Article Text

Download PDFPDF

Extended report
Phagocyte-specific S100 proteins and high-sensitivity C reactive protein as biomarkers for a risk-adapted treatment to maintain remission in juvenile idiopathic arthritis: a comparative study
  1. Joachim Gerss1,2,
  2. Johannes Roth3,
  3. Dirk Holzinger2,3,4,
  4. Nicolino Ruperto5,
  5. Helmut Wittkowski4,
  6. Michael Frosch4,
  7. Nico Wulffraat6,
  8. Lucy Wedderburn7,
  9. Valda Stanevicha8,
  10. Dimitrina Mihaylova9,
  11. Miroslav Harjacek10,
  12. Claudio Len11,
  13. Claudia Toppino12,
  14. Massimo Masi13,
  15. Kirsten Minden14,
  16. Traudel Saurenmann15,
  17. Yosef Uziel16,
  18. Richard Vesely17,
  19. Maria Teresa Apaz18,
  20. Rolf-Michael Kuester19,
  21. Maria Jesus Rua Elorduy20,
  22. Ruben Burgos-Vargas21,
  23. Maka Ioseliani22,
  24. Silvia Magni-Manzoni23,
  25. Erbil Unsal24,
  26. Jordi Anton25,
  27. Zsolt Balogh26,
  28. Stefan Hagelberg27,
  29. Henryka Mazur-Zielinska28,
  30. Tsivia Tauber29,
  31. Alberto Martini5,30,
  32. Dirk Foell2,3,
  33. for the Paediatric Rheumatology International Trials Organization (PRINTO)
  1. 1Institute of Biostatistics and Clinical Research, University of Muenster, Muenster, Germany
  2. 2Interdisciplinary Centre of Clinical Research, University of Muenster, Muenster, Germany
  3. 3Institute of Immunology, University of Muenster, Muenster, Germany
  4. 4Department of General Pediatrics, University Children's Hospital of Muenster, Muenster, Germany
  5. 5IRCCS G Gaslini, Pediatria II, Reumatologia, PRINTO, Genova, Italy
  6. 6Department of Paediatric Immunology and Rheumatology, Wilhelmina Children's Hospital, Utrecht, The Netherlands
  7. 7Rheumatology Unit, Institute of Child Health UCL, London, UK
  8. 8Riga Stradins University, Pediatric, Riga, Latvia
  9. 9University Children Hospital, Department of Paediatric Rheumatology, Sofia, Bulgaria
  10. 10Children's Hospital Zagreb, Department of Pediatrics, Immmunology/Rheumatology, Zagreb, Croatia
  11. 11Universitade Federal de Sao Paolo, Dep. De Pediatria, Sao Paulo, Brazil
  12. 12Clinica Pediatrica Università di Torino, Dipartimento di Scienze Pediatriche e dell'Adolescenza, Torino, Italy
  13. 13Clinica Pediatrica, Università di Bologna, Policlinico S.Orsola Malpighi, Bologna, Italy
  14. 14Charite University Hospital Berlin, Kinderklinik, Rheumatologie, Berlin, Germany
  15. 15University Children's Hospital, Pediatric Rheumatology, Zurich, Switzerland
  16. 16Meir Medical Centre, Dept of Pediatrics, Kfar Saba, Israel
  17. 17Detska Fakultna Nemocnica, 1st Pediatric Dept, Kosice, Slovakia
  18. 18Universidad Catolica, Clinica Reina Fabiola – GESER (Rheumatology), Cordoba, Argentina
  19. 19Rheumaklinik Bad Bramstedt, Norddeutsches Zentrum für Kinder- und Jugendrheumatologie, Bad Bramstedt, Germany
  20. 20Hospital de Cruces, Unidad de Reumatología Pediátrica, Bilbao Vizcaya, Spain
  21. 21Hospital General de Mexico, Servicio de Reumatologia, Mexico City, Mexico
  22. 22M. Iashvili Children's Central Clinic, Division of Rheumatology, Tbilisi, Georgia
  23. 23Fondazione IRCCS Policlinico S. Matteo, S.C. Pediatria Ospedaliera, Pavia, Italy
  24. 24Dokuz Eylul University Medical Faculty, Division of Pediatric Rheumatology-Immunology, Balcova, Izmir, Turkey
  25. 25Hospital Sant Joan de Deu, Unidad de Reumatología Pediatrica, Esplugues (Barcelona), Spain
  26. 26National Institute of Rheumatology and Physiotherapy (ORFI), III General and Pediatric Rheumatology Department, Budapest, Hungary
  27. 27Karolinska University Hospital, Pediatric Rheumatology Unit, Stockholm, Sweden
  28. 28Medical University of Silesia, Department of Pediatrics, Zabrze, Poland
  29. 29Asaf Harofe Medical Center, Pediatric Rheumatology Clinic, Zrifin, Israel
  30. 30Università degli studi di Genova, Dipartimento di Pediatria, Genova, Italy
  1. Correspondence to Professor Dirk Foell, Institute of Immunology, University of Muenster, Roentgenstr. 21, D-48149 Muenster, Germany; dfoell{at}uni-muenster.de

