Giant cell arteritis in clinical practice: beyond GiACTA

F. Regola1,2, J. Mora1,2, M. Riva3, G. Fontana1,2, A. Gatti1,2,

I. Cavazzana1,2, F. Franceschini1,2, P. Toniati1
Rheumatology and Clinical Immunology Unit, ASST Spedali Civili di Brescia;
1
2
Department of Clinical and Experimental Sciences, University of Brescia;
3
Department of Medical and Surgical Specialties, Radiological Sciences and Public Health,
University of Brescia, Italy.

Abstract
Objective
Giant cell arteritis (GCA) therapy relies on high-dose glucocorticoids (GCs), which are associated with a high
incidence of side effects and GCA relapses, highlighting the need for steroid-sparing agents such as tocilizumab (TCZ)
and methotrexate (MTX). The aims of this study were to analyse GC side effects and to assess the steroid-sparing
efficacy of TCZ and MTX in a real-life cohort of patients with GCA through the application of the Glucocorticoid
Toxicity Index (GTI) version 2.0.

Methods
This retrospective cohort study included patients with a new diagnosis of GCA made in our Centre and classified
according to therapy, respectively GCs alone, GCs plus MTX, and GCs plus TCZ. GTI was calculated over a 5-year
follow-up period.

Results
We enrolled 150 patients, with a median follow-up of 21 (11-39) months. During this period, 88% experienced at
least one GC side effect. The cumulative GC dose was an independent predictor of the GTI-cumulative worsening
score (CWS), regardless of treatment group or follow-up time. As first-line therapy, TCZ reduced GC dose by 25%
compared to GCs alone, leading to fewer side effects (65% vs. 90%), less GC-induced damage, and no GCA relapses
(0% vs. 38%). TCZ also independently protected against relapses, regardless of GC dose or follow-up time.
In contrast, MTX did not show similar benefits in any aspect.

Conclusion
GCs represent a cornerstone in GCA therapy, but their cumulative dose correlates with induced damage, as
quantified by the GTI. TCZ demonstrated steroid-sparing effect and clinical efficacy in a large real-life cohort.

Key words
giant cell arteritis, tocilizumab, methotrexate, glucocorticoids

Clinical2022
Clinical and Experimental Rheumatology and Experimental Rheumatology 2026; 44: 723-732.
Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Francesca Regola, MD, PhD* Introduction Methods

Jacopo Mora, MD* Giant cell arteritis (GCA) is a large- Study design
Matteo Riva, MD vessel vasculitis (LVV), typically in- A retrospective cohort study of patients
Giulia Fontana, MD
volving the extracranial branches of with GCA diagnosed at our Centre be-
Alessia Gatti, MD
Ilaria Cavazzana, MD the carotid artery and/or the aorta and tween January 1st, 2000, and December
Franco Franceschini, MD its major branches. First-line therapy 31st, 2021 was performed. Patients were
Paola Toniati, MD includes high doses of glucocorticoids enrolled in the study if fulfilling the
*Contributed equally. (GCs), followed by gradual tapering. 2022 ACR/EULAR criteria for GCA
Please address correspondence to: However, more than 40% of patients (12). Patients with a follow-up of less
Francesca Regola experience relapses or vascular progres- than 6 months and those who received
U.O. di Reumatologia e Immunologia sion during tapering, and over 80% suf- treatments other than GCs, TCZ, or
Clinica, fer from GC-related adverse events (1). MTX were excluded. The study was
Spedali Civili of Brescia, For these reasons, both the European approved by the local institutional eth-
Piazzale Spedali Civili 1, League Against Rheumatism (EULAR) ics committee and conducted accord-
25123 Brescia, Italy.
E-mail: [Link]@[Link]
and the American College of Rheuma- ing to the Declaration of Helsinki.
tology (ACR) recommend the use of
Received on June 11, 2025; accepted in
revised form on September 29, 2025.
steroid-sparing agents (2, 3). Conven- Patients
tional synthetic disease-modifying anti- Patients were stratified according to
© Copyright Clinical and
Experimental Rheumatology 2026.
rheumatic drugs (csDMARDs), such as two distinct criteria: the GCA clinical
methotrexate (MTX), have shown only phenotype and the medical therapeutic
limited clinical benefit (4). In contrast, regimen administered. Concerning the
tocilizumab (TCZ), a humanised mono- GCA clinical phenotype, patients were
clonal antibody targeting the interleu- classified into three groups in accord-
kin (IL)-6 receptor, has demonstrated ance with the 2022 ACR/EULAR cri-
efficacy and steroid-sparing properties teria for GCA: cranial (C)-GCA, large-
in randomised controlled trials (RCTs) vessel (LV)-GCA and LV-C-GCA.
(5, 6). Regarding the classification of patients
Despite these findings, there remains a based on the treatment received, the fol-
significant gap in the literature regard- lowing groups were identified: patients
ing the real-life use of TCZ in GCA (7, treated with GC alone, patients treated
8). Specifically, there is a lack of data with GC+MTX, and patients treated
on its steroid-sparing effect and overall with GC+TCZ. Patients receiving MTX
impact on GC toxicity outside of RCT or TCZ were further subdivided into
settings. Moreover, the Glucocorticoid two subgroups: MTX-1 and MTX-2,
Toxicity Index (GTI) version 2.0 has and TCZ-1 and TCZ-2, depending on
recently been developed as a standard- whether the immunosuppressive agent
ised tool to quantify GC-induced toxic- was initiated within three months or af-
ORCID iD ity. However, it has been validated in ter three months from diagnosis. This
F. Regola: 0000-0001-6560-1112 conditions such as asthma and ANCA- time interval was selected because it
J. Mora: 0009-0001-6193-6326 associated vasculitis (9-11), but it has corresponds to the average period re-
M. Riva: 0009-0000-3444-2489 not yet been applied or validated in quired for the comprehensive clinical
G. Fontana: 0009-0002-7298-8962 GCA. assessment at diagnosis and the estab-
A. Gatti: 0009-0000-9672-780X
I. Cavazzana: 0000-0002-2757-7120
To address these gaps, we conducted lishment of a complete diagnostic-ther-
F. Franceschini: 0000-0003-3678-6124 a retrospective cohort study involving apeutic pathway (Table I).
P. Toniati: 0009-0007-1953-7093 patients with newly diagnosed GCA The response to treatment was assessed
Funding: no specific funding was received followed over a 5-year period at our by an expert rheumatologist at each
from any bodies in the public, commercial Centre. Treatment strategies included visit through evaluation of improve-
or not-for-profit sectors to carry out the GC monotherapy, GC combined with ment in clinical symptoms and signs of
work described in this article. MTX, or GC combined with TCZ. GCA, as well as reduction of inflam-
The intellectual property of the The aims of this study were to analyse matory markers, when appropriate.
Glucocorticoid Toxicity Index (GTI) is
the risk of GC side effects and to assess Relapses were defined according to the
owned by the Massachusetts General
Hospital (Copyright © 2016, 2018. the steroid-sparing efficacy of TCZ and 2018 EULAR recommendations for the
Massachusetts General Hospital. All rights MTX in a well-characterised, real-life management of LVVs as the recurrence
reserved). The use of the Glucocorticoid cohort of patients with GCA through of signs or symptoms of GCA and/or
Toxicity Index 2.0 was provided free of the application of the GTI 2.0, to quan- worsening findings on radiological
fees with an academic license. tify GC-related toxicity and the inci- imaging. In particular, relapses were
Competing interests: none declared. dence of relapses. further classified as major or minor. A

