View/Open

advertisement
Contribution of pulmonary function tests to diagnosis in patients with respiratory
complaints: a prospective cohort study
Marc Decramer1 MD, Wim Janssens1 MD, Eric Derom2 MD, Guy Joos2 MD, Vincent
Ninane3 MD, René Deman4 MD, Dirk Van Renterghem5 MD, Giuseppe Liistro6 MD, Kris
Bogaerts 7 PhD and the Belgian Pulmonary Function Study Investigators*
1Respiratory
Division, University Hospital, University of Leuven, Belgium
2Respiratory
Division, University Hospital, University of Ghent, Belgium
3Respiratory
Division, St.Pierre Hospital, Brussels, Belgium
4Respiratory
Division, Groeninge Hospital, Kortrijk, Belgium
5Respiratory
Division, AZ St Jan Brugge-Oostende, Belgium
6Respiratory
Division, St Luc University Hospital, Université Catholique de Louvain,
Brussels, Belgium
7I-Biostat,
University of Leuven and University of Hasselt Belgium
Address for correspondence:
Marc Decramer, MD, PhD
Professor of Medicine
Respiratory Division
University Hospital
Herestraat 49
3000 Leuven
Belgium
Marc.Decramer@uzleuven.be
1
Summary
Background. Few studies on the diagnostic value of pulmonary function testing are
currently available.
Methods, therefore, we undertook the present study (NCT01297881) to determine the
diagnostic contribution of the four basic pulmonary function tests: spirometry, lung
volumes, airway resistance and diffusing capacity. 1,023 new outpatients without clear
diagnosis, presenting to a pulmonologist with respiratory complaints in one of the 33
participating hospitals, were included and eventually 979 patients were analyzed. They
underwent the four abovementioned pulmonary function tests and all other tests
required to establish a firm diagnosis. The clinical history and pulmonary function data
were presented to local focus groups, who established differential diagnoses and a
preferred diagnosis after each step, being blinded to the other additional investigations.
The final diagnosis was established by the attending physician on the basis of all the
investigations performed, and validated as gold standard diagnosis by the local focus
group. The primary outcome was a score determined by one divided by the number of
differential diagnoses corrected for the accuracy of the diagnosis. Secondary outcomes
were the number of differential diagnoses and the percentage of correct diagnoses. The
present publication represents the final analysis.
Findings. The primary outcome score averaged 0.226 after spirometry, and increased to
0.296, 0.373 and 0.540 after measurement of lung volumes, resistance and diffusing
capacity, respectively (P<0.0001 for each step). This score indicated a progressive
reduction of the number of differential diagnoses and an increased accuracy of the
diagnosis. This was further confirmed by the observed reduction in the number of
differential diagnoses (4.2, 3.4, 3.0, and 2.4, respectively, P<0.0001 for each step) and the
increase in the percentage correct preferred diagnoses (61, 65, 70, and 77%, respectively,
P<0.0001 for each step).
Interpretation. This study demonstrates that each of the classical pulmonary function
tests contributes significantly and independently to the final diagnosis in patients with
2
respiratory complaints. As a consequence, pulmonary physicians should perform these
four tests as a battery when confronted with new patients.
Funded by the Belgian Society of Pneumology.
Word count: 333
Key words: Pulmonary function testing, diagnosis, differential diagnosis, respiratory
disease, sensitivity, specificity
3
Introduction
Although pulmonary function testing is routinely used in the diagnosis of
respiratory disease, few methodologically well-designed studies are presently available on
the diagnostic value of these tests. This was evidenced by a recent report of the Belgian
Expert Center for Health Care published in 20071, showing that for most of the routinely
used pulmonary function tests, no symptoms based studies i.e. studies examining a group
of patients defined on the basis of their symptoms, supporting their diagnostic value and
reaching sufficient methodological quality were available. Those that were available were
generally of limited quality2. In addition, only few disease based studies i.e. studies
defining a group of patients on the basis of their disease, of sufficient quality were found
with systematic literature review. Hence, use of pulmonary function tests was found to be
primarily based on expert opinion and on international disease oriented guidelines,
predominantly dealing with asthma3-6 and COPD7-10 and consequently, reimbursement
was questioned by the health authorities. The view that few well designed studies were
performed examining the contribution of different pulmonary function tests to the
diagnosis of pulmonary disease is widely held by experts in this field11-13. This is in
contrast with the relatively large body of evidence demonstrating the prognostic value of
these tests14-18.
The present study was designed to critically assess the diagnostic accuracy in
pulmonary practice of four classically used pulmonary function tests: spirometry, absolute
lung volumes measured with a body plethysmograph, airways resistance and diffusing
capacity. This was done by performing the abovementioned tests in a total of 1,023 new
outpatients presenting with respiratory complaints to one of the participating pulmonary
practices and subsequently, examining their contribution to the reduction of the number
of differential diagnoses and to the accuracy of the diagnosis.
4
Materials and methods
Overall design
The study was a prospective cohort study in which all consecutive new outpatients
with respiratory complaints and without established diagnosis seen in two particular time
periods, were enrolled. These time periods ran from June 6th till September 12th 2011, and
from January 16th till June 12th 2012, respectively. Respiratory complaints were recorded
using a symptom list, consisting of standard complaints and questions on how long these
were present and how often they occurred. Other relevant matters could be added in
text. These patients underwent spirometry, plethysmographic measurement of absolute
lung volumes, airway resistance and measurement of diffusing capacity, always in that
order. In further follow-up, the attending pulmonologist performed all required tests to
establish a firm diagnosis including other pulmonary function tests (reversibility,
histamine challenge test, exhaled NO, pulmonary compliance,…), biochemistry, imaging,
ECG, echocardiography etc… The study was registered as NCT01297881 on
clinicaltrials.gov.
The cases were subsequently presented to local focus groups consisting of
practicing pulmonologists with experience in pulmonary function testing. This was done
using the existing LOK/GLEM network in Belgium which is a network of 24 groups of 20-25
pulmonologists who meet 4 times per year to discuss issues that relate to quality and
cost-effectiveness of care. The cases were presented to these groups in four steps,: first,
the history, clinical examination and the results of spirometry; second, results of absolute
lung volumes; third, results of measurements of airway resistance; fourth, results of
measurement of the diffusing capacity. For all variables, both absolute values and values
expressed as a percentage of predicted were presented. For diffusing capacity, both DLCO
and KCO were offered.
After each step, the local focus group that was blinded to the additional clinical
data, was asked to determine the possible differential diagnoses and a preferred
