| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
a Author for correspondence. Fax 919-541-3527; e-mail PLEIL.JOACHIM{at}EPAMAIL.EPA.GOV
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Key Words: indexing terms: exhaled breath sampling • exposure assessment
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
For volatile organic compounds (VOCs), sampling and analysis of breath is preferred to direct measurement from blood samples because breath collection is noninvasive, potentially infectious waste is avoided, and the measurement of gas-phase analytes is much simpler in a gas matrix rather than a complex biological tissue such as blood.1 To exploit these advantages for the assessment of exposure to environmental pollutants, the US Environmental Protection Agency (EPA), academic research institutions, and the medical community have been studying exhaled breath with a variety of sampling and analytical methods. An overview of such work is available in recent review articles by Wallace et al. (3)(4); a more general history of the use of breath measurement in medicine has been written by Phillips (5).
For a given scenario or human activity, a pre- and post- set of breath samples is sufficient to confirm the occurrence of an exposure to specific VOCs. To quantify the exposure and to glean additional information regarding residence time in the body, a series of postexposure breath samples can describe the time-dependent elimination of the VOC from the subject and thus be used to infer total body burden and the distribution of target chemicals in the blood and other body tissues. Some examples of this approach and the attendant theory can be found in the literature (6)(7)(8)(9). Clearly one of the critical issues in the use of elimination kinetics is the collection of samples in an appropriate time frame. For example, the residence times of VOCs in blood are on the scale of a few minutes, so to properly model this behavior requires a series of samples collected rapidly after an exposure. Second, the sample should preferably consist of alveolar air, i.e., expired breath involved in the blood gas interface deep in the alveoli with minimal contribution from the tracheal dead volume.
To address these concerns, we have developed a simple, direct collection method for individual alveolar breath samples and adapted conventional gas chromatography–mass spectrometry (GC-MS) analytical methods for trace-concentration VOC analysis. The "single breath canister" (SBC) sampling method is based on direct exhalation of alveolar air into a 1-L volume stainless steel canister with an internally passivated surface. The SBC technique requires minimal instruction for an untrained subject, is self-administered, and can (in theory) be used to collect individual samples of adjacent breaths; a 30-s sample-to-sample time frame was found to be the practical limit. Subsequent analysis is performed in the laboratory with GC-MS methods especially modified to accommodate the 1-L samples and the high concentrations of water and carbon dioxide in breath. Refs. (10) and (11) present the detailed sampling and analytical procedures (which are beyond the scope of this paper) as well as brief discussions and listings of other relevant studies concerning exhaled breath measurement.
Here we briefly describe the sampling method and present some examples of exhaled breath measurement as a diagnostic tool for determining exposure to microenvironmental pollutants. Specifically, we use the same experimental procedures to explore three different scenarios. The SBC method is used to objectively demonstrate the occurrence and relative magnitude of a recent exposure and to show how the resulting breath concentration vs elimination time data can be interpreted to estimate the magnitude and duration of the resulting blood-borne dose. Individual demonstrations of different scenarios were performed on a few subjects in the realm of methods development for sampling and analysis. As such, no generic interpretation of typical exposures can (or should) be made, nor can we interpret biological responses in context of body type, sex, or age. The intent is to demonstrate the SBC breath measurement method as a tool for exposure assessment, and, by inference, to future clinical applications.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
1) Rockford scenario (exposure to vinyl chloride and cis-1,2-dichloroethene from contaminated well water): Illinois Department of Public Health, Springfield, IL; contact: Ken McCann
2) Montana scenario (exposure to chloroform and bromodichloromethane from high-intensity swimming in chlorinated pool water): University of Montana, Missoula, MT; contact: Brent Ruby
3) Chamber scenario (exposure to trichloroethene from a controlled exposure chamber experiment): US Air Force, Wright Patterson AFB, OH; contact: Jeff Fisher; Research Triangle Institute, Research Triangle Park, NC; contact: Paul Kizakevich
In total, we collected a variety of breath samples from 17 healthy adult subjects who were exposed in the various scenarios (9 men, 8 women, ages 22 to 42). Detailed elimination studies were performed on a subset of 10 subjects for demonstration of the SBC techniques.
sampling equipment and procedures
The sampling apparatus consists of an evacuated 1-L canister
fitted with a small Teflon tube used as a mouthpiece as shown in Fig. 1
. The subject closes her lips on the tube, and as she exhales,
she opens the canister valve and the breath is collected, filling the
evacuated volume. The subject is instructed and trained to begin sample
collection at the "bottom" (or end) of a normal resting tidal
breath to achieve an alveolar sample. The tracheal dead volume is
expelled before the canister sample valve is opened. We note that the
typical "at rest" tidal volume is ~500 mL and the typical dead
volume is ~150 mL; as such, the expiratory reserve volume that is
collected is as close as possible to purely alveolar. An investigation
of the alveolar nature of an SBC sample in contrast to other techniques
is available in ref. 10. Fig. 2
depicts a subject self-administering a sample collection. The
SBC method can be used to collect a timed series of samples easily with
30-s sample-to-sample resolution as described in detail in ref.
