Annals ofthe Rheumatic Diseases 1995; 54: 815-819 815

Use of high resolution computed tomography of Ann Rheum Dis: first published as 10.1136/ard.54.10.815 on 1 October 1995. Downloaded from the lungs in patients with

Bernard Cortet, Rene-Marc Flipo, Martine Remy-Jardin, Pascal Coquerelle, Bernard Duquesnoy, Jacques Remy, Bernard Delcambre

Abstract .2The prevalence of lung mani- Objective-To assess the usefulness of festations in RA is not known with accuracy, high resolution computed tomography and varies according to the method of diag- (HRCT) of the lungs in patients with nosis. The chest radiograph, for example, may rheumatoid arthritis (RA) with and with- be abnormal in 2-5% ofpatients with RA,2 but out respiratory symptoms. plain chest radiography is not sensitive enough Patients and methods-Eighty eight RA for the diagnosis of lung involvement. High patients with a mean duration of disease resolution computed tomography (HRCT) is a 12 (SD 8) years were evaluated. Eleven non-invasive method of assessing interstitial patients were excluded because of pre- lung disease (ILD), in particular, which has vious exposure to silica. The 77 remaining recently been shown to be useful in systemic patients formed two groups according sclerosis.3" The use of HRCT in RA is poorly to the absence (group I, n = 38) or the documented, but the technique appears to be presence (group II, n = 39) of chronic useful when there is suspected clinical and respiratory symptoms. A control group radiological ILD.5 6 The aim of the present consisted of 51 non-smoking, healthy study was to assess the usefulness of HRCT in patients. RA patients with and without respiratory Results-The most frequent abnormalities symptoms, compared with a control group of observed in the 77 RA patients were non-smoking volunteers. bronchiectasis or bronchiolectasis (n = 23, 30%), pulmonary nodules (n = 17, 22%), subpleural micronodules or pseudoplaques Patients and methods (n = 13, 17%/), ground glass opacities PATIENTS (n =11, 14%), and honeycombing (n = 8, Eighty eight patients fulfilling the revised 10%). Bronchiectasis or bronchiolectasis criteria for RA of the American Rheumatism

(p = 0-012), rounded opacities (p = 0'016), Association were reviewed.7 All had undergone http://ard.bmj.com/ ground glass attenuation (p = 0.004), and HRCT examination of the thorax between honeycombing (p = 0.002) were found more 1987 and 1993. Among them, 11 patients who often in RA group II (with respiratory had been exposed to silica were excluded symptoms) than in group I (no respiratory because of possible confusion between rheu- symptoms). Non-linear septal opacities matoid and pneumoconiotic lung lesions. were more frequent in group I than in the HRCT had been performed in the remaining 77 because of of associated

suspicion on September 27, 2021 by guest. Protected copyright. Department of control group, but other HRCT findings patients , did not differ statistically significantly pulmonary disease, on the grounds of pul- CHRU Lille, between group I and the control group. monary symptoms or systematic evaluation of H6pital B, Conclusion-Bronchiectasis may be a lung changes. In the latter case, the patients 2 Avenue Oscar Lambret, characteristic lung change in RA patients. were selected consecutively from one depart- 59037, Lille Cedex, Abnormalities on HRCT are less fre- ment ofrheumatology, the systematic evaluation France quently observed in the absence of respir- of their lung changes was approved by the B Cortet R-M Flipo atory symptoms than in the presence of Hospital Ethics Committee, and all patients P Coquerelle such symptoms (29% versus 69%). gave written informed consent to participate. B Duquesnoy The 77 patients comprised 51 women and 26 B Delcambre (Ann Rheum Dis 1995; 54: 815-819) men, mean age 57 (SD 10) years (range 36-79). Department of one were non-smokers who had never Radiology, Seventy CHRU Lille, smoked and six were current smokers, mean Hopital Calmette, Rheumatoid arthritis (RA) is a common cigarette consumption 24 (20) packs/year Boulevard Jules chronic inflammatory disease affecting about (range 1-60 packs). The mean duration of RA Leclerc, 59037, Lille Cedex, 1% of the white population. Extra-articular at the time of chest HRCT was 12 (8) years. France manifestations may involve, in particular, the Subcutaneous rheumatoid nodules were noted M Remy-Jardin skin (rheumatoid nodules), eyes, heart, and in 17 patients (22%) and rheumatoid factor in J Remy lungs. Since the first description oflung disease 44 (57%). Nine patients (12%) were suffering Correspondence to: Bernard Cortet, associated with RA by Ellman and Ball in from Sjogren's syndrome (presence of sicca Department of 1948,1 several forms ofpleuropulmonary disease syndrome and positive labial salivary gland Rheumatology, CHRU Lille, H6pital B, have been established in RA: / biopsy (focus score> 1, stage III or IV on 2 Avenue Oscar Lambret, pleuritis, rheumatoid lung nodules, Caplan's Chisolm's classification8)). Functional capacity 59037, bille Cedex, France. syndrome, fibrosing alveolitis, lymphoid hyper- was evaluated by Steinbrocker classification:9 Accepted for publication plasia with germinal centres, pulmonary 12 were class I, 28 class II, and 37 class III. 26 June 1995 , constrictive bronchiolitis, and At the time of the HRCT examination, 63 of 816 Cortet, Flipo, Remy-Jardin, et al

