Introduction
GCA is a large vessel vasculitis (LVV) affecting individuals older than 50 years, more frequently women [1]. The worldwide annual incidence of GCA ranges from 2.2 to 30 cases per 100 000 persons over 50 years of age, with predominance in Scandinavia and North America compared with countries in Southern locations [2], such as Spain, where the annual incidence has been rated at 2.2–10.1 cases per 100 000 persons aged 50 years and older [2, 3].
The main clinical manifestations of GCA include new onset headache accompanied by other craniofacial symptoms, visual disturbances, PMR, fever or low-grade fever and constitutional symptoms [4]. Acute phase reactants, such as ESR and CRP, are typically increased [4]. The systemic inflammatory response score (SIR) has been identified as a predictor of ischaemic complications and glucocorticoid (GC) requirements [5–8].
Patients with a weak SIR at disease onset exhibit higher risk of presenting with cranial ischaemic manifestations [5, 6] and typically require shorter therapy with lower cumulative GC doses [7, 8].
GC is the mainstay of treatment for GCA. The current European Alliance of Associations for Rheumatology (EULAR) and American College of Rheumatology (ACR) consensus treatment guidelines for induction of remission in patients with active GCA are considered the standard of care (SOC), both recommending starting prednisone (or equivalent) at 0.5–1 mg/kg/d (generally 40–60 mg/d) [9, 10]. After achieving control (usually after one month), GC doses are recommended to be tapered within 2–3 months until 15–20 mg/d and then to ≤5 mg/d after 1 year [9].
The total duration of GC therapy differs among patients. GC can be discontinued in almost half of the patients within 2 years [9, 11, 12]. However, most patients need prolonged GC therapy and some of them require low-dose GC indefinitely [11, 12].
In order to reduce GC-related secondary effects in patients with GCA, several conventional immunosuppressive drugs have been tried as GC-sparing agents. Azathioprine and leflunomide seem to exert GC-sparing effects in patients with difficult-to-treat and GC-refractory GCA [13, 14]. The efficacy of adjunctive therapy with MTX has been explored in patients with newly diagnosed GCA in three randomized clinical trials with controversial results [15–17].
However, an individual patient data meta-analysis of the three clinical trials demonstrated that the additional use MTX reduces the relapse risk and cumulated GC doses [18]. Anti-tumor necrosis factor (TNF) agents, such as infliximab and etanercept have been used in two randomized placebo-controlled trials in newly diagnosed patients with GCA and patients with GC-related adverse effects, with negative and inconsistent results, respectively [19, 20]. Tocilizumab was effective in patients with active GCA (newly diagnosed and relapsing disease) in sustaining GC-free remission in a 52-week randomized, double-blind, placebo-controlled trial [21] with maintained effects during three years [22].
The 2022 EULAR management guidelines recommend adding tocilizumab or MTX in GCA patients with refractory or relapsing disease, or with an increased risk of GC-related adverse effects or complications [9]. More recently, upadacitinib has also shown efficacy in a phase 3 trial and has been approved for the treatment of GCA [23].
The administration of intravenous methylprednisolone (IVMP) at high doses has been scarcely explored with controversial results [24, 25]. Consequently, the use of IVMP pulses has been formally limited by the 2022 EULAR consensus task force to patients with GCA presenting with ischaemic complications [9].
This observational study aimed to analyse two GC regimens for the treatment of GCA. The SOC regimen with prednisone alone at initial doses of 1 mg/kg/d (or 40–60 mg/d) was compared with a combination of IVMP pulses followed by prednisone at lower doses and MTX as GC-sparing agent, in order to examine the efficacy and safety of both therapeutic plans. The study also tried to identify clinical profiles in which the use of IVMP and MTX could result more useful in avoiding GC-associated toxicity and relapses.
Patients and methods
Study design
This is a study of usual clinical practice aiming to compare two remission induction and maintenance regimens for newly diagnosed patients with GCA by using the observational databases of three Autoimmune Diseases Units at Internal Medicine Departments in three Spanish academic hospitals: Hospital Clinic de Barcelona (HCLI, in Barcelona), Hospital Universitario Cruces (HCRU, in Barakaldo) and Hospital Universitario 12 de Octubre (H12-O, in Madrid).
Patients, data collected and outcomes
The diagnosis of GCA was based on clinical, imaging and histological findings [1] and patients fulfilled ACR/EULAR 2022 classification criteria [4]. All patients with GCA were diagnosed and received initial treatment in the participating sites, with a minimum regular follow-up period of 2 years.
