Pyoderma Gangrenosum

1. Disease Information

1.1 Overview

PG is a primarily sterile inflammatory neutrophilic dermatosis characterized by recurrent, rapidly progressive, exquisitely painful cutaneous ulceration with undermined violaceous borders and a mucopurulent or hemorrhagic exudate. Despite the name, it involves neither infection nor gangrene — a historical misnomer dating to the era when a streptococcal etiology was assumed.

The authoritative modern overview is the Nature Reviews Disease Primers article (PMID:33033263, Maverakis et al., 2020, DOI 10.1038/s41572-020-0213-x), whose abstract states [verbatim]:

"Pyoderma gangrenosum (PG) is a rare neutrophilic dermatosis that presents with rapidly developing, painful skin ulcers hallmarked by undermined borders and peripheral erythema. Epidemiological studies indicate that the average age of PG onset is in the mid-40s, with an incidence of a few cases per million person-years. PG is often associated with a variety of other immune-mediated diseases, most commonly inflammatory bowel disease and rheumatoid arthritis. The cause of PG is not well understood, but PG is generally considered an autoinflammatory disorder."

1.2 Key identifiers

Source: OLS4 / MONDO term record for MONDO_0018824 (xref table retrieved from the EBI OLS4 API).

1.3 MONDO subtype children (relevant to has_subtypes curation)

Related syndromic entities (candidate Grouping members rather than subtypes of PG proper):

The four-variant classification is anchored in PMID:8609250 (Powell FC, Su WP, Perry HO, J Am Acad Dermatol 1996) [verbatim]: "Pyoderma gangrenosum (PG) has four distinctive clinical and histologic variants… PG often occurs in association with a systemic disease, and the specific clinical features of the skin lesion may provide a clue to the associated disease."

1.4 Synonyms

Pyoderma gangraenosum; PG; "phagedenic pyoderma" (historical); "dermatitis ulcerosa" (historical); peristomal PG (PPG), postsurgical PG (PSPG), malignant pyoderma and pyostomatitis vegetans are related-but-distinct named presentations.

1.5 Data provenance

Both individual-patient and aggregate sources exist. Individual/EHR-derived: the UK General Practice Research Database cohort (PMID:22534879), the US National Inpatient Sample analyses (PMID:29334018, PMID:29438762), and the Israeli Clalit Health Services population-based case-control series (the Kridin cohort, n=302 PG cases). Aggregated/disease-level: Orphanet ORPHA:48104, MONDO, HPO, and the systematic reviews and meta-analyses cited throughout.

2. Etiology

2.1 Disease causal factors

PG has no single cause. It is best modeled as a three-input system: (i) genetic susceptibility, (ii) an associated systemic immune-mediated or hematologic disease, and (iii) a proximate trigger. The 2025 pathogenesis review (PMID:39718519, Becker SL, Vague M, Ortega-Loayza AG, J Invest Dermatol, DOI 10.1016/j.jid.2024.09.023) states [verbatim]:

"Pyoderma gangrenosum (PG) is a neutrophilic dermatosis of unclear etiology. Numerous theories of its underlying pathogenesis have been proposed, including external triggers, neutrophilic dysfunction, complement activation, and autoimmunity, as well as a possible component of underlying genetic susceptibility."

The 2022 treatment review (PMID:35606650, Maronese CA, Pimentel MA, Li MM, Genovese G, Ortega-Loayza AG, Marzano AV, Am J Clin Dermatol, DOI 10.1007/s40257-022-00699-8) frames the mechanism sharply [verbatim]:

"Pathogenesis involves dysregulation of innate and adaptive immunity in genetically predisposed individuals, with the follicular unit as a putative initial target. T helper 17/1-skewed inflammation and exaggerated inflammasome activation produce dysregulated neutrophil-dominant milieu with elevated tumor necrosis factor-α, IL-1β, IL-1α, IL-8, IL-12, IL-15, IL-17, IL-23, and IL-36."

The follicular unit as the putative initial target is a curatable mechanistic claim worth its own pathophysiology node — it explains the pustular prodrome (a papule/pustule/vesicle ulcerating within four days, one of the Delphi minor criteria) and links PG mechanistically to hidradenitis suppurativa and acne in the PASH/PAPASH spectrum.

2.2 Genetic risk factors

Monogenic (syndromic) forms — the clearest mechanistic window.

PSTPIP1 (also called CD2BP1; HGNC:9580) is the canonical PG-associated gene. PMID:11971877 (Wise CA et al., Hum Mol Genet 2002, DOI 10.1093/hmg/11.8.961) established that [verbatim] "PAPA syndrome (pyogenic sterile arthritis, pyoderma gangrenosum, and acne, OMIM #604416)…are rare inherited disorders of early onset, primarily affecting skin and joint tissues," identifying disease-causing CD2BP1 mutations and proposing classification as autoinflammatory. PMID:14595024 (Shoham NG, …, Kastner DL, PNAS 2003, DOI 10.1073/pnas.2135380100) supplied the mechanism: PSTPIP1/CD2BP1 binds pyrin (MEFV), and PAPA-associated mutations A230T and E250Q markedly increase pyrin binding, are hyperphosphorylated when coexpressed with c-Abl kinase, and are associated with "increased IL-1beta production by peripheral blood leukocytes from a clinically active PAPA patient." This defines FMF and PAPA as disorders in the same pathway — the pyrin inflammasome.

The 2024 systematic review of inborn errors of immunity in PG (PMID:38951460, Oprea Y, Antohi DR, Vague M, Delbourgo Patton C, Wu B, Ortega-Loayza AG, Am J Clin Dermatol, DOI 10.1007/s40257-024-00875-y) states [verbatim]: "Genetic mutations such as BTK, IL1RN, ITGB2, LPIN2, MEFV, NFkB1, NLRP3… were identified in the presence of either idiopathic or syndromic PG." It identified 74 cases of PG occurring with an inborn error of immunity [paraphrase].

The genetics systematic review (PMID:25350484, DeFilippis EM, Feldman SR, Huang WW, Br J Dermatol 2015, DOI 10.1111/bjd.13493) analyzed 823 cases and reported [paraphrase] "65.2% cases were associated with inflammatory bowel disease, 16.1% with polyarthritis and 12.5% with haematological disorders," plus mutations in MTHFR and JAK2.

Polygenic / complex susceptibility. No published genome-wide association study of idiopathic PG at genome-wide significance is available as of this report — a genuine knowledge gap. The closest evidence is indirect: PMID:24487271 (Weizman A et al., Inflamm Bowel Dis 2014) reported IBD-cohort associations with IL8RA (CXCR1), PRDM1, USP15, TIMP3 for PG and erythema nodosum [paraphrase]; and PMID:42123319 (Yao H, Wu Y, Zhang R, Int J Mol Sci 2026, DOI 10.3390/ijms27093733) reports that [verbatim] "Genetic analysis confirmed IBD as a causal risk factor for PG, precisely identifying six shared genetic loci" and identified "a cross-tissue conserved inflammatory module centered on the JAK-STAT pathway, with JAK2 and STAT3 identified as network hubs."

Marzano's PASH study is the strongest evidence that PG-spectrum disease carries autoinflammatory-gene burden even without a single Mendelian lesion (PMID:25501066, Medicine 2014) [verbatim]: "Four out of our 5 PASH patients presented genetic alterations typical of well-known AIDs, including inflammatory bowel diseases, and the only patient lacking genetic changes had clinically evident Crohn disease."

2.3 Environmental and acquired risk factors

Associated systemic disease is the dominant risk determinant. The meta-analysis of 21 studies / 2,611 patients (PMID:29721816, Kridin K, Cohen AD, Amber KT, Am J Clin Dermatol 2018, DOI 10.1007/s40257-018-0356-7) reports [verbatim]:

"the overall random-effects pooled prevalence of associated systemic diseases was 56.8% (95% confidence interval 45.5–67.4)"

with IBD 17.6%, arthritis 12.8%, hematological malignancies 8.9%, solid malignancies 7.4%, and — critically for the mechanism — [verbatim] pathergy accounting for disease onset in "16.3% (95% confidence interval 7.7–27.1) of cases."

Quantified population-based effect sizes from the Israeli Clalit cohort (302 PG cases vs. matched controls), all by Kridin and colleagues:

The negative solid-malignancy result (PMID:34076886) is an important curated refutation: the older literature's 7.4% solid-malignancy prevalence figure reflects background prevalence, not excess risk. Curate it with supports: REFUTE against any claim of solid-tumor causation.

Lifestyle and metabolic. Nicotine dependence increases risk of PG among 23 of 38 chronic inflammatory diseases studied in 881,192 EHRs, overall [verbatim] "hazard ratio 2.12, confidence interval 2.10–2.14, p < 0.0001" (PMID:40012715, Kridin K, Papara C, Bieber K, et al., Front Psychiatry 2025). Overweight/obesity is a risk factor for chronic inflammatory disease broadly (PMID:39963282, HR 1.52, 95% CI 1.509–1.521, 3.1 million individuals) and high BMI is an independent risk factor for peristomal PG specifically (OR 9.895, 95% CI 1.970–43.704, p=0.005; PMID:22959399).

