Bacteraemia
Taking and reading blood cultures, what each organism says about the source, Staphylococcus aureus bacteraemia under the 2026 IDSA/ESCMID statements and SNAP, and 7-day courses for Gram-negatives.
Tier 1ConditionFree sample
Written with AI assistance and fact-checked against the sources below. For exam study, not clinical decisions. How notes are made
At a glance
- Bacteraemia is viable bacteria in the blood, shown by a positive blood culture. It may be transient (after dental work), or a bloodstream infection with systemic features. Sepsis is the dysregulated host response that can follow.
- Blood culture technique decides whether the result means anything: enough blood, two or more sets, before antibiotics.
- The organism tells you where to look. Find and control the source: drain pus, remove infected lines.
- Staphylococcus aureus bacteraemia (SAB) is a separate disease with 15 to 25% mortality. Every case gets repeat cultures, an echo, a hunt for deep foci and, ideally, infectious diseases input.
- Most Gram-negative bacteraemias with source control need 7 days (BALANCE).
Classification
| Term | Meaning |
|---|---|
| Community-onset | Positive culture taken within 48 hours of admission |
| Healthcare-associated | Onset after 48 hours in hospital, or linked to recent procedures, dialysis, lines or residential care |
| Primary | No source found, or an intravascular catheter source |
| Secondary | From an identified focus: urinary, biliary, lung, skin, bone |
| Persistent | Positive cultures 48 hours or more (S. aureus) or beyond 72 hours (other organisms) despite active therapy |
Healthcare-associated SAB is a national safety and quality indicator in Australia, reported by every public hospital.
Taking blood cultures
- Two sets (each an aerobic and an anaerobic bottle) from separate venepunctures before antibiotics. Two sets detect about 90% of bacteraemias, three sets about 98%.
- Volume is the biggest factor in yield: 8 to 10 mL per adult bottle.
- Skin antisepsis with alcoholic chlorhexidine and a dried site cuts contamination.
- Draw from a new venepuncture rather than an existing cannula. If line infection is suspected, take one set through the line and one peripherally at the same time.
- Time to positivity helps: growth within 12 to 16 hours suggests a high bacterial load or an endovascular source.
Reading the result
Is it real?
- Always significant: S. aureus, Streptococcus pneumoniae, beta-haemolytic streptococci, Enterobacterales, Pseudomonas, anaerobes such as Bacteroides, and Candida.
- Often contaminants: coagulase-negative staphylococci, Cutibacterium, Corynebacterium, Bacillus species, Micrococcus. One bottle among several sets, with no line or prosthesis, is usually contamination.
- Coagulase-negative staphylococci in two or more sets with a line, prosthetic valve, joint or pacemaker are real until proven otherwise.
The Gram stain comes first
| Gram stain | Likely organisms | Empirical step while waiting |
|---|---|---|
| Gram-positive cocci in clusters | S. aureus, coagulase-negative staphylococci | Flucloxacillin or cefazolin, plus vancomycin if MRSA risk or severe sepsis |
| Gram-positive cocci in chains or pairs | Streptococci, enterococci, pneumococcus | Benzylpenicillin or amoxicillin; ceftriaxone if pneumococcal meningitis possible |
| Gram-negative bacilli | E. coli, Klebsiella, Pseudomonas | Gentamicin or ceftriaxone by source and severity; broader cover with ESBL risk or neutropenia |
| Yeast | Candida | Echinocandin |
Rapid molecular tests on positive bottles (for example mecA for methicillin resistance) now shorten time to targeted treatment in many Australian labs.
What the organism says about the source
| Organism | Look for |
|---|---|
| S. aureus | Skin and soft tissue, lines, injecting drug use, bone and joint, endocarditis |
| Streptococcus gallolyticus, Clostridium septicum | Colorectal cancer (colonoscopy), endocarditis for S. gallolyticus |
| Streptococcus anginosus group | Abscess: liver, brain, lung, empyema |
| E. coli | Urinary or biliary tract |
| Klebsiella pneumoniae (hypervirulent) | Liver abscess, especially in people of East Asian background, with endophthalmitis and metastatic spread |
| Enterococcus faecalis | Urinary or biliary source, endocarditis in older people |
| Non-typhoidal Salmonella (over 50) | Infected aortic aneurysm or vascular graft |
| Listeria | Pregnancy, older age, immunosuppression; check for meningitis |
| Burkholderia pseudomallei | Melioidosis in the tropical north; prostate and spleen abscesses |
| Candida | Central lines, parenteral nutrition, abdominal surgery, broad-spectrum antibiotics |
Staphylococcus aureus bacteraemia
The IDSA/ESCMID 2026 consensus statements (Part 1, September 2026) replace the old complicated versus uncomplicated split with low-risk versus increased-risk for deep or metastatic infection or relapse. Antibiotic choice will come in later parts.
Risk stratification
- Main increased-risk features (named in the guideline): community onset, a positive culture 48 hours or more after the first, an intracardiac device.
- Other important features: predisposing valve disease or prior endocarditis, injecting drug use, an endovascular graft, SAB in the previous 90 days, embolic events, more than one non-contiguous focus, signs of a deep focus, or no identified source.
Evaluation
- Repeat blood cultures: at least 2 sets at 48 hours after the first positive, then every 24 to 48 hours until negative. Day 1 of therapy is the first negative day.
