Fluid resuscitation in the emergency department

Fluid resuscitation is the rapid restoration of the circulating volume to reverse tissue hypoperfusion, and it is the commonest therapeutic intervention performed in the resuscitation bay — and yet it is also among the most over-used and least-thought-through. The Fellowship candidate must master three things: the choice of fluid (a balanced crystalloid is the default, colloids are abandoned, albumin is a rescue agent with a brain-injury caveat), the dose and the target (an aliquoted bolus given only to a fluid-responder, titrated to a mean arterial pressure, a lactate clearance and a urine output), and the evidence (SMART, SAFE, Perel, FEAST).[2][4][5]

[1]

Definition, classification and the compartments

Fluid resuscitation is the rapid administration of an intravenous fluid to restore the intravascular volume and so the cardiac output and the oxygen delivery. It is distinct from maintenance fluid (the long-term water and electrolyte requirement) and from replacement fluid (the matching of ongoing measurable losses). The resuscitation itself is phased — rescue (the life-saving bolus in shock), optimisation (the titrated phase of stabilisation), stabilisation, and de-resuscitation (the later diuretic or ultrafiltration phase to remove the excess).[4]

The total body water is around 60 per cent of the body weight — roughly 42 L in a 70 kg adult — and it is distributed two-thirds intracellular (28 L) and one-third extracellular (14 L). The extracellular fluid splits further into an interstitial compartment (around 10.5 L) and an intravascular plasma compartment (around 3.5 L). The critical implication for resuscitation: an isotonic crystalloid redistributes across the whole extracellular space within minutes, so only around 20 to 30 per cent of a crystalloid bolus remains in the vessels at 30 minutes. This is the physiological basis for the colloid appeal (colloids hold fluid intravascularly) — and, as the evidence shows, also the basis for their failure to improve the outcome. [1]

[1]

The fluid types — crystalloid, colloid, balanced, saline

The first decision is the agent. The crystalloids are the default; the colloids are reserved and largely abandoned. [1]

[1]

The take-home is simple: a balanced crystalloid is the resuscitation fluid of choice in 2026. Saline is reserved for the disease-specific niches (the DKA first hour, the hyponatraemia). Albumin is a rescue agent with a brain-injury contraindication. Starch is gone.[2][3][5]

The targets — what a successful resuscitation looks like

The resuscitation is titrated to perfusion endpoints, not to a single number. [1]

[1]

The capillary refill time, the mottling and the mental state are the bedside perfusion signs; the mean arterial pressure, the lactate clearance and the urine output are the quantitative targets. The venous-to-arterial carbon-dioxide gap (over 6 mmHg) flags ongoing hypoperfusion even when the lactate has normalised. The candidate must avoid the trap of treating a number: a normal lactate with a cold periphery is not a resuscitated patient. [1]

Fluid responsiveness — the gatekeeper of every bolus

Only around half of haemodynamically unstable patients are fluid-responsive, and a bolus given to a non-responder does harm without benefit. The static markers — the central venous pressure, a single inferior-vena-cava diameter — do not predict responsiveness and should not be used. The dynamic tests do. [1]

The passive leg raise is the most reliable bedside test: the legs are lifted to 45 degrees, auto-transfusing around 300 mL of venous blood as a reversible self-challenge, and the cardiac output (or a surrogate — the pulse pressure, the end-tidal CO2) is watched for a rise of 10 per cent or more over the next minute. A positive response predicts that a fluid bolus will raise the stroke volume. The pulse-pressure variation and the stroke-volume variation apply to the ventilated, deeply sedated patient: a variation over 12 to 13 per cent predicts responsiveness. The inferior vena cava variability on ultrasound — a caval index over 50 per cent in the spontaneously breathing patient, or a distensibility over 18 per cent in the ventilated patient — is a third dynamic test. The simplest of all is the fluid challenge itself: a 250 mL aliquot of crystalloid over 5 to 10 minutes with the response watched — and stopped the moment responsiveness is lost.[4]

