Management of pyoderma gangrenosum

Abstract

Pyoderma gangrenosum (PG) is an uncommon ulcerative skin disease often associated with underlying systemic diseases. Its pathogenesis is unknown, although immune pathways have been implicated. Targeted therapy is therefore lacking and currently treatment is largely empirical and consists of corticosteroids and ciclosporin first line. This paper reviews the current and emerging knowledge about PG. PG occurs with an incidence of 3–10 per million per year. The ulcers are exquisitely painful and characteristically have a necrotic centre with irregular overhanging bluish borders. Around half of cases are associated with underlying systemic disease, most commonly inflammatory bowel disease, rheumatoid arthritis and haematological malignancies; the remaining cases are idiopathic. The pathogenesis is unknown, but the most widely supported theory is immunological, and biopsies of lesions show a predominantly neutrophilic infiltrate. Several aberrant immune processes have been described, with neutrophils and their recruitment to sites of inflammation by cytokines taking an apparently important role. Topical and systemic therapies are both vital aspects of treatment, and in recent years, immune modulators have been used with increasing success, with an emerging role for anti-tumour necrosis factor alpha agents such as the monoclonal antibody infliximab. Although uncommon, PG causes significant morbidity to those it affects. Further research is needed into the disease pathogenesis, and adequate targeted treatment.

Introduction

Pyoderma gangrenosum (PG) is an uncommon ulcerative skin disease, first described over 80 years ago. Ulcers can occur anywhere on the body, most commonly pre-tibially, and often leave unpleasant cribriform scars. There are no specific serological or histological markers, and diagnosis is predominantly clinical. The aetiopathogenesis of PG is poorly understood, and various topical and systemic therapies are used empirically, but no targeted or specific treatment exists. This paper reviews what is known about this elusive disease and its treatment, new developments in our understanding and what remains to be explored.

Methods

PubMed, Embase and the Cochrane Library were searched for journal articles containing the term ‘pyoderma gangrenosum’ within the title or abstract. After identifying relevant articles, all abstracts were read and eligible articles were retrieved. The ‘related articles’ function in PubMed was also used to identify articles not found in the original search. Only articles published in English were included. All types of clinical and non-clinical papers were included, including letters.

Definition and history

PG is a rare neutrophilic dermatosis, first described by Broque in 1916¹ as phagedenisme geometrique. In 1930, Brunsting et al.² named it pyoderma gangrenosum, believing it to be disseminated streptococcal infection causing cutaneous gangrene. Although often mimicking bacterial infection, PG is non-infectious, but lesions may become colonised by skin organisms.

Clinical features

PG is characterised by recurrent painful cutaneous ulcers, typically occurring on the pre-tibial region (Figure 1), but may be found anywhere on the body, having been described on the trunk, head and neck, upper limb and genitalia. Ulcers may arise on different anatomical sites simultaneously. PG can occur in a peristomal distribution (Figure 2), and may account for up to 50% of chronic parastomal ulcers.

Figure 1.

Pre-tibial ulcerative pyoderma gangrenosum, showing typical features.

Figure 2.

Peri-stomal pyoderma gangrenosum.

Lesions begin as small follicular pustules or papules, which expand, becoming deep painful ulcers with well-defined, raised, over-hanging, purple-bluish borders (Figure 1). Erythema and induration surround the ulcer, separating it from healthy skin. The ulcer base is necrotic, with involvement of skin, subcutaneous tissue and muscle layers, and filled with granulation tissue, blood and purulent exudate. Re-epithelialisation begins from the wound edge, healing to leave cribriform scars, which can be disfiguring, particularly if diagnosis is delayed.

Patients with PG are often systemically unwell, with arthralgia, myalgia, fever and malaise. Other extracutaneous manifestations have been reported, including mucosal involvement of the oropharynx, upper airway, eye and genitalia; sterile pulmonary, spleen, liver and bone infiltrates; and neutrophilic myositis.

Atypical cases of PG occur, resembling erythema nodosum, purpura fulminans, vasculitis, necrotising fasciitis and other neutrophilic dermatoses such as Sweet’s syndrome.

