Efficacy and safety of subcutaneous infliximab versus adalimumab, etanercept and intravenous infliximab in patients with rheumatoid arthritis: a systematic literature review and meta-analysis

ABSTRACT

Objectives

There are few comparative data for tumor necrosis factor inhibitors in patients with rheumatoid arthritis (RA).

Methods

Historical data for reference product/biosimilar intravenous infliximab, or adalimumab and etanercept, were pooled and compared with phase 3 study results for a subcutaneous (SC) formulation of the infliximab biosimilar CT-P13, in a systematic review and meta-analysis (PROSPERO: CRD42019149621).

Results

The authors identified 13 eligible controlled trials that randomized over 5400 participants to prespecified treatments of interest. Comparison with pooled historical data suggested a numerical advantage for CT-P13 SC over intravenous infliximab for almost every prespecified efficacy outcome evaluated, including Disease Activity Score in 28 joints (C-reactive protein/erythrocyte sedimentation rate), Clinical/Simplified Disease Activity Index scores, American College of Rheumatology responses, and multiple measures of disease remission and low disease activity; for the majority of outcomes, there was no overlap in 95% confidence intervals between groups. A numerical advantage for CT-P13 SC was also observed for safety outcomes (adverse events, infections, and discontinuations). Similar, but less marked, trends were observed for comparison with historical efficacy and safety data for adalimumab/etanercept.

Conclusion

CT-P13 SC offers an improved or similar benefit-to-harm ratio compared with infliximab (intravenous) and adalimumab/etanercept, for the treatment of moderate-to-severe RA.

KEYWORDS:

• Adalimumab
• biosimilar
• CT-P13
• etanercept
• infliximab
• rheumatoid arthritis

1. Introduction

Rheumatoid arthritis (RA) is the most common form of autoimmune arthritis, with a global, age-standardized annual incidence of 14.9 per 100,000 [1]. RA predominantly affects the synovial joints, potentially involving cartilage damage, bone erosion, and systemic consequences [2]. The impact of RA on health-related quality of life is substantial and can result in disability characterized by severe functional limitation and comorbidities [3].

Treatment options for RA include non-steroidal anti-inflammatory drugs, glucocorticoids, and several classes of disease-modifying antirheumatic drugs (DMARDs) (conventional synthetic DMARDs [csDMARDs], biologic DMARDs [bDMARDs], and targeted synthetic DMARDs [tsDMARDs]) [4–6]. The use of bDMARDs (including tumor necrosis factor [TNF] inhibitors [anti-TNFs]) or tsDMARDs is considered when disease activity remains moderate to high despite csDMARD therapy (i.e. when a treatment target is not met) [4–6]. Of the five currently available anti-TNFs, infliximab, etanercept, and adalimumab are the most commonly prescribed, collectively accounting for around 90% of anti-TNF prescriptions for RA [7–10]. Reference product adalimumab (Humira; AbbVie Inc., North Chicago, IL), reference product infliximab (Remicade; Janssen Biotech, Horsham, PA), and reference product etanercept (Enbrel; Immunex Corporation, Thousand Oaks, CA) are approved by the US Food and Drug Administration and the European Medicines Agency for indications including RA [11–16]. For each of these anti-TNFs, several biosimilars are approved for use in Europe and other regions [17].

At present, the evidence base to guide selection of bDMARDs is limited [18]. Despite the growing number of available anti-TNFs, there are few comparative data on the relative efficacy and safety of the different anti-TNF treatment options. To the best of our knowledge, the EXXELERATE study is the only head-to-head randomized controlled trial (RCT) conducted to date to compare anti-TNFs (specifically, certolizumab vs. adalimumab) in patients with RA. The results showed that certolizumab pegol was not superior to adalimumab, in patients receiving concomitant methotrexate (MTX) [19].

A subcutaneous (SC) formulation of CT-P13 (Remsima SC; Celltrion, Inc., Incheon, Korea), representing a new mode of administration for infliximab, received European Union marketing authorization in 2019; CT-P13 SC is currently approved in Europe for the treatment of the same indications (except juvenile) as reference infliximab, including RA [20,21]. In a phase 1/3 study of patients with active RA, CT-P13 SC administered every 2 weeks at a dose of 120 mg (after dose-loading with intravenous [IV] CT-P13; CT-P13 IV) was shown to have non-inferior efficacy to CT-P13 IV (3 mg/kg administered every 8 weeks) at Week 22, as assessed by change from baseline in Disease Activity Score in 28 joints (DAS28)-C-reactive protein (CRP) [22,23]. The two infliximab formulations demonstrated at least comparable efficacy, with post hoc analysis suggesting that the SC formulation was associated with significantly better efficacy outcomes compared with the IV formulation [24].

The primary objective of this systematic review and meta-analysis was to compare the efficacy and safety of CT-P13 SC with historical clinical data for infliximab IV, adalimumab, and etanercept, and their respective biosimilars, in patients with moderate-to-severe RA.

2. Methods

The protocol for this systematic review was prospectively registered (PROSPERO number: CRD42019149621) [25].

2.1. Systematic literature search

We conducted systematic electronic searches of Embase.com (comprising Embase, MEDLINE, and PubMed) and the Cochrane Library (comprising the Cochrane Central Register of Controlled Trials [CENTRAL], Cochrane Database of Systematic Reviews [CDSR], Database of Abstracts of Reviews of Effects [DARE] and the Health Technology Assessment database). Search strategies were constructed using relevant Medical Subject Headings (MeSH) and free-text terms, as presented in Supplementary Tables 1 and 2. The searches were restricted to English-language publications, human studies and dates from January 2009 to August 2019.

2.2. Criteria for considering studies for this review

2.2.1. Study design

We included parallel-group RCTs and open-label extensions of RCTs. Studies were required to be reported as full-text articles.

2.2.2. Participants

We included adults (aged ≥18 years) with moderate-to-severe RA, defined as having either ≥6 swollen and ≥6 tender joints, a DAS28-erythrocyte sedimentation rate (ESR) ≥3.2, or a DAS28-CRP ≥3.2, plus at least one of the following criteria: morning stiffness lasting ≥45 min; serum CRP >1.0 mg/dL; ESR >28 mm/h. Participants were required to be receiving background treatment with MTX.

2.2.3. Interventions

We included trials that studied infliximab IV, adalimumab, etanercept, and their biosimilars. Treatments were required to be administered at the dose and regimen specified in the relevant Summary of Product Characteristics.

2.2.4. Outcome measures

We included trials that reported one or more of the following outcome measures between Weeks 22 and 54 (inclusive). Efficacy outcomes: change from baseline in DAS28-CRP/ESR; low disease activity (LDA) and remission based on DAS28-CRP/ESR criteria; European League Against Rheumatism (EULAR) response (based on CRP/ESR criteria); 20%, 50% or 70% improvement in American College of Rheumatology (ACR) response (ACR20, ACR50, and ACR70, respectively); change from baseline in Clinical Disease Activity Index (CDAI) and Simplified Disease Activity Index (SDAI); LDA and remission based on CDAI and SDAI; and remission based on ACR/EULAR criteria and Boolean clinical criteria. Safety outcomes included rates of adverse events (AEs; any or serious), infections (any or serious), and treatment discontinuations due to AEs or lack of efficacy.

