Article Text
PDF
Abstract
Background Immune and vascular ageing are proposed risk factors for giant cell arteritis (GCA). Data on the impact of age at diagnosis of GCA on the clinical presentation and course of the disease are scarce.
Methods Patients with GCA followed at referral centres within the Italian Society of Rheumatology Vasculitis Study Group were enrolled up to November 2021. Patients were grouped according to age at diagnosis: ≤64, 65–79 and ≥80 years old.
Results The study included 1004 patients, mean age 72.1±8.4, female 70.82%. Median follow-up duration was 49 (IQR 23–91) months. Patients in the oldest group (≥80 years) had significantly more cranial symptoms, ischaemic complications and risk for blindness compared with the groups 65–79 and ≤64 years (blindness: 36.98% vs 18.21% vs 6.19%; p<0.0001). Large-vessel-GCA was more frequent in the youngest group (65% of patients). Relapses occurred in 47% of patients. Age did not influence the time to first relapse, nor the number of relapses. Older age was negatively associated with the number of adjunctive immunosuppressants. Patients >65 years old had 2–3 fold increased risk for aortic aneurysm/dissection up to 60 months follow-up. Serious infections, but not other treatment-related complications (hypertension, diabetes, osteoporotic fractures), were significantly associated with older age. Mortality occurred in 5.8% of the population with age >65, cranial and systemic symptoms as independent risk factors.
Conclusions The highest risk of ischaemic complications, aneurysm development, serious infections and the possible undertreatment make of GCA a very challenging disease in the oldest patients.
- Giant Cell Arteritis
- Systemic vasculitis
- Outcome Assessment, Health Care
Data availability statement
Data are available on reasonable request. Study protocol and raw data are accessible on reasonable request.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
The impact of age at disease onset in patients with giant cell arteritis (GCA) is still largely unknown and could potentially provide informative evidence to stratify the risk of disease complications in individual patients’ groups.
WHAT THIS STUDY ADDS
This study demonstrates that age at disease onset has a significant impact on the type of clinical manifestations, risk for ischaemic sequelae, treatment strategies, disease-related and treatment-related complications in patients with GCA.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
To the best of our knowledge, with over 1000 patients enrolled at a national level, this is the largest real-life cohort of patients with GCA ever described to assess the predictors of a number of outcomes of interest. The study contributes to inform on the risk profile of patients and provides valuable prognostic information according to simple information such as age groups at diagnosis as a first step towards a tailored approach to the management of GCA.
Introduction
Giant cell arteritis (GCA) is characterised by significant morbidity derived from the disease itself or its treatment.1 Permanent visual loss (PVL) is the most dreadful complication usually occurring during the initial phases of the disease.2 3 Since the introduction of specialised clinics for the diagnosis of large-vessel (LV) vasculitis (LVV), incorporating the use of ultrasound into fast-track pathways for urgently referred patients, the visual prognosis of GCA has significantly improved.4 5 Nevertheless, early diagnosis of GCA has not been coupled with a reduction in the relapse rates, nor with a reduction of the long-term consequences of the disease, including aneurysm formation.4–6 A better understanding of the influence of the patient-onset and disease-onset characteristics on relevant clinical outcomes could allow to improve a more tailored management in the future. The impact of age at disease onset in patients with GCA is still largely unknown and could potentially provide informative evidence to stratify the risk of disease complications in individual patients’ groups. Ageing has a definitive, although largely obscure, aetiopathogenetic role, accounting for the typical epidemiological distribution of GCA. Several hypotheses have been made to justify the different age onset and anatomical tropism that differentiate diseases along the spectrum of LVV. Immunosenescence and vascular changes at different anatomical sites occurring over time are possible explanations. On the clinical side, available evidence suggests a role for age in influencing the clinical phenotype, with patients with LV-GCA being generally younger and with less frequent ocular involvement than patients with cranial disease. In turn, LV-GCA has been recognised as a risk factor for relapsing disease and aortic structural damage development.7 Given the lack of a reliable biomarker reflecting the extent and severity of disease, further studies on large cohorts of patients are needed to better stratify patients according to disease subtypes and prognosis. Moreover, the management of elderly patients has significant treatment implications that deserve more data-driven evidence to support better care. The aim of this observational studygca-age (GCAGE study) was to evaluate the impact of age on the clinical manifestations, risk and time to relapse, treatment choices, and main outcomes in one of the largest cohort of patients with GCA collected to date under the guidance of the Italian Society of Rheumatology Vasculitis Study Group.
