Pulmonary fibrosis (PF) is a rare disease that is found in the course of a large variety of lung diseases. A diagnosis of idiopathic PF, which mostly occurs in patients over 50,1 is made after ruling out these causes of PF. In the course of idiopathic PF autoantibodies may be found, but they have no clinical relevance.2 However, detection of antineutrophil cytoplasmic antibodies (ANCA) in severe multiorgan dysfunction and parenchymal pulmonary disorders is highly specific for systemic vasculitides3 such as Wegener’s granulomatosis (WG), microscopic polyangiitis (MPA) and Churg–Strauss syndrome (CSS).

ANCA-positive vasculitides (ANCA-V) and Goodpasture syndrome are the major causes of pneumorenal syndrome. The spectrum of pulmonary symptoms is broad: intra-alveolar haemorrhage (IAH), granulomas and endobronchial stenosis. However, PF has been reported in only a few cases.

We retrospectively investigated 12 patients with evidence of ANCA-V associated with diffuse PF. The aim of our study was to determine the clinical and immunological characteristics of these patients.

METHODS

Between 1994 and 2004, we retrospectively investigated 12 patients, who presented with PF fulfilling the international consensus statement criteria,1 and associated with ANCA-V, fulfilling the Chapel Hill Classification criteria.4

The data were collected retrospectively from the database of the French Vasculitis Study Group, which included 517 patients.

For the follow-up, we defined pulmonary aggravation as an increase in dyspnoea, according to New York Heart Association (NYHA) stages and/or the death because of a pulmonary complication without any other cause.

RESULTS

Table 1 shows the characteristics of the patients.

The 12 patients comprised nine men and three women, with a mean age of 70.7 years (range 64–78) at ANCA-V diagnosis. Eight patients were smokers, but none had a known toxic environmental exposure, cardiac failure or a family history of PF.

According to the Chapel Hill criteria,4 10 patients had MPA and two patients had WG. In three cases, PF preceded ANCA-V by several years and in one case, PF was diagnosed several months after the demonstration of ANCA-V. In the other eight patients, PF and ANCA-V were diagnosed at the same time.

The main respiratory manifestations were coughing (10 patients) and NYHA II–IV dyspnoea (seven patients). Fine crackles were present in seven patients and no patients had clubbing.

Chest radiographs showed diffuse and bilateral interstitial opacities in all cases. A high-resolution computed tomographic scan was available for all patients, confirming symmetrical PF in all cases. There were signs of usual interstitial pneumonia in six cases and of non-specific interstitial pneumonia in one case. The type of interstitial diffuse pneumonia was unspecified in five cases. Transbronchial pulmonary biopsies were performed on three patients, disclosing fibrosis without infection, granulomas, vasculitis or IAH. Bronchoalveolar lavage was performed in five cases, disclosing a non-specific alveolitis, with 5% eosinophils in two cases. No patient had IAH with an insignificant Golde score.

Pulmonary function tests disclosed a restrictive syndrome. Initial functional pulmonary tests were available for six patients, showing a total lung capacity from 45% to 101% and a carbon monoxide transfer factor (Tlco) from 46% to 87%. During follow-up, we observed a progressive alteration of Tlco. Echocardiography, performed in seven patients, did not disclose pulmonary hypertension.

In all patients, at least one extrapulmonary organ was affected (table 1), the most common being the kidney (eight cases).

ANCA had a perinuclear pattern with anti-myeloperoxidase specificity (ELISA) in all cases. Antinuclear, antiproteinase 3, antiphospholipid and antiglomerular basement membrane antibodies were negative. Blood eosinophilia was present in seven patients ((0.5–13.0)×10⁹/l) without evidence of parasitic infection, drug hypersensitivity, or CSS.

The median follow-up was 49.2 months (range 7–116). All 12 patients were treated with steroids and cyclophosphamide as induction therapy; two patients died after 7 months of follow-up. In one case the ANCA-V was responsible for death (multiorgan failure) and in the other case, death was due to an opportunistic infection (cytomegalovirus pneumonia). For the 10 surviving patients, induction therapy allowed for remission of the ANCA-V and remission was maintained with azathioprine (six cases) or mycophenolate mofetil (one case) and steroids alone (three cases). Despite this treatment, four of 10 patients had a relapse of ANCA-V, which required an intensification of their immunosuppressive treatment. Neither induction nor maintenance therapy led to recovery of pulmonary function.

