ResusDoc Devices & Lines ← All tools
FOR REGISTERED PHYSICIANS ONLYClinical decision support only. Final responsibility for all clinical decisions lies with the treating physician. Software may contain errors — always apply independent clinical judgement.
← All devices

LVAD / ventricular assist device

No pulse, no cuff reading, no useful sats - and that can all be normal. Decide on perfusion, not on a pulse, and call the VAD centre immediately.

AdultCardiac & circulatory supportGuideline

Also called: LVAD, VAD, ventricular assist device, HeartMate, HeartWare, HVAD, mechanical circulatory support, MCS, left ventricular assist device

What it is

A left ventricular assist device (LVAD) is an implanted continuous-flow pump that takes blood from the left ventricle and drives it into the aorta, for a patient with end-stage heart failure (as a bridge to transplant or as a permanent, 'destination' therapy). A driveline tunnels out through the abdomen to an external controller and batteries.

The ED-critical fact is physiological. Because the pump flow is continuous, most LVAD patients have no palpable pulse, an often-unobtainable blood pressure cuff reading, and unreliable pulse oximetry when well. The AHA statement calls this hemodynamically stable pulseless electric activity (PEA), which we refer to in this population as pseudo-PEA. A normal heart sound is replaced by a continuous pump hum. You therefore cannot use a pulse to decide whether this patient has arrested. This page is for adult continuous-flow LVADs; paediatric and pulsatile devices (e.g. the Berlin Heart, where a pulse may be present) behave differently, and some children are discharged on support - for a child, call their VAD centre and apply their centre's advice rather than this page.

There is no UK national ED guideline for this. The guidance below is the American Heart Association scientific statement, which is the standard reference; in the UK every LVAD patient belongs to an implanting centre whose coordinator you call.

How to recognise it

What goes wrong

Pump stopped - power, controller or driveline

Collapse or deterioration with a silent pump (no hum) and alarms (which cease once batteries are flat). The statement names the two commonest causes: disconnection of power and driveline failure.

Red flag
This is the reversible emergency. The first move is mechanical, not pharmacological: the first step in assessing an unresponsive, mentally altered, or hypotensive patient with a VAD is to ensure that all connections are secure and an adequate power source is connected. Reconnect power, swap to the backup controller (a trained family member may do this faster than you), and call the VAD centre.

Pump running but the patient is shocked

A running hum but poor perfusion: altered, cold, mottled. Causes include RV failure, hypovolaemia or GI bleed, tamponade, PE, sepsis, pump thrombosis and arrhythmia. The displayed 'flow' is a calculation, not a measurement, and reads falsely high in pump thrombosis.

Red flag
Assess perfusion, not pulse. A Doppler over the brachial artery with a manual cuff gives a mean arterial pressure. Treat the cause: these patients bleed (acquired von Willebrand disease, GI angiodysplasia) and stroke at meaningful rates, so consider haemorrhage and neurology early.

VT or VF in a conscious patient

Sustained ventricular arrhythmia that the LVAD is partly tolerating - the patient may be surprisingly alert because the pump maintains some forward flow.

Red flag
Not all LVAD patients in VT/VF are unconscious. The statement advises that cardioversion and defibrillation should be performed for the same indications as used in the general population, with the understanding that perfusion in these patients does not equate to a palpable pulse. Most LVAD patients also have an ICD - see that page for pads and magnet.

Unconscious with a normally running pump

A flat, pulseless-looking patient whose pump is humming normally - commonly a stroke (ischaemic or haemorrhagic), hypoglycaemia, overdose or sepsis.

Red flag
The statement warns that a stroke patient may be unconscious with completely adequate flow yet appear as though they are in cardiac arrest because they may have no palpable pulse. Do not start compressions on a perfusing patient - check perfusion first.