Twenty to 40% of cases of PG occur following minor trauma, over previous scars or burns, or as complications of surgery (most commonly breast and abdominal). This phenomenon is known as pathergy, a pathologic hyper-reactivity to ‘normal’ stimuli. Surgical debridement is therefore contraindicated.

Epidemiology

Accurate epidemiological data on PG are limited. The relative rarity of PG poses difficulty, with estimates based on case reports, series and cohort studies, mostly in patients with inflammatory bowel disease (IBD). The general incidence is estimated at 3–10 per million per year. Langan et al.³ recently published the first large UK population-based cohort study, which estimated incidence at 6 per million per year for a European standardised population. PG may occur at any age, most frequently between the third and fifth decades of life. Some reports suggest equal gender distribution, while others propose higher prevalence in females. Just 4% of cases occur in children,⁴ usually associated with systemic disease or immunosuppression, and primarily involving the buttocks, perianal and genital regions, as well as the head,⁴ which is unusual in adults. PG has also been reported in pregnancy. PG may have a higher associated mortality than previously recognised, Langan et al.³ estimating a threefold increase in mortality compared to the general population, and 74% increase in PG associated with IBD compared to IBD controls. The reasons for this significantly increased mortality are not fully understood.

Disease associations

Approximately 50% of cases of PG are associated with a systemic disease, the rest being idiopathic. The most commonly associated diseases are IBD, rheumatoid arthritis and haematological malignancies. The recent UK cohort study found that these commonly reported disease associations were present in 33% of patients.³ Other, more rarely, associated diseases are wide ranging (Table 1). Poorer prognosis has been suggested in PG ulcers associated with arthritis,⁵ with slower healing and longer treatment duration compared to PG ulcers in general. PAPA syndrome (pyogenic arthritis, PG and acne) is a rare autosomal dominant autoinflammatory syndrome, caused by mutations in the proline-serine-threonine phosphatase-interacting protein 1 (PSTPIP1) gene on chromosome 15q,⁶ with several mutations identified. Recently, a new syndrome was described in two unrelated patients⁷: PASH syndrome consists of PG, acne and suppurative hidradenitis, without the arthritis component of PAPA syndrome.

Table 1.

Diseases associated with pyoderma gangrenosum (PG).

		Comments
Gastrointestinal
Commonly associated	Ulcerative colitis Crohn’s disease Regional enteritis	15–35% of cases of PG PG occurs in <3% of IBD Skin disease activity unrelated to IBD activity
Rarely associated	Chronic active hepatitis Primary biliary cirrhosis
Arthritides
Commonly associated	Seronegative rheumatoid arthritis Seropositive symmetrical polyarthritis	20–33% of cases of PG Most commonly seronegative rheumatoid arthritis Unrelated to arthritis activity – may occur during remission
Rarely associated	Spondyloarthropathies Osteoarthritis Juvenile idiopathic arthritis Psoriatic arthritis PAPA syndrome	PAPA syndrome: see text for description
Haematological
Commonly associated	Leukaemia – acute myeloid leukaemias Hairy cell leaukaemia Monoclonal gammopathy (usually IgA)	Up to 20% of cases of PG Leukaemia most commonly (mostly myeloid) Most often with atypical PG PG may appear before other features of the blood disorder, or its occurrence may indicate malignant transformation
Rarely associated	Myelofibrosis Myeloma Lymphoma Polycythaemia rubra vera Paroxysmal nocturnal Haemoglobinuria
Immunological
Rarely associated	Systemic lupus erythematosus (SLE) Immunodeficiencies: AIDS; complement deficiencies; hypogammaglobulinaemia Sarcoidosis
Miscellaneous
Rarely associated	Acne conglobata Takayasu’s arteritis Solid tumours Chronic obstructive pulmonary disease (COPD)

Variants

Four distinct classes of PG are suggested by Powell et al.⁸: ulcerative, pustular, bullous and vegetative. Classification is useful in recognition, but variants often overlap.

Ulcerative PG, also known as classic PG, is most common, presenting as described previously, and associated most commonly with IBD, arthritis or monoclonal gammopathies.

