Clinical and laboratory manifestations and treatment of children with TNFRSF1A gene variants

Abstract

BACKGROUND

The TNFRSF1A gene encodes TNFR1, which regulates inflammation and apoptosis. Variants in this gene present with an autoinflammatory phenotype and are linked to tumor necrosis factor receptor associated periodic syndrome (TRAPS), an orphan monogenic autoinflammatory disease (AID) that can mimic systemic juvenile idiopathic arthritis (sJIA) and other rheumatic diseases. Despite known mechanisms and classification criteria, diagnosing and choosing therapy for patients with TNFRSF1A variants remains challenging. Assessing phenotypic characteristics and considering alternative therapies for these patients is important.

AIM

To evaluate the clinical and laboratory features of sJIA associated with the TNFRSF1A gene and treatment outcomes.

METHODS

A single-center, retrospective, cross-sectional cohort study with longitudinal follow-up was performed. A total of 66 patients with fever and a referral diagnosis of sJIA were included: 33 (50%) had TNFRSF1A gene variants identified by molecular genetic testing, and 33 (50%) had no TNFRSF1A or other autoinflammatory gene variants. Demographic, clinical, laboratory, and treatment data were collected before and after verification of AID, and again at 3, 6, and 12 months.

RESULTS

Molecular genetic testing identified several TNFRSF1A variants. Of 33 patients, 4 (12%) had likely pathogenic variants, 1 (3%) had a pathogenic variant, and 3 (9.1%) had variants of uncertain significance; all were heterozygous. The most common, identified in 27 (82%), was a variant of uncertain significance (c.362G>A, p.Arg121Gln) with incomplete penetrance. These patients had paroxysmal (48.5%) or continuous fever (48.5%), rash (66.7%), joint syndrome (78.8%), abdominal pain (33.3%), gastrointestinal symptoms (27.2%), eye involvement (24.2%), chest pain (12.1%), hearing loss (9.1%), periorbital edema (9.1%), arrested development (6.1%), aseptic meningitis (3%), hydrocephalus (3%), and amyloidosis (3%). Endoscopic signs of intestinal damage were found in 20 patients (60.6%): 11 showed upper gastrointestinal tract issues; 9 had intestinal pathology. All patients with TNFRSF1A variants received antirheumatic therapy before genetic testing. Treatments included intravenous glucocorticoids (GC) (45.5%), oral GC (45.5%), intravenous immunoglobulin (18.1%), methotrexate (48.5%), cyclosporine (18.1%), azathioprine (3%), mesalazine (3%), sulfasalazine (3%), and biological (b) disease-modifying antirheumatic drugs (bDMARDs) (72.7%). Of those on bDMARDs, 39.4% received tocilizumab, 9.1% abatacept, 6.1% canakinumab, 6.1% infliximab, 3% rituximab, 3% etanercept, 3% adalimumab, and 3% certolizumab pegol. Among 24 patients with three prior bDMARD switches before genetic diagnosis, 17 (70.8%) achieved remission: 7 on tocilizumab, 6 on canakinumab, 2 on adalimumab, 1 on infliximab, and 1 on rituximab. After confirming AID, all patients received biological therapy: 9 previously untreated patients started canakinumab (3), tocilizumab (3), or etanercept (3); 7 switched drugs after ineffective prior therapy. After all therapy adjustments, all patients achieved remission: 13 on tocilizumab, 11 on canakinumab, 3 on adalimumab, 2 on etanercept, 1 on golimumab, 1 on rituximab, and 1 with GC and colchicine followed by withdrawal. Patients started on bDMARDs after genetic testing required fewer drug switches than those treated empirically before diagnosis (0.23 ± 0.42 vs 0.9 ± 0.6, P = 0.001).

CONCLUSION

Variants in the TNFRSF1A gene may cause a wide range of clinical symptoms, including eye and intestinal lesions that are not typical of sJIA. All patients with fever and unusual sJIA symptoms should have molecular genetic testing for TNFRSF1A variants to confirm or exclude AID early. Early diagnosis enables timely therapy and prevents complications. Tocilizumab (39.3%), canakinumab (33.3%), and TNF inhibitors (21.2%) achieved remission in children with TNFRSF1A variants. Some patients required multiple bDMARD switches to achieve remission, highlighting the complexity of treatment decisions in this group.

Key Words: Autoinflammatory diseases; Genetic; Biological disease modifying antirheumatic drugs; Rheumatology; Pediatric; Children; TNFRSF1A gene

Core Tip: In this study, we perform a comparative analysis of clinical and laboratory data from patients with TNFRSF1A variants and from patients with systemic juvenile idiopathic arthritis without variants in autoinflammation-associated genes, as well as an analysis of drug therapy. As a result of the study, the phenotypic features of patients with tumor necrosis factor receptor-associated periodic syndrome were identified, allowing them to suspect the disease before receiving genetic test results and to identify the most effective drugs for their treatment.

INTRODUCTION

Autoinflammatory diseases (AID) are a heterogeneous group of rare genetic conditions characterized by periodic attacks of inflammation, with fever and clinical symptoms imitating rheumatic diseases[1]. Tumor necrosis factor (TNF) receptor-associated periodic syndrome (TRAPS) is an orphan monogenic AID with autosomal-dominant inheritance associated with a change in the nucleotide sequence of the TNFRSF1A gene encoding the type 1 TNF receptor[2,3]. The altered TNFRSF1A gene affects the structure of the extracellular domain of the TNF receptor, disrupting its ability to bind to the TNF ligand. As a result, incorrectly folded receptors stimulate the hyperproduction of proinflammatory cytokines, including interleukin (IL)-1β, IL-6, TNF, and chemokines, with the development of inflammatory attacks[4-6]. The symptoms of TRAPS, such as fever, rash, ocular involvement, musculoskeletal symptoms, and serous membrane involvement, as well as changes in laboratory parameters, are similar to those of systemic juvenile idiopathic arthritis (sJIA)[7]. According to the Eurofever registry, the median delay in diagnosing TRAPS exceeds 10 years[3]. The average delay in diagnosis verification, according to foreign studies, is 15.6 ± 11.6 (range 2.0-31.0) years in patients with adult-onset TRAPS and 23.4 ± 16 (range 1.0-44.0) years in patients with pediatric-onset TRAPS[8]. Late diagnosis and administration of targeted therapy to patients with TRAPS can lead to severe complications, such as macrophage activation syndrome (MAS) and amyloidosis, with the development of multiple organ failure and even death[9,10]. Due to the Eurofever/PRINTO 2019 criteria and the potential of genetic diagnostics, the frequency of diagnosed TRAPS cases has increased[8]. The search for TNFRSF1A gene variants in a genetic study is an important diagnostic criterion for verifying the diagnosis of TRAPS[11]. Clinical manifestations characteristic of AID and not characteristic of sJIA can occur not only in carriers of pathogenic variants of the TNFRSF1A gene, but also in patients with variants of uncertain significance and polymorphisms, which indicates the genetic heterogeneity of the syndrome and the presence of factors that modify the development of the disease in carriers of the pathogenic gene[12].In addition to variants in autoinflammation-associated genes, the phenotype of a patient with AID can change under the influence of epigenetic factors, including external factors that affect gene expression and alter cellular functions without altering the genomic sequence[13]. The purpose of our study was: (1) To identify clinical and laboratory manifestations that differentiate patients with variants in the TNFRSF1A gene from patients with sJIA; (2) To identify new symptoms and organ damage in AID associated with variants in the TNFRSF1A gene; and (3) To evaluate the effectiveness of therapy in children with different TNFRSF1A gene variants.

MATERIALS AND METHODS

Study design and setting

This single-center retrospective cross-sectional cohort study with longitudinal follow-up included 66 patients with fever and a referral diagnosis of sJIA, admitted between March 1, 2013 and August 31, 2022 in the Division of Pediatric Rheumatology of the National Medical Research Center of Children’s Health, Moscow, Russian Federation, whose TNFRSF1A genic variants were subsequently identified.