2.3. Selection of studies for inclusion in the review

Two reviewers (TSK, JSC) independently screened titles and abstracts of the records (retrieved through the electronic searches) for potential inclusion in the review. These reviewers then independently conducted an assessment of the full-text articles to determine whether to ‘include’ or ‘exclude,’ recording reasons for exclusion. A third reviewer (MinYoung Jang, Celltrion Healthcare) arbitrated in the event of disagreement between the two independent screenings. A randomly selected sample of excluded studies was considered by the third reviewer to verify appropriate application of the exclusion criteria. Multiple reports of the same study were identified and collated so that each study was the unit of interest in the review. The selection process was recorded in sufficient detail to complete a PRISMA flow diagram [26].

2.4. Data extraction and management

Study characteristics and outcome data were collected using a Microsoft Excel template. The following study characteristics were extracted: study design (randomization method, blinding, duration, number of centers, dates performed; geographic location); population (number of participants randomized, demographics, prior anti-TNF use, concomitant medication, baseline disease activity); intervention and comparators (type, dose, regimen); and outcome measures (prespecified as of interest to this review, as previously described).

2.5. Measures of treatment effect and data synthesis

Data from multiple infliximab studies and data from multiple adalimumab and etanercept studies were combined in two separate analyses. The relatively low number of adalimumab and etanercept RCTs conducted in RA during the previous 10 years (i.e. the inclusion timeframe) prevented meaningful meta-analysis of adalimumab and etanercept data separately. For binary outcomes, a generalized linear mixed model was used for pooling with logit transformation. For continuous outcomes, the inverse variance method was used for pooling without transformation. For completeness, both random-effect and fixed-effects models were conducted, and the results are presented together in the forest plots.

Heterogeneity was assessed using the Chi-squared-based Q-test [27]. In the presence of significant heterogeneity (i.e. the p-value of the Q-test < 0.05 or I² >50%), the random-effects model [28] was used to calculate pooled odds ratios and 95% confidence intervals (CI). Otherwise, the fixed-effects model was used [29].

Pooled historical data for (i) infliximab IV and (ii) adalimumab and etanercept were compared with recent data for CT-P13 SC [24]. All statistical analyses were performed using R software version 3.6.1 [30] and the ‘meta’ package for R, version 4.9–7 [31].

2.6. Quality assessment

The included studies were assessed for risk of bias (i.e. internal validity) according to National Institute for Health and Care Excellence guideline criteria [32]. Assessments were conducted by one author with independent verification by a second author.

3. Results

3.1. Results of the search

A PRISMA diagram summarizing the flow of information through this systematic review is presented in Figure 1. We identified 821 records through the electronic searches. After removal of duplicates, 809 records were screened (778 records excluded) and a total of 31 full-text articles were assessed for eligibility. Fifteen articles were excluded for the following reasons: study population did not meet eligibility criteria (n = 6); reported time points did not meet eligibility criteria (n = 5); duplicate of results (n = 3); and other (n = 1) [33–47].

Figure 1. Flow diagram of rheumatoid arthritis studies, illustrated per PRISMA guidelines [26]

^aStudy involved rescue therapy.

Thirteen RCTs (reported in 16 publications) were included in the qualitative and quantitative syntheses, as follows. Infliximab (n = 6 studies): NCT01936181 [48,49]; NCT02222493 [50,51]; NCT01217086 (PLANETRA) [52,53]; JapicCTI-111620 [54]; NCT00202852 and open-label extension (OLE; NCT00732875) [55]; NCT00691028 (RISING) [56]. Adalimumab (n = 4 studies): NCT02167139 [57]; NCT02137226 (VOLTAIRE-RA) [58]; NCT03052322 (AURIEL-RA) [59]; NCT01710358 (RA-BEAM) [60]. Etanercept (n = 3 studies): NCT01895309 [61]; NCT02357069 [62]; NCT01270997 (HERA) [63].

3.2. Description of the included studies

3.2.1. Infliximab IV or infliximab biosimilars

The design and eligibility criteria of the six included studies were generally consistent (Table 1) [48–56]. All six studies were RCTs, included a double-blind period, and had a duration of ≥54 weeks. In two of the six studies, open-label treatment was administered from either Weeks 0–14 or from Week 30 onwards. One of the six studies included a switching phase where participants receiving the reference product anti-TNF were re-randomized to remain on their current treatment or switch to the biosimilar treatment (NCT02222493 [50,51]). Two of the studies (JapicCTI-111620 [54]; NCT00691028 [56]) were conducted in Japan, and one study (NCT00202852 and OLE NCT00732875 [55]) was conducted in Korea; the other studies were multinational [48–53].

Table 1. Characteristics of the included studies (infliximab IV and infliximab biosimilars)