724 Clinical and Experimental Rheumatology 2026

 Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Table I. Patient classification used in the present study. (GraphPad Software, Inc, CA, USA).
Multiple comparisons were assessed
C-GCA Patients classified as GCA according to the 2022 ACR/EULAR criteria with exclusive
cranial involvement with post-hoc tests. When performed
multiple comparisons, level of signifi-
LV-GCA Patients classified as GCA according to the 2022 ACR/EULAR criteria with exclusive cance was corrected according to the
large-vessel involvement
Benjamini-Hochberg procedure. A two-
LV-C-GCA Patients classified as GCA according to the 2022 ACR/EULAR criteria with both tailed p-value less than alpha reference
cranial and large-vessel involvement
of 0.05 was considered statistically sig-
GC Patients treated with GC monotherapy nificant.
MTX-1 Patients started on MTX < 3 months after diagnosis due to high risk of GCA relapses
or GC side effects Results
Demographic and clinical features
MTX-2 Patients started on MTX ≥3 months after diagnosis due to GCA relapses or GC side
effects
One hundred and fifty patients were
enrolled in the study: 104 patients
TCZ-1 Patients started on TCZ < 3 months after diagnosis due to high risk of GCA relapses or were females (69%) and 46 were males
GC side effects
(31%) with a median age at diagnosis
TCZ-2 Patients started on TCZ ≥3 months after diagnosis due to GCA relapses or GC side of 73 (67–78) years. Patients were cat-
effects egorised into three groups according
to GCA clinical phenotype: 96 with C-
major relapse was defined as the reap- List consists of 11 domains that de- GCA (64%), 22 with LV-GCA (15%),
pearance of disease activity associated scribed GC-related toxicities not meas- and 32 with C-LV-GCA (21%) (Table
with risk of organ damage or progres- urable through the Composite Index II). At diagnosis, 130 patients (87%)
sion of vascular inflammation (e.g. se- and therefore not suitable for statistical had at least one comorbidity, with a
vere ischaemic complications such as analysis due to their descriptive nature. mean modified rheumatic disease co-
visual loss and jaw claudication, aortic The minimal clinically important dif- morbidity index (mRDCI) value of 1
aneurysm or dissection, or vascular ste- ference for GTI score was 10 (11). (0–2) and a mean Charlson comorbidity
nosis). A minor relapse was defined as Clinical and laboratory data of all pa- index (CCI) value of 4 (3–5) (Table III).
the reappearance of signs or symptoms tients were collected from clinical The median follow-up period of the
attributable to GCA that did not meet charts and GTI 2.0 was calculated 3 entire cohort was 21 (11–39) months.
the severity criteria of a major relapse months (t3), 6 months (t6), 12 months Fifty-two patients (35%) were treated
(e.g. new temporal headache and in- (t12), 24 months (t24), 36 months with GCs alone, 20 (14%) with MTX-
creased inflammatory markers without (t36), 48 months (t48), and 60 months 1, 26 (17%) with MTX-2, 26 (17%)
ischaemic complications) (2). (t60) after diagnosis, for a total of 5 with TCZ-1, and 26 (17%) with TCZ-2.
The GC-induced toxicity was assessed years of follow-up. The median follow-up periods of each
using the GTI 2.0, an algorithm de- category were 12 (6–24) months, 18
veloped by a multispecialty medical Statistical analysis (6–36) months, 18 (6–36) months, 12
team, which allowed to quantify the Data were presented as median (1st-3rd (6–24 months), and 18 (6-30) months,
GC-induced toxicity. GTI 2.0 included interquartile) or percentage composi- respectively. Among patients who
two sections: the Composite Index and tion for continuous and categorical var- started DMARDs during follow-up, the
the Specific List. The Composite Index iables, respectively. The Fisher’s exact median time to initiation was 12 (6–12)
comprised 9 domains representing the test was employed to compare qualita- months in MTX-2 and 6 (6-6) months
main GC side effects, including varia- tive variables, while the Mann-Whitney in TCZ-2. The median MTX doses
tions in body mass index (BMI), glu- U-test was used to compare continuous were 10 (10–12.5) mg/week in MTX-
cose tolerance, blood pressure, lipid variables between groups. Variations 1 and 15 (10–15) mg/week in MTX-2.
panel, GC-induced myopathy, bone from baseline to different timepoints In TCZ groups, 20 patients in TCZ-1
mineral density (BMD), infections, were assessed using the Wilcoxon’s and 10 patients in TCZ-2 received sub-
GC-induced cutaneous toxicity, and signed rank test for paired samples. cutaneous TCZ (162 mg/week), while
GC-induced neuropsychiatric mani- Correlations were assessed with Pear- 6 patients in TCZ-1 and 16 in TCZ-2
festations. The Composite Index was son, Kendall, or Spearman coefficient received intravenous TCZ [median
further subdivided into two scores: the (according to variable type) to evaluate of 560 (540–620) mg/month and 520
Cumulative Worsening Score (CWS), associations between variables. Sta- (470–560) mg/month, respectively].
which reflected the cumulative GC-re- tistical analysis was performed either
lated damage over time, and the Aggre- using the R environment with a mixed GC cumulative dose and side effects
gate Improvement Score (AIS), which linear model or with a generalised lin- Figure 1 displays the temporal trend
captured potential improvements be- ear model with the ‘lme4’ package (R of median GC cumulative doses in the
tween two timepoints, since some side Core Team, 2024) (13), or using the different groups, divided according to
effects may be transient. The Specific GraphPad statistical software package medical therapy. In comparing TCZ-1

