American Journal of Respiratory and Critical Care Medicine

Interstitial lung disease (ILD) is associated with rheumatoid arthritis (RA); however, the prevalence and natural history are undefined. Our aim was to determine the prevalence of ILD associated with RA using a number of sensitive techniques in patients with joint disease of less than 2-yr duration. Patients who met ARA criteria for RA were recruited from community-based and hospital rheumatologists and assessed using the following measures: clinical, lung physiology, radiology (chest X-ray, high resolution CT [HRCT]), bronchoalveolar lavage (BAL) and 99mTc-DTPA nuclear scan. Thirty-six patients (25 female and 11 male) of joint disease duration of (mean ± SD) 13.2 ± 8.6 mo were studied. Abnormalities consistent with ILD were found in one or more investigations in 21 of 36 (58%), which were in lung physiology in 22%, CXR in 6%, HRCT in 33%, BAL in 52%, and 99mTc-DTPA nuclear scan in 15%. Based on the results, they were categorized as having clinically significant ILD (Group 1), abnormalities compatible with ILD, but no clinically significant ILD (Group 2) and no abnormalities compatible with ILD (Group 3). Five of 36 (14%) were in Group 1, 16 of 36 (44%) in Group 2, and 15 of 36 (42%) in Group 3. The only risk factor for the presence of abnormalities compatible with ILD was male gender (p < 0.04, Student's t test). In conclusion, changes consistent with ILD in early RA are frequent. The significance of these changes is being determined in a longitudinal study.

Pulmonary disease is found in association with rheumatoid arthritis (RA), the most common clinical manifestation being diffuse interstitial lung disease (ILD) (1-3). This is often a progressive disease which may result in disabling symptoms and respiratory failure (4). The reported prevalence of ILD in RA varies according to the criteria used to define disease, ranging from low rates based on chest X-ray (CXR) abnormalities to 41% based on abnormal physiology and up to 71% based on the presence of an alveolitis (1, 5, 6). Studies have not been able to predict who will develop disease nor defined the natural history of RA associated ILD.

Three investigations offer the potential for sensitive detection of both the presence and type of lung abnormalities in the early phase of disease. These are the high resolution CT scan (HRCT), bronchoalveolar lavage (BAL), and 99mTechnetium diethylene triamine pentacetate radionuclide scanning (99mTc-DTPA). HRCT is more sensitive and accurate than conventional radiography in detecting early changes of pulmonary fibrosis (7, 8). BAL is widely used to investigate ILD and is a sensitive indicator of inflammatory processes within the alveoli (1, 6). Clearance of 99mTc-DTPA from the lung is a measure of epithelial permeability (9) and is abnormal in several interstitial lung diseases (10) although has not been well studied in RA. In patients with scleroderma, 99mTc-DTPA clearance may be abnormally rapid in patients without clinically significant ILD (11) and a persistently rapid 99mTc-DTPA clearance appears to predict progressive disease (12). Therefore 99mTc-DTPA clearance in RA may define patients with subclinical disease as well as predict those in whom ILD is likely to progress.

Previous studies have not looked at patients with recent onset RA and the prevalence of ILD in this group is unknown. As development of pulmonary disease usually postdates the onset of joint symptoms by only 0–5 yr (1-3), surveillance of patients with recent onset of RA offers the potential for early detection of lung disease. This may increase our understanding of the natural history of ILD and determine which markers predict the presence of or progression of disease.

We therefore investigated a cohort of patients with recent onset RA to determine the prevalence of ILD defined by abnormalities in clinical assessment, lung physiology, plain radiology, HRCT, BAL, and 99mTc-DTPA radionuclide scanning.

Subjects

Thirty-six adult patients were recruited after referral from rheumatologists in Perth, from either hospital or community practices. All patients satisfied ARA criteria for RA (13). Duration of joint symptoms was < 2 yr; pregnant or lactating women were excluded. The study was approved by the Ethics Committee of Royal Perth Hospital and informed consent was obtained.

The characteristics for the group are shown in Table 1. There were 25 women and 11 men; 61% (22 of 36) had a positive rheumatoid factor. The majority (83%, 30 of 36) were on disease modifying anti-rheumatic drug (DMARD) therapy with 58% (21 of 36) taking methotrexate.

Table 1. DEMOGRAPHICS OF STUDY POPULATION

Total number36
Gender, F:M* 25:11
Age, years 51.8 ± 16.0 (20–86)
Smoking status*
 non16 (44%)
 ex16 (44%)
 current 4 (11%)
DMARD treatment*,
 Number of agents
  0 6 (17%)
  125 (69%)
  2 5 (14%)
Agent used
 Methotrexate21 (58%)
 Sulphasalazine 9 (25%)
 Hydrochloroquine 3 (8%)
 Gold 2 (6%)
Markers of disease activity*
 Rheumatoid factor positive23 (64%)
 Active/severe disease§ 13 (36%)
 Extra-articular manifestations 4 (11%)

*Data are frequency (%).

Mean ± standard deviation (range).

DMARD = disease modifying anti-rheumatic drug.

