a b s t r a c t

Introduction: Previous studies have shown that carbon monoxide, which is endogenously produced, is increased in community-acquired pneumonia . However, it has not been studied enough whether severity of pneumonia is correlated with increased carboxyhemoglobin (COHb) concentrations in CAP. The aim of this study was to determine whether endogenous carbon monoxide levels in patients with CAP were higher compared with the control group and, if so, to determine whether COHb concentrations could predict severity in CAP.

Materials and methods: Eighty-two patients with CAP were evaluated in this cross-sectional study during a 10- month period. Demographic data, pneumonia severity index and confusion, uremia, rate respiratory, pressure blood, age N 65 (CURB-65) scores, hospital admission or discharge decisions, and 30-day hospital mortality rate were recorded. In addition, 83 Control subjects were included to study. The COHb concentration was measured in arterial blood sample.

Results: The levels of COHb in patients with CAP were 1.70% (minimum-maximum, 0.8-3.2), whereas those in control subjects, 1.40% (minimum-maximum, 0.8-2.9). The higher COHb concentrations in patients with CAP were statistically significant (P b .05). Concentration of COHb correlated with pneumonia severity index (P =

.04, r = 0.187); however, it did not correlate with CURB-65 (P = .218, r = 0.112).

Conclusion: Although COHb concentrations show an increase in patients with pneumonia, it was concluded that this increase did not act as an indicator in diagnosis process or prediction of clinical severity for the physicians.

(C) 2013

Introduction

Background

Community-acquired pneumonia is an Acute infection of the pulmonary parenchyma that is associated with a considerably high morbidity and mortality [1]. In emergency settings, after establishing the diagnosis of pneumonia, almost all major decisions regarding management of pneumonia is to determine its severity. These decisions steer the physician in determining the patient’s hospitali- zation and selecting empirical antibiotic therapy as well as the patient’s estimated risk of mortality.

? Conflict of interest: none declared.

?? Financial support: none.

* Corresponding author. Gazi University Faculty of Medicine Emergency Medicine Department Besevler-Yenimahalle, Ankara, 06500 Ankara, Turkey. Tel.: +90 543

7656176, +90 312 2025517; fax: +90 312 2024162.

E-mail addresses: [email protected] (S.K. Corbacioglu), [email protected] (I. Kilicaslan), [email protected] (F. Bildik), [email protected] (A. Guleryuz), [email protected] (B. Bekgoz), [email protected] (A. Ozel), [email protected] (A. Keles), [email protected] (A. Demircan).

Carbon monoxide (CO) is produced endogenously during the reduction of the HEME protein to bilirubin by the enzyme HEME oxygenase (HO). Of the 3 isoforms of HO enzyme, HO-1 is identified as inducible form. It is mostly stimulated by oxidative stress, hypoxia, inflammation, and heavy metals [2-4]. HEME oxygenase-1 expression occurs particularly in respiratory epithelial cells in lung tissues, Endothelial cells, and alveolar macrophages [5-7]. In previous studies, Carboxyhemoglobin levels were reported to be significantly higher compared with those in control groups in pneumonia-related inflammation and hypoxic stress [8,9].

Importance

Scoring systems such as pneumonia severity index and CURB-65 [10,11], which are developed to determine the severity of pneumonia objectively, frequently cannot be used efficiently due to overcrowding in emergency department and difficulties in immediate decision making by emergency physicians, thereby leading physicians to search for scores and indicators that tend to be easily measured. In this respect, studies regarding the role of endogenous CO in pulmonary diseases and its diagnostic value have been studied [8,9].

0735-6757/$ - see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2012.10.005

S.K. Corbacioglu et al. / American Journal of Emergency Medicine 31 (2013) 520-523 521

However, it has not been studied enough whether severity of pneumonia is correlated with increased COHb concentrations in CAP.

Goals of this investigation

This study aimed to determine whether endogenous CO levels in CAP patients were higher compared with the control group and, if so, to determine whether COHb concentrations could predict severity of pneumonia in CAP patients by means of comparing pneumonia scoring systems.

Materials and methods

Study design

This prospective cross-sectional study was conducted at an urban tertiary care university hospital with an annual census of about 55000 visits from June 2011 to March 2012 during a 10-month period. The local ethics committee approved the study. Written informed consent was obtained from the patients or from legally authorized relatives.

Inclusion criteria included patients 18 years or older with CAP and the absence of any exclusion criteria (Table 1). Community-acquired pneumonia was defined as a new onset infiltrates on chest radiographs, fever (tympanic temperature >=37.8?C), chest symptoms (shortness of breath, coughing, and increase in sputum production or purulence), and abnormal pulmonary physical examination signs (crepitations, bronchial breathing, or pleural effusion). Control sub- jects were selected from healthy individuals who had no respiratory or cardiovascular disease and were currently nonsmokers.

