You are here

  • Home
  • Archive
  • Volume 10, Issue 4
  • Inflammatory biomarker analysis confirms reduced disease severity in heterozygous patients with familial Mediterranean fever

Article Text

Article menu

Download PDFPDF

Autoinflammatory disorders
Original research
Inflammatory biomarker analysis confirms reduced disease severity in heterozygous patients with familial Mediterranean fever
  1. Inès Elhani1,2,
  2. Stefan Backes3,
  3. Tilmann Kallinich4,5,6,
  4. Gayane Amaryan7,
  5. Alexandre Belot8,9,
  6. Rainer Berendes10,
  7. Thomas Berger11,
  8. Frank Dressler12,
  9. Dirk Foell13,
  10. Sabrina Fühner3,
  11. Arnd Giese14,
  12. Claas Hinze3,
  13. Anna Lisa Hitzegrad5,
  14. Gerd Horneff15,
  15. Annette Jansson16,
  16. Jens Klotsche17,
  17. Elke Lainka18,
  18. Tim Niehues19,
  19. Prasad Oommen20,
  20. Johannes-Peter Haas21,
  21. Christoph Rietschel22,
  22. Katerina Theodoropoulo23,
  23. Caroline Vinit24,25,
  24. Elisabeth Weissbarth-Riedel26,
  25. Véronique Hentgen1 and
  26. Helmut Wittkowski3,27
  1. 1Department of General Pediatrics, Hospital Centre Versailles, Le Chesnay, France
  2. 2Department of Internal Medicine, Hôpital Tenon, Paris, France
  3. 3Pediatric Rheumatology and Immunology, University Medicine, Albert-Schweitzer-Campus 1, Munster, Germany
  4. 4Pediatric Pneumology and Immunology, Charite Medical Faculty Berlin, Berlin, Germany
  5. 5Charité University Medicine Berlin, Social Pediatric Center, Berlin, Germany
  6. 6German Rheumatism Research Centre Berlin, a Leibniz Institute, Berlin, Germany
  7. 7National Pediatrics Center for Familial Mediterranean Fever, "Arabkir" Medical Complex-Institute of Child and Adolescent Health, Yerevan, Armenia
  8. 8Pediatric Nephrology, Rheumatology and Dermatology, Centre Hospitalier Universitaire de Lyon, Lyon, France
  9. 9Innate Immunity in Infectious and Autoimmune Diseases, Hospices Civils de Lyon, Lyon, France
  10. 10Pediatric Rheumatology, Children's Hospital St. Marien, Landshut, Germany
  11. 11Department of Pediatric Rheumatology, Witten/Herdecke University, Witten, Germany
  12. 12Kinderklinik, Medizinische Hochschule Hannover, Hannover, Germany
  13. 13Pediatric Rheumatology and Immunology, University Medicine, Albert-Schweitzer-Campus 1, Muenster, Germany
  14. 14Department of Internal Medicine I, Marienhospital Herne Medical Center of Ruhr University Bochum, Herne, Germany
  15. 15Department of General Pediatrics, Asklepios Hospital Group, Hamburg, Germany
  16. 16Division of Pediatric Rheumatology and Immunology, Department of Pediatrics, University Hospital of Munich, Munchen, Germany
  17. 17Epidemiology Unit, German Rheumatism Research Center Berlin, Berlin, Germany
  18. 18Department of Paediatrics II, University Hospital Essen, Essen, Germany
  19. 19Helios Klinikum Krefeld, Center of Pediatrics and Youth Medicine, Krefeld, Germany
  20. 20Department of Pediatric Oncology, Hematology and Clinical Immunology, Division of Pediatric Rheumatology, Heinrich-Heine-Universitat Dusseldorf, Dusseldorf, Germany
  21. 21German Center for Pediatric and adolescent Rheumatology, Hospital Garmisch-Partenkirchen, Garmisch-Partenkirchen, Germany
  22. 22Kinder- und Jugendrheumatologie, Clementine Children's Hospital—Dr Christ'sche Foundation, Frankfurt am Main, Germany
  23. 23Département Femme-Mère-Enfant (DFME), Unité Romande d'Immuno-rhumatologie Pédiatrique (URIRP), CHUV, University of Lausanne, Lausanne, Switzerland
  24. 24Department of General Pediatrics, Hôpital Jean Verdier, Bondy, France
  25. 25General Pediatrics and Internal Medicine, Robert-Debré Mother-Child University Hospital, Paris, France
  26. 26Department of Pediatrics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
  27. 27Institute of Immunology, University Hospital Munster Department of Internal Medicine D, Munster, Germany
  1. Correspondence to Dr Inès Elhani; ines.elhani{at}aphp.fr

Abstract

Introduction Familial Mediterranean fever (FMF) is a genetic disease leading to recurrent episodes of inflammation. Two pathogenic variants are required for classical disease, but the disease can occur in heterozygous patients. Patients are treated continuously with colchicine to prevent amyloid A (AA) amyloidosis, including heterozygous patients who display a moderate form of FMF and rarely develop AA amyloidosis. The need for lifelong colchicine treatment in heterozygous FMF is therefore controversial. We aimed to characterise genotype-specific levels of inflammatory biomarkers, and to focus on heterozygous patients who discontinued colchicine.

