Evaluation of CO exposure in active smokers while smoking using breath analysis technique

Chemosphere 53 (2003) 207–216 www.elsevier.com/locate/chemosphere Evaluation of CO exposure in active smokers while smoking using breath analysis technique Wan-Kuen Jo *, Jung-Wook Oh Department of Environmental Engineering, Kyungpook National University, 1370 Sankeok-dong, Bukgu, Daegu 702-701, South Korea Received 14 November 2002; received in revised form 17 April 2003; accepted 28 April 2003 Abstract The current study evaluated the personal CO exposure of active smokers while smoking under controlled conditions, decay rate of CO in the body following active smoking, and CO accumulation in the body from repeated active smoking using a novel device for the direct measurement of alveolar breath CO. Prior to this evaluation, the proposed alveolar CO measurement device was successfully evaluated as regards the effect of humidity, CO recovery, carryover effect, and in comparison with the bag sampling method. The breath concentrations prior to and after a single cigarette were measured using a repeated measure design. Under the controlled conditions employed in the present study, active smoking was found to cause a significant body burden of CO. The post-exposure breath CO level was 1.6–2.0 times higher than the background breath level, depending on the subject and cigarette brand. In addition, the pre- and postexposure breath concentrations were both significantly different among the subjects, yet the ratios of post-exposure to pre-exposure breath concentrations did not differ significantly between the different cigarette brands. The time-series alveolar breath concentrations measured following active smoking showed that the post-exposure alveolar CO concentrations decreased slowly even in the early phase of the decay curves, indicating a mono-compartment uptake and elimination model for the human body. The half-lives estimated in the present study (301, 315, and 385 min) were longer than or comparable to those in previous studies. The breath measurements prior to and after repeated active smoking exhibited a significant increasing trend for both the pre- and post-exposure concentrations. The changes in the pre- and post-exposure breath CO concentrations with repeated smoking ranged from 7% to 23% and from 10% to 15%, respectively, with half-hour intervals between cigarettes, and from 4% to 11% and from 6% to 8%, respectively, with hour intervals between cigarettes. Accordingly, most of the current results indicated that CO was accumulated in the human body with repeated active smoking.
2003 Elsevier Ltd. All rights reserved. Keywords: Alveolar CO measurement; Controlled conditions; Decay curve; Personal exposure; Repeated smoking 1. Introduction There has been a growing public concern over personal exposure to carbon monoxide (CO), as CO is a highly toxic pollutant that binds hemoglobin and inhibits the uptake of oxygen emitted from incomplete * Corresponding author. Fax: +82-53-950-6579. E-mail address: wkjo@knu.ac.kr (W.-K. Jo). combustion (USEPA, 1991). When exposed to moderate amounts of CO, an individual can suffer from headaches, nausea, and incorrect judgment of time intervals. In addition, exposure to CO also increases the risk of coronary artery disease (Apte et al., 1999). Cigarette smoking is a major contributor to personal exposure to CO (Cox and Whichelow, 1985; L€ ofroth et al., 1989; Klepeis et al., 1996). Personal CO exposure from cigarette smoking can be assessed based on the carboxyhemoglobin (COHb) levels, as inspired CO is rapidly 0045-6535/$ - see front matter
2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0045-6535(03)00516-2 208 W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 transferred to the blood and most CO in the body is present in the form of COHb (Lambert et al., 1988). However, even though COHb is a biological indicator of the amount of CO in the body, measuring COHb levels is not always possible due to the invasive nature of collecting blood samples. As such, the measurement of exhaled breath CO can overcome these disadvantages and provide an adequate estimation of blood COHb levels (Cohen et al., 1971; Stewart et al., 1976; Jabara et al., 1980). Therefore, the cost-effective and non-intrusive nature of a breath analysis technique has promoted its use in gauging personal CO exposure associated with cigarette smoking in many community and occupational settings (Verhoeff et al., 1983; Cox and Whichelow, 1985; Akland et al., 1985; Lambert et al., 1988; Wallace et al., 1988). For example, Cox and Whichelow (1985) found that smokers were exposed to elevated CO levels compared to non-smokers, based on measuring the levels of breath CO in 69 smokers, which ranged from 3 ppm to over 100 ppm, 74% being above 10 ppm, whereas the mean levels in 99 non-smokers were lower than those in the smokers, 79% being below 6 ppm. However, previous studies have not evaluated the effect of the exact