Peritonsillar abscess (quinsy)
EM/ENT-framed engine for the peritonsillar cellulitis → abscess continuum. Airway-emergency management, deep-neck-space surgical source control, the sepsis bundle, and the EBV-mononucleosis pathway are recognised then routed OUT by engine_id (ent.epiglottitis.core.v1, ent.retropharyngeal-abscess.core.v1, id.sepsis.core.v1, ent.pharyngitis.core.v1) — not re-authored here. RxCUIs validated live against RxNav 2026-05-17: ampicillin/sulbactam 1009148, amoxicillin/clavulanate 19711, clindamycin 2582, metronidazole 6922, penicillin G 7980, dexamethasone 3264. NOTE: the build-prompt-supplied rxcui 1659131 for ampicillin-sulbactam was found on live RxNav to resolve to "piperacillin 2000 MG / tazobactam 250 MG Injection" — the correct ampicillin/sulbactam ingredient RxCUI 1009148 (RxNorm Name "ampicillin / sulbactam") was substituted; no fabricated codes. No single society "peritonsillar abscess guideline" exists; the operating authority set (Cochrane Chang 2016, Kim Acad Emerg Med 2023 ultrasound meta-analysis, Hur Laryngoscope 2018 steroid review, AAO-HNS Tonsillectomy 2019, Klug Fusobacterium microbiology, Wright/Tiwari Lemierre) was WebSearch/PubMed-reconciled 2026-05-17 and confirmed current — no superseding Cochrane PTA update or society PTA guideline as of the retrieval date. Bayesian linkage (peritonsillar-cellulitis-vs-abscess pre-test priors by exam; LR+/LR− for uvular deviation, trismus, fluctuance, POCUS; decision thresholds aspiration-vs-CT-vs-OR; conditional dependencies; cross-dossier routing edges by engine_id) is documented in the co-located _design-brief.md + _research-bundle.md; first-class TS LR fields remain schema-blocked (same constraint as derm.cellulitis.core.v1). Effect sizes (≥5): I&D-vs-aspiration recurrence RR 3.74 (95% CI 1.63-8.59, Chang Cochrane 2016); POCUS for PTA pooled sensitivity 86% (95% CI 78-91), specificity 76% (95% CI 67-82), LR+ 3.51, LR− 0.19 (Kim Acad Emerg Med 2023); clinical exam alone ~75% sensitive / ~50% specific (Kim 2023); aspiration repeat rate ~46% vs I&D ~10% and median LOS 3.0 vs 2.0 days (Mansour Eur Arch Otorhinolaryngol 2019); post-I&D bleeding ~3.6% (Chang Cochrane 2016); descending mediastinitis untreated mortality up to 85% with CRP 340.9 vs 190.1 mg/L mediastinal vs non-mediastinal (Hu Medicina 2022); GAS recovered in ~20% / Fusobacterium ~23-58% of PTA (Klug Dan Med J 2017).