Fluid responsiveness in depth — the four dynamic tests

The static indices are abandoned because they fail the cardinal test of a predictor: the area under the receiver-operating curve for the central venous pressure against a subsequent fluid response is around 0.55 — no better than a coin toss — and a single CVP reading, or a single IVC diameter, cannot place the patient on the Frank-Starling curve. The four dynamic tests each impose a reversible change in the venous return and observe the cardiac-output response; the test is the surrogate for the bolus, and a positive test predicts that a real bolus will raise the stroke volume.[15]

[1] [1] [1]

Differential diagnosis — when is fluid not the answer?

The shocked patient has a differential, and the bedside echocardiogram, the fluid response and the history resolve it. The cardinal question before every bolus: is the patient on the steep (responsive) or the flat (unresponsive) part of the Starling curve, and is fluid even the right therapy? [1]

[1]

Immediate management — the resuscitation ladder

Stabilise the airway and the breathing, and establish two large-bore cannulae (or an intraosseous line if collapse). Then run the resuscitation ladder: an aliquoted fluid challenge, a responsiveness reassessment, and an early vasopressor. The cardinal principle — give fluid only to a responder, and start the vasopressor early. [1]

[1]

The fluid of choice is a balanced crystalloid — Hartmann's, lactated Ringer's or Plasma-Lyte — because the SMART trial showed a lower 30-day in-hospital mortality and less acute kidney injury than saline.[4] The vasopressor of choice is noradrenaline, started early; peripheral noradrenaline in a large proximal vein is safe for short periods and should not delay resuscitation while a central line is sited. The colloid rescue: albumin 4%, 250 to 500 mL aliquots, is reasonable in the septic-shock patient with a low albumin, but offers no mortality benefit over saline (SAFE)[2] and is harmful in traumatic brain injury (SAFE-TBI).[3] Hydroxyethyl starch is not used.[5]

The SMART trial and the balanced-crystalloid evidence

The SMART trial (Semler 2018) randomised over 15,000 critically ill adults to balanced crystalloids versus saline and showed a lower 30-day in-hospital mortality (10.3 per cent versus 11.1 per cent) and less renal-replacement therapy or persistent renal dysfunction with the balanced crystalloid.[4] The mechanism is the saline-induced hyperchloraemia: 0.9 per cent saline carries 154 mmol/L of chloride, which causes a hyperchloraemic metabolic acidosis, renal afferent arteriolar vasoconstriction, and an acute kidney injury. The SALT-ED trial in non-critically-ill adults reached the same conclusion. The contemporary default is therefore a balanced crystalloid, with saline reserved for the DKA first hour and the symptomatic hyponatraemia.

The balanced-crystalloid signal is large in the critically ill (SMART) but the dedicated balanced-versus-saline trials have been more equivocal, and the Fellowship candidate must hold both truths at once. SPLIT (Young 2015), a New Zealand single-centre-before-after cluster trial in 2278 patients, found no difference in AKI between buffered crystalloid and saline — but it was underpowered and its saline arm was small. BaSICS (Zampieri 2021), a 11 052-patient Brazilian multicentre trial of Plasma-Lyte 148 versus saline, found no difference in 90-day mortality — but its intervention was low-volume (mostly 1 L or less) and the trial's generalisability to the high-volume resuscitation of shock is debated. PLUS (Finfer 2022), the 5037-patient Australian trial of a balanced multielectrolyte solution versus saline, likewise found no mortality difference. The reconciliation: in the high-volume resuscitation of the critically ill (SMART, SALT-ED), the chloride load matters and the balanced crystalloid wins; in the low-volume maintenance setting (BaSICS, PLUS), the difference washes out. The pragmatic bottom line — default to a balanced crystalloid for resuscitation, because it never causes the hyperchloraemic harm that saline can, and the dedicated trials have not overturned the SMART signal.[9][10][11]

