MULTIPLE BREATH WASHOUT: A NONINVASIVE TOOL FOR IDENTIFYING LUNG DISEASE IN SYMPTOMATIC MILITARY DEPLOYERS

Abstract

RATIONALE

Military deployments to austere environments since 9/11/2001 may put “deployers” at risk for respiratory disease. Sensitive, noninvasive tools for detecting large and small airways injury are needed to identify early disease and help inform management for this at-risk population.

OBJECTIVES

We examined multiple breath washout (MBW) as a tool for identifying deployment-related airways disease and assessed host and exposure risk factors compared to healthy controls.

METHODS

Between March 2015 and March 2020, 103 healthy controls and 71 symptomatic deployers with asthma and/or distal lung disease completed a questionnaire, spirometry and MBW testing. SAS v. 9.4 was used to compare MBW parameters between deployers and controls via univariate analyses and adjusted for demographic factors using multiple linear regression.

MEASUREMENTS AND MAIN RESULTS

Deployers were significantly more likely than controls to have an abnormal lung clearance index (LCI) score indicating global ventilation inhomogeneity. Adjusting for sex, smoking status, smoking pack-years and body mass index, LCI scores were significantly more abnormal among those with deployment-related asthma and distal lung disease compared to controls. The unadjusted variable Sacin (a marker of ventilation inhomogeneity in the acinar airways) was higher and thus more abnormal in those with both proximal and distal airways disease. Deployers who reported more frequent exposure to explosive blasts had significantly higher LCI scores.

CONCLUSIONS

This study demonstrates the utility of MBW in evaluating exposure-related airways disease in symptomatic military personnel following deployment to austere environments, and is the first to link exposure to explosive blasts to measurable small airways injury.

INTRODUCTION

Following deployment to Southwest Asia and Afghanistan since September 11^th, 2001, an unknown number of military personnel (“deployers”) developed persistent and sometimes disabling respiratory symptoms, some due to bronchiolitis and asthma (1–5). While deployed, many report exposure to high concentrations of particulate matter from sandstorms, diesel combustion, and burning waste (6–10) that can cause airway injury. Structural damage to the airway may also occur from physical trauma following combat-related explosions (11, 12). Deployment-related respiratory disease clinical findings are often nonspecific, particularly in those with indolent constrictive bronchiolitis on thoracoscopic lung biopsy (1, 4). Because lung biopsy is expensive, invasive and poses risks, noninvasive tools for diagnosis of indolent small airways injury are needed.

Ventilation inhomogeneity measured using multiple breath washout (MBW) and assessed by lung clearance index (LCI) score may be a useful hallmark of early peripheral airways injury. MBW testing has utility in detecting early disease in cigarette smokers (13) and in quantifying ventilatory changes in both mild and uncontrolled asthma (14, 15) and chronic obstructive pulmonary disease (COPD) (16). Additionally, MBW is helpful in detecting early inflammatory distal lung disease in children and adults with cystic fibrosis (CF) (17–19) and in those with post-transplant bronchiolitis obliterans (20). It has also shown utility as a marker of small airways disease in patients with bronchiectasis (21), pulmonary hypertension (22), and primary ciliary dyskinesia (23). In a pilot study of symptomatic military deployers, we found that MBW/LCI may be a useful noninvasive tool for diagnosis of distal airways disease (24). However, conclusions were limited by small numbers of participants and by differences in body mass index (BMI) between deployers and healthy controls.

This study builds on our previous findings by substantially expanding the number of deployers and controls. We also explore the utility of additional MBW parameters reflecting ventilation inhomogeneity in the acinar and conducting airways. These parameters — referred to as phase III slope analyses — are used to separate ventilation heterogeneity into components based on flow in the distal conducting airways (Scond) versus that arising from obstruction closer to the gas diffusion-convection front in the lung acinus (Sacin) (16). While LCI is a global measure of ventilation inhomogeneity, Scond and Sacin indicate the amount of ventilation inhomogeneity localized in the conducting and acinar airways where bronchiolitis occurs (Figure 1).

Figure 1.

