Tocilizumab therapy in rheumatoid arthritis with interstitial lung disease: a multicentre retrospective study

Interstitial lung disease (ILD) is the most severe extra‐articular manifestation of rheumatoid arthritis (RA). Although it is responsible of 10–20% of all RA mortality, no controlled studies are available for the treatment of RA‐ILD and its therapeutic approach is still debated.


Introduction
Rheumatoid arthritis (RA) is a chronic inflammatory disease, characterised by synovial joint swelling and tenderness, with progressive disability and joint destruction. 1 The interstitial lung disease (ILD) is the most severe extra-articular manifestation of RA, with impact on both therapeutic strategy and overall prognosis and survival. 2 Approximately 10% of the RA population develops a clinically significant ILD that is responsible for decreased quality of life and progressive chronic disability, but also of 10-20% of all mortality associated with the disease, with a mean survival of 5-8 years. [3][4][5][6] Although both genetic and environmental factors have been investigated, the pathogenesis of RA-associated ILD (RA-ILD) remains unclear. 7,8 been associated with the development or progression of ILD. 9,10 For these reasons, since no controlled studies are available, the therapeutic approach to RA-ILD is still debated and often empirical. 7,[11][12][13] Some limited reports have evaluated the safety and the efficacy of tocilizumab (TCZ), a humanised anti-interleukin 6 (IL6) antibody, in the treatment of patients with RA-ILD. 14,15 In this retrospective study, we analysed the evolution of ILD in a population of RA patients treated with TCZ.

Methods
In a national multicentre study, we retrospectively collected patients with ILD associated to RA treated with TCZ. All RA patients attending the Rheumatology Units of six Italian centres after 2008 and treated with TCZ for at least 6 months were retrospectively evaluated to identify patients with ILD.
RA was diagnosed according to the 1987 or 2010 classification criteria depending on the year of diagnosis. 16,17 The study was approved by the local Institutional Review Board.
The different patterns of interstitial lung involvement, defined by pulmonary biopsy or chest high-resolution computed tomography (HRCT), were classified according to the standardised criteria of the American Thoracic Society/European Respiratory Society International Multidisciplinary Consensus Classification of the Idiopathic Interstitial Pneumonias 18 as follows: (i) definite or probable usual interstitial pneumonia (UIP), (ii) non-specific interstitial pneumonia (NSIP) and (iii) organising pneumonia (OP) and mixed patterns.
RF was determined by nephelometry; anti-cyclic citrullinated peptides antibodies (ACPA) were detected by standard commercial enzyme-linked immunosorbent assays.
The results of pulmonary function tests (PFT) were expressed as percentages of the predicted value of each parameter and corrected for age, gender and height. Pulmonary function was considered as abnormal if forced vital capacity (FVC) was <80% of predicted values. Single-breath diffusing capacity of the lung for carbon monoxide (DLCO-SB) and DLCO adjusted by the alveolar volume (DLCO-VA) were used to assess gas transfer. The last HRCT and the last PFT performed before starting TCZ were recorded as baseline.

Outcome variables
A variation of 10% of FVC or DLCO compared to baseline was considered clinically significant. 19 Improvement, worsening or stability of HRCT was centrally evaluated in a blinded manner by an experienced thoracic radiologist (GDC). PFT were collected at baseline and periodically assessed and for all patients the last available value (within 3 months from the end of follow up) was recorded. All patients but one repeated HRCT at the end of follow up.

Statistical analysis
Results were expressed as median and interquartile range (IQR). Continuous variables were compared using unpaired or paired nonparametric tests (Mann-Whitney or Wilcoxon test respectively). A P-value <0.05 was considered significant. Statistical analyses were performed using the SPSS statistical software, v17.0 (SPSS Inc., Chicago, IL, USA). 20

Results
We included 28 RA-ILD patients (18 females and 10 males, median age 64, IQR 15) treated with TCZ. The drug was administered at the standard dose, both intravenous (8 mg/kg every 4 weeks) and subcutaneously (162 mg weekly). For all patients, HRCT was available in the previous 12 months before the beginning and at the end of the therapy with TCZ, while PFT were available in 25/28 patients.
Baseline characteristics of RA-ILD patients included in the study are summarised in Table 1.
The median follow up was 30 months (range 6-90). All patients were positive for RF and all but 2 for ACPA.
An UIP pattern was described in 14/28 (50%) patients, while a NSIP pattern was identified in other 13 patients. In the last case, a combined pattern pulmonary fibrosis and emphysema was recorded.
Twenty-three patients had taken TCZ as monotherapy, while in 5 patients TCZ was associated with MTX; on the other hand, 20 patients were treated with a low dose of prednisone (5 mg daily or equivalent). All patients but three were treated with subcutaneous TCZ, and six patients were switched from intravenous to the subcutaneous route of administration.
The evolution of lung function and radiology is summarised in Figure 1.
PFT were available at baseline and at the end of the follow up in 25 patients. After a median follow up of 30 months, FVC remained stable in 14 (56%) patients, improved in 5 (20%) and worsened in 6 (24%). Mean FVC was stable during follow up (99%, IQR 27 at baseline and 96%, IQR 26 at the end of follow up).
DLCO showed a similar trend, remaining stable in 14 (56%) patients, improving in 5 (20%) and worsening in 6 (24%), even though in 3 patients DLCO and FVC showed an opposite trend. Also mean DLCO remained stable during follow up (58.5%, IQR 23 at baseline and 57%, IQR 31.5 at the end of follow up).
HRCT was performed at the end of the follow up in all 28 patients, it was stable in 25 cases, worsened in 2 and improved in the latter. The worsening was recorded in patients with UIP pattern, while the only case of improvement was observed in a patient with NSIP pattern.
No differences were recorded according to the duration of follow up or to the previous therapies.
During the follow-up period, TCZ was withdrawn in 6 of 28 patients: for inefficacy in 3 cases and for adverse events in the other 3 (no adverse events were correlated to the treatment). No withdrawals were recorded for a worsening of ILD or for infections.

