Study of Subcutaneous Fat in Children WithJuvenile DermatomyositisSUMIT VERMA, SURJIT SINGH, ANIL KUMAR BHALLA, AND MADHU KHULLAR

Objective. Lipodystrophy is a recently recognized complication of juvenile dermatomyositis (juvenile DM). Until now,the diagnosis has been based only on the physical appearance of the patient. We quantified the patterns of fat distributionin a cohort of patients with juvenile DM.Methods. Twenty patients with juvenile DM were enrolled along with a matching number of controls. Both groupsunderwent standard anthropometric measurements including assessment of skinfold thickness using Harpenden Skin-fold Caliper (Holtain, Dyfed, UK). Glucose tolerance test and serum lipid estimates were performed in the study group butnot in controls.Results. Patients with juvenile DM had lower mean weight, height, and mid–upper-arm circumference as compared withcontrols; these differences were statistically significant (P < 0.05). Eight (40%) of 20 patients had lipodystrophy onphysical appearance. When assessed by skinfold caliper, there was loss of subcutaneous fat at the mid-axillary site in 65%of patients, at the subscapular in 60%, and at the suprailiac site in 55%. Serum triglyceride levels were increased in 12of the 18 patients who underwent this test. Oral glucose tolerance test results were normal in all 20 patients.Conclusion. Sixty-five percent of our patients with juvenile DM were found to have loss of subcutaneous fat onquantification compared with 40% on physical appearance alone. Maximum fat loss occurred at the mid-axillary skinfoldsite. A significant number of patients with juvenile DM (66%) had hypertriglyceridemia. We hypothesize that lipodys-trophy and hypertriglyceridemia could well be integral components of what may be an expanded juvenile DM syndrome.

KEY WORDS. Juvenile dermatomyositis; Lipodystrophy; Skinfold thickness; Hypertriglyceridemia; Hirsutism.

INTRODUCTION

Juvenile dermatomyositis (juvenile DM) is the most com-mon inflammatory myopathy in children (1). Onset ofdisease usually occurs between ages 4 and 10 years, andthe female:male ratio is 2.7:1.0 among North Americanwhites (2). The cardinal features of this disease are proxi-mal muscle weakness and characteristic skin changes suchas heliotrope rash of the upper eyelids, Gottron’s papules,and periungual erythema with capillary loop abnormali-ties.

Recently, 2 new clinical findings have been reported inthe literature: lipodystrophy and insulin-resistant diabetes(3–10). In 1990, Tucker et al (3) reported 3 cases of lipo-

dystrophy among a cohort of 30 patients with juvenile DM.This report was followed by isolated case reports of pa-tients with juvenile DM with lipodystrophy and associatedfeatures such as hepatomegaly, acanthosis nigricans, hy-pertrichosis, and insulin-resistant diabetes (4–10). Hue-mer et al (10) studied 20 patients with juvenile DM andfound that lipodystrophy was present in 25%, whereaselevated triglyceride levels and insulin resistance werepresent in 50% of their cohort.

The diagnosis of lipodystrophy has previously beenmade only on the basis of clinical appearance of the pa-tient, and no objective method has been used to quantifyfat in these patients. In this study, we quantified sub-cutaneous body fat and evaluated oral glucose tolerancetests and lipid profiles in our cohort of children withjuvenile DM.

PATIENTS AND METHODS

Patients. The study sample comprised 20 patients (13boys, 7 girls) ages 3.5–17 years who were diagnosed withjuvenile DM according to the criteria of Bohan and Peter(11,12) (Table 1). These patients were enrolled consecu-

Sumit Verma, MD, Surjit Singh, MD, Anil Kumar Bhalla,PhD, Madhu Khullar, PhD: Post Graduate Institute of Med-ical Education and Research, Chandigarh, India.

Address correspondence to Surjit Singh, MD, AdditionalProfessor of Pediatric Allergy & Immunology, Departmentof Pediatrics, Advanced Pediatric Centre, Post GraduateInstitute of Medical Education and Research, Chandigarh-160012, India. E-mail: surjitsinghpgi@rediffmail.com.

Submitted for publication July 3, 2005; accepted in re-vised form December 1, 2005.

