Hospital Medicine Unplugged

In this episode of Hospital Medicine Unplugged, we unpack fever of unknown origin (FUO) in hospitalized adults—define precisely, chase clues relentlessly, use PET-CT wisely, avoid reflex antibiotics.
We open with the definitions that matter: classic FUO (≥38.3°C on several occasions, ≥3 weeks, no diagnosis after appropriate evaluation) and nosocomial FUO (fever >48 hrs into hospitalization with nondiagnostic initial workup). Debate continues, but “reasonable diagnostic evaluation” beats rigid clocks when clinical sense demands it.
Patterns you’ll actually see: atypical presentations of common problems—catheter infections, pneumonia, UTI, intra-abdominal sources, drug fever, VTE/PE, postop complications—plus the big buckets: infection, malignancy, NIID/autoimmune, miscellaneous. Remember: up to one-third of ICU fevers are noninfectious.
The do-firsts: confirm true fever (repeat reliable temps), then history/physical that hunts clues—travel/exposures, procedures/devices, new meds, immunosuppression, focal symptoms, timing vs. admissions, and serial re-exams. Minimum workup: CBC, chem/LFTs, ESR/CRP, two sets blood cultures (central + peripheral if CVC), UA/urine culture, CXR, LDH/CK, RF/ANA, HIV, region-specific serologies (TB/CMV/EBV), and targeted abd/pelvic imaging when signs point that way. Clue-directed testing > shotgun testing.
ICU and postop twist: start with line cultures, CXR, and targeted CT/US if recent thoracic/abdominal surgery or localizing symptoms. Consider bedside lung ultrasound when the CXR is off. Keep a wide net for drug, neurogenic, and thromboembolic fevers.
Advanced moves (when baseline is unrevealing): FDG PET-CT shines at localizing occult infection/inflammation/malignancy, especially with elevated ESR/CRP and nondiagnostic CT—great for finding biopsy targets and sometimes reassuring when negative. Molecular diagnostics (multiplex PCR, broad-range 16S, NGS) help in culture-negative and immunocompromised hosts. Biopsy (lymph node, liver, temporal artery, bone marrow) when imaging or labs suggest a focal process.
Management principles we actually follow: avoid empiric antimicrobials or steroids unless neutropenic, immunocompromised, or clinically crashing—they mask diagnoses and rarely fix undifferentiated FUO. Individualize based on clues, consult ID/Rheum/Heme-Onc, and monitor closely for turns.
Prognosis is kinder than you think: ~75% resolve without a definitive label; mortality is low in undiagnosed FUO. But don’t miss the bad actors—endocarditis, abscess, TB, lymphoma, vasculitis—where timely diagnosis prevents damage.
Pitfalls you don’t want to meet: calling it FUO before a real exam, ignoring the med list (drug fever!), over-imaging without clues, emps that sterilize cultures, skipping central + peripheral cultures with CVC, and forgetting VTE/PE or thyroiditis on the differential.
We close with the system bundle: (1) standardized FUO admission set (cultures, ESR/CRP, CXR, UA) + clue-driven add-ons; (2) daily re-exam & temp review; (3) automatic line evaluation if CVC; (4) PET-CT when baseline + CT fail and markers are high; (5) image-guided biopsy of any focality; (6) hard stop on empiric therapy unless high-risk; (7) multidisciplinary huddles at 48–72 hrs; (8) documented de-labeling plan when benign course emerges.
Clue-first, PET-smart, biopsy-guided—resist reflex antibiotics, escalate thoughtfully, and let the story lead the testing.

In this episode of Hospital Medicine Unplugged, we blitz through VTE prevention in hospitalized patients—risk-stratify early, anticoagulate wisely, protect the bleeders, and hard-wire the system so the right patient gets the right prophylaxis at the right time.
We open with the do-firsts: Padua (medical), IMPROVE (medical/bleeding), Caprini (surgical) embedded in the admission workflow. High VTE risk + acceptable bleed risk = pharmacologic prophylaxis. High bleed risk = mechanical prophylaxis until it’s safe to anticoagulate.
