Abstract
Objective
To determine if the efficacy of biologics differ based on magnetic resonance imaging (MRI) and C‐reactive protein (CRP) findings.
Methods
We compared four subgroups (MRI+/CRP+, MRI+/CRP−, MRI−/CRP+, MRI−/CRP−) from randomized controlled trials (RCTs). A comprehensive database search was performed to include axial spondylarthritis (axSpA; both radiographic axSpA [r‐axSpA] and nonradiographic axSpA [nr‐axSpA]) RCTs with treatment efficacy reported by different MRI and CRP subgroups. Study‐specific disease activity scores (at 12‐16 weeks) were pooled using a random‐effects model and compared between the four subgroups.
Results
Five trials (all nr‐axSpA) were included: three with tumor necrosis factor inhibitors (TNFi, N = 729) and two with interleukin‐17 inhibitors (IL‐17i, N = 794). TNFi and IL‐17i showed efficacy based on the Assessment of Spondyloarthritis International Society 40 (ASAS40) and Bath Ankylosing Spondylitis Disease Activity Index 50 (BASDAI50) in all MRI and CRP subgroups, except the CRP−/MRI− subgroup, which had a single study with only 39 patients. There was no statistically significant difference between the four subgroups in terms of patients achieving ASAS40 (P = 0.60, I2 = 0%) or BASDAI50 (P = 0.27, I2 = 23.9%). The number needed to treat was three for the CRP+/MRI+ and CRP+/MRI− subgroups and six for the CRP−/MRI+ and CRP−/MRI− subgroups. All trials had a low risk of bias. Between‐study heterogeneity was low to moderate. Sensitivity analyses comparing TNFi or IL‐17i versus placebo similarly showed no difference between subgroups in terms of ASAS40 (TNFi, P = 0.57; IL‐17i, P = 0.28) and BASDAI50 (TNFi, P = 0.37; IL‐17i, P = 0.18).
Conclusion
In this systematic review, there was no statistically significant difference between the four subgroups in terms of efficacy based on ASAS40 or BASDAI50.
INTRODUCTION
Although the use of biologics over the last decade has revolutionized treatment for patients with axial spondyloarthritis (axSpA), only up to 60% and 40% of patients reach partial remission and inactive disease, respectively (1). This is partly because axSpA is a heterogeneous disease with a wide range of clinical presentations, disease severity, and inflammatory burden, often posing diagnostic and management challenges. In the absence of objective markers that reliably predict which subgroup of patients with axSpA will respond to a particular therapy, we currently rely on a “trial and error” approach for treatment. Among the available markers, C‐reactive protein (CRP) testing and magnetic resonance imaging (MRI) are increasingly being performed at baseline in patients with axSpA. Leveraging this readily available information to aid our treatment decision would be optimal for clinical practice.
It is plausible that a combination of CRP levels and MRI findings at baseline may provide better insight as to the inflammatory burden of axSpA than either factor alone. Elevated CRP levels are associated with a favorable treatment response in axSpA; however, the association of inflammation on MRI with treatment response is unclear (2, 3). The CRP level is elevated in only 40% to 50% of patients with axSpA; however, when present, it correlates well with disease activity (4, 5). CRP levels do not always correlate with MRI findings, and MRI inflammation could be noted in patients with normal CRP levels (6, 7, 8). Furthermore, the correlation between inflammation on MRI and disease activity is low, so MRI may provide additional information about disease activity (9). A more nuanced understanding of the relationships between CRP levels, MRI findings, and treatment response could help identify specific subpopulations of patients who might benefit from targeted therapeutic approaches. Understanding differences in treatment response across subpopulations with differing inflammatory burdens could pave the way for personalized medicine strategies in axSpA.
Although some randomized controlled trials (RCTs) have examined treatment efficacy based on CRP and MRI subgroups, the trials themselves are not powered to examine subgroups. A systematic review and meta‐analysis of subgroup data from these trials can yield better insight. Therefore, we performed a systematic review and meta‐analysis of subgroup data (four subgroups: MRI+/CRP+, MRI+/CRP−, MRI−/CRP+, MRI−/CRP−) from available RCTs to determine if treatment efficacy differed based on MRI and CRP findings.
PATIENTS AND METHODS
Search strategy and study selection
The systematic review was conducted according to a protocol submitted to the Spondyloarthritis Research and Therapy Network. A comprehensive search of several databases and abstracts from inception to January 11, 2022, was performed using a search strategy designed and conducted by an experienced librarian with input from the study's principal investigator (Supplementary Material 1).