Methods All patients with FMF from the European databases AIDnet and JIRcohort who received colchicine during follow-up were included. Demographics, C reactive protein (CRP), serum amyloid A (SAA), S100A8/A9 and S100A12 levels, leucocyte and neutrophil counts were extracted. Visits were classified as active, subclinical or inactive according to symptoms, CRP and SAA levels.

Results Data from 747 patients were extracted (233 homozygous, 201 compound heterozygous, 224 heterozygous patients, 49 heterozygous with one class III variant and 40 compound heterozygous with two class III variants). During active visits, all biomarker levels were higher compared with inactive visits (p<0.001). Heterozygous patients showed lower levels of CRP, SAA, S100A8/A9 and S100A12 during inactive and subclinical visits than patients with two class IV-V variants. Colchicine was discontinued in 52 heterozygous patients and reintroduced in 23 of them (44%).

Conclusion S100A8/A9 and S100A12 proteins are biomarkers that can be used to assess disease activity. Heterozygous patients have lower levels of inflammatory biomarkers and some of them can sustainably discontinue colchicine treatment.

  • Familial Mediterranean Fever
  • Inflammation
  • Child

Data availability statement

Data are available upon reasonable request. The data that support the findings of this study are available from the corresponding author, IE, upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/rmdopen-2024-004677

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Familial Mediterranean fever is an autoinflammatory disease that classically requires two pathogenic variants to manifest. Subjects with only one variant may develop a moderate form of the disease, but are usually treated with lifelong colchicine. Classical inflammatory markers (C reactive protein and serum amyloid A) are not specific for the disease.

    WHAT THIS STUDY ADDS

    • S100A8/A9 and S100A12 levels are elevated in familial Mediterranean fever following a gene-dose effect and reflect disease activity. Colchicine was sustainably discontinued in 44% of patients in whom it was attempted.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • Colchicine could be discontinued in some patients with heterozygous familial Mediterranean fever. S100A8/A9 and S100A12 could be used to monitor disease activity in familial Mediterranean fever.

    Introduction

    Familial Mediterranean fever (FMF) is a long-known genetic disease primarily affecting patients from the Mediterranean region.1 Patients display periodic fever beginning in childhood, associated with systemic inflammatory symptoms, especially aseptic serositis, joint inflammation and pseudoerysipelas.1 Colchicine is the mainstay of treatment used in FMF,2 preventing inflammatory attacks and complications. Indeed, if uncontrolled, FMF can lead to AA amyloidosis, which is caused by chronic and/or recurrent systemic inflammation.1 FMF is classically seen as recessive disease, meaning that two pathogenic variants in MEFV must be present for the disease to develop. However, the possibility of diagnosing FMF in patients with a heterozygous mutation of MEFV is accepted. The current EUROFEVER/PRINTO (Paediatric Rheumatology INternational Trials Organisation) diagnostic criteria consider the clinical diagnosis of FMF in patients with only one pathogenic variant (non-confirmatory genotype) and compatible clinical manifestations.3 In FMF, disease severity and presentation are strongly influenced by genotype.4 5 Most (>98%) individuals carrying a single mutation—even with a highly pathogenic one—are clinically healthy,6 but some may have increased subclinical inflammation.7 Less than 2% of heterozygous carriers develop an FMF-like disease with a relative risk of 6.3–8.1 in comparison to healthy controls.7 8 These FMF-like patients tend to have a different clinical presentation: first, the disease seems less severe, with less-intense organ involvements.9–12 Second, a subset of heterozygous patients may experience remission of FMF symptoms after childhood and perhaps some of these patients could be more accurately classified as having Periodic Fever, Aphthous Stomatitis, Pharyngitis, and Adenitis (PFAPA) syndrome.13 Finally, heterozygous FMF-like patients scarcely, if ever, develop AA amyloidosis.14 Therefore, the relevance of lifelong colchicine use in heterozygous patients is questionable, as shown in several cohort studies.15–18

    The main objectives of colchicine treatment are to reduce inflammatory episodes and normalise inflammatory parameters in order to prevent amyloidosis.2 According to European recommendations, C reactive protein (CRP) and serum amyloid A (SAA) should be used to determine disease activity and to adapt treatments.2 However, both parameters increase during inflammatory processes, regardless of whether they might be infectious or related to a flare of FMF.19 Moreover, in most cases, both CRP and SAA return to normal levels between attacks and cannot be used as a means of monitoring overall disease control. The search for biomarkers specific to FMF has been ongoing for two decades. Among the candidate biomarkers, proteins from the S100 family have been of interest. The Damage-Associated Molecular Pattern (DAMP) proteins S100A12 and S100A8/A9 (the later complex also called calprotectin) are suggested to be specifically elevated in patients with autoinflammatory diseases. Both biomarkers have been shown to be elevated in patients with FMF compared with healthy subjects, even when out of flare.20–23

    The objectives of our study were to investigate inflammatory biomarkers, including the phagocyte-specific S100-proteins, in patients with FMF in and out of flares and according to genotype, and to analyse the influence of gene dosage on biomarkers in heterozygous patients for which colchicine was sustainably stopped.

