- Split View
-
Views
-
Cite
Cite
Tobias Seckler, Dominik Stunder, Christian Schikowsky, Stephan Joosten, Matthias Daniel Zink, Thomas Kraus, Nikolaus Marx, Andreas Napp, Effect of lead position and orientation on electromagnetic interference in patients with bipolar cardiovascular implantable electronic devices, EP Europace, Volume 19, Issue 2, 1 February 2017, Pages 319–328, https://doi.org/10.1093/europace/euv458
- Share Icon Share
Electromagnetic interferences (EMIs) with cardiovascular implantable electronic devices (CIEDs) are associated with potential risk for patients. Studies imply that CIED sensitivity setting and lead's tip-to-ring spacing determine the susceptibility of CIEDs with bipolar leads to electric and magnetic fields (EMFs); however, little is known about additional decisive parameters affecting EMI of CIEDs. We therefore investigated the influence of different patient-, device-, and lead-depending variables on EMIs in 160 patients.
We ran numerical simulations with human models to determine lead-depending variables on the risk of EMI by calculating the voltage induced in bipolar leads from 50/60 Hz EMF. We then used the simulation results and analysed 26 different patient-, device-, and lead-depending variables with respect to the EMI threshold of 160 CIED patients. Our analyses revealed that a horizontal orientation and a medial position of the bipolar lead's distal end (lead-tip) are most beneficial for CIED patients to reduce the risk of EMI. In addition, the effect of CIED sensitivity setting and lead's tip-to-ring spacing was confirmed.
Our data suggest that in addition to the established influencing factors, a medial position of the lead-tip for the right ventricular lead as achievable at the interventricular septum and a horizontal orientation of the lead-tip can reduce the risk of EMI. In the right atrium, a horizontal orientation of the lead-tip should generally be striven independent of the chosen position. Still important to consider remains a good intrinsic sensing amplitude during implant procedure.
Placement of the bipolar lead's distal end (lead-tip) during the implant procedure does significantly influence the susceptibility of cardiovascular implantable electronic devices (CIEDs) to electric and magnetic 50 Hz fields.
Lead-tip's position and orientation together with CIED sensitivity setting and lead's tip-to-ring spacing are the four major parameters affecting the risk of electromagnetic interference (EMI) in bipolar leads.
A medial position of the lead-tip is beneficial to avoid EMI in the ventricular lead.
A horizontal orientation of the lead-tip makes CIEDs less susceptible to EMI in the atrial and ventricular channel.
Introduction
Cardiovascular implantable electronic devices (CIEDs) such as pacemakers (PMs) and implantable cardioverter-defibrillators (ICDs) are increasingly implanted in patients with cardiovascular diseases with over 1 million CIEDs implanted every year.1,2 Seventy-six percentage of the implantations were performed in Europe plus the USA and in these countries, 99% of the implanted pacing leads in the right ventricle (RV) and the right atrium (RA) have bipolar sensing capabilities.1 In Germany, ∼50% of new ICDs and 15% of new PM implantations are annually conducted on patients who are younger than 70 years.3 Thus, there are a high percentage of CIED patients in the working age when first receiving a CIED.
Parallel to this development, the number of electric and magnetic fields (EMFs) in our environment increases due to technological advances producing electrical devices that all emit EMFs. Especially in occupational environment, patients may be exposed to strong EMFs.4,5 For CIED patients, this poses the risk of CIED malfunction due to the fact that exogenous EMFs induce voltages in the human body and in the implanted lead(s), which may subsequently result in misinterpretation by the CIED signal processing.6 This effect is known as electromagnetic interference (EMI) and can be hazardous for CIED patients as reported in several case studies.7–9
According to a French survey, physicians are challenged by EMI; 23% of 410 questioned cardiologists indicated that they are managing CIED malfunctions due to EMI at least once a year.10 FDA's MAUDE database revealed 165 cases of CIED malfunctions attributable to EMF in the last 4 years.11 These data do not take into account those CIEDs which do not record episodes and EMI incidents that remain unreported by patients or physicians. Thus, EMI with CIEDs creates a rare but severe problem.
In clinical practice, the problem of CIEDs' susceptibility to EMF has been approached by increasingly using bipolar leads instead of unipolar leads in the last 15 years. However, CIEDs with bipolar leads remain vulnerable to strong EMFs.12,13 To date, only two parameters are known to affect EMI of bipolar CIEDs—the CIED sensitivity setting and leads' tip-to-ring spacing.4,5,14,15 The current literature, therefore, suggests noise detection algorithms, higher sensitivity values, and shorter tip-to-ring spacing for patients with foreseeable exposure to strong EMF.4,6,14,16 However, an in vivo provocation study from our group in a cohort of 110 ICD patients showed that the threshold of EMI widely differs even for patients having the same device, sensitivity setting, and tip-to-ring spacing.17 Further influencing factors have not yet been identified for bipolar leads, although the position and orientation of lead's distal end (lead-tip) are discussed as parameters which could help in prevention of EMI.18,19 However, there is a lack of experimental data, especially on in vivo investigations for confirmation.
