Lung Clearance Index is Increased in Patients with COPD-LCI Measurements in the Daily Routine

Results: 18 patients were not able to finish measurements successfully because of significant leaks, cough and irregular breathing (GOLD I n=1 (5.6%); GOLD II n=6 (33.3%); GOLD III n=5 (27.7%); GOLD IV n=6 (33.3%). The mean LCI was significantly higher in COPD patients as compared to young and healthy volunteers (12.55 ± 3.50 vs 7.00 ± 1.02, p<0.05). Although LCI correlated with FEV1% of predicted (r2=-0.540, p<0.01) and Rtot (% pred.) (r2=0.504, p<0.01) the method failed to discriminate between GOLD II-IV classes. LCI correlated with the volume of trapped gas in elderly patients with COPD and young and healthy controls [FRC (% pred.) (r2=0.191, p<0.01), ITGV (% pred.) (r2=0.478, p<0.01), and RV (% pred.) (r2=0.462, p<0.01)]. Moreover, the results did not correlate with the 6-MWT, a validated clinical outcome parameter.


Introduction
Chronic obstructive pulmonary disease (COPD) is a progressive lung disease caused by chronic smoke inhalation and characterized by persistent airflow obstruction to the lungs that is not fully reversible [1]. Severity classification has been based on FEV 1 , which correlates poorly with clinically relevant outcomes such as health-related quality of life, breathlessness, and exercise capacity [2][3][4][5][6][7].
Therefore, there is the need of new methods for better assessment of disease severity and health status. Our present study aimed to evaluate the feasibility of multiple breath wash out (MBW) for clinical routine. We asked whether LCI correlates with already well proven clinical outcome parameters for COPD SGRQ, number of exacerbations and 6-MWT.
This method has been developed as a noninvasive tool to obtain insight into ventilation homogeneity [8]. The principle of this technique has been described more than 50 years ago and involves the measurement of the exhalation of an inert tracer gas [9]. In case of nitrogen washout, the test is performed by using 100% oxygen. If helium or SF6 are used as tracer gases, the washout is usually performed by switching to air breathing and measurement of the tracer gas. The lung clearance index (LCI) is calculated from the washout measurements and represents the multiple of the functional residual capacity (FRC) that is necessary to obtain an end tidal concentration of the tracer gas of 1/40 of the initial concentration (LCI=cumulative expired volume/ FRC) [10]. Several studies describe the use of LCI measurements to characterize lung disease, most studies were performed in children as the LCI measurement is non-invasive and does not need active respiratory maneuvers. Studies in infants or children with cystic fibrosis (CF) revealed that MBW and LCI measurements are feasible and able to detect lung disease in an early stage [11][12][13][14][15]. LCI measurements were associated with airway infection, inflammation, pulmonary function tests (PFTs), and ventilation inhomogeneity. Studied in children with asthma showed that the LCI is elevated in this disease [16]. Whereas a recent study showed that LCI is increased in patients with COPD [17], Page 2 of 5 cough or irregular breathing occurred.
The functional residual capacity was also measured using the EasyOne ProLab™ by analyzing the expired N 2 trace. Calculation of the LCI was according to standard guidelines of the ERS/ATS [8].

Limitations of the method
We stated a high drop-out rate with 18 patients who were not able to finish measurements successfully because of significant leaks, cough and irregular breathing (GOLD I n=1 (5.6%); GOLD II n=6 (33%); GOLD III n=5 (27.7%); GOLD IV n=6 (33%). The time for each measurement took more than 20 minutes (dependent on the grade of ventilation inhomogeneity). Repeats of the measurements were very often refused by the patients because the inhalation of dry oxygen gas seemed to be uncomfortable. Moreover, measurements were very often interrupted by cough which lead-as well as leakages at the mouth piece-to invalid measurements. A further limitation was that wash-out measurements were not possible in patients with the need of oxygen.
Only a small number of patients with GOLD IV (n=6) completed the study (50% of patients with COPD GOLD IV). Except for two subjects, most of our young and healthy controls were fit to complete the measurements without any disturbances (n=30).

Statistical analysis
Data are displayed as mean ± standard deviation. Data were analyzed using SPSS (Version 22, IBM, New York, USA). Overall group comparison between GOLD II-IV and healthy controls was performed by ANOVA-test/Kruskal-Wallis-test. In order to test pairwise median group differences, we used the Mann-Whitney-U-test. Pearson linear correlation analysis between LCI and other continuous variables were performed and the results are presented as r 2 and resulting p-values. A P value of 0.05 was regarded as significant.

Study population
We obtained LCI data of 23 (out of 41) patients with COPD (BMI 24.62 ± 4.32). We determined smoking exposure. Patients who complete LCI measurements were former smokers and 14 COPD patients were found to be current smokers (pack years of COPD patient's 59.55 ± 38.43 years). The mean FEV 1 % pred. was 40.41 ± 14.65 (female patients 40.33% ± 15.13%, male patients 40.45% ± 14.50%). Patients with successful measurements were classified according to FEV 1 -based GOLD stages as II (n=5) moderate, III (n=12) severe, IV (n=6) very severe ( Figure 1). Patients reported no exacerbation in the past 4 weeks. In 15 (of 23) patients with COPD, the number of exacerbations was 1.1 ± 2.6 last 12 months. The 30 healthy individuals which finished the wash out measurements successfully (2 drop-outs because of cough) were more frequently females (n=19, males n=11) with a mean age of 44.66 ± 14.61 years. All control subjects (BMI 26.50 ± 6.18 kg/m 2 ) were never-smokers with a normal FEV 1 % pred. (105.45% ± 15.07%). Data on the patients and controls are summarized in Table 1.

