Goal-directed haemodynamic therapy

Contents

Principles and Rationale

Goal-directed haemodynamic therapy (GDHT) uses haemodynamic monitoring to individualise fluid and vasopressor management towards predefined physiological targets, rather than applying fixed-volume protocols or clinical estimation.

Rationale

Perioperative physiological stress leads to occult tissue hypoperfusion — even when vital signs appear normal. Inadequate DO₂ (oxygen delivery) → anaerobic metabolism → systemic lactate accumulation → organ dysfunction. Alternatively, excessive fluid resuscitation → gut oedema, anastomotic breakdown, delayed return of GI function, and cardiac/pulmonary complications.

Oxygen Delivery (DO₂) as the Target

DO₂ = CO × CaO₂, where:

  • CaO₂ = (Hb × SaO₂ × 1.34) + (0.003 × PaO₂)
  • CO = HR × SV

Targeting a supranormal DO₂ (>600 mL/min/m²) was the original GDHT hypothesis (Shoemaker 1988). More recent algorithms target normal DO₂ while avoiding hypovolaemia.

Tools for Haemodynamic Assessment

Invasive

  • Pulmonary artery catheter (PAC/Swan-Ganz): gold standard CO measurement via thermodilution; direct measurement of PCWP, CVP, mixed venous O₂ saturation (SvO₂); now less used in non-cardiac peri-operative GDHT due to invasiveness and lack of mortality benefit when used routinely
  • Arterial line (radial/femoral): continuous BP; pulse contour analysis systems (LiDCO, Flo-Trac/Vigileo) derive stroke volume and SVV from the arterial waveform

Minimally Invasive / Non-Invasive

  • Oesophageal Doppler (CardioQ): measures aortic blood flow velocity; derives stroke volume and FTc (corrected flow time); real-time beat-to-beat; widely used in UK for intraoperative GDHT; requires sedation/anaesthesia for probe positioning
  • Pulse pressure variation (PPV) and stroke volume variation (SVV): derived from arterial line during controlled mechanical ventilation; predicts fluid responsiveness when PPV >13% or SVV >10%; requires: sinus rhythm, tidal volume ≥8 mL/kg (less reliable at lower TV), no spontaneous breathing effort
  • Near-infrared spectroscopy (NIRS/rSO₂): cerebral and somatic tissue oxygenation; non-invasive; used in cardiac and major vascular surgery

Fluid Responsiveness Predictors

A key goal of GDHT is giving fluid only when the patient is fluid responsive — i.e., when a fluid bolus will increase stroke volume by ≥10–15%. Giving fluid to a non-fluid-responsive patient causes harm (oedema, pulmonary complications) without haemodynamic benefit.

Static Markers (poor predictors)

  • CVP: does not reliably predict fluid responsiveness; still commonly used but extensively debunked
  • PCWP: slightly better but still unreliable
  • Cardiac filling pressures are affected by ventricular compliance, not just volume

Dynamic Markers (superior predictors)

Marker Principle Limitation
PPV/SVV (arterial line) Respiratory variation in pulse pressure/SV during controlled MV Requires CMV, sinus rhythm, TV ≥8 mL/kg; unreliable with RV failure or arrhythmia
IVC collapsibility index (echo) IVC diameter variation with respiration Requires spontaneous breathing; less reliable in MV
SV change post-fluid challenge Direct measurement; give 250 mL over 10 min; SV ↑ ≥15% = fluid responsive Useful but requires a fluid bolus each time
Passive leg raise (PLR) Elevating legs autotransfuses ~300 mL; assess SV/CO response in real time Non-invasive; applicable to spontaneously breathing patients; requires rapid CO measurement
End-expiratory occlusion test 15-second pause at end expiration → transient preload increase; CO increase ≥5% = fluid responsive Requires brief apnoea; needs MV; CO measured precisely
FTc (oesophageal Doppler) Corrected aortic flow time; FTc <0.35 s suggests preload deficiency Requires oesophageal Doppler probe; not applicable in awake patients

Algorithms and Targets

Typical Intraoperative GDHT Algorithm (Oesophageal Doppler-Based)

  1. Give a fluid challenge (250 mL colloid or crystalloid over 5 min)
  2. Measure stroke volume (SV) response:
    • SV ↑ ≥10% → fluid responsive → give another fluid challenge
    • SV ↑ <10% → not fluid responsive → stop; reassess for vasopressor need
  3. Repeat until SV is optimised (plateau)
  4. If MAP <65 mmHg despite optimised SV → vasopressor (noradrenaline/vasopressin)
  5. If CO/DO₂ remains low despite optimal SV and MAP → consider inotrope (dobutamine)

Haemodynamic Targets

Parameter Target
MAP ≥65 mmHg (individualise to pre-morbid BP)
SV Optimised by fluid challenge algorithm
Lactate <2 mmol/L
ScvO₂ >70%
Urine output >0.5 mL/kg/h

Evidence Base

OPTIMISE Trial (Pearse, JAMA 2014)

  • UK multicentre RCT, n=734 high-risk surgical patients
  • Oesophageal Doppler-guided GDHT (haemodynamic optimisation + dopexamine infusion) vs control
  • Result: no significant difference in 30-day complication rate or mortality; trend to benefit in surgical complications; haemodynamic optimisation was feasible
  • Post-hoc meta-analysis of pooled data suggested benefit

Older Evidence

  • Pearse 2005 (prospective pilot study): cardiac output monitoring-guided therapy in high-risk surgery → reduced postoperative complications
  • Hamilton et al (meta-analysis, 2011): 29 RCTs; perioperative GDHT associated with significant reduction in complication rates and LOS; mortality reduction in earlier studies (probably driven by greater care variability in older trials)

Sepsis Context (Early Goal-Directed Therapy — EGDT)

Rivers 2001 (original EGDT trial): protocolised resuscitation in septic shock targeting ScvO₂ ≥70%, MAP ≥65, CVP 8–12 → reduced mortality vs usual care. The three subsequent large RCTs (ARISE, ProCESS, PROMISE) found no mortality benefit from EGDT compared to usual care in septic shock — reflecting improved standard of care (not that the individual targets were wrong). See septic shock management page.