Abstract

Objectives Juvenile idiopathic arthritis (JIA) is a chronic inflammatory joint disease affecting children. Even if remission is successfully induced, about half of the patients experience a relapse after stopping anti-inflammatory therapy. The present study investigated whether patients with JIA at risk of relapse can be identified by biomarkers even if clinical signs of disease activity are absent.

Methods Patients fulfilling the criteria of inactive disease on medication were included at the time when all medication was withdrawn. The phagocyte activation markers S100A12 and myeloid-related proteins 8/14 (MRP8/14) were compared as well as the acute phase reactant high-sensitivity C reactive protein (hsCRP) as predictive biomarkers for the risk of a flare within a time frame of 6 months.

Results 35 of 188 enrolled patients experienced a flare within 6 months. Clinical or standard laboratory parameters could not differentiate between patients at risk of relapse and those not at risk. S100A12 and MRP8/14 levels were significantly higher in patients who subsequently developed flares than in patients with stable remission. The best single biomarker for the prediction of flare was S100A12 (HR 2.81). The predictive performance may be improved if a combination with hsCRP is used.

Conclusions Subclinical disease activity may result in unstable remission (ie, a status of clinical but not immunological remission). Biomarkers such as S100A12 and MRP8/14 inform about the activation status of innate immunity at the molecular level and thereby identify patients with unstable remission and an increased risk of relapse.

Statistics from Altmetric.com

Introduction

Juvenile idiopathic arthritis (JIA) is a chronic disease characterised by relapsing joint inflammation in children. Patients can be subdivided into at least three major subgroups (oligoarticular, polyarticular and systemic JIA) with a total of seven subcategories.1 Treatment strategies include non-steroidal anti-inflammatory drugs, corticosteroids, immunosuppressive agents and biological agents. In most patients clinical remission is successfully induced, but others have a refractory course with severe destructive arthritis and complications related to treatment.2 Even if remission is induced, about half of the patients experience a relapse after drug withdrawal.3,,5 It is therefore important to define the status of remission before stopping medication. According to an international consensus on preliminary JIA remission criteria, six continuous months of inactive disease while taking anti-inflammatory drugs defines clinical remission on medication.6 It is unclear, however, whether a treatment maintaining this status will improve the chances of an individual to reach the final goal of clinical remission off medication, currently defined as 12 months of inactive disease after all anti-inflammatory drugs have been withdrawn. We have shown that the stability of remission off medication is not generally improved by prolonged maintenance therapy with methotrexate but rather depends on the individual immunological predisposition of the patient.7