diagnosis. Finally, the attending pulmonologist established a final diagnosis and if
5
necessary also a side diagnosis, on the basis of all the clinical data he had obtained. If e.g.
a patient would have kyphoscoliosis and asthma, then asthma could be the final diagnosis
and kyphoscoliosis the side diagnosis or vice versa. This final diagnosis and a potential
side diagnosis were also validated by the local focus group and then coined gold standard
diagnosis. Eighteen out of the existing 24 local quality groups participated in the present
study.
The differential diagnoses, preferred diagnoses, side diagnoses and gold standard
diagnoses were chosen by ticking from a list of 14 predefined diagnoses. This list is shown
in Appendix 1. The design of the study is schematically represented in Figure 1. The study
protocol was approved by the Ethics Committee of the University Hospital Leuven, which
acted as the leading Ethics Committee and by all the Ethics Committees of the
participating hospitals which acted as subsidiary Ethics Committees. The list of the
participating hospitals is shown in Appendix 2.
For the interpretation of pulmonary function tests the ATS/ERS interpretation
strategy was followed19. As the most frequent diagnoses were expected to be asthma and
COPD, special attention was devoted to the diagnosis of these conditions. Diagnosis of
COPD required demonstration of airflow obstruction defined as an FEV1/FVC ratio below
lower limit of normal or alternatively, a reduced FEV1 associated with normal total lung
capacity, TLC. Diagnosis of asthma in the absence of airflow obstruction and full
reversibility and in addition to a history compatible to asthma, required clearly elevated
exhaled NO fractions or hyperreactivity to histamine. If the latter were absent and all
pulmonary function tests were normal we concluded to “no primary pulmonary
abnormality” or “other disease” if another disease e.g. sinusitis was diagnosed. The term
“other obstructive disease” encompassed patients with “bronchiectasis”, “cystic fibrosis”
or “small airway disease”. The latter diagnosis was made in the presence of reduced
expiratory flow rates with normal FEV1 and FEV1/FVC ratios > LLN, but lower than 0.7.
Other diseases were diagnosed according to international guidelines (see study protocol,
Appendix 3). These rules were developed in a preliminary study performed by the
Steering Committee (a committee consisting of 10 leading investigators, who initiated the
6
study) in their centers, in which 79 patients were enrolled. This preliminary study was
completely separate from the present study and these patients were not included in the
present study. The primary outcome variable (see below) was also developed in that
study. The present rules were agreed beforehand and their application was supervised by
a member of the Steering Committee, who assisted to all the meetings of the local focus
groups.
Outcomes
As the primary outcome, we used a score that combined both the number of
differential diagnoses and the accuracy of the preferred diagnosis, in one. This score was
developed in the abovementioned preliminary study. It was defined as 1/number of
differential diagnoses, provided the gold standard diagnosis was part of the differential
diagnoses. Hence, if there were 4 differential diagnoses, the score would be 0.25 and if
there were only 2, the score would be 0.5. In addition, this score was corrected for the
accuracy of the diagnosis. This was done by subtracting a penalty if the gold standard
diagnosis was part of the set of differential diagnoses, but not the preferred diagnosis.
The correction consisted of 1/28, that is half of 1 over the 14 differential diagnoses. We
divided by 2, because otherwise the penalty could result in a score of 0. Hence the lowest
score needed to be 1/14 (=2/28) minus if indicated, 1/28 penalty, yielding a score of 1/28.
Coming back to the example above: if there were 4 differential diagnoses and the
preferred diagnosis was the correct diagnosis, the score remained 0.25. If there were only
two differential diagnoses but the preferred diagnosis was not the correct diagnosis, the
score would be 0.5-1/28=0.464. Furthermore, if there was only one differential diagnosis,
which was by definition also the preferred diagnosis, and it was also the correct diagnosis
a score of 27/28 was attributed (close to 1). This was done to avoid a score of 1. If the
correct diagnosis was not part of the differential diagnoses, the lowest possible score
1/28 (close to 0) was given. This was done to avoid scores of 0.
7
Hence, the score penalized a high number of differential diagnoses and a wrong
preferred diagnosis. The secondary outcomes were the % correct preferred diagnoses
corresponding to the gold standard diagnosis after each step and the number of
differential diagnoses, but now limited to the patients with a correct preferred diagnosis
only. Sensitivity analyses were all of the above analyses considering also the side
diagnosis as being correct, if it corresponded to the gold standard diagnosis. Predefined
subgroup analyses were done for patients with “COPD”, “asthma”, “ILD”, and “no primary
pulmonary abnormality”. For these subgroups sensitivity, specificity, positive predictive
value and negative predictive value were also calculated in a posthoc analysis.
Patients
All consecutive new caucasian outpatients between 18 and 75 yrs. old were
enrolled in the study. Patients needed to have respiratory symptoms (dyspnea, cough,
sputum, wheezing etc..), but no diagnosis at the time of consultation. They were required
to be able to perform spirometry and gave written informed consent. Patients were
excluded if they had diagnosed respiratory disease, a heart attack during the last 6
months, pneumothorax, hemoptysis, suspicion of tuberculosis or were not capable of
performing pulmonary function testing. A total of 1,023 patients were recruited in the
study.
Demographic characteristics
At the outpatient visit classical demographic characteristics such as age, gender,
weight, height, BMI, smoking history, pack years, the presence of allergy/atopy, previous
pulmonary disease, reflux and ear nose throat problems were recorded. All variables were
captured in a web-based application which was accessible to the participating centers,
although their access was limited to the patients recruited by number of this particular
center.
8
Symptoms
Symptoms were recorded via the web-based application. The symptoms recorded
included: cough, wheezing, dyspnea, fatigue, sputum production, thoracic pain and
hemoptysis. Symptoms were also recorded using the COPD Assessment Testquestionnaire20, which was developed for COPD, but was also used for the other diseases
in the present study. This questionnaire captures cough, sputum, chest pain, dyspnea
during moderate exercise, limitation of activities of daily living, feeling of security when
leaving the house, sleep, and energy level.
Pulmonary function testing
All pulmonary function tests were performed using equipment that was available
in the participating pulmonary function laboratory. This consisted of a variety of
equipment including: Sensormedics Whole Body Plethysmograph, Care fusion, Vilvoorde,
Belgium; Masterscreen Jaeger, Care fusion, Vilvoorde, Belgium; Medisoft boybox 5500,
Sorinnes, Belgium; Ganshoren-Medisoft Hyp’air compact, Chaumont-Gistoux, Belgium.
All tests were performed according to the ATS/ERS standards21. Lung volumes and airway
