pacing in magnetic resonance imaging environment

European Heart Journal (2001) 22, 113–124doi:10.1053/euhj.2000.2149, available online at http://www.idealibrary.com on

Review Article

Pacing in magnetic resonance imaging environment:

Clinical and technical considerations on compatibility

F. Duru1, R. Luechinger2, M. B. Scheidegger2, T. F. Luscher1, P. Boesiger2 andR. Candinas1

1Division of Cardiology, Department of Internal Medicine, University Hospital of Zurich and 2Institute ofBiomedical Engineering and Medical Informatics, University of Zurich and Swiss Federal Institute of Technology,

Zurich, Switzerland

Introduction

Magnetic resonance imaging is a widely accepted toolfor the diagnosis of a variety of disease states. However,the presence of an implanted pacemaker is considered tobe a strict contraindication to magnetic resonance imag-ing in the vast majority of medical centres, precluding asubstantial and growing number of patients from thediagnostic advantages of this imaging modality.

The potential effects of magnetic resonance imagingon cardiac pacemakers are multiple (Table 1)[1–5]. Thestatic magnetic field may close the reed switch, resultingin asynchronous pacing. The radiofrequency field mayinduce an undesirable fast pacing rate. The time-varyingmagnetic gradient fields may induce currents in thepacemaker system of sufficient magnitude to pace theheart. There are other theoretical concerns, includingpotential thermal injury at the lead tip. However,the most influential factor that supports the currentpractice is the reported lethal consequences of magneticresonance imaging in patients with implantedpacemakers[6–8].

Despite the above-mentioned concerns, the effects ofmagnetic resonance imaging on cardiac pacemakersremain controversial. Most of the previous studies thatprohibit magnetic resonance imaging in pacemakerpatients were based on in vitro and animal model datafrom the 1980’s using older pacemaker and leadtechnology[1–3]. More recent anecdotal reports describe asmall series of pacemaker patients who have safely

0195-668X/01/220113+12 $35.00/0

undergone magnetic resonance scanning (Table 2)[9–14].These studies were performed in patients for whommagnetic resonance imaging was considered to be anabsolute diagnostic necessity.

The potential effects of magnetic resonance imaging inpatients with implanted pacemakers is outlined in thispaper, which provides an extensive review of previous invitro and animal studies, as well as recently publishedobservations in humans. Based on technical innovationsand clinical developments in magnetic resonance imag-ing and pacemaker systems, safety and compatibilityissues are discussed.

Revision submitted 24 January 2000, and accepted 27 January 2000

Key Words: Magnetic resonance imaging, pacing,pacemakers, implantable cardioverter defibrillators,electromagnetic interference.

Correspondence: Firat Duru, MD, Cardiac Arrhythmia UnitUniversity Hospital of Zurich, Raemistr. 100, CH-8091 Zurich,Switzerland.

Table 1 Potential effects of magnetic resonance imagingon pacemaker systems

1. Static magnetic field(a) Reed-switch closure(b) Pacemaker displacement(c) Changes in electrocardiograms

2. Radiofrequency field(a) Heating(b) Alterations in pacing rate(c) Pacemaker reprogramming or reset

3. Time-varying magnetic gradient field(a) Induction voltage(b) Heating(c) Reed-switch closure

Clinical considerations

Basic principles of magnetic resonanceimaging

Since the first live human images reported in 1977, thedevelopment of clinical magnetic resonance imaging hasbeen rapid[15,16]. Continuous improvement in magneticresonance methodology, hardware and software has

� 2001 The European Society of Cardiology

114 F. Duru et al.

Eur Heart J, Vol. 22, issue 2

Tab

le2

Pub

lishe

dre

port

sde

scri

bing

the

non-

leth

alco

nseq

uenc

esof

mag

neti

cre

sona

nce

imag

ing

inpa

cem

aker

pati

ents

Aut

hor

(ref

.)Y

ear

nIn

dica

tion

for

MR

scan

ning

Pac

emak

erm

odel

Dua

l/sin

gle

cham

ber

Lea

dpo

lari

tyP

acem

aker

mod

eM

agne

tic

field

Out

com

e

Ala

gona

[9]

1989

1B

rain

tum

our

AF

PII

(283

)D

ual

cham

ber

Uni

pola

rO

OO

1·5

Tel

saN

orm

alP

Mfu

ncti

onIn

bar[1

0]

1993

1C

ereb

ello

-pon

tine

Par

agon

II(2

016T

)D

ual

cham

ber

Bip

olar

OO

O1·

5T

esla

Nor

mal

PM

func

tion

angl

esy

ndro

me

Gim

bel[1

2]

1996

1H

eart

valv

eA

FP

(261

)Si

ngle

cham

ber

Uni

pola

rO

OO

1·5

Tes

laM

Rim

age

arti

fact

1B

rain

tum

our

Gen

esis

(285

)D

ual

cham

ber

Uni

pola

rD

OO

0·5

Tes

laP

ause

(2s)

1C

IAP

arag

onII

(201

6)D

ual

cham

ber

Bip

olar

OO

O0·

35T

esla

Nor

mal

PM

func

tion

1P

itui

tary

tum

our

Sync

hron

y(2

020)

Dua

lch

ambe

rB

ipol

arD

OO

15

Tes

laN

orm

alP

Mfu

ncti

on1

Cer

vica

ldi

skSy

nchr

ony

II(2

022)

Dua

lch

ambe

rB

ipol

arD

DD

15

Tes

laN

orm

alP

Mfu

ncti

onF

onta

ine[4

9]

1998

1D

izzi

ness

The

raD

R(7

940)

Dua

lch

ambe

rB

ipol

arV

VI

1·5

Tes

laR

apid

vent

ricu

lar

paci

ngG

arci

a-B

olao

[11

]19

981

Cra

nial

nerv

epa

lsy

Met

a(1

254)

Dua

lch

ambe

rB

ipol

arA

OO

1T

esla

Asy

nchr

onou

spa

cing

Som

mer

[14

]19

9818

CN

S(1

3)E

lite

(707

7)(3

x)D

ual

cham

ber

n.a.

DD

D0·

5T

esla

Asy

nchr

onou

spa

cing

Car

diac

tum

our

(2)

Elit

e(7

085)

(2x)

Dua

lch

ambe

rn.

a.D

DD

0·5

Tes

laA

sync

hron

ous

paci

ngP

erip

rost

heti

c(a

sc.

aort

a)E

lite

(707

6)D

ual

cham

ber

n.a.

DD

D0·

5T

esla

Asy

nchr

onou

spa

cing

Pse

udoa

neur

sym

(1)

The

raD

R(7

940)

(2x)

Dua

lch

ambe

rn.

a.D

DD

0·5

Tes

laA

sync

hron

ous

paci

ngP

arav

alvu

lar

(pro

sthe

tic

Rel

ay(2

94-0

3)(2

x)D

ual

cham

ber

n.a.

VO

O0·

5T

esla

Nor

mal

PM

func

tion

Aor

tic

valv

e)(1

)V

ista

(941

)(2

x)D

ual

cham

ber

n.a.

DD

D0·

5T

esla

Asy

nchr

onou

spa

cing

Con

stri

ctiv

epe

rica

rdit

is(1

)Sy

nerg

yst

(702

7,70

71)

Dua

lch

ambe

rn.

a.D

DD

0·5

Tes

laA

sync

hron

ous

paci

ngD

ialo

g(7

00)

(2x)

Sing

lech

ambe

rn.

a.V

VI

0·5

Tes

laA

sync

hron

ous

paci

ngSp

ectr

ax(5

985)

Sing

lech

ambe

rn.

a.V

VI

0·5

Tes

laA

sync

hron

ous

paci

ngC

eryx

3Si

ngle

cham

ber

n.a.

