Cardiovascular Collapse Caused by Electrocardiographically Silent 60-Hz Intracardiac Leakage Current
Implications for Electrical Safety
Background—The national standard for safe 60-Hz intracardiac leakage current under a single-fault condition is 50 μA. This standard is intended to protect patients from alternating current (AC) at levels below the threshold for sensation, but the minimum unsafe level for AC in closed-chest humans is not known. To determine this value, we studied 40 patients at testing of implantable cardioverter-defibrillators using a programmable source of 60-Hz AC.
Methods and Results—We applied AC for 5-second test periods in increasing strengths until ventricular fibrillation (VF) was induced or 1 mA was reached. Two current paths were tested: bipolar, between tip and ring electrodes of a right ventricular pacing catheter, and unipolar, from tip to a remote electrode. We observed a characteristic sequence of 3 responses as AC was increased: (1) intermittent ventricular capture with QRS morphology identical to pacing through the electrodes (minimum value, 20 μA); (2) continuous capture at cycle length 282±88 ms (minimum value, 32 μA); and (3) VF persisting after AC termination (minimum value, 49 μA). Continuous capture caused loss of pulsatile arterial pressure and cardiovascular collapse (mean arterial pressure, 32±8 mm Hg) for the duration of AC with no ECG evidence of AC stimulation. Thus, the clinical picture was that of hypotensive ventricular tachycardia (VT). The continuous-capture threshold was ≤50 μA in 9 patients (22%) for bipolar AC and in 5 (12%) for unipolar AC. All patients showed continuous capture over a wide range for both bipolar AC (68±18 to 216±238 μA) and unipolar AC (84±27 to 278±226 μA).
Conclusions—Leakage current causes cardiovascular collapse at levels below the VF threshold. Stimulation by silent AC that is neither felt nor visible on the ECG presents as hypotensive VT. In patients with intracardiac electrodes, leakage current less than or equal to the present standard of 50 μA may cause VT or VF. The safety standard for leakage current lasting ≥5 seconds should be ≤20 μA. This standard should be based on the continuous-capture threshold.
Power-line–operated electromedical equipment, connected to patients for monitoring or therapeutic purposes, may permit accidental flow (leakage) of weak alternating current (AC) through patients to ground. An intracardiac catheter may provide a low-resistance path to ground through the patient’s heart and thereby place the patient at risk for electrically induced ventricular tachycardia (VT) or ventricular fibrillation (VF).
In 1993, the American national standard for leakage current through the heart under a single-fault condition in mains-operated electrical equipment was increased from 10 to 50 μA,1 the value of the European standard since 1988.2 Both the 10-μA standard3 4 and the 50-μA standard2 were based on estimates of the risk of AC-induced VF. However, AC may cause cardiovascular collapse at levels that are below the VF threshold.5 6 7 8 9 This adverse response to AC was not considered in the selection of either safety standard. Furthermore, safe levels of AC have not been determined in closed-chest humans.
The 10-μA standard was adopted in 1967 to ensure patient safety during cardiac catheterization10 and pacemaker11 procedures. The annual number of invasive cardiac procedures in the United States has increased from <60 000 when the 10-μA standard was adopted to >3 million today. The potential number of adverse outcomes from leakage current has increased correspondingly.
Electromedical devices contain electrical isolation circuits and insulation to limit leakage current. Manufacturers continue to comply with the original 10-μA standard, but they may realize substantial cost savings by equipment designs that comply only with the newer 50-μA standard.12 However, the American Heart Association continues to recommend the 10-μA standard.13 14 15 We therefore determined the minimum unsafe levels for AC in closed-chest humans.
We studied 40 patients under propofol anesthesia at postoperative testing of transvenous implantable defibrillators. All patients gave written informed consent according to a protocol approved by the Human Subjects Committee of Cedars–Sinai Medical Center. Table 1⇓ shows clinical characteristics of the study patients. At the time of study, all antiarrhythmic drugs had been discontinued for 5 half-lives, except for digoxin (3 patients) and β-blockers (7 patients). The 3 patients in whom atrial fibrillation was the indication for a dual-chamber implantable defibrillator had neither spontaneous nor inducible VT or VF.
