Fluid responsiveness: when more fluid helps

Assessment

Before giving more fluid, the ICU team checks whether the heart will actually pump more blood with it, because extra fluid can cause harm.

Updated Nov 13, 2026

What it is

A patient is fluid responsive when a fluid bolus raises stroke volume or cardiac output meaningfully, usually by about ten to fifteen per cent. This happens when both ventricles work on the steep part of the Frank-Starling curve, where more filling still produces more output. Only about half of unstable critically ill patients respond, and giving fluid to the rest only adds oedema in the lungs, gut, kidneys and skin, which is linked to more days on the ventilator, kidney injury and death. Static values such as central venous pressure or a single ultrasound image of the inferior vena cava predict responsiveness poorly. Dynamic tests work better: the passive leg raise moves blood from the legs into the chest and acts as a reversible fluid challenge; pulse pressure and stroke volume variation measure how much output changes with each mechanical breath; and a small fluid challenge with real-time measurement of output tests the question directly. Each test has conditions that make it unreliable. Being responsive does not mean fluid is needed: fluids are given only when there is hypoperfusion, a likely response and enough tolerance. The choice of fluid matters too: balanced crystalloids are preferred for most patients, large amounts of saline cause hyperchloraemic acidosis, and starches harm the kidneys.

What to assess

Is there hypoperfusion that fluid could fix (low pressure, high lactate, slow capillary refill, oliguria)?
Is the patient likely to respond (dynamic test positive)?
Can the patient tolerate more fluid (no rising oxygen need, B-lines, venous congestion or rising abdominal pressure)?
Which fluid: balanced crystalloid for most, blood for bleeding, albumin in selected cases; avoid starches.
Which phase of illness: early resuscitation, optimisation, stabilisation or removal of excess fluid.

How it is done (and why)

Confirm the problem is hypoperfusion, not a high lactate from another cause — fluid only helps when the circulation is short of volume.
Check tolerance: oxygen needs, lung ultrasound B-lines, venous congestion and cumulative balance — a responder can still be harmed.
Set up a real-time measure of stroke volume or cardiac output (echo or pulse contour) — the test is read on flow, not pressure.
Do a passive leg raise from the semi-recumbent position and read the change within about a minute — it is reversible and gives no fluid.
If ventilated with controlled breaths in sinus rhythm, read pulse pressure or stroke volume variation — large swings suggest a responder.
If unsure, give a small, rapid fluid challenge and measure output again — this tests the question directly with limited volume.
Give fluid in boluses only while the patient keeps responding, and stop when the response is lost — after that fluid only adds oedema.
Prefer a balanced crystalloid and record every bolus in the fluid balance — saline causes hyperchloraemic acidosis and hidden fluid adds up.

Dynamic tests of fluid responsiveness

Test	How it works	When unreliable
Passive leg raise	Moves blood from legs to chest briefly	Raised abdominal pressure, leg injury, no output monitor
Pulse pressure variation	Output swings with mechanical breaths	Spontaneous breaths, arrhythmia, small tidal volumes, open chest
Fluid mini-challenge	Small bolus with output measured before and after	Imprecise monitor, repeated without limits
End-expiratory occlusion	Holding expiration briefly raises venous return	Patient breathing against the hold

Findings that change the plan

Non-responder still hypotensive: move to a vasopressor or inotrope rather than more fluid.
Spontaneous breathing, arrhythmia, low tidal volumes or an open chest: pulse pressure variation is unreliable; use leg raising or a mini-challenge.
Raised abdominal pressure can make leg raising falsely negative; measure bladder pressure if suspected.
Rising oxygen need, new B-lines or a positive balance over days: start de-resuscitation with diuretics or renal replacement.
Haemorrhage: replace with blood products, not crystalloid alone.

Next step

Reassess responsiveness after each bolus and whenever shock worsens; responders become non-responders.
Call a senior when the patient does not respond and remains hypotensive, to decide on vasopressors or inotropes.
Review cumulative fluid balance every day and plan removal once shock has resolved.
Follow chloride, bicarbonate and kidney function when large volumes of crystalloid are given.
Document which test was used, the result and the decision taken.

Test yourself

Think of the answer, then tap the question to see it.

What is fluid responsiveness?

A meaningful rise in stroke volume or cardiac output, about 10–15%, after a fluid bolus or passive leg raise, showing the heart is on the steep part of its Frank-Starling curve.

Why should a passive leg raise be judged on cardiac output and not blood pressure?

Blood pressure can change little even when output rises, and can rise for other reasons; a direct measure of flow is needed to read the test correctly.

When is pulse pressure variation unreliable?

With spontaneous breathing, arrhythmias, low tidal volumes, an open chest, raised abdominal pressure or right heart failure.

Why are balanced crystalloids preferred over saline for most patients?

Large volumes of saline cause hyperchloraemic metabolic acidosis and may harm the kidneys; balanced solutions are closer to plasma.

Add to my cards (4)

Elsewhere in DurgiPedia

Connections

In Pharmacology2

← All of anaesthesia

Educational content; it does not replace supervised training, your institution's protocols or the anaesthetist's or intensivist's decision. It contains no doses and no ventilator or infusion settings.

Sources: Fluid Management (StatPearls (NCBI Bookshelf), 2025); Crystalloid Solutions in Intravenous Fluid Therapy (StatPearls (NCBI Bookshelf), 2025).

Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake

Representative model, for educational purposes only.