Antimicrobials in sepsis

Contents

Antibiotic Timing

The 1-Hour Bundle

The Surviving Sepsis Campaign (2018 update) Hour-1 Bundle includes administration of broad-spectrum antibiotics within 1 hour of recognition of sepsis or septic shock. This recommendation remains in the 2021 SSC guidelines.

Evidence for Early Antibiotics

Retrospective studies in septic shock (Kumar et al, CCM 2006) showed each hour of delay in appropriate antibiotic administration was associated with a ~7% increase in mortality. While prospective RCT data is limited, the biological rationale is strong: early bacterial killing reduces the duration of endotoxaemia, reduces cytokine release, and prevents further tissue invasion.

Caveats

  • The 1-hour window applies to septic shock; in sepsis without shock, a 3-hour window allows more accurate syndrome identification
  • Do not delay antibiotics for cultures that require prolonged procedures, but do draw blood cultures (2 sets, aerobic and anaerobic) immediately before antibiotic administration
  • Antibiotic appropriateness matters more than speed: a delayed appropriate antibiotic may be more beneficial than an early inappropriate one, though ideally both goals are met

Empirical Regimen Selection

Principles

Empirical regimens should be chosen based on:

  1. Suspected source of infection (community vs healthcare-associated, anatomical site)
  2. Local resistance patterns (hospital antibiogram)
  3. Patient risk factors for resistant organisms (recent antibiotics, healthcare exposure, colonisation status, immunosuppression)
  4. Severity: septic shock warrants broader-spectrum coverage

Regimens by Suspected Source

Source Empirical regimen (UK example)
CAP (community-acquired pneumonia) Co-amoxiclav + clarithromycin OR amoxicillin + doxycycline (mild/mod); piperacillin/tazobactam + clarithromycin (severe)
HAP/VAP (hospital/ventilator-associated) Piperacillin/tazobactam OR meropenem; add vancomycin if MRSA risk
Urinary tract (urosepsis) Co-amoxiclav or cephalosporin (community); meropenem or gentamicin if ESBL risk
Intra-abdominal/biliary Piperacillin/tazobactam or meropenem + metronidazole
Skin/soft tissue Flucloxacillin (streptococcal/staphylococcal); add clindamycin if necrotising fasciitis (toxin inhibition)
Meningitis/encephalitis Ceftriaxone 2 g IV 12-hourly + amoxicillin (cover Listeria >50 years) + dexamethasone
Unknown source / septic shock Meropenem OR piperacillin/tazobactam (if low resistance risk) ± vancomycin

MRSA Risk Factors

Add vancomycin (or teicoplanin, daptomycin, linezolid) if:

  • Known MRSA colonisation
  • Recent MRSA infection
  • Healthcare-associated skin/soft tissue infection
  • HAP/VAP in high-prevalence unit

ESBL / Carbapenemase Risk

  • Meropenem for known ESBL colonisation or high ESBL risk (recent healthcare exposure, known colonisation)
  • Carbapenemase-producing organisms (CPO): seek microbiology input urgently; options include ceftazidime-avibactam, colistin, or combination therapy depending on mechanism

Pharmacokinetics and Pharmacodynamics (PK/PD)

Time-Dependent Killing (Beta-Lactams)

Beta-lactams (penicillins, cephalosporins, carbapenems, aztreonam): bactericidal activity depends on the time the free drug concentration exceeds the MIC (minimum inhibitory concentration). Target: free drug concentration > MIC for ≥40–50% of the dosing interval (% T>MIC).

PK/PD optimisation strategies:

  • Extended infusion: meropenem given over 3 hours rather than 30 minutes → maintains drug levels above MIC for longer → improved target attainment, particularly for organisms with high MIC
  • Continuous infusion: maintain steady-state levels; complex drug stability considerations
  • Dosing frequency: frequent dosing (e.g., meropenem 1g TDS) rather than less frequent high doses for resistant organisms

Concentration-Dependent Killing (Aminoglycosides, Fluoroquinolones)

Bactericidal activity correlates with the peak:MIC ratio — the higher the peak, the better the kill. Optimal strategy: once-daily dosing achieves higher peaks with equivalent (or lower) trough levels, reducing nephrotoxicity. Target: peak gentamicin:MIC ratio >10.

