https://orcid.org/0000-0002-3626-0694
Abstract
The objective of this study was to identify and summarize the efficacy and safety of systemic glucocorticoids (GCs) and local injections of GC in SpA.
PubMed (Medline) and EMBASE were searched with pre-defined keywords for relevant articles in English reporting randomized controlled trials (RCTs), non-randomized interventional studies and non-randomized observational studies of the efficacy of GC in SpA, with five or more patients, for inclusion in a systematic literature review. Local injections of GC included IA and entheseal injections, but excluded SI joint injections.
Out of 9657 records identified, there were 14 studies on the use of systemic GCs in SpA (364 patients), including two RCTs of oral prednisolone. On pooling data from two placebo-controlled RCTs (≤24 weeks), BASDAI 50 was 4.2 times more likely (95% CI: 1.5, 11.5) and Ankylosing Spondylitis Assessment Group (ASAS) 20 was twice more likely (95% CI: 1.1, 3.64) to occur in patients on high-dose oral prednisolone (± taper). Pulsed GCs led to dramatic improvements that lasted a few weeks to a few months. There were no deaths or major adverse events. There were 10 studies (560 patients) on local GCs in SpA. IA injection was effective in achieving a sustained response in 51.5–90% of joints at 6 months. Entheseal injections led to reduced pain and improved US parameters.
There were limited studies on either systemic or local injections of GCs in SpA. However, there was good evidence of efficacy with the use of high-dose systemic GCs in the short term (≤6 months) in SpA. Both IA and entheseal injections seemed safe and effective.
Pooled analysis showed a significantly higher chance of BASDAI 50, ASAS 20 and ASAS 40 with high-dose (± tapered) prednisolone in axial SpA.
Systemic GCs were safe in the short term, with no major adverse effects; however, the studies were small and short in duration.
Local GCs given into joints or entheses were efficacious, with minimal adverse effects in SpA.
Introduction
Spondyloarthritis (SpA) is a group of inflammatory arthritides affecting the axial skeleton (SI joints and spine) with/without peripheral joints and is often associated with HLA-B27. AS (or radiographic axial SpA) is the prototype disease that is characterized by sacroilitis on plain radiographs with/without bridging syndesmophytes [1]. SpA is estimated to occur in 0.2–1.6% of the population [2], having a similar or higher prevalence than RA [3]. Current therapeutic strategies include physical exercise, NSAIDs and biologics targeting TNFα or IL-17, with a limited role for conventional DMARDs [4, 5].
Glucocorticoids (GCs) are potent anti-inflammatory drugs used extensively in rheumatic autoimmune diseases, like RA, SLE, vasculitis, PMR, etc. As they inhibit both prostaglandin synthesis (like NSAIDs) and TNFα by trans-repression of its gene [6], mechanisms similar to the existing front-line therapies, they have a fair chance of efficacy in SpA. Indeed, soon after the discovery of cortisone, initial reports described a good response to cortisone and adrenocorticotrophic hormone (ACTH) in AS, Reiter’s and PsA [7, 8]. Subsequently, reviews in the 1960 and 1970s on AS hinted that their use was not uncommon [9–11].
There was some research on pulse-dose GCs in the 1980s, but subsequently research was limited. A review 10 years ago commented on the lack of randomized controlled trials (RCTs) [12], and most contemporary treatment guidelines cite a lack of good quality evidence. Despite a lack of evidence either way, clinicians carried a perception that systemic GCs do nothing or were useless in SpA, and use in the clinic also declined [13, 14]. On the other hand, local injections of GCs continued to be used in the clinic for both swollen peripheral joints as well as painful entheses.
In the last decade, we are aware of at least two RCTs; however , there are no recent reviews on systemic or local injections of GC. Thus, we planned this systematic literature review (SLR) to evaluate their efficacy and safety. Additionally, to gain an insight into use in the clinic and views of experts, we examined real-world use from registries and case reports, and summarized recommendations from contemporary guidelines.
Methods
This SLR was conducted in accordance with the principles of ‘Preferred reporting of systematic reviews and metanalysis’ [15]. The patient, intervention, control and outcome (PICO) question was whether GCs, systemic or local injections (IA or entheseal) were efficacious in SpA.
Search
We searched PubMed (Medline) and EMBASE databases by using both free text and medical subject headings (MeSh terms) for diseases (ankylosing spondylitis, spondylarthritis, spondylarthropathy, reactive arthritis, enteropathic arthritis, psoriatic arthritis, psoriatic arthropathy, etc.) and interventions (corticosteroid, glucocorticoid, steroid, prednisolone, prednisone, triamcinolone, etc.). These are detailed in Supplementary Table S1, available at Rheumatology online. Searches were limited to the English language, but were not limited by any time period. In addition, bibliographies of the selected articles, and relevant reviews were searched for additional publications.
Study selection, data extraction and appraisal for the SLR
Studies that reported efficacy or effectiveness of systemic GCs or local injections of GCs in SpA and included five or more subjects were included in this SLR. Apart from RCTs, we also included non-randomized interventional studies and observational prospective or retrospective studies. Studies on both IA and entheseal injections were included, but those on SI joint injections were excluded as that has been reviewed recently [16]. Two independent reviewers screened titles and abstracts of retrieved citations and selected relevant studies. The full text of these articles was then reviewed, and data was extracted from studies that met the inclusion criterion (Supplementary Fig. S1, available at Rheumatology online). The data was entered into Microsoft Excel and reviewed for accuracy. Each study was appraised for quality using the Cochrane risk tool (for RCTs) (Supplementary Fig. S2, available at Rheumatology online) and the Newcastle–Ottawa scale for non-randomized and observational studies (Supplementary Table S2, available at Rheumatology online) [17, 18].
Narrative review on real-world usage and guidelines
We summarized real-world usage from case reports, case series and data from registries. Contemporary guidelines published within the last 10 years were examined for recommendations on GCs in SpA. Use of GCs in animal models of SpA was also reviewed.
Statistical analysis
Data from observational studies were summarized qualitatively in terms of categorical measures of efficacy like BASDAI 50, ASAS 20, ASAS 40 and continuous measures (delta change) in BASDAI, Ankylosing Spondylitis Disease Activity Score (ASDAS)-CRP, BASFI, visual analogue scale (VAS) pain, VAS early morning stiffness and change in patient global assessment. For local injections, outcomes were in terms of sustained response (lack of relapse), improvement in VAS pain, and improvement in parameters on US. Results of RCTs were pooled using the meta-analytic software Revman 5.1 (Review Manager, version 5.1, The Nordic Cochrane Centre, The Cochrane Collaboration) in terms of relative risk (RR) of BASDAI 50, ASAS 20 and ASAS 40. Study heterogeneity was estimated by calculating the I2 test (with a value of >40% indicating significant heterogeneity) and the Cochran Q statistics (with a P < 0.1 being considered significant). In the case of no heterogeneity, we used the fixed-effects model to pool the individual studies [19]. Graphs were made using Graph-Pad Prism (version 8 for Windows, La Jolla, CA, USA).
Results
We screened a total of 9657 records, out of which 107 full-text articles were evaluated and 24 studies were included in the SLR (Supplementary Fig. S1, available at Rheumatology online).
Studies on systemic GCs in axSpA
There were 14 studies on the use of systemic GCs in axSpA (364 patients)—9 on pulsed GCs and 5 on low- to high-dose GCs (Tables 1 and 2).