Thus, the present study investigated, for the first time, the effect of various patient-, device-, and lead-depending parameters on the risk of EMI by using in vivo and in silico methods. While patient's physiques are usually given, device- and lead-depending parameters can be changed during the implant procedure, thus potentially providing new options for perioperative management to reduce the risk of EMI.
Methods
We initially conducted an in vivo provocation study with 160 CIED patients to evaluate the threshold of EMI of the devices (the ICD thresholds were already published17). Subsequently, we ran numerical simulations with human models to determine the influence of patient- and lead-depending parameters on the induced voltage at bipolar leads. We then used the results from the in vivo study and the simulations for regression analyses to identify parameters influencing the susceptibility of CIEDs to EMF.
In vivo provocation study
The clinical in vivo provocation study was conducted to determine the so-called interference threshold of CIEDs under worst-case conditions, i.e. the lowest field strength at which EMI occurs. Therefore, patients with ICDs or PMs were exposed to homogeneous EMFs with strengths up to 30 kV m−1 and 2550 µT at a frequency of 50 Hz.
The electric field is reproduced by a direct current injection and the electric field strength is recalculated from the measured current through the patient. The magnetic field is generated by Helmholtz coils. The patients were exposed under worst-case conditions, i.e. worst-case field direction, maximal inspiration, maximum sensitivity, and sustained pacing. The details about the test set-up and procedure are described elsewhere.17
The study design was approved by the Ethics Committee at the RWTH Aachen Faculty of Medicine and complies with the Declaration of Helsinki. The study was registered at ClinicalTrials.gov (Identifier NCT01626261).
Patient population
In the period from September 2009 to December 2014, we screened all patients presenting in the outpatient PM/ICD clinic of our department for the study. Patients who met the inclusion/exclusion criteria were invited to participate. Inclusion criteria were: age between 18 and 80 years and device implantation >4 weeks ago. Exclusion criteria were: PM dependency, hyperthyroidism, comorbidity which impedes emergency assistance (e.g. morbus bechterew and glaucoma), serum electrolyte disorders on the trial day, clinically manifest infection, acute myocardial infarction (<30 days), pregnancy, and breastfeeding.
One hundred and sixty patients (81 ICD and 79 PM) were included in the present study. All patients had bipolar leads implanted and analysable chest radiographs (i.e. the radiograph was straight recorded and all leads were visible). All patients gave written informed consent. For detailed patients' characteristics, see Supplementary material online, Table S1.
Numerical simulations
The objective of the simulations was to ascertain the influence on the risk of EMI of different lead placements, heart positions in the thorax, and tissue conductivities of the organs by numerically calculating the voltage induced in bipolar leads. We therefore used two kinds of human models: A XCAT-male model and a simplified body model (cf. Supplementary material online, Figure S1).
The XCAT-male model was developed by Simpleware Ltd using the XCAT phantom data set from Duke University.20,21 It represented an anatomically detailed normal male body taken from transverse CT, MR, and cryosection images. The simplified body model we developed consisted of 26 basic three-dimensional solids such as cubes, spheres, and cylinders. Its advantage was that model size, organ positions, and tissue properties can be freely modified.
The influence of the heart position in the thorax was scrutinized at four different positions in the simplified body model that represents individual heart positions and size.
The tissue conductivities are individual due to physiological causes and vary over time.22 We therefore assigned four different conductivity sets to the body, the blood in the heart, and the lung in the simplified body model.
The influence of the position of the bipolar lead's distal end was evaluated in the apex of the XCAT-male's RV and in seven different positions within the heart of the simplified body model. The influence of lead-tip's orientation was evaluated at each position by rotating the ring electrode around the tip electrode between 0 and 180° in 10° steps in all three anatomical planes.
The induced voltage was determined with respect to the lead-tip's orientation for all lead-tip's positions, heart positions, and tissue conductivities—in total 113 separate model configurations. Detailed information about the numerical simulations can be found in Supplementary material online, Methods.
Statistics
Regression analyses were performed to identify parameters which affect the susceptibility of CIEDs to EMF using potentially influencing patient-, device-, and lead-depending variables as independent variables and the interference threshold of CIEDs determined in the in vivo provocation study as a dependent variable.
Variables potentially influencing risk of electromagnetic interference
We defined 26 variables that potentially influence the susceptibility of CIEDs to 50/60 Hz EMF. They were derived using the results from the in vivo provocation study, current literature, and the numerical simulations, and can be categorized in patient-, device-, and lead-depending variables. The variables and their references are summarized in Table 1.