Entry points (5)
- symptomSevere unilateral sore throat with trismus, dysphagia, and odynophagia worsening over 2-5 days (classic peritonsillar abscess presentation — Klug Dan Med J 2017; Kim Acad Emerg Med 2023)severe_unilateral_sore_throat_with_trismus
- symptom"Hot-potato"/muffled voice with uvular deviation, soft-palate fullness, and contralateral tonsillar displacement (cardinal abscess sign — Chang Cochrane 2016)hot_potato_muffled_voice_uvular_deviation
- symptomDrooling / pooled secretions with trismus and inability to tolerate oral secretions — airway + dehydration entry (Klug Dan Med J 2017)drooling_pooling_secretions_trismus
- history≥1 prior peritonsillar abscess or recurrent tonsillitis — recurrence / quinsy-tonsillectomy entry (AAO-HNS Tonsillectomy 2019 — >1 PTA is a tonsillectomy modifying factor)recurrent_peritonsillar_abscess_or_tonsillitis
- symptomAcute tonsillitis on antibiotics now developing unilateral peritonsillar bulge / worsening trismus — cellulitis→abscess progression (Klug Dan Med J 2017)partially_treated_tonsillitis_not_improving
Required inputs (15)
- sore_throat_lateralityrequiredsymptom • used at ENTRYUnilateral, asymmetric throat pain with referred otalgia favours peritonsillar abscess over symmetric pharyngitis/tonsillitis (Kim Acad Emerg Med 2023 — clinical exam ~75% sensitive)
- trismus_presentrequiredsymptom • used at CONTEXTTrismus from medial-pterygoid irritation is a key abscess discriminator vs simple tonsillitis and worsens drainage access (Klug Dan Med J 2017)
- uvular_deviation_soft_palate_fullnessrequiredsymptom • used at CONTEXTUvular deviation away from the lesion + soft-palate fullness raises the abscess (vs cellulitis) post-test probability (Kim Acad Emerg Med 2023)
- voice_quality_hot_potatorequiredsymptom • used at CONTEXTMuffled "hot-potato" voice localises a supratonsillar/oropharyngeal mass; stridor/quiet voice instead pivots toward epiglottitis (Klug Dan Med J 2017)
- airway_compromise_featuresrequiredsymptom • used at RED_FLAGSStridor, tripod posture, air hunger, severe drooling with cyanosis → airway emergency; recognise then route to ent.epiglottitis.core.v1 (Klug Dan Med J 2017 — airway obstruction is a PTA complication)
- neck_swelling_torticollis_extensionrequiredsymptom • used at RED_FLAGSLateral neck swelling, torticollis, or trismus out of proportion suggests parapharyngeal/retropharyngeal spread → route to ent.retropharyngeal-abscess.core.v1 (Hu Medicina 2022 — mediastinitis precursor)
- temperaturerequiredvital • used at CONTEXTFever supports a suppurative process and severity class; high fever + rigors raises the bacteraemia/Lemierre concern (Klug Dan Med J 2017)
- sbprequiredvital • used at RED_FLAGSHypotension with a deep-neck source → systemic toxicity; recognise then route to id.sepsis.core.v1 (SSC — sepsis pathway not authored here)
- immunocompromisehistory • used at CONTEXTDiabetes / immunosuppression broadens pathogen spectrum, lowers drainage/CT and admission threshold, and raises complication risk (Gawel Int Arch Otorhinolaryngol 2025 — PTA tonsillectomy patients have higher diabetes/sepsis rates)
- anticoagulation_or_bleeding_diathesishistory • used at TREATMENTAnticoagulation / bleeding diathesis changes the drainage risk-benefit (post-procedure bleeding) and may favour aspiration over incision (Chang Cochrane 2016)
- pregnancyhistory • used at TREATMENTAntibiotic-safety gating for the regimen (penicillin-class/clindamycin preferred; metronidazole considerations) (WebSearch guideline review 2026-05-17)
- penicillin_allergyrequiredhistory • used at TREATMENTSevere penicillin allergy switches first-line β-lactam/β-lactamase-inhibitor to clindamycin (anaerobe + GAS cover) (Klug Dan Med J 2017 — clindamycin preferred over macrolide given Fusobacterium)
- wbclab • used at INITIAL_WORKUPLeukocytosis supports a suppurative process and feeds the severity / mediastinitis-risk chain (Hu Medicina 2022)
