DOI: 10.1111/j.1610-0387.2011.07831.x Original Article 1

© The Authors • Journal compilation © Blackwell Verlag GmbH, Berlin • JDDG • 1610-0379/2011 JDDG | 2011 (Band 9)

Keywords• panniculitis• biopsy• histopathology

SummaryBackground: Clinical and histologic diagnosis of panniculitis may be difficult.The patients usually present with erythematous subcutaneous nodules with orwithout additional symptoms. If a skin biopsy does not include enough subcu-taneous fat, histopathologic assessment is limited and the correct diagnosismay be delayed and require further sampling.Patients and Methods: To illustrate the difficulties in the diagnosis of panniculi-tis, we performed a retrospective examination of four patients with differentforms of panniculitis.Results: In two patients with subcutaneous panniculitis-like T cell lymphomaand lupus panniculitis, the correct diagnosis could only be ascertained after adelay of several months because repeated biopsies had to be obtainedthroughout the course of disease. In two further patients with cold panniculitisand pancreatic panniculitis, clinicians did not even suspect an inflammatoryprocess in the subcutaneous tissue. The correct diagnosis was made with adeep punch biopsy that included subcutaneous fat.Conclusions: On the one hand, these examples demonstrate the importance ofsampling subcutaneous tissue when obtaining routine punch biopsies. On theother hand, in cases where the diagnosis is uncertain, it is necessary to performlarge and deep incisional biopsies.

Histopathology of panniculitis – aspects of biopsytechniques and difficulties in diagnosisChristian Rose1,2, Martin Leverkus3, Mariella Fleischer2, Iakov Shimanovich1,2

(1) Dermatohistologic Laboratory of Lübeck, Germany(2) Department of Dermatology, Allergology, and Venereology, Schleswig-Holstein University Hospital, Lübeck,

Germany(3) Department of Dermatology, Allergology, and Venereology, Medical Faculty of Heidelberg University, Germany

JDDG; 2011 • 9 Submitted: 19.7.2011 | Accepted: 21.9.2011

IntroductionMost skin diseases involve the epidermisand dermis and can generally be diagnosedbased on the results of histological analysisand clinical examination. Performing asmall punch biopsy, which usually includesan adequate amount of skin tissue, is aquick and technically simple procedureperformed in everyday clinical practice [1].Yet when the adipose tissue is involved,clinical and histological diagnosis is moredifficult. Patients often have uncharacte-ristically indurated, poorly-borderederythematous lesions which may or maynot be related to generalized symptoms.Luis Requena called the diagnosis of pan-niculitis the “ugly duckling” of dermato-

pathology, capturing the particular diffi-culty of diagnosing the disorder [2].By far the most common form of panni-culitis is erythema nodosum which ismore common than all other forms ofpanniculitis taken together [3]. Charac-teristic features include red-brown nodu-les that are sensitive to pressure, usuallyappearing on the extensor surfaces of thelower legs, and often affecting youngerwomen. Erythema nodosum may be dueto various causes, often bacterial or viralinfections. Less common triggers includesarcoidosis, inflammatory bowel diseases(e.g., Crohn disease), and pregnancy. Inpatients with a typical clinical presenta-tion and a known cause, there is no need

for histopathological analysis and an expectant approach may be taken. Thehistology of erythema nodosum showsseptal mixed-cell panniculitis, but the results do not indicate disease etiology, ashas been underscored by two large caseseries [4, 5]. Other forms of panniculitismay have a less typical clinical pictureand such lesions must be biopsied forcertain diagnosis [6]. If, over the courseof disease, there is any uncertainty as towhether it is truly erythema nodosum, aprompt biopsy is warranted. The adipose tissue is structured in lobuleswhich are separated by thin connectivetissue septae containing blood vessels.Panniculitis may be divided histologically

into forms that primarily affect either thelobules or the septae. It should be notedthat the fine connective tissue septae areeasily destroyed by inflammation withpus formation, preventing their asses-sment. Histological analysis must assessboth the fatty tissue architecture, inclu-ding the blood vessels, as well as focal tis-sue changes such as emboli, and thus asufficiently large and deep biopsy is ne-cessary [7–9]. Yet taking a large biopsyfrom inflamed/edematous tissue carriesthe risk of abnormal wound healing andscar dehiscence with resulting cosmeticdisfigurement. Physicians are often un-derstandably reluctant to perform suchprocedures [2]. Along with the sample,any relevant clinical or pre-clinical infor-mation must also be communicated tothe histologist, including information onprior disease or immunosuppression as

this aids a reliable assessment of possibleskin disease and is particularly importantin panniculitis [10].Based on four patients with panniculitis,we describe various aspects of biopsytechnique and histopathological challen-ges in diagnosing panniculitis.