Transition between conducting and acinar airways: H&E high-power view shows the transition or respiratory zone between conducting and acinar airways. Terminal bronchioles are the last segment of the conducting portion of the respiratory tract and are lined with ciliated columnar epithelium. They divide to form the respiratory bronchioles (RB) which marks the beginning of the acinus. The respiratory bronchioles (lined with more cuboidal epithelium that may or may not have cilia) communicate with the alveolar ducts leading to multiple alveolar spaces, where gas exchange in the lung occurs. The Scond parameter, determined from multiple breath washout of nitrogen or an inert gas, is a measurement of ventilation heterogeneity in the conducting airways while the Sacin measurement indicates ventilation heterogeneity in the acinar region of the lung.

Our aims were two-fold: to explore the potential role of MBW as a tool for identifying deployment-related lung disease and to evaluate associations between MBW and both demographic and deployment exposure variables. We hypothesized that, compared to spirometry or diffusion capacity for carbon monoxide (DLCO), LCI and phase III slope analyses would be superior for identifying airway injury following military deployment, and that deployment duration and/or exposure to explosive blasts would confer risk for airways dysfunction detectable by MBW.

METHODS

Study Populations

With institutional review board approval and informed consent (HS-2851/HS-2985), deployers were recruited for study participation following clinical evaluation at National Jewish Health in Denver, Colorado. Participants had deployed for at least 30 days to Iraq, Afghanistan, and/or other southwest Asia areas of conflict between 09/2001 and 08/2017 and were self-referred or referred by treating physicians for evaluation of new onset, persistent post-deployment respiratory symptoms (including exertional dyspnea, chest tightness, wheezing and cough) that limited their ability to meet military physical fitness requirements. Deployers were classified as those with deployment-related asthma (DRA), deployment-related distal lung disease (DDLD), or with both diagnoses based on published case definitions (4). Briefly, we defined definite DDLD as the presence of histopathologic findings of hyperinflation/emphysema, bronchiolitis, small airways inflammation, peribronchiolar fibrosis, or granulomatous pneumonitis on surgical lung biopsy. Those defined as having probable DDLD did not undergo surgical lung biopsy, but had two or more findings on chest CT scan of centrilobular nodularity, air trapping, mosaicism, or bronchial wall thickening (25). DRA diagnosis required a post-bronchodilator increase in FEV1 ≥ 12% and ≥ 200 milliliters, or airways hyper-responsiveness based on a 20 percent drop in FEV1 (PC[−20] FEV1) ≤16 mg/mL on methacholine challenge (26). We collected questionnaire information on deployment duration and frequency of exposure to explosive blasts. Healthy controls had no chronic respiratory disease and no acute respiratory illness within four weeks before testing. For all subjects, demographic and medical information (including age, sex, race/ethnicity, smoking status, smoking pack-years, and BMI) was collected.

Pulmonary Function Testing

Pre-bronchodilator spirometry was performed using American Thoracic Society guidelines approximately 15 minutes prior to MBW testing for control subjects (27, 28). All clinical pulmonary function testing including those requiring forced maneuvers was performed on deployer subjects at a minimum of four hours before or after MBW testing and, in the majority, on different days in the same week. Normal values for forced vital capacity percent predicted (FVCPP), forced expiratory volume in one second percent predicted (FEV₁PP), FEV1/FVC ratio, and forced expiratory flow at 25–75% of FVCPP (FEF_25–75%PP) were above the lower limits of normal (LLN) based on published reference values (29). Healthy controls with abnormal spirometry were excluded. DLCO was measured on 62 of 71 participating deployers but not on controls. DLCOPP was defined using the Global Lung Function Initiative reference values (30).

Multiple Breath Washout/Lung Clearance Index Testing

MBW was performed with Exhalyzer D and Spiroware software (EcoMedics AG). Subjects were seated, and wore a noseclip during testing. Results are the average of at least two acceptable trials. Trials were deemed unacceptable if 1) the tracer gas did not re-equilibrate between trials; 2) there was clear evidence of a leak; 3) the breathing pattern was erratic; or 4) the trial did not meet end of test criteria described by Jensen et. al. (31). Subjects began normal breathing on room air to establish tidal volume. The washout phase used 100% oxygen and concluded when subjects’ nitrogen concentrations were below 1/40 or 2.5% of their initial concentration.