We included RCTs (phase II and III) fulfilling the following criteria: 1) adults (≥18 years) with axSpA (radiographic axSpA [r‐axSpA] and nonradiographic axSpA [nr‐axSpA]), 2) treatment response assessed using quantitative scoring, 3) treatment efficacy reported for a particular therapy compared with placebo, 4) treatment efficacy reported by different MRI and CRP subgroups, and 5) MRI sacroiliac joint (SIJ) and CRP testing performed at baseline for more than 50% of participants. Data from long‐term extension studies and experimental medications not approved by the US Food and Drug Administration for any other disease were excluded. The Preferred Reporting Items for Systematic Reviews and Meta‐Analyses statement for reporting systematic reviews was followed (Figure 1) (10). Two authors (PK and RS) screened abstracts for eligibility, retrieved full texts, and excluded irrelevant articles. Disagreements were resolved by a discussion on eligibility. Bibliographies belonging to the included studies, reviews, and relevant articles were screened for additional studies. No language restrictions were made.
Figure 1.
Flowchart describing systematic search and study selection process. CRP, C‐reactive protein; MRI, magnetic resonance imaging; nr‐axSpA, nonradiographic axial spondylarthritis; r‐axSpA, radiographic axial spondylarthritis.
Data extraction
Relevant data were extracted by one investigator (PK) and checked by two other authors (SG and RS). Full texts were considered in preference to conference abstracts unless the abstracts contained updated results. Additional summary data on subgroups were requested from pharmaceutical companies funding the RCTs and authors of the articles for investigator‐initiated RCTs, where required.
Outcomes
Study outcomes were the difference in the Assessment of Spondyloarthritis International Society 40 (ASAS40) and Bath Ankylosing Spondylitis Disease Activity Index 50 (BASDAI50) in the drug versus placebo group. ASAS40 and BASDAI50 were chosen based on our preliminary review, which showed that most studies reported these measures. Additionally, these measures did not include the CRP, which is a criterion for our subgroup classification. We compared these measures between drug and placebo groups in the four different subgroups based on MRI positivity (MRI+) and elevated CRP levels (CRP+) at baseline: 1) MRI+/CRP+, 2) MRI+/CRP−, 3) MRI−/CRP+, and 4) MRI−/CRP−. The fourth subgroup (MRI−/CRP−) was not included in most trials because this group typically would not fulfill the ASAS40 classification criteria for axSpA (11).
Statistical analysis
Data from included studies were combined for meta‐analysis using Review Manager (RevMan) version 5.4. Study‐specific risk ratios (RRs) (at 12‐16 weeks) were pooled using a random‐effects model (12). The number needed to treat (NNT) was calculated as the inverse of the risk difference. Subgroup analyses were performed by treatment groups (eg, tumor necrosis factor inhibitors [TNFi] and interleukin‐17 inhibitors [IL‐17i]).
Assessment of risk of bias, certainty in the evidence, and heterogeneity
Studies were independently evaluated by two reviewers for risk of bias using the revised Cochrane risk of bias tool (RoB 2.0) for RCTs (13). We reviewed the subgroup analysis (post hoc) study together with its original RCT and substudies while assessing the risk of bias. The certainty in evidence was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation methodology (14). Publication bias was assessed visually using funnel plots (Supplementary Material 2). Between‐study heterogeneity was assessed using I2 statistics (ie, <30% = low heterogeneity, 30%‐60% = moderate, and >60% = high) (15). Sensitivity analyses were performed to explore the likely causes of heterogeneity.
RESULTS
Characteristics of the included studies
Of 2952 studies screened, five RCTs fulfilled our inclusion criteria. Among the included studies, there were three abstracts (16, 17, 18), one RCT article (19), and one post hoc analysis of an RCT (20) (Table 1). All studies included patients with nr‐axSpA, and we did not find any r‐axSpA studies fulfilling our inclusion criteria. Post hoc analysis of Ixekizumab in bDMARD‐Naive and TNF Inhibitor Experienced Participants With Radiographic Axial Spondyloarthritis (COAST‐V and COAST‐W respectively) in r‐axSpA could not be included because this study only provided data on treatment efficacy based on the MRI and CRP subgroups alone, and data on the four exclusive groups were not available (21). nr‐axSpA was defined in the included studies based on ASAS classification criteria for nr‐axSpA (11). MRI positivity (MRI+) was defined by the presence of inflammatory lesions on the SIJ according to the ASAS definition (22). All studies had a central reader for MRI interpretation, except the study comparing etanercept against a placebo for etanercept on a background nonsteroidal anti‐inflammatory drug in the treatment of early spondyloarthritis patients who do not have x‐ray structural changes (EMBARK) study, in which a single reader (one of the two radiologists) read the MRI. Only one study used a Spondyloarthritis Research Consortium of Canada (SPARCC) score of 2 or more as a criterion for defining MRI+ (21). CRP positivity (CRP+) was defined as a value above the upper limit of normal, and the cutoffs varied from 0 to 10 mg/l. MRI and CRP measures were obtained during the initial screening period for all the included studies.
Table 1.
