Critical care Topic

Fluid responsive is not fluid depleted.

Four cards on one decision. Responsiveness is a position on a curve, not a diagnosis. The bolus works by widening a venous gradient, most of it leaves the capillary within hours, and a responsive patient can still be intolerant. Tolerance is the permission, and it's asked first.

Reviewed September 2026 · verify against current guidelines

Circulation · 1 of 4

Fluid responsive is not fluid depleted.

Responsiveness is a position on the cardiac function curve, not a diagnosis. It says the ventricle would eject more if you filled it — not that filling it is the right thing to do.

The same bolus, two places on the curve

On the ascending limb the same 500 mL raises stroke volume by 10% or more — responsive. On the plateau the same 500 mL raises it by less than 10% — not responsive, and the volume goes to the lungs.

Tests that find the limb — each needs a stroke volume monitor, not a blood pressure

Test	Threshold	Notes
Passive leg raise	ΔSV ≥ 10%	A reversible 300 mL autotransfusion. Works in arrhythmia and in spontaneous breathing, which is why it is the default.
Pulse pressure variation	> 13%	Only valid in sinus rhythm, fully passive on the ventilator, at tidal volumes of 8 mL/kg or more.
Fluid challenge	250–500 mL	The direct answer, and the only one you cannot take back. Give it fast and measure before and after.

Why a filling pressure cannot answer this. A pressure reads the ventricle's compliance curve. Responsiveness lives on its contractility curve. A stiff ventricle shows a high CVP at low volume; a dilated one a normal CVP at high volume. So the same CVP of 8 mmHg appears in patients who respond and patients who do not. Decades of studies put its predictive value near a coin toss, and it is still the most-used number for the decision.

Two traps — a healthy volunteer is fluid responsive, and so is roughly half of the unstable ICU population; responsiveness is a normal state, not a deficit to be corrected. And the test answers only what you asked — whether stroke volume will rise. Why the bolus raises output is a venous question; where the fluid ends up is a capillary one; whether the patient can absorb it is tolerance.

Circulation · 2 of 4

The gradient, not the pressure.

Blood returns to the heart down a pressure difference. A bolus works only if it raises the upstream pressure more than the downstream one — and the heart decides which.

The equation

VR = (MSFP − CVP) / R_v. Venous return runs on mean systemic filling pressure — 8–10 mmHg in a stopped circulation — minus the central venous pressure it drains into, over venous resistance.

The same bolus, two hearts — a bolus shifts the return line right; the heart decides the rest

Steep curve	The bolus buys flow. Flow rises sharply; CVP barely moves.
Flat curve	The bolus buys pressure. CVP rises sharply; flow barely moves.

Two volumes, only one of which does anything

About 30% of blood volume is arterial; the veins hold about 70%, most of it unstressed. Unstressed volume fills the vessel without stretching it — it generates no pressure and adds nothing to return. Stressed volume alone sets MSFP. A pressor converts unstressed into stressed.

Three consequences that change what you do

A rising CVP is the failure signal. If the bolus moved the pressure and not the flow, you narrowed the gradient and added congestion for nothing.
A vasopressor is a bolus you can take back. Norepinephrine squeezes unstressed into stressed volume, raising MSFP without adding a drop.
CVP is what the organs drain against. Kidney and liver perfusion depend on the gradient across them, so a high CVP injures from behind.

Two traps — MSFP is not on your monitor — it takes an inspiratory hold, so treat it as reasoning rather than a number to chase. And this is what card one left open: a static CVP cannot predict responsiveness, but a rising CVP during a bolus says the gradient is closing.

Microcirculation · 3 of 4

Where the fluid goes.

The classical model had fluid leaving the capillary at one end and returning at the other. It does not. Filtration runs the whole length, the lymphatics are the only way back, and the barrier is a gel you can strip off.

The revised Starling equation

J_v = L_pA [ (P_c − P_i) − σ(π_p − π_sg) ]. One term changed and the whole model with it: the oncotic pull is set by the subglycocalyx space, not the interstitium.

Where the barrier actually sits

Glycocalyx, then the subglycocalyx space, then endothelium. Filtration continues the whole length of the capillary — it never reverses, and there is no venous reabsorption.