PatientsPatient 1A 71-year-old man was admitted to ourintensive care unit (Figure 1) after a shar-ply-bordered, palm-sized, red patch sud-denly appeared on his left inguinal region.The patient had undergone a cystectomyto treat bladder cancer and had been givenmultiple antibiotics to combat post-ope-rative sepsis. Based on clinical suspicion ofan early-stage fixed drug eruption, a 4mm large punch biopsy was taken. The results of histology showed a normaljunction zone, ruling out a fixed drugeruption. Yet in the included adipose tis-sue, neutrophilic, lobular panniculitiswas identified. Over the course of the disease, a later examination found a circumscribed redmacule on the opposite inguinal region.The patient reported that the previousnight he had placed an ice pack on thesite to lower his temperature. The disor-der was diagnosed as cold panniculitis,based on the clinical/pathological corre-lation. The skin changes resolved sponta-neously, gradually healing fully as is typical for this form of panniculitis [11].

Patient 2A cachectic 39-year-old man in pooroverall health came to our clinic withasymptomatic red patches that had ap-peared two weeks earlier and had sincebecome redder (Figure 2). The first lesi-ons had appeared on the legs, followed bylesions on the trunk. Based on the clinical

appearance, a presumptive diagnosis ofleukocytoclastic vasculitis was made andtwo of the lesions were biopsied. Histological analysis of the punch biop-sies showed a normal epidermis and der-mis. In the adipose tissue, there werecentral areas of basophilic necrosis withcalcification surrounded by a neutrophil-rich inflammatory infiltrate. On further examination, cirrhosis of theliver and chronic pancreatitis were dia-gnosed. Serum levels of lipase and amy-lase were dramatically elevated against abackdrop of chronic alcoholism. Pancrea-tic panniculitis was diagnosed. This rarecomplication of pancreatitis often invol-ves significantly increased serum lipase,while amylase levels may be normal [12].Occasionally, serum pancreatic enzymesmay be normal, making diagnosis moredifficult [13]. The prognosis for patientswith pancreatic disease is serious. Our pa-tient died 6 weeks after the skin lesionsappeared, despite intensive medical care.

Patient 3A 48-year-old woman had an 8-week history of recurrent fever, joint pain, andincreasing pancytopenia with hemopha-gocytosis. Multiple, slightly painfulbrown nodules were found on both legs.The results of a punch biopsy revealedminimal lobular, lymphocytic panniculi-tis, which could not be more preciselyclassified. The hemophagocytosis of uncertain etiology was treated with high-dose steroids and cyclosporine. Thenodules on the legs were interpreted aserythema nodosum and treated withcompression dressings.One year later the woman was referred toour clinic again, this time with 5 cmlarge, red-brown nodules on both legs(Figure 3). Two large and deep spindle-shaped biopsies were performed.

2 Original Article Histopathology of panniculitis

JDDG | 2011 (Band 9) © The Authors • Journal compilation © Blackwell Verlag GmbH, Berlin • JDDG • 1610-0379/2011

Figure 1: (a) Cold panniculitis. Diffuse erythe-ma on the proximal upper thigh. (b) Punchbiopsy with lobular panniculitis. (c) Denseinflammatory infiltrate with neutrophils (c).

Figure 2: (a) Pancreatic panniculitis. Multiple red macules on the knee. (b) Punch biopsy with a sig-nificant inflammation in the subcutaneous lobule. (c) Basophilic fatty tissue residues (ghost cells) withdense infiltrate composed of many neutrophils.

Histological analysis of both samples re-veled lobular infiltration with atypicallymphocytes, arranged in a ring aroundsingle adipocytes (rimming). The atypi-

cal lymphocytes expressed CD3, CD8,TIA-1 and the beta chain of the T-cell re-ceptor but were negative for CD56 andCD4. The proliferation rate (MiB-1) was

50 %. Molecular biological studies sho-wed monoclonality of the T-cell infil-trate. A diagnosis of subcutaneous pan-niculitis-like T-cell lymphoma was made.The patient was treated with multiplechemotherapy drugs and autologousstem cell transplantation. The disease in-itially went into partial remission but re-curred later with pancytopenia, and 2.5years after the first skin lesions had ap-peared the woman died of sepsis. About 20 % of patients with this form oflymphoma develop hemophagocytosissyndrome, which has a high rate of fata-lity [14].