Participants

All study patients had fever and a referral diagnosis of sJIA upon admission to the hospital. Among 33 patients diagnosed with sJIA according to the International League of Associations for Rheumatology criteria, subsequent genetic analysis did not reveal variants in genes associated with AID. An additional 27 patients met the criteria for sJIA but also exhibited symptoms suggestive of AID, while 6 patients presented only with symptoms consistent with AID. The inclusion criteria were: (1) For the first group (n = 33), patients with referral diagnosis of sJIA and variants in the TNFRSF1A gene; (2) For the comparison group (sJIA, n = 33), patients with confirmed sJIA and no variants in genes associated with autoinflammatory diseases according to genetic testing; and (3) Age ≤ 18 years at enrollment. The exclusion criteria were the presence of variants in genes characteristic of other AIDs and insufficient data for a complete analysis. The diagnosis of TRAPS was established according to the Eurofever/PRINTO criteria; the sJIA diagnosis was established according to the International League of Associations for Rheumatology criteria. Figure 1 illustrates patient selection and grouping.

Figure 1 Flowchart of patient selection and grouping by diagnosis. AID: Autoinflammatory disease; TRAPS: Tumor necrosis factor receptor associated periodic syndrome; sJIA: Systemic juvenile idiopathic arthritis.

Genetic analysis

DNA isolation: Genomic DNA was isolated from dried blood spots with the MagPure Universal DNA Kit (Magen, Guangzhou, China) using the automated Auto-Pure 96 DNA isolation station (Allsheng, Hangzhou, China) and from whole blood using a DNA Blood Mini Kit (QIAGEN, Germany) on a QIAcube automated station (QIAGEN, Germany) according to the manufacturer’s protocol. DNA quality and concentration were measured either by spectrophotometry with the NanoPhotometer N60 (Implen Gmbh, Munich, Germany) or with the Qubit dsDNA HS Assay Kit and Qubit 3.0 fluorometer (Invitrogen, Waltham, MA, United States).

Sanger sequencing

All primers used for TNFRSF1A (NM_001065.4) sequencing were designed with “Beacon Designer 8.10” (PREMIER Biosoft International), and their specificity was confirmed using “Primer-BLAST”. Oligonucleotides were synthesized by JSC “Evrogen”. DNA amplification was performed on a Bio-Rad T100 (Bio-Rad, United States) and a ProFlex (Thermo Fisher Scientific, Waltham, MA, United States) thermocycler. Sequencing reactions used the BigDye^® Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific, Waltham, MA, United States) according to the manufacturer’s protocol. Capillary electrophoresis was conducted on ABI 3500XL and ABI 3500 genetic analyzers (Thermo Fisher Scientific, Waltham, MA, United States). The obtained sequences were compared against RefSeq Gene references from the National Center for Biotechnology Information database.

Next generation sequencing: The target regions of the 83 genes were analyzed. Hybridization probes for the next generation sequencing (NGS) panel were created with the Hyper Design Tool (Roche Diagnostics, Indianapolis, IN, United States) to cover all coding and splice regions of the listed genes. Library preparation followed the KAPA HyperPlus Kit protocol, using a 15-minute DNA fragmentation to yield 350 bp fragments. Target enrichment used KAPA HyperCap probes (Roche), and sequencing was performed on the MiSeq platform (Illumina Inc., San Diego, CA, United States) with V3 chemistry (600 cycles, paired-end). Each run generated an average of 51.3 million reads, 88% above Q30, with 99% of target regions covered at least 15 times and a mean read depth of 100 ×.

Bioinformatic analysis: Bioinformatic analysis was performed according to the guidelines of Genome Analysis Toolkit Best Practices (https://gatk.broadinstitute.org). Briefly, raw reads were trimmed with Trimmomatic (version 0.39), and sequence alignment was performed with bwa-mem² (version 2.3) using the GRCh38 genome assembly as a reference. Next, duplicate reads were marked using Picard tools, and base quality score recalibration was performed. Then, genetic variations (SNPs and indels) were identified with GATK HaplotypeCaller (version 4.4). Next, gene annotation was performed using an in-house script to identify variations present in the ClinVar, OMIM, and HGMD databases. All sequence variants obtained from Sanger sequencing and NGS with minor allele frequency of < 1% according to the Genome Aggregation Database (gnomAD v4.1.2) (available at: https://gnomad.broadinstitute.org) were subjected to bioinformatics analysis in the program “Alamut Visual” (Interactive Biosoftware). In-silico tools SIFT, PolyPhen HDIV, PolyPhen HVAR, Mutation Taster, FATHMM, CADD13, DANN, M-CAP, and REVEL, as well as Alamut with the built-in software modules, were used for the prediction of the pathogenicity of the missense sequence variants. The guidelines of the American College of Medical Genetics and Genomics manual were used for variant interpretation[14].

Definitions and outcomes

The following indicators were assessed in all patients: (1) Demographic: Gender, onset age, age at the time of sJIA diagnosis, age at the time of verification of TRAPS diagnosis; hereditary history of recurrent fevers, presence/absence of surgical interventions related to abdominal pain; (2) Clinical manifestations: Duration, frequency, and characteristics of fever, rash characteristics, presence of hepatomegaly, splenomegaly, lymphadenopathy, joint syndrome, eye involvement, myalgia, serositis, abdominal and chest pain, nervous system damage, hearing loss, and periorbital edema; clinical and endoscopic signs of gastrointestinal tract involvement; (3) Laboratory parameters: Hemoglobin level, number of red blood cells, white blood cells, neutrophils, platelets, erythrocyte sedimentation rate (ESR), activity of alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), concentration of ferritin and C-reactive protein (CRP) in blood serum; functional activity was assessed using a questionnaire, Childhood Health Assessment Questionnaire (CHAQ; 0-3)[15]. Disease activity in all patients before the TRAPS diagnosis was assessed by the juvenile arthritis disease activity score 71 (JADAS71) index, and after the TRAPS diagnosis was verified, by the AID activity index (AIDAI)[16]; (4) Methods of genetic test: High-throughput sequencing of a 15-gene panel (MVK, MEFV, TNFRSF1A, NLRP3, NLRP12, CECR1, IL1RN, TMEM173, NLRC4, LACC1, NOD2, PSTPIP, LPIN2, PSMB8, TRNT), n = 22; high-throughput sequencing of an 83-gene panel (ABCB1, ABCG2, ACP5, ADAM17, ADAR, AP1S3, AP3B1, ATG16 L1, BLOC1S6, CARD14, CD27, CECR1, COPA, DDX58, EGFR, ELANE, GPBAR1, IFIH1, IL10, IL10RA, IL10RB, IL12B, IL18, IL18BP, IL1RN, IL21, IL23A, IL23R, IL36RN, IL6, IRF5, IRGM, ISG15, ITK, LACC1, LPIN2, LYN, LYST, MEFV, MVK, NFAT5, NLRC4, NLRP1, NLRP12, NLRP3, NOD1, NOD2, OTULIN, PLCG2, POMP, PRF1, PSMA3, PSMB4, PSMB8, PSTPIP1, PTGER4, RAB27A, RBCK1, RNASEH2A, RNASEH2B, RNASEH2C, SAMHD1, SH2D1A, SH3BP2, SLC22A5, SLC29A3, SLC2A9, SLC7A7, STAT3, STAT4, STX11, STXBP2, TMEM173, TNF, TNFAIP3, TNFRSF1A, TNFRSF1B, TNFSF14, TNFSF15, TREX1, TTC7A, UNC13D, XIAP), n = 31. Sanger sequencing (analysis of all coding and adjacent intronic regions of a single copy of the TNFRSF1A gene), n = 1. Sanger sequencing (partial analysis of the TNFRSF1A gene), n = 12; and (5) Treatment history: The presence and duration of biological (b) and non-biological (nb) disease-modifying antirheumatic drugs (DMARDs), glucocorticoids (GC), and intravenous immunoglobulin (IVIG) before and after diagnosis of TRAPS.