Clinical and Experimental Rheumatology 2026 725

Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Table II. Signs and symptoms of patients at diagnosis. mon (20%), followed by pneumonias
(11%), acute bronchitis (7%), and her-
Signs and symptoms Cohort C-GCA LV-GCA C-LV-GCA p
(n=150) (n=96) (n=22) (n=32) pes zoster (6%).
To identify factors potentially influenc-
Temporal headache 117 (78%) 88 (92%) 0 (0%) 29 (91%) < 0.0001* ing the increased risk of GC-related
Scalp tenderness 64 (43%) 46 (48%) 0 (0%) 18 (56%) < 0.0001* side effects, an analysis of the overall
Jaw claudication 64 (43%) 51 (53%) 0 (0%) 13 (41%) < 0.0001*
Visual impairment 55 (37%) 48 (50%) 0 (0%) 7 (22%) < 0.0001* cohort was conducted using the GTI
Fatigue 110 (73%) 65 (68%) 16 (73%) 29 (91%) 0.0398* 2.0. First, to assess the potential role
Rheumatic polymyalgia 70 (47%) 47 (49%) 5 (23%) 18 (56%) 0.0398* of comorbidities at diagnosis, a corre-
Abdominal claudication 5 (3%) 0 (0%) 5 (23%) 0 (0%) < 0.0001*
lation analysis was performed between
Thoracic pain 5 (3%) 0 (0%) 3 (14%) 2 (6%) 0.0033*
Limb claudication 4 (3%) 0 (0%) 1 (4%) 3 (9%) 0.0144* CWS scores in the follow-up period
Weight loss 70 (47%) 37 (38%) 14 (64%) 19 (59%) 0.0278* and comorbidity indices at diagnosis.
Stroke or transient ischaemic 4 (3%) 2 (2%) 0 (0%) 2 (6%) 0.3148 While no statistically significant corre-
attack
Fever 63 (42%) 35 (36%) 11 (50%) 17 (53%) 0.1814
lation was found with baseline mRDCI,
Arthritis 9 (6%) 5 (5%) 2 (9%) 2 (6%) 0.7855 a high baseline CCI (>3) was associated
with significantly higher CWS scores
Data are expressed as n (%) values. during the follow-up period compared
*p≤0.05.
to low baseline CCI (≤3) (Fig. 2). Sec-
ond, to analyse the potential correla-
Table III. Comorbidities of patients at diagnosis. Data are expressed as n (%).
tions between the GC cumulative dose
Comorbidities Cohort C-GCA LV-GCA C-LV-GCA p and the development of GC-induced
(n=150) (n=96) (n=22) (n=32) side effects, a correlation analysis was
performed between the GC cumulative
Arterial hypertension 74 (49%) 56 (58%) 5 (23%) 13 (41%) 0.0058*
Dyslipidaemia 37 (25%) 21 (22%) 5 (23%) 11 (34%) 0.3552 dose and the GTI 2.0 across different
Malignant neoplasms 19 (13%) 14 (15%) 2 (9%) 3 (9%) 0.6419 timepoints, finding a positive correla-
Cardiovascular diseases 17 (11%) 13 (13%) 1 (4%) 3 (9%) 0.4501 tion between the GC cumulative dose
Cataract 17 (11%) 10 (10%) 1 (4%) 6 (19%) 0.2416
and the GTI 2.0 CWS when accounted
Diabetes mellitus 16 (11%) 9 (9%) 3 (14%) 4 (12%) 0.7848
Obesity 11 (7%) 8 (8%) 2 (9%) 1 (3%) 0.5841 for repeated measures (p<0.0001) (Fig.
Chronic lung diseases 11 (7%) 7 (7%) 1 (4%) 3 (9%) 0.7993 3) (14). The analysis was extended us-
Osteoporosis 10 (7%) 6 (6%) 0 4 (12%) 0.1875 ing a mixed linear model, which high-
Major depression 8 (5%) 4 (4%) 1 (4%) 3 (9%) 0.5165
Peptic ulcer 5 (3%) 3 (3%) 1 (4%) 1 (3%) 0.9429
lighted the GC cumulative dose as a
Chronic renal failure 5 (3%) 4 (4%) 1 (4%) 0 0.4939 predictor of the GTI-CWS indepen-
Insufficiency fractures 1 (1%) 1 (1%) 0 0 0.7534 dently of the treatment group and fol-
mRDCI§ at diagnosis 1 (0-2) 1 (1-2) 1 (0-1) 1 (0-1) 0.0351* low-up time (an increase of 1 mg/day of
CCI† at diagnosis 4 (3-5) 4 (4-5) 3 (3-5) 4 (3-4) 0.0043*
GC causes an increase of 0.0081 points
§
Modified Rheumatic Disease Comorbidity index; †Charlson Comorbidity index. in the GTI-CWS) (b: 0.0081; Std. Er-
*p≤0.05. ror: 0.0003; p<0.0001).