Characteristics of patients included in studies examining the use of systemic GCs in axSpA
| Serial number | References | Type of study | Country | No. of patients (N) | GC type and dose | Disease | NSAID response | Baseline active disease | Age, mean (range), years | Disease duration, mean (range), years | M:F | HLA-B27, n (%) | Peripheral arthritis |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||||
| 1 | Mintz et al. [20] | Prospective | Mexico | 5 (10 rx) | i.v. MP 1000 mg (1–4 pulses) | AS | NR | + | (24–43) | 5.8 (1–12) | 5:0 | 4 (80) | 1 (20) |
| 2 | Richter et al. [21] | Prospective | UK | 8 | i.v. MP 1000 mg × 3 days | AS | NR | + | 45 | 13 | 1:1 | – | |
| 3 | Ejstrup and Peters [22] | Prospective | Denmark | 7 | i.v. MP 1000 mg × 3 days | AS | NR | + | 33 (28–47) | 13 (6–26) | 2.5:1 | – | |
| 4 | Evard et al. [23] | Cross-over RCTa | UK | 10 | i.v. MP 1000 mg × 3 days + Phys vs Phys | AS | NR | + | 45 | 17.5 | 4:1 | – | |
| 5 | Peters and Ejstrup [24] | RCT | Denmark | 17 | i.v. MP 375 mg × 3 days vs 1000 mg × 3 days | AS | NR | + | 39–41 (25–61) | 7–10 (1–22) | 3:1 | 14 (82) | |
| 6 | Malaviya et al. [25] | Prospective | India | 46 | i.v. MP 500 mg x 3 days every month for 6 monthsb | AS | NR | BASDAI ≥ 4 in 29 | 28 (18–47) | 5 (1–15) | 6:1 | – | |
| 7 | Karaaslan et al. [26] | Retrospective cohort | Turkey | 27 (39 rx) | i.v. MP 500–1000 mg × 1–3 days | AS | NR | BASDAI > 5 | 33 | 5 | 2:1 | (74) | |
| 8 | Gaydukova et al. [27] | Prospective | Russia | 20 | i.v. MP 500 mg × 1 day | AS | NR | BASDAI ≥ 4 | 35.4 | 10.2 | |||
| 9 | Rihl et al. [28] | Retrospective | Germany | 15 | i.v. MP 250–500 mg x 3–5 days | AS | NR | BASDAI ≥ 4 | 47 | 11 | 14:1 | – | – |
| Low, medium or high dose | |||||||||||||
| 1 | Bagnall et al. [29] | Prospective | Canada | 11 | Cortisone 100 mg OD × 1–1.5 months, then for few days every month | AS | − | + | – | – | 11:0 | – | 8 (72) |
| 2 | Haibel et al. [30] | RCT | Germany | 36c | Oral prednisolone 50 mg, 20 mg OD, placebo × 2 wks | AS | NR | BASDAI ≥ 4 | 35–43 | 13–16 | 5:1 | 29 (81) | 6 (17) |
| 3 | Bandinelli et al. [31] | Retrospective | Italy | 57 | Oral prednisone modified release 5 mg OD × 4 wks | AxSpA | NR | BASDAI ≥ 4 | 54.3 | 5.3 | – | – | 31 (54) |
| 4 | Haroon et al. [32] | Prospective | Ireland | 40 | i.m. triamcinolone acetonide 80 mg × 1 dose | 15 AxPsA, 15AS, 10 mech LBP | NR | + | 33.9–38.6 | 4.5–5.4 | 1:1 | – | |
| 5 | Mishra et al. [33] | RCT | India | 65 | Oral prednisolone 60 mg taper → 5 mg OD × 24 wks vs Placebo | AxSpA | NR | BASDAI ≥ 4 | 28.3 | 5.3–5.9 | 21:1 | 56 (86) | 4 (6) |
| Serial number | References | Type of study | Country | No. of patients (N) | GC type and dose | Disease | NSAID response | Baseline active disease | Age, mean (range), years | Disease duration, mean (range), years | M:F | HLA-B27, n (%) | Peripheral arthritis |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||||
| 1 | Mintz et al. [20] | Prospective | Mexico | 5 (10 rx) | i.v. MP 1000 mg (1–4 pulses) | AS | NR | + | (24–43) | 5.8 (1–12) | 5:0 | 4 (80) | 1 (20) |
| 2 | Richter et al. [21] | Prospective | UK | 8 | i.v. MP 1000 mg × 3 days | AS | NR | + | 45 | 13 | 1:1 | – | |
| 3 | Ejstrup and Peters [22] | Prospective | Denmark | 7 | i.v. MP 1000 mg × 3 days | AS | NR | + | 33 (28–47) | 13 (6–26) | 2.5:1 | – | |
| 4 | Evard et al. [23] | Cross-over RCTa | UK | 10 | i.v. MP 1000 mg × 3 days + Phys vs Phys | AS | NR | + | 45 | 17.5 | 4:1 | – | |
| 5 | Peters and Ejstrup [24] | RCT | Denmark | 17 | i.v. MP 375 mg × 3 days vs 1000 mg × 3 days | AS | NR | + | 39–41 (25–61) | 7–10 (1–22) | 3:1 | 14 (82) | |
| 6 | Malaviya et al. [25] | Prospective | India | 46 | i.v. MP 500 mg x 3 days every month for 6 monthsb | AS | NR | BASDAI ≥ 4 in 29 | 28 (18–47) | 5 (1–15) | 6:1 | – | |
| 7 | Karaaslan et al. [26] | Retrospective cohort | Turkey | 27 (39 rx) | i.v. MP 500–1000 mg × 1–3 days | AS | NR | BASDAI > 5 | 33 | 5 | 2:1 | (74) | |
| 8 | Gaydukova et al. [27] | Prospective | Russia | 20 | i.v. MP 500 mg × 1 day | AS | NR | BASDAI ≥ 4 | 35.4 | 10.2 | |||
| 9 | Rihl et al. [28] | Retrospective | Germany | 15 | i.v. MP 250–500 mg x 3–5 days | AS | NR | BASDAI ≥ 4 | 47 | 11 | 14:1 | – | – |
| Low, medium or high dose | |||||||||||||
| 1 | Bagnall et al. [29] | Prospective | Canada | 11 | Cortisone 100 mg OD × 1–1.5 months, then for few days every month | AS | − | + | – | – | 11:0 | – | 8 (72) |
| 2 | Haibel et al. [30] | RCT | Germany | 36c | Oral prednisolone 50 mg, 20 mg OD, placebo × 2 wks | AS | NR | BASDAI ≥ 4 | 35–43 | 13–16 | 5:1 | 29 (81) | 6 (17) |
| 3 | Bandinelli et al. [31] | Retrospective | Italy | 57 | Oral prednisone modified release 5 mg OD × 4 wks | AxSpA | NR | BASDAI ≥ 4 | 54.3 | 5.3 | – | – | 31 (54) |
| 4 | Haroon et al. [32] | Prospective | Ireland | 40 | i.m. triamcinolone acetonide 80 mg × 1 dose | 15 AxPsA, 15AS, 10 mech LBP | NR | + | 33.9–38.6 | 4.5–5.4 | 1:1 | – | |
| 5 | Mishra et al. [33] | RCT | India | 65 | Oral prednisolone 60 mg taper → 5 mg OD × 24 wks vs Placebo | AxSpA | NR | BASDAI ≥ 4 | 28.3 | 5.3–5.9 | 21:1 | 56 (86) | 4 (6) |
Published as a letter to editor.
All patients also received monthly i.v. pamidronate + MTX, SSZ
Randomized 39, analysed 36. AxSpA: axial SpA; AxPsA: axial PsA; i.v. MP: i.v. methylprednisolone pulse; mech LBP: patients of mechanical low-back pain; rx: number of treatments; NR: non-responder; OD: once daily; Phys: physiotherapy, wk: week.
