Archives ofDisease in Childhood 1991; 66: 1057-1060

Panniculitis: a report of four cases and literaturereview

S M Randle, M B Richter, R G Palmer, A Price, B M Ansell

AbstractPanniculitis is a disease with many causes andassociations. The classification of clinicalsubtypes is unsatisfactory and hampered bythe use of eponyms.Four children with recurring panniculitis

are described and their histology presented.Three had subcutaneous fat atrophy withlobular panniculitis on biopsy; all respondedwell to corticosteroids. The fourth child had aseptal panniculitis with no atrophy of sub-cutaneous tissues and only a partial responseto treatment with corticosteroids.A widely 'accepted precise histological

classification of panniculitis is needed toenable accurate predictions of the outcome ofthis serious disorder.

Department ofRheumatology,Northwick Park Hospital,HarrowS M RandleM B RichterR G PalmerA PriceB M AnsellCorrespondence to:Dr S M Randle,Mercy Consulting Suites,Suite 27, 141 Grey Street,East Melbourne 3002,Australia.

Accepted 2 April 1991

Panniculitis is an uncommon but fascinatingcondition in childhood. Inflammation of the fatmay arise spontaneously or occur as a feature ofdistinct illnesses such as pancreatitis,' bacterialsepticaemia,2 systemic lupus erythematosis,3polyarteritis,l and malignancy.4 The literatureabounds with isolated case reports usingeponyms such as 'Weber-Christian syndrome'5and 'Rothman-Makai panniculitis'.6 Althoughattempts have been made to classify histologicalsubtypes of panniculitis,' 7 no universallyrecognised classification is in use to aid retro-spective studies. The various forms of lipo-dystrophy are often included in the classificationof panniculitis,8 which then becomes furtherconfused.

Figure I Case 1: histological appearance oflobular panniculitis (arrows). (Haematoxylinand eosin stain x 10.)

We report four cases of spontaneously occur-ring panniculitis, one in an infant and three inolder children. Follow up of our four childrenover several years suggests that lobular panni-culitis is associated with significant fat atrophyat the site of previously active lesions and a goodresponse to corticosteroids. Septal panniculitis,however, has no associated atrophy of fat andonly a partial response to corticosteroids.

Case reportsCASE 1A 6 year old girl first presented in 1975 with anupper respiratory infection and painful lumpson both thighs. Both problems resolved spon-taneously. The painful lumps reappeared threeyears later and biopsy specimen of a noduledemonstrated lobular panniculitis (fig 1).

Other investigations included a haemoglobinconcentration of 94 g/l, total white cell count of11 9x 109/l, erythrocyte sedimentation rate(ESR) of 60 mm/hour, a significant rise inantistreptolysin 0 titre (ASOT) over twomonths, a positive rheumatoid factor, andraised immune complexes. Serum C3 concen-tration was 1-65 g/l (normal range 0-81-70 g/l)and C4 0-23 g/l (normal range 0-15-0 45 g/l).Serum IgG was 28 g/l (normal range 6-5-16-5g/l). Direct Coombs test, the Venereal DiseaseResearch Laboratory test, Treponema pallidumhaemagglutination assay, and tests for anti-nuclear and anti-DNA antibodies gave negativeresults. No cryoglobulins or cold agglutininswere detected, and serum IgA, IgM, euglobulinlysis and fibrin plate lysis were normal.

Prednisolone was given in a dose of 30mg/day orally from January 1979 tapering to 30mg and 5 mg on alternate days by February1979. The prednisolone dose was steadilyreduced over the next five months, ceasing inJuly 1979 at which time there was resolution ofthe lesions with appreciable loss of subcutaneousfat at the sites of the previous nodules. Thenodules reappeared three months after cessationof treatment. Treatment with prednisolone wasagain successful in controlling the symptomswhen given in a dose of 30 mg/day orally,initially, and decreased steadily as before overthe next nine months until ceasing in July 1982.Currently the child remains well and free of newlesions.

CASE 2This girl was completely normal until the age of5 months when she developed red hot lumps onboth legs. By 16 months of age the lesions had

1057

on March 2, 2021 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.66.9.1057 on 1 S

eptember 1991. D

ownloaded from

Randle, Richter, Palmer, Price, Ansell

spread to involve her face, arms, and trunk.There were no systemic symptoms. A skinbiopsy specimen showed a septal panniculitis(fig 2) that responded temporarily to pred-nisolone in small doses.At 3-5 years of age she developed fever, joint

pains, and stiffness with increasing numbers ofnodules appearing after bouts of tonsillitis. Atrial of prophylactic penicillin decreased thenumber of infections but did not appear tomodify the disease process.