The treatment efficacy before TRAPS verification was assessed using the criteria for inactive disease/remission for both patients with sJIA, as defined by C. Wallace criteria. Drug remission was recorded while maintaining inactive disease for 6 consecutive months during therapy. The effectiveness of therapy after TRAPS verification and after 3, 6 and 12 months of therapy was evaluated based on criteria for achievement of inactive disease/remission: (1) Absence of clinical manifestations; (2) Absence of laboratory activity: Reduction of ESR < 20 mm/hour and CRP < 5 mg/L; (3) Absence of disease activity (AIDAI < 9 points); and (4) Absence of disease [physician’s visual analogue scale (VAS) ≤ 10 mm].

The criteria for treatment response included: (1) Achievement of clinical and laboratory remission; (2) Partial response meant persistence of low-grade inflammation (increased ESR from 20 to 30 mm/hour and CRP from 5 to 10 mg/L) or clinical disease symptoms; and (3) The lack of response mentioned the persistence of clinical and laboratory activity of the disease.

Statistical analysis

R-Studio software was used for data analysis. For quantitative features, the distribution of variable values across the observation groups was checked using the Shapiro-Wilk test. The sample was considered normal when P > 0.05. Quantitative indicators are described using the median and interquartile range; qualitative indicators are reported as absolute n (%). For paired comparisons of quantitative variables between groups in the absence of a normal distribution, the nonparametric Mann–Whitney criterion was used; for paired comparisons of quantitative variables within the same group (at the time of the onset of the disease and at the moment of diagnosis of rheumatic disease), a nonparametric Wilcoxon criterion was used. Fischer’s exact test was used for paired intergroup comparisons of qualitative variables. To compare clinical manifestations within the same group at disease onset and at rheumatic disease diagnosis, the McNemar test with Edwards correction was used. The differences were considered statistically significant at P < 0.05.

Ethics

This study was conducted in accordance with the Declaration of Helsinki. The Ethics Committee of the I.M. Sechenov First Moscow State Medical University (Sechenov University) approved the study protocol (No. 12-25, 22 May 2025). Written informed consent was obtained from all patients or their legal guardians (for patients aged < 15 years). All patient data were anonymized to protect confidentiality.

RESULTS

Genetic testing

Genetic testing identified a wide variety of TNFRSF1A gene variants in 33/66 (50%) patients, among whom 27 (82%) had a variant with low penetrance-c.362G>A. In 4 patients (12.1%), there was a likely pathogenic variant, in 3 (9.1%)-a variant of uncertain significance (VUS); in 1 (3%)-a combination of VUS and a pathogenic variant, in 1/33 (3%)-a variant with incomplete penetrance and VUS. TNFRSF1A gene variants were detected by NGS using a panel of 83 genes in 12/33 (36.3%) patients, a panel of 15 genes in 9 (27.2%), by Sanger sequencing with partial analysis of the TNFRSF1A gene in 1/33 (3%), or by Sanger sequencing with the full analysis of the TNFRSF1A gene in 11/33 (33.3%). The data are presented in Table 1[17-22].

Table 1 Variants of the TNFRSF1A gene identified based on molecular genetic research data, n (%).

Number of patients (n = 33)	Number of options (n = 35)	Nucleotide and protein variants of the gene TNFRSF1A (NM_001065.4)	Population frequency of the minor allele, gnomAD v.4.1.0	Description in the literature in patients with AIDs
1 (3)	1 (2.9)	c.337_339del, p.Glu113del in a heterozygous state	No data	Not described
1 (3)	1 (2.9)	c.362G>A, p.Arg121Gln in a homozygous state	1.5%	[17-19]
24 (73)	24 (68.3)	c.362G>A, p.Arg121Gln in a heterozygous state	1.5%	[17-19]
1 (3)	1 (2.9)	c.792del, p.Lys265Serfs*87 in a heterozygous state	No data	Not described
1 (3)	2 (5.7)	c.362G>A, p.Arg121Gln in a heterozygous state	1.5%	[17-19]
		c.242G>T, p.Cys81Phe in a heterozygous state	No data	Not described
1 (3)	1 (2.9)	c.596T>C, p.Ile199Thr in a heterozygous state	0.006%	[20,21]
1 (3)	1 (2.9)	c.374G>A, p.Cys125Tyr in a heterozygous state	No data	[3]
1 (3)	1 (2.9)	c.184T>A, p.Cys62Ser in a heterozygous state	No data	Not described
1 (3)	2 (5.7)	c.472+6С>Т, p.? in a heterozygous state	0.003%	[3]
		c.362G>A, p.Arg121Gln in a heterozygous state	1.5%	[17-19]
1 (3)	1 (2.9)	c.194-14G>A, p.? in a heterozygous state	0.02%	[22]

AID: Autoinflammatory disease.

Based on the presence/absence of variants in the TNFRSF1A gene, patients were divided into two groups: Group 1 (TRAPS) included 33/66 (50%) patients with variants in the TNFRSF1A gene, for whom the diagnosis of sJIA was replaced with TRAPS after molecular genetic testing; group 2 (sJIA) included 33/66 (50%) patients without variants in the TNFRSF1A gene, who continued observation with the sJIA diagnosis. The median age at diagnosis verification in patients in group 1 was 8.9 (5.07-11.2) years; the duration from onset to diagnosis verification was 1.55 (0.6-5.1) years.

Patients’ characteristics at disease onset

In both groups, girls predominated-21/33 (63.6%). At the time of disease onset, the median age was 5.3 (1.6-7.6) years and 4.3 (2.6-6.7) years, respectively. A family history of periodic febrile syndromes and appendectomy was found in 3% and 12.1% patients with TRAPS, respectively. Fever was observed in 98.5% patients of the total cohort and 97% of patients with TRAPS: Subfebrile-only in 15.2% patients with TRAPS, febrile-in 75.8% and 81.9%, pyretic-6.1% and 18.1% patients with TRAPS and sJIA, respectively. Episodes of fever were present only in patients with TRAPS (48.5%); continuous in 48.5% and 100% with TRAPS and sJIA, respectively (P = 0.000). The duration of fever in patients with TRAPS was 5-6 days (45.4%) and > 7 days (48.5%). Rash was observed in 55/66 (83.3%) patients of the total cohort: In carriers of the TNFRSF1A gene variants, rash was significantly less frequent (66.7%) than in the sJIA group (100%), P = 0.000: Maculopapular-significantly more often in patients diagnosed with sJIA than with TRAPS, P = 0.000; small-spotted, large-spotted and vesicular-only in carriers of the TNFRSF1A gene variants. Joint damage was observed in 65.1% of patients in the total cohort: In carriers of TNFRSF1A variants, significantly more often (78.8%) than in patients diagnosed with sJIA (51.5%), P = 0.03. Hepatomegaly (60.6%), lymphadenopathy (45.5%), and splenomegaly (42.4%) occurred approximately equally in both the TRAPS and sJIA groups and did not differ significantly between the groups. Nervous system damage was found in 39.4% patients of the total cohort, where it was more frequently observed in children with TNFRSF1A gene variants than in patients with sJIA (54.5% vs 24.2%, P = 0.02). Headache was observed in 42.4% and 24.2% patients with TRAPS and sJIA, respectively, but delay in psychomotor development (6.1%), hydrocephalus (3%), aseptic meningitis (3%), and hearing loss (3%) were observed only in patients with TNFRSF1A gene variants. Myalgia was recorded in 25.8% of patients in the total cohort, significantly more often in carriers of TNFRSF1A gene variants (42.4% vs 9.1%, P = 0.004). Serositis was diagnosed in 9.7% of patients in the total cohort, and it occurred approximately equally in both the TRAPS and sJIA groups and did not differ significantly between the groups. Patients with TRAPS infrequently, but specifically, experienced eye lesions (15.1%), periorbital edema (9.1%), and aphthous stomatitis (6.1%). Clinical manifestations of gastrointestinal damage were observed in 16.7% patients of the total cohort, more common in the TRAPS group (27.2% vs 6.1%, P = 0.04); liquid stools with blood (6.1%) and vomiting combined with liquid stools (3%) were observed only in patients with TNFRSF1A gene variants. MAS in carriers of TNFRSF1A gene variants was observed significantly less frequently than in sJIA (3% vs 21.2%, P = 0.05). The data are presented in Table 2.

