Staphylococcus aureus bacteremia — fellow's quick read (consolidated)

Personal study digest for a new ID fellow. Consolidated file — it merges the two prior SAB digests (2026-06-16 therapy digest + 2026-07-06 Part-1 digest) into one and supersedes both. The 2026 IDSA/ESCMID guideline (Part 1) governs risk stratification, diagnostics, and duration (consensus statements, not GRADE); it deliberately does NOT cover antibiotic selection, so the Antibiotic selection section below is drawn from the IDSA MRSA 2011 guideline and the JAMA 2025 review (Tong et al.). The “What’s changed” section is reviewer synthesis, each claim cited. Not a substitute for the full documents.

In one line

Retire “complicated vs uncomplicated.” Stratify every adult as low-risk or increased-risk of a deep/metastatic focus, then let that drive workup and duration. TTE in everyone, pull CVCs promptly, get an ID consult. No deep focus found → 14 days either way. Drug: cefazolin for MSSA; vancomycin, daptomycin, or ceftobiprole for MRSA; don’t add a second agent.

◆How do I stratify the patient?

SAB is the leading infectious cause of death from bloodstream infection; case-fatality 15–30%, ~300,000 deaths/yr worldwide. Occult metastatic seeding is present in >⅓ at presentation — endocarditis (~12%), septic arthritis (~7%), vertebral osteomyelitis (~4%), epidural/psoas/splenic abscess, septic pulmonary emboli, device seeding — which is why “uncomplicated” by gestalt is unsafe.
Three KEY risk factors for a deep/metastatic focus or relapse: (1) community-onset SAB, (2) positive blood culture ≥48 h after the first positive, (3) intracardiac device (prosthetic valve, pacemaker/ICD, LVAD).
Other risk factors: predisposing valve conditions, injection drug use, endovascular graft, SAB in prior 90 days, signs/symptoms of a deep focus, embolic events, >1 non-contiguous focus, unknown focus.
Low-risk = none of these; increased-risk = ≥1. Risk is a moving target — reassess with repeat exams; a low-risk patient can convert. S. aureus bacteriuria should prompt a hunt for SAB (uncommon primary UTI organism).

◆Workup the fellow drives

Follow-up blood cultures: ≥2 sets at 48 h after the first positive, then 1–2 sets q24–48 h until negative. Intensify (q24 h, clearance on 2 consecutive days) with ongoing symptoms, a known/suspected endovascular focus, or FUBC still positive at ≥48 h. Persistence ≥48 h is tied to ~39% 90-day mortality.
TTE in ALL adults — the panel could not define a group at low-enough IE risk to skip it.
TEE is risk-stratified, not automatic. After a negative TTE, do TEE with any endocarditis increased-risk feature (intracardiac device, predisposing valve/prior IE, positive FUBC ≥48 h, embolic event, >1 non-contiguous focus); it may be omitted after a good-quality negative TTE with none of these.
Occult focus / increased-risk / unknown source: whole-body [18F]FDG-PET/CT or a targeted combination (thoraco-abdominal CT, duplex venous US, symptom-directed MRI — e.g. gadolinium spine MRI for back pain).
Remove central venous catheters promptly (delay → more hematogenous complications and relapse). ID consultation for every SAB — associated with lower mortality.

◆Antibiotic selection — from IDSA MRSA 2011 + JAMA 2025 (Part 1 defers this)

The 2026 Part-1 guideline does not address agent choice. This is the current therapy standard from the MRSA 2011 guideline and the JAMA 2025 review.