10. The canisters used were from two commercial
manufacturers, Scientific Instrumentation Specialists, Moscow, ID, and
Biospherics, Hillsboro, OR. They are constructed of 306 stainless steel
with internal surfaces passivated with an electropolishing technique
generally referred to as the "SUMMATM" process.
Although the breath samples were all collected in 1-L volume canisters,
microenvironmental samples (where volume is noncritical) were collected
in a variety of canister types including 1.8-, 2.8-, 3-, and 6-L
volumes, subject to availability. Both "integrated" whole-air
samples (averages over a specific time period) and "grab" whole-air
samples were used to assess the inspired air during and after the
exposures.
|
|
analytical equipment and procedures
Although subsequent laboratory analysis can be performed with any
of a variety of GC-MS methods for air, for our purposes here, the
standard EPA Method TO-14 (12) and the concentrator
equipment were somewhat modified to accommodate some of the more
reactive species found in exhaled breath; this is described in detail
in ref. 11. For some quality-assurance (QA) analyses, where
the compound(s) of interest allowed, the standard TO-14 method and
commercially available instrumentation were also used. Briefly, each
breath sample is transported to the laboratory, pressurized with a
neutral gas, and a dilution factor is calculated on the basis of pre-
and postpressurization absolute pressure. The analytical
instrumentation is fully automated to extract an aliquot from the
canister, cryogenically concentrate, thermally desorb/inject onto a
capillary column, and then analyze with a mass spectrometer. Samples of
microenvironmental air from the exposure area and control samples of
inspired air after the exposure were also collected and analyzed with
the same equipment. Carbon dioxide assays of breath samples were
performed also by GC-MS; however the injection technique was adjusted
to use only a tiny (typically 50-µL) aliquot via injection loop, or
with a short cycle of the "valveless concentrator" as described in
ref. 11.
For the Rockford scenario, microenvironmental samples were analyzed with TO-14 methodology with a Nutech 320–01 cryoconcentrator (Graseby-Nutech, Smyrna, GA) and a Hewlett-Packard GC-MS system GC5880 and MS5970 (Hewlett-Packard, Avondale, PA and Palo Alto, CA, respectively). The GC column was an SPB-1 60 m x 0.32 mm (i.d.) x 1.0 µm film thickness (Supelco, Bellefonte, PA). Exhaled breath samples were analyzed with a prototype "valveless concentrator" (patent 5447556) developed under a cooperative research and development agreement (CRADA 0026–92) between Graseby-Nutech and the EPA that was interfaced to an ITS40 (Magnum) GC-MS ion trap instrument (Finnigan MAT, San Jose, CA). The analytical column was an XTI-5 30 m x 0.25 mm (i.d.) with 1.0 µm stationary phase (Restek Corp., Bellefonte, PA).
Preliminary microenvironmental and exhaled breath samples for the Montana study were analyzed with the valveless concentrator interfaced to a Saturn II GC-MS ion trap instrument (Varian, Walnut Creek, CA) by using an RTX-5 30 m x 0.25 mm (i.d.) with 1.0 µm stationary phase (Restek). For the main body of the Montana samples and all of the Chamber samples, analyses were performed with a Graseby-Nutech 3550A cryoconcentrator with a 16-canister autosampler. This was interfaced to the Finnigan ITS40 instrument detailed above.
Quantification was achieved with external calibrators prepared for each sample set for all analytes. System linearity was always confirmed over the sample range with five-point calibration. Daily response factors and system integrity were determined via single-point calibration and canister blanks. A minimum of 25% replicate analyses (of real samples) was performed to continually assess system precision. Calibrators were independently prepared and assessed by our on-site contractor, ManTech Environmental Technology, by using certified calibrators from Alphagaz, Morrisville, PA, and Scott Specialty Gases, Plumsteadville, PA.
data postprocessing and interpretation
Raw GC-MS extracted ion peak area data for analytes of interest
were corrected to reflect current calibration response factors.
