The Journal of Bone and Joint Surgery 81:1013-18 (1999)
© 1999 The Journal of Bone and Joint Surgery, Inc.
Pyoderma Gangrenosum Mimicking Postoperative Infection in the Extremities. A Report of Two Cases*
CRAIG R. BENNETT, M.D. ,
MICHAEL E. BRAGE, M.D. and
DANIEL P. MASS, M.D. , CHICAGO, ILLINOIS
Investigation performed at the Section of Orthopaedic Surgery, Department of Surgery, University of Chicago Hospitals, Chicago
 |
Introduction
|
|---|
Pyoderma gangrenosum is an uncommon skin disorder involving ulcerations and necrosis that may occur following an operation and mimic a postoperative infection. We report on two patients who had pyoderma gangrenosum that was not diagnosed initially. Both patients were managed with repeated operative débridement because of a suspected postoperative infection. The diagnosis of pyoderma gangrenosum was eventually established; however, this disorder is not commonly treated by orthopaedic surgeons and is often initially misdiagnosed.
 |
Case Reports
|
|---|
CASE 1. A thirty-two-year-old woman was first managed at an outside institution because of pain in the heel and tarsal tunnel syndrome on the right side. At our request, the referring physician provided the patient's earlier clinical record so that we could better describe the pertinent history and management in the present report. According to the office notes, the patient was first seen in November 1995 because of a several-month history of severe pain in the right heel. The pain was deep and aching and was worse after walking or prolonged rest. The patient had no relevant medical history and had had no previous operations. There was no family history of rheumatological disorders.
Physical examination showed a well perfused foot with palpable dorsalis pedis and posterior tibial pulses. The plantar fascia on the plantar and medial aspects of the heel was tender to palpation. Palpation of the tarsal tunnel was painful, and percussion over the posterior tibial nerve elicited a positive Tinel's sign that radiated distally into the plantar aspect of the foot and digits.
The symptoms persisted despite treatment with anti-inflammatory medications, orthotics, and a trial of non-weight-bearing with use of crutches, but they did improve temporarily following the use of a Medrol Dosepak (a dose of twenty-eight milligrams of methylprednisolone was given on the first day, and the dose then was decreased by four milligrams per day over seven days). When the symptoms recurred, immobilization of the foot with a cast-brace was attempted, but the patient found the device too difficult to use.
At this point, the treating physician offered an operative approach, as will be described. The preoperative workup included radiographs, magnetic resonance imaging scans, an electromyogram, and nerve-conduction-velocity studies. According to the report supplied by the referring physician, the results of these examinations were normal.
On December 19, 1995, the patient was managed, at an outside hospital, with a tarsal tunnel release, release of the deep fascia overlying the abductor hallucis muscle, and a plantar fasciectomy. A small portion of the abductor hallucis muscle was removed in order to fully decompress the posterior tibial nerve. Nine days after the index operation, there was purulent drainage from the wound. The patient was readmitted to the hospital, and the wound was incised and drained. According to the operative report, twenty to thirty milliliters of serosanguineous fluid was expressed from the wound. Pustules and areas of necrotic tissue were noted along the margins of the operative wound and on all digits of the ipsilateral foot. The lesions were debrided sharply and washed. Staphylococcus aureus and Enterobacter cloacae grew on initial culture of specimens from the wound. An infectious-disease consultation was obtained, and the patient was managed with intravenous administration of antibiotics (eighty milligrams of gentamicin every eight hours and three grams of Unasyn [ampicillin-sulbactam] every six hours). A home program of intravenous antibiotic therapy was arranged, and the patient was discharged.
The wound worsened despite the initial débridement and the antibiotic therapy. The outpatient notes revealed that the pain increased and blistering of the operative wound progressed. Cultures of specimens obtained in the clinic were negative. Oral administration of Augmentin (amoxicillin-clavulanate) was added to the antibiotic regimen. The patient was readmitted to the outside institution on January 19, 1996, and intravenous administration of vancomycin (one gram every twelve hours) was begun. On January 23, a consulting plastic surgeon performed a second incision and drainage, with skin-grafting. The operative report lacks details, indicating only that the right foot was swollen and that the wound had a necrotic surface as well as some purulent drainage. The necrotic tissue was removed sharply. A skin graft was taken from the right thigh, meshed, and fixed with staples to the right leg and ankle. Cultures of intraoperative specimens were negative.
