Clinical Rheumatology

, Volume 26, Issue 2, pp 225–230

Pulmonary involvement in ankylosing spondylitis

Authors

    • Unit of Rheumatology, Department of Internal MedicineState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • Elza Maria F. P. Cerqueira
    • Department of RadiologyState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • Sílvio M. Rezende
    • Unit of Pneumology, Department of Internal MedicineState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • Lucimara Maeda
    • Department of RadiologyState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • Roseneide A. Conde
    • Unit of Rheumatology, Department of Internal MedicineState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • Verônica A. Zanardi
    • Department of RadiologyState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • Manoel Barros Bértolo
    • Unit of Rheumatology, Department of Internal MedicineState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • José Ribeiro de Menezes Neto
    • Department of RadiologyState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
  • Adil M. Samara
    • Unit of Rheumatology, Department of Internal MedicineState University of Campinas Faculty of Medical Sciences (FCM / UNICAMP)
Original article

DOI: 10.1007/s10067-006-0286-2

Cite this article as:
Sampaio-Barros, P.D., Cerqueira, E.M.F.P., Rezende, S.M. et al. Clin Rheumatol (2007) 26: 225. doi:10.1007/s10067-006-0286-2

Abstract

This is a prospective study analyzing 52 asymptomatic, consecutive patients with ankylosing spondylitis (AS), who submitted to a pulmonary investigation that included plain chest radiography, pulmonary function test (PFT), and thoracic high-resolution computed tomography (HRCT). The results were compared according to sex, race, dorsal spine involvement, thoracic diameter, smoking status, and HLA-B27. There were four patients (8%) with an altered plain chest radiograph. PFT presented a restrictive pattern in 52% of the patients. Thoracic HRCT showed abnormalities in 21 patients (40%), predominantly nonspecific linear parenchymal opacities (19%), lymphadenopathy (12%), emphysema (10%), bronchiectasis (8%), and pleural involvement (8%). Linear parenchymal opacities were associated with a smoking history (p=0.026) and dorsal spine involvement (p=0.032). HLA-B27 was not associated with any abnormality. A lower thoracic diameter was observed in patients with dorsal spine involvement (p=0.0001), restrictive pattern at PFT (p=0.023), and linear parenchymal opacities (p=0.015). The study concluded that nonspecific subclinical pulmonary involvement is frequent in AS.

Keywords

Ankylosing spondylitisHigh-resolution computed tomographyLung involvementPulmonary function test

Introduction

Lung involvement in ankylosing spondylitis (AS) was initially described in 1941 [1], but it was considered as an extra-articular manifestation of the disease only since 1965 [2]. In 1977, Rosenow et al. [3], upon analyzing the records of 2,080 AS patients attended at the Mayo Clinic between 1966 and 1975, found 28 patients (1.3%) with pleuropulmonary manifestations, predominating upper lobe fibrosis, frequently asymptomatic and with insidious evolution in patients with long disease duration. Upper lobe cavitation could be a late complication in these patients with secondary fungal or mycobacterial infection [4, 5]. Between the hypothesis to explain this upper lobe fibrosis, it was suggested that a regional disturbance of lung ventilation [6, 7] or the genetic influence of HLA-B27 [8, 9] could be involved.

The pulmonary function tests in AS have disclosed a high prevalence of restrictive defect, characterized by a low forced vital capacity frequently associated to a low thoracic expansibility [1013]. The occurrence of alveolitis in AS is controversial. Early studies did not show alterations in the bronchoalveolar lavage in AS [14, 15], although a subsequent French study, with a larger number of patients and a control population, showed the presence of a subclinical lymphocyte alveolitis not correlated with smoking, medication use, laboratory tests for disease activity, and the presence of restrictive ventilatory dysfunction [16].