Pathergy / mechanical trauma is the single most curatable environmental trigger. Post-surgical PG (PMID:25589459, Zuo KJ, Fung E, Tredget EE, Lin AN, JPRAS 2015) analyzed 220 cases [verbatim]: "PSPG occurred most commonly after breast (25%), cardiothoracic (14%), abdominal (14%), and obstetric (13%) surgeries… Signs of wound complication occurred on average 7.0 days after surgery." Prior PG history was present in 16.8%, hematologic disorder 8.6%, IBD 5.9%, RA 3.6%.

Ostomy formation is a distinct mechanical/chemical trigger for peristomal PG (PPG). The Mayo series of 44 patients (PMID:27473454, Barbosa NS et al., J Am Acad Dermatol 2016) reports [verbatim]: "A total of 44 patients had PPG (mean age, 46 years; 32 women [73%]); 41 (93%) had inflammatory bowel disease. Mean time to PPG onset after stoma surgery was 5.2 months."

Drugs. A 2026 FAERS disproportionality analysis (PMID:42310248, Woods RH, Clin Rheumatol, DOI 10.1007/s10067-026-08237-1) found 1,316 PG reports of 13.3M total, 868 (66%) linked to antirheumatic biologics, with [verbatim] "All four interleukin (IL)-17 inhibitors exhibited disproportionate pyoderma gangrenosum reporting" — brodalumab PRR 23.02 (95% CI 8.64–61.36), bimekizumab PRR 9.10 (4.08–20.29). This is a paradoxical drug reaction: IL-17 blockade is simultaneously a candidate PG treatment (secukinumab/ixekizumab trials) and a reported PG trigger. Curate the paradox explicitly; do not resolve it silently. Broader context in PMID:30971924 (Garcovich S, …, Marzano AV, Front Pharmacol 2019), which lists PG among paradoxical skin reactions to biologics.

Other reported triggers from the wider literature: G-CSF, isotretinoin, propylthiouracil, cocaine adulterated with levamisole, and immune checkpoint inhibitors (see PMID:32382051 for the irAE framework — but note PG-specific checkpoint-inhibitor evidence is case-level).

2.4 Protective factors

No validated genetic or environmental protective factor for PG has been identified. This is a real absence, not a search failure — no gnomAD protective allele, no dietary or lifestyle protective exposure, and no vaccine has been shown to reduce PG risk. The only demonstrated prophylactic intervention is pharmacological and tertiary: perioperative corticosteroid cover in at-risk patients undergoing breast surgery (PMID:25589459 [verbatim]: "Nineteen patients (8.6%) at risk for PSPG received perioperative corticosteroids during skin grafting or later surgeries with a favorable outcome").

2.5 Gene–environment interaction

The mechanistically explicit model is: a genetically primed inflammasome (PSTPIP1–pyrin axis, or polygenic inflammasome-gene burden) sets a lowered threshold for sterile neutrophilic inflammation; minor trauma that would resolve normally instead triggers a self-amplifying IL-1 → IL-8 → neutrophil → NET → IL-1 loop. Pathergy is the gene–environment interaction, observable at the bedside. Marzano's neutrophilic-disease review (PMID:28688013, Clin Rev Allergy Immunol 2018) argues these should be regarded as polygenic autoinflammatory conditions [paraphrase]: "Gene mutations involved in autoinflammatory diseases likely contribute to neutrophilic disease pathogenesis, warranting their consideration as polygenic autoinflammatory conditions."

3. Phenotypes

3.1 Core phenotype table with suggested HPO terms

Curation caution on frequency (§7 of CLAUDE.md). Only three frequency values above have quantitative support in an abstract: pathergy 16.3%, pruritus 69%, and the systemic-disease pooled prevalence 56.8%. The frequency: slot should be omitted for the rest rather than assigned by inference.

3.2 Phenotype characteristics

Age of onset. Mid-40s on average (PMID:33033263); UK cohort median 59 years, IQR 41–72 (PMID:22534879); US inpatient mean 56 years (PMID:29334018); Australian inpatient mean 62.8 years, range 30–89 (PMID:25374597). StatPearls records onset range 11–89 years with <5% of cases in children [paraphrase]. Pediatric PG occurs and is disproportionately associated with IBD and with immunodeficiency (PMID:9875964, PMID:2370611).

Suggested onset annotation: onset_category: ADULT_ONSET at disease level, with a has_subtypes/notes acknowledgment of pediatric cases.

Severity. Highly variable — from a single small leg ulcer manageable with topical therapy (43.8% healed by 6 months with topical clobetasol alone; PMID:27502313) to fulminant multifocal disease with in-hospital death (3.2% of 2,273 US inpatient admissions; PMID:29334018; and 5/23 deaths in one Australian series, PMID:25374597).

Progression. Classically acute onset, rapidly progressive expansion over days, then a chronic phase with slow healing over months. Median time to healing on topical therapy was 145 days (95% CI 96 days to ∞) (PMID:27502313). Peristomal PG mean time to complete response was 10.7 weeks (PMID:27473454). The disease course is relapsing–remitting: recurrence after any treatment in 23 of 38 (61%) peristomal cases (PMID:27473454), and 28–30% recurrence at 6 months in the STOP GAP randomized trial (PMID:26071094).

Anatomic distribution. Lower legs predominate (a Delphi minor criterion is "multiple ulcerations, at least 1 on an anterior lower leg"). In one inpatient series (PMID:25374597) [verbatim]: "Lesions were localised to lower limb in 13 patients, peristomal region in four, breast in three, upper limb in one, and two patients had PG at multiple sites." Lesions are typically asymmetric and may be multifocal; bilateral involvement occurs but is not the rule.

3.3 Quality-of-life impact

Pain is the dominant QoL driver and was a prespecified secondary outcome in STOP GAP (PMID:26071094). The best recent per-phenotype QoL data concern pruritus (PMID:42472079, Becker SL, Zhang R, Latour E, Downey K, Roland-McGowan J, Gillespie J, Ortega-Loayza AG, JID Innovations 2026, DOI 10.1016/j.xjidi.2026.100500) [verbatim]:

"We analyzed data from 136 patients with 178 ulcers. At baseline, 69% of the patients reported moderate pruritus with a mean severity of 3.3 (0–10 scale, 95% confidence interval = 2.9–3.8), which decreased with healing (from 3.7 to 2.6). Quality of life scores improved in parallel with healing. Higher pruritus severity was associated with younger age and inflammatory arthritis."

Opioid burden is a secondary QoL harm; a small prospective case series of topical cannabis in three PG patients reported [verbatim] "Clinically significant analgesia that was associated with reduced opioid utilization was noted in all three cases" (PMID:28818631) — low-quality evidence, curate as IN_VITRO/OTHER-tier at best, or omit.

Hospitalization burden is severe: mean length of stay 47 days (range 5–243) in one inpatient series (PMID:25374597).

4. Genetic / Molecular Information

4.1 Causal genes

There is no causal gene for idiopathic (non-syndromic) PG. This must be stated explicitly in the entry — PG is a Complex disease, and asserting a causal gene would be wrong. The Mendelian genetics belong to the syndromic entities:

PSTPIP1 (CD2BP1), HGNC:9580, OMIM *606347; PAPA syndrome OMIM #604416. Autosomal dominant. Encodes proline-serine-threonine phosphatase-interacting protein 1, an F-BAR adaptor that binds PTP-PEST and pyrin.

• Variant class: missense. Canonical: p.Ala230Thr (A230T) and p.Glu250Gln (E250Q) (PMID:11971877, PMID:14595024); also p.Glu250Lys (E250K) (PMID:25845478), p.Gly403Arg (G403R) (PMID:25683018), G258A and aberrant splicing variants (PMID:21790734).
• Functional consequence: These are best described as gain-of-function with respect to pyrin binding exerting a dominant-negative effect on pyrin's inhibitory regulation of the inflammasome — the net result is IL-1β overproduction. Per PMID:14595024 [paraphrase]: PAPA-associated mutations "markedly increased pyrin binding and were hyperphosphorylated when coexpressed with c-Abl kinase," with "increased IL-1beta production by peripheral blood leukocytes from a clinically active PAPA patient."
• Curation note (dismech GAIN_OF_FUNCTION decision tree): use GeneticContext.functional_impact_category: GAIN_OF_FUNCTION (or DOMINANT_NEGATIVE, arguably more accurate w.r.t. pyrin) for the variant, and separately modifier: GAIN_OF_FUNCTION on the inflammasome/IL-1β biological_processes node for the pathway state. These are two different claims.
• Origin: germline, autosomal dominant.
• Allele frequency: PAPA variants are private/ultra-rare; not meaningfully represented in gnomAD.
• Note the reported PSTPIP1-negative PAPA phenotype (PMID:19700023) — locus heterogeneity is real.