The study databases were approved by the local Institutional Review Boards of the participating sites (Hospital Clinic de Barcelona [HCB/216/0274], Hospital Universitario Cruces [CEIC E18/31] and Hospital Universitario 12 de Octubre [CEI 16/331]) in accordance with Spanish legislation and the Declaration of Helsinki. Patients signed an informed consent form for the use of their data for clinical research at the time of entering the cohort.
Data collected at disease onset included sex, age, headache and other craniofacial symptoms, transient or irreversible ischaemic manifestations, PMR, fever (>37°C), weight loss (>4 kg or 4% of body weight) and time of follow-up. Laboratory markers recorded included ESR, CRP and haemoglobin levels. All baseline acute phase reactants were collected before or within 3 days of GC therapy.
Clinical and laboratory information was also collected during the course of the disease when relapses were identified. As in previous studies, the SIR was evaluated by the presence of four parameters such as fever (>37°C), weight loss (>4 kg or 4% of total body weight), ESR ≥85 mm/h and haemoglobin <11 g/dl [5–8], and patients were considered to have a weak SIR when they had two or fewer inflammatory parameters and a strong SIR when three or four parameters were present [5–8].
The main outcomes analysed were the achievement of disease remission, defined as the absence of clinical signs and symptoms of GCA with the normalization of ESR (<30 mm/h) and CRP (<10 mg/l), and the occurrence of relapses, which were considered as the recurrence of signs and symptoms for GCA or PMR with an increase in ESR or CRP, following a period of sustained remission [9, 10]. For the purposes of this study, imaging techniques were not used to define remission or relapse.
Secondary outcomes included time (weeks) since diagnosis until achieving a prednisone dose ≤5 mg/d, cumulative dose of oral GC and GC-attributable toxicity, which included the new onset or worsening of a previous hypercholesterolaemia requiring changes in therapy, or the new onset of diabetes mellitus, cataracts, symptomatic osteoporotic fractures, symptomatic osteonecrosis and severe infections (defined as those requiring admission and/or resulting in patient’s death). No baseline assessment for asymptomatic manifestations (osteoporotic fractures, osteonecrosis, cataracts) was routinely made.
Therapeutic schemes
Two different therapeutic schemes for remission induction of GCA were analysed. The EULAR/ACR SOC, based on monotherapy with prednisone [9, 10], was compared with a regimen of IVMP pulses and MTX (IVMP/MTX-based). By protocol, all patients at HCLI were treated with the SOC regimen and all patients at HCRU received the IVMP/MTX-based regimen.
Patients at H12-O received one or other scheme at the decision of the attending physician.
In the SOC regimen, starting doses of oral prednisone 40–60 mg/d (or equivalent) during 4 weeks were followed by a progressive tapering until achieving a maintenance dose of 5–2.5 mg/d. In this SOC group, pulses of IVMP at doses ranging from 125 to 1000 mg/d during 3 days were reserved only for patients with severe ischaemic manifestations.
All patients in the IVMP/MTX-based regimen were treated with three initial pulses of IVMP 125, 250 or 500 mg/d (125 or 250 mg/d depending on the severity of systemic symptoms and 500 mg/d pulses being reserved for patients with severe ischaemic manifestations, like visual loss or stroke) and MTX. IVMP pulses were followed by prednisone at initial doses of 20–30 mg/d with a fixed tapering scheme: weekly reductions to 20 and 15 mg/d; once 15 mg/d were reached, prednisone was reduced within 2 weeks to 10 mg/d and then every 2–4 weeks to 10/7.5, 7.5 and 7.5/5 mg/d, reaching a maintenance dose of 5 mg/d in 12–16 weeks. Reduction to prednisone 2.5 mg/d was accomplished within eight additional weeks.
Additional immunosuppressive therapy with MTX at a weekly dose of 7.5–15 mg (either oral or subcutaneous) was started from the beginning in all patients, except for those with severe refractory ischaemic complications in whom IV CYC 500 mg every two weeks up to 3–6 doses was started due to the rapidity of action, followed by MTX after CYC completion.
Statistical analysis
Descriptive data were generated, using percentages, means and standard deviations (SD), for the whole cohort and for the two therapeutic approaches. Data were compared by Chi-squared test or Student’s t test as appropriate.
