1

Pre-LVAD: findings that raise risk

Labs · clinical state · echo
01

Labs / end organs

Historical pre-LVAD cohort risk flags
  • Total bilirubin ≥2.0 mg/dL.
  • Creatinine ≥2.3 mg/dL or renal replacement therapy.
  • AST ≥80 U/L.
  • BUN >39 mg/dL.
  • Rising bilirubin / creatinine, oliguria, ascites and persistent edema add concern.
02

Clinical state

Less reserve before the operation
  • Vasopressor requirement or multiple inotropes.
  • Preoperative mechanical ventilation.
  • Shock, low venous saturation or worsening organ perfusion.
  • Prior RV failure / RV infarction.
  • Previous cardiac surgery.
03

Echo

A vulnerable RV
  • Severe RV systolic dysfunction.
  • RV dilation, especially with poor contraction.
  • Severe tricuspid regurgitation.
  • TAPSE ≤1.7 cm or FAC ≤35% flags dysfunction. Read with global RV function.
  • Pulmonary hypertension / high RV afterload.
2

Hemodynamic risk flags, with context

Cohort cutoffs · document support
MeasureRisk thresholdUse at the bedside
PAPi
(PASP − PADP) / mean RAP
<1.85 before LVADAssociated with early RV failure in a 132-patient cohort. Not a universal implant or explant threshold.
PAPi after optimization≤3.33 before LVADResidual-risk signal in a 315-patient cohort after PAC-guided optimization. Compare measurements under documented support.
PAPi in acute RV infarction≤0.9Severe RV dysfunction and risk of RV support / death in acute inferior MI.
RAP / CVP>15 mmHgCongestion, especially when paired with low CI and a dilated RV.
RAP / PCWP>0.63Disproportionate right-sided congestion before LVAD.
Cardiac index≤2.2 L/min/m²Low forward output. Concern increases when RAP stays high.
RV stroke work index≤250 mmHg·mL/beat/m²
= ≤3.4 g·m/beat/m²
RVSWI = (mPAP − RAP) × SVI.
SVI = CI × 1000 / HR (mL/beat/m²). Use CI in L/min/m² and HR in beats/min. Multiply RVSWI by 0.0136 for g·m units.
Read the clusterLow PAPi + high RAP + low CI + rising bilirubin / creatinine + a poor RV should prompt a planned RV-support strategy. These cohort risk flags are not stand-alone pump triggers. Recheck after optimization. Record rhythm, ventilation and support when trending PAPi.
3

RV power output: equation and interpretation

Pressure in mmHg · CO in L/min
RVPO = mPAP × CO / 451 WIndex: mPAP × CI / 451 W/m². Example: 30 × 4 / 451 = 0.27 W. Net hydraulic power uses (mPAP − RAP) × CO / 451. No established RVPO-alone cutoff predicts pre-LVAD RV failure. Falling flow / power with rising RAP is concerning. Higher power may reflect pulmonary load. Systemic CPO 0.6 W is not an RVPO cutoff.
4

Fick versus thermodilution in RV dysfunction

Before a CO / CI result drives treatment
SituationPractical rule
Choose / name the methodWithout shunt or circuit distortion, prefer thermodilution (TD) over estimated Fick. Direct Fick uses measured oxygen consumption. Estimated Fick assumes it. CI = CO/BSA.
TR / reliable TDSevere TR does not automatically invalidate TD. Average 3 values within 10% with consistent injectate and respiratory phase. Repeat poor curves.
Low flow / disagreementDecision-changing discrepancy: recheck Hb, paired arterial / PA samples and technique. Adjudicate with direct Fick. Do not average disagreeing methods.
Shunt or RVADShunt: saturation run and shunt-specific Fick flows, not routine TD. RVAD: map drainage, return and sampling sites. See page 4.
5

Before CABG or valve surgery

Items that raise concern
A

Patient / substrate

Poor RV + greater pulmonary load
  • RV dysfunction, severe TR, pulmonary hypertension.
  • sPAP >55 mmHg, LVEF <50%, AF were associated with postoperative RV failure.
  • Renal/hepatic dysfunction, congestion and prior RV infarction.
B

Operation / perioperative stress

A smaller margin for error
  • Mitral or tricuspid surgery.
  • Redo / complex operation, prolonged CPB or ischemic time.
  • RV ischemia, hypoxemia, hypercapnia / acidosis, excessive intrathoracic pressure and loss of AV synchrony.
CTICU
priorities
FillRelieve congestion. Give small, reassessed volume only for underfilling.
PerfuseSupport systemic pressure and forward output together.
UnloadLower pulmonary load and ensure adequate LV unloading.
SynchronizePreserve sinus rhythm / AV synchrony. Consider pacing when it improves output.
1