Data collection

Investigators collected demographic characteristics, medical comorbidities, smoking history, vital signs, physical examination findings, excluding criteria, laboratory results, results of imaging studies, taking antibiotic regimens, admissions or discharge de- cisions, Glasgow Coma Scale scores, PSI, and CURB-65 scores in patients with CAP.

Measurement of COHb

Arterial blood sample is obtained by puncture of the radial or Brachial artery, and the COHb level was measured with blood gases analyzer (Roche Diagnostics GmbH., Mannheim, Germany) for 5 minutes in an emergency laboratory setting.

Statistical analysis

All statistical calculations were performed using the SPSS software program (version release 13.0; SPSS, Inc, Chicago, IL). Demographic data related to patients and control subjects were expressed as number, percentage, and mean and SD.

The Kolmogorov-Smirnov test was used to evaluate whether the demographic data and COHb levels of the patients and control subjects were normally distributed. Because the results were within

Table 1

Exclusion criteria of the study

• • • Age b18 years
    • Current smokers
    • Hematological disorders
    • autoimmune diseases
    • CO poisoning
    • Health care-associated pneumonia
    • Hospital-acquired pneumonia
    • Ventilator-acquired pneumonia

the normal distribution, demographic data of the patients were reported as mean +- SD. With respect to age averages between both groups, the difference was measured by the Student t test. Carboxyhemoglobin levels were expressed as median (minimum, maximum values) because of nonnormal distribution. Statistical analysis of these data was performed between each group using Mann-Whitney U test. Although the difference between PSI and CURB-65 groups regarding COHb levels was measured by Kruskal- Wallis analysis test, the correlation between COHb levels and PSI and CURB-65 scores was assessed using Kendall correlation test. All the hypotheses were constructed as 2 tailed, and an ? critical value of .05 was considered as significant.

Results

A total of 138 patients with pneumonia were studied during the study period. Eighty-two patients (42 females-40 males; mean age, 68

+- 15 years) met the criteria and were included in the study. Eighty- three control subjects (39 females-44 males; mean age, 57 +- 14 years) were also included to the study for comparing COHb levels of patients groups. Flow chart of the patients was given in Fig. 1.

The levels of COHb were significantly higher in patients with pneumonia (n = 82; COHb, 1.70%; minimum-maximum, 0.8-3.2) compared with those in control subjects (n = 83; COHb, 1.40%; minimum-maximum, 0.8-2.9). Baseline characteristics of the study population and their levels of COHb are shown in Table 2. Pneumonia severity index and CURB-65 scores of patients and COHb levels are shown in Table 3. Distributions of COHb levels of the patients according to the PSI and CURB-65 groups are shown in Fig. 2. carboxyhemoglobin concentrations correlated within PSI groups (P

=.04, r = 0.187); however, it did not correlate within CURB-65 groups (P = .218, r = 0.112).

The cutoff levels were obtained regarding the differences between groups requiring hospitalization (moderate and high) and groups not requiring hospitalization (low) according to PSI score for COHb by means of the receiver operating characteristic curve analysis. According to these levels, when COHb value was considered 1.650%, sensitivity was found 64% and specificity was 58%, but when COHb value was noted as 1.750%, sensitivity was measured 56% and specificity was 69%. The area under the curve was measured as 0.636 (P = .034) (95% confidence interval [0.514-0.759]).

Of all patients, 26.9% (n = 22) were discharged within 24 hours of admission, and 73.1% were hospitalized. Two patients died in the hospitalized group during a 30-day period.

Discussion

Because the enzyme HEME oxygenase pathway has started to be depicted and its isoform HO-1 is proven to be induced by numerous inflammatory, hypoxic, and oxidative stress [12,13], several studies aiming to identify the relationship of endogenous CO, a side product of this pathway, with various diseases, mainly pulmonary diseases that account for potentially hypoxic disease group, have been published. How to use COHb levels in clinical practice, which were indicated higher particularly in pulmonary diseases in these studies, was our starting point in this study. In the study conducted by Yasuda et al, on the importance of COHb levels in some pulmonary diseases, although COHb mean value was found 0.65% +- 0.03% in the control group, it was measured 1.13% +- 0.14% in asthma patients, 1.05% +- 0.01% in patients with pneumonia, and 0.93% +- 0.03% in patients with idiopathic pulmonary fibrosis, and it was reported statistically significant. Moreover, a significant decrease in COHb levels was observed between the initial levels and after Steroid treatment in asthma patients and antibiotic treatment in pneumonia.

