*2.1. Subjects' Characteristics*

A total cohort of 50 subjects was enrolled. At HRCT scan, two patients showed a "probable UIP" pattern and were consequently excluded from the study. The remaining 48 subjects, 37 female (77.1%) and 11 males (22.9%) were included in the study. The three study groups were composed by a total of 21, 10 and 17 ACPA-positive subjects without evidence of arthritis (ND), early RA (ERA) and long-standing RA (LSRA), respectively. Patients did not suffer from other significant comorbidities. The demographic, clinical and laboratory characteristics are summarized in Table 1. Within the LSRA group, eight patients (47.1%) were already on treatment with methotrexate (MTX), four patients (23.5%) were taking sulphasalazine and three patients (17.6%) hydroxychloroquine, either as monotherapy or as combination therapy. For the laboratory evaluation, 22 healthy control subjects (HC) were also enrolled, 15 females (68.2%) and 7 males (31.8%), the mean age being 47.3 ± 10.5 years.


**Table 1.** Demographic, clinical features and autoantibodies of enrolled subjects.

Data are reported as mean ± SD, unless stated otherwise. *p* values intended for comparisons between ND, ERA, LSRA and HC (whenever applicable) subgroups of participants. †: post hoc test *p* < 0.05 vs. ERA. Abbreviations: ND, no disease subjects; ERA, early rheumatoid arthritis patients; LSRA, long standing rheumatoid arthritis patients; HC, healthy controls; BMI, body mass index; ACPA, anti-citrullinated proteins antibodies; RF, rheumatoid factor.

#### *2.2. Lung Function and Physiological Responses to Exercise*

The results of the main PFT and CPET parameters are shown in Table 2.


**Table 2.** Selected pulmonary functional responses measured at rest and during exercise testing.

Data are reported as mean ± SD. *P* values intended for comparisons between ND, ERA and LSRA subgroups of participants. \*: post hoc test *p* < 0.05 vs. ND; Reduced DLCO: DLCO < 80% of predicted value; reduced exercise tolerance: V'O2 peak < 80% of predicted value; impaired ventilatory efficiency: V'E/V'CO2 a θ<sup>L</sup> > 34 and/or V'E/V'CO2 slope > 30. Abbreviations: ND, no disease subjects; ERA, early rheumatoid arthritis patients; LSRA, long standing rheumatoid arthritis patients; FEV1, forced expiratory volume; FVC, forced vital capacity; TLC, total lung capacity; DLCO, diffusing lung capacity for carbon monoxide; KCO, transfer coefficient of the lung; V'O2, oxygen uptake; V'CO2, carbon dioxide output; V'E, minute ventilation; θL, lactate threshold; eMVV, estimated maximal voluntary ventilation; SpO2, peripheral capillary oxygen saturation; ΔSpO2, peak-rest change in peripheral capillary oxygen saturation.

Reduced DLCO (i.e., <80% predicted value) was observed in 57.1%, 50% and 64.7% of ND, ERA and LSRA subjects, respectively (*p* = 0.747). There were no significant differences between groups in the examined PFT parameters, even after correction for smoking status. FEV1, when expressed as percentage of predicted value, was significantly lower in patients who were current or former smokers compared to non-smokers (*p* = 0.018). DLCO values in current and former smokers were not different from those of never smoker subjects.

All patients achieved maximal effort during CPET. Mean work rate at peak exercise was 105 ± 32.5 W (76.1 ± 15.4% predicted value). Reduced exercise tolerance, defined as V'O2 peak < 80% predicted value, was found in 19 out of 48 (39.6%) patients. Even if not significantly different, ND and ERA subjects showed lower rates of exercise intolerance than LSRA. Anticipated θ<sup>L</sup> (i.e., V'O2 at θ<sup>L</sup> < 40% predicted V'O2 peak), indicating impaired aerobic fitness, was observed in six (12.5%) subjects, with no difference among subgroups. No abnormalities in cardiovascular responses were found, neither within the whole group nor among each subgroup.