[1]

The colloid evidence — SAFE, SAFE-TBI and the abandoned starch

The SAFE trial (Finfer 2004) settled the albumin question: albumin 4 per cent and saline produced the same 28-day mortality in a heterogeneous ICU population, so albumin offered no advantage as a routine resuscitation fluid.[2] The SAFE-TBI substudy (2007) then revealed the one place albumin is actively harmful — in traumatic brain injury, albumin resuscitation carried a higher mortality than saline, and albumin is contraindicated in the head-injured patient.[3] The synthetic colloids fared worse still: the CHEST and 6S trials showed that hydroxyethyl starch increased the acute kidney injury and the need for renal-replacement therapy with no mortality benefit, and the Perel Cochrane review confirmed that no colloid offers an outcome advantage over crystalloid.[5] Starch is abandoned; gelatin and dextrans share the no-benefit and the harm profile.

The Fellowship candidate should be able to recite the two starch trials by name and finding, because the starch question is a recurring viva staple and its answer is unambiguous. CHEST (Myburgh 2012, the Crystalloid versus Hydroxyethyl Starch Trial) randomised 7000 ICU patients to 6 per cent HES 130/0.4 versus saline and found no mortality difference but a significant increase in the need for renal-replacement therapy (7.0 per cent versus 5.8 per cent). 6S (Perner 2012, the Scandinavian Starch for Severe Sepsis/Septic Shock trial) randomised 800 severe-sepsis patients to 6 per cent HES 130/0.42 versus Ringer's acetate and found a higher 90-day mortality (51 per cent versus 43 per cent) and more RRT. Together with the Perel Cochrane review, these two trials killed the synthetic colloid: HES was withdrawn from the European market for use in sepsis in 2013, and is not stocked for resuscitation in modern practice.[7][8]

Differential fluid strategies by disease

The dose and the agent are tailored to the disease. Sepsis: a balanced crystalloid 30 mL/kg in aliquots within the first three hours (the Surviving Sepsis starting dose), then noradrenaline; this aggressive-fluid paradigm of Rivers EGDT has since been tempered by ProCESS, ARISE and ProMISe, which showed that protocolised EGDT was not superior to usual care — the modern practice is aliquoted fluid with early responsiveness assessment.[1][14] Trauma: damage-control — minimise crystalloid, give blood products in a 1:1:1 ratio, target permissive hypotension (SBP 80 to 90) until the bleeding is controlled, and tranexamic acid within three hours. DKA: 0.9 per cent saline 10 to 20 mL/kg in the first hour (the one place saline is first-line, to address the sodium deficit), then switch to a balanced crystalloid once the sodium stabilises; fluids run with the insulin and the potassium, not as a separate step. Burns: the Parkland formula — lactated Ringer's (Hartmann's) at 3 to 4 mL per kilogram per per cent of total body surface area in the first 24 hours, half in the first 8 hours from the time of the burn; the ANZ modification uses 2 to 4 mL/kg/% with the early escharotomy cues. Paediatric sepsis: 10 to 20 mL/kg aliquots — the FEAST trial showed that fluid boluses increased mortality in resource-limited African children without shock, a caution against the blind bolus in the non-shocked febrile child.[6]

Permissive hypotension and damage-control resuscitation

In the actively bleeding trauma patient, the resuscitation paradigm inverts: the goal is not a normal blood pressure, and the goal is not the crystalloid. Damage-control resuscitation (DCR) minimises the crystalloid, delivers the blood products in a balanced ratio, permits a lower-than-normal blood pressure until the bleeding is surgically controlled, and corrects the traumatic coagulopathy with the blood products themselves rather than with crystalloid. The physiological rationale: the uncontrolled haemorrhage patient who is resuscitated to a normal pressure before surgical control simply bleeds faster — the raised pressure blows off the soft clot, dilutes the clotting factors with crystalloid, and cools the patient, all of which worsen the coagulopathy and the death. [1]