The first ten minutes

  1. Call the patient's VAD centre / coordinator immediately - before you are sure what is wrong. The statement is explicit that Prompt notification of VAD center and its personnel (ie, VAD coordinator) is strongly recommended. The family and the equipment carry the answers; keep them with the patient.
  2. Assess perfusion, not pulse: mental status, breathing, skin colour and temperature, capillary refill. A pulse is expected to be absent.
  3. Listen for the pump hum over the left chest and upper abdomen, and check the device: alarms, secure driveline and controller connections, and adequate battery power. If the pump has stopped, restore power and swap the controller - when changing batteries, swap one power source at a time and never disconnect both at once, or the pump stops.
  4. Get a Doppler MAP: BP can be obtained noninvasively with a manual sphygmomanometer and a Doppler ultrasound probe. An MAP above roughly 50-60 mmHg with a running pump points away from arrest.
  5. Deciding on chest compressions is the hard part. If there are no signs of life and the patient is genuinely not perfusing, start compressions - and have a second person check power, connections and the controller at the same time; do not make CPR wait on troubleshooting. Capnography helps: the statement says a PETCO2 <20 mm Hg in an unresponsive, correctly intubated, pulseless patient with an LVAD would seem to be a reasonable indicator of poor systemic perfusion and should prompt rescuers to initiate chest compressions. Historic fears about dislodging the pump no longer withhold CPR from a patient with no perfusion - but do not compress a patient who is perfusing.
  6. Treat the reversible cause (bleeding, sepsis, arrhythmia, tamponade, PE, RV failure) with standard therapy, remembering they are anticoagulated, and arrange transfer to the VAD centre.

Do not

Do not call it an arrest because there is no pulse.
Continuous-flow LVAD patients are normally pulseless. The statement states that lack of a pulse cannot be used as a means to determine whether a patient is in cardiac arrest or a low-flow, low-perfusion state. Decide on perfusion.
Do not withhold chest compressions from a pulseless patient who is genuinely not perfusing.
The consensus is that if an LVAD is definitively confirmed by a trained person and there are no signs of life, bystander CPR, including chest compressions, should be recommended. A non-perfusing patient needs CPR; the old worry about the device does not override that. Equally - do not compress a perfusing patient.
Do not separate the patient from their controller, spare batteries and backup controller.
Receiving hospitals usually have no spares - the statement notes replacement equipment may be limited or nonexistent at receiving hospitals, particularly at non-VAD centers. The equipment and a trained family member must travel with the patient.
Do not read the controller's flow number as a true cardiac output, and do not forget the anticoagulation.
Displayed flow is estimated from pump speed and power and reads falsely high in pump thrombosis. And every LVAD patient is anticoagulated, so a hypotensive LVAD patient may be bleeding - the GI tract especially.

Who owns it

The implanting VAD centre owns the patient; its VAD coordinator is the first call and often directs care. Locally, cardiology, critical care and the resuscitation team manage the emergency. Tell the centre: pump running or stopped, alarms, perfusion/Doppler MAP, rhythm, and what you have already done. Transfer to the VAD centre whenever possible.

Defibrillation and pacing with an implanted device

You still defibrillate an arresting patient who has a device - you just keep the pads away from the generator. Standard ALS applies; the device changes pad position, not the algorithm. These pad rules apply to a child in arrest too; paediatric pad size and energy are not covered here.

Keep pads off the generator
The guidance: external electrodes should not be placed over or close to implanted electronic devices. A pad over the box risks damaging the device and the myocardium where the leads sit, and wastes energy.
How far, and which way
A distance of at least 10-15 cm between the edge of the device and the edge of the defibrillator electrode is recommended. Placing the pads approximately perpendicular to the device and its leads reduces current entering the device; an anterior-posterior position is a clean way to achieve this with a pectoral or a left-lateral (S-ICD) generator.
Do not withhold the shock
In a shockable arrest, if the ICD has not fired or its shocks have not worked, external defibrillation should be attempted in the usual way. The device is a reason to move the pad, never a reason to delay defibrillation.
External (transcutaneous) pacing
Can be used over an implanted device. If the implanted device is still emitting pacing spikes, set the external pacer faster than the programmed rate and/or out of demand mode, or its output may be inhibited by the implanted device's spikes.
Loop recorders and chest neurostimulators
Avoid placing a pad directly over a loop recorder. Similar clearance precautions apply to a neurostimulator implanted in the chest wall - the guidance advises Similar precautions are advised in people with neurostimulators implanted in the chest.

What a magnet does to each device

This is the single most-confused fact in the whole tool, and it is dangerous because the same magnet does opposite things depending on the device. The rule beneath the table: a magnet suspends shocks on anything that can shock (ICD, CRT-D, S-ICD) and does nothing to stop pacing on anything that only paces - on a pacemaker it forces fast, asynchronous pacing. Always identify the device from its card before reaching for a magnet. Two caveats: the magnet response is not universal - a leadless pacemaker (Micra/Aveir) has no magnet sensor and the response can be programmed off, so no magnet response does not prove the device is inactive; and these rules apply to a child with one of these devices too (paediatric pad size, energy and specialist advice are not covered here).