    Methods

    Patients

    Patients were identified from the JIR (Juvenile Inflammatory Rheumatism) cohort, an ongoing international multicentre data that started in 2013 (http://www.fondationres.org/fr/jircohorte—NTC02377245), and AID-net, a German retrospective registry that was in place from 2008 to 2018, with patient recruitment to the registry and patient material deposition to biomaterial banks.24 25 These registries collect clinical and biological data at each medical visit throughout the follow-up period. We considered for inclusion all patients with a registered diagnosis of FMF with genetic status provided, who had been included in the registries before the age of 18 years old, and who needed long-term medication with colchicine due to clinical FMF. The export took place in March 2023 and included all eligible visits since the start of both databases. Clinical data including sex, mutational status, disease history and medication history were extracted.

    Ethical approval for AID-Net was obtained from University Muenster (Ref.: 2009-031 f-S) and from University Duisburg-Essen (08-3866). The JIRcohort protocol was approved by the French Ethics Committee (CCTIRS) on 21 April 2015 (decision number 14.302). The electronic form of the JIRcohort was approved by the National Commission of data processing and liberties (CNIL) on 27 March 2015 (decision number DR-2015-218).

    Genetic classifications

    Variant pathogenicity was defined according to the Infevers registry ranging from class I (benign) to class V (pathogenic) (https://infevers.umai-montpellier.fr/web/search.php?n=1). Homozygous and compound heterozygous FMF was defined by the presence of two class IV-V variants (likely pathogenic and pathogenic variants) in exon 10 of MEFV, and heterozygous patients were defined by the presence of one class IV/V variant in exon 10 of MEFV. Patients with one class III variant and one class IV-V variant were also classified as heterozygous.26 Patients with variants described as causative of dominant FMF (exon 8) as well as patients with no variant ≥III were excluded. Habits of genetic testing vary according to the year in which the diagnosis was made, but also according to local habits. Since 2013, according to the national recommendations for diagnosis and care (PNDS) in France (https://www.has-sante.fr/upload/docs/application/pdf/2013-02/pnds_-_fievre_mediterraneenne_familiale.pdf), Exon 2 variant E148Q is not reported to physicians as it is considered a benign variant, although it is a class III variant in the Infevers database.27 Furthermore, since 2020, the international recommendations for the genetic diagnosis of autoinflammatory diseases stipulate that these class III variants should no longer be reported in laboratory reports even if they used to be.28 To harmonise patient data, we have based our definitions on these more recent restrictive recommendations.

    Clinical visits

    Visits were taken into account if they included the result of at least one CRP or SAA value. For each patient, each visit was classified into three categories according to disease activity: ‘active disease visits’ (ongoing inflammatory attacks, fever, serositis or pseudoerysipelas during the visit), ‘subclinical disease visits’ (raised CRP and/or SAA levels without clinical signs related to inflammatory attacks) or ‘inactive disease visits’ (normal CRP and SAA levels, outside an inflammatory flare). Categorisation of patients was similar for our AID-Net analysis of patients with FMF from May 2014, that included a minor group of patients (n=128), that is now part of the whole study group.29 Raised inflammatory markers were defined as CRP >5 mg/L and SAA >7 mg/L.

    Determination of inflammatory markers

    Inflammatory markers (erythrocyte sedimentation rate (ESR), CRP, SAA, leucocyte count, neutrophils count, Proteins S100A8/A9 and S100A12) were obtained from registry visits as inserted by the participating centres. S100A8/A9 and S100A12 levels from AID-Net patients were collected centrally at University of Muenster and measured with Bühlmann MRP8/14 Calprotectin ELISA (Bühlmann Laboratories AG) according to manufacturer’s instructions, or for S100A12 using an inhouse ELISA (normal level <150 ng/mL) as previously reported.23 S100A8/A9 and S100A12 concentrations in the JIRcohort were assessed at individual participating centres and obtained as part of routine visits.

    For each patient, the median values of biomarkers were calculated for each category of visits (active, subclinical and inactive). This means that each patient has a maximum of three values per biomarker that are included in the analysis. Only if patients presented in all three disease activity categories could three values (median for multiple presentations) be analysed, otherwise fewer.

    Statistics

    Numeric variables were expressed as median (IQR) and discrete outcomes as absolute and relative (%) frequencies. Normality and heteroskedasticity of continuous data were assessed with Shapiro-Wilk and Levene’s test, respectively. Continuous outcomes were compared with Anova, Welch Anova or Kruskal-Wallis tests according to data distribution. Discrete outcomes were compared with χ2 or Fisher’s exact test accordingly. The alpha risk was set to 5% and two-tailed tests were used. Statistical analysis was performed with EasyMedStat (V.3.28; www.easymedstat.com).