Ideally, physicians would define a patient's remission status on the basis of the synovial inflammation. In clinical remission, a low degree of inflammatory activity may remain present even though this is not detectable by clinical or routine laboratory tests. Relapses might evolve from subclinical disease activity, for example, a status of clinical but not immunological remission. In this situation, remission off medication will not be achievable and treatment withdrawal will result in relapsing disease. While commonly used laboratory parameters have no predictive value for the further course of the disease, our recent data showed that biomarkers of innate immune activation may detect non-apparent disease activity and thereby inform about the risk of relapse.7

Three promising molecular biomarkers of inflammation are the neutrophil activation marker S100A12, the phagocyte activation marker myeloid-related proteins 8 and 14 heterocomplex (MRP8/14; S100A8/A9; calprotectin) and the acute phase reactant high-sensitivity C reactive protein (hsCRP). The neutrophil activation marker S100A12 (EN-RAGE; calgranulin C) is expressed and secreted mainly by granulocytes and exhibits strong proinflammatory activity. Intracellular signalling via various pathways induces secretion of cytokines by monocytes and upregulation of adhesion molecules on endothelial cells.8 ,9 MRP8/14 are calcium-binding proteins of the S100 family expressed by early recruited monocytes and granulocytes. The MRP8/14 complex is secreted by infiltrating phagocytes in synovial inflammation and acts as a proinflammatory ligand of Toll-like receptor 4.10 ,11 Serum concentrations of these proteins correlate with synovial inflammation in JIA, and in particular with the risk of developing a relapse after withdrawal of methotrexate.7 ,12 ,13 CRP is a member of the class of acute phase reactants that respond to rising concentrations of circulating cytokines. The hsCRP test sensitively measures low levels of CRP that are below the threshold of standard CRP analyses.

Based upon previous findings on the predictive value of MRP8/14, we performed a comparative investigation to determine whether patients with JIA at risk of relapse despite clinical remission are also identified by the molecular biomarkers S100A12 and hsCRP.

Methods

Patients

In a previously published controlled trial, patients with JIA with inactive disease for at least 3 months were randomised to withdraw all medication after a further 6 or 12 months of continuous inactive disease status.7 Thus, in total, study patients' inactive disease status continued for 6+3=9 or 12+3=15 months before medication was withdrawn. Stable remission was confirmed clinically (physicians' evaluation of disease activity, absence of systemic signs/symptoms, active arthritis, uveitis) and by erythrocyte sedimentation rate (ESR) as well as standard CRP.6 Of the 364 enrolled patients, samples were available for 188 (52%). The biomarkers S100A12, MRP8/14 and hsCRP were analysed at the time of treatment withdrawal (ie, in confirmed clinical remission on medication). These patients were included in the present study. The demographic data of the study patients have been published previously.7 Patients were followed up for 1 year after treatment withdrawal. Inactive disease was defined according to the provisional criteria for defining clinical inactive disease in JIA;6 the recently published modified criteria were not available at the time the current project was planned.14

Laboratory measures

Serum concentrations of MRP8/14 and S100A12 were determined by sandwich enzyme-linked immunosorbent assays as previously reported.15 Samples were analysed once in clinical remission, at the time when all anti-inflammatory therapies were stopped. The analysing laboratory in Muenster was blinded to the patients' characteristics. A hsCRP test measuring low levels of CRP using laser nephelometry was applied.

Statistical analysis

Biomarker levels were compared between different patient groups using the non-parametric Mann–Whitney U test. The comparison between patients who experienced a flare within 6 months and patients with stable remission over at least 6 months after treatment withdrawal constitutes the primary statistical analysis that provides confirmatory statistical evidence. Moreover, receiver operating characteristic (ROC) analyses were performed in order to evaluate the predictive value of all biomarkers for the risk of disease flares. All biomarkers were dichotomised, applying optimal cut-off levels as per ROC analysis. Specifically, cut-off levels were determined so that predicting flares within 6 months after treatment withdrawal a maximal Youden index was obtained.16 The time from treatment withdrawal to disease flare in a maximum 12-month follow-up period was evaluated by univariate and multivariate methods for survival data. Patients without flare within 12 months were censored and considered to be in stable remission. Based on a proportional hazards model, the HR of each molecular biomarker was estimated. Point estimates are provided with 95% CI. p Values ≤0.05 are considered statistically significant. Statistical analyses were performed using SAS Version 9.2 for Windows (SAS Institute, Cary, North Carolina, USA). The study complied with the Standards for Reporting of Diagnostic Accuracy (STARD) guidelines.17