Measurements of carbon dioxide concentrations were then used to assess
a subject's data set and used to normalize to a nominal 5% for
alveolar breath to assure internal consistency for that subject. For
the healthy adult subjects studied in this work, 5% was a reasonable
normalizing factor for the "true" alveolar concentration, although
this technique may require revision in the case of ill or impaired
subjects with pulmonary disorders. Because each subject was seated,
calm, and at rest during the sample collection, there was no
variability due to exercise or hyperventilation; we used the carbon
dioxide correction only as a method to account for slight variations
from the sample pressurization step in the laboratory and from the
subject's technique in filling the canister. We found through many
trials that even the most experienced subject can occasionally entrain
some ambient air into the sample, or prematurely close the valve, both
of which would result in a dilution of alveolar air in the finally
processed sample. In either case, the assay of carbon dioxide provides
an accurate correction factor for the overall "alveolar" nature of
an individual sample and thus is used to scale the concentration of the
analyte to correctly reflect the breath concentration. Issues
concerning the assay of carbon dioxide in breath, exercise- and
breathing technique-related perturbations, and the consistency of data
sets from one individual are discussed in refs. 10 and
11.
Time after exposure vs analyte concentration data were modeled to
generate a mathematical approximation of the breath elimination of
VOCs. We chose a multiterm exponential decay model by Wallace et al.
(7) and applied it to the data with GraphPad Prism
(GraphPad Software, San Diego, CA), a nonlinear modeling program
optimized for a least-squares fit (no weighting). Specifically, the
model takes the form:
 | (1) |
Cair (µg/m3) is the
contribution of the inspired air during elimination (in our
experiments, Cair = 0). The resulting
equation is integrated from time = 0 to infinity, then multiplied
by the alveolar breathing rate R (m3/min) to
establish the mass of contaminant Mcont (µg)
eliminated via exhalation. The resulting quantity takes the form:
 | (2) |
Additionally, if Eq. 1
is evaluated at time t = 0, and
we have access to the blood/breath partition coefficient P
(m3/1000 L), we can generate an estimate of the lower bound
of the highest blood concentrations caused by the exposure as follows:
 | (3) |
Finally, the estimated parameters ki
from Eq. 1
are used to estimate the half-life
t1/2i (min) of the contaminant
in the "ith" modeled bodily compartment by:
 | (4) |
to a specific
biological system or compartment, we generally consider the 1st term to
represent the blood, the 2nd "highly perfused tissues" (liver,
other organs, some muscle), the 3rd "moderately perfused tissues"
(muscle, organs, liquid reservoirs), and the 4th "poorly perfused
tissues" (adipose and connective tissue). Some well-known examples of
pharmacokinetic modeling of VOCs and interpretations of bodily
compartments are listed in the references
(8)(9)(13)(14)(15). |
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Exposure parameters for the three experiments are given in Table 1
. These include the compound name, the duration of the exposure,
the measured average values for the air concentration, and the water
concentration. Note that a variety of sequences are represented:
short-term, low-concentration; moderate-term, low- and
moderate-concentration; and long-term, high-concentration. Also, we
have included conversion equivalency between µmol/L and µg/L for
the compounds of interest for reference.
|
Table 2
presents the numeric values for the parameters that result in
the best nonlinear least-squares fit for each data set for the
exponential elimination function as described in Eq. 1
. Each data set
is identified by the experiment name, subject/experiment identification
code, and compound measured; the number of measurements and elimination
times are given to assist interpretation of the model parameters. The
number of compartments (or exponential terms) most appropriate to the
data set was empirically determined with a "residuals runs"
analysis to set the lower limit, and occurrence of redundant terms to
set the upper limit. For these data, a time frame >2 h is necessary to
be able to mathematically discern the elimination from the 3rd
compartment. In all cases the model fit to the concentration data was
excellent, with most R2 test limits >0.99.
|
Graphs of representative data of VOC elimination from each experiment
are presented in Fig. 3
. Fig. 3A
–C shows data and model for R1ABL1 (vinyl chloride)
and R1ABL2 (cis-1,2-dichloroethene) as an example of a
short-term, low-concentration exposure; M1FMS1 (bromodichloromethane)
and M1FMS2 (chloroform) as an example of moderate-term, low- and
moderate-concentration exposure; and C3F110 and C3M210
(trichloroethene) as examples of long-term, high-concentration
exposure, respectively.
