Published ahead of print on September 4, 2003, doi:10.1164/rccm.200302-248OC
© 2003 American Thoracic Society End-Tidal Carbon Monoxide Corrected for Lung Volume Is Elevated in Patients with Cystic FibrosisDepartment of Pediatric Pulmonology, University Medical Center, Utrecht, The Netherlands; and Department of Pediatrics, Stanford University Medical Center, Stanford, California Correspondence and requests for reprints should be addressed to Suzanne W. Terheggen-Lagro, M.D., Department of Pediatric Pulmonology, University Medical Center, Internal Postal Code KH 01.419.0, PO-Box 85090 3508, AB Utrecht, The Netherlands. E-mail: s.terheggen{at}wkz.azu.nl
Several factors influence levels of end-tidal carbon monoxide (ETCO). We studied determinants of ETCO corrected for inhaled CO (ETCOc) levels in healthy control subjects and compared ETCOc levels and determinants between healthy control subjects and patients with cystic fibrosis (CF). Thirty healthy control subjects (mean ± SD age, 23 ± 6 years) and twenty clinically stable patients with CF, aged 13.5 ± 3.5 years were included. ETCO was measured with the CO-STAT End-Tidal Breath Analyzer (Natus Medical, Inc., San Carlos, CA), and determinants included lung volume (measured with the multiple-breath helium wash-in method), CO-diffusion capacity, and different expiratory flow rates. In healthy control subjects we found a significant correlation between ETCOc and lung volume (r = 0.64, p < 0.05) and with CO-diffusion capacity uncorrected for VA (r = 0.48, p = 0.02). There was no expiratory flow rate dependency in either group. Patients with CF showed no difference in ETCOc levels compared with control subjects (mean 1.2 ± 0.4 ppm vs. 1.3 ± 0.4 ppm, p = 0.32), but patients with CF had lower total lung capacity–helium than healthy control subjects. ETCOc corrected for lung volume was significantly higher in patients with CF compared with control subjects (p < 0.001). We hypothesize that a possible increase in breath CO caused by airway inflammation might be masked by differences in lung volumes between control subjects and patients with CF.
Key Words: ETCOc • inflammatory marker • lung volume • cystic fibrosis Exhaled carbon monoxide (CO) concentration, usually quantitated in end-tidal breath CO (ETCO), has been described in several studies as a candidate marker for airway inflammation in lung diseases like cystic fibrosis (CF) (1–7). Increased levels of ETCO described in these studies might reflect increased oxidative stress, which can cause an induction of the enzyme heme oxygenase (HO-1). This induced HO-1 has its origin in the respiratory epithelium of the bronchi and in airway macrophages and is considered to be an antioxidant enzyme (6–9). However, in a recent study, no increase in ETCO in patients with CF or asthma was found, and there seem to be several pitfalls in measurements of exhaled CO that need to be considered (10). ETCO can be quantified by a number of different techniques. Most techniques are based on electrochemical CO sensors that are inexpensive and give reproducible results but are susceptible to interference from a number of different breath components, for example, hydrogen (H2) (8). Therefore, breath analyzers insensitive to H2 or with separate H2-sensors need to be used for obtaining reliable results (11, 12). CO levels in exhaled air are influenced by levels of CO in inhaled ambient air and by active and passive tobacco smoking. Thus, it is important that ETCO measurements are corrected for inhaled air represented by room air CO (ETCOc) and that smoking is excluded. Under normal physiologic conditions ETCO is largely produced by the oxidative degradation of heme and diffuses from the blood stream to alveolar air. This alveolar origin is supported by the lack of expiratory airflow dependency of ETCO concentration (10). It is unclear whether in patients with inflammatory diseases, like CF, an increase in ETCO concentration is caused by increase of the systemic CO production or by production of CO in the lungs due to induction of epithelial HO-1 or possibly due to lipid peroxidative processes (13–15). Furthermore, a decrease of CO diffusion capacity (DLCO) or alveolar surface area might influence the amount of CO transferred to the alveolar air. In patients with CF, a decline in lung volume and DLCO occurs during the course of the disease because of fibrosis, and this might influence ETCOc (15–17). Finally, ETCOc can be influenced by a number of pathologic and nonpathologic conditions that increase the rate of hemoprotein degradation like anemia, hematomas, and fasting (18). The aim of this study was to investigate correlation of determinants such as lung volume, DLCO, and expiratory flow rate, with ETCOc measurements in healthy subjects using the relatively H2-insensitive CO-STAT End-Tidal Breath Analyzer (Natus Medical, Inc., San Carlos, CA). Furthermore we studied possible differences between healthy subjects and patients with CF with regard to these determinants.
