Evaluation of diffusing capacity of the lung for carbon monoxide normalized per liter alveolar volume as a parameter for assessment of interstitial lung diseases

One of the most important clinical indications of DLCO-SB technique is assessing interstitial lung diseases (ILDs), as there is thickening of the alveolar membrane and a diminished total lung capacity (TLC) due to interstitial processes, which may lead to a severe decline in transfer factor. Th e acinus is disrupted and the diff usion pathway is lengthened. Typical diseases are extrinsic allergic alveolitis, pulmonary vasculitis syndromes, systemic lupus erythematosus, and of course interstitial fi brosis [2].


Introduction
Th e single-breath (SB) test using carbon monoxide (CO) is the most widely used method to measure the pulmonary diff using capacity. Th e result is usually expressed for the whole lung [diff using capacity for carbon monoxide (DLCO)] or per unit alveolar volume (DLCO/VA) [1].
One of the most important clinical indications of DLCO-SB technique is assessing interstitial lung diseases (ILDs), as there is thickening of the alveolar membrane and a diminished total lung capacity (TLC) due to interstitial processes, which may lead to a severe decline in transfer factor. Th e acinus is disrupted and the diff usion pathway is lengthened. Typical diseases are extrinsic allergic alveolitis, pulmonary vasculitis syndromes, systemic lupus erythematosus, and of course interstitial fi brosis [2]. DLCO/VA represents the diff using capacity in the available alveolar spaces. In other words, DLCO/VA determines whether the currently available alveolar spaces are functioning normally [3]. In healthy adults, DLCO/VA is ∼4-5 ml CO transferred/min/l of lung volume [4].
during the period between May 2011 and May 2012 were recruited. ILDs diagnosis was based on the the clinical history, radiographic abnormalities, low DLCO, and 6-min walk testing; according to an offi cial American Th oracic Society (ATS), the European Respiratory Society (ERS), the Japanese Respiratory Society ( JRS), the Latin American Th oracic Association (ALAT) statement, idiopathic pulmonary fi brosis requires evidence-based guidelines for diagnosis and management [6].

Methods
Th e patients underwent spirometry including forced expiratory volume in the fi rst second (FEV1), forced vital capacity (FVC), FEV1/FVC, an d maximal midexpiratory fl ow (MMEF) (Master Lab; Jaeger, Wurzburg, Germany). Th e best results were chosen from three eff orts following the ATS/ERS guidelines 2005 [7].
Th ey also performed DLCO-SB (Master Lab; Jaeger) with the following technique after determination of ambient conditions by ambient unit and ga s analyzer calibration (helium sensor, 2.5-10%; CO sensor, 0.15-0.30%). Th e patient was asked to approach the mouthpiece and to close his nose with nose-clip. Th e patient was instructed to breathe quite normally. After at least three breaths, the patient was instructed to exhale as deeply as possible from normal breathing. After maximal expiration, the patient was requested to inhale fast as deeply as possible according to the ATS/ERS recommendations [7]; inspiration was completed within 2-4 s. Th e patient inhaled a gas mixture of 0.3% CO and a tracer gas helium 10%. Th e occlusion time automatically starts after 1/3 of inspiration. At the end of inspiration, the patient was prevented from expiration for the period of time set as occlusion time. Th e patient was asked to keep the mouthpiece in his mouth and hold his breath for 10 s. Th e pressure curve displayed during the occlusion showed whether the patient had held his breath or whether he had tried to expire or inspire despite the occlusion. After the set occlusion time had expired, the shutter was opened and the patient exhaled smoothly, without hesitation or interruption. Discard volume and sampling volume were exhaled by sampling tube. Th e gas sample collected for analysis remained in the tube. Th e remaining air was exhaled by the opened shutter. Th e sampling valve closed and the patient left the mouthpiece. Th e measuremen t program allows the measurement of DLCO and the following additional parameters: Krogh factor (KCO) (DLCO/VA), VA, TLC, residual volume (RV) , RV/TLC%, and functional residual capacity (FRC).

Statistical analysis
Analysis of data was performed using statistical program for the social sciences (SPSS, version 20; SPSS Inc., Chicago, Illinois, USA) as follows: (1) Description of quantitative variables as mean, SD, and range. (2) Correlation coeffi cient test was used to rank variables positively or inversely using Pearson's correlation, as all variables are parametric (SD <50% mean). (3) Regression linear analysis was performed to compare quantitative variables in parametric data (SD <50% mean).
Th e level of signifi cance was set as: P value greater than 0.05 was considered a nonsignifi cant statistical result. P value less than 0.05 was considered statistically signifi cant result.