The seminal evidence is Bickell 1994 (NEJM): 598 patients with penetrating torso trauma and a systolic blood pressure below 90 were randomised to immediate versus delayed (until operating-room control) fluid resuscitation. The delayed group received less fluid, had fewer complications, and a lower mortality (62 per cent versus 70 per cent). The trial has caveats — it is penetrating trauma in a single urban centre, and its message does not extend to the blunt trauma or the head-injured patient — but it established the principle that the crystalloid before surgical control is not always beneficial.[17]

[1] [1]

Fluid creep, positive balance and de-resuscitation

The most important conceptual shift in fluid therapy over the last decade is the recognition that the fluid that helps at hour one harms at hour twelve. Fluid creep — the insidious accumulation of small, individually defensible boluses, drug-dilution volumes and carrier infusions — drives a cumulative positive balance that is an independent predictor of mortality in the critically ill. The modern resuscitation is therefore phased: a rescue phase of aliquoted boluses to a responsiveness ceiling, a stabilisation phase of minimal maintenance, and a de-resuscitation phase of active fluid removal once the shock has resolved. [1]

Two trials reframed the sepsis fluid target. CLASSIC (Meyhoff 2022) randomised 1554 septic-shock ICU patients to a restrictive versus a liberal fluid strategy after the initial resuscitation; the restrictive group received a median of 1.2 L versus 3.0 L in the days after randomisation, and the restrictive strategy was safe — no increase in death, kidney failure, or ischaemic events, and a consistent trend toward less harm. CLOVERS (Shapiro 2023, the PETAL Network) randomised 1563 patients with sepsis-induced hypotension to an early-restrictive (less fluid, earlier vasopressors) versus a liberal fluid strategy in the first 24 hours; the restrictive strategy was non-inferior for 90-day mortality, again with no excess of adverse events. The synthesis: after the initial 1 to 2 L of resuscitation, additional fluid does not help, and an earlier vasopressor with less fluid is an acceptable and probably preferable strategy.[12][13]

[1] [1] [1] [1]

Paediatric fluid resuscitation — the FEAST reframing

The paediatric fluid bolus was once reflexive — every febrile, tachycardic child with a suspected infection received the 20 mL/kg bolus. The FEAST trial (Maitland 2011) overturned that reflex in the resource-limited setting and reshaped the paediatric resuscitation globally.[6]

The FEAST caveats matter for the exam. The trial was conducted in resource-limited African hospitals without the mechanical ventilation, the intensive care, or the rapid transfusion available in a high-income emergency department; the children were largely in compensated shock (not the overt hypotensive shock of the western ICU); and the harm signal has not been reproduced in the ventilated, ICU-resourced setting. The contemporary paediatric practice (the Surviving Sepsis Pediatric 2020 guideline) is therefore: identify the true septic shock (the cold, poorly perfused, weak-pulse, altered-mental-state child), give the 10 to 20 mL/kg balanced-crystalloid bolus in aliquots to the responder, watch for the fluid creep and the pulmonary oedema, and do not reflexively bolus the febrile child with the isolated tachycardia. The lessons for the Fellowship viva — the fluid bolus is a therapy with a harm profile, not a reflex; the responsiveness assessment applies to the child as to the adult; and the over-resuscitation of the child is as dangerous as the under-resuscitation. [1]

[1] [1]