Subsequent treatment included chloroquine,indomethacin, naproxen, and sulphasalazine;none of these drugs resulted in any lastingbenefit. Prednisolone was used in small doses(for example, 5 mg) daily until the age of 5 5years when alternate day treatment was begunup to a dose of 20 mg at 11 years old. Increases

Figure 2 In contrast tofig I this biopsy shows relatively normalfat lobules withinflammation ofthe surrounding septa (arrow). (Haematoxylin and eosin stain x 10.)

Figure 3 Case 2: calfnodules ofseptal panniculitis inrelapse.

in prednisolone dose led to partial improvementsbut relapses still occurred (fig 3). Temporaryimprovement was seen after three pulses ofintravenous methylprednisolone were given inresponse to an acute exacerbation of the disease.Dapsone was started at 13 years of age withsome improvement of fevers and nodules.During the course of this girl's illness her

haemoglobin has remained at 100-111 g/l andher ESR has ranged from 13 to 30 mm/hour.

Tests for antinuclear antibody, immunecomplexes, autoantibodies to gastric parietalcells, smooth muscle, mitochondria, reticulinand thyroglobulin gave negative results. ASOT,throat swab, and urine and blood cultureswere also negative. Plasma urea, electrolytes, Creactive protein, triglycerides, cholesterol, redcell folate, serum complement, immuno-globulins, and B12 concentrations and bonemarrow examination were all normal.

CASE 3In December 1985 an 11 year old girl developedpainful erythematous indurated lesions over herfeet and legs (fig 4) after recovering from a sorethroat. Shewas not febrile butappeared lethargic.At the same time the metatarsophalangeal jointsof both feet were tender.There was no improvement with penicillin,

aspirin, other non-steroidal anti-inflammatorydrugs, and bed rest. Her haemoglobin concen-tration was 113 gIl, white cell count 6 8 x 109/l,and ESR 60 mm/hour. Rheumatoid factor waspositive (RAHA 1:1280) initially, becomingnegative by November 1987. Plasma C reactiveprotein was 27 000 ,tg/l (normal <8000). The

Figure 4 Case 3: swelling and erythema ofthe legs in theacute phase oflobular panniculitis.

1058

on March 2, 2021 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.66.9.1057 on 1 S

eptember 1991. D

ownloaded from

Panniculitis: a report offour cases and literature review1

plasma urea and electrolytes, liver functiontests, amylase activity, ASOT, serum immuno-globulins and complement, urine analysis, noseand throat swabs, and chest radiography allgave normal results. Autoantibodies to parietalcells, mitochondria, smooth muscle, reticulin,thyroid microsomes, and thyroglobulin were notdetected. Antinuclear and anti-DNA antibodieswere negative and antibodies to extractablenuclear antigen were not detected.One month after onset of the illness nodules

began to appear on the child's arms. A skinbiopsy specimen demonstrated lobular panni-culitis and prednisolone was started at a dose of40 mg alternating with 10 mg daily. She becamemore energetic and the skin lesions graduallyresolved, leaving large areas of subcutaneous fatloss (fig 5). Prednisolone intake was cautiouslyreduced as three small lesions persisted andoccasional lesions appeared at sites of minorskin trauma. By August 1988 the prednisolonedose was 25 mg on alternate days, the child feltwell, and the ESR had fallen to 6 mm/hour.

CASE 4A girl weighing 4500 g was born by electivecaesarean section after a normal pregnancy inAugust 1985. She had signs of ventricular andatrial septal defects leading to congestive cardiacfailure by 5 weeks of age. Corrective cardiacsurgery was performed three weeks later aftermedical treatment had failed to control hercondition.

Figure 5 Case 3: appreciable subcutaneous fat loss seeWafter resolution ofthe acute inflammatory stage oflobularpanniculitis.

Postoperatively the baby required intensiverespiratory and cardiovascular support but wasable to go home after two weeks in hospital.Shortly afterwards she developed red tenderlumps on her trunk, thighs, arms, and scalp.These lumps resolved spontaneously but re-appeared five weeks later on the baby's arms,shoulder, and face where they were sufficientlytender to interfere with feeding.A biopsy specimen of an active lesion

showed lobular panniculitis that responded wellto prednisolone in a dose of 15 mg/kg/day. Thedose of steroid was reduced over the next fivemonths and there has been no recurrence ofdisease off treatment.

Investigations at the time of biopsy showed ahaemoglobin concentration of 95 g/l, white cellcount of 12 2x 109/1, and an ESR of 13 mm/hour.ESR at presentation was 40 mm/hour. Serumamylase and creatine phosphokinase activitiesand ASOT were normal. No immune complexeswere detected.