§Active/severe disease based on combination of rheumatoid factor status, ESR, CRP, presence of joint erosions and Ritchie index.

Clnical Assessment

Patients were asked a modified American Thoracic Society Respiratory Questionnaire (14) relating to cough, dyspnea, sputum production, wheeze, past medical problems, and risk factors for respiratory disease such as smoking, medications, occupation, and recreational history. Cigarette consumption was evaluated in pack years (one pack year = 20  /day for 1 yr). Current smokers were those who had smoked during the previous 6 mo; never smokers had smoked less than 20 packs during their lifetime. Patients were examined for the presence of basal crackles and extra-articular manifestations (e.g., vasculitis, nodules, Sjögren's syndrome); joint activity was assessed by Ritchie index of the number of tender and swollen joints (15). Blood was taken from each patient for rheumatoid factor (RF titre; latex agglutination, n = < 50 IU/ml), hemoglobin, ESR, and C-reactive protein (CRP).

To determine if the character of the joint disease was associated with pulmonary abnormalities, the patients were classified as having active and/or severe disease, as has been previously validated (16) by the presence of three or more of the following; (1) rheumatoid factor positive; (2) ESR ⩾ 30 OR CRP ⩾ 20; (3) > 2 joint erosions; (4) Ritchie index > 12. Based on these criteria, 36% (13 of 36) were classified as having active/severe rheumatoid disease.

Pulmonary Function Tests

Lung function was measured using a standard protocol and included spirometry (PK Morgan spiroflow; PK Morgan Ltd, Kent, UK), lung volumes (PK Morgan helium dilution analyzer) and transfer factor measured by single breath diffusing capacity (PK Morgan autolink breath system). The highest of three reproducible measurements was used, corrected for haemoglobin (Hb adjusted-Dl CO) and expressed as % predicted for age, height and gender according to standardized tables (17). A 6 min walk test was performed after familiarization and distance walked and SaO2 at the beginning and end of the test (finger pulse oximetry [18]) were measured.

Lung physiology was considered abnormal and compatible with ILD if the Hb adjusted-Dl CO was ⩽ 75% predicted in the absence of airflow obstruction or desaturation on exercise of ⩾ 4% to a SaO2 ⩽ 95%. Isolated reduction in FVC or TLC to below 80% predicted without a reduction in Hb adjusted-Dl CO was not considered physiological evidence of ILD.

Radiology

Hand and foot radiographs (looking for the presence of erosive joint disease), CXR, and HRCT were performed. The HRCT was performed with 1 mm thick sections at 10 mm intervals with prone slices to exclude dependent changes. The radiology was analysed by two radiologists blinded to the history. The HRCT scans were analysed according to the method of Remy-Jardin (19) and given a semi-quantitative score for the presence or absence of changes compatible with fibrosis or ground glass attenuation (0 = absent, 1 = mild, 2 = moderate and 3 = severe [honeycombing for fibrosis]).

Bronchoalveolar Lavage and Cell Analysis

Bronchoscopic examination and lavage was performed using a standard method. Patients were pre-medicated with atropine and omnopon intramuscularly. The upper airway was anaesthetized with 4% lignocaine, the vocal cords with 2% lignocaine. The bronchoscope (Olympus B2; Olympus, Japan) was wedged into a subsegmental bronchus in the right middle lobe. In no patients were the airways reddened, however lignocaine or 20 ml of saline was instilled and discarded prior to the lavage which was performed with 240 ml of warm saline (4 × 60 ml aliquots). The lavage was filtered through two layers of sterile gauze to remove mucus, centrifuged, washed twice and resuspended in RPMI 1640 with HEPES. Cells were resuspended at 2 × 10 5/ml in PBS. Cytospin preparations were made and cell differential count performed on slides fixed with 95% alcohol and stained with Papanicolaou (Kinetik, Brisbane, Australia), and air dried slides fixed with Diff-Quik stain set (Lab Acids Pty. Ltd., Narambeen, Australia). No significant differences were found between the two methods. The mean for four fields of 200 cells each was taken. Bronchial cells were less than 5% in all patients and were excluded from the final differential cell count.

Twenty-nine patients underwent a BAL. Four patients declined the procedure and three others were excluded on the basis of age and premorbid condition (20). Based on previous local measurements of control subjects, lavage cell differential was considered normal if values were as follows: (1) Nonsmokers: lymphocytes < 15%, neutrophils <  2%, eosinophils < 0.5%; (2) Smokers: lymphocytes < 10%, neutrophils < 4%, eosinophils < 1%.