Results
Fifty-three patients with ILD (mean age 47.9 ± 13.7 years) participated in this study. Of these patients, 20 were men and 33 were women, and thei r BMI was 29.25 ± 7.16.
Th e mean ± SD spirometric parameters were as follows: Th ere was statistically signifi cant positive correlation between age and FVC and statistically signifi cant negative correlation between age and RV using Pearson's correlation (Table 2).
Height showed statistically signifi cant positive correlation with both DLCO and FRC using Pearson's correlation (Table 3).
Th ere was statistically signifi cant positive correlation between VA and each of the following parameters, TLC, FVC, RV, and FRC, whereas there was statistically signifi cant negative correlation between VA and KCO using Pearson's correlation (Table 4). Th ere was statistically signifi cant relationship between FVC and RV using regression linear analysis (Table 6).
Th ere was statistically signifi cant positive correlation between DLCO and KCO using Pearson's correlation ( Table 7).
Th ere was statistically signifi cant relationship between DLCO and MMEF, RV/TLC, TLC, and KCO using regression linear analysis. In contrast, there was no signifi cant correlation between DLCO and the following parameters, FVC, RV, FRC, and BMI (Table 8).      Th e results showed that KCO had statistically signifi cant positive correlation with DLCO and statistically signifi cant negative correlation with TLC using Pearson's correlation (Table 9).
Th ere was statistically signifi cant relationship between KCO and DLCO, MMEF, RV/TLC, and TLC using regression linear analysis Table 10.