Characteristics of patients included in studies examining the use of systemic GCs in axSpA
| Serial number | References | Type of study | Country | No. of patients (N) | GC type and dose | Disease | NSAID response | Baseline active disease | Age, mean (range), years | Disease duration, mean (range), years | M:F | HLA-B27, n (%) | Peripheral arthritis |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||||
| 1 | Mintz et al. [20] | Prospective | Mexico | 5 (10 rx) | i.v. MP 1000 mg (1–4 pulses) | AS | NR | + | (24–43) | 5.8 (1–12) | 5:0 | 4 (80) | 1 (20) |
| 2 | Richter et al. [21] | Prospective | UK | 8 | i.v. MP 1000 mg × 3 days | AS | NR | + | 45 | 13 | 1:1 | – | |
| 3 | Ejstrup and Peters [22] | Prospective | Denmark | 7 | i.v. MP 1000 mg × 3 days | AS | NR | + | 33 (28–47) | 13 (6–26) | 2.5:1 | – | |
| 4 | Evard et al. [23] | Cross-over RCTa | UK | 10 | i.v. MP 1000 mg × 3 days + Phys vs Phys | AS | NR | + | 45 | 17.5 | 4:1 | – | |
| 5 | Peters and Ejstrup [24] | RCT | Denmark | 17 | i.v. MP 375 mg × 3 days vs 1000 mg × 3 days | AS | NR | + | 39–41 (25–61) | 7–10 (1–22) | 3:1 | 14 (82) | |
| 6 | Malaviya et al. [25] | Prospective | India | 46 | i.v. MP 500 mg x 3 days every month for 6 monthsb | AS | NR | BASDAI ≥ 4 in 29 | 28 (18–47) | 5 (1–15) | 6:1 | – | |
| 7 | Karaaslan et al. [26] | Retrospective cohort | Turkey | 27 (39 rx) | i.v. MP 500–1000 mg × 1–3 days | AS | NR | BASDAI > 5 | 33 | 5 | 2:1 | (74) | |
| 8 | Gaydukova et al. [27] | Prospective | Russia | 20 | i.v. MP 500 mg × 1 day | AS | NR | BASDAI ≥ 4 | 35.4 | 10.2 | |||
| 9 | Rihl et al. [28] | Retrospective | Germany | 15 | i.v. MP 250–500 mg x 3–5 days | AS | NR | BASDAI ≥ 4 | 47 | 11 | 14:1 | – | – |
| Low, medium or high dose | |||||||||||||
| 1 | Bagnall et al. [29] | Prospective | Canada | 11 | Cortisone 100 mg OD × 1–1.5 months, then for few days every month | AS | − | + | – | – | 11:0 | – | 8 (72) |
| 2 | Haibel et al. [30] | RCT | Germany | 36c | Oral prednisolone 50 mg, 20 mg OD, placebo × 2 wks | AS | NR | BASDAI ≥ 4 | 35–43 | 13–16 | 5:1 | 29 (81) | 6 (17) |
| 3 | Bandinelli et al. [31] | Retrospective | Italy | 57 | Oral prednisone modified release 5 mg OD × 4 wks | AxSpA | NR | BASDAI ≥ 4 | 54.3 | 5.3 | – | – | 31 (54) |
| 4 | Haroon et al. [32] | Prospective | Ireland | 40 | i.m. triamcinolone acetonide 80 mg × 1 dose | 15 AxPsA, 15AS, 10 mech LBP | NR | + | 33.9–38.6 | 4.5–5.4 | 1:1 | – | |
| 5 | Mishra et al. [33] | RCT | India | 65 | Oral prednisolone 60 mg taper → 5 mg OD × 24 wks vs Placebo | AxSpA | NR | BASDAI ≥ 4 | 28.3 | 5.3–5.9 | 21:1 | 56 (86) | 4 (6) |
| Serial number | References | Type of study | Country | No. of patients (N) | GC type and dose | Disease | NSAID response | Baseline active disease | Age, mean (range), years | Disease duration, mean (range), years | M:F | HLA-B27, n (%) | Peripheral arthritis |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||||
| 1 | Mintz et al. [20] | Prospective | Mexico | 5 (10 rx) | i.v. MP 1000 mg (1–4 pulses) | AS | NR | + | (24–43) | 5.8 (1–12) | 5:0 | 4 (80) | 1 (20) |
| 2 | Richter et al. [21] | Prospective | UK | 8 | i.v. MP 1000 mg × 3 days | AS | NR | + | 45 | 13 | 1:1 | – | |
| 3 | Ejstrup and Peters [22] | Prospective | Denmark | 7 | i.v. MP 1000 mg × 3 days | AS | NR | + | 33 (28–47) | 13 (6–26) | 2.5:1 | – | |
| 4 | Evard et al. [23] | Cross-over RCTa | UK | 10 | i.v. MP 1000 mg × 3 days + Phys vs Phys | AS | NR | + | 45 | 17.5 | 4:1 | – | |
| 5 | Peters and Ejstrup [24] | RCT | Denmark | 17 | i.v. MP 375 mg × 3 days vs 1000 mg × 3 days | AS | NR | + | 39–41 (25–61) | 7–10 (1–22) | 3:1 | 14 (82) | |
| 6 | Malaviya et al. [25] | Prospective | India | 46 | i.v. MP 500 mg x 3 days every month for 6 monthsb | AS | NR | BASDAI ≥ 4 in 29 | 28 (18–47) | 5 (1–15) | 6:1 | – | |
| 7 | Karaaslan et al. [26] | Retrospective cohort | Turkey | 27 (39 rx) | i.v. MP 500–1000 mg × 1–3 days | AS | NR | BASDAI > 5 | 33 | 5 | 2:1 | (74) | |
| 8 | Gaydukova et al. [27] | Prospective | Russia | 20 | i.v. MP 500 mg × 1 day | AS | NR | BASDAI ≥ 4 | 35.4 | 10.2 | |||
| 9 | Rihl et al. [28] | Retrospective | Germany | 15 | i.v. MP 250–500 mg x 3–5 days | AS | NR | BASDAI ≥ 4 | 47 | 11 | 14:1 | – | – |
| Low, medium or high dose | |||||||||||||
| 1 | Bagnall et al. [29] | Prospective | Canada | 11 | Cortisone 100 mg OD × 1–1.5 months, then for few days every month | AS | − | + | – | – | 11:0 | – | 8 (72) |
| 2 | Haibel et al. [30] | RCT | Germany | 36c | Oral prednisolone 50 mg, 20 mg OD, placebo × 2 wks | AS | NR | BASDAI ≥ 4 | 35–43 | 13–16 | 5:1 | 29 (81) | 6 (17) |
| 3 | Bandinelli et al. [31] | Retrospective | Italy | 57 | Oral prednisone modified release 5 mg OD × 4 wks | AxSpA | NR | BASDAI ≥ 4 | 54.3 | 5.3 | – | – | 31 (54) |
| 4 | Haroon et al. [32] | Prospective | Ireland | 40 | i.m. triamcinolone acetonide 80 mg × 1 dose | 15 AxPsA, 15AS, 10 mech LBP | NR | + | 33.9–38.6 | 4.5–5.4 | 1:1 | – | |
| 5 | Mishra et al. [33] | RCT | India | 65 | Oral prednisolone 60 mg taper → 5 mg OD × 24 wks vs Placebo | AxSpA | NR | BASDAI ≥ 4 | 28.3 | 5.3–5.9 | 21:1 | 56 (86) | 4 (6) |
Published as a letter to editor.
All patients also received monthly i.v. pamidronate + MTX, SSZ
Randomized 39, analysed 36. AxSpA: axial SpA; AxPsA: axial PsA; i.v. MP: i.v. methylprednisolone pulse; mech LBP: patients of mechanical low-back pain; rx: number of treatments; NR: non-responder; OD: once daily; Phys: physiotherapy, wk: week.
Outcomes in studies examining the use of systemic GC in axSpA
| Serial number | References | GC type | Time outcome assessed | ΔPain, 10-cm VAS | ΔEMS, 10-cm VAS | ΔBASDAI | ΔASDAS-CRP | ΔBASFI | BASDAI 50 % | ASAS 20 % | ASAS 40 % |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||
| 1 | Mintz et al. [20] | i.v. MP | 1 day | ↓ 9/10 | ↓ | ||||||
| >10 mo | ↓ 4/5 | ↓5/5 | |||||||||
| 2 | Richter et al. [21] | i.v. MP | 1 wk | −3.5a,* | −80 min* | ||||||
| 4 wks | −3.8a,** | −63.5 min** | |||||||||
| 12 wks | −3.4a,* | −39 min | |||||||||
| 3 | Ejstrup and Peters [22] | i.v. MP | 3 days | −3.8a | |||||||
| 2 wks | −3.5a | ||||||||||
| 12 wks | −2.0a | ||||||||||
| 4 | Evard et al. [23] | i.v. MP + Phys vs Phys | 24 wks | ↓ 7/8 vs 4/10 | |||||||
| 5 | Peters and Ejstrup [24] | i.v. MP 375, 1000 | 3 days | −2.6, −3.2a | −56, −41 mina | ||||||
| 8 wks | −1.8, −2.2a | −49, −27 mina | |||||||||
| 24 wks | −0.8, −1.5a | +11, −11 mina | |||||||||
| 6 | Malaviya et al. [25] | i.v. MP + othersb | 24 wks | − | − | − | − | − | 85 | 85 | − |
| 7 | Karaaslan et al. [26] | i.v. MP | 4 wks | −1.3*** | |||||||
| 8. | Gaydukova et al. [27] | i.v. MP | 2 wks | −2.9*** | −1.27*** | 35 | 55 | 45 | |||
| 12 wks | −3.4*** | −1.18*** | 40 | 65 | 60 | ||||||
