Clinical Commander

All dossiers
derm.pyoderma-gangrenosum.core.v1

Pyoderma gangrenosum (neutrophilic ulcerative dermatosis)

dermatologyacutesubacuteadultoutpatientacuteinpatient

DERMATOLOGY-framed pyoderma gangrenosum engine — "the great misdiagnosis". Two cardinal framings own this engine: (1) PG is a DIAGNOSIS OF EXCLUSION confirmed by clinical criteria (Maverakis Delphi / PARACELSUS) with NO confirmatory test — biopsy from the ulcer EDGE EXCLUDES mimics (infection/vasculitis/malignancy/vascular); a neutrophilic infiltrate is supportive but non-specific and confirms nothing; (2) PATHERGY — surgical debridement / aggressive surgery enlarges PG catastrophically, and the dominant clinical error is misdiagnosis as infection followed by debridement (the no-debridement rule is encoded as the first regimen step + a contraindication rule + a severity trigger). Evidence refreshed (not merely tagged) 2026-05-18 via PubMed MCP — all 8 PMIDs PubMed-verified this session with DOIs: Maverakis Delphi 29450466 (10.1001/jamadermatol.2017.5980); PARACELSUS 29388188 (10.1111/bjd.16401); STOP GAP 26071094 (10.1136/bmj.h2958); Brooklyn infliximab RCT 16188920 (10.1136/gut.2005.074815); Partridge systematic review 29478243 (10.1111/bjd.16485); Shaigany claims cohort 34714405 (10.1007/s00403-021-02278-z); Flynn postsurgical-PG-as-necrotising-infection 31498172 (10.1097/01.ASW.0000579692.74662.bb); Lytvyn biologic-paradoxical-PG review 35293377 (10.1097/01.ASW.0000820252.96869.8e). PubMed is the source for these guideline/trial facts. RxCUIs validated live against RxNav 2026-05-18 (forward name→cui + reverse cui→RxNorm Name, ingredient-level): clobetasol propionate 21245, triamcinolone acetonide 10761, tacrolimus 42316, prednisone 8640, cyclosporine 3008, methylprednisolone 6902, infliximab 191831, adalimumab 327361, immune globulin/IVIG 1426680 (reverse "immunoglobulin G, human"), mycophenolate mofetil 68149, azathioprine 1256, cyclophosphamide 3002, dapsone 3108, doxycycline 3640. No hand-authored codes. Non-pharmacologic entries (no-debridement pathergy rule, exclusion-workup gate, moist non-adherent wound care, compression, treat-the-associated-disease) carry non_pharm:true and are rxcui-exempt. Schema-blocked: formal PG diagnostic scoring (Maverakis Delphi 1-major-8-minor / PARACELSUS ≥10-point) and PG severity/extent scores are NOT in the clinical-tools-registry as TS calculators — captured narratively in CONTEXT/BRANCHING_WORKUP/RISK_STRATIFICATION and in the companion _design-brief.md. Decision surface is satisfied by the regimen ladder + workup.chronic_pruritus/le_edema/cvi + calc.ckd_epi_2021. Schema-blocked calc tickets surfaced to the design brief. Bayesian linkage (exclusion-differential pre-test priors, LR+/LR− for ≥8 distinguishing findings incl. the infection/cellulitis pivot, ≥4 conditional dependencies incl. biopsy|edge-sampling and antibiotic-non-response|misdiagnosed-infection, T_treat/T_test for empiric immunosuppression vs continued exclusion, ≥4 cross-engine routing edges by engine_id to derm.cellulitis/derm.drug-eruption/rheum.vasculitis/gi.ibd/id.necrotising-fasciitis) is documented in the co-located _design-brief.md + _research-bundle.md; first-class TS LR fields remain schema-blocked (same constraint as the cellulitis/atopic-dermatitis gold templates). Diagnosis-of-exclusion emphasis: biopsy excludes mimics, it does NOT confirm PG. Effect sizes (≥10, exclusion/treatment target): STOP GAP — no difference ciclosporin vs prednisolone, ~47% (28/59 vs 25/53) healed by 6 mo, ~15-20% healed by 6 wk, recurrence ~28-30%, serious infections more common with prednisolone (PMID 26071094); Brooklyn infliximab — 46% (6/13) vs 6% (1/17) improved at wk2 (p=0.025), 69% (20/29) open-label beneficial response, 21% (6/29) remission wk6, 31% non-response (PMID 16188920); Maverakis ≥4/8 minor criteria Sn 86% / Sp 90% (PMID 29450466); PARACELSUS violaceous border in 98%, ≥10 pts discriminates PG from venous leg ulcers (PMID 29388188); Shaigany — only 9.4% biopsy / 8% tissue culture / 1.4% both before PG diagnosis, IBD 16.3% comorbidity, systemic corticosteroid 17% (misdiagnosis-rate proxy, PMID 34714405); ~50-70% PG associated systemic-disease prevalence; biologic-paradoxical PG 71.9% rituximab / 21.1% TNF-α (PMID 35293377). Full numerics + PMID anchors in _research-bundle.md.