Especially patients with GOLD IV failed in valid measurements
The time for each measurement took more than 20 minutes (dependent on the grade of ventilation inhomogeneity). Repeats of the measurements were very often refused by the patients because the inhalation of dry oxygen gas seemed to be uncomfortable. Moreover, measurements were very often interrupted by cough which lead-as well as leakages at the mouth piece -to invalid measurements. A further no data exist concerning suitability for the daily use in elderly patients with COPD. An abnormal LCI reflects ventilation inhomogeneity due to airway pathology, suggesting that this measurement could be a suitable noninvasive outcome measure in COPD. Until now, there exist no data about the mechanical handling of the method in the clinical routine of elderly patients with COPD and the relationship to the clinical outcome parameters SGRQ, exacerbation rate, and 6-MWT has not been investigated.
The aim of the present study was to investigate whether LCI measurements are feasible in patients with different FEV 1 stages of COPD and whether the LCI is correlated with classical PFTs. We hypothesized that LCI would be elevated in COPD and correlate with clinical outcome parameters such as SGRQ, rate of exacerbations, and 6-MWT.

Subjects
In this prospective, cross-sectional, and observational study, patients were recruited from the Department of Internal Medicine V from January 2013 to January 2014. The diagnosis of COPD in all patients was based according to the GOLD, definition on a postbronchodilator FEV 1 /FVC<70% pred., history of >10 pack years smoking, and exclusion of other obstructive lung diseases. As healthy controls we included never smokers with normal pulmonary function tests (PFTs). Exclusion criteria for all individuals were: A respiratory infection within 4 weeks and the need of oxygen. This study was approved by the ethics committee of the Ärztekammer des Saarlandes and informed consent was obtained.

Clinical measurements
Pulmonary function tests (PFTs) by spirometry and body plethysmography were performed following the ATS/ERS guidelines [18]. Individuals performed a minimum of three FVC and FEV 1 maneuvers, maximum expiratory flows at 75%, 50%, 25% of FVC (MEF 75, MEF 50, MEF 25) was recorded. The residual volume (RV), total lung capacity (TLC), and gas volume in the thorax (ITGV) were determined by body plethysmography. The diffusion capacity for carbon monoxide (DLCO) was determined by the single breath technique following the ERS/ATS guidelines [19].
Smoking status (current or former) and smoking history (packyears) were determined. In a subset of patients, we recorded data on COPD related symptoms (St. George's Respiratory Questionnaire), the presence of comorbidities (Charlson comorbidity index, CCI) [20], the exercise capacity by the 6 minute walking test (6-MWT), and the number of exacerbations during the last 12 months. Exacerbations were defined by use of antibiotics, steroids, or both or admission to the hospital related to worsening respiratory symptoms.

Measurement of the LCI by MBW
MBW measurements were performed by using an EasyOne ProLab™ (ndd, Switzerland) according to the manufacturer's description with ambient N 2 as tracer gas. Patients with the need of continuous oxygen supplementation were excluded because the pre-test dilution of the ambient N 2 in the lungs makes the assessment of the start N 2 concentration difficult. MBW/LCI measurement was integrated into a clinical workflow. After taking the medical history, measurement of classical PFTs, the individuals were subjected to MBW testing. N 2 was determined by indirect method using the molar mass and the CO 2 signal. An MBW measurement was declared invalid if significant leaks, Figure 1: The LCI is increased in patients with COPD (all COPD stages II-IV). There was no difference between the spirometric stages II-IV.*P<0.01. limitation was that wash-out measurements were not possible in patients with the need of oxygen.

Patients with COPD
Only a small number of patients with GOLD IV (n=6) completed the study (50% of patients with COPD GOLD IV). Except for two subjects most of our young and healthy controls were fit to complete the measurements without any disturbances (n=30).

The LCI correlated with airway obstruction but failed to discriminate between GOLD I-IV
The LCI of patients with COPD was significantly higher as compared to healthy and young controls (p<0.05) (Figure 1). Correlation analysis between spirometric airflow obstruction and the LCI showed negative association between the LCI and the FEV 1 (% pred.) (r²=0.540, p<0.01) (Figure 2a), MEF75 (r²=0.458, p<0.01) (Figure 2b), MEF50 (r²=0.403, p<0.01) (Figure 2c), MEF25 (r²=0.309, p<0.01) (Figure 2d), and a positive correlation between the LCI and R tot (% pred.) (r²=0.504, p<0.01) in the whole study population (young and healthy plus elderly patients with COPD) (Figure 2e). Correlation analysis limited to  p<0.01), and the R tot % (total resistance, percent of predicted). R 2 is indicated in the graph COPD patients showed a significant correlation between the LCI and R tot (% pred.) (r²=0.295, p<0.05) (not shown). There was no significant correlation between the LCI and FEV 1 (% pred.), MEF 75%, MEF 50%, and MEF 25% in the COPD group. These data show that LCI was significantly increased in patients with COPD compared to young and healthy controls and this was correlated with spirometric measures of airway obstruction and airway resistance. Overall group comparison of median LCI between healthy controls and the three groups GOLD II-IV was significant p<10 -3 . Within the elderly patients with COPD there were no significant differences of median LCI between GOLD II, III and IV (pairwise comparisons between GOLD II and III as well as GOLD III and IV). GOLD I was not considered because of the small sample size n=1 (no sample representation for the population).