VV

I0·

5T

esla

Asy

nchr

onou

spa

cing

n=

num

ber

ofpa

tien

tsst

udie

d;n.

a.=

data

not

avai

labl

e;P

M=

pace

mak

er

, January 2001

Review 115

now resulted in whole body imaging systems that arecapable of producing high contrast images with a spatialresolution of under 1 mm, and in total imaging times ofunder 15 min, which permit a physician to complete adiagnostic evaluation[17]. A magnetic resonance imageis the relative response of specific nuclei to absorbedradiofrequency energy. Most magnetic resonance imagesare designed to observe the hydrogen nucleus becauseof its relative abundance in the body. A magneticresonance image is usually a tomographic map of thedistribution of protons in the imaged sample. However,image contrast is influenced more by other physicalfactors, including differences in the ability to re-emit theabsorbed radiofrequency signal and flow phenomena.Magnetic resonance imaging is similar to computerizedtomography as a cross-sectional imaging modality, butit is non-ionizing and offers additional soft tissue con-trast. The imaging sequences can be modified to visual-ize blood flow and to compensate for the blurring effectsof cardiac or respiratory motion. Magnetic resonancealso offers the unique ability to acquire images invirtually any orientation, without repositioning thepatient. This provides greater convenience for the medi-cal staff and minimizes patient discomfort. In addition,magnetic resonance imaging provides chemical infor-mation not measurable with conventional imagingmodalities. The above-mentioned unique abilities haveaccelerated the acceptance of magnetic resonanceimaging in current medical practice.

Magnetic resonance imaging in cardiovascular diseasesMagnetic resonance imaging provides substantial infor-mation on cardiac anatomy and measurements ofcardiac function, such as flow, perfusion, wall motion,and metabolism[19]. Special magnetic resonance taggingsequences allow non-invasive assessment of myocardial

deformation during contraction/relaxation[20–25]. Mag-netic resonance imaging may also be an invaluable toolwith which to study the influence of pacing site andmultisite pacing on cardiac function. While clinicallyused pacing produces considerable deterioration ofmechanical performance[26–32], pacing from various siteshave shown that more synchronous activation of theventricles is associated with improvements in cardiachaemodynamic function[33–35], which may be demon-strated by magnetic resonance imaging. Cardiac stressstudies using magnetic resonance compatible cathetersmay detect abnormalities in cardiac function and flowthat may become apparent only during increaseddemand[36–39]. In addition, a stable heart rate imposedby pacing would offer alternative means of preventingartifacts in all magnetic resonance imagingtechniques[40–42].

Effects of magnetic resonance imaging oncardiac pacemakers

In vitro and animal studiesThe potential hazardous effects of magnetic resonanceimaging in patients with cardiac pacemakers have beenstudied since 1983. Pavlicek and colleagues were the firstto report the effects of magnetic resonance imaging onpacemaker function[1]. They showed that radiofrequencyfields present in an magnetic resonance unit could poss-ibly inhibit demand pacemakers and time-varying mag-netic fields could generate pulse amplitudes to mimiccardiac activity. The threshold for initiating the asyn-chronous mode of a pacemaker was reported to be aslow as 17 gauss. The possibility of altering pacemakerparameters was presented as a serious limitation ofmagnetic resonance imaging. Fetter et al. showed thatpacemakers reverted from the demand to the asynchro-nous mode within the magnetic field of the scanner (0·15Tesla), but microscopic testing showed no evidence ofreed switch sticking or magnetizing, or damage to otherdiscrete pacemaker components[2].

Other investigators studied the feasibility of dual-chamber pacing systems in the magnetic resonanceenvironment. Erlebacher et al. tested different DDDpacemakers in a saline phantom, and showed thatduring scanning at 0·5 Tesla, all units malfunctioneddue to radiofrequency interference which caused totalinhibition of atrial and ventricular output, or resulted inatrial pacing at very high rates[3]. The potential for rapidcardiac stimulation during magnetic resonance scanningwas also reported in animal studies[43]. Lauck et al.investigated the performance of different stimulationmodes (VVI, VVIR, VOO, DDD, DDDR and DOO)during magnetic resonance scanning at 0·5 Tesla[44].Reversible activation of the reed switch with consecutiveasynchronous stimulation was observed in all pace-makers. Pacemakers in the asynchronous mode werenot affected with regard to stimulation rate andcapture during scanning. In contrast, pacemakers with

Potential benefits of magnetic resonanceimaging in paced patients

General indicationsIn patients with permanent pacemakers, magneticresonance imaging may provide important diagnosticbenefits, since this technique has advantages over otherimaging modalities[18]. Magnetic resonance imaging hasnow become the procedure of choice for all congenital,traumatic, hereditary, vascular, infectious, autoimmune,metabolic, and neoplastic disorders of the centralnervous system. It is the procedure of choice for furtherevaluation of the entire range of musculoskeletal dis-orders, when the physical examination or plain radiogra-phy suggests a serious abnormality. Magnetic resonanceimaging shows excellent contrast between tumour andother tissues, therefore it has become an indispensabletool in oncology. Cardiovascular magnetic resonanceimaging is useful in the evaluation of congenital andacquired diseases of the heart and great vessels and hasbeen a rapidly advancing area of clinical research[19].

Eur Heart J, Vol. 22, issue 2, January 2001

116 F. Duru et al.

automatic mode switching to demand pacing or pro-grammed inactivation of the reed switch were triggeredin the dual chamber mode and were inhibited in thesingle chamber mode. Thus, the investigators recom-mended programming into the asynchronous modeprior to scanning, and in those without permanentpacemaker dependency, complete inactivation of thesystem, if possible.

The effects of more powerful magnetic resonancescanners (i.e., 1·5 Tesla) on cardiac pacemakers wereinitially reported by Hayes et al.[45]. In vivo evaluationof different single and dual-chamber pacemakers showedreversion into asynchronous mode and transient reedswitch inhibition. Seven of the eight pulse generatorspaced rapidly when exposed to the radiofrequency signalassociated with a marked decrease in blood pressure.Stimulation cycle length was 200 ms (300 beats . min�1)corresponding to the frequency of pulsing. It was pro-posed that rapid pacing was the result of an ‘antenna’effect that couples the radiofrequency energy backinto the pacemaker output circuits. More recently,Achenbach et al. showed in a phantom study on 11pacemakers and 25 leads that no pacemaker malfunc-tion was observed in asynchronous pacing mode (VOO/DOO), whereas inhibition and rapid pacing wereobserved during spin-echo imaging if the pacemakerswere set to VVI or DDD mode[46]. The authors sug-gested that rapid pacing was caused by induction ofcurrents above the sensing threshold in the atrial leadand consequent triggering of ventricular stimulation.Direct interference with the pacemaker electronicsseemed to be an unlikely explanation, because the rapidpacing rate was always equal to the programmed fre-quency limit. The heating effect of pacemaker leads wasalso investigated in this study. Continuous registrationof the temperature at the lead tip using an opticaltemperature sensor showed a maximal temperatureincrease of 63·1 �C during 90 s of scanning. In sevenelectrodes, the temperature increase exceeded 15 �C.Resultant myocardial necrosis could be demonstrated inhistological studies.