Category . | Variables . | References . |
---|---|---|
Patient | Sex, weight, height, circumference of the thorax, abdomen, hip, and upper arm, distance between the shoulders, diameter of heart, and thorax | Joosten et al.23 found the influence of body measurements for unipolar PMs. Therefore, we decided to scrutinize these also for bipolar PMs and ICDs. |
Maximal inspiration (yes/no) | Napp et al.17 and Joosten et al.23 | |
Device | Manufacturer, time period since market release | Napp et al.17 |
Sensitivity setting | Toivonen et al.16 and Scholten and Silny14 | |
Insertion site | Irnich6 found the influence of insertion sites for unipolar CIEDs. | |
Type, operating mode (pacing or sensing) | Device manuals explain that some types change their sensitivity threshold in the event of pacing. | |
Lead | Manufacturer, fixation type | Hille et al.18 published the suspicion of the influence of the lead design on EMI. |
Tip-to-ring spacing | Irnich6 | |
Distance from lead-tip to vertebra Th1 and lateral rib cage, distance between tip and ring along the transverse (medio-lateral) axis, longitudinal (cranio-caudal) axis, and the sagittal (antero-posterior) axis, polar angle (θ) | Those variables were the results of the numerical simulations. |
Category . | Variables . | References . |
---|---|---|
Patient | Sex, weight, height, circumference of the thorax, abdomen, hip, and upper arm, distance between the shoulders, diameter of heart, and thorax | Joosten et al.23 found the influence of body measurements for unipolar PMs. Therefore, we decided to scrutinize these also for bipolar PMs and ICDs. |
Maximal inspiration (yes/no) | Napp et al.17 and Joosten et al.23 | |
Device | Manufacturer, time period since market release | Napp et al.17 |
Sensitivity setting | Toivonen et al.16 and Scholten and Silny14 | |
Insertion site | Irnich6 found the influence of insertion sites for unipolar CIEDs. | |
Type, operating mode (pacing or sensing) | Device manuals explain that some types change their sensitivity threshold in the event of pacing. | |
Lead | Manufacturer, fixation type | Hille et al.18 published the suspicion of the influence of the lead design on EMI. |
Tip-to-ring spacing | Irnich6 | |
Distance from lead-tip to vertebra Th1 and lateral rib cage, distance between tip and ring along the transverse (medio-lateral) axis, longitudinal (cranio-caudal) axis, and the sagittal (antero-posterior) axis, polar angle (θ) | Those variables were the results of the numerical simulations. |
Category . | Variables . | References . |
---|---|---|
Patient | Sex, weight, height, circumference of the thorax, abdomen, hip, and upper arm, distance between the shoulders, diameter of heart, and thorax | Joosten et al.23 found the influence of body measurements for unipolar PMs. Therefore, we decided to scrutinize these also for bipolar PMs and ICDs. |
Maximal inspiration (yes/no) | Napp et al.17 and Joosten et al.23 | |
Device | Manufacturer, time period since market release | Napp et al.17 |
Sensitivity setting | Toivonen et al.16 and Scholten and Silny14 | |
Insertion site | Irnich6 found the influence of insertion sites for unipolar CIEDs. | |
Type, operating mode (pacing or sensing) | Device manuals explain that some types change their sensitivity threshold in the event of pacing. | |
Lead | Manufacturer, fixation type | Hille et al.18 published the suspicion of the influence of the lead design on EMI. |
Tip-to-ring spacing | Irnich6 | |
Distance from lead-tip to vertebra Th1 and lateral rib cage, distance between tip and ring along the transverse (medio-lateral) axis, longitudinal (cranio-caudal) axis, and the sagittal (antero-posterior) axis, polar angle (θ) | Those variables were the results of the numerical simulations. |
Category . | Variables . | References . |
---|---|---|
Patient | Sex, weight, height, circumference of the thorax, abdomen, hip, and upper arm, distance between the shoulders, diameter of heart, and thorax | Joosten et al.23 found the influence of body measurements for unipolar PMs. Therefore, we decided to scrutinize these also for bipolar PMs and ICDs. |
Maximal inspiration (yes/no) | Napp et al.17 and Joosten et al.23 | |
Device | Manufacturer, time period since market release | Napp et al.17 |
Sensitivity setting | Toivonen et al.16 and Scholten and Silny14 | |
Insertion site | Irnich6 found the influence of insertion sites for unipolar CIEDs. | |
Type, operating mode (pacing or sensing) | Device manuals explain that some types change their sensitivity threshold in the event of pacing. | |
Lead | Manufacturer, fixation type | Hille et al.18 published the suspicion of the influence of the lead design on EMI. |
Tip-to-ring spacing | Irnich6 | |
Distance from lead-tip to vertebra Th1 and lateral rib cage, distance between tip and ring along the transverse (medio-lateral) axis, longitudinal (cranio-caudal) axis, and the sagittal (antero-posterior) axis, polar angle (θ) | Those variables were the results of the numerical simulations. |
Three variables represented conditions of each interference threshold in the in vivo study: sensitivity setting and operating mode (pacing/sensing) as well as the status of inspiration (maximal inspiration yes/no) of the patient.
Patient-depending variables were gender, body weight, and height, which were measured in light clothing with shoes. The circumferences of the thorax, the abdomen, the hip, and the upper arm as well as the distance between the shoulders were taken.
Devices' manufacturer and type (PM or ICD), the insertion site, and the lead manufacturer were taken from the CIED identity card. The lead's tip-to-ring spacing and active/passive fixation type were noted from the technical manual. The date of market release was recorded as time period related to December 2014.
Regression analysis
For regression analyses, the Cox and Tobit models24–26 were applied because patients' devices with interference thresholds above the tested limits could thus be included in the analysis and were considered as censored observations.
