Case to highlight a rare differential diagnosis of necrotising fasciitis in the presence of a stoma: peristomal pyoderma gangrenosum

Background

Ostomies are required for the management of conditions such as congenital anomalies, colonic obstruction, inflammatory bowel disease (IBD), intestinal trauma or gastrointestinal malignancy. Peristomal skin complications may arise such as mechanical trauma, dermatitis, parastomal ulcerations, granulomas and rarely peristomal pyoderma gangrenosum (PPG).

Pyoderma gangrenosum (PG) is a neutrophilic dermatosis with unknown aetiology first described by Brunsting et al in 1930.1 It is characterised by chronic, recurrent and painful cutaneous ulcerations. A subtype—PPG—is very rare, occurs in 0.5–1.5 per million people annually and accounts for 15% of all PG cases.2

Due to its low incidence, there exists a potential for misdiagnosis as other more common conditions such as peristomal dermatitis, ulcerations and necrotising fasciitis (NF). Importantly, treatment of NF and PPG is very different: the surgical debridement used to eradicate infected tissue in NF is harmful in PPG, and debridement causes further trauma and exacerbates pathergy. Thus, this case report serves as a cautionary tale to highlight the importance of differentiating PPG from NF.

Case presentation

A 65-year-old woman, non-smoker, non-drinker, with no significant medical or surgical history, presented with bleeding per-rectum and change in bowel habits for a year. She was diagnosed to have a low rectal tumour with no other pathologies in the colon on colonoscopy. Biopsies of the tumour confirmed adenocarcinoma. There were no distant metastases on staging CT of the thorax, abdomen and pelvis. MRI of the rectum showed T2N0 disease with uninvolved circumferential resection margin. She underwent robotic abdominoperineal resection for low rectal cancer, recovered well and was discharged home on postoperative day 5.

On postoperative day 8, she returned to the emergency department with worsening pain and redness around the stoma and port sites. There were no associated symptoms of systemic sepsis, such as fever, chills and rigours. She reported normal stoma function. On clinical examination, she was afebrile at 36.9°C; her blood pressure was 100/56 mm Hg and pulse rate was 89 beats/min. Her abdomen was examined; stoma plate was removed revealing erythema and ulceration around the stoma, supraumbilical port site and ex-drain site. No crepitus was noted. The stoma lumen was patent with small area of mucocutaneous dehiscence at medial aspect (figure 1).

From the initial clinical picture, concern was that of an infective process and thus blood investigations and CT imaging of the abdomen and pelvis were performed to exclude any NF, peristomal or intra-abdominal pathologies.

Investigations

Blood investigations showed haemoglobin of 103 g/L (120–160), leucocytosis of 31.72×10⁹/L (4.0–10.0) and platelets 409×10⁹/L (140–440). Inflammatory markers C reactive protein (CRP) was raised at 106 mg/L (0.2–9.1) and procalcitonin was normal at 0.42 µg/L (<0.50).

CT of the abdomen and pelvis reported extensive dermal thickening surrounding the stoma with mild adjacent subcutaneous fat stranding, suggesting cellulitis or inflammatory changes. No subcutaneous or intra-abdominal free air or collections were detected (figure 2).

Differential diagnosis

The initial differential diagnosis was that of NF. However, the absence of crepitus clinically and lack of subcutaneous gas at the affected area on the CT precluded this diagnosis. Thus, the patient was started on intravenous antibiotics and monitored closely. The patient had persistent leucocytosis and raised CRP, and on review of the wound 2 days post-admission showed progression of the ulcerations, raising concern of possible microperforation of the stoma with leakage of contents into the subcutaneous layer and resultant NF (figure 3). Decision was made to perform emergent surgical debridement. The stoma would be examined intraoperatively to exclude any perforations into the subcutaneous space.

Intraoperatively, the stoma was healthy with no defects, thus was left in situ and not refashioned. The subcutaneous layer of the affected ulcerated area was also healthy with no evidence of NF (figure 4). The ulcerated skin was debrided and tissue was sent for histopathology and culture.

Histology showed florid acute suppurative inflammation with abscess formation. Sections of skin show hyperkeratosis, parakeratosis and acanthosis. There were ulceration and a florid dermal acute neutrophilic infiltrate with abscess formation which extended into subcutaneous tissue. Areas of fat necrosis were also noted. Periodic acid-Schiff and Grocott methenamine-silver stains for fungi were negative. No granulomata, dysplasia or malignancy was seen. Tissue culture grew Escherichia coli and Candida glabrata.

The patient was started on culture-directed antibiotic therapy, but the wound worsened (figure 5), with uptrending inflammatory markers—leucocytosis 86.15×10⁹/L and CRP of 333 mg/L. A dermatology consult was made and the patient was diagnosed to have PPG after the review of pre-debridement photographs of the ulcerations (figure 1) coupled with the progression of ulcerations at the edges of the debridement corresponding to the clinical manifestation of pathergy (figure 5). In addition to the primary wound findings, the patient also developed PG with a bullous plaque with purpuric edges and erythematous borders at another right iliac fossa port site uninvolved at the point of initial presentation. In view of the positive wound cultures and suppurative inflammation noted on histology, the dermatologists concluded that there was likely superimposed infection of the PPG contributed by peristomal leakage of contents prior to presentation.

Treatment

She was subsequently managed with a multidisciplinary team involving the colorectal surgeons, dermatology, infectious disease and our stoma and wound care specialist. She was continued on intravenous antibiotics to cover for superimposed infection and started on systemic oral corticosteroids (1 mg/kg/day—50 mg prednisolone once a day) with clobetasol (topical steroid) ointment to be applied to the areas of PG.

Post-debridement, the abdominal wall defect spanned 7.8×19.5×1.0 cm. Daily dressing change caused great discomfort to the patient despite adequate analgesia, the wound produced high exudates which precluded the application of the stoma appliance. Decision was made, after consult with dermatology, to initiate Kinetic Concepts, Inc (KCI) vacuum-assisted closure (VAC) therapy with gradual uptitration of the negative pressures with concomitant steroid administration to reduce the pathergy response. Clobetasol ointment was applied onto the wound bed as a non-adherent layer, followed by mepilex transfer to reduce irritation, before VAC granulofoam (white foam before grey foam) was applied to the wound bed and pressures kept at 50 mm Hg for the first 2 weeks with subsequent increase to 125 mm Hg. After the seal was achieved with the aid of mepilex foam, the stoma pouch was applied over the VAC dressing. A 20Fr flatus tube was inserted into the stoma to encourage drainage of stoma contents into the pouch and reduce peristomal leakage (figure 6). VAC dressing was changed at an interval of 3 days with adequate analgesia prescribed by our pain team colleagues, in the presence of the dermatology and colorectal teams.

Outcome and follow-up

The wound showed marked improvement with the above treatment (figure 7). Her inflammatory markers downtrended to leucocytes of 12.4×10⁹/L and CRP of 2.7 µg/L. She was subsequently discharged home with biweekly follow-up with our stoma and wound care specialist, and outpatient dermatology follow-up. VAC therapy was weaned off after 1 month from debridement, and the abdominal wound was nursed with Dermacyn gauze, clobetasol ointment and mepilex foam, after which the stoma wafer and pouch was applied. After 3 months of dedicated wound care, her wound has granulated and begun epithelialisation (figure 8).