Call the risk and act:• Padua ≥4 (medical) or Caprini ≥7 (surgical) → pharmacologic ppx.• IMPROVE-Bleed high or active bleeding/procedure → hold drugs, start IPC.• Reassess daily with changes in mobility, procedures, renal function, platelets.
Treatment—build the anticoagulant backbone:• LMWH first-line: enoxaparin 40 mg SC daily (consider 40 mg BID if BMI >40); adjust for renal impairment.• UFH 5,000 units SC q8–12h when CrCl <30 mL/min or rapid on/off is needed.• Fondaparinux 2.5 mg SC daily (avoid if severe renal impairment).• DOACs (betrixaban, rivaroxaban): consider only in select high-VTE/low-bleed patients; not routine for all-comers.
When not to anticoagulate—don’t miss the red flags: active major bleeding, recent high-risk procedure, platelets <50K, prior HIT, severe coagulopathy, intracranial lesions. If any are present, go mechanical and re-evaluate daily.
Mechanical moves:• IPC/SCDs preferred; ensure correct sizing, skin checks, and adherence.• GCS are less effective and higher skin-injury risk—reserve for select cases.• Ambulation helps but is not prophylaxis.
Duration that sticks:• Medical inpatients: during immobilization/acute stay (~6–14 days); avoid routine post-discharge extension.• General surgery: ≥7–10 days; longer if cancer or limited mobility.• Major ortho (THA/TKA/hip fracture): 10–35 days (LMWH/fondaparinux/DOAC per protocol).
Extended prophylaxis reality check:• Benefits: fewer symptomatic VTEs in carefully selected high-risk patients.• Costs: more major bleeding, no clear mortality win.• Bottom line: use selectively (persistent immobility, active cancer, prior VTE, high Padua/IMPROVE) only when bleed risk is low.
System plays that move the needle:• EHR-embedded scores + default order sets (LMWH/UFH with renal/weight dosing).• Auto-alerts for missed ppx, nurse-driven IPC adherence checks, and discharge prompts for ortho/cancer pathways.• Education for patients & teams to counter underuse/overuse.• Dashboards & feedback tied to appropriate prophylaxis, not just “any.”
Monitoring & safety:• Track Hgb/bleeding, platelets (for HIT on heparins), renal function, and drug–drug interactions.• Consider anti–Xa in morbid obesity or extremes of renal function when dosing uncertainty threatens efficacy.• For mechanical ppx, inspect skin daily; avoid with severe PAD or limb wounds.
Special populations quick hits:• ICU: LMWH preferred; mechanical if bleeding.• Obesity: consider higher LMWH dosing; monitor.• Renal failure: UFH over LMWH.• Cancer: favor pharmacologic ppx in-hospital if bleed risk is acceptable.• Stroke: start heparin after hemorrhage excluded; SCDs if bleeding risk.
Medication pitfalls you don’t want to meet: underdosing in obesity, LMWH in severe CKD, unnecessary anticoagulation in low-risk patients, skipping daily reassessment, forgetting PPI in high GI-risk aspirin/NSAID use, and missing HIT.
We close with the bundle: (1) auto-score Padua/IMPROVE/Caprini at admission; (2) default LMWH/UFH orders with renal/weight logic; (3) auto-hold + IPC for high bleed risk; (4) daily mobility/bleed/VTE re-check; (5) ortho/cancer extended-duration pathways; (6) HIT surveillance + anti–Xa when needed; (7) patient education on meds, IPC, and red flags.
Risk-led, safety-first, system-powered VTE prevention—anticoagulate the right patients, protect the bleeders, and make it automatic.

In this episode of Hospital Medicine Unplugged, we cut through pericardial effusion in the inpatient setting—spot it fast with TTE, act early on tamponade, spare procedures when safe, and tailor to cause.