Cox regression was used for electric field interference thresholds as they are censored individually, i.e. the exposed maximum of the electric field differed between patients due to calculation from the injected current. Tobit regression was used for magnetic field interference thresholds as they are fixed censored, i.e. all participants were exposed up to a maximum magnetic field of 2550 µT.17
For a more precise regression analysis, the device type was used as a stratification factor in order to consider the system differences in sensing algorithm between PMs and ICDs.
All data were analysed using the R statistical software (Version 3.1.0, www.r-project.org). The analysis was separately done for the atrial and ventricular lead, each with its interference thresholds for EMF. The models were fit through a stepwise selection algorithm, which combines aspects of forward and backward selection. The criterion used for model selection was the Bayesian information criterion (BIC) and the selection process stops when BIC does not further decrease, meaning that the remaining (unselected) variables do not contribute statistically significant information to the model. The values for P and pseudo-R2 were taken from the likelihood ratio test and maximum-likelihood estimation, respectively.
Results
Influence of the lead-tip's orientation on the risk of electromagnetic interference (numerical simulations)
Field type . | Plane . | Simplified body model . | XCAT-male . | ||
---|---|---|---|---|---|
θVMax (°) . | θVMin (°) . | θVMax (°) . | θVMin (°) . | ||
Electric | Sagittal | 90.36 ± 1.86 | 0.09 ± 0.94 | 92 | 0 |
Frontal | 91.61 ± 3.67 | 1.61 ± 3.93 | 99 | 9 | |
Magnetic | Sagittal | 90.71 ± 11.29 | 0.71 ± 2.91 | 92 | 0 |
Frontal | 105.18 ± 19.41 | 18.13 ± 15.74 | 100 | 10 |
Field type . | Plane . | Simplified body model . | XCAT-male . | ||
---|---|---|---|---|---|
θVMax (°) . | θVMin (°) . | θVMax (°) . | θVMin (°) . | ||
Electric | Sagittal | 90.36 ± 1.86 | 0.09 ± 0.94 | 92 | 0 |
Frontal | 91.61 ± 3.67 | 1.61 ± 3.93 | 99 | 9 | |
Magnetic | Sagittal | 90.71 ± 11.29 | 0.71 ± 2.91 | 92 | 0 |
Frontal | 105.18 ± 19.41 | 18.13 ± 15.74 | 100 | 10 |
The values for the simplified body model are averaged over all lead-tips' positions, heart positions, and tissue conductivity sets.
Max, maximum; Min, minimum; V, volt.
Field type . | Plane . | Simplified body model . | XCAT-male . | ||
---|---|---|---|---|---|
θVMax (°) . | θVMin (°) . | θVMax (°) . | θVMin (°) . | ||
Electric | Sagittal | 90.36 ± 1.86 | 0.09 ± 0.94 | 92 | 0 |
Frontal | 91.61 ± 3.67 | 1.61 ± 3.93 | 99 | 9 | |
Magnetic | Sagittal | 90.71 ± 11.29 | 0.71 ± 2.91 | 92 | 0 |
Frontal | 105.18 ± 19.41 | 18.13 ± 15.74 | 100 | 10 |
Field type . | Plane . | Simplified body model . | XCAT-male . | ||
---|---|---|---|---|---|
θVMax (°) . | θVMin (°) . | θVMax (°) . | θVMin (°) . | ||
Electric | Sagittal | 90.36 ± 1.86 | 0.09 ± 0.94 | 92 | 0 |
Frontal | 91.61 ± 3.67 | 1.61 ± 3.93 | 99 | 9 | |
Magnetic | Sagittal | 90.71 ± 11.29 | 0.71 ± 2.91 | 92 | 0 |
Frontal | 105.18 ± 19.41 | 18.13 ± 15.74 | 100 | 10 |
The values for the simplified body model are averaged over all lead-tips' positions, heart positions, and tissue conductivity sets.
Max, maximum; Min, minimum; V, volt.
Influence of the lead-tip's position on the risk of electromagnetic interference (numerical simulations)
The most medial of the seven scrutinized lead-tip positions within the simplified body model's heart revealed the lowest induced voltage. From this position to the most lateral lead-tip position, which was 6 cm apart along the transversal axis, the induced voltage from magnetic fields increased by 185% (cf. P2 and P3 in Supplementary material online, Table S6). Furthermore, between the most anterior and the most posterior lead-tip position, which was 6 cm apart along the sagittal axis, the induced voltage differed only by 2.7% (cf. P6 and P7 in Supplementary material online, Table S6). Therefore, in magnetic fields, the lead-tip's position along the transversal axis is a decisive parameter concerning the risk of EMI, whereas the position along the sagittal axis is negligible. Thus, a medial position of the lead-tip minimizes the risk of EMI in magnetic fields. In electric fields, different lead-tips' positions within the heart had negligible influence on the induced voltage (maximum difference 7%). Thus, the numerical simulation showed no influence of the lead-tip's position for electric fields.