Results

Overall performance of biomarkers

Of 188 enrolled patients, 120 remained in stable remission over 1 year while 15, 20 and 33 patients experienced a flare within 0–3, 3–6 and 6–12 months after treatment withdrawal, respectively. Median follow-up of all 188 patients was 25.7 months from treatment withdrawal (mean 28.8, range 0.5–67.7 months), and median follow-up of those 120 patients who remained in stable remission over 1 year was 25.7 months from treatment withdrawal (mean 28.8, range 11.7–67.7 months). Comparing patients who stopped treatment at 6 months versus 12 months after randomisation, there were no relevant differences in mean and median biomarker values. Similarly, the two patient groups did not differ significantly with respect to the flare rate within 3 months after treatment withdrawal. Clinical or standard laboratory parameters could not differentiate between patients at risk of relapse and those not at risk. Comparing patients who experienced a flare within 6 months with those in stable remission after treatment withdrawal, the former had significantly higher S100A12 concentrations (median 130 vs 68 ng/ml, p=0.0023) and significantly higher MRP8/14 concentrations (median 850 vs 440 ng/ml, p=0.0061) but not significantly different hsCRP levels (median 0.04 vs 0.04 mg/dl, p=0.1215). Figure 1 shows the median biomarker levels of patients who experienced a flare within 0–3, 3–6 and 6–12 months and of those with stable remission more than 1 year after treatment withdrawal. Particularly high levels of S100A12 and MRP8/14 were associated with early flares within 0–3 months after biomarker analysis. Later flares within 3–6 months and 6–12 months after treatment withdrawal were also associated with elevated levels of S100A12 and MRP8/14, but the association was less distinct. These markers, analysed at the time of stopping treatment in remission, indicate subclinical inflammatory activity—that is, a status of unstable remission with a high risk of disease flares within 3 months. A monotone relation of biomarker levels and time to disease flare exists: with increasing time to disease flare, median S100A12 and MRP8/14 levels decreased monotonically. These consistent results indicate that the observed association with the risk of flare represents a valid and reliable finding. In the case of the biomarker hsCRP, no such consistent results were observed.

Figure 1

Biomarker levels of patients who experienced a flare within 0–3, 3–6 and 6–12 months and of those with stable remission over 1 year. The upper and lower bounds of each box indicate the 25th and 75th percentile, respectively, and heavy lines within the box represent the median. Whiskers are drawn to the nearest value not beyond a standard span from the 25th and 75th percentile, respectively, with the standard span being defined as1.5 times (IQR); values beyond the whiskers (outliers) are drawn individually. p Values report a comparison of patients who experienced a flare within 0–3/3–6/6–12 months with patients in remission over at least 3/6/12 months after discontinuing medication (Mann–Whitney U test).

In ROC analyses the above results regarding the close association of the biomarkers S100A12 and MRP8/14 with the risk of flare as well as the lesser association of hsCRP were confirmed. Differentiating flares within 6 months versus stable remission for at least 6 months after treatment withdrawal, the resulting areas under the ROC curves of S100A12, MRP8/14 and hsCRP were 0.6600 (CI 0.5502 to 0.7698), 0.6455 (CI 0.5360 to 0.7550) and 0.5845 (CI 0.4679 to 0.7010), respectively. The corresponding figures for flares within 3 months were 0.6987 (CI 0.5324 to 0.8650), 0.7510 (CI 0.5991 to 0.9030) and 0.4697 (CI 0.2681 to 0.6712), respectively.

