The risk of depression and anxiety is not increased in individuals with juvenile idiopathic arthritis – results from the south-Swedish juvenile idiopathic arthritis cohort

Juvenile idiopathic arthritis (JIA) is a chronic inflammatory joint disease affecting children and adolescents under the age of 16, causing joint swelling with mobility limitations and pain as possible short-term and long-term consequences [1]. The mean incidence of JIA 2002 – 2010 was 12.8/100 000 children in the south-Swedish region of Skåne [2], with reported numbers internationally ranging from 3.1 – 15.0/100 000 children [3‐8].

According to the definition from the International League of Associations for Rheumatology (ILAR), JIA serves as an umbrella term for seven subcategories, defined by the number of inflamed joints, additional organ involvement and presence of systemic features. Six of the subcategories have adult equivalents and only oligoarticular JIA, where the majority of patients have antinuclear antibodies (ANA) and disease onset before the age of six, is very uncommon in adults [9]. Since the subcategories of JIA includes patients with heterogenous as well as overlapping phenotypes, such as ANA-positivity, early onset and asymmetrical joint involvement, there are suggestions of a new classification system where ANA positive children with early onset of arthritis (irrespective of joint count) should be a defined subtype. This subtype is specific for arthritis in children. The suggestions also include other refinements of the subtypes to make the pediatric arthritides further resemble their adult counterparts [10, 11].

Although JIA is considered a childhood disease, 45.6% of the patients in the Nordic JIA cohort still had active disease 18 years after diagnosis [12] and 41% of the patients still had active disease or needed pharmacological treatment in a Norwegian study 30 years after diagnosis [13]. The pharmacologic treatment of JIA has evolved considerably during the last decades, especially with the introduction of biological disease-modifying antirheumatic drugs (DMARDs) in the start of the twenty-first century. Before biological DMARDs (bDMARDs) were available, other synthetic DMARDs such as gold, sulfasalazine and chloroquine phosphate were used and are today termed conventional synthetic DMARDs (csDMARDs) [1]. Methotrexate is today the most widely used csDMARD and was introduced as treatment for juvenile arthritis in Sweden in the mid 1980s.

Studies have demonstrated that children with chronic diseases are more likely to develop psychiatric comorbidities than the general population [14, 15]. In recent years more attention is given to the possible effect of inflammation for the pathophysiology of mood-disorders and anxiety. In a stressful situation, secreted stressors activate inflammatory pathways in circulating mononuclear cells, leading to increased levels of pro-inflammatory cytokines that through different mechanisms exert their effects on the central nervous system, for example by decreasing the concentration of monoamines [16, 17]. Clinical symptoms of depression have been reported in 7 – 36% of children with JIA and 7 – 64% reported significant anxiety during childhood [18].

In the study area of southern Sweden, the prevalence of depression in the general population in 2019 was 1.09%. Prevalence for anxiety was 1.04%, with female predominance (sex ratio 3:2 and 2:1 respectively) [19].

Swedish children with the chronic, autoimmune disorders childhood-onset inflammatory bowel disease (IBD), type one diabetes mellitus and celiac disease have significantly increased risk for mood-disorders and anxiety compared to the general population in long-term follow-up studies [20‐22]. Since studies of psychiatric comorbidities in JIA report conflicting results, the risk of depression and anxiety for individuals with JIA is not known.

By using a population-based cohort of individuals diagnosed with JIA 1980 – 2010 with a follow-up time up to 39 years, our main aim was to investigate if individuals with JIA have increased risk of developing depression or anxiety as compared to references without JIA. We also aimed to investigate if JIA individuals are diagnosed with depression and anxiety at an earlier age as compared to sex- and age-matched references without JIA, and if the risk of depression and anxiety differed between the defined groups of JIA patients; ANA-positive disease with early onset, and DMARD treated.

In this study we showed that individuals with JIA in our population based south-Swedish cohort did not have an increased risk of being diagnosed with depression or anxiety compared to age- and sex matched references. Neither did subgroup-analyses, based on disease onset or DMARD treatment, show any increased risk.

Other studies have shown that children with chronic diseases, inflammatory as well as non-inflammatory, are more likely to develop psychiatric comorbidities than the general population [14, 15, 20‐22]. There was therefore reason to believe that this also applied to JIA. Indeed, a chronic pediatric disease may serve as a stressor as the child or adolescent must cope with managing the disease alongside with normal bodily and behavioural changes growing up and since JIA often has a remitting course, also in periods of disease remission patients are psychologically under the threat of a flare. In addition, the cardinal symptom of JIA is arthritis, a condition causing pain and limited motion. For some patients, systemic symptoms of inflammation such as fatigue and malaise are also present. Pain and fatigue are factors known to negatively affect quality of life [18, 27], possibly in part due to reduced physical activity and social exclusion. Presence of depressive symptoms can also further aggravate pain and disability in JIA [28]. However, possible protective factors may be present in our study-cohort compared to in individuals without JIA, such as team-centred care for patients with rheumatic diseases comprising preventive psychologic counselling to develop coping strategies for a life with chronic disease and easy access to psychological support in times of despair. A rheumatologist might also not register a diagnosis of depression or anxiety in a patient expressing symptoms of malaise or unease, rather interpreting this as crisis reaction after diagnosis or an effect of chronic pain, which differs in individuals without a chronic diagnosis. Our results do not exclude the possibility of individuals with JIA having more distress symptoms or that the potential impact of these symptoms have greater impact on their quality of life than in the references population, since our study does not include patient reported data and symptoms. Follow-up of the psychological well-being should still be carried out in the JIA patient group.