We open with the do-firsts: focused history/exam (dyspnea, chest pressure/fullness, pulsus paradoxus), ECG, CRP/ESR, CBC/chemistry ± troponin if myopericarditis suspected. TTE is first-line to confirm effusion, size it, and look for tamponade physiology (RA/RV diastolic collapse, plethoric IVC, respiratory inflow variation). If TTE is inconclusive, use CMR/CCT/TEE selectively for complex/loculated effusions or to chase secondary causes—routine CMR/CCT for tamponade assessment isn’t recommended.
Call tamponade when the clinical picture and echo line up; remember the physiology: rate of accumulation beats absolute volume—rapid fluid overwhelms pericardial compliance.
Action when it tips:• Cardiac tamponade or hemodynamic compromise → urgent pericardiocentesis, ideally echo-guided. Gentle IV fluids while prepping; avoid PPV if possible.• Send fluid: cell count, protein/glucose/LDH, Gram & cultures, AFB/TB PCR, cytology ± triglycerides if chylous suspected.
If there’s no tamponade:• Don’t tap by default. Observe with serial TTE for moderate/large effusions.• Recurrent large effusions or recurrence after prior tap? Pericardial window is reasonable—lower reintervention vs repeat taps but higher procedural risk; pick based on etiology, recurrence risk, and patient goals.
When effusion rides with acute pericarditis, build the anti-inflammatory backbone:• NSAID (e.g., ibuprofen 600–800 mg q6–8h or aspirin 650–1000 mg q8h) + colchicine 0.5–0.6 mg BID (>70 kg) or QD (≤70 kg) for 3 months; add PPI.• Reserve corticosteroids (e.g., pred 0.2–0.5 mg/kg/day) for refractory or contraindications to NSAID/colchicine; taper slowly, CRP-guided.
Etiology plays—treat the cause:• TB: full RIPE; consider steroids if constrictive/inflammatory phenotype.• Uremia: intensify dialysis; drain if tamponade/large symptomatic.• Malignancy: drain for symptoms/diagnosis; high recurrence → window or extended catheter drainage; consider intrapericardial therapy in select cases.• Post–cardiac injury/post-op: standard NSAID+colchicine; window for loculated/recurrent.
Who stays and how we watch:• Large or symptomatic but stable: close inpatient monitoring, daily clinical checks, frequent TTE (daily or q1–2 days).• Small-to-moderate, stable: outcome is usually benign—watchful waiting with interval TTE; educate on red flags (worsening dyspnea, presyncope, hypotension).• Chronic, large, asymptomatic, idiopathic: watchful waiting is often safe; consider drainage if >3 months or new symptoms/inflammation—evidence is limited, so individualize.
Pearls & pitfalls you don’t want to meet:• Don’t chase every effusion with a needle—match intervention to hemodynamics and suspected cause.• Early steroids in pericarditis → higher recurrence; use as a bridge only when needed.• Light’s criteria don’t apply to pericardial fluid—normal fluid is exudative.• Cytology can be falsely negative—a benign result doesn’t exclude malignancy; keep hunting if suspicion is high.• Trend CRP to steer NSAID taper and to judge inflammation quieting down.
We close with the inpatient bundle that sticks: (1) TTE first ± selective CMR/CCT/TEE; (2) tamponade pathway (STAT echo → urgent echo-guided pericardiocentesis → full fluid studies); (3) no routine taps for stable effusions; (4) NSAID + colchicine for inflammatory cases, steroid-sparing strategy; (5) serial TTE for moderate/large effusions; (6) etiology-specific tracks (TB, uremia, malignancy, post-op) with multidisciplinary consults; (7) shared decision-making for pericardial window in recurrent/malignant disease.
Bottom line—imaging-led and risk-stratified: use TTE to steer, drain when unstable, treat the cause, and surveil the rest without over-proceduring.

In this episode of Hospital Medicine Unplugged, we sprint through inpatient gout—confirm the diagnosis, kill the flare fast (safely), and treat to target so patients stop bouncing back.