99mTc-DTPA Scan

Subjects inhaled 2 ml of an aerosol containing 99mTc-DTPA (pentastan RM5 ARI). The aerosol was produced using a nebulizer (Mallinkrodt Ultravent; Mallinkrodt Medical, Inc., MO) with compressed air flow rate of 11.5 L/min. Previous tests in our laboratory have shown the resultant aerosol to consist of particles of MMAD 1.4 (± 0.2) mm. Subjects inhaled the aerosol whilst supine during normal tidal breathing through a mouthpiece with a one-way valve and noseclip in place until approximately 2.0 mm Ci were retained in the lung. Subjects remained supine and radioactivity was measured over the posterior thorax continuously for 30 min with a gamma camera (Siemens Orbiter; Siemens, Aktiengesellschafterlangen, Germany). Regions of interest were defined for each lung using a 10% edge cutout. Background subtraction was employed. Clearance rates were found to be mono-exponential (best fit); rates were calculated over 30 min in either lung separately and both lungs together and as the difference between these two measures was minimal for analysis the clearance rate for both lungs together was used. Clearance rate was defined as when radioactivity reached half of its initial radioactivity (t1/2) and was expressed in minutes. Based on previous work in our laboratory, an abnormal clearance was < 45 mins which is > 2 SD below the mean.

Patient Classification

Patients were classified into three groups according to the presence and type of abnormality as follows: Group 1–clinically significant ILD if they had: (1) Symptoms or signs unexplained by another disease; and (2) Radiological (CXR or HRCT) evidence of ILD; and (3) Restrictive lung physiology or abnormal BAL.

Group 2–abnormalities compatible with ILD but no clinically significant ILD if they had: (1) Restrictive lung physiology; or (2) Radiological (CXR or HRCT) evidence of ILD; or (3) Abnormal BAL.

Group 3–No abnormality compatible with ILD if none of the abnormalities described above were present.

The significance of rapid 99mTc-DTPA clearance in the absence of any abnormality in other tests is not clear. In this situation (one patient), this was not included as an abnormality compatible with ILD.

Statistical Analysis

For two group comparisons, where the characteristics of the groups with and without pulmonary abnormalities were compared, a Student t-test (parametric) or Wilcoxon rank sum test (nonparametric) for continuous variables and chi-square analysis for categoric variables were used. A correction for multiple testing (Tukey) was applied. To determine parameters which were associated with abnormalities in lung physiology, univariate regression analysis was used with Hb adjusted-Dl CO (dependent variable) and other continuous variables (age, % neutrophils, % eosinophils, and % lymphocytes in BAL, 99mTc-DTPA clearance time, ESR, CRP, and Ritchie index). Potentially significant parameters were then tested for possible interrelationship by multivariate regression analysis which included the non continuous variables, gender, presence of active and/or severe rheumatoid disease, and use of methotrexate. All analyses were performed with SPSS software package (SPSS, Chicago, IL).

Presence of Disease

Of the 36 patients studied, 21 (58%) had an abnormality in one or more investigation that was consistent with ILD. The number of abnormalities and tests in which they were detected in individual patients is summarized in Table 2. This demonstrates a wide range of patterns of abnormalities with no consistent correlation between tests for the presence of abnormalities. Changes consistent with ILD were found by physiology in eight (22%), CXR in two (6%), HRCT in 12 (33%), BAL in 15 of 29 (52%), and DTPA scan in 4 of 27 (15%, nonsmokers only) subjects. No occupational or other risk factors for the development of ILD, other than RA, were present in any patient.

Table 2. NUMBER OF ABNORMALITIES COMPATIBLE WITH ILD ON SCREENING TESTS

Number of Patients (total = 36)11121127311114
Abnormalities*
 PhysiologyXXXXXX
 CXRXX
 HRCTXXXXXXXX
 BALXXXXXXXX
99mTc-DTPAXXXX

*X = abnormality detected; — = no abnormality.

Lung Physiology

Lung physiology compatible with ILD was present in 22% of the total group. Results of TLC, FVC, and Hb adjusted-Dl CO (% predicted) are shown in Figure 1. Although only seven (19%) had a Hb adjusted-Dl CO < 75% predicted in the absence of airflow limitation (Table 4), if a less stringent cutoff of < 80% is taken, 12 (33%) had abnormal values. Desaturation on exercise was present in three (8%). TLC or FVC <  80% predicted in the absence of airflow obstruction was present in four (11%) and was always associated with a Hb adjusted-Dl CO < 80% of predicted.

Table 4. ABNORMALITIES DETECTED COMPATIBLE WITH ILD*

Patient Groups
Group 1 (n = 5)Group 2 (n = 16 )Group 3 (n = 15)Total (n = 36 )
Signs and symptoms5 (100%) 0 (0%)0 (0%) 5 (14%)
Dl CO < 75% pred3 (60%) 4 (25%)0 (0%) 7 (19%)
Dl CO < 80% pred4 (80%) 7 (44%)1 (7%)12 (33%)
Total compatible with ILD 4 (80%) 4 (25%)0 (0%)19 (53%)
CXR - Interstitial pattern2 (40%) 0 (0%)0 (0%) 2 (6%)
HRCT
 Ground glass3 (60%) 7 (44%)10 (28%)
 Thickened septal/   nonseptal lines5 (100%) 4 (27%) 9 (25%)
 Total abnormal5 (100%) 7 (44%)0 (0%)12 (33%)
BAL alveolitis
 Number performedn = 5n = 13n = 11n = 29
 Neutrophilic alone2 (40%) 3 (23%) 5 (17%)
 Eosinophil alone0 (0%) 2 (15%) 2 (7%)
 Lymphocytic alone1 (20%) 0 (0%) 1 (3%)
 Mixed2 (40%) 5 (38%) 7 (14%)
 Total abnormal5 (100%)10 (77%)0%15 (52%)
DTPA
 Number performedn = 4n = 14n = 9n = 27
 < 45 min 2 (50%) 1 (7%)1 (11%) 4 (15%)

*Data expressed as frequency (% of subgroup).

