https://orcid.org/0000-0001-8670-1745
Abstract
To assess the effect of rituximab (RTX) on the lung function parameters in SSc interstitial lung disease (SSc-ILD) patients.
PubMed and Embase were searched to identify studies on SSc-ILD treated with RTX, confined to a predefined inclusion and exclusion criteria. A systematic review and meta-analysis were performed on the included studies on changes in forced vital capacity (FVC) and diffusion capacity of carbon monoxide (DLCO) from baseline to 6 and 12 months of follow-up.
A total of 20 studies (2 randomized controlled trials, 6 prospective studies, 5 retrospective studies and 7 conference abstracts) were included (n = 575). RTX improved FVC from baseline by 4.49% (95% CI 0.25, 8.73) at 6 months and by 7.03% (95% CI 4.37, 9.7) at 12 months. Similarly, RTX improved DLCO by 3.47% (95% CI 0.99, 5.96) at 6 months and 4.08% (95% CI 1.51, 6.65) at 12 months. In the two studies comparing RTX with other immunosuppressants, improvement of FVC by 6 months in the RTX group was 1.03% (95% CI 0.11, 1.94) greater than controls. At the 12 month follow-up, RTX treatment was similar to controls in terms of both FVC and DLCO. Patients treated with RTX had a lower chance of developing infections compared with controls [odds ratio 0.256 (95% CI 0.104, 0.626), I2 = 0%, P = 0.47).
Treatment with RTX in SSc-ILD was associated with a significant improvement of both FVC and DLCO during the first year of treatment. RTX use was associated with lower infectious adverse events.
Rituximab improved forced vital capacity (FVC) and diffusion capacity at the 6 and 12 month endpoints.
Improvement of FVC by 6 months was greater in the rituximab group compared with controls.
SSc-ILD patients treated with rituximab had a lower chance of developing infections compared with controls.
Introduction
SSc is a chronic connective tissue disease (CTD) manifested by widespread vasculopathy and progressive fibrosis of the skin and internal organs [1]. Pathologically it is characterized by extracellular matrix remodelling, inflammation, impaired tissue healing, microvascular injury and tissue hypoxia [1]. Significant internal organ involvement is in the form of interstitial lung disease (ILD) [2]. Up to 90% of patients with SSc may exhibit radiological changes of ILD in high-resolution CT (HRCT) of the thorax, whereas 40–75% of patients may show pulmonary function test (PFT) abnormalities. Clinically significant lung fibrosis is said to be present in ∼25% of all SSc patients, but significant heterogeneity exists with regard to clinical subsets, clinical manifestations and autoantibody profiles [3].
The recently published European League Against Rheumatism Scleroderma Trial and Research (EUSTAR) report also cited a relatively high prevalence of clinically significant ILD (∼50%) among patients with SSc [4]. Development of ILD in SSc, with or without pulmonary hypertension, contributes to major morbidity and the risk of death. The presence of ILD, pulmonary hypertension and cardiac involvement are the leading causes of death in patients with SSc [1]. Notably, SSc shares the worst standardized mortality ratio among all systemic rheumatic diseases [5].
Recent years have witnessed a number of immunosuppressive and antifibrotic therapeutic strategies in SSc-related ILD (SSc-ILD). Options range from CYC, AZA, MMF, pirfinedone, rituximab (RTX) and tyrosine kinase inhibitors like nintedanib and autologous stem cell therapy [6–11]. The use of CYC is limited by its adverse effects, including infection, gonadal failure, solid organ malignancy and haematological toxicity. Some of these manifestations are related to cumulative drug exposure, precluding prolonged or recurrent use [9]. Alternatives like MMF, while better tolerated, have significant gastrointestinal toxicity as well as issues like cost and pill burden [9]. Importantly the issue of high dropout rates was discussed as one of the main results in a previous meta-analysis of CYC [12].
RTX, a chimeric anti-CD20 monoclonal antibody, has shown encouraging results in some studies conducted globally on both cutaneous and lung fibrosis in patients with SSc-ILD [1]. The utility of RTX as a potential therapeutic target is based on the central role of B cells in the pathogenesis of SSc [9]. RTX has been used off-label in patients with SSc who did not respond to conventional therapy [13]. However, the evidence for improvement of lung function as measured by forced vital capacity (FVC) is not uniform [1, 9]. The present systematic review and meta-analysis was conducted to assess the effect of RTX on lung function parameters in SSc-ILD.
Materials and methods
Literature search
We searched the existing literature for any previous systematic review and found none. Next we searched PubMed and Embase with the following keywords: (systemic sclerosis OR scleroderma) AND (interstitial lung disease OR ILD OR diffuse parenchymal lung disease OR DPLD) AND (rituximab OR anti-CD 20) without any restrictions on language. The last date of the search was 31 January 2020. The selected titles and abstracts were organized on Zotero and exported to Covidence (Covidence, Melbourne, VIC, Australia). The references of all retrieved articles were scanned for additional relevant citations. After removal of duplicates, abstracts were screened and redundant articles were removed. Full texts were retrieved for the included articles.
Eligibility criteria
We included studies if relevant information on patients’ characteristics, baseline pulmonary functional parameters, treatment interventions and outcomes were available. There was no restriction on language or study design and reports in abstract form were also included. However, we excluded single case reports or case series with fewer than five patients. The inclusion criteria were classification of SSc based on either 1980 American Rheumatism Association criteria or 2013 American College of Rheumatology/European League Against Rheumatic Diseases criteria [14, 15], HRCT-proven ILD before initiation of RTX, at least one subgroup of patients were treated with RTX with or without other concomitant immunosuppressive agents and there was outcomes assessment by FVC (% predicted) both at baseline and at the end of either 6 or 12 months. Exclusion criteria were studies in which RTX group–specific outcomes were not reported and where exact values of PFT parameters were not reported. Narrative reviews were excluded. Previous treatment with immunosuppressive agents was not considered an exclusion criterion. All reported adverse events were included.
Study selection and data abstraction
Two reviewers (R.P.G. and A.R.) carried out the searches, study selection and data extraction independently. Disagreements in the process of selection were resolved by discussion until a consensus was achieved. Both independently performed screening and assessed studies for inclusion according to the above-mentioned eligibility criteria. The full texts of all potentially relevant articles were retrieved for detailed review. The two reviewers used a predesigned data collection spreadsheet to extract data. The following data were extracted: authors; year of publication; study design [prospective, retrospective, randomized controlled trial (RCT), conference abstract]; baseline characteristics of patients, including the number of patients that met the inclusion criteria, age, gender, duration of SSc, proportion of dcSSc and lcSSc; treatments before and concurrent with RTX; total dose and regimen of RTX; baseline FVC, diffusion capacity of carbon monoxide (DLCO) and the modified Rodnan skin score (mRSS); 6 and 12 month FVC, DLCO and mRSS; presence or absence of a control group and the control drug (actual drug or placebo); baseline, 6 month and 12 month FVC, DLCO and mRSS in the control arm; and adverse events associated with RTX.