Table 2 Demographic and clinical characteristics of patients with gene variants TNFRSF1A and systemic juvenile idiopathic arthritis at disease onset and at the time of diagnosis of systemic juvenile idiopathic arthritis, n (%)/median (interquartile range).

Indicator	At onset				At time of sJIA diagnosis
	Total group (n = 66)	Patients with gene variants TNFRSF1A(n = 33)	Patients with sJIA (n = 33)	P value	Total group (n = 66)	Patients with gene variants TNFRSF1A(n = 33)	Patients with sJIA (n = 33)	P value
Boys	24 (36.4)	12 (36.4)	12 (36.4)	1.0	24 (36.4)	12 (36.4)	12 (36.4)	1.0
Girls	42 (63.6)	21 (63.6)	21 (63.6)	1.0	42 (63.6)	21 (63.6)	21 (63.6)	1.0
Age at the time of disease onset (years)	4.73 (2.0-7.4)	5.3 (1.6-7.6)	4.3 (2.6-6.9)	0.97	5.8 (3.39.3)	6.1 (3.49.0)	4.8 (3.27.7)	0.7
Duration of period between onset and verification of the diagnosis of rheumatic disease (years)					0.22 (0.1-0.6)	0.33 (0.2-0.8)	0.18 (0.1-0.4)	0.08
Burdened heredity for periodic febrile syndromes	1 (1.5)	1 (3)	0 (0)	1.0	1 (1.5)	1 (3)	0 (0)	1.0
History of appendectomy	2 (3)	2 (6.1)	0 (0)	0.49	4 (3)	4 (12.1)	0 (0)	0.11
Fever	65 (98.5)	32 (97)	33 (100)	1.0	61 (92.4)	33 (100)	28 (84.8)	0.07
Fever characteristics
Subfebrile (37.1 °C-38.0 °C)	5 (7.6)	5 (15.1)	0 (0)	0.053	4 (6.1)	1 (3)	3 (9.1)	0.48
Febrile (38.1 °C-39.0 °C)	52 (78.8)	25 (75.8)	27 (81.9)	0.76	51 (77.2)	28 (84.8)	23 (69.6)	0.24
Pyretic (39.1 °C-40 °C)	8 (12.1)	2 (6.1)	6 (18.1)	0.25	6 (9.1)	4 (12.1)	2 (6.1)	0.43
Number of attacks
Once a month	1 (1.5)	1 (3)	0 (0)	1.0	0 (0)	0 (0)	0 (0)	1.0
2 times a month	4 (6.1)	4 (12.1)	0 (0)	0.11	2 (3)	2 (6.1)	0 (0)	0.49
3 times a month	4 (6.1)	4 (12.1)	0 (0)	0.11	23 (34.8)	23 (69.7)	0 (0)	0.0001
> 4 times a month	7 (10.7)	7 (21.2)	0 (0)	0.01	5 (7.6)	5 (15.1)	0 (0)	0.053
Continuous fever	49 (74.2)	16 (48.5)	33 (100)	0.0001	36 (54.6)	3 (9.1)	28 (100)	0.0001
Duration of fever
1-3 days	1 (1.5)	1 (3)	0	1.0	0 (0)	0 (0)	0 (0)	1.0
5-6 days	15 (22.7)	15 (45.5)	0	0.0001	18 (27.2)	18 (54.5)	0 (0)	0.0001
> 7 days	49 (74.2)	16 (48.5)	33 (100)	0.0001	48 (72.8)	15 (45.5)	28 (100)	0.002
Rash	55 (83.3)	22 (66.7)	33 (100)	0.0001	52 (78.8)	19 (57.6)	29 (87.8)	0.01
Rash characteristics
Maculopapular	43 (65.1)	10 (30.3)	33 (100)	0.0001	41 (62.1)	8 (24.2)	29 (87.8)	0.0001
Small-spotted	7 (10.6)	7 (21.2)	0 (0)	0.01	6 (9.1)	6 (18.1)	0 (0)	0.02
Large-spotted	4 (6.1)	4 (12.1)	0 (0)	0.11	4 (6.1)	4 (12.1)	0 (0)	0.11
Vesicular	1 (1.5)	1 (3)	0 (0)	1.0	1 (1.5)	1 (3)	0 (0)	1.0
Joint syndrome	43 (65.1)	26 (78.8)	17 (51.5)	0.03	47 (71.2)	27 (81.8)	20 (60.6)	0.1
Characteristics of joint syndrome
Oligoarthritis	29 (44)	18 (54.5)	11 (33.3)	0.21	33 (50)	18 (54.5)	15 (45.5)	0.47
Polyarthritis	14 (21.2)	8 (24.2)	6 (18.1)	0.76	14 (21.2)	9 (27.2)	5 (15.1)	0.36
Hepatomegaly	40 (60.6)	21 (63.6)	19 (57.6)	0.8	54 (81.8)	23 (69.7)	31 (94)	0.025
Lymphadenopathy	30 (45.5)	17 (51.5)	13 (39.4)	0.45	37 (56)	18 (54.5)	19 (57.6)	1.0
Splenomegaly	28 (42.4)	14 (42.4)	14 (42.4)	1.0	37 (56)	16 (48.5)	21 (63.6)	0.32
Nervous system lesion	26 (39.4)	18 (54.5)	8 (24.2)	0.02	18 (27.2)	14 (42.4)	4 (12.1)	0.01
Characteristics of neurological symptoms
Headache	22 (33.3)	14 (42.4)	8 (24.2)	0.19	15 (22.7)	11 (33.3)	4 (12.1)	0.07
Aseptic meningitis	1 (1.5)	1 (3)	0 (0)	1.0	0 (0)	0 (0)	0 (0)	1.0
Hydrocephalus	1 (1.5)	1 (3)	0 (0)	1.0	1 (1.5)	1 (3)	0 (0)	1.0
Delay in psycho-speech development	2 (3)	2 (6.1)	0 (0)	0.49	2 (3)	2 (6.1)	0 (0)	0.49
Hearing loss	1 (1.5)	1 (3)	0 (0)	1.0	3 (4.5)	3 (9.1)	0 (0)	0.23
Myalgia	17 (25.8)	14 (42.4)	3 (9.1)	0.004	18 (27.2)	17 (51.5)	1 (3)	0.0001
Pericarditis	10 (15.1)	5 (15.1)	5 (15.1)	1.0	12 (18.2)	7 (21.1)	5 (15.1)	0.75
Pleurisy	3 (4.5)	2 (6)	1 (3)	1.0	4 (6.1)	2 (6.1)	2 (6.1)	1.0
Abdominal pain	8 (12.1)	6 (18.1)	2 (6.1)	0.25	10 (15.1)	10 (30.3)	0 (0)	0.001
Chest pain	4 (6)	4 (12.1)	0	0.11	4 (6.1)	4 (12.1)	0 (0)	0.11
Eye afflictions	5 (7.5)	5 (15.1)	0 (0)	0.053	8 (12.1)	8 (24.2)	0 (0)	0.004
Characteristics of eye damage
Unilateral uveitis	1 (1.5)	1 (3)	0 (0)	1.0	5 (7.5)	5 (15.1)	0 (0)	0.053
Bilateral uveitis	3 (4.5)	3 (9.1)	0 (0)	0.23	1 (1.5)	1 (3)	0 (0)	1.0
Retinal angiopathy	1 (1.5)	1 (3)	0 (0)	1.0	2 (3)	2 (6.1)	0 (0)	0.49
Periorbital edema	3 (4.5)	3 (9.1)	0 (0)	0.23	4 (6.1)	4 (12.1)	0 (0)	0.11
Aphthous stomatitis	2 (3)	2 (6.1)	0 (0)	0.49	2 (3)	2 (6.1)	0 (0)	0.49
Clinical manifestations of gastrointestinal tract lesions	11 (16.7)	9 (27.2)	2 (6)	0.04	8 (12.1)	8 (24.2)	0 (0)	0.004
Clinical manifestations of gastrointestinal tract lesions
Liquid stool	5 (7.5)	4 (12.1)	1 (3)	0.35	3 (4.5)	3 (9.1)	0 (0)	0.23
Liquid stool with an admixture of blood	2 (3)	2 (6.1)	0 (0)	0.49	2 (3)	2 (6.1)	0 (0)	0.49
Vomiting	3 (4.5)	2 (6.1)	1 (3)	1.0	2 (3)	2 (6.1)	0 (0)	0.49
Vomiting combined with liquid stool	1 (1.5)	1 (3)	0 (0)	1.0	1 (1.5)	1 (3)	0 (0)	1.0
Changes in gastrointestinal endoscopic examination data					36 (54.5)	20 (60.6)	16 (48.5)	0.45
Changes in gastrointestinal endoscopic examination data
Gastritis	17 (25.8)	9 (27.2)	8 (24.2)	1.0	3 (4.5)	3 (9.1)	0 (0)	0.49
Erosive gastritis					1 (1.5)	1 (3)	0 (0)	1.0
Erosive esophagitis + duodenitis					1 (1.5)	0 (0)	1 (3)	1.0
Erosive esophagitis + bulbitis					1 (1.5)	0 (0)	1 (3)	1.0
Erosive proctitis + erosive gastritis + duodenitis + jejunit					1 (1.5)	1 (3)	0 (0)	1.0
Erosive ileitis + erosive gastritis + duodenitis					1 (1.5)	1 (3)	0 (0)	1.0
Erosive sigmoiditis + gastritis + jejunitis					1 (1.5)	1 (3)	0 (0)	1.0
Gastritis + bulbitis	1 (1.5)	1 (3)	0 (0)	1.0	1 (1.5)	0 (0)	1 (3)	1.0
Gastritis + duodenitis + jejunitis					3 (4.5)	3 (9.1)	0 (0)	0.23
Gastritis + bulbitis + duodenitis					4 (6.1)	0 (0)	4 (12.1)	0.11
Gastritis + duodenitis					11 (16.7)	3 (9.1)	8 (24.2)	0.185
Bulbitis + duodenitis	1 (1.5)	1 (3)	0 (0)	1.0	1 (1.5)	1 (3)	0 (0)	1.0
Bulbitis					2 (3)	2 (6.1)	0 (0)	0.49
Duodenitis	2 (3)	1 (3)	1 (3)	1.0	2 (3)	1 (3)	1 (3)	1.0
Sigmoiditis					3 (4.5)	3 (9.1)	0 (0)	0.23
MAS	8 (12.1)	1 (3)	7 (21.2)	0.052	15 (22.7)	7 (21.2)	8 (24.2)	1.0
Amyloidosis	0 (0)	0 (0)	0 (0)	1.0	1 (1.5)	1 (3)	0 (0)	1.0