Arthritis & Rheumatism (Arthritis Care & Research)Vol. 55, No. 4, August 15, 2006, pp 564–568DOI 10.1002/art.22108© 2006, American College of Rheumatology

ORIGINAL ARTICLE

564

tively from the Pediatric Rheumatology and ImmunologyClinic of the Advanced Pediatric Centre, Postgraduate In-stitute of Medical Education and Research (PGIMER),Chandigarh, India. An equal number of healthy childrenmatched on age, sex, and socioeconomic status (13) wereenrolled from the Growth Clinic of Advanced PediatricCentre, PGIMER as controls. The study was conductedover a period of 14 months from July 2003 to August 2004.The Institute Ethics Committee approved the studyprotocol.

Oral prednisolone was used in all patients except 1(patient 4), with the mean duration of treatment being 1.77years. Patient 4 had presented very late and had severewasting. Corticosteroids were not used in this patient. Shealso had severe osteoporosis and calcinosis, requiringalendronate therapy (14). Two patients with severe diseaseand dysphagia at onset (patients 8 and 20) were givenintravenous methylprednisolone (30 mg/kg/day) for 5 daysfollowed by oral prednisolone. Prednisolone was admin-istered in a dosage of 2 mg/kg/day (maximum 60 mg) for 1month, and was then tapered over a period of �2 years.Methotrexate, azathioprine, and hydroxychloroquine wereused in selected patients depending on the clinical re-quirements.

The mean � SD followup period for patients was 2.2 �1.7 years (range 0.5–7 years). During the study period, 2patients (patients 9 and 12) finished treatment successfullyand were taken off drugs. One of our study patients (pa-tient 7) died at home during followup, possibly due toaspiration pneumonia secondary to pharyngeal muscleweakness. Two of our patients (patients 2 and 6) hadrelapses during followup and had to be restarted on ste-roids along with methotrexate.

Study methods. Patients and controls underwent thefollowing anthropometric measurements using standard-ized techniques and instruments in the growth laboratoryon a preappointed date and time: body weight (in kg),crown-heel length/height (in cm), head circumference (incm), chest circumference (in cm), mid–upper-arm circum-ference (in cm), biceps skinfold thickness (in mm), tricepsskinfold thickness (in mm), subscapular skinfold thick-ness (in mm), mid-axillary skinfold thickness (in mm),medial calf skinfold thickness (in mm), medial thigh skin-fold thickness (in mm), suprailiac skinfold thickness (inmm), paraumbilical skinfold thickness (in mm), and bodymass index (BMI; in kg/m2). Body weight was measured onan electronic weighing scale (Avery Kolkata, West Bengal,India; least count 50 gm). Height was measured with astadiometer (Holtain, Dyfed, UK; least count 1 mm). Cir-cumference was measured with a fiberglass tape measure(least count 1 mm). Harpenden Skinfold Caliper (Holtain;least count 0.2 mm) was used to measure skinfold thick-ness. All skinfold measurements were taken on the leftside of the body. The magnitude of interobserver errorwith regard to skinfold thickness varied between 0 mmand 0.4 mm, whereas it varied between 0 mm and 2 mm forheight, 0 gm and 50 gm for body weight, and 0 mm and 2mm for circumferential measurements.

Glucose tolerance tests (15) and serum lipid estimations(16) were carried out only in patients and not in controls.All investigations were conducted in the HypertensionResearch Laboratory of the Department of ExperimentalMedicine and Biotechnology, PGIMER.

The mean � SD and coefficient of variation were calcu-lated for each body parameter. In view of non-Gaussiandistribution of skinfold thickness, all absolute skinfold

Table 1. Clinical and biochemical profile of patients with juvenile dermatomyositis*

PatientsAge,

years/sexMuscle

weakness Lipodystrophy HirsutismHeliotrope

rash CalcinosisGottron’spapules

Triglyceride level(mg, %)

1 7.5/M � � � – – � 77.52 12/M �† � � – – – 154.33 14/M � � � � � � 142.24 9/F � � � � � � 325.85 6.5/F � � – � – � 193.16 13/M � � – – – – 82.87 13/M �‡ � � � – � 219.78 4.5/F � � � � – � 155.89 17/M � – – � – – 95.8