Quality assessment
Quality assessment was based on the Cochrane Risk of Bias tool for RCTs and the Newcastle–Ottawa Scale for the cohort studies. Two reviewers (A.M. and G.B.S.) independently assessed the quality of studies and any disagreements were resolved with the discussion of three reviewers (A.M., G.B.S. and P.G.) [16, 17]. The quality assessment results are presented in Supplementary Tables S1 and S2, available at Rheumatology online.
Statistical analysis
Patients’ demographic characteristics and PFT results were analysed from those studies enrolling at least five patients. Those presenting data as mean and s.d. were kept as such and those representing data as median and range were transformed into estimated mean and CI with the Hozo’s transform [18]. The effects of RTX were examined in terms of FVC, DLCO and mRSS at 6 and 12 months and pooled mean differences (FVC 6 or 12 months − FVC baseline) were determined along with the 95% CI. Comparison of the effect of RTX with matched controls was done at the said time points in terms of relative change in parameters (ΔFVCRTX – ΔFVCControl, etc.). For the matched control analysis, the available data were mean and s.d. at baseline and at the designated time points. A pooled estimate of the changes over time was obtained based on the Hedges’ g method [19, 20]. Heterogeneity was calculated using the I2 statistic representing the percentage of total variation across studies [20]. If I2 was zero, then a fixed effects model was used and if I2 > 0, a random effects model was used [21]. Two-step sensitivity analysis was carried out. We repeated all results, in the first step deleting the conference abstracts and in the second step deleting the retrospective studies. Begg and Mazumdar tests and funnel plots were used for estimation of publication bias and a P-value ≤0.05 was taken as significant publication bias [22].
Results
Search results and characteristics of included studies
A flow diagram of the search results is presented in Fig. 1. Overall, 598 studies were identified and after removal of duplicate, overlap and irrelevant studies, 75 were eligible for full-text review. After full-text screening, 20 were included in this meta-analysis (n = 575 for RTX) [1–4, 9, 13, 23–37].
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram of study selection process for this systematic review and meta- analysis
Baseline patient characteristics are presented in Table 1. Among these, seven studies were published as conference abstracts (n = 263) [23, 24, 27, 30, 32, 33, 36]. The abstracts were carefully scrutinized and it was ensured that redundant publications were excluded. Two studies (39 RTX and 39 controls) compared treatment effects with those of matched controls after 6 months [3, 8] and four studies (194 RTX group and 535 controls) compared treatment effects with those of matched controls after 12 months [1, 13, 28, 29].
Study characteristics
| Study | Design | Age, years | Duration, years | dcSSc/lcSSc | F/M | Dosea | Follow-upb | Baseline FVCc | Baseline DLCOc |
|---|---|---|---|---|---|---|---|---|---|
| Thiebaut et al. [13] | Retrospective | 47.3 (4.9) | 12 (4.1) | 7/6 | NR | NR | 24 months | 76.8 (5.5) (7) | 52.3 (4.3) (7) |
| Sari et al. [27] | Retrospective | 50.1 (11.4) | 9.8 (4.8) | NR | 13/1 | 5.7d | 15 months | 53.7 (12.6) (14) | NR |
| Lepri et al. [34] | Retrospective | 58.8 (10.4) | 7.9 (7.7) | NR | 20/3 | 2.8 | 24 months | 83 (9.5) (23) | 53.6 (7.1) (23) |
| Jordan et al. [4] | Prospective | 47.8 (4.6) | 7 (2.3) | NR | 5/4 | NR | 6 months | 60.6 (2.4) (9) | 41.1 (2.8) (9) |
| Fraticelli et al. [31] | Prospective | 54.3 (10.4) | 2.6 (2.2) | 9/6 | 9/6 | 2.8d | 12 months | 82.4 (17.3) (15) | 61.2 (16.6) (15) |
| Daoussis et al. [29] | Prospective | 54.3 (14.3) | 5.7 (6.7) | 30/3 | 25/8 | ad | 48 months | 80.6 (21.2) (33) | 59.2 (18.2) (33) |
| Daoussis et al. [28] | RCT | 53.4 (7.4) | 6.9 (4.9) | 8/0 | 6/2 | a | 12 months | 68.1 (19.7) (8) | 52.2 (20.7) (8) |
| Bosello et al. [25] | Prospective | 41.4 (13.1) | 2.5 (2.9) | 6/0 | NR | 2d | 48.5 months | 64.3 (13.5) (6) | 46 (9.4) (6) |
| Sircar et al. [9] | RCT | 34.67 (8.1) | 1.8 (0.7) | 30/0 | 25/5 | 2 | 6 months | 61.3 (11.3) (30) | NR |
| Elhai et al. [1] | Prospective | NR | NR | NR | NR | NR | 24 months | 76.3 (19.1) (146) | 54.4 (18.8) (146) |
| Fitzgerald et al. [2] | Retrospective | 54 (15.8) | NR | 1/5 | 1/5 | 2 (4)d, a (2) | 12 months | 74.9 (8.7) (6) | 47.5 (20.2) (6) |
| Ananieva et al. [23]e | Prospective | 46 (8.2) | 5.6 (4.4) | 42/29 | NR | 1.43 | 13 months | 77.4 (19.9) (71) | NR |
| Ananieva et al. [24]e | Prospective | 45.7 (13) | 7.6 (4) | NR | 26/1 | 1.48 | 12 months | 70 (17.8) (27) | 38.3 (14) (27) |