Methods
Patients with a confirmed clinical diagnosis of GCA between 1 January 1988 and 1 November 2020, satisfying the 1990 American College of Rheumatology classification criteria8 and/or Chapel Hill Conference9 definitions were included in this retrospective and prospective observational cohort study. Following the publication of the 2022 ACR/EULAR classification criteria for GCA,10 88% of the patients included in the cohort were confirmed to fulfil the new criteria. Patients were diagnosed and followed at 18 referral centres for the management of vasculitides participating in the Italian Society of Rheumatology Vasculitis Study Group. Medical records of the included patients were reviewed from the date of GCA diagnosis to the end of the study follow-up (1 November 2021), last visit, migration or death. Patients with at least 6 months follow-up were included. Data were recorded using a standardised electronic data collection form. It was not appropriate or possible to involve patients or the public in the design, conduct or reporting of our research.
The list of collected variables, definitions and details on the year of diagnosis are available in online supplemental material. ESR and C reactive protein (CRP) were defined as very elevated if above the median for the whole population (erythrocyte sedimentation rate >77 mm/hour; CRP>64 mg/L). Relapses were defined as recurrence of signs and/or symptoms of GCA (including new or progressive LV stenosis or aneurysm on imaging) leading to an increase in GC or immunosuppressive treatment. The following clinical outcomes were assessed at 12 and 60 months from diagnosis: aortic aneurysm or dissection, serious infections, and a combined outcome of new-onset hypertension, diabetes, osteoporotic fractures. Survival was assessed.
Supplemental material
Endpoints: The primary endpoint was to assess the association of age at disease onset with the number of relapses. Secondary endpoints were association of age with time to first relapse; association of age at disease onset with the time to reach a low-dose GC (GC dose ≤5 mg/day); association of age at disease onset with the time to start another immunosuppressive drug (ID) and ID number; association of age at disease onset with disease-related and treatment-related outcomes at 12 and 60 months and overall survival.
Statistical analysis
We analysed data with Stata V.17 (StataCorp). A two-sided p<0.05 was considered statistically significant. We set significance at 0.017 for post hoc comparisons between age groups; these corresponded to supposedly clinically relevant age group categorisation (≤64, 65–79 and ≥80 years). We used the mean and SD or the median and quartiles (IQR) to describe continuous variables and compared them with the Kruskall-Wallis test; we described categorical variables with counts and percent and compared them with the Fisher’s exact test. We computed incidence rates as the ratio of counts and total follow-up times together with their 95% Poison CIs. We computed median survival and IQR with the reverse Kaplan-Meier method. We analysed the primary endpoint using both a univariable and a multivariable zero-inflated Poisson model with age group adjusted for a series of predefined confounders (age, sex, phenotype at disease onset (cranial vs LV-GCA); ischaemic onset, systemic symptoms at disease onset, very elevated acute phase reactants at onset, initial GC dose, Hb levels at onset) selected according to potential clinical significance and available literature evidence.11 12 We computed incidence rate ratios (IRR) and 95% CI. For the secondary endpoints, we analysed the time to event endpoints with univariable and multivariable Cox models; we computed HRs and 95% CI. We plotted Kaplan-Meier curves by age group. We analysed the number of new ID with negative binomial regression; we report IRR and 95% CI. We used logistic regression to model disease-related and treatment-related outcomes at 12 and 60 months; we computed ORs and 95% CI. Sensitivity analyses are presented in online supplemental material.
Results
General characteristics
The cohort consisted of 1004 patients with GCA, mean age 72.1±8.4 years, female 711 (70.82%). Median follow-up duration was 49 (IQR 23–91) months.
When stratifying the population according to age groups at diagnosis, there were 195 patients in the group of patients diagnosed ≤64 years of age, 617 patients in the group diagnosed between 65 and 79 years of age, and 192 patients with a diagnosis ≥80 years of age.
The general characteristics of the population according to age groups are presented in table 1. The distribution of symptoms at disease onset according to age groups is presented in figure 1. Patients in the oldest group (≥80 years old) had significantly more cranial symptoms compared with younger patients. Similarly, PVL was significantly more frequent in the oldest age group with 71 (36.98%) presenting with blindness, compared with 112 (18.21%) in the age group 65–79, and 12 (6.19%) in the ≥64 years old group; p<0.0001. There were no new cases of blindness at 12 months follow-up, and one adjunctive patient developing PVL at 60 months follow-up. A sensitivity analysis excluding patients with a clinical diagnosis only did not change the results on PVL.
