Clinical remission and subsequent relapse in patients with juvenile idiopathic arthritis: predictive factors according to therapeutic approach

Juvenile idiopathic arthritis is the most common chronic rheumatic disease in childhood. This term comprises an heterogenous group of arthritis of unknown aetiology, each of which has differential genetics, etiopathogenesis, onset age and disease outcomes [1, 2]. It constitutes a significant cause of disability and quality of life impairment in JIA pediatric and adult patients [3‐5].

Recent advances in JIA treatment addressing more specific targets have led to better short and long-term disease outcomes. Evidence throughout the last decade has shown that an early and tight treatment to target approach of the disease increases the likelihood of achieving and sustaining clinical remission over time [6‐9]. Nevertheless, a notable proportion of patients relapse, either while still on medication or after its withdrawal [10‐12].

On the other hand, it seems reasonable to withdraw all medications in those patients that already have achieved CR to avoid the costs and possible adverse events derived of maintaining it over time. In this scenario, the main concern to most physicians is the occurrence of a relapse, which has shown to be frequent in most studies [13‐15]. In those patients who attain CR despite having a more aggressive course of the disease, maintaining tapered DMARD doses could increase the likelihood of sustaining a quiescent status of the disease over time, and avoid relapse after treatment withdrawal [9, 13, 15, 16].

No consistent evidence or guidelines are available regarding how or when to withdraw DMARD, or whether to do it abruptly or after the dose has been gradually tapered. Some studies have analysed the possible predictors of flare in patients in CR on and off medication, obtaining heterogeneous and conflicting results. Moreover, those predictors have not been validated to address therapeutic strategies in a real practice setting. This fact is partly explained by the heterogeneity within JIA categories, which consequently generates great variability in the definition and composition of study populations. In addition, evidence in literature has shown heterogeneity in clinical remission and inactive disease definitions, as well as significant variability in therapeutic strategies carried out by pediatric rheumatologists in JIA treatment [13, 15, 17, 18].

Few studies have analysed the prevalence and likelihood of CR regarding the therapeutic approach, which stands the aim of our study. Administering synthetic and biologic DMARD alone or combined is of significant relevance due to its influence on its tolerability, adherence, safety, efficacy and on subsequent disease outcomes [9, 19, 20].

We conducted a single-center, observational, retrospective-prospective study of JIA patients who were attended at Vall d’Hebrón University Hospital’s Pediatric Rheumatology Unit. This tertiary care hospital is a reference centre for JIA patients in Catalonia (Spain), where the mean JIA incidence was estimated to be of 6.9 per 100,000 children by 2010 [21].

A retrospective cohort was created on January 1st, 2012, in which patients diagnosed of JIA, treated with DMARD, and followed-up to December 31st, 2013, were included. Thereafter, a two-year prospective follow-up was performed from January 1st, 2014 to December 31st, 2015.Patients were included if: (a) the symptoms started before age 16 and lasted for at least 6 weeks, (b) fulfilled the ILAR Edmonton 2001 criteria [22], and (c) were receiving synthetic and/or biologic DMARD.

Patients were evaluated every 3 months according to study protocol and local clinical practice. Demographic, clinical, immunologic and treatment data were collected from paper charts and electronic medical records. Uveitis and other non-uveitis comorbidities were considered as separate variables taking into account the clinical relevance of JIA-associated uveitis. Non-uveitis comorbidities were defined as coexisting disorders or diseases, other than uveitis, and not causally related to drug toxicity. DMARDs were initiated in therapeutic doses according to patient’s body surface when an optimal control of disease activity could not be attained or when corticosteroids could not be withdrawn along the first 3 months from disease or follow-up onset. Treatment adjustments were performed at scheduled appointments or earlier, if needed, in flare occurrence. DMARD tapering was defined as performing either a dose reduction (below therapeutic dose) or an increase in the interval between doses. When dose tapering was not performed, the DMARD was abruptly withdrawn.

This study was done in accordance with Ethics Research Committees of Vall d’Hebrón University Hospital and Mútua Terrassa University Hospital, in the corresponding phases of the study. Written consent was obtained from patients and/or parents.

A total of 264 patients meeting the ILAR diagnostic criteria for JIA were included, of which 254 had received DMARD (synthetic or biologic). 206 patients were included in the retrospective study, which received 764 DMARD treatments. After 47 subsequent exclusions, 159 patients were included in the prospective study, which received 267 DMARD. At the end of follow-up, 136 patients remained on treatment, with 202 DMARD (Fig. 1).