|
Table 3
documents the calculated biological parameters as described in
Materials and Methods. For the lower-bound estimate of the
total dose, Mcont, as described in Eq. 2
, we
used a conservative breathing rate estimate of 0.007 m3/min
(7 L/min). For the estimates of the highest blood concentration
generated by the exposure (see Eq. 3
), we used blood/breath partition
coefficients published by Gargas et al. (16): 1.16 for
vinyl chloride, 9.85 for cis-1,2-dichloroethene, 6.85 for
chloroform, and 8.11 for trichloroethene. Though unavailable in the
literature, we estimated the partition coefficient for
bromodichloromethane at 29.9 through interpolation of coefficient vs
boiling point for trihalomethanes.
|
Note that even for a brief 10-min shower exposure of 25 µg/m3 (inhalation) and 4 µg/L (dermal contact in water), we calculate 0.9 µg absorbed dose of vinyl chloride and a blood concentration of 0.01 µg/L. The 2-h swimming pool exposure (for trained athletes) resulted in an estimated dose of 100 µg and a blood concentration of ~2 µg/L. The 4-h exposure to 100 ppmv (550 000 µg/m3) of trichloroethene,which (for comparison) represents half of the integrated Occupational Safety and Health Administration permissible exposure limit of 100 ppm for occupational exposure for an 8-h workshift (17), resulted in estimated absorbed doses from 23 to 90 mg with peak blood concentrations averaging ~1200 µg/L.
Calculated half-lives (as defined in Eq. 4
) in the 1st compartment
(blood) were calculated from 0.5 to 4 min, the 2nd compartment (highly
perfused tissues) from 8 to 62 min, and the 3rd compartment (moderately
perfused tissues) from 1.5 to 19 h across all subjects and all
types of exposures to halogenated VOCs.
Overall, these experiments gave reasonable estimates of absorbed dose
and blood concentrations in line with the cited literature. Critical
examination of the time-dependence of the elimination curves as graphed
in Fig. 3
and the model results in Tables 2
and 3
show that the first
few minutes after the exposure are critical in estimating the dose and
peak blood concentration because the breath concentration is decreasing
by half in the first 1–2 min. Gathering representative data during
this time is complicated by several factors. First, it is
experimentally difficult to assign a precise time (t =
0) for when the elimination begins because some time does elapse during
the transition of the subject from exposure to "clean" air. Second,
the physical movement of the subject and the increased activity
associated with sample collection tend to affect the ventilation rate
during this time and may perturb the "alveolar" nature of the
sample. Finally, the logistics of collecting a quick series of samples
in a few minutes require some coordination of effort in sample
container handling and records keeping. The behavior of the
"slower" tissues (2nd and 3rd compartments) is more easily deduced
because accurate sample timing is not as critical.
On the basis of the results from these experiments, we are investigating modifications to our techniques to get a more accurate representation of the 1st-compartment elimination. We will experiment with the use of a portable clean-air supply and "on-demand" breathing regulators (modified scuba diving gear) to precisely decouple the exposure period from the elimination period. Second, we will attempt to extend the model into a hypothetical "zeroth" compartment that could be considered a very fast, low-capacity compartment comprising the mucous membranes of the mouth and tracheal airway by collecting adjacent breaths during the first 30 s of the elimination time.
The analysis of exhaled breath is an excellent exposure assessment tool for halogenated VOCs that can unambiguously demonstrate that an exposure has occurred, and with a time series of samples after the exposure, gives the possibility of modeling the washout of the contaminant from the body. The appropriate multiterm exponential decay model has been used to establish approximate residence times for pollutants in the body, and also to estimate minimum blood concentrations and integrated dose, all without an invasive medical procedure. The techniques presented here could be extended to the broader clinical setting where the analyte could be a volatile bioresponse marker for a disease state, for an administered drug or anesthetic, or for changes subsequent to a medical procedure. Future work should focus on individual differences among a statistically significant number of subjects sorted by body type, sex, and age to further develop the method for eventual mainstream clinical use.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
The information in this document, funded by the EPA, has been subjected to Agency review and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
1 Nonstandard abbreviations: VOCs, volatile organic compounds; EPA, Environmental Protection Agency; GC-MS, gas chromatography–mass spectrometry; SBC, single breath canister; QA, quality assurance; and QC, quality control. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Z-2..
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  W. Cao and Y. Duan Breath Analysis: Potential for Clinical Diagnosis and Exposure Assessment Clin. Chem., May 1, 2006; 52(5): 800 - 811. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. von Basum, H. Dahnke, D. Halmer, P. Hering, and M. Murtz Online recording of ethane traces in human breath via infrared laser spectroscopy J Appl Physiol, December 1, 2003; 95(6): 2583 - 2590. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