General characteristics of the population according to age groups
Distribution of onset symptoms according to age groups.
Primary endpoint
Association of age at disease onset with the number of relapses
Relapses occurred in 473 (47%) patients over a median follow-up period of 4.1 years (IQR 1.9–7.6), accounting for an overall rate of 15 relapses per 100 person-year (95% CI 14 to 16).
The incidence of relapses according to age groups was as follows, group ≤64 years: rate of 14 relapses per 100 person-year (95% CI 12 to 16); group 65–79 years: rate of 16 relapses per 100 person-year (95% CI 14 to 17); group ≥80 years: 12 relapses per 100 person-year (95% CI 10 to 14); p=0.001.
In a multivariable model (table 2), age groups at disease onset did not influence the number of relapses during follow-up. On the other hand, having cranial manifestations at disease onset was an independent risk factor for the number of future relapses. Moreover, disease onset with ischaemic manifestations, and having systemic symptoms at onset were associated with a lower number of relapses. Adding the use of ID initiated at the time of diagnosis to the model was associated with a significant reduction of the number of relapses, but did not change the effect of age.
Multivariable analysis for the predictors of the number of relapses and time to first relapse
In a sensitivity analysis excluding patients with a diagnosis of GCA based on clinical diagnosis only, the results did not significantly change.
Secondary endpoints
Association of age at disease onset with the time to first relapse
Age at disease onset did not influence the time to first relapse (figure 2). The rate of relapses (patient level) was 14 per 100 person years (95% CI 11 to 16) for the age group ≤64; 15 per 100 person-years (95% CI 14 to 17) for the age group 65–79; 13 per 100 person-years (95% CI 10 to 16) for the group ≥80 years, p=0.055. The multivariable analysis for the association of age with the time to first relapse is shown in table 2. None of the assessed parameters significantly predicted the time to first relapse.
Kaplan-Meier for relapse-free survival estimate by age group.
Association of age at disease onset with the time to reach low-dose glucocorticoids
Patients reaching a dose ≤5 mg/day were 155 (83%) in the group ≤64 years of age; 473 (80%) in the group 65–79, and 155 (83%) in the group ≥80 years old.
The Cox model assessing the impact of age on the time to reach GC≤5 mg/day confirmed the significant impact of age; p=0.004 (figure 3).
Kaplan-Meier curve for the estimate of reaching a low glucocorticoid (GC) dose ≤5 mg/day according to age group.
However, in the multivariable analysis, age was not confirmed to be an independent risk factor for the time to reach a lower GC dose (table 3). On the other hand, having ischaemic manifestations was inversely correlated with the time to reach GC≤5 mg/day. Hb levels at baseline were associated with the time to reach GC≤5 mg/day. The initiation of an ID from diagnosis significantly influenced the time to reach GC≤5 mg/day but did not change the effect of age.
Multivariable analysis for the predictors of the time to reach lower glucocorticoid (GC) dose and the need to start and number of immunosuppressive drugs
Association of age at disease onset with the time to start another ID and their number
Patients in the group ≤64 years of age received an ID in 129 (71%) cases; group 65–79 years in 308 (61%) cases; group ≥80 years in 59 (37%) cases. The median time to ID initiation was 4.61 (IQR 0.46–12.57) months. Overall, 231 (45%) of patients received ID at the time of GCA diagnosis, without significant differences according to age groups, p=0.938. The remaining patients were treated with ID after ≥1 relapse. The majority of patients (47%) received one ID during follow-up, with 12% of patients needing a second-line ID over time. Few patients (9%) received ≥3 ID.
Methotrexate (MTX) was the most frequently prescribed ID (364 patients received MTX, of whom 167 from diagnosis), followed by IL-6 inhibitors (178 patients received tocilizumab, of whom 46 from diagnosis). Other ID were as follows: leflunomide (n=3), azathioprine (n=34), mycophenolate mofetil (n=11), cyclophosphamide (n=14), TNF-inhibitors (n=12), ustekinumab (n=2).