LCI is calculated as the cumulative expired volume (CEV) during the number of turnovers required to reach washout divided by the functional residual capacity (FRC) (32). We defined an abnormal LCI score as greater than 7.5 (33). Phase III slope parameters normalized for tidal volume (Sacin*VT and Scond*VT) were calculated based on published literature (34). For all MBW parameters, higher values indicate more ventilation heterogeneity from small airways abnormalities.

Data Analysis

Data were recorded in a Research Electronic Data Capture database (35, 36). Analyses were conducted in SAS v. 9.4. To compare demographics, spirometry, and MBW parameters between deployers (overall and between diagnostic groups) and controls, we used t-tests and ANOVA for continuous variables and Chi-square or Fisher Exact tests for categorical variables. Pairwise testing was conducted when the overall ANOVA F-test was statistically significant (p<0.05), using Tukey-adjusted t-tests and Bonferroni-adjusted Chi-square tests. Adjusted analyses were conducted using multiple linear regression to compare MBW parameters between deployer diagnostic groups and controls while adjusting for demographic characteristics that were significantly different between groups in univariate analyses. [See Appendix A for full description of statistical methods].

RESULTS

Between March 2015 and March 2020, we performed LCI testing on 71 subjects with deployment-related lung disease and 103 healthy controls. Age, race/ethnicity and total pack-years of smoking were similar between deployers and controls (Table 1). There were significantly more males and former smokers among deployers, while controls had significantly more never smokers. Mean BMI for deployers was significantly higher than controls.

Table 1:

Comparisons of demographic, spirometric, diffusion, and multiple breath washout parameters between healthy controls and deployers with deployment-related asthma (DRA) only, deployment distal lung disease (DDLD) only, or both DDLD and DRA

	Controls n = 103	All Deployers n = 71	p-value, deployers vs. controls*	DDLD and DRA n = 18	DDLD Only n = 44	DRA Only n = 9	p-value, diagnosis groups vs. controls†
Demographics and Clinical Characteristics
Male	52 (50.5)	61 (86.0)	<0.0001	17 (94.4)	38 (86.4)	6 (66.7)	<0.0001
Age (years)	39.3 (±13.5)	42.4 (±9.5)	0.08	41.9 (±6.2)	44.1 (±10.3)	35.0 (±7.5)	0.07
Non-Hispanic white	87 (84.5)	56 (78.9)	0.34	13 (72.2)	35 (79.6)	8 (88.9)	0.59
Smoking status†		<0.0001					0.003
Never smoker	88 (85.4)	43 (60.6)	0.0002	13 (72.2)	25 (56.8)	5 (55.6)	0.0008
Former smoker	12 (11.7)	24 (33.8)	0.0004	5 (27.8)	16 (36.4)	3 (33.3)	0.003
Current smoker	3 (2.9)	4 (5.6)	0.45	0 (0.0)	3 (6.8)	1 (11.1)	0.26
Pack-years	9.8 (±10.3)	10.7 (±12.2)	0.82	13.1 (±11.5)	11.7 (±13.2)	2.5 (±2.6)	0.49
BMI (kg/m2)	26.3 (±5.0)	31.4 (±5.4)	<0.0001	32.1 (±6.0)	31.5 (±4.4)	29.1 (±7.9)	<0.0001
Pulmonary Function Testing‡
FEV1PP	101.3 (±10.6)	90.1 (±14.2)	<0.0001	89.3 (±15.5)	88.1 (±13.5)	101.6 (±10.0)	<0.0001
FVCPP	101.2 (±11.0)	89.8 (±12.5)	<0.0001	91.8 (±10.5)	86.9 (±13.2)	100.0 (±5.2)	<0.0001
FEV1/FVC ratio	81.2 (±4.7)	80.0 (±5.8)	0.16	77.1 (±6.1)	80.4 (±5.2)	83.3 (±6.5)	0.02
FEF 25–75PP	105.7 (±24.0)	96.5 (±30.3)	0.05	87.2 (±31.9)	96.6 (±28.0)	113.9 (±34.2)	0.03
Abnormal spirometry	0 (0)	15 (23.8)	<0.0001	3 (20.0)	12 (30.0)	0 (0)	<0.0001
DLCOPP	--	--	--	124.2 ±19.7	113.2 ±18.7	122.5 ±9.3	0.09
Multiple Breath Washout Parameters
LCI Score	7.14 (±0.74)	8.08 (±1.43)	<0.0001	8.49 (±1.96)	7.98 (±1.24)	7.72 (±0.84)	<0.0001
LCI Score > 7.5	28 (27.2)	41 (57.8)	<0.0001	11 (61.1)	26 (59.1)	4 (44.4)	0.0006
Sacin*VT	0.09 (±0.08)	0.12 (±0.10)	0.03	0.16 (±0.16)	0.10 (±0.05)	0.12 (±0.09)	0.01
Scond*VT	0.019 (±0.014)	0.025 (±0.020)	0.02	0.028 (±0.025)	0.025 (±0.018)	0.020 (±0.014)	0.06