Pustular PG is more superficial, mostly associated with IBD. Lesions appear as small discrete pustules surrounded by normal skin, and do not progress, remaining in the pustular stage for several months. Occasionally, ulceration may occur, giving a picture of ulcerative PG, or pustular and ulcerative variants may occur simultaneously. Often, appearance of pustules is related to the activity of the bowel disease, and treatment of this may lead to remission of the PG. Pyostomatitis vegetans is a pustular eruption associated with IBD, affecting the oral mucosa. This may be associated with pustular PG in up to half of cases.

Bullous PG is another superficial variant, characterised by painful bullae, occurring more commonly on the upper limbs and face, usually associated with haematological malignancy. Bullae spread concentrically, progressing to superficial erosions and ulcers. Sweet’s syndrome, another neutrophilic dermatosis which may occur in patients with leukaemia, presents with lesions of a similar nature. Bullous PG occurring with leukaemia often has a poorer prognosis than other variants.

Vegetative PG is usually less aggressive than other variants, following a more indolent course with fewer symptoms, and generally responds to milder therapies. Lesions are often solitary and do not show the violaceous undermined border or purulent base of ulcerative PG. Vegetative PG usually occurs in otherwise well individuals, but rare associations have been reported with Behçet’s disease, rheumatoid arthritis, diabetes mellitus and haematological disorders.

Peristomal PG may occur after ileostomy or colostomy in patients with IBD, probably as a result of pathergy evoked either by the initial surgical procedure itself or by subsequent irritation of the skin by stoma apparatus or leakage of faeces. Peristomal PG ulcers are similar to classic PG ulcers but may be distinguished by the presence of bridges of normal epithelium. Adhesion of the stoma appliance to the skin may be prevented, and the painful nature of the ulcers makes treatment difficult.

Histopathology

Histopathology of PG is non-specific, but biopsy may be necessary to exclude other causes of ulceration. The stage of the ulcer, site of sampling and the type of PG affect findings.

Early lesions show deep folliculitis with neutrophil infiltration and pus. The erythematous border exhibits dermal oedema and infiltration of chronic inflammatory cells, predominantly lymphocytes, but also macrophages, plasma cells and mast cells. At the ulcer edge, there is mild-to-moderate lymphocytic infiltration, and associated endothelial swelling and fibrinoid necrosis of dermal blood vessels. The ulcer centre shows inflammation throughout the dermis and subcutaneous tissue, with dense neutrophilic infiltration and accumulation of histiocytes and macrophages, leading to abscesses, matrix disintegration and necrosis.

Later, lymphocytes predominate, with dermal and epidermal infiltration, and more prominent haemorrhage, infarction, necrosis and fibrosis. The ulcer border shows epidermal cell hyperplasia, and further vascular changes occur in the surrounding skin, including endothelial swelling and angioplasia. As lesions regress, macrophages and plasma cells invade the dermis, leading to fibrosis.

Aetiopathogenesis

The precise aetiology of PG is unclear, and it is unknown why some conditions give rise to PG, while others give different skin lesions, such as erythema nodosum, although genetic predispositions have recently been identified, at least in IBD-associated cases.⁹ Cases of PG induced by various drugs have been reported, including iodine contrast media, propylthiouracil, isotretinoin, the antipsychotic sulpiride and the tyrosine kinase inhibitor sunitinib. Pathergy, as previously described, may account for a proportion of cases, suggesting dysregulated inflammatory responses to common stimuli.

The most popular theory is one of an immune aetiology, supported by the association of PG with immunologic disorders such as rheumatoid arthritis and IBD, and its response to treatment with immunosuppressive therapy. Several aberrant immune processes have been described, including abnormal monocyte and neutrophil chemotaxis and phagocytosis, lymphocyte dysfunction, complement defects and circulating cytokines. Protein deposition in the vasculature of affected sites suggests immune complex formation. Cross-reacting antibodies against antigens present in the site of associated disease (such as the bowel) and the skin have also been suggested.