A detailed compilation of the results for all scrutinized tissue conductivities, heart positions, and lead-tips' positions is provided in Supplementary material online, Results.
Parameters affecting electromagnetic interference with the ventricular lead (in vivo results)
For the ventricular lead, 213 observations, respectively interference thresholds, were evaluated. The results are presented in Table 3 for the electric and magnetic field. The influencing variables of interference with the ventricular lead can be summarized to device sensitivity setting, lead-tip's position, and orientation.
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.673a | 0.456 to 0.995 | 0.047 | 0.10 (n = 108) |
Magnetic field | |||||
1. Operation mode | Sensing to pacing | −1673b | −2310 to −1035 | <0.001 | 0.34 (n = 105) |
2. Distance lead-tip to lateral rib cage | 1 mm | +19.9b | +2.14 to +37.6 | 0.028 | |
3. Sensitivity setting | 0.1 mV | +240b | +101 to +380 | <0.001 | |
4. Transverse tip-to-ring distance | 1 mm | +159b | +53.1 to +265 | 0.003 |
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.673a | 0.456 to 0.995 | 0.047 | 0.10 (n = 108) |
Magnetic field | |||||
1. Operation mode | Sensing to pacing | −1673b | −2310 to −1035 | <0.001 | 0.34 (n = 105) |
2. Distance lead-tip to lateral rib cage | 1 mm | +19.9b | +2.14 to +37.6 | 0.028 | |
3. Sensitivity setting | 0.1 mV | +240b | +101 to +380 | <0.001 | |
4. Transverse tip-to-ring distance | 1 mm | +159b | +53.1 to +265 | 0.003 |
CI, confidence interval.
aHR from Cox regression.
bRegression coefficient in µT from Tobit regression.
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.673a | 0.456 to 0.995 | 0.047 | 0.10 (n = 108) |
Magnetic field | |||||
1. Operation mode | Sensing to pacing | −1673b | −2310 to −1035 | <0.001 | 0.34 (n = 105) |
2. Distance lead-tip to lateral rib cage | 1 mm | +19.9b | +2.14 to +37.6 | 0.028 | |
3. Sensitivity setting | 0.1 mV | +240b | +101 to +380 | <0.001 | |
4. Transverse tip-to-ring distance | 1 mm | +159b | +53.1 to +265 | 0.003 |
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.673a | 0.456 to 0.995 | 0.047 | 0.10 (n = 108) |
Magnetic field | |||||
1. Operation mode | Sensing to pacing | −1673b | −2310 to −1035 | <0.001 | 0.34 (n = 105) |
2. Distance lead-tip to lateral rib cage | 1 mm | +19.9b | +2.14 to +37.6 | 0.028 | |
3. Sensitivity setting | 0.1 mV | +240b | +101 to +380 | <0.001 | |
4. Transverse tip-to-ring distance | 1 mm | +159b | +53.1 to +265 | 0.003 |
CI, confidence interval.
aHR from Cox regression.
bRegression coefficient in µT from Tobit regression.
For the analysis of the electric field, the interference thresholds of 91% of the patients had to be considered as censored observations because they could not be disturbed in the in vivo study. The model fit by Cox regression reached therefore 0.1 for pseudo-R2 and only the sensitivity setting was determined as an influencing variable. The sensitivity setting reduced the risk of interference by 32.7% for every increase of 0.1 mV as the hazard ratio (HR) of 0.673 shows.
In the magnetic field, the Tobit regression model seemed to show a good fit with a pseudo-R² of 0.344 because pseudo-R² cannot reach one, even for a perfect model fit.27 The influencing variables for the ventricular lead contributed in the following priority order: the operation mode, the distance from lead-tip to lateral rib cage, the sensitivity setting, and the transverse tip-to-ring distance. If the CIED acted in the pacing mode, the interference threshold dropped by 1673 µT. The higher the distance was in transverse direction from lead-tips to lateral rib cage, in other words, the more medial the lead-tip was positioned, the lower the risk of interference was—interference threshold increased by 20 µT for each mm. If the sensitivity was set 0.1 mV higher, the interference threshold increased by 240 µT. The transverse tip-to-ring distance influenced the susceptibility by an increase in the interference threshold by 159 µT with every additional mm, suggesting that a more horizontal orientation of the lead-tip reduces the risk of EMI.
Parameters affecting electromagnetic interference with the atrial lead (in vivo results)
For the atrial lead, 238 observations, respectively interference thresholds, were evaluated. The results are presented in Table 4 for the electric and magnetic field. The two most influencing variables for the susceptibility to 50 Hz fields were the device sensitivity setting and the lead's polar angle (θ).