Volume of Distribution in Critical Illness

  • Sepsis causes capillary leak → ↑ volume of distribution (Vd) for hydrophilic drugs (aminoglycosides, beta-lactams, glycopeptides)
  • Higher Vd → lower drug concentrations for a given dose → risk of subtherapeutic levels
  • Higher loading doses may be needed in septic shock; therapeutic drug monitoring (TDM) essential

Renal Clearance

  • AKI reduces elimination of renally cleared drugs → dose reduction needed (gentamicin, glycopeptides, beta-lactams)
  • But: augmented renal clearance (ARC) occurs in young patients with hyperdynamic sepsis → increased creatinine clearance → rapid drug elimination → subtherapeutic levels even with normal dosing
  • ARC affects: meropenem, pip-tazo, glycopeptides; TDM or higher dosing may be needed

Therapeutic Drug Monitoring (TDM)

  • Vancomycin: target AUC/MIC ratio ≥400 (Bayesian-guided dosing recommended); troughs alone are inadequate for optimising dosing
  • Gentamicin: pre-dose (trough) <1 mg/L; extended interval dosing with Hartford nomogram
  • Meropenem/pip-tazo: beta-lactam TDM increasingly used in critically ill; target free drug > MIC for 100% T>MIC in severe infections

De-escalation and Stewardship

When to De-escalate

  • When cultures and sensitivities are available (typically 48–72h): narrow spectrum based on identified organism
  • If initial broad coverage was empirical and cultures are negative: reassess clinical syndrome; consider clinical cure rather than a fixed duration
  • Do not de-escalate prematurely in immunocompromised or severely unwell patients

Duration of Antibiotics

Shorter courses are effective for most infections:

  • Sepsis (no clear source identified, culture negative, clinical response): 5–7 days
  • CAP: 5 days (IDSA/BTS)
  • VAP: 7 days (non-fermenter pathogens may require longer)
  • IE: 4–6 weeks (depends on organism and valve type — see infective endocarditis page)
  • Abdominal: source controlled → 4 days sufficient (IDSA guidelines)

Procalcitonin-Guided Therapy

Procalcitonin (PCT) rises in bacterial infection, falls with treatment response. PCT-guided antibiotic stewardship:

  • High PCT at baseline (>0.5 mcg/L): supports bacterial sepsis diagnosis; start/continue antibiotics
  • PCT fall >80% from peak (or absolute value <0.25 mcg/L): supports antibiotic discontinuation
  • Multiple RCTs (PRORATA, ProGUARD) show PCT guidance reduces antibiotic duration without increasing mortality or treatment failure
  • PRORATA trial (Bouadma, Lancet 2010): PCT-guided strategy reduced antibiotic exposure by 2.7 days in ICU patients; no harm

Antimicrobial Stewardship Principles

  1. Right drug: based on suspected pathogen and local resistance
  2. Right dose: PK/PD-guided; TDM where indicated
  3. Right duration: stop when no longer needed
  4. Review at 48–72h: de-escalate or stop if appropriate
  5. Culture before antibiotics (but do not delay treatment)
  6. Restrict broad-spectrum agents: carbapenems, glycopeptides require justification and approval; reduce selection pressure for resistance

Source Control

Definition

Source control refers to physical measures to drain abscesses, debride necrotic tissue, remove infected foreign bodies, or otherwise eliminate the anatomical focus of infection — it is a core component of sepsis management alongside antibiotics.

Principles

  • Timing: "as soon as medically and logistically practical" (SSC 2021) — delays in source control allow ongoing bacterial seeding
  • Least invasive effective method: percutaneous drainage (CT/US-guided) preferred over open surgery where feasible
  • Examples: abdominal collections → CT-guided drain; cholangitis → ERCP + sphincterotomy/biliary drain; empyema → thoracic drain/VATS; infected prosthetic valve → surgical valve replacement; necrotising fasciitis → surgical debridement (cannot be substituted by antibiotics alone)
  • Removal of infected central lines, urinary catheters

Antifungals in Sepsis

Candida and ICU

Invasive candidiasis (IC) is the most common fungal infection in ICU:

  • Risk factors: prolonged ICU stay, broad-spectrum antibiotics, TPN, renal replacement therapy, immunosuppression, abdominal surgery, Candida colonisation
  • Candida score: risk stratification tool; score ≥3 suggests empirical antifungal consideration

Empirical Antifungals

  • Not recommended routinely — empirical antifungals increase antifungal resistance and Clostridium difficile; increase cost; most sepsis in ICU is bacterial
  • Consider in high-risk patients (multiple risk factors above) with unexplained sepsis not responding to antibiotics, particularly with Candida colonisation