Human studiesThere are few anecdotal reports of unexpected deaths inpatients undergoing magnetic resonance imaging[6–8]. Inone case, the patient had no escape ventricular rhythmand apparently died due to asystole. Another patientdeveloped ventricular fibrillation during the imagingprocedure that was not recognized immediately becauseECG monitoring was not used[47]. It is likely that manypotential complications are unreported in the literaturefor various reasons (liability, etc.). On the other hand,there are also reports of pacemaker patients who under-went magnetic resonance imaging safely (Table 2).Therefore, differences exist among clinicians regardingthe perceived safety of scanning paced patients[48].

In patients who underwent magnetic resonanceimaging of the head, no pacemaker malfunction wasobserved with the pacemaker turned off or programmedto an asynchronous pacing mode prior to magnetic


resonance exposure[9–11]. In another study on fivepatients with pacemakers, Gimbel et al. reported normalpacemaker performance in four patients during mag-netic resonance imaging (0·35 and 1·5 Tesla)[12]. Onepatient had a pause of approximately 2 s in durationnear the completion of magnetic resonance imaging, thecause of which could not be determined. This occurredin a pacemaker dependent patient with a unipolar dual-chamber device programmed to the DOO mode. Norapid cardiac pacing occurred and no patient reported atorque or heating sensation. Fontaine et al. reported acase of rapid cardiac pacing during magnetic resonanceimaging (1·5 Tesla) in a patient with a dual chamberpacemaker[49]. The patient developed an irregular ven-tricular rhythm during radiofrequency pulsing whichterminated with the cessation of radiofrequency pulsing.

Magnetic resonance imaging at 0·5 Tesla was shownto have no influence on atrial and ventricular stimula-tion thresholds, P and R wave amplitudes, electrodeimpedance, battery voltage, current, and impedancemeasurements in patients with implanted pacemakers[13].In the largest reported series to date, Sommer et al.studied 14 pacemaker models in vitro and 18 patientsundergoing magnetic resonance scanning at 0·5 Teslawith standard spin, turbo spin, and gradient echosequences[14]. All pacemakers switched to the asynchro-nous mode due to activation of the reed switch in thestatic magnetic field in vitro. Atrial and ventricularstimulation thresholds remained unchanged. Pacemakerprogramme changes, damage to pacemaker compo-nents, dislocation/torque of the pacemaker and rapidpacing of the pacemaker were observed neither in vitronor in vivo. The investigators concluded that thepresence of an implanted pacemaker should not beconsidered an absolute contraindication to magneticresonance imaging at 0·5 Tesla.

Safety issues in patients with retained pacingleads

Many patients have endocardial pacemaker leads left inplace after pulse generator removal. The safety of mag-netic resonance imaging in patients with retained endo-cardial pacemaker wires has not been systematicallyinvestigated to date. Likewise, the behaviour of pacingpulmonary artery catheters during magnetic resonanceimaging is not known. On the other hand, the effect oftransoesophageal pacing leads during magnetic reso-nance imaging were investigated in animal studies whichshowed heating with consequential necrosis[50].

Magnetic resonance imaging in patients with retainedepicardial wires after cardiac surgery was previouslyconsidered to be a relative contraindication[51]. Tem-porary pacing wires, usually made of braided stainlesssteel, are sutured to the epicardial surface of the heartover the right ventricle and right atrium after cardiacsurgery, and connected to an external pacemaker if thepatient develops bradycardia or atrioventricular block.

Review 117

Theoretical calculations using a circuit formed by epi-cardial pacing wires showed induction of currents up to80 �A using a magnetic field strength of 1·5 Tesla[52].Hartnell et al. investigated the safety of 1 or 1·5 Teslamagnetic resonance systems operating with conventionalpulse sequences in 51 patients with retained epicardialpacing wires, cut short at the skin, after cardiac sur-gery[53]. None of the patients reported symptoms sug-gesting arrhythmia or other cardiac dysfunction duringmagnetic resonance imaging, and there were no changesfrom the baseline ECG rhythms. Therefore, retainedepicardial wires do not seem to present a hazard topatients in the magnetic resonance environment. How-ever, this conclusion applies mostly to non-cardiacmagnetic resonance examinations[54]. A survey amongneuroradiologists regarding magnetic resonance imag-ing in patients with retained epicardial pacemakerwires after cardiac surgery yielded varying practicepatterns[55]. Some physicians would perform magneticresonance examinations even if they know that wires arepresent, while others would not even screen for epi-cardial pacemaker wires. Despite the many exami-nations, no respondents had experienced a problemrelated to the presence of epicardial wires in themagnetic resonance environment.

Technical considerations

Electromagnetic fields in magnetic resonanceimaging

There are three types of electromagnetic fields used in anmagnetic resonance imaging unit:

1. The main static magnetic field is used to align theprotons. An intense static magnetic field (B0) is alwayspresent even when the scanner is not imaging. The fieldstrength is 0·5–1·5 Tesla in most of the currently avail-able magnetic resonance imaging units for clinical use.Current state-of-the-art technology is pushing this upperlimit to 4 or 5 Tesla in research magnetic resonanceimaging systems. This is about 100 000 times the mag-netic field strength of the earth. In addition to magneticfield strength, magnetic field homogeneity — themeasure of field uniformity within the measurement areaof the magnet — is an important parameter in themagnet system. Optimal field homogeneity is crucial togenerating images free from distortion and with themaximum possible signal-to-noise ratio.2. The pulsed radiofrequency field, generated by thebody-coil or the head-coil, is used to change the energystate of the protons and elicit magnetic resonance signalsfrom tissue. The radiofrequency field is homogeneous inthe central region and has two components: (a) Themagnetic field is circularly polarized in the axial plane,and (b) the electric field is related to the magnetic fieldby Maxwell’s equations. The radiofrequency field isswitched on and off during measurements and has a

frequency of 21–64 MHz depending on the magneticfield strength.3. The time-varying magnetic gradient fields, used forspatial localization, change their strength along differentorientations and operate at frequencies in the order of1 kHz. The vectors of the magnetic field gradients in thex, y, and z directions are produced by three sets oforthogonally positioned coils and are switched on onlyduring the measurements.

Potential problems in magnetic resonanceimaging equipments

The potential benefits of magnetic resonance imagingare numerous; however, there are hazards intrinsic to themagnetic resonance environment. These hazards may beattributed to one or to a combination of the three maincomponents that make up the magnetic resonanceenvironment.

1. Static magnetic field(a) Reed-switch closure. When a pacemaker is broughtclose to an magnetic resonance scanner, the reed switchcloses, and asynchronous pacing occurs, which maycompete with the underlying cardiac rhythm. For thisreason, it is common practice for patients with pace-makers to be prevented from undergoing magneticresonance examinations; they are also physicallyrestricted to remaining at, or outside of, at most, the5-gauss line surrounding the magnetic resonance imag-ing environment[56]. In patients with unstable condi-tions, such as myocardial ischaemia, there is substantialrisk for ventricular fibrillation during asynchronouspacing. A physician has to decide if asynchronouspacing over a longer time period, e.g., up to 1 h, isjustifiable or not. In most modern pacemakers themagnet function is programmable. If the magnet re-sponse is switched off, synchronous pacing is possibleeven in strong magnetic fields. In addition, the reedswitch may not close inside the homogeneous magneticfield in some special orientations. Programming thepacemaker to the asynchronous mode will avoid theinfluence of the reed switch during magnetic resonanceimaging.