the 77 patients were receiving decision. Multiple radiological criteria for the

or second line drugs: glucocorticoids (n = 49); diagnosis of lung involvement were assessed, Ann Rheum Dis: first published as 10.1136/ard.54.10.815 on 1 October 1995. Downloaded from (n = 24); tiopronine (n= 10); including site and severity. The major abnor- sodium aurothiomalate (n = 7); sulphasalazine malities screened for were: rounded opacities (n = 7); hydroxychloroquine (n = 1). The patients (parenchymal micronodules = rounded lesions were allocated to two groups according to less than 3 mm in diameter; nodules = rounded the absence (group I, n = 38) or presence lesions greater than 3 mm in diameter; (group II, n = 39) of respiratory symptoms subpleural micronodules = areas of hyper- such as (in the morning or all day), attenuation less than 3 mm in diameter); sputum production (in the morning or all day), ground glass attenuation (bronchi and vessels or dyspnoea. visible); dependent areas of attenuation A control group comprised 51 healthy (bronchial walls and vessels obscured); septal subjects who had never smoked and who lines and non-septal lines; honeycombing were previously enrolled in a prospective study (areas of cystic spaces with thickened walls); assessing HRCT of the lungs in healthy adult bronchiectasis (abnormal visualisation of volunteers.'0 There were 34 women and 17 proximal airways); bronchiolectasis (abnormal men (mean age 33 (8) years), all urban visualisation of airways in peripheral locations: dwellers and recruited from workers in our dilated bronchiolar division visualised along hospital. None had a past history of lung their length when horizontal, or peripheral disease and, in particular, none had previously signet ring signs when coursing in a vertical undergone chest surgery or suffered from a direction; architectural distortion; emphysema respiratory illness such as bronchiolitis of any characterised by areas of decreased attenuation origin in previous years or in infancy. Some and disruption of the vascular pattern. control patients had respiratory symptoms: Statistical comparisons between the different cough (n = 4; morning, n = 3; all day, n = 1), groups of patients were made using the x2 test sputum production in the morning (n = 2), and with Yates' correction if necessary. Correlation dyspnoea occurring after strenuous activity coefficients were calculated by linear regression such as climbing three flights of stairs, heavy analysis. housework, or walking more than one mile on level ground (n = 8). Results The table summarises the results. METHODS HRCT ofthe thorax was performed with either an Elscint 2400 (Hackensak, NJ) or a Siemens HRCT IN RA PATIENTS Somatom Plus (Erlangen, Germany). Serial HRCT was abnormal in 38 (49%) of the 77 RA slices were taken through the chest, each 1 mm patients- 1 from group I (29%) (no respir- in and 10 mm apart. factors width Technical atory symptoms) and 27 from group II (69%) http://ard.bmj.com/ were 130 kV and 420 mA (Elscint CT unit) or (with respiratory symptoms) (p < 0 001). 137 kV and 255 mA (Siemens CT unit). The most frequent abnormality on HRCT of Images were reconstructed using a high spatial the lungs was bronchiectasis or bronchiolectasis frequency algorithm for parenchymal analysis (figure), observed in 23 patients (30%)-three and a standard algorithm for mediastinal patients in group I (8%), and 20 patients in evaluation. HRCT studies were performed at group II (51%) (p = 0 0 12)-and consisted of suspended end inspiratory volume with one honeycombing in seven ofthem. Among the 16 second (Siemens) or two seconds (Elscint) with bronchiectasis or bronchiolectasis but no on September 27, 2021 by guest. Protected copyright. scan time with patients in the supine position. honeycombing, 14 patients were non-smokers In cases of limited joint mobility, especially who had never smoked. Specific evaluation involving the shoulders, HRCT scan was per- of these 14 non-smoking patients showed formed with the patient's arms positioned bronchiectasis or bronchiolectasis associated alongside the body, without any effect on the with diffuse bronchial thickening in all but image quality. The HRCT examinations were one; in four, airway changes were the sole interpreted by two radiologists blind to the abnormality on HRCT examination, and in clinical history, who reached a concensus 10 the airway changes were associated with parenchymal micronodules (n = 6), dependent High resolution computed tomography (HRCT) findings in patients with rheumatoid areas of attenuation (n = 4), subpleural micro- arthritis (RA) without (group I) and with (group II) respiratory symptoms nodules or pseudoplaques (n = 4) and emphy- HRCTfinding Group I Group II Control group sema (n = 2). (n = 38) (n = 39) (n = 51) Rounded opacities were observed in 17 patients Bronchial wall thickening 2 (5) 11 (28) 9 (18) (22%). Three types were identified according Bronchiectasis or bronchiolectasis 3 (8) 20 (51)* 0 Rounded opacities 4 (11) 13 (33)* 11(22) to their size and location: parenchymal micro- Septal lines 0 0 4 (8) nodules (n = 6), nodules (n = 3), and subpleural Non-septal linear opacities 5 (13) 9 (23) 0* = Ground glass attenuation 1 (3) 10 (26)** 0 micronodules or pseudoplaques (n 13). Honeycombing 0 8 (23)** 0 In the absence of respiratory symptoms Dependant areas of attenuation 1 (3) 4 (10) 6 (12) in Distortion 1 (3) 4 (10) 0 (group I), rounded opacities were observed Emphysema 2 (5) 2 (5) 0 four patients (111%), peripheral parenchymal Enlarged lymph nodes 2 (5) 5 (13) 0 in and micro- Enlarged pulmonary arteries 0 1 (3) 0 micronodules two, subpleural Pleural abnormalities 3 (8) 9 (23) 0 nodules in two. In group II (with respiratory 13 had rounded n = Number of patients with HRCT findings. Number in parentheses are percentage. *p < 0-05, symptoms), patients (33%) **p < 0-01 compared with group I. opacities: parenchymal micronodules in four, High resolution computed tomography ofthe lungs in patients with RA 817