resistance were measured with a constant volume body plethysmograph, at a breathing
frequency of 35-40/min. Diffusing capacity was measured by single-breath breath-holding
carbon monoxide gas transfer method22. Values were corrected for alveolar volume, but
not for hemoglobin concentration. Further technical details of the measurements are
summarized in the study protocol (Appendix 3). Normal values were those of Quanjer et
al23. The interpretation strategy of ATS/ERS was used19. All original data were entered in
the web-based application, which calculated reference values, % predicted values and
lower limits of normal values. These values were provided to the focus groups for their
discussion. Consistency of the data was checked by a central study coordinator, before
meetings of the focus groups and data analysis were started.
9
Other tests
After pulmonary function testing, the attending pulmonologist performed all tests
required to establish a diagnosis. These tests included: chest x-ray and CT-scan,
reversibility testing, measurement of exhaled NO, histamine challenge test, skin prick
testing, biochemistry, blood gases, cardiac echography, ECG etc.. The attending
pulmonologist indicated which tests he had performed and which in his view contributed
most to his final diagnosis. The tests were used at the discretion of the attending
pulmonologist. No algorithms specifically designed for this study were used.
Statistical Analysis
The statistical analysis plan is attached as Appendix 4. All analyses were done using
SAS version 9.2.The primary outcome was analyzed after logistic transformation in order
to stabilize its variance and normalize its distribution, using a linear mixed model 24. The
main structure contained 4 time points and an unstructured variance-covariance matrix
was used in order to properly account for the dependence between the different
measurements. The mixed effect model used the logit of the score as dependent variable
and the indicators for the different tests as independent variables. An unstructured
variance-covariance matrix was used. The primary hypothesis was H0:µ1=µ2=µ3=µ4 vs. H1:
otherwise with µi, i=1…, 4 was the mean transformed score after the i-th pulmonary
function test. In case of a significant difference at the 0.05 level, all pairwise tests
between the different time points were performed with a Bonferroni-Holm correction for
multiple testing25.
The secondary endpoint, the percentage of correct preferred diagnoses at each
point in time was modeled using a logistic generalized estimating equation (GEE) model.
The GEE model for the percentages of correct final diagnoses modeled a binary variable
for correct diagnosis as dependent variable and the indicators for the different tests as
independent variables using a logit link and an independent working correlation matrix.
10
The number of differential diagnoses in the subgroup of patients with a correct preferred
diagnosis was analyzed using a Poisson generalized estimating equation (GEE) model. The
GEE Poisson model for the number of differential diagnoses modeled the number of
differential diagnoses as dependent variable and the indicators for the different tests as
independent variables using a log link and an unstructured working correlation matrix. In
case of a significant difference at the 0.05 level, pairwise comparisons were performed for
both models with a Bonferroni-Holm correction for multiple testing. No formal sample
size calculation was performed before the study.
As a sensitivity analysis the above analyses were repeated considering the
preferred diagnosis as correct if it or its side diagnosis corresponded to the gold standard
diagnosis. Analyses were also repeated in the four predefined subgroup analyses defined
above.
Role of the funding source
This study was supported by the Belgian Society of Pneumology. This support was
limited to a grant of 10,000 euro to support the statistical analysis of the data. The
funding source had no other role. MD, WJ and KB had access to the raw data. The
corresponding author had full access to all the data and the final responsibility to submit
for publication.
Results
Patient disposition and baseline data
Patient disposition is schematically represented in Figure 2. Of the 1,285 patients
screened, 1,023 patients were enrolled and eventually, 979 patients were analyzed.
Baseline data are represented in Table 1. Symptoms are summarized in Table 2 and
duration of symptoms is shown in Table 3. As can be seen the most prevalent presenting
symptoms were dyspnea and cough, followed by sputum production and wheezing. 84 %
of the patients already had symptoms for a few months or over 1 year, before they saw a
pulmonologist. Baseline pulmonary function results are shown in Table 4. As can be seen
11
on the average the pulmonary function results were within normal limits, although
obviously in a fraction of the patients either all or some of the tests were clearly abnormal
as indicated by the median and lower and upper quartile values.
The most frequently used and the, in the view of the attending pulmonologist,
most contributing additional tests are shown in Table 5. The top three most performed
tests were: chest x-ray (66%), reversibility testing (50%) and exhaled NO (32%). The most
contributing tests in the judgment of the pulmonologist were reversibility testing (26%),
provocation testing (25%) and CT-scan of the chest (15%).
Diagnoses
The distribution of the preferred diagnoses was as follows: Asthma (39%), COPD
(23%), No primary pulmonary abnormality (16%), Other Obstructive disease (7%), ILD
(4%), Hyperventilation syndrome (3%), Neuromuscular disease (2%), Obesity (2%), Cardiac
failure (2%), Thoracic/Pleural disease (1%), Pulmonary Vascular Disease (1%), Systemic
sclerosis (0%) and other diseases (3%). The relative frequency distribution of the
differential diagnoses was similar. Table 6 summarizes the frequency of the preferred and
differential diagnoses after each step. The gold standard diagnosis distribution was also
similar and the final diagnosis of the attending pulmonologist was confirmed by the focus
group in 838/979 (86%) of the cases. This figure was 200/222 (90%), 321/369 (87%),
35/41 (85%) and 142/158 (90%), for COPD, asthma, ILD, and no primary pulmonary
abnormality, respectively. The discordant cases were virtually all localized in the disease
categories asthma, COPD and other obstructive diseases. Side diagnoses were established
by the attending pulmonologist in 465 out of 979 or 48 % of the cases, and by the focus
groups in 411 out of 979 or 42 % of the cases.
Primary outcome
The results of the primary analysis are shown in Table 7. After clinical history and
spirometry the back-transformed mean score was 0.226. It increased to 0.296, 0.373, and
12
0.540 after measurement of lung volumes, resistance and diffusion capacity, respectively.
The increase with each of these steps was highly statistically significant (P <0.0001).
Similar results were obtained in the sensitivity analysis. The score was higher with each
step in this analysis as now also a side diagnosis corresponding to the gold standard
diagnosis was considered as a correct preferred diagnosis. Similar results were found in
the four predefined subgroups (Table 8 A-D). In each of the subgroups, the increase in
score was highly significant after each step (P<0.001). As can be seen in Table 8 the
relative contribution of the different tests tended to differ in the different diagnostic
groups. E.g. diffusing capacity tended to contribute more to the diagnosis of COPD and
