Clinical Commander

All dossiers
pulm.aspiration-pneumonia.core.v1

Aspiration Pneumonia

pulmonologyacutesubacuteadultacuteinpatientoutpatient

THE PIVOT (§5.5.2): pneumonitis-vs-pneumonia is encoded as a self-row in sibling_differentiation (8 discriminators with PMIDs: mechanism, onset-relative-to-event, fever curve, witnessed event, infiltrate behaviour, response-without-antibiotics, ARDS potential, antibiotic decision), the FRAME phase purpose/advance_when, the pneumonitis_observe regimen step, and 2 severity_triggers (pneumonitis_resolving_stop_antibiotics; massive_acid_aspiration_ards_mendelson). Classify first, then deprescribe if pneumonitis. GUIDELINE UPDATE encoded: routine anaerobic coverage is NO LONGER recommended for community aspiration pneumonia (Metlay ATS/IDSA 2019 PMID 31573350 conditional rec AGAINST unless lung abscess/empyema); modern oral-flora data (Bartlett 2013 PMID 23398871; Mandell NEJM 2019 PMID 30763196) — header docblock, regimen step rationales, contraindication_rules, and CAP sibling all rewritten to reflect this. Anaerobic agents (metronidazole/clindamycin) are role:add_on gated on abscess/empyema/necrotizing/severe-periodontal triggers ONLY. Effect-size / epidemiology numbers with verified PMIDs (≥5): dysphagia screening in stroke pooled pneumonia OR 0.57 (95% CI 0.45-0.72) and mortality OR 0.52 (0.35-0.77) (Sherman/Martino JAHA 2021 PMID 34096328); ED dysphagia screen reduced HAP in ischemic stroke 13.8%→8% (p=0.007) and hemorrhagic 19%→15% (p<0.001) (Schrock Am J Emerg Med 2018 PMID 29685361); chemical pneumonitis self-limited resolving 24-48 h without antibiotics, ~50% of severe acid aspiration progresses to ARDS in classic Mendelson series (Marik NEJM 2001 PMID 11228282); anaerobes recovered in <10-15% of modern aspiration pneumonia and almost never without abscess/empyema (Bartlett 2013 PMID 23398871); lung-abscess medical cure with appropriate antibiotics ~80-90% over 3+ weeks (Bartlett 2013 PMID 23398871; Mandell 2019 PMID 30763196). Cross-dossier routing via workups[].branches_to + sibling_differentiation[].sibling_engine_id using real on-disk engine_ids: pulm.cap.core.v1, pulm.hospital-acquired-pneumonia-non-covid.v1, pulm.ards.core.v1 (Mendelson→ARDS), id.sepsis.core.v1, pulm.pe.core.v1 (look-alike), neuro.ischaemic-stroke.v1 (post-stroke dysphagia secondary prevention). Special populations covered in inpatient playbook non_drug_actions + triggers: post-stroke/neuro dysphagia (bedside screen before oral intake; route neuro.ischaemic-stroke.v1), elderly/dementia (hand-feeding > PEG — PEG does not reduce aspiration/mortality), tube-fed (position, post-pyloric, HOB), intubated/peri-procedural (NPO, cuff, post-extubation screen), alcohol/overdose (airway protection during reduced consciousness), GERD/achalasia (reflux precautions, treat motility), palliative (goals-of-care for recurrent aspiration; PEG does not prevent recurrence). Regimen axis: when-to-treat fork (pneumonitis observe/deprescribe) → community outpatient amox-clav → ward amp-sulbactam → hospital-acquired/severe pip-tazo ± MRSA, all WITHOUT a routinely added anaerobic agent (metronidazole/clindamycin add_on gated on abscess/empyema/necrotizing/severe-periodontal). Renal/hepatic adjust, de-escalation at 48-72 h, duration 5-7 d uncomplicated / 3 wk+ abscess, stop-if-pneumonitis deprescribing — matches pulm.pe.core.v1 regimen depth. RXCUI CORRECTIONS (RxNav REST verified 2026-05-16): amoxicillin-clavulanate 723→19711 (was amoxicillin); ampicillin-sulbactam 733→1009148 (was ampicillin); metronidazole 6851→6922 (was METHOTREXATE — serious error); piperacillin-tazobactam 18631→74169 (was AZITHROMYCIN — serious error); meropenem 74169→29561 (was mislabeled pip-tazo). Confirmed-correct unchanged: ceftriaxone 2193, clindamycin 2582, cefepime 20481, vancomycin 11124, linezolid 190376, moxifloxacin 139462, oxygen 7806. SCHEMA-GAP NOTES: (1) _types.ts has no first-class field for a pneumonitis-vs-pneumonia discriminator table — encoded via a sibling_differentiation self-row (sibling_engine_id === engine_id), phase purpose/advance_when, severity_triggers, and the .depth.md table; (2) RequiredWorkup.branches_to is the only cross-engine routing field — used for the 7-engine branch list; (3) no notes/rationale/contraindication fields on RegimenDrug beyond rationale — anaerobic-restriction logic encoded in drug.triggers + rationale + axis contraindication_rules; (4) PMID provenance lives only in evidence.pmids + inline comments + rationale strings (no per-claim citation field). PRODUCTION blockers: (1) speech-swallow evaluation tools (FEES/MBS scoring) not in clinical-tools-registry; (2) no engine-specific test file under tests/ (uses shared dossier-contract.test.ts); (3) terminology codes not yet RxNav/terminology-service validated. Legacy WRONG PMIDs (11297706 complement, 30699320 neuroscience, 11876581 cocaine-tx) removed and replaced with verified anchors.