Subsequently, a total of six operative débridements were performed at the outside institution. In addition to the plastic-surgery and infectious-disease consultants mentioned earlier, the primary physician and other general medical physicians assisted in the care of the patient. The patient continued to have so-called cellulitis of the ankle as well as pustular and vesicular lesions of the foot. She refused the option of a below-the-knee amputation and was transferred to the University of Chicago Hospitals on February 6.
As an infection was still the primary suspicion, the patient was immediately taken to the operating room for débridement on arrival at our institution (Fig. 1-A). At the time of the operation, the foot was swollen and there was evidence of cellulitis. Numerous pustules and multiple areas of necrosis were observed throughout the region. Areas that had been debrided previously were covered with yellow coagula. The foot and ankle were debrided sharply of all necrotic tissue. After all dead and affected tissue had been removed down to the subcutaneous tissues, the subcutaneous tissue bled and appeared to be uninvolved. The skin over the heel and metatarsal fat pads appeared viable and was retained. All wounds were irrigated with six liters of saline solution. Cultures for bacteria and fungi were negative. The patient was also seen by a plastic surgeon, a rheumatologist, and an infectious-disease specialist, who recommended that she be managed with intravenous administration of Zosyn (piperacillin-tazobactam) (3.375 grams every six hours). She had daily therapy sessions, which involved use of a whirlpool and treatment with hyperbaric oxygen. Another operative débridement was performed on February 13. Only minimum necrotic tissue was noted during the operation, and no new pustules were identified. Several centimeters of skin was removed from the anterolateral aspect of the ankle because it appeared to be erythematous. A curet was used to lightly debride the coagula covering the remaining wounds. The wounds appeared to stabilize and began to granulate. Plastic surgeons performed split-thickness skin-grafting on February 19. The skin graft had a 90 percent take, and the patient was discharged on February 29.

View larger version (96K):
[in this window]
[in a new window]
|
Figs. 1-A, 1-B, and 1-C: Case 1.
Fig. 1-A: Photograph, made when the patient was first seen at our institution, demonstrating severe necrotic ulceration throughout the medial aspect of the right foot.
|
|
The patient was readmitted on March 7, 1996, because of continued severe pain in the foot, with new areas of ulceration and blistering (Fig. 1-B). New pustular lesions also were observed on the ipsilateral thigh near the skin-graft donor site. The plastic-surgery consultant unroofed these pustules at the bedside and obtained specimens for culture. Again, all cultures were negative. Magnetic resonance imaging of the foot and ankle revealed no evidence of deep infection. A dermatology consultation was obtained, and pyoderma gangrenosum was immediately suspected. Antibiotic therapy was stopped, and oral administration of prednisone (eighty milligrams per day) was started. The third toe had become avascular secondary to severe involvement, and a partial amputation was performed. The rest of the foot began to heal, and, on the recommendation of the dermatologist, no additional operative intervention was attempted. The foot gradually healed over the next three months (Fig. 1-C). The steroid medications were slowly tapered off until the patient was no longer taking them. She continued to report sharp pain over the plantar surface of the foot, with the exception of the heel, which remained numb. The patient had no recurrence of pyoderma gangrenosum for five months, at which point she was lost to follow-up.

View larger version (88K):
[in this window]
[in a new window]
|
Fig. 1 Photograph, made when the patient was readmitted to the hospital on March 7, 1996, demonstrating failure of the skin graft and continued necrotic ulceration throughout the entire foot. This photograph was made before the diagnosis of pyoderma gangrenosum was established.
|
|

View larger version (117K):
[in this window]
[in a new window]
|
Fig. 1 Photograph made approximately two months after the initiation of steroid treatment. The dorsum of the foot is completely closed. Note the partial amputation of the third toe, which was performed after the toe became avascular during the initial phase of treatment.
