Unmet needs in RA-ILD
ILD, a severe fibrotic disease of the lung parenchyma, is one of the most common causes of death in RA patients [
19]. However, our knowledge regarding the precise prevalence, pathogenesis, and natural history of RA-ILD is poorly understood. Heterogeneous disease course and the lack of optimal screening tools and guidelines make early diagnosis and proper intervention challenging. Early RA-ILD patients are frequently asymptomatic [
20,
21]. When cough and dyspnea appear (associated with ILD and not due to other causes), the disease may already be advanced and is associated with a poor prognosis [
11]. In addition, the current mainstay treatment regimens of RA such as methotrexate, leflunomide, and anti-tumor necrosis factor-α (TNF-α) agents might play a possible role in lung injury and exacerbation of existing ILD [
22‐
24]. However, the use of methotrexate in RA-ILD is controversial and has recently also been associated with increased survival and a decreased risk of ILD in RA [
25,
26]. Finally, to our knowledge, there are few ongoing and completed randomized, double-blind, placebo-controlled therapeutic trials including patients with RA-ILD [
27,
28], and no treatment recommendations exist. Treatment approaches in clinical practice for RA-ILD are largely based on data derived from SSc-ILD or IIM-ILD, individual physician’s experience, and published case reports and series [
21,
29].
Genetic, environmental, and demographic risk factors for RA-ILD
In the past decade, studies found that genetic variants have been implicated in the development of RA-ILD. A gain-of-function promoter variant (rs35705950) in the mucin 5B (MUC5B) gene was associated with RA-ILD, more specifically associated with evidence of UIP [
30]. HLA-B54, HLA-DQ1B*0601, HLA-B40, and the site encoding α-1 protease inhibitor are associated with an increased risk of RA-related ILD [
31]. An excess of mutations was observed in telomere maintenance genes (TERT, PARN, RTEL1) and in SFTPC, involved in surfactant homeostasis, with increased odds ratios (OR) of 3.17 (95% CI 1.53–6.12;
p = 9.45 × 10
−4) for ILD as compared with controls [
32].
Published data, mostly from retrospective studies, has identified several environmental and demographic risk factors that predict RA-ILD development, including male sex, older age, older age at RA diagnosis, tobacco smoking, high disease activity, seropositivity and titer of rheumatoid factor (RF) and/or anti-cyclic citrullinated peptide (CCP) antibodies, and the presence of other non-pulmonary extra-articular manifestations [
20,
33,
34]. Other RA-ILD-associated antibodies have also been described to be associated with ILD, including anti-carbamylated proteins and other post-translational modified proteins [
35].
Because respiratory symptoms occur often late and are frequently unspecific, it is not advised to screen patients for ILD based on symptomatology. Usually, a chest X-ray is the initial examination performed to evaluate suspected lung involvement. It is frequently performed in RA patients prior to initiating methotrexate and biologic agent treatment. X-ray has a good specificity for ILD diagnosis, but a very low sensitivity limiting its role for ILD screening purposes [
36,
37].
HRCT is the gold standard for ILD diagnosis and evaluation of disease severity of ILD. HRCT can identify even subtle ILD changes. Serial HRCT scans can be performed to monitor existing diseases. However, radiation exposure and high cost are two limiting factors in the use of HRCT [
38,
39], especially for screening purposes in younger patients and for monitoring over time. Recently, to reduce radiation dose, a newly proposed 9-slice HRCT protocol showed good accuracy and sensitivity of 93% and 88%, respectively, compared with the standard whole-chest HRCT (64-slice or 128-slice) in 205 systemic sclerosis (SSc) patients [
40]. However, in less developed countries, the availability of low radiation protocols and HRCT facilities may be limited, and the high cost may restrict its use for screening and monitoring purposes. Also, as RA is a very prevalent disease, it may not be appropriate to screen all RA patients for ILD with HRCT.
Serial PFTs have been used to monitor patients with CTD-ILD and are frequently used in clinical practice as well as in clinical trials [
41,
42]. Forced vital capacity (FVC) and diffusing capacity for carbon monoxide (DLco) are important pulmonary function parameters for assessing lung physiology, including in RA-ILD [
43]. The predicted FVC% and DLco% value could help guide management strategies and predicts mortality. However, its role in screening for early asymptomatic ILD is controversial. Recently, a study by Suliman et al. demonstrated a high risk of missing significant SSc-related ILD when relying solely on PFTs [
44]. Among 102 SSc patients, 64 (63%) showed significant ILD on HRCT, while only 27 (26%) had an FVC < 80% predicted, and 54 (53%) had a decrease in the results of at least 1 PFT. Forty (62.5%) of 64 patients with significant ILD on HRCT had normal FVC, translating into a high false-negative rate [
44]. These results were confirmed by Hoffmann-Vold et al. in a prospective cohort study including 305 SSc patients [
45]. There exists no data in this regard for RA-ILD, but it is highly probable that comparable results hold true for RA. Therefore, more sensitive and repeatable methods for screening RA patients for ILD are highly on demand.
