INTRODUCTION — Interstitial lung disease (ILD) is a frequent complication of systemic sclerosis (SSc) that often has a poor prognosis. In a retrospective study of 619 patients with SSc, 40 percent of patients had a restrictive ventilatory defect (suggesting interstitial lung disease, ILD) either alone or in combination with pulmonary arterial hypertension [1]. Herein, we discuss the prognosis and treatment of SSc-associated ILD.

The clinical presentation and diagnosis of SSc lung disease and the treatment of SSc and SSc-associated pulmonary arterial hypertension are discussed separately. (See "Clinical manifestations of systemic sclerosis (scleroderma) lung disease" and "Evaluation for and diagnosis of lung disease in systemic sclerosis (scleroderma)" and "Overview of the treatment and prognosis of systemic sclerosis (scleroderma) in adults" and "Prognosis and treatment of pulmonary vascular disease in systemic sclerosis (scleroderma)".)

DEFINITIONS — The term ILD is broadly used to describe a heterogeneous group of disorders that are classified together because of similar clinical, radiographic, physiologic, or pathologic manifestations.

In the vast majority of patients with SSc-associated ILD, the lung injury is characterized by a pattern termed nonspecific interstitial pneumonia (NSIP) [2]. Histopathologically, NSIP is characterized by varying degrees of pulmonary inflammation and fibrosis, with some forms being primarily inflammatory (cellular NSIP) and others primarily fibrotic (fibrotic NSIP). Most investigators believe that cellular NSIP is the early stage of fibrotic NSIP. Although NSIP may have significant fibrosis, it is usually of uniform temporality. Fibroblastic foci and honeycombing, if present, are rare. (See "Idiopathic interstitial pneumonias: Clinical manifestations and pathology", section on 'Nonspecific interstitial pneumonia'.)

In a minority of patients with SSc-associated ILD, the histopathologic pattern is that of usual interstitial pneumonia (UIP). This pattern is characterized by a non-uniform distribution of alternating zones of dense fibrosis, fibroblast foci, scant inflammation, normal lung, and honeycomb change.

In this topic review, we refer broadly to SSc-associated ILD and do not distinguish among the histopathologic subtypes. Historically, SSc-associated ILD has been referred to by various terms, such as fibrosing alveolitis, interstitial pulmonary fibrosis, and/or idiopathic interstitial pneumonitis.

PROGNOSIS — In patients with systemic sclerosis (SSc), interstitial lung disease (ILD) predicts increased mortality [3-9]. This was illustrated by a retrospective study of 953 patients with SSc; patients with severe ILD had a nine-year survival rate of approximately 30 percent, whereas patients with SSc who did not have severe involvement of an organ system had a nine-year survival rate of 72 percent [4]. The most rapid decline in forced vital capacity (FVC) occurred within the initial three years of disease onset, indicating that lung injury and fibrosis are early complications.

The prognostic value of lung biopsy in patients with SSc-associated ILD is limited; symptomatic and physiologic severity of the ILD are better predictors of outcome than histopathologic subtype. In a retrospective histopathological evaluation of 80 patients with SSc and biopsy proven ILD, 76 percent had nonspecific interstitial pneumonia (NSIP) and 11 percent had usual interstitial pneumonia (UIP) [5]. The five-year survival rate for patients with NSIP and UIP was similar, 82 and 91 percent, respectively. Markers of a worse prognosis included a lower diffusion capacity (DLCO) and a more rapid decline of DLCO over three years.

The prognostic value of BAL is uncertain [5,10-12]. Although one retrospective study reported that increased eosinophils in the BAL fluid was associated with a poorer prognosis, a larger prospective cohort study of 141 patients with SSc-associated ILD found that the proportion of eosinophils in the BAL fluid did not correlate with increased mortality, rate of functional deterioration, or progression-free survival [5,10]. An increased proportion of neutrophils in BAL fluid was associated with more extensive lung disease on high resolution computed tomography (HRCT), a greater reduction in diffusing capacity (DLCO), and early mortality (HR 8.40, 95% CI 1.91-36.95), but it did not predict the rate of functional deterioration or progression-free survival [10]. In a prospective clinical trial of 158 patients with early-stage SSc and symptomatic lung involvement, the presence or absence of BAL fluid neutrophilia did not predict rate of worsening or response to therapy [13].

