Fellow's quick read · Infectious Diseases

Community-acquired pneumonia

ATS/IDSA 2019 (anchor) + ATS 2025 update · reviewed 2026-06-19

Personal study digest for a new ID fellow. Recommendations are from the ATS/IDSA 2019 guideline (Metlay et al) and the local our institution inpatient CPG; the "What's changed since 2019" section is reviewer synthesis of newer evidence (ATS 2025 update, the IDSA non-endorsement, and landmark trials), each claim cited. Not a substitute for the full guideline.

In one line

Most CAP is treated empirically without a pathogen ever identified. The fellow's leverage points: get the severity right (it drives site of care, cultures, and regimen), cover MRSA/Pseudomonas only when locally-validated risk factors are present, de-escalate at 48 h, and stop early — and know that since 2019 steroids are back in for severe CAP.

When to suspect / how to grade severity

Workup the fellow drives

Empiric therapy (ATS/IDSA 2019 + local our institution)

ScenarioDrug / dose / routeNotes
Outpatient,
healthy
Amoxicillin 1 g PO TID · or doxycycline 100 mg PO BID · or a macrolide (only if local pneumococcal resistance <25%)Macrolide monotherapy downgraded to conditional in 2019
Outpatient,
comorbidities
(Amox/clav or cephalosporin) + (macrolide or doxycycline) · or respiratory FQ monotherapy (levo 750 / moxi 400)Comorbidities = chronic heart/lung/liver/renal dz, DM, EtOH, malignancy, asplenia
Inpatient,
nonsevere
β-lactam + macrolide (local: ceftriaxone 1 g IV q24h + azithromycin 500 mg q24h) · or respiratory FQStop azithromycin after a 1.5 g total (3 d) atypical course
Inpatient,
severe
β-lactam + macrolide (preferred) · or β-lactam + respiratory FQ. FQ monotherapy is suboptimal in severe CAPAdd steroids in severe CAP — see "What's changed"
MRSA risk
(add-on)
Vancomycin (AUC/level-guided; 15 mg/kg q12h) or linezolid 600 mg q12hDaptomycin is contraindicated in pneumonia. Negative nasal MRSA PCR → de-escalate
Pseudomonas risk
(add-on)
Cefepime 2 g IV q8h · pip-tazo 4.5 g q6h · ceftazidime 2 g q8h · meropenem 1 g q8h · aztreonam 2 g q8h (severe β-lactam allergy)Cover empirically only with locally-validated risk factors; cultures to de-escalate

Duration & stopping

Key decisions a fellow owns

What's changed since 2019

Reviewer synthesis of newer evidence — not the 2019 guideline. Each claim cited; flags where 2019 (and the local CPG) now lag practice.

  • Steroids are back IN for severe CAP — the headline reversal. 2019 recommended against steroids in severe CAP; CAPE COD then showed hydrocortisone 200 mg/d IV in ICU severe CAP cut 28-day mortality to 6.2% vs 11.9% (absolute −5.6%, 95% CI −9.6 to −1.7; P=0.006), with less intubation and vasopressor use. An individual-patient-data meta-analysis (8 RCTs) confirmed a 30-day mortality benefit (OR 0.72, 0.56–0.94), concentrated in high-CRP (>204 mg/L) patients; hydrocortisone (not dexamethasone) carried the signal. ATS 2025 now suggests systemic steroids in severe CAP — but NOT in influenza pneumonia, and not in nonsevere CAP. Recency-sensitive / local lag: our inpatient CPG (rev. May 2025) still limits steroids to refractory septic shock — confirm before quoting locally. CAPE COD, NEJM 2023;388:1931-41 (PMID 36942789) · Smit IPD meta-analysis, Lancet Respir Med 2025;13:221-233 · See, BMJ Open Respir Res 2024;11:e002141 · ATS 2025, doi:10.1164/rccm.202507-1692ST
  • Shorter is fine: ≤3 days in stable, nonsevere disease. The PTC trial (hospitalized, moderately-severe CAP, stable at day 3) found stopping β-lactam after 3 days non-inferior to 8 days. ATS 2025 now suggests <5 days (minimum 3) for outpatients and nonsevere inpatients who reach clinical stability, while keeping ≥5 days for severe CAP. This relaxes the 2019 "minimum 5 days" floor. Dinh PTC, Lancet 2021;397:1195-1203 (PMID 33773631) · ATS 2025
  • The live ATS↔IDSA split: antibiotics when a virus is positive. ATS 2025 suggests empiric antibacterials for virus-positive patients who are outpatients with comorbidities or nonsevere inpatients (fear of bacterial coinfection). IDSA declined to endorse the guideline over exactly these 2 of 10 recommendations, arguing most virus-positive patients have no bacterial coinfection, brief withholding in nonsevere illness is safe, and reflexive antibiotics fuel overuse/harm. Both societies agree: no antibiotics for virus-positive outpatients without comorbidities; do treat severe CAP. As the ID consultant, drive individualized, trajectory-based decisions — not reflexive coverage. IDSA position (Klompas et al), CID 2026;82:622-624, doi:10.1093/cid/ciaf625
  • Lung ultrasound is now an accepted imaging alternative. ATS 2025: where appropriate POCUS expertise exists, lung ultrasound is a reasonable alternative to chest radiography for diagnosing CAP (conditional, low quality). ATS 2025, doi:10.1164/rccm.202507-1692ST
  • Still current: the severe-CAP criteria, "abandon HCAP," risk-factor-driven MRSA/Pseudomonas coverage with nasal-PCR de-escalation, no routine procalcitonin to withhold therapy, and no routine anaerobic coverage for aspiration all remain standard. ATS/IDSA 2019, doi:10.1164/rccm.201908-1581ST

Anki cards minted this run

  1. Steroids in severe CAP — CAPE COD reversal; exclude influenza (newest practice change).
  2. ATS↔IDSA 2025 split — antibiotics for virus-positive nonsevere CAP (the controversy).
  3. Duration ≤3 days in stable nonsevere CAP — PTC trial.
  4. Severe CAP definition — 1 major / ≥3 minor (drives cultures, regimen, steroids).

Sources: ATS/IDSA CAP guideline 2019 — Metlay et al, Am J Respir Crit Care Med 2019;200(7):e45-e67 (doi:10.1164/rccm.201908-1581ST); ATS CAP guideline 2025 — Jones et al, Am J Respir Crit Care Med 2025;212(1):24-44 (doi:10.1164/rccm.202507-1692ST, PMID 40679934); IDSA non-endorsement — Klompas et al, Clin Infect Dis 2026;82(4):622-624 (doi:10.1093/cid/ciaf625); CAPE COD — Dequin et al, N Engl J Med 2023;388:1931-1941 (doi:10.1056/NEJMoa2215145, PMID 36942789); IPD steroid meta-analysis — Smit et al, Lancet Respir Med 2025;13:221-233 (doi:10.1016/S2213-2600(24)00405-3); corticosteroid systematic review — See et al, BMJ Open Respir Res 2024;11(1):e002141 (doi:10.1136/bmjresp-2023-002141, PMID 38262670).