Article Text

Download PDFPDF

Original research
Adipokine profile in long-term juvenile dermatomyositis, and associations with adipose tissue distribution and cardiac function: a cross-sectional study
  1. Henriette Schermacher Marstein1,2,3,
  2. Birgit Nomeland Witczak1,3,
  3. Kristin Godang4,
  4. Nicoleta Christina Olarescu4,5,
  5. Thomas Schwartz1,3,
  6. Berit Flatø5,6,
  7. Øyvind Molberg6,7,
  8. Jens Bollerslev4,5,
  9. Ivar Sjaastad1,3,8 and
  10. Helga Sanner2,6
  1. 1Institute of Experimental Medical Research, Oslo University Hospital Ullevaal, Oslo, Norway
  2. 2Department of Health Sciences, Oslo New University College, Oslo, Norway
  3. 3KG Jebsen Center for Cardiac Research, Oslo University Hospital, Oslo, Norway
  4. 4Section of Specialized Endocrinology, Department of Endocrinology, Morbid Obesity and Preventive Medicine, Medical Clinic, Oslo University Hospital, Oslo, Norway
  5. 5Institute for Clinical Medicine, Medical Faculty, University of Oslo, Oslo, Norway
  6. 6Department of Rheumatology, Oslo University Hospital, Oslo, Norway
  7. 7Faculty of Medicine, University of Oslo, Oslo, Norway
  8. 8Department of Cardiology, Oslo University Hospital, Oslo, Norway
  1. Correspondence to Henriette Schermacher Marstein; henriette.marstein{at}


Objectives In long-term juvenile dermatomyositis (JDM), altered adipose tissue distribution and subclinical cardiac dysfunction have been described. Our aims were to compare adipokine levels in patients with JDM after long-term disease with controls, and explore associations between adipokines and (1) adipose tissue distribution and (2) cardiac function.

Methods The study cohort included 59 patients with JDM (60% female, mean age 25.2 years, mean disease duration 16.9 years), and 59 age/sex-matched controls. Updated Pediatric Rheumatology International Trials Organization criteria for clinically inactive JDM were used to stratify patients into active (JDM-active) or inactive (JDM-inactive) disease groups. Lipodystrophy was clinically assessed in all patients. In all study participants, we measured adipose tissue distribution by dual-energy X-ray absorptiometry and cardiac function by echocardiography. Serum adipokines (adiponectin, apelin-12, lipocalin-2, leptin, visfatin and resistin) were analysed using ELISA.

Results Patients with JDM had higher leptin levels compared with controls (p≤0.01). In JDM-active, apelin-12 and visfatin were higher compared with JDM-inactive (p≤0.05). In JDM-total and JDM-active, lower adiponectin correlated with lipodystrophy and total fat mass. Also, systolic dysfunction correlated with: lower adiponectin in JDM-total, JDM-inactive and JDM-active, and with lower apelin-12 in JDM-total and JDM-active and resistin in JDM-active (all p≤0.05). Lower adiponectin correlated with diastolic dysfunction in JDM-total and JDM-active.

Conclusion After long-term disease, leptin levels were unfavourably regulated in patients with JDM compared with controls, and apelin-12 and visfatin in JDM-active versus JDM-inactive. We found associations between adipokines and both adipose tissue distribution and cardiac systolic function in all patients with JDM, which was most prominent in patients with active disease.

  • autoimmune diseases
  • cardiovascular diseases
  • dermatomyositis
  • inflammation
  • cytokines

Data availability statement

Data are available upon reasonable request. The data that underlie this article cannot be shared publicly due to the privacy of individuals who participated in the study. The data will be shared if a reasonable request is made to the corresponding author.

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:

Statistics from

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.


  • After long-term disease, patients with juvenile dermatomyositis (JDM) show signs of subclinical cardiac dysfunction and unfavourable adipose tissue distribution.

  • In JDM, little is known about adipokine profile and associations with adipose tissue distribution and cardiac function.