ILD, than to the diagnosis of asthma and no primary pulmonary abnormality.
Secondary analyses
Similarly, the mean number of differential diagnoses in patients with a correct final
diagnosis was reduced from 4.2 after spirometry, to 3.4, 3.0 and 2.4, after measurement
of lung volumes, resistance and diffusion capacity, respectively. Also each of these
reductions was highly statistically significant (P <0.0001). Similar results were obtained in
the sensitivity analysis. All of the pairwise comparisons showed a highly significant
difference (P < 0.0001), in the secondary analysis as well as in the sensitivity analysis. The
pairwise ratios in mean number of differential diagnoses ranged from 0.56 to 0.88. The
GEE analysis of the number of correct diagnoses demonstrated 597/979 61% (95%CI 5864%) correct diagnoses after clinical history and spirometry. This increased to 637/979
65% (95%CI 62-68%), 683/979 70% (95%CI 67-73%) and 755/979 77% (95%CI 75-80%),
after measurement of lung volumes, resistance and diffusing capacity, respectively. These
data are shown in Table 9. Each of these steps reached statistical significance (P < 0.0001).
Similar step increases were obtained in the sensitivity analysis. The number of correct
diagnoses was 688/979 70%, 723/979 74%, 766/979 78% and 815/979 83%, after each
step respectively. Here again the number is higher after each step, because also a correct
side diagnosis was considered as a correct preferred diagnosis, if corresponding to the
13
gold standard. Similar results were found in the predefined subgroups. In patients with
COPD the mean number of differential diagnoses in patients with a correct final diagnosis
decreased from 4.3 after spirometry to 3.1, 2.8 and 2.0 after measurement of lung
volumes, resistance and diffusing capacity, respectively (P<0.0001). Similarly in patients
with asthma it decreased from 3.9 to 3.1, 2.7 and 2.3, respectively (P<0.0001). The
number of correct diagnoses increased from 276/369 75% (95%CI 70-79%) to 285/369
77% (95%CI 73-82%), 293/369 79% (95%CI 75-84%) and 303/369 82% (95%CI 78-86%)
(P=0.0154), in patients with asthma, while in patients with COPD it only clearly increased
after measurement of diffusing capacity [184/222 83% (95%CI 78-88%) to 185/222 83%
(95%CI 78-88%), 182/222 82% (95%CI 77-87%) and 207/222 93 % (95%CI 90-97%)
(p=0.0005), respectively].
For the four subgroups sensitivity, specificity, negative and positive predictive value all
increased after each step (P-value ranging from 0.00017 to 0.0161) except for specificity
in the ILD group (P=0.34). For patients with asthma e.g. sensitivity increased from
276/369 75% (95%CI 70-79%) to 285/369 77% (95%CI 73-82%), 293/369 79% (95%CI 7584%) and 303/369 82% (95%CI 78-86%), respectively (P= 0.0161); specificity from 451/610
74% (95%CI 70-77%) to 456/610 75% (95%CI 71-78%), 498/610 82% (95%CI 79-85%) and
528/610 87% (95%CI 84-89%), respectively (P<0.0001); negative predictive value from
451/544 83% (95%CI 80-86%) to 456/540 84% (95%CI 81-88%), 498/574 87% (95%CI 8490%) and 528/594 89% (95%CI 86-91%), respectively (P=0.0001) and positive predictive
value from 276/435 63% (95%CI 59-68%) to 285/439 65% (95%CI 60-69%), 293/405 72%
(95%CI 68-77%) and 303/385 79% (95%CI 75-83%), respectively (P<0.0001).
Discussion
The present study clearly demonstrated that in the present stepwise comparative
cohort study each of the four classical pulmonary function tests (spirometry, lung
volumes, airway resistance and diffusing capacity) significantly contributed to reduction
14
of the number of differential diagnoses and the increase in the percentage correct
preferred diagnoses. The relative reduction in the mean number of differential diagnoses
ranged from 12% to 20% per test, and the increase in the percentage of correct final
diagnoses ranged from 4 % to 7% per test. Hence, each of these tests had an independent
and significant diagnostic value. Moreover, after the four tests on average, only 2
differential diagnoses were left and the % correct diagnoses amounted to 77%, showing
that the four pulmonary function tests combined had an important diagnostic yield. We
believe that this is an important clinical finding as at present very few well-designed
studies were published supporting the additional diagnostic value of other pulmonary
function tests besides spirometry, in an unselected group of patients without prior
diagnosis11-13. Similar results were obtained in the predefined subgroups COPD, asthma,
ILD and no primary pulmonary abnormality. We are not aware of any similar study
conducted before.
This study also clearly shows that after history, clinical examination and
spirometry, a diagnosis is rarely certain. This would be the case if there only would be one
differential diagnosis. The latter occurred only in 3 out of 222 cases (1.4%) with a
diagnosis of COPD and in 11 out of 369 cases (3%) in patients with a diagnosis of asthma.
After the 4 tests these numbers were 114 out of 222 cases (51%) with a diagnosis of COPD
and 105 out of 369 cases (28%) with a diagnosis of asthma. In the remaining cases the
diagnosis of asthma was further corroborated by exhaled NO and histamine provocation
testing, whereas the diagnosis of COPD was further supported by CT-scan. Hence, the
present data clearly demonstrate that in these frequent lung diseases more extensive
pulmonary function testing clearly contributes to diagnosis. The strengths of this study
include its size, prospective design, validation of the diagnosis by focus groups of
pulmonary physicians, the inclusion of nearly all consecutive patients seen by the
participating pulmonary physicians and its real life circumstances in which an unselected
group of patients was examined in a normal outpatient setting. However, the present
study also suffers from a number of limitations. First, it only examines the four “classical”
pulmonary function tests and does not address other tests that were shown to be of
15
significant diagnostic value in specific disease entities. These include: exhaled NO and
histamine or methacholine challenge testing, considered to be useful in the diagnosis of
asthma26,27, reversibility testing which has been challenged repeatedly28,29 and
measurement of pulmonary compliance shown to be useful in the diagnosis of interstitial
lung disease and sarcoidosis30-33. Second, the outcomes measured essentially were a
score related to the number of differential diagnoses and the percentage correct
preferred diagnoses, but we did not estimate the degree of certainty of a diagnosis which
clearly also will vary with the stepwise application of pulmonary function tests. This was
largely due to the fact that we could not find a reliable and validated way to measure this
degree of certainty, although a scheme was developed for idiopathic interstitial
pneumonia34. However, it is also clear that the number of differential diagnoses and the
degree of certaintly of a diagnosis do not necessarily bear a simple relationship.
Third, although the diagnoses and differential diagnoses were validated by focus
groups of pulmonary physicians, a certain degree of bias based on the expected outcome
of the study, may have been present. However, some degree of bias will always be
present in such studies as we cannot eliminate the competence and experience of the
pulmonary physicians in the interpretation of the clinical and pulmonary function data. In
addition, we believe that this potential bias was minimal since the validation occurred in