Entry points (4)

  • symptom
    Witnessed aspiration / choking event
    witnessed_aspiration
  • symptom
    Cough, fever, dyspnea after dysphagia/altered LOC
    cough_fever_dyspnea
  • imaging
    Dependent-lobe infiltrate (RLL, posterior segments) on CXR/CT
    dependent_lobe_infiltrate
  • problem_list
    High-aspiration-risk substrate — stroke, dementia, ALS, esophageal dysmotility, NG tube, intoxication
    dysphagia_high_risk

Required inputs (15)

  • agerequired
    demographic • used at CONTEXT
    Age stratifies severity and antibiotic dosing
  • spo2required
    vital • used at CONTEXT
    Hypoxemia and severity grading
  • temprequired
    vital • used at CONTEXT
    Bacterial superinfection vs chemical pneumonitis time course
  • rrrequired
    vital • used at CONTEXT
    Tachypnea + work-of-breathing severity
  • sbprequired
    vital • used at CONTEXT
    Septic shock screening; CURB-65 element
  • aspiration_event
    history • used at CONTEXT
    Witnessed acid/gastric aspiration event timestamp anchors the pneumonitis-vs-pneumonia clock (Marik NEJM 2001 PMID 11228282)
  • symptom_onset_relative_to_event
    history • used at FRAME
    Hours-after-event abrupt + improving = pneumonitis; >24-48 h with persistent fever/leukocytosis = bacterial pneumonia (Marik 2001 PMID 11228282; Mandell NEJM 2019 PMID 30763196)
  • dysphagia_or_altered_los
    history • used at CONTEXT
    Risk substrate (stroke, dementia, intoxication, recent extubation, NG/PEG tube) — drives secondary prevention (Mandell 2019 PMID 30763196)
  • periodontal_disease_or_abscess
    history • used at CONTEXT
    Severe periodontal disease/putrid sputum/abscess is now the ONLY routine indication for anaerobic coverage (Metlay ATS/IDSA 2019 PMID 31573350; Bartlett 2013 PMID 23398871)
  • recent_hospitalization
    history • used at CONTEXT
    Hospital-acquired aspiration → broaden coverage to gram-negatives + Pseudomonas
  • cxrrequired
    imaging • used at INITIAL_WORKUP
    Dependent-lobe infiltrate confirms pneumonia vs uncomplicated pneumonitis
  • ct_chest
    imaging • used at BRANCHING_WORKUP
    Lung abscess, empyema, foreign body aspiration, cavitation
  • wbc
    lab • used at INITIAL_WORKUP
    Leukocytosis/bandemia for bacterial pneumonia
  • creatininerequired
    lab • used at INITIAL_WORKUP
    eGFR for renal-dosed antibiotics
  • lactate
    lab • used at INITIAL_WORKUP
    Sepsis severity

12-phase flow (12)