.'4112121211 HRCT IN THE CONTROL GROUP

Abnormalities were evident on HRCT in 22 Ann Rheum Dis: first published as 10.1136/ard.54.10.815 on 1 October 1995. Downloaded from control patients (43%). The most frequent abnormality was subpleural micronodules (n = 11, 22%); others were: bronchial wall thickening without any other bronchial r abnormality (n 9, 18%), dependent areas of attenuation (n =6, 12%) and septal lines (n = 4, 8%). A significant correlation between each respiratory symptom and HRCT findings was found only for bronchial wall thickening - ~ (p - 0 009). C5 ti-n: COMPARISON OF HRCT IN RA PATIENTS AND CONTROLS Bronchiectasis or bronchiolectasis (p < 0001), septal lines (p = 0 05), non-septal linear opacities (p < 0 01), ground glass attenuation (p < 0 01), honeycombing (p < 0-05), and pleural abnor- malities (p < 0O01) occurred more frequently in RA patients than in the control group. Non-septal linear opacities was the sole abnor- 11 63 mality observed more frequently on HRCT in C i RA patients without respiratory symptoms Bronchiectasis: cylindral dilatation ofperipheral bronchi in the lower lobes and in tihe (group I) than in control patients (p < 001). lingula, with mild bronchial wall thickening in the posterior segment of the right k9wer lobe. Discussion parenchymal nodules in three, and sut)pleural The most frequent abnormality depicted on micronodules in 11. This difference betwteen the HCRT in our study was bronchiectasis or groups was statistically significant (p = (D-016). bronchiolectasis-found in 23 patients (30%). Non-septal linear opacities were observed in While seven of these patients had lung changes 14 patients (18%) and were always loc:ated in (honeycombing) considered as indirect signs of peripheral lung tissue. Such abnornmalities , the remaining 16 RA were the sole HRCT findings in six Ipatients patients with abnormalities in the absence of and were associated with honeycomlbing in features of lung fibrosis included 14 non- eight; they were exclusively or predomlinantly smokers who had never smoked. This strongly http://ard.bmj.com/ in the lower lung zones in all cases. suggests that neither smoking nor an increased Five patients (13%) had non-septa1 linear susceptibility to its effects were the main opacities in the absence of respiratory symp- causes for the development of bronchiectasis toms (group I) and nine (23%) wiith this or bronchiolectasis. In addition, we failed feature had respiratory symptoms (group II) to observe bronchiectasis/bronchiolectasis in (difference between groups not significant). our control group of non-smokers, among Ground glass attenuation representin.g early whom bronchial abnormalities consisted active alveolitis was found in 11 patientss (14%) solely of bronchial wall thickening which could on September 27, 2021 by guest. Protected copyright. -one in group I (3%) and 10 in giroup II be explained by occupational exposure to (10%) (p = 0 004). It was usually bilate~ral and pollutants or other environmental factors symmetrical. (urban dwellers). An association between Honeycombing highly suggestive ofpuiImonary bronchiectasis and RA is well known," but it fibrosis was detected in eight patients (10%) remains unclear if bronchiectasis occurs in and comprised a microcystic pattern (n = 5), or severe RA and if it is accompanied by other a combination of microcystic and macirocystic extra-articular manifestations.'2 13 McDonagh, honeycombing (n = 3). Honeycombirng was et al, using HCRT, found bronchiectasis in always bilateral and asymmetrical, aind was four of 20 RA patients (20%), but none was associated with bronchiectasis or br(onchio- suffering from clinical symptoms of respiratory lectasis in seven patients, ground1 glass disease.6 In the same study, the authors attenuation in seven, non-septal linear observed bronchiectasis in 30% of the patients opacities in all cases, and pleural thickening or with interstitial lung disease. On the basis of pleural effusion in five. Evidence of honey- pulmonary function tests, it has been shown combing was always associated with pIresence that airway disease may be the commonest of respiratory symptoms (significant difEference form of RA lung involvement;'4 in other RA between groups (p = 0.002)). patients, low values of forced expiratory The presence of respiratory symptoms (cough, volume in one second (FEVI), forced vital dyspnoea or sputum production) was signifi- capacity (FVC), FEV1/FVC, and a significant cantly correlated with the following HRCT high prevalence of bronchial reactivity to findings: rounded opacities (p = 0-016), inhaled methacholine have been reported. 15 ground glass attenuation (p = 0 004), honey- The most common explanation is that combing (p = 0-002), and bronchiect:asis or bronchiectasis in RA patients results from bronchiolectasis (p < 0-001). frequent respiratory tract infections.'6 The 818 Cortet, Flipo, Remy-jardin, et al

predominant location of bronchiectasis or toms likewise differed between these studies.