virtually public hearings of the focus groups, whose members in majority did not enroll
patients in the present study. The attending physician presenting the cases was not
involved in the final decision of the focus group. Fourth, the present study only provides
information on the diagnostic value in the present setting i.e. new patients with an
uncertain diagnosis at the time of the study. Hence, it does not clarify how pulmonary
function tests perform in providing the basis for the adaptation of therapy, when
repeated in patients with a known diagnosis. It also does not give insight in the prognostic
value of the tests that may be present in a given clinical setting. It does not provide us
with information on the diagnostic performance of other pulmonary function tests such
as compliance measurements or provocation testing. Fifth, the present study only gives us
16
information on how the routinely used pulmonary function tests perform in a specialized
population of patients seen by a pulmonologist. If this population were to be diluted in a
larger group of patients with diagnoses to which pulmonary function tests cannot
contribute directly e.g. sinusitis or cardiomyopathy, the final number of correct preferred
diagnoses would be expected to be lower. The present population may also clearly differ
from a population seen in general practice. Finally, the present study does not allow us to
conclude which specific diagnostic paths are shortened by performance of more
pulmonary function tests nor what the effect on the cost of a specific diagnostic path
would be.
It is also important to emphasize that the diagnostic information obtained from
pulmonary function testing in the present study, may potentially be obtained from other
tests as well. The diagnosis of asthma can be supported by the measurement of exhaled
NO26 and CT-scan can contribute to the diagnosis of airway disease and emphysema35,36,
while it definitely contributes to the diagnosis of ILD37 and sarcoidosis38. In the present
study, the latter techniques were used to establish a final diagnosis, but their
performance was not compared to the performance of the different pulmonary function
tests. Hence, the present study does not demonstrate that these four pulmonary function
tests are the only way to obtain a correct diagnosis, but rather that they represents a way
to do this. In addition, it does not appear to be a bad way as eventually 77%, and if also a
correct side diagnosis was taken into account even 83% of correct diagnoses were
obtained, after performance of the four classical pulmonary function tests.
In conclusion, the present study was designed as a prospective cohort study in
new outpatients with respiratory complaints and an uncertain diagnosis. Diagnoses were
established by a focus group of pulmonary physicians and compared with the gold
standard for which all available diagnostic information was used. It demonstrated that in
addition to spirometry, measurement of lung volumes, airway resistance and diffusing
capacity significantly and independently contributed to the reduction of the number of
differential diagnoses and establishment of the correct diagnosis.
17
Research in context
Systematic Review
The authors searched PubMed, Embase and Cochrane d-base, using the terms
“pulmonary function testing and diagnosis and respiratory complaints” and “pulmonary
function testing and differential diagnosis and respiratory complaints”. The search was
restricted to publications in English and was performed in March 2013.All publications
dealing with the subject of the present study were selected.Also references of retrieved
papers and commonly used references to guidelines and statements were used. As stated
in the article, no evidence is presently available on the contribution of different
pulmonary function tests to diagnosis in respiratory disease.
The present trial is to our knowledge the first systematic trial on how different pulmonary
function tests contribute to the diagnosis in new patients with respiratory complaints.
Interpretation
The study shows that spirometry, measurement of lung volumes, airway resistance and
diffusing capacity all independently contribute to diagnosis by reducing the number of
differential diagnoses and increasing diagnostic accuracy. As a consequence, pulmonary
physicians should use these four tests as a battery when confronted with new patients
with respiratory complaints.
Acknowledgements
The writers thank the attending pulmonologists, the participating hospitals and the
patients for their participation in the study. We thank the Belgian Society of Pneumology
for their financial and moral support. We thank Mrs. Kristien De Bent, Ms. Erica Balligand
and Mr. Geert Celis for their expert technical support and Mr. Jurgen Silence for the
18
development of the web-based application. Finally, we thank all the pulmonary function
technicians of the participating hospitals.
Conflict of interest statement
None of the authors had any financial interests or received fees or grant support from any
of the pulmonary function equipment companies.
Contribution of the authors
All of the authors contributed to the design of the study and the performance of the
study. They also all were involved in the interpretation of the results. The first draft of the
manuscript was written by the first author (MD). All of the other authors contributed to
the further development of the manuscript. The decision to submit the manuscript was
borne by the first author (MD).
The Belgian Pulmonary Function study investigators
R.Vanherreweghe (Algemeen Stedelijk Ziekenhuis, Aalst); R.Deman, S.Deryke, B.Ghesyens,
M.Haerens, S.Maddens (AZ Groeninge, Kortijk); W.Bultynck, W.Temmerman, L.Van
Zandweghe (AZ Sint Blasius, Dendermonde); R.De Pauw, C.Depuydt, C.Haenebalcke,
V.Ringoet, D. Van Renterghem (AZ Sint Jan, Brugge); A.Carlier, D.Colle, C.De Cock,
J.Lamont (AZ Maria Middelares, Gent); P.Brancaleone (CH Jolimont-Lobbes); M.Vander
Stappen (CHR Haute Senne); R.Peché, Ph.Pierard, P.Quarré, A.Van Meerhaeghe (CHU,
Charleroi); D.Cataldo, JL.Corhay, B.Duysinx, V.Heinen, M.Naldi, D.Nguyen, L.Renaud (CHU,
Liège); M.Cotils, P.Duchatelet, A.Gocmen, C.Lenclud (Clinique Louis Caty-Baudour);
Ph.Lebrun, J.Noel (Clinique Saint-Pierre, Ottignies); A.Frémault (Grand Hopital de
Charleroi); Ph.Bertrand, B.Bouckaert, I.Demedts, P.Demuynck (Heilig Hart Ziekenhuis,
Roeselare); E.Frans, A.Heremans, T.Lauwerier, J.Roelandts (Imelda, Bonheiden); G.Bral,
19
I.Declercq, I.Malysse (Jan Ypermanziekenhuis, Ieper); V.Van Damme (St. Andries
Ziekenhuis, Tielt); Dr.Martinot (St. Elisabeth, Namur); P.Vandenbrande (St. Maarten
Duffel, Mechelen); M.Bruyneel, I.Muylle, V.Ninane, (St. Pierre, Bruxelles); Ph.Collard,
G.Liistro, B.Mwenge Gimbada, Th.Pieters, C.Pilette, F.Pirson, D.Rodenstein (UCL,
Bruxelles); L.Delaunois, E.Marchand, O.Vandenplas (UCL, Mont-Godinne); W.De Backer,
P.Germonpre, A.Janssens, A.Vrints (UZA, Antwerpen); M.Decramer, W.Janssens, P.Van
Bleyenbergh, G.Verleden, W.Wuyts (UZ Gasthuisberg, Leuven); G.Brusselle, E.Derom, G.
Joos, K.Tournoy, K.Vermaelen (UZ, Gent); P.Alexander, C.Gysbrechts (Ziekenhuis Ronse);
L.Bedert, Dr.Bomans, Dr.De Beukelaar, E.De Droogh, D.Galdermans, Dr.Ingelbrecht,
Dr.Lefebure, H.Slabbynck, Dr.Van Mulders, Dr.Van Schaardenburg, Dr.Verbuyst (ZNA,
Antwerpen); M.Daenen (ZOL Genk).
Tables
Table 1. Baseline characteristics.
Table 2. Symptoms at presentation.
Table 3. Duration of symptoms.
Table 4. Pulmonary function results.
Table 5. Most frequently used and most contributing additional tests according of the