Discussion
In our study, we evaluated the efficacy and the safety of TCZ in a retrospective cohort of patients affected by RA-ILD, showing good safety and efficacy. The treatment of ACPA, anti-cyclic citrullinated peptides antibodies; cDMARD, conventional diseases modifying anti-rheumatic drugs; CPFE, combined pulmonary fibrosis and emphysema; HRCT, high-resolution computed tomography; IQR, interquartile range; NSIP, nonspecific interstitial pneumonia; TCZ, tocilizumab; UIP, usual interstitial pneumonia; these patients is challenging, due to the possible role of DMARD in the progression of the disease and in the development of acute exacerbation (AE). In fact, both MTX and leflunomide have been associated to ILD progression and development; recently, Conway demonstrated a mild increased risk of respiratory adverse events (but not ILD) in RA patients treated with MTX compared with other conventional and biologic DMARD. 21 Moreover, some authors described a possible class effect of all anti-tumour necrosis factor inhibitors (TNFi) in the new onset or exacerbation of ILD secondary to RA. [22][23][24][25][26][27][28][29] Perez-Alvarez et al. 27 and the British Society of Rheumatology have specifically cautioned prescribing TNFi to patients with RA-ILD for the supposed increased risk of exacerbation of the ILD. 30 Nakashita did not observe an increase of the prevalence of ILD progression in patients with TCZ and abatacept, whereas a prevalence of 3% of new ILD appearance and 24% of ILD worsening were described in TNFi users. 24,29,31 On the other hand, Curtis et al. found no significant differences in the risk of ILD incidence and its related complications between patients exposed to TCZ, rituximab, or abatacept compared with TNFi therapies. 32 Moreover, some case reports have reported acute worsening of pre-existing RA-ILD in patients treated with TCZ. 33,34 Despite the lack of evidence, the use of MTX is poorly indicated in patients with RA-ILD. [35][36][37] In this regard, TCZ could represent a possible safe drug in these patients, considering its efficacy in RA also as monotherapy. 38 In our population, the majority of the patients showed a stability of pulmonary function and about 20% showed an improvement of PFT and less than a quarter of patients showed a deterioration of lung function. At the same time, only 2 of 28 patients showed a worsening of HRCT.
On the whole, TCZ demonstrated a good safety profile in patients with RA-ILD and a good efficacy on the stabilisation of lung involvement. 14,15 In small case series, other authors observed similar efficacy in RA-ILD patients treated with other biologic DMARD. In particular, Md Yusof showed the improvement or the stability of ILD in 30/44 RA patients treated with rituximab, however describing a high number of infectious adverse events; 39 more recently, Fernández-Díaz et al. described a high rate of improvement of ILD in 63 RA patients treated with abatacept with or without conventional DMARD. 40 Data on abatacept were confirmed in 55 Japanese RA-ILD patients, despite the authors observing a deterioration of lung function in patients treated with a combination therapy with MTX. 41 Finally, no reports have been reported until now about possible involvement in ILD appearance or deterioration for the Janus Kinases inhibitors. 42,43 We cannot exclude that some biologic DMARD, such as TCZ, abatacept and rituximab, poorly influence the natural clinical history of ILD, and our results, in line with recent literature data, could reflect the natural evolution of lung involvement in RA. It should be essential, to minimise the risk of progression and AE of ILD and reduce the risk of infection, to provide the available vaccinations to all patients.
The management of RA-ILD patients remains a critical unmet medical need. Waiting for prospective controlled studies, in patients with RA-ILD should be preferred biologic DMARD, namely IL-6 inhibitors, abatacept and probably Janus kinases inhibitors, that have demonstrated a good safety profile in this specific population. 3, 11,15,42,43 Conclusions Despite the low number of patients investigated, TCZ should be considered a possible therapeutic option in patients with RA-ILD. Despite increasing data tend to exclude a role for MTX in development and progression of ILD, 44 the possibility to use TCZ in monotherapy without a significant loss of efficacy 35 allows us to use this drug also in cases with ILD and high articular disease activity where the use of MTX is poorly recommended.
Finally, an early diagnosis of ILD in RA patients is mandatory to understand the natural history of ILD, its possible predictive factors, and to evaluate the real involvement of some DMARD, such as MTX, in the development and progression of this severe extraarticular complication. 45 References