Values presented are the mean (±standard deviation) for continuous variables or the count (%) for categorical variables.

^*P-values are from t-tests for continuous variables using the Satterthwaite result for unequal variances and pooled result for equal variances between groups. Categorical variables were compared via Chi-Square test or Fisher Exact test. Only p-values <0.003 are bolded, adjusting for multiple comparisons via Bonferroni correction (0.05/17 tests).

^†P-values for overall ANOVA F-test between controls, DDLD &DRA, DDLD only and DRA only for continuous variables and Chi-Square or Fisher Exact tests for categorical variables. Bolded values were significantly different (p<0.05) from the control group using Tukey pairwise t-tests if the overall ANOVA test was statistically significant. The overall test for smoking status is presented in the header row, with individual Chi-Square or Fisher Exact test results presented on the individual category lines. Pairwise comparisons for categorical variables were adjusted via Bonferroni correction (0.05/12 tests), with only p-values <0.004 bolded.

^‡Spirometry data available from 78 controls and 63 deployers.

Abbreviations: BMI = body mass index, FEV₁PP = percent predicted forced expiratory volume in 1 second, FVCPP = percent predicted forced vital capacity, FEV₁/ FVC ratio = forced expiratory volume in 1 second/forced vital capacity, FEF 25–75PP = percent predicted forced expiratory flow at 25–75% of lung volume, DLCOPP = percent predicted diffusion capacity for carbon monoxide, LCI = lung clearance index, VT = tidal volume, DDLD = deployment-related distal lung disease, and DRA = deployment-related asthma.

As expected, mean FEV₁PP and FVCPP were significantly higher in healthy controls compared to deployers (Table 1). FEV1/FVC ratio and FEF25–75PP were not significantly different between these groups. Means for all spirometric values were normal for both deployers and controls, and only 15 (24%) deployers had any abnormal spirometric value.

Among the 71 deployers, 18 had both DDLD and DRA, 44 had DDLD alone, and nine had DRA alone (Table 1). The majority (58%) of deployers with DDLD and/or DRA had an abnormal LCI (>7.5), and deployers as a whole were significantly more likely than controls to have an abnormal LCI score (Table 1). In controls (n=70) who never smoked and did not have an obese BMI, mean LCI score for males was 7.13 ±0.80 and females was 6.97 ±0.61 (37). Comparing demographic, clinical, and pulmonary function findings between controls and the three deployment lung disease diagnostic groups, trends were similar to results for all deployers. However, those with DDLD alone or both DDLD and DRA had significantly lower FEV₁PP and FVCPP (Table 1).

Among 63 deployers who completed spirometry, 24 (38%) with abnormal LCI had no abnormalities on spirometry, while 12 (19%) had abnormalities on both methods, and three (5%) had only spirometric abnormalities. DLCO was normal (≥80% predicted) in 61/62 symptomatic deployers for whom measurements were available. Mean DLCOPP measurements were not significantly different (p=0.09) between the three deployment lung disease diagnostic groups in unadjusted comparisons.