The study was approved by the ethics committee of the University Medical Center Utrecht where the studies were conducted. Thirty (19 female) healthy, nonsmoking volunteers, aged 22.8 ± 6.1 years (range 10–38 years), were included. Twenty patients with CF (nine female) in clinically stable condition, i.e., without actual signs of pulmonary infections, defined as an increase in sputum production or cough, fever, anorexia, or decline in lung function, (especially percent predicted FEV1) were recruited from the Pediatric and Adult Cystic Fibrosis Clinic at the University Medical Center, Utrecht. Patients with CF were aged 13.5 ± 3.5 years (range 7–21 years), percent predicted FEV1 was 73 ± 25, percent predicted FVC% was 84 ± 20, and percent residual volume as part of total lung capacity (RV/TLC%) was 37 ± 16. All patients were colonized with Pseudomonas aeruginosa and/or Staphylococcus aureus bacteria. ETCOc measurements were assessed with the CO-STAT. This is a computer-controlled instrument containing an infrared optical bench for the measurement of CO2 and an electrochemical sensor for the measurement of CO and H2. A single-use patient sampler, consisting of a flexible, 5-French nasal catheter with a filter cartridge is attached to the instrument and inserted 5 mm into the nostril. During 90 seconds of normal nasal breathing, the subject's expired air is continuously sampled by the instrument for quantitation of the mean ETCO and CO2. At the completion of the test the catheter is disconnected from the filter, and the room air CO concentration is measured to correct ETCO for inspired CO (ETCOc). With this device, ETCOc can be measured easily and reproducibly, even in young infants, because the only requirement is spontaneous breathing (12). Expiratory flow rate dependency was studied in a subgroup of 11 control subjects and 4 patients with CF, to distinguish between an alveolar and bronchial origin of ETCOc. Different expiratory flow rates were achieved by using an adjustable expiratory flow resistance, in series with a Lilly-type pneumotachometer. Measurements were performed without resistance and with increasing resistances, resulting in continuous expiratory flow rates of 0.05 and 0.2 L/second. Lung volume (TLC-helium [He]) was assessed in all healthy subjects and patients with CF using the multiple-breath helium wash-in method (Masterscreen FRC; Erich Jaeger, Würzburg, Germany). In 23 out of 30 control subjects and in 5 out of 20 patients with CF, DLCO was measured using a standardized single-breath technique (Masterlab; Erich Jaeger). All lung function measurements were performed according to the American Thoracic Society/European Respiratory Society standards. Other lung function parameters in patients with CF and in control subjects were measured with spirometry and plethysmography (Masterscreen CS and Masterlab systems) and contained the following parameters: percent predicted FEV1, percent predicted FVC, RV/TLC%, TLC-box, and TLC-box %. For reference values, the data of Zapletal and coworkers were used (19).
Statistical Analysis Correlation between ETCOc and TLC-He and DLCO was assessed with Pearson's correlation coefficient. Univariate regression analysis was used to assess the regression equation of ETCOc and TLC-He. Mean values of ETCOc and of ETCOc corrected for TLC-He of control subjects and patients with CF were compared using the nonparametric Mann–Whitney U test. Statistical analysis was performed using the Statistical Package for the Social Science (SPSS version 10.1, Chicago, IL).
Group characteristics are described in Table 1 . Patients with CF were significantly younger than healthy control subjects and had significantly lower TLC-He values (also in percent predicted).
ETCOc Correlation with Parameters in Healthy Control Subjects ETCOc in healthy control subjects showed a significant correlation with TLC-He (r = 0.64, p < 0.05), with TLC-box (r = 0.73, p < 0.05), and with DLCO uncorrected for VA (r = 0.48, p = 0.021). Linear regression analysis was performed, and Figure 1 shows the regression line and equation for ETCOc and TLC-He. Figure 2 shows correlation between ETCOc and DLCO for control subjects and patients with CF.