The cohort was 71% female and 93% Caucasian, 52% had an oligoarticular JIA, 79% of which were persistent. Patients included had a median age at disease onset of 3 years [Inter quartile range (IQR) 4 years], at diagnosis of 4 years (IQR 5 years) and median disease duration at first DMARD onset of 4 years (IQR 8 years). Antinuclear antibodies (ANA) were detected in 68% of patients, and HLA B27 in 15% [67% of the enthesitis-related arthritis (ERA)]. Uveitis was present in 21% of patients (n = 43). Non-uveitis comorbidities were registered in 20% of patients (n = 41). Of these, 32% were other autoimmune diseases (n = 13), 32% genetic mutations or syndromes (n = 13), 10% chronic infections (n = 4), 10% congenital malformations (n = 4), and 16% miscellanea (n = 7). Thirty-two patients associated non-uveitis comorbidities with clinical significance (78%), being mild in 38% of the cases (n = 12), moderate in 34% (n = 11) and severe in 28% (n = 9).

Patients who achieved CR at least once throughout study follow-up (Table 1) were predominantly Caucasian (94%), had a persistent oligoarticular JIA (84%), did not associate non-uveitis comorbidities (87%), had not presented aprevious DMARD adverse event (75%), and had not required biologic DMARD (58%) (p < 0.05). When adjusting the analysis according to the JIA category, these results were confirmed for oligoarticular JIA. ERA patients achieved CR more frequently in the absence of non-uveitis comorbidities (88%, n = 15) (p = 0.035). The significant variables related to ever achieving CR obtained in the multivariate analysis were the absence of both non-uveitis comorbidities β 3.12 (CI 95% 1.5–6.51) (p = 0.002) and DMARD adverse event β 2.68 (CI 95% 1.42–5.07) (p = 0.002).

Clinical remission of the disease, based on Wallace criteria, was achieved at least once by 70% of the JIA patients included in the cohort along a median of 9-year follow-up, and was the main reason of DMARD withdrawal in both studies.

Greater CR rates were observed along retrospective study, which might be explained by the longer follow-up duration. In accordance with this observation, two studies have hypothesized, based on study results observed throughout last decade, that CR on medication increases with longer disease duration [9, 17].

In our study, patients were predominantly Caucasian females, with early JIA onset (< 5 years), oligoarticular persistent JIA, ANA positivity and early age at DMARD onset (< 6 years). MTX was the most frequently administered sDMARD and the most commonly administered alone.

Ever attaining CR was related to persistent oligoarticular subtype, absence of coexisting comorbidities (other than uveitis) and not having undergone an adverse event with DMARD therapy. These results were confirmed for oligoarticular JIA patients, among which CR predominated in those treated with Methotrexate alone and as first DMARD. Patients with oligoarticular JIA, particularly those with persistent subtype, have achieved the highest CR rates on and off medication in various studies, being treated in most cases with MTX as DMARD of choice [17, 25‐27].

On the other hand, coexisting comorbidities in JIA patients, including other autoimmune diseases, genetic syndromes, or chronic infections, among others, could lead to both: a higher occurrence of DMARD toxicity and inefficacy. Raab et al. studied associated comorbidities in young adult JIA patients of the JuMBO register and concluded that comorbidities have a significant impact on disease activity and global health status [28]. Furthermore, Simon et al. have recently put in evidence that patients with JIA are more likely to have concurrent autoimmune diseases, and this association might have an important role in JIA treatment decisions and outcomes [29].

Lastly, the negative influence of DMARD toxicity in ever achieving CR in JIA patients could be explained in those having a more aggressive course of the disease. In these, higher DMARD doses and therapeutic targets are combined, leading to a greater likelihood of DMARD withdrawal due to drug toxicity and subsequent poorer control of disease activity.

According to therapeutic approach, CR was observed in patients treated with synthetic DMARD alone more than half of the times, being Methotrexate the most administered sDMARD and being CR the predominant reason of its withdrawal. These findings support that Methotrexate is still the DMARD of choice in most JIA categories, even in the biologic era [10, 30, 31]. Within this group CR was associated with female sex, persistent oligoarticular subtype, absence of comorbidities (other than uveitis) and Methotrexate treatment as first DMARD course. ANA positivity, the absence of HLA B27, uveitis and DMARD toxicity, and having performed DMARD dose tapering were related to higher CR rates within all treatment groups, but more significantly to patients treated with sDMARD alone. Furthermore, ANA association with CR was confirmed for oligoarticular and psoriatic subtypes and HLA B27 for oligoarticular and ERA subtypes in those patients treated with sDMARD alone. However, the results must be considered cautiously, since the number of patients in some categories was scarce.