bronchiolectasis in the lower zones ofthe lungs Among our 38 patients with respiratory symp- Ann Rheum Dis: first published as 10.1136/ard.54.10.815 on 1 October 1995. Downloaded from supports this hypothesis. In addition, alpha-1 toms, abnormalities on HRCT were observed antitrypsin deficiency has been suggested to in 29 (69%), while only 11 (29%) of our 39 increase the risk of lung involvement in RA.'7 patients without respiratory symptoms had Rounded opacities, found in 22% of our HRCT abnormalities. In contrast, McDonagh patients, were the second most frequently et al observed on HRCT a prevalence of lung observed abnormality. Their location and size involvement in up to 75% of asymptomatic were consistent with the characteristic sub- rheumatoid patients.6 However, most of these pleural location of rheumatoid nodules.'8 patients were current smokers and the HRCT Pulmonary nodules are usually symptomless, abnormalities probably reflected a greater or contribute little to pulmonary dysfunction; influence of tobacco than of RA. The signifi- however, they can grow in size and tend to cance of the abnormalities we demonstrated rupture into the pleura causing pneumo- in the absence of respiratory symptoms is thorax, hydropneumothorax, or even pyo- unknown, as we did not perform pulmonary pneumothorax, as observed in one of our function tests. While McDonagh's group did patients.19 20 The frequency of parenchymal not find any significant difference on pul- nodules in our patients (three in 77-4%) was monary function testing between RA patients lower than that reported from pathological with clinical evidence ofinterstitial lung disease studies18 and on previous HRCT scan studies,6 and control RA patients, when the pulmonary even though HRCT is able to demonstrate function of the group of 20 patients with ILD lung nodules in the rheumatoid population was compared with that of 10 patients with more frequently than plain radiographs normal HRCT, or nodules or pleural disease (fewer than 1% of rheumatoid nodules are alone, reductions in FEV, and lung carbon estimated to be radiographically detectable21). monoxide transfer factor became significant.6 We observed subpleural micronodules in 17% Apart from the non-septal linear opacities of our RA patients; although subpleural micro- which were more frequent in our RA patients nodules were identified in 22% of our control without respiratory symptoms (group I) than in group, several CT findings suggest their control patients, the frequency of HRCT possible relationship with rheumatoid lung abnormalities was similar in these two groups. disease. Subpleural lesions in RA patients are However, the fact that these groups were not observed in the upper, mid and lower lung age matched should be taken into account zones, whereas in healthy adults subpleural when interpreting this result. Finally, in the micronodules are exclusively located in the absence ofrespiratory symptoms, slight ground upper part of the lungs."' glass opacification representative of active Ground glass opacification representing an alveolitis was demonstrated in only one early stage of active alveolitis was the fourth patient, and the honeycombing considered to most frequent HCRT abnormality we detected indicate lung fibrosis was not observed. (14%). Discovery