Damage to the reed switch by the strong static mag-netic field is a theoretical concern, but it was notreported in any of the published studies[2,43]. Afterexposure to the static magnetic field of magnetic reso-nance equipment, the reed switch of various pacemakersclosed at the same field strength as before. Likewise,extended exposure to static magnetic fields (10 hexposure to 1·5 Tesla) did not result in any detectablechange on the reed switch, the telemetric coils, or in thepacemaker software settings[57].

(b) Pacemaker displacement. Some parts of pace-makers, such as batteries and reed switches, containferromagnetic materials; thus, mechanical forces occur


118 F. Duru et al.

and the pacemaker may be displaced. Pacemaker dis-placement may occur in response to magnetic force ormagnetic torque. There exists a magnetic force only ifthe magnetic field changes from place to place. It will bestrongest in an area with the greatest field change over ashort distance. In the centre of a magnetic resonancescanner, where the magnetic field is maximal, no mag-netic force can be measured. Magnetic force increaseswith increased distance from the magnet isocenter. Incontrast, magnetic torque tries to turn ferromagnetic orparamagnetic material parallel to the magnetic field. Ithas a linear association with the magnetic field strengthand will be strongest in the middle of the magneticresonance scanner.

The magnetic forces exerted on the pacemakers arerelatively small. In a preliminary study, we evaluated themagnetic force and torque on 32 pacemaker models(15 dual-chamber and 17 single-chamber units) in a1·5 Tesla magnetic resonance unit[58]. The measuredmagnetic force was in the range of 0·05–3·60 Newton.The newer generation of pacemakers had significantlylower magnetic force values, even lower than the gravity,as compared to the older devices. Likewise, the torquelevels were significantly reduced in modern pacemakers.Pacemaker displacement by the magnetic force of themagnetic resonance equipment was not reported in anyof the published studies[9–14].

Pacemakers are paramagnetic, leading to a preferredorientation parallel to a strong magnetic field. If apatient lies parallel to the main magnetic field, as in mostmagnetic resonance scanners (some ‘open’ magneticresonance scanners are exceptions), the orientation ofthe pacemaker is nearly optimal, so the magnetic torquethat rotates the pacemaker may be relatively small(Fig. 1). However, exceptional situations may arisewhen pacemakers are near the entrance of the magnetic


resonance scanner. The position of the thorax (where thepacemaker and the corresponding leads are located)with respect to the magnet bore is also important.

The metallic parts of the leads are usually composedof MP35N. This alloy of nickel, cobalt, chromium, andmolybdenum is non-ferromagnetic; therefore, there is noconcern that such leads will move or dislodge[59].

Pacemaker

m

B0

No torque

m

F

–F

Strong torque

B0

Figure 1 When the orientation of the pacemaker is optimal, the magnetic torque is relatively small (left).Stronger torque occurs in the presence of an orthogonal position of the pacemaker relative to the mainmagnetic field (B0) of the magnetic resonance scanner (right).

(c) Changes in electrocardiograms. An electric potentialis produced by a moving conductor, such as flowingblood, beating heart, or respiring lungs in a staticmagnetic field. The induced electrical potentials fromthis motion are signals superimposed on the surfaceelectrocardiogram[60,61]. Polarization of the flowingblood in the magnetic field may often result in increasedT-wave amplitude. Gaffey et al. demonstrated reproduc-ible T wave alterations with a threshold near 0·3 Tesla,the extent of which increased linearly with fieldstrength[62]. Electrocardiogram alterations have alsobeen observed at field strengths as low as 0·1 Tesla[63].These observations are in agreement with theoreticalcalculations that the electrocardiographic changes arefrom the induced electromagnetic force associated withhigh velocity blood flow perpendicular to the staticmagnetic field and presumed to be caused when thenegatively charged red blood cells course around theaortic arch[64]. The induced electrical potentials fromthe blood motion can be formulated as:

Potential (mV)=1

10 Magnetic field (T) · Blood velocity (cm/s) ·

Vessel diameter (cm)

It was concluded theoretically that the impact ofmagnetohaemodynamic effects is not important physio-logically, because the current densities associated with

Review 119

these potentials are much lower than that needed forelectrical excitation of the heart[65]. No arrhythmias orchanges in heart rate or blood pressure were inducedin vivo, and the normal electrocardiogram appearedimmediately on cessation of magnetic field exposure[66].However, the existence of T wave alterations may beconfused with signs of disease if the magnetic influence isnot recognized. Occasionally, there may be sufficientelevation of T waves to produce technical difficultieswhile attempting to gate or trigger the radiofrequencyand the gradient pulse timing of the magnetic resonanceimaging procedure relative to the timing of the QRScomplex of the cardiac cycle.

The strong static magnetic field of an magneticresonance scanner may also slightly change the intra-cardiac electrograms. However, pacemakers use electro-gram deflections which correspond to P and R waves onsurface electrocardiogram for sensing. The shape ofthese deflections are not greatly altered, because in thistime period blood flow and heart motion are relativelysmall.

2. Radiofrequency field(a) Heating. At the frequencies of interest in magneticresonance imaging, some of the radiofrequency energy isabsorbed and converted to heat. In fact, the mainbiological effects associated with exposure to radiofre-quency radiation are related to the thermogenic qualitiesof the radiofrequency field[67–74]. The power depositedby radiofrequency pulses during magnetic resonanceimaging is dependent upon multiple factors, includingthe power and duration of the radiofrequency pulse, thetransmitted electromagnetic field frequency, the numberof radiofrequency pulses applied per unit time, the typeand configuration of the radiofrequency transmitter coilused, the volume of tissue imaged, the electrical resis-tivity of the tissue, the configuration of the anatomicalregion imaged, etc[56]. The causes of heating in themagnetic resonance environment are twofold: (a) radio-frequency field coupling to the lead can occur, inducingsignificant local heating, and (b) currents induced duringthe radiofrequency transmission can cause local Ohm’sheating next to the tip of the lead.

Guidelines have been established on the allowablewhole body exposure, expressed as specific absorptionrate (SAR) given as watts per kilogram. The maximumtemperature elevation (ignoring cooling effects) can bedetermined by the following formula:

�T (�C)=SAR (W/kg) · time (s)

specific heat (J/g · �C)

For a specific absorbed power of 4 W . kg�1 for10 min, the maximum temperature rise equals0·7 �C using the soft tissue specific heat of0·83 kcal . kg�1 . �C�1. Because of specific absorptionrate limitations, magnetic resonance for diagnosticimaging has not been associated with any detrimentaleffects, and it appears that this level of whole-bodyspecific absorption rate is acceptable. However, thelocal electric field can be amplified near conducting

instruments making the peak specific absorption ratedifficult to predict. The radiofrequency field in an mag-netic resonance scanner has sufficient energy to causelocal heating of long conductive wires, such as pace-maker leads, which could destroy parts of the adjacentmyocardial tissue. The effects of electrode heating arenot detectable by monitoring during magnetic resonanceimaging. However, an increase in pacing threshold,myocardial perforation and lead penetration, or evenarrhythmias caused by scar tissue are among thepotential concerns long after scanning, but such long-term heating effects of magnetic resonance imaging havenot been studied yet.

Theoretical calculations, in agreement with in vitroexperimental findings, showed that no heatingproblem occurs in patients with large compact metallicimplants[75]. Therefore, pulse generators without leadsare not expected to cause heating problems. In a prelimi-nary study, we investigated the heating effects of mag-netic resonance imaging at 1·5 Tesla on 10 various leads(screw-in/passive, uni/bi-polar, coaxial/coradial) invitro[76]. Temperature increase exceeded 10 �C in eightleads and 20 �C in five leads. Under special conditionswith the lead tip at the surface of the saline tank,maximal temperature increase of 69 �C was observed.None of the tested leads appeared to be particularly safeagainst heating. The findings were similar when the leadswere isolated or connected to pulse generators. There-fore, retained pacemaker leads (yet without a pace-maker) may present a hazard to patients in the magneticresonance environment.