10 5/M � – – � – – 85.911 7/F � – � � – – 80.612 11/F �‡ – – � – � 120.913 5.5/F � – – � – � NA14 3.5/M � – – � � � 253.215 9/M �‡ – – � – � 196.416 14/M � – – � – � 128.517 14/M �‡ – – � – � 329.218 15/M � – – � – – 349.819 13/F �‡ – – – – � NA20 12.5/M � – – � – � 82.1

* � � present; — � absent; NA � not applicable.† Cardiac involvement (ventricular ectopics).‡ Weak gag reflex.

Subcutaneous Fat in Juvenile DM 565

thickness measurements were transformed into their log-arithms according to the formula given by Edwards et al(17) to contain the higher magnitude of variability. Themagnitude of intergroup difference was quantified for eachabsolute as well as log-transformed measurement usingWilcoxon’s signed rank test.

RESULTS

The clinical and biochemical profiles of all 20 patientscomprising the study group are detailed in Table 1. Boys(n � 13) comprised 65% of the sample and had a mean �SD age of 11.5 � 3.7 years, whereas girls (n � 7) had amean � SD age of 8.0 � 3.6 years (18).

Five patients (20%) had severe muscle weakness in theform of weak gag reflex and dysphagia and required pro-longed tube feeding. One patient had ventricular ectopicsand arrhythmia. Heliotrope rash was observed in 16 pa-tients (80%), Gottron’s papules in 14 (70%), hirsutism in 7(35%), and lipodystrophy in 8 (40%). Three children hadcalcinosis (15%), 1 patient had vasculitic ulcers, and 1patient had acanthosis nigricans (5%).

Patients with juvenile DM had lower weight, shorterheight, and lower mid–upper-arm circumference whencompared with healthy controls (Table 2); these differ-ences were statistically significant (P � 0.05). The mean �SD difference in weight between controls and patients was7.6 � 4.4 kg, and 75% of patients with juvenile DMweighed less than controls. The mean � SD difference inheight between controls and patients was 13.2 � 6.3 cm,with 90% of patients with juvenile DM being shorter thancontrols. Similar trends with respect to mid–upper-armcircumference were found, with 65% of patients havinglower measured values (mean � SD 3.7 � 2.2 cm) whencompared with controls. The mean � SD BMI in patientswith juvenile DM was less compared with controls (13.7 �1.4 kg/m2 versus 15.5 � 1.4 kg/m2). This difference, how-ever, was not statistically significant.

Of the 8 skinfold sites considered (Table 3), the mid-axillary was found to be the most affected (65%), followedby the suprailiac (60%), subscapular (55%), medial calf(50%), paraumbilical (50%), triceps (45%), medial thigh(35%), and biceps (25%). The mean � SD net subcutane-ous fat loss was 1.5 � 0.7 mm in the biceps, 2.3 � 1.4 mmin the triceps, 1.7 � 0.8 mm in the subscapular, 1.8 � 1.2mm in the mid-axillary, 2.4 � 2.3 mm in the medial calf,6.2 � 3.8 mm in the medial thigh, 2.5 � 1.9 mm in thesuprailiac, and 1.8 � 1.7 mm in the paraumbilical. Ontransforming absolute skinfold thickness into logarithms,the magnitude of relative variability or coefficient of vari-ability was found to be reduced. The difference betweenabsolute and logarithm transformed values ranged from59% to 72%.

None of the patients had an impaired or abnormal glu-cose tolerance test result or clinical features suggestive ofdiabetes mellitus. Lipid profile testing was performed in18 patients. Triglyceride levels were found to be high in 12(66%) patients (Table 1). Five patients had high serumcholesterol levels and 3 had low high-density lipoprotein(HDL).