    Results

    FMF genotypes in the cohorts

    A total of 747 individuals diagnosed with FMF and treated with colchicine were identified from registries, including 455 patients from AIDnet and 292 from JIRcohort (figure 1A). No gender disparity was observed between the groups. Nevertheless, genotype distribution was significantly different between the two databases (p<0.001), with more heterozygous and patients with class III variants in AIDnet, while there were more compound heterozygous patients in JIRcohort (figure 1B). The distribution of homozygous (n=233), compound heterozygous (n=201) and heterozygous patients (n=224) differed between the registries, with heterozygous patients being more prevalent in AIDnet and compound heterozygous patients being more prevalent in JIRcohort (p<0.001).

    Figure 1
    Figure 1

    Characteristics of included patients according to database. (A) Flow chart of included patients and corresponding mutation status. (B) Dot plot of included patients according to genotype and database (one patient is represented by one dot). WT, wild type; III, class III variant.

    Heterozygous patients with FMF

    The percentage of heterozygous patients who discontinued colchicine, either temporarily or sustainably, was evenly distributed between the two databases (table 1). In heterozygous patients who discontinued colchicine, the median total duration of colchicine use was 5 years longer in the JIR cohort compared with AIDnet (p=0.006). However, the age at which colchicine was discontinued did not exhibit a significant difference, suggesting that patients from JIRcohort initiated treatment at an earlier age.

    View this table:
    Table 1

    Patients’ characteristics according to database

    Differences between the registers in terms of data collection

    The content of both databases was also heterogenous. For instance, while 37%–60% (n=170–137) of patients from AIDnet had at least one measurement of S100A8/A9 and S100A12, because of the included biorepository, it only concerned 1%–7% (n=2–20) of patients from JIRcohort. Likewise, SAA and leucocytes count were more frequently entered into Aidnet than JIR cohort. Overall, not all biomarkers were entered at all visits for both registries.

    All biomarkers differentiate between active attacks and inactive disease states

    All tested biomarkers were higher when measured during an active visit compared with subclinical and inactive visits and increased gradually. While CRP, SAA, ESR, Neutrophil count and S100A12 were significantly different for all activity groups, S100A8/A9 and leucocyte count were not significantly different between subclinical and active visits (table 2).

    View this table:
    Table 2

    Inflammatory biomarkers according to disease activity in all included patients

    Biological differences between FMF genotypes in patients with class IV-V variants

    Mutational status is described in table 3. During follow-up, homozygous patients were more likely to register active visits or subclinical visits (p<0.001, respectively). Conversely, heterozygous patients were more frequently classified as inactive than homozygous and compound heterozygous patients (p<0.001, table 4). In inactive visits, median SAA, ESR, S100A8/A9 and S100A12 levels were higher in homozygous patients. Pairwise analyses further suggested that values of SAA, S100A8/A9 and S100A12 levels were specifically lower in heterozygous patients than in compound heterozygous patients during inactive visits (table 4, figure 2A,B). Moreover, in subclinical visits, ESR, S100A8/A9 and S100A12 levels were higher in homozygous than in heterozygous patients. Leucocyte and neutrophil counts were not different between genotypes in neither inactive nor subclinical visits. We could not perform statistical tests for active visits because there were too few datapoints reported.

    View this table:
    Table 3

    Mutational status in MEFV gene of included patients

    View this table:
    Table 4

    Visits characteristics and inflammatory biomarkers according to genotype

    Figure 2
    Figure 2

    S100 proteins levels measured during inactive visits levels according to genotype. (A) Decimal logarithm conversion (10n) of median S100A8/A9 levels in patients with class III variants, heterozygous, compound heterozygous and homozygous patients. (B) Decimal logarithm conversion (10n) of median S100A12 levels in patients with class III variants, heterozygous, compound heterozygous and homozygous patients.

    In a substantial proportion of heterozygous patients with FMF colchicine could be stopped

    Colchicine was stopped during follow-up in 52 patients (23%). Among these, discontinuation was sustained up to last visit in 29 patients, and colchicine was restarted in 23 patients (figure 1A). Reasons for discontinuation in the patients were not reported in both registries. Median age at colchicine stop was 9.7 years old (2.5–17.2). Median follow-up time after colchicine stop was 12 months (0–79) and was >2 years in 13 patients (45%). In patients who restarted colchicine, it was restarted after 1.92 years (0.23–3.86). There was no difference in genotype between patients who had sustainably stopped and those who had restarted colchicine (p=0.3). The most frequent variant was M694V (69%), followed by V726A (23%), R761H (4%), M680I (2%) and M694I (2%).

    S100 proteins are key biomarkers to differentiate patients with class IV-V variants from those with class III variants

    As the clinical significance of class III variants (such as E148Q) is controversial, and as described in the Methods section, we excluded these patients from our overall analysis and from the colchicine discontinuation analysis. Nevertheless, we were interested in an additional comparison of these patients, as shown in table 4, in which we compared 89 patients with only class III variants with the other genotypes. We found that S100A12 levels were specifically lower in patients with only class III variants compared with heterozygous patients at both inactive and subclinical visits (table 4). In addition, S100A8/A9 levels were significantly lower in inactive disease. No other biomarker was different between the two groups. Furthermore, they are significantly elevated in homozygous and compound heterozygous patients when compared with heterozygous patients in inactive disease.