For the multivariate adjustment, a propensity score (PS) analysis was performed [26]. The individual PS expresses the probability of being treated with the SOC prednisone vs IVMP/MTX-based protocols, and was calculated for each patient by using a logistic regression model, in which age at the time of diagnosis of GCA, gender, centre (HCLI, HCRU, H12-O), baseline clinical presentation (e.g. headache, jaw claudication, PMR, fever, weight loss, neuro-ophthalmologic ischaemic events) and laboratory (ESR, CRP and haemoglobin) values, biopsy and/or imaging techniques confirming GCA, presence of diabetes mellitus and the SIR at disease onset were all entered as independent predictors. The discriminatory capacity of this model was assessed by the area under the receiver operating characteristic (ROC) curve, with a resulting value of 0.85.
Since all patients in both the groups achieved remission, we compared the time to remission in the SOC prednisone and IVMP/MTX-based groups by Kaplan–Meier curves and log-rank test. For the multivariate analysis, we used Cox regression models adjusted for PS. All the efficacy analyses were performed for the whole cohort, but also separately for patients with different SIR (scoring ≤2 or ≥3), and for those with or without ischaemic events.
For patients developing ischaemic complications, the effect of therapy on preventing permanent ischaemic damage was also calculated.
Similarly, the influence of SOC prednisone and IVMP/MTX-based schemes on the occurrence of relapses was assessed by Chi-squared test in the whole cohort and in the two SIR (weak and strong) subgroups. The multivariate analysis was completed by logistic regression models adjusted by PS.
The effects of SOC prednisone and IVMP/MTX-based therapeutic modalities on GC cumulative doses were assessed by Student’s t test. For this purpose, several GC-related continuous variables were used, including the initial prednisone dose, cumulative prednisone dose at 3, 6, 12 and 24 months and time to achieve a prednisone dose of 5 mg/d. To assess the effects of both treatment regimens on GC-attributable toxicity, univariate analyses (for each of the individual toxicity items and for presenting with any of them) were performed by Chi-squared test.
The multivariate analysis was completed by using a logistic regression model adjusted by variables potentially influencing the occurrence of GC-related side effects (sex, age at GCA onset, diabetes and/or hypertension previous to GCA diagnosis) as well as the baseline SIR and the presence of ischaemic complications as markers of disease severity. The statistical analysis was done using STATA/MP 18.0 for Mac (StataCorp LP, TX, USA).
Results
Clinical presentation
One hundred and fifty-one patients with GCA were included in the study (72 [48%] from HCLI, 59 [39%] from HCRU and 20 [13%] from H12-O). Of them, 79 (52%) were treated with the SOC prednisone-based scheme and 72 (48%) patients with the IVMP/MTX-based regimen. All HCLI patients were treated with prednisone alone, those from HCRU received the IVMP/MTX-based regime, and H12-O patients followed a mixed protocol: seven (35%) with prednisone alone and 13 (65%) with IVMP/MTX.
The main demographic and clinical characteristics of the whole cohort and the study groups are shown in Table 1. There were no significant differences between groups regarding age, sex, clinical manifestations and laboratory parameters (ESR, CRP and haemoglobin levels) at disease onset. The SIR index was also similar in both the groups.
Neuro-ophthalmologic ischaemic manifestations tended to be less frequent, although not statistically significant, in patients receiving SOC prednisone-based therapy (33% vs 47% in IVMP/MTX group; P = 0.073).
Differences between the groups on GC use
All data regarding the therapeutic regimens used are illustrated in Table 2. While all patients in the IVMP/MTX-based group received IVMP, 13 (16.5%) patients in the SOC prednisone group were also initially treated with IVMP because of the development of neuro-ophthalmologic ischaemic events. Most patients in the IVMP/MTX-based group received doses of either 125 mg (31 [43%] cases or 250 mg (29 [40%] cases).
On the other hand, 1000 mg IVMP pulses were only used in five patients, all in the SOC prednisone group. As per protocol, prednisone initial doses were higher in the SOC prednisone than in the IVMP/MTX-based group (mean 56.3 vs 22.3 mg/d; P < 0.001), and MTX was used in all but one (98.6%) patient in the IVMP/MTX-based group vs none in the SOC group. In the same IVMP/MTX group, CYC was used in six (8%) patients due to severe refractory ischaemic manifestations.
Tocilizumab was marginally used in two patients (one in each group) as rescue therapy.