We open with the do-firsts: aspirate the joint or tophus when feasible to see MSU crystals under polarized light—doubles as a septic arthritis rule-out. If aspiration’s a no-go, lean on classic features and, when needed, ultrasound/DECT—but don’t anchor on serum urate alone (can be normal in flares). Map comorbidities/meds (CKD, HF, diabetes; diuretics, low-dose ASA, CNIs; CYP3A4/P-gp interactions).
Acute flare playbook—individualize, start early:• Colchicine (low dose): 1.0–1.2 mg, then 0.5–0.6 mg 1 hr later; then 0.5–0.6 mg once/twice daily for 7–10 days. Avoid with strong CYP3A4/P-gp inhibitors or severe renal/hepatic impairment.• Glucocorticoids: intra-articular for mono/oligoarticular disease (favorite in hospitalized patients). Systemic steroids if needed (e.g., prednisone ~0.5 mg/kg/day with short taper) when NSAIDs/colchicine won’t fly.• NSAIDs at anti-inflammatory doses only if no CKD/HF/ulcer/CV risk. In many inpatients, avoid.• Combination (short course) if pain control lags and no contraindications (e.g., colchicine + steroid).• Refractory/intolerant? Consider IL-1 blockade (anakinra off-label, canakinumab where available) after infection is excluded.
Hospital pitfalls you don’t want to meet: delayed treatment, stopping chronic urate-lowering therapy (ULT) on admission without cause, wrong colchicine dose, NSAIDs in CKD/HF, or forgetting drug–drug checks (macrolides, azoles, cyclosporine).
Treat-to-target that sticks (after the flare settles or continue if already on):• Indications for ULT: recurrent flares, tophi, erosive disease (and often CKD).• Allopurinol first-line: start low (≤100 mg/day; lower with CKD) and titrate q3–6 weeks to serum urate <6 mg/dL (≤5 mg/dL if heavy tophi). CKD is not a reason to cap the dose—titrate with labs.• Prophylaxis during ULT changes: colchicine 0.5–0.6 mg daily (or low-dose NSAID/pred if colchicine isn’t an option) for 3–6 months, and at least 1 month after hitting target and flare-free.• Febuxostat if allopurinol-intolerant—use caution in ASCVD. Uricosurics if good renal function; pegloticase for refractory tophaceous disease.• Safety moves: consider HLA-B*58:01 testing in high-risk ancestries (e.g., many Asian and Black populations) or CKD to reduce AHS risk; do med rec for colchicine interactions.
Patient & system moves: teach that gout is chronic, crystal-driven, and fixable with urate lowering; lifestyle tweaks help a bit but medication is the engine. Use the admission to (1) confirm gout, (2) start/titrate ULT when appropriate, (3) prescribe prophylaxis, (4) schedule labs/follow-up, and (5) hand over a clear flare plan.
Fast, crystal-proven, and target-driven—aspirate when you can, tailor the anti-inflammatories, and lock in SU <6 mg/dL so flares fade and tophi melt.

In this episode of Hospital Medicine Unplugged, we sprint through septic shock—treat it like the medical emergency it is: move fast, restore perfusion, hit early antibiotics, control the source, and individualize hemodynamics.
We open with the do-firsts: rapid recognition via Sepsis-3 (infection + organ dysfunction; shock = vasopressors to keep MAP ≥65 and lactate ≥2 after fluids). Fire the Hour-1 moves: obtain cultures (don’t delay therapy), start broad-spectrum antibiotics within 1 hour, and begin crystalloids—≥30 mL/kg within 3 hours for hypoperfusion/shock (weak, low-quality but time-critical). Track lactate and consider capillary refill as adjunct.
Fluid strategy that sticks: use dynamic measures (passive leg raise, stroke volume variation, echo) over static numbers to decide more vs no more. Reassess often to avoid fluid overload—resuscitate to perfusion, then go conservative.