Based on physical appearance and results of tests ofserum triglyceride levels, the study sample was catego-rized into 3 subgroups arbitrarily: group A consisted ofpatients with juvenile DM with severe loss of subcutane-ous fat on physical appearance and increased serum tri-glyceride levels, group B consisted of patients with normalphysical appearance and increased serum triglyceride lev-els, and group C consisted of patients with DM depictingvariable fat deposition on physical appearance and normalserum triglyceride levels. Patients in group A had lowermid-axillary, subscapular, and medial calf skinfold thick-nesses as compared with controls; these differences werestatistically significant (P � 0.05). Patients in group B hadlower fat at the mid-axillary skinfold site. However, thedifference was not statistically significant.

Table 2. Mean � SD different parameters of patients with juvenile dermatomyositis and controls

Body parametersGroup A(n � 6)

Group B(n � 6)

Group C(n � 6)

Total(n � 20)

Weight, kgPatients 21.0 � 9.2 24.7 � 14.4 27.8 � 10.7 24.5 � 10.8Controls 28.2 � 10.6 29.2 � 12.4 31.5 � 11.1 29.2 � 11.0

Height, cmPatients 120.4 � 19.5 115.4 � 23.1 130.4 � 18.8 122 � 19.4Controls 133.7 � 21.2 127.6 � 26.7 140.2 � 18.9 132.9 � 22.9

Head circumference, cmPatients 50.6 � 1.9 51.2 � 2.3 52.3 � 2.3 51.4 � 2.0Controls 50.0 � 2.1 50.3 � 2.7 53.7 � 3.0 51.4 � 2.9

Chest circumference, cmPatients 57.7 � 7.1 61.1 � 13.6 64.1 � 8.6 60.7 � 9.6Controls 58.9 � 7.0 66.6 � 11.5 62.7 � 8.4 60.5 � 8.8

Mid–upper-arm circumference, cmPatients 13.4 � 2.5 18.0 � 5.4 17.7 � 2.9 16.5 � 4.0Controls 17.4 � 2.3 19.5 � 3.5 18.2 � 2.3 18.2 � 2.7

Body mass index, kg/m2

Patients 13.8 � 1.6 17.2 � 4.5 15.7 � 2.5 13.7 � 1.4Controls 15.1 � 1.27 16.9 � 1.7 15.5 � 1.4 15.5 � 1.4

566 Verma et al

DISCUSSION

Juvenile DM has recently been associated with a numberof seemingly unrelated abnormalities such as lipodystro-phy, insulin-resistant diabetes, and hypertriglyceridemia(3–10). Our study suggests that lipodystrophy may be oneof the cardinal features of this condition. Lipodystrophyhas previously been diagnosed only on the basis of phys-ical appearance. This is the first study to quantify loss ofbody fat in patients with juvenile DM.

The work of Huemer et al (10) suggests that patientswith juvenile DM have compromised height and weightattainments. Our study reiterates this point. Skinfoldthicknesses were significantly lower in study patients ascompared with controls, but loss of body fat was not uni-form at all sites. Our results suggest that the mid-axillary,suprailiac, and subscapular skinfolds are the most af-

fected. This observation is difficult to explain and may berelated to the release of inflammatory mediators and cyto-kines in this disease.

Glucose tolerance test results were normal in all patientsin our study. This is in contrast to the results of Huemer etal (10) who demonstrated impaired and abnormal glucosetolerance test results in 2 patients, each with lipodystro-phy. The followup period in the latter study was longerthan ours, and it is possible that some children in ourstudy may go on to develop impaired glucose tolerancelater in life. Other possible explanations for these differ-ences could be related to genetic, dietary, and lifestyledifferences. Measurement of insulin levels may have pro-vided additional information, but investigation of insulinlevels could not be carried out due to technical reasons.

Serum lipid estimations showed that 12 of 18 study

Table 3. Mean � SD skinfold thicknesses of patients with juvenile dermatomyositis and controls*

SkinfoldsGroup A(n � 6)

Group B(n � 6)

Group C(n � 6)

Total(n � 20)

Biceps, mmPatients 4.4 � 2.0

(125.8 � 46.1)8.4 � 7.2

(163.5 � 42.1)7.2 � 6.4

(149.1 � 59.7)6.9 � 5.4

(149.0 � 47.6)Controls 4.1 � 1.2

(128.4 � 35.8)6.6 � 3.0

(161.5 � 25.1)3.1 � 0.6

(113.4 � 28.4)4.5 � 2.2

(132.5 � 34.0)Triceps, mm

Patients 4.9 � 1.6(140.6 � 34.2)