Group 1 = Clinical significant ILD; Group 2 = Abnormalities without significant ILD; Group 3 = no ILD.

Five further scans < 45 min but confounded by smoking. Of abnormal scans in nonsmokers, one of four occurred in Group 3 (no evidence of ILD).

Defined by Dl CO < 75% predicted and/or desaturation on exercise.

Radiology

As has been previously demonstrated, the CXR was insensitive, detecting abnormalities compatible with ILD in only two patients (6%) whereas the HRCT detected abnormalities in 12 (33%). Additionally, emphysema was seen on the CXR in two (6%) and on the HRCT in five (14%) (Table 3). Emphysema and ILD coexisted in two. The most common abnormalities consistent with ILD detected were ground-glass attenuation (28% of patients) and thickened nonseptal (25%) or septal (14%) lines.

Table 3. ABNORMALITIES DETECTED ON THORACIC HRCT*

AbnormalitiesTotal (n = 36)
Compatible with ILD
 Thickened nonseptal lines 9 (25%)
 Thickened septal lines 5 (14%)
 Ground glass attenuation10 (28%)
 Honeycombing2 (6%)
 Bronchiectasis2 (6%)
 Traction bronchiectasis2 (6%)
Other abnormalities
 Air space consolidation1 (3%)
 Emphysema 5 (14%)
Total abnormaln = 16
 Compatible with ILD alone 9 (25%)
 Compatible with emphysema3 (8%)
 Co-existent ILD and emphysema2 (6%)

*Data presented as frequency (% of total).

Bronchoalveolar Lavage

Of the 29 patients who underwent a BAL, 15 (52%) had an alveolitis, with 12 (41%), seven (24%), and seven (24%) having elevated neutrophils, eosinophils and lymphocytes, respectively. In seven patients there was elevation of more than one cell type (mixed alveolitis). The degree of abnormality was highly variable as shown by the distribution of individual results (Figure 2).

99mTc-DTPA Clearance Scan

A 99mTc-DTPA clearance scan was performed in 35 patients; eight of them were smoking during the assessment and, although asked to refrain for the 3 d prior to the scan, compliance with this could not be independently assessed. In the 27 nonsmokers, clearance was abnormally rapid in four (Figure 3). In the eight smokers, the values ranged from very abnormal to normal (range 19–74 min, normal > 45 min), which may reflect either the timing of the cigarette prior to the scan or actual underlying lung disease.

Other Respiratory Disease

Other respiratory problems detected included airflow obstruction (physiology, 14%) and emphysema (HRCT, 14%). No pleural effusions, pleural disease nor rheumatoid nodules were seen radiologically.

Classification of Clinical Subgroups

Patients were subdivided into three groups according to the presence and type of abnormality as previously defined (Table 4). Five (14%) had clinically significant ILD (Group 1), 16 (44%) had abnormalities compatible with ILD but no clinically significant ILD (Group 2), and 15 (42%) did not have any abnormality compatible with ILD (Group 3). Of those in Group 1, all had symptoms (dry cough or dyspnea) or signs (fine crackles) and evidence of ILD on HRCT. Additionally four had an alveolitis detected on BAL. Interestingly, two did not have a Hb-adjusted Dl CO < 75% predicted at the time of testing although one patient desaturated with exercise and had a Hb-adjusted Dl CO < 80%. Three had abnormally rapid 99mTc-DTPA clearance times (Table 4) although one patient was currently smoking.

In 16 (44%, Group 2), there were abnormalities compatible with ILD in between 1–3 investigations but no clinically significant ILD as defined was present. In many patients, these changes were mild. The abnormalities compatible with ILD consisted of abnormal lung physiology in four, HRCT in seven, and an alveolitis in 10 (Table 4). Of the 16 in Group 2, 11 had an abnormality in one investigation, four in two investigations, and one in three investigations. Fifteen (47%, Group 3) had no abnormalities compatible with ILD on any test.

In comparing the results in the three clinically defined groups, lung physiology (% predicted) was significantly more abnormal in Group 1 than either Group 2 or 3 for Hb-adjusted Dl CO (Group 1, 69% ± 5.7; Group 2, 87% ± 3.7; Group 3, 86% ± 3.5; p < 0.05, Student t-test), FVC (Group 1, 91% ± 8.1; Group 2, 111% ± 3.8; Group 3, 113% ± 4.2; p < 0.05), or TLC (Group 1, 83% ± 8.6; Group 2, 103% ± 3.5; Group 3, 105% ± 3.1; p < 0.02).