Vasopressors—clean and quick: if hypotension persists, start norepinephrine first-line (peripheral large-bore acceptable while placing a central line). Target MAP 65 mm Hg; higher targets don’t help and may harm. Add vasopressin for catecholamine-sparing or if NE is climbing; add epinephrine if still unstable. Consider inotropes if cardiac output is low.
Source control saves lives: drain it, remove it, debride it, fix it—ASAP. Persistent infection defeats even perfect antibiotics.
Adjuncts with judgment: hydrocortisone 200 mg/day for refractory shock (faster shock reversal; uncertain mortality benefit). Vitamin C “cocktails” and routine β-blockade aren’t supported for outcomes—not routine. Nail the basics: VTE prophylaxis, lung-protective ventilation when needed, glucose 140–180 mg/dL, nutrition, stress-ulcer prophylaxis when indicated.
What the trials taught us: EGDT (Rivers 2001) started the era; ProCESS/ARISE/ProMISe showed no survival edge for rigid protocols over high-quality usual care. The lesson—individualize, guided by physiology and frequent re-assessment, not fixed checkboxes.
Monitoring that matters: continuous MAP, urine output, mentation, skin perfusion; serial lactate for trend (clearance beats single checks). Watch for iatrogenics—volume overload, arrhythmias, extravasation with peripheral pressors, and steroid-related hyperglycemia/infection. Early ICU admission when ICU-level care is needed.
Antibiotics—start broad, de-escalate early: pick empirics by source + local resistance + patient risks, then narrow with cultures and response. Shorter, targeted courses once stable and controlled source reduce harm.
Pitfalls you don’t want to meet: delaying antibiotics, under-resuscitating early then over-hydrating late, chasing CVP, ignoring dynamic tests, skipping source control, overshooting MAP, and forgetting goals-of-care in high-mortality phenotypes.
We close with the system moves: a sepsis bundle that (1) auto-pages a response team; (2) launches cultures → antibiotics ≤1 hr → crystalloids; (3) embeds dynamic fluid checks and MAP 65 default; (4) standardizes NE first-line, early vasopressin add-on; (5) routes to source control within hours; (6) sets lactate-clearance/cap refill targets; (7) flags refractory shock → hydrocortisone; (8) hard-wires de-escalation and duration review at 48–72 hours.
Fast, physiologically guided, and source-focused—antibiotics now, perfusion now, source control now—then tailor the rest.

In this episode of Hospital Medicine Unplugged, we sprint through hyperosmolar hyperglycemic state (HHS)—spot early, rehydrate hard (safely), fix electrolytes, start insulin after fluids, and hunt the trigger.
We open with the do-firsts: ABCs, tele, frequent vitals, bedside neuro checks, and labs that matter—glucose, BMP with corrected Na⁺, calculated effective osmolality (>320 mOsm/kg), serum/urine ketones (minimal/absent), β-hydroxybutyrate, VBG/ABG, CBC, lactate, UA/cultures, CXR/ECG as indicated. Think precipitant early: infection (most common), meds (steroids, thiazides, antipsychotics), MI/stroke, nonadherence, undiagnosed DM.
Call the diagnosis when glucose >600 mg/dL, effective osmolality >320 mOsm/kg, minimal/absent ketosis, and marked dehydration. Mental status change is common but not required. Distinguish from DKA—no significant acidosis/ketosis and often older type 2 patients.
Risk reality check: mortality often >10%; danger flags include hypotension, AKI, corrected hypernatremia, severe hyperosmolality, infection, and altered mental status. Most patients need ICU.