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.527a | 0.412 to 0.675 | <0.001 | 0.49 (n = 120) |
2. Polar angle (θ) | 1° | 1.026a | 1.007 to 1.046 | 0.008 | |
3. Maximal inspiration | No to yes | 2.449a | 1.344 to 4.465 | 0.004 | |
4. Tip-to-ring spacing | 1 mm | 1.335a | 1.164 to 1.532 | <0.001 | |
5. Time period since market release | 1 year | 0.883a | 0.799 to 0.976 | 0.015 | |
Magnetic field | |||||
1. Sensitivity setting | 0.1 mV | +274b | +193 to +354 | <0.001 | 0.35 (n = 118) |
2. Polar angle (θ) | 1° | −32.7b | −47.7 to −17.7 | <0.001 | |
3. Longitudinal tip-to-ring distance | 1 mm | −70.2b | −123 to −16.6 | 0.01 |
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.527a | 0.412 to 0.675 | <0.001 | 0.49 (n = 120) |
2. Polar angle (θ) | 1° | 1.026a | 1.007 to 1.046 | 0.008 | |
3. Maximal inspiration | No to yes | 2.449a | 1.344 to 4.465 | 0.004 | |
4. Tip-to-ring spacing | 1 mm | 1.335a | 1.164 to 1.532 | <0.001 | |
5. Time period since market release | 1 year | 0.883a | 0.799 to 0.976 | 0.015 | |
Magnetic field | |||||
1. Sensitivity setting | 0.1 mV | +274b | +193 to +354 | <0.001 | 0.35 (n = 118) |
2. Polar angle (θ) | 1° | −32.7b | −47.7 to −17.7 | <0.001 | |
3. Longitudinal tip-to-ring distance | 1 mm | −70.2b | −123 to −16.6 | 0.01 |
CI, confidence interval.
aHR from Cox regression.
bRegression coefficient in µT from Tobit regression.
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.527a | 0.412 to 0.675 | <0.001 | 0.49 (n = 120) |
2. Polar angle (θ) | 1° | 1.026a | 1.007 to 1.046 | 0.008 | |
3. Maximal inspiration | No to yes | 2.449a | 1.344 to 4.465 | 0.004 | |
4. Tip-to-ring spacing | 1 mm | 1.335a | 1.164 to 1.532 | <0.001 | |
5. Time period since market release | 1 year | 0.883a | 0.799 to 0.976 | 0.015 | |
Magnetic field | |||||
1. Sensitivity setting | 0.1 mV | +274b | +193 to +354 | <0.001 | 0.35 (n = 118) |
2. Polar angle (θ) | 1° | −32.7b | −47.7 to −17.7 | <0.001 | |
3. Longitudinal tip-to-ring distance | 1 mm | −70.2b | −123 to −16.6 | 0.01 |
Variable . | Step . | Estimate(s) . | 95% CI . | P-value . | Pseudo-R2 . |
---|---|---|---|---|---|
Electric field | |||||
1. Sensitivity setting | 0.1 mV | 0.527a | 0.412 to 0.675 | <0.001 | 0.49 (n = 120) |
2. Polar angle (θ) | 1° | 1.026a | 1.007 to 1.046 | 0.008 | |
3. Maximal inspiration | No to yes | 2.449a | 1.344 to 4.465 | 0.004 | |
4. Tip-to-ring spacing | 1 mm | 1.335a | 1.164 to 1.532 | <0.001 | |
5. Time period since market release | 1 year | 0.883a | 0.799 to 0.976 | 0.015 | |
Magnetic field | |||||
1. Sensitivity setting | 0.1 mV | +274b | +193 to +354 | <0.001 | 0.35 (n = 118) |
2. Polar angle (θ) | 1° | −32.7b | −47.7 to −17.7 | <0.001 | |
3. Longitudinal tip-to-ring distance | 1 mm | −70.2b | −123 to −16.6 | 0.01 |
CI, confidence interval.
aHR from Cox regression.
bRegression coefficient in µT from Tobit regression.
When exposed to electric fields, increasing the sensitivity setting by 0.1 mV reduced the risk of interference by 47.3% (HR 0.527). With every additional degree of the lead's polar angle (θ) towards a more vertical position, the risk of interference rose by 2.6% (HR 1.026). The third, fourth, and fifth variables affecting the interference threshold in electric fields were the status of inspiration (at maximal inspiration the susceptibility rose by 144.9%; HR 2.449), tip-to-ring spacing (increase of 1 mm rose the susceptibility by 33.5%; HR 1.335), and time period since market release (per additional year, the susceptibility was reduced by 11.7%; HR 0.883).
For magnetic fields, three variables were added: sensitivity setting, polar angle (θ), and longitudinal tip-to-ring distances. The interference threshold was increased by 274 µT if the sensitivity was set 0.1 mV higher. An increase in the leads polar angle (θ) by 1° towards a more vertical position reduced the interference threshold by 33 µT. The longitudinal tip-to-ring distance, as a third variable, decreased the interference threshold by 70 µT for each additional mm, indicating that a more horizontal orientation of the lead-tip minimizes the risk of EMI.
Overall, the determined variables seem to be good predictors for EMI with the atrial lead reaching a pseudo-R2 of 0.49 for the electric field (Cox regression) and a pseudo-R2 of 0.35 for the magnetic field (Tobit regression).27
Discussion
The present study demonstrates that lead-tip's position and orientation are decisive parameters regarding the susceptibility of bipolar CIEDs to exogenous EMF. The main goals of standard CIED implant procedures remain high local sensing amplitude, good pacing threshold, and low risk of dislodgement or perforation. For patients with foreseeable exposure to strong EMF, however, the following two lead placement objectives should be considered to prevent patients from EMI:
more medial position of the lead-tip (only RV lead, the RA lead-tip already has an anatomically inherent medial position).
more horizontal orientation of the lead-tip (RA and RV lead).