In unadjusted analyses, LCI score and Sacin*VT were significantly different among all groups, while Scond*VT was not (Table 1). Pairwise testing showed that those with both DDLD and DRA and those with DDLD alone had significantly higher mean LCI scores than controls. Those with DRA alone had higher (more abnormal) LCI scores than controls, but the sample size in this group was small and this difference was not statistically significant. Those with both DDLD and DRA had significantly higher Sacin*VT compared to controls, while those with DDLD or DRA alone did not significantly differ from controls. Scond*VT was not significantly different between controls and any of the deployment-related lung disease groups (Figure 2).

Figure 2.

Distribution of multiple breath washout parameters by diagnostic groups. A. LCI score was significantly higher in both the DDLD only and DDLD and DRA groups compared to controls in unadjusted analyses and the combined group remained significant in adjusted analyses. B. The group with DDLD and DRA had a significantly higher Sacin*VT compared to controls in unadjusted analyses. C. There was no significant difference in Scond*VT between the deployment lung disease groups and controls.

Dark gray shaded groups were significantly different (p<0.05) from the control group using Tukey pairwise comparisons if the overall ANOVA test was statistically significant.

To account for demographic differences between groups in univariate analyses, linear regression models were constructed for the MBW parameters to adjust for sex, smoking status, smoking pack-years and BMI. BMI was included as a continuous covariate in the models after exploratory analysis showed linearly increasing MBW with BMI categories for both controls and deployers. Mean differences, 95% confidence intervals, and Tukey-adjusted p-values comparing each of the diagnostic groups both before and after adjustment are presented in Table 2. We found that LCI score remained significantly different (p=0.005) between those with both DDLD and DRA compared to controls after adjustment, but the relationship diminished for those with DDLD alone. After adjustment, those with both DDLD and DRA no longer had significantly more acinar heterogeneity compared to controls, though Sacin*VT measurements remained higher. Full model parameter estimates are included in Appendix Table A1. As a sensitivity analysis (see Appendix Table A2), we matched controls and deployers (n=27) based on age, sex, race, smoking status, smoking pack-years and BMI and found similar results to adjusted analyses. Mean LCI scores were 0.36 units (95% CI: 0.03, 0.68) higher for symptomatic deployers compared to controls (p=0.03).

Table 2:

Estimated differences in MBW parameters between groups in unadjusted analyses and in analyses adjusted for smoking status, smoking pack-years, BMI and sex

	Unadjusted			Adjusted
Comparison	Difference in Means	95% Confidence Interval	p-value*	Difference in Means	95% Confidence Interval	p-value†
LCI			<0.0001			<0.0001
DDLD and DRA vs Control	1.35	0.64, 2.06	<0.0001	0.89	0.21, 1.56	0.005
DDLD only vs Control	0.84	0.34, 1.34	0.001	0.39	−0.11, 0.9	0.18
DRA only vs Control	0.58	−0.38, 1.54	0.40	0.56	−0.31, 1.42	0.34
DDLD and DRA vs DDLD only	0.51	−0.27, 1.28	0.33	0.49	−0.19, 1.18	0.24
DDLD and DRA vs DRA only	0.77	−0.36, 1.9	0.30	0.33	−0.69, 1.35	0.83
DDLD only vs DRA only	0.26	−0.75, 1.28	0.91	−0.16	−1.07, 0.74	0.97
Sacin*VT			0.01			0.06
DDLD and DRA vs Control	0.07	0.01, 0.13	0.008	0.06	−0.005, 0.12	0.08
DDLD only vs Control	0.01	−0.03, 0.05	0.84	−0.005	−0.05, 0.04	0.99
DRA only vs Control	0.03	−0.05, 0.11	0.71	0.02	−0.06, 0.1	0.89
DDLD and DRA vs DDLD only	0.06	0, 0.12	0.08	0.06	−0.001, 0.13	0.05
DDLD and DRA vs DRA only	0.04	−0.05, 0.13	0.68	0.04	−0.06, 0.13	0.75
DDLD only vs DRA only	−0.02	−0.1, 0.06	0.93	−0.03	−0.11, 0.06	0.84
Scond*VT			0.06			0.44
DDLD and DRA vs Control	0.009	−0.002, 0.02	0.14	0.01	−0.005, 0.02	0.44
DDLD only vs Control	0.006	−0.002, 0.01	0.17	0.003	−0.005, 0.01	0.71
DRA only vs Control	0.001	−0.01, 0.02	0.99	−0.0002	−0.015, 0.01	1.00
DDLD and DRA vs DDLD only	0.003	−0.009, 0.01	0.91	0.003	−0.008, 0.01	0.89
DDLD and DRA vs DRA only	0.008	−0.009, 0.03	0.64	0.01	−0.01, 0.02	0.73
DDLD only vs DRA only	0.005	−0.01, 0.02	0.85	0.004	−0.012, 0.02	0.92

