To develop multidisciplinary recommendations to improve the management of rheumatoid arthritis-related interstitial lung disease (RA-ILD).
MethodsClinical research questions relevant to the objective of the document were identified by a panel of rheumatologists and pneumologists selected based on their experience in the field. Systematic reviews of the available evidence were conducted, and evidence was graded according to the Scottish Intercollegiate Guidelines Network (SIGN) criteria. Specific recommendations were made.
ResultsSix PICO questions were selected, three of which analysed the incidence and prevalence of RA-ILD, associated risk factors, and predictors of progression and mortality. A total of 6 specific recommendations on these topics, structured by question, were formulated based on the evidence found and/or expert consensus.
ConclusionsWe present the first official SER-SEPAR document with specific recommendations for RA-ILD management developed to resolve some common clinical questions and facilitate decision-making for patients.
Elaborar unas recomendaciones multidisciplinares para mejorar el manejo de la enfermedad pulmonar intersticial difusa asociada a la artritis reumatoide (EPID-AR).
MétodosUn panel de reumatólogos y neumólogos expertos identificó preguntas clínicas de investigación relevantes para el objetivo del documento. Se realizaron revisiones sistemáticas de la evidencia, que se graduó de acuerdo con los criterios del Scottish Intercollegiate Guidelines Network (SIGN). Tras ello, se formularon las recomendaciones.
ResultadosSe seleccionaron seis preguntas PICO, tres de las cuales específicamente evaluaron la incidencia y prevalencia de esta complicación, los factores de riesgo para su desarrollo, y los factores pronósticos de mortalidad y de progresión de la EPID-AR. Se formularon un total de 6 recomendaciones específicas sobre estos aspectos, estructuradas por pregunta, con base en la evidencia encontrada y/o consenso de expertos.
ConclusionesSe presenta el primer documento oficial SER-SEPAR de recomendaciones específicas para el abordaje la EPID-AR, con el fin de ayudar en la toma de decisiones a los clínicos directamente implicados en su manejo y aproximar la práctica asistencial a la mejor evidencia posible.
Rheumatoid arthritis (RA) is an immune-mediated systemic inflammatory disease that primarily affects the joints and causes joint destruction, impaired functional capacity, and quality of life, and may even reduce life expectancy. It is the most prevalent form of chronic polyarthritis in Spain with the greatest social and health impact. According to data from the EPISER study, RA has an estimated prevalence of .82% in the Spanish adult population.1
In addition to joint damage, in a variable percentage of cases the disease presents with different extra-articular manifestations, including pulmonary manifestations. The spectrum of respiratory complications caused by RA is very broad and includes pleural involvement, rheumatoid pulmonary nodules (including Caplan syndrome), interstitial lung disease (ILD) and follicular or obliterative bronchiolitis.2,3 Bronchiectasis, primary pulmonary hypertension, secondary amyloidosis, thromboembolic disease, and cricoarytenoid arthritis have been reported more rarely.
ILD is one of the most frequent and severe pulmonary complications of RA.4,5 The risk of RA patients developing this complication is almost 9 times higher than in the general population.6 Furthermore, although there has been a significant improvement in prognosis in the last 25 years, 7 ILD is still the second leading cause of death from the disease, accounting for 10%–20% of deaths. This clinical fact warrants a common strategy between rheumatologists and pneumologists to improve the prognosis and quality of life of patients with this complication.
For this reason, the Spanish Societies of Rheumatology (SER) and Pneumology and Thoracic Surgery (SEPAR) promoted the drafting of a joint document of recommendations for the management of RA-ILD based on the best available evidence, with the aim of answering most of the common clinical questions, reducing the variability in medical care and facilitating decision-making for clinicians directly involved in the care of the condition.
MethodsA qualitative synthesis of the scientific evidence and consensus techniques was used in the drafting of these recommendations, which reflect the agreement of experts based on the available evidence and their clinical experience.8
The procedure for developing the project was as follows:
- 1
Creation of the working group. An interdisciplinary working group was formed by 5 rheumatologists from SER (JN, GB, IC, NMV, and AMO) and 5 pneumologists from SEPAR (MA, ECJ, MAN, CV, and JARP). The participants were endorsed by their society to participate in this document. The clinical and methodological aspects were coordinated, respectively, by one of the rheumatologists (JN) and a pneumologist (JARP), as principal investigators (PI), and two methodology specialists from the SER Research Unit.
- 2
Identification of key areas. The content and key aspects of the document were defined, and the clinical research questions with the greatest impact on clinical practice were formulated. The questions were reformulated in patient, intervention, comparison, and outcome (PICO) format.
- 3
Literature search. A search for published scientific evidence was conducted and progressively expanded until October 2020. The databases PubMed (MEDLINE), EMBASE, and Cochrane Library (Wiley Online) were used for this purpose. The process was completed with a manual search of the references of the studies identified, as well as posters and conference abstracts that the reviewers and experts considered to be of interest.
- 4
Analysis and synthesis of the scientific evidence. Rheumatologists from SER who are experts in evidence review conducted the systematic reviews and synthesis of the scientific evidence. The level of scientific evidence was assessed using Scottish Intercollegiate Guidelines Network (SIGN)9 criteria (Appendix B Annex).