As indicated in Figure 3 , no significant differences in ETCOc at the two flow rates of 0.2 and 0.05 L/second could be seen in the exhaled air of 11 healthy control subjects and 4 patients with CF.
Comparison of Control Subjects and Patients with CF Patients with CF showed no difference in ETCOc levels compared with control subjects (1.2 ± 0.4 ppm vs. 1.3 ± 0.4 ppm, p = 0.32, Figure 4) . Because there was a strong relation between ETCOc and TLC-He in healthy control subjects and TLC-He was significantly lower in patients with CF, we assessed ETCOc corrected for TLC-He (in percent) in both groups by using the regression equation ETCOc = 0.2043 x TLC-He + 0.1035 (Figure 1). ETCOc corrected for TLC-He values were significantly higher in the patients with CF compared with control subjects (ETCOc [%] 143 ± 41% vs. 100 ± 25%, p < 0.001, Figure 5) .
In our study we found no significant difference in ETCOc levels between healthy control subjects and patients with CF. Using the CO-STAT analyzer, we found that several physiologic factors influenced ETCOc levels in healthy subjects. Most striking was the relation between ETCOc and total lung volume as measured with the multiple-breath helium wash-in method. Apparently, subjects with larger lung volumes and probably larger lung surface areas produce higher concentrations of breath CO. Our study is the first to describe this relationship. Another physiologic factor that might influence exhaled CO level, like DLCO, was significantly related to ETCOc in healthy volunteers and also in patients with CF, although we must state that we only measured DLCO in five patients with CF. DLCO corrected for VA was not significantly related to ETCOc. This is plausible because both ETCOc and CO diffusion are dependent on VA. In this study we found no expiratory airflow rate dependency in healthy control subjects, which suggests an alveolar origin of CO as has been implied previously by Zetterquist and colleagues and by Kharitonov and colleagues (10, 20). Our findings in healthy control subjects suggest that the lack of difference in ETCOc levels between control subjects and patients with CF might be due to important differences in baseline characteristics. When ETCOc in both groups was corrected for lung volume, we found a significantly higher ETCOc predicted in the patients with CF compared with control subjects (Figure 5). Although values for ETCOc corrected for TLC-He were significantly higher in patients with CF, the range of values for patients with CF and healthy control subjects have significant overlap. In healthy subjects ETCOc also strongly correlated with TLC-box, but because airtrapping (shown by significantly higher RV/TLC% in patients with CF) was increased in patients with CF we chose to correct ETCOc for TLC-He, which better reflects the actual ventilating part of the total lung volume. Progressive disease in CF is characterized by a decrease of functional VA and surface because of mucus plugging and air trapping. The progressive differences in TLC-He between patients and control subjects might mask differences in alveolar CO diffusion and/or production. This is in line with the data of Togores and colleagues who showed that levels of ETCO can be underestimated in smokers with severe airflow obstruction (21). Previous studies have shown both increased ETCO levels (1–6) as well as equal levels of ETCO in patients with CF compared with healthy control subjects (10). Differences in ETCOc found in our study groups and the groups described in previous studies might be caused by the previous use of both different electrochemical sensors and/or measuring techniques, as well as the lack of correction for inhaled CO. ETCOc in our study was analyzed with the CO-STAT Analyzer that has a separate sensor for H2 and measures exhaled CO during spontaneous breathing (11, 22). Levels of ETCO measured with other electrochemical sensors like the EC50 (Bedfont Instruments, Kent, UK) were obtained after the following maneuver. Subjects inspired maximal from functional residual capacity, held their breath at TLC for 15 to 20 seconds, and then exhaled into the mouthpiece of the analyzer. Zetterquist and colleagues have shown that exhaled CO levels increased by about 80% from baseline after a 10-second breath-hold and that exhaled CO levels did not increase further after a longer breath-hold (10). It is possible that especially in patients with a ventilation–perfusion mismatch (as in CF) a 15-second breath-hold facilitates a prolonged diffusion of CO from the blood to the alveoli and thus results in higher levels of exhaled CO than when measured with the CO-STAT Analyzer. It has also been reported that other breath components