Measurement Variability in SingleBreath Diffusing Capacity of the Lung*
Naresh M. Punjabi, MD, PhD, FCCP; David Shade, JD; Anshul M. Patel, MD;
and Robert A. Wise, MD, FCCP
Study objectives: The single-breath diffusing capacity of the lung (DLCO) is a commonly
performed pulmonary function test. The current American Thoracic Society (ATS) recommendations for reproducibility of DLCO measurements suggest that two measurements for the DLCO
agree within 10% or 3 mL/min/mm Hg of the average value. The European Respiratory Society
(ERS) recommends that two measurements should agree within 10%. The objectives of the
present study were to examine whether the current reproducibility criteria were met in a general
pulmonary function laboratory and to determine whether alternative criteria might be appropriate.
Design: Cross-sectional study.
Setting: University-based pulmonary function laboratory.
Patients or Participants: Patients referred for spirometry, helium lung volumes, and DLCO
measurement.
Interventions: None.
Measurements and results: In a sample of 6,193 patients referred for clinical testing, 98.3% had
two DLCO values that fulfilled the current ATS criteria for reproducibility. The coefficient of
variation (CV) and the percentage difference between two repeat measurements were inversely
associated with the baseline DLCO and the FEV1. As the baseline DLCO (percentage of predicted)
or FEV1 (percentage of predicted) decreased, there was an increase in the CV and the percentage
difference. In contrast, the absolute difference between repeat measurements was relatively
stable irrespective of the baseline DLCO or FEV1 values. Other patient factors, such as gender and
race, were not associated with measurement variability. Using an absolute difference of 2 to 2.5
mL/min/mm Hg between two DLCO measurements as alternative criteria for reproducibility,
91.5% and 95.8% of the patient sample fulfilled these criteria, respectively.
Conclusions: Reproducibility of the DLCO measurement is generally much better than current
standards allow. Future standards should consider an absolute difference rather than a percent(CHEST 2003; 123:1082–1089)
age difference criterion for DLCO reproducibility.
Key words: diffusing capacity; measurement variability; pulmonary function testing; reproducibility; transfer factor
Abbreviations: ATS ⫽ American Thoracic Society; CI ⫽ confidence interval; CV ⫽ coefficient of variation;
Dlco ⫽ single-breath diffusing capacity of the lung for carbon monoxide; ERS ⫽ European Respiratory Society;
FRC ⫽ functional residual capacity; PFT ⫽ pulmonary function test; RV ⫽ residual volume; STPD ⫽ standard temperature and pressure, dry; SVC ⫽ slow vital capacity; TLC ⫽ total lung capacity; Vi ⫽ inspired volume
ingle-breath diffusing capacity of the lung for
S carbon
monoxide (Dlco) is the most common
method for measuring diffusing capacity of the lung.
*From the Division of Pulmonary and Critical Care Medicine,
Johns Hopkins University, Baltimore, MD.
Supported by National Institutes of Health grant No. HL04065.
Manuscript received March 27, 2002; revision accepted September 12, 2002.
Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail:
[email protected]).