There is increasing evidence for the role of neutrophils in PG. Neutrophil trafficking abnormalities have been observed, which may be due to alterations in expression of cell surface integrins CR3 and CR4 and inappropriate integrin-receptor interactions, along with increased metabolic oscillations, responsible for cell locomotion.¹⁰ Immunohistochemistry of skin specimens from PG ulcers demonstrates high expression of myeloperoxidase, a neutrophil marker, and of interleukin-8 (IL-8), a potent chemoattractant that recruits neutrophils to sites of infection. In human skin xenografts, IL-8 overexpression induced subepidermal neutrophil accumulation and formation of ulcers resembling PG clinically and histologically.¹¹ Furthermore, cultured fibroblasts isolated from a PG ulcer produced high levels of IL-8, while fibroblasts from surrounding unaffected skin produced no detectable IL-8,¹² indicating local production from cells within the lesion. Raised serum concentrations of IL-8 have been reported in patients with active PG,¹² as have raised serum concentrations of cytokines IL-6,¹³ and granulocyte colony-stimulating factor (G-CSF), which stimulates proliferation, differentiation and survival of neutrophils. Not only is G-CSF overexpressed in patients with pre-existing PG, PG has reportedly been induced by prophylactic G-CSF use during chemotherapy.¹⁴ IL-23 is further involved in neutrophil recruitment via T_H17 cells and is also highly expressed in PG lesions. IL-23 is implicated in pathogenesis of Crohn’s disease and may similarly play a role in PG. Treatment of a patient with the anti-IL-23 monoclonal antibody ustekinumab resulted in complete healing of her PG.¹⁵

Neutrophils may also play a role in PG through sustained inappropriate generation of reactive oxygen species (ROS) promoting angiogenesis in ulcers. In samples from PG lesions, hypoxia inducible factor 2-alpha (HIF2a) and vascular endothelial growth factor (VEGF) were constitutively highly expressed.¹⁶ Stimulation of HIF2a by ROS leads to the production of VEGF and angiopoietin-2 (Ang-2), which promote angiogenesis, perpetuating lesions.

Rare familial cases of PG occur as part of PAPA syndrome, as discussed, due to PSTPIP1 gene mutations. This encodes a cytoskeletal adaptor protein which participates in regulation of the inflammatory response by forming part of the inflammasome involved in IL-1β production.¹⁷ Mutations in the gene lead to hyperphosphorylation of the protein, altering its interactions in the inflammasome, and leading to overproduction of IL-1β, which may be another pathway engaged in the pathogenesis of PG.

Diagnosis

PG is a ‘diagnosis of exclusion’, and should be considered in any patient with non-healing ulceration, particularly in the context of associated systemic diseases. There are no specific serologic markers, so diagnosis relies on clinical history and examination. Lesion biopsy should be taken to exclude other cutaneous diseases presenting similarly, importantly malignancy and vasculitic conditions. Undetected associated systemic disease should be sought. Suspected PG should be referred urgently to a physician experienced in the disease. Late or missed diagnosis of PG may lead to extensive ulceration and tissue damage, or to inappropriate, potentially harmful therapy.

Numerous conditions mimic PG, of which there are several groups (Table 2).

Table 2.

Differential diagnosis of pyoderma gangrenosum (PG).

	Examples
Infection	Deep fungal infections – sporotricosis, aspergillosis, cryptococcosis, zygomycosis Bacterial – erysipelas or gangrene; cutaneous tuberculosis, syphilis Viral – herpes simplex Parasitic – amoebiasis cutis
Vascular occlusive and venous disease	Antiphospholipid syndrome Venous stasis ulceration Other less common vasculopathies
Vasculitis	Wegener’s granulomatosis Polyarteritis nodosa Takayasu’s arteritis
Malignancy	Malignant processes involving the skin, in particular leukaemias and lymphomas
Drug reaction	Ulcerating or pustular drug reactions
Exogenous tissue injury	Insect or spider bites Factitious disorder Münchhausen syndrome (dermatitis artefacta)
Other inflammatory conditions	Cutaneous Crohn’s disease SLE Necrotising fasciitis Sweet’s syndrome

Because of difficulties in diagnosis, Su et al.¹⁸ proposed diagnostic criteria, which when used together can strongly support a PG diagnosis. Two major and four minor criteria are proposed, with both major and at least two minor required for diagnosis:

Major criteria:  A painful rapidly progressing cutaneous ulcer;

         Exclusion of other causes of ulceration.

Minor criteria:  Presence of systemic diseases associated with PG;

         History suggestive of pathergy;

         Characteristic histopathological findings;

         Response to systemic steroids or immunosuppression.