like H2 and hydrogen peroxide can interfere with ETCO measurements (11). Hydrogen peroxide in exhaled air of children with CF with an infectious exacerbation is elevated but is not elevated in patients with stable disease (23, 24). Other substances like ethane and 8-isoprostane are increased in exhaled air in patients with CF, and these correlate well with ETCO levels and may interfere with CO measurements (4, 25). Another possible explanation might be differences in inhaled CO in the different studies, although in all recent studies ETCO is corrected for background CO. Finally, ETCOc is not measured orally but nasally with the CO-STAT Analyzer. There is some discussion about whether or not CO is endogenously produced in the nose and paranasal sinuses. Andersson and colleagues described increased CO levels by breathing through the nose (26, 27). Lundberg and colleagues however were unable to detect any CO signal in nasal air (28). The CO-STAT Analyzer has been validated against the Vitalograph Breath CO-monitor (Vitalograph Inc., Lenexa, KS), and when measuring ETCOc with different expiratory flow rates in this study, ETCOc was measured orally (22). We found no difference between nasally measured ETCOc and orally measured ETCOc with different flow rates. The use of another measuring method therefore cannot explain the differences in ETCOc found between our study and previous studies. Elevated levels of ETCO in inflammatory lung diseases are believed to be caused by increased production in airway epithelium and are attributed to upregulation of HO-1. Although induction of HO-1 is predominantly caused by increased gene transcription rates, there has been no evidence yet that upregulation of HO-1 mRNA in vivo leads to a quantitative increase in enzyme activity or CO production (10, 29–32). In healthy control subjects and in patients with CF, we found no expiratory flow dependency in ETCOc levels. Expiratory flow independence makes a contribution of CO from the bronchi less likely and confirms the alveolar origin of ETCO (10, 20). The predominant source of endogenously produced CO (about 85%) is produced by enzymatic degradation of heme and we cannot exclude increased hemolysis in patients with CF, resulting in higher carboxyhemoglobin and exhaled CO levels because hemolysis or carboxyhemoglobin was not investigated in this study. Finally, endogenously produced CO also originates from nonheme-related release like lipid peroxidation and the release of CO by some bacteria, but under normal physiologic conditions this accounts for only 15% of the endogenously produced CO (33, 34). The lungs of all patients with CF in this study were colonized with Pseudomonas and/or S. aureus bacteria. CO release however by these bacteria has not been described and it is not clear if bacterial CO release plays a role of importance in the endogenous production of CO in a disease like CF. Lipid peroxidation might play a role of importance in additional production of ETCO, which is suggested by an increase of markers of lipid peroxidation in exhaled air and in plasma (4, 25–26, 35–36). In summary, in this study we have shown that levels of ETCOc are strongly related to lung volume in healthy control subjects and are expiratory flow–independent in both control subjects and patients with CF. We found no difference in ETCOc between control subjects and patients with CF, although ETCOc corrected for lung volume was significantly higher in patients with CF. An increase in CO caused by airway inflammation might be masked by differences in lung volumes between control subjects and patients with CF. To assess the utility of ETCOc as a biomarker of airway inflammation in CF, future studies are needed. ETCOc and ETCOc corrected for TLC-He need to be measured longitudinally in the same group of patients with CF, in clinically stable condition and in times of pulmonary exacerbation.
The authors thank Ingeborg Prins for performing all lung function tests.
Supported by the Netherlands Organization for Scientific Research (NWO) Medical Research Foundation (grant 940-37-020). Conflict of Interest Statement: S.W.T-L. has no declared conflict of interest; M.W.B. has no declared conflict of interest; H.J.V. has been a paid consultant ($8,400 per year) with Natus Medical, Inc., of San Carlos, CA, during 2001, 2002, and 2003 and owns 1,782 shares obtained through exercised stock options, has remaining stock options (2,000 shares), and owns additional shares (1,000 shares) bought on the open market; C.K.v.d.E. has no declared conflict of interest. Received in original form February 20, 2003; accepted in final form August 30, 2003
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