- 5
Formulation of recommendations. Once the critical reading had been completed, the PIs, together with the members of the expert group reviewing the evidence for each of the PICO questions, formulated the recommendations, based on formal evaluation or "reasoned judgement" of the evidence for each of the questions. The quality, quantity, and consistency of the scientific evidence, the generalisability, applicability, and clinical impact of the results were considered. The strength of the recommendations was graded using the SIGN7 system. For questions where the evidence was insufficient, recommendations were formulated based on consensus of the expert group.
The recommendations have been divided into five main areas: incidence and prevalence of ILD in RA, risk factors for the onset of ILD in RA, prognostic factors for mortality and pulmonary progression, as well as the safety of drug treatment in patients with RA-ILD, and its efficacy in managing this complication.
External review. The draft document was subjected to external review to ensure the validity and accuracy of the recommendations and, subsequently, to public exposure so that other SER and SEPAR members, as well as different groups and potentially interested entities, could evaluate the document, and offer criticism or suggestions.
The most common histological types of RA-ILD are usual interstitial pneumonia (UIP) and non-specific interstitial pneumonia (NSIP).2–7 The UIP pattern is as frequent or more frequent than NSIP. Other patterns described are organising pneumonia (OP), lymphoid interstitial pneumonia (LIP), diffuse alveolar damage (DAD), acute fibrinous and organising pneumonia (AFOP), desquamative interstitial pneumonia (DIP), respiratory bronchiolitis associated with interstitial lung disease and the combined pulmonary fibrosis + emphysema (CPFE) pattern.2–7,10 These last three patterns are included in the spectrum of smoking-related lung lesions,11 although they have also been described in patients with RA who are non-smokers.12,13
The ILD pattern must be correctly identified because of its therapeutic and prognostic implications, since the more inflammatory patterns (cellular NSIP, OP, LIP) will respond better to anti-inflammatory and immunosuppressive treatment than UIP.
Diagnosis of RA-ILDIf ILD is suspected (presence of cough and/or dyspnoea of more than 3 months duration or velcro-like crackles on respiratory auscultation), a chest X-ray and lung function test (LFT) including spirometry and diffusing capacity of the lungs for carbon monoxide (DLCO) are usually performed for initial screening. In general, chest X-ray is a very insensitive technique in the early stages. In ILD, spirometry shows restrictive ventilatory impairment together with decreased DLCO. However, in early stages, it is not uncommon for decreased DLCO to be the only abnormality present.
High-resolution computed tomography (HRCT) of the chest to confirm ILD is the gold standard for the diagnosis of this complication. The radiological pattern on HRCT correlates well with the histological pattern in most cases, and therefore a lung biopsy is usually not necessary. HRCT is also useful to quantify the extent of parenchymal involvement (alveolitis/fibrosis), to determine the potential reversibility of the lesions (alveolitis/fibrosis), to establish a prognosis (UIP vs. non-UIP) and to assess response to treatment.
Disease activity and response to treatment is monitored with LFTs (spirometry and plethysmography to assess forced vital capacity [FVC], DLCO and total lung capacity), the 6-minute walk test and assessment of dyspnoea with one of the different clinical scales. Chest HRCT is usually only repeated in cases where it is necessary to assess functional deterioration, treatment readjustments or if exacerbation, infection, or other respiratory complications, such as lung cancer, are suspected.
In advanced stages, Doppler echocardiography is also useful to detect onset of secondary pulmonary arterial hypertension.
Bronchoalveolar lavage is not routinely performed. It is only indicated if it helps in the differential diagnosis, primarily with infections. Lung biopsy is only performed if there is suspicion of non-RA-ILD, doubts about the presence of other entities such as neoplasms or infection, or to establish a histo-specific diagnosis in atypical cases.
The differential diagnosis of ILD is mainly infection, pulmonary drug toxicity, thromboembolic disease, heart failure, or lung cancer (fundamentally in smokers).
Definition of progressive fibrosing ILDA variable percentage of patients with RA-ILD (around 40% at 5 years after ILD diagnosis)14–16 develop progressive pulmonary fibrosis with severe deterioration of lung function, rapid progression to chronic respiratory failure, and high mortality. There is no standard or agreed definition of what is considered progressive fibrosing ILD. The randomised clinical trial INBUILD17 used the following four criteria to define progression in patients with fibrosing ILD: 1) a decrease in baseline FVC > 10% of the percentage of the estimated theoretical value; 2) a decrease in FVC of between 5%–10% with evidence of progressive fibrosing on chest HRCT; 3) a decrease in FVC of between 5%–10% with worsening respiratory symptoms (dyspnoea and dry cough), or 4) a worsening of dyspnoea with progressive fibrosing on HRCT in the previous 24 months despite treatment.
Some experts also define progression as a decrease in FVC between 5%–10% with a worsening of DLCO greater than 15%.18
According to the INBUILD study,17 fibrosing ILD is defined as fibrosis changes affecting more than 10% of the lung parenchyma.