An individualized RV-protection plan

Pilot-derived targets · individualize to response
Filling
Aim for CVP about 8-14 mmHg with improving congestion and output.
Systemic perfusion
Aim for MAP about 70-90 mmHg. Follow lactate, urine output and venous saturation.
Left filling
Aim for PCWP <18 mmHg when measured. Check LVAD unloading and septum.
Rate / rhythm
Sinus rhythm + AV synchrony. The pilot used HR >100/min. Pace only when filling and output improve.
Lungs
SpO₂ >95% · PaCO₂ <50 mmHg
Keep plateau pressure <30 cm H₂O.
Geometry / oxygen delivery
Use echo-guided LVAD speed. Treat bleeding and anemia that limits oxygen delivery.
A

Underfilled?

Use a small, observed preload trial
  • Give 100–250 mL, then reassess flow, CVP and echo. Stop if pressure rises without improved output.
B

Congested?

Read MAP together with CVP
  • MAP 70 − CVP 25 leaves a perfusion gradient of only 45 mmHg. Support arterial pressure and relieve venous congestion.
2

Consider pacing before escalating catecholamines

Rate and AV timing are treatments
A

Sinus bradycardia

Intact AV conduction
  • With reliable atrial wires, use atrial demand pacing to raise rate while preserving native ventricular activation.
  • Test arterial waveform, CI/venous saturation, filling and echo. Keep the rate that improves output.
B

AV block / lost synchrony

Atrial + ventricular wires
  • Use AV sequential pacing when indicated. Optimize AV timing for filling and forward flow.
  • With AF/flutter, avoid rapid atrial tracking. Correct triggers and restore a useful rhythm promptly.
C

Dependable pacing

Each shift and after a change
  • Confirm sensing, electrical capture and mechanical response. Record mode, rate, AV delay and thresholds.
  • Keep backup available. New hypotension, arrhythmia or loss of capture needs immediate reassessment.
3

Vent and drips: give each intervention a purpose

Pressure · filling · flow
01

Ventilator

Recruit without overdistension
  • Tidal volume 6-8 mL/kg predicted body weight. Check plateau pressure and auto-PEEP.
  • Start with low PEEP, often 5 cm H₂O. Increase when recruitment / oxygenation improves and the RV tolerates it.
  • Correct hypoxemia, hypercapnia, acidosis, pain and dyssynchrony. Aim for early extubation when perfusion and gas exchange permit.
PVRCollapseLung volumeOverdistension
Low and high lung volumes both raise pulmonary vascular load.
02

Drips

Treat the limiting physiology
ProblemPractical choice
Low outputDobutamine 2.5–20 mcg/kg/min. Start low, titrate to perfusion. Watch tachyarrhythmia.
Low output + pulmonary loadMilrinone 0.125–0.75 mcg/kg/min. Start low. Watch hypotension and renal accumulation. Reduce for renal impairment.
HypotensionNorepinephrine ± vasopressin. Support coronary/systemic perfusion while correcting the cause.
Pulmonary loadiNO 20 ppm or iEPO 50 ng/kg/min were used in INSPIRE-FLO. Ensure uninterrupted delivery, including transport / extubation. Wean stepwise to avoid rebound.
CongestionIV loop diuretic. Escalate decongestion if venous pressure stays high. Consider renal support for refractory overload.
4

Reassess after every important change

One team · one support plan
A

Trend perfusion and filling

Record the support state
  • RAP/CVP, PA/PCWP when available, CI, lactate, venous saturation, urine output and renal/hepatic trajectory.
  • Pair these with RV/LV echo, rhythm, ventilator settings, vasoactive doses and LVAD speed.
B

Escalate early

Congestion + poor perfusion despite treatment
  • Exclude tamponade, bleeding, ischemia, PE, arrhythmia and device/outflow obstruction.
  • Persistent RV failure needs an RV-support decision before progressive organ injury.
Why use a coordinated bundle?The RV Protection pilot randomized 20 patients, 10 per group. At 24 weeks, survival free of hemodynamic or hemocompatibility events was 80% vs 40%. Severe RV failure was 0/10 vs 3/10 (P=0.20). This supports further study, not a proven universal bundle.