- crplab • used at INITIAL_WORKUPVery high CRP (e.g. >190-340 mg/L) is associated with descending mediastinitis risk in deep-neck infection (Hu Medicina 2022)
- pta_pocus_or_ct_neckimaging • used at BRANCHING_WORKUPIntraoral/transcervical POCUS (or contrast CT if POCUS equivocal or deep-space spread suspected) resolves the cellulitis-vs-abscess question and guides aspiration (Kim Acad Emerg Med 2023 — pooled US sens 86% / spec 76%)
12-phase flow (12)
- 1FRAMEFrame as the peritonsillar cellulitis → abscess continuum: severe unilateral sore throat + trismus + muffled voice + uvular deviation. Distinguish from symmetric pharyngitis/tonsillitis, infectious mononucleosis, epiglottitis, retropharyngeal abscess, and odontogenic deep-space infection. Airway, deep-space surgical source control, sepsis, and EBV-mono pathways are routed OUT, not authored here.advance: PTA scope confirmed; airway/deep-space/sepsis/mono concerns flagged for routing by engine_id
- 2ENTRYRecognise the severe-unilateral-sore-throat-with-trismus, hot-potato-voice/uvular-deviation, drooling, recurrent-PTA, or partially-treated-tonsillitis-not-improving entry; record laterality up front (symmetric throat argues against PTA).inputs: sore_throat_lateralityactions: calc.centoradvance: entry trigger present; laterality and Centor recorded
- 3CONTEXTBuild the cellulitis-vs-abscess pre-test prior: trismus, uvular deviation + soft-palate fullness, hot-potato voice, fever; comorbidity (diabetes/immunocompromise), recurrence history. This phase assigns the pre-POCUS abscess probability.inputs: trismus_present, uvular_deviation_soft_palate_fullness, voice_quality_hot_potato, temperature, immunocompromiseactions: workup.lymphadenopathyadvance: cardinal-feature set + pretest abscess prior assigned
- 4RED_FLAGSRecognise the dangerous extensions: airway compromise (stridor/tripod/air-hunger) → route to ent.epiglottitis.core.v1; parapharyngeal/retropharyngeal spread / torticollis / neck swelling → route to ent.retropharyngeal-abscess.core.v1; descending mediastinitis (very high CRP, chest/back pain) → deep-space + thoracic; Lemierre (rigors, neck-vein tenderness, septic pulmonary emboli); systemic toxicity / qSOFA≥2 / hypotension → route to id.sepsis.core.v1. Recognised here, NOT managed here.inputs: airway_compromise_features, neck_swelling_torticollis_extension, sbpactions: workup.airway_distress, calc.qsofa, calc.news2advance: airway / deep-space / mediastinitis / Lemierre / sepsis screened and routed by engine_id if positive
- 5INITIAL_WORKUPTargeted labs only: CBC + CRP (suppuration + mediastinitis-risk anchor — Hu Medicina 2022); CMP for hydration/renal dosing in the drooling/poor-intake patient; coagulation panel if drainage planned in an anticoagulated/bleeding-diathesis patient; consider Monospot/EBV serology if bilateral tonsillar exudate + posterior cervical nodes + splenomegaly suggest coexisting mononucleosis. Throat/aspirate culture only at drainage (Fusobacterium-aware — Klug Dan Med J 2017).inputs: wbc, crpactions: panel.cbc, panel.inflammation, panel.cmp, panel.coagadvance: baseline labs sent; hydration + bleeding risk assessed
- 6BRANCHING_WORKUPCellulitis-vs-abscess decision tree: clinical exam alone is only ~75% sensitive / ~50% specific (Kim Acad Emerg Med 2023) → intraoral or transcervical POCUS (pooled sens 86% / spec 76%; LR+ 3.5, LR− 0.19) as the discriminator and aspiration guide; contrast CT neck if POCUS equivocal, trismus precludes intraoral exam, or parapharyngeal/retropharyngeal/mediastinal extension suspected. Peritonsillar cellulitis (no drainable collection) → antibiotics + observe, no drainage.inputs: pta_pocus_or_ct_neckactions: workup.cellulitis_necfasc, workup.bacterial_meningitisadvance: abscess confirmed/excluded by POCUS or CT, or empiric cellulitis pathway entered