Performance of dichotomised biomarkers

The determined cut-off levels of the biomarkers S100A12, MRP8/14 and hsCRP for optimal prediction of disease flares within 6 months after treatment withdrawal were 175 ng/ml, 690 ng/ml and 0.3 mg/dl, respectively. Applying these cut-off levels, the Youden index for the biomarkers S100A12, MRP8/14 and hsCRP for predicting disease flares within 6 months was 0.3134, 0.2822 and 0.2183, respectively. Table 1 shows further diagnostic statistics of the dichotomised biomarkers. The biomarkers S100A12 and MRP8/14 performed particularly well in predicting early flares within 3 months after treatment withdrawal. In contrast, the biomarker hsCRP may add some information for predicting later flares. This suggests that a combined marker may be beneficial—that is, a predictive algorithm that results from a logical combination of single markers. Among the possible combinations, combining the markers S100A12 and hsCRP attained the best predictive performance. Specifically, the predicted outcome of the combined marker ‘S100A12 or hsCRP’ was positive if S100A12 ≥175 ng/ml or hsCRP ≥0.3 mg/dl. As shown in table 1, the combined marker ‘S100A12 or hsCRP’ brings together the different specific predictive advantages of the single biomarkers and outperforms all single biomarkers with respect to the prediction of early as well as later flares. The combined marker ‘MRP8/14 or hsCRP’ performs almost as well as the combined marker ‘S100A12 or hsCRP’. Table 2 shows the number of flares of patients with high and low values of the dichotomised biomarkers at different time points after treatment withdrawal.

Table 1

Diagnostic statistics of dichotomised biomarkers

Table 2

Number of flares after treatment withdrawal of patients with high and low values of the dichotomised biomarkers

In the above analyses, the predictive performance of a (single or combined) marker is evaluated, assigning equal weights to sensitivity and specificity. Alternatively, one may search for a marker with especially high sensitivity (accepting a lower specificity), or vice versa. In table 3, three successively connected strategies are evaluated, with the decision to stop or continue treatment based on: (1) the single marker S100A12; (2) the combination ‘S100A12 or hsCRP’; and (3) a further combination of all markers ‘S100A12 or hsCRP or MRP8/14’. If combined marker sets are used, the decision to stop treatment would be more restrictive; however, treatment may be continued unnecessarily in more patients despite the lack of a flare risk.

Table 3

Comparison of three different strategies to use (combined) biomarkers for the decision to stop or continue treatment

In univariate survival analyses, the time from treatment withdrawal to disease flare in a maximal 12-month follow-up period was evaluated. For S100A12, the HR of flare was 2.81 (CI 1.70 to 4.65, p<0.0001), for MRP8/14 it was 2.24 (CI 1.39 to 3.62, p=0.0009) and for hsCRP it was 2.25 (CI 1.27 to 4.00, p=0.0058). To illustrate the predictive performance of the biomarkers S100A12 and hsCRP, figure 2 shows Kaplan–Meier curves of the time from treatment withdrawal to disease flare of patients with high versus low S100A12 and hsCRP values, respectively. The corresponding Kaplan–Meier curve for MRP8/14 has been published previously.7 In addition to univariate survival analyses, all three biomarkers were included simultaneously in a multivariate model. Multivariate analyses revealed that, in the presence of S100A12 and hsCRP, MRP8/14 did not provide useful additional information for the prediction of the time to disease flare (p=0.1049). For the best combined marker ‘S100A12 or hsCRP’, the resulting HR of flare was 2.96 (CI 1.83 to 4.79, p<0.0001).

Figure 2

Kaplan–Meier curves of the time from biomarker analysis to disease flare. The graphs show a follow-up of 12 months after withdrawal of treatment, which was the time when the biomarker analysis was performed. The y-axis shows the proportion of patients without flares after discontinuing medication.

