trophilia and extensive SSc-ILD. Based on ndings in two large patient cohorts, it appears that the link between disease progression and a BAL neutrophilia merely refects the fact that more extensive SSc-ILD is more intrinsically progressive. In over 140 patients with SSc-ILD, BAL neutrophil levels were linked to global disease severity, as judged by PFTs and HRCT disease extent, and had no independent prognostic value with regard to disease progression or long-term mortality [50]. In the SLS, PFT follow-up was such shorter short in duration but was carried out at strictly standardized time intervals. BAL ndings were predictive neither of a treatment effect nor of disease progression in the placebo arm [51]. Following these studies, BAL is no longer routinely performed in the prognostic evaluation of SSc-­ ILD. BAL continues to be performed in selected patients with disproportionate upper-lobe abnormalities (to exclude pulmonary tuberculosis) or when HRCT evaluation suggests a coexisting disease process such as smoking-related interstitial lung disease.

Management

Historically, the core principle in the management of SSc-­ ILD has been to suppress infammation with corticosteroid or immunosuppressive therapy. This approach, based on a disease model in which infammation precedes and leads tobrosis, has been supported only by anecdotal reports and uncontrolled treatment effects in small groups of patients. The key limitation is the low prevalence of SSc-ILD in routine practice. For this reason, treatment statements were determined by clinical experience at single-referral centres for many years. However, since the millennium, multi-centre treatment studies in SSc-ILD (or in broader SSc populations with SSc-ILD evaluated using a key secondary endpoint), have proven to be possible. Pivotal data include placebo-­ controlled trials of oral cyclophosphamide [18], intravenous cyclophosphamide [52], bosentan [53], tocilizumab [54], nintedanib [55], a pivotal comparison between oral cyclophosphamide and mycophenolate mofetil [56], and a recent placebo-controlled trial of rituximab [57].

Cyclophosphamide was a logical trial therapy in the rst controlled treatment trials in SSc-ILD because partial regression with treatment was seen in some patients in small pilot series. In the landmark placebo-controlled SLS-1 trial of oral cyclophosphamide, statistically signi cant treatment effects were apparent at 1 year on FVC levels, dyspnoea, skin thickening and quality of life [19]. The SLS-1 trial was followed by a UK placebo-controlled trial of intravenous cyclophosphamide (given once a month for 6 months, followed by maintenance therapy with oral azathioprine) [52]. The study was under-powered due to recruitment dif culties that are

now regarded as inescapable in this eld. The FVC treatment effect was similar to that seen in the SLS-1 trial, in spite of only marginally signi cant (p = 0.08) due to the small cohort sizes (n = 45) [50]. Taken together, the two studies prompted EULAR to conclude that cyclophosphamide was an appropriate therapy in SSc-ILD [58].

However, this conclusion has not been uniformly accepted, and at the least, it is clear that cyclophosphamide should not be introduced indiscriminately in all patients with SSc-ILD. In both trials, the average FVC treatment bene t was less than 5% of baseline values and in the SLS-1 trial, although not in the UK trial, the small gain in FVC came at the cost of a signi cant prevalence of adverse effects. Importantly, many patients with mild lung disease were enrolled in both studies. This is understandable: the risk that an individual patient may receive a placebo, when open therapy is available, is likely to be more acceptable to patients and referring physicians alike when lung disease was not overtly progressive or severe. In keeping with this limitation, it is salutary that after patients in the SLS-1 trial had completed treatment and returned to routine follow-up, less than 15% were prescribed open therapy by their primary physicians [59]. Crucially, there was no treatment effect in the SLS-1 trial in patients with mild disease on HRCT. By contrast, there was a striking treatment effect on FVC (>10%) in extensive brotic disease, providing a useful clue as to which patients are likely to bene t in clinical practice [18].

At the time, it came as a surprise that in both trials, the treatment effect mostly represented stabilization with active treatment, rather than improvement. Regression of disease had been seen more frequently in previous smaller retrospective reports and received more focus than disease stabilization though, in the largest case series, FVC levels increased by an average of 4% in 39 patients receiving cyclophosphamide but fell by 7% in 30 untreated patients [60]. Based on the reversibility of disease in these and other pilot series, BAL and HRCT ndings considered to be indicative of “alveolitis” were inclusion criteria in the SLS study [18], although, as discussed earlier, reversible disease is identi ed reliably in SSc-ILD by neither test.