HRCT was abnormal in all five with clinically significant ILD (Group 1), and changes were radiologically severe with evidence of honeycombing in two (Table 5). Of the seven patients in Group 2 with abnormal HRCT scans, in three it was the only abnormality on any test. Most of the abnormalities detected in Group 2 were mild (Table 5).

Table 5. SEVERITY OF ABNORMALITIES COMPATIBLE WITH ILD*

Patient Groups
Group 1 (n = 5)Group 2 (n = 16 )
Fibrosis
 Mild3 (60%)4 (25%)
 Moderate
 Severe (honeycombing)2 (40%)
 Total5 (100%)4 (25%)
Ground glass attenuation
 Mild2 (40%)6 (38%)
 Moderate1 (20%)1 (6%)
 Severe
 Total3 (60%)7 (44%)

*Data presented as frequency (% of subgroups).

Group 1 = clinically significant ILD; Group 2 = abnormalities but not significant ILD.

BAL demonstrating an increased proportion of inflammatory cells compatible with ILD was present in 100% (5/5) of those in Group 1 and 77% (10/13) in Group 2 (Table 4). For seven of 10 patients with an abnormal BAL in Group 2, this was the sole abnormality on all tests (Table 2) and ranged from a minimal to marked abnormality, one patient having 79% lymphocytes comprising the BAL cells. Patients with clinically significant ILD in Group 1 had a higher proportion of neutrophils and eosinophils than the other two groups. The values (median ± range) in each group are as follows; Neutrophils: Group 1, 5% (2–10); Group 2, 3% (0–10); Group 3, 2% (0–3.4); (p < 0.05, Group 1 versus Groups 2 or 3, Wilcoxon Rank Sum); Eosinophils: Group 1, 1.5% (0–32); Group 2, 0% (0–3); Group 3, 0% (0–2.5); (p = 0.06, Group 1 versus Group 2; p < 0.001, Group 1 versus Group 3, Wilcoxon Rank Sum). Group 3 had a lower proportion of lymphocytes in comparison with the other groups (lymphocytes: Group 1, 17% (2.5– 38); Group 2, 7% (1.8–79); Group 3, 2.9% (1-10); p = 0.06, Group 1 versus Group 3; p < 0.03, Group 2 versus Group 3, Wilcoxon Rank sum).

For the 99Mtc-DTPA clearance scan, clearance times in Groups 1, 2, and 3 were (mean ± SD) 54 min ± 24, 72 min ± 12, and 72 min ± 27, respectively, which were not statistically significantly different (Group 1 versus Groups 2 and 3, p = 0.22).

Predictors of Disease

The patients with clinically significant ILD (Group 1, n = 5) were compared with those without clinically significant disease (Groups 2 and 3 separately and together) to determine the characteristics that predicted the presence of clinically significant ILD (Table 6). Only male gender was associated with clinically significant disease (p < 0.04). There were no significant differences between Groups 2 and 3. To determine which measurements were associated with restrictive lung physiology, regression analysis was performed with Hb adjusted-Dl CO% predicted as the dependent variable. On both univariate and multivariate analysis, only % neutrophils was significantly associated with a low Hb adjusted-Dl CO% predicted (r = −0.34; p = 0.02).

Table 6. PREDICTORS OF CLINICAL ILD AND ABNORMALITIES COMPATIBLE WITH ILD*

Patient Groups
Group 1 (n = 5)Group 2 (n = 16)Group 3 (n = 15)Group 2 + 3 (n = 31)
Age, yrs54.4 ± 2.152.4 ± 3.349.4 ± 3.351.4 ± 3.1
Male gender4 (80%)3 (19%)  4 (27%)  7 (23%)
Smoking status, ever smoked2 (40%)8 (50%)10 (67%)18 (58%)
Positive RF2 (40%)9 (56%)12 (80%)21 (68%)
Active/severe rheumatoid disease 2 (40%)4 (25%) 7 (47%)11 (33%)
Methotrexate2 (40%)9 (56%)10 (67%)19 (61%)

*Data expressed as mean (% of subgroup).

* Unchanged from original.

Group 1 = clinically significant ILD; Group 2 = abnormalities without significant ILD; Group 3 = no evidence of ILD.

p = 0.04, chi square test for Group 1 versus Group 2.

p = 0.06, chi square test for Group 1 versus Group 3.

p = 0.04, chi square test for Group 1 versus Group 2 and 3. No significant difference between groups for other parameters

The frequency of pulmonary abnormalities found in association with RA has been shown to vary widely and is likely to depend on multiple factors, notably the stage in the disease patients are studied, the source of patient referral and the parameters used to define disease. Our study differs from others in that we have reviewed a group of patients who are homogeneous with respect to the duration of RA, all having symptoms for less than 2 yr. In most other studies, RA has been present for longer and for greatly varying periods (21, 22). Second, our study was limited to patients with confirmed RA defined by ARA criteria who were studied irrespective of respiratory symptoms. They were referred mainly from community-based rheumatologists and therefore are more likely to reflect the spectrum of rheumatoid disease in the community rather than a tertiary hospital cohort (5, 19, 21, 22). Finally, we have used a comprehensive group of investigations sensitive for the detection of early disease to detect abnormalities.