DiscussionOur four cases demonstrate the two histologicaltypes of panniculitis previously described,lobular and septal.' Although these types havebeen recognised in the past no attempt has beenmade to correlate histological type with responseto treatment. Our review of the literaturesupports this paper's finding that lobularpanniculitis responds better to treatment withprednisolone than the septal type.Two of our children were infants at the onset

of the illness. Most of the infants describedpreviously have had significant systemicsymptoms and in some cases a fatal outcome.6Our infant with lobular panniculitis has donewell on steroids alone. She now has no fatatrophy and is offtreatment with no recurrences.Taylor9 and more recently Conway et allo havealso described cases of panniculitis with onset ininfancy where the clinical course has beenrelatively mild and there has been a goodresponse to treatment. Our child with septalpanniculitis whose disease began at 5 months ofage, although still having problems, has notsuccumbed and has not developed symptoms ofperivisceral disease.The nature of the inflammatory infiltrate

found on histology varies according to thedisease stage at which the biopsy is performed. "It is important to obtain adequate biopsyspecimens early in the disease as once the endstage is reached, subcutaneous fibrosis of thepanniculis may be all that is seen and thelobular architecture of the fat is distorted.Punch biopsies are not recommended becausethey may not obtain adequate samples of fattytissue because of inflammation of the sub-cutaneous tissue, which leads to decreasedadherence of the dermis to the fat in the earlystages.'

Unless an acute lesion is biopsied histologycan be meaningless so that children presentingwith a loss of subcutaneous fat are particularlydifficult to diagnose. In the four patientsdescribed by Peters and Winkelmann noduleswere not present at the beginning of the disease

1059

on March 2, 2021 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.66.9.1057 on 1 S

eptember 1991. D

ownloaded from

1060 Randle, Richter, Palmer, Price, Ansell

or if present were obvious only briefly. 2 Threeof their patients were thought to have lipo-dystrophy until biopsies showed panniculitis.Upper respiratory tract infections preceded

the panniculitis in two of our cases with lobularpanniculitis. The child with septal panniculitissuffered frequent exacerbations of her diseaseafter bouts of tonsillitis as did a child withlobular panniculitis described by Gupta andRasmussen. 3 Septal panniculitis has beenrecently associated with Lyme disease. 4 Thereis a similarity between this form of panniculitisand erythema nodosum. Erythema nodosum isconsidered to be a hypersensitivity reaction to avariety of antigenic stimuli and is seen duringseveral infectious diseases including thosecaused by , haemolytic streptococci. It is clearfrom the literature that the presence of infectionis not predictive of the histological type ofpanniculitis nor the clinical outcome.The infant who underwent corrective cardiac

surgery was initially thought to have 'coldpanniculitis' resulting from the hypothermiaduring surgery. A recent case report describesan infant with similar problems after cardiacsurgery and the diagnosis of subcutaneous fatnecrosis of the newborn was made on thehistological findings.'5 Cold panniculitis haspreviously been described but is localised to thearea affected by the hypothermia, begins hoursto at most a few days after the insult, and doesnot spontaneously recur without further ex-posure to cold.'6 Our child's lesions recurredtoo long after cardiac surgery to be consistentwith cold panniculitis. Nor is this child'scondition compatible clinically or histologicallywith sclerema neonatorum or subcutaneous fatnecrosis of the newborn.Immunological abnormalities have previously

been reported in association with panniculitis.Allen-Mersh described a 7 year old girl whofirst presented with a panniculitis of the footand who then went on to develop diabetesmellitus, hepatic cirrhosis, red cell surfaceantibodies, and antibodies to mitochondria andgastric parietal cells.'7 Recently Billings et alreported three cases of lobular panniculitis."8One of the patients had insulin dependentdiabetes and Hashimoto's thyroiditis, thesecond later developed juvenile chronic arthritis,and the third insulin dependent diabetes.

Hendricks et al describe a child whose diseasebegan after immunisation and another who didnot develop new lesions while he was lymphocytedepleted.6Our two older children with lobular panni-

culitis had positive rheumatoid factors. One hada raised ASOT and immune complexes. Allchildren had negative autoantibody screens.The significance of these isolated findings is notclear and a review of the literature does notsuggest any correlation between the nature ofthe immunological disturbance and the histo-logical type of panniculitis.