Estimation of the heating problem in leads. A thin linearwire has a maximal absorption rate with half the lengthof �, where � represents the wavelength of the electro-magnetic waves. In a vacuum, the wavelength of theradiofrequency field in a 1·5 Tesla magnetic resonancescanner is approximately 4·7 m, hence maximal absorp-tion will occur in a wire of 2·35 m. In case of salinewater, the critical length will be reduced to 30 cm. For ashorter wire (l<�/2), the absorption of energy is reducedby the factor [l/(�/2)]3. However, there are other factorsthat may influence the degree of heating, such as thegeometrical structure of the wires, the placement in thebody, the insulation of the wires. The strongest heatingoccurs in areas with high changes of the electric field.Tissue next to sharp edges and points will be exposed tothe highest dose of thermal energy. On the other hand,the constant flow of blood around the leads may have acooling effect. The impact of magnetic resonance scan-ning at 0·5 Tesla with respect to heating should be lessthan that of 1·5 Tesla, since the power absorption of alinear antenna is proportional to the square of thefrequency of the radiofrequency field[77]. Hence, theuse of a 0·5 Tesla system reduces the power absorp-tion considerably[78]. Likewise, magnetic resonancesequences with a low specific absorption rate (e.g.gradient echo preferred over spin echo) may decreaseheating. An accurate calculation of the heating problemcould not been performed to date.


120 F. Duru et al.

(b) Alterations in pacing rate. The radiofrequency fieldmay induce an undesirable fast pacing rate[3,44,45]. Twomechanisms have been proposed to explain rapid pac-ing: direct interference with pacemaker electronics maybe possible, because pacing of up to 300 beats . min�1

synchronized to the radiofrequency pulses has beenobserved in several studies[3,44]. In contrast, rapid pacingat the upper tracking limit in DDD pacemakers may becaused by induction of currents above the sensingthreshold in the atrial lead and consequent triggering ofventricular stimulation[45].

(c) Pacemaker reprogramming or reset. Changing theprogrammed parameters (phantom programming) orresetting a pacemaker during magnetic resonance inves-tigation may also be a concern. The newer generation ofpacemakers have ‘security checks’ at each sequence ofpacemaker programming that may avoid changingparameters. In addition, the telemetry frequencies ofmost manufacturers (32–175 kHz) are out of the rangeof the frequencies of the radiofrequency and gradientfields.

3. Time-varying magnetic gradient fields(a) Induction voltage. The contribution of the time-varying gradient fields to the total strength of themagnetic field is negligible; however, an effect may occurbecause these fields are rapidly applied and removed.The important factors are the rate with which themagnetic field changes and the length of time thischanging field is applied. Budinger has calculated thatan electrical current density of 1 �A . cm�2 will beinduced for a time-varying field of 1 Tesla . s�1[66]. Evenusing today’s gradient systems with a time-varying fieldup to 50 Tesla . s�1, the induced currents are likelyto stay below the biological thresholds for cardiacfibrillation (in the range of 100 to 1000 �A . cm�2).

Using the induction law:

V=dB/dt · � · r2

where V=induced voltage (volts), dB/dt=time rate ofmagnetic field change (Tesla . s�1), and �r2=area ofwire loop (m2), the amount of induced voltage is pro-portional to the area of the wire loop. The gradient fieldcan induce a critical voltage in unipolar leads. In bipolarleads this danger should be smaller.

Estimation of the induced voltage in unipolar leads. Thearea to be considered in the induction law has beenreported to be between 200 to 500 cm2[79]. A theoreticalupper limit for the induced voltage is 20 V. Such avoltage during more than 0·1 ms will be enough to pacethe heart. If the regular gradients are applied and thez-projection of the area is taken into account, theinduced voltage will be reduced.

Estimation of the induced voltage in bipolar leads. Ifbipolar leads are used, the area between the two wires ismuch smaller. Stimulation of the heart should no


longer be possible. The induced voltage, however, mayinfluence sensing.

(b) Heating The induced current may lead to localheating. However, assuming that the magnetic field ofthe gradients and the radiofrequency are orthogonal, thecalculated heating effect of the gradient fields is muchless compared to that caused by the radiofrequency field,and therefore, may be neglected.

(c) Reed-switch closure. It may be possible to open andreclose the reed switch in the main magnetic field by thegradient field. Theoretically, few reed switch orienta-tions inside the homogeneous main magnetic field exist,in which case it does not close. In such a case, it ispossible that the gradient fields could close and reopenthe reed switch.

Magnetic resonance imaging of differentbody structures

A critical issue concerning safety of patients with pace-makers undergoing magnetic resonance imaging is thestructure of the body to be scanned (e.g. thorax vs heador extremities). Most of the data collected in this regardhas been of the brain. In such cases, the heart andpacemaker system are outside the isocentre of the mag-netic resonance scanner, leading to a reduction in theradiofrequency field strength at the device. With the useof a transmit/receive head coil, the radiofrequency fieldat the area of the lead and the pacemaker will be furtherreduced. On the other hand, the pacemaker will be closeto the portal of the magnetic resonance scanner wherethe highest magnetic forces on the device occur, asdiscussed previously[58].

A similar reduction of the radiofrequency field inter-actions will occur with magnetic resonance imaging ofthe lower extremities. Recently, a specially designed, lowfield strength (0·2 Tesla magnetic resonance system,Artoscan, Lunar Corp., Madison, WI/Esaote, Genoa,Italy) magnetic resonance system has become availablefor imaging of extremities[80]. This magnetic resonancesystem has a small magnet bore where only the extremitywill be placed (while the rest of the body remainsoutside), leading to very low radiofrequency, gradientand static magnetic fields at the thorax. Because ofthe unique design features of the extremity magneticresonance system, it has been suggested that it may bepossible to perform extremity magnetic resonance imag-ing in patients with cardiac pacemakers and implantablecardioverter defibrillators[81].

The clinical use of magnetic resonance for cardio-vascular imaging is much less common today than‘non-cardiac’ magnetic resonance, which may be neededin a patient with an implanted pacemaker. The safety ofusing magnetic resonance imaging for cardiac evaluationin pacemaker recipients has not been tested on anysignificant scale to date.

Review 121

Image artifacts due to pacemaker systemcomponents

Image artifacts can be caused by the presence of pace-maker system components that are in or near theimaging field of view in the magnetic resonance environ-ment. This is due to the static magnetic field of theimplanted components that can perturb the relation-ship between the position and frequency essential foraccurate image reconstruction. The pulse generatorshave a magnetic susceptibility that is significantly differ-ent from that of tissue, therefore marked distortion mayresult (Fig. 2). The amount of image artifact producedby the leads may only be of concern if imaging anatomyis in the immediate vicinity.

Additional problems with implantablecardioverter-defibrillators

Problems with implantable cardioverter-defibrillators inthe magnetic resonance environment may be expected tobe similar to those observed with pacemakers. However,implantable cardioverter-defibriliators use different andlarger batteries that may cause higher magneticforces[58]. Despite a dramatic reduction in size andweight, new generation implantable cardioverter-defibriliators may still pose problems due to strongmagnetic torque[58]. The sensitivity of implantablecardioverter-defibriliators is normally much higherthan that of pacemakers. Therefore, implantablecardioverter-defibriliator devices may falsely detect aventricular tachyarrhythmia and subsequently deliverantitachycardia pacing, cardioversion or defibrillation

therapies, which can lead to an actual ventricular tachy-arrhythmia. In addition, magnetic fields may preventdetection of ventricular tachycardia or fibrillation.The heating problem of implantable cardioverter-defibriliator leads can be expected to be the same as inpacemakers.

