septic shock

in review 8 min read Updated 2026-10-30
septic shock

Sepsis with vasopressors needed for MAP ≥65 mmHg and lactate >2 mmol/L despite fluid. Cultures → antibiotics within 1 h → ≥30 mL/kg balanced crystalloid over 3 h → norepinephrine (peripheral is fine) to MAP 65 (60–65 if ≥65 y). Escalate norepinephrine → + vasopressin → + epinephrine; septic cardiomyopathy with persistent hypoperfusion → add dobutamine. Hydrocortisone 200 mg/day for any vasopressor-dependent septic shock. After the first 30 mL/kg, liberal and restrictive strategies are both acceptable; guide further fluid with dynamic measures (conditional). Antibiotic choice, source hunt and duration: sepsis of unknown source.

septic shock — first 6 hours
  • blood cultures × 2, lactate → empiric antibiotics within 1 h — see sepsis of unknown source
  • balanced crystalloid (Ringer’s lactate) ≥30 mL/kg within 3 h — actual body weight; adjusted/ideal weight if BMI >30
  • unstable (mottled, altered, severe hypotension): start norepinephrine with the fluid, not after it
  • norepinephrine peripherally via good large-bore IV while central access is arranged
  • MAP target 65 mmHg; 60–65 if age ≥65
  • arterial line if escalating vasopressors or cuff readings unreliable
  • repeat lactate at 2–4 h; capillary refill time with each reassessment
  • guide further fluid by responsiveness: passive leg raise or 500 mL bolus with stroke-volume check
  • norepinephrine rising (~0.25–0.3 µg/kg/min) → add vasopressin 0.03 U/min (fixed dose)
  • hydrocortisone 50 mg IV q6h once on vasopressors
  • POCUS/echo if escalating: LV/RV function, effusion, IVC
  • source control (drain, debride, remove line) within 6 h of diagnosis
  • ICU within 6 h
  • goals of care conversation early
septic shock — vasopressor order

MAP <65 despite initial fluid. Unstable → start norepinephrine with the fluid, not after it.

first
  • norepinephrine to MAP 65 (60–65 if ≥65 y) — peripheral IV is fine to start
then / if needed
  • add vasopressin 0.03 U/min, fixed, as norepinephrine climbs (~0.25–0.3 µg/kg/min)
  • add epinephrine if MAP is still inadequate
  • LV dysfunction with persistent hypoperfusion → add dobutamine
  • hydrocortisone 50 mg IV q6h once on vasopressors
avoid

first hour

Sepsis is vasoplegia + venodilation + capillary leak ± myocardial depression. The first litres restore venous return; beyond preload responsiveness, extra fluid leaks into lungs and gut. Vasopressors fix the vascular tone that fluid can’t.

  • screen with NEWS, MEWS or SIRS, not qSOFA alone (strong): qSOFA is too insensitive to screen. Lactate supports the diagnosis but no single test rules sepsis in or out
  • hypoperfusion = MAP <65, SBP <90, SBP well below baseline, or raised lactate. Lactate 2–4 mmol/L also counts (it carries excess mortality); ≥4 is not the threshold for fluid
  • ≥30 mL/kg crystalloid in 3 h (conditional, low certainty; observational basis). Harm from fluid mostly appears beyond ~50 mL/kg. Reassess after each bolus
  • fluid then vasopressor is the default; in unstable shock run them together
  • no protocolised CVP/ScvO₂ targets: early goal-directed therapy was no better than usual care (ProCESS (2014))
  • antibiotics within 1 h of recognising shock; cultures first only if that adds no delay
  • source control within 6 h once an anatomical source is found (abscess, obstructed tract, infected line, necrotising soft-tissue infection)

fluids after the first 30 mL/kg

  • liberal vs restrictive: no mortality difference (CLOVERS (2023): early vasopressors vs more fluid; CLASSIC (2022): restriction in the ICU). Either is acceptable. In practice, reach for norepinephrine early rather than chase MAP with litres
  • guide with dynamic measures over exam or CVP: passive leg raise, stroke-volume change on echo or arterial waveform after 500 mL (responsive = ≥10–15% rise), pulse-pressure variation (only in passive, fully ventilated patients in sinus rhythm)
  • capillary refill time (normal <3 s) and serial lactate as adjuncts. Epinephrine raises lactate through β2 stimulation, not hypoperfusion, so read the trend with that in mind
  • balanced crystalloid over 0.9% saline (conditional, moderate certainty): meta-analysis probably reduces mortality (OR 0.94) and new RRT (OR 0.86). SMART (2018) showed fewer major adverse kidney events; PLUS (2022) found no difference. Ringer’s lactate costs the same as saline
  • albumin: crystalloid alone is preferred. Albumin adds cost without a mortality gain (SAFE, ALBIOS). Consider it after large crystalloid volumes or in cirrhosis; avoid in traumatic brain injury. Canadian Blood Services: alternatives first in critical illness
  • no starches (strong: more death and RRT) and no gelatin
  • after the acute phase: once shock resolves, de-resuscitate with diuretics or ultrafiltration