Age groups were not statistically associated with the need for a first adjunctive ID (table 3). In a multivariable analysis, having the cranial subset of disease was inversely correlated with the need for the first ID. Moreover, having ischaemic symptoms was a risk factor for the need of a first ID. Onset Hb values were associated with the need for ID.
The age group ≥80 years of age received a significant smaller number of ID: HR 0.58 (95% CI 0.37 to 0.91), p=0.019.
Post hoc analysis confirmed that the group ≥80 received less ID compared with the group 65–79: 0.60 (0.41; 0.86); p=0.002.
The Cox model assessing the impact of age on the number of steroid-sparing agent used during follow-up is shown in online supplemental figure 1.
In a multivariable analysis (table 3) older age at disease onset (≥80 years) was negatively associated with the total number of ID; on the other hand, initial GC dose >50 mg/day was significantly associated with the total number of ID.
Association of age at disease onset with disease-related and treatment-related outcomes at 12 months
Aortic aneurysm/dissection at 12 months follow-up
At 12 months from diagnosis, there were 11 (6.83%) cases of aortic aneurysm/dissection in the group ≤64 years, 20 (4.18%) in the group 65–79, and 11 (7.86%) in the group ≥80 years of age; p=0.136.
At multivariable analysis, group ≥80 had a significantly higher risk of developing aortic complications compared with both younger groups (online supplemental table 1). The comparison of the age groups 65–79 and ≤64 was not significant. Male sex was identified as an independent risk factor for the development of aortic complications at 12 months. Cranial phenotype and cranial onset symptoms were inversely correlated.
Serious infections at 12 months follow-up
Serious infections occurred in 48 (6.19%) patients distributed as follows: 3 (1.88%) in the group ≤64, 36 (7.32%) in the group 65–79, 10 (7.25%) in the group ≥80 years old; p=0.025. At multivariable analysis, the two older age groups were confirmed to be risk factors for serious infections, together with higher initial GC dose and onset with systemic symptoms (online supplemental table 1). Compared with age group ≤64 years, patients 65–79 had an OR of 11.90 (95% CI 1.69 to 83.63); p=0.013; group ≥80 years compared with ≤64 had an OR of 13.71 (95% CI 1.59 to 117.91); p=0.017.
Hypertensions, diabetes, fractures at 12 months follow-up
The combined outcome of treatment-induced complications (hypertension, diabetes, osteoporotic fractures) at 12 months was recorded for 26 (15.95%) patients in the group ≤64 years, 111 (22.93%) in the group 65–79 and 33 (23.91%) in the group ≥80 years old; p=0.131. At multivariable analysis, age was not associated with the risk to develop such treatment-induced complications. Systemic symptoms at onset were an independent risk factor (online supplemental table 1).
Association of age at disease onset with disease-related and treatment-related outcomes at 60 months
Aortic aneurysm/dissection at 60 months follow-up
At 60 months from diagnosis, there were 10 (9.90%) cases of aortic aneurysm/dissection in the group ≤64 years, 16 (6.87%) in the group 65–79, and 5 (12.50%) in the group ≥80 years of age; p=0.326.
At multivariable analysis, age was a risk factor for aortic complications (online supplemental table 2). A higher rate of aortic complications was showed in the group ≥80 vs group 65–79 years old: OR 3.60 (95% CI 1.77 to 7.31); p<0.0001. Male sex was identified as an independent risk factor for the development of aortic complications at 60 months. Cranial phenotype and cranial onset symptoms were inversely correlated.
Serious infections at 60 months follow-up
Serious infections occurred in 31 (8.61%) patients distributed as follows: 5 (5.21%) in the group ≤64, 22 (9.78%) in the group 65–79, 4 (10.26%) in the group ≥80 years old; p=0.386. At multivariable analysis, older age groups were found to be risk factors for serious infections at 60 months (online supplemental table 2). Compared with age group ≤64 years, patients 65–79 had an OR of 3.95 (95% CI 1.72 to 9.09); p=0.001; group ≥80 years compared with ≤64 had an OR of 6.41 (95% CI 1.90 to 21.61); p=0.003.
Hypertensions, diabetes, fractures at 60 months follow-up
The combined outcome of treatment-induced complications (hypertension, diabetes, osteoporotic fractures) at 60 months was recorded for 15 (15%) patients in the group ≤64 years, 68 (29.57%) in the group 65–79 and 9 (23.08%) in the group ≥80 years old; p=0.016.