Treatment—build the HHS bundle (sequence matters):• Fluids first (cornerstone): start 0.9% saline to restore perfusion; target a gradual fall in osmolality (≤3 mOsm/kg/hr) and glucose drop ~50–75 mg/dL/hr. Switch to 0.45% saline if corrected Na⁺ is high and the patient is hemodynamically improved. Add D5 when glucose <300 mg/dL to prevent rapid osm shifts.• Potassium next: total body K⁺ is depleted—start replacement once urine output is adequate; hold insulin if K⁺ <3.3 until corrected.• Insulin after partial volume repletion: IV regular insulin 0.1 U/kg/hr (or 0.14 U/kg/hr without bolus). Titrate to keep glucose ~200–300 mg/dL until osmolality normalizes, then transition to subQ when eating and stable.• Electrolytes/thiamine: replace Mg²⁺/PO₄³⁻ if severe/symptomatic; consider thiamine in malnourished to prevent Wernicke’s.• Treat the trigger: early antibiotics if infection suspected; manage MI, stroke, pancreatitis, offending meds.
Monitoring that saves: strict I/O, weights, mental status, q1–2h glucose, q2–4h electrolytes/osmolality early, continuous cardiac monitoring (arrhythmias from K⁺/Mg²⁺ shifts). Avoid overcorrection—aim for steady, not speedy tonicity change.
If the backbone buckles: refractory hyperosmolality or hemodynamic fragility → slow down insulin, reassess fluids/Na⁺ strategy, check for occult infection, consider ICU protocols for glycemic variability reduction. Watch for cerebral edema (rare in adults; risk in children)—worsening headache, bradycardia, neuro decline → pause rapid correction, escalate care.
HHS vs DKA—management contrasts that matter:• HHS = bigger fluid deficit (8–12 L), slower insulin start, tonicity-driven targets; DKA = earlier insulin, acidosis focus.• Ketones minimal in HHS; bicarbonate rarely helpful (reserve for severe acidosis from another cause).
Medication pitfalls you don’t want to meet: insulin before K⁺ correction, missing corrected Na⁺, swinging osmolality too fast, skipping thiamine in high-risk patients, or undertreating infection.
We close with the system moves: an HHS pathway that (1) triages to ICU when altered or unstable; (2) defaults to fluids → K⁺ → insulin with nurse-driven titration; (3) embeds q2–4h labs and osmolality targets; (4) auto-orders trigger workup + early antibiotics; (5) flags anticoagulation/VTE prophylaxis; (6) standardizes electrolyte/PO₄³⁻/Mg²⁺/thiamine replacement; (7) transitions to basal–bolus insulin with diabetes education, sick-day rules, meter/CGM review, and follow-up.
Fast, protocolized, and osmolality-wise—fill the tank, fix the salts, then insulin; treat the cause and prevent the next hit.

In this episode of Hospital Medicine Unplugged, we sprint through thyrotoxicosis and thyroid storm—recognize early, stabilize in the ICU, stop hormone effects fast, and line up definitive therapy.
We open with the do-firsts: high-acuity triage, tele + frequent vitals, broad labs (TSH↓, free T4/T3↑), cultures/CXR/UA if infection suspected, and an immediate search for triggers (infection, surgery, trauma, stopped ATDs, iodinated contrast, amiodarone). Don’t miss mimics—sepsis, sympathomimetic toxicity, serotonin syndrome.
Call the diagnosis clinically and score it: BWPS ≥45 or JTA TS1/TS2 = storm until proven otherwise. Use TRAb for Graves’, ultrasound/Doppler when scintigraphy is out, and thyroglobulin if exogenous hormone is on the table.
Risk reality check: mortality 3.6–25%. High-risk clues—atrial fibrillation, heart failure, hyperpyrexia, CNS dysfunction, hepatic injury, and clear precipitants (especially Graves’, toxic nodules, or abrupt ATD withdrawal). Admit to ICU.