  • We here showed that after long-term disease, patients with clinically active JDM had higher levels of apelin-12 and visfatin compared with patients with inactive disease.

  • Cardiac systolic function was positively associated with apelin-12 in all patients with JDM, and more strongly in patients with clinically active disease.


  • Our findings indicate involvement of the adipokines: leptin, apelin-12, visfatin and adiponectin in JDM, other than a relation to obesity, which should be studied further.


Multiorgan pathologies are found in idiopathic inflammatory myopathies (IIMs) including juvenile dermatomyositis (JDM). The main characteristics of JDM are skin rashes and proximal muscle weakness, probably caused by vasculopathy.1 Furthermore, the heart can be affected although mostly subclinical.1 Another frequent consequence to JDM is altered adipose tissue distribution with the presence of lipodystrophy2 and increased abdominal adipose tissue3 as recently reported from our study group.4

Adipokines are cytokines produced mainly in adipose tissue cells, with an effect on fat distribution, blood pressure, metabolism and vascular homeostasis.5 Adipose tissue consists of adipocytes and immune cells, and when expanding, adipocytes increase in size and more immune cells accumulate. In this context, both cell types generate an inflammatory milieu due to altered proinflammatory cytokine release.6 In the general population, most adipokines such as resistin,7 apelin-12,8 visfatin,9 leptin10 and lipocalin-2 (also known as neutrophil gelatinase-associated lipocalin (NGAL)11 12 increase in obesity, with the exception of adiponectin,13 which is decreased. In autoimmune diseases, the situation might be more complex. In fact, there is evidence that adipokines associate with disease activity and systemic inflammation independent of total body mass index (BMI),14–19 also shown in JDM.14 16 However, to our knowledge, adipokine levels and their relation to the adipose tissue distribution have not been studied in long-term JDM.

In the general population, it is well established that obesity-related inflammation is associated with cardiovascular diseases with impaired cardiac contractile (systolic) and relaxation (diastolic) function, especially in individuals with central adiposity.5 Subclinical cardiac diastolic dysfunction has been found in adult-onset DM20 21 after short-term disease. Both diastolic and systolic dysfunction (mostly subclinical) have been reported in our JDM cohort of adults and children clinically examined after long-term disease.22 23 Studying the same cohort of patients with JDM and controls, we recently described an altered body composition with increase in body fat percentage and abdominal fat accumulation (this included increased android to gynoid fat mass ratio and visceral adipose tissue (VAT)) with associations with cardiac and metabolic dysfunction.24 25

In view of the unfavourable adipose tissue distribution associated with impaired cardiometabolic function in long-term JDM, we aimed to explore the role of adipokines in this clinical setting. Aims of the study were to compare the adipokine profile in patients with JDM versus controls and in patients with active versus inactive disease. Further, we aimed to explore associations between the selected adipokines and adipose tissue distribution as well as cardiac systolic and diastolic function.

Patients and methods

Study population

In this cross-sectional study, the eligibility criteria for the Norwegian nationwide JDM cohort have been described in detail.4 Briefly, patients were included if they had age <18 years at disease onset, age ≥6 years at the follow-up examination, disease duration >24 months, and a probable or definitive diagnosis of DM according to the Bohan and Peter criteria.26 These inclusion criteria were met by the 66 patients diagnosed in Norway between January 1970 and June 2006. Four of those were deceased and 62 were tracked through the National Population Register; 59 of 62 (95%) patients participated in the JDM cohort study.4 Sex and age-matched controls (1:1) were randomly drawn from the National Population Register as previously described in detail.27 Exclusion criteria for the controls were among others cardiac and other autoimmune diseases including JDM.27