Baseline characteristics of the included studies
| Author, year | Clinical trial, ID | Study population | Drug | Concomitant medications allowed | Duration (weeks) a | Number randomized (Drug/PBO) | MRI and CRP positivity definitions | Outcomes reported |
|---|---|---|---|---|---|---|---|---|
| Sieper et al, 2016 (abstract) (16) | GO‐AHEAD, NCT01453725 | nr‐axSpA | GOL | NSAIDs | 16 | 197 (97/100) |
MRI+: ASAS definition of a positive MRI finding, assessed by a central reader. CRP+: >ULN |
Primary: ASAS20 Secondary: ASAS40, BASDAI50, and ASAS PR |
| Brown et al, 2018 (20) | EMBARK, NCT01258738 | nr‐axSpA | ETN | NSAIDs, CS (<10 mg/day), MTX, HCQ, SSZ | 12 | 215 (106/109) |
MRI+: ASAS definition of a positive MRI finding, single reader (one of the two radiologists) CRP+: >3 mg/l |
Primary: ASAS40 Secondary: ASDAS‐CRP, BASDAI50 |
| Deodhar et al, 2020 (abstract) (18) | C‐axSpAnd, NCT02552212 | nr‐axSpA | CZP | NSAIDs, conventional DMARDs | 12 | 317 (159/158) |
MRI+: ASAS definition of a positive MRI finding, read centrally by two readers (and by an adjudicator if necessary) CRP+: >ULN (defined as 10.0 mg/l) |
ASDAS MI, ASAS40 |
| Maksymowych et al, 2020 (abstract) (21) | COAST‐X, NCT02757352 | nr‐axSpA | IXE | NSAIDs, CS (<10 mg/day), MTX, HCQ, SSZ | 16 | 239 (155/84) |
MRI+: MRI SPARCC ≥2 CRP+: CRP > 5 mg/l |
Primary: ASAS40 Secondary: ASDAS, BASDAI50, SF‐36 PCS |
| Braun et al, 2021 (19) | PREVENT, NCT02696031 | nr‐axSpA | SEC | NSAIDs, SSZ, MTX, CS (<10 mg/day) | 16 | 555 (369/186) |
MRI+: ASAS definition of a positive MRI finding, assessed by a central reader CRP+: >ULN (hsCRP >5 mg/l) by the central laboratory |
Primary: ASAS40 Secondary: BASDAI50, ASAS PR, and ASDAS‐CRP ID |
Abbreviations: ASAS, Assessment of Spondyloarthritis International Society; ASDAS, Ankylosing Spondylitis Disease Activity Score; BASDAI50, Bath Ankylosing Spondylitis Disease Activity Index; C‐axSpAnd, Certolizumab Pegol Compared to Placebo in Subjects With axSpA Without X‐ray Evidence of AS; COAST‐X, Ixekizumab (LY2439821) in Participants With Nonradiographic Axial Spondyloarthritis; CRP, C‐reactive protein; CS, systemic corticosteroids; CZP, certolizumab pegol; DMARD, disease‐modifying antirheumatic drug; EMBARK, Study Comparing Etanercept Against a Placebo for Etanercept on a Background Nonsteroidal Anti‐Inflammatory Drug in the Treatment of Early Spondyloarthritis Patients Who do Not Have X‐ray Structural Changes; ETN, etanercept; GO‐AHEAD, Effect of Golimumab in Participants With Active Axial Spondyloarthritis; GOL, golimumab; HCQ, hydroxychloroquine; hsCRP, high‐sensitivity C‐reactive protein; IXE, ixekizumab; LDA, low‐disease activity; MI, major improvement; MRI, magnetic resonance imaging; MTX, methotrexate; nr‐axSpA, nonradiographic axial spondylarthritis; NSAID, nonsteroidal antiinflammatory drug; PBO, placebo; PR, Partial Response; PREVENT, Study of Efficacy and Safety of Secukinumab in Patients With Non‐radiographic Axial Spondyloarthritis; ULN, upper limit of normal; SEC, secukinumab; SF‐36 PCS, Short Form 36 physical component summary; SPARCC, Spondyloarthritis Research Consortium of Canada; SSZ, sulfasalazine.
Controlled phase duration.
All included studies were double‐blind RCTs using placebo as the control group. The drugs studied were TNFi in three studies (N = 729) (16, 18, 20) and IL‐17i in two studies (N = 794) (17, 19). The effect of golimumab in participants with active axial spondyloarthritis (GO‐AHEAD) study using golimumab (GOL) allowed concomitant use of only nonsteroidal antiinflammatory drugs (NSAIDs) (16); however, the other four studies continued to receive concomitant medications, including NSAIDs, corticosteroids of less than 10 mg, and conventional disease‐modifying antirheumatic drugs. Among the TNFi, the follow‐up duration was 12 weeks for etanercept and certolizumab studies (18, 20) and 16 weeks for the GOL study (16). In the IL‐17i studies, both studies had a controlled phase duration of 16 weeks (17, 19). Detailed baseline characteristics by MRI and CRP subgroups were not available for most studies; however, the number of patients seemed balanced in the drug and placebo groups by different subgroups (Table 2).
Table 2.