RCT	Patients	Randomized arms	N	Arm in meta-analysis	Double-blind phase	Open-label phase	Reference
NCT01936181	RA (ACR) Age 18–75 years Disease duration ≥6 months Active disease: ≥6 SJC; ≥6 TJC and ≥1 of: ESR ≥28 mm/h or CRP ≥1 mg/dL Exclusion: any prior biologic agent; any DMARD/systemic immunosuppressant agent other than MTX in 4 weeks prior to randomization	3 mg/kg SB2 IV (Weeks 0, 2, 6 and q8w) + MTX	291	rINF; SB2	54 weeks + 24-week switching study Baseline corticosteroids and NSAIDs were allowed	None	[48,49]
		3 mg/kg rINF IV (Weeks 0, 2, 6 and q8w) + MTX	293
NCT02222493	Age ≥18 years RA (ACR/EULAR 2010) Disease duration ≥4 months ACR functional class I–III Moderate-to-severe RA: ≥6 SJC; ≥6 TJC and hsCRP ≥10 mg/dL Prior DMARD required washout prior to study start	3 mg/kg GP1111 IV (Weeks 0, 2, 6 and q8w) + MTX (10–25 mg/wk)	324	GP1111; rINF rINF/rINF; rINF/GP1111; GP1111/GP1111	30 weeks + 24-week switching study Any second DMARD and folic/folinic acid continued	Second switching period from Weeks 54 to 78	[50,51]
		3 mg/kg rINF IV (Weeks 0, 2, 6 and q8w) + MTX (10–25 mg/wk)	326
NCT01217086 PLANETRA	Age 18–75 years Disease duration ≥1 year Active disease: ≥6 SJC; ≥6 TJC and ≥2 from: morning stiffness ≥45 min; ESR ≥28 mm/h or CRP ≥2 mg/dL despite MTX ≥3 months No prior biologic use permitted	3 mg/kg CT-P13 IV (Weeks 0, 2, 6 and q8w) + MTX (12.5–25 mg/wk) + folic acid (≥5 mg/wk)	302	CT-P13; rINF	54 weeks Baseline corticosteroids and NSAIDs could be continued	Weeks 54–102 (maintain CT-P13 IV vs. switch from rINF to CT-P13 IV)	[52,53]
		3 mg/kg rINF IV (Weeks 0, 2, 6 and q8w) + MTX (12.5–25 mg/wk) + folic acid (≥5 mg/wk)	304
JapicCTI-111620	Age 20–75 years RA (ACR 1987) Disease duration ≥1 year Active disease: ≥6 SJC; ≥6 TJC and ≥2 from: morning stiffness ≥45 min; ESR ≥28 mm/h or CRP ≥2 mg/dL Prior MTX treatment for ≥12 weeks (stable dose for 4 weeks prior to study drug administration)	3 mg/kg CT-P13 IV (Weeks 0, 2, 6 and q8w) + MTX (6–16 mg/wk) + folic acid (≤5 mg/wk)	51	CT-P13, rINF	54 weeks	None	[54]
		3 mg/kg rINF IV (Weeks 0, 2, 6 and q8w) + MTX (6–16 mg/wk) + folic acid (≤5 mg/wk)	53
NCT00202852 and OLE NCT00732875	RA (ACR 1987) Active disease: ≥6 SJC; ≥6 TJC and ≥2 from: morning stiffness ≥45 min; ESR ≥28 mm/h or CRP ≥2 mg/dL despite MTX ≥3 months Oral/parenteral MTX for ≥3 months (stable 12.5 mg/wk dose ≥4 weeks before screening); oral corticosteroid/NSAID doses stable ≥4 weeks before screening; other DMARDs stopped 4 weeks prior to study drug administration	3 mg/kg rINF IV (Weeks 0, 2, 6 and q8w) + MTX (≥12.5 mg/wk)	71	rINF	30 weeks Concomitant NSAIDs and corticosteroids permitted	84 weeks OLE	[55]
		Placebo + MTX (≥12.5 mg/wk)	72
NCT00691028 (RISING)	Age 18–75 years RA (ACR 1987) Active disease: ≥6 SJC; ≥6 TJC and ≥2 from: morning stiffness ≥45 min; ESR ≥28 mm/h or CRP ≥2 mg/dL MTX-IR	3 mg/kg rINF IV (q8w from W14) + MTX (8 mg/wk)	99	rINF 3 mg/kg IV open-label from Weeks 0 to 14	Weeks 14 to 54	Weeks 0–14	[56]
		6 mg/kg rINF IV (q8w from W14) + MTX (8 mg/wk)	104
		10 mg/kg rINF IV (q8w from W14) + MTX (8 mg/wk)	104

ACR, American College of Rheumatology; CRP, C-reactive protein; DAS28, Disease Activity Score in 28 joints; DMARD, disease-modifying antirheumatic drug; EULAR, European League Against Rheumatism; ESR, erythrocyte sedimentation rate; hsCRP, high sensitivity CRP; IV, intravenous; MTX, methotrexate; MTX-IR, inadequate response to prior methotrexate; NSAID, non-steroidal anti-inflammatory drug; q8w, every 8 weeks; RA, rheumatoid arthritis; RCT, randomized controlled trial; rINF, reference product infliximab; SJC, swollen joint count; TJC, tender joint count.

Across studies, inclusion criteria required participants to have active RA according to ACR (1987) [64] or EULAR/ACR (2010) criteria [65] and to have received a stable dose of MTX prior to baseline (permitted doses ranged from 6 to 25 mg/week). Prior bDMARD use was not permitted in two studies (NCT01936181 [48,49]; NCT01217086 [52,53]), and required a washout period in two studies (NCT02222493 [50,51]; NCT00202852 [55]).

All six studies included a treatment arm where reference product infliximab was administered IV at a dose of 3 mg/kg at Weeks 0, 2, and 6, and every 8 weeks thereafter. The comparator arms included the biosimilars SB2 (NCT01936181 [48,49]), GP1111 (NCT02222493 [50,51]), and CT-P13 (NCT01217086 [52,53], JapicCTI-111620 [54]). These agents were administered at the same dose and dosing interval as the reference product infliximab (i.e. IV at an initial dose of 3 mg/kg at Weeks 0, 2, and 6, and every 8 weeks thereafter).

A total of 2342 participants were initially randomized/enrolled to relevant treatment arms in the included studies (Supplementary Table 3). Where reported, the mean or median age of participants ranged from 49.3 to 54.5 years, 79–90% of participants were female, mean/median body weight ranged from 53.4 to 74.2 kg, and mean/median disease duration ranged from 6.3 to 8.2 years.

3.2.2. Reference product/biosimilar adalimumab or etanercept

The design and eligibility criteria of the seven included studies were generally consistent (Table 2) [57–63]. All seven studies were RCTs, included a double-blind period, and had a duration of ≥48 weeks. Additionally, two studies included an open-label extension and two studies included a switching period where participants were re-randomized to either remain on the current treatment or switch to the biosimilar (NCT02167139 [57] and NCT02137226 [58]). Of the etanercept studies, one (NCT01270997 [63]) was conducted in Korea and one (NCT02357069 [62]) was conducted in Japan and Korea; the other etanercept [61] and adalimumab [57–60] studies were multinational.

Table 2. Characteristics of the included studies for reference product or biosimilar adalimumab and etanercept