| 9 | Rihl et al. [28] | i.v. MP | 1 day | −3.5*** | |||||||
| 24 wks | −70 min*** | −2.1*** | |||||||||
| 48 wks | −54 min*** | −1.9*** | |||||||||
| Low, medium or high dose | |||||||||||
| 1 | Bagnall et al. [29] | Cortisone | 2 yrs | Function improved 8/11 | |||||||
| 2 | Haibel et al. [30] | Pred 50, 20, placebo | 2 wks | −3.25 −1.83, −1.22 | −3.41*, −1.27, −1.03 | −2.39*, −1.19, −0.66 | −1.56**, −1.16**, −0.34 | −1.76, −0.94, −0.74 | 33, 27, 8 | 58, 39, 23 | 33, 27, 15 |
| 3 | Bandinelli et al. [31] | Pred MR | 12 wks | −1.8*** | −3.2*** | −2.5*** | 52.6 | ||||
| 4 | Haroon et al. [32] | i.m.TA (AxPsA, AS, LBP) | 2 wks | −2.46,*, −1.66, −1.0 | −1.93*, −1.13, −0.84 | −−1.43*, −1.03, −0.81 | −2.38*, −0.93, −0.44 | ||||
| 4 wks | −2, −1.33, −1.3 | −1.57**, −0.85, −0.62 | −1.09***, −0.77, −0.73 | −1.76, −0.78, −0.48 | |||||||
| 5 | Mishra et al. [33] | Pred, Placebo | 12 wks | −1.5, −1.3 | −1.0, −0.8 | −1.1, −1.0 | |||||
| 24 wks | −1.9**, −0.5 | −1.1**, −0.3 | −1.3, −0.5 | 37.5**, 9.1 | 43.8, 24.2 | 37.5, 15.2 | |||||
| Serial number | References | GC type | Time outcome assessed | ΔPain, 10-cm VAS | ΔEMS, 10-cm VAS | ΔBASDAI | ΔASDAS-CRP | ΔBASFI | BASDAI 50 % | ASAS 20 % | ASAS 40 % |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||
| 1 | Mintz et al. [20] | i.v. MP | 1 day | ↓ 9/10 | ↓ | ||||||
| >10 mo | ↓ 4/5 | ↓5/5 | |||||||||
| 2 | Richter et al. [21] | i.v. MP | 1 wk | −3.5a,* | −80 min* | ||||||
| 4 wks | −3.8a,** | −63.5 min** | |||||||||
| 12 wks | −3.4a,* | −39 min | |||||||||
| 3 | Ejstrup and Peters [22] | i.v. MP | 3 days | −3.8a | |||||||
| 2 wks | −3.5a | ||||||||||
| 12 wks | −2.0a | ||||||||||
| 4 | Evard et al. [23] | i.v. MP + Phys vs Phys | 24 wks | ↓ 7/8 vs 4/10 | |||||||
| 5 | Peters and Ejstrup [24] | i.v. MP 375, 1000 | 3 days | −2.6, −3.2a | −56, −41 mina | ||||||
| 8 wks | −1.8, −2.2a | −49, −27 mina | |||||||||
| 24 wks | −0.8, −1.5a | +11, −11 mina | |||||||||
| 6 | Malaviya et al. [25] | i.v. MP + othersb | 24 wks | − | − | − | − | − | 85 | 85 | − |
| 7 | Karaaslan et al. [26] | i.v. MP | 4 wks | −1.3*** | |||||||
| 8. | Gaydukova et al. [27] | i.v. MP | 2 wks | −2.9*** | −1.27*** | 35 | 55 | 45 | |||
| 12 wks | −3.4*** | −1.18*** | 40 | 65 | 60 | ||||||
| 9 | Rihl et al. [28] | i.v. MP | 1 day | −3.5*** | |||||||
| 24 wks | −70 min*** | −2.1*** | |||||||||
| 48 wks | −54 min*** | −1.9*** | |||||||||
| Low, medium or high dose | |||||||||||
| 1 | Bagnall et al. [29] | Cortisone | 2 yrs | Function improved 8/11 | |||||||
| 2 | Haibel et al. [30] | Pred 50, 20, placebo | 2 wks | −3.25 −1.83, −1.22 | −3.41*, −1.27, −1.03 | −2.39*, −1.19, −0.66 | −1.56**, −1.16**, −0.34 | −1.76, −0.94, −0.74 | 33, 27, 8 | 58, 39, 23 | 33, 27, 15 |
| 3 | Bandinelli et al. [31] | Pred MR | 12 wks | −1.8*** | −3.2*** | −2.5*** | 52.6 | ||||
| 4 | Haroon et al. [32] | i.m.TA (AxPsA, AS, LBP) | 2 wks | −2.46,*, −1.66, −1.0 | −1.93*, −1.13, −0.84 | −−1.43*, −1.03, −0.81 | −2.38*, −0.93, −0.44 | ||||
| 4 wks | −2, −1.33, −1.3 | −1.57**, −0.85, −0.62 | −1.09***, −0.77, −0.73 | −1.76, −0.78, −0.48 | |||||||
| 5 | Mishra et al. [33] | Pred, Placebo | 12 wks | −1.5, −1.3 | −1.0, −0.8 | −1.1, −1.0 | |||||
| 24 wks | −1.9**, −0.5 | −1.1**, −0.3 | −1.3, −0.5 | 37.5**, 9.1 | 43.8, 24.2 | 37.5, 15.2 | |||||
Values are approximations from graph. bAlso gave monthly pamidronate + DMARDs. *P < 0.05, **P < 0.01, ***P < 0.001. AxSpA: axial SpA; AxPsA: axial PsA; EMS: early morning stiffness; LBP: low back pain; MP: methylprednisolone pulse; mech LBP, patients of mechanical low back ache; rx, number of treatments; NR, Non-responder; Phys, Physiotherapy; TA: triamcinolone acetonide; wk, week; mo, months; yrs, years.
Outcomes in studies examining the use of systemic GC in axSpA
| Serial number | References | GC type | Time outcome assessed | ΔPain, 10-cm VAS | ΔEMS, 10-cm VAS | ΔBASDAI | ΔASDAS-CRP | ΔBASFI | BASDAI 50 % | ASAS 20 % | ASAS 40 % |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||
| 1 | Mintz et al. [20] | i.v. MP | 1 day | ↓ 9/10 | ↓ | ||||||
| >10 mo | ↓ 4/5 | ↓5/5 | |||||||||
| 2 | Richter et al. [21] | i.v. MP | 1 wk | −3.5a,* | −80 min* | ||||||
| 4 wks | −3.8a,** | −63.5 min** | |||||||||
| 12 wks | −3.4a,* | −39 min | |||||||||
| 3 | Ejstrup and Peters [22] | i.v. MP | 3 days | −3.8a | |||||||
| 2 wks | −3.5a | ||||||||||
| 12 wks | −2.0a | ||||||||||
| 4 | Evard et al. [23] | i.v. MP + Phys vs Phys | 24 wks | ↓ 7/8 vs 4/10 | |||||||
| 5 | Peters and Ejstrup [24] | i.v. MP 375, 1000 | 3 days | −2.6, −3.2a | −56, −41 mina | ||||||
| 8 wks | −1.8, −2.2a | −49, −27 mina | |||||||||
| 24 wks | −0.8, −1.5a | +11, −11 mina | |||||||||
| 6 | Malaviya et al. [25] | i.v. MP + othersb | 24 wks | − | − | − | − | − | 85 | 85 | − |
| 7 | Karaaslan et al. [26] | i.v. MP | 4 wks | −1.3*** | |||||||
| 8. | Gaydukova et al. [27] | i.v. MP | 2 wks | −2.9*** | −1.27*** | 35 | 55 | 45 | |||
| 12 wks | −3.4*** | −1.18*** | 40 | 65 | 60 | ||||||
| 9 | Rihl et al. [28] | i.v. MP | 1 day | −3.5*** | |||||||
| 24 wks | −70 min*** | −2.1*** | |||||||||
| 48 wks | −54 min*** | −1.9*** | |||||||||
| Low, medium or high dose | |||||||||||
| 1 | Bagnall et al. [29] | Cortisone | 2 yrs | Function improved 8/11 | |||||||
| 2 | Haibel et al. [30] | Pred 50, 20, placebo | 2 wks | −3.25 −1.83, −1.22 | −3.41*, −1.27, −1.03 | −2.39*, −1.19, −0.66 | −1.56**, −1.16**, −0.34 | −1.76, −0.94, −0.74 | 33, 27, 8 | 58, 39, 23 | 33, 27, 15 |
| 3 | Bandinelli et al. [31] | Pred MR | 12 wks | −1.8*** | −3.2*** | −2.5*** | 52.6 | ||||
| 4 | Haroon et al. [32] | i.m.TA (AxPsA, AS, LBP) | 2 wks | −2.46,*, −1.66, −1.0 | −1.93*, −1.13, −0.84 | −−1.43*, −1.03, −0.81 | −2.38*, −0.93, −0.44 | ||||
| 4 wks | −2, −1.33, −1.3 | −1.57**, −0.85, −0.62 | −1.09***, −0.77, −0.73 | −1.76, −0.78, −0.48 | |||||||
| 5 | Mishra et al. [33] | Pred, Placebo | 12 wks | −1.5, −1.3 | −1.0, −0.8 | −1.1, −1.0 | |||||
| 24 wks | −1.9**, −0.5 | −1.1**, −0.3 | −1.3, −0.5 | 37.5**, 9.1 | 43.8, 24.2 | 37.5, 15.2 | |||||
| Serial number | References | GC type | Time outcome assessed | ΔPain, 10-cm VAS | ΔEMS, 10-cm VAS | ΔBASDAI | ΔASDAS-CRP | ΔBASFI | BASDAI 50 % | ASAS 20 % | ASAS 40 % |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Pulse dose | |||||||||||
| 1 | Mintz et al. [20] | i.v. MP | 1 day | ↓ 9/10 | ↓ | ||||||
| >10 mo | ↓ 4/5 | ↓5/5 | |||||||||
| 2 | Richter et al. [21] | i.v. MP | 1 wk | −3.5a,* | −80 min* | ||||||
| 4 wks | −3.8a,** | −63.5 min** | |||||||||
| 12 wks | −3.4a,* | −39 min | |||||||||
| 3 | Ejstrup and Peters [22] | i.v. MP | 3 days | −3.8a | |||||||
| 2 wks | −3.5a | ||||||||||
| 12 wks | −2.0a | ||||||||||
| 4 | Evard et al. [23] | i.v. MP + Phys vs Phys | 24 wks | ↓ 7/8 vs 4/10 | |||||||