ASAS40 and BASDAI50 responses in the MRI/CRP subgroups of the included studies
| Author, year | Duration (weeks) a | Number randomized (drug/PBO) | Drug | Subgroups | n (drug, PBO) | ASAS40 | BASDAI50 | ||
|---|---|---|---|---|---|---|---|---|---|
| Drug, n (%) | PBO, n (%) | Drug, n (%) | PBO, n (%) | ||||||
| Sieper et al, 2016 (abstract) (16) | 16 | 197 (97/100) | GLM | CRP+/MRI+ | 53 (26, 27) | 18 (69%) | 7 (26%) | 18 (69%) | 10 (37%) |
| CRP+/MRI− | 27 (13, 14) | 10 (77%) | 3 (21%) | 11 (85%) | 4 (29%) | ||||
| CRP−/MRI+ | 78 (39, 39) | 19 (49%) | 8 (21%) | 17 (44%) | 9 (23%) | ||||
| CRP−/MRI− | 39 (19, 20) | 8 (42%) | 5 (25%) | 10 (53%) | 7 (35%) | ||||
| Brown et al, 2018 (20) | 12 | 215 (106/109) | ETN | CRP+/MRI+ | 77 (41, 36) | 19 (46%) | 8 (22%) | 26 (63%) | 11 (30%) |
| CRP+/MRI− | 15 (7, 7) | 4 (57%) | 0 (0%) | 4 (58%) | 0 (0%) | ||||
| CRP−/MRI+ | 97 (46, 50) | 10 (22%) | 8 (16%) | 14 (30%) | 13 (26%) | ||||
| CRP−/MRI− | 26 (11, 14) | 2 (18%) | 0 (0%) | 2 (18%) | 2 (14%) | ||||
| Deodhar et al, 2020 (abstract) (18) | 12 | 317 (159/158) | CZP | CRP+/MRI+ | 87 (45, 42) | 26 (58%) | 6 (15%) | NA | NA |
| CRP+/MRI− | 80 (40, 40) | 16 (40%) | 2 (5%) | NA | NA | ||||
| CRP−/MRI+ | 150 (74, 76) | 34 (46%) | 10 (13%) | NA | NA | ||||
| Maksymowych et al, 2020 (abstract) (21) | 16 | IXE Q4W | CRP+/MRI+ | 58 (30, 28) | 12 (40%) | 5 (18%) | 13 (43%) | 5 (18%) | |
| CRP+/MRI− | 51 (22, 29) | 7 (32%) | 3 (10%) | 5 (23%) | 3 (10%) | ||||
| CRP−/MRI+ | 52 (25, 27) | 12 (48%) | 6 (22%) | 10 (40%) | 5 (19%) | ||||
| 16 | 239 (155/84) | IXE Q2W | CRP+/MRI+ | 52 (24, 28) | 10 (42%) | 5 (18%) | 8 (33%) | 5 (18%) | |
| CRP+/MRI− | 51 (22, 29) | 10 (46%) | 3 (10%) | 10 (46%) | 3 (10%) | ||||
| CRP−/MRI+ | 59 (32, 27) | 11 (34%) | 6 (22%) | 9 (28%) | 5 (19%) | ||||
| Braun et al, 2021 (19) | 16 | 555 (369/186) | SEC | CRP+/MRI+ | 166 (111, 55) | 47 (52%) | 12 (22%) | 50 (45%) | 7 (13%) |
| CRP+/MRI− | 154 (103, 51) | 34 (33%) | 15 (29%) | 34 (33%) | 13 (26%) | ||||
| CRP−/MRI+ | 235 (155, 80) | 57 (37%) | 25 (31%) | 54 (35%) | 19 (24%) | ||||
Abbreviations: ASAS, Assessment of Spondyloarthritis International Society; BASDAI50, Bath Ankylosing Spondylitis Disease Activity Index; CRP, C‐reactive protein; CZP, certolizumab pegol; ETN, etanercept; GLM, golimumab; IXE, ixekizumab; MRI, magnetic resonance imaging; NA, not applicable; Q2W, every 2 weeks; Q4W, every 4 weeks; SEC, secukinumab.
Controlled phase duration.
Treatment efficacy: ASAS40 and BASDAI50
All studies included ASAS40 as their primary outcome, except the subgroup analysis from the certolizumab pegol compared to placebo in subjects with axspa without x‐ray evidence of as (C‐axSpAnd) study (18). This study also reported the Ankylosing Spondylitis Disease Activity Score (ASDAS) major improvement; however, no studies reported BASDAI50. All other studies reported BASDAI50 as their secondary outcome (Table 1). Stratified analyses on baseline MRI and CRP were prespecified for all the included studies.