Correspondence to: Naresh M. Punjabi, MD, PhD, FCCP, Division of Pulmonary and Critical Care Medicine, Johns Hopkins
Asthma and Allergy Center, 5501 Hopkins Bayview Circle,
Baltimore, MD 21224; e-mail: [email protected]
In 1987, the American Thoracic Society (ATS) published a consensus statement on the single-breath
technique for measuring Dlco and recommended
that two acceptable measurements should be within
10% or 3 mL CO (standard temperature and pressure, dry [STPD])/min/mm Hg of the average
Dlco.1 The consensus statement was motivated by
the need to standardize the measurement of Dlco
and decrease interlaboratory and intralaboratory
variability. In 1993, the European Respiratory Society (ERS) published a similar report for standardization of measuring Dlco and recommended that two
measurements should agree within 10%.2 In contrast
to the ATS report, the ERS statement did not
1082
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
Clinical Investigations
include absolute difference as part of the acceptability criteria. The consensus statement required that
the “average of two determinations which agree to
within 10%” should be reported. The ERS statement
is ambiguous because it is not clear whether the two
Dlco measurements should be within 10% of the
mean or 10% of the smallest or the largest value. To
address new technological developments in the
equipment used to measure Dlco and clarify the
differences between the ATS and ERS statements,
the ATS published an updated report in 1995. The
revised statement maintained the standards for
Dlco reproducibility that were originally professed
in the 1987 ATS report.3 Since that revision, there
have been no studies testing the utility of the
published criteria for Dlco variability in a general
pulmonary function laboratory. Moreover, although
it is generally expected that underlying lung disease
can increase measurement variability, the effects of
obstructive and restrictive ventilatory defects on
variability of Dlco measurement have not been
thoroughly investigated. Therefore, the objectives of
the present study were as follows: (1) to determine
the percentage of patients that meet the current
recommendations for Dlco variability in a large
urban university clinical laboratory, (2) to assess the
feasibility of alternative reproducibility criteria, and
(3) to explore the effect of varying degrees of
obstructive and restrictive ventilatory impairment on
Dlco measurement variability.
Materials and Methods
We conducted a retrospective review of pulmonary function
test (PFT) results for patients referred to the pulmonary function
laboratory over a consecutive 3-year period. Approval for use of
the pulmonary function data were obtained from the institutional
review board on human research. Patient confidentiality was
maintained by removing all identifying information from the
clinical data. PFT records were reviewed to select patients who
underwent spirometry and had Dlco measured during the same
visit. To eliminate inclusion of multiple measurements, only the
first PFT for each patient was selected for the present study.
Standard techniques for PFTs, in general accordance with the
current guidelines, were used.4 – 6 Equipment was manufactured
by W.E. Collins (Braintree, MA) or was custom designed for the
laboratory. Spirometry was performed to obtain the FEV1 and
the FVC. Lung volume measurements were available on a subset
of the patient sample. Functional residual capacity (FRC) in this
subset was measured with the multiple-breath helium dilution
method. Slow vital capacity (SVC) and expiratory reserve volume
were measured in triplicate at the end of each multiple-breath
procedure. Residual volume (RV) was calculated by subtracting
the average expiratory reserve volume from FRC. Total lung
capacity (TLC) was determined by adding the largest SVC to the
calculated RV. Alveolar volume and Dlco were measured using
the single-breath method in accordance with ATS guidelines with
the following exceptions.3 First, the algorithm used to calculate
breath-hold time measured the time from the onset of inspiration
to the beginning of sample collection. Second, alveolar volume
was not corrected for anatomic dead space. Patients attempted a
maximum of five maneuvers to obtain two acceptable and
reproducible Dlco measurements. An acceptable maneuver was
defined as an inspiratory breath of at least 90% of FVC,
breath-hold time of 9 to 11 s, and a rapid inspiration, as defined
in the ATS statement.3 Reproducible measurements were defined as any two measurements whose difference is ⱕ 10% of the
average value. If a patient could not meet these criteria, the two
acceptable measurements were then recorded. All patients,
however, had a Dlco reported as long as one acceptable
maneuver was performed. Predicted values for spirometry, helium lung volumes, and Dlco were based on prediction equations derived from the literature.4,7
Patients with lung volume data were characterized based on
the type of ventilatory impairment. The following criteria were
used: restrictive ventilatory impairment (TLC ⱕ 80% and FEV1/
FVC ⱖ 0.80), obstructive ventilatory impairment (TLC ⬎ 80%
and FEV1/FVC ⱕ 0.80), mixed obstructive and restrictive ventilatory impairment (TLC ⱕ 80% and FEV1/FVC ⬍ 0.80), or no
ventilatory impairment (TLC ⬎ 80% and FEV1/FVC ⱖ 0.80).