Entry points (5)

  • symptom
    Rapidly enlarging, exquisitely painful ulcer with a violaceous, undermined, overhanging border (PARACELSUS — violaceous border in 98% of PG; Jockenhöfer Br J Dermatol 2019 PMID 29388188)
    rapidly_progressive_painful_violaceous_ulcer
  • history
    A papule/pustule/vesicle that ulcerated within ~4 days of appearing (Maverakis Delphi minor criterion; JAMA Dermatol 2018 PMID 29450466)
    pustule_ulcerating_within_4_days
  • history
    Ulcer that enlarged after surgical debridement, biopsy, or minor trauma — PATHERGY; commonly misdiagnosed as wound/necrotising infection (PMID 31498172)
    wound_worsening_after_debridement_or_trauma
  • history
    Painful ulcer not responding to antibiotics with sterile cultures — the dominant infective-misdiagnosis entry (Shaigany Arch Dermatol Res 2021 PMID 34714405)
    ulcer_not_responding_to_antibiotics
  • history
    Ulceration around a stoma or at a recent surgical/trauma site (peristomal / postsurgical pathergy PG — not a stoma-care or surgical-site infection problem; PMID 31498172)
    peristomal_or_postsurgical_ulcer

Required inputs (17)

  • ulcer_pain_severityrequired
    symptom • used at ENTRY
    Pain markedly out of proportion to ulcer appearance (>4/10 VAS) is a core PG discriminator vs vascular/infective ulcers and a PARACELSUS minor criterion (Jockenhöfer Br J Dermatol 2019 PMID 29388188)
  • ulcer_border_morphologyrequired
    symptom • used at CONTEXT
    Violaceous, undermined, overhanging, irregular border (present in ~98% of PG) is the single strongest morphologic pointer and anchors the criteria (PARACELSUS PMID 29388188; Maverakis Delphi PMID 29450466)
  • ulcer_progression_raterequired
    symptom • used at CONTEXT
    Rapid expansion over days is a PARACELSUS major criterion; an indolent ulcer argues a vascular/neoplastic mimic (Jockenhöfer Br J Dermatol 2019 PMID 29388188)
  • pathergy_or_trauma_localisationrequired
    history • used at CONTEXT
    Lesion arising/worsening at a site of trauma/surgery/debridement (pathergy) is a Maverakis minor criterion AND the critical management-safety flag (no debridement) (PMID 29450466; PMID 31498172)
  • associated_systemic_diseaserequired
    history • used at CONTEXT
    IBD / inflammatory arthritis / haematologic malignancy or IgA monoclonal gammopathy is present in ~50-70%; a Maverakis minor criterion and drives mandatory associated-disease workup (PMID 29450466; Shaigany PMID 34714405)
  • antibiotic_nonresponse_with_sterile_culturesrequired
    history • used at BRANCHING_WORKUP
    Failure to respond to adequate antibiotics with negative tissue cultures shifts the prior away from infection toward PG — the dominant misdiagnosis pivot (Shaigany Arch Dermatol Res 2021 PMID 34714405)