Discussion
Reduction in VA by disease processes is the largest potential source of error in interpreting DLCO. Correction for the eff ect of altered VA has been attempted by reporting the ratio of DLCO/VA [8]. DLCO/VA was introduced in clinical practice mainly to allow for reductions in VA brought about by a loss of pulmonary tissue, as for example, following pneumonectomy [9]. Englert [10] showed, in 74 patients, that pneumonectomy resulted in a reduction of TLC to 58% of predicted, with DLCO and DLCO/ VA being 70 and 114% of predicted, respectively. It is clear that, in such instances, a correction for DLCO by the participant's VA is warranted, as the decrease in DLCO following pneumonectomy is of a totally diff erent nature than that caused by a thickened alveolar capillary membrane, as in lung fi brosis, or by lung destruction, as in emphysema. However, such simple correction for DLCO by VA may not be appropriate in all circumstances [9].
Ayers et al. [9] considered interstitial fi brosis to be an example of loss of lung units, leading to the maintenance of a normal DLCO/VA ratio.
Cotes et al. [11] stated that the predictions for DLCO depend on age, sex, and height. VA depends on sex and height but not on age. In adults, KCO depends inversely on age and height but, in a review of the literature, hardly at all on sex [12].
In this study, we did not fi nd any statistical relationship or correlation between DLCO, DLCO/VA, and age or sex; however, height showed statistically signifi cant positive correlation with DLCO.
Stam et al. [13] stated that DLCO increases and DLCO/VA decreases exponentially with height. As TLC is also exponentially related to height, both DLCO and DLCO/VA are linearly related to TLC.
Th e variability between our results and the other studies may be related to the limited number of patients in our studies, and they based their statements on healthy population.
In this study, there was a statistically signifi cant positive correlation between FVC and TLC-SB, RV/ TLC, and FRC, but there was no correlation of FVC with DLCO and DLCO/VA. Th ese are in agreement with several researchers results [9,14], which show no statistical relationship between DLCO, DLCO/VA, and the parameters of spirometry.
Agusti et al. [2] and Frans et al. [15] observed that, in patients with a restrictive pattern of pulmonary function, DLCO/VA is proportionally less decreased   than DLCO. Th eir results are in acceptance with our results, as the mean of DLCO/VA and DLCO was 76.51 and 45.62, respectively. In addition, they reported that the opposite trend has been observed in patients with an abnormally high VA.
Although our study reported a statistically signifi cant positive correlation between DLCO and KCO, a study on 2313 patients showed large diff erences and much variability between the two parameters [16].
However, such variability between our study and Johnson's study related to diff erent subgroups in his study; there were patients with asthma, emphysema, extrapulmonary lung disease, ILD, and lung resection in his study, whereas our study was on a single group with ILD. In addition, our study was on a limited number of patients compared with his study; Johnson [16] also stated that, as VA decreased, DLCO decreased linearly and KCO increased.
In searching for the validity of the DLCO test as providing an assessment of lung volume, a study [16] was conducted comparing VA with TLC determined by plethysmography. Th e VA provides the lung volume in which helium is distributed during the DLCO test. Th e ILD group had low lung volumes, but their VA was near th eir TLC (VA 91 ± 17% of TLC), which was not the case in groups of moderate-to-severe obstruction where TLC was increased and the VA was lower than TLC, being 58 ± 15% of TLC. In our study, the mean of TLC and VA was 64.8 ± 14.21 and 63.2 ± 14.5, respectively; VA represents 97.5% of TLC.
One of the advantages of the DLCO-SB occurs at TLC level, which is a reproducible reference point. It was reported [17] that VA lays within 10% of TLC and as the diff erence is related to the anatomic dead space and the gas mixing in 10 s breath hold is incomplete, it can be concluded that the VA and TLC-SB can be a good guide for lung volume in patients with ILD.
It is worth to mention the relationship between DLCO and lung volume; however, it is not linear and markedly less than 1 : 1. Hence, these simple ratios as traditionally reported do not provide an appropri ate way to normalize DLCO for lung volum e [18].
In criticizing the DLCO/VA, Forster [19] mentioned that the changes in DLCO with lung volume in patients with mixed airway and alveolar disease are complex, which can make it potentially misleading to use DLCO/VA as an index. Hughes and Pride [20] noted that DLCO/VA does not correct either for failure to reach maximal lung volume or for pathologically reduced lung volumes.
We have to point that DLCO and DLCO/VA are usually compared with predicted values, which are determined in healthy volunteers, who by defi nition have a normal TLC. Th us, the current predicted values relate to measurements made at normal TLC [11].
In patients with a restrictive ventilatory defect (i.e. a reduced TLC) or with a larger than normal TLC, a comparison with predicted values at predicted TLC can lead to erroneous conclusions. A decrease in lung volume will cause a decrease in surface area, and consequently in DLCO. However, DLCO/VA is higher at reduced VAs compared with predicted values estimated at a normal TLC [17]. Stam et al. [12] suggested that, in restrictive pulmonary disease, DLCO and DLCO/VA should be compared with predicted values at a lung volume equal to the patients actual TLC. Th erefore, they derived reference values for DLCO/VA as a function of VA.
Th eir results were corroborated by Chinn et al. [18] and Frans et al. [15], who found a comparable relationship between DLCO and VA. However, a disadvantage of such a method is that both predicted values of DLCO and DLCO/VA at predicted TLC and the volume correction procedure have their own variability [17].
Hughes and Pride [20] stated that KCO enhances understanding of DLCO. It is clear that the nonlinear relationship between KCO and lung volume precludes DLCO/VA from being a volume correction for the DLCO when VA is reduced, but KCO remains a true refl ection of alveolar CO uptake effi ciency at a given volume. Th ey mentioned that the emphasis on DLCO/VA as a correction factor for lung volume is misconceived and refl ects a misapprehension of the physiology. Hence, they believe the term DLCO/VA should be replaced by the more informative term, KCO.
Our fi ndings must be considered in the context of the limitations of this study. First, we did not correct the DLCO values for hemoglobin concentration, as this information was not available in all patients. Th is certainly may have changed the DLCO and DLCO/ VA values; however, it should have changed both equally, and thus not aff ected the primary purpose of our study, which was to compare the two values. Second, we accounted on TLC-SB technique rather than TLC measured by body plethysmography, which is more accurate, because it was not requested by the patients physicians to limit the costs. However, the TLC-SB and VA at low lung volumes give reproducible results we can account on.

Conclusion
In interpretation of DLCO-SB, the DLCO/VA ratio should not be neglected and should be in coherence with the interpretation of DLCO, as decreased DLCO/VA strongly suggests parenchymal lung disease. However, alone it does not provide a valid index of the eff ect of changes in VA; it may lead to errors in interpretation of the diff using capacity. VA and TLC-SB could be good indicators of lung volume in patients with ILD, which need further investigations on a w ide scale.