Overall, there was no sign of ventilatory limitation, average V'E peak being 49.8 ± 12% eMVV. In this regard, none of the recruited patients showed V'E peak values > 85% eMVV. Mean VT was 1.8 ± 0.5 L at peak exercise, corresponding to 50.0 ± 10.0% of FVC, confirming the absence of ventilatory constraints during exercise. Apparently, no differences in breathing patterns were observed between subgroups.

Peak SpO2 fell within normal values for all groups, although a significant trend in progressive reduction was found between groups (97.7 ± 1.1% vs 97.6 ± 1.2% vs 96.8 ± 1.7% among ND, ERA and LSRA patients, respectively; *p* = 0.008). Significant haemoglobin desaturation (peak—rest change in SpO2 > 4%) was found only in one patient with LSRA.

Impaired ventilatory efficiency, defined as V'E/V'CO2 at θ<sup>L</sup> > 34 and/or V'E/V'CO2 slope > 30, was found in about one third of the study group. The highest frequency was registered among LSRA patients (41.2%); however, no significant differences were observed in the rate of reduced ventilatory efficiency among the three subgroups.

#### *2.3. Laboratory Results*

There were no significant differences in ACPA and RF levels between the groups. Significantly higher values of RF were registered among subjects from all the three groups with reduced exercise tolerance (263.8 ± 216.5 IU/L vs 119.4 ± 171.71 IU/ L; *p* = 0.019) and with impaired ventilatory efficiency (326.6 ± 282.9 IU/ L vs 90.8 ± 98.6 IU/ L; *p* = 0.015). No further relations were observed between ACPA and RF levels and pulmonary function at rest and during exercise.

SPD serum levels were significantly higher among study group subjects compared with HC (158.5 ± 132.3 ng/mL vs. 61.27 ± 34.11 ng/mL; *p* < 0.0001). Similarly, subgroup analysis revealed higher SPD levels within the ND (132.1 ± 125.1 ng/mL; *p* = 0.023), ERA (181.1 ± 151.3 ng/mL; *p* = 0.003) and LSRA subgroups (176.1 ± 131.8 ng/mL; *p* < 0.0001) compared with HC (61.2 ± 34.1 ng/mL) (Figure 1). No difference was observed between patient groups; nonetheless, a significant trend in increasing levels of SPD was noticed from ND to LSRA (*p* = 0.004). There were no differences in SPD levels based on smoking history (i.e., current, former or never smoker). However, significantly higher levels of SPD were present in ACPA-positive never smokers (ND+ERA+LSRA) compared with healthy controls (142.2 ± 94.1 ng/mL vs 61.2 ± 34.1 ng/mL; *p* < 0.0001). In the LSRA group, no difference was found in SPD serum levels concerning treatment with MTX.

**Figure 1.** Comparison of SPD serum levels between ACPA-positive subjects and healthy controls. \*: *p* < 0.05; \*\*: *p* < 0.005; \*\*\*: *p* < 0.0005. Abbreviations: SPD, surfactant protein D; HC, healthy controls; ND, no disease subjects; ERA, early rheumatoid arthritis patients; LSRA, long standing rheumatoid arthritis patients.

SPD levels were inversely related to V'O2 peak (expressed as percentage of predicted value; *p* = 0.024; rho = −0.32), V'O2 at θ<sup>L</sup> (percentage of predicted V'O2 peak; *p* = 0.013; rho = −0.36) and peak SpO2 (*p* = 0.008; rho = −0.38). As observed for RF, SPD levels were also shown to be significantly higher among patients with reduced exercise ventilatory efficiency (V'E/V'CO2 a θ<sup>L</sup> > 34) (250.8 ± 155.3 ng/mL vs. 141.2 ± 121.9 ng/mL; *p* = 0.007). A similar result was also observed when limiting the analysis to ND patients (351.15 ± 219 ng/mL vs 93.5 ± 44.9 ng/mL; *p* = 0.01).
