Severe interstitial lung disease in connective tissue disease: rituximab as rescue therapy

RESULTS

At the time of rituximab treatment, all patients had extremely severe ILD, with a median D_L,CO of 25% (range 16–32%) and FVC of 45% (range 37–59%), and an anticipated survival of less than 6–12 months (based on clinical experience and survival data in patients with idiopathic fibrotic ILD of similar severity) [1]. Following rituximab, we observed a significant categorical improvement in clinical status and/or PFTs (D_L,CO and/or FVC) in seven out of eight patients (p=0.008) according to the criteria outlined in the methods section. In one patient (patient 8), disease severity did not appear to change following treatment, other than mild symptomatic improvement. In six patients with serial PFTs, there was a median improvement in D_L,CO of 22% (range 0–119%; p=0.04) and FVC of 18% (range 0–100%; p=0.03) within 9–12 months of rituximab administration. This improvement occurred following a median decline in D_L,CO of 16% (range 8–67%) and in FVC of 29% (range 3–45%) in the 9–12 months preceding rituximab treatment (fig. 1a and b). Improvements in PFTs and/or clinical status tended to occur rapidly (within 2–3 months in most patients), and were also statistically significant (p<0.05) when analysed at the earlier time point of 6±1 months post-rituximab treatment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

a) Change in diffusion capacity of the lung for carbon monoxide (D_L,CO) % predicted in the 9–12 months before and after rituximab administration. Following rituximab, there was a median improvement in D_L,CO % pred of 22% (range 0–119%; p=0.04), in contrast to a median decline of 16% (range 8–67%) prior to rituximab. b) Change in forced vital capacity (FVC) % pred in the 9–12 months before and after rituximab administration. Following rituximab, there was a median improvement in FVC % pred of 18% (range 0–100%; p=0.03), in contrast to a median decline of 29% (range 3–45%) prior to rituximab.

Five patients had a striking response to rituximab, with dramatic improvements in lung function (patients 3, 4 and 5) and/or successful extubation (patients 1 and 7). Change in clinical status and PFTs are summarised in table 2. Supplementary figure S1a–f shows lung function responses to treatment in individual patients. Additional clinical information can be found in the online supplementary material, including complete functional follow up.

Polymyositis/dermatomyositis-associated ILD

Patient 1

This 45-yr-old male was transferred to the ICU at our hospital after failed extubation following a quadriceps muscle biopsy procedure. Recently diagnosed dermatomyositis with associated organising pneumonia had deteriorated despite i.v. methylprednisolone and, at transfer, HRCT demonstrated a pattern of organising pneumonia with admixed fibrosis and diffuse alveolar damage (fig. 2a). The combination of life-threatening respiratory failure, HRCT findings and the underlying autoimmune process resulted in the decision to treat concurrently with i.v. cyclophosphamide 750 mg and rituximab. A dramatic improvement in HRCT appearances occurred over the following six weeks (fig. 2a and b), and improvements in PFTs have continued during follow-up (with a D_L,CO of 50% and FVC 99% predicted at most recent review).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2–

Patient 1. A 45-yr-old male with dermatomyositis. a) Contrast-enhanced thin section CT shows patchy consolidation and ground glass opacification in the lower lobes, consistent with organising pneumonia and diffuse alveolar damage. A pneumomediastinum is present. b) 7 weeks following commencement of rituximab. The lower lobe consolidation and ground glass changes have almost completely resolved. Some of the reticular changes anteriorly, in the nondependent lung, were thought to be attributable to ventilator-induced damage.

Patient 2

This 60-yr-old male with longstanding PM (Jo-1 antibody positive)-associated fibrotic NSIP, experienced gradual respiratory deterioration and persisting myositis activity despite concurrent mycophenylate, cyclosporine and prednisolone. Within 9 months of rituximab, serum creatine kinase had decreased from 4,500 to 370 IU·L⁻¹, and there was a significant subjective improvement in muscle strength and dyspnoea, with a 19% improvement in D_L,CO (table 2).