MAS: Macrophage activation syndrome; sJIA: Systemic juvenile idiopathic arthritis.

At disease onset, both groups showed increased inflammatory markers. In patients diagnosed with sJIA, the number of leukocytes, neutrophils, and platelets in the blood, as well as the concentration of ferritin in the blood serum, were significantly higher than in carriers of TNFRSF1A gene variants, and the hemoglobin level was significantly lower. The ESR indicator in both groups exceeded the reference values and did not significantly differ between groups. Serum CRP concentrations in carriers of TNFRSF1A gene variants were significantly lower than in sJIA patients and amounted to 47.5 (26.7-68) mg/L and 84.7 (60-128) mg/L, respectively (P = 0.001). The data are presented in Table 3.

Table 3 Laboratory parameters of patients with TNFRSF1A gene variants and systemic juvenile idiopathic arthritis at disease onset the and at the time of systemic juvenile idiopathic arthritis diagnosis, median (interquartile range).

Laboratory indicators	Normal values	Total group (n = 66)	Patients with gene variants TNFRSF1A(n = 33)	Patients with sJIA (n = 33)	P value	Total group (n = 66)	Patients with gene variants TNFRSF1A(n = 33)	Patients with sJIA (n = 33)	P value
Red blood cells, (million/mL)	4-5.2	4.0 (3.6-4.2)	4.0 (3.8-4.2)	3.9 (3.4-4.0)	0.06	4.4 (3.9-4.8)	4.27 (3.8-5.01)	4.4 (4.01-4.7)	0.97
Hemoglobin (g/L)	115-145	100 (94-104)	101 (95-106)	98 (90-102)	0.03	110 (101-119.2)	108 (101-118)	112 (100-120)	0.45
Leukocytes (thousand/mL)	4.5-11.5	17.9 (12.3-22.7)	12.6 (11.2-18.8)	20.3 (17.8-24.2)	0.0001	12.7 (8.9-17)	11.2 (8.9-15.2)	13.5 (9.9-18.6)	0.27
Neutrophils (thousand/mL)	1.1-9.9	14.5 (8.5-18.2)	9.1 (6.7-14.6)	17.6 (14.5-19.1)	0.0001	8.8 (4.8-12.2)	7.6 (4.3-10.2)	9.9 (5.7-13.2)	0.13
Platelets (thousand/mL)	150-450	456 (353.3-599)	415 (334-456)	556 (456-660)	0.0001	410 (340-521.5)	372 (316-469)	425 (353-528)	0.19
ESR (mm/hour)	0-20	53 (37-61)	45 (35-60)	55 (43-60)	0.2	34.5 (15-52)	34 (15-52)	35 (24-52)	0.93
СRP (mg/L)	< 5	66.5 (34.1-115.7)	47.5 (26.7-68)	84.7 (60-128)	0.001	37.8 (10.5-102)	43 (7.8-102.4)	40.5 (11.2-92.8)	0.86
ALT (U/L)	< 40	24 (20-34)	23 (20-30)	24 (20-46)	0.31	16 (11.8-24.5)	19 (11-26)	15 (12-24)	0.89
AST (U/L)	< 42	27.5 (23-36)	26 (20-35)	30 (24-40)	0.19	25.5 (22-33.2)	28 (23-37)	24 (20-30)	0.1
LDH (U/L)	91-295	243.5 (203-289)	228 (167-284)	246 (225-289)	0.09	243 (198-301.5)	237 (173-301)	247 (200-298)	0.46
Ferritin (ng/mL)	12-84	113.5 (75.7-259)	102.3 (67.6-227.4)	167.9 (98.6-357.9)	0.03	120.3 (53.3-473.2)	129.3 (54.2-349.2)	118.9 (65.7-450.4)	0.65
Physician's VAS, (mm)	0-100					56.5 (33.8-70)	65 (48-76)	74 (56-80)	0.005
Parent/patient's VAS (mm)	0-100					64.5 (44-74.3)	70 (54-78)	78 (40-70)	0.26
CHAQ	0-3					0.75 (0.125-1)	0.75 (0.5-1.5)	0.5 (0.125-0.75)	0.002
JADAS71						17.4 (13-22.6)	16.2 (12-21.5)	18.6 (15.5-25)	0.37

sJIA: Systemic juvenile idiopathic arthritis; ESR: Erythrocyte sedimentation rate; CRP: C-reactive protein; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase; LDH: Lactate dehydrogenase; VAS: Visual analogue scale; CHAQ: Childhood Health Assessment Questionnaire; JADAS71: Juvenile arthritis disease activity score 71.

Drug therapy prescribed after the onset and before the diagnosis of sJIA in patients with TNFRSF1A gene variants

Before the diagnosis of rheumatic disease, all patients were prescribed non-steroidal anti-inflammatory drugs (NSAIDs), antibiotics were prescribed to 75.8% of patients, oral glucocorticosteroids (GCS) to 30.3%, intravenous GCS to 39.3%, intramuscular GCS to 6.1%, and intra-articular GCS to 6.1%. IVIG was used in 21.2% of patients, and azathioprine and mesalazine were used by 6.1% of patients.