The SLS-1 and UK cyclophosphamide trials did not provide guidance on best longer-term management. Indeed, the SLS data underlined the need for maintenance therapy without providing any answers in this regard. Analyses of lung function trends in the SLS cohort showed that therapeutic bene ts had entirely been lost 12 months after treatment cessation [59]. For many years, azathioprine or methotrexate was used empirically until mycophenolate mofetil has gained in popularity based on a perception of greater ef cacy and lower toxicity. This anecdotal impression was evaluated formally in a meta-analysis of safety and ef cacy drawn from six eligible studies: outcomes were evaluated using trends in

FVC and DLCO% [61]. Overall, mycophenolate was associated with stabilization of disease, with no signi cant improvement in either pulmonary function variables.

These and other data prompted a comparison of the ef - cacy and tolerability of mycophenolate and cyclophosphamide in SSc-ILD, the SLS-2 study, undertaken by the USA SLS group [56]. Treatment with mycophenolate for 2 years or cyclophosphamide for 1 year was associated with similar improvements in measures of lung function: mycophenolate was better tolerated and associated with less toxicity.

High-dose corticosteroid therapy is viewed by many as absolutely contraindicated in SSc-ILD, due to an association between renal crisis and the use of prednisolone doses in excess of 15 mg daily [62, 63].

Following the SLS-1 and SLS-2 studies, routine therapeutic intervention in SSc-ILD most commonly consisted of mycophenolate, preferred to cyclophosphamide based on tolerability, and variably given in combination with low-dose corticosteroid therapy. However, the therapeutic landscape has now changed in SSc-ILD with the recent trials of tocilizumab [54] and nintedanib [55].

The focuSSced trial of tocilizumab [54] in the treatment of patients with diffuse cutaneous SSc can be applied to routine SSc-ILD management with two caveats: the primary endpoint used and the study enrolment strategy. Active treatment was associated with signi cant attenuation of decline in FVC; however, FVC change was a secondary endpoint— the primary end-point (change in the modi ed Rodnan skin score) was not met. Reassuringly, in an earlier phase II trial of tocilizumab in SSc, the same pattern of endpoint responsiveness was observed: with active treatment, a change in skin score (the primary endpoint) was not seen, but there was a reduction in the frequency of pulmonary function decline [64]. Furthermore, the study did not selectively enrol patients with SSc-ILD, although the treatment effect was robust in the majority SSc-ILD subgroup. Pulmonary function tests at baseline were mostly normal or mildly reduced. From the nature of the study population, it can be argued that the treatment bene t applies to patients with diffuse cutaneous SSc and limited SSc-ILD, when pro-infammatory pathways may be more prominent.

Nintedanib, a tyrosine kinase inhibitor with antibrotic effects, was evaluated in the SENSCIS nintedanib SSc-ILD trial, in patients with brosis affecting at least 10% of the lungs. The annual rate of FVC decline, the primary endpoint, was 52.4 mL in the nintedanib arm and 93.3 mL in the placebo arm, a relative reduction in FVC decline of approximately 45%. No bene t of was observed for other

manifestations of systemic sclerosis. Treatment was generally well tolerated with gastrointestinal adverse events, including diarrhoea, occurring more frequently with active treatment. Patients on mycophenolate for at least 6 months before randomization were able to participate and made up 49% of the total cohort. Importantly, the ndings indicated that the combination of mycophenolate and nintedanib is a safe treatment option [65]. The bene ts of mycophenolate and nintedanib in attenuating FVC decline appeared to be additive, but this requires further evaluation as the use of mycophenolate was not randomized.

Nintedanib was also evaluated in the INBUILD trial of non-IPF brotic lung disease, progressing despite management, including a subgroup of patients with SSc-ILD [66]. The treatment effect in attenuating FVC decline was highly statistically signi cant with a proportional reduction in decline of over 50% associated with active treatment. Taken together, the SENSCIS and INBUILD nintedanib data provide a basis for nintedanib therapy in SSc-ILD, both as initial treatment in selected patients and in patients with progressive lung disease despite immunosuppressive therapy.

Taken together, pivotal trials in SSc-ILD have validated the use of cyclophosphamide, mycophenolate, nintedanib and tocilizumab in patients at risk of ILD progression. Immunosuppressive therapy is likely to remain rst-line treatment in the majority of SSc-ILD patients as neither tocilizumab nor nintedanib has had bene ts in extra-­ pulmonic disease. The optimal selection of patients to receive initial combination therapy with mycophenolate and nintedanib has yet to be de ned.

Rituximab shows promise as an SSc-ILD intervention, despite yet to be subjected to controlled evaluation in a large patient cohort. In a pilot evaluation of anti-topoisomerase-­ positive SSc-ILD patients, there were improvements in pulmonary function data initially [67] and further improvements at 2 years [68]. Uncontrolled treatment bene ts have also been reported in patients with polymyositis lung [69] and in a mixed group of patients with connective tissue disease and life-threatening lung disease [70]. In a small randomized control trial, comparing Rituximab with standard treatment, signi cant bene ts were observed in the Rituximab arm in serial FVC and serial DLco [71]. The most compelling data come from a placebo-controlled evaluation of Rituximab, primarily focusing on skin disease, but supporting a signi cant bene t in SSc-ILD based on a key secondary endpoint [57].