Comparing different imaging methods in the evaluation of the lung involvement in AS, Fenlon et al. [17] showed radiographic alterations in 15% of the patients, while the chest high-resolution computed tomography (HRCT) was altered in 69% of the patients, predominating nonspecific findings (interstitial lung disease, bronchiectasis, paraseptal emphysema, mediastinal lymphadenopathy). Turetschek et al. [18], studying the frequency and the distribution of early pulmonary lesions in 25 AS patients with a normal chest X-ray, found abnormalities on thin-section CT scan in 71% of the patients. The most frequent abnormalities were thickening of the interlobular septa, a mild bronchial wall, pleural and linear septal thickening. Senocak et al. [19], analyzing the spectrum of lung parenchyma changes in 20 AS patients, found chest HRCT alterations in 85% of the cases, predominantly emphysema (45%), septal and pleural thickening (45%), nodules (40%) and subpleural band formation (35%). Kiris et al. [20], investigating HRCT findings in 28 early AS patients, found nonspecific abnormalities in 18 patients (64%) with normal chest radiography results. El Maghraoui et al. [21], studying thoracic HRCT in 55 AS patients without respiratory symptoms, found alterations in 4% of the chest X-rays and 53% of the thoracic HRCT, predominating nonspecific interstitial changes and emphysema, with some cases of interstitial lung disease (7%) and upper lobe fibrosis (9%). Souza Jr. et al. [22], studying pulmonary manifestations on inspiratory and expiratory HRCT in 17 AS patients, found abnormalities in 88% of the cases, predominantly airway disease (82%), interstitial abnormalities (65%), and emphysema (35%). Altin et al. [23], analyzing thoracic HRCT in 38 AS patients, observed nonspecific abnormalities in 73%, which is higher in patients with early AS. HRCT has been widely used for the analysis of the lung interstitium and in the morphological characterization of localized and diffuse abnormalities of the lung parenchyma, and it is superior to the radiographic and conventional CT studies [24].

In the present study, the authors analyze the subclinical pulmonary involvement in a series of 52 AS patients.

Materials and methods

Patients

This prospective study included 52 consecutive patients, diagnosed with AS according to the modified New York criteria [25], who were attended at the outpatient clinic of the Hospital de Clínicas of the State University of Campinas between January 1998 and December 1999. All the patients were asymptomatic about lung complaints (dyspnea, chest pain, chronic cough, hemoptysis, cyanosis) and did not fulfill standard criteria for any other disease in the group of spondyloarthropathies or connective tissue diseases. Every patient gave written informed consent before being included in the study.

The patients were submitted to a complete clinical examination, with special emphasis on the spine (Schober test, thoracic diameter, occiput-to-wall distance, hands-floor distance) and lung examination. Patients with a clinical history of inflammatory pain at the dorsal spine, accentuation of dorsal kyphosis and a low thoracic diameter (less than 3 cm) were considered as presenting thoracic AS involvement. HLA-B27 was typed at low resolution using polymerase chain reaction-amplified DNA, hybridized with sequence-specific oligonucleotide primers (One-Lambda, Canoga Park, CA), and HLA-B*27 alleles were typed using a high-resolution HLA-B27 typing kit (Dynal, Oslo, Norway).

Pulmonary investigation

Patients were submitted to plain chest radiography, pulmonary function test (PFT), and thoracic HRCT. The patients underwent all the exams at the same day. Patients were also asked about their smoking status. A Mantoux skin test was performed at the same month of the pulmonary investigation.

Pulmonary function tests

Forty-six patients underwent PFT at the same day of the imaging exams; however, six patients were unable to perform PFT. The main parameters evaluated were forced vital capacity (FVC), total lung capacity (TLC), residual volume (RV) and 1-s forced expiratory volume (FEV1). Observed values were compared with those predicted for age, sex, and height for each patient. Based on these measurements, the following categories were defined: (1) normal, FVC>80% and FEV1/FVC>80%; (2) restrictive defect, FVC<80% and TLC<80%; (3) obstructive defect, FEV1/FVC<70% and RV>100%.

Imaging

All 52 patients underwent plain chest radiography (posteroanterior and lateral). At the same sitting, thin-section, high-resolution chest CT was also done. CT scans were obtained on Somaton AR (Siemens) in all patients. The scanning parameters were 2.0-mm collimation at 10-mm intervals, 3-s scanning time, 140 kVp, 70 mA, reconstructed by means of a high-spatial frequency algorithm from the lung apex to the diaphragm in supine position. Six CT scans in prone position were also obtained (two each in apex, medium, and basal lung regions). The images were obtained at a window level of −700 and width of 1,400.

The results were compared according to sex, race, dorsal spine involvement, thoracic diameter, smoking status, and HLA-B27.

Statistical analysis

Data were analyzed, in cases of association, using the chi-square test (x2) and the Fisher’s exact test. Nonparametric Mann–Whitney test (Wilcoxon test) was used to compare quantitative data. It required p values ≤0.05 as evidence of statistical significance; a statistical trend was considered to be present when 0.05<p≤0.10.