patients with both GC and MTX-1 pa- not reach statistical significance. On the TCZ and MTX effects on
tients, a notable trend emerged, indi- other hand, comparisons between GC GC side effects and GCA relapses
cating a lower GC cumulative dose for vs. TCZ-2 and GC vs. MTX-2 did not The analysis of GC side effect incidence
TCZ-1 patients from the timepoint t24 show a statistically significant differ- showed significant variation among
onwards. Specifically, TCZ-1 patients ence (Table IV). treatment groups (Table V). A higher
accumulated a mean GC dose of 8747 Throughout the follow-up period, GC proportion of patients experienced at
(6961–9780) mg at t24 and 9060 (8326– side effects were observed in 132 pa- least one side effect in the GC (90%),
11367) mg at t36, reflecting reductions tients (88%). A total of 451 GC-related MTX-1 (95%), MTX-2 (100%), and
of 20% (p=0.0327) and 34% (p=0.0222) side effects were documented with TCZ-2 (88%) groups compared to the
respectively, compared to GC patients. an average occurrence of about 3.4 TCZ-1 group (65%) (p=0.0017). Ad-
Similarly, compared to MTX-1 patients, GC-related side effects per patient. ditionally, patients in these groups re-
TCZ-1 patients showed reduction of Notably, 369 (82%) side effects oc- ported more GC side effects per patient,
23% (p=0.0323) and 28% (p=0.0208) curred within the initial 2 years fol- with ≥3 side effects in 44% of GC, 70%
at the same timepoints. By t60, TCZ-1 lowing diagnosis and the initiation of of MTX-1, 46% of MTX-2, and 50% of
patients accumulated a mean GC dose GC treatment. The most prevalent GC TCZ-2 patients compared to 19% in the
of 12320 (12193–13259) mg, indicat- side effects included arterial hyperten- TCZ-1 group (p=0.0151). In patients
ing reductions of 25% compared to GC sion (42%) and infections (43%) (Ta- who started DMARDs at least 3 months
patients and 14% compared to MTX-1 ble V). Among infections, urinary tract after diagnosis, only 30% of TCZ-2 pa-
patients; however, these differences did infections (UTIs) were the most com- tients developed GC side effects after

726 Clinical and Experimental Rheumatology 2026

 Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Fig. 1. GC cumulative dose at different timepoints categorised by the type of medical therapy. Data are expressed as median values.

Table IV. GC cumulative dose at different timepoints categorised by the type of medical therapy. Data are expressed as median (1st-3rd interquartile)
values.

GC (n=52) MTX-1 (n=20) MTX-2 (n=26) TCZ-1 (n=26) TCZ-2 (n=26) p p p p p p

GC vs. GC vs. GC vs. GC vs. MTX-1 MTX-2
MTX-1 MTX-2 TCZ-1 TCZ-2 vs. TCZ-1 vs. TCZ-2

t3 3664 (2526-4222) 3875 (3296-4098) 3436 (2067-4284) 3784 (3089-4145) 3499 (1809-4106) 0.8751 0.3564 0.5457 0.3620 0.7988 0.8049

t6 5876 (4289-7086) 6369 (5063-6990) 5944 (3467-7138) 5671 (4918-6991) 5811 (3811-7419) 0.5546 0.6035 0.9366 0.7829 0.5133 0.8620

t12 7923 (6313-9928) 9075 (7281-9868) 8725 (5736-10173) 7416 (6499-9089) 7689 (6255-11063) 0.4546 0.8981 0.4371 0.9772 0.1696 0.9175

t18 9738 (7639-12316) 10594 (8484-11363) 10379 (7756-12663) 8319 (6724-10343) 9378 (7941-14531) 0.6574 0.3133 0.2135 0.7040 0.1089 0.7576

t24 10885 (8664-13472) 11386 (9150-12529) 12563 (8565-14556) 8747 (6961-9780) 10854 (9531-18910) 0.9272 0.3697 0.0327* 0.5881 0.0323* 0.9900

t36 13691 (10229-16345) 12505 (10689-14103) 14676 (10642-16908) 9060 (8326-11367) 13971 (10696-20353) 0.7818 0.3792 0.0222* 0.6210 0.0208* 0.9113

t48 15176 (11510-19424) 13504 (11859-15763) 16848 (12629-20370) 11545 (11023-12871) 16965 (11363-20590) 0.3866 0.5838 0.1571 0.6451 0.3463 0.9694

t60 16457 (13218-21001) 14253 (13218-20703) 18650 (13065-21761) 12320 (12193-13259) 19670 (12232-23801) 0.6276 0.6543 0.0903 0.6740 0.2091 0.7398