In the study conducted by Paredi et al [14], on the CO levels exhaled (eCO) by the patients with cystic fibrosis, although eCO levels

522 S.K. Corbacioglu et al. / American Journal of Emergency Medicine 31 (2013) 520-523

Fig. 1. Flow chart of the patients

of the patients with cystic fibrosis were found 6.7 +- 0.6 ppm, in the control group, this value was indicated as 2.4 +- 0.4 ppm. In another study conducted by Yamara et al [15], on eCO levels in asthma patients, eCO levels were observed 1.4 +- 0.2 ppm in patients with stable asthma, whereas these levels were found 4.6 +- 0.4 ppm in patients with asthma attacks. In addition, the study showed a statistically significant decline in the weekly postattack values of patients with asthma attacks. In the study by Yasuda et al regarding chronic obstructive pulmonary disease patients, they determined COHb levels as 0.55% +- 0.02% in the control group and 0.081% +- 0.02% in patients with stable COPD, and they indicated this difference as statistically significant [16]. In yet another study by Yasuda et al, in 2004 involving 33 patients with pneumonia and a control group with 22 subjects, although COHb mean distribution for the pneumonia group varied between 1.0% and 1.2%, COHb mean value for the control group was found between 0.6% and 0.8%, and this difference was considered significant.

Similarly, in our study, COHb levels in patients with pneumonia regarding the control subjects were significantly higher (1.7% vs 1.4%). When the values mentioned in the above studies and in our study are closely evaluated, it can be concluded that there is a genuine relationship between increased COHb levels and pulmonary diseases with hypoxic and Inflammatory conditions. Yet, although the differences are found statistically significant, it is still not very clear to us whether we can use them in patient management or how we are going to make use of them because these levels are regarded to be very close to the ones in clinical practice. We at least suggest that merely evaluating COHb levels alone is not enough to diagnose pneumonia and/or support the diagnosis.

Yet, in our study, we also analyzed whether the COHb levels that are known to be higher in patients diagnosed with pneumonia compared with the healthy population can give an idea about the clinical severity of the pneumonia or whether these levels have a correlation with PSI and CURB-65 scores that are assumed to suggest a more objective idea about the clinical severity of pneumonia.

Table 2

Patients’ characteristics and relation to COHb concentrations

Table 3

Distribution of PSI and CURB-65 score categorizations according to patients’

Comorbidity n (%) COHb (%) median P

(minimum-maximum)

hospitalization/discharge and COHb levels

DM Yes 26 (31.7) 1.55 (1.0-3.2) .221

	No	56 (68.3)	1.70 (0.8-3.2)
Hypertension	Yes	34 (41.5)	1.65 (1.0-2.3)	.406	PSI
	No	48 (58.5)	1.70 (0.8-3.2)		Low	24 (55.8)	19 (44.2)	43 (52.4)	1.60 (0.8-3.2)
Heart failure	Yes	11 (13.4)	1.70 (1.2-2.5}	.886	Moderate	30 (90.9)	3 (8.1)	33 (40.2)	1.80 (1.1-3.1)
	No	71 (86.6)	1.70 (0.8-3.2)		High	6 (100)	0 (0)	6 (7.4)	1.85 (1.1-2.3)
COPD	Yes	23 (28.0)	1.70 (1.0-3.1)	.333	CURB-65
	No	59 (72.0)	1.60 (0.8-3.2)		Low	31 (60.7)	20 (39.3)	51 (62.2)	1.60 (0.8-3.2)
Malignancy	Yes	8 (9.8)	1.95 (1.0-2.3)	.402	Moderate	22 (91.6)	2 (8.4)	24 (29.3)	1.75 (1.2-3.2)
	No	74 (90.2)	1.65 (0.8-3.2)		High	7 (100)	0 (0)	7 (8.5)	1.80 (1.1-2.3)
					Total	60 (73.1)	22 (26.9)	82 (100)

Hospitalization, n (%)

Discharged, n (%)

Total, n (%)

COHB (%) median (minimum-maximum)

Abbreviations: COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus.

S.K. Corbacioglu et al. / American Journal of Emergency Medicine 31 (2013) 520-523 523

Fig. 2. Distributions of the COHb levels of the patients according to the PSI and CURB-65 groups.

Although there exists no direct study regarding the correlation between COHb levels and PSI when the clinical severity of pneumonia with regard to the relationship between COHb levels is studied in the literature, in the study by Yasuda et al, it was indicated that there was a significant correlation between the increased COHb levels of patients with pneumonia and C-reactive protein (CRP) values. The sensitivity of CRP values in predicting clinical severity of pneumonia was studied by Chalmers et al [17] in 570 patients with pneumonia. They showed that, in cases where the critical limit value of CRP was considered as 100 mg/dL, the sensitivity was as delicate as PSI for prediction of 30-day mortality (96.3%) and higher than PSI sensitivity (97.6%-81.3%) as an indicator for complicated pneumonia.