10.8 � 7.7(186.2 � 28.5)

9.6 � 7.5(175.3 � 43.2)

8.7 � 6.2(169.4 � 37.7)

Controls 6.0 � 1.2(160.8 � 12.0)

9.7 � 3.5(187.0 � 17.3)

4.9 � 0.6(150.2 � 8.5)

6.8 � 2.8(165.1 � 19.3)

Subscapular, mmPatients 3.7 � 0.7

(126.8 � 16.5)9.8 � 7.6

(174.8 � 38.9)9.5 � 8.1

(172.1 � 46.2)7.9 � 6.5

(149.0 � 47.6)Controls 5.9 � 2.6

(153.9 � 29.0)8.0 � 3.2

(174.8 � 21.4)4.7 � 0.9

(146.9 � 14.3)6.1 � 2.6

(157.4 � 23.3)Mid-axillary, mm

Patients 2.8 � 0.4(99.7 � 17.2)

7.1 � 5.9(150.7 � 45.5)

6.6 � 4.7(156.2 � 52.9)

6.1 � 4.9(131.1 � 45.8)

Controls 4.5 � 0.9(140.8 � 15.9)

8.1 � 4.7(171.7 � 28.5)

3.7 � 0.7(127.7 � 22.6)

5.3 � 3.1(145.1 � 27.6)

Suprailiac, mmPatients 8.0 � 4.4

(166.5 � 40.4)11.3 � 5.7

(190.5 � 27.7)9.1 � 5.8

(172.9 � 33.3)9.2 � 5.0

(176.2 � 32.4)Controls 7.4 � 2.4

(176.6 � 19.8)9.5 � 2.0

(187.8 � 11.8)5.8 � 3.1

(156.9 � 22.6)7.5 � 2.8

(170.5 � 22.5)Paraumbilical, mm

Patients 10.5 � 5.1(188.1 � 43.4)

15.9 � 7.2(209.9 � 23.8)

15.9 � 9.3(211.0 � 24.8)

15.0 � 7.5(203.5 � 33.3)

Controls 31.8 � 6.4(200.9 � 29.9)

15.4 � 8.6(206.7 � 26.0)

7.4 � 3.6(178.7 � 33.8)

12.6 � 7.1(194.4 � 29.6)

Medial calf, mmPatients 4.6 � 1.1

(142.1 � 15.5)11.3 � 8.8

(180.0 � 42.2)12.2 � 11.6

(181.8 � 146.6)9.9 � 8.4

(172.2 � 39.8)Controls 7.1 � 2.1

(169.4 � 20.7)11.3 � 6.2

(19.1 � 25.7)3.7 � 1.8

(156.7 � 1.7)7.7 � 4.3

(170.1 � 24.6)Medial thigh, mm

Patients 7.1 � 3.2(166.2 � 24.9)

11.2 � 9.9(176.7 � 40.0)

11.5 � 11.6(175.5 � 49.7)

10.6 � 8.5(177.2 � 39.7)

Controls 6.0 � 1.7(159.2 � 19.1)

10.2 � 4.9(186.1 � 25.5)

5.9 � 2.8(153.8 � 27.9)

7.2 � 3.6(165.8 � 26.1)

* Values within parentheses represent log-transformed values.

Subcutaneous Fat in Juvenile DM 567

group patients had elevated triglyceride levels, 5 patientshad high serum cholesterol, and 3 patients had low HDL.Children with juvenile DM may need lipid-loweringagents as part of their initial treatment regimen as well asat followup, even when the muscle weakness has remitted.When subcategorizing the study patients based on physi-cal appearance and serum triglyceride levels, we foundthat patients in group A (n � 6) had lower mid-axillary,subscapular, and suprailiac skinfold thickness as com-pared with controls; these differences were statisticallysignificant (P � 0.05). Hirsutism was seen in 5 patients inthis group. Patients in group B (n � 6) had lower mid-axillary skinfold thickness, but this was not statisticallysignificant. Four patients in group B had hirsutism. Noneof the patients in group C (n � 6) had decreased skinfoldthickness or hirsutism. These findings suggest that hyper-triglyceridemia is the first metabolic abnormality to appearin children with juvenile DM, followed by the appearanceof hirsutism and lipodystrophy.