Data collection and clinical measurements

Patients with JDM and controls were comprehensively clinically examined during a follow-up examination visit of 1–2 days at Oslo University Hospital (OUS) (September 2005–May 2009). The examination included echocardiography, dual-energy X-ray absorptiometry (DXA) scanning for body composition and blood sampling as previously described in detail.4 23 24 Muscle strength/endurance was assessed by unilateral Manual Muscle Test 8 (MMT-8) (0–80) and Childhood Myositis Assessment Score (CMAS) (0–52). In the patients, we assessed disease activity by the Disease Activity Score for JDM (0–20) and cumulative organ damage by the Myositis Damage Index (0–40).4 Clinically inactive disease was defined according to updated Pediatric Rheumatology International Trials Organization (PRINTO) criteria as the presence of physician’s global assessment of overall disease activity ≤0.2, and at least three of the following: (a) serum creatine kinase ≤150 units/L, (b) MMT-8 ≥78 and (c) CMAS ≥48.28 Patients with clinically inactive disease were defined as JDM-inactive and the remaining were defined as having active disease or JDM-active. A self-reporting questionnaire was used to assess average physical activity inducing sweating or breathlessness and reported as physical activity hours/week.

Serum analyses

Non-fasting blood samples were drawn at the follow-up examination. Lipid profile and erythrocyte sedimentation rate were measured at the accredited medical biochemistry laboratory, Department of Medical Biochemistry, OUS, Rikshospitalet according to a standard protocol. The remaining serum blood samples were stored in smaller batches at −80°C.

Adipokines were analysed in serum batches in 2014 by ELISA according to the manufacturer’s protocol. The chosen adipokines were: resistin (DY1359), adiponectin (DY1065), leptin (DY398), lipocalin-2/NGAL (DY1757) (all R&D Systems Minneapolis, Minnesota, USA), visfatin (EK-003-80) and apelin-12 (EKE-057-23) (both from Phoenix Pharmaceutical Burlingame, California, USA). Samples from all patients and controls were analysed at the same time to minimise the run-to-run variability.

Adipose tissue distribution

We analysed adipose tissue data retrieved from the original DXA scans which were performed at the follow-up examination. For this study, the data were reanalysed in the same software version, as described earlier.24 In brief, in all study participants, total fat mass was measured by a narrow fan-beam densitometer scan according to the manufacturer’s protocol (GE Healthcare Lunar Prodigy Advance, Madison, Wisconsin, USA). VAT was quantified in the android region of the trunk using the application (CoreScan; GE Healthcare). The android region of interest (ROI) contained both VAT and subcutaneous adipose tissue (SAT). The software estimated the quantity of SAT in the android ROI, and the amount of VAT was computed by subtracting SAT from the total android fat in study participants ≥18 years (38 patients and 36 controls) and is presented as total mass (kg).25

Cardiac function measures

Cardiac function was assessed by echocardiography (Vivid 7 ultrasound scanner, GE-Vingmed Ultrasound, Horten, Norway) as described earlier.23 Mitral annulus (MA) displacement was assessed by anatomical M-mode in lateral and septal position using the four-chamber apical view. Left ventricular (LV) length was assessed as the distance from the endocardial border of the apex to the mid-portion of the MA plane. Systolic function was measured as long-axis strain (LAS) and calculated as MA displacement expressed as a percentage of LV end-diastolic length. Diastolic function was recorded as early diastolic tissue velocity (e’).22


For normally distributed continuous data, differences between patients and controls were tested by paired samples t-test, and differences between JDM subgroups by independent samples t-test. The Χ2 test was used for categorical data. We used Spearman’s or Pearson’s correlation coefficients to determine correlations when appropriate. Strengths of correlations were defined as weak rsp=0.1–0.30, moderate rsp=0.31–0.60 and strong rsp=0.61–1.0; only moderate and strong correlations were presented and discussed. P values of ≤0.05 were considered statistically significant. Due to the hypothesis-generating nature of the study, we did not correct for multiple comparisons.