Characteristics of patients at the time of sJIA diagnosis

Fever at the time of sJIA diagnosis was observed in 92.4% of patients in the overall cohort; the periodicity of fever attacks was observed only in patients with the TNFRSF1A gene variants: Lasting up to 5-6 days in 54.5% and > 7 days in 45.5% children; in 84.8% children, the second group with sJIA had continuous fever. In patients with TRAPS and sJIA, most had febrile fever (84.8% and 69.6%, respectively), less often subfebrile fever (3% in the TRAPS group), and pyretic fever (12.1% in the TRAPS and 6.1% in the sJIA group). Rash was registered in 48/66 (72.7%) patients of the total cohort, significantly less frequently in the TRAPS group than in patients with sJIA (57.6% vs 87.8%, P = 0.01): Maculopapular rash was still significantly more common in patients with sJIA (24.2% vs 87.8%, P = 0.000); small-spotted, large-spotted, and vesicular rashes were observed in 18.1%, 12.1%, and 3% of carriers of the TNFRSF1A gene variants. Articular involvement was present in 71.2% of patients in the overall cohort: 81.8% in the TRAPS group and 60.6% in the sJIA, primarily oligoarthritis. Both at the onset and at the time of diagnosis of rheumatic disease, articular syndrome was observed predominantly in the TNFRSF1A gene variant carriers compared with patients diagnosed with sJIA. Hepatomegaly was detected in 81.8% patients of the overall cohort: In patients diagnosed with TRAPS, it was significantly less common than in those with sJIA (69.7% vs 94%, P = 0.025); lymphadenopathy (54.5%) and splenomegaly (57.6%) were detected at similar frequencies in both patients with TRAPS and sJIA. Nervous system damage at the time of sJIA diagnosis was observed in 27.2% of patients in the total cohort, more often in carriers of TNFRSF1A variants than in patients with sJIA (42.4% vs 12.1%, P = 0.01). Headache was observed in 33.3% and 12.1% patients with TRAPS and sJIA, respectively, and less frequently in both groups compared to the onset; aseptic meningitis, registered at the time of clinical manifestations onset in 1 (3%) carrier of the variant in the TNFRSF1A gene, was stopped by the time of diagnosis of rheumatic disease, P = 1. The distribution of such signs as hydrocephalus and delays in psycho-speech development did not differ from the indicators at the onset. Myalgia was observed in 27.2% patients of the total cohort, 51.5% of patient with TRAPS and 3% of patients with sJIA, significantly mostly at the time of diagnosis, as well as at the onset of disease in patients in the TRAPS group. Serositis was detected in 24.2% of patients in the total cohort (19.7% at the onset). Complaints of abdominal (15.1%) and chest (6.1%) pain were reported only in patients with TNFRSF1A gene variants; eye damage increased from 7.5% to 12.1% in TRAPS. Aphthous stomatitis persisted in 6.1% of patients in the TRAPS group. Clinical manifestations of gastrointestinal damage were observed in 12.1% of patients in the total cohort and, unlike the onset, were noted only in carriers of TNFRSF1A gene variants (P = 0.004). MAS development at the time of rheumatic disease diagnosis was observed in 21.2% of patients with TRAPS, more often than at disease onset (P = 0.038). Amyloidosis at the time of diagnosis of rheumatic disease was detected in 1/33 (3%) patients with TRAPS, P = 1 (Table 2).

Compared with onset of the disease, in patients with TRAPS, the hemoglobin level significantly increased (P = 0.04), while the number of neutrophils decreased (P = 0.02); in patients with sJIA, the number of erythrocytes significantly increased (P = 0.000), while the number of leukocytes, neutrophils, and platelets in the blood significantly decreased (P = 0.000; P = 0.0001; P = 0.001, respectively). The blood hemoglobin level was slightly reduced; the Me activities of AST, ALT, and LDH in the blood serum corresponded to reference values, and the ferritin concentration exceeded them. Hyperferritinemia, significant increases in serum AST, ALT, and LDH activity, and a significant decrease in hemoglobin were observed in 31.8% of patients in the total cohort. The median ESR values and serum CRP concentrations exceeded the reference values in patients with TRAPS and sJIA, respectively, and did not differ significantly (P = 0.93 and P = 0.86, respectively). Compared with the onset of disease, in the TRAPS group, the median hemoglobin level significantly increased (P = 0.04), the number of neutrophils, ESR and ALT activity decreased (P = 0.02; P = 0.004; P = 0.04, respectively); in patients with sJIA, the median of the number of erythrocytes and hemoglobin level in the blood significantly increased; the median number of leukocytes, neutrophils (P = 0.000) and platelets in the blood (P = 0.001), the ESR (P = 0.001), the concentration of CRP (P = 0.005) and ALT activity in the blood serum significantly decreased (P = 0.006). All patients had high disease activity according to the physician and the JADAS71 index, and poor health according to the parent/patient’s VAS. Functional impairment, as measured by the CHAQ index, was 0.75 and 0.5 points, respectively (P = 0.002). The VAS and JADAS71 scores in patients with sJIA were higher than those in carriers of the TNFRSF1A gene variants. The data are presented in Table 3.

Endoscopic examination of patients at the time of sJIA diagnosis

The incidence of gastrointestinal tract lesions according to endoscopic examination data at the time of rheumatic disease diagnosis was 60.6% and 48.5% in patients with TRAPS and sJIA, respectively. In patients with sJIA, only lesions of the upper intestine (esophagitis, gastritis, duodenitis, and their combinations) were recorded; colon damage was recorded exclusively in patients with TRAPS (27.2%) and none in patients with sJIA (P = 0.001). The data are presented in Table 2.

Drug therapy for patients with TNFRSF1A gene variants prescribed after sJIA diagnosis

After sJIA diagnosis, 81.9% patients in the first group received NSAIDs, antibacterial drugs-36.3%, oral GC-45.5%, intravenous GC (methylprednisolone, dexamethasone)-45.5%, intra-articular GC-6.1%, IVIG-18.1%; methotrexate-48.5%, cyclosporine-18.1%, azathioprine-3%, mesalazine-3%, sulfasalazine-3%.

Biological DMARDs were prescribed after sJIA diagnosis to 24/33 (72.7%) patients in the first group with TNFRSF1A gene variants. Tocilizumab was used in 13/24 (54.2%) patients, abatacept in 3/24 (12.5%), canakinumab in 2/24 (8.3%), infliximab in 2/24 (8.3%), adalimumab in 1/24 (4.2%), etanercept in 1/24 (4.2%), rituximab in 1/24 (4.2%), and certolizumab pegol in 1/24 (4.2%).

The remaining 9/33 (27.2%) patients with identified TNFRSF1A gene variants received immunosuppressants in combination with GC and IVIG: Methotrexate in combination with intravenous and oral GC-6.1%; methotrexate in combination with oral GC-6.1%; methotrexate in combination with cyclosporine-6.1%; methotrexate in combination with cyclosporine and intravenous GC-6.1%; methotrexate in combination with cyclosporine, intravenous and oral GC and IVIG-3%.

Efficacy of bDMARDs administered to patients with TNFRSF1A gene variants before verification of the TRAPS diagnosis

Remission against the background of first bDMARD use was recorded in 11/24 (45.8%) patients in the first group. With tocilizumab as first-line biological therapy, remission was observed in 7/11 (63.7%) patients with TNFRSF1A variants; with canakinumab in 2/11 (18.1%); with infliximab in 1/11 (9.1%); and with adalimumab in 1/11 (9.1%). An incomplete response was recorded in 3/24 (12.5%) patients: To tocilizumab therapy in 1/13 (7.7%), to rituximab therapy in 1/13 (7.7%), and to etanercept in 1/13 (7.7%). Lack of response was recorded in 10/24 patients (41.7%) with variants in the TNFRSF1A gene: During therapy with tocilizumab in 5/10 (50%), abatacept in 3/10 (30%), certolizumab pegol in 1/10 (10%), and rituximab in 1/10 (10%).

Taking into account the persistent disease activity, 8/13 (61.6%) patients in the first group were switched to a second biological drug: To canakinumab in 4/8 (50%), to adalimumab in 2/8 (25%), to etanercept in 1/8 (12.5%), to rituximab in 1/8 (12.5%); 5/13 (38.5%) continued treatment with the first biological drug (3/5 (60%) with an incomplete response, 2/5 (40%) with no response during therapy with tocilizumab) and subsequent escalation of the GC dose with an effect, which subsequently allowed the reduction of oral GC dose to a maintenance dose (0.3 ± 0.2 mg/kg/day).