Complications and pitfalls

The complications of over-resuscitation are now recognised as a major cause of morbidity: a pulmonary oedema, an abdominal compartment syndrome (an intra-bladder pressure over 20 mmHg that compresses the kidneys and the gut), a tissue oedema (the gut wall, the brain — fluid worsens the intracranial pressure in TBI), and an acute kidney injury from the renal oedema. The hyperchloraemic metabolic acidosis from saline (chloride 154, pH 7.28, base excess minus 8) mimics the lactic acidosis and worsens the AKI; the treatment is to switch to a balanced crystalloid, not to give bicarbonate. The dilutional coagulopathy from crystalloid in haemorrhage is part of the rationale for the damage-control blood-product strategy. The pitfalls: giving fluid to a non-responder; chasing a central-venous-pressure target; saline when a balanced crystalloid is available; under-resuscitating the bleeding patient with crystalloid instead of activating the massive-haemorrhage protocol; withholding noradrenaline until a fluid ceiling is reached; and forgetting that albumin is harmful in TBI. [1]

Prognosis, disposition and special populations

The outcome is the disease's outcome, but the fluid strategy modulates it — a persistently positive fluid balance over the first 48 to 72 hours of septic shock is independently associated with mortality, and de-resuscitation (the later diuresis or ultrafiltration) is now standard critical-care practice. The shocked patient is admitted to resus then the intensive care; the DKA patient to a monitored bed with an insulin-infusion protocol; the major burn to a burns centre. The elderly have less reserve and are easily overloaded — smaller aliquots and an early vasopressor. The chronically hypertensive patient needs a higher MAP target (80 to 90). The cardiac patient needs a bedside echo first and small 250 mL aliquots. The traumatic-brain-injury patient must stay euvolaemic and must not receive albumin. The pregnant patient is managed with a left-lateral tilt to relieve the aortocaval compression. The paediatric patient receives weight-based aliquots with the FEAST caution against the blind bolus in the non-shocked febrile child.[6]

Evidence and regional guidelines

The contemporary framework rests on six trials. Rivers EGDT (2001) introduced the aggressive early goal-directed therapy and the early fluid targets.[1] The modern Surviving Sepsis 2021 guideline retains at least 30 mL/kg of crystalloid in the first three hours (a weak, low-quality recommendation) but balances it with the early noradrenaline and the responsiveness assessment. SMART (2018)[4] and SALT-ED established the balanced-crystalloid default. SAFE (2004)[2] settled the albumin question (no benefit) and SAFE-TBI (2007)[3] added the brain-injury contraindication. Perel (2013)[5] confirmed no colloid advantage. FEAST (2011)[6] tempered the paediatric bolus reflex.

[1] [1]

Exam pearls

• The fluid of choice is a balanced crystalloid (SMART) — Hartmann's, lactated Ringer's, Plasma-Lyte.
• Give fluid only to a responder — test with a passive leg raise; the static markers (CVP, a single IVC) do not predict it.
• The sepsis starting dose is 30 mL/kg crystalloid in aliquots, then noradrenaline early — do not chase a fluid ceiling.
• Saline in volume causes a hyperchloraemic metabolic acidosis (chloride 110-plus, pH 7.28) — recognise it and switch.
• Albumin equals saline overall (SAFE) but is harmful in traumatic brain injury (SAFE-TBI).
• Hydroxyethyl starch is abandoned (CHEST, 6S, Perel) — more AKI and renal-replacement therapy, no mortality benefit.
• In trauma, minimise the crystalloid — blood products 1:1:1, permissive hypotension, tranexamic acid within three hours.
• DKA — saline 10 to 20 mL/kg in the first hour, then balanced; fluids run with insulin and potassium.
• Burns — Parkland: Hartmann's 3 to 4 mL/kg/%TBSA, half in the first 8 hours from the time of the burn.
• A positive fluid balance predicts mortality — de-resuscitate later.
• The dedicated balanced-versus-saline trials (SPLIT, BaSICS, PLUS) were null — but the pragmatic default remains the balanced crystalloid, because it is never more harmful than saline.
• The fluid challenge is a true bolus (250 mL over 5 to 10 minutes) with a declared target and a measured response — a slow infusion is not a challenge.
• Permissive hypotension (SBP 80 to 90) is for the bleeding patient UNTIL surgical control — never for the traumatic brain injury, which needs a MAP of at least 80 (Bickell 1994).
• Fluid creep is the slow killer — CLASSIC and CLOVERS showed a restrictive strategy with an earlier vasopressor is safe and preferable after the initial resuscitation.
• FEAST did not abolish the paediatric bolus — it reframed it: bolus the true shock, not the febrile-compensated child; reassess after every aliquot.
• The CVP is dead as a fluid target (ROC 0.55) — use the dynamic tests (PLR, fluid challenge, IVC variability, pulse-pressure variation). [1]