Treatment—build the anti-thyroid storm bundle (give in this sequence):• β-blocker first for adrenergic control: propranolol PO/IV (also ↓ T4→T3). If decompensated LV failure or bronchospasm, consider esmolol (titrate) or diltiazem for rate control.• Thionamide to block new synthesis: PTU (LD 500–1000 mg, then 250 mg q4h) or methimazole (60–80 mg/day divided). PTU adds peripheral T4→T3 blockade, but recent data show no mortality advantage vs methimazole—individualize by contraindications and availability.• Iodide (SSKI or Lugol’s) ≥1 hour after thionamide to block release.• Glucocorticoid: hydrocortisone 100 mg IV q8h (or dexamethasone) for adrenal support and ↓ T4→T3.• Cholestyramine 4 g q6–8h as an adjunct in severe or refractory cases.• Supportive care that matters: acetaminophen + cooling (avoid aspirin), aggressive fluids, nutrition, respiratory support, treat precipitating infection, VTE prophylaxis.
If the backbone buckles (refractory or ATD-intolerant):• Therapeutic plasma exchange to rapidly lower T3/T4 and bridge to definitive therapy—watch coagulopathy/bleeding.• Emergent thyroidectomy in experienced hands when medical therapy fails or is contraindicated, or when life-threatening thyrotoxicosis coexists with urgent non-thyroid surgery.
Etiology plays—treat the cause:• Graves’: ATDs first; long-term low-dose methimazole is reasonable for selected patients.• Toxic multinodular goiter / toxic adenoma: control with ATDs, then RAI (preferred) or surgery for large goiter/compression.• Destructive thyroiditis: β-blockers ± steroids; ATDs don’t help.• Exogenous hormone: stop the source; expect low thyroglobulin.• Amiodarone-induced thyrotoxicosis (AIT):– Type 1 (↑ synthesis): high-dose methimazole (± potassium perchlorate in refractory cases).– Type 2 (destructive): prednisone 40–60 mg/day taper.– Mixed/unclear: combine thionamide + steroid.– Continuing amiodarone is individualized with cardiology.
Cardiac complications—don’t miss them:• Atrial fibrillation is common; prioritize rate control with β-blocker unless contraindicated. Anticoagulate per CHA₂DS₂-VASc, not “because thyrotoxicosis.” Rhythm strategies work better once euthyroid.
Monitoring & de-escalation that sticks:• Trend free T4/T3 q12–24h early, then space out as the patient settles; watch LFTs/CBC for hepatotoxicity or agranulocytosis.• Reassess need for iodide after several days; taper β-blocker and steroids with clinical and biochemical improvement.• For AIT or ongoing amiodarone: check TSH/free T4 every 1–2 weeks, then q4–6 weeks; involve cardiology.
Special situations:• Pregnancy: PTU in 1st trimester, consider switch to methimazole thereafter; plan definitive therapy pre-conception when possible.• Elderly/apathetic hyperthyroidism: low-signal presentations—have a low threshold for storm bundle.
Medication pitfalls you don’t want to meet:• Delaying β-blockade or giving iodide before thionamide (can worsen synthesis).• Over-β-blockade in shocky patients; use esmolol if tenuous.• Missing drug-induced causes (amiodarone, immune checkpoint inhibitors, lithium) or biotin assay interference.
We close with the system moves: a thyroid storm pathway that (1) triages to ICU; (2) fires β-blocker → thionamide → iodide → steroid → cholestyramine with weight/renal/liver checks; (3) auto-treats precipitants; (4) flags AF anticoagulation by CHA₂DS₂-VASc; (5) escalates to plasma exchange or surgery when refractory; (6) routes to etiology-specific tracks (Graves’, nodular, thyroiditis, AIT) with multidisciplinary consults; and (7) embeds CBC/LFT safety monitoring.
Fast, guideline-aligned, and complication-aware—recognize storm, sequence therapy right, fix the trigger, and plan definitive control.

In this episode of Hospital Medicine Unplugged, we blitz through bacterial meningitis—recognize fast, give antibiotics now, add steroids early, and never delay care for tests.
We open with the do-firsts: minutes matter. Draw blood cultures → start empiric IV antibiotics immediately (don’t wait for CT/LP) → add dexamethasone before or with the first dose (stop if Listeria). Rapid risk screen for CT-before-LP: altered mental status, focal deficits, papilledema, immunocompromise, known CNS disease, or new-onset seizures. If none, LP now.