With the use of the second bDMARD, remission was achieved in 5/8 (62.5%) patients of the first group: Those receiving canakinumab-3/5 (60%), adalimumab-1/5 (20%), rituximab-1/5 (20%); in 3/8 (37.5%), disease activity persisted (with the use of canakinumab-in 1/3, adalimumab-in 1/3 and etanercept-in 1/3 of patients).

Switching to a third bDMARD before TRAPS verification was successfully performed in 1/3 patients (33.3%) in the first group on canakinumab, achieving remission.

As a result of two empirical switches (three bDMARDs) before genetic verification of diagnosis, remission was achieved in 17/24 (70.8%) patients of the first group: Those receiving canakinumab-6/17 (35.3%), tocilizumab-7/17 (41.2%), adalimumab-2/17 (11.8%), infliximab 1/17 (5.8%) and rituximab - 1/17 (5.8%).

Disease activity persisted in 16/33 (48.5%) patients in the first group: In 9/16 (56.2%) “biologically naive” patients who received immunosuppressants, IVIG in combination with or without GC, or bDMARDs; and in 7/16 (43.8%) “biologically non-naive” patients who received bDMARDs.

To achieve remission in 17/33 (51.5%) patients of the first group before verification of TRAPS, an average of 0.8 ± 0.57 (min: 0; max: 2) empirical switching between bDMARDs (three bDMARDs) was required.

Drug therapy prescribed after verification of TRAPS diagnosis

Azathioprine and mesalazine were continued in 3% of patients with TNFRSF1A gene variants; methotrexate, 18.1%; cyclosporine, 15.1%. Colchicine was prescribed to 3% of patients; none received sulfasalazine.

After genetic verification of TRAPS diagnosis, all 33 patients with TNFRSF1A variants received bDMARDs. A total of 9/33 patients (27.2%) who did not receive biological therapy before TRAPS diagnosis verification (“biologically naive”) were prescribed bDMARDs based on the genetic testing: 3/9 (33.3%)-canakinumab, 3/9 (33.3%)-tocilizumab and 3/9 (33.3%)-etanercept; 17/33 (51.5%) patients continued to receive biological drugs, against which they developed remission; 5/33 (15.1%) patients continued to receive biological drugs in combination with GC at a low dose of 0.2 ± 0.1 mg/kg/day: Tocilizumab-3/5 (60%), canakinumab-2/5 (40%); 2/33 (6.1%) underwent biological therapy correction (1/2-on canakinumab switched to adalimumab, 1/2-on tocilizumab switched to canakinumab). No patients received abatacept or certolizumab pegol after verification of TRAPS diagnosis.

Subsequently, 3/33 (9.1%) patients required correction of bDMARDs: 1/33 (3%) patient on etanercept therapy with secondary inefficiency, switching to rituximab with secondary ineffectiveness, switching to tocilizumab with subsequent achievement of remission lasting 2 years and subsequent development of secondary inefficiency and switching to golimumab with achievement of remission; 1/33 (3%) patient on adalimumab therapy with switching to canakinumab-secondary inefficiency, switching to rituximab-secondary inefficiency, switching to tocilizumab with achievement of disease remission; 1/33 (3%) patient discontinued rituximab due to achievement of long-term remission, colchicine therapy was continued. To achieve remission after TRAPS verification, it took on average 0.62 ± 1.11 (min: 0; max: 3) empirical switching between bDMARDs (four bDMARDs).

Efficiency of biological therapy after TRAPS verification

Therapy effectiveness was assessed by achieving a complete response and inactive disease/remission at 3 and 6 months after the start of treatment. In total, before and after verification of TRAPS diagnosis, clinical and laboratory remission was registered in 100% of patients with TRAPS as a result of bDMARDs switches: During therapy with tocilizumab-in 13/33 (39.3%), during therapy with canakinumab-in 11/33 (33.3%), during therapy with adalimumab-in 3/33 (9.1%), during therapy with etanercept-in 2/33 (6.1%), during therapy with rituximab-in 1/33 (3%), during therapy with infliximab-in 1/33 (3%), during therapy with golimumab-in 1/33 (3%). Figure 2 demonstrates the effectiveness of bDMARDs in patients with TRAPS after diagnosis verification.

Figure 2 Efficacy of biological disease-modifying anti-rheumatic drugs in patients with tumor necrosis factor receptor-associated periodic syndrome as result of all switches. TRAPS: Tumor necrosis factor receptor associated periodic syndrome.

The dynamics of clinical manifestations in patients included in the study were assessed after the diagnosis of sJIA and therapy initiation, after verification of the AID and the appointment/correction of pathogenetic therapy. The effect was assessed at 3, 6, and 12 months after administration or continued use of bDMARDs. Three months after TRAPS verification, polyserositis resolved in all patients; after 6 months, fever, rash, lymphadenopathy, and myalgia resolved; after 12 months, articular syndrome resolved. The dynamics of clinical manifestations in patients with TRAPS are presented in Figure 3.

Figure 3 Dynamics of clinical manifestations of patients with tumor necrosis factor receptor-associated periodic syndrome (n = 33). ^aP < 0.05. TRAPS: Tumor necrosis factor receptor associated periodic syndrome; sJIA: Systemic juvenile idiopathic arthritis.

The dynamics of disease activity indicators were assessed after establishing sJIA diagnosis and initiating bDMARD therapy, and after verifying TRAPS diagnosis and changing/initiating biological therapy. Three months after JA diagnosis and biological therapy initiation, the median ESR and CRP in patients with TNFRSF1A gene variants statistically significantly decreased and amounted to 20 (10-38.5) mm/hour (P = 0.003) and 13 (0.4-28) mg/L (P = 0.05), respectively, but exceeded the reference values; a significant improvement in functional activity and a decrease in disease activity according to the physician were also recorded, as evidenced by a statistically significant decrease in the median CHAQ index and physician’s VAS to 0.5 (0.15-0.9) and 25 (12-35) points (P = 0.002 and P = 0.001). After 6 months, the median ESR indicator and CRP concentration exceeded the reference values but were significantly lower than the background, at 24 (20;42) mm/hour (P = 0.05) and 15 (6;36) mm/hour (P = 0.01), respectively. The functional activity of patients continued to improve, and disease activity according to the doctor decreased: The CHAQ index median and the physician’s VAS decreased to 0.5 (0.1-1) and 14 (5-29) points, respectively (P = 0.002 and P = 0.001, respectively). After 12 months, the median ESR and CRP concentration in patients with variants in the TNFRSF1A gene increased, exceeded the reference values and amounted to 22 (16-40) mm/hour and 12.5 (0.8-24) mg/L, but was significantly lower compared with background values (P = 0.05 and P = 0.01, respectively); the median CHAQ index and physician’s VAS in patients with variants in the TNFRSF1A gene increased to 0.5 (0.1-1.2) and 20 (8-34) points, respectively.

After verification of TRAPS diagnosis and initiation/correction of biological therapy, disease activity was assessed using the AIDAI index and the physician’s VAS score. After 3 months, the median AIDAI index and physician’s VAS significantly decreased and amounted to 5 (4-8) and 8 (5-10) points, respectively, P = 0.001; after 6 and 12 months, the AIDAI index corresponded to a low degree of disease activity and amounted to 5 (4-6) and 4 (3-5) points, respectively (P = 0.001). The median ESR and CRP concentrations significantly decreased 3 months after diagnosis verification, were consistent with reference values, and were 4 (2-6) mm/hour and 1.23 (0.45-2.48) mg/L (P = 0.001 and P = 0.001, respectively); they remained within the normal range for 6 and 12 months.

