Give the right empirical antibiotics immediately (ideally within 1 h) in shock, chosen by source, MDR risk and the local antibiogram, with full loading doses and prolonged-infusion beta-lactams (now a strong SSC 2026 recommendation).
At the bedside
Timed steps- 1
Time zero
0 minStart the clock at triage or recognition. In septic shock each hour of delay is associated with higher mortality.
- 2
Cultures first, if fast
0-45 minBlood cultures ×2 (or 1 set where resources are limited) and likely-source samples (sputum or endotracheal aspirate, urine, pus, CSF if safe) before the first dose. SSC 2026 recommends cultures as soon as possible and ideally before antibiotics, but never delay antibiotics for them.
- 3
Choose by source and MDR risk
≤ 1 hSSC 2026: cover a specific MDR organism only if the risk is high (colonisation or infection with it, prolonged broad-spectrum antibiotics, long stay in a unit where it is common); do not cover it if the risk is low. In India also weigh dialysis, ICU-onset infection and local ESBL/CRE rates. Use the template table and your antibiogram.
- 4
Septic shock with unclear focus: combination
≤ 1 hICMR prefers a carbapenem (imipenem or meropenem) ± amikacin, adding MRSA, CRE or antifungal cover only with risk factors. Avoid pip-taz in shock until ceftriaxone-resistant bacteraemia is excluded (MERINO).
- 5
MRSA cover only with risk
≤ 1 hPrior MRSA, line or skin/soft-tissue source, haemodialysis, post-operative infection. SSC 2026 suggests no empirical cover for an MDR organism (including MRSA) when the risk is low.
- 6
Anaerobic and antifungal cover only with risk
≤ 1 hSSC 2026: no routine anaerobic cover (e.g. for lung or urinary sources). Add it for intra-abdominal, deep pelvic or obstetric, necrotising soft-tissue, head and neck, CNS abscess or empyema sources. Pip-taz or a carbapenem already cover anaerobes. No routine empirical antifungal; consider case by case (immunosuppression, long antibiotics or stay, abdominal source).
- 7
Tropical cover in Indian febrile illness
≤ 1 hAdd doxycycline (or azithromycin) if scrub typhus, rickettsia or leptospirosis is plausible; artesunate if malaria is possible.
- 8
Full loading dose regardless of kidney function
First doseSepsis expands the volume of distribution. Adjust only maintenance doses for kidney function after the first 24 h.
- 9
Prolonged beta-lactam infusion
MaintenanceAfter a full first dose over 30 min, give meropenem or pip-taz as an extended (3-4 h) or continuous infusion. SSC 2026 now recommends this (strong, moderate certainty; was weak in 2021). Use therapeutic drug monitoring case by case where available.
- 10
Review cultures and de-escalate
48-72 hSee the de-escalation topic.
Empirical, septic shock or high ESBL risk
Meropenem
- Dose
- 1 g every 8 h over 3 h (2 g every 8 h for meningitis or high MIC)
- Preparation
- 1 g in 50-100 mL NS, prepared fresh
First dose over 30 min.
Empirical, lower ESBL risk, not in shock until ESBL bacteraemia excluded
Piperacillin-tazobactam
- Dose
- 4.5 g every 6 h; first dose over 30 min, then each over 3-4 h
- Preparation
- 4.5 g in 100 mL NS
Second gram-negative agent in shock
Amikacin
- Dose
- 15-20 mg/kg once daily
- Worked example
- 70 kg: 1050-1400 mg once daily (round to 1 g or 1.5 g by local policy)
MRSA cover when risk factors present
Vancomycin
- Dose
- Load 20-35 mg/kg (max 3 g), then 15-20 mg/kg every 8-12 h
- Target
- AUC 400-600 mg·h/L (trough 15-20 mg/L if AUC unavailable)
- Worked example
- 70 kg: load 1.75-2 g over 2 h, then 1 g every 12 h; adjust by levels
Carbapenem-resistant organisms (non-urinary)
Polymyxin B
- Dose
- Load 2-2.5 mg/kg over 1 h, then 1.25-1.5 mg/kg every 12 h
- Worked example
- 70 kg: load 175 mg (1.75 million units), then 100 mg (1 million units) every 12 h
Use in combination; no renal adjustment.
