Pyoderma Gangrenosum in the Head and Neck

A 72-year-old white woman presented to her local ear, nose, and throat (ENT) physician with a painful swelling of the parotid gland. A wide incision of the parotid was made, and purulent drainage could be expressed. She was treated with systemic antibiotic drugs. The woman's condition failed to improve, and the ENT physician admitted her to the hospital with a persistent purulent salivary fistula and suspected pyogenic parotitis.

On examination, a dense swelling at the tail of the right parotid and a salivary fistula at the right mandibular angle with purulent secretion were found. No other local signs of inflammation were obvious. The laboratory results of routine blood tests (eg, erythrocyte sedimentation rate and glucose level) seemed to be within reference ranges; the C-reactive protein level and white blood cell count were also within reference ranges. All branches of the facial nerve were fully intact. On palpation of the face and neck, no enlarged lymph nodes or other masses could be found. The excretory ducts of the salivary glands showed no abnormalities, and all other otorhinolaryngologic examinations and results were within reference range.

Smear tests were performed by carefully rotating the swab to ensure that a good inoculum had been obtained from skin ulcers in an area approximately 1 cm in diameter. Bacteria or mycoid structures could not be detected. Fine-needle aspiration cytologic examination findings of the parotid were nonspecific and showed an increase of lymphocytes, histiocytes, and neutrophilic granulocytes. Broad spectrum antibiotics such as amoxicillin/clavulanate potassium (Augmentin; Beragena Arzneimittel GmbH, Baden Baden, Germany) were administered. Unsuccessful attempts were made to reduce salivary secretions with cauterization of the wound with polycresulene and the additional application of Scopoderma TTS plaster (Novartis AG, Basel, Switzerland). The salivary secretions were treated with injection of botulinum toxin type A into the parotid with ultrasonographic guidance.

Four weeks later, a parotidectomy, including excision of the fistula, was performed. The pathological report described moderate sialadenitis and severe inflammation of the skin with infiltration of lymphocytes, monocytes, and giant cells. There was no indication for a specific infection (eg, sarcoidosis or tuberculosis) or for malignancy. The patient was discharged on the fourth day after surgery with a closed wound and further antibiotic treatment (levofloxacin).

After 4 weeks, the patient was readmitted to the hospital with wound healing problems. The patient displayed a painful wound dehiscence with a yellow secretion in the infra-aural suture area. The edge of the wound was violet with underlying skin necrosis (Figure). Although systemic signs of inflammation such as pyrexia or elevated erythrocyte sedimentation rate were again missing, a profound abscess was considered. Different systemic antibiotic regimens (piperacillin sodium, metronidazole, clindamycin phosphate, clarithromycin, and ceftazidime pentahydrate) and topical antibiotics (fusidinate) were administered for weeks.

An adverse effect from the suture material or infection with atypical mycobacterium was suspected. Eight weeks after the parotidectomy, a revision wound debridement, wound lavage, and primary wound closure were performed.

Four weeks later, the patient was again admitted with wound dehiscence and unchanged secretion. The edge of the wound was again painful, with aggravated livid skin necrosis and bullous-pustular ulcers. Following the advice of the Institute of Microbiology at the University Hospital Cologne (Cologne, Germany), smear testing was repeated extensively. This led to the cultural detection of Mycobacterium abscessus, Corynebacterium species, Staphylococcus lugdunensis, and Staphylococcus epidermidis, all of which are sensitive to clarithromycin. Unfortunately, closure of the dehiscence and reduction of secretion could not be achieved by systemic administration of clarithromycin.

During follow-up, no improvement occurred. The suspicion of a microbiological genesis thus was rejected, and the antibiotic therapy was stopped. Münchhausen syndrome was suspected. However, with 24 hours of occlusion dressing during 1 week, the symptoms worsened greatly. The patient's medical history was presented at an interdisciplinary infectious disease meeting where, among other diagnoses, one of autoimmune disease was proposed. Findings from rheumatoid factor, antinuclear antibody, and antineutrophil cytoplasmic antibody testing were negative for organisms. Because the patient did not have other symptoms of vasculitis, a diagnosis of PG was proposed. Therapy with 80 mg of oral prednisolone daily was initiated. Only 3 days later, a clear improvement of the lesion was observed, and the patient was discharged from hospital with the tentative diagnosis of PG. After 2 weeks, the wound was closed and dry with far fewer signs of inflammation (Figure). After 2 months, the dosage of prednisolone was reduced stepwise. One year later the patient was free of symptoms and without further recurrence.

Occurrence of PG is uncommon in the head and neck area. It has been described as being present in this region in only 5% of cases.^{4 - 6} Usually, PG occurs in the lower limbs, especially the pretibial region, where numerous links to rheumatic diseases and diabetes mellitus have been shown. It is frequently (in 40%-50% of cases) associated with inflammatory bowel disease (either ulcerative colitis or regional enteritis [Crohn disease]). Another link exists to rheumatoid arthritis. It is rarely seen in association with malignant tumors such as plasmocytomas, gammopathies, leukosis, or mucosa-associated lymphoma tissue lymphoma. However, in the present case, none of these associated diseases could be identified.