Also in another study conducted by Yasuda et al, in patients with pneumonia comparing arteriovenous COHb levels with white blood cell count, positive correlation was reported. With regard to these studies, although it may be deduced that the clinical severity of pneumonia can be correlated with CRP and COHb levels that are correlated with white blood cell count can be associated with the severity of pneumonia, this interpretation seems to be indirect. In our study where the correlation of COHb levels with direct PSI and CURB-65 scores was evaluated, no correlation was observed with CURB-65 score (P = .218), yet a weak positive correlation was found with PSI score (P = .04, r = 0.187). When the limit value, which is assumed to assist in hospitalization decisions by the help of receiver operating characteristic curve, was set as 1.650%,

sensitivity 64%, and specificity 58%, we may conclude that COHb levels proven to be increased in patients with pneumonia is not a reliable indicator in predicting clinical severity.

Limitations

Our most glaring limitation is the difference in mean ages of patients and control groups. We strived to compensate for this difference during the period of picking patients. However, this problem was due to the patient population presenting to our emergency unit. The population had more comorbidity compared with the other emergency units in the same city, and therefore, we hardly had healthy patients who were older than 70 years. In the study by De Siqueira et al [18] regarding COHb reference levels, no relationship between age and COHb levels was found. The mean value for the 15-to-25 age group was indicated as 1.12%; for 25-to-35 age group, 0.78%; and for older patients, it was measured as 0.90%. However, in our study, we did not find any relationship between age and COHb value. This limitation seems to have minor significance with regard to the above data in the literature.

Conclusion

In conclusion, although COHb levels show an increase in patients with pneumonia, we cannot conclude that this increase acts as an indicator in diagnosis process or prediction of clinical severity for the physician.

References

1. Pimentel L, McPherson SJ. Community-acquired pneumonia in the emergency department: a practical approach to diagnosis and management. Emerg Med Clin North Am 2003;21(2):395-420.
2. Maines MD, Kappas A. Cobalt induction of hepatic heme oxygenase; with evidence that cytochrome P-450 is not essential for this enzyme activity. Proc Natl Acad Sci USA 1974;71(11):4293-7.
3. Maines MD, Trakshel GM, Kutty RK. Characterization of two constitutive forms of rat liver microsomal heme oxygenase. Only one molecular species of the enzyme is inducible. J Biol Chem 1986;261(1):411-9.
4. McCoubrey Jr WK, Huang TJ, Maines MD. Isolation and characterization of a cDNA from the rat brain that encodes hemoprotein heme oxygenase-3. Eur J Biochem 1997;247(2):725-32.
5. Lee PJ, Camhi SL, Chin BY, Alam J. AP-1 and STAT mediate hyperoxia-induced gene transcription of heme oxygenase-1. Am J Physiol Lung Cell Mol Physiol 2000;279(1):L175-82.
6. Donnelly LE, Barnes PJ. Expression of heme oxygenase in human airway epithelial cells. Am J Respir Cell Mol Biol 2001;24(3):295-303.
7. Li N, Venkatesan MI, Miguel A, et al. Induction of heme oxygenase-1 expression in macrophages by diesel exhaust particle chemicals and quinones via the antioxidant-responsive element. J Immunol 2000;165(6):3393-401.
8. Yasuda H, Yamaya M, Yanai M, et al. Increased blood carboxyhaemoglobin concentrations in inflammatory pulmonary diseases. Thorax 2002;57(9):779-83.
9. Yasuda H, Sasaki T, Yamaya M, et al. Increased arteriovenous carboxyhemoglobin differences in patients with inflammatory pulmonary diseases. Chest 2004;125(6):2160-8.
10. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336(4):243-50.
11. Lim WS, Van Der Eerden M, Laing R, et al. Defining community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax 2003;58:377-82.
12. Tenhunen R, Marver HS, Schmid R. Microsomal heme oxygenase: characterization of the enzyme. J Biol Chem 1969;244(23):6388-94.
13. Tenhunen R, Marver HS, Schmid R. The enzymatic conversion of heme to bilirubin by microsomal hemeoxygenase. Proc Natl Acad Sci USA 1968;61(2):748-55.
14. Paredi P, Shah PL, Montuschi P, et al. Increased carbonmonoxide in exhaled air of patients with cystic fibrosis. Thorax 1999;54(10):917-20.
15. Yamara M, Sekizawa K, Ishizuka S, et al. Exhaled carbon monoxide levels during treatment of acute asthma. Eur Respir J 1999;13:757-60.
16. Yasuda H, Yamaya M, Nakayama K, et al. Increased arterial carboxyhemoglobin concentrations in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2005;171:1246-51.
17. Chalmers JD, Singanayagam A, Hill AT. C-reactive protein is an independent predictor of severity in community-acquired pneumonia. Am J Med 2008;121(3): 219-25.
18. De Siqueira ME, Martins I, Costa AC, Lima SA. Reference values for carboxyhe- moglobin. Rev Saude Publica 1997;31(6):618-23.