Patient 3

This 60-yr-old female with anti-synthetase antibody (Jo-1)-associated fibrotic NSIP experienced a 12-month history of worsening symptoms and deteriorating PFTs despite i.v. cyclophosphamide. Following rituximab, there was a striking improvement in respiratory and extra-pulmonary symptoms within 3 months, coupled with improvements in PFTs (table 2).

Patient 4

This 29-yr-old female was referred with rapidly progressive fibrotic NSIP associated with anti-synthetase syndrome (Jo-1), and deterioration despite prednisolone and mycophenylate, followed by i.v. cyclophosphamide. Rituximab resulted in major improvements in symptoms and PFTs within 2–3 months of administration, with improvements continuing over the subsequent 12 months of follow-up.

Patient 5

This 51-yr-old male was referred with very rapidly progressive fibrotic NSIP, associated with anti-synthetase syndrome (Jo-1), and deteriorating interstitial lung involvement despite prednisolone, mycophenylate and ultimately i.v. cyclophosphamide. Treatment with rituximab resulted in rapid and striking improvements in symptoms and PFTs (supplementary figure S1d), however he died in his local hospital 18 months following rituximab therapy (with the cause of death reported as a lower respiratory tract infection). During follow-up at our hospital 8 days before death, he remained well with stable PFTs (D_L,CO of 38% and FVC of 68%) and normal inflammatory markers. The most recent lymphocyte subset analysis performed 12 months following rituximab (and 6 months prior to death) revealed persisting B cell depletion although infections had not been problematic over this time. Cyclosporine had recently been commenced (in addition to a long-standing mycophenolate and prednisolone regimen), due to persisting myositis activity, and to allow further steroid reduction.

DISCUSSION

The observation that rituximab is an effective rescue therapy in some patients with severe CTD-ILD has important clinical implications. If these findings are validated by further experience, it can no longer be assumed that patients with severe progressive CTD-ILD have the same bleak treated outcome as those with end-stage fibrotic idiopathic interstitial pneumonia. We report objective improvement in seven of eight patients (with very striking improvements in five patients) in whom the only realistic outcomes appeared to be lung transplantation or early death.

Although this study can be criticised as observational, it appears highly unlikely that a placebo-controlled evaluation could be performed in this context. In the first instance, patients and many clinicians will find a placebo-controlled trial unacceptable in the setting of life-threatening disease for which there is open therapy available with a good conceptual basis for its use. The selective bias against patients with more severe disease, when open treatment is accessible, has been highlighted as one of the key obstacles to observing clear-cut benefits in placebo-controlled trials in SSc-ILD [22], and in chronic pulmonary sarcoidosis [23]. Secondly, careful clinical observational studies based on treatment by proof of concept, rather than evidence base, may be the only feasible option in rare and/or clinically heterogeneous ILDs as those described here. Finally, severe end-stage disease is less attractive to pharmaceutical and other funding bodies, in view of a greater risk of non-efficacy and applicability only to a small patient subset, as highlighted by the exclusion of patients with more severe disease in most of the large recently completed clinical trials in IPF [24–29].

While our report is observational (with the inherent limitations), our analysis was a cross-over design, and thus the treatment effect was intrinsically more convincing than an evaluation of de novo therapy with no reference to previous lung function trends. Two series have previously reported the efficacy of rituximab in CTD-ILD [18, 19]. Importantly, both series included patients with much less severe interstitial lung involvement than our own cohort. Sem et al. [18] reported a positive effect of rituximab in seven of 11 patients with anti-synthetase syndrome-associated ILD, the majority of whom had failed conventional immunosuppression [18]. Although complete lung function data is not reported, graphical representation of available baseline measurements suggests that D_L,CO and FVC were only mildly to moderately impaired (D_L,CO ≥60% predicted in eight of nine patients, FVC ≥50% in six patients), in contrast with the markedly more severe lung disease in our PM/DM cohort (table 2). In SSc-associated ILD, Daoussis et al. [19] and colleagues reported a significant improvement in PFTs following rituximab treatment (compared to standard therapy). Again, rituximab-treated patients in this cohort had less severe ILD (mean % predicted D_L,CO of 52±21% and FVC of 68±20%).

