Overview — The Only Reason to Give Fluid Is to Increase Stroke Volume
"The decision to give a fluid bolus should be answered by one question: will this fluid increase the patient's stroke volume? Roughly half of haemodynamically unstable patients are not fluid-responsive, and in them a bolus only causes harm — oedema, worsened oxygenation, and a positive fluid balance that independently predicts mortality. Modern haemodynamic monitoring exists to predict fluid responsiveness, not merely to measure pressures."
Washington Manual of Critical Care, 4th Ed. Wolters Kluwer 2023. Chapter: Haemodynamic Monitoring.Fluid Responsiveness — The Central Concept
A patient is fluid-responsive if a fluid bolus increases stroke volume/cardiac output by ≥10–15%. Only responders benefit from fluid. Both ventricles must be on the steep (preload-dependent) part of the Frank–Starling curve for fluid to help; on the flat part, fluid adds volume overload without raising output.
Marino Physiology — Frank–Starling & Heart–Lung Interaction
"The Frank–Starling relationship is the physiological heart of fluid management: increasing preload increases stroke volume only while the ventricle operates on the ascending limb of its function curve. The art of haemodynamic monitoring is to determine, at the bedside and in real time, where on that curve a given patient's ventricle is sitting — because that, not any single pressure, dictates whether volume will help or harm."
Marino PL. The ICU Book, 5th Ed. Haemodynamic Monitoring & Fluid Responsiveness. Wolters Kluwer; 2025.In a mechanically ventilated, passive patient in sinus rhythm, each positive-pressure breath transiently alters venous return and stroke volume. If both ventricles are preload-dependent, this produces a large cyclical swing in arterial pressure and stroke volume.
- Pulse Pressure Variation (PPV) >13% or Stroke Volume Variation (SVV) >13% predicts fluid responsiveness
- It works because the heart–lung interaction "stress-tests" the Starling curve every breath
The catch — when PPV/SVV is NOT valid: spontaneous breathing, cardiac arrhythmia (AF), low tidal volume (<8 mL/kg), open chest, very high respiratory rate, raised intra-abdominal pressure, or RV failure. In these (common) situations, use the passive leg raise or a fluid challenge instead.
CVP and PCWP are static pressures — they reflect a point on the venous/cardiac pressure–volume relationship, not the slope of the Starling curve. Because venous compliance, ventricular compliance, intrathoracic pressure and right-heart function all vary widely between patients, a given CVP corresponds to wildly different volume states in different people. This is why a single CVP cannot predict whether fluid will increase cardiac output.
Abandoning the CVP Number
A very high CVP suggests volume overload / RV failure / tamponade / fluid intolerance (a reason NOT to give fluid). A rising CVP during resuscitation flags developing venous congestion (linked to AKI). But for the everyday question "should I give a bolus?", use a dynamic test, not the CVP.
The Tests That Actually Work
| Test | Threshold | Use when | Caveats |
|---|---|---|---|
| Passive Leg Raise (PLR) | ↑CO/SV ≥10% | Almost any patient — the most robust, reversible "auto-bolus" | Need a real-time CO/SV monitor (or VTI on echo); not in ↑ICP; measure within 60–90 s |
| PPV / SVV | >13% | Ventilated, passive, sinus rhythm, TV ≥8 mL/kg | Invalid in spontaneous breathing, arrhythmia, low TV, open chest, RV failure |
| Fluid challenge | ↑SV/CO ≥10–15% | When dynamic tests unavailable | Give a defined small bolus (e.g. 4 mL/kg / 250 mL over 5–10 min) & measure response; commits volume |
| End-expiratory occlusion | ↑CO ≥5% | Ventilated patient who tolerates a 15-s breath hold | Needs precise CO monitor |
| IVC distensibility (echo) | >~12–18% | Ventilated; quick bedside screen | Less reliable in spontaneous breathing, high intra-abdominal pressure, RV failure |
PLR transfers ~300 mL of venous blood from the legs and splanchnic bed to the central circulation — a reversible, internal fluid bolus. Start the patient semi-recumbent (trunk 45°), then lay the trunk flat and raise the legs to 45°, and measure the change in cardiac output/stroke volume (not blood pressure) within 60–90 seconds. A rise ≥10% predicts the patient will respond to real fluid. Its great advantage: it works in spontaneously breathing and arrhythmic patients where PPV/SVV fail, and it commits no actual fluid.