RCT	Patients	Randomized arms	N	Arm in meta-analysis	Double-blind phase	Open-label phase	Reference
Studies of reference product or biosimilar adalimumab
NCT02167139	Age 18–75 years RA (ACR 1987) Disease duration ≥6 months to 15 years Active disease: ≥6 SJC; ≥6 TJC and either ESR ≥28 mm/h or CRP ≥1 mg/dL No prior biologic use permitted	40 mg SB5 SC (eow) + MTX (10–25 mg/wk)	269	rADA/rADA; SB5/SB5; rADA/SB5	0–24 weeks and switching phase from 24–52 weeks (continue with rADA or switch to SB5) Baseline NSAIDs and other analgesics could be continued	None	[57]
		40 mg rADA SC (eow) + MTX (10–25 mg/wk)	273
NCT02137226 (VOLTAIRE-RA)	Age 18–80 years Disease duration ≥6 months Active disease: ≥6 SJC; ≥6 TJC and either ESR ≥28 mm/h or CRP ≥1 mg/dL No prior ADA use; ≤1 prior bDMARD	40 mg BI-695501 SC (eow) + MTX (10–25 mg/wk)	324	rADA/rADA; BI/BI rADA/BI	0–24 weeks and switching phase from 24–58 weeks (continue with rADA or switch to BI-695501)	Weeks 54–78	[58]
		40 mg rADA SC (eow) + MTX (10–25 mg/wk)	321
NCT03052322 (AURIEL-RA)	Age ≥18 years RA (ACR 1987 or EULAR/ACR 2010) Disease duration ≥6 months Active disease: ≥6 SJC; ≥6 TJC and either ESR ≥28 mm/h or CRP ≥1 mg/dL No prior ADA use; ≤1 prior anti-TNF	40 mg MSB11022 SC (eow) + MTX (10–25 mg/wk)	143	MSB11022; rADA	52 weeks	None	[59]
		40 mg rADA SC (eow) + MTX (10–25 mg/wk)	145
NCT01710358 (RA-BEAM)	Age ≥18 years Active disease: ≥6 SJC; ≥6 TJC; hsCRP ≥6 mg/dL No prior bDMARD use	4 mg baricitinib qd + MTX (15–25 mg/wk)	487	rADA	52 weeks (patients receiving placebo switched to baricitinib at Week 24) Concomitant stable doses of csDMARDs, NSAIDs, analgesics, or glucocorticoids permitted	None	[60]
		40 mg rADA SC (eow) + MTX (15–25 mg/wk)	330
		Placebo	488
Studies of reference product or biosimilar etanercept
NCT01895309	Age 18–75 years RA (ACR 1987) Disease duration ≥6 months to 15 years Active disease: ≥6 SJC; ≥6 TJC and either ESR ≥28 mm/h or CRP ≥1 mg/dL No prior bDMARD use	50 mg SB4 SC (qw) + MTX (10–25 mg/wk) + folic acid (5–10 mg/wk)	299	SB4; rETN	52 weeks Baseline NSAIDs and glucocorticoids could be continued	OLE extension Weeks 52 to 100	[61]
		50 mg rETN SC (qw) + MTX (10–25 mg/wk) + folic acid (5–10 mg/wk)	297
NCT02357069	Age 20–75 years RA (ACR 1987); FC I–III (ACR 1991) Disease duration ≥6 months Active disease: ≥6 SJC; ≥6 TJC; ESR ≥28 mm/h; or CRP ≥1 mg/dL and DAS28-ESR ≥3.2 No prior etanercept or ≥2 prior bDMARDs	50 mg LBEC0101 SC (qw) + MTX (7.5–15 mg/wk in Korea; 6–16 mg/wk in Japan)	187	LBEC0101; rETN	54 weeks Baseline NSAIDs, analgesics and oral/topical/nasal corticosteroids could be continued	None	[62]
		50 mg rETN SC (qw) + MTX (7.5–15 mg/wk in Korea; 6–16 mg/wk in Japan)	187
NCT01270997 (HERA)	Age ≥20 years RA (ACR 1987); FC I–III (ACR 1991) Active disease: ≥6 SJC; ≥6 TJC; ESR ≥28 mm/h; or CRP ≥1 mg/dL RF^+ or ACP^+ or bone erosions of hand or feet on X-ray	25 mg HD203 SC (twice weekly) + MTX (7.5–25 mg/wk)	147	HD203; rETN	48 weeks Baseline NSAIDs, corticosteroids could be continued	None	[63]
		25 mg rETN SC (twice weekly) + MTX (7.5–25 mg/wk)	147

ADA, adalimumab; ACP, anti-citrullinated protein; ACR, American College of Rheumatology; anti-TNF, tumor necrosis factor inhibitor; bDMARD, biologic DMARD; CRP, C-reactive protein; csDMARD, conventional synthetic DMARD; DAS28, Disease Activity Score in 28 joints; DMARD, disease-modifying antirheumatic drug; EULAR, European League Against Rheumatism; eow, every other week; ESR, erythrocyte sedimentation rate; FC, functional class; hsCRP, high sensitivity CRP; INF, infliximab; MTX, methotrexate; NSAID, non-steroidal anti-inflammatory drug; qd, once daily; qw, once weekly; RA, rheumatoid arthritis; rADA, reference product adalimumab; RF, rheumatoid factor; RCT, randomized controlled trial; rETN, reference product etanercept; SC, subcutaneous; SJC, swollen joint count; TJC, tender joint count.

Across studies, inclusion criteria required participants to have active RA according to ACR (1987) [64] or EULAR/ACR (2010) criteria [65] and to have received a stable dose of MTX prior to baseline; permitted doses ranged from 6 to 25 mg/week. Prior bDMARD use was not permitted in three of the studies (NCT02167139 [57]; NCT01710358 [60]; NCT01895309 [61]), one prior anti-TNF was permitted in one study (NCT03052322 [59]), and one prior bDMARD was permitted in one study (NCT02137226 [58]).

Four of the studies included a treatment arm where reference product adalimumab was administered SC at a dose of 40 mg every other week. Comparator arms included SB5 (NCT02167139 [57]), BI-695501 (NCT02137226 [58]) and MSB11022 (NCT03052322 [59]) and placebo/baricitinib (NCT01710358 [60]). Comparators were also administered SC at a dose of 40 mg every other week. Three of the studies included a treatment arm where reference product etanercept was administered SC at a dose of 50 mg once weekly or 25 mg twice weekly; the comparator arms included the etanercept biosimilars SB4 (NCT01895309 [61]), LBEC0101 (NCT02357069 [62]), and HD203 (NCT01270997 [63]), administered SC at doses of 50 mg weekly or 25 mg twice weekly (HD203).

A total of 3069 participants were initially randomized/enrolled to relevant treatment arms in the included studies (Supplementary Table 4). Where reported, the mean age of participants ranged from 49.8 to 55.5 years, 76–89% of participants were female, mean body weight ranged from 56.3 to 75.1 kg and mean disease duration ranged from 5.4 to 10 years.

3.2.3. CT-P13 SC

Participants were required to have active RA according to ACR/EULAR 2010 criteria [65]. A total of 167 individuals were randomized to receive CT-P13 SC at a dose of 120 mg every 2 weeks; participants were aged 19 to 74 years old, 77.8% were female and mean (standard deviation) disease duration was 6.82 (7.15) years [23].

3.3. Risk of bias in the included studies

All studies included in this systematic literature review were rated as having a low risk of selection, performance, attrition, and detection bias (Table 3). Internal and external validity were confirmed for all included studies (Table 3).

Table 3. Quality assessment results

Anti-TNF	Study		Reference		Risk of bias				Validity
					Selection Bias	Performance Bias	Attrition Bias	Detection Bias	Internal Validity		External Validity
Infliximab	NCT01936181		[49]		Low	Low	Low	Low	++		++
	NCT02222493		[50]		Low	Low	Low	Low	++		++
	NCT01217086		[52]		Low	Low	Low	Low	++		++
	JapicCTI-111620		[54]		Low	Low	Low	Low	++		+
	NCT00202852		[55]		Low	Low	Low	Low	++		+
	NCT00691028		[56]		Low	Low	Low	Low	++		+
Adalimumab/etanercept	NCT02167139		[57]		Low	Low	Low	Low	++		++
	NCT02137226		[58]		Low	Low	Low	Low	++		++
	NCT03052322		[59]		Low	Low	Low	Low	++		++
	NCT01710358		[60]		Low	Low	Low	Low	++		++
	NCT01895309		[61]		Low	Low	Low	Low	++		++
	NCT02357069		[62]		Low	Low	Low	Low	++		+
	NCT01270997		[63]		Low	Low	Low	Low	++		+

+, checklist partially fulfilled, conclusions are unlikely to change; ++, checklist fulfilled, conclusions are unlikely to change.

anti-TNF, tumor necrosis factor inhibitor.