The LCI correlated with hyperinflation but failed to show significant differences between elderly patients with COPD GOLD I-IV
Pulmonary hyperinflation and ventilation inhomogeneity is a hallmark of COPD and is correlated to clinical outcomes [8]. We asked whether the LCI is correlated with measurements of hyperinflation obtained from body plethysmography. FRC (% pred.), ITGV (% pred.) and RV (% pred.) were significantly elevated in patients with COPD as compared to normal controls (p<0.01). We found that LCI was correlated with the volume of trapped gas in patients with COPD. LCI showed significant correlation with FRC (% pred.) (r²=0.191, p<0.01) (Figure 3a), ITGV (% pred.) (r²=0.477, p<0.01) ( Figure 3b) and RV (% pred.) (r²=0.462, p<0.01) (Figure 3c) in the whole study population. Correlation analysis limited to elderly COPD patients showed no significant correlation between the LCI and FRC (% pred.), ITGV (% pred.), and RV (% pred.) (data not shown).

The LCI was not correlated to patient related outcomes
We next asked whether the LCI was correlated with other outcomes such as the Charlson comorbiditiy index, the number of exacerbations, the SGRQ scores, and the walking distance in the 6-MWT. The LCI neither showed a significant correlation to the Charlson comorbiditiy index (n=23, r²=0.020, p=n.s.) nor to the number of exacerbations (n=15, r²=0.00, p=n.s.). LCI measurements showed a trend to increase with higher SGRQ (n=8, r²=0.402, p=0.092). No significant relationship between the LCI and the walking distance of the 6-MWT could be detected (n=16, r 2 =0.027, p=n. s.).

Discussion
The main findings of the present study are that LCI measurements are feasible in patients with COPD and that the LCI is increased as compared to normal individuals. LCI measurements were associated with airway obstruction (FEV 1 , MEF75, MEF50, MEF25, and R tot ) and with pulmonary hyperinflation (FRC%, ITGV%). Even though LCI measurement is possible in patients with COPD and capable of detecting inhomogeneous ventilation (airway obstruction and hyperinflation), the method has some disadvantages: Compared to body plethysmography, the LCI measurements are more time consuming: Each measurement took more than 20 minutes (dependent on the grade of ventilation inhomogeneity). Moreover, this method excludes patients with severe respiratory failure: Measurements of patients with concurrent oxygen application resulted in invalid measurements. Especially patients with high grade impairment of the lung function refused LCI measurements because the inhalation of dry oxygen gas seemed to be uncomfortable. Moreover, measurements were interrupted by cough several times which lead-as well as leakages at the mouth piece-to invalid measurements. This explained the high drop-out rate of COPD patients (18 out of 41) in our study. Only a small number of patients with GOLD IV completed the study which led to a too small number of patients to draw definite conclusions. This could explain why we could not find any significant correlations between LCI and the different FEV 1 stages of COPD (LCI was stable and did not increase with decreasing FEV 1 ). In addition, there was no significant correlation between LCI and 6-MWT, an important clinical outcome parameter. Currently the data on LCI measurements in adults and patients with COPD are limited. As one of the advantages of LCI measurement is its non-invasiveness, the recent studies focus on children with cystic fibrosis or asthma [11][12][13][14][15][16]. The results of the present investigation on healthy adults (7.58 ± 1.57) are in concordance with data from a N 2 based LCI study in 284 healthy subjects [21], which showed a mean LCI for the age groups: 7-19 yrs: 6.54 ± 0.28; 20-39 yrs; 6.70 ± 0.36; 40-59 yrs: 7.28 ± 0.43; 60-70 yrs: 7.78 ± 0.62. Other studies described the LCI in healthy adults with 6.94 ± 0.64 (n=40, age 40.7 ± 10.3 yrs) [17] and 6.7 ± 0.4 (n=48, age 33 (19-58)) [13]. According to other studies, we could detect a positive correlation between the LCI and the age as described earlier [22].
In conclusion, the present study gives information about the feasibility of LCI measurement in patients with COPD in daily routine. The LCI is increased in all FEV 1 stages investigated (II-IV) and correlates with airway obstruction and hyperinflation. However, lung clearance measurements require a large effort by the elderly COPD patients, which may be difficult for application in daily routine. Nevertheless, LCI might be a tool to gain insight into lung pathophysiology in patients with COPD. Our preliminary study encourages further investigations for the future.