| Campochiaro et al. [27]e | Retrospective | 50.5 (11.8) | 7.6 (4.8) | NR | 20/1 | 2 | 6 months | 89.7 (21.1) (21) | 62.3 (21) (21) |
| Desinova et al. [30]e | NR | 50.6 (9.7) | 5.2 (4.7) | 8/1 | 7/2 | 2.72 | 12 months | 76.6 (9.6) (9) | 48.5 (9.5) (9) |
| Koneva et al. [33]e | NR | 48 (2) | 6.6 (5.9) | 25/17 | 36/6 | 2.5 | 29 months | 73.2 (18.8) (42) | 42.6 (15.7) (42) |
| Koneva et al. [32]e | NR | 48.5 (12.9) | 6.5 (5.7) | 31/23 | 44/10 | 2.1 | 25 months | 74.5 (19.6) (54) | 47.4 (22.1) (54) |
| Onat et al. [36]e | NR | 58.1 (11.8) | 2.5 (0.4) | 21/18 | 37/2 | 2 | 12 months | 86.4 (21) (39) | 64.2 (17.9) (39) |
| Moazedi-Fuerst et al. [35] | Prospective | 43 (6.4) | 5.3 (2.1) | 5/0 | 5/0 | 2 | 12 months | 72 (5.2) (5) | 48.5 (6.7) (5) |
| Vilela et al. [26] | Retrospective | 42.9 (11.6) | 6.4 (4.1) | 10/0 | 9/1 | 2 | 6 months | 66.4 (17.9) (10) | NR |
| Study | Design | Age, years | Duration, years | dcSSc/lcSSc | F/M | Dosea | Follow-upb | Baseline FVCc | Baseline DLCOc |
|---|---|---|---|---|---|---|---|---|---|
| Thiebaut et al. [13] | Retrospective | 47.3 (4.9) | 12 (4.1) | 7/6 | NR | NR | 24 months | 76.8 (5.5) (7) | 52.3 (4.3) (7) |
| Sari et al. [27] | Retrospective | 50.1 (11.4) | 9.8 (4.8) | NR | 13/1 | 5.7d | 15 months | 53.7 (12.6) (14) | NR |
| Lepri et al. [34] | Retrospective | 58.8 (10.4) | 7.9 (7.7) | NR | 20/3 | 2.8 | 24 months | 83 (9.5) (23) | 53.6 (7.1) (23) |
| Jordan et al. [4] | Prospective | 47.8 (4.6) | 7 (2.3) | NR | 5/4 | NR | 6 months | 60.6 (2.4) (9) | 41.1 (2.8) (9) |
| Fraticelli et al. [31] | Prospective | 54.3 (10.4) | 2.6 (2.2) | 9/6 | 9/6 | 2.8d | 12 months | 82.4 (17.3) (15) | 61.2 (16.6) (15) |
| Daoussis et al. [29] | Prospective | 54.3 (14.3) | 5.7 (6.7) | 30/3 | 25/8 | ad | 48 months | 80.6 (21.2) (33) | 59.2 (18.2) (33) |
| Daoussis et al. [28] | RCT | 53.4 (7.4) | 6.9 (4.9) | 8/0 | 6/2 | a | 12 months | 68.1 (19.7) (8) | 52.2 (20.7) (8) |
| Bosello et al. [25] | Prospective | 41.4 (13.1) | 2.5 (2.9) | 6/0 | NR | 2d | 48.5 months | 64.3 (13.5) (6) | 46 (9.4) (6) |
| Sircar et al. [9] | RCT | 34.67 (8.1) | 1.8 (0.7) | 30/0 | 25/5 | 2 | 6 months | 61.3 (11.3) (30) | NR |
| Elhai et al. [1] | Prospective | NR | NR | NR | NR | NR | 24 months | 76.3 (19.1) (146) | 54.4 (18.8) (146) |
| Fitzgerald et al. [2] | Retrospective | 54 (15.8) | NR | 1/5 | 1/5 | 2 (4)d, a (2) | 12 months | 74.9 (8.7) (6) | 47.5 (20.2) (6) |
| Ananieva et al. [23]e | Prospective | 46 (8.2) | 5.6 (4.4) | 42/29 | NR | 1.43 | 13 months | 77.4 (19.9) (71) | NR |
| Ananieva et al. [24]e | Prospective | 45.7 (13) | 7.6 (4) | NR | 26/1 | 1.48 | 12 months | 70 (17.8) (27) | 38.3 (14) (27) |
| Campochiaro et al. [27]e | Retrospective | 50.5 (11.8) | 7.6 (4.8) | NR | 20/1 | 2 | 6 months | 89.7 (21.1) (21) | 62.3 (21) (21) |
| Desinova et al. [30]e | NR | 50.6 (9.7) | 5.2 (4.7) | 8/1 | 7/2 | 2.72 | 12 months | 76.6 (9.6) (9) | 48.5 (9.5) (9) |
| Koneva et al. [33]e | NR | 48 (2) | 6.6 (5.9) | 25/17 | 36/6 | 2.5 | 29 months | 73.2 (18.8) (42) | 42.6 (15.7) (42) |
| Koneva et al. [32]e | NR | 48.5 (12.9) | 6.5 (5.7) | 31/23 | 44/10 | 2.1 | 25 months | 74.5 (19.6) (54) | 47.4 (22.1) (54) |
| Onat et al. [36]e | NR | 58.1 (11.8) | 2.5 (0.4) | 21/18 | 37/2 | 2 | 12 months | 86.4 (21) (39) | 64.2 (17.9) (39) |
| Moazedi-Fuerst et al. [35] | Prospective | 43 (6.4) | 5.3 (2.1) | 5/0 | 5/0 | 2 | 12 months | 72 (5.2) (5) | 48.5 (6.7) (5) |
| Vilela et al. [26] | Retrospective | 42.9 (11.6) | 6.4 (4.1) | 10/0 | 9/1 | 2 | 6 months | 66.4 (17.9) (10) | NR |
F: female; FVC: percentage of predicted FVC at the time of initiation of RTX therapy; M: male; NR: not reported. aDose of RTX was either reported in total dose received over the first year (entered as a numeral indicating grams/person of RTX) or regimen of RTX given (375 mg/m2, 4 such) which is entered as ‘a’. Numbers within parentheses indicate the number of subjects receiving a certain dose. bThis represents the duration of follow-up reported in the article. For the meta-analysis, data were extracted for the 12 month time point or the end of follow-up, whichever was earlier. cNumbers within parentheses indicate the number of subjects for whom data could be extracted.
These studies had patients with re-treatment with RTX. Specifically the following: Sari et al. [27]: 12 patients received re-treatment at 6 month intervals; Daoussis et al. [29]: although all patients were re-treated, none were re-treated with RTX before the first year of follow-up was complete; Bosello et al. [25]: 8 patients were re-treated (timing not mentioned); Fraticelli et al. [31]: all patients re-treated at 6 months; Fitzgerald et al. [2]: 2 patients were re-treated (exact time point not mentioned). eConference abstract.