of such an abnormality on In summary, HRCT is a useful diagnostic http://ard.bmj.com/ HRCT could indicate early aggressive treat- tool in the assessment of lung involvement in ment to prevent irreversible fibrosis. patients with respiratory symptoms and offers Honeycombing was present in eight patients information not readily obtained from plain (10%), always bilateral and asymmetrical, chest radiography. The significance of the with a predominant peripheral and inferior several abnormalities which were observed in distribution. These morphological features are the absence of respiratory symptoms remains in agreement with those previously reported in unclear and requires subsequent longitudinal on September 27, 2021 by guest. Protected copyright. the literature,22 23 and are indistinguishable studies. from pulmonary fibrosis caused by other connective tissue diseases.24 Although inter- stitial lung disease is a well known extra- 1 Ellman P, Ball R E. Rheumatoid disease with joint and pulmonary manifestations. BMJ 1948; 2: 816-20. articular manifestation of RA, its prevalence is 2 Kelly C A. Rheumatoid arthritis: classical lung disease. In: closely related to the diagnostic tool used for Kelly C A, ed. Lung disease in rheumatic disorders. London: Bailliere Tindall, 1993; 1-17. its recognition. Pathological studies have 3 Hansell D M, Kerr I H. The role of high resolution shown interstitial lung disease in up to 80% of computed tomography in the diagnosis of interstitial lung 25 disease. Thorax 1991; 46: 77-85. patients,18 whereas interstitial lesions are 4 Harrison N K, Glanville N R, Strickland B, et al. Pulmonary radiographically identified in fewer than 5%.26 involvement in systemic sclerosis: the detection of early changes by thin section CT, bronchoalveolar lavage Although the aim of this study was not to and 99m Tc-DTPA clearance. Respir Med 1989; 83: compare chest radiography with lung HRCT, 403-14. 5 Fewins H E, McGowan I, Whitehouse G H, Williams J, we are able to confirm the lack of sensitivity of Mallya R. High definition computed tomography in plain chest radiography, as 90% of our patients rheumatoid arthritis associated pulmonary disease. Br J Rheumatol 1991; 30: 214-6. suffering from lung fibrosis had a normal chest 6 McDonagh J, Greaves M, Wright A R, Heycock C, radiograph. Owen J P, Kelly C. High resolution computed tomography was seen in 5% of our RA of the lungs in patients with rheumatoid arthritis and Emphysema interstitial lung disease. Br J Rheumatol 1994; 33: patients, whereas McDonagh et a76 found it 118-22. 7 Amett F C, Edworthy S M, Bloch D A, et al. The American in 20%. Tobacco smoking may explain this Rheumatism Association 1987 revised criteria for the difference, as 80% of their patients were classification of rheumatoid arthritis. Arthritis Rheum 1988;31: 315-24. current smokers, whereas 90% of those in our 8 Chisolm D M, Mason D K. Labial salivary gland biopsy in study were non-smokers and had never smoked. Sjogren's disease. J Clin Pathol 1968; 21: 656-60. 9 Steinbrocker 0, Trager C H, Batterman R C. Therapeutic The prevalence of HRCT findings according criteria in rheumatoid arthritis. J7AMA 1949; 37: to the absence or presence ofrespiratory symp- 393-406. High resolution computed tomography ofthe lungs in patients with RA 819