Patient 4A 53-year-old woman had a one-year hi-story of gradually progressing atrophy ofthe adipose tissue on both upper arms(Figure 4). A punch biopsy taken 10 yearspreviously showed a circumscribed lym-phocytic inflammatory reaction in thesmall amount of included fatty tissue. Nospecific diagnosis was made at that time.Magnetic resonance imaging (MRI) stu-dies of the upper arm now showed an in-flammatory process in the fatty tissue ex-tending to the fascia. The radiologist’sreport advised further histological analy-sis. The excision biopsy showed primarilylobular lymphocytic panniculitis with in-creased amounts of interstitial mucin.The patient was diagnosed with lupuspanniculitis. She had no other manifesta-tions of lupus erythematosus. Therapywith hydroxychloroquine was begun, andthere was no further progression of skinsymptoms. Lupus panniculitis often oc-curs in association with discoid lupuserythematosus and more rarely with sy-stemic lupus erythematosus [15].

DiscussionThis retrospective description of four pa-tients is intended to highlight importantaspects of biopsy procedures and histolo-gical analysis of panniculitis.As the patients with cold panniculitisand pancreatic panniculitis demonstrate,early forms of disease in particular mayhave clinically atypical manifestations. Inthese patients, panniculitis was not con-sidered in the early differential diagnosis.Diagnosis was confirmed by includingadipose tissue in the punch biopsy andsubsequent clinicopathological correla-tion. This underscores the importance ofincluding some of the fatty tissue in anypunch biopsy.

Histopathology of panniculitis Original Article 3

© The Authors • Journal compilation © Blackwell Verlag GmbH, Berlin • JDDG • 1610-0379/2011 JDDG | 2011 (Band 9)

Figure 3: (a) Subcutaneous pannuculitis-like T-cell lymphoma. Multiple red-brown nodules on bothlegs. (b) Punch biopsy with little subcutaneous fat. (c) Incisional biopsy with a dense infiltrate in thesubcutaneous tissue taken at the same microscopic magnification. (d) Atypical lymphocytes surrounda adipocyte (rimming).

Figure 4: (a) Lupus panniculitis. Marked atrophy of the fatty tissue on the left upper arm. (b)Magnetic resonance tomography of the upper arm with contrast agent to demonstrate the inflamma-tory process in the fatty tissue and fascia. (c) Punch biopsy with focal collection of lymphocytes. (d)Incisional biopsy with a dense predominantly lymphocytic panniculitis taken at the same microscop-ic magnification.

Punch biopsies are quick and easy, andare thus preferred by clinicians over scalpel biopsies. At body sites with a thindermis, such as the forearm or extensorsurface of the lower leg, a 6 mm punchbiopsy, even for confirmation of a pre-sumptive diagnosis of panniculitis, caninclude a sufficient amount of fatty tissue. Depending on the affected site,the sex of the patient and nutritional status, the thickness of the dermis variesfrom 0.5 mm to more than 8 mm. Whentaking the sample, care must be taken toavoid crush artifacts caused by the for-ceps. Small punch biopsies generally donot include sufficient fatty tissue and arethus inadequate.The other two patients illustrate the limits and the risks of taking a biopsythat is too small or too superficial. Theresult was a considerably delayed diagno-sis of subcutaneous panniculitis-like T-cell lymphoma in the one and lupuspanniculitis in the other. The initial punch biopsies did not enablea specific diagnosis; rather the histopa-thologist wrote only a descriptive report.In the patient with lymphoma, only mi-nimal lobular, lymphocytic panniculitiswas observed, while in the patient withlupus panniculitis there was only focallymphocytic panniculitis.In the lymphoma patient, a large biopsycould only be taken after recurrence ofthe skin lesions and hemophagocytosiswith fever. In the patient with lupus pan-niculitis a further biopsy was advised bya radiologist based on the results of MRI. If there is any uncertainty, and if an initialbiopsy fails to confirm the diagnosis, aspindle-shaped biopsy should be perfor-med with a scalpel. The same applies toparts of the body with a thick dermissuch as the thighs and back. A sufficientamount of tissue should be obtained. It isdifficult to distinguish on histology bet-ween lymphoma and an inflammatoryprocess. Sufficient tissue should also beobtained for immunohistological andmolecular biological studies or even con-sultation with a specialized center. Thistype of biopsy is much more difficultthan a punch biopsy. The physician mustplan to allow for more time to performthe procedure. The specimen shouldmeasure 2–3 cm in length. When using ascalpel to take the sample, care should betaken to first tilt it inwardly as otherwisean inadequate amount of fatty tissue willbe included. The tissue should immedia-

tely be placed in a container with an ade-quate amount of formalin solution. Thetissue should not be first placed on cottongauze as this causes significant artifactswhich prevent an evaluation of cellularmorphology. This is especially importantin regard to lymphoma [16]. After clo-sure of the wound, the site should be co-vered with a strong compression dressing. In summary, the four patients presentedhere demonstrate that panniculitis maycertainly be diagnosed using a punchbiopsy. If there is any uncertainty, and ifprevious biopsies could not confirm thediagnosis, a large and deep spindle-sha-ped biopsy should be taken.