Study characteristics
| Study | Design | Age, years | Duration, years | dcSSc/lcSSc | F/M | Dosea | Follow-upb | Baseline FVCc | Baseline DLCOc |
|---|---|---|---|---|---|---|---|---|---|
| Thiebaut et al. [13] | Retrospective | 47.3 (4.9) | 12 (4.1) | 7/6 | NR | NR | 24 months | 76.8 (5.5) (7) | 52.3 (4.3) (7) |
| Sari et al. [27] | Retrospective | 50.1 (11.4) | 9.8 (4.8) | NR | 13/1 | 5.7d | 15 months | 53.7 (12.6) (14) | NR |
| Lepri et al. [34] | Retrospective | 58.8 (10.4) | 7.9 (7.7) | NR | 20/3 | 2.8 | 24 months | 83 (9.5) (23) | 53.6 (7.1) (23) |
| Jordan et al. [4] | Prospective | 47.8 (4.6) | 7 (2.3) | NR | 5/4 | NR | 6 months | 60.6 (2.4) (9) | 41.1 (2.8) (9) |
| Fraticelli et al. [31] | Prospective | 54.3 (10.4) | 2.6 (2.2) | 9/6 | 9/6 | 2.8d | 12 months | 82.4 (17.3) (15) | 61.2 (16.6) (15) |
| Daoussis et al. [29] | Prospective | 54.3 (14.3) | 5.7 (6.7) | 30/3 | 25/8 | ad | 48 months | 80.6 (21.2) (33) | 59.2 (18.2) (33) |
| Daoussis et al. [28] | RCT | 53.4 (7.4) | 6.9 (4.9) | 8/0 | 6/2 | a | 12 months | 68.1 (19.7) (8) | 52.2 (20.7) (8) |
| Bosello et al. [25] | Prospective | 41.4 (13.1) | 2.5 (2.9) | 6/0 | NR | 2d | 48.5 months | 64.3 (13.5) (6) | 46 (9.4) (6) |
| Sircar et al. [9] | RCT | 34.67 (8.1) | 1.8 (0.7) | 30/0 | 25/5 | 2 | 6 months | 61.3 (11.3) (30) | NR |
| Elhai et al. [1] | Prospective | NR | NR | NR | NR | NR | 24 months | 76.3 (19.1) (146) | 54.4 (18.8) (146) |
| Fitzgerald et al. [2] | Retrospective | 54 (15.8) | NR | 1/5 | 1/5 | 2 (4)d, a (2) | 12 months | 74.9 (8.7) (6) | 47.5 (20.2) (6) |
| Ananieva et al. [23]e | Prospective | 46 (8.2) | 5.6 (4.4) | 42/29 | NR | 1.43 | 13 months | 77.4 (19.9) (71) | NR |
| Ananieva et al. [24]e | Prospective | 45.7 (13) | 7.6 (4) | NR | 26/1 | 1.48 | 12 months | 70 (17.8) (27) | 38.3 (14) (27) |
| Campochiaro et al. [27]e | Retrospective | 50.5 (11.8) | 7.6 (4.8) | NR | 20/1 | 2 | 6 months | 89.7 (21.1) (21) | 62.3 (21) (21) |
| Desinova et al. [30]e | NR | 50.6 (9.7) | 5.2 (4.7) | 8/1 | 7/2 | 2.72 | 12 months | 76.6 (9.6) (9) | 48.5 (9.5) (9) |
| Koneva et al. [33]e | NR | 48 (2) | 6.6 (5.9) | 25/17 | 36/6 | 2.5 | 29 months | 73.2 (18.8) (42) | 42.6 (15.7) (42) |
| Koneva et al. [32]e | NR | 48.5 (12.9) | 6.5 (5.7) | 31/23 | 44/10 | 2.1 | 25 months | 74.5 (19.6) (54) | 47.4 (22.1) (54) |
| Onat et al. [36]e | NR | 58.1 (11.8) | 2.5 (0.4) | 21/18 | 37/2 | 2 | 12 months | 86.4 (21) (39) | 64.2 (17.9) (39) |
| Moazedi-Fuerst et al. [35] | Prospective | 43 (6.4) | 5.3 (2.1) | 5/0 | 5/0 | 2 | 12 months | 72 (5.2) (5) | 48.5 (6.7) (5) |
| Vilela et al. [26] | Retrospective | 42.9 (11.6) | 6.4 (4.1) | 10/0 | 9/1 | 2 | 6 months | 66.4 (17.9) (10) | NR |
| Study | Design | Age, years | Duration, years | dcSSc/lcSSc | F/M | Dosea | Follow-upb | Baseline FVCc | Baseline DLCOc |
|---|---|---|---|---|---|---|---|---|---|
| Thiebaut et al. [13] | Retrospective | 47.3 (4.9) | 12 (4.1) | 7/6 | NR | NR | 24 months | 76.8 (5.5) (7) | 52.3 (4.3) (7) |
| Sari et al. [27] | Retrospective | 50.1 (11.4) | 9.8 (4.8) | NR | 13/1 | 5.7d | 15 months | 53.7 (12.6) (14) | NR |
| Lepri et al. [34] | Retrospective | 58.8 (10.4) | 7.9 (7.7) | NR | 20/3 | 2.8 | 24 months | 83 (9.5) (23) | 53.6 (7.1) (23) |
| Jordan et al. [4] | Prospective | 47.8 (4.6) | 7 (2.3) | NR | 5/4 | NR | 6 months | 60.6 (2.4) (9) | 41.1 (2.8) (9) |
| Fraticelli et al. [31] | Prospective | 54.3 (10.4) | 2.6 (2.2) | 9/6 | 9/6 | 2.8d | 12 months | 82.4 (17.3) (15) | 61.2 (16.6) (15) |
| Daoussis et al. [29] | Prospective | 54.3 (14.3) | 5.7 (6.7) | 30/3 | 25/8 | ad | 48 months | 80.6 (21.2) (33) | 59.2 (18.2) (33) |
| Daoussis et al. [28] | RCT | 53.4 (7.4) | 6.9 (4.9) | 8/0 | 6/2 | a | 12 months | 68.1 (19.7) (8) | 52.2 (20.7) (8) |
| Bosello et al. [25] | Prospective | 41.4 (13.1) | 2.5 (2.9) | 6/0 | NR | 2d | 48.5 months | 64.3 (13.5) (6) | 46 (9.4) (6) |
| Sircar et al. [9] | RCT | 34.67 (8.1) | 1.8 (0.7) | 30/0 | 25/5 | 2 | 6 months | 61.3 (11.3) (30) | NR |
| Elhai et al. [1] | Prospective | NR | NR | NR | NR | NR | 24 months | 76.3 (19.1) (146) | 54.4 (18.8) (146) |
| Fitzgerald et al. [2] | Retrospective | 54 (15.8) | NR | 1/5 | 1/5 | 2 (4)d, a (2) | 12 months | 74.9 (8.7) (6) | 47.5 (20.2) (6) |
| Ananieva et al. [23]e | Prospective | 46 (8.2) | 5.6 (4.4) | 42/29 | NR | 1.43 | 13 months | 77.4 (19.9) (71) | NR |
| Ananieva et al. [24]e | Prospective | 45.7 (13) | 7.6 (4) | NR | 26/1 | 1.48 | 12 months | 70 (17.8) (27) | 38.3 (14) (27) |
| Campochiaro et al. [27]e | Retrospective | 50.5 (11.8) | 7.6 (4.8) | NR | 20/1 | 2 | 6 months | 89.7 (21.1) (21) | 62.3 (21) (21) |
| Desinova et al. [30]e | NR | 50.6 (9.7) | 5.2 (4.7) | 8/1 | 7/2 | 2.72 | 12 months | 76.6 (9.6) (9) | 48.5 (9.5) (9) |
| Koneva et al. [33]e | NR | 48 (2) | 6.6 (5.9) | 25/17 | 36/6 | 2.5 | 29 months | 73.2 (18.8) (42) | 42.6 (15.7) (42) |
| Koneva et al. [32]e | NR | 48.5 (12.9) | 6.5 (5.7) | 31/23 | 44/10 | 2.1 | 25 months | 74.5 (19.6) (54) | 47.4 (22.1) (54) |