Call the diagnosis with CSF that screams bacteria: WBC >2,000/µL with neutrophil predominance, protein >2.2 g/L, glucose <34 mg/dL or CSF:serum glucose ratio <0.4, ± elevated CSF lactate. Send Gram stain & culture (variable sensitivity), and PCR/multiplex panels to speed organism ID.
Know your pathogens & hosts: adults—S. pneumoniae (~72%), N. meningitidis (~11%); neonates—E. coli & GBS; Listeria risk with neonates, age >50, pregnancy, alcoholism, or immunosuppression. Global reality check: incidence ranges 0.9 → 80/100,000; mortality up to 54% in low-resource settings; ~24% survivors with sequelae.
Empiric therapy—hit hard, cover right:• Adults (community-acquired): vancomycin + ceftriaxone (or cefotaxime).• Add ampicillin for Listeria risks (neonates, >50, immunocompromised, pregnant).• Ceftriaxone-resistant pneumococcus regions: vancomycin is essential.• Healthcare-associated/neurosurgical: vancomycin + anti-pseudomonal β-lactam (cefepime/ceftazidime/meropenem). Tailor to susceptibilities; consider intraventricular/IT therapy for MDR gram-negatives (e.g., colistin for CRAB).• Penicillin/cephalosporin allergy: aztreonam or ciprofloxacin for gram-negatives (plus vanc), or meropenem if appropriate.
Adjuncts & supportive care that save brains:• Dexamethasone 0.15 mg/kg IV q6h × 4 days—give before/with first antibiotic in suspected pneumococcus; discontinue if Listeria. Benefits strongest in high-income settings for mortality & severe hearing loss.• Fluids: maintenance, not restriction; avoid hypovolemia.• ICP management when indicated: head elevation, hyperosmolar therapy (HTS/mannitol), neurocritical monitoring; EVD if hydrocephalus.• Seizures: treat promptly; consider prophylaxis if cortical involvement or prior seizures.• Glucose/Na monitoring (SIADH common); manage fever, hypotension, and shock aggressively.
De-escalate with data: narrow antibiotics once culture/PCR returns; respect local resistance. Typical durations (once pathogen proven): N. meningitidis 7 days, H. influenzae 7–10, S. pneumoniae 10–14, GBS/E. coli 14–21, Listeria ≥21—adjust to course & complications.
Complications you don’t want to meet: septic shock/DIC, cerebral edema, hydrocephalus, stroke, hearing loss, cognitive deficits. Every hour of antibiotic delay raises mortality—build systems that prevent it.
Prevention plays:• Vaccines (pneumococcal, meningococcal, Hib) curb incidence, but non-vaccine serotypes & resistance are rising—stay updated.• Chemoprophylaxis for N. meningitidis contacts (rifampin, ciprofloxacin, or ceftriaxone) and droplet precautions until 24 h of effective therapy.
Pitfalls & pro-tips:• Do not delay antibiotics for CT/LP when no CT indications.• Don’t forget ampicillin when Listeria risk is present.• Stop steroids once Listeria is confirmed.• Under-resuscitation worsens outcomes—avoid fluid restriction.• Order sets for dosing (ceftriaxone 2 g q12h; vanc targeting 15–20 μg/mL; ampicillin 2 g q4h).
We close with the system moves: a meningitis bundle that (1) auto-fires blood cultures → antibiotics → dexamethasone within 60 minutes; (2) screen-flags CT-before-LP criteria; (3) standardizes LP labs + PCR; (4) routes empiric regimens by age/risk/setting; (5) locks in neurocritical care & ICP protocol for severe cases; (6) hard-wires de-escalation & duration on day 2–3; (7) triggers vaccines & contact prophylaxis at discharge.
Speed, steroids, spinal fluid, and stewardship—treat first, confirm fast, and prevent the next case.