|
|
CASE 2. A twenty-five-year-old woman who had a history of juvenile rheumatoid arthritis and Crohn disease was first seen at the University of Chicago Hospitals in 1990 for the treatment of severe deformities and pain in both wrists and hands. The rheumatoid arthritis had caused volar subluxation of the second through fifth metacarpophalangeal joints of the right hand. The involved digits were not functional and were rigidly flexed at 50 degrees. The second through fourth metacarpophalangeal joints of the left hand were painful, stiff, and swollen, with erosions and narrowing of the joint space seen on radiographs, but they were not subluxated. Both wrists were painful and had a restricted range of motion. Radiographs of both wrists showed severe destruction of the radiocarpal articulations and osseous erosions of the radial and ulnar styloid processes. On July 6, 1992, the patient was managed with a left radiocarpal arthrodesis; metacarpophalangeal arthroplasties of the left index and long fingers, performed with use of silicone implants; and a synovectomy of the metacarpophalangeal joint of the left ring finger. In November 1992, metacarpophalangeal arthroplasties with silicone implants were performed in all four fingers of the right hand. In March 1993, the patient had a right radiolunate arthrodesis and resection of the right ulnar head. In October 1995, the silicone implants in the index and long fingers of the left hand were revised because they had broken and were causing pain. Arthroplasties with silicone implants were also performed in the ring and little fingers of the left hand because of pain and progressive disease.
The silicone implants in the right hand failed, and revision arthroplasty of the third, fourth, and fifth metacarpophalangeal joints was performed on August 4, 1997. Ten days after the operation, the patient was seen at an outside emergency room because of increasing pain in the right hand as well as a fever of 39.5 degrees Celsius. She was diagnosed as having cellulitis and was given one 500-milligram dose of vancomycin intravenously. She was referred back to the operating surgeon for care. When the patient returned to the University of Chicago Hospitals the next day, the operative wound appeared to be grossly infected, with drainage of purulent material. On August 15, after removal of the sutures in the operating room, the wound spread apart easily and there was drainage of a purulent exudate. The involved skin and subcutaneous tissues were debrided, and the specimens were sent for culture and gram-staining. The capsules of the metacarpophalangeal joints were intact but were opened because of concern about deep infection. The joint fluid appeared to be purulent, and all of the silicone implants were removed. The wound was lavaged thoroughly with saline solution and dressed (Fig. 2-A). Postoperatively, the patient was managed with intravenous administration of vancomycin (750 milligrams every twelve hours).

View larger version (124K):
[in this window]
[in a new window]
|
Figs. 2-A and 2-B: Case 2.
Fig. 2-A: Photograph of the right hand, made after operative débridement, demonstrating the retraction of the skin edges and the necrotic appearance of the wound, which could be mistaken for a severe and rapidly spreading infection.
|
|
On August 18, the patient was in excruciating pain and had a fever of 39.8 degrees Celsius. She was taken to the operating room for repeat débridement. The wound appeared worse, with copious drainage, formation of bullae throughout the hand, and retracted skin edges, which had a purple hue. There was also what appeared to be a carbuncle along the lateral aspect of the right thigh. Incision of this lesion yielded purulent material, which extended to the iliotibial band; the area was debrided and the wound was packed open.
Cultures and gram stains of specimens that had been obtained during the first débridement were found to be negative. Because of the suspicious appearance of the wound at the time of the second débridement, infectious-disease and dermatology consultations were ordered. Pyoderma gangrenosum was immediately suspected. Antibiotic therapy was stopped, and intravenous administration of SoluMedrol (methylprednisolone sodium succinate; one gram per day) was started. The intravenous therapy was discontinued after three days, at which time oral administration of methylprednisolone (100 milligrams per day) was begun. The fever subsided within twenty-four hours after the initiation of steroid therapy, and the patient remained afebrile for the rest of her hospital stay. The resultant wound encompassed the entire dorsum of the hand. Additional operative intervention was avoided, and the wound contracted over the next three months (Fig. 2-B). The metacarpophalangeal joints achieved an arc of motion from 0 to 40 degrees, and the steroid medication was slowly tapered off over three months as the disease went into remission. At the time of the twelve-month follow-up, the patient had had no recurrence.