attending pulmonologist.
Table 6. Frequency of the preferred and differential diagnoses after each step.
Table 7. Primary outcome of the study.
Table 8. Primary outcome of the study in four predefined subgroups.
Table 9. Secondary outcome, % correct final diagnoses using a logistic GEE (Generalized
Estimating Equation) model.
20
Figure Legends
Figure 1. Schematic overview of the study design.
Figure 2. Overview of patient disposition.
Figure 3. Histogram of the number of differential diagnoses in cases with COPD (left) and
asthma (right), after the first step (upper panels) and the fourth step (lower panels).
Number of words in the body of the text: 4,514
References
1. Van den Bruel A, Gailly J, Devriese S. Pulmonary function tests 2007. KCE report
60A. www.kce.fgov.be last accessed July 3rd 2013
2. Pratter MR, Curley FJ, Dubois J, Irwin RS. Cause and evaluation of chronic dyspnea
in a pulmonary disease clinic. Arch Intern Med 1989; 149: 2277-2282.
3. Institute for Clinical Systems Improvement (ICSI). Diagnosis and outpatient
management of asthma 2005. www.icsi.org last accessed July 3rd 2013
4. Global Initiative for Asthma (GINA). Global strategy for asthma management and
prevention. 2012 Update. www.ginasthma.org last accessed July 3rd 2013
5. Scottish Intercollegiate Guidelines Network (SIGN), British Thoracic Society. British
guideline on the management of asthma 2005. www.sign.ac.uk last accessed July
3rd 2013
6. Finnish Medical Society Duodecim. Long-term management of asthma. In: EBM
Guidelines. Evidence-Based Medicine, Helsinki Finland, Duodecim Medical
Publications Ltd. 2004. www.ebm-guidelines.com last accessed July 3rd 2013
7. Institute for Clinical Systems Improvement 2005. Chronic obstructive pulmonary
disease. www.ICSI.org last accessed July 3rd 2013
21
8. National Collaborating Centre for Chronic Conditions/ National Institute for Health
and Clinical Excellence 2004. National guideline on management of chronic
pulmonary disease in adults in primary and secondary care. Thorax 2004; 59 Suppl
1: 1-232.
9. Global initiative for Chronic Obstructive Lung Disease. Global strategy for the
diagnosis, management, and prevention of chronic obstructive pulmonary disease.
2013 Update. www.goldcopd.com last accessed July 3rd 2013
10. Finnish Medical Society Duodecim. Chronic obstructive pulmonary disease (COPD).
www.ebm-guidelines.com last accessed July 3rd 2013
11. Miller A and Enright PL. PFT interpretive strategies: American Thoracic Society/
European Respiratory Society 2005 guideline gaps. Respiratory Care 2012; 57: 127133.
12. Ruppel GL. What is the clinical value of lung volumes? Respiratory Care 2012; 57:
26-35.
13. Enright PL. Spirometer + body box = VW beetle + Mercedes? Respiratory Medicine
2011; 105: 957-58.
14. Kanner RE, Renzetti AD Jr, Stanish WM, Barkman HW Jr and Klauber MR.
Predictors of survival in subjects with chronic airflow limitation. Am J Med 1983;
74: 249-55.
15. Anthonisen NR, Wright EC and,Hodgkin JE. Prognosis in chronic obstructive
pulmonary disease. Am Rev Respir Dis 1986; 133: 14-20.
16. Hansen EF, Phanareth K, Laursen LC, Kok-Jensen A and Dirksen A. Reversible and
irreversible airflow obstruction as predictor of overall mortality in asthma and
chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999; 159:
1267-71.
17. Wells AU, Desai SR, Rubens MB, Goh NS, Cramer D, Nicholson AG et al. Idiopathic
pulmonary fibrosis: a composite physiologic index derived from disease extent
observed by computed tomography. Am J Respir Crit Care Med 2003; 167: 962-9.
22
18. Casanova C, Cote C, de Torres JP, Aguirre-Jaime A, Marin JM, Pinto-Plata V et al.
Inspiratory-to-total lung capacity ratio predicts mortality in patients with chronic
obstructive pulmonary disease. Am J Respir Crit Care Med 2005; 171: 591-597.
19. Pellegrino R, Viegi G, Brusasco V, Crapo RO, Burgos F, Casaburi R et al.
Interpretative strategies for lung function tests. Eur Respir J 2005; 26: 948-68
20. Jones PW, Harding G, Berry P, Wiklund I, Chen WH, Kline Leidy N. Development
and first validation of the COPD Assessment Test. Eur Respir J 2009; 34: 648-54.
21. Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A et al.
Standardisation of spirometry. Eur Respir J 2005; 26: 319-38.
22. Rosenberg E. The 1995 update of recommendations for a standard technique for
measuring the single-breath carbon monoxide diffusing capacity (transfer factor).
Am J Respir Crit Care Med 1996; 154: 265-66.
23. Quanjer PH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, Yernault JC. Lung
volumes and forced ventilatory flows. Report Working Party Standardization of
Lung Function Tests, European Community for Steel and Coal. Official Statement
of the European Respiratory Society. Eur Respir J 1993; 6: Suppl 16, 5-40.
24. Holm S. A simple sequentially rejective multiple test procedure. Scandinavian
Journal of Statistics 1979; 6: 65-70.
25. Liang K and Zeger S. Longitudinal data analysis using generalized linear models.
Biometrika 1986; 73: 13-22.
26. Hewitt RS, Modrich CM, Medlicott T, Cowan JO, Taylor DR. Supporting the
diagnosis of non-specific respiratory symptoms in primary care: the role of exhaled
nitric oxide measurement and spirometry. Primary Care Respiratory Journal 2008;
17: 97-103.
27. Sterk PJ, Fabbri LM, Quanjer PH, Cockcroft DW, O'Byrne PM, Anderson SD et al.
Airway responsiveness. Standardized challenge testing with pharmacological,
physical and sensitizing stimuli in adults. 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: 53-83.
23
28. Calverley PM, Burge PS, Spencer S, Anderson JA, Jones PW. Bronchodilator
reversibility testing in chronic obstructive pulmonary disease. Thorax 2003; 58:
659-64.
29. Albert P, Agusti A, Edwards L, Tal-Singer R, Yates J, Bakke P et al. Bronchodilator
responsiveness as a phenotypic characteristic of established chronic obstructive
pulmonary disease. Thorax 2012; 67: 701-708.
30. Boros PW, Enright PL, Quanjer PH, Borsboom GJ, Wesolowski SP, Hyatt RE.
Impaired lung compliance and DL,CO but no restrictive ventilatory defect in
sarcoidosis. Eur Respir J 2010; 36: 1315-1322.
31. Jodoin G, Gibbs GW, Macklem PT, McDonald JC, Becklake MR. Early effects of
asbestos exposure on lung function. Am Rev Respir Dis 1971; 104: 525-35.
32. De Troyer A and Yernault JC. Inspiratory muscle force in normal subjects and
patients with interstitial lung disease. Thorax 1980; 35: 92-100.
33. Fulmer JD, Roberts WC, von Gal ER, Crystal RG. Morphologic-physiologic correlates
of the severity of fibrosis and degree of cellularity in idiopathic pulmonary fibrosis.
J Clin Invest 1979; 63: 665-676.
34. Flaherty KR, King TE Jr, Raghu G, Lynch JP 3rd, Colby TV, Travis WD et al. Idiopathic
interstitial pneumonia: what is the effect of a multidisciplinary approach to
diagnosis? Am J Respir Crit Care Med 2004; 170: 904-910.
35. McDonough JE, Yuan R, Suzuki M, Seyednejad N, Elliott WM, Sanchez PG et al.
Small-airway obstruction and emphysema in chronic obstructive pulmonary
disease. N Eng J Med 2011; 365: 1567-75.
36. Galbán CJ, Han MK, Boes JL, Chughtai KA, Meyer CR, Johnson TD et al. Computed
tomography-based biomarker provides unique signature for diagnosis of COPD
phenotypes and disease progression. Nat Med 2012; 18: 1711-5.
37. Gulati M. Diagnostic assessment of patients with interstitial lung disease. Prim
Care Respir J 2011; 20: 120-7.
38. Miller BH, Rosado-de-Christenson ML, McAdams HP, Fishback NF. Thoracic
sarcoidosis: radiologic-pathologic correlation. Radiographics 1995; 15: 421-37.
24
Appendices
Appendix 1. List of differential diagnoses.