| 5 | Peters and Ejstrup [24] | i.v. MP 375, 1000 | 3 days | −2.6, −3.2a | −56, −41 mina | ||||||
| 8 wks | −1.8, −2.2a | −49, −27 mina | |||||||||
| 24 wks | −0.8, −1.5a | +11, −11 mina | |||||||||
| 6 | Malaviya et al. [25] | i.v. MP + othersb | 24 wks | − | − | − | − | − | 85 | 85 | − |
| 7 | Karaaslan et al. [26] | i.v. MP | 4 wks | −1.3*** | |||||||
| 8. | Gaydukova et al. [27] | i.v. MP | 2 wks | −2.9*** | −1.27*** | 35 | 55 | 45 | |||
| 12 wks | −3.4*** | −1.18*** | 40 | 65 | 60 | ||||||
| 9 | Rihl et al. [28] | i.v. MP | 1 day | −3.5*** | |||||||
| 24 wks | −70 min*** | −2.1*** | |||||||||
| 48 wks | −54 min*** | −1.9*** | |||||||||
| Low, medium or high dose | |||||||||||
| 1 | Bagnall et al. [29] | Cortisone | 2 yrs | Function improved 8/11 | |||||||
| 2 | Haibel et al. [30] | Pred 50, 20, placebo | 2 wks | −3.25 −1.83, −1.22 | −3.41*, −1.27, −1.03 | −2.39*, −1.19, −0.66 | −1.56**, −1.16**, −0.34 | −1.76, −0.94, −0.74 | 33, 27, 8 | 58, 39, 23 | 33, 27, 15 |
| 3 | Bandinelli et al. [31] | Pred MR | 12 wks | −1.8*** | −3.2*** | −2.5*** | 52.6 | ||||
| 4 | Haroon et al. [32] | i.m.TA (AxPsA, AS, LBP) | 2 wks | −2.46,*, −1.66, −1.0 | −1.93*, −1.13, −0.84 | −−1.43*, −1.03, −0.81 | −2.38*, −0.93, −0.44 | ||||
| 4 wks | −2, −1.33, −1.3 | −1.57**, −0.85, −0.62 | −1.09***, −0.77, −0.73 | −1.76, −0.78, −0.48 | |||||||
| 5 | Mishra et al. [33] | Pred, Placebo | 12 wks | −1.5, −1.3 | −1.0, −0.8 | −1.1, −1.0 | |||||
| 24 wks | −1.9**, −0.5 | −1.1**, −0.3 | −1.3, −0.5 | 37.5**, 9.1 | 43.8, 24.2 | 37.5, 15.2 | |||||
Values are approximations from graph. bAlso gave monthly pamidronate + DMARDs. *P < 0.05, **P < 0.01, ***P < 0.001. AxSpA: axial SpA; AxPsA: axial PsA; EMS: early morning stiffness; LBP: low back pain; MP: methylprednisolone pulse; mech LBP, patients of mechanical low back ache; rx, number of treatments; NR, Non-responder; Phys, Physiotherapy; TA: triamcinolone acetonide; wk, week; mo, months; yrs, years.
Pulsed GCs
The first study by Mintz in 1981 reported dramatic improvement within 24 h after a pulse of methylprednisolone (MP) in NSAID-refractory active AS patients, but found it to be short-lived, lasting only 2 weeks. However, with multiple pulses on consecutive days, the improvement was longer lasting: with three-pulses it lasted for 10 months or more [20]. Based on this, Richter and Ejstrup in two separate studies used a uniform strategy of giving pulses of MP (1000 mg) on 3 consecutive days, and reported significant improvements in pain and mobility that lasted up to 3 months [21, 22].
A letter reported a cross-over RCT between pulsed MP (1000 mg), on alternate days with physiotherapy, to physiotherapy alone, and reported improvements in stiffness but not pain with the former (without giving many details) [23]. In the only RCT published to date on pulsed GCs in SpA, Peters in 1992 compared low-dose with high-dose pulses of MP in axSpA patients, but did not include any placebo arm. They randomized nine patients to 375 mg and eight patients to 1000 mg MP for 3 consecutive days and found immediate pain reductions, with an NSAID-free period lasting a median of 8 and 25 days, respectivelywith gradual increase in pain to pre-treatment levels by 8–12 months, with no significant difference between the two arms [24].
The more recent studies have reported using contemporary indices for disease activity. Malaviya et al. in 2007 studied a regimen of monthly cycles of pulses of MP (500 mg × 3 days) for 6 months, along with pamidronate and DMARDs, and reported an excellent improvement, with BASDAI 50 and ASAS 20 for 85% of the patients in this study [25]. A retrospective study from Germany compared MP pulses with TNFi and reported similar efficacy over a period of 12 months [28]. Two abstracts on pulsed GCs (one a retrospective analysis [26], the other a prospective study [27]) have also been reported. The first reported that patients receiving pulsed MP (1–3 pulses) had significantly reduced BASDAI at 1 month, whereas the other was a prospective study of a single pulse of MP (500 mg) that reported ASAS 20 and 40 in 60% and 65% of patients, respectively, at 12 weeks.
The adverse effects of pulsed GCs were metallic or bitter taste, palpitations, insomnia or irritability, but no serious adverse effects were reported. One study reported two fractures 2–3 years later (Supplementary Table S3, available at Rheumatology online). Another study, which did not do an efficacy analysis, found no rhythm disturbances on holter monitoring after pulsed GCs in axSpA patients [34].
Low-, medium- or high-dose GCs
Soon after the discovery of cortisone, Bagnall et al. conducted a study in 1953 on 11 World War II veterans with AS who were disabled due to disease activity and treated them with cortisone that was given daily for 1–1.5 months and then intermittently (up to a few days every month). There was functional improvement in a majority, with many rejoining active duties [29].
After that, there was mainly research on pulsed GCs in the 1980s–1990s, with no studies on low- to high-dose GCs till recently. In the last decade, two randomized double-blind controlled trials were performed that compared oral prednisolone with placebo over 2–24 weeks (Tables 1 and 2). Haibel et al. randomized 36 axial SpA patients to three groups—daily prednisolone 50 mg, 20 mg or placebo for 2 weeks—and found significant decreases in the BASDAI with 50 mg, but not with 20 mg at 2 weeks; however, ASDAS significantly improved in both GC groups. None of the two doses reached the primary outcome of higher BASDAI 50 [30]. Based on that, we developed a protocol using step-down oral prednisolone and randomized 65 axial SpA patients to prednisolone or placebo for 24 weeks. A significantly higher proportion of patients in the prednisolone group achieved BASDAI 50, clinically significant and major improvements for ASDAS, and ASAS partial remission compared with placebo at 24 weeks. However, there was no significant difference in BASFI [33].
Apart from RCTs, a retrospective Italian study reported that modified-release prednisolone 5 mg daily, led to improvements in both axial and peripheral pain at 12 weeks [31]. In a prospective study, a single injection of i.m. triamcinolone acetonide (80 mg) was administered to patients with either axial PsA, AS, or mechanical low back pain. ASDAS significantly improved in axial PsA, but not in patients with AS [32].
In all the studies on low- to high-dose GCs, only minor side effects like fluid retention, cushingoid facies, weight gain, or hyperglycaemia were reported, and there were no serious adverse effects. (Supplementary Table S3, available at Rheumatology online) There were no studies of systemic GCs for peripheral arthritis in SpA.