  • necrotising_infection_red_flagsrequired
    symptom • used at RED_FLAGS
    Crepitus, gas, fulminant sepsis, true rapidly advancing necrotising infection must be excluded BEFORE attributing a non-response to PG — necrotising fasciitis is a surgical emergency (route OUT) (IDSA SSTI; PMID 31498172)
  • systemic_toxicity_extent_of_diseaserequired
    symptom • used at RED_FLAGS
    Fulminant/multifocal PG with systemic toxicity or rapidly progressive disease triggers admission + IV immunosuppression (Partridge Br J Dermatol 2018 PMID 29478243; STOP GAP PMID 26071094)
  • tissue_culture_including_atypicalsrequired
    lab • used at INITIAL_WORKUP
    Tissue culture (bacterial, mycobacterial, fungal) from the ulcer is required to EXCLUDE infection — a Maverakis minor criterion; sterile cultures support PG (PMID 29450466)
  • edge_biopsy_to_exclude_mimicsrequired
    imaging • used at INITIAL_WORKUP
    Biopsy from the ulcer EDGE excludes vasculitis, infection, malignancy (the Maverakis MAJOR criterion is a neutrophilic infiltrate, which is supportive but NON-confirmatory) — biopsy excludes, it does not confirm (PMID 29450466)
  • anca_cryoglobulin_antiphospholipid
    lab • used at BRANCHING_WORKUP
    ANCA / cryoglobulins / antiphospholipid antibodies exclude vasculitic and thrombo-occlusive ulcer mimics on the exclusion pathway (Maverakis Delphi exclusion principle PMID 29450466)
  • cbc_with_differential
    lab • used at INITIAL_WORKUP
    Cytopenias / leukaemic picture screen for an associated haematologic malignancy (worse prognosis) and provide a pre-immunosuppression baseline (Shaigany Arch Dermatol Res 2021 PMID 34714405)
  • spep_serum_immunofixation
    lab • used at BRANCHING_WORKUP
    SPEP / immunofixation screens for IgA monoclonal gammopathy and myeloma — a recognised PG association with prognostic weight (Partridge Br J Dermatol 2018 PMID 29478243)
  • lft
    lab • used at INITIAL_WORKUP
    Baseline + on-treatment hepatotoxicity monitoring for ciclosporin / azathioprine / methotrexate-class agents (STOP GAP PMID 26071094; Partridge PMID 29478243)
  • creatinine
    lab • used at TREATMENT
    Ciclosporin nephrotoxicity surveillance + immunosuppressant renal dose-adjust; CKD-EPI 2021 race-free eGFR (STOP GAP PMID 26071094; Inker NEJM 2021)
  • infection_screen_tb_hbv_hcv
    lab • used at INITIAL_WORKUP
    Latent-TB + hepatitis screen before infliximab / adalimumab or other TNF-class biologic initiation (Brooklyn Gut 2006 PMID 16188920; Partridge PMID 29478243)
  • pregnancy_status
    demographic • used at TREATMENT
    Mycophenolate / cyclophosphamide / methotrexate are teratogenic and contraindicated; prednisone or ciclosporin preferred if systemic therapy needed in pregnancy (Partridge Br J Dermatol 2018 PMID 29478243)

12-phase flow (12)