| Onat et al. [36]e | NR | 58.1 (11.8) | 2.5 (0.4) | 21/18 | 37/2 | 2 | 12 months | 86.4 (21) (39) | 64.2 (17.9) (39) |
| Moazedi-Fuerst et al. [35] | Prospective | 43 (6.4) | 5.3 (2.1) | 5/0 | 5/0 | 2 | 12 months | 72 (5.2) (5) | 48.5 (6.7) (5) |
| Vilela et al. [26] | Retrospective | 42.9 (11.6) | 6.4 (4.1) | 10/0 | 9/1 | 2 | 6 months | 66.4 (17.9) (10) | NR |
F: female; FVC: percentage of predicted FVC at the time of initiation of RTX therapy; M: male; NR: not reported. aDose of RTX was either reported in total dose received over the first year (entered as a numeral indicating grams/person of RTX) or regimen of RTX given (375 mg/m2, 4 such) which is entered as ‘a’. Numbers within parentheses indicate the number of subjects receiving a certain dose. bThis represents the duration of follow-up reported in the article. For the meta-analysis, data were extracted for the 12 month time point or the end of follow-up, whichever was earlier. cNumbers within parentheses indicate the number of subjects for whom data could be extracted.
These studies had patients with re-treatment with RTX. Specifically the following: Sari et al. [27]: 12 patients received re-treatment at 6 month intervals; Daoussis et al. [29]: although all patients were re-treated, none were re-treated with RTX before the first year of follow-up was complete; Bosello et al. [25]: 8 patients were re-treated (timing not mentioned); Fraticelli et al. [31]: all patients re-treated at 6 months; Fitzgerald et al. [2]: 2 patients were re-treated (exact time point not mentioned). eConference abstract.
The average age of patients was 48.8 years (95% CI 46.5, 51.22) from 19 studies (n = 429) [2, 4, 9, 13, 23–37]. The average duration of disease was 5.74 years (95% CI 4.87, 6.61) from 18 studies (n = 423) [4, 9, 13, 23–37]. The proportion of female patients was 83.6% (95% CI 77.3, 90) from 16 studies (n = 345) [2, 4, 9, 24, 26–37]. Fourteen studies reported on the subtype of scleroderma: 233 with dcSSc and 108 with lcSSc [2, 9, 13, 23, 25, 26, 28–33, 35, 36].
Failure of previous immunosuppression before introduction of RTX was reported in 12 studies: 83 patients received CYC, 27 patients received MMF, 9 received MTX, 4 received AZA and one each received tocilizumab and etanercept before initiation of RTX therapy [2, 9, 26–31, 34–37]. One study reported that seven patients were treated with previous immunosuppression, but the nature was not disclosed [26]. One study reported that none of the patients were previously treated with any immunosuppressive agents [9].
Eight studies (n = 133) reported that concomitant prednisolone was used [2, 9, 27–29, 31, 33, 37]. Two studies reported that no concomitant steroids were used [26, 35]. No concomitant immunosuppressive agents were used in three studies [9, 31, 37]. Eleven other studies reported the use of other immunosuppressive agents [2, 9, 26–31, 34–37]. Among these, two studies (n = 25) did not specify the agents [33, 34] and the rest reported the following: MMF in 63, CYC in 11, AZA in 10, MTX in 9 and HCQ in 1. These studies were conducted in 12 countries (Italy, France, Spain, Switzerland, Austria, Australia, Turkey, India, Greece, Russia, Ireland and Brazil) [2, 9, 13, 23–37] and two were reports from the EUSTAR group collecting data from multiple participating European Union countries [1, 4]. Two were RCTs [9, 28], six were prospective cohorts [1, 4, 25, 29, 31, 35], five were retrospective studies [2, 13, 26, 34, 37] and seven were conference abstracts [23, 24, 27, 30, 32, 33, 36]. The efficacy of RTX could be compared with matched controls at the 6 month time point in two studies and at the 12 month time point in four studies [1, 4, 9, 13, 28, 29]. Among these, three studies reported on concomitant immunosuppressive use: MMF in 49, MTX in 2 and HCQ in 1 [1, 28, 29]. Sircar et al. [8] reported that no other immunosuppressive agent was used concomitantly. Control drugs were as follows: only prednisolone in 3, CYC in 32, MMF in 12, AZA in 2 and MTX in 6 [8, 27, 28]. However, three studies did not mention the exact agent used in the control arms [1, 4, 9].
Efficacy of RTX at 6 months on PFT
Five studies (n = 75) reported data on the change in FVC from baseline to 6 months [4, 9, 26, 27, 35]. The pooled meta-analysis indicated that there was moderate statistical heterogeneity between the included studies (I2 = 44%, P = 0.13), so a random effects model was adopted to determine the effect size. The pooled mean difference of FVC (FVC at 6 months − FVC at baseline; Fig. 2) showed that the administration of RTX increased FVC by 4.49% (95% CI 0.25, 8.73). Two studies (n = 30) reported data on the change in DLCO from baseline to 6 months (Supplementary Fig. S1, available at Rheumatology online) [4, 27]. In the absence of any significant statistical heterogeneity (I2 = 0%, P = 0.43), a fixed effects model was used that yielded a significant increase in DLCO of 3.47% (95% CI 0.99, 5.96).