NCSTN, HGNC:17836. First nicastrin mutation in PASH reported by PMID:25601011 (Duchatelet S, …, Hovnanian A, Br J Dermatol 2015) — LOF in the γ-secretase complex, the same mechanism as familial HS.

JAK2 V617F. Somatic, in the context of PG arising with a myeloproliferative neoplasm (PMID:25350484). This is the only well-supported somatic variant in the PG literature and should be curated with variant_origin: SOMATIC.

4.2 Modifier genes

Not established. Candidate loci from the IBD-cohort study (PMID:24487271): IL8RA/CXCR1, PRDM1, USP15, TIMP3 — these are association signals in IBD patients with cutaneous EIMs, not validated modifiers, and should be curated with relationship_type: SUSCEPTIBILITY at most, with an explicit KNOWLEDGE_GAP discussion.

4.3 Epigenetics

No PG-specific DNA-methylation, histone-modification, or chromatin study has been published. Searches of ENCODE/Roadmap-indexed literature return nothing PG-specific. This is a documented gap and should be recorded as a discussions entry with kind: KNOWLEDGE_GAP.

4.4 Chromosomal abnormalities

Trisomy 8 is the one recurrent cytogenetic association, arising through the MDS route: PG with myelodysplastic syndrome and trisomy 8 (PMID:28943508, Fujiwara D et al., Eur J Dermatol 2017). Trisomy 8 MDS is independently linked to Behçet-like and neutrophilic inflammation. This should be curated as a comorbid hematologic driver, not as a germline chromosomal abnormality of PG.

5. Environmental Information

5.1 Environmental factors and ECTO grounding

The environmental exposures that matter in PG are mechanical and iatrogenic, not toxicological. Suggested influences_mechanisms links:

ECTO binding caution. Per the dismech environmental-term audit guidance, ECTO has good coverage for chemical exposures (ECTO: tobacco-smoke terms exist) but poor coverage for surgical trauma and ostomy effluent. Search ECTO before binding; if no term fits, leave exposure_term unbound with a notes: line recording that ECTO was searched — that is a correct outcome, not a gap.

5.2 Lifestyle factors

Nicotine dependence (HR 2.12 for chronic inflammatory disease generally, PG among the 23 diseases with elevated risk; PMID:40012715) and obesity (PMID:39963282, PMID:22959399) are the two with population-scale support. No dietary factor is established.

5.3 Infectious agents

None. This is a defining negative. PG is sterile: "exclusion of infection" is a Delphi minor criterion (PMID:29450466), and wound cultures are characteristically negative. The critical clinical corollary is that PG is misdiagnosed as infection: PMID:12409543 (Weenig RH, Davis MD, Dahl PR, Su WP, N Engl J Med 2002, DOI 10.1056/nejmoa013383) found that 10% of consecutive patients treated for PG had an alternative diagnosis — including infection, vasculitis, malignancy, and vascular occlusive disease [paraphrase]. And in the other direction, PG patients presenting to infectious-disease clinics receive inappropriate antibiotics and delayed immunosuppression (PMID:42517131).

Curate this as an evidence item with supports: REFUTE against any infectious-etiology claim, and reference NCBITaxon nowhere.

6. Mechanism / Pathophysiology

6.1 The causal chain (suggested pathophysiology node graph)

[Genetic susceptibility]                    [Trigger: trauma / associated systemic disease / drug]
   PSTPIP1-pyrin axis, inflammasome-gene burden          pathergy, IBD flare, MDS clone
    \                                   /
     v                                 v
(1) Inflammasome Dysregulation and IL-1 Overproduction   [MOLECULAR]
                  |
                  v
(2) Type-1 / Type-17 Cytokine Skewing (TNF-α, IL-17, IL-23, IL-36, IL-12, IL-15)  [MOLECULAR]
                  |
                  v
(3) Complement C5a Generation and C5aR1 Signaling        [MOLECULAR]
                  |
                  v
(4) Chemokine-Driven Neutrophil Recruitment (IL-8/CXCL8, CXCL1/2/3, CXCL16, RANTES)  [CELLULAR]
                  |
        +---------+---------+
        v                   v
(5) GSDMD-Dependent NETosis     (6) T-cell Infiltration at Wound Margin  [CELLULAR]
        |                   |
        +---------+---------+
                  v
(7) MMP-2/MMP-9-Mediated Extracellular Matrix Destruction  [TISSUE]
                  |
                  v
(8) Sterile Neutrophilic Dermal Abscess and Ulceration     [TISSUE]
                  |
                  v
(9) Painful Non-Healing Ulcer with Undermined Border       [ORGANISM]

Nodes 1→5 constitute a feed-forward amplification loop: NETs release IL-1α/IL-1β and DNA-associated DAMPs that re-trigger the inflammasome, which is the mechanistic basis of pathergy. Node 3 is the rate-limiting node for the C5a-directed therapies.

6.2 Molecular pathways

IL-1 / inflammasome axis — the core. IL-1β and its receptors are significantly overexpressed in PG lesional skin. PMID:24903614 (Marzano AV, Fanoni D, Antiga E, Quaglino P, Caproni M, Crosti C, Meroni PL, Cugno M, Clin Exp Immunol 2014, DOI 10.1111/cei.12394) is the flagship comparative study (16 PG, 6 Sweet, 6 controls) [paraphrase]: "IL-1β and its receptor I were significantly elevated in both PG (P=0.0001) and SS (P=0.004–0.040). In PG, chemokines including IL-8 (P=0.0001), CXCL1/2/3 (P=0.002), CXCL16 (P=0.003), and RANTES (P=0.005) were overexpressed… Fas/Fas ligand and CD40/CD40 ligand systems were overexpressed in PG (P=0.0001–0.012)."

GO terms: GO:0050701 interleukin-1 secretion; GO:0072559 NLRP3 inflammasome complex (CC); GO:0141201 positive regulation of NLRP3 inflammasome complex assembly (verify current label); GO:0006954 inflammatory response.

Pyrin pathway. PSTPIP1–pyrin binding (PMID:14595024) links PG to the FMF axis. GO: GO:0005515 protein binding (too generic — prefer GO:0140632 inflammasome complex assembly, verify).

Type-17 / IL-23 axis. IL-17 and IL-23 are elevated in lesional skin (PMID:20636397, PMID:35606650). GO: GO:0072538 interleukin-17-mediated signaling pathway; GO:0038155 interleukin-23-mediated signaling pathway.

IL-36 axis. Named among the elevated mediators in PMID:35606650 and the rationale for spesolimab (anti-IL-36R). See PMID:38779986 (Sugiura K et al., JEADV 2024) for the IL-36 pathway argument.

Complement C5a. PMID:37516310 (Wang Z, Hornick N, Vague M, Yang D, Keller J, Kody S, Leachman S, Ortega-Loayza AG, Liu Y, J Invest Dermatol 2024, DOI 10.1016/j.jid.2023.06.204), "NETosis Is Induced by Complement Component 5a: Implications in the Pathogenesis of Pyoderma Gangrenosum," supplies the mechanistic bridge between complement activation and neutrophil dysfunction and is the direct scientific rationale for vilobelimab. GO: GO:0006956 complement activation; GO:0038178 complement component C5a signaling pathway.

JAK-STAT. PMID:42603447 (Liu W, Peng L, Wang R, Fan J, Chen L, Shen Z, Mol Immunol 2026, DOI 10.1016/j.molimm.2026.08.009) used single-cell RNA-seq and multiplex IHC to show JAK/STAT overactivation in PG lesions, with [verbatim]: "In vitro cell experiments further demonstrated that the JAK inhibitor tofacitinib suppresses STAT phosphorylation in myeloid and T cells, myeloid NETosis, and IL-17A production." GO: GO:0007259 cell surface receptor signaling pathway via JAK-STAT. PMID:42123319 independently nominates JAK2 and STAT3 as network hubs shared between PG and IBD.

6.3 Cellular processes

NETosis is the central effector cell-death program. The landmark mechanistic paper is PMID:40034857 (Li S, Ying S, Fang H, Qiao J, iScience 2025, DOI 10.1016/j.isci.2025.111925), "Gasdermin D-dependent neutrophil extracellular traps exacerbate cytokine storm contributing to pyoderma gangrenosum pathogenesis" [verbatim]:

"In this study, we discovered that the serum levels of NETs were elevated in PG patients compared to healthy controls. Injection of serum from PG patients into the dorsal skin of wild-type mice led to the formation of localized cutaneous ulcers. Furthermore, subsequent modeling demonstrated a significant increase of NETs and GSDMD in skin lesions and peripheral blood serum of wild-type mice. In GSDMD-/- mice, the severity of skin ulcers after modeling was significantly diminished."

GO: GO:0140447 cytokine precursor processing (verify); GO:1990266 neutrophil migration; GO:0036102 leukotriene B4 metabolic process (peripheral); GO:0044130/NET-formation terms should be verified against current GO — the canonical is GO:1990266 neutrophil migration plus a NET term.

