Anifrolumab induces a fast improvement in ultrasound-detected synovitis in systemic lupus erythematosus: Results from a 6-months follow-up ultrasound study

Introduction

The advent of biologic drugs (bDMARDs) represents a substantial innovation in the management of Systemic Lupus Erythematosus (SLE) patients, thanks their ability to target disease-specific mechanisms. Data from randomized controlled trials (RCTs) and observational studies have clearly demonstrated the efficacy and safety of the two biologic drugs approved for SLE, belimumab (BLM) and anifrolumab (ANF). ¹ Accordingly, the updated EULAR recommendations for SLE managing suggest the early treatment with bDMARDs, even before conventional immunosuppressant drugs. ¹

ANF has been recently approved to treat SLE patients, after the publication of the results derived from the TULIP-1 and TULIP-2 RCTs.^2,3 The drug has clearly demonstrated its efficacy in treating extra-renal SLE, controlling disease activity and facilitating glucocorticoid dose sparing and discontinuation, with a significant impact in chronic damage progression. ⁴ To date, data from real-life scenario underline the excellent efficacy and rapidity of ANF in SLE-related skin involvement, but very little is known about the drug’s effectiveness on other SLE-related manifestations. ⁴ Among these, joint involvement is one of the most frequent features in SLE patients, potentially involving up to 90% of subjects, with a significant impact on quality of life. ⁵ It is characterized by clinical heterogeneity, ranging from mild to more severe phenotypes. In the latter cases, the presence of erosive and/or deforming arthritis can be observed. ⁵ The application of more sensitive imaging tool, such as ultrasound (US), has demonstrated the presence of erosive damage in about a quarter of SLE patients with clinically evident arthritis. ⁵ Although these data clearly suggest the need for targeted treatment to prevent the development of joint damage, there are no specific indications on which drugs to use.

Observational studies demonstrated the efficacy of BLM to treat SLE-related articular manifestations. The multicentre study published by Zen and colleagues in 2023 demonstrated that BLM led to clinical improvement in a significant proportion of patients with joint involvement in a real-life setting, as assessed by clinimetric indices. ⁶ This result had previously been highlighted in one of our studies, in which we confirmed the effectiveness of BLM using US. Indeed, we observed the significant reduction of US inflammatory score during a 12-months follow-up. ⁷

Moving on the efficacy of ANF in SLE-related joint involvement, post hoc analyses of TULIP-1 and TULIP-2 RCTs demonstrated significant improvements in the articular domain, in terms of swollen and tender joints count. ⁸ On the contrary, real-life data focused on the efficacy of ANF in SLE-related joint involvement are extremely limited. We previously described a case-series including three SLE patients treated with ANF for prevalent joint involvement. Alongside the improvement in clinimetric indices, we observed a rapid reduction of US inflammatory score already after 1 month of treatment, with further improvement at 3-months follow-up. ⁹ This result has been recently confirmed by analysing other four SLE patients described by Cassone and colleagues: all the assessed patients achieved complete remission of arthritis and lupus disease activity within 4 months, as assessed by using DAS28 and SLEDAI-2k. However, in these patients ultrasound evaluation did not represent an outcome since it was not performed regularly. ¹⁰

Moving from these premises, here we described the results of a prospective longitudinal analysis aimed at evaluating the efficacy of ANF in SLE-related joint involvement by using ultrasound assessment and clinimetric indices.

Methods

We conducted a prospective longitudinal study including SLE patients who were receiving ANF for musculoskeletal manifestations. All patients, followed-up at Lupus Clinic of Sapienza University of Rome, were diagnosed according to 2019 ACR/EULAR classification criteria. ¹¹ The drug was prescribed according to the current clinical practice and administered intravenously at a dose of 300 mg every 4 weeks.

The study design included the evaluation at four time-points, namely at baseline (T0) and after 1 (T1), 3 (T3) and 6 (T6) months of ANF treatment.

The study was approved by the local ethics committee (Sapienza University of Rome, Policlinico Umberto I, Rome, Italy). All patients gave their informed consent.

Clinical and laboratory data were reported in a standardized electronic form, including demographics, past medical history with date of diagnosis, co-morbidities, and previous and concomitant treatments. A complete physical examination was performed on all the SLE patients.

Overall disease activity was assessed by using SLEDAI-2k and SLE-DAS.^12,13 Furthermore, joint assessment was conducted, including tender/swollen joint count (0–28) and patients global health assessment, to calculate DAS28crp.