Pyoderma gangrenosum yields extracutaneous manifestations such as pulmonary infiltration in which the culture findings are negative for organisms. Other organs and systems that may be involved include the heart, central nervous system, gastrointestinal tract, eyes, liver, spleen, bone, and lymph nodes. In the area of ENT, the association between PG and Wegener granulomatosis is of importance. This should be specifically excluded in all cases of PG affecting the head and neck.⁷ The appearance of PG is relatively consistent with large ulcers that can spread rapidly showing underlying violaceous borders and a necrotic, purulent base.⁸ In most cases, a skin trauma is present in the patient's medical history. It starts as a bullous-pustular lesion with a small, red papule changing and developing ulcerations with slight bleeding from the ulcer bed. The lesion changes to necrotic, hemorrhagic, suppurative erosions and livid, irregular, undermined borders within a few weeks. In early stages, follicular and perifollicular inflammation with intradermal abscess formation are described.^{2 ,8 - 10} The ulcers can be very painful on contact. The diagnosis of PG is primarily based on clinical symptoms and its characteristic appearance whereas the microbiological test results that are negative for organisms are characteristic. However, in practice, the diagnosis is commonly made by excluding other causes of similar-looking cutaneous ulcerations.

The etiology of PG is still uncertain inasmuch as its pathophysiologic characteristics are poorly understood. It is thought to be an autoimmune disorder, which is in line with the links to the rheumatic diseases mentioned at the beginning of this section. Pyoderma gangrenosum affects both sexes, although a female predominance may exist. All ages can be affected, but PG occurs mainly in the fourth and fifth decades of life. Untreated, the ulcers may pursue an acute progressive course, continue to enlarge, persist unchanged, or may slowly heal. If they are persistent, treatment must be nonsurgical. Removal of the necrotic tissue and surgical debridement is contraindicated and results in exacerbation of the ulcer. Pyoderma gangrenosum responds to immunosuppressive agents, and the first-line treatment is oral corticosteroids. For adults, prednisolone is given (0.5-2 mg/kg daily by mouth [the pediatric dose is 4-5 mg/m² daily by mouth]). In cases of incomplete response to treatment with corticosteroids, cyclosporine may be added for remission induction.^{2 - 3 ,8 ,11} Topical therapies may be added to the treatment, including local wound care and topical corticosteroids. The prognosis of PG is generally good; however, recurrences are frequent, and residual scarring is common, especially if adequate treatment is delayed.

Other causes of ulcerations should be excluded to ensure adequate and sufficient therapy. The differential diagnoses include cutaneous abscesses, vasculitis, necrotizing fasciitis, necrotizing vasculitis, tick bites, atypical mycobacteriosis, cutaneous tuberculosis, Sweet syndrome, and ulcers of any etiology (venous, thrombotic, vasculitis, neuropathic, and atherosclerotic). Malignant disease, especially basal cell carcinoma and squamous cell carcinoma, must be excluded. Of utmost importance is the distinction between PG and (because of its possibly fatal outcome) necrotizing fasciitis.¹² Necrotizing fasciitis necessitates early surgical exploration.

In conclusion, apparent wound infections and nonhealing wounds that do not respond to treatment with antibiotics must raise suspicion toward a noninfectious etiology such as PG. Systemic corticosteroids are currently the mainstay of therapy for patients with PG. Surgery is contraindicated because it leads to an aggravation of the lesion.² Minimal wound debridement and skin grafting can be employed only if PG ulcers are nonprogressive and immunosuppression is given in parallel treatment. Because of the possible fatal outcome, necrotizing fasciitis must be primarily excluded.

Correspondence: Claus Wittekindt, MD, Department of Otorhinolaryngology, Friedrich-Schiller-Universität Jena, Lessingstraße 2, D-07743, Jena, Germany (claus.wittekindt@med.uni-jena.de).

Submitted for Publication: September 21, 2005; final revision received May 2, 2006; accepted May 9, 2006.

Author Contributions: Drs Wittekindt, Lüers, and Klussmann had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Wittekindt, Lüers, and Klussmann. Acquisition of data: Wittekindt and Lüers. Analysis and interpretation of data: Wittekindt, Lüers, and Klussmann. Drafting of the manuscript: Wittekindt and Lüers. Critical revision of the manuscript for important intellectual content: Wittekindt, Lüers, Klussmann, and Hüttenbrink. Statistical analysis: Lüers. Administrative, technical, and material support: Wittekindt. Study supervision: Wittekindt, Klussmann, and Hüttenbrink.

Financial Disclosure: None reported.