View larger version (116K):
[in this window]
[in a new window]
|
Fig. 2 Photograph demonstrating contraction and healing of the wound after local wound care and intravenous administration of SoluMedrol (methylprednisolone sodium succinate) followed by oral administration of methylprednisolone.
|
|
 |
Discussion
|
|---|
Pyoderma gangrenosum is an ulcerative, necrotic dermatosis of unknown etiology9. It was first described in 1930 by Brunsting et al.3, who believed that streptococcal infection was a major component leading to secondary cutaneous gangrene. This theory subsequently was shown to be false11, but, because of the clinical appearance of the disease, the name pyoderma gangrenosum has persisted.
The prevalence of pyoderma gangrenosum is difficult to determine. In one report, fifteen patients with this condition were seen over a ten-year period at a regional dermatology center serving a population of 0.5 million12. Pyoderma gangrenosum mainly affects adults who are between twenty-five and fifty-four years old12, but it can occur in childhood, when it is often associated with a systemic disease such as inflammatory bowel disease, immunodeficiency, leukemia, or infection with the human immunodeficiency virus12,13. In adults, pyoderma gangrenosum is also often associated with systemic illness such as inflammatory bowel disease, rheumatoid arthritis, chronic active hepatitis, leukemia, or monoclonal gammopathy of the IgA type. However, it may occur in healthy people13 and can develop following trauma or an operative procedure1,2,5,6,8,9.
Several current theories suggest that pyoderma gangrenosum may be caused by serum immune complexes or defects in immune mechanisms. Powell et al.12 observed that 1.9 percent (twenty-one) of 1132 patients with Crohn disease had pyoderma gangrenosum and found that the activity of the skin disease often paralleled that of the inflammatory bowel disease. Those authors noted that the increased bowel permeability associated with Crohn disease results in increased bacterial antigenemia and subsequent formation of serum immune complexes. Use of direct immunofluorescence in patients who have pyoderma gangrenosum has demonstrated a number of immunopathological findings, including the presence of IgM, C3, fibrin, and, occasionally, IgG and IgA in the superficial and deep dermal blood vessels at the periphery of the lesions15. It is possible that pyoderma gangrenosum results from the immune response to the deposition of these complexes in the cutaneous vessels6. However, the endogenic or ectogenic antigens to which these immunoglobulins bind have not been identified14.
Defects in both humoral and cell-mediated aspects of the immune system have been described in patients with pyoderma gangrenosum12. Cell-mediated abnormalities include cutaneous anergy to various substances such as Candida or purified protein derivative12. Other studies have demonstrated an intrinsic defect involving neutrophils, with impaired chemotaxis and abnormal phagocytosis16. Rand et al.13 postulated that an abnormal cell-mediated response to damaged tissue may cause an intradermal inflammatory process that results in the lesions of pyoderma gangrenosum.
Pyoderma gangrenosum can localize at the site of trauma or an operative procedure. This type of response, termed pathergy, may represent a misdirected, host-mediated cellular response to tissue that has been antigenically modified as a result of trauma or an operation in a person who demonstrates altered immune reactivity13. Because pyoderma gangrenosum is often seen in patients who are immunosuppressed, either as a result of treatment of a disease or as a direct result of disease, the development of this skin disorder might simply be a reflection of a generalized immunological defect. Although many theories regarding the cause of pyoderma gangrenosum are centered around an alteration in immunity, no one immune deficiency common to these patients has ever been demonstrated16.
The appearance of the lesion is characterized by erythema and edema as well as necrotic ulceration. The rapidity with which the borders may expand, the temporal proximity to an initial operative procedure, and the fact that cultures that may be positive as soon as the wound is colonized often lead the surgeon to the mistaken belief that the process is a rapidly spreading infection that needs urgent débridement. Because pyoderma gangrenosum is often a pathergic process, the trauma associated with the operation only exacerbates the condition; therefore, additional procedures are contraindicated until the disease is in a quiescent stage9.