Forest plot of the change in FVC in the RTX group by 6 months
MD: mean difference.
Efficacy of RTX at 12 months on PFT
Sixteen studies (n = 505) reported data on the change in FVC from baseline to 12 months [1, 2, 13, 23–25, 28–37]. There was a significant increase in FVC at 12 months compared with baseline [7.03% (95% CI 4.37, 9.7), I2 = 30%, P = 0.13; Fig. 3] in the random effects model. Fourteen studies (n = 415) reported data on the change in DLCO from baseline to 12 months (Supplementary Fig. S2, available at Rheumatology online) [1, 2, 13, 23–25, 28–36]. The random effects model yielded a significant increase in DLCO of 4.08% (95% CI 1.51, 6.65; I2 = 32%, P = 0.12).
Forest plot of the change in FVC in the RTX group by 12 months
MD: mean difference.
Efficacy of RTX compared with matched controls in terms of FVC at 6 and 12 months
Two studies compared the relative change in FVC (ΔFVCRTX − ΔFVCControl) from baseline to 6 months in the RTX group compared with a matched control group (n = 39 for both) [4, 9] (Fig. 4). There was a significant relative increase in FVC at 6 months in the RTX group compared with the control group [1.03% (95% CI 0.11, 1.94), I2 = 93%, P < 0.001). Four studies (n = 194 in the RTX group and n = 535 in the control group) reported data on the relative change in FVC (ΔFVCRTX − ΔFVCControl) and DLCO (ΔDLCORTX − ΔDLCOControl) from baseline to 12 months (Supplementary Figs S3 and S4, available at Rheumatology online) [1, 13, 28, 29]. The random effects model yielded a relative increase in FVC of 0.13% (95% CI −0.5, 0.31; I2 = 64%, P = 0.04) from baseline to 12 months and a relative decrease in DLCO of 0.44% (95% CI −1.09–0.2; I2 = 98%, P ≤ 0.01). No studies compared the relative change of DLCO from baseline to 6 months.
Forest plot of the change in FVC from baseline in the RTX vs control group at 6 months
MD: mean difference.
Efficacy of RTX on mRSS
Five studies compared the change in mRSS from baseline to 6 months on RTX (n = 96) [9, 25–27, 31] (Supplementary Fig. S5, available at Rheumatology online). A significant decrease in the mRSS was observed in the fixed effects model [−6.92 (95% CI −9.61, −4.24), I2 = 0%, P = 0.47). Eight studies compared the change in the mRSS from baseline to 12 months on RTX (n = 141) [13, 25, 28–30, 35–37] (Supplementary Fig. S6, available at Rheumatology online). A significant decrease in the mRSS was observed in the random effects model [−6.31 (95% CI −9.62, −2.99), I2 = 68%, P < 0.0.01).
Sensitivity analysis
We performed a sensitivity analysis of the PFT results described above by removing conference abstracts in the first step and then retrospective studies in the second step. The results are summarized in Table 2 and in Supplementary Figs S7–S14, available at Rheumatology online.
Outcome of the sensitivity analysis
| Outcome measure | Sensitivity steps | References | RTX group, n | Control group, n | MD | 95% CI | I2, P-value |
|---|---|---|---|---|---|---|---|
| Change of FVC from baseline to 12 months | 1 | [1, 2, 13, 25, 28–30, 34, 35, 37] | 260 | NA | 6.92 | 2.28, 11.54 | 56%, 0.01 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 7.23 | −0.14, 14.59 | 75%, <0.01 | |
| Change of DLCO from baseline to 12 months | 1 | [1, 2, 13, 25, 28, 29, 31, 34, 35] | 244 | NA | 5.21 | 0.91, 9.52 | 54%, 0.03 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 5.46 | −1.25, 12.16 | 71%, <0.01 | |
| Change of mRSS from baseline to 12 months | 1 | [13, 25, 28, 29, 37] | 93 | NA | −7.02 | −11.36, −2.68 | 65%, 0.01 |
| 2 | [25, 28, 29, 35] | 66 | NA | −8.78 | −12.84, −4.73 | 48%, 0.12 | |
| Change in FVC from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | 0.1 | −0.09, 0.29 | 68%, 0.04 |
| Change in DLCO from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | −0.68 | −1.33, −0.03 | 97%, <0.01 |
| Outcome measure | Sensitivity steps | References | RTX group, n | Control group, n | MD | 95% CI | I2, P-value |
|---|---|---|---|---|---|---|---|
| Change of FVC from baseline to 12 months | 1 | [1, 2, 13, 25, 28–30, 34, 35, 37] | 260 | NA | 6.92 | 2.28, 11.54 | 56%, 0.01 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 7.23 | −0.14, 14.59 | 75%, <0.01 | |
| Change of DLCO from baseline to 12 months | 1 | [1, 2, 13, 25, 28, 29, 31, 34, 35] | 244 | NA | 5.21 | 0.91, 9.52 | 54%, 0.03 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 5.46 | −1.25, 12.16 | 71%, <0.01 | |
| Change of mRSS from baseline to 12 months | 1 | [13, 25, 28, 29, 37] | 93 | NA | −7.02 | −11.36, −2.68 | 65%, 0.01 |
| 2 | [25, 28, 29, 35] | 66 | NA | −8.78 | −12.84, −4.73 | 48%, 0.12 | |
| Change in FVC from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | 0.1 | −0.09, 0.29 | 68%, 0.04 |
| Change in DLCO from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | −0.68 | −1.33, −0.03 | 97%, <0.01 |
Sensitivity steps: 1: meta-analysis after removal of conference abstracts; 2: meta-analysis after removal of retrospective studies. MD: mean difference.