Treatment

There is no targeted or specific therapy, and current treatment is largely empirical, depending on severity and extent of ulceration, associated disease and patient factors. The goals of treatment are to reduce lesion inflammation and pain and promote healing. Control of the systemic disease, particularly IBD, often results in control of skin lesions. Treatment therefore combines local wound care and systemic treatment. In ulcerative colitis-associated PG, total colectomy may alleviate the skin disease, although this is often not curative, and risks pathergy following surgery. However, successful surgical procedures with background immunotherapy have occurred without inducing PG. Current treatments are summarised in Table 3.

Table 3.

Current treatment of pyoderma gangrenosum (PG).

Topical	Moist wound management	Regular lavage; foam dressings
	Topical antimicrobials	Prevent wound colonisation and superinfection
	Topical analgesics	Including topical opioids
	Surgical	Split skin grafts; allogeneic cultured skin grafts Debridement contraindicated – potential pathergy
	Corticosteroids	Topical cream or intralesional injection
	Ciclosporin	Particularly successful in peristomal lesions Use with caution to avoid promoting bacterial overgrowth and potential systemic effects
	Steroid-sparing agents	Tacrolimus and pimecrolimus; mesalazine Monotherapy or adjuvants to systemic treatments
	Others	Disodium cromoglycate; benzoyl peroxide; nicotine
Systemic	Corticosteroids	High dose (e.g. prednisolone 1–2 mg/kg/day) to induce remission, tapered only once healing Pulsed intravenous methylprednisolone is an alternative for aggressive or extensive disease
	Ciclosporin	First-line agent as alternative to, or in combination with, steroids High oral doses (3-10 mg/kg/day) dependant on clinical response, adverse effects, and plasma levels Monitor plasma levels, blood pressure and creatinine
	Antimicrobials and anti-inflammatories	Adjuncts to topical/systemic therapies E.g. dapsone and sulfasalazine
	Immunosuppressants	Used for their steroid-sparing benefits Particularly beneficial in patients with steroid-resistant IBD. E.g. azathioprine, methotrexate, cyclophosphamide, chlorambucil, mycophenolate mofetil
	Immune modulators	Adjuncts or alternatives to conventional therapy Including thalidomide, interferon-alpha (IFNα), intravenous immunoglobulin (IVIg), leukocyte apheresis Anti-TNFα monoclonal antibodies: Infliximab, Etanercept, Adalimumab

Topical therapy is important in all cases of PG. Maggots, used to treat other chronic ulcers, have been shown to be ineffective in PG, due to reduced survival in these ulcers.¹⁹ Split-thickness skin grafts and allogeneic-cultured skin equivalents²⁰ have been employed, usually after induction of remission with immunomodulators. Hyperbaric oxygen therapy is reportedly helpful, even in refractory cases, and may permit reduction of medication doses.

Systemic therapy is required in all but localised superficial cases. The mainstay is immunosuppression, systemic corticosteroids and/or ciclosporin being first line. Aggressive disease may require high doses of both.

Corticosteroids are required at high doses, with pulsed therapy at suprapharmacological doses (e.g. 1 g/day for 5 days) as an alternative. Recently, inhaled corticosteroids were used with success in a case of peristomal PG refractory to other treatments.²¹

Ciclosporin A, a calcineurin antagonist causing inhibition of T-lymphocyte activation, is the most widely used steroid alternative. Clinical improvement can be seen within weeks, with complete healing often within three months; a maintenance dose may be required. Ciclosporin requires careful monitoring: side effects include nephrotoxicity, hypertension, central nervous system effects and development of lymphoma. Reversible ciclosporin myopathy is another potential complication and has been noted in PG patients.

Antimicrobial and anti-inflammatory agents are used as adjuncts. Dapsone (diaminodiphenylsulfone) and other sulpha drugs are the most widely used and effective, possessing several anti-inflammatory effects, including inhibition of neutrophil migration and ROS production. Sulfasalazine is particularly beneficial in patients with colitis. Clofazimine, colchicine and minocycline are other examples.