What the revision costs you at the bedside

~15%	Roughly the share of a crystalloid bolus still intravascular within a few hours of giving it in inflammation.
σ falls	Sepsis sheds the glycocalyx, so the reflection coefficient drops and the colloid advantage largely goes with it.
Lymph is the exit	Flow can rise several-fold, but not indefinitely. Edema is filtration outrunning that ceiling.

The bolus damages the barrier that would have retained it. Stretching the atrium releases natriuretic peptide, which strips the glycocalyx. A rapid bolus into a full patient therefore raises its own permeability, and the next bolus works less well than the last. The practical reading: give volume in the smallest aliquot that answers the question, at the point where the gradient is genuinely low — and stop as soon as the flow stops rising.

Two traps — this is not the curve from card one. The Frank–Starling curve describes the ventricle; the Starling equation describes the capillary. Same name, different organ, different decision. And "third spacing" names a space that does not exist — it is ordinary filtration outrunning lymphatic return, which is why the fluid comes back on its own once the inflammation settles.

Circulation · 4 of 4

Responsive, but not tolerant.

Responsiveness asks whether the output will rise. Tolerance asks what the volume will cost the lung, the kidney and the gut. They are independent, and only tolerance is a permission.

Reading tolerance — congestion upstream of each organ, not a single number

Lung	B-lines in multiple zones, falling oxygenation, more oxygen needed during the bolus itself.
Kidney and gut	Venous congestion on Doppler, a dilated inferior vena cava that does not vary, rising abdominal pressure.
Right heart	A dilated right ventricle turns each further bolus into septal shift and less left-sided filling.

Two questions, four patients

	Tolerant	Intolerant
Responsive	Give it. The only quadrant where a bolus is the answer. Small aliquot, measure, repeat while the flow rises.	Buy the flow another way. The output would rise, but congestion already costs more than the flow is worth. Pressor or inotrope instead.
Not responsive	No benefit to buy. The volume is absorbed without harm and without gain. Look elsewhere for the cause of the shock.	De-resuscitate. Nothing to gain, harm accruing. Stop maintenance fluids, then diuretic or ultrafiltration. Most late ICU patients are here.

Benefit falls, harm rises

The first hours buy the most. The gradient is genuinely low and the barrier largely intact, so volume does what you intend.
By day two it inverts. The same bolus mostly buys congestion, and a cumulative positive balance tracks with worse outcomes.

Two traps — tolerance is a decision, not a measurement: no index tells you how much congestion these organs will accept, so read it from several sites and revise after every aliquot. And at the bedside the two questions come in one order only — tolerance first, because an intolerant patient does not need the responsiveness test performed. These cards teach the reverse, because mechanism has to precede decision.

SV stroke volumeCVP central venous pressureMSFP mean systemic filling pressureVR venous returnPLR passive leg raisePPV pulse pressure variation

Caveats exist; this is a general framework. Educational reference only — verify against local protocol.

Sources

Verify against current guidelines and local protocol before acting.

Monnet X, Teboul JL. Passive leg raising: five rules, not a drop of fluid! Crit Care 2015;19:18.
Marik PE, Baram M, Vahid B. Does central venous pressure predict fluid responsiveness? A systematic review of the literature and the tale of seven mares. Chest 2008;134:172–178.
Michard F, Teboul JL. Predicting fluid responsiveness in ICU patients: a critical analysis of the evidence. Chest 2002 (pulse pressure variation).
Guyton AC. Determination of cardiac output by equating venous return curves with cardiac response curves. Physiol Rev 1955.
Woodcock TE, Woodcock TM. Revised Starling equation and the glycocalyx model of transvascular fluid exchange. Br J Anaesth 2012;108:384–394.
Malbrain MLNG, et al. Principles of fluid management and stewardship in septic shock: the four D's and the four phases. Ann Intensive Care 2018.
Boyd JH, et al. Fluid resuscitation in septic shock: a positive fluid balance and elevated CVP are associated with increased mortality. Crit Care Med 2011.

Downloads

Every card for this topic — carousels and tables, print-ready for the wards or for sharing.

Carousel4 slides