    Discussion

    In this study, we (1) successfully pooled data from two European databases of patients with well-characterised FMF disease during their childhood follow-up. We (2) showed that the levels of inflammatory biomarkers in FMF are genotype dependent, reflect gene dosage and correlate with the clinically described severity of the variants. Patients with heterozygous variants had lower levels of both classical (CRP, SAA and ESR) and novel (S100A8/A9, S100A12) biomarkers than compound heterozygous and homozygous patients. In addition, we (3) showed that colchicine could be discontinued in a significant number of heterozygous patients with a treatment-free follow-up of one to 2 years, adding to the growing body of knowledge about disease outcomes in heterozygous FMF. Finally, (4) we found that patients with only class III variants could be clearly distinguished from class IV/V variants based on S100 protein analysis.

    The first step of our study—combining two different databases—revealed different practices in the management of FMF. As the diagnosis of FMF is a combination of phenotype and genotype, a definitive diagnosis may be difficult in heterozygous patients. 6 7 Differences in genotype ordering and interpretation may account for the higher proportion of patients with class III variants in AIDnet than in JIRcohort. This was also influenced by the national genotype reporting practices as described in the Methods section. Genotype reporting could also differ according to the genetic analysis performed (next-generation sequencing panel, exon 10 Sanger or whole gene Sequencing). Overall, the merging of two databases required additional revisions with the exclusion of patients with class III variants for the global analysis and a careful interpretation.

    The role of MEFV genotype for the FMF phenotype is very well known.4 While M694V variants are associated with a more severe disease with a dose-dependent effect, they additionally result in higher frequency of attacks, higher biomarkers and higher rates of AA amyloidosis.4 30 A previous report had already shown higher CRP levels in patients with exon 10 homozygosity than in those with compound heterozygosity and heterozygosity.31 In addition, healthy carriers have elevated levels of inflammatory biomarkers compared with non-carriers.32 Conversely, patients with two M694V variants have higher levels than those heterozygous for M694V.32

    S100A8/A9 and S100A12 have been candidate biomarkers to sensitively assess inflammation in FMF before.22 29 33 34 In this study, we confirm that median S100A8/A9 and S100A12 levels are higher during flares than during subclinical visits and inactive visits.29 We also found that heterozygous patients displayed lower inflammatory biomarkers in both inactive and subclinical state than patients with FMF with two class IV/V variants. Taken together, S100-protein evaluation according to disease activity status in patients with FMF clearly reflects gene dosage of MEFV. We had recently identified pyroptosis as a key event inherent of FMF pathophysiology, to be responsible for S100A12 secretion, especially in inactive patients with FMF.29 30 33 35 This is the first study, that also identifies S100A8/A9 as mirror of MEFV-gene dosage in a real-world setting, not surprisingly, as all three proteins are biochemically closely related with intracellular and extracellular functions according to their DAMP activity.36 The exact place of these two biomarkers in the management of FMF still needs to be determined, particularly in the therapeutic management and in classifying patients in terms of risk of colchicine resistance and/or disease severity. However, a single cut-off for all patients with FMF may not be appropriate as genotype affect their levels. In addition, the added value of these biomarkers in predicting the risk of AA amyloidosis remains to be determined.

    Lifelong colchicine medication is the rule in FMF because of the risk of the life-threatening complication AA amyloidosis in homozygous patients.2 27 As previously mentioned, this risk is minimal in heterozygous patients.37 In this study, we found that colchicine stop was sustained until last visit in 56% of heterozygous patients for which colchicine stop was tried. Colchicine discontinuation in FMF has been up to now investigated in four studies,15–18 which included patients with class III variants and without MEFV variants depending on the study. This approach differs from ours, which excluded patients with FMF without genetic confirmation and separated those with a class III variant. High-dose colchicine medication prior to withdrawal, high attack severity, compound heterozygosity and M694V variants were associated with higher relapse rates in previous studies.16 Although the median follow-up of 2 years in our study may be short, compared with a lifetime perspective, the relevance of continuous colchicine treatment in patients who can go several months to years without flares should be questioned. Heterozygous patients may experience successive periods of active and inactive disease influenced by genetic modifiers and/or environmental factors. Therefore, the need to restart colchicine may indicate the re-emergence of a transient active phase of the disease rather than a definitive need for treatment. In previous studies, colchicine was discontinued for 6 months to 3 years in patients with inactive disease.15–18 Further studies will need to be conducted to specifically determine the duration without colchicine restart in these patients. Definite treatment discontinuation may however apply to a subgroup of heterozygous patients whose characteristics remain to be determined. In addition, it is controversial whether these patients, especially during infancy, should be reclassified as having PFAPA or syndrome of undifferentiated recurrent fever rather than classical FMF.

    International recommendations do not consider genotype when determining treatment strategy. However, our results suggest that the management of these patients should probably be different from that of patients with two pathogenic variants. Factors that may be considered in the decision to discontinue colchicine in heterozygous FMF include the exact genotype (M694V variant), family history of AA amyloidosis, predictable phase of active disease (attacks triggered by life events, pregnancy…), and other biomarkers that still need to be determined.

