Patients and methods

Thirty patients were recruited from a specialist clinic and fulfilled objective criteria for “definite” or “probable” primary SS as proposed by Fox et al,8 with no evidence of another autoimmune disease. All were white women with a median age at diagnosis of 54 years (range 36–76). A clinical examination for glandular and extraglandular manifestations was undertaken, including Schirmer's tests, rose bengal staining of the conjunctiva and minor salivary gland biopsy specimen. Subjective xerostomia supported by the absence of a salivary pool under the tongue was accepted as a criterion. Serological tests were performed for antinuclear antibodies (ANA), rheumatoid factor (RF), anti-Ro and anti-La antibodies. Symptoms and signs of possible pulmonary disease were recorded together with a smoking history. A chest radiograph and pulmonary function tests (forced expiratory volume in one second (FEV₁), vital capacity (VC), carbon monoxide transfer factor (Tlco), and transfer coefficient (Kco); results expressed as standardised residuals9) were performed within six months of referral. Patients were reassessed after a median duration of four and then 10 years, including a full clinical examination, chest radiograph, and pulmonary function tests.

Standardised residuals9 were used to express changes in pulmonary function with time, to take account of differences in predicted values with age and between patients. The pairedt test was used to assess changes in lung function with time and the unpaired t test to compare anti-Ro positive and anti-Ro negative groups.

Results

Table 1 gives biographical details of the 30 patients studied. At baseline 13/30 patients reported dyspnoea. Five patients were current smokers. Bilateral fine basal crackles were present in four patients. Two had evidence of fibrosing alveolitis on plain chest radiograph. Five patients had a Tlco more than two standardised residuals below the predicted value (fig 1). After four years' follow up a further four patients developed dyspnoea. Three were now smokers, two having stopped since the initial assessment. Bilateral fine basal crackles were noted in nine patients, and one had features of a pleural effusion. The chest radiographs had become abnormal in a further two patients. All showed evidence of alveolitis, with associated persistent pleural effusions in one patient. There were significant reductions in Tlco (p<0.02) and Kco (p<0.02) compared with the baseline measurements (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Transfer coefficient (Kco) and carbon monoxide transfer factor (Tlco) expressed as standardised residuals in patients with primary Sjögren's syndrome within six months of diagnosis and at four and 10 years' follow up.

At 10 years, of the original 30 patients, four had died and four were lost to follow up. One patient died from bronchopneumonia complicating fibrosing alveolitis; two from cerebrovascular events, and the other from metastatic spread of a transitional cell carcinoma of the bladder. A total of 16/22 patients reported dyspnoea, of whom 12 had a persistent non-productive cough; three were current smokers. There were no new cases of pulmonary fibrosis identified on plain chest radiograph. The Tlco and Kco did not deteriorate further from the results at four years' follow up, with significant improvement in both Tlco and Kcobetween baseline and 10 years (p<0.001 and p<0.05 respectively) and between four years' and 10 years' follow up (p<0.001).

SEROLOGY

At initial documentation, 10 patients were anti-Ro or anti-La antibody positive, or both, seven ANA positive (but anti-Ro and anti-La antibody negative), two RF positive, and 11 seronegative. There were no changes in the patients' serology when repeated after 10 years. The one patient who died of an infective complication of fibrosing alveolitis was ANA positive. Anti-Ro antibody positive patients otherwise had significantly lower transfer factors (p<0.02) and transfer coefficients (p<0.01) than patients with primary SS without this serological marker (fig 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Transfer coefficient (Kco) and carbon monoxide transfer factor (Tlco) expressed as standardised residuals in Ro positive and Ro negative patients with primary Sjögren's syndrome at 10 years' follow up.

Discussion

The clinical significance of pulmonary disease in primary SS is uncertain and the reported prevalence from cross sectional studies ranges from 9% to 60%.10-12 This wide range is partially due to variation in the sensitivity of tests employed. High resolution computed tomography can identify parenchymal abnormalities in patients free of respiratory symptoms,5 suggesting subclinical pulmonary disease, whereas plain chest radiography and simple tests of pulmonary function, such as Tlco and Kco, are relatively insensitive techniques.

The significance of early parenchymal changes is best assessed by longitudinal studies. In our original report7 we described a significant reduction in Tlco and Kco over four years. Other follow up studies have found that lung function does not deteriorate to the same extent. Although Mialonet al demonstrated a statistically significant decrease in Tlco and Kco over 55 months' follow up of 18 patients with primary SS,13 the reduction was not considered to be large enough to be clinically significant. Linstow et al followed up 27 patients with primary SS previously reported to have reduced pulmonary diffusing capacities and found the Tlco and Kcoincreased significantly after seven years.6

Our 10 year follow up data suggest that the observed fall in Tlco and Kco seen at four years has not continued. There were statistically significant increases in both the Tlco and Kco between baseline and four year follow up results and those at 10 years' follow up. Radiological evidence of fibrosing alveolitis was present in four patients at year four. One patient with fibrosing alveolitis died of bronchopneumonia, but no new cases were identified between four and 10 years' follow up. However, respiratory symptoms remained prevalent, with 16/22 reporting breathlessness on minimal exertion and 12/22 a chronic non-productive cough.

The wide range in the reported prevalence of pulmonary disease in primary SS may also result from differences in patient groups. We have previously found that the anti-Ro antibody identifies a subgroup of patients with primary SS with more systemic disease,1 and the Tlco and Kco were significantly lower in our anti-Ro antibody positive patients at 10 years' follow up.

It is difficult to explain the apparent improvement in lung function in these patients beyond four years' follow up. This improvement was seen both in patients treated with oral corticosteroids, where there was radiological evidence of interstitial lung disease, and in those where there was no therapeutic intervention. The improvement in lung function found by Linstow et al 6 was also independent of treatment. We have recently reported14a reduction towards normal in serum IgG levels in patients with primary SS followed up for 10 years, again in the absence of treatment, suggesting a reduction in disease activity. Longitudinal studies involving bronchoalveolar lavage are required to see if the natural history of the lung disease in these patients is characterised by a reduction in infiltrating inflammatory cells with time.

Although our original paper7 had concluded that pulmonary complications of primary SS represented a significant threat to health and accounted for considerable morbidity, our long term follow up of these patients is reassuring, and suggests that most patients with primary SS do not develop progressive lung disease. Pulmonary complications occur predominantly in anti-Ro antibody positive patients and present early in the course of the disease.