LUS for CTD-ILD
In the past two decades, LUS has evolved as a promising tool in the assessment of pulmonary parenchymal diseases [
46]. The inherent characteristics of ultrasound, including that it is non-ionizing, non-invasive, at low cost, repeatable, and easily accessible, make LUS a possible initial screening tool. LUS has been proposed to assess the extent and severity of ILD by detecting and quantifying the number of lung comet tail signs (B-lines) that originate from thickened septa [
47]. B-lines are defined as discrete laser-like vertical hyperechoic reverberation artifacts that arise from the pleura, extend to the bottom of the screen without fading, and move synchronously with respiration [
48]. B-lines are visible when the lung parenchymal air content is partially decreased and/or the interstitial space is volumetrically expanded, such as in pulmonary edema and/or ILD [
49‐
51]. Despite LUS with the assessment of B-lines is appealingly simple to use, to learn, and to teach, sufficient theoretical and practical skills and training are prerequisites [
52]. International evidence-based consensus recommendations for point-of-care lung ultrasound from Volpicelli et al. are helpful in guiding the implementation, development, and standardization of LUS across a variety of clinical settings [
48].
Currently, different LUS scoring systems (total lung scanning sites range from 10 to 72) to assess and quantify the severity of CTD-ILD have been developed [
53]. More scanning sites will undoubtedly be more accurate; however, this is at the cost of requiring more time to do the scanning. Although the full validation of LUS in CTD-ILD has not been completed, the data on early screening and diagnosis of ILD are encouraging [
14,
54].
Recently, LUS and HRCT were used to screen early ILD change in 64 asymptomatic RA patients. LUS examination revealed that 18 patients (28%) had sonographic ILD (Table
1). In 16 (89%) LUS-positive patients, HRCT scans confirmed the ILD diagnosis [
55]. More data are derived from ILD in SSc patients. Barskova et al. first used LUS for the screening of ILD in very early SSc patients. The concordance rate between B-lines and HRCT for the assessment of ILD was 0.83, and B-line numbers were significantly different in patients with and without HRCT-ILD (57 ± 53 vs 9 ± 9,
p < 0.05) and with and without ground glass opacity (GGO) on HRCT (63 ± 47 vs 33 ± 40,
p < 0.05). These data suggest that LUS has high sensitivity to identify ILD in SSc [
56]. A prospective study in 133 SSc patients showed that LUS findings correlated with HRCT (
p = 0.001) and these preliminary data revealed high sensitivity and specificity of LUS to detect ILD [
57]. Çakir et al. investigated the ability of LUS to assess ILD severity in 48 SSc patients, showing a good correlation between B-lines, HRCT (
r = 0.89,
p = 0.0001), and the Medsger disease scale (
r = 0.55,
p = 0.0001), and a negative correlation with DLco (
r = −0.56,
p = 0.0001) and FVC (
r = −0.46,
p = 0.001). The diagnostic accuracy of LUS was comparable to HRCT [
58]. In another study, LUS was performed in 104 patients undergoing HRCT for suspected ILD. According to HRCT, ILD was present in 50 patients. The false-negative and false-positive numbers of LUS were 4 (8%) and 11 (22%), respectively, compared to HRCT as the gold standard. The study concluded that LUS could be a sensitive tool for ILD detection, although the data point to the need for follow-up when LUS is abnormal [
37]. In addition, LUS could detect alveolar-interstitial involvement (an early sign of ILD) in 31 patients with CTD [
59]. These promising data support LUS as a screening tool for the diagnosis of early ILD in RA (Table
1).