  — Systemic sclerosis (SSc)-associated interstitial lung disease (ILD) probably results from parenchymal lung injury, followed by inflammation and subsequent fibrosis. Inflammation is the target of therapy since fibrosis of the lung parenchyma is irreversible. Thus, the ideal candidate for treatment is the patient who has active alveolar and interstitial inflammation.

The potential benefit of treatment must be weighed against the risks of therapy on a case-by-case basis. Unfortunately, the decision about whether to initiate therapy can be difficult because the benefits of therapy appear to be modest and the toxicities can be significant.

We suggest that immunosuppressive therapy be initiated in patients with SSc-associated ILD who have respiratory symptoms, abnormal and/or declining pulmonary function, evidence of active (ie, progressive) disease, and no contraindications, because these patients have high risk of progressive lung disease and increased mortality.

Active disease — Signs of active (ie, progressive) disease include an early disease stage, abnormal and/or declining pulmonary function, and possibly ground glass opacity on high resolution computed tomography (HRCT).

• Disease stage — Patients with early SSc-associated ILD are more likely to have active inflammation and to respond to pharmacologic intervention than those with long-standing disease. Although there is no uniform agreement on how early disease should be defined, it generally refers to the initial 12 to 24 months of disease, before significant irreversible organ fibrosis is established. This concept is supported by a retrospective study that compared the response to different therapies in 122 patients with SSc-associated ILD [14]. Regardless of the therapy, patients with early disease were more likely to have improvement of pulmonary function than those with longer disease duration.
• Pulmonary function — Pulmonary function tests demonstrating a restrictive ventilatory defect and decreased diffusing capacity for carbon monoxide (DLCO) are consistent with SSc-associated ILD. Serial tests that demonstrate worsening pulmonary function are an indicator of active disease.
• High-resolution computed tomography (HRCT) - Early studies suggested that ground-glass opacification on HRCT is due to inflammatory infiltration of the interstitium and alveoli (picture 1), while reticular or nodular abnormalities on HRCT are more likely due to parenchymal fibrosis. Subsequent studies however failed to demonstrate a correlation between ground-glass opacification and alveolar inflammation. (See "High resolution computed tomography of the lungs".)

• Bronchoalveolar lavage (BAL) — In some centers, BAL is performed as a part of the evaluation to exclude pulmonary infection, which may be important in patients already receiving immunosuppressive therapy for SSc. However, findings on BAL analysis do not appear to predict progression of lung disease or response to therapy [10,12]. There is, therefore, insufficient evidence to recommend performing BAL for assessment of disease activity or likelihood of treatment response. (See 'Prognosis' above.)

CONTRAINDICATIONS — Active or suspected infection is the most common contraindication to immunosuppressive therapy. Pregnant, lactating, or neutropenic patients should also avoid immunosuppressive therapy. Patients with a history of cyclophosphamide-related hemorrhagic cystitis should not be prescribed cyclophosphamide again.

DRUG THERAPY — Only a small number of randomized, controlled therapeutic trials have been performed in patients with systemic sclerosis (SSc)-related interstitial lung disease (ILD). It is difficult to recruit patients for placebo-controlled studies because of the high risk of disease progression. In addition, clinical heterogeneity (eg, diffuse versus limited SSc, early- versus late-stage disease), varying disease severity, and varying rates of progression complicates analysis of results. Therefore, there is a paucity of outcome data to guide treatment decisions.

General approach — Once the decision has been made to initiate therapy, we advocate intravenous monthly cyclophosphamide together with low dose oral glucocorticoids (equivalent of prednisone ≤10 mg/day). Liberal fluid intake is encouraged during therapy. Azathioprine plus glucocorticoid is an alternative regimen that can be considered for patients with contraindications to cyclophosphamide or who decline cyclophosphamide. Data supporting these regimens are presented below.

Cyclophosphamide — Cyclophosphamide can be administered orally or intravenously. Randomized trials have been conducted using both routes of administration and suggest that cyclophosphamide imparts modest benefit in SSc patients with early, symptomatic disease, whether administered orally or intravenously [13,15]. General guidelines for the use of cyclophosphamide are presented separately. (See "General principles of the use of cyclophosphamide in rheumatic and renal disease".)