There was no significant difference in ASAS40 (P = 0.65, I2 = 0%) or BASDAI50 (P = 0.27, I2 = 24.1%) between the MRI and CRP subgroups. Treatment efficacy compared with placebo was statistically significant for the subgroups, except for the CRP−/MRI− subgroup (Figures 2 and 3), which included only two studies with 65 patients (Table 2) (20, 23). The NNT was lower for the CRP+ compared with the CRP− subgroup. For ASAS40, NNTs were three for the CRP+/MRI+ and CRP+/MRI− subgroups and six for the CRP−/MRI+ and CRP−/MRI− subgroups. For BASDAI50, NNTs were three for the CRP+/MRI+ and CRP+/MRI− subgroups, eight for the CRP−/MRI+ subgroup, and 10 for the CRP−/MRI− subgroup. A sensitivity analysis on TNFi and IL‐17i studies separately showed no difference between subgroups in terms of ASAS40 (TNFi, P = 0.88; IL‐17i, P = 0.28) or BASDAI50 (TNFi, P = 0.34; IL‐17i, P = 0.18) (Supplementary Materials 4 and 5). Similarly, when we examined subgroups based on either MRI or CRP, there were no statistically significant differences in MRI+ versus MRI− subgroups (P = 0.46, I2 = 0%) and CRP+ versus CRP− subgroups (P = 0.27, I2 = 26.1) (Supplementary Materials 6 and 7). However, numerically higher efficacy was noted in the CRP+ compared with the CRP− subgroup, and NNTs for CRP+ and CRP− were three and six, respectively. NNTs for MRI+ and MRI− were the same (four for both groups).
Figure 2.
Forest plot of ASAS40 response in the study drug (TNFi and IL‐17i) versus placebo in the four MRI/CRP subgroups. ASAS40, Assessment of Spondyloarthritis International Society 40; CI, confidence interval; CRP, C‐reactive protein; df, degrees of freedom; IL‐17i, interleukin‐17 inhibitor; IV, Inverse variance; MRI, magnetic resonance imaging; PBO, placebo; TNFi, tumor necrosis factor inhibitor.
Figure 3.
Forest plot of BASDAI50 response in the study drug (TNFi and IL‐17i) versus placebo in the four MRI/CRP subgroups. BASDAI50, Bath Ankylosing Spondylitis Disease Activity Index 50; CI, confidence interval; CRP, C‐reactive protein; df, degrees of freedom; IL‐17i, interleukin‐17 inhibitor; IV, Inverse variance; MRI, magnetic resonance imaging; PBO, placebo; TNFi, tumor necrosis factor inhibitor.
Statistical evaluation of publication bias was not feasible because only five trials were available. The included studies were judged to have low risk of bias (Supplementary Material 3A). Between‐study heterogeneity within each subgroup assessed using I2 statistics was low to moderate. Certainty in evidence for the outcome of efficacy rate when comparing all biologics with placebo was moderate despite some indirectness due to variation in MRI approaches and not taking into account the degree of inflammation (eg, SPARCC score), and some inconsistency between the estimates (Supplemental Material 3B).
DISCUSSION
This systematic review of five RCTs showed that all MRI and CRP subgroups in nr‐axSpA showed treatment efficacy (in terms of ASAS40 or BASDAI50), except the CRP−/MRI− subgroup, which was underpowered and reported in only two studies. No significant difference in outcomes was noted between the subgroups based on MRI and CRP positivity. Although somewhat numerically higher treatment efficacy was noted in the CRP+ group, treatment efficacy was not different based on MRI positivity.
Our study showed numerically higher treatment efficacy in the CRP+ subgroup compared with the CRP− subgroup (RR 2.5 vs. 1.9), and the NNT was lowest for the CRP+ subgroup (NNT 3 vs. 6). However, the difference in treatment efficacy among the CRP subgroups was not statistically different (P = 0.24). Several studies have shown that baseline CRP can be a predictor of treatment response to biologics in axSpA (24). A real‐world study of 1250 patients with ankylosing spondylitis (AS) showed a 7% increase in the likelihood of achieving ASAS40 response at week 12 of TNFi therapy with each 1‐mg/dl increase in the CRP level (25). Similarly, the treatment response seen in the RAPID‐axSpA trial, which required an elevated CRP level (CRP > 7.9 mg/l) even in the r‐axSpA group, was much better compared with other studies (26). Most studies included in our analysis did not show a significant difference between treatment responses in drug versus placebo groups in the CRP− subgroup (16, 19, 20, 21). However, our pooled analysis showed drug efficacy in this CRP− subgroup as well (Supplemental Materials 7A and B). The individual studies had a large confidence interval for their RRs and were most likely underpowered to show a difference within this CRP− subgroup.