  1. 1FRAME
    Frame PG as a rapidly progressive, exquisitely painful neutrophilic ulcerative dermatosis that is (a) a DIAGNOSIS OF EXCLUSION confirmed by clinical criteria (Maverakis Delphi / PARACELSUS) with NO confirmatory test — biopsy EXCLUDES mimics, it does not confirm — and (b) governed by PATHERGY: surgical debridement / aggressive surgery enlarges it catastrophically. The dominant error is misdiagnosis as infection and surgical debridement. ~50-70% have an associated systemic disease.
    advance: exclusion-diagnosis + criteria framing set; pathergy / no-debridement rule flagged; associated-disease mandate noted
  2. 2ENTRY
    Recognise the rapidly enlarging, disproportionately painful violaceous ulcer (with or without a preceding pustule, pathergy, antibiotic non-response, or peristomal/postsurgical context); capture pain-out-of-proportion up front as the leading discriminator.
    inputs: ulcer_pain_severity
    actions: workup.chronic_pruritus
    advance: entry trigger present; disproportionate pain recorded
  3. 3CONTEXT
    Build the criteria + safety context: violaceous undermined border morphology, rapid progression rate, pathergy/trauma localisation (and the explicit no-debridement flag), and a structured associated-systemic-disease inventory (IBD, inflammatory arthritis, haematologic malignancy / IgA gammopathy). Anchor against Maverakis Delphi minor criteria and the PARACELSUS items.
    inputs: ulcer_border_morphology, ulcer_progression_rate, pathergy_or_trauma_localisation, associated_systemic_disease
    actions: workup.le_edema
    advance: criteria features tallied; pathergy/no-debridement flagged; associated-disease context captured
  4. 4RED_FLAGS
    Before attributing a non-healing ulcer to PG, EXCLUDE a true necrotising soft-tissue infection (crepitus, gas, fulminant sepsis, genuinely rapidly advancing necrosis) — that is a surgical emergency and routes OUT (id.necrotising-fasciitis.core.v1). Conversely, recognise fulminant / multifocal / systemically toxic PG that mandates admission + IV immunosuppression. The trap: a misdiagnosed PG sent to surgical debridement enlarges (pathergy) — STOP surgery, immunosuppress.
    inputs: necrotising_infection_red_flags, systemic_toxicity_extent_of_disease
    actions: panel.cbc, panel.inflammation
    advance: true necrotising infection excluded/routed; fulminant PG recognised + escalated; pathergy-driven enlargement recognised and surgery halted
  5. 5INITIAL_WORKUP
    PG has NO confirmatory test. Workup is for EXCLUSION + associated disease: tissue culture (bacterial/mycobacterial/fungal) to exclude infection; biopsy from the ulcer EDGE to exclude vasculitis/infection/malignancy (a neutrophilic infiltrate is the Maverakis MAJOR criterion but is supportive, NOT diagnostic); CBC; LFT/renal pre-immunosuppression baseline; latent-TB + HBV/HCV if a biologic is anticipated. Edge biopsy is taken cautiously given pathergy.
    inputs: tissue_culture_including_atypicals, edge_biopsy_to_exclude_mimics, cbc_with_differential, lft, infection_screen_tb_hbv_hcv
    actions: panel.cbc, panel.lft, panel.renal
    advance: infection-exclusion cultures + edge biopsy sent; pre-immunosuppression safety labs drawn
  6. 6BRANCHING_WORKUP
    Exclusion decision tree: sterile cultures + antibiotic non-response → away from infection; vascular studies (ABI/venous duplex) + workup.le_edema/workup.cvi → exclude arterial/venous ulcer; ANCA / cryoglobulins / antiphospholipid → exclude vasculitis and thrombo-occlusive mimics; biopsy-edge histology → exclude malignancy (SCC/Marjolin) and infection; and the MANDATORY associated-disease workup — colonoscopy/IBD evaluation, SPEP/immunofixation, haematology, inflammatory-arthritis screen. If all mimics excluded and ≥4/8 Maverakis minor criteria met → PG.
    inputs: antibiotic_nonresponse_with_sterile_cultures, anca_cryoglobulin_antiphospholipid, spep_serum_immunofixation
    actions: workup.le_edema
    advance: mimics excluded; criteria threshold met OR an alternative diagnosis assigned + routed; associated-disease workup initiated
  7. 7DIFFERENTIAL
    Terminal exclusion differential with named pivots: PG vs infection/cellulitis/necrotising STI (sterile culture + antibiotic non-response + pathergy pivot — route derm.cellulitis.core.v1 / id.necrotising-fasciitis.core.v1) vs venous ulcer (CEAP/stasis pivot — workup.cvi) vs arterial ulcer (ABI pivot) vs vasculitis (ANCA + biopsy pivot — route rheum.vasculitis.core.v1) vs antiphospholipid/calciphylaxis (thrombo-occlusive + calcium-phosphate/biopsy pivot) vs malignancy/Marjolin (chronicity + edge-biopsy pivot) vs factitial ulcer (geometric/access pivot) vs Sweet syndrome (overlapping neutrophilic, plaques>ulcer pivot) vs ecthyma gangrenosum (Pseudomonas/neutropenia pivot) vs cutaneous Crohn (perianal/biopsy granuloma pivot) vs brown-recluse/loxoscelism (bite history pivot) vs Martorell HYTILU (hypertensive ischaemic supramalleolar pivot). PG remains a diagnosis of exclusion.