Three measures of Dlco reproducibility were examined: (1) absolute difference, (2) percentage difference, and (3) the coefficient of variation (CV). Absolute difference between two repeated measures was defined as follows:
absolute difference ⫽ 兩Dlco1 ⫺ Dlco2兩
Percentage difference was calculated as a ratio of the absolute
difference to the average of the two repeated measures:
兩Dlco1 ⫺ Dlco2兩averageDlco ⫻ 100
CV was determined as follows:
冑冘
2
(Dlco[i] ⫺ average Dlco)2
CV ⫽
i⫽1
average Dlco
⫻ 100
The number and percentage of patients meeting different
reproducibility criteria were determined. Patients were grouped
according to the type of ventilatory impairment (obstructive,
restrictive, and mixed) to assess the impact of abnormal lung
function on Dlco variability. Analysis was performed using
statistical software (SAS 8.0; SAS Institute; Cary, NC).
Results
Demographic and Pulmonary Function
Characteristics
A total of 6,193 unique PFT records were selected
for the current study with availability of spirometry
and Dlco measurement during the same visit. Patient characteristics and pulmonary function data for
the study sample are summarized in Table 1. The
average age was 53.2 years (SD, 15.4 years). The
sample consisted of 2,903 women (46.9%) and 3,290
men (53.1%). There were 3,840 white patients
(62.0%) and 2,352 African-American patients
(38.0%). Of the 6,193 patients with spirometry and
Dlco measurements, 3,294 patients (53.2%) also
had measurement of static lung volumes with the
multiple-breath helium dilution method. The type of
www.chestjournal.org
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
CHEST / 123 / 4 / APRIL, 2003
1083
Table 1—Demographic and Pulmonary Function Data
of the Study Sample (n ⴝ 6,193)
Mean
(SD)
Variables
Age, yr
Body mass index
Spirometry
FEV1, L
FVC, L
FEV1/FVC, %
Dlco, mL/min/mm Hg
Helium lung volumes
(n ⫽ 3,294)
TLC, L
FRC, L
SVC, L
RV, L
Median
(Interquartile Range)
53.2 (15.4)
28.1 (7.4)
53.0 (42.0–65.0)
26.9 (23.1–31.1)
2.27 (0.95)
2.99 (1.09)
74.6 (11.7)
18.1 (7.13)
2.19 (1.53–2.89)
2.87 (2.15–3.69)
77.3 (69.3–82.5)
17.8 (13.1–22.5)
4.98 (1.46)
2.67 (0.94)
3.13 (1.14)
1.85 (0.76)
4.82 (3.90–5.97)
2.53 (1.98–3.19)
3.00 (2.25–3.86)
1.71 (1.33–2.22)
ventilatory impairment in patients with measurements of lung volume (n ⫽ 3,294) was determined.
Eight hundred twenty-one patients (25.0%) had no
evidence of a ventilatory defect, 1,368 patients
(41.5%) had an obstructive defect only, 555 patients
(16.8%) had a restrictive defect only, and 550 patients (16.7%) had a mixed obstructive-restrictive
defect.
Reproducibility of DLCO
The distribution of actual number of Dlco trials
performed per individual for the study sample was as
follows: 39.3% (two tests), 22.5% (three tests), 10.5%
(four tests), 17.5% (five tests), and 10.1% (missing
data on number of trials). Of the 6,193 patients,
98.3% patients met the current ATS criteria and had
two Dlco values that were within 10% or 3 mL CO
(STPD)/min/mm Hg of the average value. Figure 1
shows the percentage of patients that met other
candidate criteria for Dlco reproducibility. Above an
absolute difference of 2 mL CO (STPD)/min/mm Hg
between two measurements, the proportion of patients that met alternate criteria remained relatively
unchanged despite varying degrees of percentage
difference requirement. At or below an absolute
difference of 2 mL CO (STPD)/min/mm Hg between two measurements, the number of patients
that met criteria decreased as the percentage difference criteria was made more stringent (Fig 1).