^*P-values for overall ANOVA F-test between controls, DDLD &DRA, DDLD only and DRA only are included in the header with p-values from individual Tukey pairwise t-tests included in each row. Bolded values were significantly different (p<0.05).

^†P-values for the overall F-test between controls, DDLD &DRA, DDLD only and DRA only are included in the header with p-values from individual comparisons included in each row. Bolded values were significantly different (p<0.05).

Abbreviations: MBW = multiple breath washout, BMI = body mass index, LCI = lung clearance index, VT = tidal volume, DDLD = deployment-related distal lung disease, and DRA = deployment-related asthma.

Since many deployers had both DDLD and DRA, we were interested in exploring whether one or both diagnoses drove increases in MBW parameters noted in our analyses. First, we investigated whether effects of combined DDLD and DRA are additive or if the effect of DDLD is modified by DRA, resulting in a larger increase in MBW parameters than expected. We explored these effects by testing for an interaction between DDLD and DRA in linear regression models (see Appendix A). We found that the effect of DDLD did not significantly differ depending on the presence of DRA for any MBW parameters. Therefore, we removed interaction terms from our models. In adjusted analyses, controlling for concomitant DDLD diagnosis, those with DRA had significantly elevated LCI scores (p=0.01) and Sacin*VT (p=0.01), but not Scond*VT (see Appendix Table A3). Additionally, controlling for concomitant DRA diagnosis, those with DDLD had significantly elevated LCI scores (p=0.03) in adjusted analyses, while Sacin*VT and Scond*VT were not significantly elevated (see Appendix Table A3). This suggests that both DDLD and DRA contributed to increases in LCI scores observed in deployers with both DRA and DDLD; however, DRA was the stronger driver of elevated Sacin*VT levels in this group as demonstrated by the very small change in Sacin*VT based on DDLD diagnosis.

As the relationship between demographic factors and MBW parameters are not well-characterized in the literature, we examined these associations in adjusted linear regression models (see Appendix Table A1). We found significant relationships between BMI, LCI and Scond*VT, with both MBW measurements increasing with higher BMI (p<0.0001 and p=0.002 respectively). Sacin*VT was not significantly associated with BMI (p=0.8). None of the MBW parameters significantly differed between sexes, but LCI scores (p=0.8) and Sacin*VT (p=0.05) were generally higher in males compared to females, while Scond*VT (p=0.22) was generally lower in males.

We also examined smoking effects on MBW parameters. Overall, the numbers of both current smokers (7/174, 5%) and former smokers (36/174, 21%) were small. However, we did find differences in LCI between smoking groups (p=0.02), with former smokers having significantly higher (p=0.01) mean LCI scores than never smokers. Current smokers did not have significantly different mean LCI scores compared to never smokers (p=0.96) or former smokers (p=0.37). Additionally, we found significant increases in LCI score with longer smoking duration measured by pack-years (p<0.0001). Sacin*VT and Scond*VT did not significantly differ by smoking status (p=0.42 and p=0.13, respectively) and were not associated with pack-years (p=0.48 and p=0.56, respectively).

Because of concerns about small airways effects from cumulative deployment exposures, we examined whether total deployment duration had a measurable impact on any MBW parameters. The mean total deployment duration in our study population was 22 months (range 2 months to 10 years). Deployment duration was not significantly correlated with LCI score or Sacin*VT, nor was it significant for Scond*VT when two outliers were removed from analysis (see Appendix A).