At multivariable analysis, age was not significantly associated with the risk of treatment-related complications at 60 months. On the other hand, systemic symptoms at onset were associated with the risk of hypertension/diabetes/fractures (online supplemental table 2).
Survival
Overall there were 58 deaths (5.8%). Mortality rate was 0.08 (95% CI 0.012 to 0.6) per 100 person-year for the age group ≤64, 1.26 (95% CI 0.92 to 1.72) per 100 person-year for the group 65–79; 1.94 (1.21 to 3.13) per 100-person year for the group ≥80 years old.
Mortality was significantly associated with the age at disease onset (online supplemental figure 2). The risk was significantly higher for the group 65–79 vs group aged ≤64 (p<0.0001), and those aged ≥80 vs ≤64 (p<0.0001).
At multivariable analysis, age at disease onset was confirmed to be a risk factor for mortality, together with cranial phenotype or cranial symptoms at onset, and systemic symptoms at onset (online supplemental table 3).
Discussion
Our study shows that age at disease onset has a significant impact on the type of clinical manifestations, risk for ischaemic sequelae, treatment strategies, disease-related and treatment-related complications in patients with GCA. To the best of our knowledge, with over 1000 patients enrolled at a national level, this is the largest real-life cohort of patients with GCA ever described to assess the predictors of a number of outcomes of interest. Age, not only influences the susceptibility to GCA, but also acts as a relevant determinant for the clinical phenotype of disease. We confirmed the previously recognised predominant distribution of LV-GCA in younger patients.7 Moreover, we demonstrated that typical cranial symptoms including headache, scalp tenderness, jaw claudication and amaurosis fugax were more likely to occur in patients belonging to the two older age groups at diagnosis compared with the younger group. Importantly, severe ischaemic complications including stroke and PVL were significantly more frequent in older patients (≥80 years old). In the oldest group, the frequency of blindness was as high as 36.98% of patients, compared with 18.21% in the group 65–79, and 6.19% in the youngest group. While the prognostic advantages in terms of reduced risk of blindness brought by the introduction of fast-track clinics for the early diagnosis of GCA are unquestionable, our data suggest that further efforts to improve the knowledge of the disease should be dedicated to elderly patients who carry the highest risk of referral after PVL has already occurred. In our cohort, all but one case of blindness occurred at the time of diagnosis rather than during follow-up, confirming that visual loss is usually an early manifestation of GCA and rarely occurs once the GCs are started.2 3 Previous studies have reported that advanced age at disease onset is an independent risk factor for blindness.3 13 The reasons leading to the different disease phenotypes in different age groups and to a higher ischaemic risk in older patients are still unknown. Previous evidence has implicated traditional cardiovascular risk factors, thrombocytosis, constitutional symptoms and low inflammatory response as risk factors for blindness in GCA.2 3 14–16 Generalised endothelial dysfunction, enhanced by traditional cardiovascular comorbid conditions, may be responsible for the increased risk for ischaemic complications, especially in the older age groups.17 Moreover, immunosenescence phenomena may contribute to perpetuate chronic inflammatory tissue damage and impair the vascular reparative mechanisms.17 18 Age-induced abnormalities of immune and non-immune tissue-residing cells in the vessel walls may lead to a higher susceptibility to the action of systemic proinflammatory cytokines in the context of vasculitis and enhance ischaemic complications, particularly in this specific older age group.17 Our study demonstrates that age at diagnosis of GCA is an independent risk factor for aortic complications both at 12 and 60 months follow-up, regardless of the presence of a recognised LV-GCA phenotype. Indeed, patients ≥80 years of age at diagnosis carry a 2–3 fold higher risk of developing aortic aneurysms/dissection compared with younger groups. Male gender was another confirmed risk factor for aortic complications, in line with previous evidence.19 It is possible that the same underlying cardiovascular risk factors, and vascular ageing changes leading to a higher risk of ischaemic manifestations during the early phases of the disease may contribute in time to the increased risk of aortic complications. However, it is also possible that given the higher prevalence of typical cranial onset in older patients, LV imaging is not performed systematically in all patients for diagnostic purposes therefore underestimating the frequency of LV-GCA in this group of patients. Aortic inflammation at disease onset is indeed one of the major risk factors for future aortic complications in GCA.19 20