T cells at the wound margin. PMID:33033263 [verbatim]: "Studies have focused on the role of T cells, especially at the wound margin; these cells may support the destructive autoinflammatory response by the innate immune system." This is architecturally important: PG is not purely innate. PMID:20636397 quantified the spatial gradient [paraphrase]: "In ulcerative PG, CD3 and CD163 were significantly higher in wound edge than wound bed, while myeloperoxidase was expressed more in wound bed." That edge-vs-bed gradient is a curatable spatial mechanism and the reason the Delphi criteria specify biopsy of the ulcer edge.

Clonal T-cell proliferation in lesions has been described (StatPearls [paraphrase]), and immunoprofiling has suggested T-cell exhaustion.

6.4 Protein dysfunction

• PSTPIP1: mutant protein shows increased pyrin binding and c-Abl-dependent hyperphosphorylation, disrupting normal PTP-PEST interaction (PMID:11971877, PMID:14595024). UniProt: O43586 (PSTPIP1_HUMAN). Pyrin: O15553 (MEFV).
• Gasdermin D (GSDMD): pore-forming executioner; UniProt P57764. Required for NET formation and ulcer severity in the mouse model (PMID:40034857).
• MMP-2 / MMP-9: matrix-degrading effectors, overexpressed in PG lesional skin, more so than in Sweet syndrome and amicrobial pustulosis (PMID:20636397, PMID:21658319). UniProt P08253 (MMP-2), P14780 (MMP-9).
• Myeloperoxidase (MPO): UniProt P05164; expressed maximally in the wound bed.

6.5 Metabolic changes

No PG-specific metabolomic or lipidomic signature has been published. No entry in MetaboLights, Metabolomics Workbench, or HMDB is PG-specific. Record as a KNOWLEDGE_GAP.

6.6 Immune system involvement

PG is classified as an autoinflammatory — not autoimmune — disease. The definitional argument is in PMID:24903614 and PMID:25501066: recurrent sterile inflammation without circulating autoantibodies and without autoreactive T cells. PMID:25501066 adds a key compartmental finding [verbatim]:

"In peripheral blood, serum levels of the main proinflammatory cytokines, that is, IL-1β, tumor necrosis factor-α, and IL-17, were within the normal range, suggesting that in PASH syndrome, the inflammatory process is mainly localized into the skin."

This skin-localized, serum-normal pattern is a mechanistically load-bearing fact: it explains why serum cytokine panels are useless as PG biomarkers and why lesional-tissue assays are required. (Note that the PG-proper serum proteome may be broader — PMID:37909252 reports that "the serum proteome of pyoderma gangrenosum is more expansive than that of hidradenitis suppurativa" — so do not over-generalize the PASH finding to all PG.)

Contrasting counterpoint worth curating: the framework paper on autoinflammatory classification is PMID:19302049 (Masters SL, Simon A, Aksentijevich I, Kastner DL, Annu Rev Immunol 2009), "Horror autoinflammaticus."

6.7 Tissue damage mechanisms

Proteolytic (MMP-2/MMP-9), oxidative (MPO-derived reactive oxygen and halogenated species), and NET-mediated cytotoxicity — all downstream of the neutrophil. PMID:20636397 concludes [paraphrase] that the study "identifies PG as a paradigm of neutrophil-mediated inflammation with proinflammatory cytokines/chemokines and MMPs as important tissue damage effectors." Ischemia and fibrosis are not primary mechanisms — this distinguishes PG from Martorell ulcer and arterial ulcers in the differential.

6.8 Biochemical abnormalities

No enzyme deficiency, no ion-channel defect, no receptor loss. The abnormality is regulatory: a lowered activation threshold of the pyrin/NLRP3 inflammasome and of the C5a-neutrophil axis.

6.9 Molecular profiling

Transcriptomics. Two Ortega-Loayza studies anchor this: - PMID:28734003 — "Dysregulation of inflammatory gene expression in lesional and nonlesional skin of patients with pyoderma gangrenosum" (Br J Dermatol 2018, DOI 10.1111/bjd.15837). Note the nonlesional finding: dysregulation is present in clinically normal skin, consistent with a systemic predisposition rather than a purely local event. - PMID:34536481 — "Molecular and Cellular Characterization of Pyoderma Gangrenosum: Implications for the Use of Gene Expression" (J Invest Dermatol 2022, DOI 10.1016/j.jid.2021.08.431). - PMID:39098048 — dHACM interventional transcriptomics (NCT05120726), 4 patients, RNA-seq pre/post treatment [verbatim]: "We observed varied changes to the local expression of inflammatory response, positive regulators of cellular proliferation, and extracellular matrix disassembly cytokines. All PG wounds produced granulation tissue following treatment and were closed using split-thickness skin grafts."

Proteomics. PMID:37909252 (Flora A, Pham J, Jepsen R, Frew JW, JEADV 2024, DOI 10.1111/jdv.19611) — the PG serum proteome is more expansive than that of HS.

Single-cell and spatial. The most recent frontier: - PMID:42603447 — scRNA-seq + multiplex IHC demonstrating JAK/STAT overactivation, NETosis, and aberrant Th17 differentiation. - "IL-12/IL-23 blockade reveals patterns of asynchronous inflammation in pyoderma gangrenosum" — J Invest Dermatol 2024/2025 (bioRxiv preprint 2024.04.26.591387). Asynchronous inflammation — different regions of the same ulcer at different inflammatory stages — is a mechanistically important and currently under-curated concept: it explains treatment-response heterogeneity within a single lesion and argues against single-biopsy sampling.

Datasets. No PG-specific GEO series was confirmed during this search. If curating a datasets: block, run just discover-datasets Pyoderma_Gangrenosum and just verify-datasets — and apply the Named Entity Confusion triage the CLAUDE.md warns about, since "pyoderma" searches will surface veterinary canine pyoderma (a bacterial folliculitis, a completely different disease).

Functional genomics screens. None PG-specific. Gap.

6.10 Suggested CL terms

7. Anatomical Structures Affected

7.1 Organ level

Primary: Skin — UBERON:0002097 skin of body; specifically UBERON:0002199 dermis (the site of the neutrophilic infiltrate) and UBERON:0001003 skin epidermis (secondarily destroyed). Predilection sites: UBERON:0000975 anterior region of leg / pretibial skin; peristomal abdominal skin (UBERON:0001416 skin of abdomen); breast skin (UBERON:0001868, UBERON:0000310 breast).

The hair follicle (UBERON:0002073 hair follicle) deserves a node given the "follicular unit as putative initial target" hypothesis (PMID:35606650).

Secondary / extracutaneous. PG is overwhelmingly cutaneous, but sterile neutrophilic infiltrates in extracutaneous organs are documented and clinically important. Reported sites: lung (the commonest extracutaneous site, presenting as sterile pulmonary infiltrates or nodules), spleen, psoas muscle, bone, and eye. PMID:15888172 (Hubbard VG, Friedmann AC, Goldsmith P, Br J Dermatol 2005) describes idiopathic PG with splenic and psoas muscle involvement [paraphrase], and notes these extracutaneous manifestations are extremely rare.

UBERON: UBERON:0002048 lung; UBERON:0002106 spleen; UBERON:0001369 psoas major muscle (verify); UBERON:0001474 bone element.

Body systems: integumentary (primary); immune/hematopoietic (both mechanism and comorbidity); musculoskeletal (syndromic arthritis); digestive (IBD comorbidity).

7.2 Tissue and cell level

Dermis (connective tissue) is the primary compartment. The infiltrate is dense, sterile, and predominantly neutrophilic, forming dermal abscesses; with epidermal ulceration and, in the ulcerative variant, an undermined edge where the epidermis is separated from the underlying dermis.

The 86-patient Mayo review (PMID:3889978, Powell FC, Schroeter AL, Su WP, Perry HO, QJM 1985) records the histologic zonation [paraphrase]: "Lymphocytic vasculitis predominated peripherally; neutrophilic infiltrates centrally." This matches Marzano's later immunohistochemical gradient (PMID:20636397): CD3+ T cells and CD163+ macrophages at the wound edge; MPO+ neutrophils in the wound bed.

7.3 Subcellular level

• Inflammasome complex: GO:0072559 NLRP3 inflammasome complex; GO:0140738 pyrin inflammasome complex (verify current label/ID).
• Plasma membrane pore (GSDMD): GO:0005886 plasma membrane.
• Azurophilic granule: GO:0042582 azurophil granule (MPO, elastase source).
• Extracellular NET: GO:0005576 extracellular region.

7.4 Localization and lateralization

Asymmetric and often multifocal; lower legs most common. Bilateral presentations occur but are atypical enough to be reported as such (PMID:41614012). Peristomal, breast, and post-surgical-incision distributions are trigger-determined, not intrinsic to the disease. Notably, in post-surgical breast PG, PMID:17966539 records that the disease "affects any anatomical location except the nipple-areolar complex" [paraphrase] — an interesting anatomically specific sparing that would be worth verifying before curating.