In further multivariate survival analyses, in addition to the molecular biomarkers S100A12, MRP8/14 and hsCRP, the demographic and clinical variables gender, age at disease onset, disease subtype (persistent oligoarthritis, polyarticular group, systemic JIA) and methotrexate treatment (6 vs 12 months after induction of disease remission) were included. The above results were confirmed qualitatively.

Discussion

This study provides a thorough comparison of predictive biomarkers for the risk of relapse of JIA in a cohort of patients from a controlled trial on treatment withdrawal in JIA remission.7 We hypothesised that three biomarkers (S100A12, MRP8/14 and hsCRP) might be useful for such prediction. It has previously been shown that MRP8/14 may provide valuable information about clinically occult disease activity.18 This result was confirmed in another study which concluded that MRP8/14 may indicate residual activity even in the absence of other laboratory or clinical signs of continuing inflammation.19 Finally, MRP8/14 was confirmed as a marker for the risk of relapse in JIA.7 For S100A12, on the other hand, cross-sectional data of 124 patients show that it may also indicate synovial inflammation even when other signs of arthritis are absent.20 Standard CRP was evaluated in all three of the above studies but no association was found with disease activity as it had to be below the detection level by definition to comply with the remission status. No data on hsCRP are available in these previous studies. The present study outperforms previous studies in that it was a prospective follow-up study with a prespecified objective of evaluating the predictive value of molecular biomarkers. Moreover, the number of patients was larger than in previous studies and, for the first time, all three molecular markers S100A12, MRP8/14 and hsCRP were evaluated simultaneously, thus enabling an adjusted multivariate analysis to determine which of the markers provides the best predictive value for the risk of disease flares.

This study revealed that each of the single phagocyte activation markers S100A12 and MRP8/14 provides good diagnostic performance in the prediction of disease flares if they are used separately. Both biomarkers are especially useful in predicting early flares within 3 months after treatment withdrawal, which confirms previous findings.7 The biomarker hsCRP applied separately does not provide reliable prediction of disease flares, but this result may be influenced by the study design. Patients were withdrawn from treatment and were included in the study only if ESR and/or standard CRP confirmed stable remission. This inclusion criterion led to a selection bias of patients roughly according to their CRP levels. Thus, patients with high CRP levels were not included since they did not meet the criteria for stable remission; only those with normal CRP levels entered the study and the follow-up phase. Patients with high hsCRP levels might have developed flares shortly after the biomarker analysis but the patients were not documented in the present study. Missing flares of patients with higher CRP levels in the study documentation may explain why the diagnostic performance (particularly the sensitivity) of the biomarker hsCRP in the prediction of early flares is low.

However, the acute phase reactant hsCRP seems to provide some independent additional information on later flares. This may be the reason why the predictive performance of the marker S100A12 and MRP8/14 was improved by adding hsCRP. The best combined marker panel (as shown by the best Youden index) was ‘S100A12 or hsCRP’. However, the results can be interpreted from different angles: the combination with the best trade-off between sensitivity and specificity is not the most restrictive with regard to the negative likelihood ratio. In clinical practice this could mean that the decision to withdraw treatment in these cases could put patients at higher risk of having a flare afterwards. Therefore, if a strategy to minimise the risk of flare is preferred, then the addition of the marker MRP8/14 could be considered. However, with regard to the number of patients who would potentially be unnecessarily treated further because at least one of the three markers would return a positive result despite stable remission, the combination of all three markers cannot be recommended.

Nevertheless, the rationale for using biomarker analyses is to base the decision to stop or continue treatment on indicators of immunological disease remission. However, no inflammatory marker test was 100% sensitive or specific. Therefore, the markers can only be used according to the specifications expected by the physician. For this reason and due to the limited number of patients, the above considerations need to be interpreted with caution. Physicians, patients and/or parents may still prefer a more restrictive or a more permissive strategy based upon personal experience.