Age at diagnosis was not found to be significantly associated with the time to first relapse and number of future relapses. Our study failed to identify any other independent risk factor for the time to first relapse. The lack of evidence (in terms of clinical manifestations, biomarkers or imaging findings) guiding the risk stratification of patients at diagnosis, and consequently tailored treatment, is a major unmet need in the management of GCA. As a consequence, current international recommendations21 22 suggest to treat patients similarly, regardless of disease phenotype. Nevertheless, our study demonstrates that in clinical practice, older age influences the time to reach a low-dose maintenance dose of GC and reduces the probability of receiving adjunctive ID to control the disease. It is possible that older patients are undertreated, and this might contribute to the higher risk of aortic complications demonstrated by our study. Nonetheless, the treatment decisions applied to older populations are certainly the result of a challenging risk/benefit balance accounting for safety issues, including the risk for serious infections. Our study showed that patients ≥80 years of age at diagnosis have a 13-fold increase in the risk of serious infections during the first year of disease. The risk reduces to a sixfold increase over the following 5 years of follow-up, but still remains highly significant, even if patients in the oldest group reach low doses of GC more rapidly than younger patients. A significant risk for serious infections was found also for the group 65–79 years of age. These findings have important implications for treatment choices in elderly patients with GCA. Our findings demonstrate that the risk for GC-induced complications including new-onset hypertension, diabetes, osteoporotic fractures are increased in patients with GCA, regardless of age at diagnosis. The development of such new cardiovascular risk factors often occurs early in the disease23 and adds on to the already increased cardiovascular burden of these patients. Moreover, the presence of diabetes in patients with GCA has been implicated in increasing the risk of mortality in case of serious infections24 generating a vicious circle of interconnected complications, adverse events and increased healthcare costs.25 Finally, our study demonstrated a higher mortality risk for older patients, together with cranial and systemic symptoms at onset as further independent risk factors. Besides the expected higher mortality risk related to older age and comorbidities, the finding of a significant association with aortic complications, known as the major risk factor for mortality in patients with GCA,26 and the higher risk for serious infections pose further threats to patients diagnosed with GCA at an older age.
Our study has some limitations including the observational nature, and the potential referral bias of more severe or complicated cases to reference centres in Italy. However, our cohort is one of the largest ever described and was not restricted to an individual specialist centre, providing greater generalisability than prior single-centre studies. Moreover, despite the high expertise level of the centres participating in the study, and the adherence to the international recommendations, we cannot exclude some degree of individual centre variation regarding the use of imaging to detect LV-GCA and treatment choices.
In conclusion, our study demonstrates that, given the higher risk for severe ischaemic manifestations, aortic complications, serious infections, death, and potential undertreatment, elderly patients with GCA represent a unique and challenging subgroup of patients that deserves further dedicated research to optimise their management and improve prognosis.
Data availability statement
Data are available on reasonable request. Study protocol and raw data are accessible on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by Comitato Etico Policlinico San Matteo di Pavia (REF. P20200015359), Italy. Participants gave informed consent to participate in the study before taking part.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Josef S Smolen
Twitter @tomelleri_a, @DrAleeG
Collaborators List of collaborators for the Italian Society of Rheumatology Vasculitis Study Group:-Alice Bartoletti (1. dipartimento di medicina interna e terapia medica, Università di Pavia, 2 Reumatologia, Fondazione IRCCS Policlinico San Matteo, Pavia);-Giovanni Zanframundo (1. dipartimento di medicina interna e terapia medica, Università di Pavia, 2 Reumatologia, Fondazione IRCCS Policlinico San Matteo, Pavia);-Elena Maria Baldissera (Unit of Immunology, Rheumatology, Allergy and Rare Diseases; IRCCS San Raffaele Hospital; Milan; Italy);-Chiara Asperti (Unit of Immunology, Rheumatology, Allergy and Rare Diseases; IRCCS San Raffaele Hospital; Milan; Italy);-Riccardo Bixio (Rheumatology Unit, Department of Medicine, University of Verona);-Gattamelata Angelica (Rheumatology Unit, Department of Clinical Internal, Anesthesiological and Cardiovascular Sciences, Sapienza University of Rome, Italy),-Giacomo Bagni (Department of Experimental and Clinical Medicine, University of Firenze, Firenze, Italy).
Contributors All authors contributed to the development of the research project, collection of data and revision of the manuscript. All collaborators contributed to the collection of data and revision of the manuscript. SM
acted as guarantor accepting full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.