vasopressors and inotropes

flowchart TD
  A[Septic shock after initial fluid] --> B[Norepinephrine<br/>peripheral OK · MAP 65]
  B -->|rising dose ~0.25–0.3 µg/kg/min| C[Add vasopressin 0.03 U/min]
  C -->|MAP still inadequate| D[Add epinephrine]
  B --> E{Persistent hypoperfusion<br/>despite MAP + volume?}
  E -->|LV dysfunction on echo| F[Add dobutamine<br/>or switch to epinephrine alone]
  D --> G[Refractory: check source, echo,<br/>steroids given? goals of care]
  • norepinephrine first line (strong): fewer arrhythmias than dopamine (SOAP II (2010)) and better data than epinephrine, vasopressin or angiotensin II. Peripheral start is preferred to delay. There is no evidence on safe dose or duration; check the site often and move centrally if it’s needed beyond the first hours
  • MAP 65, not higher (strong). Age ≥65: 60–65 (conditional). 65 Trial (2020): permissive MAP 60–65 cut vasopressor exposure, with a non-significant mortality trend favouring it (larger in chronic hypertension). Arterial line or cuff both acceptable (conditional); most place a line once pressors escalate
  • vasopressin as second agent (conditional, moderate certainty): meta-analysis RR 0.89 for mortality, less AF, possibly more digital ischaemia. VASST (2008): no overall difference; possible benefit in less severe shock, so add early rather than as a last resort. Fixed 0.03 U/min (= 1.8 U/h), not titrated
  • epinephrine third (conditional, very low certainty). Without vasopressin, add it straight to norepinephrine
  • septic cardiomyopathy: global LV hypokinesis with ongoing hypoperfusion despite adequate MAP and filling. Add dobutamine to norepinephrine, or use epinephrine alone (conditional). Dobutamine often raises MAP through cardiac output despite dropping SVR. A plethoric IVC and B-lines mean more fluid will only cause harm. Milrinone vasodilates more and has a long, renally cleared half-life, so it’s a poor fit for vasoplegia. No levosimendan
  • avoid: dopamine (norepinephrine preferred: strong, high certainty); terlipressin and β-blockers to control rate (conditional against). Methylene blue: insufficient evidence; it raises BP but hasn’t been shown to improve survival, so rescue only. Angiotensin II is not first line. Midodrine to wean low-dose pressors is unproven

Doses and receptor profiles: vasoactive drugs.


corticosteroids

  • hydrocortisone 50 mg IV q6h (200 mg/day) for septic shock on vasopressors at any dose (conditional, low certainty; SCCM 2024 and SSC 2026). The older “norepinephrine ≥0.25 µg/kg/min for >4 h” trigger has been dropped. Most panellists start within 4 h or below ~0.3 µg/kg/min
  • no high-dose, short courses (>400 mg/day for <3 days; strong against)
  • mechanism: restores vascular responsiveness to catecholamines, and treats relative adrenal insufficiency without needing a cortisol level or ACTH test
  • effect: faster shock reversal (RR ~1.25, high certainty) and fewer days of organ dysfunction. Mortality gain is small and uncertain (short-term RR 0.93). Costs: hypernatraemia, hyperglycaemia, possibly neuromuscular weakness
  • trials: ADRENAL (2018) (n=3800, hydrocortisone alone): no 90-day mortality difference, faster shock resolution. APROCCHSS (2018) (hydrocortisone + fludrocortisone 50 µg, sicker patients): lower 90-day mortality. Whether the gain came from fludrocortisone or from the sicker population is unresolved; no guideline recommends fludrocortisone, so adding it is optional
  • duration: trials used 5–7 days. Once off vasopressors, most stop without a taper
  • sepsis without shock: no steroids, unless the patient also has severe CAP or ARDS, which carry their own steroid indications (community-acquired pneumonia, ards)

supportive care

issueapproach
oxygenliberal or conservative SpO₂ targets both acceptable; HFNC over conventional O₂ and over NIV for hypoxaemic failure → noninvasive respiratory support
ventilationlung-protective → ards
daily ICU bundlerestrictive transfusion, insulin for hyperglycaemia, LMWH prophylaxis, stress ulcer prophylaxis if at risk, early enteral feeding → icu supportive care
acidaemiano bicarbonate for lactic acidosis; consider only if pH ≤7.2 with AKI stage 2–3 → sodium bicarbonate
AKIRRT only for a definitive indication; CRRT or IHD equally → acute kidney injury
feverno antipyretics or cooling to improve outcome (comfort is fine)
no benefitvitamin C, IVIG, vitamin D, polymyxin B haemoperfusion, blood purification, probiotics

refractory shock

Escalating pressors despite the above → go back to basics before adding drugs:

  • source: undrained collection, necrotising infection, infected device, wrong antimicrobial (resistance, fungal, viral) → re-image, repeat cultures, broaden. See sepsis of unknown source
  • echo: septic cardiomyopathy, RV failure, tamponade, dynamic LVOT obstruction (hyperdynamic, underfilled heart on inotropes → stop inotrope, give fluid)
  • other diagnosis or co-shock: haemorrhage, adrenal crisis (steroids already cover it), anaphylaxis, abdominal compartment syndrome, toxins
  • acidaemia, hypocalcaemia: both blunt the vasopressor response
  • goals of care: revisit early; involve palliative care where survival is unlikely

Prepared 2026-10-30. Not yet verified — check every number against the sources.

Key references

+9 more sources

Suggest a correction septic shock