��

Figure 2 Magnetic resonance image artifacts caused by pacemakers and endocardial pacing leads: (a)Marked image disturbances around the pacemakers may be observed (pacemaker positions are shown by therectangles). Bending and obliteration of contours within the imaging phantom (saline container) are noted.(b) The non-ferromagnetic leads show only mild artifacts appearing as small signal-void areas surrounded byor adjacent to high signal zones (arrows).

Future directions for magnetic resonanceimaging compatible pacing

Any component of a pacing system, which is introducedinside the patient, must fulfil a number of stringentrequirements. Instruments to be designed must satisfygeneral safety considerations unique to the magneticresonance environment. In addition, they should distortthe image as little as possible. It is possible that themagnetic resonance environment exerts less magneticforce and torque on newer pacemakers with smallerbattery and size, as compared to the older pacemakers.Future studies are needed to investigate the safety ofmagnetic resonance imaging in patients with modernpacemakers.

Currently available pacing leads contain metal parts,which may cause hazardous side effects and distortimage quality, therefore they are not suitable for usein the magnetic resonance environment. Non-ferromagnetic materials such as copper, with suscepti-bilities close to body tissue are desirable for use in pacingcatheters. Copper is an excellent conductor and repre-sents an almost ideal material for pacemaker leads.However, it is not biocompatible, a problem which canbe solved by appropriate insulation. Several investi-gators have developed magnetic resonance imaging


122 F. Duru et al.

compatible leads and reported the initial in vitro andin vivo results. Jerzewski et al. designed a pacing leadby modifying a commercially available 6F pacingcatheter[82]. The stainless steel shaft braiding was omit-ted and 90% platinum/10% iridium electrodes were usedat the tip. The electric wiring was made of nearly purecopper. In phantom experiments, the lead was observedto cause only minor image artifacts, and in vivo, externalpacing with this lead proved to be effective. Recently,Halperin et al. have also shown that diagnostic electro-physiological studies and catheter ablation under mag-netic resonance guidance may be feasible using leadsmade from non-magnetic materials[83]. However, thepotential for tissue heating around the leads was notconsidered in these studies. Therefore, absolute compat-ibility of these leads remains of concern and requiresfurther investigation.

Conclusions

Previous studies showed that magnetic resonance imag-ing in patients with implanted pacemakers may behazardous. However, there are newer anecdotal reportsof safe magnetic resonance imaging when this diagnostictechnique was an absolute necessity. While it is possibleto image a person safely, one must be aware of allvariables to ascertain safety. The effects of magneticresonance systems on the function of pacemakers andimplantable cardioverter-defibriliators depend on vari-ous factors including the strength of the static magneticfield, the pulse sequence used, the anatomic region beingimaged, and many other factors. In addition, eachmanufacturer’s pacemakers and each pacing modalitymay behave differently. Therefore, knowledge of thebehaviour of the specific pacemaker model in vitro maybe helpful. More extensive in vivo testing using differentmodels of pacemakers from various manufacturers isneeded to identify device-specific susceptibility to theelectromagnetic fields induced by magnetic resonanceimaging.

At present, magnetic resonance imaging should beavoided in patients with pacemakers and implantablecardioverter-defibriliators until more information (e.g.results of large studies/registries) is known. If non-magnetic resonance imaging modalities are not adequateto make a diagnosis, after a careful risk–benefit assess-ment, magnetic resonance imaging at low field strengths(0·5 Tesla or less) with continuous monitoring (e.g.ECG/pulse oximetry) during the procedure may beconsidered but only in experienced centres. Body scansand magnetic resonance sequences with a high specificabsorption rate should be avoided whenever possible.Changes in pacemaker programming (ideally OOOmode or programming off in non-pacemaker dependentpatients or asynchronous mode in pacemaker dependentpatients) may be performed prior to scanning. Introduc-ing the patient slowly into the magnet with monitoring,and starting magnetic resonance imaging with graded


scanning sequences (single slice, low resolution) andthen eventually progressing to conventional sequencesshould be considered. In any case, clinicians should keepin mind the potential for severe adverse effects withpacemaker use in magnetic resonance systems. There-fore, experienced cardiology assistance and full resusci-tation facilities should be available during scanning.While a few patients (and physicians) may have beenfortunate in the past, it does not mean that magneticresonance scanning of pacemaker patients is safe, since afuture patient may present with a pacemaker configura-tion that corresponds to one of the extreme scenarios.Even with the same pacemaker configuration and mag-netic resonance settings, a safe procedure at one timedoes not guarantee a safe repeat examination.

With continuing progress in magnetic resonanceimaging technology, higher gradient changes over timeare being routinely applied to acquire thinner slices andmore rapid images. Therefore, the feasibility and safetyof magnetic resonance imaging in some studies may notbe extrapolated to techniques using faster field strengthsor other imaging protocols that are in use or yet to bedeveloped. However, innovations in pacemaker and leadtechnology may enable magnetic resonance-compatiblepacing systems in the foreseeable future.

This work is supported by a grant from the Bakken ResearchCenter, Medtronic Inc., Maastricht, the Netherlands.

References

[1] Pavlicek W, Geisinger M, Castle L, Borkowski GP, MeaneyTF, Bream BL, Gallagher JH. The effects of nuclear magneticresonance on patients with cardiac pacemakers. Radiology1983; 147: 149–53.

[2] Fetter J, Aram G, Holmes DR, Gray JE, Hayes DL. Theeffects of nuclear magnetic resonance imagers on external andimplantable pulse generators. PACE 1984; 7: 720–7.

[3] Erlebacher JA, Cahill PT, Pannizzo F, Knowles RJ. Effect ofmagnetic resonance imaging on DDD pacemakers. Am JCardiol 1986; 57: 437–40.

[4] Shellock FG, Kanal E. SMRI Report: Policies, guidelines andrecommendations for MR imaging safety and patient manage-ment. J Magnetic Resonance Imaging 1992; 2: 247–8.

[5] Hayes DL, Vlietstra RE. Pacemaker malfunction. Ann InternMed 1993; 119: 828–35.

[6] Avery JE. Loss prevention case of the month. Not myresponsibility! J Tenn Med Assoc 1988; 81: 523.

[7] US Food and Drug Administration, Center for Devices andRadiological Health MR product reporting program andmedical device report program. MDR 175218, 351516,125938.

[8] Schiebler M, Kaut-Watson C, Williams DL. Both sedated andcritically ill require monitoring during MRI. MR 1994;Winter: 41–5.

[9] Alagona P, Toole JC, Maniscalco BS, Glover MU, AbernathyGT, Prida XE. Nuclear magnetic resonance imaging in apatient with a DDD pacemaker (letter). PACE 1989; 12: 619.

[10] Inbar S, Larson J, Burt T, Mafee M, Ezri MD. Case report:nuclear magnetic resonance imaging in a patient with apacemaker. Am J Med Sci 1993; 305: 174–5.

[11] Garcia-Bolao I, Albaladejo V, Benito A, Alegria E, ZubietaJL. Magnetic resonance imaging in a patient with adual-chamber pacemaker. Acta Cardiol 1998; 53: 33–5.