Our study did not find a difference in the treatment efficacy based on MRI positivity. In fact, the NNTs for both MRI+ and MRI− groups were the same (NNT = 4). Studies examining the association between MRI positivity and treatment response in axSpA show disparate results. This is likely related to the definition of MRI+ used in the studies. MRI+ was defined by fulfillment of ASAS classification criteria in all but one study (21) in our systematic review. All nr‐axSpA clinical trials require fulfillment of ASAS classification criteria for MRI if CRP is negative. However, this does not necessarily translate to a higher burden of axSpA. Similar to our study, Rudwaleit et al did not find a statistically significant association of MRI SIJ inflammation, and only high CRP levels and widespread inflammation of the spine were associated with a BASDAI50 response in this study of patients from two different RCTs with longstanding AS (mean disease duration 14 years) (2). Similarly, a subgroup analysis of patients from the adalimumab in patients with non‐radiographic axial spondyloarthritis (ABILITY‐1) study did not show a difference in ASAS40 response based on the presence of MRI SIJ inflammation or a SPARCC score of 2 or more with TNFi efficacy in axSpA (2, 27). The latter study, however, did show a numerically higher response to adalimumab with a SPARCC score of 2 or more. More recently, a post hoc analysis of the ABILITY‐3 study showed that higher SPARCC MRI SIJ scores predicted clinical remission (ASDAS ID (inactive disease) and ASAS ID (inactive disease) in active nr‐axSpA at weeks 12 and 28 following treatment with adalimumab (28, 29). Therefore, validated quantitative MRI scoring systems, such as the SPARCC scoring method for inflammation, are likely better predictors of treatment efficacy compared with MRI positivity according to ASAS criteria (30).
Some studies have suggested a graded treatment response based on MRI and CRP status, with the best response seen in double‐positive (MRI+/CRP+) patients (18, 20). The differences between the MRI and CRP subgroups were smaller when ASAS40 and BASDAI50 were examined as opposed to ASDAS (18), which includes CRP as one of the main contributing factors. Similarly, our pooled analysis did not show a significant difference in the ASAS40 and BASDAI50 responses among the MRI and CRP subgroups. The sensitivity analysis with subgroups based on either MRI or CRP alone (MRI+ vs. MRI− and CRP+ vs. CRP−) also suggested no significant difference. Therefore, our study showed that TNFi and IL‐17i efficacy in nr‐axSpA might not be significantly different among the MRI and CRP subgroups. Beyond the included studies, post hoc analysis of the COAST‐V/W trials in r‐axSpA did not find any difference in treatment efficacy among the MRI and CRP subgroups, and treatment efficacy was noted irrespective of baseline serum CRP levels or spinal MRI score (21). Of note, the study examined MRI spinal SPARCC scores and not SIJ. This study did not report exclusive MRI and CRP subgroups as outlined in our inclusion criteria and hence excluded. Only two of the studies in our systematic review included patients with MRI− and CRP− fulfilling the clinical arm of the ASAS classification (HLA‐B27+/MRI−). No superior efficacy was noted in the treatment arm compared with the placebo arm in this subgroup, which was likely underpowered to note a difference (20, 23). Also, a possibility of a wrong axSpA diagnosis should be considered in this subgroup, as is recommended with nonresponse by the 2022 ASAS European alliance of associations for rheumatology (EULAR) guidelines (31). Therefore, there is an unmet need to study this subgroup.
Our systematic review has several limitations. First, our study was limited to patients with nr‐axSpA, and it is unclear if the study results are generalizable to r‐axSpA. Next, we were not able to study patients in the CRP−/MRI− subgroup, and only two studies reported data on this subgroup and were underpowered. Third, the definitions of MRI and CRP positivity varied among the included studies, and a binary classification may not accurately capture the spectrum of CRP or MRI changes. The ASAS definition of MRI+ used in most studies may not amount to significant bone marrow edema; therefore, future studies should define MRI+ in terms of SPARCC scores with a uniform cutoff. Lastly, between‐study heterogeneity was low to moderate. The sensitivity analysis for studies with TNFi and IL‐17i compared with placebo did not show any statistically significant difference between the groups in terms of ASAS40 or BASDAI50. This stratification by therapy reduced some within‐group heterogeneity, but it does not explain all of it.
In conclusion, our systematic review showed that there was no difference in outcomes between subgroups based on MRI and CRP characteristics. Although numerically higher treatment responses were noted with high baseline CRP levels, treatments were effective regardless of MRI and CRP positivity. Future studies should investigate the association of the degree of MRI inflammation (eg, SPARCC scores) and CRP elevation with treatment response in axSpA.
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article and revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Karmacharya had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design
Gensler.
Acquisition of data
Karmacharya, Gensler.
Analysis and interpretation of data
Karmacharya, Gupta, Shahukhal, Khanal, Murad, Gensler.
Supporting information
Disclosure Form
Supplementary File 1.
Supplementary File 2.
Supplementary File 3.
Supplementary File 4A.
Supplementary File 4B.
Supplementary File 5A.
Supplementary File 5B.
Supplementary File 6A.
Supplementary File 6B.
Supplementary File 7A.
Supplementary File 7B.
ACKNOWLEDGMENT
We would like to thank our librarian, Larry Prokop, for his help with the extensive search strategy.
Funded by the fellowship grant from the Spondyloarthritis Research and Therapy Network. Dr. Karmacharya's work was supported by the Rheumatology Research Foundation Scientist Development Award.
Author disclosures are available at https://onlinelibrary.wiley.com/doi/10.1002/acr2.11581.