*p≤0.05.

starting TCZ compared to 70% who had two Composite Index scores, AIS and tients vs. MTX-1 patients, respective-
developed them before (p<0.0001). In CWS, in relation to the type of medical ly at timepoints t6, t12, t18, t24, t36,
contrast, 45% of MTX-2 patients devel- therapy. Regarding GTI 2.0 AIS, com- t48, and t60. In the same way, GTI 2.0
oped GC side effects after MTX intro- paring TCZ-1 patients to GC patients CWS between TCZ-1 patients and GC
duction versus 55% before (p=0.2225). and MTX-1 patients, a trend to statis- patients or MTX-1 patients presented a
Additionally, specific GC side effects, tical significance was showed, with trend to statistical significance, show-
particularly infections, neuropsychiat- lower scores in TCZ-1 group. In par- ing a difference of 43%, 51%, 38%,
ric symptoms, and dyslipidaemia, were ticular, there was a difference of 52%, 49%, 62%, 67%, and 76% for TCZ-1
significantly lower in the TCZ-1 group 90%, 61%, 69%, 78%, 86%, and 86% patients vs. GC patients, and a differ-
compared to the GC group. for TCZ-1 patients vs. GC patients, and ence of 64%, 65%, 56%, 63%, 61%,
Figure 4 displays the temporal trend a difference of 67%, 90%, 68%, 77%, 68%, and 67% for TCZ-1 patients vs.
of the GTI 2.0, calculated using the 79%, 86%, and 85% for TCZ-1 pa- MTX-1 patients, respectively at time-

Clinical and Experimental Rheumatology 2026 727

Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Table V. GC side effects categorised by the type of medical therapy. Data are expressed as n (%) or mean values.

GC side effects Cohort GC MTX-1 MTX-2 TCZ-1 TCZ-2 p

(n: 150) (n: 52) (n: 20) (n: 26) (n: 26) (n: 26)

Infections, including grade 4 and 64 (43%) 26 (50%) 9 (45%) 10 (38%) 5 (19%) 14 (54%) 0.0736*
grade 5 infections
Arterial hypertension 63 (42%) 23 (44%) 12 (60%) 11 (42%) 7 (27%) 10 (38%) 0.2553
Neuropsychiatric manifestations, 53 (35%) 23 (44%) 12 (60%) 9 (35%) 3 (11%) 6 (23%) 0.0041*
including GC-induced psychosis
Diabetes mellitus and its compli- 42 (28%) 15 (29%) 8 (40%) 6 (23%) 7 (27%) 6 (23%) 0.7197
cations (diabetic retinopathy,
nephropathy, and neuropathy)
Cataract 38 (25%) 15 (29%) 3 (15%) 10 (38%) 5 (19%) 5 (19%) 0.3019
Dyslipidaemia 37 (25%) 13 (25%) 8 (40%) 5 (19%) 2 (8%) 9 (35%) 0.0791*
Dermatologic manifestations, 29 (19%) 9 (17%) 7 (35%) 6 (23%) 3 (11%) 4 (15%) 0.3094
including grade 4 dermatologic
manifestations
Cardiovascular diseases other than 26 (17%) 12 (23%) 2 (10%) 6 (23%) 2 (8%) 4 (15%) 0.3667
arterial hypertension
Weight gain 25 (17%) 11 (21%) 5 (25%) 4 (15%) 3 (11%) 2 (8%) 0.4395
Osteoporosis 25 (17%) 6 (11%) 3 (15%) 8 (31%) 3 (11%) 5 (19%) 0.2520
GC-induced myopathy 19 (13%) 6 (11%) 4 (20%) 4 (15%) 1 (4%) 4 (15%) 0.5236
Insufficiency fractures 16 (11%) 5 (10%) 0 4 (15%) 1 (4%) 6 (23%) 0.0741
Cushingoid features 6 (4%) 1 (2%) 1 (5%) 1 (4%) 1 (4%) 2 (8%) 0.8155
Glaucoma 4 (3%) 3 (6%) 0 1 (4%) 0 0 0.4006
Avascular necrosis 2 (1%) 0 0 1 (4%) 0 1 (4%) 0.4309
Tendon rupture 1 (< 1%) 1 (2%) 0 0 0 0 0.7546
Gastrointestinal symptoms 1 (< 1%) 0 0 0 0 1 (4%) 0.3083
Total GC side effects 451 169 74 86 43 79 N/A
Mean of GC side effects per patient 3.4 3.6 3.9 3.3 2.5 3.4 N/A
Number of patients with at least 1 132 (88%) 47 (90%) 19 (95%) 26 (100%) 17 (65%) 23 (88%) 0.0017*
GC side effect
Number of patients with 3 or more 67 (45%) 23 (44%) 14 (70%) 12 (46%) 5 (19%) 13 (50%) 0.0151*
GC side effects

*p≤0.05.