Characteristics and disease variables in patients

Mean (SD) age of patients at follow-up was 25.2 (12.5) years and the mean disease duration 16.9 (10.6) years (table 1). According to the updated PRINTO criteria, 21 (35.6%) had inactive disease and 38 (64.4%) had active disease (table 1). All the 10 (16.9%) patients who received prednisolone at follow-up were JDM-active (table 1). Patients had lower low-density lipoprotein-cholesterol, total cholesterol and high-density lipoprotein-cholesterol than controls, while triglyceride was higher in patients than controls (table 1). Further, both systolic/LAS% and diastolic/e’ function were reduced in patients versus controls (p≤0.01) (table 1). Lipodystrophy was found in 10 (16.9%) patients and none of the controls. VAT was measured in all study participants ≥18 years and was higher in patients compared with controls (p≤0.05) (table 1). There were no significant differences in total fat mass, VAT, cardiac function, or lipid levels between JDM-active and JDM-inactive patients.

Table 1

Characteristics and disease variables in patients with JDM at follow-up and controls, and in patients with active and inactive disease

Serum adipokine levels in study participants

Leptin levels were higher in JDM-total than controls (p≤0.05) (table 2). In JDM-active, visfatin and apelin-12 were both higher compared with JDM-inactive (p≤0.05 and p≤0.01, respectively) (table 2). When stratifying JDM-total according to the occurrence of lipodystrophy, patients with lipodystrophy had lower adiponectin levels compared with patients without (mean (SD) 2.4 µg/mL (2.8) vs 5.4 µg/mL (1.7); p≤0.01).

Table 2

Adipokines in patients with JDM and controls, and in patients with active or inactive disease

Correlations between adipokines and fat distribution variables in study participants

In JDM-total and JDM-active, adiponectin correlated moderately and negatively with the occurrence of lipodystrophy and with total fat mass (table 3). Also, there was a moderate and negative correlation between adiponectin and VAT in JDM-total and controls, and more strongly in JDM-active (table 3). Further, lipocalin-2 correlated moderately and positively with total fat mass in JDM-active and controls and with VAT in JDM-inactive (table 3). Resistin correlated moderately and positively with total fat mass in JDM-total and JDM-inactive (table 3). In JDM-inactive, both apelin-12 and visfatin correlated moderately and negatively with VAT. A similar correlation between apelin-12 and VAT was found in controls (table 3). Leptin correlated strongly and positively with total fat mass in all groups (table 3).

Table 3

Correlations between adipokines and fat measures in patients with JDM and controls, and in patients with active and inactive disease

Associations between adipokines and cardiac function in study participants

In JDM-total, we found both adiponectin and apelin-12 to correlate moderately and positively with LAS% (systolic function) (figure 1A,B). No correlations were found between LAS% and lipocalin-2 and resistin in neither JDM-total nor controls (figure 1C,D). When correlating adipokines with LAS% in JDM-active patients, we found moderate positive correlations with adiponectin and apelin-12, and moderate negative correlations with lipocalin-2 and resistin (figure 1E,H). Also, in JDM-inactive, we found a moderate positive correlation between LAS% and adiponectin (figure 1E). In both JDM-total and JDM-active, adiponectin correlated moderately and positively with e’ (diastolic function) (figure 2A,B).

Figure 1

Correlations between systolic function and adipokines. (A–D) Correlations in all patients with juvenile dermatomyositis (JDM-total) compared with controls. (E–H) Correlations in patients with active compared with patients with inactive disease. (A and E) adiponectin, (B and F) apelin-12, (C and G) lipocalin-2 and (D and H) resistin. *P≤0.05, **p≤0.01. LAS, long-axis strain; rsp, Spearman’s correlation coefficient.

Figure 2

Correlations between diastolic function and adiponectin. (A) Correlations in all patients with juvenile dermatomyositis (JDM) compared with controls. (B) Correlations in patients with active versus inactive disease. *P≤0.05, **p≤0.01. e’, early diastolic tissue velocity; rsp, Spearman’s correlation coefficient.