Table 2 summarizes the measures of Dlco reproducibility for the entire study sample. Subgroup
analyses for patients without lung volume measurements showed no significant differences in the measures of Dlco reproducibility compared to the full
Figure 1. Percentage of patients meeting Dlco reproducibility criteria as a function of percentage
difference and absolute difference between two repeat measurements. Each line reflects the
percentage of patients who meet a specific absolute difference between two measurements as the level
of percentage difference is varied. The percentage difference between two measurements reflects the
percentage difference of the mean.
1084
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
Clinical Investigations
Table 2—Indices of DLCO Reproducibility (n ⴝ 6,193)
Variables
Mean (SD)
CV, %
Percentage difference, %
Absolute difference, mL CO
(STPD)/min/mm Hg
4.07 (5.37)
5.76 (7.75)
0.90 (0.96)
Median
(Interquartile Range)
2.87 (1.29–5.17)
4.05 (1.82–7.32)
0.70 (0.30–1.20)
study sample. In patients with lung volume measurements, the measures of Dlco reproducibility were
also examined across categories of ventilatory impairment (Table 3). Patients with either an obstructive,
restrictive, or mixed defect had more variability in
measurement as reflected by the mean percentage
difference and the mean CV compared to those
without any ventilatory impairment. However, the
absolute difference between two tests remained
relatively stable across patients with and without
ventilatory impairment. Measurement variability was
also examined across different levels of inspired
volume (Vi). As the ratio of Vi to vital capacity
(largest of FVC or SVC) increased, variability in
Dlco measurement decreased (Table 4).
To examine other determinants of Dlco reproducibility, we examined the associations between
several patient factors and each index of Dlco
reproducibility. As the degree of airflow obstruction
increased, as assessed by the FEV1 percentage of
predicted, there was an increase in the percentage
difference (Fig 2, top) and CV (data not shown).
However, there was relatively minimal change in the
absolute difference between two Dlco measurements with decreasing FEV1 percentage of predicted (Fig 2, bottom). Measurement variability was
also inversely related to the baseline average Dlco
value. As the Dlco (percentage of predicted) decreased, the CV (data not shown) and the percentage
difference progressively increased (Fig 3, top). Although the absolute difference between two Dlco
measurements increased with increasing Dlco (percentage of predicted), the change was relatively
small and insignificant (Fig 3, bottom).
We also examined whether other patient factors,
including gender and race, would be predictive of
variability in Dlco measurement. Men, on average,
had a higher mean Dlco (20.1 mL/min/mm Hg vs
16.4 mL/min/mm Hg, p ⬍ 0.0001) and a higher
mean absolute difference (0.93 mL CO [STPD]/min/
mm Hg vs 0.88 mL CO [STPD]/min/mm Hg,
p ⬍ 0.03) compared to women. Thus, the mean CV
for men compared to women was lower (3.8% vs
4.3%, p ⬍ 0.0001) as was the mean percentage
difference (5.3% vs 6.2%, p ⬍ 0.001). Multivariable
linear regression models of each reproducibility index (absolute difference, percentage difference, and
CV) revealed no associations with gender after adjusting for the average Dlco. Bivariate analyses also
demonstrated that African Americans had a lower
average Dlco (16.8 mL/min/mm Hg vs 19.0 mL/
min/mm Hg, p ⬍ 0.0001) than whites. The mean CV
and the mean absolute difference between the two
Dlco measurements in African Americans was 4.5%
and 0.94 mL CO (STPD)/min/mm Hg, respectively.