We found that pulmonary abnormalities compatible with ILD are present in a significant proportion of patients with recent onset rheumatoid arthritis. Abnormalities in at least one parameter were present in 56% with changes most commonly being present on BAL (52%) and HRCT (33%) and less commonly on physiology (22%) and CXR (6%), the latter frequently used in clinical practice as screening tests.

Additionally, we have categorized patients into three groups based on a combination of the pattern and extent of abnormality, such as would be done in clinical practice. We have defined those in Group 1 as having clinically significant disease based on both accepted cutoffs for defining abnormality and the pattern of abnormality from a group of tests rather than the results from one single test. Although in severe interstitial lung disease abnormalities are usually present in most investigations, the definition of “clinically” versus “not clinically” significant ILD in milder disease is difficult. Thus, Group 2 is likely to contain a mixture of patients who may or may not subsequently develop significant ILD, but who in this cross-sectional portion of the study, could not be differentiated.

Clinically significant ILD was present in 14%. It is difficult to compare the observed prevalence of clinically significant ILD in our study with that reported by others. In most studies RA had been present for long periods than in our study. In those studies however, the prevalence of significant pulmonary disease ranged from 2.5% (22) to 41% (23). In an autopsy series, 35% had pulmonary fibrosis (24) and 10% died of respiratory failure. Of our five patients, one required immediate therapy with immunosuppressive drugs because of life threatening pulmonary disease, in two therapy was seriously considered but withheld as there were contraindications to therapy and in the other two, disease was mild. At this stage (12–24 mo follow-up) none of our patients have progressively deteriorated (data not shown). It has been suggested ILD presents early in the course of RA and occasionally precedes joint symptoms (1-3). Whether a greater proportion of our patients will eventually develop ILD remains to be determined.

We observed a statistically significant association between male gender and the presence of clinically significant ILD, consistent with previous reports (2). The explanation for the male preponderance is unclear but it was not associated with age, smoking history, activity, or severity of rheumatoid disease nor occupational exposure. Additionally, we found no association between rheumatoid disease severity or extra-articular manifestations and the presence of ILD. Although this may reflect the small numbers, the absence of such associations is consistent with previous reports (5, 21). Use of methotrexate in the management of RA did not differ across the groups and was not associated with the level of Hb-adjusted Dl CO. Although in our study, the period of use of methotrexate was short (< 2 yr), previous work has also found that in the absence of acute pneumonitis use of methotrexate is not associated with the development of restrictive physiology or deterioration in gas exchange (25).

When the group as a whole was reviewed, lung physiology was compatible with ILD in 22% of our patients. In surveys that have used similar physiological criteria but where hospital-clinic based patients with RA of varying duration were studied, Popper (23%) (21) and Banks (10.4%) (22) observed a similar frequency of abnormalities whereas Frank (5) detected a higher frequency (40%). As expected, in our study, patients with clinically significant ILD had lower lung volumes and Hb-adjusted Dl CO than the other groups, reflecting the greater severity of the ILD. Of the 31 patients without clinically significant ILD, four had a Hb-adjusted Dl CO < 75%, and a further five had a Hb-adjusted Dl CO < 80%, all with normal lung volumes. Others have shown that reduced transfer factor is the first abnormality seen in CFA (26) and isolated reduction in transfer factor may indicate early disease. It has also been shown that patients with CFA with a better prognosis have a normal FVC (27). Thus the significance of an isolated reduction in transfer factor and normal lung volumes in patients with RA is not known.

As expected, we found the CXR was insensitive in comparison with HRCT for the detection of ILD (28). The CXR did not show changes of ILD in 3 of 5 of those with clinically significant ILD, nor were there abnormalities in any other patient. In the seven patients in Group 2 where HRCT showed changes compatible with ILD, most of the abnormalities were mild suggesting detection of early disease. However, although HRCT may prove to be a useful investigation in new onset RA, detecting early lung disease, the specificity of the changes seen at this early stage is not known. In three subjects, the abnormal HRCT was the only abnormality in any test, three had abnormalities in one other test and one had abnormalities in two tests. It is anticipated that the follow-up component of this study will provide valuable additional information which will hopefully enable us to understand the importance of these early changes.

Contrary to Remy-Jardin's retrospective study (19) we did not find a high prevalence of bronchial abnormalities on HRCT. However their population was different. In 50% of patients the HRCT scan was performed to analyse suspected rheumatoid lung and in 45%, respiratory symptoms were already present. We detected no nodules, pleural disease or effusions. However this may too reflect the early stage of rheumatoid disease.

There was an abnormal lavage profile in 45% of our patients which ranged from minimal to marked. Others have found an abnormal BAL in 14–60% of patients with RA (25, 29, 30). There was a significantly higher proportion of neutrophils and eosinophils in those with clinically significant ILD. Furthermore, a higher proportion of neutrophils was predictive of a lower Hb-adjusted Dl CO. Garcia performed a BAL on 24 consecutive patients with RA and defined three groups; those with clinical ILD who had a predominant neutrophilic alveolitis; those with an abnormal lavage in the absence of clinically significant ILD who had a predominant lymphocytic alveolitis and a third group who by definition had a normal lavage (23). In CFA, patients with a lymphocytic alveolitis tend to have a better prognosis than those with a neutrophilic alveolitis (31). In our study however, lymphocytes were elevated in both groups 1 and 2 in comparison with group 3, although levels varied markedly. In one patient in Group 2 for example, lymphocytes comprised 79% of the cells of the BAL. Whether a higher proportion of lymphocytes suggests a better prognosis is still not clear (32).