Table 1
Overview of included studies
Moazedi-Fuerst et al. [ 55] | 64 RA patients | To screen subclinical RA-ILD | HRCT | N/A | N/A | 97.1 | 97.3 | 98.6 | 94.3 | N/A |
| 58 SSc patients, including 32 VEDOSS | To screen early SSc-ILD | HRCT | 100% | > 5 ≥20 | 100 83 | 55 96 | 100 N/A | 78 N/A | 94 N/A |
| 133 SSc patients | To detect and predict asymptomatic SSc-ILD | HRCT | N/A | N/A | 91.2% | 88.6% | N/A | N/A | N/A |
| 48 SSc patients | To evaluate the severity of SSc-ILD | HRCT | N/A | ≥6 > 24 | 100 79.3 | 84.2 94.7 | 100 N/A | 90.6 N/A | 93.7 94.8 |
| 104 suspected ILD patients | To evaluate the accuracy of LUS detection of ILD | HRCT | 100% | > 5 > 10 | 92 92 | 53 66 | 87 90 | 64 71 | 90 N/A |
| 31 suspected rheumatoid lung involvement patients | To investigate the utility of LUS | HRCT | N/A | > 5 | 73.5 | 88.2 | 51.7 | 95.1 | N/A |
Recently, Tardella et al. explored the optimal cutoff values of numbers of B-line to predict the presence of significant ILD in 40 SSc patients [
60]. An excellent correlation between the LUS B-line number and HRCT Warrick score was confirmed (Spearman rho 0.958,
p = 0.0001). The receiver operating characteristic curve analysis revealed that the presence of 10 B-lines is the cutoff point with the greatest positive likelihood rate (12.52) for the presence of significant SSc-ILD (Warrick score ≥ 7). The value represents the best compromise between the best sensitivity (96.3%) and specificity (92.31%) by HRCT [
60].
Gargani et al. preliminarily assessed the prognostic value of LUS B-lines to predict new development or worsening of pulmonary involvement in a total of 396 consecutive patients with SSc enrolled from three rheumatology departments. In the multi-variable analysis, the number of posterior B-lines ≥5 was associated with new development or worsening of ILD (hazard ratio, 3.378; 95% CI, 1.137–9.994;
p = 0.028), performed better than anti-topoisomerase I antibody positivity [
61].
These promising findings support LUS both as a screening and following tool for ILD in SSc and, hopefully, also other CTDs including RA.
KL-6 for CTD-ILD
A biomarker may be defined as “any substance, structure, or process that can be measured in the body or its products and influences or predicts the incidence of outcome or disease” [
62]. There is growing evidence that some circulating serological and alveolar biomarkers could reflect pathological processes, from early alveolar epithelial cell damage to advanced fibrosis. Among them, KL-6 has been extensively studied and it has emerged as a potentially sensitive surrogate marker of the presence of CTD-ILD and its severity [
63‐
65]. KL-6 antigen is a mucin-like, high molecular weight glycoprotein expressed on the surface membrane of alveolar epithelial cells and bronchiolar epithelial cells, which increases following cellular injury and/or regeneration [
66]. Its pathogenic role in pulmonary fibrosis is suggested by its pro-fibrotic, anti-apoptotic effects on fibroblasts [
67].
Multiple studies indicated that serum and bronchoalveolar lavage fluid KL-6 levels were significantly correlated with HRCT findings and PFT variables in CTD-ILD [
68,
69]. KL-6 concentrations were significantly higher in patients with ILD compared to those without ILD and showing a correlation to the ILD course [
70,
71]. Furthermore, KL-6 had predictive value for the development and progression of ILD [
72‐
75]. High KL-6 levels (≥ 640 U/mL) were independently associated with a UIP pattern (OR, 5.173;
p = 0.05) in RA-ILD [
76]. KL-6 levels could reflect early pulmonary epithelial cell injury and increased alveolar-capillary permeability [
77]. Serum KL-6 levels were also associated with alveolitis in 66 SSc patients. The receiver operating characteristic curve analysis to evaluate the accuracy of KL-6 for the diagnosis of active alveolitis showed that 500 U/mL was the best cutoff value with a sensitivity of 78.8% and specificity of 90% (area under the curve (AUC) = 0.90) [
78].
In a retrospective study, comprised of 60 confirmed CTD-ILD patients, including 11 patients with RA-ILD, circulating KL-6 levels correlated positively with the LUS B-line score (
r = 0.54,
p < 0.0001) [
17]. The significant relationship was confirmed in 38 patients with IIM-ILD (
r = 0.43,
p < 0.01) [
18]. In a recent case, B-lines and KL-6 were utilized to closely detect and follow a patient with anti-MDA-5-positive, clinically amyopathic dermatomyositis associated with rapidly progressive ILD [
79]. Changes in B-line numbers and serum levels of KL-6 were consistent with the changes in HRCT findings and clinical presentation. Basing the treatment decision on B-line scores and KL-6 values, the patient was successfully rescued and avoided excessive radiation exposure. The case indicated that combining lung ultrasound and serum biomarkers might be a possibility for monitoring rapidly progressive ILD.