duration and intensity of smoking. In addition, the time period between smoking and the taking of a breath sample has not been controlled in previous studies. Accordingly, the present study was designed to evaluate the personal CO exposure in active smokers while smoking under controlled conditions. The controlled conditions included the smoking room, breath measurement room, duration and intensity of smoking, and time period between the smoking and the breath measurement. Furthermore, the decay rate of CO in the body following active smoking and CO accumulation in the body after repeated active smoking were also evaluated to provide the necessary data for the development of pharmacokinetic models associated with CO exposure from active smoking. Prior to these evaluations, a device specifically designed for the direct measurement of alveolar breath CO was tested. 2. Experimental methods 2.1. Study protocol The current study included three different experiments associated with CO exposure from active smoking, and prior to conducting the experiments, a novel device for measuring alveolar CO was evaluated. The alveolar breath-sampling device designed by Raymer et al. (1990a,b) for the measurement of expired volatile organic compounds (VOC) was modified by inserting a real-time CO dosimeter into the system. In the first personal CO exposure experiment, the breath concentrations were measured prior to and after the active smoking of one cigarette. The second experiment was conducted to estimate the decay half-life of CO following active smoking, based on measuring a time series of breath concentrations. The final experiment examined the effects of repeated active smoking on CO accumulation in the body based on measuring the alveolar CO concentrations for smokers prior to and after every cigarette according to two specified time intervals. Two separate rooms were utilized for the smoking and breath measurements. The smoking was conducted in an empty 90.2 m3 room at Kyungpook National University, while the breath measurements were conducted in another empty room (61.1 m3 ) next to the smoking room. The only furniture in the rooms was a few tables and chairs, and there were no combustion appliances in either room. Before each experiment, all the windows and doors of both rooms were left open for a minimum of 1 h to minimize any residual levels in the rooms. The windows and doors were then closed and the background CO level in each room measured for 30 min immediately before the experimental smoking. The background CO concentrations in the rooms were less than 0.7 ppm. Five apparently healthy male student smokers living in the student dormitories at Kyungpook National University (KNU) in Korea were recruited for the three exposure studies (Table 1). The subjects were only selected if they were not employed in occupations involving potential exposure to CO. On the survey days, the subjects were asked to walk from their dormitories to the smoking room and not use any motor vehicle, thereby avoiding potential exposure to CO from transportation or other sources. In addition, on the survey days, the subjects were asked to only smoke the experimental cigarettes and not smoke prior to the experiment. The subjects were trained in the breath measurement maneuver at least five times prior to the study. Training was also given to 10 additional nonsmoking subjects who were used to test the alveolar CO measurement device. Informed consent was obtained from all the subjects. The five smoking subjects were asked to smoke at a prescribed rate: 5.0 cm of a cigarette (marked prior to smoking) for 4.5 min. The smoking duration was determined based on a preliminary study and the subjects were trained to control the smoking duration. Immediately after smoking, the subjects moved next door to the breath measurement room. The breath CO measurements were performed for 1 min. Prior to smoking, all the cigarettes were conditioned for at least 48 h at 20 C and 60% relative humidity. 2.2. Construction and evaluation of alveolar CO measurement device A diagram of the proposed alveolar CO measurement device is presented in Fig. 1. The major compo- 209 W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 Table 1 Information on subjects Subject Sex Age Weight (kg) Height (m) kg/m No. of cigarettes smoked per daya Study(s) involved in S1 S2 S3 S4 S5 M M M M M 20 21 21 22 22 71 69 74 70 65 1.75 1.72 1.76 1.78 1.73 41 40 42 39 38 10 20 15 10 15 I and II I and II I, II, and III I and III I and III a Average number of cigarettes smoked per day by each subject. Fig. 1. Device for measuring CO in alveolar breath. nents of the device included a mouthpiece, non-rebreathing two-way valve (Laerdal Medical Co.), Teflon tube (1.3 cm i.d. · 760 cm), and CO dosimeter (CMCD-10P, GASTEC Co.). In the alveolar CO measurement procedure, the subject was asked to inspire ambient room air for 5 s through a new mouthpiece into a non-rebreathing two-way valve, hold the breath