Outcome of the sensitivity analysis
| Outcome measure | Sensitivity steps | References | RTX group, n | Control group, n | MD | 95% CI | I2, P-value |
|---|---|---|---|---|---|---|---|
| Change of FVC from baseline to 12 months | 1 | [1, 2, 13, 25, 28–30, 34, 35, 37] | 260 | NA | 6.92 | 2.28, 11.54 | 56%, 0.01 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 7.23 | −0.14, 14.59 | 75%, <0.01 | |
| Change of DLCO from baseline to 12 months | 1 | [1, 2, 13, 25, 28, 29, 31, 34, 35] | 244 | NA | 5.21 | 0.91, 9.52 | 54%, 0.03 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 5.46 | −1.25, 12.16 | 71%, <0.01 | |
| Change of mRSS from baseline to 12 months | 1 | [13, 25, 28, 29, 37] | 93 | NA | −7.02 | −11.36, −2.68 | 65%, 0.01 |
| 2 | [25, 28, 29, 35] | 66 | NA | −8.78 | −12.84, −4.73 | 48%, 0.12 | |
| Change in FVC from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | 0.1 | −0.09, 0.29 | 68%, 0.04 |
| Change in DLCO from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | −0.68 | −1.33, −0.03 | 97%, <0.01 |
| Outcome measure | Sensitivity steps | References | RTX group, n | Control group, n | MD | 95% CI | I2, P-value |
|---|---|---|---|---|---|---|---|
| Change of FVC from baseline to 12 months | 1 | [1, 2, 13, 25, 28–30, 34, 35, 37] | 260 | NA | 6.92 | 2.28, 11.54 | 56%, 0.01 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 7.23 | −0.14, 14.59 | 75%, <0.01 | |
| Change of DLCO from baseline to 12 months | 1 | [1, 2, 13, 25, 28, 29, 31, 34, 35] | 244 | NA | 5.21 | 0.91, 9.52 | 54%, 0.03 |
| 2 | [1, 25, 28, 29, 31, 35] | 213 | NA | 5.46 | −1.25, 12.16 | 71%, <0.01 | |
| Change of mRSS from baseline to 12 months | 1 | [13, 25, 28, 29, 37] | 93 | NA | −7.02 | −11.36, −2.68 | 65%, 0.01 |
| 2 | [25, 28, 29, 35] | 66 | NA | −8.78 | −12.84, −4.73 | 48%, 0.12 | |
| Change in FVC from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | 0.1 | −0.09, 0.29 | 68%, 0.04 |
| Change in DLCO from baseline to 12 months compared with controls | 2 | [1, 28, 29] | 187 | 518 | −0.68 | −1.33, −0.03 | 97%, <0.01 |
Sensitivity steps: 1: meta-analysis after removal of conference abstracts; 2: meta-analysis after removal of retrospective studies. MD: mean difference.
Publication bias
Publication bias was estimated with funnel plots and Begg and Mazumdar test (Supplementary Table S3 and Supplementary Figs S15–S17, available at Rheumatology online). All P-values were >0.05, suggesting an absence of significant publication bias.
Adverse events
There were five studies including 112 patients that reported adverse events related to RTX treatment [9, 24, 28–30, 37]. The pooled incidence of infectious adverse events was 13.9% (95% CI 7.8, 20; I2 = 0%, P = 0.91; Supplementary Fig. S18, available at Rheumatology online). Of these, respiratory tract infections were most common, occurring in 12 patients, followed by herpes zoster reactivation in 3 patients, urinary tract infection in 2 patients and hepatitis B virus reactivation in 1 patient. Infusion reaction was reported in four studies, occurring in 7 of 104 patients [9, 24, 29, 30, 37]. Among the three patients with herpes zoster reactivation, one patient was on 10 mg/day of prednisolone [9] and another on 10–15 mg/day of prednisolone [30]. The third study did not report on the use of concomitant glucocorticoids [24]. One study reported two events of ECG changes (ST depression), although attribution to RTX was not clear [24]. Death was reported in 6 of 39 patients in two studies [2, 29]. Five deaths were reported during the follow-up period, none of which could be reliably attributed specifically to RTX. Two studies reported a comparison of adverse events in patients on RTX vs matched controls (63 patients in the RTX group and 48 in the control group) [9, 29]. The odds ratio (OR) of infectious adverse effects was significantly in favour of RTX compared with controls [OR 0.256 (95% CI 0.104, 0.626), I2 = 0%, P = 0.47; Supplementary Fig. S19, available at Rheumatology online].
Discussion
Effective treatment of SSc-ILD remains a challenge despite advancements in the current therapeutic armamentarium. MMF and CYC are widely used and acceptable choices [6, 7, 38]. CYC, generally the treatment of choice, stabilizes FVC. However, its impact on the DLCO is less remarkable [12, 38, 39]. Nintedanib is the first drug to be approved by the US Food and Drug Administration for SSc-ILD after publication of the SENSCIS trial [10], in which a combination of immunosuppressive and anti-fibrotic therapy (rather an immunomodulatory therapy) was shown to be useful in the management of SSc-ILD. However, although nintedanib arrested the progression of ILD, an annual decrease of 52.4 ml/year in FVC was noted in the nintedanib group [10]. RTX has been used in numerous clinical studies, including two RCTs for the treatment of both aggressive treatment-resistant and naïve ILD in patients with SSc. However, its exact role in the management of SSc-ILD is yet to be determined. A few studies showed that RTX could actually increase FVC in patients with SSc-ILD [9, 28, 29]. Our meta-analysis aimed to quantitatively determine the effect of RTX on pulmonary function in patients with SSc-ILD and whether this effect is more than that of other drugs within the first year of treatment.
The meta-analysis demonstrated that RTX improved FVC from baseline by 4.49% (95% CI 0.25, 8.73) at 6 months and by 7.03% (95% CI 4.37, 9.7) at 12 months. Similarly, RTX improved DLCO by 3.47% (95% CI 0.99, 5.96) at 6 months and 4.08% (95% CI 1.51, 6.65) at 12 months. Both these results were supported by a commensurate decrease in the mRSS. These results were stable in the sensitivity analyses as well. Improvement of FVC at 6 months over baseline was 1.03% (95% CI 0.11, 1.94) more compared with controls. However, at the 12 month time point, RTX treatment had similar effects on both FVC and DLCO compared with controls, although the effect on FVC approached statistical significance. However, the effect of RTX on DLCO was not stable in the sensitivity analysis and, compared with controls, was worse with RTX and therefore should be regarded with caution. There was very high heterogeneity in the matched control analysis for DLCO (97%) in sensitivity step 2. This makes interpretation difficult. In terms of infectious adverse events, RTX in pooled analysis had a significantly lower OR compared with controls, with very little heterogeneity. Regarding comparison of adverse effects, including infectious adverse events, two studies were included, namely Sircar et al. [9] and Doussis et al. [29]. In the former, both the control and RTX groups were matched and both groups received 10 mg/day prednisolone and the control group received i.v. CYC. In the latter study, a disproportionately higher proportion of patients received prednisolone (94.4% vs 54.5%, P = 0.0037), MTX (33.3% vs 6.1%, P = 0.011), MMF (55.5% vs 30.3%, P = 0.08) and AZA (11.1% vs 0, P = 0.05). Therefore it appears that increased infectious adverse effects observed in the control arms were mostly due to increased receipt of glucocorticoids and cytotoxic agents like CYC, MMF or AZA.