Other interventions have not been studied in detail in SSc-­ ILD. Bone marrow transplantation has been used in small groups of patients with severe SSc, including in some

patients with severe SSc-ILD [72, 73]. Three randomized studies have shown that autologous hematopoietic stem cell transplantation is superior to conventional treatment in SSc, resulting in longer overall survival, longer disease-free survival and higher quality of life [72, 74, 75]. However, although signi cant improvements in pulmonary function tests have been documented, this treatment approach has yet to be widely viewed as standard therapy in severe SSc-ILD.

Clinical Vignette

A case of a 71-year-old man who presented with a 1-year history of Raynaud phenomenon, limiting dyspnoea and limited cutaneous scleroderma. He was a former smoker, stopping at age 30 with a ve pack-­ year smoking dose. There was no other past medical history of note. At presentation, his exercise tolerance was unlimited on the fat, walking at his own pace, but he was compelled to rest on climbing two fats of stairs. On auscultation of his chest, crackles were audible to the mid-zones.

No abnormalities were present on routine blood tests. Initial autoimmune serology showed strongly positive anti-nuclear antibodies but no speci c serological abnormalities and, in particular, he was anti-­ Scl70 antibody negative. Pulmonary function tests revealed FVC 60% of predicted, DLco 34% and Kco 68%. The calculated alveolar-arterial oxygen gradient was at the upper limit of normal (3.1 kPa). An echocardiogram was unremarkable. Representative HRCT sections (Fig. 12.4a–c) were in keeping with the overall conclusion, on rapid evaluation of all images between the main carina and the higher diaphragm that disease extent was “intermediate” overall (i.e. not clearly either <20% or >20% on rapid evaluation). Based on the FVC level of 60% of predicted, his lung disease was staged as extensive. There was also enlargement of the main pulmonary artery with the diameter greater than the aortic diameter (Fig. 12.4b).

His initial lung-speci c therapy consisted of Prednisolone 10 mg daily, intravenous cyclophosphamide 650 mg/m squared at four weekly intervals. After six cycles of cyclophosphamide, he continued on low-­

dose prednisolone and mycophenolate for 6 years of follow-up with complete stability of FVC levels. However, there were major changes in measures of gas transfer and gas exchange (Fig. 12.5a–c). It is instructive to note that following a normal echocardiogram, serial Kco levels (Fig. 12.5b) provided the clearest signal of an increasing pulmonary vasculopathy whereas DLco change (Fig. 12.5a) was less clear cut, possibly due to the confounding effects of interstitial lung disease. Serial measures of pO2 and the A-a gradient were less useful in early progression of vasculopathy but changed strikingly prior to overt right ventricular decompensation. Even before pulmonary hypertension was diagnosed, the patient had received intermittent prostanoids for Raynaud phenomenon and warfarin was added follow right heart study. Following right ventricular decompensation, the patient was oxygen dependent and sildena l was introduced.

The case is presented for several reasons. The use of the “mild/extensive” severity staging system, discussed in the text, is illustrated: a clear conclusion that the patient had extensive disease, despite a duration of disease of only 1 year, led to vigorous immunomodulation, with complete stabilization of interstitial lung disease during 6 years of follow-up. Had progression of interstitial lung disease occurred, possible treatment options, based on recent data, would have included nintedanib. The autoantibody pro le was nonspeci c, possibly in keeping with the parallel development of interstitial lung disease and a disproportionate pulmonary vasculopathy (in association with marginally extensive lung disease but no hypoxia until pulmonary hypertension was established). Overall, despite targeted therapy, there was an insidious downward in measures of gas transfer and gas exchange over 6 years, despite treatment. However, from our knowledge of the usual outcome in pulmonary hypertension in SSc, a worthwhile treatment bene t cannot be excluded. Finally, the case does illustrate the use of serial pulmonary function indices in helping to increase suspicion of worsening pulmonary vasculopathy, leading to earlier invasive evaluation in appropriate cases.

Fig. 12.4  (a–c) Representative HRCT sections in a patient who was staged as having extensive interstitial lung disease, based on an “indeterminate” disease extent and an FVC level of 60% of predicted. In basal sections, disease extent on HRCT was clearer greater than 20%,

but this was counter-balanced by much less extensive disease between the main carina and the pulmonary venous confuence. Note also, in (a), that the ratio of the pulmonary artery to the aorta was increased, suggesting the presence of pulmonary vasculopathy