Review 123

[12] Gimbel R, Johnson D, Levine P, Wilkoff B. Safe perform-ance of magnetic resonance imaging on five patients withpermanent cardiac pacemakers. PACE 1996; 19: 913–9.

[13] Vahlhaus C, Sommer T, Lewalter T et al. Magnetic resonanceimaging at 0·5 tesla in patients with cardiac pacemakers.Contraindication or justifiable risk? (Abstr) Eur Heart J 1998;19(Abstr Suppl): 208.

[14] Sommer T, Lauck G, Schimpf R et al. MRT bei Patienten mitHerzschrittmachern: In-vitro- und In-vivo-Evaluierung bei 0,5Tesla. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr1998; 168: 36–43.

[15] Mansfield P, Maudsley AA. Medical imaging by NMR. Br JRadiol 1977; 50: 188–94.

[16] Hinshaw WS, Bottomley PA, Holland GN. Radiographicthin-section image of the human wrist by nuclear magneticresonance. Nature 1977; 270: 722–3.

[17] An introduction to magnetic resonance imaging. In: BasicPrinciples of MR Imaging. Philips Medical Systems.

[18] Stark DD. Clinical indications for MRI. In: Magin RL,Liburdy RP, Persson B, eds. Biological effects and safetyaspects of nuclear magnetic resonance imaging and spec-troscopy. New York: The New York Academy of Sciences:1992: 332–4.

[19] Task Force of the European Society of Cardiology, in collab-oration with the Association of European Pediatric Cardiolo-gists. The clinical role of magnetic resonance in cardiovasculardisease. Eur Heart J 1998; 19: 19–39.

[20] Rademakers FE, Rogers WJ, Guier WH et al. Relation ofregional cross-fiber shortening to wall thickening in the intactheart: Three-dimensional strain analysis by NMR tagging.Circulation 1994; 89: 1174–82.

[21] Palmon LC, Reichek N, Yeon SB et al. Intramural myocardialshortening in hypertensive left ventricular hypertrophy withnormal pump function. Circulation 1994; 89: 122–31.

[22] Fischer SE, McKinnon GC, Scheidegger MB, Prins W, MeierD, Boesiger P. True myocardial motion tracking. Magn ResonMed 1994; 31: 401–13.

[23] Young AA, Imai H, Chang C, Axel L. Two-dimensionalleft ventricular deformation during systole using magneticresonance imaging with spatial modulation of magnetization.Circulation 1994; 89: 740–52.

[24] Fischer SE, McKinnon GC, Maier SE, Boesiger P. Improvedmyocardial tagging contrast. Magn Reson Med 1993; 30:191–200.

[25] Maier SE, Fischer SE, McKinnon GC, Hess OM,Krayenbuehl HP, Boesiger P. Acquisition and evaluation oftagged magnetic resonance images of human left ventricle.Computerized Med Imaging 1992; 16: 73–80.

[26] Tse HF, Lau CP. Long-term effect of right ventricular pacingon myocardial perfusion and function. J Am Coll Cardiol1997; 29: 744–9.

[27] Prinzen FW, Augustijn CH, Allessie MA, Arts T, Delhaas T,Reneman RS. The time sequence of electrical and mech-anical activation during spontaneous beating and ectopicstimulation. Euro J Cardiol 1992; 13: 535–43.

[28] Delhaas T, Arts T, Prinzen FW, Reneman RS. Relationbetween regional electrical activation time and subepicardialfiber strain in the canine left ventricle. Euro J Physiol 1993;423: 78–87.

[29] Prinzen FW, Cheriex EC, Delhaas T et al. Asymmetricthickness of the left ventricular wall resulting from asynchro-nous electric activation: a study in dogs with ventricularpacing and in patients with left bundle branch block. AmHeart J 1995; 130: 1045–53.

[30] Augustijn CH, Arts T, Prinzen FW, Reneman RS. Mappingthe sequence of contraction of the canine left ventricle. Euro JPhysiol 1991; 419: 529–33.

[31] van Oosterhout MF, Prinzen FW, Arts T et al. Asynchronouselectrical activation induces asymmetrical hypertrophy of theleft ventricular wall. Circulation 1998; 98: 588–95.

[32] McVeigh ER, Prinzen FW, Wyman BT, Tsitlik JE,Halperin HR, Hunter WC. Imaging asynchronous mechanical

activation of the paced heart with tagged MRI. Magn ResonMed 1998; 39: 507–13.

[33] Buckingham TA, Candinas R, Schlapfer J et al. Acute hemo-dynamic effects of atrioventricular pacing at different sites inthe right ventricle individually and simultaneously. PACE1996; 20: 909–15.

[34] Buckingham TA, Duru F, Candinas R. Hemodynamic effectsof alternate and multisite pacing. In: Santini M, ed. Progressin Clinical Pacing: Armonk, NY: Futura Media Services, Inc.,1996: 77–82.

[35] Buckingham TA, Candinas R, Duru F et al. Acute hemody-namic effects of alternate and combined pacing in patientsafter cardiac surgery. PACE 1999; 22: 887–93.

[36] Schaefer S, Peshock RM, Parkey RW, Willerson JT. A newdevice for exercise MR imaging. Am J Roentgenol 1986; 147:1289–90.

[37] Pennell DJ, Underwood SR, Manzara CC et al. Magneticresonance imaging during dobutamine stress in coronaryartery disease. Am J Cardiol 1992; 70: 34–40.

[38] Baer FM, Smolarz K, Jungehulsing M et al. Feasibility ofhigh-dose dipyridamole-magnetic resonance imaging fordetection of coronary artery disease and comparison withcoronary angiography. Am J Cardiol 1992; 69: 51–6.

[39] Hofman MB, de Cock CC, van der Linden JC et al.Transesophageal cardiac pacing during magnetic resonanceimaging: feasibility and safety considerations. Magn ResonMed 1996; 35: 413–22.

[40] Lenz GW, Haacke EM, White RD. Retrospective cardiacgating: a review of technical aspects and future directions.Magn Reson Imaging 1989; 7: 445–55.

[41] Spraggins TA. Wireless retrospective gating: application tocine cardiac imaging. Magn Reson Imaging 1990; 8: 675–81.

[42] Prato FS, Drost DJ, King M et al. Cardiac T1 calculationsfrom MR spin-echo images. Magn Reson Med 1987; 4:227–43.

[43] Holmes DR, Hayes DL, Gray JE, Merideth J. The effects ofmagnetic resonance imaging on implantable pulse generators.PACE 1986; 9: 360–70.

[44] Lauck G, von Smekal A, Wolke S et al. Effects of nuclearmagnetic resonance imaging on cardiac pacemakers. PACE1995; 18: 1549–55.

[45] Hayes DL, Holmes DR, Gray JE. Effect of 1·5 T nuclearmagnetic resonance imaging scanner on implanted permanentpacemakers. PACE 1987; 10: 782–6.

[46] Achenbach S, Moshage W, Diem B, Bieberle T, Schibgilla V,Bachmann K. Effects of magnetic resonance imaging oncardiac pacemakers and electrodes. Am Heart J 1997; 134:467–73.

[47] Pohost GM, Blackwell GG, Shellock FG. Safety of patientswith medical devices during application of magnetic resonancemethods. Ann N Y Acad Sci 1992; 649: 302–12.

[48] Gimbel JR, Lorig RJ, Wilkoff BL. Survey of magnetic reso-nance imaging in pacemaker patients. HeartWeb (periodicalonline, http:/www.heartweb.org/) Accessed Oct. 1, 1996.