Finally, US assessment was performed by using a MyLab Eight Exp (Esaote, Genoa, Italy) machine equipped with a multifrequency linear array transducer (6–18 MHz) with the application of Power Doppler (PD, PFR 750 Hz, Doppler frequency 11.1 MHz, gain 50% and low filters). US evaluation was performed at level of bilateral metacarpophalangeal and proximal interphalangeal joints of the hands. According to the OMERACT Definitions, ¹⁴ we evaluated the presence of US inflammatory features (synovial effusion, synovial hypertrophy and Power Doppler - pD), each measured with a semiquantitative score (0–3), resulting in a total US inflammatory score (0–180). We also evaluated the total pD score (0–60). Moreover, we assessed the presence of erosive damage according to OMERACT Definitions. ¹⁴ The US examination was performed by a trained rheumatologist (FCe) with 20 years of experience in the field of musculoskeletal US, blinded to the clinical data of evaluated patients. Intra-reader reliability was assessed, yielding a Cohen’s kappa of 0.8.

Statistical analyses were performed using Graph Pad Prism Version 5 (La Jolla, CA). Continuous data were presented as means with standard deviations (SDs) or medians with interquartile range (IQR), depending on the distribution of the data (tested with the Kolmogorov-Smirnov test). Categorical data were presented as proportions. Wilcoxon test was performed. Univariate comparisons between nominal variables were calculated using the chi-square test or Fisher’s exact test where appropriate.

The Friedman test for multiple time-point comparisons was applied.

Two-tailed values were reported. A value <0.05 was considered statistically significant.

Results

Our cohort comprised 13 female SLE patients [median age 51 years (IQR 17.5), median disease duration 16 years (IQR 16.5)].

Table 1 summarizes the main demographics, clinical and laboratory features of enrolled patients; furthermore, data about disease treatments are reported. All data refereed to overall disease.

OPEN IN VIEWER

Table 1.

Disease related manifestations, laboratory features and treatments on enrolled patients (N = 13). All data refereed to overall disease.

Clinical manifestations, N (%)
Joint involvement	13 (100)
Skin involvement	8 (61.5)
Haematological manifestations	6 (46.1)
Renal involvement	3 (23.1)
Constitutional manifestations	3 (23.1)
NPSLE	1 (7.7)
Lung involvement	1 (7.7)
Laboratory abnormalities, N (%)
anti-dsDNA	11 (84.6)
anti-SSA	5 (38.5)
anti-SSB	2 (15.3)
anti-Sm	3 (23.1)
anti-RNP	4 (30.8)
Lupus anticogulant	3 (23.1)
anti-Cardiolipin (IgG/IgM)	3 (23.1)
anti-β2 glicoprotein I	1 (7.7)
Ratest	4 (30.8)
anti-CCP	6 (46.1)
Low C3/C4 serum levels	10 (76.9)
Treatments, N (%)
Glucocorticoids	13 (100)
Hydroxychloroquine	13 (100)
Methotrexate	7 (53.8)
Mycophenolate mofetil	5 (38.5)
Cyclosporin A	1 (7.7)
Cyclophosphamide	1 (7.7)
Belimumab	10 (76.9)

Notably, 6 and 4 patients showed positivity for anti-CCP and Ratest respectively.

Eight patients (61.5%) showed erosive damage in at least one evaluated joint. The presence of erosions was significantly associated with positivity for anti-CCP antibodies (p = 0.03).

Furthermore, the majority of enrolled patients have been previously treated with biological drugs (10 patients treated with belimumab before starting ANF, stopped for lack of efficacy in 2 patients, loss of efficacy in 7 patients and adverse event in one). At the time of study enrolling (at ANF baseline) 12 patients (92.3%) were treated by hydroxychloroquine and 10 (76.9%) by immunosuppressant drugs (5 patients by micophenolate, 4 by methotrexate and 1 by cyclosporin A). Moving on glucocorticoids, we found a significant decrease in median daily dosage during follow-up (baseline: 10 mg, IQR 6.25; 6 months 2.5 mg, IQR 2.5, p = 0.0004).