  1. 1FRAME
    THE PIVOT — classify aspiration PNEUMONITIS (chemical/Mendelson: abrupt onset hours after a witnessed acid/gastric event, infiltrate evolving and improving, fever early and resolving, RESPONDS WITHOUT ANTIBIOTICS in 24-48 h; severe acid load → ARDS) vs aspiration PNEUMONIA (bacterial: subacute >24-48 h, persistent/late fever + leukocytosis, fixed dependent-segment infiltrate, REQUIRES ANTIBIOTICS) (Marik NEJM 2001 PMID 11228282; Mandell & Niederman NEJM 2019 PMID 30763196)
    inputs: symptom_onset_relative_to_event
    advance: Pneumonitis-vs-pneumonia working classification assigned with timing/fever-curve rationale
  2. 2ENTRY
    Triggered by witnessed aspiration/choking, dependent-segment infiltrate (RLL, posterior segments, RUL-posterior if supine), or sepsis in a high-aspiration-risk substrate patient (Mandell 2019 PMID 30763196)
    inputs: age
    advance: Aspiration risk substrate or imaging finding present
  3. 3CONTEXT
    Capture risk substrate (dysphagia, stroke, dementia, ALS, intoxication/overdose, NG/PEG tube, recent extubation/peri-procedural, GERD/achalasia), oral health (severe periodontal disease — now the key anaerobic-coverage trigger), prior antibiotics, recent hospitalisation (≥48 h → hospital-acquired pattern)
    inputs: spo2, temp, rr, sbp, aspiration_event, dysphagia_or_altered_los, periodontal_disease_or_abscess, recent_hospitalization
    advance: Substrate, oral health, hospital-acquired status, and severity captured
  4. 4RED_FLAGS
    Severe sepsis/septic shock, hypoxemic respiratory failure, ARDS from massive acid aspiration (Mendelson), lung abscess, empyema, large-volume vomiting with ongoing aspiration risk, foreign-body aspiration with airway obstruction
    inputs: spo2, sbp
    actions: calc.qsofa
    advance: No emergent escalation needed OR sepsis bundle / airway protection initiated
  5. 5INITIAL_WORKUP
    CXR (dependent-segment pattern — RLL/posterior segments upright, RUL-posterior/superior-RLL supine); blood cultures + sputum Gram stain/culture if pneumonia suspected; lactate, CBC, creatinine. KEY: an evolving/improving infiltrate that clears in 24-48 h with NO new fever supports pneumonitis (observe, no abx); a fixed/progressing infiltrate with persistent fever supports bacterial pneumonia (Marik 2001 PMID 11228282; Mandell 2019 PMID 30763196)
    inputs: cxr, wbc, creatinine, lactate
    actions: aspiration_pneumonia
    advance: Imaging + cultures obtained and pneumonitis-vs-pneumonia trajectory clarified
  6. 6BRANCHING_WORKUP
    CT chest if poor response at 48-72 h or abscess/necrotizing/empyema concern (cavitation, air-fluid level, loculated pleural collection); bronchoscopy if foreign-body aspiration or post-obstructive pneumonia (rule out endobronchial malignancy in non-resolving); modified barium swallow / FEES once acute illness settles
    inputs: ct_chest
    advance: Complications and mimics excluded or characterised
  7. 7DIFFERENTIAL
    Aspiration pneumonitis vs aspiration pneumonia (THE pivot) vs CAP (no aspiration substrate) vs HAP/VAP (≥48 h hospitalised → pulm.hospital-acquired-pneumonia-non-covid.v1) vs ARDS (Mendelson can cause it → pulm.ards.core.v1) vs PE (pleuritic, hypoxia disproportionate to infiltrate → pulm.pe.core.v1) vs lung abscess/empyema vs sepsis from another source (id.sepsis.core.v1)
    advance: Working diagnosis identified with discriminators documented
  8. 8RISK_STRATIFICATION
    CURB-65 / PSI for severity + disposition; decide anaerobic-coverage (ONLY if lung abscess, empyema, necrotizing pneumonia, or severe periodontal disease with putrid sputum — NOT routine, per Metlay ATS/IDSA 2019 PMID 31573350); decide community vs hospital-acquired empiric spectrum
    actions: calc.curb65
    advance: Severity score + anaerobic-coverage decision + community-vs-hospital spectrum documented
  9. 9TREATMENT
    PNEUMONITIS (chemical, witnessed event, infiltrate clearing, no late fever/leukocytosis) → SUPPORTIVE ONLY, withhold/stop antibiotics, reassess 48 h. PNEUMONIA → empiric antibiotics targeting oral aerobes / S. aureus / enteric Gram-negatives (community: ampicillin-sulbactam or amoxicillin-clavulanate; hospital-acquired/severe: pip-tazo ± MRSA) WITHOUT routine added anaerobic agent — add metronidazole/clindamycin ONLY for abscess/empyema/necrotizing/severe periodontal disease (Metlay 2019 PMID 31573350; Mandell 2019 PMID 30763196). Renal/hepatic dose adjust; speech-swallow consult + NPO until assessed
    inputs: creatinine
    actions: calc.ckd_epi_2021
    advance: Treat-vs-observe decision made; if treating, antibiotic + duration + anaerobic decision + dysphagia consult documented
  10. 10DISPOSITION
    Pneumonitis improving → home with swallow follow-up; mild community pneumonia CURB-65 0-1 → outpatient; CURB-65 2 / hypoxia → ward; CURB-65 ≥3 / septic shock / hypoxemic respiratory failure / ARDS → ICU
    advance: Disposition decided
  11. 11MONITORING
    Clinical response at 48-72 h (de-escalate / stop antibiotics if pneumonitis confirmed by resolution); repeat imaging only if non-resolving (then CT for abscess/empyema); swallow re-evaluation pre-discharge
    advance: Response trend documented; antibiotics stopped if pneumonitis, narrowed if pneumonia; aspiration-prevention plan in place
  12. 12FOLLOWUP
    Secondary prevention bundle — speech-language pathology swallow rehab, diet/texture modification, oral-hygiene programme, head-of-bed elevation, sedative/anticholinergic deprescribing, vaccination, goals-of-care discussion for recurrent aspiration (PEG does NOT eliminate aspiration risk) (Mandell 2019 PMID 30763196; Sherman/Martino JAHA 2021 PMID 34096328)
    advance: Aspiration-prevention + secondary-prevention bundle in place