Radial arterial pressure was recorded continuously. We gave AC through a standard 6F temporary pacing catheter with a distal electrode area of 15 mm2 and ring-electrode area of 7 mm2 (Explorer, Boston Scientific). It was inserted via the right femoral vein and positioned with its tip near the right ventricular apex. AC was delivered from a battery-operated, custom-built, constant-current source (Medtronic model 2917 60-Hz fibrillator, IDE No. G970011) that permitted us to program the duration and strength of AC. We measured the root-mean-square value of delivered AC using a digital ammeter (model 87 True RMS Multimeter, Fluke Inc). Fixed-duration AC was applied in increasing strengths until sustained VT or VF was induced or the maximum output of 999 μA was reached. Sustained VT was defined as VT requiring termination by cardioversion or pacing. Sustained VF was defined as VF requiring defibrillation. We recorded the 12-lead ECG and right ventricular electrogram on optical disk using a multichannel, electrophysiology data–acquisition system (Cardiolab 3.1 or 4.0, Prucka, Inc). We also recorded single-channel ECGs on a monitor designed for intraoperative or intensive care unit use (model M1094A or M1094B, Hewlett Packard) and on the monitor of an external defibrillator (model PD1200, Zoll Medical). The latter 2 monitors meet American Heart Association guidelines for 60-Hz notch filters4 on ECG equipment; the former does not.
The initial programmed strength of AC was 40 μA. We increased it by 10-μA increments up to 200 μA, by 50 μA between 250 and 600 μA, and by 100 μA thereafter. A characteristic sequence of 3 responses occurred as AC was increased (Figure 1⇓): (1) intermittent ventricular capture with QRS morphology identical to pacing through the electrodes; (2) continuous ventricular capture with the same QRS morphology; and (3) initial continuous capture progressing to sustained VT or VF that persisted after termination of AC. We defined continuous capture as a consecutive series of stimulated QRS complexes without intervening spontaneous QRS complexes for the duration of AC.
AC was increased if intermittent capture was not stimulated by a level of 40 μA. If intermittent capture was stimulated, this threshold procedure was performed: AC was decreased by 2 levels and was applied up to 5 times at this and the next-higher level unless intermittent capture was induced. This process was iterated until 5 consecutive AC applications at a given level did not result in intermittent capture. The weakest AC strength that stimulated any QRS complexes was defined as the intermittent capture threshold. We then repeated the threshold process until we determined the continuous-capture threshold and the VF threshold or until we reached the maximum output of the fibrillator. If AC induced sustained VT, the sustained-VT threshold was accepted as an end point, and the VF threshold was not determined. We defined the sustained-VT/VF threshold as the weakest current that induced either sustained VT or VF. No threshold procedure required >1 decrease (step down) in AC. There was a 15-second rest period after each application of AC that produced continuous capture and a 4-minute rest after each episode of sustained VT or VF.
In all patients, we applied AC for 5-second test intervals over 2 current paths in random order: bipolar, between tip and ring electrodes; and unipolar, from tip to an ECG patch electrode positioned on the skin near the left or right shoulder. In the first 20 patients, we tested reproducibility of the continuous-capture and VF thresholds. Unipolar and bipolar thresholds were determined twice. Randomization for the first and second trials was done independently. In the final 20 patients, we determined a strength-duration relationship for AC. In addition to determining thresholds for unipolar and bipolar AC of 5-second duration, we determined thresholds for bipolar AC synchronized to the QRS complex with durations of 0.5 and 1 second. The order of testing was determined randomly. We recorded only the capture threshold and sustained-VT/VF threshold for 0.5- and 1-second applications because intermittent and continuous capture could not be distinguished.
We compared bipolar and unipolar thresholds for intermittent capture, continuous capture, and sustained VT/VF using paired t tests. We analyzed possible correlations between selected clinical variables and threshold values. We analyzed prior myocardial infarction, left ventricular ejection fraction, and heart failure class to test the hypothesis that thresholds are lower for patients who have more advanced heart disease. We performed correlations between thresholds and the clinical variables of body-surface area, age, and sex to test the null hypothesis.
To analyze reproducibility, we used the Lin concordance coefficient16 between first and second determinations of each threshold value in the first 20 patients. We used a previously described17 t test based on absolute differences normalized by median value to determine whether continuous-capture threshold or VF threshold was more reproducible.