3.4. Effects of the interventions

3.4.1. Infliximab versus CT-P13 SC

A summary of findings for the meta-analyses and comparison with CT-P13 SC is presented in Table 4. Selected forest plots for the meta-analyses of the historical data are presented in Supplementary Figures 1 to 9. A numerical advantage (with non-overlapping 95% CIs) was observed for CT-P13 SC over infliximab IV at the 30-week and 54-week time points for most outcomes: mean change from baseline in DAS28-CRP score; DAS28-CRP remission rates; mean change from baseline in DAS28-ESR score; DAS28-ESR remission rate (54 weeks only); the proportion of patients achieving a EULAR good response (CRP criteria); the proportion of patients achieving ACR20, ACR50 and ACR70 responses; mean change from baseline in CDAI score; CDAI remission rate (54 weeks only); mean change from baseline in SDAI score; SDAI remission rate; Boolean clinical remission rate; ACR/EULAR remission rate, the proportion of patients experiencing any AE; the proportion of patients experiencing any serious adverse event (SAE); the proportion of patients who discontinued due to AEs; and the proportion of patients who discontinued due to lack of efficacy (no patients discontinued due to lack of efficacy for CT-P13 SC at either time point); data for these outcomes are presented in Table 4. A visual presentation of the results for selected outcomes is shown in Figure 2.

Figure 2. A comparison of infliximab IV historical data with CT-P13 SC results in terms of remission rates (a), ACR response rates (b), and key safety outcomes (c)

Error bars present 95% confidence intervals.

ACR, American College of Rheumatology; ACR##, ##% improvement in ACR criteria; AE, adverse event; CDAI, Clinical Disease Activity Index; CRP, C-reactive protein; DAS28, Disease Activity Score in 28 joints; EULAR, European League Against Rheumatism; IV, intravenous; N/A, not available; SAE, serious adverse event; SC, subcutaneous; SDAI, Simplified Disease Activity Index.

Table 4. Summary of 30-week and 54-week findings for infliximab IV, infliximab biosimilars, and CT-P13 SC

Outcome	Time point	Participants, N	Studies (arms), N	Infliximab and infliximab biosimilars, Estimate (95% CI)	CT-P13 SC, Estimate (95% CI) [24]
Mean change from BL in DAS28-CRP score	30 weeks	751	1 (2)	−2.01 (−2.28, −1.74)^a	−2.94 (−3.14, −2.73)
	54 weeks	766	2 (3)	−1.94 (−2.49, −1.39)^b	−3.14 (−3.34, −2.93)
Mean change from BL in DAS28-ESR score	30 weeks	684	2 (4)	−2.26 (−2.37, −2.15)^a	−3.02 (−3.23, −2.80)
	54 weeks	101	1 (2)	−1.76 (−2.05, −1.46)^a	−3.28 (−3.49, −3.06)
DAS28-CRP remission^c	30 weeks	1084	2 (4)	0.21 (0.19, 0.24)^a	0.37 (0.30, 0.45)
	54 weeks	665	2 (4)	0.25 (0.22, 0.28)^a	0.46 (0.38, 0.54)
DAS28-ESR LDA^c	30 weeks	1012	2 (4)	0.11 (0.10, 0.13)^a	0.16 (0.11, 0.22)
	54 weeks	584	1 (2)	0.12 (0.10, 0.15)^a	0.16 (0.11, 0.23)
DAS28-ESR remission^c	30 weeks	1012	2 (4)	0.14 (0.12, 0.16)^a	0.18 (0.12, 0.24)
	54 weeks	583	1 (2)	0.18 (0.15, 0.21)^a	0.30 (0.23, 0.38)
Boolean remission, clinical^c	30 weeks	606	1 (2)	0.04 (0.03, 0.06)^a	0.18 (0.13, 0.25)
	54 weeks	606	1 (2)	0.05 (0.04, 0.08)^a	0.25 (0.18, 0.32)
ACR/EULAR remission^c	30 weeks	592	1 (2)	0.09 (0.07, 0.12)^a	0.22 (0.16, 0.29)
	54 weeks	566	1 (3)	0.13 (0.10, 0.16)^a	0.31 (0.24, 0.39)
EULAR moderate response (CRP)^c	30 weeks	1145	2 (4)	0.46 (0.43, 0.49)^a	0.43 (0.35, 0.51)
	54 weeks	665	2 (4)	0.40 (0.37, 0.44)^a	0.35 (0.27, 0.42)
EULAR good response (CRP)^c	30 weeks	1145	2 (4)	0.34 (0.30, 0.39)^b	0.53 (0.45, 0.61)
	54 weeks	665	2 (4)	0.39 (0.35, 0.43)^a	0.61 (0.53, 0.68)
EULAR moderate response (ESR)^c	30 weeks	1012	2 (4)	0.57 (0.54, 0.60)^a	0.61 (0.53, 0.68)
	54 weeks	–	–	–	0.46 (0.38, 0.54)
EULAR good response (ESR)^c	30 weeks	1012	2 (4)	0.25 (0.23, 0.28)^a	0.33 (0.26, 0.41)
	54 weeks	–	–	–	0.47 (0.39, 0.55)
ACR20^c	30 weeks	1691	5 (9)	0.68 (0.65, 0.72)^b	0.87 (0.81, 0.92)
	54 weeks	1790	4 (10)	0.69 (0.66, 0.72)^b	0.88 (0.83, 0.93)
ACR50^c	30 weeks	1149	4 (7)	0.40 (0.37, 0.43)^a	0.65 (0.58, 0.73)
	54 weeks	1790	5 (10)	0.44 (0.41, 0.48)^b	0.72 (0.64, 0.78)
ACR70^c	30 weeks	1078	3 (6)	0.20 (0.17, 0.22)^a	0.41 (0.34, 0.49)
	54 weeks	1790	4 (10)	0.25 (0.22, 0.29)^b	0.48 (0.41, 0.56)
Mean change from BL in CDAI score	30 weeks	596	2 (4)	−21.18 (−24.74, −17.62)^b	−31.69 (−33.66, −29.72)
	54 weeks	101	1 (2)	−15.05 (−17.55, −12.54)^a	−33.29 (−35.25, −31.33)
Mean change from BL in SDAI score	30 weeks	596	2 (4)	−22.18 (−25.42, −18.94)^b	−32.81 (−34.91, −30.72)
	54 weeks	101	1 (2)	−15.73 (−18.40, −13.07)^a	−34.43 (−36.50, −32.37)
CDAI LDA^c	30 weeks	–	–	–	0.41 (0.33, 0.49)
	54 weeks	584	1 (2)	0.34 (0.30, 0.38)^a	0.36 (0.29, 0.44)
CDAI remission^c	30 weeks	–	–	–	0.16 (0.11, 0.23)
	54 weeks	584	1 (2)	0.12 (0.10, 0.15)^a	0.29 (0.22, 0.37)
SDAI LDA^c	30 weeks	–	–	–	0.38 (0.31, 0.46)
	54 weeks	584	1 (2)	0.31 (0.27, 0.35)^a	0.36 (0.28, 0.44)
SDAI remission^c	30 weeks	606	1 (2)	0.08 (0.06, 0.11)^a	0.18 (0.13, 0.25)
	54 weeks	1190	2 (4)	0.10 (0.07, 0.13)^b	0.30 (0.23, 0.37)
Any AE^c	30 weeks	1903	4 (7)	0.58 (0.56, 0.60)^a	0.32 (0.25, 0.40)
	54 weeks	1392	4 (7)	0.81 (0.66, 0.90)^b	0.55 (0.47, 0.62)
Any infection^c	30 weeks	718	2 (3)	0.26 (0.22, 0.32)^b	0.15 (0.10, 0.22)
	54 weeks	–	–	–	0.29 (0.22, 0.37)
Any SAE^c	30 weeks	1903	4 (7)	0.08 (0.07, 0.09)^a	0.02 (0.00, 0.05)
	54 weeks	1392	4 (7)	0.11 (0.10, 0.13)^a	0.04 (0.01, 0.08)
Any serious infection^c	30 weeks	583	1 (2)	0.03 (0.02, 0.04)^a	0.01 (0.00, 0.03)
	54 weeks	786	3 (5)	0.03 (0.02, 0.05)^a	0.02 (0.00, 0.05)
Any discontinuation due to AE^c	30 weeks	1838	3 (6)	0.07 (0.06, 0.08)^a	0.01 (0.00, 0.03)
	54 weeks	1393	4 (7)	0.11 (0.09, 0.14)^b	0.04 (0.01, 0.08)
Any discontinuation due to lack of efficacy^c	30 weeks	1838	3 (6)	0.01 (0.01, 0.02)^a	0.00
	54 weeks	1393	4 (7)	0.03 (0.02, 0.03)^a	0.00