Carbapenem-resistant organisms (including UTI)
Colistin (colistimethate sodium)
- Dose
- Load 9 MU, then 4.5 MU every 12 h (normal kidney function)
CRE (KPC or OXA-48); with aztreonam for NDM
Ceftazidime-avibactam
- Dose
- 2.5 g every 8 h (+ aztreonam 2 g every 8 h for metallo-beta-lactamase)
Empirical templates by source (adapt to the local antibiogram)
| Source | Community, low MDR risk | Healthcare-associated or high MDR risk | Add if |
|---|---|---|---|
| Pneumonia | Ceftriaxone 2 g OD + azithromycin 500 mg OD (or doxycycline) | Meropenem 1 g q8h EI (or pip-taz 4.5 g q6h EI if ESBL risk low) | MRSA risk: vancomycin or linezolid; influenza season: oseltamivir |
| Urinary | Ceftriaxone 2 g OD | Meropenem 1 g q8h ± amikacin | Obstruction: urgent decompression |
| Intra-abdominal | Ceftriaxone 2 g OD + metronidazole 500 mg q8h | Meropenem ± echinocandin if recurrent perforation | Source control, ideally within 6 h (SSC 2026) |
| Necrotising soft tissue | Pip-taz or meropenem + clindamycin 600-900 mg q8h | Meropenem + vancomycin + clindamycin | Surgery now |
| Meningitis | Ceftriaxone 2 g q12h + vancomycin (+ ampicillin 2 g q4h if > 50 y or immunocompromised) | Meropenem 2 g q8h + vancomycin | Dexamethasone 0.15 mg/kg q6h with or before first dose |
| Septic shock, focus unclear (ICMR) | Carbapenem ± amikacin | Carbapenem ± amikacin ± polymyxin B/colistin ± vancomycin/teicoplanin | Doxycycline for tropical cover; echinocandin if Candida risk |
Templates only. Many Indian ICUs report ESBL rates above 50% in hospital-onset infection and frequent carbapenemases (NDM, OXA-48-like). EI = extended infusion.
Calculators for this topic
References
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
Empirical antibiotic decision
Step 1 of ~8
Decision: Shock or probable sepsis?
Based on Prescott HC, Antonelli M, Alhazzani W, et al; Indian Council of Medical Research. Educational use; follow local protocol.
Evidence
- Kumar 2006: in septic shock, each hour of delay after hypotension onset was associated with ~7.6% lower survival (observational).
- Seymour 2017 (New York sepsis mandate): each hour to bundle completion and antibiotics increased risk-adjusted mortality.
- SSC 2026 timing (unchanged from 2021): antibiotics immediately, ideally within 1 h, for possible, probable or definite septic shock and for probable or definite sepsis without shock (strong, very low certainty). Possible sepsis without shock: time-limited rapid investigation, antibiotics within 3 h if concern persists (conditional). Low likelihood of infection without shock: defer and monitor (conditional). Procalcitonin should not decide whether to start (conditional).
- Prolonged beta-lactam infusion upgraded to strong (moderate certainty): an 18-RCT meta-analysis (9,108 patients, including BLING III) found lower short-term mortality (RR 0.91; about 25 fewer deaths per 1,000). Give a loading dose first; mind line compatibility.
- New in 2026 (all conditional, very low or low certainty): empirical cover for a specific MDR pathogen only when its risk is high; no anaerobic cover without risk factors; no routine empirical antifungal; no Candida biomarkers (e.g. beta-D-glucan) to decide on starting antifungals; pathogen-specific rapid tests and antimicrobial TDM case by case; prehospital antibiotics for septic shock if evaluation is > 60 min away.
- MERINO (2018): pip-taz was inferior to meropenem for ceftriaxone-resistant E. coli or Klebsiella bacteraemia (30-day mortality 12.3% vs 3.7%).
- BLING III (2024): continuous infusion did not significantly reduce 90-day mortality; a companion meta-analysis suggested lower mortality with prolonged infusion.
- Polymyxins (2019 consensus): polymyxin B preferred for systemic infection; colistin for lower UTI.
Controversies
- Empirical double gram-negative cover (e.g. adding amikacin) is reasonable in very high-resistance units until cultures return, but adds nephrotoxicity.
- Empirical antifungals: only with risk factors (prolonged broad-spectrum antibiotics, TPN, recurrent abdominal perforation, multisite Candida colonisation, central lines).
- Beta-lactam therapeutic drug monitoring: SSC 2026 suggests it case by case, but it is rarely available in India.
- SSC 2021 suggested two gram-negative agents only when MDR risk is high. SSC 2026 frames this instead as cover for a specific MDR organism by risk; the guideline leaves agent choice to IDSA/ESCMID guidance and local data.
🇮🇳 India notes
- ICMR AMR surveillance (AMRSN) reports high carbapenem resistance in Klebsiella pneumoniae and Acinetobacter baumannii in Indian ICUs, with NDM and OXA-48-like enzymes predominant. This makes ceftazidime-avibactam + aztreonam important for NDM producers.
- ICMR 2019 suggests 7-10 days for most sepsis and allows procalcitonin to support shortening.
- Cost matters: ceftazidime-avibactam is expensive; polymyxin B is cheaper but more toxic.
References
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
Test yourself
Module quiz · score 70% to complete the module
Educational reference for trained clinicians. Check doses against your hospital protocol and the patient in front of you.