for 20 s, then completely blow out for 5 s into a temporary storage Teflon tube. Holding the breath for 20 s after inspiring a full lung capacity has been found to allow enough time for an equilibrium of CO to be established in the arterial blood and alveolar air (Jones et al., 1958). The exhaled breath was then withdrawn into a portable calibrated CO dosimeter for monitoring. The alveolar CO measurements obtained using the proposed device were evaluated as regards the effect of humidity, CO recovery, carryover effect (defined as any interference of residues from the previous measurement), and in comparison with the bag sampling method employing a bag attached with a non-rebreathing twoway valve and CO dosimeter. The humidity test was undertaken by comparing the CO concentrations in clean dry test air and humidified test air. To simulate the humidity of real breath, the humidified test air was prepared by passing dry air (1 or 40 ppm CO) through a humidity generation system in which distilled water was heated and bubbled to generate water vapor. Meanwhile, the dry test air by-passed the humidity generation system. The CO concentrations in the dry and humidified air, plus the relative humidity were recorded every minute for 10 min after a warm-up of 30 min. The CO recovery from the device was determined by comparing the synthetic breath CO concentrations at the inlet and measuring port of the spirometer, while a stream of synthetic breath containing 1 ppm of CO flowed through the alveolar CO measurement device. The whole procedure was repeated 10 times with a time interval of at least 1 h. Another area of interest was how well the spirometer would perform when used to measure a low concentration (1 ppm) of CO shortly after measuring a high concentration (40 ppm). Thus, any possible carryover of CO in the device was examined by generating and measuring synthetic breath containing a high concentration of CO. After the high-level measurement, the CO input concentration was then reduced to 1 ppm and measured in the same manner as for the high level. As with the recovery experiment, the whole procedure was repeated 10 times with a time interval of at least 1 h. The proposed direct alveolar CO measurement method employing a breath collection device and CO dosimeter was also compared with the bag sampling method employing a bag attached with a non-rebreathing two-way valve and CO dosimeter as regards the alveolar breath CO levels of 10 non-smoking male subjects. For smokers, the breath CO concentrations can vary depending on the time interval between the tests of the two methods, thereby interfering with a proper comparison, whereas the breath CO levels of nonsmokers are less likely to vary with time. Consequently, the current comparison was performed using only nonsmokers to avoid a confounding factor due to the time interval between the tests of the two methods. The subjects were apparently healthy graduate students at KNU. Each subject provided a breath sample for the direct measurement device and consecutive breath sample for the bag sampling. In the bag sampling procedure, the subjects were asked to inspire low CO-level room air for 5 s through a new mouthpiece into a non-rebreathing two-way valve, hold the breath for 20 s, then blow out for 5 s into a 1 l capacity Teflon bag. This breathing procedure was continued until about 70% of the bag capacity was filled. This breath pattern (inspiring for 5 s, holding breath for 20 s, and blowing out for 5 s) was consistent with that used for the direct measurement 210 W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 method. The alveolar breath CO content in the sampling bag was then analyzed using the same calibrated CO dosimeter as used for the direct measurement method. A more detailed discussion of the alveolar CO measurement device and evaluation procedures can be provided on request. 2.3. Breath measurement after active smoking: exposure study I The five male subjects were asked to smoke five different commercial cigarette brands on five different days, respectively. The experiments conducted on the same day used different subjects with at least 1 h between each cigarette. The breath CO concentrations were measured prior to and 1 min after each cigarette. The cigarette brands were randomly assigned to the subjects on the survey days. The commercial filtered cigarette brands, identified here as A–E, included two Korean brands (cigarettes A and B), two imported American brands (cigarettes C and D), and an imported Japanese brand (cigarette E), all of which are currently popular in Korea. For each brand, the average mass of tobacco per cm of length was determined by cutting off the butt section of five cigarettes and weighing the tobacco in the remaining portion (Chortyk and Schlotzhauer, 1989; Daisey et al., 1998). The results were as follows: 119.1 ± 1.4 mg/cm, cigarette A; 116.5 ± 1.1 mg/cm, cigarette B; 114.8 ± 0.6 mg/cm, cigarette C; 113.7 ± 0.9 mg/cm, cigarette D; 113.1 ± 1.3 mg/cm, cigarette E. The total length of a cigarette was 8.3 cm for all the brands. 