From Arterial Line to Cardiac Output
Point-of-care echo is the most accessible and cost-effective haemodynamic tool in most Indian ICUs — VTI-based PLR/fluid-response assessment needs only a probe and is far more informative than a CVP line.
Advanced pulse-contour monitors are limited to tertiary centres; the passive leg raise + serial lactate + clinical perfusion triad is a robust, low-cost resuscitation strategy everywhere.
"Should I Give Fluid?" — Step by Step
Is there evidence of hypoperfusion?
- Lactate, mentation, urine, skin, MAP. If perfusion is adequate → stop giving fluid.
Assess fluid responsiveness (dynamically)
- Ventilated, passive, sinus, TV ≥8 → PPV/SVV >13%.
- Otherwise → passive leg raise (measure CO/SV) or a small defined fluid challenge.
Responder → give a defined bolus & reassess
- e.g. 250–500 mL balanced crystalloid; re-measure SV/CO and perfusion.
- Repeat the responsiveness test before the next bolus — responsiveness disappears as you fill.
Non-responder (or fluid-intolerant) → stop fluids
- Switch to vasopressors / inotropes per shock type; consider de-resuscitation once stable (a positive cumulative balance worsens outcomes).
- Look for fluid intolerance: rising CVP, new B-lines, worsening oxygenation.
Common Mistakes in Haemodynamic Monitoring
A single CVP value does not predict fluid responsiveness — the evidence is unequivocal. Use dynamic tests (PLR, PPV/SVV, fluid challenge with SV measurement). Reserve CVP for trends and recognising venous congestion/RV failure.
PPV/SVV are only valid in ventilated, passive, sinus-rhythm patients with adequate tidal volume. In spontaneous breathing, AF, or low TV they mislead — switch to the passive leg raise.
The passive leg raise must be assessed by the change in cardiac output/stroke volume (or aortic VTI on echo). Blood pressure alone is insensitive and will miss true responders/non-responders.
Being fluid-responsive is normal physiology — most healthy people are. Fluid is only indicated when there is BOTH hypoperfusion AND responsiveness. Chasing responsiveness in a perfusing patient just causes oedema.
A persistently positive fluid balance independently predicts mortality, AKI and prolonged ventilation. Once the patient is stable, shift from resuscitation to maintenance and consider active de-resuscitation.
Trials (PAC-Man, ESCAPE) show no routine mortality benefit and potential harm. Reserve the PAC for selected complex cases (severe/mixed RV failure, pulmonary hypertension) and prefer echo + dynamic tests otherwise.
References
- Marino PL. The ICU Book, 5th Ed. Haemodynamic Monitoring; Fluid Responsiveness. Wolters Kluwer; 2025.
- Washington Manual of Critical Care, 4th Ed. Chapter: Haemodynamic Monitoring. Wolters Kluwer 2023.
- Marik PE, Cavallazzi R. Does the central venous pressure predict fluid responsiveness? An updated meta-analysis. Crit Care Med 2013;41:1774–1781.
- Monnet X, Marik PE, Teboul JL. Passive leg raising for predicting fluid responsiveness: a systematic review and meta-analysis. Intensive Care Med 2016;42:1935–1947.
- Cecconi M, De Backer D, Antonelli M et al. Consensus on circulatory shock and haemodynamic monitoring (ESICM Task Force). Intensive Care Med 2014;40:1795–1815.
- Harvey S, Young D, Brampton W et al. (PAC-Man). Assessment of the clinical effectiveness of pulmonary artery catheters. Lancet 2005;366:472–477.