points t6, t12, t18, t24, t36, t48, and (p=0.0002 and p<0.0001, respective- groups, a summary table reporting the
t60. ly). Furthermore, only 4 TCZ-2 pa- incidence of relapses, the incidence of
Within the entire cohort, 63 patients tients (16%) developed a minor relapse GC-related adverse events, the median
(42%) experienced a GCA relapse ac- after initiation of TCZ compared to GC cumulative dose at t60, the median
counting for a total of 94 relapses. 12 MTX-2 patients (43%) after initia- GTI scores (both AIS and CWS), and
Dividing the cohort based on therapy tion of MTX, of which 7 experienced the corresponding percentage variation
type, 20 GC patients developed 26 re- major relapses (p=0.0410). In logistic for the TCZ- and MTX-groups com-
lapses (38%), of which 8 major and 18 regression for longitudinal data, TCZ pared to GC group is reported in Ta-
minor, 9 MTX-1 patients 15 relapses was shown to be a protective factor ble VI. The table highlights that TCZ-
(45%), of which 5 major and 10 minor, against relapses, independently of GC 1 patients had the lowest cumulative
19 MTX-2 patients 28 relapses (73%), cumulative dose and follow-up time (b: GC exposure (percentage reduction vs
of which 13 major and 15 minor, TCZ- -21.81; Std. Error: 7.97; p=0.0062), a GC group of 25%) and the most pro-
1 patients none, and 15 TCZ-2 patients finding not observed for MTX. nounced reductions in both GTI-AIS
25 relapses (58%), of which 14 major (86%) and GTI-CWS (76%) scores,
and 11 minor. Specifically, significant Summary of outcomes by with no patients experiencing disease
differences were observed between GC treatment group relapses. In contrast, MTX-2 patients
and TCZ-1 (p<0.0001 and p<0.0001, To provide a concise overview of showed the highest relapse rate (73%)
respectively) and MTX-1 vs. TCZ-1 the main outcomes across treatment and only moderate reductions in GTI

728 Clinical and Experimental Rheumatology 2026

 Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Fig. 2. Correlation analysis

between the Glucocorticoid
Toxicity Index (GTI) 2.0
Cumulative Worsening Score
(CWS) and the Charlson
Comorbidity index (CCI) at
diagnosis. A baseline CCI
value ≤3 was categorised as
low (low CCI), while a base-
line CCI value >3 was cat-
egorised as high (high CCI).

Fig. 3. Correlation Chronic GC therapy is recognised as a

analysis between the risk factor for numerous diseases and
GC cumulative dose
(DCS) and the Glu- for most of these side effects, a dose-
cocorticoid Toxicity dependent effect of the cumulative GC
Index (GTI) 2.0 Cu- dose has been demonstrated (17). Dur-
mulative Worsening ing follow-up, approximately 90% of
Score (CWS) when
accounted for repeat- patients developed at least one GC-re-
ed measures. lated side effect (average of 3.4 events
per patient), with arterial hypertension
and infections being the most frequent.
The increased risk of severe infections
(grade ≥3 according to GTI 2.0) ob-
served in this study was an expected
finding due to the immunosuppressive
effect of steroid therapy, and it aligned
with previously published data (18).
scores. The incidence of GC-related agents, especially MTX and TCZ. The Neuropsychiatric side effects, including
side effects was lowest in the TCZ-1 aims of this study were to analyse the anxiety, insomnia, and depression, were
group (65%), while the other groups risk of GC side effects and to assess also commonly reported, although they
experienced higher rates, particularly the steroid-sparing efficacy of TCZ and may partly stem from the emotional
MTX-2 (100%). These findings under- MTX in a well-characterised, real-life burden of a GCA diagnosis, especially
score the superior steroid-sparing ef- cohort of patients with GCA through when visual complications are present
fect and relapse prevention associated the application of the GTI 2.0 (15) to at onset, rather than being solely attrib-
with early TCZ therapy compared to quantify GC-related toxicity and the utable to GC use.
MTX therapy and GC monotherapy. incidence of relapses. Furthermore, our data confirm that
In this study, the cumulative GC dose patients with multiple baseline co-
Discussion at 1 and 5 years (7,923 mg and 16,457 morbidities and higher cumulative GC
The treatment of GCA relies on the use mg, respectively) was higher compared doses were at an increased risk of de-
of high-dose GCs, but around 40% of to RCTs, yet aligned with the limited veloping new comorbidities or worsen-
patients still experience relapses, and available data in the literature (6) (16). ing pre-existing ones. Indeed, using a
chronic GC use leads to side effects, This difference is likely due to the per- mixed linear model, we identified the
impacting prognosis. Consequently, sonalised GC tapering protocol in the GC cumulative dose as an independent
recent focus has shifted toward com- outpatient setting, as opposed to the predictor of the GTI-CWS, regardless
bining GCs with immunosuppressive standardised approaches in RCTs. of the treatment group or the follow-up

Clinical and Experimental Rheumatology 2026 729

Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Fig. 4. Glucocorticoid Toxicity Index (GTI) 2.0 Aggregate Improvement Score (AIS) and GTI 2.0 Cumulative Worsening Score (CWS) at different time-
points categorised by the type of medical therapy. Data are expressed as median values.

time (an increase of 1 mg/day of GC duce GC requirements and toxicity in scores, with a reduction in both AIS, the
caused an increase of 0.0081 points in GCA (19). However, there are no pub- index of the active damage at each time-
the GTI-CWS) (b: 0.0081; Std. Error: lished studies on its steroid-sparing ef- point, and CWS, the index of the cumu-
0.0003; p<0.0001). fect in cohorts other than those from the lative patient damage, when compared
TCZ demonstrated its efficacy in two two RCTs. Our analysis revealed that to GC therapy alone, with a difference
RCTs (5, 6), ensuring disease remission TCZ, when used as first-line therapy of 86% and 76% at t60, respectively.
after 52 weeks of treatment and result- and started within 3 months after diag- These results were further supported
ing in a reduced cumulative GC dose. nosis (TCZ-1), significantly reduced the by the lower incidence of GC-related
These results were anticipated by a first cumulative GC dose by 25% compared side effects in the TCZ groups. In the
open-label multicentre study, which to GC therapy alone at t60. TCZ treat- entire cohort, approximately 40% of
suggested the potential of TCZ to re- ment also resulted in improved GTI 2.0 patients experienced at least one dis-

730 Clinical and Experimental Rheumatology 2026

 Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

Table VI. Main outcomes across treatment groups. Δ: percentage variation.