The role of ANA positivity in the likelihood of achieving and sustaining CR, and therefore, in short and long-term outcomes, has not yet been clarified. Some studies suggest that ANA negativity would lead to a poorer response to MTX throughout the first 6 months of the disease, a higher active joint count along the first 6 to 24 months, regardless on active joint count, and to a persistent active disease through the first 3 years from JIA onset [25, 32‐34]. The possible association between ANA positivity and a better MTX response would explain that oligoarticular patients included in the study, and not having coexistent comorbidities (uveitis or others), would have solely required treatment with Methotrexate alone to attain CR. These results would support what has been recently suggested in two studies which showed that JIA patients with early onset, ANA positivity and lower active joint count, constitute a homogeneous JIA subtype with less aggressive course of the disease, higher likelihood of achieving CR and subsequently better outcome [35, 36]. Nevertheless, our results contravene the results obtained in two other studies which suggest that ANA status does not alter remission rates nor JIA outcome [37, 38].

Regarding HLA B27, several studies have put in evidence that HLA B27 presence in JIA patients leads to a higher risk of bone erosions and radiologic structural damage appearance [39, 40], a more persistent and extended course of the disease [41, 42], and a lower likelihood of achieving clinical remission [12, 25].

Dose tapering was performed in two thirds of patients in CR on medication. Relapse throughout dose tapering was scarce in both studies. Two thirds of DMARD treatments administered along CR of the disease were withdrawn throughout our retrospective study, with a subsequent relapse observed in two thirds of the cases, after a median time to relapse of about 12 months with comparable results in all treatment groups. Significant lower rates of relapse after withdrawal were observed in patients with persistent oligoarticular JIA, treated with sDMARD alone (mostly Methotrexate) in which the drug dose was previously tapered. Nonetheless, no differences in method, order (sDMARD or bDMARD first) or duration (median longer than 12 months in both studies) of dose tapering were observed within treatment groups in both studies, and neither was associated with CR of the disease nor relapse after withdrawal. Regarding patients in CR on medication, our therapeutic strategy is comparable to the one performed by most rheumatologists participating in two survey studies, who would wait at least 12 months while in CR on medication to start considering DMARD tapering or withdrawal [14, 15]. Furthermore, given that CR was more frequently achieved in patients treated with MTX alone in our study, these results would support that maintaining MTX at least for 12 months after achieving CR before withdrawal is associated with a higher likelihood of maintaining remission off treatment, as suggested by Klotscheet al [43]. One of the limitations of our study is that the exact date of clinical remission start was unavailable in most patients, and therefore, total duration of CR on treatment could not be measured. Duration of CR on medication remains currently of significant interest given that its association with relapse after treatment discontinuation has shown conflicting results [13].

When the study finalised after a median follow-up time of 9 years, more than two thirds had attained CR at least once, and nearly one third of patients were in clinical remission off all medication, which is slightly lower than the 42% observed in Nordalet al. study [44]. Considering that in our study persistent oligoarticular JIA is the most prevalent JIA subtype, we might have underestimated CR rates through follow-up losses, given that patients lost to follow up would more likely be in remission than those who maintain it.

On the other hand, two thirds were still on DMARD therapy. It should be noted that in those patients treated with sDMARD and bDMARD combined, tapering duration and number of times the dose was tapered were greater, whereas subsequent withdrawal and following flares were less frequently observed in the prospective study. This fact would support the recently growing belief that in some patients, especially those with persistent and extended poliarticular RF positive, systemic or ERA JIA, the disease behaves more aggressively, needing more specific therapeutic targets and a sustained treatment over time, even when CR is achieved. In those scenarios, maintaining a minimum and progressively tapered dose for at least 12 months (before even considering withdrawal) would appear to be a good therapeutic approach to minimize subsequent relapses.

The main limitation of our study is its observational mono-centric and non-controlled design. This fact might have led to selection bias, having a lower number of patients included in a predominantly oligoarticular JIA cohort. Consequently, the depicted results should be applied cautiously to other JIA categories. That being said, the coincidence of the results obtained in both retrospective and prospective studies grants coherence to the conclusions provided. Lastly, this study fails to describe total duration of CR both on and off medication as the exact date of CR onset was unavailable in most patients. Clinical remission’s survival in JIA patients remains an issue of great concern given the complexity of establishing and defining when CR itself begins.

More than two thirds of JIA patients attained CR along the 9 years of follow-up, and nearly one third achieved CR off medication. Females with early JIA onset, lower active joint count and ANA positivity were the ones achieving and sustaining remission more frequently, especially when receiving synthetic DMARD alone and in the absence of HLA B27, comorbidities or previous DMARD toxicity. In patients who already have attained remission, a progressive DMARD dose tapering for at least 1 year before considering withdrawal would minimize the probability of subsequent relapse, particularly in those who required a biologic DMARD to achieve a quiescent state of the disease.