2.4. Time-series breath measurements following active smoking: exposure study II Time-series breath concentrations were measured following the smoking of a domestic brand cigarette (cigarette A) by three male subjects (Table 1). Each subject smoked a cigarette on a different day. The breath concentrations were then monitored 1, 5, 10, 20, 30, 40, 50, 60, 90, 120, 180, 300, and 420 min after smoking. The subjects stayed relaxed in the breath measurement room for most of the non-smoking time, primarily working on a computer. During the experiments, none of the subjects was exposed to any other potential sources of CO and the restroom was the only other place visited. The measured breath concentrations and their corresponding times (post-exposure) were fit to the equations Y ¼ AeBt and Y ¼ CeDt þ EeFt , corresponding to one-compartment and two-compartment pharmacokinetic decay, respectively, where Y is the concentration in the breath at any time t, A, C, and E are constants, and B, D and F are the exponential constants that determine the rate of decay. This curve fitting was accomplished using the non-linear curve fitting routine incorporated into Sigma Plot software (Jandel Scientific Software). The geometric mean of the time between the start and end of the breath sample collection was paired with the breath concentration when fitting the empirical data. The half-life ðt1=2 Þ of the one-compartment model was calculated by dividing the exponential constant into the natural log of 2. The half-lives of the two-compartment model were identified in the same manner using D and F independently. 2.5. Breath measurements prior to and after repeated active smoking: exposure study III The experiment was conducted by asking three male subjects (Table 1) to smoke six cigarettes of a domestic brand (cigarette A) during 3 h with half-hour intervals between each cigarette and eight cigarettes of a domestic brand (cigarette A) during 8 h with hour intervals between each cigarette. Only one subject was tested at a time. The activities of the subjects after each cigarette were similar to those in exposure study II. Breath measurements were conducted prior to and 1 min after every cigarette. 3. Results and discussion 3.1. Evaluation of alveolar CO measurement device The alveolar CO measurement device was evaluated as regards the effect of humidity, CO recovery, carryover effect, and in comparison with the bag sampling method. The humidity test showed that the CO concentrations in the clean dry test air and humidified test air were not significantly different from each other for the two CO levels tested (Table 2). For the low standard gas level (1 ppm), the mean and standard deviation values were 1.1 ± 0.1 ppm for both the clean dry test air and humidified test air. The mean and standard deviation relative humidity values for the low standard gas level test were 95 ± 2%. For the high standard gas level (40 ppm), the mean and standard deviation values were 40.9 ± 0.4 and 41.0 ± 0.3 ppm for the clean dry test air and humidified test air, respectively. The mean and standard deviation relative humidity values for the high standard gas level test were 96 ± 1%. The results of the experimental testing for any CO loss from the alveolar measurement device are shown in Table 3. The mean recoveries for the first and second runs were 104% and 102%, respectively, with the average value of the two means being 103%, indicating no major CO loss from the alveolar measurement device. Table 3 also shows the results of the carryover experiments. In the low level experiment, the mean recoveries for the first and second runs were 102% and 96%, respectively, with the average value of the two 211 W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 Table 2 CO concentrations (ppm) in clean dry test air and humidified test aira Number Low level 1 2 3 4 5 6 7 8 9 10 Mean SD %RSD High level Dry air Humidified air Dry air Humidified air 1.1 1.2 1.2 1.1 1.2 1.1 1.0 1.2 1.2 1.1 1.1 0.1 9 1.0(97) 1.1(96) 1.1(94) 1.2(97) 1.1(95) 1.0(98) 1.1(96) 1.1(93) 1.2(95) 1.1(93) 1.1(95) 0.1(2) 9 41.2 40.7 41.3 40.8 41.2 40.9 40.3 40.9 41.5 40.6 40.9 0.4 <1 40.5(96) 41.5(95) 41.1(98) 41.3(97) 40.7(96) 41.2(95) 40.5(94) 41.3(97) 41.0(96) 41.1(98) 41.0(96) 0.3(1) <1 a Low and high levels are the concentrations prepared to provide 1 and 40 ppm, respectively. The values in parenthesis are the relative humidity of the humidified air (%). The water vapor concentration in the dry air was less than 2 ppm. Table 3 Percent recovery of CO related to system loss and carryover experiments for two experimental runs for each experimental conditiona Number Loss experimentb Carryover experiment Low level 1 2 3 4 5 6 7 8 9 10 Mean SD %RSD Avg High level Run 1 Run 2 Run 1 Run 2 Run 1 Run 2 100 109 100 109 111 100 93 110 100 109 104 6 6 92 100 111 110 100 100 108 100 93 108 102 7 7 91 109 100 