Studies on the efficacy of local glucocorticoid injections, both IA and entheseal, in SpA
| Serial number | References | Type of study | Country | Patients (joints/entheses) | Disease, site of injection | Age, yrs | Disease duration, yrs | M:F | GC, US guidance | Follow up | Response ratea | Change in pain | Others |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| IA injections | |||||||||||||
| 1 | Eder et al. [36] | Retrospective, cohort | Canada | 220 (245) | PsA, many jointsb | 45.8 | 9.5 | 1.6:1 | MPA 5–80 mg | 3 months 3 or 6 months | 41.6 51.5 | − | − |
| 2 | Thilaak et al. [37] | Prospective | India | 120 | PsA | 50 − 52 | − | 1:1, 1.7:1 | MPA 20 mg vs etoricoxib 120 mg | 6 weeks | − | −3.1, −2*** | − |
| 3 | Kumar et al. [38] | RCT | India | 100 | 11 SpA, 89 RA, knees | 41.2 − 42.6 | 3.4–5.0 | 1:3 | MPA 80 mg vs TA 80 mg | 6 months | 90 (RA 81.5) | − | − |
| 4 | Weitoft et al. [39] | RCT | Sweden | 159 | 57 PsA, 102 RA, knees | 52.5 − 55 | 14.5 − 14.7 | 1:1.2, 1:2.5 | THA 20 vs 40 mg | 6 months | 66.7 (RA 69.6) | − | − |
| Entheseal injections | |||||||||||||
| 1 | Cunnane et al. [40] | Prospective | Ireland | 8 | Most SpA, heel painc | 42 | 8 | 1:1 | TA 10 mg, USG+ | 4 − 6 weeks | ↓ | Normal gait (except 1) | |
| 2 | Huang et al. [41] | Prospective | China | 12 | Most SpA, Achilles enthesitis | 23.1 | 6.1 | 6:1 | Beta 7 mg vs Eta, USG+ | 12 weeks | −3.8*, −3.7* | ↓BASDAI, USG: ↓ thickness ↓blood flow | |
| 3 | Di Geso et al. [42] | Prospective | Italy | 21 | 12 SpA, 9 RA, tenosynovitis | 44.5 | − | 2:1 | MPA 20 mg, USG+ | 2 weeks | ↓ *** | ↓ tenderness ***, USG: ↓ GS score***, ↓PD*** | |
| 4 | Shrivastava et al. [43] | Prospective | India | 18 (27) | AS, achilles tendonitis | 26.1 | − | 8:1 | MPA 20 mg, USG+ | 6 weeks | −4*** | USG: ↓ thickness, ↓oedema, ↓ PD | |
| 5 | Gutierrez et al. [44] | Prospective | Italy | 114 | 56 SpA tenosynovitisd | − | − | − | MPA 20 mg, USG+ vs − | 1 − 4 weeks | USG: ↓PD | ||
| 6 | Girolimetto et al. [45] | Prospective | Italy | 46 (73) | PsA dactylitis | 46.2 | 4.4 | 29:17 | MPA 20 mg vs NSAID | 12 weeks | −5.5***, −1# | VAS-FI, LDI-b, FC better in GC | |
| Serial number | References | Type of study | Country | Patients (joints/entheses) | Disease, site of injection | Age, yrs | Disease duration, yrs | M:F | GC, US guidance | Follow up | Response ratea | Change in pain | Others |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| IA injections | |||||||||||||
| 1 | Eder et al. [36] | Retrospective, cohort | Canada | 220 (245) | PsA, many jointsb | 45.8 | 9.5 | 1.6:1 | MPA 5–80 mg | 3 months 3 or 6 months | 41.6 51.5 | − | − |
| 2 | Thilaak et al. [37] | Prospective | India | 120 | PsA | 50 − 52 | − | 1:1, 1.7:1 | MPA 20 mg vs etoricoxib 120 mg | 6 weeks | − | −3.1, −2*** | − |
| 3 | Kumar et al. [38] | RCT | India | 100 | 11 SpA, 89 RA, knees | 41.2 − 42.6 | 3.4–5.0 | 1:3 | MPA 80 mg vs TA 80 mg | 6 months | 90 (RA 81.5) | − | − |
| 4 | Weitoft et al. [39] | RCT | Sweden | 159 | 57 PsA, 102 RA, knees | 52.5 − 55 | 14.5 − 14.7 | 1:1.2, 1:2.5 | THA 20 vs 40 mg | 6 months | 66.7 (RA 69.6) | − | − |
| Entheseal injections | |||||||||||||
| 1 | Cunnane et al. [40] | Prospective | Ireland | 8 | Most SpA, heel painc | 42 | 8 | 1:1 | TA 10 mg, USG+ | 4 − 6 weeks | ↓ | Normal gait (except 1) | |
| 2 | Huang et al. [41] | Prospective | China | 12 | Most SpA, Achilles enthesitis | 23.1 | 6.1 | 6:1 | Beta 7 mg vs Eta, USG+ | 12 weeks | −3.8*, −3.7* | ↓BASDAI, USG: ↓ thickness ↓blood flow | |
| 3 | Di Geso et al. [42] | Prospective | Italy | 21 | 12 SpA, 9 RA, tenosynovitis | 44.5 | − | 2:1 | MPA 20 mg, USG+ | 2 weeks | ↓ *** | ↓ tenderness ***, USG: ↓ GS score***, ↓PD*** | |
| 4 | Shrivastava et al. [43] | Prospective | India | 18 (27) | AS, achilles tendonitis | 26.1 | − | 8:1 | MPA 20 mg, USG+ | 6 weeks | −4*** | USG: ↓ thickness, ↓oedema, ↓ PD | |
| 5 | Gutierrez et al. [44] | Prospective | Italy | 114 | 56 SpA tenosynovitisd | − | − | − | MPA 20 mg, USG+ vs − | 1 − 4 weeks | USG: ↓PD | ||
| 6 | Girolimetto et al. [45] | Prospective | Italy | 46 (73) | PsA dactylitis | 46.2 | 4.4 | 29:17 | MPA 20 mg vs NSAID | 12 weeks | −5.5***, −1# | VAS-FI, LDI-b, FC better in GC | |
LDI-b: Leeds’ dactylitis index-basic; FC: finger circumference; MPA: methyl prednisolone acetate; THA: triamcinolone hexacetonide; TA: triamcinolone acetonide; PD: Power doppler signal; USG+: injected under ultrasound guidance; VAS-FI: Visual analogue scale- functional impairment.
Lack of relapse after IA injection,
Finger joints (MCP, PIP, DIP) 44%, knees 21%, wrist 9.4%,
Six Retrocalcaneal Bursitis, three plantar fascia, two posterior tibial tendon,
Finger extensor compartments, flexors, peroneus longus and brevis, tibialis posterior. *P < 0.05, ***P < 0.001.