Fluid Type

  • Balanced crystalloids (Plasmalyte, Hartmann's) preferred over normal saline for large-volume resuscitation (hyperchloraemic acidosis, AKI risk with high-chloride saline)
  • Colloids (albumin, Gelofusine) for SV optimisation intraoperatively (faster distribution into intravascular space); however, no clear outcome advantage over crystalloid in large trials

Limitations and Patient Selection

Who Benefits from GDHT?

GDHT appears most beneficial for:

  • High-risk surgical patients (major abdominal, thoracic, vascular surgery; significant comorbidity; anticipated large blood loss or fluid shifts)
  • Patients who cannot tolerate large fluid volumes (cardiac failure, renal failure) — optimised IV fluid management avoids both under- and over-hydration

Limitations

  • PPV/SVV unreliable in spontaneous breathing, low tidal volume ventilation, arrhythmia, or RV failure
  • Oesophageal Doppler requires oesophageal placement — not applicable in awake patients or those with oesophageal pathology
  • No single monitoring modality works in all situations — integrate multiple parameters
  • Operator experience and training influence outcomes with any monitoring system
  • GDHT works best when embedded in a care bundle (enhanced recovery — ERAS): anaesthetic technique, surgical approach, nutrition, early mobilisation

Viva Questions

1. What is fluid responsiveness and why is it a more useful concept than giving fluid to a fixed CVP target?

Fluid responsiveness describes whether a fluid bolus will produce a meaningful increase in stroke volume (typically defined as ≥10–15% SV increase). It is grounded in the Frank-Starling relationship: on the ascending portion of the curve, preload increases stroke volume; on the plateau portion, additional preload produces no further SV increase — and causes fluid overload instead. CVP, the traditional target for fluid resuscitation, was adopted on the premise that a "filling pressure" reflects preload and therefore guides fluid administration. However, CVP is a poor predictor of fluid responsiveness — it is affected by ventricular compliance, pericardial restraint, intrathoracic pressure, and right heart function, none of which directly tells you where on the Frank-Starling curve the patient sits. Multiple studies show CVP correlates poorly with measured fluid responsiveness. Dynamic predictors — PPV, SVV, PLR response — directly test preload-responsiveness using physiological perturbations and have superior accuracy. The benefit: giving fluid only to fluid-responsive patients avoids the harm of fluid overloading non-responders (gut oedema, anastomotic complications, pulmonary oedema) while ensuring responsive patients receive adequate resuscitation.

2. Describe how the passive leg raise (PLR) test works as a predictor of fluid responsiveness, and what its advantages are.

The PLR test elevates the patient's legs (and lowers the torso if starting semi-recumbent) to approximately 45°. This autotransfuses approximately 300 mL of venous blood from the legs and splanchnic circulation into the central circulation, temporarily increasing cardiac preload. If the patient is preload-responsive (on the ascending portion of the Frank-Starling curve), stroke volume and cardiac output will increase within 30–60 seconds. CO change ≥10–15% from baseline is considered a positive response. Advantages: (1) it is a reversible test — the legs are lowered and the haemodynamic effect reverses, so no unnecessary fluid is given; (2) it can be applied to spontaneously breathing patients (unlike PPV/SVV which require controlled mechanical ventilation); (3) it can be performed at the bedside without any additional equipment (provided a cardiac output measurement is available — even continuous Doppler or echo); (4) it has high sensitivity and specificity (~85–90% in multiple studies). The key requirement is a means of measuring CO/SV in real time (oesophageal Doppler, arterial pulse contour, echo) — a purely clinical assessment of the PLR response (HR, BP change) is insufficient and may miss a true response or falsely identify non-responders.

3. What does the OPTIMISE trial tell us about GDHT, and how does it inform current practice?

OPTIMISE (Pearse, JAMA 2014) was a UK multicentre RCT that randomised 734 high-risk surgical patients to oesophageal Doppler-guided GDHT (with a dopexamine infusion) vs standard care. The primary outcome (30-day complications) showed a non-significant trend to benefit with GDHT (36.6% vs 43.4%; odds ratio 0.71, p=0.07). No mortality difference was detected. OPTIMISE was important for demonstrating feasibility of GDHT in UK practice and showing a trend towards reduced complications, but did not provide definitive evidence of benefit. In the context of the broader literature: Hamilton's 2011 meta-analysis of 29 RCTs suggested significant benefit from GDHT in reducing complications and LOS in high-risk surgery. Current UK practice, supported by NICE guidance, recommends cardiac output monitoring for high-risk surgical patients (defined as those with expected perioperative mortality >5%). GDHT is embedded in ERAS protocols where it is one component of a broader perioperative care bundle. The most impactful application remains in patients having major colorectal, vascular, or hepatobiliary surgery with significant anticipated haemodynamic stress.