- 7DIFFERENTIALTerminal differential with pivot findings: peritonsillar abscess vs peritonsillar cellulitis (drainable collection on POCUS/CT pivot) vs severe tonsillitis/pharyngitis (symmetric, no uvular deviation, no trismus pivot) vs infectious mononucleosis (bilateral exudate + posterior cervical nodes + splenomegaly + atypical lymphocytes pivot — route to ent.pharyngitis.core.v1) vs epiglottitis (rapid airway, stridor, minimal oral findings, "thumbprint" pivot — route to ent.epiglottitis.core.v1) vs retropharyngeal abscess (neck stiffness, posterior pharyngeal bulge, CT pivot — route to ent.retropharyngeal-abscess.core.v1) vs odontogenic/dental deep-space infection (molar source, floor-of-mouth pivot).advance: single best diagnosis selected; PTA-on-mononucleosis coexistence flagged
- 8RISK_STRATIFICATIONSeverity → disposition: uncomplicated drainable PTA, tolerating oral intake, no airway/deep-space/sepsis flag → outpatient drainage + oral antibiotics; trismus precluding bedside drainage, dehydration, failed outpatient drainage, immunocompromise, suspected deep-space spread, qSOFA≥2/NEWS2 high → admit / OR. SIRS / qSOFA layered for the systemic-toxicity upgrade.inputs: temperature, sbpactions: calc.qsofa, calc.sirs, calc.news2advance: severity class + disposition track assigned
- 9TREATMENTSource control + antimicrobial + adjunct: (1) drain a confirmed abscess — needle aspiration OR incision-and-drainage (Chang Cochrane 2016 — I&D lower recurrence RR 3.74 favouring I&D; aspiration less painful); quinsy/à-chaud tonsillectomy if recurrent/failed drainage/concurrent tonsillectomy indication (Rosi-Schumacher Int J Pediatr Otorhinolaryngol 2023 — equivalent outcomes); (2) anaerobe + GAS-cover antibiotic, Fusobacterium-aware (ampicillin/sulbactam IV or amoxicillin/clavulanate PO; clindamycin if penicillin-allergic; add/substitute metronidazole for anaerobe cover; penicillin G for confirmed GAS); (3) adjunct single-dose IV dexamethasone (Hur Laryngoscope 2018 — faster fever/pain/intake recovery); (4) hydration + analgesia. Antibiotic-safety gating by pregnancy/allergy/renal.inputs: penicillin_allergy, anticoagulation_or_bleeding_diathesis, pregnancyactions: protocol.septic_shockadvance: abscess drained (or cellulitis treated medically); antibiotic + steroid + hydration started; ENT engaged
- 10DISPOSITIONOutpatient: successful bedside drainage, tolerating oral intake/analgesia, no red flag, reliable follow-up → discharge on oral antibiotics + ENT review. Admit: trismus precluding drainage, dehydration/can't tolerate PO, immunocompromise, airway/deep-space/sepsis flag, failed outpatient drainage, age extremes/unreliable follow-up. Airway/deep-space/sepsis → admit and route OUT by engine_id.inputs: temperature, sbpadvance: disposition documented; route-out completed if airway/deep-space/sepsis positive
- 11MONITORINGExpect symptomatic improvement within 24-48 h of adequate drainage + antibiotics; persistent fever, trismus, or pain at 24-48 h → re-image for re-accumulation / parapharyngeal-retropharyngeal extension / wrong-spectrum, do not silently extend antibiotics. Watch for delayed Lemierre (rigors, pleuritic chest pain, neck-vein tenderness) and post-drainage bleeding (Chang Cochrane 2016 — post-I&D bleed ~3.6%).inputs: crp, wbcactions: panel.inflammationadvance: objective improvement by 24-48 h, OR re-evaluation/re-imaging triggered
- 12FOLLOWUPSmoking-cessation counselling (smoking is an independent PTA risk factor — Klug Dan Med J 2017); ENT follow-up; recurrence counselling. Interval/quinsy tonsillectomy candidacy if ≥1 prior PTA or recurrent tonsillitis meeting AAO-HNS modifying-factor criteria (AAO-HNS Tonsillectomy 2019; Rosi-Schumacher 2023 — quinsy and interval tonsillectomy equivalent outcomes). Counsel return precautions for re-accumulation, airway, and Lemierre.inputs: recurrent_peritonsillar_abscess_or_tonsillitisadvance: follow-up + tonsillectomy-candidacy decision + return precautions documented