Our study is the first to demonstrate associations of circulating adipokines with measures of adipose tissue distribution and cardiac function in JDM. After long-term disease duration, patients with JDM had higher levels of leptin compared with controls, and in patients with JDM-active, apelin-12 and visfatin were higher compared with JDM-inactive disease. Lower levels of the anti-inflammatory adipokine adiponectin correlated with unfavourable fat distribution including the presence of lipodystrophy, increased VAT and increased total fat mass in both JDM-total and JDM-active. Higher adiponectin and apelin-12 exhibited potential cardioprotective properties in JDM, showing positive correlations with cardiac systolic function. Adiponectin was also positively correlated with diastolic function. In contrast, higher resistin and lipocalin-2 correlated with impaired cardiac systolic function, only in the JDM-active group.

The obesity-related apelin-12 was higher in JDM-active compared with JDM-inactive, although there were no significant differences in apelin-12 between JDM-total and controls. To our knowledge, this is the first time apelin-12 is studied in any autoimmune disease. Interestingly, under hypoxic conditions, cellular apelin-12 secretion increases and subsequently stimulates angiogenesis.29 30 Since vasculopathy with concomitant hypoxia is a hallmark of JDM,1 increased levels of apelin-12 in patients with active disease might be a consequence of this compensatory mechanism.

Higher levels of the proinflammatory visfatin were found in JDM-active than in JDM-inactive and higher levels of leptin found in JDM-total compared with controls. Both visfatin and leptin have been found increased in patients with IIM16 17 and in patients with rheumatoid arthritis (RA),18 31 all compared with controls. In mice, leptin deficiency protects against development of various autoimmune diseases, including systemic lupus erythematosus and RA.32 Also, a positive correlation between leptin and C reactive protein (CRP) in non-obese healthy individuals has been shown.33 Further, visfatin has been associated with disease activity in patients with IIM,16 and might be involved in regulating angiogenesis in a non-beneficial cardiovascular manner.34 Interestingly, the two adipokines have several immune-modulatory effects.34 As such, visfatin and leptin might be involved in development or maintaining inflammation and/or disease activity in JDM; however, we have no data to support this notion.

We found lower levels of the anti-inflammatory adiponectin to be associated with higher VAT, more total fat mass and the presence of lipodystrophy in JDM-total, as well as in the JDM-active group. Also, the level of adiponectin was lower in patients with lipodystrophy than in patients without lipodystrophy. It is well established that adiponectin decreases with visceral fat accumulation.6 13 Importantly, lower adiponectin could also be a result of reduced adipose tissue mass as found in patients with generalised lipodystrophy,35 which might also be the case for the patients with JDM in our study.

Higher resistin correlated with increased total fat mass in JDM-total and in JDM-inactive. A positive correlation was found between resistin and measures of fat mass in younger obese subjects when looking at changes over time.36 Interestingly, studying human cells, obese adipocytes secreted more resistin compared with lean adipocytes and subsequently contributed to impaired myogenesis, especially in older muscle cells.7

Lipocalin-2 correlated moderately and visfatin negatively with VAT in JDM-inactive. Lipocalin-2 has been found to correlate with both BMI and CRP in overweight to obese individuals,12 which indicates an involvement of lipocalin-2 in low-grade inflammation. Associations between visfatin and measures of adipose compartments are inconclusive.9 37 As such, we have no explanation for our finding in JDM.

As expected, leptin was associated with total fat mass in all patient subgroups as well as in controls. Leptin is secreted from subcutaneous fat cells, proportional with adipocyte growth,38 and has previously been shown to correlate with total fat mass, in both healthy adults10 and in patients with RA.18 39

Moving to cardiac function, both lower adiponectin and apelin-12 were associated with impaired systolic function (measured by LAS%) in all patient subgroups (but not controls and not significant for apelin-12 in JDM-inactive). Besides being upregulated in obesity and mostly associated with metabolic abnormalities in the general population,29 there is evidence that apelin-12 is beneficial to cardiac function29 similar to higher adiponectin levels.40 41 Improved systolic function correlated with higher levels of apelin-12 in patients with mild to severe heart failure42 and with higher levels of adiponectin in patients with inflammatory cardiomyopathy.40 Also, in experimental studies, cardiomyocyte contraction improved after exposure to both adipokines.43 44 Apelin-12 might provide increased angiogenesis as discussed earlier, also in the heart.