Scenario	Drug / dose / route	Notes
Empiric (MRSA risk)	Vancomycin (AUC 400–600) or daptomycin 6–10 mg/kg IV daily	Cover MRSA where prevalence >5% (e.g. US) or with MRSA risk factors, until susceptibilities return
MSSA — preferred	Cefazolin 2 g IV q8h (2 g q6h if critically ill)	Lower 30-day mortality (RR 0.70) & less nephrotoxicity (RR 0.36) than antistaphylococcal penicillins in observational data
MSSA — alternative	Nafcillin/oxacillin 2 g q4h · flucloxacillin/cloxacillin	Antistaphylococcal penicillins (ASPs); watch cefazolin’s theoretical high-inoculum effect
MRSA	Vancomycin (AUC-guided) · daptomycin 6–10 mg/kg daily (not for pneumonia) · ceftobiprole 500 mg IV q6h (q8h after day 8)	3 FDA-approved MRSA-SAB agents; ceftobiprole approved 3 Apr 2024
Severe PCN allergy	Vancomycin or daptomycin	Anaphylaxis / SJS-TEN. Cefazolin is fine for most non-severe allergy

No drug has proven superior to vancomycin; daptomycin and ceftobiprole are non-inferior options (daptomycin 44% vs 42% success; ceftobiprole 70% vs 69% vs daptomycin).
Do NOT add a second agent routinely — 8 RCTs since 2016 (rifampicin, fosfomycin, β-lactam combos) failed to improve mortality; adjuncts add toxicity (gentamicin → nephrotoxicity).

◆Diagnostic & duration decisions (adults)

Decision	2026 recommendation	Notes (all “consensus”)
Framework	Low-risk vs increased-risk of deep foci / relapse	Replaces “complicated/uncomplicated,” which the panel calls imprecise
FUBC	≥2 sets at 48 h, repeat q24–48 h until clear	Clearance = first negative with no later positives
TTE	All adults	No reliably-definable very-low-IE-risk group
TEE	Only if an IE increased-risk feature after negative TTE	Consider for community-onset or IDU; may skip if good TTE + no features
Whole-body imaging	PET/CT or multimodal for occult focus (increased-risk)	Observational evidence only; mortality/relapse impact unproven
Duration — low-risk, no focus	14 days	Not shorter, not longer
Duration — increased-risk, no focus	14 days if symptoms resolved & workup negative	>14 d only in select cases (prolonged bacteremia, incomplete workup)

◆Duration & the clock

A risk factor is not a focus. Increased-risk SAB with a thorough negative evaluation still gets 14 days — the old reflex of 4–6 weeks for anything “complicated” does not apply when no deep focus is found. A proven deep focus (endocarditis, vertebral osteomyelitis) still earns its own longer, focus-specific course.
Start the clock at blood-culture clearance (first negative), not day 1 of antibiotics. If source control happens after clearance, count from removal of the focus.
Prolonged (≥48 h) bacteremia is the trigger to extend and to re-hunt for a focus.

◆Key decisions a fellow owns

Positive FUBC ≥48 h is the single most robust predictor of a bad outcome and signals inadequate source control → escalate imaging (TEE, PET/CT) and reassess for a drainable/removable focus.
Source control early — remove infected lines/devices, drain abscesses, debride. Shorter time to source control → faster clearance and lower mortality.
Prompt line/device removal + early ID consult are the two lowest-effort mortality levers.
Persistent bacteremia (~30% >3 d): re-image for an uncontrolled source; salvage antibiotic changes/additions are not RCT-supported.

◆Special populations — pediatrics

Evaluate every child with SAB for a deep-seated focus — no low-risk pediatric group could be defined. Most community-onset pediatric SAB has a musculoskeletal source.
TTE only with structural heart disease, prolonged bacteremia, or signs/symptoms of IE — otherwise it may be omitted; TEE adds little over a good TTE in most young children.
Neonates are a separate problem — less likely to have an obvious focus, higher endocarditis rate; consider whole-body imaging if SAB persists.
No-focus duration is 14 days; extend for endovascular risk (congenital heart disease, thrombi, FUBC ≥48 h) with an incomplete evaluation.

What’s changed — and what Part 1 defers

Reviewer synthesis — how the 2026 guideline reshapes practice, and where it stays silent. Each claim cited.