In contrast, the mean CV and mean absolute difference
in whites was 3.8% and 0.88 (STPD)/min/mm Hg,
respectively. Since African Americans in this group
were also found to have a lower FEV1 percentage of
predicted and a higher proportion of women compared to whites (62.0% vs 47.7%, p ⬍ 0.0001), multivariable linear regression models were used to
examine the independent relationship between race
and Dlco reproducibility. After adjusting for gen-
Table 3—Ventilatory Impairment and Indices of DLCO Reproducibility in Patients With Lung
Volume Measurements (n ⴝ 3,294)
Index of Reproducibility
Type of Ventilatory Impairment
Normal (n ⫽ 821)
Mean (SD)
Median (interquartile range)
Obstructive (n ⫽ 1,368)
Mean (SD)
Median (interquartile range)
Restrictive (n ⫽ 555)
Mean (SD)
Median (interquartile range)
Mixed obstructive-restrictive (n ⫽ 550)
Mean (SD)
Median (interquartile range)
Absolute Difference,
mL CO(STPD)/min/mm Hg
Percentage
Difference, %
CV, %
0.93 (0.80)
0.70 (0.30–1.40)
4.37 (3.83)
3.46 (1.66–6.25)
3.09 (2.7)
2.45 (1.18–4.42)
0.90 (0.75)
0.70 (0.30–1.30)
5.74 (7.56)
4.19 (1.92–7.41)
4.05 (5.11)
2.97 (1.36–5.24)
0.85 (0.85)
0.60 (0.30–1.20)
5.79 (5.90)
4.14 (2.06–7.65)
4.09 (4.17)
2.92 (1.45–5.43)
0.78 (0.86)
0.60 (0.30–1.10)
6.29 (8.56)
4.46 (1.94–8.25)
4.44 (5.87)
3.15 (1.37–5.84)
www.chestjournal.org
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
CHEST / 123 / 4 / APRIL, 2003
1085
Table 4 —Effect of Inspired Volume Criteria on Indices of DLCO Reproducibility
Index of Reproducibility
Vi/VC Ratio*
⬍ 0.70
Mean (SD)
Median (interquartile
0.70–0.74
Mean (SD)
Median (interquartile
0.75–0.79
Mean (SD)
Median (interquartile
0.80–0.84
Mean (SD)
Median (interquartile
0.85–0.89
Mean (SD)
Median (interquartile
ⱖ 0.90
Mean (SD)
Median (interquartile
Absolute Difference,
mL CO (STPD)/min/
mm Hg
Percentage Difference, %
CV, %
183
1.36 (1.64)
0.90 (0.40–1.70)
12.51 (20.57)
6.50 (3.02–13.56)
8.76 (13.99)
4.60 (2.13–9.59)
111
1.09 (1.74)
0.70 (0.20–1.20)
8.46 (10.20)
5.19 (2.11–10.16)
5.99 (7.22)
3.65 (1.49–7.20)
193
1.08 (1.29)
0.70 (0.30–1.40)
8.42 (11.75)
5.21 (1.92–9.42)
5.96 (8.33)
3.66 (1.36–6.66)
363
0.80 (0.84)
0.50 (0.30–1.00)
6.50 (10.49)
4.03 (1.95–8.13)
4.54 (6.65)
2.86 (1.38–5.75)
876
0.87 (0.79)
0.70 (0.30–1.20)
5.70 (7.67)
4.02 (1.99–7.16)
4.04 (5.49)
2.85 (1.40–5.07)
4,467
0.89 (0.91)
0.70 (0.30–1.20)
5.26 (5.87)
3.94 (1.75–6.99)
3.72 (4.12)
2.77 (1.24–4.94)
Patients,
No.
range)
range)
range)
range)
range)
range)
*VC ⫽ larger of FVC or SVC (if available).
der, FEV1 percentage of predicted, and average
Dlco, race was not associated with either the
percentage difference or the CV. However, race
remained significantly associated with the absolute
difference after adjusting for gender, FEV1 percentage predicted, and average Dlco. The adjusted
absolute difference between African Americans and
whites was 0.06 mL CO (STPD)/min/mm Hg (95%
confidence interval [CI], 0.01 to 0.11).