108 109 90 100 108 91 110 102 8 8 100 90 100 91 89 108 100 92 100 92 96 6 6 96 103 104 102 105 104 102 107 99 94 102 4 4 101 94 106 103 99 98 97 99 99 102 100 3 3 103 99 101 a Low and high levels are the concentrations prepared to provide 1 and 40 ppm, respectively. For the two runs (Run 1 and Run 2), the position of the two CO monitors was exchanged at the inlet and measuring port of the spirometer. The percent recovery was calculated by dividing the measuring port CO concentration by the inlet CO concentration, then multiplying by 100. Avg indicates the average of two means for each experimental condition. b Loss experiment only included low level. means being 99%, indicating no problems associated with the carryover of CO. As such, the data from both of the above experiments demonstrated the good reproducibility of the proposed alveolar CO measurement system. This was further supported by the relative standard deviations (RSDs) of the data sets that were less than 7% for the loss experiment and less than 8% for the carryover experiment. The alveolar breath CO concentrations obtained from 10 non-smokers using the direct measurement method and bag sampling followed by the dosimeter measurement method were compared. For the direct measurement method, the alveolar CO concentrations for the non-smokers ranged from 2.4 to 5.0 ppm with a mean of 3.6 ppm and standard deviation of 0.8 ppm, while for the bag sampling––CO dosimeter measurement method, the concentrations ranged from 2.6 to 5.4 ppm with a mean of 3.8 ppm and standard deviation of 0.8 ppm. A paired t-test exhibited that the results did not differ significantly from each other, indicating that the 212 W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 two methods were very comparable for alveolar CO measurements. However, the field transport and use of bags is laborious, plus another disadvantage associated with bags is that they must be evacuated and cleaned with clean air prior to being used again. Consequently, the evaluation studies of the proposed alveolar CO measurement device confirm that the new device can be effectively applied to measure CO levels found in breath, thereby overcoming the disadvantages associated with the conventional bag. Table 5 Mean breath CO concentrations (ppm) ± standard deviation measured prior to and 1 min after smoking according to commercial branda Cigarette Prior to After After/prior to A B C D E 4.5 ± 1.3 4.7 ± 1.5 4.3 ± 1.5 4.4 ± 1.4 4.6 ± 1.8 8.1 ± 2.4 8.5 ± 2.6 8.6 ± 3.2 8.1 ± 2.7 8.7 ± 3.1 1.8 1.8 2.0 1.8 1.9 a The number of samples was 25 for each commercial brand (five participants and five samples from each brand). 3.2. Alveolar CO levels after active smoking Tables 4 and 5 confirm that under the controlled conditions employed in the present study, active smoking caused a significant body burden of CO. The postexposure breath CO level was 1.6–2.0 times higher than the background breath level, depending on the subject. The post-exposure (1 min after active smoking) concentrations presented clear peak concentrations that could be practically measured using the breath measurement protocol employed in this study. The increase in the breath concentrations after active smoking was due to the absorption of CO through the lungs during smoking. As such, the contaminant crossed the body membranes into the blood stream, as the CO measured in the exhaled breath was supplied by the blood stream as it passed through the lungs. Furthermore, a DuncanÕs multiple comparison test revealed significant differences in both the pre- and postexposure breath concentrations among the subjects (Table 4). When the analytes share the same letter designation, there was no statistically significant difference in the breath CO concentrations among the subjects. For example, the pre-exposure breath concentrations determined for subject S3 were statistically significantly different from those determined for subjects S1 and S2, yet not statistically significantly different from those determined for subjects S4 and S5. Subject S2 showed the highest pre- and post-exposure concentrations, whereas subject S1 had the lowest pre- and post- exposure concentrations. The mean exhaled breath CO concentrations measured prior to and 1 min after active smoking were 6.5 and 12.2 ppm, respectively, for subject S2, yet 2.9 and 4.5 ppm, respectively, for subject S1. The variation may have been due to physiological factors, differences in the ability to deliver end-expired breath representative of alveolar air, or uncertainties related to the breath measurement (Lambert et al., 1988; Wallace et al., 1988). Unlike the breath-level difference among the subjects, the ratios of post-exposure to pre-exposure breath concentrations did not significantly differ between the different commercial cigarette brands (Table 5). As described in Section 2, the current study involved a repeated measure design, whereby each subject smoked five different commercial brands, with a