Results

There was predominance of the male sex (81%), Caucasian race (75%), and HLA-B27 positive patients (75%). Among the 39 HLA-B27 positive patients, 36 patients presented HLA-B*2705 and three HLA-B*2702. The mean age was 40.1 years (varying from 18 to 68 years, with a SD of 12.4), and the mean disease duration was 16.2 years (varying from 2 to 53 years, SD 10.5). Dorsal spine involvement was present in 30 patients (58%). Mean thoracic diameter was 1.89 cm (varying from 0.5 to 4.0 cm, SD 0.8, median 1.75 cm). Thirty-three patients (63%) never smoked, 14 (27%) were current smokers, and five (10%) were ex-smokers. As the five ex-smokers had stopped smoking in the last 2 years after smoking for more than 10 years, they were considered smokers for the statistical analysis.

Four patients (8%), who were asymptomatic about lung complaints, presented altered plain chest radiographs: two had nonspecific linear opacities, one had apical opacity, and one had descendent aortic aneurysm. None of these patients referred a previous history of infectious pulmonary disease.

Mantoux skin test was positive in 12 patients (23%). Six patients presented a Mantoux value <10 mm with a negative sputum culture. Four patients presented Mantoux values between 10 and 19 mm with negative sputum culture and denied symptoms or recent contact with tuberculosis. Two patients presented a Mantoux value of 20 mm. One male patient with a negative sputum culture had a family history of tuberculosis (his mother died of tuberculosis 30 years ago), and one female patient presented a positive sputum culture and was treated for tuberculosis for 6 months with a good result. Patients with a positive Mantoux skin test and negative sputum cultures are being followed by the Unit of Pneumology. No one presented a diagnosis of tuberculosis in the 5 years subsequent to the end of the study.

Forty-six patients underwent PFT. Six patients were unable to perform the exam. Twenty-four patients (52%) presented a restrictive pattern, and six patients (13%, all of them smokers) presented a concomitant obstructive pattern; no patient had an isolated obstructive pattern. The mean FVC was 79% predicted (varying from 49 to 131%, SD 16.5), and the mean FEV1 was 87% predicted (varying from 52 to 137%, SD 18.6).

Thoracic HRCT showed abnormalities in 21 patients (40%) (Table 1). Nonspecific parenchymal opacities were the most frequent abnormalities observed at thoracic HRCT (Fig. 1); five of the ten patients with parenchymal opacities were not able to perform the HRCT in prone position due to an extreme difficulty of mobilization of the spine. The patient with the apical opacity shown in the plain chest radiograph presented apical fibrosis at the thoracic HRCT (Fig. 2). Five of the six patients with lymphadenopathy presented calcified lymph nodes. Four of the five patients with emphysema presented paraseptal emphysema (two were current smokers) (Fig. 3), and a smoker patient presented a centrilobular emphysema. Bronchiectasis (Fig. 4) was usually localized, and three of the four patients were current smokers. The pleural involvement, observed in three patients, was generally represented by a mild thickening. Between the three patients with pulmonary nodules, two had calcified granulomas, and the other presented small centrilobular nodules and was referred to the Unit of Pneumology. One patient presented asymptomatic diffuse cystic disease of the lungs. The patient with the descendent aortic aneurysm was referred to the Unit of Vascular Surgery and submitted to a successful corrective surgery.
Table 1

Abnormalities observed at the thoracic HRCT

Abnormality

Patients (N=52)

Percentage

Nonspecific linear parenchymal opacities

10

19

Lymphadenopathy

6

12

Emphysema

5

10

Bronchiectasis

4

8

Pleural involvement

4

8

Pulmonary nodules

3

6

Pulmonary cysts

1

2

Descendant aortic aneurysm

1

2

Total

21

40

HRCT High-resolution computed tomography

https://static-content.springer.com/image/art%3A10.1007%2Fs10067-006-0286-2/MediaObjects/10067_2006_286_Fig1_HTML.jpg
Fig. 1

High-resolution thoracic CT image of nonspecific parenchymal opacities. Reticular opacities from 2 to 5 cm in length, often peripheral and generally in contact with the pleural surface

https://static-content.springer.com/image/art%3A10.1007%2Fs10067-006-0286-2/MediaObjects/10067_2006_286_Fig2_HTML.jpg
Fig. 2