Discussion
The present study examined the utility of the
current recommendations on reproducibility limits
for the measurement of single-breath Dlco. Using a
consecutive sample of patients referred to a general
pulmonary function laboratory, 98.3% patients referred for clinical testing met the current ATS
reproducibility criteria for repeat Dlco measurement. The use of alternative criteria that are more
stringent did not significantly alter the number of
patients that could fulfill these criteria under routine
testing in accordance with the current standards.
The results of this study also indicate that patient
factors, such as gender and race, are not significantly
associated with measurement variability for the
single-breath method. In contrast, the presence of an
obstructive or restrictive ventilatory defect was associated with an increase in measurement variability.
Other factors associated with increased measurement variability included the following: (1) a ratio of
Vi to vital capacity ⬍ 90%, (2) a low FEV1 percentage of predicted, and (3) a low Dlco.
The current guideline for reproducibility in the
ATS standards is to perform at least two maneuvers
that agree within 10% or 3 mL/min/mm Hg of the
average value.3 Thus, two Dlco maneuvers could
differ by as much as 20% or 6 mL/min/mm Hg and
still be considered reproducible. It is the policy in
our laboratory to perform testing until two maneuvers are obtained which are within 10% of the
average up to a maximum of five maneuvers depending on the cooperation of the patient. In a patient
population referred for clinical testing, we could
successfully fulfill the current ATS reproducibility
criteria in virtually all cases where two acceptable
maneuvers could be obtained. Moreover, 95.9% of
the patient sample had two Dlco measurements
within 5% and 2.5 mL/min/mm Hg of each other.
These results are consistent with data from a Norwegian population-based study that showed that 98%
of the volunteers from the community could meet
the current ATS criteria for Dlco reproducibility.8
Thus, the current standards for Dlco reproducibility may be too liberal to meaningfully enforce good
performance.
The results of the current study also reveal that the
CV and percentage difference were largest in patients with abnormal lung function. Because the
mean Dlco is used in the calculation of the CV and
the percentage difference, the finding of higher
measurement variability in patients with lung disease
1086
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
Clinical Investigations
Figure 2. Box and whisker plots for the (top) percentage difference and (bottom) absolute difference
for repeat Dlco measurements as a function of FEV1 percentage of predicted (circles represent
outliers, boxes represents the 25th and 75th percentiles, and the whiskers represent the 10th and 90th
percentiles).
www.chestjournal.org
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
CHEST / 123 / 4 / APRIL, 2003
1087
Figure 3. Box and whisker plots for the (top) percentage difference and (bottom) absolute difference
for repeat Dlco measurements as a function of Dlco percentage of predicted (circles represent
outliers, boxes represent the 25th and 75th percentiles, and the whiskers represent the 10th and 90th
percentiles).
1088
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
Clinical Investigations
may simply reflect a lower mean Dlco. However,
the possibility that there is also inherently greater
variation due to the underlying lung disease cannot
be completely excluded. The implication of these
findings is that the use of percentage reproducibility
alone as a criterion for ending a test session would
require that more maneuvers be performed in individuals with lower Dlco. The ATS standards address this issue by allowing a session to be considered
reproducible if two acceptable values are within
3 mL/min/mm Hg of the average value. This criterion could be met in ⬎ 98% of the sample, regardless of the percentage difference, and thus is also
very liberal.
Based on the present investigation, it would be
reasonable to consider alternative criteria for Dlco
reproducibility. A variety of alternatively reproducibility criteria and the percentage of patients meeting
these criteria are presented in Figure 1. Ninety to
95% of the patients should be able to perform a
reproducible Dlco measurement if they can perform two acceptable maneuvers. Any reproducibility
criteria that take into account the percentage difference will tend to show better performance in patients with larger baseline Dlco values. Thus, we
would propose that a reasonable criterion for Dlco
reproducibility would be to consider an absolute
difference of 2 to 2.5 mL CO (STPD)/min/mm Hg
between two measurements, which were met in
91.5% and 95.8% of the patients, respectively.