sufficient time interval between successive experiments to remove any residual effect from the previous brand. As such, it was considered that the effect of the physiological variation among the subjects on the post-exposure concentrations was compensated for among the cigarette brands. Consequently, the results suggest that the commercial cigarette brands tested in the current study caused a similar exposure to CO for smokers with similar smoke habits. A further study may be conducted to confirm this assertion, since systematic elements of the experimental protocol may also have been responsible for the current results. Table 4 Mean breath CO concentrations (ppm) ± standard deviation measured prior to and 1 min after smoking and results of DuncanÕs multiple comparison testa Subject Prior to S1 S2 S3 S4 S5 2.9 ± 0.6 6.5 ± 1.3 4.1 ± 0.6 4.2 ± 1.1 4.6 ± 0.8 After 4.5 ± 0.7 12.2 ± 1.4 8.4 ± 0.9 7.6 ± 1.8 9.2 ± 1.4 After/prior to DuncanÕs grouping Prior to After 1.6 1.9 2.0 1.8 2.0 C A B B B EE AA CC DD BB a In DuncanÕs grouping, the breath CO concentrations for subjects sharing the same letter were not statistically significantly different from each other. The number of samples was 25 for each subject. W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 3.3. Expiration decay curves Fig. 2 shows the decay of CO in the body as measured in the alveolar breath after active smoking. The CO concentrations in the exhaled alveolar breath are a function of the equilibrium exchange of CO between the arterial blood and the air in the alveolar sacs. The breath concentrations after smoking were higher than the presmoking concentrations. The peak breath CO concentrations were not observed directly after smoking, but rather after 5–20 min, depending on the subject. Similar results were also observed in a previous study (Lee and Yanagisawa, 1995), where a novel breath CO sampler was applied to measure the breath samples of three subjects after smoking. In this case it was found that the alveolar CO of two subjects decreased 30 min after smoking, whereas the alveolar CO of the other subject decreased right after smoking. This pattern has also been reported in other previous VOC decay studies (Raymer et al., 1991; Weisel et al., 1992). Consequently, this pattern would appear to be realistic, although the exact cause remains unclear. The decrease in the post-exposure alveolar CO concentrations was slow even in the early phase of the decay curves (Fig. 2), indicating a mono-compartment uptake and elimination model for the human body. This was further supported based on a comparison of an adjusted coefficient of determination (Adj. R2 ) or the F -values associated with the one-and two-compartment fits of each data set, which indicated that the one-compartment fits were superior (Table 6). This pattern is not consistent with those of previous decay studies (Raymer et al., 1991; Weisel et al., 1992) conducted for several VOCs in various microenvironments. Previous studies have found an initial rapid fall in the breath VOC concentration, followed by a much slower decrease, indicating a multicompartment uptake and elimination model for the Fig. 2. CO breath level post-exposure for three subjects (S1, S2, and S3). 213 human body. The difference between previous VOC studies and the current study can be explained by fact that CO behaves differently from lipophilic VOCs, as it is so completely taken up by hemoglobin (Jones et al., 1958; Lambert et al., 1988). Thus, for CO there is no second compartment, since very little CO enters the second compartments, such as the muscles and fat tissue. The half-lives estimated using the one-compartment model (301, 315, and 385 min) were longer than or comparable to those in previous studies. Lee and Yanagisawa (1995) reported that the half-lives of alveolar CO in three male smokers were 214, 243, and 242 min when the subjects smoked while sitting quietly. Similarly, the subjects in the current study stayed relaxed in the breath sampling room during the post-exposure time, primarily working on a computer, thereby providing no significant difference between the two studies as regards the post-exposure activity that could perturb the elimination of CO. Peterson and Stewart (1970) found in 39 separate experiments that the half-life of COHb in the blood of young men ranged from 128 to 409 min. Nonetheless, it is still noted that physiological differences between the subjects employed in the current study and those used in the two previous studies may have an influence on the comparison of the half-lives. 