Studies on the efficacy of local glucocorticoid injections, both IA and entheseal, in SpA
| Serial number | References | Type of study | Country | Patients (joints/entheses) | Disease, site of injection | Age, yrs | Disease duration, yrs | M:F | GC, US guidance | Follow up | Response ratea | Change in pain | Others |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| IA injections | |||||||||||||
| 1 | Eder et al. [36] | Retrospective, cohort | Canada | 220 (245) | PsA, many jointsb | 45.8 | 9.5 | 1.6:1 | MPA 5–80 mg | 3 months 3 or 6 months | 41.6 51.5 | − | − |
| 2 | Thilaak et al. [37] | Prospective | India | 120 | PsA | 50 − 52 | − | 1:1, 1.7:1 | MPA 20 mg vs etoricoxib 120 mg | 6 weeks | − | −3.1, −2*** | − |
| 3 | Kumar et al. [38] | RCT | India | 100 | 11 SpA, 89 RA, knees | 41.2 − 42.6 | 3.4–5.0 | 1:3 | MPA 80 mg vs TA 80 mg | 6 months | 90 (RA 81.5) | − | − |
| 4 | Weitoft et al. [39] | RCT | Sweden | 159 | 57 PsA, 102 RA, knees | 52.5 − 55 | 14.5 − 14.7 | 1:1.2, 1:2.5 | THA 20 vs 40 mg | 6 months | 66.7 (RA 69.6) | − | − |
| Entheseal injections | |||||||||||||
| 1 | Cunnane et al. [40] | Prospective | Ireland | 8 | Most SpA, heel painc | 42 | 8 | 1:1 | TA 10 mg, USG+ | 4 − 6 weeks | ↓ | Normal gait (except 1) | |
| 2 | Huang et al. [41] | Prospective | China | 12 | Most SpA, Achilles enthesitis | 23.1 | 6.1 | 6:1 | Beta 7 mg vs Eta, USG+ | 12 weeks | −3.8*, −3.7* | ↓BASDAI, USG: ↓ thickness ↓blood flow | |
| 3 | Di Geso et al. [42] | Prospective | Italy | 21 | 12 SpA, 9 RA, tenosynovitis | 44.5 | − | 2:1 | MPA 20 mg, USG+ | 2 weeks | ↓ *** | ↓ tenderness ***, USG: ↓ GS score***, ↓PD*** | |
| 4 | Shrivastava et al. [43] | Prospective | India | 18 (27) | AS, achilles tendonitis | 26.1 | − | 8:1 | MPA 20 mg, USG+ | 6 weeks | −4*** | USG: ↓ thickness, ↓oedema, ↓ PD | |
| 5 | Gutierrez et al. [44] | Prospective | Italy | 114 | 56 SpA tenosynovitisd | − | − | − | MPA 20 mg, USG+ vs − | 1 − 4 weeks | USG: ↓PD | ||
| 6 | Girolimetto et al. [45] | Prospective | Italy | 46 (73) | PsA dactylitis | 46.2 | 4.4 | 29:17 | MPA 20 mg vs NSAID | 12 weeks | −5.5***, −1# | VAS-FI, LDI-b, FC better in GC | |
| Serial number | References | Type of study | Country | Patients (joints/entheses) | Disease, site of injection | Age, yrs | Disease duration, yrs | M:F | GC, US guidance | Follow up | Response ratea | Change in pain | Others |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| IA injections | |||||||||||||
| 1 | Eder et al. [36] | Retrospective, cohort | Canada | 220 (245) | PsA, many jointsb | 45.8 | 9.5 | 1.6:1 | MPA 5–80 mg | 3 months 3 or 6 months | 41.6 51.5 | − | − |
| 2 | Thilaak et al. [37] | Prospective | India | 120 | PsA | 50 − 52 | − | 1:1, 1.7:1 | MPA 20 mg vs etoricoxib 120 mg | 6 weeks | − | −3.1, −2*** | − |
| 3 | Kumar et al. [38] | RCT | India | 100 | 11 SpA, 89 RA, knees | 41.2 − 42.6 | 3.4–5.0 | 1:3 | MPA 80 mg vs TA 80 mg | 6 months | 90 (RA 81.5) | − | − |
| 4 | Weitoft et al. [39] | RCT | Sweden | 159 | 57 PsA, 102 RA, knees | 52.5 − 55 | 14.5 − 14.7 | 1:1.2, 1:2.5 | THA 20 vs 40 mg | 6 months | 66.7 (RA 69.6) | − | − |
| Entheseal injections | |||||||||||||
| 1 | Cunnane et al. [40] | Prospective | Ireland | 8 | Most SpA, heel painc | 42 | 8 | 1:1 | TA 10 mg, USG+ | 4 − 6 weeks | ↓ | Normal gait (except 1) | |
| 2 | Huang et al. [41] | Prospective | China | 12 | Most SpA, Achilles enthesitis | 23.1 | 6.1 | 6:1 | Beta 7 mg vs Eta, USG+ | 12 weeks | −3.8*, −3.7* | ↓BASDAI, USG: ↓ thickness ↓blood flow | |
| 3 | Di Geso et al. [42] | Prospective | Italy | 21 | 12 SpA, 9 RA, tenosynovitis | 44.5 | − | 2:1 | MPA 20 mg, USG+ | 2 weeks | ↓ *** | ↓ tenderness ***, USG: ↓ GS score***, ↓PD*** | |
| 4 | Shrivastava et al. [43] | Prospective | India | 18 (27) | AS, achilles tendonitis | 26.1 | − | 8:1 | MPA 20 mg, USG+ | 6 weeks | −4*** | USG: ↓ thickness, ↓oedema, ↓ PD | |
| 5 | Gutierrez et al. [44] | Prospective | Italy | 114 | 56 SpA tenosynovitisd | − | − | − | MPA 20 mg, USG+ vs − | 1 − 4 weeks | USG: ↓PD | ||
| 6 | Girolimetto et al. [45] | Prospective | Italy | 46 (73) | PsA dactylitis | 46.2 | 4.4 | 29:17 | MPA 20 mg vs NSAID | 12 weeks | −5.5***, −1# | VAS-FI, LDI-b, FC better in GC | |
LDI-b: Leeds’ dactylitis index-basic; FC: finger circumference; MPA: methyl prednisolone acetate; THA: triamcinolone hexacetonide; TA: triamcinolone acetonide; PD: Power doppler signal; USG+: injected under ultrasound guidance; VAS-FI: Visual analogue scale- functional impairment.
Lack of relapse after IA injection,
Finger joints (MCP, PIP, DIP) 44%, knees 21%, wrist 9.4%,
Six Retrocalcaneal Bursitis, three plantar fascia, two posterior tibial tendon,
Finger extensor compartments, flexors, peroneus longus and brevis, tibialis posterior. *P < 0.05, ***P < 0.001.
Qualitative summary and meta-analysis of systemic GCs in axial SpA
We made a qualitative summary of efficacy in all the studies and found that BASDAI 50 was achieved in 20–40%, and ASAS 20 in 40–60% of patients after systemic GCs (Fig. 1). On pooling data from the two RCTs that compared high-dose oral prednisolone (± taper) with placebo in the short term (≤6 months) in axial SpA, BASDAI 50 was 4.2 times more likely (95% CI 1.5, 11.5) and ASAS 20 was twice as likely (95% CI 1.1, 3.64) in patients on prednisolone (Fig. 2).
Graphical representation of effectiveness of systemic steroids in all studies (randomized controlled trial or not) at various time points in terms of BASDAI 50, ASAS 20 and ASAS 40
Forest plot of efficacy of oral steroids in high dose or step-down high dose in the short term (≤24 weeks). BASDAI 50; ASAS 20; ASAS 40
Local steroid injections in SpA
After excluding studies on injections into the SI joint, there were 10 studies on local injections of GCs—4 on IA and 6 on entheseal injections in SpA patients (Table 3).
IA GC injection in SpA
Peripheral arthritis occurs in one-third of patients with AS and non-radiographic SpA [35]. We found four studies that reported on the efficacy of IA GC injection in SpA (Table 3). In 2016, Eder et al. reported IA MP acetate (MPA) injection to be effective in PsA, with more than half of patients remaining relapse-free at 3–6 months [36]. In the same year, Thilaak et al. compared IA MPA with etorixocib in PsA, and found the former was better for pain relief at 6 weeks [37]. Two RCTs, one by our group in 2016 and the second by Weitoft et al. in 2019, examined the effect of different preparations or doses of MPA or triamcinolone in knee arthritis [38, 39]. Both of these studies included both RA and SpA patients, but were not primarily planned (or powered) to examine differences in efficacy between RA and SpA. We extracted data on relapse in SpA and RA patients; on pooling the data, there was no significant difference in the RR of relapse [0.95 (95% CI: 0.6, 1.5)] between joints of SpA as compared with RA patients (Fig. 3a). Sustained response at 6 months after IA GCs in SpA patients ranged from 51.5% to 90% (Fig. 3b). There were no major adverse effects, except some local skin hypopigmentation reported (Supplementary Table S4, available at Rheumatology online).
Efficacy of IA glucocorticoids (GCs) in SpA: (a) comparison of risk of relapse after IA GCs in RA vs SpA at 24 weeks; (b) sustained response after IA injection of GCs in SpA at 24 weeks
Local steroid injection for enthesitis, tenosynovitis and dactylitis
Enthesitis, tenosynovitis and dactylitis are important clinical features of SpA. We found six studies, most of which used US guidance for injecting GCs into entheses (Table 3). Two reported a good response in a small number of patients (the majority with SpA) with enthesitis (and bursitis) [40, 41]. Di Geso et al. found significant improvement in pain, tenderness and power doppler (PD) signal at 2 weeks in tenosynovitis in SpA patients after local GC injection [42]. Shrivastava et al. injected inflamed achilles tendons in AS patients, and found significant reduction at 6 weeks in pain, entheseal thickness, and entheseal and bursal PD signal [43]. Gutierrez et al. compared US guided injections with blind injections, and found US guided injections were more efficacious and less painful [44]. Girolimetto et al. compared GC injection with NSAIDS and found the former to be better in psoriatic dactylitis [45].
Real-world reports of GC use in SpA
In the early days after the introduction of GCs, case series of good response to cortisone in both advanced AS as well as in Reiter’s syndrome were published [7, 8, 46–48]. In the more recent past, Yoshida et al. from Japan reported excellent response to i.v. MP pulses (1 g) for 3 days in a case of AS refractory to NSAIDs [49]. Another recent report of AS with membranous glomerulonephritis also reported good pain relief after 3 days of MP pulses (500 mg/day) [50].