REFERENCES
- 1. Cruz‐Machado AR, Rodrigues‐Manica S, Silva JL, et al. Effect of biologic disease‐modifying anti‐rheumatic drugs targeting remission in axial spondyloarthritis: systematic review and meta‐analysis. Rheumatology (Oxford) 2020;59:3158–71. [DOI] [PubMed] [Google Scholar]
- 2. Rudwaleit M, Schwarzlose S, Hilgert ES, et al. MRI in predicting a major clinical response to anti‐tumour necrosis factor treatment in ankylosing spondylitis. Ann Rheum Dis 2008;67:1276–81. [DOI] [PubMed] [Google Scholar]
- 3. Maneiro JR, Souto A, Salgado E, et al. Predictors of response to TNF antagonists in patients with ankylosing spondylitis and psoriatic arthritis: systematic review and meta‐analysis. RMD Open 2015;1:e000017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Rudwaleit M, Haibel H, Baraliakos X, et al. The early disease stage in axial spondylarthritis: results from the German Spondyloarthritis Inception Cohort. Arthritis Rheum 2009;60:717–27. [DOI] [PubMed] [Google Scholar]
- 5. Rudwaleit M, Landewé R, van der Heijde D, et al. The development of Assessment of SpondyloArthritis international Society classification criteria for axial spondyloarthritis. Part I. Classification of paper patients by expert opinion including uncertainty appraisal. Ann Rheum Dis 2009;68:770–6. [DOI] [PubMed] [Google Scholar]
- 6. Tsang HH, Chung HY. The discriminative values of the Bath Ankylosing Spondylitis Disease Activity Index, Ankylosing Spondylitis Disease Activity Score, C‐reactive protein, and erythrocyte sedimentation rate in spondyloarthritis‐related axial arthritis. J Clin Rheumatol 2017;23:267–72. [DOI] [PubMed] [Google Scholar]
- 7. Machado P, Landewé RB, Braun J, et al. MRI inflammation and its relation with measures of clinical disease activity and different treatment responses in patients with ankylosing spondylitis treated with a tumour necrosis factor inhibitor. Ann Rheum Dis 2012;71:2002–5. [DOI] [PubMed] [Google Scholar]
- 8. Bredella MA, Steinbach LS, Morgan S, et al. MRI of the sacroiliac joints in patients with moderate to severe ankylosing spondylitis. AJR Am J Roentgenol 2006;187:1420–6. [DOI] [PubMed] [Google Scholar]
- 9. MacKay JW, Aboelmagd S, Gaffney JK. Correlation between clinical and MRI disease activity scores in axial spondyloarthritis. Clin Rheumatol 2015;34:1633–8. [DOI] [PubMed] [Google Scholar]
- 10. Moher D, Liberati A, Tetzlaff J, et al. Preferred Reporting Items for Systematic Reviews and Meta‐Analyses: the PRISMA statement. Ann Intern Med 2009;151:264–9. [DOI] [PubMed] [Google Scholar]
- 11. Rudwaleit M, van der Heijde D, Landewé R, et al. The Assessment of SpondyloArthritis International Society classification criteria for peripheral spondyloarthritis and for spondyloarthritis in general. Ann Rheum Dis 2011;70:25–31. [DOI] [PubMed] [Google Scholar]
- 12. DerSimonian R, Laird N. Meta‐analysis in clinical trials. Control Clin Trials 1986;7:177–88. [DOI] [PubMed] [Google Scholar]
- 13. Cohrane.org . Risk of bias 2.0 tool. URL: /risk‐bias‐20‐tool.