However, and according to the classic definitions of the American Society of Pulmonology, worsening of any type of ILD (fibrosing or non-fibrosing) is defined as a decrease in FVC (expressed as a percentage of the predicted value) greater than 10% or a decrease in DLCO greater than 15% during follow-up.19,20
ResultsA total of 18 recommendations have been formulated. This article contextualises the six recommendations made following the systematic literature review (SLR) of the epidemiology, risk factors, and prognosis of RA-ILD (Table1).
SER-SEPAR recommendations for the management of rheumatoid arthritis-associated interstitial lung disease: epidemiology, epidemiology, risk factors and prognosis of RA-ILD.
| Recommendations | Recommendation grade |
|---|---|
| Based on its incidence and prevalence, it is recommended that the possibility of interstitial lung disease should always be considered in the initial assessment and follow-up of patients with rheumatoid arthritis | √ |
| Patients with rheumatoid arthritis should be systematically screened for interstitial lung disease in cases with respiratory symptoms or if velcro-like crackles are heard | √ |
| In cases without respiratory symptoms and with normal auscultation, the need for screening should be assessed individually according to the number of risk factorsa for onset of this complication, regardless of the time since onset of rheumatoid arthritis | √ |
| The presence of interstitial lung disease should always be considered when deciding on treatment, given the potential risk of pneumonitis described with some of the drugs commonly used in rheumatoid arthritis | √ |
| In patients with rheumatoid arthritis, none of the investigational serum biomarkers tested so far has been shown to have better predictive value for the onset of interstitial lung disease than anti-citrullinated protein/peptide antibodies. Based on the current evidence, the drafting group does not recommend the use of other serum biomarkers in routine clinical practice | D |
| In patients with rheumatoid arthritis and interstitial lung disease it is recommended that the presence of prognostic factors associated with progressionb and mortalityc be considered when planning follow-up (frequency of check-ups) and treatment strategy | B |
DLCO, diffusing capacity for carbon monoxide; FVC, forced vital capacity; HRCT, high-resolution computed tomography.
The main risk factors for the development of interstitial lung disease, easily identifiable in the clinic, are male sex, older age, late disease onset, duration of rheumatoid arthritis, history of smoking, sustained moderate, or high disease activity, and rheumatoid factor and anti-citrullinated protein/peptide antibody positivity.
The main predictors of progression of interstitial lung disease are the radiological pattern of usual interstitial pneumonia, elevated anti-citrullinated protein/peptide antibody titres, degree of baseline DLCO impairment, a ≥10% decrease (percentage of estimated theoretical value) in FVC during follow-up, extensive lung involvement on chest HRCT, and elevated serum levels of interleukin-6 and Krebs von den Lungen-6 glycoprotein.
The main prognostic factors associated with increased mortality are advanced age at diagnosis of interstitial lung disease, male sex, duration of rheumatoid arthritis, moderate or high disease activity, the usual radiological pattern of interstitial pneumonia, low baseline FVC and/or DLCO, a decrease in FVC > 10% or DLCO > 15% during follow-up, extensive lung involvement on chest HRCT, GAP and CPI (composite physiologic index) indices, and elevated serum levels of Krebs von den Lungen-6 glycoprotein.
For further information on any of the sections summarised in this article, the full content of the document can be accessed on the SER website (www.ser.es).
Epidemiology of RA-ILD— What is the incidence and prevalence of RA-IDP in patients with RA?
- □
Recommendation. Based on its incidence and prevalence, it is recommended that the possibility of ILD should always be considered in the initial assessment and follow-up of patients with RA (recommendation grade √).
It is vital to establish the prevalence and incidence of ILD in RA. Knowing the magnitude of this complication is essential to define strategies for early diagnosis and treatment, and good health planning.
The scientific evidence for this recommendation was based on an SLR in which the population was adult subjects, the intervention was a diagnosis of RA according to ACR 1987 or ACR/EULAR 2010 criteria, and the outcome was the incidence or prevalence of ILD diagnosed by chest HRCT scan and/or lung biopsy.
Incidence of RA-associated ILDThe incidence of RA-associated ILD is 1.056 cases per 1000 patients (1.452 in males and .677 in females) according to a study conducted in a Japanese population (IORRA cohort) between 2004 and 2007.21
Outside the body of evidence, two other studies have been identified that evaluate the incidence in the Caucasian population, although the condition was not diagnosed with HRCT in all cases. In the study by Koduri et al.,22 conducted in an English cohort of patients with RA onset, this complication was only investigated in the presence of respiratory symptoms or signs as part of routine clinical care practice. After a follow-up of 3 years, the incidence of ILD was 4.1 per 1000 patients. In another prospective population-based cohort study (Rochester, Minnesota) the cumulative incidence of ILD at 10, 20, and 30 years was 3.5, 6.3, and 7.7%, respectively.6 In this study, the risk of RA patients developing this complication was much higher than that of the general population (HR: 8.96).
Prevalence/frequency of RA-ILDEleven articles have been identified in which all patients diagnosed with RA (consecutive unselected patients) were screened for ILD.23–33 The main results of these studies are shown in Table 2. The frequency of this complication shows great variability, ranging from 11.9% to 55.7%. This variability can be explained by the differences between the populations studied (great heterogeneity in demographic characteristics, time since onset, and degree of disease activity, frequency of smoking and percentage of patients with respiratory symptoms) and by the lack of uniformity in the HRCT findings used to establish a diagnosis of ILD.