Three deaths occurred later in the follow-up (range 37–67 months) after progressive aggravation of the PF. Deaths were related to an acute respiratory failure with fatal pneumothorax in one case and non-documented infectious pneumonia in others. For two of these patients the ANCA-V was still in remission and one patient had a relapse of the ANCA-V 4 months before death.

Long-term follow-up of the PF itself (range 18–116 months) showed that five patients had a pulmonary aggravation, whereas the others remain stable. No discriminative measures seem predictive of the worsening, except eosinophilia (>0.5×10⁹/l; range (0.585–4.350)×10⁹/l), present in the five patients whose condition worsened.

DISCUSSION

Only a few cases of association of PF and ANCA-V have been reported.5^–11 In the main ANCA-V series and specifically the pathological series,12 the prevalence of PF was low.

In our review, we analysed publications in which cases showed a clear clinical manifestation of ANCA-V associated with PF5^–11: we found only 17 eligible patients. An analysis of those patients displays some important characteristics. The age at diagnosis of ANCA-V seems often higher than expected; specifically, patients were 70.7 years of age in our series and 68.8 years of age in published reports, thus closer to the age of idiopathic PF.1 There was a male predominance in our series (75%), which has not been confirmed in previously reported studies (47%). All the patients had myeloperoxidase (MPO)-ANCA and the ANCA-V was more often MPA than WG, suggesting a pathogenic role of MPO or MPO antibodies in the PF.12 Unusual eosinophilia was present in the five severely affected patients, but we had indications for CSS according to American College of Rheumatology and Chapel Hill criteria.

The radiological presentation of PF supports signs of usual interstitial pneumonia (6/7 cases in our series and 6/7 in reported studies).

In eight of the 12 patients of our series, PF and ANCA-V were diagnosed at the same time, suggesting a close relationship between the two diseases. The published series show a slight difference as ANCA-V may appear many years after the diagnosis of PF (in 8/17 cases, PF occurred up to 6 years before ANCA-V; table 2).

In our series and as reported elsewhere, the prognosis is poor. Five patients (41.7%) died before the 67th month after diagnosis: three of the patients were in a late stage of follow-up and death was due to a fatal and refractory evolution of PF. Therefore, once the acute stage of ANCA-V is complete, survival is directly related to PF and not to ANCA-V and the prognosis seems to be no better than with idiopathic PF.1

Furthermore, we did not find a close correlation between MPO-ANCA titres, the course of the vasculitis and the evolution of PF. For the four patients whose ANCA-V relapsed, only two had worsening PF, whereas for the five patients with aggravated PF, only two of them had had a relapse of ANCA-V. Our findings corroborate those of other reported studies (the evolution of PF and the relapse of ANCA-V were unrelated in six of 10 cases).

The theory for the pathogenesis of PF is now based on fibroblastic foci. Several studies, in both idiopathic PF and ANCA-V,6 13 have suggested that IAH may play an indirect role in the aggravation of respiratory insufficiency. Other studies have shown focal fibrosis directly opposed to the pulmonary lesions seen in ANCA-V such as lung vasculitis, granulomas or necrosis.14 However, for all the patients who underwent a long-term follow-up, there were neither arguments supporting constitution of parenchymal granulomas or foci of necrosis nor IAH. Furthermore, bronchial artery vasculitis or lung capillaritis was not found histologically. Even if these lesions had occurred owing to active ANCA-V, patients should have responded positively to immunosuppressive treatment and the evolution of the PF would have been stable, or at least parallel to the ANCA-V.

For all these reasons, neither our cases, nor those previously reported, suggested that PF is only due to an abnormal wound healing of pulmonary ANCA-V lesions as an occult IAH.

One hypothesis which may explain PF in this situation concerns MPO-ANCAs. The constant positivity of anti-MPO antibodies in our series (and in other reported studies) and animal models15 suggest a putative role for MPO-ANCA in the pathogenesis of PF.

These 12 cases of PF associated with ANCA-V show that this association is remarkable for the older age of the patients, the constant positivity of MPO-ANCA and the poor prognosis of the pulmonary disease, which is independent of the course of the vasculitis. Blood eosinophilia seems common in this setting. A fortuitous association of two rare diseases seems unlikely and we hypothesise that MPO-ANCAs play a direct role in the pathogenesis of PF.