ICD (including CRT-D, S-ICD)
A magnet suspends shock therapy only while it is held in place - the guidance: Placing a strong magnet on the skin over the ICD generator will suspend tachyarrhythmia detection (and thereby suspend shock delivery). It does not stop pacing. Tape it on and check it stays: removal or displacement of the magnet will immediately restore the ability of the ICD to deliver a shock. One manufacturer (Biotronik) restores shocks after only 8 hours, so with that or an unknown device lift and replace the magnet periodically. Before applying the magnet for a shock storm, put external defibrillator pads on and get a rhythm strip - the magnet removes the safety net, so if the shocks are appropriate (VT/VF) you must be ready to defibrillate externally at once.
Pacemaker (and CRT-P)
A magnet does NOT switch it off. It will not deactivate it but will cause the pacemaker to fire at a faster than usual rate and stops it being inhibited by the patient's own beats (asynchronous pacing). The guidance adds that magnets should not be used in a misguided attempt to deactivate a pacemaker. The one time a pacemaker magnet helps: if the pacemaker is being wrongly inhibited by oversensing (surgical diathermy, myopotentials, lead noise) and the patient is bradycardic or pausing, holding a magnet on forces asynchronous pacing and restores a reliable rate.
On a device that both paces and shocks
The magnet suspends the shock function only; pacing continues. The appendix states plainly: Magnet application does NOT affect the programmed pacemaker function of the device. So a magnet makes a pacing-dependent ICD patient no less paced - it only stops shocks.
Vagus nerve stimulator (VNS)
Opposite logic to a cardiac magnet. Holding the magnet over the generator deactivates stimulation (continuous application of the magnet to the generator will deactivate the device); swiping it delivers an extra burst to abort a seizure. In an emergency without the patient's VNS magnet, any magnet with a strength over 50 gauss can be used to deactivate the device.
Loop recorder
Nothing to suspend - it delivers no therapy. A magnet is irrelevant to it.
After death with an active ICD
A magnet is only a temporary hold, not deactivation: Temporary deactivation using a magnet is not adequate before autopsy, device removal or cremation. A physiologist must deactivate the ICD, and if the patient has died with the magnet suspending shocks it must stay taped in place until they do. Note a brand that auto-reverts (e.g. Biotronik at 8 h) - after that the magnet protects no one, so tell the mortuary and device service.

Who to call

The team and the criteria are national. The phone number is local, and this tool does not know it. Find it once, now, and write it on the department's own card — that is the part no website can do for you.

Say the device name first
Open with laryngectomy, LVAD, shunt — the single word changes who comes and what they bring. Do not bury it in the history.
Say what you have already established
Whether the device is patent, whether it has been removed or disturbed, and what you have already tried. The specialist is deciding whether to walk or run.
Ask who owns it out of hours
Device ownership and the on-call route often differ from the implanting specialty, and it is frequently a regional centre rather than your own hospital.
The patient's own team is a resource, not an afterthought
Named nurse specialists, home-care companies and device helplines answer questions nobody else can, and the patient usually has the number.

When the device is not the problem

A patient with a device gets pneumonia, sepsis, pulmonary oedema and myocardial infarction at the same rate as everybody else. The device is the most visible thing about them and it attracts the whole assessment.

The NTSP says it out loud
The project's own manual warns that patients with tracheostomies can become unwell with all the problems that other patients get too. It is easy to become fixated with the tracheostomy. The same sentence applies to every device in this tool.
Clear the device, then leave it
Establish that the device is doing its job — the stoma is patent, the line draws back, the pump is running — and then assess the patient as you would anyone else.
The anchoring runs both ways
The opposite error is treating a device complication as a medical illness: a blocked gastrostomy read as vomiting, a baclofen pump failure read as sepsis.

What the patient will be carrying

LVAD patients carry a medical alert identifying the device and centre, and usually their anticoagulation details. Crucially they travel with a backup controller and spare batteries, and often a family member trained to change the controller. Keep all of it with the patient.

Scope of this page

This page covers recognising an LVAD patient and the first ED decisions: assess perfusion rather than pulse, check the pump and power, get a Doppler MAP, decide about chest compressions, and call the VAD centre. It does not teach pump reprogramming, speed changes, echo-guided VAD assessment or the management of a total artificial heart, all of which belong to the VAD centre. There is no UK national ED guideline; the source is the AHA scientific statement. It is written for adult continuous-flow LVADs; paediatric and pulsatile devices differ and are directed by the child's VAD centre.

Sources