Immunosuppressants, including azathioprine, 6-mercaptopurine, methotrexate, cyclophosphamide and chlorambucil, are used as steroid sparing agents, or as first-line adjuncts. Their therapeutic effect is less predictable; not all patients will respond. Cytotoxicity can lead to potentially severe side effects, requiring careful monitoring. Mycophenolate mofetil causes selective reduction of T- and B-lymphocyte proliferation through the inhibition of purine and pyrimidine synthesis. It is effective in PG and can lead to complete remission when given alone in some cases.

Immune modulators are relatively new in treatment of PG but have shown great success. Thalidomide has several immunomodulatory effects, including suppression of tumour necrosis factor alpha (TNFα) and neutrophil chemotaxis. It has been used to treat PG both alone and with corticosteroids.²² Its teratogenicity necessitates caution in women of child-bearing age. Interferon-alpha (IFNα) is also reportedly successful. Several cases have been treated with intravenous immunoglobulin (IVIg) with complete healing of lesions,²³ and leukocyte apheresis has also been used in PG refractory to traditional therapies.²⁴

Development of antibody technology has been a great advance in the treatment of PG. Infliximab, an anti-TNF chimeric monoclonal antibody used in many conditions including IBD and rheumatoid arthritis, is becoming increasingly popular in the treatment of PG associated with these conditions. Doses of 5 mg/kg are effective, given by intravenous infusion at intervals, although single infusions have been shown to be effective. Etanercept, a TNF receptor-IgG fusion protein, and adalimumab, a fully humanised monoclonal anti-TNF antibody, are also in use. Their efficacy may be lost due to the development of anti-drug antibodies by the patient; antibody-mediated infusion reactions also occur relatively frequently. Concurrent administration of another immunosuppressant may therefore be wise. TNFα inhibition may worsen bacterial infections while masking the signs, and septic shock can ensue, as exemplified by the recent death of a patient with PG receiving infliximab.²⁵

Further treatments include oral isotretinoin and topical phenytoin sodium (which is also used in other ulcerative conditions). Recombinant human growth factors such as EGF have shown some success when applied topically.²⁶ In another report, treatment of myelodysplastic syndrome with bone marrow transplantation resulted in remission of the associated PG, although this was potentially due to immunosuppressants used concurrently to prevent graft versus host disease.²⁷

Future directions

Limited knowledge of the pathogenesis of PG currently means there is a lack of specific or universally successful therapies. Drugs used in treatment of PG are highly potent and have significant and potentially life-threatening side effects, and are currently used based on limited evidence from case reports and small series, with a lack of randomised controlled trials. Because the disease is rare, multicentre trials are necessary to tackle this problem, and PG was recently identified by the UK Dermatology Clinical trials network as a top priority for independent research. The network’s multicentre trial of prednisolone versus ciclosporin (the STOP GAP study), with a primary objective of velocity of healing at six weeks, and multiple secondary objectives looking at overall treatment response, safety and cost-effectiveness of the two treatments, was recently completed and results are due to be published soon.²⁸

Emerging techniques such as gene therapy may be applicable to PG, particularly given the readily accessible nature of the skin as a target. However, candidate genes are yet to be elucidated, and further clarification of the disease pathogenesis will be of benefit; possible targets may include genes involved in the immune processes and recruitment of inflammatory cells, or those involved in wound healing, such as fibroblast, keratinocyte and epidermal growth factors. Progress with novel therapies should not halt investigation into the pathogenesis of PG or evaluation of currently available treatments and the search for more targeted therapies.

Declarations

Competing interests

None declared

Funding

No funding received for the work on this manuscript. RH has received grants from Royal College of Surgeons of England, and Bowel Disease Research Foundation for preclinical laboratory work which may be applicable to the treatment of pyoderma gangrenosum in future.

Ethical approval

Not required

Guarantor

AT and RH

Contributorship

AT contributed to the conception and design of the article, acquired and interpreted the data, drafted the article and revised it critically, approved the final version for publication, and agrees to be accountable for all aspects of the work; RH contributed to the conception and design of the article, interpreted the data, revised the article critically for intellectual content, approved the final version for publication, and agrees to be accountable for all aspects of the work.

Acknowledgements

None

Provenance

Not commissioned; peer-reviewed by Ray Delicata, Wael Al-Daraji and Julia Schofield