- TTE in every adult. No group is safe to skip.
- TOE after a negative TTE if there's an intracardiac device, valve disease or prior endocarditis, a positive culture at 48 hours or more, emboli, or more than one focus. Consider it in community-onset SAB and injecting drug use. See the endocarditis note.
- Whole-body imaging (FDG-PET/CT, or targeted imaging of likely sites) for increased-risk patients without a focus after initial evaluation.
- Infectious diseases consultation is strongly encouraged. Bedside ID input reduces mortality in observational studies.
- Source control: remove infected lines, drain abscesses, consider removing other prosthetic material.
Treatment
- MSSA: flucloxacillin 2 g IV 4-hourly or cefazolin 2 g IV 8-hourly. SNAP (NEJM, June 2026): cefazolin was at least as effective as flucloxacillin, with 90-day mortality of 15.0% against 17.0%, and less acute kidney injury (13.9% against 19.6%). It wasn't shown to be better for survival.
- Penicillin-susceptible S. aureus: benzylpenicillin (SNAP, Lancet 2026: fewer kidney adverse events and likely at least as effective as flucloxacillin). Use it only when the lab has confirmed penicillin susceptibility on a second test, such as a penicillin zone-edge disc test (CLSI or EUCAST; the test SNAP required) or absence of blaZ by PCR.
- MRSA: vancomycin with AUC-guided or trough dosing. CAMERA2 found that adding a beta-lactam to vancomycin increased AKI without improving outcomes.
- Duration: 14 days from the first negative culture when evaluation finds no deep focus and signs and symptoms of infection have resolved, in both risk groups (IDSA/ESCMID 2026). If source control (line removal, drainage) happens after cultures clear, the day of source control can be day 1. Consider longer for persistent bacteraemia or an incomplete work-up. Endocarditis, osteomyelitis, septic arthritis or deep abscess needs 4 to 6 weeks, guided by the focus.
- Pitfall: counting duration from the first dose rather than from the first negative culture.
Gram-negative bacteraemia
- BALANCE (NEJM 2025, 3,608 patients): 7 days of antibiotics was non-inferior to 14 days for 90-day mortality (14.5% versus 16.1%) in hospitalised patients, including ICU patients. It excluded severe immunosuppression, S. aureus, Candida and infections needing prolonged therapy.
- IDSA 2025 complicated UTI guideline also recommends 7 days for Gram-negative bacteraemia from a urinary source in patients who improve.
- Oral step-down with a well-absorbed agent (ciprofloxacin, cotrimoxazole, high-dose amoxicillin-clavulanate if susceptible) is reasonable once the patient is stable and the source is controlled.
- Repeat blood cultures aren't routinely needed when the patient is improving.
Other organisms
- Streptococci: benzylpenicillin. Group A streptococcal bacteraemia is invasive group A streptococcal disease, nationally notifiable; close contacts may need prophylaxis under the CDNA guideline.
- Enterococcus faecalis: amoxicillin. Consider endocarditis and do an echo, particularly with community onset or no other source.
- Candida: echinocandin, remove central lines, dilated eye examination, echo, repeat cultures, and 14 days after the first negative culture.
Australian specifics
- Community-associated MRSA is common in Aboriginal and Torres Strait Islander communities in northern and central Australia and among Māori and Pacific peoples, and SAB rates are several times higher. Cover MRSA empirically in these settings.
- Melioidosis in the tropical north in the wet season: meropenem for severe disease, then eradication therapy with cotrimoxazole.
Pitfalls
- Dismissing S. aureus in one bottle as a contaminant.
- Skipping the echo in SAB because the patient looks well.
- Leaving an infected line in place and wondering why cultures stay positive.
- Missing colorectal cancer after S. gallolyticus or C. septicum bacteraemia.
- Treating every Gram-negative bacteraemia for 14 days when 7 days is enough after source control.
- Vancomycin for MSSA because it was started empirically. It is inferior to a beta-lactam; switch once susceptibility is known.
Finished reading?
Sources
- IDSA/ESCMID 2026 consensus statements on Staphylococcus aureus bacteremia: risk stratification, diagnostic evaluation and management of adults and children (Part 1)Infectious Diseases Society of America and European Society of Clinical Microbiology and Infectious Diseases, 2026 (published 9 September 2026)
- Cefazolin versus flucloxacillin or cloxacillin for methicillin-susceptible Staphylococcus aureus bacteremia (SNAP trial)AU/NZNew England Journal of Medicine, 2026
- Benzylpenicillin versus flucloxacillin or cloxacillin for the treatment of penicillin-susceptible Staphylococcus aureus bacteremia (SNAP): an international, multicentre, open-label, non-inferiority randomised controlled trialAU/NZSNAP Trial Group, Lancet 2026;408(10550):141-52 (reported by UCL Institute of Clinical Trials and Methodology), 2026 (online 17 June 2026)
- Antibiotic treatment for 7 versus 14 days in patients with bloodstream infections (BALANCE trial)New England Journal of Medicine (Daneman et al.), 2025
- Therapeutic Guidelines: Antibiotic. Bloodstream infections and sepsisAU/NZTherapeutic Guidelines Limited (eTG complete), 2025 to 2026 staged update (topic version not verified)
- Last updated
- 14 November 2026
What changed
- 14 November 2026 Deploy to Cloudflare on push to main