Among symptomatic deployers, 37/67 (55%) reported exposure to explosive blasts and 48/66 (73%) to controlled detonations during deployment. Combining both variables, we calculated a Blast Exposure Intensity Score based on frequency of exposure to explosive blasts during each deployment (see Appendix A). We found that those with higher Blast Exposure Intensity Scores had significantly higher LCI scores (p=0.04) in unadjusted analyses (see Appendix Table A4). Adjusting for sex, age, smoking status, smoking pack-years and BMI, this association diminished (p=0.15) but remained higher for those with more frequent blast exposure.

DISCUSSION

This is the first study showing that, in a sizeable cohort of military deployers with persistent respiratory symptoms and normal spirometry, MBW measurements are helpful in distinguishing deployers with clinically confirmed distal lung disease and asthma from healthy controls. In unadjusted analyses, LCI was significantly higher for deployers with both DDLD and DRA and for those with distal lung disease alone (DDLD) compared to controls; these results remained significant for those with both diagnoses after controlling for differences in BMI, sex, smoking status and smoking pack-years. Notably, the majority (76%) of deployers who had higher, more abnormal LCI scores had normal spirometry, indicating that LCI is more sensitive than spirometry in the noninvasive detection of deployment-related lung disease. This study provides further evidence that LCI appears to be a useful noninvasive tool for recognition of airways disease in symptomatic military deployers with normal lung function.

We also found that unadjusted Sacin*VT, a marker of ventilation inhomogeneity in distal acinar airways, was significantly higher for deployers with both asthma and bronchiolitis compared to controls. This finding was no longer significant after controlling for differences in BMI, sex and smoking status; however, trends remained the same, with those deployers diagnosed with both DDLD and DRA having higher Sacin*VT levels than controls.

LCI is well-established in the pediatric literature as a sensitive tool to detect early distal lung abnormalities due to cystic fibrosis, and correlates better than spirometry with radiographic evidence of air trapping (19, 38). In pediatric patients with obliterative bronchiolitis (BO), elevated LCI correlated significantly with decreased FEV1 and FEF_25–75 (20, 39). However, this small cross-sectional study was unable to assess the utility of LCI in children with less severe disease and normal spirometry. There are fewer reports of MBW in adults with clinically confirmed small airways disease. Investigators examined the utility of LCI and Sacin in distinguishing bronchiolitis obliterans syndrome (BOS), with or without chronic grant-vs-host disease, following allogeneic hematopoietic stem cell transplantation (40). They found that these MBW parameters were the most sensitive markers of biopsy-proven BOS, while FEV1/FVC ratio was often normal. These conclusions support our findings in deployers that LCI may be most useful in detecting and monitoring early stages of small airways disease (41).

Studies of severe asthmatics suggest that the acinar compartment of ventilation is particularly important in contributing to asthma severity and peripheral lung inflammation, whereas Scond is likely associated with airway remodeling (15, 42). In unadjusted analyses, our study showed that Sacin*VT was significantly higher in symptomatic deployers with both asthma and distal lung disease, and that those with deployment-related asthma alone also had higher Sacin*VT than those with distal lung disease alone. Since deployers with asthma were diagnosed in adulthood following military deployment (and within the past 10 years, before long-standing asthma may have led to substantial airways remodeling), our findings are consistent with the understanding of Sacin*VT as a probable marker of peripheral airways inflammation with ventilation inhomogeneity arising in close proximity to the diffusion-convection front. Longitudinal Sacin*VT measurement in asthmatic deployers may be a helpful measure of asthma control (14).

Several investigators have shown that Sacin is increased in obesity and is unrelated to measures of lung volume (43). Although our study did not find a trend with increasing BMI and Sacin, we noted an association between Scond*VT and increasing BMI in both healthy controls and symptomatic deployers. These observations underscore the need for standardized MBW reference values that account for differences in BMI.

Published normative data for MBW parameters among healthy adults is limited (33). Our study examined 103 healthy adults, providing sufficient power to detect differences between those with and without lung disease. Moreover, healthy controls were tested in the same time period as study subjects, minimizing potential confounding from testing variability across time. We found associations between BMI, smoking, sex and LCI scores, emphasizing the importance of adjusting for these variables in future MBW studies (44). LCI remained significantly abnormal in deployers with distal lung disease even after adjusting for these differences. This is noteworthy, as military deployers must pass rigorous physical fitness tests and may be less likely than the general population to have abnormal LCI prior to deployment.