    Our study has several limitations intrinsic due to its retrospective nature. Differences in genetic testing protocols and variant classification, such as the reporting of class III variants may have introduced heterogeneity in the diagnosis and classification of patients. However, we have used the latest recommendations for variant classification to reduce this bias. In addition, some analyses could not be carried out due to missing data, especially regarding novel biomarkers and active visits. Patients were included if they had had at least one measurement of CRP and SAA. Therefore, patients with a more severe disease could be over-represented. Moreover, the reasons for colchicine discontinuation were not specified, but could include physician’s decision or non-compliance. Finally, the decision to restart colchicine was not described and may vary between physicians, centres and countries. Further prospective studies should investigate heterozygous patients with controlled colchicine discontinuation and evaluate their biochemical inflammatory activity including S100-proteins at the time of discontinuation and according to disease status.

    Interestingly, classical biomarkers (SAA, CRP and ESR) were not different between patients with only class III variants and those with a class IV-V variant. S100 proteins were the only biomarker that differed specifically between the two groups. The influence of class III variants on phenotype is still debated: A Turkish study found that patients with both class III/IV-V had more severe symptoms than those with just heterozygous class IV-V, but biomarker levels were not discussed,38 whereas an Israelian study showed the contrary.26 The low levels of S10012A measured in patients with class III variants may be similar to those found in healthy individuals or those with non-inflammasome-related diseases.39 Therefore, these results suggest that the inflammatory activity in these patients is less than in classical FMF and therefore even more supportive of a trial of colchicine withdrawal in these patients.

    Conclusion

    In conclusion, we could show that S100A8/A9 and S100A12 levels are elevated in FMF following a gene-dose effect and reflect disease activity. For patients with heterozygous FMF, we see a specific less-severe phenotype of FMF, that may require specific management. Possibly this includes discontinuation of colchicine, which could be monitored with the aid of new biomarker levels, that is, S100A8/A9 and S100A12. Their uniquely lower levels of inflammatory biomarkers should prompt clinicians to consider genotype when interpreting laboratory results in patients with FMF.

    Data availability statement

    Data are available upon reasonable request. The data that support the findings of this study are available from the corresponding author, IE, upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Ethical approval for AID-Net was obtained from University Muenster (Ref.: 2009-031-f-S) and from University Duisburg-Essen (08-3866). The JIRcohort protocol was approved by the French Ethics Committee (CCTIRS) on 21 April 2015 (decision number 14.302). The electronic form of the JIRcohort was approved by the National Commission of data processing and liberties (CNIL) on 27 March 2015 (decision number DR-2015-218). Participants gave informed consent to participate in the study before taking part.

    Acknowledgments

    Sophie Georgin-Lavialle, Jana Pashlopnik, Gerald Berthet, Diana Dan, Rahma Guedri, Etienne Merlin, Pascal Pillet, Frederica Vanoni, Isabelle Koné-Paut.