There are several limitations in the current study.
First, it should be noted that the technique for
performing the Dlco in our laboratory varies
slightly from the ATS guidelines. We have done this
to maintain stability in our reported measurements
as the standards change over time. While this may
have a slight effect on the reported Dlco values, it
is not likely that it should affect reproducibility
within individuals.9 Second, pulmonary function
testing in our laboratory is performed on two different types of testing systems, one a commercial
system (W.E. Collins) with customized software, and
the other a locally fabricated apparatus with an
automated valve. It is possible that the intrasubject
variability could differ with other types of equipment, although this is unlikely since most automated
systems perform in a similar fashion to those in use
in our laboratory. Restricting the analyses to tests
performed on one system vs another revealed statistically small differences in the measures of Dlco
reproducibility that were not clinically meaningful.
For example, the average absolute difference for
tests performed on the commercial system was 0.81
(95% CI, 0.76 to 0.85), whereas the average absolute
difference for tests performed on the locally fabricated apparatus was 0.96 (95% CI, 0.93 to 0.99).
Although these values were statistically different,
inferences regarding the use of absolute difference
instead of the percent difference between repeat
measurements did not change when the analysis was
restricted to a specific system. Despite these limitations, the current study has several strengths. Because we utilized pulmonary function data from a
general clinical laboratory, the study sample included a broad range of patients that are typical for
a metropolitan academic medical center. Furthermore, the testing was performed by at least seven
different technicians with a wide range of experience. Thus, we suspect that our results would show
larger variability then might be expected in other
more controlled settings such as a research study or
surveys of a general population.
In summary, reproducibility in the measurement
of Dlco is generally much better than the current
recommended standards. Based on the results of the
current study, we suggest that future standards
should consider setting a more stringent criteria
based on an absolute difference in test results rather
than a percentage difference.
ACKNOWLEDGMENT: The authors thank Dr. Robert O.
Crapo for his helpful comments.
References
1 American Thoracic Society. Single breath carbon monoxide
diffusing capacity (transfer factor): recommendations for a
standard technique; statement of the American Thoracic
Society. Am Rev Respir Dis 1987; 136:1299 –1307
2 Cotes JE, Chinn DJ, Quanjer PH, et al. Standardization of the
measurement of transfer factor (diffusing capacity): Report
Working Party Standardization of Lung Function Tests,
European Community for Steel and Coal; official statement
of the European Respiratory Society. Eur Respir J Suppl
1993; 16:41–52
3 American Thoracic Society. Single-breath carbon monoxide
diffusing capacity (transfer factor): recommendations for a
standard technique, 1995 update. Am J Respir Crit Care Med
1995; 152:2185–2198
4 Goldman HI, Becklake MR. Respiratory function tests: normal values at median altitudes and the prediction of normal
results. Am Rev Respir Dis 1959; 79:457– 467
5 Ferris BG. Epidemiology Standardization Project (American
Thoracic Society). Am Rev Respir Dis 1978; 118:1–120
6 American Thoracic Society. Standardization of spirometry,
1994 update. Am J Respir Crit Care Med 1995; 152:1107–
1136
7 Burrows B, Kasik JE, Niden AH, et al. Clinical usefulness of
the single-breath pulmonary diffusing capacity test. Am Rev
Respir Dis 1961; 84:789 – 806
8 Welle I, Eide GE, Bakke P, et al. Applicability of the
single-breath carbon monoxide diffusing capacity in a Norwegian Community Study. Am J Respir Crit Care Med 1998;
158:1745–1750
9 Beck KC, Offord KP, Scanlon PD. Comparison of four
methods for calculating diffusing capacity by the single breath
method. Chest 1994; 105:594 – 600
www.chestjournal.org
Downloaded From: http://publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21992/ on 06/18/2017
CHEST / 123 / 4 / APRIL, 2003
1089