99mTc-DTPA clearance times were analysed in 35 subjects but may have been confounded by smoking in 29%. No significant differences were seen which could be due to the small number with clinically significant ILD. Whether normal clearance times in an individual predict stable lung function, as has been shown with scleroderma associated ILD (12) remains to be determined.

We are the first group to have comprehensively used HRCT, BAL, and 99mTc-DTPA scans in a group of patients with early RA and we detected abnormalities compatible with ILD in a greater proportion than detected by clinical assessment, CXR, and lung physiology alone. Only in clinically significant ILD were abnormalities present on most investigations, however, even then, not in all patients was abnormal lung physiology detected. Of the 15 patients with abnormalities but without clinically significant ILD, 10 had abnormalities in only one investigation; in general the changes were mild. Additionally, there appeared to be no hierarchy in the tests as to which was the most sensitive and this lack of concordance suggests that HRCT, BAL, 99mTc-DTPA scanning and lung physiology may provide complementary information about the extent and type of lung disease. The significance of the abnormalities we have detected remains to be determined by a longitudinal study and at this stage, it is difficult to recommend which tests are appropriate for use in screening the new patient with rheumatoid arthritis.

In summary, in this cross-sectional study, we have found clinically significant ILD in 14% of community based patients with RA of less than 2 yr duration. Additionally we found a large group (41%) of asymptomatic patients in whom abnormalities compatible with ILD were present. The natural history of both these groups is not known and therefore the predictive value of these findings is unclear. This will be determined by a longitudinal study of this cohort which will hopefully provide useful information about the investigation and management of lung disease in patients with RA.

The writers thank Iris Pillay and Carol Watson for their assistance in data collection and Dr. Valerie Burke for statistical advice.

Supported by The Arnold Yeldan and Mary Raine Medical Research Foundation and Medical Research Foundation, Royal Perth Hospital.