Main characteristics of the 11 studies that have analysed the prevalence of RA-ILD included in the systematic review of the literature.
| Study | Country | N° of centres | N | Age (years) | Time since onset (years) | Smoking (%) | RF (+)/ACPA (+) (%) | Respiratory symptoms | Frequency of ILD |
|---|---|---|---|---|---|---|---|---|---|
| Bilgici et al.23 (2005) | Turkey | 1 | 52 | 53.6 ± 11.2 | 8.37 ± 8.17 | 25% | 67.3%/NS | Asymptomatic 59.6% (31/52) | 55.76% |
| Symptomatic 21% (11) | |||||||||
| Dawson et al.24 (2001) | UK | 1 | 150 | 58.9 ± 10.3 | 12.7 ± 8 | 70% | 77%/NS | Asymptomatic; 15 had crackles on respiratory auscultation (10%) | 18.6% |
| Mohd Noor et al.28 (2009) | Malaysia | 1 | 63 | 56.7 ± 10 | 14 ± 12 | 6% | 66.7%/NS | Asymptomatic 67% | 44.4% |
| Mori et al.29 (2008) | Japan | 1 | 126 | 60 ± 12.4 | 61 patients with established RA (Time since onset: 11.8 ± 9.7) | 20% | 83%/93.8% | Patients with established RA: Asymptomatic 68.8% (42/61) Symptomatic 31.2% (19) | Total: 11.9% (15/126) |
| 65 patients with time since onset: 2.6 ± 4.5 | Asymptomatic patients 83.1% (54/65) Symptomatic 16.9% (11) | 13.1% (8/61) | |||||||
| 10.7% (7/65) | |||||||||
| Skare et al.30 (2011) | Brazil | 1 | 71 | 58.7 ± 10.4 | 11.8 ± 6.9 | 35.8% | 704%/38% | Asymptomatic 44.75% (30/71) | 54.92% |
| Symptomatic 55.25% | |||||||||
| Wang y Du31 (2015) | China | 1 | 544 | NS | 5.6 ± 6.5 | 6.8% | NS/NS | NS | 15.26% |
| Crackles on auscultation: 52 (9.5%) | |||||||||
| Zhang et al.32 (2017) | China | 1 | 550 | 61 ± 13 | 8 ± 9 | 20.4% | 76.5%/NS | Asymptomatic 59% | 43.1% |
| Symptomatic 41% | |||||||||
| Zrour et al.33 (2005) | Tunisia | 1 | 74 | 48 ± 14 | 96 ± 88 m (8 years) | 14.6% | 60.6%/NS | Asymptomatic 74.7% | 28% |
| Symptomatic 25.3% | |||||||||
| Gabbay et al.25 (1997) | Australia | 1 | 36 | 51.8 ± 16 | < 2 | 55.5% | 64%/NS | Asymptomatic | 33% |
| Gochuico et al.26 (2008) | United States | 1 | 74 | NS | < 2 | 50% | NS/NS | Asymptomatic | 41.89% |
| Habib et al.27 (2011) | Saudi Arabia | 3 | 40 | 37.6 ± 10.3 | < 2 | 12.5% | 35% /NS | Asymptomatic | 27.5% |
ACPA: Anti-citrullinated Protein/peptide Antibodies; ILD: Interstitial Lung Disease; NS: Not specified; RF: Rheumatoid Factor.
Of the eleven articles identified, three included patients with RA of less than two years' duration.25–27In these studies, the variability of the prevalence of ILD is lower and ranges from 27.5% to 41.8%, with higher figures in populations with a higher frequency of smoking.25,26
In the eight studies conducted in patients with established RA (more than two years since onset)23,24,28–33 there is wide disparity in the frequency figures, even between the two studies that investigated this question prospectively.23,24 Thus, Bilgici et al.,23 in a cohort of 52 unselected RA patients from a single centre in Turkey, found chest CT changes consistent with ILD in 55.7%. Dawson et al.,24 in a cohort of 150 consecutive RA patients in two hospitals in the North of England, describe findings of "fibrosing alveolitis", a term comparable to UIP and fibrosing NSIP patterns, in 19% of cases. In general, and predictably, in these eight studies, the higher the percentage of patients included with respiratory symptoms, the higher the frequency of ILD.
Outside the body of evidence, in a prospective study conducted in Spain in patients with RA onset without respiratory symptoms in which only those with radiographic or LFT abnormalities underwent chest HRCT, the prevalence of confirmed ILD at disease diagnosis was 5%, increasing to 7.5% at 5 years since onset.34 These figures are lower than those reported in a Chinese cohort in which the frequency of ILD at diagnosis was 9.76%, increasing to 30.9% at 5-year follow-up.35
However, if we only consider the frequency of symptomatic or clinically significant RA-ILD (which is not the issue raised in the research question), the prevalence by HRCT in several landmark studies is between 10% and 29%.36–39 Finally, the prevalence of ILD in the only autopsy study to investigate this complication is 35%.40
Risk factors for onset of ILD in RA— In patients with RA, what are the risk factors for onset of ILD?