High-resolution thoracic CT image of bilateral apical fibrosis. Linear and reticular lesions of interstitial lung fibrosis on the right, as well as focal fibrotic mass in the peripheral left lung

https://static-content.springer.com/image/art%3A10.1007%2Fs10067-006-0286-2/MediaObjects/10067_2006_286_Fig3_HTML.jpg
Fig. 3

High-resolution thoracic CT image of paraseptal emphysema. Small isolated areas of destruction marginated by thin linear opacities that extend to the pleural surface

https://static-content.springer.com/image/art%3A10.1007%2Fs10067-006-0286-2/MediaObjects/10067_2006_286_Fig4_HTML.jpg
Fig. 4

High-resolution thoracic CT image of cystic bronchiectasis. A string of cysts, caused by multiple dilated bronchi lying adjacent to each other

The male sex was statistically associated with dorsal involvement (p=0.012). African-Brazilian race was associated with bronchiectasis (p=0.044) and pleural involvement (p=0.044) at the thoracic HRCT. Smoking patients presented greater frequency of linear parenchymal opacities (p=0.026) at the thoracic HRCT. HLA-B27 was not associated with any abnormality. Dorsal spine involvement was associated with restrictive pattern at PFT (p=0.001) and linear parenchymal opacities (p=0.032) at the thoracic HRCT. A lower thoracic diameter was observed in patients with dorsal spine involvement (p=0.0001), restrictive pattern at PFT (p=0.023) and linear parenchymal opacities (p=0.015) at the thoracic HRCT.

Discussion

Although symptomatic pulmonary manifestations are infrequent in AS, most studies that analyzed the association between AS and parenchymal lung abnormalities were based on the analysis of chest radiographs. Recent studies analyzing HRCT results showed that the frequency of lung abnormalities is higher and nonspecific. The present study, analyzing 52 consecutive patients who were asymptomatic about lung complaints, observed predominantly nonspecific alterations in 8% of the chest radiographs and 40% of the thoracic HRCT, with 52% of the patients presenting restrictive pattern at the PFT. It confirms that plain chest radiography usually fails to reveal pulmonary abnormalities, reflecting the increased superiority of thoracic HRCT in the detection of these interstitial abnormalities.

Between the four patients (8%) with a detected abnormality in the chest radiograph, only one patient presented an alteration (upper lobe fibrosis) that could be related to AS. The remaining three patients presented nonspecific abnormalities. In similar studies, abnormalities in the chest radiographs were observed in 4 to 16% [17, 18, 21, 22].

A positive Mantoux skin test was observed in 23% of the studied AS population. Four patients presented a Mantoux test between 10 and 19 mm. As they presented no symptoms or epidemiological evidence compatible with the diagnosis of tuberculosis, they were not treated; they are currently followed by the Unit of Pneumology. Two patients presented Mantoux test result of 20 mm. One female patient presented a positive sputum culture and was treated for 6 months, while a male asymptomatic patient with a negative sputum culture presenting positive epidemiology (his mother died of tuberculosis) is currently followed without treatment, based on the Brazilian Consensus on Tuberculosis [26]. At the present moment, the Mantoux skin test is not a routine laboratory exam in the management of AS patients, and it is not observed in an increased rate of tuberculosis in AS patients treated with sulfasalazine or methotrexate. Nevertheless, with the increased use of TNF-blocking agents in AS especially infliximab and etanercept [27, 28], a major concern regarding the reactivation of latent tuberculosis [29] would indicate the prophylactic treatment of tuberculosis in case of prescription of a TNF-blocking agent in patients with a Mantoux skin test >5 mm [30].

Despite the fact that all the 52 studied patients were asymptomatic about lung complaints, 52% of the patients presented a restrictive pattern at the PFT. This restrictive pattern was statistically associated with a dorsal involvement and a lower thoracic diameter. Previous studies have also demonstrated that the impaired lung function parameters are not necessarily related to the degree of lung changes seen in HRCT [17, 18]. The causes of these restrictive lung parameters in AS patients are related to the negative compliance of the thoracic cage caused by the increased stiffness of the spine and the ankylosis of the costovertebral joints [5, 1012, 31, 32]. The limited expansion and the fixation of the thoracic cage in a relative hyperinflation cause a reduced FVC, whereas the TLC can even be normal or slightly decreased [18, 33, 34]. This functional lung impairment is detected even in cases of juvenile spondyloarthropathy [35] and is responsible for the limited aerobic capacity [36] and the impaired exercise tolerance [37, 38] observed in AS patients.