^aFixed effects model (Cochran’s Q-test p > 0.05). ^bRandom effects model (Cochran’s Q-test p ≤ 0.05). ^cReported as proportion of participants achieving the outcome.

ACR, American College of Rheumatology; ACR##, ##% improvement in ACR criteria; AE, adverse event; BL, baseline; CDAI, Clinical Disease Activity Index; CI, confidence interval; CRP, C-reactive protein; DAS28, Disease Activity Score in 28 joints; ESR, erythrocyte sedimentation rate; EULAR, European League Against Rheumatism; IV, intravenous; LDA, low disease activity; SAE, serious adverse event; SC, subcutaneous; SDAI, Simplified Disease Activity Index.

A numerical advantage for CT-P13 SC over infliximab IV (with overlapping 95% CI) was observed for the proportion of patients achieving/reporting DAS28-ESR remission (30 weeks), DAS28 LDA (ESR criteria; 30 and 54 weeks), EULAR good response (ESR criteria; 30 weeks), any infection (30 weeks) or any serious infection (30 weeks and 54 weeks); data for these outcomes are presented in Table 4.

3.4.2. Adalimumab and etanercept versus CT-P13 SC

A summary of the findings for the meta-analyses and comparison with CT-P13 SC is presented in Table 5. Selected forest plots for the meta-analyses are presented in Supplementary Figure 1 to 9. A numerical advantage (with non-overlapping 95% CIs) was observed for CT-P13 SC at 54 weeks over adalimumab/etanercept at 1 year for most outcomes: mean change from baseline in DAS28-CRP score; mean change from baseline in DAS28-ESR score; the proportion of patients achieving ACR50 and ACR70 responses; mean change from baseline in CDAI score; mean change from baseline in SDAI score; Boolean clinical remission rate; the proportion of patients experiencing any AE; the proportion of patients experiencing any infection and the proportion of patients who discontinued due to lack of efficacy (no patients discontinued due to lack of efficacy with CT-P13 SC); data for these outcomes are presented in Table 5. A visual presentation of the results for selected outcomes is presented in Figure 3.

Figure 3. A comparison of adalimumab and etanercept historical data at 1 year with CT-P13 SC results at 54 weeks in terms of remission rates (a), ACR response rates (b), and key safety outcomes (c)

Error bars represent 95% confidence intervals.

ACR, American College of Rheumatology; ACR##, ##% improvement in ACR criteria; AE, adverse event; CDAI, Clinical Disease Activity Index; CRP, C-reactive protein; DAS28, Disease Activity Score in 28 joints; SAE, serious adverse event; SC, subcutaneous; SDAI, Simplified Disease Activity Index.

Table 5. Summary of findings for reference product or biosimilar adalimumab and etanercept (at 1 year) and CT-P13 SC (at 54 weeks)