We found better diastolic function with higher levels of adiponectin in JDM-total and JDM-active. Similar to our findings, adiponectin was found independently and positively associated with diastolic function in the general population.41 Also, lower adiponectin has been found in both humans and rodents with myocarditis and impaired diastolic function.40 45 These data support a possible beneficial role for both adiponectin and apelin-12 upon cardiac function, which may also apply for patients with JDM.

In JDM-active, higher resistin and lipocalin-2 were associated with impaired systolic function. Interestingly, both resistin and lipocalin-2 seem to be involved with the severity of heart failure,46 47 indicating an involvement for both adipokines in cardiomyocytes. Experimental studies have shown that resistin impaired cardiomyocyte contraction in rats48 and that both resistin and lipocalin-2 induced cardiomyocyte hypertrophy,12 48 which is frequently found in systolic dysfunction. Importantly, there was no difference in systolic function between patients with active versus those with inactive disease in our cohort.22 The role for resistin and lipocalin-2 upon systolic function in active disease should be studied further.

Our results might suggest that in JDM assessed after long-term disease, adipokines are released from dysfunctional adipose tissue as well as from other tissues affected by the disease. The increase in leptin might induce or maintain JDM due to an immune-modulatory mechanism besides the traditional leptin/obesity-related inflammatory effect. The novel finding of increased apelin-12 and visfatin in JDM-active might be involved in angiogenesis, which may have an impact on cardiac function. As such, it is important to keep in mind that in patients with unfavourable adipose tissue distribution, the adipokine profile is presumably altered. The adipokine profile found in long-term JDM might be an added burden on the inflammatory state of the disease.

Strengths and limitations

The present study represents a longstanding JDM outcome study in which 95% of all identified and tracked patients with JDM participated. Thus, our results are less biased toward serious cases compared with other outcome studies.49 Our age-matched and sex-matched controls represent the general population as they were randomly drawn from the Norwegian population registry. There are some limitations to our study besides those earlier mentioned. Unfortunately, lipodystrophy was not further categorised into generalised, partial or focal based on the patient’s fat loss. Leptin is highly responsive to food intake as it is a satiety hormone providing cross-talk between adipocytes and the satiety centre in the hypothalamus. For accurate measurements, blood sampling should have been done after an overnight fast. However, in our study, blood was drawn at approximately the same time point during the late morning. Also, being a cross-sectional study, no conclusions can be drawn as to the causal effect of the adipokines on cardiac function. Our study was performed some years ago and included patients who might not have been subjected to optimal treatment.


Our study demonstrates an imbalance in proinflammatory and anti-inflammatory adipokines, which suggests adipocyte dysfunction in JDM after long-term disease. Further, the unfavourable adipose tissue distribution and subclinical cardiac dysfunction are associated with these imbalanced adipokine levels, especially in patients with active disease. The finding of increased apelin-12 and its association with systolic function in JDM-active patients is novel. The underlying mechanisms need to be further explored.

Data availability statement

Data are available upon reasonable request. The data that underlie this article cannot be shared publicly due to the privacy of individuals who participated in the study. The data will be shared if a reasonable request is made to the corresponding author.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by the Regional Ethics Committee for Medical Research in South-East Norway (S-05144). Participants gave informed consent to participate in the study before taking part.


The authors would like to thank Anita Tollisen for help with patient inclusion, Ellen Nordal and Marite Rygg for patient recruitment.



  • Contributors Each author has made an important contribution to the study, is thoroughly familiar with the data and has approved this submission. HM stands as guarantor for the work and conduct of the study. Patients with JDM in Norway and their matched controls have assisted as research volunteers.

  • Funding The study was supported by the Anders Jahres Fund for the promotion of science and the Olav Raagholt and Gerd Meidel Raagholt Research Foundation.

  • Competing interests None declared.

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