Pneumonia severity & empiric therapy

Where to treat: ATS/IDSA 2019 recommends a validated prognostic rule plus clinical judgment, preferring the PSI (strong) over CURB-65 (conditional). PSI classes I–III are the low-risk classes. The PSI can underestimate severity in younger patients, and it scores any SBP below 90 the same way.

Who needs ICU-level care: PSI and CURB-65 were not built for this. Use the 2007 IDSA/ATS criteria: 1 major or ≥3 minor = severe CAP. This definition also decides the inpatient regimen, steroid use and course length.

How this page decides: one site-of-care line. New confusion, SBP <90, hypoxemia (SpO₂ <90% or PaO₂ <60 mmHg), CURB-65 ≥2, PSI IV–V or severe CAP each rule out an outpatient recommendation. When the PSI and CURB-65 disagree, the page says so. Blank items are never scored as normal for a low-risk call: core vital signs, mental status and oxygenation must be entered, and a PSI with missing labs is shown as a minimum (≥).

Still admit despite a low score if the patient cannot keep up oral intake, or has substance use, cognitive impairment, severe comorbid illness or poor functional status (ATS/IDSA 2019).

1 point each: Confusion (new) · Urea >7 mmol/L (BUN >19.6 mg/dL) · Respiratory rate ≥30 · Blood pressure: SBP <90 or DBP ≤60 mmHg · age ≥65. CRB-65 drops urea, so it works without labs.

Risk groups as given in NICE NG250 (2025). Urea in mg/dL is converted as BUN (mg/dL) = urea (mmol/L) × 2.8.

Step 1 — class I: age ≤50, none of the five comorbidities, and none of: altered mental status, pulse ≥125, RR ≥30, SBP <90, temperature <35 or ≥40 °C. Class I needs no labs.

Step 2 — points:

*PORT validation cohort (Fine 1997). Across the three cohorts, mortality ranged 0.1–0.4% (I), 0.6–0.7% (II) and 0.9–2.8% (III). With labs missing, the calculator shows the points it can count as a minimum (≥) and does not use that PSI for a low-risk call. Class I needs no labs, but does need age, comorbidities, mental status, pulse, RR, SBP and temperature.

Major: septic shock needing vasopressors · respiratory failure needing mechanical ventilation.

Minor: RR ≥30 · PaO₂/FiO₂ ≤250 · multilobar infiltrates · confusion/disorientation · BUN ≥20 mg/dL · WBC <4,000/µL (from infection alone, not chemotherapy) · platelets <100,000/µL · core temperature <36 °C · hypotension needing aggressive fluid resuscitation.

Severe = 1 major or ≥3 minor. Minor criteria you have not assessed are listed under the result. A non-severe result can change once you add them.

Outpatient, no comorbidity, no MRSA/Pseudomonas risk: amoxicillin 1 g three times daily, or doxycycline 100 mg twice daily, or a macrolide (azithromycin 500 mg day 1 then 250 mg daily; clarithromycin 500 mg twice daily or ER 1,000 mg daily). Use the macrolide only where pneumococcal macrolide resistance is <25%.

Outpatient with comorbidity (chronic heart, lung, liver or renal disease; diabetes; alcoholism; malignancy; asplenia): amoxicillin/clavulanate (500/125 mg three times daily, 875/125 mg twice daily or 2,000/125 mg twice daily) or a cephalosporin (cefpodoxime 200 mg or cefuroxime 500 mg twice daily) plus a macrolide or doxycycline. Alternative: a respiratory fluoroquinolone alone (levofloxacin 750 mg, moxifloxacin 400 mg or gemifloxacin 320 mg daily).

Inpatient, non-severe: β-lactam (ampicillin–sulbactam 1.5–3 g q6h, cefotaxime 1–2 g q8h, ceftriaxone 1–2 g daily or ceftaroline 600 mg q12h) plus a macrolide (azithromycin 500 mg daily or clarithromycin 500 mg twice daily). Alternative: levofloxacin 750 mg or moxifloxacin 400 mg daily alone. If both macrolides and fluoroquinolones are contraindicated: β-lactam plus doxycycline 100 mg twice daily.

Inpatient, severe: β-lactam plus macrolide (stronger evidence) or β-lactam plus fluoroquinolone. Same agents and doses as above.

MRSA coverage: vancomycin 15 mg/kg q12h, adjusted to levels, or linezolid 600 mg q12h. Pseudomonas coverage: piperacillin–tazobactam 4.5 g q6h, cefepime 2 g q8h, ceftazidime 2 g q8h, imipenem 500 mg q6h, meropenem 1 g q8h or aztreonam 2 g q8h. These options do not cover ESBL organisms.

Do not add routine anaerobic coverage for suspected aspiration unless you suspect lung abscess or empyema. Doses are for adults with normal kidney function.

This calculator follows the Table 4 column headings: recent hospitalization with IV antibiotics and locally validated risk factors. Either one alone is flagged, but does not trigger empiric coverage; recent hospitalization with IV antibiotics still triggers pretreatment cultures. The 2019 guideline dropped the HCAP category. Cover empirically only for prior isolation or locally validated risk factors. De-escalate at 48 h if cultures do not grow MRSA or Pseudomonas and the patient is improving.

• Imaging: lung ultrasound is an acceptable alternative to chest radiography where expertise exists (conditional).
• Positive respiratory-virus test: suggest no empiric antibiotics in outpatients without comorbidities. Suggest antibiotics in outpatients with comorbidities and in all inpatients (conditional, very low certainty). IDSA declined to endorse the update over the last two recommendations. It favours individual, dynamic decisions about bacterial co-infection.
• Duration: non-severe CAP (outpatient or inpatient) — suggest <5 days, minimum 3 days, once clinically stable (conditional). Severe CAP — ≥5 days (strong).
• Corticosteroids: recommend against in non-severe inpatient CAP (strong). Suggest systemic corticosteroids in severe CAP (conditional, low certainty), excluding influenza pneumonia. 2019 had advised against routine use in both.

The update names no corticosteroid agent, dose or duration. Its implementation notes favour giving them early, within 24 h of meeting severe-CAP criteria. Unchanged from 2019 and still in force: the regimens, MRSA/Pseudomonas logic, culture indications, anaerobic coverage, oseltamivir for influenza, and no routine follow-up imaging.

Outpatients: no routine sputum or blood cultures.

Inpatients: send pretreatment blood cultures and sputum Gram stain/culture if the patient has severe CAP (especially if intubated), is being covered empirically for MRSA or Pseudomonas, has had these organisms before (especially in the respiratory tract), or was hospitalized and given IV antibiotics in the last 90 days. Otherwise these tests are not routine.

Urinary antigens: pneumococcal — severe CAP. Legionella — severe CAP, outbreak exposure or recent travel. Influenza: test when influenza is circulating. Procalcitonin: do not use it to decide whether to start antibiotics.

• These scores predict 30-day mortality or need for intensive care. None of them rules pneumonia in or out.
• Hypoxemia, social factors and the ability to take oral medication override a low score.
• Immunocompromised patients, and pneumonia acquired in hospital, are outside these guidelines.
• A patient who received one antibiotic class recently should get a different class.
• Course lengths apply to uncomplicated CAP. Empyema, lung abscess, S. aureus bacteremia or slow response need individual decisions.