A prospective study was performed including 35 Caucasian patients (32 women) diagnosed with SSc. All patients fulfilled the ACR/EULAR 2013 classification criteria,14 had undergone a baseline HRCT, and were followed-up for 4 years. Patients were consecutively assessed in the Scleroderma Unit at our hospital and followed up by the same specialist every 4 months in moderate-severe ILD patients, or every 6 months in patients with no pulmonary involvement or mild ILD, following international guidelines.15 The study was approved by the hospital's Ethics Committee for Clinical Research, and all the patients provided informed consent for their participation, according to the Declaration of Helsinki. The exclusion criteria were the presence of other inflammatory respiratory diseases (such as asthma, COPD or others), atopic features at inclusion, or an infection of the respiratory tract during the previous month.

The study population was divided into 2 groups: the ILD group included 12 patients with ILD defined as radiological evidence of interstitial disease on HRCT, containing 1 patient with pulmonary hypertension associated to ILD (PH-ILD); and 23 SSc patients with no evidence of ILD were selected as the control group, including 2 patients with isolated pulmonary arterial hypertension (PAH). PH was defined as mean pulmonary arterial pressure (mPAP) ≥25mmHg in right heart catheterization (RHC). PH-ILD was considered if PH was diagnosed in a patient with ILD and lower than 60% predicted forced vital capacity (FVC) or moderate-severe interstitial disease in the HRCT.16 Isolated PAH was diagnosed by RHC according to international guidelines.16,17 Other SSc comorbidities were also recorded as previously described.18 All patients underwent complete PFTs, fraction of exhaled NO (FeNO), exhaled CO (eCO) and EBC collection at baseline. Subjects were instructed to refrain from food intake and smoking during the 2 and 24 h prior to sample collection, respectively. All samples were analyzed in the same laboratory.

Spirometry, static lung volumes and transfer lung studies were performed annually using MasterLab equipment (MasterLab, Jaegger, Germany) following the ERS-ATS recommendations.19 The PFTs outcomes were analyzed as the change in % predicted FVC or diffusing capacity for carbon monoxide (DLCO) compared to baseline.

FeNO measurements were performed using NIOX MINO® device (Aerocrine AB, Solna, Sweden) according to current guidelines.20 Patients sat in an upright position and wore a nose clip during the process. Patients inhaled NO-free air through the device and subsequently exhaled at a fixed flow rate of 50mL/s for approximately 10s. The first technically acceptable FeNO measurement was used for analysis.21

Exhaled CO was measured by a MicroCO Meter (MicroCO – Micro Medical Ltd, Rochester, Kent, UK), using an electrochemical sensor. Patients were instructed to breath deeply to total lung capacity, to hold the breath for 20s, and then to exhale slowly and completely through a mouthpiece. Two successive recordings were made and maximal values were used in all calculations.

EBC was collected during tidal breathing with a commercially available condenser (EcoScreen; Jaeger, Würzburg, Germany). Subjects breathed tidally through a mouthpiece connected to the condenser while wearing nose clips. A fixed volume of 150L of exhaled breath was collected per subject.22 EBC samples were processed in sampling tubes and divided into 500μL aliquots. One aliquot was used to measure the pH after deaeration, and the remaining were immediately stored at −70°C, and analyzed within 1 month of collection.

EBC pH was measured after deaeration with helium (350mL/min for 10min), using a calibrated pH meter (Model GLP 21; Crison Instruments SA; Barcelona, Spain) with an accuracy of ±0.01 pH, and a probe for small volumes (Crison 50 28; Crison Instruments SA; Barcelona, Spain). The probe was calibrated daily with standard pH 7.02 and 4.00 buffers.23

EBC nitrite (NO2−) and nitrate (NO3−) levels were determined with a colorimetric assay based on the Griess reaction (Cayman Chemical Company, USA). The concentrations were measured at 540nm absorbance with a microplate reader. Assay sensitivity was 1μM for nitrite and 2.5μM for nitrate.

IL-6 levels in EBC samples were evaluated by a commercially available high sensitivity enzyme-linked immunosorbent assay (ELISA) (Bender MedSystems GmbH, Austria). Assay sensitivity was 0.03pg/mL.

Non-parametric tests were performed to assess the presence of statistically significant differences. Categorical data were analyzed by chi-square and Fisher's exact tests. Continuous data were analyzed using the Mann–Whitney U test in the between-groups differences and Wilcoxon signed rank test in within-group comparisons. Associations between variables were analyzed using Spearman's correlation. Receiver Operating Characteristic (ROC) curves and Youden's index method were used to determine the threshold value of each biomarker, which demonstrated the best sensitivity and specificity to predict a decrease of 10% in FVC, a 15% in DLCO or death. Log Rank test was used to measure the progression-free survival (PFS) in the follow-up, whose endpoint was the occurrence of combined damaging events, defined as a 10% decrease in FVC, a 15% decrease in DLCO from baseline, or death. A P value <0.05 was considered significant. SPSS version 20.0 for Windows (SPSS Inc., Chicago) and GraphPad InStat4 (GraphPad Software Inc., San Diego) were used to carry out the statistical analyses.

Demographic data and baseline characteristics of the study population are shown in Table 1. Median (interquartile range, IQR) age was 59.0 (42.0 to 68.0) years. HRCT was consistent with non-specific interstitial pneumonia (NSIP) in all ILD patients, whereas no ILD findings were observed in controls. The median of pulmonary disease duration was 1.4 years in the ILD group. At inclusion, 7 patients were taking immunosuppressant therapy.