Recent studies have demonstrated abnormalities on thoracic HRCT varying from 53 [21] to 88% [22]. The most frequent abnormalities were frequently mild and in the lower lobes and considered nonspecific: parenchymal bands, subpleural lines, emphysema, bronchiectasis, bronchial wall, and interlobular septa thickening. The existence of interstitial lung disease in AS was not demonstrated until the advent of thoracic HRCT. Upper lobe fibrosis, with or without cavitation, is a rare HRCT finding, as it is on plain chest radiography. In the present study, we detected alterations on thoracic HRCT in 40% of the AS patients, with only one case of upper lobe fibrosis. Parenchymal bands and sparse thickening of the interlobular septa, characterized as linear opacities, were the most frequent findings (19%) considered as nonspecific interstitial lung disease. There was a statistical association between these nonspecific linear abnormalities and dorsal spine involvement, restrictive defect at PFT and the smoking habit. These pulmonary abnormalities are probably associated to the restrictive pulmonary capacity due to the rigid thoracic cage. The fact that the HRCT scans at the prone position were not possible to obtain in many AS patients due to the extreme rigidity of the dorsal and lumbar spine strengthens this hypothesis and brings to consideration the possibility of these alterations being associated with decubitus atelectasia (gravitationally dependent opacities). The authors considered the association with tabagism fortuitous because these tomographic findings are not consistently related to the habit of smoking [39, 40]. The comparison of the thoracic HRCT findings in this study with those reported in the literature is showed in Table 2.
Table 2

Thoracic HRCT findings in ankylosing spondylitis

 

Present study

Fenlon et al. [17]

Turetscheck et al. [18]

Senocak et al. [19]

Kiris et al. [20]

El-Maghraoui et al. [21]

Souza Jr. et al. [22]

N (%)

N (%)

N (%)

N (%)

N (%)

N (%)

N (%)

Patients

52

26

21

20

28

55

17

HRCT abnormalities

21 (40)

18 (69)

15 (71)

17 (85)

18 (64)

29 (53)

15 (88)

Pulmonary fibrosis

1 (2)

2 (8)

5 (9)

1 (6)

Interstitial lung disease

4 (15)

4 (19)

3 (15)

2 (7)

4 (7)

Linear opacities

10 (19)

11 (42)

6 (29)

9 (45)

5 (18)

26 (47)

10 (59)

Lymphadenopathy

6 (12)

3 (12)

Emphysema

5 (10)

4 (15)

2 (10)

9 (45)

5 (9)

6 (35)

Bronchiectasis

4 (8)

6 (23)

2 (10)

3 (15)

2 (7)

4 (7)

2 (12)

Pleural involvement

4 (8)

1 (4)

6 (29)

9 (45)

3 (18)

Pulmonary nodules

3 (6)

1 (4)

2 (10)

8 (40)

7 (25)

1 (6)

Pulmonary cysts

1 (2)

3 (18)

Descendant aortic aneurysm

1 (2)

HRCT High-resolution computed tomography

In a previous description of disease patterns in 147 AS patients attending the outpatient clinic of our institution, we did not find symptomatic pulmonary involvement related to AS [41]. The male sex presented a higher frequency of dorsal and cervical spine involvement, compatible with the results described in this study. HLA-B27 was not associated with pulmonary involvement, as hypothesized previously [8, 9]. As most abnormalities were not directly related to AS, it is comprehensible that this association was not obtained in the present study. Analyzing HLA-B*27 alleles, there was a frank predominance of HLA-B*2705, as observed in the casuistic of AS [42] and undifferentiated spondyloarthropathies [43] at our institution.

In conclusion, a subclinical pulmonary involvement is frequent in AS. The restrictive pattern at PFT is associated with dorsal spine involvement and a low thoracic diameter, while thoracic HRCT presents a series of nonspecific alterations, predominating the linear parenchymal opacities.

Acknowledgements

The authors would like to thank Helymar C. Machado and Andreia Semolini for the statistical analysis of this study.

Copyright information

© Clinical Rheumatology 2006