8. Temporal Development

8.1 Onset

• Typical age: mid-40s mean (PMID:33033263); median 59 (IQR 41–72) in the UK cohort (PMID:22534879). Peristomal PG onsets younger (mean 46, PMID:27473454), reflecting the IBD population.
• Pattern: acute to subacute. The prodrome is a papule, pustule, or vesicle that ulcerates within 4 days (Delphi minor criterion 4, PMID:29450466), then expands rapidly. Post-surgical PG shows wound complication signs at a mean of 7.0 days after surgery (PMID:25589459), with the wider reported range being 4 days to 6 weeks (PMID:17966539). Peristomal PG onsets much later — mean 5.2 months after stoma surgery (PMID:27473454), range 2 weeks to 3 years (PMID:7912923).

8.2 Progression

Stages. No formal staging system exists (unlike AJCC or WHO systems). Clinically, PG is described in two phases: an inflammatory/expanding phase (violaceous undermined border advancing) and a healing/granulating phase (cribriform re-epithelialization). This two-phase model is the design basis of trials such as NCT04274166, "Secukinumab for the Inflammatory Phase of Pyoderma Gangrenosum."

Rate. Rapid during the inflammatory phase (a defining diagnostic feature, and the reason "rapid progression" is a Su major criterion), then slow. Median time to healing on topicals: 145 days (PMID:27502313).

Course pattern. Relapsing–remitting / recurrent. Recurrence 28–30% at 6 months post-treatment in STOP GAP (PMID:26071094); 61% recurrence in peristomal PG (PMID:27473454), rising to 67% (10 of 15) after stoma relocation or revision — a key negative surgical finding.

Duration. Chronic and lifelong in susceptibility, episodic in expression. Peristomal PG healed completely in all 20 patients of one series but took a mean of 11.4 months (median 8, range 1–41) (PMID:10807281).

8.3 Remission patterns

Treatment-induced remission is the norm; spontaneous remission is uncommon but reported for the vegetative variant. Peristomal PG achieved remission in 29 of 31 (94%) patients (PMID:27473454). Stoma closure had the highest complete-response rate (4/4, no recurrences) — a mechanistically satisfying result: remove the trigger, remove the disease.

8.4 Critical periods

Two windows dominate:

1. The first ~2 weeks post-surgery — the window in which PSPG must be distinguished from wound infection. Getting this wrong is catastrophic: debridement in this window causes pathergic enlargement. PMID:25589459's operational recommendation [verbatim]: "Debridement should not be performed before dermatologic consultation to assess for PSPG."
2. The first year after diagnosis of an associated disease — the period of maximal PG risk after UC (OR 35.50 in year 1, PMID:33647909) and after hematologic malignancy (PMID:39118665). This defines a surveillance window.

9. Inheritance and Population

9.1 Epidemiology

Incidence. The authoritative population-based figure is from the UK GPRD study (PMID:22534879, Langan SM, Groves RW, Card TR, Gulliford MC, J Invest Dermatol 2012, DOI 10.1038/jid.2012.130) [verbatim]:

"The adjusted incidence rate standardized to European standard population was 0.63 (95% confidence interval (CI) 0.57–0.71) per 100,000 person-years."

Broader literature estimates 3–10 cases per million per year (PMID:25213386 [paraphrase]), i.e. 0.3–1.0/100,000/year — consistent with the UK figure.

Prevalence. The systematic review and meta-regression is PMID:40506010 (Shea M, Munoz EP, Kumar I, Zanet RA, Sengupta S, Ortega-Loayza AG, J Invest Dermatol 2025, DOI 10.1016/j.jid.2025.05.030). Its abstract could not be retrieved through any of the routes tried (PubMed cookie wall, Europe PMC null abstract field, Semantic Scholar null, JID 403). Do not curate a pooled prevalence number from this paper until the abstract has been fetched with just fetch-reference PMID:40506010 and the snippet verified. Until then, curate prevalence from the incidence data plus the Orphanet band.

Suggested dismech Prevalence records:

prevalence:
- population: United Kingdom (General Practice Research Database)
  measure_type: ANNUAL_INCIDENCE
  prevalence_class: BAND_1_9_PER_1000000
  rate_per_100000: 0.63
  rate_low: 0.57
  rate_high: 0.71
  notes: European-standard-population-adjusted incidence rate.
  evidence:
  - reference: PMID:22534879
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The adjusted incidence rate standardized to European standard population was 0.63 (95% confidence interval (CI) 0.57-0.71) per 100,000 person-years."
    explanation: Population-based incidence estimate from a representative UK primary-care database.

Mortality. From the same UK cohort [verbatim]:

"The risk of death was three times higher than that for general controls (adjusted hazard ratio=3.03, 95% CI 1.84–4.73, P<0.001), 72% higher than that for IBD controls (adjusted hazard ratio=1.72, 95% CI 1.17–2.59, P=0.013), with a borderline increase compared with RA controls (adjusted hazard ratio=1.55, 95% CI 1.01–2.37, P=0.045)."

In-hospital mortality: 3.2% (74 of 2,273 US inpatient admissions, PMID:29334018).

A 2026 signal worth tracking: PMID:42263577 (Kerniss H et al., Atherosclerosis 2026), "Pyoderma gangrenosum is associated with excess incident major atherothrombotic events."

9.2 Inheritance

For idiopathic PG: multifactorial / polygenic; not Mendelian. Recurrence risk to relatives is not quantified.

For the syndromic forms: - PAPA syndrome (MONDO:0011462, OMIM #604416): autosomal dominant, PSTPIP1. HPO inheritance term HP:0000006 Autosomal dominant inheritance. - PASH: mostly sporadic; occasional NCSTN or PSTPIP1 variants (PMID:25601011, PMID:26713508). - Penetrance/expressivity in PAPA: incomplete penetrance and highly variable expressivity — the classic triad is often incomplete, and the PG component in particular is variably present and typically post-pubertal while the pyogenic arthritis is childhood-onset. PMID:25845478 discusses this variability explicitly [paraphrase]. PSTPIP1-negative PAPA phenotypes exist (PMID:19700023). - Anticipation, germline mosaicism, founder effects, consanguinity, carrier frequency: not established for PG or PAPA.

For idiopathic PG the appropriate Inheritance annotation is HP:0010982 Polygenic inheritance or, more honestly, omit inheritance entirely and record a KNOWLEDGE_GAP — there is no polygenic architecture study to cite.

9.3 Population demographics

Sex ratio. Consistent female predominance:

The male predominance in hematologic-malignancy-associated PG (PMID:31560977) is a real subgroup inversion and should be curated on the subtype, not the disease.

Ethnicity. US NIS data: 71.1% Caucasian (PMID:29334018). Whether this reflects true susceptibility or ascertainment is unresolved — no ancestry-stratified incidence study exists. Do not curate an ethnic predisposition claim.

Geographic distribution. No endemic pattern; reported worldwide, including from resource-limited settings (PMID:42502448, Uganda). PG is not geographically clustered.

Age distribution. Peak in the 5th–6th decades; <5% pediatric; onset reported 11–89 years.

10. Diagnostics

10.1 The core problem

PG remains a clinical diagnosis with no confirmatory test. The most consequential diagnostic study remains PMID:12409543 (N Engl J Med 2002): 10% of patients treated for PG had a different disease — vascular occlusive disease, vasculitis, malignancy, infection, drug-induced ulceration, or exogenous tissue injury. Over-diagnosis exposes patients to unnecessary immunosuppression; under-diagnosis leads to pathergic debridement.

10.2 Diagnostic criteria

Three published criteria sets. Curate all three with their operating characteristics.

(a) Su criteria (2004) — PMID:15533059 (Su WP, Davis MD, Weenig RH, Powell FC, Perry HO, Int J Dermatol, DOI 10.1111/j.1365-4632.2004.02128.x). Two major + two of four minor: - Major: (1) rapid progression of a painful necrolytic cutaneous ulcer with an irregular, violaceous, undermined border; (2) exclusion of other causes of cutaneous ulceration. - Minor: (1) history suggestive of pathergy or cribriform scarring; (2) systemic disease associated with PG; (3) histopathologic findings (sterile dermal neutrophilia ± mixed inflammation ± lymphocytic vasculitis); (4) rapid response to systemic corticosteroid treatment.

(b) Delphi consensus criteria (2018) — PMID:29450466 (Maverakis E, Ma C, Shinkai K, et al., JAMA Dermatol 154(4):461–466, DOI 10.1001/jamadermatol.2017.5980). One major + ≥4 of 8 minor. Abstract [verbatim]:

"Delphi exercise yielded 1 major criterion—biopsy of ulcer edge demonstrating neutrophilic infiltrate—and 8 minor criteria: (1) exclusion of infection; (2) pathergy; (3) history of inflammatory bowel disease or inflammatory arthritis; (4) history of papule, pustule, or vesicle ulcerating within 4 days of appearing; (5) peripheral erythema, undermining border, and tenderness at ulceration site; (6) multiple ulcerations, at least 1 on an anterior lower leg; (7) cribriform or 'wrinkled paper' scar(s) at healed ulcer sites; and (8) decreased ulcer size within 1 month of initiating immunosuppressive medication(s). Receiver operating characteristic analysis revealed that 4 of 8 minor criteria maximized discrimination, yielding sensitivity and specificity of 86% and 90%, respectively."