Appendix 2. List of participating hospitals.
Appendix 3. Study protocol.
Appendix 4. Statistical Analysis Plan.
25
Table 1. Baseline characteristics.
Patient characteristic
Age, yrs
Value (N= 979)
53 (16)
Males
503 (51%)
CAT-score, units
14.8 (7.5)
BMI, kg.m2
27 (6)
Allergy/atopy
208 (21%)
Previous Pulmonary Disease
139 (14%)
Reflux
122 (12%)
Ear-Nose-Throat problems
166 (17%)
(Ex-) Smoker
586 (60%)
Pack yrs*
27 (12;40)
Data are median (IQR), n (%) or mean (SD).
*Based on 516 patients, excluding non-smokers.
26
Table 2. Symptoms at presentation.
History
Value (N= 979)
Cough
660 (67%)
Wheezing
294 (30%)
Dyspnea
727 (74%)
Fatigue
131 (13%)
Sputum production
321 (33%)
Thoracic pain
143 (15%)
Hemoptoe
12 (1%)
CAT score
14 (9;20)
Data are n (%) or median (IQR).
27
Table 3. Duration of symptoms.
History
Value (N=979)
Complaints since*
few months
337 (42%)
1 year or more
335 (42%)
few weeks
65 (8%)
one month
47 (6%)
Others (1 day, few days, 1 week)
15 (2%)
Previous consultation GP
605 (62%)
Pulmonary function by GP
34 (3%)
Referral**
GP
212 (24%)
own initiative
475 (54%)
other specialist
192 (22%)
Data are n (%).
*Based on 799 patients; **Based on 879 patients.
28
Table 4. Pulmonary function results.
Pulmonary function variable
n
Value (N= 979)
% predicted
(N=979)
FVC (L)
979
3.5 (2.7;4.3)
99 (86;111)
FEV1 (L)
979
2.6 (1.8;3.3)
89 (72;103)
FEV1/VC
979
75 (66;81)
95 (85;103)
TLC (L)
975
6.0 (1;7.1)
104 (92;115)
RV (L)
975
2.4 (1.9;3.1)
120 (97;149)
FRC (L)
967
3.5 (2.8;4.2)
113 (94;136)
Raw (kPa.L.s-1)
966
0.43 (0.2;0.4)
132 (63;186)
sGaw (s.kPa-1.L)
770
1.0 (0.7;1.4)
114 (80;161)
TL,CO (mmol.min-1.kPa-1)
966
6.8 (5.2;8.6)
79 (64;91)
KCO (mmol.min-1.kPa-1.L-1)
911
1.4 (1.2;1.6)
92 (78;106)
Data are median (IQR).
29
Table 5. Most frequently used and most contributing additional tests according of the
attending pulmonologist.
Test
Value (N=979)
Top 10 most contributing
Top 10 most frequently used
Reversibility
230 (26%)
488 (50%)
Provocation test
218 (25%)
264 (27%)
CT-scan chest
128 (15%)
270 (28%)
Chest x-ray
127 (14%)
645 (66%)
Exhaled NO
77 (9%)
315 (32%)
Ergospirometry
23 (3%)
68 (7%)
Skin prick test
23 (3%)
261 (27%)
Biopsy
11 (1%)
Cardiac echography
8 (1%)
Endoscopy
7 (1%)
91 (9%)
Biochemistry
242 (25%)
Blood gases
73 (7%)
Data are n (%).
30
Table 6. Frequency of the preferred and differential diagnoses after each step.
Differential diagnosis
Preferred diagnosis
N= 979
1
2
3
4
1
2
3
4
COPD
345
323
309
289
230
220
214
229
(35%)
(33%)
(32%)
(30%)
(23%)
(22%)
(22%)
(23%)
907
877
776
607
435
439
405
385
(93%)
(90%)
(79%)
(62%)
(44%)
(45%)
(41%)
(39%)
106
90
64
43
9
8
9
8
(11%)
(9%)
(6%)
(4%)
(1%)
(1%)
(1%)
(1%)
508
414
340
282
57
63
72
68
(52%)
(42%)
(35%)
(29%)
(6%)
(6%)
(7%)
(7%)
153
75
68
46
6
12
13
16
(16%)
(8%)
(7%)
(5%)
(1%)
(1%)
(1%)
(2%)
130
58
57
37
7
8
7
10
(13%)
(6%)
(6%)
(4%)
(1%)
(1%)
(1%)
(1%)
169
88
86
74
22
26
27
22
(17%)
(9%)
(9%)
(8%)
(2%)
(3%)
(3%)
(2%)
272
147
133
100
25
34
33
40
Asthma
Upper Airway
Other obstructive
Neuromuscular
Thoracic/Pleural
Obesity
ILD
31
(28%)
(15%)
(14%)
(10%)
(3%)
(3%)
(3%)
(4%)
Systemic
150
72
63
35
1
0
0
5
sclerosis
(15%)
(7%)
(6%)
(4%)
(<1%)
(%)
(0%)
(1%)
217
150
137
81
27
23
16
11
(22%)
(15%)
(14%)
(8%)
(3%)
(2%)
(2%)
(1%)
197
181
161
93
4
5
5
22
(20%)
(18%)
(16%)
(10%)
(<1%)
(1%)
(1%)
(2%)
515
418
356
312
112
98
126
122
(53%)
(43%)
(36%)
(32%)
(11%)
(10%)
(13%)
(12%)
147
131
109
97
26
21
29
24
(15%)
(13%)
(11%)
(10%)
(3%)
(2%)
(3%)
(2%)
245
220
192
173
18
22
23
17
(25%)
(22%)
(20%)
(18%)
(2%)
(2%)
(2%)
(2%)
Cardiac failure
Pulmonary
vascular
No abnormality
Hyperventilation
Other
Data are n (%).
32
Table 7. Primary outcome of the study.
Test
Mean transformed score
Back-transformed
P-value*
score
(standard error)
Spirometry
-1.229 (0.035)
0.226
Lung volumes
-0.868 (0.041)
0.296
< 0.0001
Resistance
-0.520 (0.054)
0.373
< 0.0001
Diffusing capacity
0.162 (0.067)
0.540
< 0.0001
*P-value comparing the difference with the previous test.
Score was defined by 1/number of differential diagnoses, provided the gold standard
diagnosis was part of the differential diagnoses. The score was corrected for the accuracy
of the diagnosis by attributing a score of 27/28 (= close to 1) if there was only one
differential diagnosis and it was also the correct diagnosis, or 1/28 (= close to 0) if the
correct diagnosis was not among the differential diagnoses. In case the gold standard was
part of the differential diagnoses, but not the preferred diagnosis 1/28 was subtracted
from the score. The transformed score corresponds to the score transformed by logistic
transformation. The back-transformed score corresponds to the score transformed back
to the original format.Table 8A. Primary outcome of the study.
Test
Mean transformed score
COPD
(standard error)
Back-transformed
P-value*
score
33
Spirometry
-1.210 (0.064)
0.230
Lung volumes
-0.755 (0.066)
0.320
< 0.0001
Resistance
-0.319 (0.108)
0.421
< 0.0001
Diffusing capacity
1.153 (0.148)
0.760
< 0.0001
* P-value comparing the difference with the previous test.
Primary outcome of the study in the predefined subgroups: COPD (panel A), asthma
(panel B), ILD (panel C) and no primary pulmonary abnormality (panel D).
Same conventions as in Table 7.
34
Table 8B. Primary outcome of the study.
Test
Mean transformed score
ASTHMA
(standard error)
Back-transformed
P-value*
score
Spirometry
-0.970 (0.050)
0.275
Lung volumes
-0.525 (0.070)
0.372
< 0.0001
Resistance
-0.081 (0.089)
0.480
< 0.0001
Diffusing capacity
0.251 (0.101)
0.562
< 0.0001
35
Table 8C. Primary outcome of the study.
Test
Mean transformed score
ILD
(standard error)
Back-transformed
P-value*
score
Spirometry
-2.035 (0.193)
0.116
Lung volumes
-1.560 (0.209)
0.174
< 0.0001
Resistance
-1.443 (0.217)
0.191
< 0.0001
Diffusing capacity
0.391 (0.317)
0.596
< 0.0001
36
Table 8D. Primary outcome of the study.
Test
Mean transformed score
Pulmonary
(standard error)
Back-transformed
P-value*
score
abnormality
Spirometry
-1.262 (0.037)
0.221
Lung volumes
-0.852 (0.045)
0.299
< 0.0001
Resistance
-0.486 (0.058)
0.381
< 0.0001
Diffusing capacity
0.303 (0.074)
0.575
< 0.0001
37
Table 9. Secondary outcome, % correct final diagnoses using a logistic Generalized
Estimation Equation (GEE) model.
Test
Proportion
%
P-value*
Correct final
diagnoses
Spirometry
597/979
61
Lung volumes
637/979
65
< 0.0001
Resistance
683/979
70
< 0.0001
Diffusing capacity
755/979
77
< 0.0001
* P-value comparing the difference with the previous test.
38
New
outpatients
with
respiratory
symptoms
but without
diagnosis
Spirometry with flow-volume
curve
Absolute volumes
Resistance
Diffusion capacity
History
Clinical Data
LOK/GLEM meeting
Frequency: 4/year
Preferred diagnosis 1
Differential diagnoses set 1
Addition of spirometry with
Flow-volume curve
Preferred diagnosis 2
Differential diagnoses set 2
Addition of absolute
volumes
Preferred diagnosis 3
Differential diagnoses set 3
Addition of resistance
Final diagnosis
+ side diagnosis
Attending
pulmonologist
Preferred diagnosis 4
Differential diagnoses set 4
Addition of diffusion
Laboratory results and
imaging
LOK/GLEM
Final diagnosis
+ side diagnosis
Gold standard
Fig 1.
39
1,285 patients
screened
262 excluded
29 disease already diagnosed
30 <18 yrs or >75yrs
49 not able to perform spirometry
64 preoperative spirometry
65 miscellaneous
1,023 patients
enrolled
44 excluded from analysis
10 inadequate flow-volume loop
34 incomplete investigation
979 patients
analysed
Fig 2.
40
COPD
Step 1
COPD
Step 4
ASTHMA
Step 1
ASTHMA
Step 4
Fig 3.
41
Appendix 1. List of differential diagnosis (checklist).

