Background

The single-breath carbon monoxide diffusing capacity (DLCO) is the product of two measurements during breath holding at full inflation: the rate constant for carbon monoxide uptake from alveolar gas (KCO [minute−1]) and the “accessible” alveolar volume (VA). Consequently, the same DLCO may result from various combinations of KCO and VA values. Changes in each of KCO and VA may reflect different injury sites and mechanisms. In theory, the decrease in DLCO may result from a fall in VA (mainly due to restrictive and/or obstructive defects) and/or a fall in KCO (due to alveolar/capillary damage or a microvascular disease). Few studies have focused on the significance of DLCO in diffuse parenchymal lung diseases (DPLD) [1–5], highlighting the prognostic value of its component KCO. No study to our knowledge has sought to assess the validity of the above mentioned theory in the context of DPLD. Our primary objective in the present study was to assess in a large cohort of distinct types of DPLD the potential variability of both VA and KCO for fixed values of DLCO. A secondary objective was to determine whether a low VA value in this context might reflect a distal airway obstruction in addition to a potential restrictive defect. To this end, we designed a retrospective, cross-sectional study of three distinct types of DPLD: idiopathic pulmonary fibrosis (IPF, the prototype for fibrotic pulmonary diseases predominantly affecting the lower lobes), stage IV sarcoidosis (predominantly affecting the upper lobes) and connective tissue disease-associated interstitial lung diseases (CTD-ILDs, which are usually characterized by diffuse, inflammatory lesions rather than fibrotic damage).

Methods

Each of three university hospitals in France provided pulmonary function test (PFT) datasets from around 80 DPLD patients (75, 80 and 87 patients, respectively). Pulmonary function tests had been performed according to international recommendations and had used similar quality criteria [6–8]. Only raw PFT data were provided and % predicted values were subsequently calculated by a single investigator (CD2) for the whole population according to Stanojevic for spirometry [9] and other international recommendations for DLCO and static lung volumes respectively [10, 11]. The PFTs (spirometry, body plethysmography and single-breath carbon monoxide transfer) using routine techniques had been performed for clinical purposes. We got approval from the Institutional Review Board of the French learned society for respiratory medicine – Société de Pneumologie de Langue française, which judged our study as fully observational and which therefore did not require any informed consent.

Two-hundred and forty-two patients with complete datasets were retrospectively assigned to IPF (n = 85), sarcoidosis (n = 73) or CTD-ILD (n = 84) groups. Patients with IPF and CTD-ILD exhibited lower values of DLCO than those with sarcoidosis (43 ± 18 % predicted (11-89 %), 44 ± 15 (12-88 %), and 56 ± 18 % (19-115 %), in IPF, CTD-ILD and sarcoidosis, respectively, p < 0.0001). Then, three PFT datasets (one per group) were matched for DLCO % predicted (agreement 5 %, by a single investigator (CD2)) to allow comparisons of the groups at similar levels of DLCO. Consequently, 77 patients were excluded from the analysis due to matching selection (for instance IPF and CTD-ILD subjects with very low DLCO % predicted values and sarcoidosis subjects with high DLCO values). Results were expressed as means ± SD. Continuous variables were compared using the Student’s t-test or the analysis of variance (ANOVA, see Table) as appropriate. The chi-squared test was used for the comparison of qualitative variables (smoking history). Statistical significance was defined by a p value <0.05. All analyses were performed using the Statview 4 package (SAS institute, Grenoble, France).

Results

One hundred and sixty-five PFT datasets (55 per group) were analysed (Table 1). The three study groups had similar mean values for KCO and VA as well as for DLCO (the matching criterion). However, on an individual patient basis, a similar DLCO could be obtained from various combinations of KCO and VA (Fig. 1). This figure clearly shows that KCO can vary from decreased (diffuse loss of units) to normal or barely increased (discrete loss of units) values. We show here that for a similar DLCO value of 50 % predicted, for instance, KCO varied from 60 to 95 % predicted and VA from 55 to 85 % predicted.

Table 1 Demographic and functional characteristics of the study participants
Full size table
Fig. 1
figure 1

Relationships between DLCO on one hand and VA (left panel) and KCO (right panel) on the other. Circles represent sarcoidosis (closed: with airflow limitation, n = 17; open: without airflow limitation). a Dotted lines describe “reduced expansion” (upper bold line) and “loss of units” effects, calculated according to Hughes and Pride [4]. Patients with DPLD lied in the discrete to diffuse loss of alveolar unit areas. b The dotted line is the identity line for the DLCO-KCO plot; patients along this line have normal VA and the reduced DLCO is related to a decrease in KCO due to microvascular pathology

Full size image

In addition, 17 patients exhibited an airflow limitation (FEV1/FVC < lower limit of normal). They all belonged to the sarcoidosis group (Table 1). The reduction in alveolar volume (measured using a dilution technique) relative to total lung volume (TLC, measured using body plethysmography), expressed as VA/TLC, was correlated with parameters of central airway obstruction (FEV1/FVC: r2 = 0.10, p < 0.001) and even more strongly with distal airway obstruction (RV/TLC: r2 = 0.25, p < 0.001). Since the VA/TLC value of the population as a whole may seem lower than expected (Table 1) even in patients without significant airflow limitation (n = 148, FEV1/FVC = 0.82 ± 0.06), we further evaluated whether some patients exhibited a small airways obstructive syndrome defined by a normal FEV1/FVC ratio and a greater reduction of both FEV1 and FVC than TLC (FVC % predicted/TLC % predicted < 0.80). We found 20 such subjects, described in Table 2. Similarly to proximal airflow limitation, small airways obstructive syndrome was predominantly present in sarcoidosis.

Table 2 Small airway obstructive syndrome (SAOS) in patients without proximal airflow limitation (FEV1/FVC > lower limit of normal)
Full size table

Discussion

Our present study confirms that an abnormally low DLCO can result from very different combinations of the primary measurements KCO and VA. This was the case for all three types of DPLD. Furthermore, the assessment of VA/TLC [12], the latter being obtained from body plethysmography, may suggest both central or peripheral airway obstruction and this was observed particularly in sarcoidosis thereby providing additional clues to the pathogenic features of this condition. We recently described diseases associated with a small airway obstructive syndrome (a non-specific pattern frequently observed in pulmonary function testing units [13]). It is noteworthy that in that study, sarcoidosis and interstitial pneumonia were two of the conditions associated with this pattern. In the present work, we extend our previous data showing that a DPLD can exhibit a mixed pattern associating both a restrictive syndrome and a small airways obstructive syndrome.

Conclusions

In conclusion, we confirmed that the components of DLCO (KCO and VA) may largely vary in DPLD while DLCO appears constant. The magnitudes of KCO and VA values might indicate distinct disease mechanisms and thereby bear a relative prognostic value in addition to giving clues to pathogenesis of these diseases. For these reasons, clinicians should take into account not only DLCO but also VA and KCO when seeking to assess DPLD, in order to provide a more informed and better care to these patients.