- □
Recommendation. Patients with RA should be systematically screened for ILD in cases with respiratory symptoms or if velcro-like crackles are heard (recommendation grade √).
- □
Recommendation. In cases without respiratory symptoms and with normal auscultation, the need for screening should be assessed individually according to the number of risk factors for onset of this complication, regardless of the time since onset of RA (recommendation grade √).
- □
Recommendation. The presence of ILD should always be considered when deciding on treatment, given the potential risk of pneumonitis described with some of the drugs commonly used in RA (recommendation grade √).
- □
Recommendation. In patients with RA, none of the investigational serum biomarkers tested so far has been shown to have better predictive value for the onset of ILD than ACPA antibodies. Based on the current evidence, the drafting group does not recommend the use of other serum biomarkers in routine clinical practice (recommendation grade D).
Despite its frequency and severity, to date there are no recommendations on screening for this complication, in either the initial assessment or in the follow-up of RA patients. However, as mentioned above, when the entire RA population was screened for ILD by HRCT, a non-negligible percentage of subclinical disease was detected (between 11.9% and 55.7%), confirming that it is underdiagnosed. In fact, in clinical practice, it is not uncommon for ILD to be diagnosed late, as it is often asymptomatic or pseudo-symptomatic in its early stages.
Therefore, knowing the main risk factors associated with the onset of RA-ILD will help us identify the subgroup of patients at greatest risk of developing this complication who may benefit from specific screening to diagnose them early (even if they do not have respiratory symptoms or signs).
According to the results of the SLR, the main risk factors for the onset of RA-ILD are:
- 1
Male sex35,41 (level of evidence 2+).
- 2
Advanced age and late onset of the disease (level of evidence 2++).32,35,36,42,43 No cut-off point is established in most studies, except two: >60 years (OR: 1.48; 95% CI: 1.01–2.18)35 and ≥65 years (RR: 4.58; 95% CI: 1.67–12.53).42
- 3
Duration of the disease, onset of ILD being more frequent during the first 5–10 years of RA progression28,32,35 (level of evidence 2+).
- 4
Smoking32,36,44,45 (level of evidence 2++). Smoking and smoking rate are strongly associated with onset of ILD, with an OR: 3.76 (95% CI: 1.59–8.88) in the case of having smoked ≥25 pack-years and 1.9 (95% CI: .68–5.24) if less than 25 pack/years.44 Several studies indicate that smokers’ increased risk of developing ILD is associated with the presence of the shared epitope, showing that there is an important gene-environment interaction between the shared epitope alleles and smoking.1,42,44
- 5
Moderate or high sustained RA activity45 (level of evidence 2++). Patients with moderate/high maintained disease activity according to DAS28-CRP scores are at twice the risk of developing ILD than patients in remission or with low disease activity (results adjusted for sex, smoking, RA duration, and serostatus: HR: 2.22; 95% CI: 1.28–3.82).45 The risk of ILD in the different disease activity statuses is as follows: low disease activity (DAS28 > 2.6–3.2) 1.41 (95% CI: .61–3.28), moderate disease activity (>3.2-5.1) 2.08 (95% CI: 1.06–4.05), and high disease activity (>5.1) 3.48 (95% CI: 1.64–7.38).45
- 6
RF positivity (level of evidence 2++) and, especially, ACPA positivity (level of evidence 1+), especially at high titres.10,32,35,46–48 In a meta-analysis (which included 7 studies with 1685 patients; I2: 0%), the risk of ILD and pulmonary fibrosis in the presence of ACPA was almost 5 times higher (OR: 4.68; 95% CI: 2.07–10.57).46
In addition to RF and ACPA, the usefulness of other potential biomarkers predictive of the onset of ILD is being investigated, most notably antibodies directed against carbamylated proteins (anti-CarP). A correlation has been demonstrated between levels of all anti-CarP specificities (foetal calf serum, fibrinogen, and chimeric fibrine/filagrine homocitrullinated peptide) and the presence of ILD, after adjustment for other known risk factors49 (level of evidence 3). Promising preliminary results have also been obtained with serum levels of extracellular matrix metalloproteinase 7 (MMP-7), interferon gamma-inducible protein-10 [IP-10] or CXCL10, interleukin-18, and 90- and 70-KDa heat shock proteins (HSP90/70)50–53 (level of evidence 2). None of these are yet available in clinical practice, nor have they as yet been shown to have greater predictive value for the development of this complication than ACPAs.
Likewise, some genetic biomarkers have been identified, including mutations in the MUC5B gene54 (which encodes one of the proteins of the mucin family) (level of evidence 2+) and mutations in the telomerase genes that cause accelerated telomere shortening55,56 (in cases with a family history of ILD) (level of evidence 2+).
Finally, onset of ILD has also been linked to some of the drugs commonly used in the treatment of this disease. This issue is reviewed in depth in the second part of the recommendations, also published in this journal.