To conclude, juvenile DM appears to be a more gener-alized disorder rather than one that involves only themuscles and skin. Many children with juvenile DM may goon to develop hypertriglyceridemia, lipodystrophy, andlater, perhaps, insulin-resistant diabetes.

REFERENCES

1. Rider LG, Miller FW. Classification and treatment of the ju-venile idiopathic inflammatory myopathies. Rheum Dis ClinNorth Am 1997;23:619–55.

2. Cassidy JT, Petty RE. Juvenile dematomyositis. In: Cassidy JT,Petty RE, editors. Textbook of pediatric rheumatology. 4th ed.Philadelphia: W. B. Saunders Company; 2001. p. 465–94.

3. Tucker LB, Sadeghi-Nejad A, Schaller JG. The association ofacquired lipodystrophy with juvenile dermatomyositis [ab-stract]. Arthritis Rheum 1990;33 Suppl 9:S146.

4. Commens C, O’Neill P, Walker G. Dermatomyositis associated

with multifocal lipoatrophy. J Am Acad Dermatol 1990;22:966–9.

5. Torrelo A, Espana A, Boixeda P, Ledo A. Partial lipodystro-phy and dermatomyositis [letter]. Arch Dermatol 1991;127:1846–7.

6. Kavanagh GM, Colaco CB, Kennedy CT. Juvenile dermatomy-ositis associated with partial lipoatrophy. J Am Acad Derma-tol 1993;28:348–51.

7. Quecedo E, Febrer I, Serrano G, Martinez-Aparicio A, AliagaA. Partial lipodystrophy associated with juvenile dermato-myositis: report of two cases. Pediatr Dermatol 1996;13:477–82.

8. Adams BS, Cemeroglu AP, Haftel HM, Mitchell JM, HopwoodNJ. Prevalence of insulin resistance in juvenile dermatomyo-sitis [abstract]. Arthritis Rheum 1996;39 Suppl 9:S192.

9. Irene NT, Parveen B, Hema OH, Rajagopalan V. Juvenile der-matomyositis associated with lipodystrophy. Indian J Derma-tol Venereol Leprol 2003;69:343–4.

10. Huemer C, Kitson H, Malleson PN, Sanderson S, Huemer M,Cabral DA, et al. Lipodystrophy in patients with juveniledermatomyositis: evaluation of clinical and metabolic abnor-malities. J Rheumatol 2001;28:610–5.

11. Bohan A, Peter JB. Polymyositis and dermatomyositis (first oftwo parts). N Engl J Med 1975;292:344–7.

12. Bohan A, Peter JB. Polymyositis and dermatomyositis (secondof two parts). N Engl J Med 1975;292:403–7.

13. Mahajan BK, Gupta MC. Social environment. In: Mahajan BK,Gupta MC, editors. Textbook of preventive and social medi-cine. 2nd ed. New Delhi: Jaypee Publishers; 1995. p. 124–38.

14. Mukamel M, Horev G, Mimouni M. New insight into calcino-sis of juvenile dermatomyositis: a study of composition andtreatment. J Pediatr 2001;138:763–6.

15. The Expert Committee on the Diagnosis and Classification ofDiabetes Mellitus. Report of the Expert Committee on thediagnosis and classification of diabetes mellitus. DiabetesCare 2003;26 Suppl 1:S5–20.

16. Henry JB. Clinical diagnosis and management by laboratorymethods. 18th ed. Philadelphia: W. B. Saunders; 1991. p.204–11.

17. Edwards DA, Hammond WH, Healy MJ, Tanner JM, White-house RH. Design and accuracy of calipers for measuringsubcutaneous tissue thickness. Br J Nutr 1955;9:133–43.

18. Singh S, Kumar L, Shankar KR. Juvenile dermatomyositis innorth India. Indian Pediatr 1997;34:193–8.

568 Verma et al