1. Helmers R., Galvin J., Hunninghake G. W.Pulmonary manifestations associated with rheumatoid arthritis. Chest1001991235238
2. Hunninghake G., Fauci A. S.Pulmonary involvement in the collagen vascular diseases. Am. Rev. Respir. Dis.1191979471496
3. Roschmann R. A., Rothenberg R. J.Pulmonary fibrosis in rheumatoid arthritis. A review of clinical features and therapy. Semin. Arthritis Rheum.161987174185
4. Monson P. R., Hall A.Mortality amongst arthritic patients. J. Chron. Dis.291976459467
5. Frank S. T., Weg J. G., Harkleroad L. E., Fitch R. F.Pulmonary dysfunction in rheumatoid disease. Chest6319732734
6. Wallaert B., Dugas M., Dansin E., Perez T., Marquette C. H., Ramon P., Tonnel A. B., Voisin C.Subclinical alveolitis in immunological disorders: transition between health and disease? Eur. Resp. J.3199012061216
7. Mathieson J. R., Mayo J. R., Staples C. A., Müller N. L.Chronic diffuse infiltrative lung disease: comparison of diagnostic accuracy of CT and chest radiography. Radiology1711989111116
8. Strickland B., Strickland N. H.The value of high definition narrow section computer tomography in fibrosing alveolitis. Clinical Radiology391988589594
9. O'Brodovich H., Coates G.Pulmonary clearance of 99mTc-DTPA: a non invasive assessement of epithelial integrity. Lung1661987116
10. Staub N. C., Hyde R. W., Crandall E.NHLBI Workshop Summary. Workshop techniques to evaluate lung alveolar—microvascular Injury. Am. Rev. Respir. Dis.141199010711077
11. Harrison N. K., Glanville A. R., Strickland B., Haslam P. L., Corrin B., Addes B. J., Lawrence R., Millar A. B., Black C. M., Turner-Warwick M.Pulmonary involvement in systemic sclerosis: the detection of early changes by thin section CT scan, bronchoalveolar lavage and 99mTc-DTPA clearance. Respir. Med.881989403414
12. Wells A. U., Hansell D. M., Harrison N. K., Lawrence R., Black C. M., duBois R. M.Clearance of inhaled 99mTc-DTPA predicts the course of fibrosing alveolitis. Eur. Respir. J.61993797802
13. Arnett F. C., Edworthy S. M., Bloch D. A., McShane D. J., Friese J. F., Cooper N. S., Healy L. A., Caplan S. R., Liang M. H., Luthra H. S., Medsger T. A., Mitchell D. M., Neustadt D. H., Pinals R. S., Schaller J. G., Sharp J. T., Wilder R. L., Hunder G. G.The American Rheumatism Association revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum.321988315324
14. Comstock G. W., Tockman M. S., Helsing K. J., Hennesy K. M.Standardised respiratory questionaires: comparison of the old with the new. Am. Rev. Respir. Dis.11919794553
15. Ritchie D. M., Boyle J. A., McInnes J. M., Jasani M. K., Dalakos T. G., Greveson P., Buchanan W. W.Clinical studies with an articular index for the assessment of joint tenderness in patients with rheumatoid arthritis. Q. J. Med.371968319406
16. Felson D. T., Anderson J. J., Boers M., Bombardier C., Chernoff M., Fried B., Furst D., Goldsmith C., Kieszaks S., Lightfoot R., Paulus H., Tugwell P., Weinblatt M., Widmark R., Williams H. J., Woolfe F.The American College of Rheumatology preliminary core set of disease activity measures for rheumatoid arthritis clinical trials. Arthritis Rheum.361993729740
17. Cotes, J. E. 1993. Lung Function: Assessment and Application in Medicine, 5th ed. Blackwell, Oxford.
18. Butland R. J. A., Pang J., Gross E. R., Woodcock A. A., Geddes D. M.Two-six-and twelve-minute walking tests in respiratory disease. B.M.J.284198216071608
19. Remy-Jardin M., Remy J., Cortet B., Mauri F., Delcombre B.Lung changes in rheumatoid arthritis: CT findings. Radiology1931994375382
20. Strumpf I. J., Feld M. K., Cornelius M. J., Keogh B. A., Crystal A. G.Safety of fiberoptic bronchoalveolar lavage in evaluation of intestitial lung disease. Chest801981268271
21. Popper M. S., Bogdonoff M. L., Hughes R. L.Interstitial rheumatoid lung disease: a reassessment and review of the literature. Chest621972243250
22. Banks, J., C. Banks, B. Cheong, V. Umachandran, A. P. Smith, J. D. Jessop, and M. H. Pritchard. 1992. An epidemiological and clinical investigation of pulmonary function and respiratory function and respiratory symptoms in patients with rheumatoid arthritis. Q. J. Med. 307– 308:795–806.
23. Garcia J. G. N., Parham N., Killam D., Garcia P. L., Keogh B. A.Bronchoalveolar lavage fluid evaluation in rheumatoid arthritis. Am. Rev. Respir. Dis.1331986450454
24. Suzuki A., Ohosone Y., Obana M., Mita A., Matsuoka Y., Irimajira S., Fukuda J.Cause of death in 81 autopsied patients with rheumatoid arthritis. J. Rheumatol.2119943336
25. Dayton C. S., Schwartz D. A., Sprince N. L., Yagla S. J., Davis C. S., Koehnke R.  K., Furst D. E., Hunninghake G. W.Low dose methotrexate may cause air trapping in patients with rheumatoid arthritis. Am. J. Respir. Crit. Care Med.151199511891193
26. Cherniack, R. M. 1993. Physiological alteration in intestitial lung disease. In M. I. Schwarz, and T. E. King, editors. Interstitial Lung Disease, 2nd ed. Mosby, St. Louis, MO. 79–90.
27. Rudd R. M., Haslem P. L., Turner-Warwick M.Cryptogenic fibrosing alveolitis. Relationships of pulmonary physiology and bronchoalveolar lavage to response to treatment and prognosis. Am. Rev. Respir. Dis.124198118
28. Fujii M., Adachi S., Tanaka K., Shimizu T., Hirota S., Sako M., Kano M., Imura S.A comparative study with CT and plain Xray film on the diagnosis of interstitial pulmonary disease of patients with rheumatoid arthritis. Jpn. J. Clin. Radiol.34198999106
29. Gilligan D. M., O'Connor C. M., Ward K., Moloney D., Breshnihan B., Fitzgerald M. X.Bronchoalveolar lavage in patients with mild and severe rheumatoid lung disease. Thorax451990591596
30. Popp W., Ritschka L., Scherak O., Braun O., Kolarz G., Rauscher H., Zwick H.Bronchoalveolar lavage in rheumatoid arthritis and secondary Sjögrens syndrome. Lung1681990221231
31. Haslam P. L., Turton C., Heard B., Lukoszek A., Collins J. V., Salsburg A. J., Turner-Warwick M.Bronchoalveolar lavage in pulmonary fibrosis: comparison of cells obtained with lung biopsy and clinical features. Thorax351980918
32. Wallaert B., Hatron P. Y., Grossbois J. M., Tonnel A. B., Devulder B., Voisin C.Subclinical pulmonary involvement in collagen vascular diseases assessed by bronchoalveolar lavage. Relationship between alveolitis and subsequent changes in lung function. Am. Rev. Respir. Dis.1331986574580
Correspondence and requests for reprints should be addressed to Dr. Fiona R. Lake, University Department of Medicine, Medical Research Foundation Building, Level 4, Rear 50 Murray Street, Perth, Western Australia 6001. E-mail:

Related

No related items
American Journal of Respiratory and Critical Care Medicine
156
2

Click to see any corrections or updates and to confirm this is the authentic version of record