While little prognostic data exists for very severe and progressive CTD-ILD, clinical experience suggests an expected survival of 6–12 months in patients with similar disease severity as those described in our cohort. Our results suggest that, in the context of underlying autoimmunity, even patients with very severe ILD (for whom we would usually plan palliative treatment) may be rescued with rituximab, despite disease that is unresponsive to all traditional treatments. Immunological over-activity appears to be a key process in progressive lung fibrosis, even when fibrosis is severe, and the effectiveness of rituximab suggests that B lymphocytes are key protagonists in disease progression.

Several important caveats should be highlighted in this report. Life-threatening disease severity in patients 1 and 7 (both of whom were mechanically ventilated at the time of referral), resulted in the decision to administer rituximab and i.v. cyclophosphamide concurrently. This does make an accurate assessment of the stand-alone effects of rituximab difficult to ascertain. However, cyclophosphamide was administered only twice, at a reduced dose (50–75% of the dose routinely used), whereas the standard regimen for ILD-CTD would have included six monthly doses (at 600 mg·m⁻² each) [30]. It therefore seems unlikely that the small dose of cyclophosphamide was effective on its own in controlling devastating ILD, although an additive effect cannot be excluded. In the other patients, the effect of concomitant or preceding immunosuppression needs to be considered as a potential confounding factor. Three received treatment with i.v. cyclophosphamide before switching to rituximab. Although a delayed effect of i.v. cyclophosphamide cannot be excluded with absolute certainty, stabilisation and/or improvement in lung function was observed in responders within 6 months of initiation of cyclophosphamide in a large retrospective cohort of patients with progressive fibrotic lung disease [31]. The remaining three patients had continued to progress while on mycophenolate, and response to rituximab seems the most plausible explanation for the observed improvement.

A negligible treatment effect was observed in one patient (patient 8) following rituximab. As described, this 60-yr-old male had SSc-associated fibrotic NSIP for >10 years and, while there was a definite improvement in symptoms and exercise capacity at 9 months, there was no significant change in PFTs.

Increased risk of infection following rituximab is a serious concern. In haematological malignancies, repeated doses of rituximab appear to be associated with an increased infection risk [32, 33], although this is not clearly apparent in rheumatological disease. In a recent meta-analysis of 745 patients with rheumatoid arthritis (RA), the rate of serious infections was not significantly different between rituximab and placebo-treated patients (2.3% versus 1.5%) [34]. In a long-term safety analysis of 2,578 patients with RA treated with at least one course of rituximab, the rate of severe infections did not increase with subsequent doses, despite an increased incidence of hypogammaglobulinaemia [35]. Progressive multifocal leukoencephalopathy has been described in rituximab-treated patients. Although the reported incidence is extremely low (less than 1:20,000 in rituximab-treated RA patients) [36], patients should be monitored closely for the development of new or worsening neurological symptoms. Rituximab induced pulmonary reactions have also been reported, with steroid responsive organising pneumonia the most common manifestation in one systematic review [37].

Two patients in our cohort had significant infectious complications following rituximab. Patient 7 experienced recurrent lower respiratory tract infections following rituximab and patient 5 died of a respiratory tract infection 18 months after rituximab. Patient 5 had commenced cyclosporine (in addition to longstanding mycophenolate) to better control his myositis, 4 months prior to the fatal respiratory infection, and the contribution of rituximab in predisposing to infection is difficult to establish. Neither patient experienced reduced immunoglobulin levels during follow-up. The possibility that rituximab-induced serious lung infections are more frequent in patients with ILD compared to autoimmune diseases without significant lung involvement will need to be evaluated further.