To avoid late diagnosis of ILD in RA patients, the drafting group recommends regular questioning about the presence of respiratory symptoms and auscultation at least once a year for dry velcro-like crackles. The presence of velcro-like crackles correlates well with the presence of pulmonary fibrosis.57 In addition, two studies have shown that conventional respiratory auscultation is acceptably diagnostically accurate to detect ILD compared to digital stethoscope.58,59
The group also considers that patients with respiratory symptoms (cough and/or dyspnoea of more than 3 months' duration in the absence of known underlying pulmonary disease or pre-existing heart disease) should be routinely screened, and all those with velcro-like crackles detected on auscultation, even if asymptomatic. In asymptomatic patients with normal auscultation, the need for ILD screening will be assessed on an individual basis according to the number of risk factors for onset of the complication, regardless of time since onset of RA.
Risk factors for onset of RA-ILD— What are the prognostic factors for mortality and progression of Lund disease in patients with RA-ILD?
- □
Recommendation: In patients with RA and ILD, it is recommended that the presence of prognostic factors associated with progression and mortality be considered when planning follow-up (frequency of check-ups) and treatment strategy (recommendation grade B).
ILD remains the second leading cause of premature death in RA, after cardiovascular complications.2–7 Approximately half of patients have stable or very slowly progressive ILD, whereas in the other half lung function deteriorates more or less rapidly.60,61 This heterogeneity in the course of the disease makes it essential to identify prognostic factors for severe disease and mortality. The SLR identified factors associated with progression of ILD and factors associated with increased mortality in patients with RA-ILD.
Predictors of RA-ILD progression- □
Radiological pattern of UIP: the risk of progression in patients with UIP is approximately three times higher than in patients with NINE16,35 (level of evidence 2+).
- □
Elevated ACPA titres (3 times above the upper limit of normal) with an OR of 4.03 (95% CI: 1.04–15.6)62 (level of evidence 2+).
- □
The degree of baseline DLCO deterioration,35,62,63 having been demonstrated with two cut-off points: DLCO < 45% (percentage of predicted value) (OR: 3.02; 95% CI: 1.13–8.03)35 and, only in patients with a progressive fibrosing phenotype, DLCO < 54%: (OR: .85; 95% CI: .74–.98)63 (level of evidence 2+).
- □
In patients with UIP pattern, a drop ≥10% in FVC during follow-up (OR: .89; 95% CI: .82–.97) (level of evidence 2+).64
- □
Overall, the risk of progression of ILD increases the lower the baseline DLCO values, and the greater the worsening of FVC and/or DLCO during follow-up16 (level of evidence 2+).
- □
Extensive distribution on fibrotic changes on HRCT, principally traction bronchiectasis (≥ 4 lung areas affected with > 50% of the subpleural total)65 (level of evidence 2++).
- □
Elevated serum levels of IL-6 and KL-6 (level of evidence 2+).64,66
Patients with RA-ILD have a 3- to 10-fold higher adjusted mortality risk than RA patients without the complication, regardless of the follow-up period and the presence of comorbidities6,67,68 (level of evidence 2+).
The main prognostic factors for mortality identified in the SLR are:
- □
Advanced age at diagnosis of ILD22,35,60,62,64,67,69–72: no cut-off point is established in most of the studies found, except three: >60 years (HR: 3.18; 95% CI: 1.27 to 7.9435 and HR: 2.32; 95% CI: 1.27 to 4.2562) and ≥65 years (OR: 1.08; 95% CI: 1.02–1.15)70 (level of evidence 2+).
- □
Male gender67,73,74: men have 3.6- to 14.5-fold higher mortality than women with this complication67 (level of evidence 2+).
- □
Duration of RA60,75: the longer the duration of disease at diagnosis of ILD, the higher the mortality (HR per decade: 1.81; 95% CI: 1.16–2.83)60 (level of evidence 2+).
- □
Moderate or high disease activity (DAS28-VSG >3.2) is also associated with increased mortality (OR: 1.6; 95% CI 1.0–2.5).75
- □
Radiological pattern of UIP69,70,73,74,76,77: according to the conclusions of a recent meta-analysis that included 10 retrospective studies with 1.256 patients with RA-ILD (I2: 76%),76 the UIP pattern has a mortality risk almost 2 times higher than non-UIP patterns (RR: 1.66; 95% CI: 1.07–2.76) (level of evidence 1++). In analysis by subgroups, when the UIP pattern was compared with the NSIP pattern the RR was 2.39 (95% CI: .86–6.68) and 1.45 (95% CI: .89 to 2.37) when the UIP was compared with other non-NSIP patterns. Increased mortality has also been described with a probable UIP pattern77 (level of evidence 2+) and, in general, with all the forms of fibrosing RA-ILD (HR: 2.1; 95% CI: 1.11–4.26)69 (level of evidence 2+).
It is important to highlight that in two studies60,78 the radiological pattern is no longer a predictor of mortality after adjustment for confounding variables in multivariate analysis, the main prognostic factors being the degree of baseline deterioration in respiratory function tests (FVC and DLCO) and the magnitude of their worsening during follow-up.
- □
A low baseline DLCO or a decrease in DLCO > 15% during follow-up (level of evidence 2++).60,61,71–73,76,79 Only the study by Zamora-Legoff et al.60 establishes a cut-off point for baseline DLCO, setting it at less than 40% (HR: 2.48; 95% CI: 1.55–3.95) (level of evidence 2+). The prognostic value of baseline DLCO is independent of the presence or absence of emphysema due to smoking76 (level of evidence 1++).