Many patients in studies evaluating the efficacy of cyclophosphamide versus placebo in SSc-associated ILD have also been on a low dose glucocorticoid. It is for this reason that we usually use a low dose glucocorticoid with cyclophosphamide. (See 'Adjuvant glucocorticoids' below.)

Oral — Oral cyclophosphamide appears to be beneficial for some patients with symptomatic SSc-associated ILD [13,16-18]. This was demonstrated in a double-blind, multicenter trial (the Scleroderma Lung Study) that randomly assigned 158 patients with early SSc-associated ILD, dyspnea, and evidence of active alveolar inflammation to receive either oral cyclophosphamide (≤2 mg/kg) or placebo daily for one year [13]. Evidence of active alveolar inflammation included ground-glass opacities on high resolution computed tomography, or elevated neutrophils or eosinophils in bronchoalveolar lavage fluid. Concomitant glucocorticoids were permitted, but only in doses equivalent to ≤10 mg/day of prednisone. Outcomes at 12 months were:

• The cyclophosphamide group had a smaller decline than the placebo group in forced vital capacity (-1.0 versus -2.6 percent of predicted) and total lung capacity (-0.3 versus -2.8 percent of predicted). Diffusion capacity was the same in both groups.
• Measures of skin thickness, dyspnea severity, and self-reported disability also favored cyclophosphamide.
• The cyclophosphamide group had modest improvement in measures of health related quality of life, compared to those in the placebo group [19].
• Adverse events occurred more frequently in the cyclophosphamide- treated group. Among the 79 patients who received cyclophosphamide, nine developed hematuria, 19 developed reversible leukopenia, seven developed neutropenia, and five developed pneumonia within the 12 month treatment period. One patient developed severe hemorrhagic cystitis requiring surgery and three patients developed cancer. The duration of this trial was insufficient to assess the true incidence of cyclophosphamide-associated cancers (bladder, hematological and skin cancers), because they may not become apparent for many years.

These results indicate that a one-year course of oral daily cyclophosphamide has modest clinical efficacy for the treatment of patients with early SSc and active symptomatic alveolitis. However, the toxicity of daily oral cyclophosphamide is considerable and the decision to initiate therapy must carefully balance risk versus potential benefit [20].

A follow-up study reported 24 month outcomes for 93 of the 109 patients who completed one year of oral cyclophosphamide or placebo [21]. Of the 48 patients who had received cyclophosphamide,12 also received low dose prednisone. The beneficial effects of cyclophosphamide on forced vital capacity (FVC) appeared to persist for six months after stopping the drug. However, by 24 months the improvement in FVC was no longer present, suggesting that the response to cyclophosphamide is not durable. Improvement in respiratory symptoms, as well as in skin induration, persisted in the cyclophosphamide-treated group at 24 months.

Details of oral cyclophosphamide dosing are provided separately. (See "General principles of the use of cyclophosphamide in rheumatic and renal disease", section on 'Use of daily oral cyclophosphamide'.)

Intravenous — The effectiveness of intravenous cyclophosphamide used in combination with low dose glucocorticoid at preventing deterioration of lung function in patients with SSc-associated ILD has been studied in small observational studies and one randomized controlled trial [15,22-26]. In the trial, 45 patients were randomly assigned to receive six monthly infusions of cyclophosphamide plus prednisolone (20 mg on alternate days) [15]. At 12 months, there was a modest improvement of FVC in the cyclophosphamide group after adjustment for baseline FVC, but this improvement did not achieve statistical significance. Neither DLCO nor measures of dyspnea showed improvement in either group.

Selection of the initial dose of cyclophosphamide for intravenous infusion is based upon estimated body surface area (BSA) expressed in square meters (m2) and is adjusted for advanced age, obesity, and renal function. Subsequent doses are based on the white blood cell nadir and response to therapy. (See "General principles of the use of cyclophosphamide in rheumatic and renal disease".)

In summary, these randomized clinical trials suggest that cyclophosphamide plus low-dose glucocorticoid imparts modest benefit in SSc patients with early-stage symptomatic lung involvement. The improvements with oral administration of cyclophosphamide achieved statistical significance, whereas intravenous administration in a smaller trial did not. However, the intravenous route results in a lower cumulative dose and may be associated with a lower rate of side effects. Therefore, we generally prefer intravenous cyclophosphamide.