[49] Fontaine JM, Mohamed FB, Gottlieb C, Callans DJ,Marchlinski FE. Rapid ventricular pacing in a pacemakerpatient undergoing magnetic resonance imaging. PACE 1998;21: 1336–39.

[50] De Cock CC, Hofman MBM, van Rossum AC, Visser FC,Visser CA. Safety and feasibility of transesophageal pacingduring magnetic resonance imaging (abstract). Eur Heart J1995; 16: 401.

[51] Shellock FG, Kanal E. Safety considerations of cardiovascu-lar magnetic resonance studies. In: Pohost GM, ed. Cardio-vascular applications of magnetic resonance. Mount Kisco,NY: Futura, 1993: 427–46.

[52] Peden CJ, Collins AG, Butson PC, Whitwam JG, Young IR.Induction of microcurrents in critically ill patients in magneticresonance systems. Crit Care Med 1993; 21: 1923–8.

[53] Hartnell GG, Spence L, Hughes LA, Cohen MC, Saouaf R,Buff B. Safety of MR imaging in patients who have retainedmetallic materials after cardiac surgery. Am J Roentgenol1997; 168: 1157–9.


124 F. Duru et al.

[54] Kanal E. Safety of MR imaging in patients with retainedepicardial pacer wires (letter). Am J Roentgenol 1998; 170:213.

[55] Murphy KJ, Cohan RH, Ellis JH. MR imaging in patientswith epicardial pacemaker wires. Am J Roentgenol 1999; 172:727–8.

[56] Kanal E. An overview of electromagnetic safety considera-tions associated with magnetic resonance imaging. In: MaginRL, Liburdy RP, Persson B, eds. Biological effects andsafety aspects of nuclear magnetic resonance imaging andspectroscopy. Ann N Y Acad Sci 1992; 649: 204–24.

[57] von Smekal A, Wolke S, Seelos KC et al. Cardiac pacemakersin magnetic resonance studies: Influence of the static magneticfield on hardware components and software (Abstr). EurHeart J 1993; 14(Abstr Suppl): 335.

[58] Luechinger R, Duru F, Scheidegger M, Luscher TF, BoesigerP, Candinas R. Evaluation of the magnetic force and torquein a magnetic resonance imaging unit on pacemakers andimplantable cardioverter-defibrillators (Abstr). PACE 1999;22: 708.

[59] Shellock FG, Morisoli SM. Ex vivo evaluation of ferro-magnetism and artifacts of cardiac occluders exposed to a1·5-T MR system. J Magn Reson Imaging 1994; 4: 213–5.

[60] Beischer DE, Knepton JC Jr. Influence of strong magneticfields on the electrocardiogram of squirrel monkeys (Saimirisciureus). Aerospace Med 1964; 35: 939–44.

[61] Togawa T, Okai O, Oshima M. Observation of blood flowEMF in externally applied strong magnetic field by surfaceelectrodes. Med Biol Eng 1967; 5: 169–70.

[62] Gaffey CT, Tenforde TS, Dean EE. Alterations in theelectrocardiograms of baboons exposed to DC magnetic fields(Abstr). Bioelectromagnetics 1980; 1: 209.

[63] Tenforde TS, Gaffey CT, Moyer BR, Budinger TF. Cardio-vascular alterations in Macaca monkeys exposed to stationarymagnetic fields: experimental observations and theoreticalanalysis. Bioelectromagnetics 1983; 4: 1–9.

[64] Kanal E, Shellock FG, Talagala L. Safety considerations inMR imaging. Radiology 1990; 176: 593–606.

[65] Budinger TF. Emerging nuclear magnetic resonance tech-nologies: Health and safety. In: Magin RL, Liburdy RP,Persson B, eds Biological effects and safety aspects of nuclearmagnetic resonance imaging and spectroscopy. Ann NY AcadSci 1992; 649: 1–18.

[66] Budinger TF. Nuclear magnetic resonance (NMR) in vivostudies: known thresholds for health effects. J Comput AssistTomogr 1981; 5: 800–11.

[67] Shuman WP, Haynor DR, Guy AW, Wesbey GE, SchaeferDJ, Moss AA. Superficial- and deep-tissue temperatureincreases in anesthetized dogs during exposure to high specificabsorption rates in a 1·5-T MR imager. Radiology 1988; 167:551–4.


[68] Adair ER, Berglund LG. On the thermoregulatory conse-quences of NMR imaging. Magn Reson Imaging 1986; 4:321–33.

[69] Adair ER, Berglund LG. Thermoregulatory consequences ofcardiovascular impairment during NMR imaging in warm/humid environments. Magn Reson Imaging 1989; 7: 25–37.

[70] Adair ER, Berglund LG. Predicted thermophysiologicalresponses of humans to MRI fields. Ann NY Acad Sci 1992;649: 188–200.

[71] Shellock FG, Crues JV. Temperature, heart rate, and bloodpressure changes associated with clinical MR imaging at 1·5 T.Radiology 1987; 163: 259–62.

[72] Shellock FG, Schaefer DJ, Grundfest W, Crues JV. Thermaleffects of high-field (1·5 tesla) magnetic resonance imaging ofthe spine. Clinical experience above a specific absorption rateof 0.4 W/kg. Acta Radiol Suppl Stockh 1986; 369: 514–6.

[73] Shellock FG, Gordon CJ, Schaefer DJ. Thermoregulatoryresponses to clinical magnetic resonance imaging of thehead at 1·5 tesla. Lack of evidence for direct effects on thehypothalamus. Acta Radiol Suppl Stockh 1986; 369: 512–3.

[74] Shellock FG, Schaefer DJ, Gordon CJ. Effect of a 1·5 T staticmagnetic field on body temperature of man. Magn Reson Med1986; 3: 644–7.

[75] Buchli R, Boesiger P, Meier D. Heating effects of metallicimplants by MRI examinations. Magn Reson Med 1988; 7:255–61.

[76] Luechinger R, Duru F, Scheidegger M, Luscher TF, BoesigerP, Candinas R. Heating effects of magnetic resonance imagingon pacemaker leads: Effect of lead types, lead positioning andreproducibility of findings (Abstr). PACE 1999; 22: 717.

[77] Jackson JD. Classical Electrodynamics, 3rd edn., New York:Wiley & Sons, 1998.

[78] Roschmann P. Comments on ’Human magnetic resonanceimaging at 8 T’. (letter) NMR in Biomed 1999; 12: 315–7.

[79] Irnich W. Interference in pacemaker. PACE 1984; 7: 1021–48.[80] Shellock FG. Safety considerations for the extremity MR

system. In: Pocket guide to MR procedures and metallicobjects: Update 1998. Philadelphia: Lippincott-Raven, 60–3.

[81] Shellock FG, O’Neil M, Ivans V et al. Cardiac pacemakersand implantable cardioverter defibrillators are unaffectedby operation of an extremity MR imaging system. Am JRoentgenol 1999; 172: 165–70.

[82] Jerzewski A, Pattynama P, Steendijk P, Doornbos J, Roos A,Baan J. Development of an MRI-compatible catheter forpacing the heart: initial in vitro and in vivo results. J MagnReson Imag 1996; 6: 948–9.

[83] Halperin H, Jumrussirikul P, Laredo A, McVeigh E, Atalar E,Lima J, Calkins H, Berger R. Magnetic resonance guidedcatheter ablation (Abst). PACE 1998; 21: 945.

pacing in magnetic resonance imaging environment

Documents