- □
A low baseline FVC during follow-up (HR: 2.57; 95% CI: 1.79–3.70) (level of evidence 1++) or a decrease in FVC > 10%61,71,76 (level of evidence 2+). These studies do not establish a cut-off point for baseline FVC.
- □
Extensive lung involvement on thoracic HRCT has been demonstrated when more than 20% of the lung parenchyma is involved10,80,81 (level of evidence 2++), and when more than 30% is involved 30%61,71 (level of evidence 2+).
- □
Elevated serum levels of KL-6, which are associated with higher mortality (3–4 times higher)64,74 (level of evidence 2+).
Finally, the usefulness of the GAP index (calculated with gender, age, and FVC and DLCO values) to predict the risk of mortality at 1, 2 and 3 years in patients with RA-ILD has been demonstrated60,82 (level of evidence 2+). The CPI score83 (level of evidence 2+) also appears to be useful for this purpose.
DiscussionRA-ILD is a complex clinical situation, both because of its frequency and potential severity, and because of the difficulty in its therapeutic management. There is little published evidence on the subject, most is of low methodological quality, and with contradictory results in some aspects. There are also no recommendations yet agreed upon by scientific societies.
This why we present the first official document drawn up by the SER and SEPAR with specific recommendations for the management of RA-ILD, which aims to serve as a guide for professionals caring for these patients and to bring clinical practice in line with the best possible evidence. These recommendations have been reached through a strict and validated methodology of systemic reviews of the scientific literature and expert consensus techniques.
Epidemiological information is essential to measure the socio-health importance of diseases and to contribute to proper health planning. Therefore, one of the main objectives of the paper was to determine the prevalence and incidence of ILD in RA patients. However, the SLR revealed great variability in the studies that have investigated this issue, and therefore its actual frequency could not be reliably established. As a guideline, the prevalence of symptomatic ILD diagnosed by chest HRCT varies between 10% and 29%,36–39 whereas its incidence varies between 1.05 and 4.1 cases per 1000 patients,21,22 and the risk of developing this complication in RA patients is much higher than in the general population.6,84
In more than three-quarters of patients, ILD develops after RA has been diagnosed, usually in the first 5–10 years of disease progression.2–7,21,22,28,32,35,85 In these cases, it is not uncommon for it to be diagnosed late, since it is often asymptomatic in its initial stages or presents with few symptoms, and goes unnoticed until advanced stages, when the extent of fibrotic changes limits therapeutic possibilities and survival (UIP is the most frequent histopathological pattern in RA). Less frequently, ILD occurs at onset of RA or precedes the joint manifestations by months or years.2–7 This circumstance has been reported in up to 10%–14% of patients with RA and ILD and is not always present and is often confused with idiopathic forms, despite their different prognosis and treatment.68,86,87
Although a significant improvement in prognosis has been observed over the last 25 years,7 RA-ILD is still the second leading cause of premature death in RA, after cardiovascular complications.2–7 Patients with RA-ILD have an adjusted mortality risk 3–10 times higher than RA patients without this complication6,67 and their median survival from diagnosis of ILS varies from 2.6 to 8.1 years.88
Although ILD is a frequent and under-diagnosed complication with high morbidity and mortality, there are as yet no recommendations for screening. In the absence of studies supporting the cost-effectiveness of universal screening, a reasonable strategy would be to design a selective screening proposal based on patients’ clinical risk. This clinical risk would be defined by the presence of respiratory symptoms (cough and/or dyspnoea of an unrelated cause of more than 3 months' duration), by the presence of dry velcro-like crackles on respiratory auscultation, or by the number of risk factors for the patient developing this complication. This initiative towards a multidisciplinary approach to selective screening criteria to identify patients with RA and ILD early was launched as a joint collaboration between the SER and SEPAR as part of the RA-EPIDSER project.89 A tool to enable early identification of these patients with good sensitivity will help improve their survival and quality of life.
Another important point is that RA-ILD behaves heterogeneously in terms of its clinical course and prognosis. ILD will be stable or progress very slowly in approximately half of patients, while in the other half lung function deteriorates rapidly, especially in patients with fibrosing ILD.60,61 This heterogeneity in progression makes it essential to identify prognostic factors for pulmonary progression and mortality to detect patients with a more severe profile early, who will be candidates for specific treatment as soon as the diagnosis of ILD is confirmed and for close clinical follow-up. The main prognostic factors include UIP pattern, extensive lung involvement on chest HRCT, the severity of deterioration of FVC and DLCO at diagnosis of RA-ILD, and the extent that these parameters have worsened during follow-up.10,60,61,69–74,76,77,79–81 Apart from the clinical and functional variables already in use, it remains to be seen how useful some of the biomarkers under investigation are, mainly IL-6 and KL-6, will be as tools to help predict progression and mortality in the short and medium term. Although the prognostic predictive capacity of these biomarkers has been validated in some studies (most with a small sample size), their use in daily clinical practice is limited and international guidelines are reluctant to recommend their use90 mainly due to the lack of efficiency studies demonstrating the cost-benefit compared to other currently used clinical and functional variables, the absence of consensus on the measurements that could be most useful, and the need for additional validation studies with a large number of patients.