The diagnosis of pyoderma gangrenosum is based on clinical findings as there are no helpful diagnostic tests. Even biopsies of the lesions are nondiagnostic and demonstrate nonspecific areas of necrosis and ulceration that are characterized by the infiltration of large numbers of acute and chronic inflammatory cells15. Nevertheless, tissue must be obtained during the workup in order to exclude other diseases. The differential diagnosis includes postoperative infection, brown recluse spider bite, cutaneous amebiasis, North American blastomycosis, Meleney gangrene, factitious ulcer, ulcerated necrobiosis lipoidica, and erythema multiforme16. If the patient has not had operative débridement, tissue may be obtained by means of a skin biopsy so that the histological features of the lesion can be examined and cultures can be performed. The possibility of exacerbating a pathergic reaction should be taken into consideration before a skin biopsy is performed. However, the necessity of establishing a diagnosis and ruling out other disease processes usually outweighs this risk12.
Once the diagnosis of pyoderma gangrenosum has been established, treatment involves a combination of local wound care and systemic therapy. Powell et al.12 recommended the use of local therapy for the reduction of pain, the prevention and treatment of secondary bacterial infections, and the maintenance of a suitable environment for wound-healing. Application of antibacterial creams, or even a topical sulfone, and gentle daily lavage with sterile saline solution are recommended10. Skin-grafting is to be avoided because the grafts often fail and are associated with the development of new lesions at the donor site5,16.
Systemic treatment often begins with the administration of corticosteroids. Usually, prednisone is given orally in daily doses of sixty to 120 milligrams7,12. The use of steroids is continued until the disease is under control, as indicated by healing of the ulcers and decreasing pain. Occasionally, corticosteroids either fail or are contraindicated, in which case a variety of second-line agents may be used. Sulfones such as dapsone act by inhibiting neutrophils; they also appear to prevent edema and may decrease acute inflammatory reactions12. Cyclosporin A is increasingly being used for patients in whom the disease is refractory to all other forms of treatment4. Cyclosporin A appears to act on T lymphocytes and to limit the production of interleukin-2. The proliferation of T-helper cells is inhibited whereas the production of T-suppressor cells is unaffected, resulting in the reversal of the ratio of T-helper cells to T-suppressor cells4.
Several factors may minimize the risk of recurrence of pyoderma gangrenosum at the site of an operation in a patient who has a history of this condition. Long et al.9 noted the development of pyoderma gangrenosum at the points of suture entry and exit in a patient who had had a total hip arthroplasty. When a subsequent total hip arthroplasty was performed on the contralateral side, subcuticular sutures were used for skin closure and the operative site healed uneventfully. Those authors concluded that the trauma of puncturing the skin with sutures may exacerbate the disease. It seems prudent, therefore, to recommend an atraumatic skin closure in patients who have a history of pyoderma gangrenosum. More importantly, those authors recommended delaying elective operative procedures until the pyoderma gangrenosum is in a quiescent state.
In summary, as reported in the literature1,5,8,10,14,16 and as demonstrated by the cases of our two patients, pyoderma gangrenosum is often mistaken for a postoperative infection and is treated with operative débridement and administration of antibiotics. Operative treatment is overwhelmingly contraindicated because of its well documented propensity to exacerbate the condition. Surgeons must, therefore, be wary of the existence of this disease.
A characteristic feature of pyoderma gangrenosum is its acute and rapid development. A small, red lesion develops at the site of the operative wound and, within hours, becomes painful, prominent pustules16. The pustules spread rapidly and, within a few days, become necrotic and ulcerated. The borders of the ulcers have a purple hue, are erythematous and irregular, and may spread one to two centimeters per day in a centrifugal pattern. It is extremely difficult to distinguish these lesions from a wound infection requiring débridement.
We believe that surgeons will continue to perform operative débridement for the treatment of an assumed infectious process when pyoderma gangrenosum develops following an orthopaedic procedure, but we hope that awareness of this entity will make it possible for repeated débridements to be avoided. We wish to emphasize that the wounds of pyoderma gangrenosum are characterized by ulcerations, purulent bases, and irregular, erythematous borders with a purple hue. If the lesion does not immediately improve after débridement and administration of antibiotics, and if cultures are negative, then a diagnosis of pyoderma gangrenosum should be considered.
 |
Footnotes
|
|---|
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
The Center for Bone and Joint Disease, 5319 Grand Boulevard, New Port Richey, Florida 34652.
Department of Orthopaedics, University of California, San Diego, 200 West Arbor Drive, MC 8894, San Diego, California 92103. E-mail address for Dr. Brage: mbrage@ucsd.edu.