It is clear that many factors are involved inthe initiation of panniculitis. It is possible thatvarious physical and chemical agents and

infections expose parts of the lipocyte increasingits vulnerability to attack by autoimmunemechanisms. This is supported by the increasedincidence of panniculitis among children withaI,-antitrypsin deficiency.'9

Steroids have been the mainstay of treatmentof panniculitis but some children do notrespond.20 Other agents tried with success havebeen chloroquine,211 azathioprine,'0 cyclo-phosphamide,22 and more recently dapsone.23 24

Panniculitis is a disease with at least twodistinct histological subtypes. Current treat-ments are often based on the results presentedin numerous case reports rather than theoutcome of prospective trials of treatment. Theincidence of panniculitis is low but for aparticular child this illness may cause seriousdisability or even death. It is for these reasons itis suggested that all children who present withrecurrent painful nodules of panniculitis havean early and adequate biopsy to ascertain thehistology. Prompt treatment may preventfurther severe illness as well as disfiguringsubcutaneous fat loss.

I Ackerman B. Panniculitis. Histologic diagnosis of inflammatoryskin diseases. Philadelphia: Lea and Febiger, 1978:779-825.

2 Bagel J, Grossman ME. Subcutaneous nodules in pseudo-monas sepsis. Am J Med 1986;80:528-9.

3 Winkelmann RK. Panniculitis in connective tissue disease.Arch Dermatol 1983;119:336-44.

4 Aronson IK, West DP, Vanakojis D, Ronan SG, lossifides I,Zeitz HJ. Panniculitis associated with cutaneous T-celllymphoma and cytophagocytic histiocytosis. BrJ' Dermatol1985;112:87-96.

5 Edge J, Dunger DB, Dillon MJ. Weber-Christian panniculitisand chronic active hepatitis. EurJf Pediatr 1986;145:227-9.

6 Hendricks WM, Ahmad M, Gratz E. Weber-Christiansyndrome in infancy. BrJ Dermatol 1978;98:175-86.

7 Niemi KM, Forstrom L, Hannuksela M, Mustakollio KK,Salo OP. Nodules on the legs. A clinical, histological andimmunohistological study of 82 patients representingdifferent types of nodular panniculitis. Acta Derm Venereol(Stockh) 1977;57:145-54.

8 Patterson JW. Panniculitis. Acta Dermatol 1987;123:1615-8.9 Taylor GA. Prolonged remission of Weber-Christian

syndrome in an infant. Clin Pediatr (Phila) 1981;20:521-3.10 Conway SP, Smithells RW, Peters WM. Weber-Christian

panniculitis. Ann Rheum Dis 1987;46:339-41.11 AronsonIK,ZeitzHJ,VariakojisD. Panniculitisin childhood.

Pediatr Dermatol 1988;5:216-30.12 Peters MS, Winkelmann RK. Localised lipoatrophy (atrophic

connective tissue disease panniculitis). Arch Dermatol1980;116: 1363-8.

13 Gupta AK, Rasmussen JE. Multiple areas of localized tissueloss in a child. Arch Dermatol 1986;122:1199-2000.

14 Kramer N, Rickert RR, Brodkin RH, Rosenstein EO. Septalpanniculitis as a manifestation of Lyme disease. Am J Med1986;81:149-52.

15 Silverman AK, Mickels EH, Rasmussen JE. Subcutaneousfat necrosis in an infant occurring after hypothermic cardiacsurgery. JAm Acad Dermatol 1986;15:331-6.

16 Hultcrantz E. Haxthausen's disease. Cold panniculitis inchildren. J Laryngol Otol 1986;100:1329-32.

17 Allen-Mersh TG. Weber-Christian panniculitis and auto-immune disease: a case report. J Clin Pathol 1976;29:144-9.

18 Billings JK, Milgraum SS, Gupta AK, Headington JT,Rasmussen JE. Lipoatrophic panniculitis: a possibleautoimmune inflammatory disease of fat. Arch Dermatol1987;123:1662-6.

19 Rubenstein HM, Jaffer AM, Kuduna JC, Lortratanakul Y.Alpha I-antitryspin deficiency with severe panniculitis.Report of two cases. Ann Intern Med 1977;86:742-4.

20 Winkelmann RK, McEvoy MT, Peters MS. Lipophagicpanniculitis of childhood. J Am Acad Dermatol 1989;21:971-8.

21 Shelley WB. Chloroquine-induced remission of nodularpanniculitis present for 15 years. J Am Acad Dermatol1981;5:168-70.

22 Kirch W, Duhrsen V, Hoensel H, Ohnhaus E. Cyclo-phosphamide-induced remission in Weber-Christianpanniculitis. Rheumatol Int 1985;5:239-40.

23 Uplekar MW, Antia NH. Dapsone dependent nodularpanniculitis. Indian J Lepr 1986;58:286-90.

24 Roth DE, Schikler KS, Callen JP. Annular atrophic connectivetissue panniculitis of the ankles. Y Am Acad Dermatol1989;21:1 152-6.

on March 2, 2021 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.66.9.1057 on 1 S

eptember 1991. D

ownloaded from