Apart from AS, there are many reports on GCs in ReA. A report from Japan described a dramatic response of oral prednisolone in a case of ReA due to Yersinia enterocolitica [51]. Cases that were refractory to oral GCs but responded to pulse GCs have also been reported. Pasricha et al. reported monthly pulsed dexamethasone to be efficacious in a patient with severe polyarthritis and fever due to Reiter’s syndrome who was refractory to oral GCs [52]. A similar case was reported from the UK of a patient with Reiter’s syndrome (HIV positive) with severe incapacitating polyarthritis despite low-dose oral prednisolone, who improved with i.v. pulses of MP [53]. Apart from these, there have been small case series focusing on the use of systemic GCs in special situations in SpA. One reported GCs to be useful in patients with active SpA with fever [54, 55], but another reported only minimal effect in pregnant patients with flare of SpA [56].
Registry data suggests that GC use in axial SpA is not uncommon in some countries in Europe and Asia. Data from a German registry revealed that oral steroids were prescribed in 20% of AS patients in 2000–01, but their use had declined to 10% by 2012; concurrent with increased use of biologics [57]. A cross-sectional Spanish study from four centres reported the use of oral GCs in 3.2% of AS patients [58], while a study from Netherlands reported higher GC use in patients with undifferentiated SpA (5%) than in patients with AS (1.4%) [59]. In a US administrative claims database, among 78 704 rheumatic patients on GCs, the ratio of SpA to RA patients was 1:16 [60]. A large single-centre study from China on 555 patients with AS who received low-dose oral GCs for 1 year or more reported only minimal adverse effects. They did not find any significant differences in bone mineral density between treated and untreated AS patients, with a RR of 1.15 (95% CI 0.74–1.7), nor significant differences in dyslipidaemia, hyperglycaemia or obesity [61].
There is data to suggest that there is higher use of systemic GC in PsA [62]. A German collaborative study found low-dose GCs in 26.6% of PsA (compared with 49.7% in RA) [63]. Similarly, a study from USA reported prior use of GCs in 14% and current use in 8% of PsA patients [64]. GC use may be dependent on clinical features, with one study reporting that 35% of patients with PsA had received GCs, finding higher use for dactylitis than axial symptoms [65].
Recommendations of contemporary guidelines on GCs in SpA
Almost all the guidelines on axial SpA in the last 10 years have given a strong recommendation against the use of systemic GCs for the long term (Supplementary Table S5, available at Rheumatology online) [4, 5, 66–72]. This was based on a relative lack of evidence of efficacy by RCTs, and based on the known risk of bone loss and other adverse effects with GCs. Only some guidelines allowed for short-term GC use in flares of SpA, but most did not make this distinction [66, 68, 71]. Some guidelines recognized that GCs may have a role in special situations in SpA, like polyarticular flare, flare in pregnancy or concomitant active IBD [4, 72]. On the other hand, all guidelines uniformly recommended the use of injectable steroids into joints, SI or peripheral, and entheses, although this was a conditional recommendation due to a lack of robust efficacy studies.
Studies on GCs in animal models of SpA
We could only find sparse data on GC use in animal models of SpA. In mice models, inhibition of inflammation was shown by the use of GCs. Dexamethasone was found to reduce inflammation in the spontaneous ankylosis model in aging male DBA/1 mice [73]. Another study in a proteoglycan-induced spondylitis mouse model found early intervention with both etanercept and prednisolone improved inflammatory symptoms and ameliorated structural damage of the spine [74]. In HLA-B27 transgenic rats, a differential in development of arthritis depending on the background strain has also been postulated to occur due to reduced corticotrophin-releasing hormone synthesis and steroid production in some strains [75].
Discussion
This systematic review and meta-analysis found good evidence for the efficacy of systemic GCs in NSAID-refractory axial SpA. In the recent past, two placebo-controlled RCTs on NSAID-refractory patients have been published, in which high-dose (one with fast taper) oral prednisolone was given for up to 24 weeks. [30, 33] The pooled analysis of these showed significantly higher chances of achieving BASDAI 50, ASAS 20 and ASAS 40 responses with prednisolone. However, the pooled effect was modest, and the studies had only a limited number of patients treated over a short duration, with no data on use over a long duration, which may be associated with significant adverse effects.
Also, multiple prospective studies and one RCT have documented dramatic responses with pulses of MP in axial SpA unresponsive to NSAIDs, lasting from a few weeks to many months. In addition, there are many case reports of GCs, both pulsed and in medium to high doses, being effective in axial and peripheral flares in SpA, specifically in patients with ReA (with or without fever). Animal model data is sparse but suggests benefit. Registry data suggests that 2–10% of patients with AS receive GCs in some centres [57–59, 63].
The caveats were that evidence was available only for high or pulse doses, with little data on low-dose GCs, and most studies were of relatively short duration with small number of patients. The only long-term prospective study on medium to high doses was published way back in 1953, and it suggested the need for repetitive short courses of GCs to maintain benefit. With pulsed MP, multiple (usually three) pulses lead to longer duration of benefit compared with a single pulse, but there are no significant differences between lower- or higher-dose MP pulses (375–500 mg vs 1000 mg). In all these studies, adverse effects were reported to be mild (e.g. weight gain with high dose, and bitter taste with pulsed therapy) and there were no serious adverse effects. However, it must be borne in mind that longer duration use is likely to lead to the adverse effects that are well-known with GCs.
Despite the evidence of efficacy, would there be any place for systemic GCs in treating SpA? There are many reasons why physicians may not be enthusiastic about using them. The most important is the risk of exacerbating bone loss in already osteopenic axial SpA patients who are at high risk of vertebral fractures [76, 77]. The second is the availability of biologics like TNFi and IL17i, for which (unlike GCs) efficacy and safety over the long term have been proven in larger and longer RCTs. Finally, an inherent ‘framing or contrast bias’ on the effect of GCs in SpA as compared with RA (where even low doses work) may lead to a negative perception.
We feel that, in light of the evidence presented above, short-term use of high-dose or pulsed systemic GCs should be part of the therapeutic armamentarium for SpA to manage severe flares of axSpA that are disabling and non-responsive to NSAIDs, when access to biologic is limited or when they are contraindicated. When used for the short term, their benefits in alleviating severe pain and stiffness, and improving mobility, would probably outweigh the risk of bone loss and other adverse effects. However, low-dose GCs should be avoided as there is practically no data on their efficacy in SpA, and GCs should not be used for long periods in SpA to avoid the risk of fractures, infections, etc. Unfortunately, most of the guidelines, except for some like the one by Canadian Rheumatology Association/Spondyloarthritis research consortium of Canada, have lumped both short-term and long-term systemic GC use in SpA together, and given a strong negative recommendation. We believe that guidelines need to look at recent evidence, be broad-based, and allow the short-term use of GCs for patients without access to expensive therapies like biologics. As most of the patients in studies had radiographic axSpA, it is currently unclear whether GCs would also be efficacious in non-radiographic axial SpA.
Apart from systemic use, local injections into the peripheral joints or entheses were also reviewed in this SLR. IA injections into the peripheral joints were found to have sustained efficacy of 51.5–90% at 6 months, with pooled data from two RCTs showing no significant difference in relapses in SpA compared with RA joints. One study found better efficacy of IA injection compared with an NSAID in the short term in PsA. Entheseal injections were efficacious in relieving pain and improving US parameters like oedema and Power Doppler signal. All studies reported minimal adverse effects, mostly local hypopigmentation. We did not include SI joint injections of GCs in this SLR, because this has been recently reviewed elsewhere [16]. Local injections have been recommended by all guidelines, reflecting their efficacy and safety, but only conditionally, with a need for larger studies. It is intuitive that local GCs would be much less likely to cause systemic adverse effects like osteopenia compared with systemic GCs, and should be preferred wherever possible. Although, studies did not report complications, the risk of local infection and tendon rupture has been well documented.
To conclude, this SLR found good evidence for efficacy of systemic GCs at a high dose, and this may be a useful addition in the armamentarium of axial SpA for short-term control of disease activity in some settings. Local injections of joints and entheses were found to be useful with minimal adverse effects. Larger and longer RCTs are required of pulsed, high-dose (with taper) and low-dose systemic GCs in axial SpA (both radiographic and non-radiographic), as well as of local injections in peripheral joints and enthuses, to further define their role and adverse effects in this setting.
Acknowledgements
Funding: No funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out work described in this article.
Disclosure statement: The authors have declared no conflicts of interest.
Data availability statement
All data generated or analysed during this study are included in this published article and its supplementary information files.
Supplementary data
Supplementary data are available at Rheumatology online.
Comments