- 14. Murad MH. Clinical practice guidelines: a primer on development and dissemination. Mayo Clin Proc 2017;92:423–33. [DOI] [PubMed] [Google Scholar]
- 15. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta‐analysis. Stat Med 2002;21:1539–58. [DOI] [PubMed] [Google Scholar]
- 16. Sieper J, van der Heijde D, Maksymowych W, et al. Efficacy of golimumab for nonradiographic axial spondyloarthritis (nr‐axSpA): subgroup analysis by baseline MRI and C‐reactive protein status [abstract]. Arthritis Rheumatol 2016;68 Suppl 10. URL: https://acrabstracts.org/abstract/efficacy‐of‐golimumab‐for‐nonradiographic‐axial‐spondyloarthritis‐nr‐axspa‐subgroup‐analysis‐by‐baseline‐mri‐and‐c‐reactive‐protein‐status/. [Google Scholar]
- 17. Maksymowych W, Marzo‐Ortega H, Ǿstergaard M, et al. Efficacy of ixekizumab on disease activity and quality of life in patients with active nonradiographic axial spondyloarthritis and objective signs of inflammation, stratified by baseline CRP/sacroiliac joint MRI status [abstract]. Arthritis Rheumatol 2020;72 Suppl 10. URL: https://acrabstracts.org/abstract/efficacy‐of‐ixekizumab‐on‐disease‐activity‐and‐quality‐of‐life‐in‐patients‐with‐active‐nonradiographic‐axial‐spondyloarthritis‐and‐objective‐signs‐of‐inflammation‐stratified‐by‐baseline‐crp‐sacroilia/. [Google Scholar]
- 18. Deodhar A, Gensler L, Hall S, et al. Certolizumab pegol efficacy in patients with non‐radiographic axial spondyloarthritis stratified by baseline MRI and C‐reactive protein status [abstract]. Arthritis Rheumatol 2020;72 Suppl 10. URL: https://acrabstracts.org/abstract/certolizumab‐pegol‐efficacy‐in‐patients‐with‐non‐radiographic‐axial‐spondyloarthritis‐stratified‐by‐baseline‐mri‐and‐c‐reactive‐protein‐status/. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Braun J, Blanco R, Marzo‐Ortega H, et al. Secukinumab in non‐radiographic axial spondyloarthritis: subgroup analysis based on key baseline characteristics from a randomized phase III study, PREVENT. Arthritis Res Ther 2021;23:231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Brown MA, Bird PA, Robinson PC, et al. Evaluation of the effect of baseline MRI sacroiliitis and C reactive protein status on etanercept treatment response in non‐radiographic axial spondyloarthritis: a post hoc analysis of the EMBARK study. Ann Rheum Dis 2018;77:1091–3. [DOI] [PubMed] [Google Scholar]
- 21. Maksymowych W, Gallo G, Bolce R, et al. Ixekizumab is effective in the treatment of radiographic axial spondyloarthritis regardless of the level of C‐reactive protein or magnetic resonance imaging scores [abstract]. Arthritis Rheumatol 2019;71 Suppl 10. URL: https://acrabstracts.org/abstract/ixekizumab‐is‐effective‐in‐the‐treatment‐of‐radiographic‐axial‐spondyloarthritis‐regardless‐of‐the‐level‐of‐c‐reactive‐protein‐or‐magnetic‐resonance‐imaging‐scores/. [Google Scholar]
- 22. Rudwaleit M, Jurik AG, Hermann KG, et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification of axial spondyloarthritis: a consensual approach by the ASAS/OMERACT MRI group. Ann Rheum Dis 2009;68:1520–7. [DOI] [PubMed] [Google Scholar]
- 23. Sieper J, van der Heijde D, Dougados M, et al. A randomized, double‐blind, placebo‐controlled, sixteen‐week study of subcutaneous golimumab in patients with active nonradiographic axial spondyloarthritis. Arthritis Rheumatol 2015;67:2702–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Baraliakos X, Szumski A, Koenig AS, et al. The role of C‐reactive protein as a predictor of treatment response in patients with ankylosing spondylitis. Semin Arthritis Rheum 2019;48:997–1004. [DOI] [PubMed] [Google Scholar]
- 25. Rudwaleit M, Claudepierre P, Wordsworth P, et al. Effectiveness, safety, and predictors of good clinical response in 1250 patients treated with adalimumab for active ankylosing spondylitis. J Rheumatol 2009;36:801–8. [DOI] [PubMed] [Google Scholar]
- 26. Landewe R, Braun J, Deodhar A, et al. Efficacy of certolizumab pegol on signs and symptoms of axial spondyloarthritis including ankylosing spondylitis: 24‐week results of a double‐blind randomised placebo‐controlled phase 3 study. Ann Rheum Dis 2014;73:39–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Sieper J, van der Heijde D, Dougados M, et al. Efficacy and safety of adalimumab in patients with non‐radiographic axial spondyloarthritis: results of a randomised placebo‐controlled trial (ABILITY‐1). Ann Rheum Dis 2013;72:815–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Maksymowych WP, Kumke T, Auteri SE, et al. Predictors of long‐term clinical response in patients with non‐radiographic axial spondyloarthritis receiving certolizumab pegol. Arthritis Res Ther 2021;23:274. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Sieper J, Landewé R, Magrey M, et al. Predictors of remission in patients with non‐radiographic axial spondyloarthritis receiving open‐label adalimumab in the ABILITY‐3 study. RMD Open 2019;5:e000917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Maksymowych WP, Inman RD, Salonen D, et al. Spondyloarthritis Research Consortium of Canada magnetic resonance imaging index for assessment of sacroiliac joint inflammation in ankylosing spondylitis. Arthritis Rheum 2005;53:703–9. [DOI] [PubMed] [Google Scholar]
- 31. Ramiro S, Nikiphorou E, Sepriano A, et al. ASAS‐EULAR recommendations for the management of axial spondyloarthritis: 2022 update. Ann Rheum Dis 2023;82:19–34. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Disclosure Form
Supplementary File 1.
Supplementary File 2.
Supplementary File 3.
Supplementary File 4A.
Supplementary File 4B.
Supplementary File 5A.
Supplementary File 5B.
Supplementary File 6A.
Supplementary File 6B.
Supplementary File 7A.
Supplementary File 7B.