3.4. Alveolar CO levels prior to and right after repeated smoking The alveolar CO levels measured prior to and after every successive cigarette are shown in Fig. 3 based on a period of 3 h with half-hour intervals between each cigarette and 8 h with hour intervals between each cigarette. Moreover, the percentage changes in the breath concentrations were computed by fitting a linear relationship to the number of cigarettes and dividing the slope of the lines by the overall average (Table 7). The breath concentrations showed significantly increasing trends for both the pre- and post-exposure concentrations. All the changes were statistically significant by at least p < 0:05, with the exception of the pre-exposure concentrations for S5 with half-hour intervals. The changes in the pre- and post-exposure breath CO concentrations with repeated smoking ranged from 7% to 23% and from 10% to 15%, respectively, with half-hour intervals between each cigarette, and 4% to 11% and 6% to 8%, respectively, with hour intervals. Therefore, these results indicate that CO was accumulated in the body after repeated smoking with both specified time intervals. This conclusion is also supported by the longer half-lives of CO in the human body than the two specified time intervals, as described in the previous section. The breath CO concentrations measured prior to the 1 h experiment were slightly higher than those prior to 2 the 1 h experiment for subjects S3 and S5, yet the result was reversed for subject S4. However, for all subjects, 214 a Y ¼ AeBt and Y ¼ CeDt þ EeFt , corresponding to the one-compartment and two-compartment model, respectively, where Y is the concentration in the breath at any time t, A, C, and E are constants, and B, D and F are the exponential constants that determine the rate of decay. The values in parenthesis are the standard error, ‘‘t1=2 ’’ represents the half-life of the one-compartment model, ‘‘t1;1=2 ’’ and ‘‘t2;1=2 ’’ represent the half-lives of the first and second compartment for the two-compartment model, respectively, and ‘‘nd’’ represents ‘‘not determined’’. b F -test value. c F -test significance. 0.0002 0.0015 0.0002 20 12 21 0.83 0.74 0.83 346 nd nd 14 173 347 0.0020(0.001) 0.0000(0.040) 0.0000(0.220) 9.4(1) 2.3(70) 3.2(28) 0.049(0.14) 0.004(0.03) 0.002(0.03) 1.0(1) 6.8(70) 5.7(146) 0.84 0.78 0.86 S1 S2 S3 9.9(0.3) 9.0(0.4) 6.3(0.2) 0.0023(0.0003) 0.0022(0.0004) 0.0018(0.0002) 301 315 385 67 44 76 0.0001 0.0001 0.0001 E D C Adj. R2 t1=2 (min) B A One-compartment model Subject Table 6 Decay parameters following subject exposurea Fb pc Two-compartment model F t1;1=2 (min) t2;1=2 (min) Adj. R2 F p W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 Fig. 3. Alveolar CO levels prior to and after every successive cigarette based on 12 and 1 h intervals between cigarettes for three subjects (S3, S4, and S5). Separate experiments were conducted for the two time intervals. the breath CO concentration differences between prior to and after smoking were somewhat larger before starting the 12 h experiment than starting the 1 h experiment. Although an exact cause for this result is unclear, one possible reason was a variation in the physiological condition of the subjects on the different experiment days. Moreover, the breath CO concentrations at the end of the 12 h experiment (sixth concentrations) for two subjects (S3 and S5) were higher than the sixth concentrations in the 1 h experiment for both the pre- and post-smoking concentrations. Accordingly, it is suggested that the CO body burden for smokers is associated with the frequency of smoking. 4. Conclusions The current study evaluated three different forms of CO exposure associated with active smoking, using a novel device for the direct measurement of alveolar breath CO that was successfully evaluated as regards the effect of humidity, CO recovery, carryover effect, and in comparison with the bag sampling method. The first personal CO exposure experiment that measured the breath concentrations prior to and after active smoking confirmed that, under the controlled conditions employed in the present study, active smoking caused a 215 W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 Table 7 Changes in alveolar CO concentration with repeated smoking based on two specified time intervals according to sampling time and subjecta Sampling time Prior After 1/2 h interval 1 h interval S3 S4 S5 S3 S4 DF -test DF -test DF -test DF -test DF -test  +16 +10  +23 +15 +7 ns +13  +11 +7 S5  +9 +6 DF -test +4 +8 Statistically significant: ns, not statistically significant. a Changes in alveolar CO concentrations, D ¼ (slope/l)100%, where l ¼ overall CO mean.  p < 0:05;  p < 0:005;  p < 0:0001: significant body burden of CO. In addition, subject-tosubject variability was found to be an important factor for both the pre- and the post-exposure breath concentrations, whereas the cigarette brand was found to be irrelevant. In the second exposure experiment, the decrease in the post-exposure alveolar CO concentrations was slow even in the early phase of the decay curves, indicating a mono-compartment uptake and elimination model for the human body. Moreover, the half-lives estimated in the present study were longer than or comparable to those in previous studies. Finally, in most cases, the alveolar CO concentrations measured prior to and after every successive cigarette indicated that CO was accumulated in the human body with repeated smoking. Acknowledgements This study could not have been accomplished without the dedicated support of five volunteers. We wish to thank the reviewers for their thoughtful corrections and valuable suggestions on our manuscript. This work was supported by grant no. R05-2001-000-01269-0 from the Basic Research Program of the Korea Science & Engineering Foundation. References Akland, G.G., Hartwell, T.D., Johnson, T.R., Whitmore, R.W., 1985. Measuring human exposure to carbon monoxide in Washington, DC and Denver, Colorado, during the winter of 1982–1983. Environ. Sci. Technol. 