Patients treated with IVMP/MTX-based scheme received lower cumulative doses of prednisone during all the follow-up time points than patients in the SOC group (Table 2). The time to achieve a reduction of prednisone to ≤ 5 mg/d was significantly lower in the IVMP/MTX group (mean 13.9 vs 56.6 weeks, P < 0.001).
Remission and relapses according to the therapeutic scheme used
All 151 patients eventually achieved remission, at a median time of 4 (IQR 3–7) weeks, similar for both the treatment groups (SOC prednisone vs IVMP/MTX, P = 0.7), as reflected in the Kaplan–Meier survival analysis (Fig. 1), comparing the time (in weeks) to remission in both the groups (P = 0.93, log-rank test), and by the multivariate Cox regression (PS-adjusted HR 1.04, 95% CI 0.67–1.61; P = 0.84).
In the analysis by clinical subgroups, no differences were seen between both therapeutic regimens in patients with different SIR (score ≤2 or ≥3) (Fig. 2a and b), and those with or without ischaemic manifestations at disease onset (Fig. 2c and d).
Among patients presenting with ischaemic manifestations, the occurrence of permanent sequelae was also similar in both the treatment groups (SOC prednisone 54% vs IVMP/MTX 50%; P = 0.78); PS-adjusted logistic regression, OR 0.54 (95% CI 0.12–2.4; P = 0.42). Similarly, no significant differences between both the treatment groups were observed in any of the subgroups (by SIR or ischaemic events) in the propensity score-adjusted Cox models (Table 3).
With regard to relapses, 13 (8.6%) patients suffered relapses during the two-year study period. Both the treatment groups had a similar proportion of relapsing patients (SOC prednisone 9% vs IVMP/MTX 8.3%; P = 0.9). Likewise, no significant differences in relapse risk were seen between both the therapeutic groups in the logistic regression model (PS-adjusted OR 0.42, 95% CI 0.11–1.58; P = 0.20).
Relapses were more frequent, although not statistically significant, in patients with a strong vs a weak SIR (14% vs 6.5%; P = 0.14). However, the occurrence of relapses did not differ in the SOC prednisone vs the IVMP/MTX groups among patients with either a weak SIR (7.2% vs 5.6%; P = 0.7) or a strong SIR (12.5% vs 15%; P = 0.8).
GC-associated secondary effects
GC-related adverse effects were recorded in 43 (28.5%) patients. The most frequent complication was new or worsening hypercholesterolaemia (15% patients), followed by new onset diabetes mellitus (6%), cataracts (4.6%) and symptomatic osteoporotic fractures (4%). No cases of osteonecrosis were recorded.
Severe infections requiring admission were recorded in six (4%) patients (two gastrointestinal, two respiratory and two urinary), none of them fatal, and all occurring in the SOC group.
Hypercholesterolemia, cataracts and severe infections tended to occur more frequently in the SOC prednisone-based group, although the differences were only significant for infections (P = 0.029). The use of the IVMP/MTX scheme was associated with a reduction in the proportion of patients suffering any GC-related complication (20% vs 37%, P = 0.019, Table 4). This reduction in the risk of GC-related complications was confirmed in the logistic regression analysis (adjusted OR 0.37, 95% CI 0.15–0.89; P = 0.027).
Discussion
Although GC remain the cornerstone of therapy of GCA, the long-term use of oral prednisone (or equivalent) is associated with a remarkable risk of potential toxicity [9, 24, 27]. Because high doses of intravenous GC have both genomic and non-genomic signalling pathways, they exert a more rapid and potent anti-inflammatory effects with a safer profile, when given over short periods of time (pulse therapy) [28]. However, due to the paucity of data on the efficacy of IVMP pulses as induction therapy in GCA, the administration of IVMP 250–1000 mg for 3–5 days is only advised by current guidelines in patients with neuro-ophthalmologic manifestations (such as amaurosis fugax, acute visual loss, diplopia or stroke) [9, 10].
The use of IVMP pulses in patients with GCA without visual disturbances has been explored with different results. A randomized, double-blind, placebo-controlled study of 27 patients with non-ischaemic GCA compared IVMP pulses (at 15 mg/kg/d) or placebo for three consecutive days, followed by oral prednisone at initial doses of 40 mg/d following the same tapering regimen. After 78 weeks, more patients receiving the extra-doses of IVMP achieved sustained remission and were also taking low doses of prednisone in a significantly higher proportion [24].