Adjuvant glucocorticoids — The optimal dose of adjuvant glucocorticoids used in combination with cyclophosphamide is uncertain. In studies evaluating the efficacy of cyclophosphamide versus placebo in SSc-associated ILD, many patients have also received low dose glucocorticoid [13,15]. It is for this reason that we usually use a low dose glucocorticoid (equivalent of ≤10 mg/day of prednisone) with cyclophosphamide.

We avoid combining high dose glucocorticoids with cyclophosphamide because of the lack of clinical trial data and the attendant risks of scleroderma renal crisis and immunosuppression. One observational series described a short-term benefit to high dose compared with low dose glucocorticoids in patients receiving cyclophosphamide for SSc-related ILD, but this has not been our experience [27].

Prophylaxis — Immunosuppressed patients are at increased risk for Pneumocystis jirovecii (PCP, previously called Pneumocystis carinii) infection; thus, we recommend that patients taking cyclophosphamide also receive prophylaxis against Pneumocystis jirovecii. In our practice, we administer trimethoprim-sulfamethoxazole (160 mg/800 mg) three times per week. (See "Treatment and prevention of Pneumocystis carinii (P. jirovecii) pneumonia in non-HIV-infected patients".)

Monitoring therapy — Cyclophosphamide has the potential for short and long term toxicity, including neutropenia, opportunistic infections, cystitis, bladder cancer, infertility, and adverse drug interactions. White blood cell count, renal function, and urinalysis should be monitored during cyclophosphamide therapy. A detailed discussion about potential side effects and monitoring during cyclophosphamide therapy is provided separately. (See "General principles of the use of cyclophosphamide in rheumatic and renal disease", section on 'Monitoring'.)

Azathioprine — Azathioprine is an alternative immunosuppressive agent for SSc patients with early stage ILD who are not candidates for cyclophosphamide or who decline cyclophosphamide. It is generally administered at a dose of 2.5 mg/kg per day to a maximum of 150 mg/day.

A retrospective analysis described 11 patients with SSc-associated ILD who received azathioprine plus prednisone for worsening pulmonary symptoms or declining lung function. The results showed that among the eight patients who received treatment for a minimum of 12 months, five patients had >10 percent improvement of FVC and three patients remained stable [28]. Three patients discontinued azathioprine due to adverse effects within the first six months of therapy.

Azathioprine appears less efficacious in SSc-associated ILD than oral daily cyclophosphamide [29]. In an unblinded trial, 60 patients with early diffuse SSc and ILD were randomly assigned to receive either cyclophosphamide (up to 2 mg/kg per day) or azathioprine (2.5 mg/kg per day). During the first six months of therapy, patients in both groups also received prednisolone (15 mg/day), which was subsequently tapered. After 18 months, FVC and diffusion capacity were stable in patients treated with cyclophosphamide, but had declined in patients treated with azathioprine. Leukopenia was more frequent in the cyclophosphamide group.

Preliminary data suggest that azathioprine may have a role as maintenance therapy in patients who have completed a course of cyclophosphamide. A retrospective series of 20 patients with SSc-associated ILD found stabilization or improvement in pulmonary function tests after a combination of six months of monthly intravenous cyclophosphamide followed by 18 months of azathioprine [26].

Monitoring — Monitoring for adverse effects during azathioprine therapy is discussed in detail elsewhere. (See "Pharmacology and side effects of azathioprine when used in rheumatic diseases", section on 'Drug dosage and monitoring'.)

D-Penicillamine — D-Penicillamine is an immunomodulatory agent that inhibits the formation of collagen crosslinks. D-Penicillamine has been used for over four decades to treat SSc, but its efficacy remains uncertain. Several retrospective studies demonstrated clinical benefits, including an improved skin score, less new organ involvement, and improved survival [30-32]. However, in a trial that randomly assigned 134 patients with diffuse SSc to receive either standard dose (750 to 1000 mg/day) or low dose (125 mg every other day) D-penicillamine, there was no difference in the extent of skin involvement or survival [33].

Glucocorticoids — Glucocorticoids, either as monotherapy or in combination with other drugs, have been widely used to treat SSc-associated ILD with variable benefit [34-41]. Given the lack of convincing benefit and the increased risk for scleroderma renal crisis, glucocorticoid monotherapy is not recommended. (See "Scleroderma renal crisis", section on 'Risk factors'.) We generally administer glucocorticoids in combination with cyclophosphamide. (See 'Adjuvant glucocorticoids' above.)