    References

    1. ↵
      1. Ozdogan H,
      2. Ugurlu S
      . Familial Mediterranean Fever. Presse Med 2019;48:e61–76. doi:10.1016/j.lpm.2018.08.014
      OpenUrl
    2. ↵
      1. Ozen S,
      2. Demirkaya E,
      3. Erer B, et al
      . EULAR recommendations for the management of familial Mediterranean fever. Ann Rheum Dis 2016;75:644–51. doi:10.1136/annrheumdis-2015-208690
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Gattorno M,
      2. Hofer M,
      3. Federici S, et al
      . Classification criteria for autoinflammatory recurrent fevers. Ann Rheum Dis 2019;78:1025–32. doi:10.1136/annrheumdis-2019-215048
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Kisaoglu H,
      2. Baba O,
      3. Kalyoncu M
      . Genotype-Phenotype Associations of Children With Familial Mediterranean Fever in a Cohort Consisting of M694V Mutation and Implications for Colchicine-Resistant Disease. J Clin Rheumatol 2023;29:207–13. doi:10.1097/RHU.0000000000001953
      OpenUrl
    5. ↵
      1. Barut K,
      2. Sahin S,
      3. Adrovic A, et al
      . Familial Mediterranean fever in childhood: a single-center experience. Rheumatol Int 2018;38:67–74. doi:10.1007/s00296-017-3796-0
      OpenUrlPubMed
    6. ↵
      1. Ben-Chetrit E,
      2. Touitou I
      . The significance of carrying MEFV variants in symptomatic and asymptomatic individuals. Clin Genet 2024;106:217–23. doi:10.1111/cge.14566
      OpenUrl
    7. ↵
      1. Lieber M,
      2. Kallinich T,
      3. Lohse P, et al
      . Increased serum concentrations of neutrophil-derived protein S100A12 in heterozygous carriers of MEFV mutations. Clin Exp Rheumatol 2015;33:S113–6.
      OpenUrl
    8. ↵
      1. Jéru I,
      2. Hentgen V,
      3. Cochet E, et al
      . The risk of familial Mediterranean fever in MEFV heterozygotes: a statistical approach. PLoS One 2013;8:e68431. doi:10.1371/journal.pone.0068431
    9. ↵
      1. Ozturk K,
      2. Cakan M
      . The analysis of genotype-phenotype correlation in familial Mediterranean fever. Pediatr Int 2022;64:e15017. doi:10.1111/ped.15017
    10. ↵
      1. Pras E,
      2. Livneh A,
      3. Balow JE, et al
      . Clinical differences between North African and Iraqi Jews with familial Mediterranean fever. Am J Med Genet 1998;75:216–9. doi:10.1002/(sici)1096-8628(19980113)75:2<216::aid-ajmg20>3.0.co;2-r
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Booty MG,
      2. Chae JJ,
      3. Masters SL, et al
      . Familial Mediterranean fever with a single MEFV mutation: where is the second hit? Arthritis Rheum 2009;60:1851–61. doi:10.1002/art.24569
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Marek-Yagel D,
      2. Berkun Y,
      3. Padeh S, et al
      . Clinical disease among patients heterozygous for familial Mediterranean fever. Arthritis Rheum 2009;60:1862–6. doi:10.1002/art.24570
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Hentgen V,
      2. Grateau G,
      3. Stankovic-Stojanovic K, et al
      . Familial Mediterranean fever in heterozygotes: are we able to accurately diagnose the disease in very young children? Arthritis Rheum 2013;65:1654–62. doi:10.1002/art.37935
      OpenUrlCrossRefPubMed
    14. ↵
      1. Giese A,
      2. Kurucay M,
      3. Kilic L, et al
      . Quality of life in adult patients with Familial Mediterranean fever living in Germany or Turkey compared to healthy subjects: a study evaluating the effect of disease severity and country of residence. Rheumatol Int 2013;33:1713–9. doi:10.1007/s00296-012-2622-y
      OpenUrlPubMed
    15. ↵
      1. Ben-Zvi I,
      2. Krichely-Vachdi T,
      3. Feld O, et al
      . Colchicine-free remission in familial Mediterranean fever: featuring a unique subset of the disease-a case control study. Orphanet J Rare Dis 2014;9:3. doi:10.1186/1750-1172-9-3
    16. ↵
      1. Cohen K,
      2. Spielman S,
      3. Semo-Oz R, et al
      . Colchicine treatment can be discontinued in a selected group of pediatric FMF patients. Pediatr Rheumatol Online J 2023;21:2. doi:10.1186/s12969-022-00780-w
    17. ↵
      1. Aviel YB,
      2. Rawan S,
      3. Fahoum S, et al
      . Discontinuation of Colchicine Therapy in Children With Familial Mediterranean Fever. J Rheumatol 2021;48:1732–5. doi:10.3899/jrheum.201158
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Sönmez HE,
      2. Batu ED,
      3. Bilginer Y, et al
      . Discontinuing colchicine in symptomatic carriers for MEFV (Mediterranean FeVer) variants. Clin Rheumatol 2017;36:421–5. doi:10.1007/s10067-016-3421-8
      OpenUrl
    19. ↵
      1. Korkmaz C,
      2. Ozdogan H,
      3. Kasapçopur O, et al
      . Acute phase response in familial Mediterranean fever. Ann Rheum Dis 2002;61:79–81. doi:10.1136/ard.61.1.79
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Asan G,
      2. Derin ME,
      3. Doğan HO, et al
      . Can Calprotectin Show Subclinical Inflammation in Familial Mediterranean Fever Patients? J Korean Med Sci 2020;35:e63. doi:10.3346/jkms.2020.35.e63
    21. ↵
      1. Kılıçaslan E,
      2. Düzenli T,
      3. Çelik S, et al
      . Assessment of Serum Resistin and Plasma Calprotectin Levels as Biomarkers of Inflammation in Patients with Familial Mediterranean Fever Disease. Mediterr J Rheumatol 2022;33:322–7. doi:10.31138/mjr.33.3.322
      OpenUrl
    22. ↵
      1. Tezcan D,
      2. Onmaz DE,
      3. Sivrikaya A, et al