Section of Orthopaedic Surgery, Department of Surgery, University of Chicago Hospitals, 5841 South Maryland Avenue, MC 3079, Chicago, Illinois 60637.
 |
References
|
|---|
-
Adam, D. J.; Nawroz, I.; and Petrie, P. W.: Pyoderma gangrenosum severely affecting both hands. J. Hand Surg., 21-B: 792-794, 1996.
-
Brown, R. E.; Lay, L.; and Graham, D.: Bilateral pyoderma gangrenosum of the hand: treatment with dapsone. J. Hand Surg., 18-B: 119-121, 1993.
-
Brunsting, L. A.; Goeckerman, W. H.; and O'Leary, P. A.: Pyoderma (echthema) gangrenosum. Clinical and experimental observations in five cases occurring in adults. Arch. Dermatol. and Syphilol., 22: 655-680, 1930.
-
Curley, R. K.; Macfarlane, A. W.; and Vickers, C. F.: Pyoderma gangrenosum treated with cyclosporin A. British J. Dermatol., 113: 601-604, 1985.[Medline]
-
Esnault, P.; Dompmartin, A.; Moreau, A.; Caraes, B.; and Leroy, D.: Recurring postoperative pyoderma gangrenosum. Internat. J. Dermatol., 34: 647-650, 1995.
-
Harland, C. C., and Millard, L. G.: Pyoderma gangrenosuma complication of chronic venous leg ulceration?. Clin. and Exper. Dermatol., 18: 545-547, 1993.
-
Kann, S. E.; Jacquemin, J.; and Stern, P. J.: Simulators of hand infections. In Instructional Course Lectures, American Academy of Orthopaedic Surgeons. Vol. 46, pp. 69-82. Rosemont, Illinois, American Academy of Orthopaedic Surgeons, 1997.
-
Laurencin, C. T., and Shoen, S. L.: Pyoderma gangrenosum affecting the hand. J. Bone and Joint Surg., 76-B(6): 985-986, 1994.
-
Long, C. C.; Jessop, J.; Young, M.; and Holt, P. J.: Minimizing the risk of post-operative pyoderma gangrenosum. British J. Dermatol., 127: 45-48, 1992.[Medline]
-
Papilion, J. D., and Bergfield, T. G.: Pyoderma gangrenosum complicating infection of the hand. A case report and review of the literature. Clin. Orthop., 254: 144-146, 1990.
-
Percival, G. H.: Pyoderma gangrenosum: the histology of the primary lesion. British J. Dermatol., 69: 130-136, 1957.[Medline]
-
Powell, F. C.; Su, W. P.; and Perry, H. O.: Pyoderma gangrenosum: classification and management. J. Am. Acad. Dermatol., 34: 395-409, 1996.[Medline]
-
Rand, R. P.; Olerud, J. E.; and Verrier, E. D.: Pyoderma gangrenosum after coronary artery bypass grafting. Ann. Thoracic Surg., 55: 1016-1018, 1993.[Abstract]
-
Shands, J. W., Jr.; Flowers, F. P.; Hill, H. M.; and Smith, J. O.: Pyoderma gangrenosum in a kindred. Precipitation by surgery or mild physical trauma. J. Am. Acad. Dermatol., 16: 931-934, 1987.[Medline]
-
Su, W. P.; Schroeter, A. L.; Perry, H. O.; and Powell, F. C.: Histopathologic and immunopathologic study of pyoderma gangrenosum. J. Cutan. Pathol., 13: 323-330, 1986.[Medline]
-
Wustrack, K. O., and Zarem, H. A.: Pyoderma gangrenosum: recognition and management. Plast. and Reconstr. Surg., 62: 423-428, 1978.

CiteULike Connotea Del.icio.us Facebook Technorati Twitter What's this?
This article has been cited by other articles:

|
 |

|
 |
 
E.L. Newell, S.I. Chaudhry, and M.M. Black
Acute Bullous Hemorrhagic Pyoderma Gangrenosum: Pitfalls of Diagnostic Delay. A Case Report
J. Bone Joint Surg. Am.,
January 1, 2008;
90(1):
174 - 177.
[Full Text]
[PDF]
|
 |
|
|