Ethical responsibilitiesProtection of humans and animals. The authors declare that no experiments on humans or animals have been conducted for this research.
Data confidentiality. The authors declare that no patient data appear in this article.
Right to privacy and informed consent. The authors declare that no patient data appear in this article.
FundingThis project was funded by the Spanish Society of Rheumatology (SER) and the Spanish Society of Pneumology and Thoracic Surgery (SEPAR).
AuthorshipAll the authors made substantial contributions to a) the conception and design of the study and the data analysis, b) drafting of the article or critical review of the intellectual content, and c) final approval of the version presented.
Conflict of interestsJosé Antonio Rodríguez Portal has received funding from Roche and Sanofi for attendance at courses/congresses; honoraria from Roche, Boehringer and Janssen for presentations; financial support from Bristol, Janssen, Roche, and Boehringer for consultancy for pharmaceutical companies and other technologies; and has participated in clinical trials funded by Gilead, Roche and Boehringer.
Noé Brito García has no conflict of interests to declare.
Petra Díaz del Campo Fontecha has no conflict of interests to declare.
Claudia Valenzuela has received funding from Boehringer and Roche for attendance at courses/congresses; honoraria from Roche for presentations; and has received financial support from Boehringer, Galapagos, and Roche for consultancy for pharmaceutical companies and other technologies.
Ana María Ortiz has received funding from Lilly, Novartis, Pfizer, Roche, Sandoz, Sanofi, and UCB for attending courses/congresses; honoraria from Abbvie, Lilly, MSD, and Roche for presentations; funding from Bristol, Gilead, MSD, and Roche for participating in research, and has received financial support from Abbvie, Gilead, Janssen, Lilly, and Pfizer for consultancy for pharmaceutical companies and other technologies. She has also participated in clinical trials for MSD/Amgen and Bristol.
María Asunción Nieto has received funding from Boehringer, MSD, Roche, and TEVA for attending courses/congresses; honoraria from Boehringer, BMS, and Roche for presentations and for educational programmes or courses; funding from Boehringer and Roche for participating in research; and has received financial support from Roche, Boehringer, and Astra for consultancy for pharmaceutical companies and other technologies.
Natalia Mena Vázquez has received funding from Abbvie, Novartis, Pfizer, and Roche for attending courses/congresses; honoraria from Abbvie, MSD, Pfizer, and Roche for presentations; and has received funding from Abbvie and MSD for educational programmes and courses.
Esteban Cano Jiménez has received funding from Boehringer and Roche for attendance at courses/congresses; honoraria from Boehringer, Chiesi, Roche, and Rovi for presentations; and has received financial support from Galapagos for consultancy for pharmaceutical companies and other technologies.
Iván Castellví has received funding from Actelion, BMS, Boheringer, Kern, Lilly, Novartis, Pfizer, and Sanofi for attendance at courses/congresses; honoraria from Actelion, BMS Boehringer, Nordic, Pfizer, and Roche for presentations; funding from Actelion for educational programmes and courses, and has received financial support from Actelion, Boehringer, Gebro, and Kern for consultancy for pharmaceutical companies and other technologies.
Myriam Aburto has received funding from Boehringer and Roche for attendance at courses/congresses; honoraria from BMS, Boehringer, and Roche for presentations; and financial support from Boehringer for consultancy for pharmaceutical companies and other technologies. She has also participated in a Boehringer clinical trial.
Gema Bonilla has received funding from Abbvie and Janssen for attendance at courses/congresses and honoraria from Abbvie, BMS, Boehringer, Novartis, Roche, and UCB for presentations.
María Vanesa Hernández has no conflict of interests to declare.
Félix Manuel Francisco Hernández has no conflict of interests to declare.
María Correyero Plaza has no conflict of interests to declare.
Isabel Castrejón has no conflict of interests to declare.
Miguel Ángel Abad Hernández has no conflict of interests to declare.
Javier Narváez has received funding from Bristol, Kern, Lilly, Pfizer, Roche, and Sanofi for attendance at courses/congresses; honoraria from Abbvie, Bristol, Boehringer, Gebro pharma, GSK, Kern, Lilly, Pfizer, Sanofi, and Sobi for presentations and consultancies for pharmaceutical companies and other technologies, and has participated in clinical trials funded by Boehringer, GSK, Janssen, Roche, and Vorso.
The group of experts in this study would like to thank Mercedes Guerra Rodríguez, documentalist at the SER, for her collaboration in the evidence search strategies.
They would also expressly like to thank Dr. Raimon Sanmartí Sala and Dr. Gustavo Enrique Zabert as expert reviewers of the document, for their critical review and contributions to it.
They would also like to thank Dr. Federico Díaz González, Director of the SER Research Unit, for his participation in the review of the final manuscript and for his help in preserving the independence of this document.
The following is Supplementary data to this article:
Please cite this article as: Rodríguez Portal JA, Brito García N, Díaz del Campo Fontecha P, Valenzuela C, María Ortiz A, Nieto MA, et al. Recomendaciones SER-SEPAR para el manejo de la enfermedad pulmonar intersticial difusa asociada a la artritis reumatoide. Parte 1: epidemiología, factores de riesgo y pronóstico. Reumatol Clín. 2022;18:443–452.