Outcome	Participants, N	Studies (arms), N	Adalimumab, etanercept and their biosimilars, Estimate (95% CI)	CT-P13 SC, Estimate (95% CI) [24]
Mean change from BL in DAS28-CRP score	595	2 (3)	−2.51 (−2.72, −2.29)^b	−3.14 (−3.34, −2.93)
Mean change from BL in DAS28-ESR score	926	2 (3)	−2.71 (−2.99, −2.43)^b	−3.28 (−3.49, −3.06)
DAS28-CRP remission^c	595	2 (3)	0.32 (0.25, 0.39)^b	0.46 (0.38, 0.54)
DAS28-ESR LDA^c	1621	4 (8)	0.17 (0.15, 0.19)^a	0.16 (0.11, 0.23)
DAS28-ESR remission^c	1621	4 (8)	0.29 (0.23, 0.34)^b	0.30 (0.23, 0.38)
Boolean remission, clinical^c	754	2 (5)	0.14 (0.12, 0.16)^a	0.25 (0.18, 0.32)
EULAR moderate response (CRP)^c	221	1 (2)	0.47 (0.40, 0.53)^a	0.35 (0.27, 0.42)
EULAR good response (CRP)^c	221	1 (2)	0.47 (0.40, 0.53)^a	0.61 (0.53, 0.68)
EULAR moderate response (ESR)^c	1885	4 (10)	0.50 (0.48, 0.52)^a	0.46 (0.38, 0.54)
EULAR good response (ESR)^c	1885	4 (10)	0.43 (0.37, 0.50)^b	0.47 (0.39, 0.55)
ACR20^c	2014	5 (10)	0.78 (0.72, 0.83)^b	0.88 (0.83, 0.93)
ACR50^c	2014	5 (10)	0.54 (0.48, 0.60)^b	0.72 (0.64, 0.78)
ACR70^c	2014	5 (10)	0.34 (0.28, 0.40)^b	0.48 (0.41, 0.56)
Mean change from BL in CDAI score	836	2 (3)	−25.77 (−28.95, −22.60)^b	−33.29 (−35.25, −31.33)
CDAI LDA^c	1326	3 (6)	0.34 (0.32, 0.37)^a	0.36 (0.29, 0.44)
CDAI remission^c	1326	3 (6)	0.20 (0.18, 0.22)^a	0.29 (0.22, 0.37)
Mean change from BL in SDAI score	835	2 (3)	−26.70 (−29.70, −23.71)^b	−34.43 (−36.50, −32.37)
SDAI LDA^c	1324	3 (6)	0.36 (0.33, 0.38)^a	0.36 (0.28, 0.44)
SDAI remission^c	1324	3 (6)	0.21 (0.18, 0.23)^a	0.30 (0.23, 0.37)
Any AE^c	2526	6 (12)	0.72 (0.63, 0.80)^b	0.55 (0.47, 0.62)
Any infection^c	997	3 (5)	0.46 (0.40, 0.52)^b	0.29 (0.22, 0.37)
Any SAE^c	2526	6 (12)	0.08 (0.06, 0.10)^b	0.04 (0.01, 0.08)
Any serious infection^c	926	2 (3)	0.01 (0.01, 0.02)^a	0.02 (0.00, 0.05)
Any discontinuation due to AE^c	1808	5 (8)	0.05 (0.04, 0.06)^a	0.04 (0.01, 0.08)
Any discontinuation due to lack of efficacy^c	1808	5 (8)	0.01 (0.00, 0.01)^a	0.00

^aFixed effects model (Cochran’s Q-test p > 0.05). ^bRandom effects model (Cochran’s Q-test p ≤ 0.05). ^cReported as proportion of participants achieving the outcome.

ACR, American College of Rheumatology; ACR##, ##% improvement in ACR criteria; AE, adverse event; BL, baseline; CDAI, Clinical Disease Activity Index; CI, confidence interval; CRP, C-reactive protein; DAS28, Disease Activity Score in 28 joints; ESR, erythrocyte sedimentation rate; EULAR, European League Against Rheumatism; LDA, low disease activity; SAE, serious adverse event; SDAI, Simplified Disease Activity Index.

A numerical advantage for CT-P13 SC over adalimumab/etanercept (with overlapping 95% CI) was observed for the proportion of patients achieving/reporting DAS28 remission (CRP/ESR criteria), EULAR good response (CRP/ESR criteria), ACR20, CDAI remission; CDAI LDA, SDAI remission, any SAE, and discontinuation due to AEs. Data for these outcomes are presented in Table 5.

4. Discussion

This is the first systematic review comparing the safety and efficacy of infliximab IV, adalimumab or etanercept versus CT-P13 SC. Our study identified 13 relevant RCTs that evaluated the efficacy and safety of infliximab IV, adalimumab, or etanercept and their respective biosimilars. Pooled data from these studies were compared with data from a pivotal study of CT-P13 SC in patients with RA [24].

Previous analyses have demonstrated comparable efficacy between infliximab, etanercept, adalimumab, certolizumab, and golimumab [66,67]; however, golimumab and certolizumab were not considered in the present review as they account for less than 10% of anti-TNFs prescribed for RA [9,10]. Aaltonen and colleagues conducted a systematic literature review and analysis of 26 RCTs and concluded that no single agent clearly rose above the others in terms of efficacy [67]. In agreement, a more recent network meta-analysis showed that infliximab, etanercept, adalimumab, certolizumab, and golimumab demonstrated a highly significant inhibitory effect on joint destruction versus placebo; results for infliximab, etanercept, adalimumab, and certolizumab did not differ significantly from each other [66]. However, the study concluded that high infliximab doses (infliximab 10 mg/kg every 8 weeks) resulted in better efficacy than the standard dose [66]. Since SC formulations may offer similar effects of dose escalation, further investigation into the potential benefits of SC infliximab formulations is warranted.

In our analysis, a numerical advantage for CT-P13 SC over infliximab IV was observed for almost every efficacy outcome evaluated, including DAS28, CDAI, and SDAI scores, ACR response, and measures of disease remission (including DAS28-CRP, DAS28-ESR, CDAI, SDAI, Boolean, and ACR/EULAR definitions) and LDA. This advantage was observed at both the 30-week and 54-week time points, and, in most instances, there was no overlap in 95% CIs between the groups. The statistical significance of the advantage was not assessed using p-values. The present findings can potentially be explained by the pharmacokinetic profile of CT-P13 SC, which maintains trough serum concentrations that are higher than the therapeutic threshold [20,52]. Considering the improved efficacy reported with high infliximab doses, compared with standard doses, in a previous meta-analysis [66], it follows that any numerical improvements in the efficacy profile of infliximab SC compared with infliximab IV might be plausibly explained by increased trough infliximab concentrations. Indeed, several studies have demonstrated a correlation between clinical efficacy and trough concentration levels of anti-TNFs [56,68]. Similarly, a numerical advantage was observed for CT-P13 SC for every safety outcome evaluated, including the proportions of participants experiencing AEs or infections, or discontinuing study treatment. Although generally less marked, similar numerical trends were observed for efficacy and safety outcomes based on comparison of the historical data for adalimumab and etanercept with the data for CT-P13 SC.

Across the studies included in our analysis, demographics of the participants were representative of a population with moderate-to-severe RA. Participants were aged 50–55 years, had active disease according to ACR or EULAR criteria despite prior treatment with stable doses of MTX, and had an average disease duration of 5–10 years. Consistent with the epidemiology of RA, the included studies enrolled a greater proportion of women. The baseline clinical and demographic characteristics of individuals receiving CT-P13 SC (mean age 50.9 years; 78% female; mean disease duration 6.8 years) were comparable to those observed in the included historical studies. Importantly, eligibility criteria for inclusion of studies in the review were reflective of the population of interest and usual care pathways for that population [4,5]. Strengths of the review process include the comprehensive nature of the electronic searches conducted and independent screening of the search results by two reviewers to determine relevant studies for inclusion. Therefore, our findings are applicable to current populations of patients with active, moderate-to-severe RA and an inadequate response to MTX.

In total, the included studies randomized over 2300 participants to receive either infliximab IV or infliximab biosimilars, and over 3000 participants to receive reference product or biosimilar adalimumab or etanercept, resulting in a large overall sample size. Half of the contributing studies enrolled over 500 participants, and most of the meta-analyses conducted included more than 500 participants in total. All prespecified outcomes of interest were well reported across the included studies, and 52/54-week data were available for most outcomes evaluated. The findings of our analysis can be considered contemporary, as the included studies were generally conducted in the previous decade and most patients were recruited to the studies from 2015 onwards. Furthermore, the included studies contributing to the historical dataset were assessed for risk of bias and determined to have a low risk of bias for all domains assessed. Based on these findings, it is acceptable to conclude that pooling studies with internal validity yields generalizable results.