Figure 1 shows the changes in the US inflammatory and pD scores during the follow-up. As graphically represented in Figure 1 (A/B), we found a significant reduction of US scores already after 1 month of treatment, with a further significant reduction after 3 and 6 months of follow-up. The application of Friedman test for multiple time-point comparisons confirmed the significant reduction in US inflammatory and pD scores (p < 0.0001 for both).OPEN IN VIEWER

Figure 1(C) represents the modifications in the proportion of pD-positive patients during the observation period: interestingly, after 6 months of ANF treatment, 90% of patients showed negative pD signal in all the joints.

At the same time, we evaluated changes in disease activity indices: all the results were summarized in Table 2. We observed a rapid reduction, already after 1 month, in global indices (SLEDAI-2k and SLE-DAS) and in DAS28crp values, with further significant improvement in all the subsequent time-points. According to DAS28crp values, after 1 month of treatment, 25% of patients reached a condition of low-disease activity; this percentage increased to 70% and 75% at 3 and 6 months of observation, respectively.

OPEN IN VIEWER

Table 2.

Changes in disease activity indices during observational period (SLEDAI-2k, SLEDAS, DAS28crp).

​	T0	T1	T3	T6	p
SLEDAI-2k median (IQR)	7 (2)	4 (4)	2 (2)	1 (3.5)	T0 vs T1 0.002
					T0 vs T3 0.0004
					T0 vs T6 0.006
SLE-DAS median (IQR)	7.9 (5.0)	4.9 (5.2)	1.3 (3.3)	1.4 (4.2)	T0 vs T1 0.01
					T0 vs T3 0.0007
					T0 vs T6 0.008
DAS28crp median (IQR)	4.9 (1.7)	4.5 (1.8)	3.2 (0.9)	3.1 (1.1)	T0 vs T1 0.002
					T0 vs T3 <0.0001
					T0 vs T6 <0.001

Discussion

Our analysis, made in a real-life scenario, underlines the fast efficacy of ANF on joint synovitis, as shown by the reduction in US scores in the early stages of treatment. Furthermore, a further improvement was recorded at subsequent time-points.

Indeed, US assessment plays an important role in the management of chronic arthropathies, including SLE-related arthritis, providing information about the joint inflammatory status. This imaging technique could help in monitoring disease activity and thus in the evaluation of treatment response. The inclusion of US in a comprehensive assessment of inflammatory arthropathies has been frequently proposed, in view of the higher sensitivity of this technique compared to the clinical evaluation. ¹⁵ In our opinion, the changes in pD US score observed in our patients during ANF treatment represent a particularly noteworthy result. In fact, US pD has been clearly described as a significant risk factor for erosive damage in inflammatory arthritis, independent of clinical disease activity, serological biomarkers, or administered treatments, as supported by multiple high-quality studies. ¹⁵ Therefore, the rapid effect of ANF treatment on US pD signal could suggest the ability of this drug to prevent the development of erosive damage. It should be emphasised that the small size of our sample does not allow us to draw definitive conclusions, but only to put forward a hypothesis that will certainly need to be confirmed by larger studies and a longer follow-up period. Nevertheless, this suggestion appears very interesting in the light of evidence that erosive damage in patients with SLE-related arthritis is not uncommon, occurring in around 25% of cases. ⁵

Our findings are in line with data from the literature. In particular, the study conducted by Nzeusseu Toukap and colleagues in 2007 demonstrated a distinct molecular signature in synovial tissue of SLE patients with arthritis. Indeed, compared to rheumatoid arthritis and osteoarthritis, SLE patients showed the up-regulation of interferon-inducible genes, suggesting the involvement of this specific signature in joint involvement. ¹⁶

Another aspect that highlights the significance of our results is the phenotype of the patients enrolled in our analysis. These are patients with long-standing disease, most of whom had previously been treated with BLM, which was discontinued due to primary or secondary failure. Despite this phenotype, ANF clearly demonstrated its efficacy and rapid action. It will certainly be interesting to evaluate the action of the drug in patients with earlier disease, in terms of duration but also in terms of therapeutic line.

Certainly, out study shows several limitations, including the small size, the monocentric design and the lack of a control group. Furthermore, due to the observational nature if the analysis we cannot evaluate the confounding effect of concomitant therapies.

Indeed, for these reasons our study does not allow for conclusive results, which cannot therefore be generalised. Obviously, it will be necessary to increase the number of patients enrolled and the follow-up period.

However, although preliminary, our results could represent an interesting suggestion for ANF treatment in SLE patients.

In conclusion, the inclusion of imaging tools, such as US, to the clinimetric evaluation provides additional information for more comprehensive and accurate management of SLE patients with joint involvement.