Methotrexate — Methotrexate has been used to treat both SSc and localized forms of scleroderma with modest improvement in skin involvement; however, a beneficial effect on lung involvement has not been demonstrated. Because methotrexate has been associated with the development of pneumonitis and, rarely, pulmonary fibrosis, we do not recommend its use in the treatment of SSc-associated ILD.

Ineffective agents — Many drugs that were previously used for treating SSc-associated ILD produced no improvement in lung function. Ineffective drugs include colchicine, para-aminobenzoic acid, chlorambucil, disodium EDTA, methysergide, and relaxin [38,42-44].

FOLLOW-UP — Follow-up of patients receiving drug therapy for systemic sclerosis (SSc)-associated interstitial lung disease (ILD) requires that the response to therapy be evaluating and the duration of therapy determined.

Assessing response — Assessing the response to therapy is challenging because:

• Improvement tends to be slow
• Small changes in symptoms, lung function, and/or radiographic appearance can be masked by the effect of existing fibrosis
• Stabilization, rather than improvement, may be the best result that can be achieved

Optimal follow-up has not been established. We initially evaluate our patients at least monthly for adverse effects of therapy. In the absence of adverse sequelae, we continue therapy for at least six months and then reassess the patient's symptoms, subjective exercise tolerance, pulmonary function, and high resolution computed tomography (HRCT). Therapy is continued if there has been improvement.

Duration of therapy — The optimal duration of therapy is unknown. We generally do not continue cyclophosphamide beyond one year, because of toxicity. Other therapies are typically continued until clinical improvement reaches a plateau, at which time gradual withdrawal of therapy can be considered.

LUNG TRANSPLANTATION — Lung transplantation may be an option for patients with severe systemic sclerosis (SSc)-associated interstitial lung disease (ILD) that is not responsive to pharmacologic interventions. (See "Lung transplantation: General guidelines for recipient selection".)

Carefully selected SSc patients undergoing lung transplantation have the same morbidity and mortality of lung transplantation as patients undergoing lung transplantation for idiopathic pulmonary fibrosis. This was illustrated in a retrospective review of the result of lung transplantation in nine patients with SSc-associated ILD [45]. None of the patients had cutaneous ulcers, recurrent episodes of aspiration, renal failure, or left ventricular dysfunction. Four year survival was approximately 70 percent.

A subsequent retrospective survey of 47 patients with SSc (mean age 46 years) who underwent lung transplantation determined that single lung transplantation was more common than double lung transplantation (57 versus 43 percent) [46]. Fifteen percent of the patients died within 30 days following transplant, and 32 percent died later. Late mortality was most commonly due to infection; other causes included respiratory failure, malignancy, and pulmonary hypertension. The one- and three-year survival rates were 68 and 46 percent, respectively, which did not differ from patients who received lung transplants for other conditions.

SUPPORTIVE CARE — Patients with systemic sclerosis (SSc)-associated interstitial lung disease (ILD) should receive the same supportive therapies used for the management of other types of ILD. These measures include supplemental oxygen, yearly influenza vaccination, periodic pneumococcal vaccinations, and pulmonary rehabilitation therapy.

INVESTIGATIONAL APPROACHES — Mycophenolate mofetil (MMF) and autologous hematopoietic cell transplantation are two therapeutic approaches that have been tried with limited success. The anti-leukemic agent, imatinib mesylate, has anti-fibrotic activity in cultured fibroblasts and prevents experimental fibrosis in mice [47].

Mycophenolate mofetil — Mycophenolate mofetil (MMF) is an inhibitor of lymphocyte proliferation that may prove to be safe and effective for the treatment of SSc-associated ILD [48-50]. In an observational cohort study, 13 patients with early SSc received anti-thymocyte globulin plus prednisolone for five days, followed by MMF (up to 2 gm/day) maintenance therapy for 12 months [48]. Long-term MMF was well tolerated. Although the extent of skin involvement improved, there was no change in mean forced vital capacity or diffusion capacity in this small trial. In two retrospective reviews of a combined total of 30 patients with scleroderma ILD, MMF treatment for up to 24 months was associated with improved or stable pulmonary function [51,52]. Larger investigations are needed before MMF can be recommended.