There is a clear need to stratify patients with JIA according to their risk profile and the potential course of the disease. Therapeutic stratification is currently performed on the basis of disease categories and severity of inflammation.21 However, these approaches focus on the initiation of therapies. There is no doubt that we also need guidance at times of low disease activity or after reaching remission. It is important to adapt the intensity of anti-inflammatory treatment in order to balance the need for sufficient control of disease activity and the potential harm of therapeutics in the developing and growing child. In this regard, we also need to stratify patients for the need of maintenance treatment versus the option to stop all medications. In an attempt to allow approaches in patients who are successfully treated, an international consensus on JIA remission was achieved. According to these preliminary remission criteria, six continuous months of inactive disease on medication defines clinical remission on medication.6 ,14 It is unclear whether treatment should be continued in remission on medication for a variable time before stopping.

We now provide data showing that biomarkers of inflammation can be used to stratify patients at this point. In the presence of low phagocyte activation markers, the likelihood of a flare is low even after stopping treatment. On the other hand, subclinical disease activity may result in unstable remission—for example, a status of clinical but not immunological remission. As biomarkers of inflammation, phagocyte activation markers from the group of calcium-binding S100 proteins (plus hsCRP) have a close correlation with actual disease activity and therefore detect subclinical inflammation at the time of the analysis, if present. Applied as snapshot analysis, the biomarkers inform about the stability of remission and the risk of flares, especially within a time frame of a few weeks up to 3 months. This is also reflected by the reduction in the number of patients in remission at around 3 months after the biomarker test (figure 2). While, on the one hand, with the determination of subclinical activity the risk of flare over the following 3 months may be assessed, the patient is also protected to some extent for another 6–12 weeks after treatment withdrawal due to the long half-life of methotrexate. The fall in the number of patients with a higher risk profile due to subclinical disease activity at around 3 months after stopping methotrexate could therefore be explained by the fact that, at this time, the drug is no longer effective in preventing the flare, leading to a peak of relapses in patients with unstable disease remission around this time point. A limitation of our study is that it is only based on biomarker data from a single time point, at withdrawal of medication. It is therefore meaningful to repeat the analysis at intervals of approximately 3 months to monitor the immunological status of JIA patients (figure 3). An international multicentre study prospectively investigating the feasibility of treatment withdrawal decisions with a stratification based on the confirmation of immunological remission is planned.

Figure 3

Proposed stratification of patients in remission. Patients with juvenile idiopathic arthritis (JIA) in clinical remission have an elevated risk of flares after withdrawal of treatment if subclinical inflammation is present. This would be a status of clinical but not immunological remission. Since S100A12 (plus high-sensitivity C reactive protein (hsCRP)) may detect subclinical disease, a stratification of patients according to their immunological status is feasible. Adapting the criteria for JIA remission by Wallace et al,6 we propose that the decision to either stop treatment or continue with drugs that maintain clinical remission should be stratified based on the biomarker result. In cases of continued treatment, the maintenance treatment may contribute to immunological remission after a time. Thus, with biomarker analyses the stratification may be repeated over time with proposed intervals of 3 months. In the colour figure available in the online version of the article, red lines represent phases of inactive disease and green lines represent phases of inactive disease.

In conclusion, the neutrophil activation marker S100A12 and the phagocyte activation marker MRP8/14—possibly supplemented by hsCRP—indicate subclinical inflammation at the molecular level and thereby may identify patients with increased risk of relapse, even if clinical signs of disease activity are absent. Molecular markers of inflammation can be used to stratify patients at times of low disease activity and to optimise strategies of personalised medicine.

Acknowledgments

The authors thank Melanie Saers and Manfred Fobker for performing biomarker assays.

References

View Abstract

Footnotes

  • Funding This work was supported by a grant from the Interdisciplinary Centre of Clinical Research at the University of Muenster (IZKF CRA04) and FP7-Network PHARMACHILD. The clinical trial was supported by grants from the non-for-profit organisation PRINTO.

  • Competing interests None.

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

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.