Ventricular lead placement
Since the introduction of transvenous cardiac pacing, the right ventricular apex (RVA) has been the preferred site for ventricular lead implantation.28,29 Alternative lead positions are the interventricular septum (IVS) and the right ventricular outflow tract (RVOT). For high RV septal positions (e.g. RVOT), an active lead fixation is needed for stable anchoring. Due to the RV anatomy, every position determines the lead-tip's orientation. Our data extend the knowledge on lead placement in the RV by suggesting that a medial position of the lead-tip may reduce the risk of EMI for RV leads. Such a medial position can be achieved by choosing placement at the RV septum instead of the RVA. In most of the cases, lead placement at RVOT allows an even more medial position. In addition, we have shown that a more vertical orientation of the lead-tip increases the risk of EMI, thus being potentially harmful for CIED patients. Considering the impact of the EMI affecting variables for RV lead placement (cf. Table 3), the orientation appears less important (fourth variable) than the lead-tip's position (second variable). However, a horizontal orientation of the lead-tip should be taken into account at every position of the lead-tip (RVA, IVS, and RVOT) as far as possible. Especially at the RVOT position, a horizontal orientation is difficult to achieve due to the RVOT anatomy.
We therefore conclude that IVS positions (as medial as possible) and an orientation of the lead-tip as horizontal as possible are most beneficial for CIED patients to reduce the risk of EMI (cf. Figure 1 for an example of IVS lead position). If IVS is not accessible, RVOT should be chosen over RVA. If an apical position remains as the only acceptable pacing site, special emphasis should be taken on a high intrinsic EGM signal. Achieving a good sensing amplitude offers the opportunity for programming higher sensitivity values, thus leading to less susceptibility to EMF. In the case of ICDs, defibrillator testing should be considered. Ensuring a proper sensing of fibrillation waves with a high sensitivity value provides options for reprogramming the sensitivity settings in the case of EMI.
Atrial lead placement
Independent of the chosen position achieving a horizontal orientation should be the general aim. However, a horizontal orientation of the lead-tip should not be chosen in abandonment of a good intrinsic sensing amplitude or risk of lead dislodgement. Particularly, the atrial sensing amplitude is of importance with regard to the risk of EMI due to the corresponding signal-to-noise ratio. Small intrinsic atrial signals cause settings of low sensitivity values that make the CIED more susceptible to EMI (cf. Tables 3 and 4). Due to smaller interference thresholds in the atrial lead, a stable anchoring, good measuring values and a horizontal orientation of the lead-tip should be the main goal for every patient.
Study consistency in the context of literature
The regression analyses of our in vivo data identified the following variables affecting the risk of EMI: lead-tip's position and orientation, the sensitivity settings, the operation mode, the tip-to-ring spacing, the time period since market release, and the status of inspiration (cf. Tables 3 and 4).
The quality of the regression analysis is specified by pseudo-R2 values. Considering the achieved values (0.34–0.49) for the RA, the quality of the regression analysis is good to very good according to Menard.27 The regression analysis for the RV in electric fields achieved a pseudo-R2 of 0.1, which is why we consider this regression analysis as sufficient but incomplete. The lower value of the pseudo-R2 is due to the small number of patients (9%) in which EMI could be provoked in the RV channel in electric fields. Whereas in the RA channel in magnetic/electric fields in 53%/43% of the patients, EMI could be provoked, subsequently increasing the number of significant variables of the regression model and the pseudo-R2.
First assumptions about the influence of the lead-tip's position and orientation on EMI were published 2009 in a conference proceeding by Hille et al.18 However, Hille et al. used a cuboid phantom model, filled with saline solution, and numerical simulations of that cuboid phantom to scrutinize five positions of the lead-tip under magnetic field exposure. The orientation of the lead-tip was only changed in the frontal plane and only one lead angle was recorded for each position. Nevertheless, this first in vitro/in silico approximation also revealed that, in magnetic fields, a more medial position of lead-tip reduces the risk of EMI. In our study, the influence of the lead-tip's position and orientation on EMI was found out by using two computational body models. This was confirmed by the regression analyses of the in vivo data, which revealed the lead-tip's position as the second and the lead-tip's orientation (transverse tip-to-ring distance) as the fourth influencing variable for RV leads in magnetic fields as Table 3 shows. The RA lead-tip already has an anatomically inherent medial position, explaining the fact that no influence of the RA lead-tip's position was found. Therefore, in return, the lead-tip's orientation (polar angle) in the RA was found to be of critical importance (second variable in magnetic and electric fields, Table 4).
The major influence of device sensitivity settings on the risk of EMI has been well described by Toivonen et al.16 (in vivo) and Scholten and Silny14 (in vitro). In three of four regression analyses performed in our study, sensitivity setting turned out to be the first significant variable. Only for the ventricular lead in magnetic fields, sensitivity setting was revealed as a third variable. However, the first variable in this case was operating mode of the device, which is strongly connected to the device sensitivity setting considering the lower starting values during pacing.5 Napp et al.5 showed that different sensitivity progressions after sensing and pacing occur.