DISCUSSION

AIDs comprise a group of orphan monogenic conditions characterized by disturbances in innate immune system regulation, which in pediatric practice can present under the “mask” of juvenile arthritis with systemic onset, complicating differential diagnosis. In our study, before molecular genetic testing and identification of TNFRSF1A variants, patients were predominantly diagnosed with “juvenile arthritis with systemic onset” (82%), which is consistent with the study by Yasumura et al[23]. Molecular genetic testing revealed pathogenic and likely pathogenic variants of the TNFRSF1A gene in 15% of patients, and VUS in 85%, of which the c.362G > A variant was found in 82% of patients. Patients with VUS were diagnosed with probable TRAPS according to the current Eurofever/PRINTO criteria (2019), provided that at least two clinical symptoms are present[11]. In their study, Lachmann et al[3] analyzed the genetic variants of 158 TRAPS patients, and the c.362G>A variant was identified in 34% of patients. Carriers of the c.362G>A variant were more likely to have later disease onset and less pronounced clinical symptoms. According to a systematic review by Cudrici et al[6] (2020), the most common clinical manifestations in 209 patients with confirmed pathogenic variants of the TNFRSF1A gene were fever (96%), abdominal pain (70%), myalgia (69%), arthralgia (69%), erythematous rash (60%), acute conjunctivitis (37%), periorbital edema (28%), chest pain (33%), arthritis (22%), lymphadenopathy (16%), headache (13%), and hepato- and/or splenomegaly (4%), which is only partially consistent with the results of our study[6]. In our cohort of patients with TNFRSF1A gene variants, active joint syndrome (81.8%), hepatomegaly (69.7%), lymphadenopathy (54.5%), and rash (57.6%) were more frequently observed. Clinical symptoms such as myalgia (51.5%), uveitis (18.2%), abdominal pain (30.3%), chest pain (12.1%), and periorbital edema (12.1%) were significantly less common. A comparative analysis of disease course characteristics in patients with TNFRSF1A gene variants and sJIA demonstrated that several clinical manifestations can develop at disease onset, including fever, rash, and articular syndrome. However, we have identified features that are more characteristic of each disease. In carriers of TNFRSF1A gene variants, fever was paroxysmal in the overwhelming majority of cases, whereas it was continuous in sJIA. Rash developed significantly less frequently than in JIA (66.7% vs 100%) and was small-spotted, large-spotted, and vesicular, while it was maculopapular in sJIA. Articular syndrome developed significantly more often than in sJIA (78% vs 51.5%) and was manifested by oligo- and polyarthritis, while arthralgia predominated in sJIA,. In patients with variants of the TNFRSF1A gene, clinical manifestations were registered that were not observed in patients diagnosed with sJIA: Paroxysmal fever in 16/33 (48.5%) patients, abdominal pain in 10/33 (33.3%), manifestations of gastrointestinal damage in 9/33 (27.2%): Diarrhea with pathological impurities in 2/33 (6.1%), diarrhea without impurities in 4/33 (12.1%), vomiting in 2/33 (6.1%); eye damage (uveitis and retinal angiopathy) in 8/33 (24.2%), chest pain in 4/33 (12.1%), hearing loss in 3/33 (9.1%), periorbital edema in 3/33 (9.1%), delay in psychomotor development in 2/33 (6.1%), aseptic meningitis in 1/33 (3%) and hydrocephalus in 1/33 (3%) patient. Vuran and Berdeli[24] examined patients with clinical manifestations of FMF but without MEFV gene variants. As a result of genetic tests, two variants with low penetrance-c.362G>A and one pathogenic variant in the TNFRSF1A gene were identified among 40 patients; among these three patients, they presented with gastrointestinal symptoms (3/3), periodic fever (3/3), maculopapular rash (1/3), arthralgia (2/3), exudative pharyngitis (1/3) and lymphadenopathy (1/3). It is worth noting that the patients did not have myalgia, ocular damage, or periorbital edema, underscoring the phenotypic variability of patients with TNFRSF1A variants[24]. Given that abdominal pain and other clinical manifestations of gastrointestinal lesions in our study were observed only in patients carrying TNFRSF1A gene variants, we analyzed changes in esophago-gastro-duodenoscopy and colonoscopy data obtained in 75.5% of children. Inflammatory changes in the upper and lower gastrointestinal tract were detected in 60.6% of patients with variants of the TNFRSF1A gene: The most common combination was gastritis, duodenitis, and jejunitis (9.1%); gastritis and duodenitis (9.1%). Sigmoiditis and erosive changes in the gastrointestinal tract were observed only in carriers of the TNFRSF1A gene variants. An analysis of changes in the gastrointestinal tract based on endoscopic examination data from patients with TRAPS, as reported in the foreign and domestic literature, has not been previously conducted. Patients with AIDs and sJIA are characterized by an increase in the inflammatory markers ESR and serum CRP concentration[2]. Analysis of laboratory parameters showed that at disease onset, inflammatory markers (ESR, leukocyte and neutrophil counts, and serum CRP concentration) were significantly elevated above reference values in both carriers of TNFRSF1A gene variants and patients with sJIA. In our cohort, leukocyte, neutrophil, platelet, serum ferritin, and CRP levels were significantly higher in patients with sJIA than in carriers of TNFRSF1A gene variants, whereas ESR did not differ significantly between the groups. Biological therapy is a pathogenetic method of treatment for most AIDs. Despite this, certain difficulties may often arise when selecting therapy when pathogenetic drugs registered for the treatment of AIDs are not effective. In a study by Gaggiano et al[8] demonstrated that biological therapy was prescribed to only 61% of patients with TNFRSF1A gene variants and TRAPS and was effective in 73% of patients. IL-1 inhibitors (anakinra and canakinumab) were used in 86% of cases; the remaining patients received TNF inhibitors: Adalimumab, infliximab, or etanercept. GCs were prescribed as monotherapy to 75% of patients before initiating bDMARDs, with a complete response in 53%[8]. In our study, the most effective treatment for periodic syndrome associated with TNFRSF1A variants was the IL-6 Receptor inhibitor tocilizumab, which induced remission in 39.3% of patients. The IL-1β inhibitor canakinumab was less effective, with disease remission in 33.3% of patients; 21.2% of patients responded to TNF-α inhibitors. Currently, data on the efficacy of IL-6 inhibitors in patients with TRAPS are quite limited. Only two clinical cases have been previously published describing the effectiveness of tocilizumab therapy in patients with TRAPS as second and third line therapy[25,26]. In their multicenter study of 44 patients, Vitale et al[27] demonstrated a complete response to tocilizumab therapy in 2/3 of patients, with a partial response in one.

Limitations

The study’s retrospective design may introduce bias, and the exclusion of specific patient groups and the small cohort size may also introduce bias. A sample size of only 33 patients with TNFRSF1A variants and 33 sJIA patients may lack statistical power and may not be generalizable to broader populations. A small sample size increases the risk of type II errors, in which actual effects may go undetected. Missing data and a heterogeneous population (different genetic variants and different ages) could influence the study results. Different types of genetic testing may miss other genetic variants in different genes and do not account for gene interactions. Our study did not evaluate the confounding variables that could affect outcomes, such as comorbidities, variations in treatment protocols over time, or differences in healthcare access. These factors could skew the results and interpretations. A more comprehensive analysis that accounts for these confounders would provide more robust and reliable findings.

CONCLUSION

Analysis of data from patients with TNFRSF1A gene variants showed a wide variety of clinical symptoms in this patient cohort. Manifestations such as fever, rash, joint involvement, damage to the nervous system, and gastrointestinal changes according to endoscopic examination, as well as high laboratory activity, were more common in patients with TRAPS. The results of molecular genetic testing, combined with phenotypic characteristics, allow verification of the TRAPS diagnosis at an early stage, prompt initiation of pathogenetic therapy, and prevention of complications. Biological DMARDs are targeted therapy for most AIDs. The pathogenesis and role of inflammatory cytokines in patients with TNFRSF1A gene variants are well established, but standard therapies are not always effective. According to the results of our study, tocilizumab (39.3%), canakinumab (33.3%), and TNF-inhibitors (21.2%) allowed remission in children with variants in the TNFRSF1A gene. Drugs from the IL-6r inhibitor group have demonstrated effectiveness in patients with TRAPS, making them an additional option in cases of ineffectiveness and/or intolerance to the generally accepted treatment regimen with IL-1 inhibitors.