Outcomes\groups GC (n=52) MTX-1 (n=20) MTX-2 (n=26) TCZ-1 (n=26) TCZ-2 (n=26)

Patients with ≥1 GCA relapse 38% 45% 73% 0 16%

Patients with ≥1 GC side effect 90% 95% 100% 65% 88%
Median GC cumulative dose at t60 16457 mg 14253 mg 18650 mg 12320 mg 19670 mg
(Δ vs. GC group) (N/A) (–13%) (+13%) (–25%) (+19%)
Median GTI-AIS score at t60 139 137 103 20 127
(Δ vs. GC group) (N/A) (–1%) (–26%) (–86%) (–9%)
Median GTI-CWS score at t60 219 162 111 53 200
(Δ vs. GC group) (N/A) (–26%) (–49%) (–76%) (–9%)

ease relapse, but none occurred in the On the contrary, one of the key key component in treatment strategies
TCZ-1 group. Similarly, in the TCZ-2 strengths of this study is its large sam- for GCA to improve outcomes and limit
group, which included patients who ple size, along with the uniformity of GC-related toxicity. In contrast, MTX
started TCZ following a disease relapse data collection due to the single-centre did not show similar results in reducing
or GC-induced side effects, only 16% design and the extended follow-up pe- GC use or preventing relapses.
of patients experienced a relapse after riod. Notably, while the GTI 2.0 score
starting TCZ. Moreover, in logistic has been validated for conditions such Take home messages
regression for longitudinal data, TCZ as bronchial asthma and ANCA-associ- • Glucocorticoids are associated with a
was shown to be a protective factor ated vasculitis (9) (10) (11), analysing high incidence of side effects in giant
against relapses, independently of GC the variation of this parameter over a cell arteritis patients.
cumulative dose and follow-up time 1-year period, its application in GCA • Glucocorticoid cumulative dose
(p=0.0062). These findings confirmed has been limited. To our knowledge, correlates with induced damage, as
the efficacy of TCZ in both inducing this study is the first to utilise the GTI quantified by the Glucocorticoid
and maintaining disease remission, re- 2.0 in a cohort followed longitudinally Toxicity Index version 2.0.
gardless of whether it is used as first- for over 12 months. Additionally, this • Tocilizumab demonstrated steroid-
line or second-line treatment. study contributes valuable insights into sparing effect and clinical efficacy in
In contrast, the steroid-sparing effect of the steroid-sparing effects of TCZ and a large real-life cohort.
MTX was modest, with limited reduc- MTX in a real-world setting, thereby
tions in GC cumulative dose and side enhanced our understanding of long- References
effects, consistent with previous ret- term treatment strategies for GCA. 1. PROVEN A, GABRIEL SE, ORCES C,
O’FALLON WM, HUNDER GG: Glucocorti-
rospective studies and meta-analyses In conclusion, this study provides fur- coid therapy in giant cell arteritis: duration
(4). Moreover, MTX did not appear to ther evidence that chronic GC use in and adverse outcomes. Arthritis Care Res
be a protective factor against relapses. GCA patients is associated with a high 2003; 49(5): 703-8.
These findings are in line with two incidence of side effects, observed in [Link]
2. HELLMICH B, AGUEDA A, MONTI S et al.:
RCTs, which similarly failed to dem- approximately 90% of patients, which 2018 Update of the EULAR recommenda-
onstrate a significant benefit of adding correlate with the GC cumulative dose tions for the management of large vessel vas-
MTX to GC therapy, either in terms of and the presence of comorbidities at culitis. Ann Rheum Dis 2020; 79(1): 19-30.
https://
steroid-sparing effect, GC side effects, diagnosis. TCZ, when used as first-line [Link]/10.1136/annrheumdis-2019-215672
or GCA relapses (20, 21). therapy, significantly reduced the GC 3. MAZ M, CHUNG SA, ABRIL A, LANGFORD
One major limitation of this study is its cumulative dose by 25% and decreased CA, GORELIK M, GUYYATT G: American
retrospective design. The GTI 2.0 score, GC-related side effects, as quantified by College of Rheumatology/Vasculitis Founda-
tion Guideline for the management of giant
originally designed for prospective the GTI 2.0. Logistic regression further cell arteritis and Takayasu arteritis. Arthritis
RCTs, was applied to a retrospective co- revealed that GC cumulative dose is an Rheumatol 2021; 73(8): 1349-65.
hort, which makes the study susceptible independent predictor of the GTI-CWS. [Link]
4. GÉRARD AL, SIMON-TILLAUX N, YORDAN-
to missing or incomplete data, potential- Moreover, TCZ, whether used as first- OV Y et al.: Efficacy and safety of steroid-
ly affecting the uniformity and statisti- line or second-line therapy, prevented sparing treatments in giant cell arteritis ac-
cal power of the results. To mitigate this, disease relapses in a real-world cohort, cording to the glucocorticoids tapering regi-
we included only patients with com- proving to be a protective factor against men: a systematic review and meta-analysis.
Eur J Intern Med 2021; 88: 96-103.
plete clinical follow-up data for at least relapses independent of GC cumulative [Link]
6 months. Additionally, the absence of dose and follow-up time. These find- 5. VILLIGER PM, ADLER S, KUCHEN S et al.:
standardised treatment protocols in the ings highlight the efficacy of TCZ as a Tocilizumab for induction and maintenance
of remission in giant cell arteritis: a phase
real-world setting introduced variabil- steroid-sparing agent in achieving re- 2, randomised, double-blind, placebo-con-
ity in treatment regimens, which might mission and maintaining long-term dis- trolled trial. Lancet 2016; 387(10031): 1921-
have acted as confounding factors. ease control, emphasising its value as a 7. https://