Another study of 164 patients with GCA compared the additional use of IVMP 240 mg followed by oral prednisone with prednisone monotherapy, both at doses of 0.7 mg/kg/d, subsequently and equally tapered in all patients. IVMP pulses did not provide significant long-term GC-sparing effects in any of the parameters analysed after one year of follow-up [25]. In a retrospective study in patients with GCA with and without visual symptoms, the administration of IVMP was not associated with visual acuity or survival improvement.
Among 419 patients, 109 (26%) were treated with IVMP at 500–1000 mg/d during 3 days, all of them having visual symptoms. Patients receiving IVMP were treated with higher initial prednisone doses (median 60 vs 50 mg/d) and maintained slightly higher daily prednisone doses at three months, compared with those treated with prednisone alone (median [IQR] 20 [17.5–30] vs 20 [15–27.5] mg/d; P = 0.04) [29]. A previous study by the HCRU group explored the effects of different doses of oral GC in the achievement of remission in patients with GCA.
Compared with patients receiving oral prednisone (at different doses), those treated with IVMP 250–500 mg/d during 3 days had a more rapid remission and reduction of prednisone to ≤7.5 mg/d, regardless of whether patients were receiving additional therapy with MTX or not [30]. On the other hand, a study of 127 patients with autoimmune diseases (mainly systemic vasculitis) treated with initial high daily and continuous GC doses and with systematic studies for bone mineral density identified the use of initial GC pulses as a major risk factor for developing vertebral fractures [31]. Given the controversial results and the limited quality of evidence, the 2022 EULAR consensus task force recommended limiting the use of IVMP to patients with GCA and ischaemic complications [9].
In the present observational study, the IVMP/MTX-based regimen consisted of daily IVMP (125, 250 or 500 mg) during 3 days immediately followed by prednisone at medium starting doses (20–30 mg/d), plus the addition of MTX (or CYC in a few patients with severe ischaemic symptoms) as a GC-sparing agent, with rapid tapering of prednisone. Compared with the SOC regimen based on the only use of prednisone at high doses, the IVMP/MTX combined treatment showed similar results in terms of remission and occurrence of relapses. In this regard, remission was achieved by all patients at a median time of 4 weeks, with both therapeutic modalities not influencing the response rates either in the whole cohort or in different subgroups of clinical presentations, such as higher/lower intensity of the SIR or presence/absence of neuro-ophthalmologic ischaemic complications.
However, the use of IVMP/MTX was associated with a shorter time to reach prednisone doses ≤5 mg/d, lower GC cumulative doses and reduced GC-related toxicity during 2 years of follow-up.
Our study has several limitations, including: (1) the variability among therapeutic schemes even within the same treating group or participating centre, such as the use of MTX or CYC in the IVMP/MTX treated patients, and the different doses of IVMP in both the groups. This potential bias was overcome by performing a PS analysis, in which the different participating centres were entered as an independent predictor, in addition to a large number of clinical and laboratory variables, for the calculation of the individual PS coefficients, all multivariate analyses being subsequently adjusted by PS; (2) the efficacy and toxicity profiles of the different doses of IVMP could not be analysed, since the small number of patients did not allow the fragmentation of the IVMP/MTX-based group; (3) the specific role of IVMP and MTX was not possible to disclose within the IVMP/MTX-based group, since all patients (except one) received both drugs; (4) the attribution to GC of the different items of toxicity considered in this study is arguable, since all of them can be indeed part of the usual multifactorial risk factors in an elderly population, as expected in a cohort of patients with GCA; and (5) the relatively short follow-up time.
In conclusion, this study emphasizes the rapid response of patients with GCA to GC. Both GC (IVMP/MTX and SOC prednisone)-based regimens provided similar results in terms of achieving disease remission and controlling relapses at 2 years of follow-up. The use of IVMP/MTX-based modality does not seem to influence response according to the different forms of disease presentation (different intensity of the SIR or development of ischaemic complications).
However, the IVMP/MTX-based regimen was associated with a much more rapid reduction and a much lower cumulative GC dose, with a lower risk to experience GC-related side effects than the SOC prednisone-based therapy.
Although these results should be prospectively assessed in clinical trials and larger cohorts of patients with GCA, the initial treatment with high doses of IVMP during 3 days followed by reduced doses of prednisone plus MTX could be a reasonable GC-sparing therapeutic strategy for remission induction and maintenance in patients with GCA, especially those at higher risk for GC toxicity.