AcknowledgementWe thank Dr. M. Beese, Radiology Prac-tice at Tesdorpfhaus, Lübeck, Germany,for performing and interpreting theMRI. <<<

Conflict of interestNone.

Correspondence to

Priv.-Doz. Dr. med. Christian RoseDermatohistologisches Labor LübeckMaria-Goeppert-Straße 1D-23562 LübeckTel.: +49-451-50-270-50Fax: +49-451-50-270-55 E-mail: rose@dermatohistologie-luebeck.de

References1 Weyers W, Diaz C, Weyers I, Borghi S.

Die Hautbiopsie. Hautarzt 1999; 50:145–58.

2 Requena L. Indroduction. SeminCutan Med Surg 2007; 26: 65.

3 Requena L, Sánchez Yus E. Erythemanodosum. Semin Cutan Med Surg2007; 26: 114–25.

4 Cribier B, Caille A, Heid E, GrosshansE. Erythema nodosum and associateddiseases. A study of 129 cases. Int JDermatol 1998; 37: 667–72.

5 Bohn S, Büchner S, Itin P. Erythemanodosum: 112 Fälle. Epidemiologie,Klinik und Histopathologie. SchweizMed Wochenschr 1997; 127:1168–76.

6 Chopra R, Chhabra S, Thami GP, Punia RPS. Panniculitis: clinical over-lap and the significance of biopsy fin-dings. J Cutan Pathol 2010; 37: 49–58.

7 Requena L, Sánchez Yus E. Panniculi-tis. Part I. Mostly septal panniculitis. JAm Acad Dermatol 2001; 45: 163–83.

8 Requena L, Sánchez Yus E. Panniculi-tis. Part II. Mostly lobular panniculitis.J Am Acad Dermatol 2001; 45:325–61.

9 Diaz-Cascajo C, Borghi S, Weyers W.Pannikulitis. Definition of terms anddiagnostic strategy. Am J Dermatopa-thol 2000; 22: 530–49.

10 Kerl H, Cerroni L. „Standard of reason-able care“ in der Dermatopathologie. JDtsch Dermatol Ges 2011; 9: 721–2.

11 Wagner G, Sachse MM. Die Kältepan-nikulitis bei Reiterinnen. Akt Dermatol2010; 36: 403–7.

12 Hasenöhrl K, Stein A, Wozel G, Meu-rer M. Pankreopathische Pannikulitisbei neuroendokrinen Karzinom desPankreas. Hautarzt 2006; 57: 237–41.

13 Jacobson-Dunlop E, Takiguchi R,White CR, White KP. Fatal pancreatitispresenting as pancreatic panniculitiswithout enzyme elevation. J Cutan Pathol 2011; 38: 456–7.

14 Willemze R, Jansen PM, Cerroni L,Berti E, Santucci M, Assaf C, Can-ninga-van Dijk MR, Carlotti A, GeertsML, Hahtola S, Hummel M, Jeskanen L, Kempf W, Massone C, Ortiz- Romero P, Paulli M, Petrella T, RankiA, Peralto JLR, Robson A, Senff NS,Vermeer MH, Wechsler J, Whittaker S,Meijer CJLM. Subcutaneous pannicu-litis-like T-cell lymphoma: definitionclassification, and prognostic factors:an EORTC cutaneous group sudy of83 cases. Blood 2008; 111: 838–45.

15 Kuhn A, Sticherling M, Bonsmann G.Klinische Manifestation des kutanenLupus erythematodes. J Dtsch Derma-tol Ges 2007; 5: 1124–40.

16 Kerl H, Cerroni L, Fink-Puches R,Wolf IH. Primäre kutane B-Zell- Lymphome. J Dtsch Dermatol Ges2004; 2: 605–21.

4 Original Article Histopathology of panniculitis

JDDG | 2011 (Band 9) © The Authors • Journal compilation © Blackwell Verlag GmbH, Berlin • JDDG • 1610-0379/2011