Regarding the influence of lead's tip-to-ring spacing, Irnich15 published, in 2002, an analytical approach suggesting a longer spacing to increase the risk of EMI. Our in vivo data confirmed this finding by revealing the tip-to-ring spacing as the fourth variable for atrial leads in electric fields and indirectly in terms of the horizontal/vertical component of the tip-to-ring spacing in magnetic fields (fourth variable in RV and third variable in RA).
In eight ICD patients, integrated bipolar leads from Boston Scientific were implanted in the RV. The integrated bipolar lead's ring electrode is also the defibrillation electrode. The missing dedicated ring electrode results in a longer tip-to-ring spacing of integrated bipolar leads than the average tip-to-ring spacing of other ICD leads—in our case, 12 mm at Boston Scientific ICD leads vs. 10 mm for the remaining ICD leads. The regression analyses, however, revealed no significant influence of any particular lead manufacturer. The lead model was not included in statistics (Table 1) due to the high diversity of lead models, which is why a direct statement regarding the integrated bipolar leads is not possible. However, the percentage of integrated bipolar leads in which EMI could be provoked was 12.5%, compared with 11.9% in the remaining ICD patients with true bipolar leads. This is most likely due to the revealed influence of the tip-to-ring spacing. Therefore, we assume that using the defibrillation electrode as a ring electrode is not of importance for the risk of EMI.
Given novel developments in device technology, the time period since market release also seemed to have an influence on EMI (Table 4). Further investigations indicated, however, that this result is probably related to the lower programmable maximum sensitivity of modern devices. Thus, direct comparison of device generations is not feasible to estimate the quality of EMF noise suppression.
Joosten et al.23 found in 2009 that status of inspiration, physique, and insertion site are affecting the risk of EMI in electric fields for unipolar PMs. The importance of inspiration status for EMI risk is consistent with the regression analysis for electric fields presented in our study, revealing the maximal inspiration as a third variable for the RA channel. For the RV channel, no influence was found. Possibly due to a lag of observations, this variable did not reached statistical significance in the regression model. Further research on this particular topic is needed. Regarding the influence of physique and insertion site, our regression analysis did not reveal a significant influence to the risk of EMI with bipolar CIEDs. The insertion site, which is one of the most influencing variables for unipolar CIEDs, was also considered insignificant for bipolar CIEDs in Irnich's prior discussed analytical approach.15 The influence of the physique on the risk of EMI for bipolar CIEDs has to be further investigated in subsequent research.
Our study applied two different regression models (Cox and Tobit) each with two in vivo data sets (RA and RV), which sum up to four independent regression analyses (cf. Tables 3 and 4). In addition, two independent computational body models (XCAT-male and simplified body model) were scrutinized in a total of 113 different configurations in magnetic and electric fields. The results of our study not only provide systematical in vivo proof of prior published influencing factors,14,15,18,23 but also extend our current knowledge by elaborating the influence of the lead placement on the risk of EMI and by providing novel information how patients could benefit from this.
Limitations
The orientation of the lead is changing within one heart cycle. However, the chest radiograph is just a snapshot which is not synchronized to the heart beat. The angle information taken from the radiograph data may have uncertainties, but given the cohort size (n = 160) we believe that this error is compensated.
The interference thresholds collected with the provocation study are found for frontal magnetic field exposure and vertical electric field exposure, which is the worst-case exposure. Other exposure set-ups could lead to different influencing variables, but will also lead to a reduction in the risk of EMI. Therefore, it is assumed that the variables found for the worst-case exposure also cover other exposure set-ups.
The EMFs generated during the provocation study are operating at powerline frequency (50 Hz). The resulting interference thresholds can only be scaled to certain frequencies (16–1000 Hz). Indeed, 50/60 Hz fields are very common, given the fact that those frequencies are used for power lines.
Conclusions
Our data show that orientation and position of the lead-tip influences the susceptibility to EMF in CIED patients with a potential beneficial effect of more horizontal orientation and a medial position of the atrial and ventricular lead-tip. In case of foreseeable strong EMF exposition, this fact has to be taken into account and during the implant procedure, special attention should be paid to the orientation of the lead-tip. Still, as recommended, highest possible intrinsic signals should be obtained at the implant procedure and under these conditions, a high value for the sensitivity should be programmed as clinically reasonable.
Supplementary material
Supplementary material is available at Europace online.
Funding
This work was supported by the German Social Accident Insurance Institution for the Energy, Textile, Electrical, and Media Products Sectors (BG ETEM); the research unit for electropathology (FFE); and a research grant from the B. Braun Foundation, Melsungen, Germany.
Conflict of interest: A.N. and M.D.Z. received travel grants by Biotronik, Boston Scientific, Medtronic, and St Jude Medical.
Acknowledgements
The authors thank the volunteers who participated in this study.
References
Author notes
The first two authors contributed equally to the work.