Red flags

[1]

References

1. [1]Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock N Engl J Med, 2001.PMID 11794169
2. [2]Finfer S, Bellomo R, Boyce N, et al. A comparison of albumin and saline for fluid resuscitation in the intensive care unit N Engl J Med, 2004.PMID 15163774
3. [3]SAFE Study Investigators. Saline or albumin for fluid resuscitation in patients with traumatic brain injury N Engl J Med, 2007.PMID 17761591
4. [4]Semler MW, Self WH, Wanderer JP, et al. Balanced Crystalloids versus Saline in Critically Ill Adults N Engl J Med, 2018.PMID 29768150
5. [5]Perel P, Roberts I, Ker K. Colloids versus crystalloids for fluid resuscitation in critically ill patients Cochrane Database Syst Rev, 2013.PMID 23450531
6. [6]Maitland K, Kiguli S, Opoka RO, et al. Mortality after fluid bolus in African children with severe infection N Engl J Med, 2011.PMID 21615299
7. [7]Myburgh JA, Finfer S, Bellomo R, et al. Hydroxyethyl starch or saline for fluid resuscitation in intensive care N Engl J Med, 2012.PMID 23075127
8. [8]Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl starch 130/0.42 versus Ringer's acetate in severe sepsis N Engl J Med, 2012.PMID 22738085
9. [9]Young P, Bailey M, Beasley R, et al. Effect of a Buffered Crystalloid Solution vs Saline on Acute Kidney Injury Among Patients in the Intensive Care Unit: The SPLIT Randomized Clinical Trial JAMA, 2015.PMID 26444692
10. [10]Zampieri FG, Machado FR, Veiga VC, et al. Effect of Intravenous Fluid Treatment With a Balanced Solution vs 0.9% Saline Solution on Mortality in Critically Ill Patients: The BaSICS Randomized Clinical Trial JAMA, 2021.PMID 34375394
11. [11]Finfer S, Micallef S, Hammond N, et al. Balanced Multielectrolyte Solution versus Saline in Critically Ill Adults N Engl J Med, 2022.PMID 35041780
12. [12]Meyhoff TS, Hjortrup PB, Wetterslev J, et al. Restriction of Intravenous Fluid in ICU Patients with Septic Shock N Engl J Med, 2022.PMID 35709019
13. [13]National Heart, Lung, and Blood Institute PETAL Network; Shapiro NI, et al. Early Restrictive or Liberal Fluid Management for Sepsis-Induced Hypotension N Engl J Med, 2023.PMID 36688507
14. [14]The ProCESS Investigators; Yealy DM, Kellum JA, Huang DT, et al. A randomized trial of protocol-based care for early septic shock N Engl J Med, 2014.PMID 24635773
15. [15]Marik PE, Cavallazzi R, Vasu T, Hirani A. Dynamic changes in arterial waveform derived variables and fluid responsiveness in mechanically ventilated patients: a systematic review of the literature Crit Care Med, 2009.PMID 19602972
16. [16]Jabot J, Teboul JL, Richard C, Monnet X. Passive leg raising for predicting fluid responsiveness: importance of the postural change Intensive Care Med, 2009.PMID 18795254
17. [17]Bickell WH, Wall MJ Jr, Pepe PE, et al. Immediate versus delayed fluid resuscitation for hypotensive patients with penetrating torso injuries N Engl J Med, 1994.PMID 7935634