19, 911–918. Apte, M.G., Cox, D.D., Hammond, S.K., Gundel, L.A., 1999. A new carbon monoxide occupational dosimeter: results from a worker exposure assessment survey. J. Expos. Anal. Environ. Epidemiol. 9, 546–559. Chortyk, O.T., Schlotzhauer, W.S., 1989. The contribution of low tar cigarettes to environmental tobacco smoke. J. Anal. Toxicol. 13, 129–134. Cohen, S.I., Perkins, N.M., Ury, H.K., Goldsmith, J.R., 1971. Carbon monoxide uptake in cigarette smoking. Arch. Environ. Health 22, 55–60. Cox, B.D., Whichelow, M.J., 1985. Carbon monoxide levels in the breath of smokers and nonsmokers: effect of domestic heating systems. J. Epidemiol. Comm. Health 39, 75–78. Daisey, J.M., Mahanama, K.R.R., Hodgson, A.T., 1998. Toxic volatile organic compounds in simulated environmental tobacco smoke: emission factors for exposure assessment. J. Expos. Anal. Environ. Epidemiol. 8, 313–334. Jabara, J.W., Keefe, T.J., Beaulieu, H.J., Buchan, R.M., 1980. Carbon monoxide: dosimetry in occupational exposures in Denver, Colorado. Arch. Environ. Health 35, 198–204. Jones, J.G., Ellicott, M.F., Cadigan, J.B., Gaensler, E.A., 1958. The relationship between alveolar and blood carbon monoxide concentrations during breathholding. J. Lab. Clin. Med. 51, 553–564. Klepeis, N.E., Ott, W.R., Switzer, P., 1996. A multiple-smoker model for predicting indoor air quality in public lounges. Environ. Sci. Technol. 30, 2813–2820. Lambert, W.E., Colome, S.D., Wojciechowski, S.L., 1988. Application of end-expired breath sampling to estimate carboxyhemoglobin levels in community air pollution exposure assessments. Atmos. Environ. 22, 2171–2181. Lee, K., Yanagisawa, Y., 1995. Sampler for measurement of alveolar carbon monoxide. Environ. Sci. Technol. 29, 104–107. L€ ofroth, G., Burton, R.M., Forehand, L., Hammond, S.K., Seila, R.L., Zweidinger, R.B., Lewtas, J., 1989. Characterization of environmental tobacco smoke. Environ. Sci. Technol. 23, 610–614. Peterson, J.E., Stewart, R.D., 1970. Absorption and elimination of carbon monoxide by inactive young men. Arch. Environ. Health 21, 165–171. Raymer, J.H., Thomas, K.W., Cooper, S.D., Whitaker, D.A., Pellizzari, E.D., 1990a. A device for sampling of human alveolar breath for the measurement of expired volatile organic compounds. J. Anal. Toxicol. 14, 337–344. Raymer, J., Thomas, K., Cooper, S., Pellizzari, E., 1990b. VOC breath measurement study: alveolar breath method. Final Report on EPA Contract No. 68-02-4544, W.A. II-40, Task 1. Raymer, J.H., Pellizzari, E.D., Thomas, K.W., Cooper, S., 1991. Elimination of volatile organic compounds in breath after exposure to occupational and environmental microenvironments. J. Expos. Anal. Environ. Epidemiol. 1, 439–452. Stewart, R.D., Stewart, R.S., Stamm, W., Seelen, R.P., 1976. Rapid estimation of carboxyhemoglobin level in firefighters. J. Am. Med. Ass. 235, 390–392. United States Environmental protection Agency (USEPA). 1991. Air quality criteria for carbon monoxide. Washington, DC, EPA-600/8-90-045A. 216 W.-K. Jo, J.-W. Oh / Chemosphere 53 (2003) 207–216 Verhoeff, A.P., van der Velde, H.C.M., Boleij, J.S.M., Lebret, E., Brunekreef, B., 1983. Detecting indoor CO exposure by measuring CO in exhaled breath. Int. Arch. Occup. Environ. Health 53, 167–173. Wallace, L., Thomas, J., Mage, D., Ott, W., 1988. Comparison of breath CO, CO exposure, and Coburn model predictions in the US EPA Washington-Denver (CO) study. Atmos. Environ. 22, 2183–2193. Weisel, C.P., Jo, W.K., Lioy, P.J., 1992. Utilization of breath analysis for exposure and dose estimates of chloroform. J. Expos. Anal. Environ. Epidemiol. Suppl. 1, 55–69. Wan Kuen Jo is a full professor in the Department of Environmental Engineering at Kyungpook National University in Korea. He has a M.S. in Environmental Science from the New Jersey Institute of Technology, USA, and a Ph.D. from the Graduate Program of Exposure Assessment at Rutgers University, USA. He has published 41 environmental exposurerelated research articles in several professional journals, including 23 international professional journals. Jung-Wook Oh is a chief researcher in the Department of Environmental Research Laboratory, Daegu Industry, Korea. He has a B.S. in Environmental Science from Kyungsan University, Korea, and an M.S. from the Department of Environmental Engineering at Kyungpook National University.