      . Assessment of serum neopterin and calprotectin as biomarkers for subclinical inflammation in patients with familial Mediterranean fever. Ir J Med Sci 2023;192:2015–22. doi:10.1007/s11845-022-03173-w
      OpenUrl
    23. ↵
      1. Kessel C,
      2. Holzinger D,
      3. Foell D
      . Phagocyte-derived S100 proteins in autoinflammation: putative role in pathogenesis and usefulness as biomarkers. Clin Immunol 2013;147:229–41. doi:10.1016/j.clim.2012.11.008
      OpenUrlCrossRefPubMed
    24. ↵
      1. Cabrera N,
      2. Lega J-C,
      3. Kassai B, et al
      . Safety of biological agents in paediatric rheumatic diseases: A real-life multicenter retrospective study using the JIRcohorte database. Joint Bone Spine 2019;86:343–50. doi:10.1016/j.jbspin.2018.08.003
      OpenUrl
    25. ↵
      1. Lainka E,
      2. Bielak M,
      3. Hilger V, et al
      . Translational research network and patient registry for auto-inflammatory diseases. Rheumatology (Oxford) 2011;50:237–42. doi:10.1093/rheumatology/keq270
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Tirosh I,
      2. Yacobi Y,
      3. Vivante A, et al
      . Clinical significance of E148Q heterozygous variant in paediatric familial Mediterranean fever. Rheumatology (Oxford) 2021;60:5447–51. doi:10.1093/rheumatology/keab128
      OpenUrlCrossRef
    27. ↵
      1. Georgin-Lavialle S,
      2. Savey L,
      3. Buob D, et al
      . French practical guidelines for the diagnosis and management of AA amyloidosis. Rev Med Intern 2023;44:62–71. doi:10.1016/j.revmed.2022.12.004
      OpenUrl
    28. ↵
      1. Shinar Y,
      2. Ceccherini I,
      3. Rowczenio D, et al
      . ISSAID/EMQN Best Practice Guidelines for the Genetic Diagnosis of Monogenic Autoinflammatory Diseases in the Next-Generation Sequencing Era. Clin Chem 2020;66:525–36. doi:10.1093/clinchem/hvaa024
      OpenUrlCrossRef
    29. ↵
      1. Gohar F,
      2. Orak B,
      3. Kallinich T, et al
      . Correlation of Secretory Activity of Neutrophils With Genotype in Patients With Familial Mediterranean Fever. Arthritis Rheumatol 2016;68:3010–22. doi:10.1002/art.39784
      OpenUrl
    30. ↵
      1. Stoler I,
      2. Freytag J,
      3. Orak B, et al
      . Gene-Dose Effect of MEFV Gain-of-Function Mutations Determines ex vivo Neutrophil Activation in Familial Mediterranean Fever. Front Immunol 2020;11:716. doi:10.3389/fimmu.2020.00716
    31. ↵
      1. Kelesoglu FM,
      2. Aygun E,
      3. Okumus NK, et al
      . Evaluation of subclinical inflammation in familial Mediterranean fever patients: relations with mutation types and attack status: a retrospective study. Clin Rheumatol 2016;35:2757–63. doi:10.1007/s10067-016-3275-0
      OpenUrl
    32. ↵
      1. Lachmann HJ,
      2. Sengül B,
      3. Yavuzşen TU, et al
      . Clinical and subclinical inflammation in patients with familial Mediterranean fever and in heterozygous carriers of MEFV mutations. Rheumatology (Oxford) 2006;45:746–50. doi:10.1093/rheumatology/kei279
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      1. Kallinich T,
      2. Wittkowski H,
      3. Keitzer R, et al
      . Neutrophil-derived S100A12 as novel biomarker of inflammation in familial Mediterranean fever. Ann Rheum Dis 2010;69:677–82. doi:10.1136/ard.2009.114363
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Taylan A,
      2. Gurler O,
      3. Toprak B, et al
      . S1000A12, Chitotriosidase, and Resolvin D1 as Potential Biomarkers of Familial Mediterranean Fever. J Korean Med Sci 2015;30:1241–5. doi:10.3346/jkms.2015.30.9.1241
      OpenUrl
    35. ↵
      1. Varga G,
      2. Schleifenbaum S,
      3. Koenig U, et al
      . Phagocytic cell death leads to enhanced release of pro-inflammatory S100A12 in familial Mediterranean fever. Mol Cell Pediatr 2023;10:19. doi:10.1186/s40348-023-00173-3
    36. ↵
      1. Foell D,
      2. Wittkowski H,
      3. Roth J
      . Mechanisms of Disease: a “DAMP” view of inflammatory arthritis. Nat Rev Rheumatol 2007;3:382–90. doi:10.1038/ncprheum0531
      OpenUrl
    37. ↵
      1. Atoyan S,
      2. Hayrapetyan H,
      3. Sarkisian T, et al
      . MEFV and SAA1 genotype associations with clinical features of familial Mediterranean fever and amyloidosis in Armenia. Clin Exp Rheumatol 2016;34:72–6.
      OpenUrl
    38. ↵
      1. Sarı A,
      2. Bodakçi E,
      3. Armağan B, et al
      . Phenotypic characterization of Familial Mediterranean Fever patients harboring variants of uncertain significance. Turk J Med Sci 2021;51:1695–701. doi:10.3906/sag-2011-273
      OpenUrl
    39. ↵
      1. Wittkowski H,
      2. Frosch M,
      3. Wulffraat N, et al
      . S100A12 is a novel molecular marker differentiating systemic-onset juvenile idiopathic arthritis from other causes of fever of unknown origin. Arthritis Rheum 2008;58:3924–31. doi:10.1002/art.24137
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • VH and HW contributed equally.

    • Contributors Conceptualisation and supervision: VH, HW. Methodology: VH, HW and IE. Investigation: SB, TK, GA, AB, RB, TB, FD, DF, SF, AG, CH, ALH, GH, AJ, JK, EL, TN, PO, J-PH, CR, KT, CV, EW-R, VH and HW. Formal analysis: IE. Draft-original draft: IE and SB. Writing - review and editing: SB, TK, GA, AB, RB, TB, FD, DF, SF, AG, CH, ALH, GH, AJ, JK, EL, TN, PO, JPH, CR, KT, CV, EW-R, VH, HW. IE is the guarantor of this work.

    • 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.

    • Provenance and peer review Not commissioned; externally peer reviewed.