Clinical and Experimental Rheumatology 2026 731

Real-life tocilizumab efficacy in giant cell arteritis / F. Regola et al.

[Link]/10.1016/S0140-6736(16)00560-2 tion of glucocorticoid-associated morbidity 17. WILSON JC, SARSOUR K, COLLINSON N,

6. STONE JH, TUCKWELL K, DIMONACO S et in severe asthma using the Glucocorticoid TUCKWELL K, MUSSELMAN D, KLEARMAN
al.: Trial of tocilizumab in giant cell arteritis. Toxicity Index. J Allergy Clin Immunol Pract M: Serious adverse effects associated with
N Engl J Med 2017; 377(4): 317-28. 2021; 9(1): 365-72. https:// glucocorticoid therapy in patients with giant
[Link] [Link]/10.1016/[Link].2016.11.006 cell arteritis (GCA): a nested case-control
7. QUINN KA, DASHORA H, NOVAKOVICH E, 12. PONTE C, GRAYSON PC, ROBSON JC et analysis. Semin Arthritis Rheum 2019; 46(6):
AHLMAN MA, GRAYSON PC: Use of 18F- al.: American College of Rheumatology/ 819-27. https://
fluorodeoxyglucose positron emission to- EULAR Classification Criteria for Giant Cell [Link]/10.1016/[Link].2016.11.006
mography to monitor tocilizumab effect on Arteritis. Arthritis Rheumatol 2022; 74(12): 18. SCHMIDT J, SMAIL A, ROCHE B, GAY P,
vascular inflammation in giant cell arteritis. 1881-89. [Link] SALLE V, PELLET H: Incidence of severe in-
Rheumatology (Oxford) 2021; 60(9): 4384- 13. R CORE TEAM R: A language and environment fections and infection-related mortality dur-
89. https:// for statistical computing [Internet]. Vienna, R ing the course of giant cell arteritis: a mul-
[Link]/10.1093/rheumatology/keaa894 Foundation for Statistical Computing, 2024. ticenter, prospective, double-cohort study.
8. REGOLA F, CERUDELLI E, BOSIO G et al.: Available from: [Link] Arthritis Rheumatol 2016; 68(6): 1477-82.
Long-term treatment with tocilizumab in 14. BAKDASH JZ, MARUSICH LR: Repeated [Link]
giant cell arteritis: efficacy and safety in a measures correlation. Front Psychol 2017; 19. LORICERA J, BLANCO R, HERNANDEZ JL et
monocentric cohort of patients. Rheumatol 7(8): 456. al.: Tocilizumab in giant cell arteritis: multi-
Adv Pract 2020; 4(2): 1-9. [Link] center open-label study of 22 patients. Semin
[Link] 15. STONE JH, MILOSLAVSKY EM, NADEN RP, Arthritis Rheum 2015; 44(6): 717-23. https://
9. MCDOWELL PJ, STONE JH, ZHANG Y et al.: BIJLSMA JWJ, BROGAN PA, BROWN ES: [Link]/10.1016/[Link].2015.05.006
Glucocorticoid toxicity reduction with me- Development of a Glucocorticoid Toxic- 20. SPIERA HJ, MITNICK M, KUPERSMITH M et
polizumab using the glucocorticoid toxicity ity Index (GTI) using multi-criteria decision al.: A prospective, double-blind, randomized,
index. Eur Respir J 2022; 59(1): 2100160. analysis. Ann Rheum Dis 2017; 76(3): 543- placebo-controlled trial of methotrexate in
[Link] 46. [Link] the treatment of giant cell arteritis (GCA).
10. JAYNE DRW, MERKEL PA, SCHALL TJ, BEK- 16. BEST JH, KONG AM, UNIZONY S, TRAN O, Clin Exp Rheumatol 2001; 19(5): 495-501.
KER P: Avacopan for the treatment of AN- MICHALSKA M: Risk of potential glucocor- 21. HOFFMAN GSC, HELLMAN MC, GUILLEVIN
CA-Associated vasculitis. N Engl J Med ticoid-related adverse events in patients with DB et al.: A multicenter, randomized, double-
2021; 384(7): 599-609. giant cell arteritis: results from a USA-based blind, placebo-controlled trial of adjuvant
[Link] electronic health records database. Rheuma- methotrexate treatment for giant cell arteri-
11. MCDOWELL PJ, STONE JH, ZHANG Y, HON- tol Ther 2019; 6(4): 599-610. tis. Arthritis Rheum 2002; 46(5): 1309-18.
EYFORD K, DUNN L, LOGAN RJ: Quantifica- [Link] [Link]

732 Clinical and Experimental Rheumatology 2026