Treatment of Confirmed Candida

  • Echinocandins (caspofungin, anidulafungin, micafungin): first-line for IC; fungicidal activity; covers most Candida species including C. glabrata and C. krusei; IV only
  • Fluconazole: appropriate for susceptible Candida (C. albicans) once stable; step-down from echinocandin; avoid empirically if azole-resistant species possible
  • Cryptococcal meningitis: amphotericin B deoxycholate (or liposomal) + flucytosine; then fluconazole maintenance
  • Aspergillosis: voriconazole or isavuconazole; high-risk: prolonged neutropenia, lung transplant, haematological malignancy

Viva Questions

1. Explain the pharmacodynamic concepts of time-dependent and concentration-dependent killing and how they guide antibiotic prescribing in the ICU.

Beta-lactam antibiotics (penicillins, cephalosporins, carbapenems) kill bacteria in a time-dependent manner — bactericidal activity is proportional to the proportion of the dosing interval during which free drug concentrations exceed the MIC. Once concentrations exceed 4–5× MIC, further increases do not improve killing. For these drugs, the goal is to maintain free drug levels above the MIC for the longest possible fraction of the dosing interval (typically target ≥40–50% T>MIC, or ≥100% for serious infections with less susceptible organisms). In the ICU, this is achieved by extended or continuous infusion, or more frequent dosing. Contrast with aminoglycosides and fluoroquinolones, which are concentration-dependent: bactericidal activity depends on the peak drug concentration relative to the MIC (peak:MIC ratio target ≥10). Higher peaks achieve better kill. For aminoglycosides, this means once-daily dosing (achieves high peaks with lower cumulative dose) rather than divided doses — which also reduces nephrotoxicity. Additionally, critical illness alters pharmacokinetics: expanded volume of distribution reduces peak concentrations; augmented renal clearance (ARC) increases elimination; AKI causes drug accumulation. TDM is therefore essential for vancomycin, aminoglycosides, and increasingly for beta-lactams in severe infection.

2. How does VANISH (see VANISH journal club) inform your choice of vasopressor in septic shock, and how does this relate to antibiotic management?

The VANISH trial (Gordon, JAMA 2016) was a 2×2 factorial RCT testing vasopressin vs noradrenaline and hydrocortisone vs placebo in septic shock (n=409). Vasopressin did not improve kidney failure-free days compared to noradrenaline, and there was no significant mortality difference. VANISH does suggest vasopressin may reduce the rate of renal replacement therapy initiation compared to noradrenaline, though the primary endpoint was not met. The factorial design also confirmed no significant interaction between vasopressin and hydrocortisone. In terms of antibiotic relevance: vasopressor choice impacts renal perfusion, and renal impairment affects antibiotic pharmacokinetics dramatically. Patients on noradrenaline with concurrent AKI require dose reductions of renally cleared antibiotics (aminoglycosides, carbapenems, glycopeptides); patients with vasopressin-preserved renal function may have different PK profiles. Restoration of renal perfusion with vasopressors also influences whether augmented renal clearance (ARC) or AKI is the dominant PK perturbation. Thus, vasopressor management, renal function, and antibiotic dosing are deeply connected — daily review of creatinine clearance and TDM is essential.

3. When should you de-escalate antibiotics in sepsis, and what is the role of procalcitonin?

De-escalation is appropriate when: culture and sensitivity results identify an organism susceptible to a narrower-spectrum agent (narrow to the most targeted drug); when cultures are negative and the patient is improving clinically (suggests lower bacterial burden or non-bacterial diagnosis); or when the agreed clinical endpoint has been reached (temperature, inflammatory markers, organ function normalising). De-escalation should occur at 48–72 hours as a systematic review point — not passively. Procalcitonin provides a biomarker framework: a rise in PCT supports bacterial infection; a fall ≥80% from peak (or absolute value <0.25 mcg/L) supports stopping antibiotics if clinically consistent. The PRORATA trial showed PCT-guided antibiotic cessation reduced antibiotic exposure by ~2.7 days without increased mortality or treatment failure. Caveats: PCT does not distinguish colonisation from infection; PCT can rise in trauma, surgery, and some non-infectious inflammatory states; PCT is less reliable in localised infection without bacteraemia. De-escalation decisions should always integrate clinical assessment — improving vital signs, resolving organ dysfunction, white cell count trend — with biomarker guidance. Steroids suppress PCT, which must be accounted for in patients receiving corticosteroids.