In the Cochrane meta-analysis of CYC in SSc-ILD, an increase in FVC of 2.83% was demonstrated compared with placebo, but no significant difference compared with MMF was noted [12]. In the Scleroderma Lung Study II (SLS II) trial, improvement in FVC at the end of trial was 2.19% for MMF and 2.88% for CYC [7]. In the current analysis, the improvement in FVC on RTX (6 months, 4.49%; 12 months, 7.03%) appears to be higher. It is worth noting that SLS II was an RCT comparing MMF and CYC, with a predefined patient population and extensive inclusion and exclusion criteria. Hence it is imperative that the degree of improvement in FVC as noted in our meta-analysis needs to be corroborated independently by a robustly designed RCT.
The magnitude of increase in FVC by 12 months is also clinically significant. The range of the minimal clinically important difference for FVC at 12 months is reported as 3.0–5.3%, which is attained with RTX therapy at 6 months and is further improved until 12 months [40].
The use of RTX in SSc-ILD is supported by the pathophysiological aspects of the disease. B cells play a central role in SSc and its various manifestations. Antibodies like anti-Scl-70 and anti-centromere are both diagnostically and prognostically useful. The tight-skin mouse model (tsk1) of human SSc demonstrates overexpression of hyperresponsive, autoantibody-producing CD19+ B cells [41]. In humans, an increased number of naïve B cells and reduced numbers of memory B cells and plasmablasts are said to be characteristic of SSc. The memory B cells overexpress CD80 and CD86, which are co-stimulatory molecules and activation markers [42]. Abnormal B cell signalling in tsk1 and the presence of B cells and B cell gene expression in skin biopsies from human SSc patients also suggest this pathogenic model. B cells, a source of IL-6, also contribute to the pathogenesis of fibrosis by several other mechanisms [43].
In the present study, the increase in FVC after initial RTX was first appreciated at 6 months and increased further until the end of first year. However, the magnitude of increase of DLCO achieved at the 6 month time point did not further increase at 12 months. After an initial exposure, B cells are nearly completely depleted by 3 months and nearly completely repopulated within 6 and 12 months [43]. Therefore it is expected that some blunting of the effect of initial RTX should be observed at the 12 month time point. The DLCO is generally considered to be more sensitive to pulmonary involvement compared with the FVC but its interpretation is confounded by other factors like concomitant obstructive airway disease or the presence of pulmonary arterial hypertension (PAH), which was not analysed (due to underreporting) in this study [44]. It is not known whether PAH, if present at baseline, improves in patients with SSc with immunosuppression. An RCT previously reported no beneficial effect of RTX on PAH in SSc [9]. However, this study was not powered to detect a change in PAH. A recently published study reported on adjuvant B cell depletion therapy and concluded that B cell depletion is a potentially effective and safe treatment for SSc-PAH [45]. Asymmetric representation of PAH in the RTX arm may explain the behaviour of the DLCO. This also brings into question whether the RTX should be repeated at 6 month intervals, considering the B cell repopulation effect seen after the 6 month time point.
It may be argued that in the present meta-analysis, all studies indicating statistically significant improvement in FVC and DLCO are at a group level and often may not fully reflect clinically significant changes. However, when considering longitudinal spirometric measurements, a wide variability in within-person FVC often exists. It is also known that FVC can decrease by ∼0.2 l per decade, even for healthy people who have never smoked, and DLCO may decrease by 5% per decade [46, 47]. Therefore, taking this into consideration, it is reasonable to infer from the present study that RTX is associated with the stability of ILD in the short term and may be associated with a modest improvement at a group level.
Regarding adverse events, the association of fewer side effects, especially infectious ones, is important, considering that SSc-ILD patients are at high risk of pulmonary infections. CYC and MMF, broad-spectrum immunosuppressives, could lead to a reduced number of both B and T cells and increase the risk of serious infections. On the other hand, RTX reduces the number of B cells up to the stage of plasmablasts and does not alter the number of T cells or memory B cells [48]. Therefore infectious side effects should be less compared with broad-spectrum immunosuppressives.
This meta-analysis had several limitations. First, the number of patients included was small and the follow-up duration was kept within 1 year. The data on the duration of follow-up was not uniform and may add to bias. We could not compare between various treatment regimens of RTX and this might add some variability. Some of the studies used immunosuppressive agents, including steroids, with RTX and individual data were not available for analysis. This could potentially introduce bias. The duration of disease varied from study to study and it is acceptable that a drug might appear more efficacious in an early disease compared with its relatively poor showing with a longer duration. There was a dearth of RCTs. However, our results on FVC remained stable even after multiple sensitivity analyses.
Conclusion
To conclude, in this meta-analysis, RTX was associated with an improvement of both FVC and DLCO during the first year of treatment of SSc-ILD. The effect of RTX appeared to be superior compared with other immunosuppressives during the first 6 months of treatment in terms of FVC. RTX use was associated with lower ORs of infectious adverse events. Considering the limitations, further studies, including RCTs, comparing the effectiveness of the different regimens of RTX and combination therapy of RTX with another immunosuppressive agent vs RTX alone should be performed. Importantly, repeated courses of RTX given every 6 months and its relationship with DLCO in the subsequent 6 months should also be investigated.
Acknowledgements
R.P.G. takes responsibility for (is the guarantor of) the content of the article, including the data and analysis. R.P.G. was responsible for the study design; data acquisition, analysis and interpretation; writing and revision of the manuscript, final approval and accountability. A.R. was responsible for the study design, data interpretation, writing and revision of the manuscript, final approval and accountability. M.C. was responsible for data analysis and interpretation, writing and revision of the manuscript, final approval and accountability. A.M. was responsible for the study design, data interpretation, writing and revision of the manuscript, final approval and accountability. G.B.S. was responsible for the study design, data interpretation, writing and revision of the manuscript, final approval and accountability. P.G. was responsible for the study design, data interpretation, writing and revision of the manuscript, final approval and accountability.
Funding: No specific funding was received from any funding bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.
Disclosure statement: None of the authors have declared any conflicts of interest.
Data availability statement
Data is available upon request.
Supplementary data
Supplementary data are available at Rheumatology online.
Comments