The present review has several limitations. Firstly, for some of the meta-analyses the p-values for Cochran’s Q-test ‘may be representative’ of substantial heterogeneity. We used both fixed and random-effects meta-analysis models, the latter of which assumes that the effects being estimated in the different studies are not identical, but follow a normal distribution [69]. Thus, random-effects models assist in controlling for heterogeneity that cannot be readily explained by other factors. It remains possible, however, that the observed statistical heterogeneity may be a consequence of clinical or methodological diversity among the included studies. For example, permitted prior or concomitant treatments varied across the included studies (possibly depending on the year that the study was performed and available treatments at that time). Secondly, two studies employing placebo as the sole control group (NCT00202852; reference product infliximab vs. placebo [55]), or as one of the control groups (NCT01710358 [RA-BEAM] [60]; reference adalimumab vs. baricitinib or placebo), may have had relatively unfavorable results in terms of the observed effectiveness or safety of the active comparator, compared with the trials evaluating a biosimilar and its reference drug. This is because in clinical trials involving placebo, the psychological impact of the possibility of receiving placebo can lead to a negative effect on observed efficacy and safety of the drug under evaluation [70]. Besides these two studies, all included studies employed active controls, whereby a reference product was compared with either a biosimilar or an alternative dose of the reference product. In contrast to placebo-controlled trials, clinical trials using a reference drug that has already been demonstrated to be effective and well tolerated as an active control may result in relatively more favorable findings for the drug under evaluation, in terms of the observed effectiveness or safety. For this reason, biosimilar studies have shown better efficacy compared to those from historical studies with placebo. Thirdly, while most of the included studies were multinational, several infliximab and etanercept studies were locally focused in Korea and/or Japan. There were some numerical differences in efficacy findings between these types of study. However, it is important to note that the number of patients included in the locally conducted studies represents a relatively small proportion of the total number of patients included in this analysis. Taken together, we believe that the inclusion of the locally focused studies will not have significantly impacted the overall conclusions of this analysis. A fourth limitation of the present review is that data were pooled for adalimumab and etanercept. Interestingly, despite the pooling of data for two different drugs, heterogeneity within the associated meta-analyses did not appear to be consistently greater than that observed for the infliximab analyses; this finding is consistent with the results of a meta-analysis suggesting similar clinical efficacy between etanercept and adalimumab [67]. Fifth, the time points reported varied across the adalimumab and etanercept studies (48–58 weeks) and were compared with the longest-term, 54-week data currently available for CT-P13 SC. However, we are satisfied that the comparison is robust and reflective of long-term treatment with anti-TNFs. In addition, outcomes relating to structural damage measurements and immunogenicity were not included in this review. Finally, as the historical data were compared with findings for CT-P13 SC originating from a single study, the comparative findings might not be applicable to any other infliximab SC formulations that become available in the future.

In the context of the limitations discussed above, our findings suggest that CT-P13 SC administration offers a similar or improved benefit-to-harm ratio compared with infliximab IV or adalimumab/etanercept, for the treatment of moderate-to-severe RA. In terms of clinical practice, adalimumab and etanercept are commonly used in RA because SC is preferred by patients and physicians. However, the clinical benefit of IV infliximab is well demonstrated and is thus still commonly used. The present meta-analysis suggests that infliximab SC displays an improved (or, at least similar) clinical profile compared with infliximab IV, etanercept, and adalimumab. Therefore, CT-P13 SC may bring improved patient outcomes in a given clinical context.

5. Conclusions

The present systematic literature review and meta-analyses have shown that CT-P13 SC offers an improved, or at least similar, benefit-to-harm ratio for all outcome measures compared with infliximab IV and adalimumab/etanercept, for the treatment of moderate-to-severe RA.

Declaration of interest

R Caporali has received speaker’s fees and consultation fees from AbbVie, Bristol-Myers Squibb, Celltrion, Fresenius-Kabi, Gilead-Galapagos, Lilly, Merck Sharp & Dohme, Pfizer, Roche, Samsung-Bioepis, Sanofi, and UCB. Y Allanore has no competing interests to disclose. R Alten has received honoraria from AbbVie, Bristol-Myers Squibb, Celltrion, Gilead, Janssen, Lilly, Merck, Novartis, Pfizer, Roche-Chugai and UCB; and research grants from Novartis, Pfizer, and Roche. B Combe has received honoraria from AbbVie, Bristol-Myers Squibb, Gilead, Janssen, Lilly, Merck, Novartis, Pfizer, Roche-Chugai, Sanofi and UCB; and research grants from Novartis, Pfizer, and Roche. P Durez has received speaker’s fees from AbbVie, Bristol-Myers Squibb, Galapagos, Lilly, and Sanofi. F Iannone has received speaker’s fees and consultation grants from AbbVie, Actelion, Bristol-Myers Squibb, Biogen, Lilly, Merck Sharp & Dohme, Pfizer, Roche, Sanofi, and UCB. MT Nurmohamed has received consulting fees from AbbVie, Celgene, Celltrion, Eli Lilly, Janssen, and Sanofi; speaker’s fees from AbbVie, Bristol-Myers Squibb, Eli Lilly, Roche, and Sanofi; and research funding from AbbVie, Bristol-Myers Squibb, Celgene, Eli Lilly, Janssen, Merck Sharp & Dohme, Mundipharma, Novartis, Pfizer, Roche, and Sanofi. SJ Lee and G Park are employed full time by Celltrion Inc. TS Kwon and JS Choi are employed full time by Celltrion Healthcare. DH Yoo has received speaker’s fees and consultation fees from Celltrion Healthcare and Celltrion Inc.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

RC, YA, RA, BC, PD, FI, MTN, SJL, GP and DHY contributed to data analysis or interpretation. TSK and JSC contributed to study design, data collection and data analysis or interpretation. All authors reviewed and critically revised drafts of the manuscript, approved the final draft for submission, and are accountable for the accuracy and integrity of the article.

Acknowledgments

We would like to thank MinYoung Jang (Celltrion Healthcare), who acted as an independent reviewer for the selection of studies included in the review. Medical writing support (including development of a draft outline and subsequent drafts in consultation with the authors, assembling tables and figures, collating author comments, copyediting, fact checking and referencing) was provided by Beatrice Tyrrell, DPhil, at Aspire Scientific Limited (Bollington, UK), and funded by Celltrion Healthcare (Incheon, Republic of Korea).

Supplemental material

Supplemental data for this article can be accessed on the publisher’s website.

Data availability statement

The data that support the findings of this analysis are available from TSK upon reasonable request.

Additional information

Funding

The paper was funded by Celltrion Healthcare (Incheon, Republic of Korea).