First dedicated IDSA/ESCMID SAB guideline. It replaces “complicated/uncomplicated” with a low-risk vs increased-risk framework on 3 key risk factors. Consensus statements, not GRADE. IDSA/ESCMID 2026, Consensus Statements 1–7 (idsociety.org); not yet PubMed-indexed as of 2026-07-20.
14 days is now the anchor even for many increased-risk patients — a risk factor without a demonstrated focus does not buy 4–6 weeks. Operationalizes the algorithm-based-therapy RCT and the low-risk oral-switch data. Holland et al., JAMA 2018;320:1249-1258; SABATO, Kaasch et al., Lancet Infect Dis 2024;24:523-534 (PMID 38244557; doi).
SAB is not a 7-day disease. The short-course BSI trial changing gram-negative practice excluded S. aureus, so its 7-day result does not extend to SAB. BALANCE, NEJM 2025 (7 vs 14 d non-inferior; S. aureus excluded).
Part 1 defers therapy → use MRSA 2011 + JAMA 2025. For MSSA, cefazolin is now preferred over antistaphylococcal penicillins (meta-analysis of 14 studies: 30-day mortality RR 0.70 [0.54–0.91], nephrotoxicity RR 0.36 [0.21–0.59]). Weis et al., Clin Microbiol Infect 2019;25:818-827; per Tong et al., JAMA 2025 (doi; PMID 40193249).
Ceftobiprole — 3rd FDA-approved MRSA-SAB agent — non-inferior to daptomycin (day-70 success 70% vs 69%); FDA-approved 3 Apr 2024. Recency-sensitive — confirm formulary. ERADICATE, Holland et al., NEJM 2023;389(15):1390-1401 (PMID 37754204; doi).
Combination therapy is out — 8 RCTs since 2016 negative, including adjunctive rifampicin. ARREST, Thwaites et al., Lancet 2018;391:668-678 (N=758).
Open / unresolved: the risk framework is unvalidated (the panel says so); no low-risk pediatric group could be defined; PET/CT rests on observational data — its effect on mortality/relapse vs targeted imaging is unproven.

◆Anki

No new cards this run — consolidation only; both source digests' cards are already in the deck. Already carded: (Part-1 review) low-risk vs increased-risk framework + 3 key risk factors, TTE-all/TEE-if-feature, 14-day duration for low- and increased-risk without a focus, FUBC ≥48 h = strongest adverse predictor; (therapy review) ceftobiprole/ERADICATE, cefazolin>ASP for MSSA, combination-out/ARREST, SABATO oral step-down.

Sources: IDSA/ESCMID 2026 SAB guideline — Executive Summary + Consensus Statements 1–7 (Liu, Chambers, Vandenesch, Kern; idsociety.org; not yet PubMed-indexed). Tong SYC, Fowler VG Jr, Skalla L, Holland TL. Management of S. aureus Bacteremia: A Review. JAMA 2025 (doi:10.1001/jama.2025.4288; PMID 40193249). ERADICATE/ceftobiprole — Holland TL et al., NEJM 2023;389(15):1390-1401 (PMID 37754204; doi:10.1056/NEJMoa2300220). Daptomycin — Fowler VG Jr et al., NEJM 2006;355:653-665. Cefazolin vs ASP meta-analysis — Weis S et al., Clin Microbiol Infect 2019;25:818-827. Vancomycin AUC — Rybak MJ et al., Am J Health Syst Pharm 2020;77:835-864. Combination therapy — ARREST, Thwaites GE et al., Lancet 2018;391:668-678; CAMERA2, Tong SYC et al., JAMA 2020;323:527-537. Duration/oral switch — Holland et al., JAMA 2018;320:1249-1258; SABATO, Kaasch et al., Lancet Infect Dis 2024;24:523-534 (PMID 38244557); BALANCE, NEJM 2025 (doi:10.1056/NEJMoa2404991). Therapy anchor: IDSA MRSA 2011 (Liu C et al., Clin Infect Dis 2011). Therapy claims verified against the JAMA 2025 source-of-record; SAB-guideline currency via IDSA + PubMed (July 2026).