    advance: mimics actively excluded; PG diagnosed by criteria OR alternative assigned + routed by engine_id
  8. 8RISK_STRATIFICATION
    Stratify by extent/tempo: localised/mild (single, limited, slowly progressive) → topical/intralesional ladder; moderate-severe or rapidly progressive/multifocal → systemic corticosteroid or ciclosporin first-line; fulminant / systemically toxic / pathergy-driven enlargement after misdirected surgery → admit + IV immunosuppression + multidisciplinary. Layer the prognostic modifier: associated haematologic malignancy / IgA gammopathy → worse prognosis, heme/onc engagement. (PG severity scores are schema-blocked TS calculators; captured narratively.)
    inputs: systemic_toxicity_extent_of_disease, ulcer_progression_rate
    advance: localised vs moderate-severe vs fulminant tier + prognostic-modifier overlay assigned
  9. 9TREATMENT
    CARDINAL RULE: NO surgical debridement / aggressive surgery (pathergy → catastrophic enlargement); manage peristomal/postsurgical PG conservatively with immunosuppression; graft/essential surgery only under immunosuppressive cover. Ladder by extent: localised → ultrapotent topical corticosteroid (clobetasol) or intralesional triamcinolone, topical tacrolimus + meticulous moist non-adherent wound care + analgesia + compression if leg oedema (non-pharm). Moderate-severe / rapidly progressive → systemic prednisone 0.5-1 mg/kg OR ciclosporin (STOP GAP — comparable efficacy, choose on toxicity); pulse methylprednisolone for fulminant. Refractory / steroid-sparing → infliximab (RCT-supported, esp. IBD-associated), adalimumab, IVIG, mycophenolate, azathioprine, cyclophosphamide. TREAT THE ASSOCIATED DISEASE (IBD/arthritis/heme drives the skin). Gate on pregnancy, renal/hepatic function, TNF infection screen.
    inputs: creatinine, pregnancy_status
    advance: no-debridement rule enforced; extent-appropriate immunosuppression + wound care + analgesia started; associated disease addressed; agent gated on comorbidity/pregnancy
  10. 10DISPOSITION
    Outpatient for localised disease (criteria-based diagnosis, exclusion + associated-disease workup, topical/intralesional ± oral immunosuppression, dermatology follow-up). Admit for rapidly progressive / fulminant / systemically toxic PG, IV immunosuppression, severe uncontrolled pain, or a pathergy-driven enlargement after misdirected surgery requiring multidisciplinary stabilisation. Route associated-disease ownership OUT by engine_id (GI/heme/rheum) while retaining skin + immunosuppression ownership.
    inputs: systemic_toxicity_extent_of_disease
    advance: disposition documented; admission only for fulminant/toxic/uncontrolled-pain criteria; specialty co-management arranged
  11. 11MONITORING
    Disease: ulcer size + pain trajectory + healing (response within ~1 month of starting immunosuppression is itself a Maverakis minor criterion); STOP GAP benchmark ~47% healed by 6 months on prednisolone OR ciclosporin. Drug safety: prednisone (glucose/BP/bone), ciclosporin (BP + creatinine, CKD-EPI), azathioprine/MMF (CBC/LFT), biologics (infection/TB reactivation). Taper guided by response with transition to a steroid-sparing agent. Track associated-disease activity (IBD/arthritis/heme) as it drives the skin.
    inputs: creatinine, lft
    actions: panel.cbc, panel.lft, panel.renal
    advance: ulcer/pain trajectory assessed at the agent-appropriate interval; drug-class safety labs on schedule; associated-disease activity tracked
  12. 12FOLLOWUP
    Chronic surveillance: continued meticulous moist non-adherent wound care + analgesia, slow steroid taper with steroid-sparing maintenance, recurrence vigilance (PG recurs in ~28-30% — STOP GAP), pathergy education (avoid elective surgery/debridement; flag PG to any future surgeon; perioperative immunosuppressive cover if surgery unavoidable), cribriform-scar counselling, and ongoing associated-disease monitoring (repeat heme screen if new cytopenia; IBD/arthritis activity). Dermatology continuity for any systemic agent.
    inputs: associated_systemic_disease, pathergy_or_trauma_localisation
    actions: workup.chronic_pruritus
    advance: wound-care + taper + pathergy education + recurrence + associated-disease surveillance plan documented