Pediatric rheumatologists today work with “treat-to-target” strategies with frequent visits and treatment adjustments to achieve remission as fast as possible. The goal is to keep periods with inflammatory activity short to prevent adverse somatic and psychiatric effects. Prior to the introduction of methotrexate and bDMARDs, the inflammatory activity in JIA was often difficult to effectively reduce. One could speculate that modern and more “aggressive” treatment strategy influences the risk of psychiatric co-morbidities, as it is suggested that chronic inflammation may serve as a stressor for psychiatric disease [16, 17]. It would have been interesting to compare measures of disease activity to the presence of depression or anxiety in our study, but unfortunately, we did not have active joint count and laboratory markers of inflammation in all cases. We did however choose to study the risk of depression and anxiety in subgroups of individuals in need of DMARD treatment, since this may serve as surrogate marker for more severe disease and high inflammatory activity, but in neither of these subpopulations the risk was different compared to matched references. However, there are other reasons than high inflammatory activity to start DMARD treatment, and the lack of a good marker for disease activity is a limitation to our study.

The subgroup of ANA-positive disease with early onset is especially interesting as this diagnostic subgroup is believed to have unique pathophysiological mechanisms and is typical for pediatric onset arthritis. Many of the patients with this subtype have oligoarticular disease, which is associated with good prognosis of remission and treatment response [1]. Nevertheless, with symptom debut before the age of six, these patients have the longest disease duration, which was associated with the development of psychiatric illness in the study by Kyllönen et al. [29]. However, we did not find such an association in our study and there was no difference in median age at diagnosis with depression or anxiety between individuals with JIA and references, which further strengthens that disease duration have little impact to the risk for depression or anxiety in JIA.

The finding of increased HR for depression and anxiety in the group of individuals with diagnosis/study inclusion 1980 – 1990 should be interpreted with caution since it might have other explanations than change of treatment regime. We have studied the cumulative incidence and since the risk of being diagnosed with depression or anxiety increases with age and there is greater awareness of depression and anxiety in the general population as well as at centres for primary care 2019 compared to 1998, the oldest individuals in this study are the ones with the highest risk of comorbid diagnosis. Another possible explanation to why the HR decreases over time in, at least in the female subgroups, is that symptoms of depression and anxiety might be interpreted as a reaction to the JIA diagnosis during the first years of disease, while they are considered and treated as a true diagnosis later in the disease course.

Our study has some limitations. In the creation of the cohort, we only had full regional coverage of inpatient visits because registration of ICD codes for outpatient visits were not mandatory in Sweden until 2002. We had access to outpatient visits from Lund University Hospital (2010 renamed as Skåne University Hospital) for the total study period. Thus, there is the potential risk of us having missed to include JIA cases with mild disease diagnosed prior to 2002 with only outpatient visits at healthcare facilities other than the Lund University Hospital. However, we believe that number to be limited since most children needed inpatient care for rehabilitation and treatment in the 1980’s, and intra-articular corticosteroid injections in younger children were administered during inpatient care in the 1980’s and 90’s.

We only had access to registered ICD codes for depression and anxiety since 1998. This may result in a possible overestimation of incident cases since we did not exclude individuals that might have had such a diagnosis before the study period and hence have an increased risk of a repeated diagnosis. The references are selected on having at least one healthcare consultation during follow-up. Thus, we might also overestimate the existence of depression and anxiety in the control group. The lack of data may on the other hand underestimate the frequency of psychiatric comorbidities since we lack 18 years of data. The subgroup analysis among the individuals diagnosed with JIA between 1998 – 2010 did in fact find a decreased HR for anxiety in JIA males, as in the subgroup analysis based on year of JIA diagnosis 2001—2010. However, the cases are few, CI is wide and just below the significance level, and the clinical relevance of this finding is thus questionable.

The SHR did not have full coverage of diagnosis codes registered in outpatient care, mainly from primary care, until 2004 which also contributes to the risk of underestimation. However, the presence of depression and anxiety as well as the cumulative incidence of these conditions in our study is higher, both in the JIA cohort and references, than reported numbers in other Swedish studies on comorbid psychiatric disorders in childhood-onset IBD (6.6% mood disorders and 10.4% anxiety) and celiac disease (6.2% mood disorders and 8.2% anxiety) [20, 21]. In these studies, they considered one registered ICD code as a true comorbid event and did also not validate the diagnosis through medical record review or personal interview. Therefore, the requirement in our study for two codes if registered only in outpatient care would rather have reduced the number of cases. However, the above-mentioned studies had shorter median follow-up time, 9 and 12.2 years respectively, and did not have access to ICD codes registered from Swedish primary care facilities, where mild and moderate variants of depression and anxiety often are treated. Our increased presence and cumulative incidence can partly be explained by the longer follow-up time (13.3 years for females and 15.1 years for males) and inclusion of primary care diagnoses, at least since 2004.

There are also strengths to our study. We have a population-based cohort with validated diagnoses of JIA in a well-defined region of Sweden, thus including all severities of the disease from mild and self-limiting to remitting and functional disabling disease course compared to the Finnish study by Kyllönen et al. [29]. Most other published studies on mental health in JIA patients have cross-sectional study design and patient reported outcome [18, 28, 30]. Our study is longitudinal with follow-up up to 39 years after JIA onset and based on diagnosis assessed by medical care givers.

In conclusion, we have studied the long-term risk of depression and anxiety in a south-Swedish population-based cohort of individuals with JIA diagnosed 1980 – 2010. This is, to our knowledge, the largest study with longitudinal follow-up of psychiatric comorbidities in JIA. Individuals with juvenile arthritis do not appear to be diagnosed with depression or anxiety at an earlier age or more often than sex- and age-matched references. We neither found any increased risk in subgroups with longer disease duration or more severe disease. Thus, we conclude that growing up with inflammatory arthritis do not appear to increase the risk of being diagnosed with depression or anxiety disorders later in life.