Hematopoietic cell transplantation — Immunoablative therapy followed by autologous hematopoietic stem cell infusion has been used in some patients with severe SSc. However, the cost and treatment-related morbidity and mortality of these procedures currently relegate them to the investigative setting. Several randomized trials are underway. (See "Immunomodulatory and antifibrotic approaches to the treatment of systemic sclerosis (scleroderma)", section on 'High dose immunosuppression with autologous stem cell support'.)

Imatinib — Imatinib mesylate, an inhibitor of the c-abl protein tyrosine kinase, is highly effective for the treatment of chronic myeloid leukemia and gastrointestinal stromal tumors. The selective inhibition of tyrosine kinase interferes with the signaling of both platelet-derived growth factor (PDGF) and transforming growth factor (TGF)-beta, two pivotal mediators of the fibrotic process in SSc [53,54].

Fibroblasts explanted and cultured from lesional skin and bronchial tissue from patients with SSc were stimulated with TGF-beta and PDGF, then exposed to imatinib [55,56]. This treatment led to the inhibition of Type I collagen production in a dose-dependent manner. Clinical trials of imatinib in the treatment of SSc are underway.

SUMMARY AND RECOMMENDATIONS

• Interstitial lung disease (ILD) predicts poor outcome in patients with systemic sclerosis (SSc). (See 'Prognosis' above.)

• The purpose of treating patients with SSc-associated ILD is to reduce alveolar and interstitial inflammation, in the hope that less interstitial fibrosis will develop. (See 'Prognosis' above.)

• The initial evaluation of patients with SSc-associated ILD includes high resolution computed tomography (HRCT) and pulmonary function testing, including diffusion capacity (DLCO). These tests are used to determine the extent and severity of disease. Worsening pulmonary function demonstrated by serial testing may be the best indicator of progressive impairment. (See 'Active disease' above.)

• Bronchoalveolar lavage may be useful for ruling out infection, and for assessing the severity of lung involvement, but appears to have little utility in predicting progression or response to therapy. (See 'Active disease' above.)

• The clinical benefits of therapy appear to be modest and associated with substantial toxicity; therefore, the decision to initiate therapy must be made on a case-by-case basis after carefully balancing risk versus potential benefit and the importance of each to the patient. (See 'Prognosis' above.)

• We suggest treatment with intravenous monthly cyclophosphamide therapy plus low dose glucocorticoids (equivalent of ≤10 mg/day of prednisone) instead of other immunosuppressive agents for patients with SSc-associated ILD who have respiratory symptoms, abnormal and/or declining pulmonary function, and no contraindications to immunosuppressive therapy (Grade 2B). (See 'Prognosis' above and 'General approach' above.)

• Selection of the initial dose of cyclophosphamide for intravenous infusion is based upon estimated body surface area (BSA) expressed in square meters (m2) and is adjusted for advanced age, obesity, and renal function. Subsequent doses are based on the white blood cell nadir and response to therapy. White blood cell count, renal function, and urinalysis should be monitored during cyclophosphamide therapy. (See 'General approach' above and 'Monitoring therapy' above.)

• A randomized, controlled trial has shown that azathioprine is inferior to cyclophosphamide. We suggest azathioprine plus glucocorticoids for patients with contraindications to cyclophosphamide or who decline cyclophosphamide (Grade 2C). Azathioprine is typically administered at a dose of 2.5 mg/kg per day to a maximum of 150 mg/day. (See 'Azathioprine' above.)

• We recommend prophylaxis against Pneumocystis jirovecii (previously called Pneumocystis carinii) for patients receiving immunosuppressive therapy (Grade 1B). In our practice, we administer trimethoprim-sulfamethoxazole (160 mg/800 mg) three times per week. (See 'Prophylaxis' above.)

• We suggest that immunosuppressive therapy be continued for at least six months (Grade 2C). Changes in the patient's symptoms, subjective exercise tolerance, pulmonary function, and HRCT should then be assessed. Therapy should be continued only if there has been improvement. Typically, therapy is continued until the improvement reaches a plateau, at which time gradual withdrawal of therapy can be considered. (See 'Assessing response' above and 'Duration of therapy' above.)

• Selected SSc patients who have severe ILD that is unresponsive to therapy may be referred for lung transplantation (Grade 2C). (See 'Lung transplantation' above.)