Baseline imaging, PFTs and biomarkers are described in Table 2. Significant differences at baseline were found between the study groups in the FVC, TLC, DLCO and KCO, with lower values in the ILD group. EB and EBC biomarkers were analyzed in both groups of patients at baseline. No statistical differences were found in FeNO or eCO levels. However, FeNO was positively associated with baseline FVC (r=0.41, P=0.03). In the ILD group, a negative correlation was found between eCO levels and baseline FVC (r=−0.75, P<0.01). No significant differences in EBC biomarkers were found between the 2 study groups. However, analyzing 11 patients with ground-glass opacities in the HRCT, the median of EBC pH was significantly lower than in patients without this radiographic pattern (7.4 vs. 8.0, P=0.02). No other association was observed in terms of HRCT findings with baseline biomarkers.

A substudy excluding the 7 patients taking immunosuppressive therapies found the same differences in baseline PFTs and EB and EBC biomarkers.

A patient from the control group with baseline EBC pH of 8.09 and FeNO of 14.5ppb developed mild ILD diagnosed by HRCT with FVC higher than 70% at 45 months of follow-up, and thereafter presented PH-ILD identified by RHC at 47 months of follow-up. Four patients died during follow-up, 2 from the ILD group and 2 controls: one due to PAH and due to a stroke.

Figs. 1 and 2 show FVC and DLCO values, respectively, in the 2 study groups during the follow-up period.

Fig. 1.

FVC values at baseline and during follow-up. A and B show the unadjusted variation of % predicted FVC during follow up (each point represents a patient) for ILD patients and controls, respectively. In the ILD group, we observed a decrease of 6.4% in the mean unadjusted FVC % predicted values during the 4-year period compared to baseline (95% CI: −11.2 to −1.6; P=0.01) (A), whereas no significant differences were found in the control group (B). A higher decline of FVC % was found in ILD group than controls (P<0.01).

(0.15MB).

Fig. 2.

DLCO values at baseline and during follow-up. A and B show the unadjusted variation of % predicted DLCO during follow up (each point represents a patient) for ILD patients and controls, respectively. DLCO values were significantly lower at 4 years of follow-up compared to baseline in the ILD group, with a mean decrease of 11.1% (95% CI: −18.0 to −4.3; P<0.01) (A), and in controls, with a mean decline of 5.5% (95% CI: −10.0 to −1.1; P=0.01) (B).

(0.13MB).

Correlations between EBC biomarkers and PFTs are shown in Fig. 3. In all patients, low pH values were correlated with lower DLCO at 4 years of follow-up (Fig. 3A). Furthermore, low FeNO levels were associated with lower FVC at 4 years of follow-up (Fig. 3B). Low baseline FeNO levels were related to a decrease of FVC and DLCO at 4 years of follow-up (Fig. 3C and D). In ILD patients, a negative correlation was found between high eCO levels and lower FVC at the end of follow-up (r=−0.66; P=0.03). In addition, we observed a trend towards correlation between high IL-6 levels and lower DLCO at end of study in SSc-ILD patients (r=−0.53; P=0.10). There were no statistical differences in the associations of the remaining EBC biomarkers.

Fig. 3.

Correlation between EBC pH and FeNO with pulmonary function tests. The figure shows correlations including all patients. Variations of FVC and DLCO % were calculated in respect of baseline parameters (PFTs at 4 years minus baseline PFTs). FeNO, Fraction of exhaled NO.

(0.22MB).

ROC curves were used to quantify the accuracy of each baseline biomarker to predict functional decline or death. The area under the curve (AUC) for EBC pH was 0.65 (95% CI: 0.41 to 0.89; P=.23), and a threshold value of 7.88 was established using Youden's index method, with a sensitivity of 0.71 and a specificity of 0.58. AUC for FeNO was 0.81 (95% CI: 0.65 to 0.96; P=0.01), and showed a threshold of 10.75ppb, with a sensitivity and specificity of 0.85 and 0.77, respectively. Kaplan–Meier curves illustrate the PFS according to the former cut-off values of pH and FeNO (Fig. 4). EBC pH lower than 7.88 showed a decrease in lung function parameters or death during follow-up (log Rank P=0.03). FeNO levels lower than 10.75ppb were related to worse PFS (log Rank P<0.01). There were no statistical differences in survival related to baseline eCO levels.

Fig. 4.

Kaplan–Meier survival plots representing progression-free survival since EBC collection. The end point was defined as a decrease from baseline FVC of at least 10%, or a 15% decrease in DLCO, or death during follow-up. (A) The survival plot found a better progression-free survival in patients with EBC pH values equal to or higher than 7.88. (B) The survival plot showed a higher progression-free survival in patients with FeNO levels equal to or greater than 10.75ppb.

(0.13MB).

In the substudy excluding patients taking immunosuppressive therapies, the AUC for EBC pH was 0.66 (95% CI: 0.44 to 0.88; P=0.18), with the same threshold value of 7.88, although there were no statistical differences in PFS (log Rank P=0.09). Regarding FeNO assays, the AUC was 0.88 (95% CI: 0.74 to 1.0; P=0.18), with a threshold of 10.25ppb, with a sensitivity and specificity of 0.80 and 0.84, respectively. Furthermore, worse PFS was identified in patients with FeNO levels under 10.25ppb (log Rank P<0.01).