COPD

Asthma

Upper airway obstruction

Other obstructive diseases including Bronchiectasis,
Cystic fibrosis and Bronchiolitis.

Neuromuscular disease

Thoracic deformity

Lung function abnormalities due to obesity*

Interstitial lung disease

Systemic sclerosis and other vasculitis

Cardiac failure*

Pulmonary vascular disease*

No primary pulmonary abnormality

Hyperventilation syndrome*

Other diagnosis
*Only if clinical evidence was present
42
Appendix 2. Participating Hospitals.
1
Algemeen Stedelijk Ziekenhuis, Aalst
18
Jan Yperman Ziekenhuis, Ieper
2
AZ Groeninge, Kortrijk
19
O.L.Vrouwziekenhuis, Aalst
3
AZ Sint Blasius, Dendermonde
20
St. Andreas Ziekenhuis,Tielt
4
AZ St. Jan, Brugge
21
St. Elisabeth, Namur
5
AZ St. Jozef, Turnhout
22
St. Jean, Bruxelles
6
AZ Maria Middelares, Gent
23
St. Maarten Duffel, Mechelen
7
CHU Brugman, Bruxelles
24
St. Pierre, Bruxelles
8
CHU, Charleroi
25
UCL, Mont-Godinne
9
CHU, Liège
26
UZA, Antwerpen
10
CHR de Namur
27
UZ Gasthuisberg, Leuven
11
Clinique Louis Caty-Baudour
28
UZ, Gent
12
Clinique Saint-Pierre, Ottignies
29
Virga Jesse, Hasselt
13
Cliniques Universitaires St Luc,
30
Ziekenhuis Jolimont
Bruxelles
14
Grand Hospital de Charleroi
31
Ziekenhuis Ronse
15
Heilig Hart Ziekenhuis, Roeselare
32
ZNA, Antwerpen
16
Hopital de la Citadelle, Liège
33
ZOL, Genk
17
Imelda, Bonheiden
43
Download