients are followed to identify which laboratory markers accurately predict morbidity and mortality. It is obvious that no single pathway explains the wide range of abnormalities found with CO poisoning. Indeed, because the pathways overlap, there may be competing processes such that abnormalities linked to one CO dose and/or duration of exposure may not be found with an alternative pattern of poisoning. The high frequency of myocardial enzyme elevations has highlighted the risk for cardiac injuries. S100 beta seems to correlate with CO-induced coma, but that is not a difficult physical finding to establish. Studies to date have been helpful for elucidating pathophysiological events in clinical CO poisoning, but for the most part sophisticated laboratory measurements have not yet improved diagnostic or clinical management decisions for patients.
References
1. Raub JA et al. Toxicology 2000; 145: 1-14.
2. Juurlink DN et al. Cochrane Database Syst Rev 2005: CD002041.
3. Logue C. The cochrane library - feedback.Available at:Http://www.Cochranefeedback.Com/cf/cda/citation.Do?Id=9569#9569. 2006:
4. Rasmussen LS et al.Acta Anaesthesiol Scand 2004; 48: 469-473.
5. Brvar M et al. Resuscitation 2004; 61: 357-360.
6. Yardan T et al. Am J Emerg Med 2009; 27: 838-842.
7. Cakir Z et al. Am J Emerg Med 2006; 28: 61-67.
8. Thom S et al. Am J Respir Crit Care Med 2006; 174: 1239-1248.
9. Schnittger V, Rosendahl K, Lind F, Palmblad J. Investig Med 2004; 52: 523-530.
10. Satran D et al. J Am Coll Cardiol 2005; 45: 1513-1516.
11. Henry CR et al. Jama 2006; 295: 398-402.
12. Pach D et al. Przegl Lek 2005; 62: 465-467.
13. Davutoglu V et al. Inhal Toxicol 2006; 18: 155-158.
14. Kalay N et al. Am J Cardiol 2007; 99: 322-324.
15. Unal E, Yazar A, Oran B. Inhal Toxicol 2007; 19: 587; author reply 589.
16. Teksam O et al. Eur J Emerg Med 2010.
17. Hampson NB, Rudd RA, Hauff NM. Crit Care Med 2009; 37: 1941-1947.
18. Thom S et al. Clin Toxicol 2010; in press:
19. Ruckerl R et al. Am J Respir Crit Care Med 2006; 173: 432-441.
20. Ruckerl R et al. Environ Health Perspect 2007; 115: 1072-1080.
21. Pekkanen J et al. Occup Environ Med 2000; 57: 818-822.
22. Steinvil A et al. Environ Res 2008; 106: 51-61.
23. Liao D et al. J Expo Anal Environ Epidemiol 2005; 15: 319-328.
24. Baccarelli A et al. J Thromb Haemost 2007; 5: 252-260.
25. Kamijo Y, Soma K, Ide T. Am J Emerg Med 2007; 25: 483-485.
26. Ide T, Kamijo Y. Am J Emerg Med 2008; 26: 908-912.
27. Ide T, Kamijo Y. Am J Emerg Med 2009; 27: 992-996.
The authors
Dr Phi-Nga Jeannie Le
University of Pennsylvania
and
Dr Stephen R. Thom
Professor of Emergency Medicine
Chief, Hyperbaric Medicine
Medical Director, PennStar Flight
University of Pennsylvania
3620 Hamilton Walk
Philadelphia PA 19104-6068
Tel: +1 215-898-9102
Fax: +1 215-573-7037