(c) PARACELSUS score (Jockenhöfer et al., 2019) — a weighted point score; the most sensitive instrument, with historically limited specificity. The definitive validation is PMID:41785996 (Moelleken M, Ortega-Loayza AG, Busch D, …, Dissemond J, J Am Acad Dermatol 2026, DOI 10.1016/j.jaad.2026.02.101), an international multicenter study of 1,403 cases from 14 institutions in 7 countries [verbatim]:

"Among 1403 cases (57.0% women, 43.0% men; mean age 62 years), 85 wound entities were identified, including 180 cases of PG. Raising the diagnostic cut-off from ≥10 to >10 points improved specificity (93.2% vs 96.8%; P < .001), positive predictive value (68.4% vs 81.9%; P < .001) and accuracy (94.1% vs 97.0%; P < .001). The false-positive rate was decreased (6.8% vs 3.2%; P < .001) with a non-significant reduction in sensitivity (100.0% vs 98.3%; P = .25)."

This is the single most important 2026 diagnostic update and should be curated as a definitions entry with definition_type: PHENOTYPE_ALGORITHM, derivation_basis: ESTABLISHED_CRITERIA, and validation_status.status: VALIDATED_AGAINST_GOLD_STANDARD.

10.3 Biopsy — and the argument against it

Delphi makes biopsy the sole major criterion. But biopsy is both risky (pathergy) and non-specific. PMID:41923959 (Moore AM, Karch JL, Bradley KE, Stevanovic M, Salem I, Parker DJ, Simmons BJ, Skin Health Dis 2026, DOI 10.1093/skinhd/vzaf087) [verbatim]:

"Among 58 patients, 26 (45%) underwent biopsies, with only 10 (38%) contributing to a PG diagnosis… Given risk of pathergy, nonspecific histopathological findings and low sensitivity, in our opinion, based on this small sample size, biopsies have limited diagnostic value for PG."

This is a genuine, live controversy between the Delphi and PARACELSUS camps and belongs in a discussions block with kind: KNOWLEDGE_GAP, not silently resolved in favor of one side.

Histopathology when performed: dense dermal neutrophilic infiltrate with abscess formation, epidermal ulceration, and an undermined edge; often a peripheral lymphocytic vasculitis with central neutrophilia (PMID:3889978). Biopsy the ulcer edge, not the base — the CD3/CD163 edge vs. MPO base gradient (PMID:20636397) is the histological reason.

10.4 Laboratory tests

There is no diagnostic biomarker. Tests are performed to (a) exclude mimics and (b) find the associated systemic disease:

Serum cytokine measurement is not diagnostically useful — PASH data show serum IL-1β, TNF-α and IL-17 within the normal range despite florid lesional overexpression (PMID:25501066).

10.5 Imaging and functional testing

No imaging is diagnostic. CT/MRI is used to define extent in extracutaneous PG and to exclude deep infection/osteomyelitis. Vascular imaging excludes arterial insufficiency. No role for PET, EEG, EMG, or electrophysiology.

10.6 Genetic testing

Not indicated for sporadic adult-onset PG. Indicated when: - PG presents in childhood (consider PSTPIP1, and an inborn-error-of-immunity panel — PMID:38951460) - The PAPA/PASH/PAPASH/PASS/PsAPASH phenotype is present → PSTPIP1 single-gene or targeted panel; NCSTN for PASH - There is recurrent sterile inflammation suggesting a hereditary periodic fever syndrome → MEFV, NLRP3, IL1RN, NFKB1, LPIN2 panel - PG occurs with cytopenias/MPN → somatic JAK2 V617F on blood/marrow (a somatic test, not germline)

WES/WGS have a defined role in unexplained childhood or syndromic PG (PMID:38951460 assembled its 74 cases largely from such workups). CMA, karyotype, FISH, mtDNA testing, and repeat-expansion testing have no role except FISH/karyotype for suspected MDS (e.g. trisomy 8 — PMID:28943508).

10.7 Omics-based diagnostics

None validated for clinical use. Research-stage: lesional transcriptomics (PMID:34536481, PMID:28734003), serum proteomics (PMID:37909252), serum NET levels (PMID:40034857 — elevated in PG vs. healthy controls, the most promising candidate biomarker to date). Liquid biopsy: N/A.

10.8 Differential diagnosis

10.9 Screening

No population screening exists or is warranted (prevalence far too low).

Targeted case-finding, however, is standard of care in both directions: - Screen every new PG patient for an underlying systemic disease. Justified by the 56.8% pooled prevalence (PMID:29721816) and by the mortality gradient: PMID:29438762 (Kaffenberger BH, Hinton A, Krishna SG, J Am Acad Dermatol 2018) [verbatim]: "vasculitis and hematologic malignancy/dyscrasia, when compared with inflammatory bowel disease, were associated with a 4-fold to 6-fold increased risk of in-hospital mortality." - Screen specifically for hematologic malignancy. PMID:31560977 (Montagnon CM, …, Tolkachjov SN, J Am Acad Dermatol 2020) [verbatim]: "patients with PG should be evaluated for hematologic malignancies, with MDS being the most common." - Genetic counseling and cascade testing apply only to PAPA-spectrum families.

11. Outcome / Prognosis

11.1 Survival and mortality

• All-cause mortality HR 3.03 (95% CI 1.84–4.73) vs. matched general-population controls (PMID:22534879). Excess mortality persists after adjusting for IBD (HR 1.72) and RA (HR 1.55) comparators, so it is not fully explained by comorbidity.
• In-hospital mortality 3.2% across 2,273 US admissions (PMID:29334018); 14% (4/29) and 22% (5/23) in two small tertiary inpatient series (PMID:23903083, PMID:25374597) — reflecting severity selection.
• No published 5-/10-year survival curve specific to PG. Gap.
• Death is usually attributable to sepsis complicating the ulcer, complications of long-term immunosuppression, or the underlying systemic disease — not to the ulcer per se. In PMID:26071094, serious adverse reactions, "particularly infections, were more prevalent in the prednisolone group."

11.2 Morbidity and function

• Prolonged hospitalization: mean LOS 47 days (range 5–243) (PMID:25374597).
• Complications of therapy are near-universal: 66% of admissions had complications from medical therapy, most commonly poor glycemic control (17%) and steroid-induced diabetes (14%) (PMID:23903083).
• Permanent cribriform atrophic scarring in most healed patients.
• Chronic pain and pruritus (PMID:42472079); opioid dependence risk.
• No PG-specific validated QoL instrument exists; studies use generic dermatology instruments (DLQI) and pain NRS. STOP GAP included QoL as a secondary outcome (PMID:26071094).

11.3 Disease course and recovery potential

• 47% healed at 6 months on either prednisolone or ciclosporin in STOP GAP (PMID:26071094) — i.e., more than half of adequately treated patients still have an open ulcer at 6 months. This is the most sobering number in PG and should be curated verbatim.
• 43.8% healed by 6 months on topical therapy alone in the parallel cohort (PMID:27502313).
• Recurrence: 30% (ciclosporin) / 28% (prednisolone) in STOP GAP; 61% in peristomal PG.
• Remission is achievable: 94% of peristomal PG patients reached remission (PMID:27473454).

11.4 Prognostic factors

11.5 Prognostic biomarkers

None validated. Serum NET level (PMID:40034857) is the leading research candidate.

12. Treatment

12.1 The central fact

There is no FDA-approved therapy for pyoderma gangrenosum. Stated directly in PMID:39720859 (Keum H, Zhivov EV, Ortega-Loayza AG, Expert Rev Clin Pharmacol 2025, DOI 10.1080/17512433.2024.2447776) [paraphrase]: the disease "lacks an FDA-approved treatment." Every therapy below is off-label.

12.2 First-line systemic therapy — the STOP GAP evidence

The only adequately powered head-to-head RCT is PMID:26071094 (Ormerod AD, Thomas KS, Craig FE, Mitchell E, Greenlaw N, Norrie J, Mason JM, Walton S, Johnston GA, Williams HC, BMJ 2015, DOI 10.1136/bmj.h2958), 121 patients across 39 UK hospitals [paraphrase, verify against cached abstract before curating]:

"At six weeks, ciclosporin showed mean speed of healing of −0.21 (1.00) cm²/day versus −0.14 (0.42) cm²/day for prednisolone, with no significant between-group difference (0.003 cm²/day, 95% CI −0.20 to 0.21; P=0.97). By six months, ulcer healing occurred in 28/59 (47%) ciclosporin participants and 25/53 (47%) prednisolone participants. Recurrence rates were similar: 30% with ciclosporin and 28% with prednisolone. Adverse reactions were comparable (68% versus 66%), though serious adverse reactions, particularly infections, were more prevalent in the prednisolone group."