10 Remy-Jardin M, Remy J, Boulenguez C, Sobaszek A, 18 Yousem S A, Colby T V, Carrington C B. Lung biopsy in Edme J-L, Furon D. Morphologic effects of cigarette rheumatoid arthritis. Am Rev Respir Dis 1985; 131:

smoking on airways and pulmonary parenchyma in 770-7. Ann Rheum Dis: first published as 10.1136/ard.54.10.815 on 1 October 1995. Downloaded from healthy adult volunteers: CT evaluation and correlation 19 Portner M M, Gracie W A. Rheumatoid lung disease with with pulmonary function tests. Radiology 1993; 186: cavitary nodules, pneumothorax and eosinophilia. N Engl 107-15. Med 1966; 275: 697-700. 11 Solanki T, Neville E. Bronchiectasis and rheumatoid 20 McConnoche K, O'Sullivan M, Khalh J F, Purchard M H, disease: is there an association? BrJ Rheumatol 1992; 31: Gibbs A R. Aspergillus colonization of pulmonary 691-3. rheumatoid nodule. Respir Med 1989; 83: 157-60. 12 McMahon M J, Swinson D R, Shettar S, et al. 21 Walker W C, Wright V. Pulmonary lesions and rheumatoid Bronchiectasis and rheumatoid arthritis: a clinical study. arthritis. Medicine (Baltimore) 1968; 47: 501-20. Ann Rheum Dis 1993; 52: 776-9. 22 Steinberg D L, Webb R W. CT appearances of rheumatoid 13 Shadick N A, Fanta C H, Weinblatt M E, Odonnell W, lung disease. ComputAssist Tomogr 1984; 8: 881-4. Coblyn J S. Bronchiectasis. A late feature of severe 23 Fuji M, Adachi S, Shimizu T, Hirota S, Sako M, Kono M. rheumatoid arthritis. Medicine (Baltimore) 1994; 73: Interstitial lung disease in rheumatoid arthritis: 161-70. assessment with high-resolution computed tomography. 14 Geddes D, Webley H, Emerson P. Airway obstruction Thorac Imaging 1993; 8: 54-62. in rheumatoid arthritis. Ann Rheum Dis 1979; 38: 24 Remy-Jardin M, Remy J, Wallaert B, Bataille D, 222-5. Hatron P Y. Pulmonary involvement in progressive 15 Hassan W U, Keaney N P, Holland C D, Kelly C A. systemic sclerosis. Sequential evaluation with CT, Bronchial reactivity and airflow obstruction in pulmonary function tests and bronchoalveolar lavage. rheumatoid arthritis. Ann Rheum Dis 1994; 53: 511-4. Radiology 1993; 188: 499-506. 16 Bamji A, Cooke N. Rheumatoid arthritis and chronic 25 Cervantez-Perez P, Toro-Perez A H, Rodriguez-Jurado P. bronchial suppuration. Scand Rheumatol 1985; 14: Pulmonary involvement in rheumatoid arthritis. JAMA 15-21. 1980; 243: 1715-7. 17 Geddes D M, Webley M, Turton C W, et al. Alpha 1 26 Walker W C, Wright V. Diffuse interstitial pulmonary antitrypsin phenotypes in fibrosing alveolitis and fibrosis and rheumatoid arthritis. Ann Rheum Dis 1969; rheumatoid arthritis. Lancet 1977; 2: 1049-50. 28: 252-9. http://ard.bmj.com/ on September 27, 2021 by guest. Protected copyright.