To our knowledge, this is the first study to explore combined host and exposure-related determinants of all MBW parameters including ventilation inhomogeneity in conducting and acinar airways. In unadjusted analyses, we found that frequency of exposure to explosive blasts may result in damage to distal airways as indicated by higher LCI scores. Future investigations should consider effects of blasts on small airways integrity.

We found that MBW is more sensitive than DLCO in detecting deployment-related small airways disease, particularly in those who also had asthma. Of the deployers with combined asthma and distal lung disease, 70% had abnormal LCI scores while only one had decreased DLCO. The limited utility of DLCO in this clinical setting may be due to the “pseudo-normalizing” effect of asthma on DLCO, attributed to greater lung capillary surface area and higher blood volume (45). In those with combined DDLD and DRA, the impaired diffusion capacity sometimes associated with small airways disease may be offset by the supra-normal DLCO seen with asthma. As described in patients with COPD, our study found effects on acinar airways among deployers with normal diffusion capacity (16). However, since nearly all of the symptomatic deployers had normal DLCO, we were unable to determine whether Sacin*VT increased as DLCO decreased.

A 2015 American Thoracic Society (ATS) Workshop Report examined the readiness of LCI for use in multicenter clinical trials and clinical care of patients with cystic fibrosis (46). The expert panel concluded that gaps in practice guidelines for LCI interpretation, the need for more standardized operating procedures, and limited normative data were problematic, but that recommendations for MBW use in clinical settings were likely in the near future. In 2019, Driskel et al (47) found that MBW was both feasible and reproducible in detecting early BOS in adults following lung transplant. Our study underscores the feasibility of LCI as a sensitive, noninvasive tool in an adult outpatient setting that requires less than 60 minutes to perform.

Our study has a number of strengths. First, it includes a sizeable number of deployers with persistent and disabling respiratory symptoms despite normal resting lung function. Our sample size for healthy controls was also substantial, and we were able to adjust for multiple potential confounders including age, sex, BMI, and both smoking status and pack-years. Diagnostic criteria for asthma and distal lung disease in the deployer cohort were well-defined and based on the most sensitive clinical tools available. Our study design enabled reliable comparisons of spirometry to LCI, Sacin*VT and Scond*VT as testing was performed on the same day, assuring that data were comparable.

Potential limitations of our study include self-reported smoking status, deployment duration and exposure to explosive blasts. However, the significant associations between blast exposure and abnormal LCI as well as with smoking suggest that self-reporting of both exposure variables was reliable. A minority of study subjects performed only two rather than three MBW maneuvers. However, duplicate measurements have been shown to be sufficient in adult patients with pulmonary disease and in healthy controls (48). Additionally, we did not explore the role of allergic rhinitis/rhinosinusitis as a possible confounder in risk for small airways dysfunction in either deployers or controls (49, 50). Lastly, the small number of deployers with DRA alone limited our power to detect differences between this group and others.

In summary, this is the first study to clearly demonstrate the utility of LCI as a tool for identifying exposure-related airways disease in symptomatic military personnel following deployment. Our findings suggest that MBW measurements may be valuable supplemental noninvasive tools for clinical diagnosis and decision-making in the at-risk population of military deployers. This is particularly promising given costs and potential risks associated with surgical lung biopsy for diagnosis of the patchy, heterogeneous abnormalities typical for bronchiolitis. Our study is also the first to link exposure to explosive blasts to risk for small airways abnormalities using MBW.

Future directions will focus on monitoring longitudinal changes in LCI to track lung disease progression, reversibility or stability in symptomatic deployers with asthma, bronchiolitis or both. Additionally, longitudinal MBW measurements may be useful to assess responses to treatment, particularly in symptomatic deployers using inhaled bronchodilators and steroids. Such testing may also be helpful in showing early small airway improvements in those who stop smoking and in monitoring damage related to persistent smoking or from other inhalational insults.

Highlights.

• Multiple breath washout useful to detect airways disease in symptomatic deployers

• Ventilation inhomogeneity in acinar airways suggests distal lung injury

• Small airways effects from military blast-related exposures implicated

• Normative data on healthy adult multiple breath washout in heterogeneous group