Curated conclusion: prednisolone and ciclosporin are therapeutically equivalent; choose by comorbidity and adverse-effect profile (avoid ciclosporin in renal impairment/hypertension; avoid prednisolone in diabetes and in the immunosuppression-naive elderly).

12.3 Treatment table with suggested NCIT annotations

A target_mechanisms note. Several of these treatments should carry target_mechanisms links into the pathophysiology graph with an evidence-bearing INHIBITS edge — vilobelimab → the C5a node, spesolimab → the IL-36 node, canakinumab/anakinra → the IL-1β node, infliximab/adalimumab → the TNF-α node, baricitinib → the JAK-STAT node. This is exactly the drug-target pattern the dismech modules use.

12.4 The infliximab RCT

PMID:16188920 (Brooklyn TN, Dunnill MG, Shetty A, Bowden JJ, Williams JD, Griffiths CE, Forbes A, Greenwood R, Probert CS, Gut 2006, DOI 10.1136/gut.2005.074815), the only placebo-controlled biologic RCT with a positive result [verbatim]:

"significantly more patients in the infliximab group had improved (46% (6/13)) compared with the placebo group (6% (1/17); p = 0.025)"

Overall clinical response 69%; complete remission 21% at week 6 [paraphrase].

12.5 The trial graveyard — a curatable pattern

An unusually high proportion of PG trials have been terminated or withdrawn: three gevokizumab Phase III trials, the vilobelimab Phase III (NCT05964413), the spesolimab Phase II (NCT06092216), etrasimod Phase II, hyperbaric oxygen Phase III, the dHACM Phase IV, two adalimumab Phase II trials (withdrawn), deucravacitinib Phase I (withdrawn), secukinumab (withdrawn), and PRP (withdrawn).

This is not incidental — it reflects (a) recruitment difficulty in an ultra-rare disease, (b) the absence of a validated primary endpoint, and (c) the high spontaneous/steroid-induced healing rate that swamps drug effect. PMID:39927907 (Becker SL, Ortega-Loayza AG, "The Changing Landscape of Clinical Research in Pyoderma Gangrenosum," J Invest Dermatol 2025) addresses exactly this. Curate it as a KNOWLEDGE_GAP discussion on trial methodology, and do not curate a terminated trial's drug as an effective treatment.

12.6 Treatment algorithm

1. Confirm the diagnosis (PARACELSUS >10, or Delphi 1 major + ≥4 minor) and actively exclude mimics — 10% misdiagnosis rate.
2. Search for the associated systemic disease (CBC, SPEP, colonoscopy if GI symptoms, joint assessment) — and stratify prognosis on it.
3. Localized/mild disease: superpotent topical corticosteroid or topical tacrolimus ± intralesional triamcinolone. 43.8% heal by 6 months.
4. Extensive or rapidly progressive disease: systemic prednisolone 0.75 mg/kg/day or ciclosporin 4 mg/kg/day — equivalent; choose by comorbidity.
5. Refractory or steroid-dependent: add/switch to anti-TNF (infliximab has RCT support; adalimumab has Phase III + 52-week real-world data). In IBD-associated PG, anti-TNF treats both compartments.
6. Anti-TNF failure: ustekinumab, IL-1 blockade (especially if PAPA/PASH), IL-23 blockade, or a JAK inhibitor. Use IL-17 blockade with awareness of the paradoxical-PG signal.
7. Throughout: meticulous non-debriding wound care, aggressive pain control, infection surveillance, and avoid surgical debridement during the inflammatory phase.
8. Surgery only when disease is quiescent and under immunosuppressive cover. PMID:23903083 [verbatim]: "All 3 patients who underwent split skin grafting under immunosuppressive cover (with 2 having hyperbaric oxygen therapy) had no postoperative graft failure or pathergy."
9. Peristomal PG: treat the ulcer and the underlying IBD; consider stoma closure (best response), avoid stoma relocation (67% recurrence).

12.7 Real-world adalimumab data (2026)

PMID:42107018 (Yamamoto T, Tanizaki H, Yamasaki K, Matsubara N, Nakayama M, Iwashita E, Yamanaka K, Dermatol Ther 2026, DOI 10.1007/s13555-026-01772-4), 67 patients, 52 weeks [paraphrase]: PGA 0/1 in 36.0% at week 12, 46.2% at week 26, 57.7% at week 52; pain score 0 in 45.7% at week 26 and 52.4% at week 52; infection AEs 14.9%, serious reactions 9.0%; no relapses among patients discontinuing for improvement.

12.8 Pharmacogenomics

None established for PG. Generic pharmacogenomic considerations apply to the drugs used (TPMT/NUDT15 for azathioprine; CYP3A4/ABCB1 for ciclosporin) but no PG-specific PGx evidence exists. Record as a gap.

13. Prevention

13.1 Primary prevention

Not possible for a first episode of idiopathic PG. No modifiable exposure has been shown to prevent PG. The only defensible primary-prevention statements are indirect: smoking cessation (HR 2.12 for chronic inflammatory disease, PMID:40012715) and weight management (PMID:39963282, PMID:22959399).

13.2 Secondary prevention (early detection)

The actionable secondary-prevention target is early recognition of PSPG in the post-operative window, because the intervention (withhold debridement, start immunosuppression) is time-critical and the harm from missing it is severe.

13.3 Tertiary prevention — the strongest evidence in this section

1. Avoid pathergy-inducing procedures. No elective debridement, no needle biopsy of an active edge without cause, no stoma relocation. This is the single most effective preventive intervention in PG.
2. Perioperative corticosteroid prophylaxis in at-risk patients. PMID:25589459 [verbatim]: "Patients at risk of PSPG undergoing breast surgery may benefit from perioperative prednisone to prevent PSPG which can lead to destructive wound enlargement and significant scarring." Risk group: prior PG, RA, IBD, or hematologic malignancy undergoing breast, cardiothoracic, or abdominal surgery.
3. Control the underlying disease. Treating active IBD prevents PG flares; in peristomal PG, PPG onset "usually heralds active CD" (PMID:10807281 [paraphrase]).
4. Maintenance biologic therapy to prevent recurrence (PMID:12907338: ten of thirteen patients maintained healing with infusions every 4–12 weeks [paraphrase]).
5. Infection prophylaxis and monitoring during immunosuppression — the leading cause of serious adverse events (PMID:26071094).

13.4 Immunization

No vaccine prevents PG. Standard immunosuppression-related vaccination (pneumococcal, influenza, zoster; live vaccines contraindicated on biologics) applies as supportive care, not as PG prevention.

13.5 Screening, risk stratification, counseling, public health

• Population screening: not warranted.
• Genetic screening: only in PAPA-spectrum families (NCIT:C15240 Genetic Counseling). Prenatal/PGD is theoretically available for a known PSTPIP1 variant but is not standard practice given the treatable phenotype.
• Risk stratification: the Kridin ORs (UC 14.6×, CD 28×, hematologic malignancy 7.9×, gout 5.2×, RA 3.3×, GPP 5.1×) constitute an implicit risk model, though no validated PG risk-prediction calculator exists.
• Public health / environmental interventions: not applicable.

14. Other Species / Natural Disease

This section is largely a negative, and the negative is important — it is a Named Entity Confusion trap.

• NCBI Taxonomy: NCBITaxon:9606 Homo sapiens. PG as defined here is a human disease.
• Critical NEC warning: the term "pyoderma" in veterinary medicine (canine superficial/deep pyoderma) denotes a bacterial folliculitis, usually Staphylococcus pseudintermedius, which is mechanistically the opposite of PG — infectious rather than sterile, and treated with antibiotics rather than immunosuppression. Any literature search, deep-research report, or dataset-discovery run on "pyoderma" will surface large volumes of canine pyoderma literature. Do not curate any of it into this entry. Run just preflight-dr <report> MONDO:0018824 on any DR report before use; note that PG has no MONDO causal gene, so the preflight will likely return SKIP and the manual synonym/OMIM checks must be done by hand.
• Naturally occurring PG in other species: no OMIA entry corresponds to human PG. There is no established naturally occurring animal counterpart.
• Orthologous genes (relevant only to the syndromic forms): mouse Pstpip1 (MGI), Mefv, Gsdmd. Human PSTPIP1 HGNC:9580; GSDMD HGNC:25697; MEFV HGNC:6998.
• Zoonotic potential / cross-species transmission: not applicable — PG is non-infectious and non-transmissible.
• Comparative biology: the inflammasome/pyrin/GSDMD machinery is deeply conserved across mammals, which is what makes the mouse models below informative; the disease is not.

15. Model Organisms

15.1 The flagship model — GSDMD/serum-transfer mouse

PMID:40034857 (Li S, Ying S, Fang H, Qiao J, iScience 2025, DOI 10.1016/j.isci.2025.111925) established the first purpose-built PG animal model [verbatim]:

"Injection of serum from PG patients into the dorsal skin of wild-type mice led to the formation of localized cutaneous ulcers. Furthermore, subsequent modeling demonstrated a significant increase of NETs and GSDMD in skin lesions and peripheral blood serum of wild-type mice. In GSDMD-/- mice, the severity of skin ulcers after modeling was significantly diminished. Overall, our findings shed light on the role of GSDMD in regulating the production of NETs by neutrophils and the release of inflammatory factors in the pathogenesis of PG and establish an animal model for studying PG."

Suggested dismech animal_models entry:

animal_models:
- name: PG-patient-serum transfer model in wild-type and Gsdmd-/- mice
  species: Mouse
  genotype: C57BL/6 wild type; Gsdmd knockout
  publication: PMID:40034857
  modeled_mechanisms:
  - target: GSDMD-Dependent NETosis
    relationship: RECAPITULATES
    fidelity: MODERATE
    description: >-
      Intradermal injection of PG patient serum induces localized cutaneous
      ulceration with lesional NET and GSDMD accumulation; genetic GSDMD
      deletion attenuates ulcer severity, establishing NETosis as causally
      required rather than merely correlated.
    limitations: >-
      A passive serum-transfer model rather than a genetic model of the human
      disease; it reproduces the effector arm (NET-driven ulceration) but not
      the upstream genetic susceptibility, the associated systemic diseases,
      or the chronic relapsing course. Ulcers are induced and localized, not
      spontaneous. The transferable serum factor is not identified.
    readouts:
    - name: Cutaneous ulcer severity
      target: GSDMD-Dependent NETosis
      direction: DECREASED
      interpretation: Ulcer severity is reduced in Gsdmd-/- versus wild-type mice.

15.2 PSTPIP1 / PAPA models

Mechanistic work is largely in vitro and ex vivo rather than in dedicated knock-in mice. PMID:14595024 used yeast two-hybrid screening, co-expression in monocytes and granulocytes, and patient PBMC IL-1β measurement — [paraphrase] "increased IL-1beta production by peripheral blood leukocytes from a clinically active PAPA patient" carrying A230T. Published Pstpip1 mouse work exists in the wider autoinflammation literature but was not retrieved as PG-specific in this search; verify directly in MGI before curating a specific mouse line.

Related and better-characterized models for the shared IL-1 axis: DIRA (Il1rn-deficient) mice, relevant via PMID:19494218 (Aksentijevich I et al., N Engl J Med 2009) [verbatim]: "We identified homozygous mutations of IL1RN in nine affected children" with "neonatal onset of sterile multifocal osteomyelitis, periostitis, and pustulosis." These recapitulate sterile neutrophilic skin inflammation but not PG's ulcer morphology.

15.3 In vitro and human-derived systems (NAMs)

These belong in experimental_models:, not animal_models::

15.4 Model limitations — the honest summary

PG has no model that recapitulates the human disease. No mouse spontaneously develops chronic, relapsing, pathergy-responsive ulceration with an undermined violaceous border. The GSDMD model is an induced effector-arm model; the PSTPIP1 work is molecular. This is a genuine HUMAN_MODEL_MISMATCH (not merely a KNOWLEDGE_GAP) in the dismech sense: evidence exists in models, but its translational validity to human PG is the open question. It is a substantial reason PG therapeutics have advanced by clinical serendipity and mechanism-borrowing from psoriasis/HS rather than by target validation.

15.5 Model resources

MGI (mouse Pstpip1, Gsdmd, Mefv, Il1rn), IMPC/KOMP for knockout availability, Alliance of Genome Resources for orthology, Cellosaurus for any cell lines. No PG-specific model repository or registry exists.

Cross-cutting notes for the dismech curator

1. This entry should conform to existing modules. Candidates: - A new or existing neutrophilic-inflammation/inflammasome module would be the natural home for the IL-1β → IL-8 → neutrophil chain. Check kb/modules/ for an inflammasome module before authoring one; the cellular_senescence/granuloma_formation precedents show the shape. - granuloma_formation is not the right module — PG is abscess-forming, not granuloma-forming (except the superficial granulomatous variant, which is the exception that proves the rule). - The Xogenesis convention does apply conceptually: PG forms a pathological structure (sterile dermal abscess/ulcer). If a sterile_neutrophilic_abscess_formation module is ever authored, PG is its flagship conformer.

2. Groupings. PG is a natural member of a Neutrophilic_Dermatoses grouping alongside Sweet syndrome, amicrobial pustulosis of the folds, and the syndromic PG entities — with grouping_basis: [SHARED_MECHANISM, SHARED_PHENOTYPE]. Marzano's three-tier classification (deep/hypodermal → PG; plaque-type/dermal → Sweet; superficial/epidermal; plus syndromic PG as a fourth subset) in PMID:28688013 is the ready-made rationale.

3. Disease-like phenotype. PG carries both HP:0025452 and MONDO:0018824 — exactly the pattern the CLAUDE.md "disease-like phenotypes" module family describes (osteoporosis, glaucoma). Many other disorders will want to annotate PG as a phenotype; this entry is the mechanism they should point at.

4. Claims to curate with supports: REFUTE. (a) Infectious etiology — sterile by definition; (b) solid malignancy as a risk factor — PMID:34076886 found none; (c) stoma relocation as treatment — 67% recurrence, PMID:27473454; (d) IL-1β/TNF-α/IL-17 serum levels as biomarkers — normal in PASH, PMID:25501066.

5. Verify before committing. Every NCIT, CHEBI, GO, CL, and UBERON term marked "verify" above needs just validate-terms. The MONDO xref table came from the OLS4 API and should be re-derived rather than trusted. The Shea 2025 prevalence abstract (PMID:40506010) and the Ortega-Loayza transcriptomics abstracts (PMID:34536481, PMID:28734003) were not retrievable in this session and must be fetched before any snippet from them is used.

Sources

Primary literature (Europe PMC / PubMed): PMID:33033263 · PMID:39718519 · PMID:22534879 · PMID:29450466 · PMID:26071094 · PMID:16188920 · PMID:29721816 · PMID:24903614 · PMID:20636397 · PMID:21658319 · PMID:28688013 · PMID:25501066 · PMID:23571383 · PMID:11971877 · PMID:14595024 · PMID:25601011 · PMID:25350484 · PMID:24487271 · PMID:38951460 · PMID:37516310 · PMID:40034857 · PMID:42603447 · PMID:42123319 · PMID:41785996 · PMID:41923959 · PMID:15533059 · PMID:8609250 · PMID:3889978 · PMID:12409543 · PMID:27502313 · PMID:27473454 · PMID:29288099 · PMID:22959399 · PMID:10807281 · PMID:12171681 · PMID:7912923 · PMID:25589459 · PMID:17966539 · PMID:31560977 · PMID:29334018 · PMID:29438762 · PMID:23903083 · PMID:25374597 · PMID:42472079 · PMID:42107018 · PMID:41638422 · PMID:42310248 · PMID:39720859 · PMID:39927907 · PMID:41255587 · PMID:37610614 · PMID:35606650 · PMID:39098048 · PMID:37909252 · PMID:34536481 · PMID:28734003 · PMID:28943508 · PMID:32634344 · PMID:33647909 · PMID:39118665 · PMID:32481527 · PMID:32613390 · PMID:34076886 · PMID:41379726 · PMID:40012715 · PMID:39963282 · PMID:19494218 · PMID:19302049 · PMID:17655751 · PMID:31092515 · PMID:12907338 · PMID:30971924 · PMID:42517131

Ontology and database resources: - OLS4 / MONDO term MONDO_0018824 - ClinicalTrials.gov API — pyoderma gangrenosum trials - StatPearls: Pyoderma Gangrenosum (NCBI Bookshelf NBK482223) - Europe PMC REST API

Web sources consulted: - Insights into the Pathogenesis of Pyoderma Gangrenosum — J Invest Dermatol - Pyoderma Gangrenosum: An Updated Literature Review — Am J Clin Dermatol - Diagnostic Criteria of Ulcerative Pyoderma Gangrenosum — JAMA Dermatol - Prevalence of Pyoderma Gangrenosum: Systematic Review and Meta-Regression — J Invest Dermatol (abstract not retrievable; do not curate from this until fetched) - Systemic associations of pyoderma gangrenosum: a systematic review — Skin Health Dis - Genetic mutations in pyoderma gangrenosum, hidradenitis suppurativa, and associated autoinflammatory syndromes — PMC - IL-12/IL-23 blockade reveals patterns of asynchronous inflammation in pyoderma gangrenosum (bioRxiv preprint) - Exploratory Study of IFX-1 in Patients With Pyoderma Gangrenosum — NCT03971643

Reference Validation

Checked with linkml-reference-validator 0.2.1.

All extracted references resolved successfully.