Table 2⇓ shows that group mean values for intermittent-capture, continuous-capture, and sustained-VT/VF thresholds were lower for bipolar AC than for unipolar AC (P≤0.001).
Cardiovascular Collapse Caused by Continuous Capture
Continuous capture caused loss of pulsatile arterial pressure and cardiovascular collapse for the duration of AC (Figure 2⇓). Mean arterial pressure fell from 80±14 mm Hg before AC to 32±8 mm Hg at the end of AC. Figure 3A⇓ is a percentile plot of the continuous-capture threshold for 5-second applications of AC. Table 2⇑ shows minimum, mean, median, and maximum values. The ventricular cycle length at the continuous-capture threshold was not significantly different for unipolar AC versus bipolar AC (285±92 versus 274±101 ms; P=0.45).
Relation of Pacing Threshold to Continuous-Capture Threshold
The pacing threshold for single 0.5-ms, constant-voltage pulses was not significantly different for unipolar pulses versus bipolar pulses (1.01±0.89 versus 0.96±0.38 mA; P=0.65), but these values were higher than the corresponding continuous-capture thresholds (P<0.001). The ratio of the pacing threshold for single 0.5-ms pulses to the continuous-capture threshold was 13±13 for unipolar AC and 15±7 for bipolar AC.
ECG Findings at Continuous Capture
At the bipolar continuous-capture threshold, AC interference could not be detected on any of the 3 monitors in any patient. At the unipolar continuous-capture threshold, AC interference ≥0.5 mV was recorded in 4 patients (10%) on the electrophysiological recording system in leads I, aVR, or aVL. Six other patients (15%) had subtle thickening of the baseline <0.5 mV. With the other 2 monitors, AC interference could not be detected for either unipolar or bipolar AC ≤500 μA. Thus, the typical ECG appearance of continuous capture was that of VT with morphology identical to pacing through the electrode catheter but no AC interference.
Sustained VT or VF was induced in 34 patients (85%) for bipolar AC and in 32 patients (80%) for unipolar AC. Figure 3B⇑ is a percentile plot of sustained-VT/VF thresholds. Table 2⇑ shows minimum, mean, median, and maximum values. The end-point arrhythmia (VT or VF) was the same for bipolar and unipolar AC in all patients. It was VF in 25 patients and sustained monomorphic VT (cycle length, 274±51 ms) in 9. Both the group mean threshold and the lowest individual patient threshold were lower for bipolar AC than unipolar AC. The continuous-capture and VF thresholds were equal for 2 patients with bipolar AC and for 1 with unipolar AC.
Ranges of Intermittent and Continuous Capture
Intermittent capture occurred over a narrow range (14±13 μA for bipolar AC versus 22±21 μA for unipolar AC; P<0.02). Continuous capture occurred over a wider range than intermittent capture (P<0.001). In patients in whom sustained VT/VF was induced, the range of continuous capture was less for bipolar AC than for unipolar AC (146±236 versus 193±222 μA; P=0.001). In those patients in whom sustained VT/VF could not be induced, the difference between the maximum output of 999 μA and the continuous-capture threshold (bipolar, 60±7 μA [n=6]; unipolar, 77±18 μA [n=8]) provides a lower limit for this range.
Clinical Correlates of Threshold Values
None of the clinical variables evaluated correlated with the continuous-capture threshold or the VF threshold for unipolar or bipolar AC. The patient with the lowest continuous-capture threshold (32 μA) was a 60-year-old woman who had hypertensive cardiovascular disease, normal coronary arteries, and a left ventricular ejection fraction of 60%. The patient with the lowest VF threshold (49 μA) was a 72-year-old man who had valvular heart disease and an ejection fraction of 20%. There were too few patients without spontaneous or inducible ventricular arrhythmias to provide a meaningful analysis of this subgroup. Of these patients, the lowest continuous-capture thresholds occurred in a 73 year-old man who had coronary artery disease, paroxysmal atrial fibrillation, and a left ventricular ejection fraction of 58%. These values were 50 μA for bipolar AC and 60 μA for unipolar AC.
Figure 4⇓ shows reproducibility of the continuous-capture and sustained-VT/VF thresholds in the first 20 patients. The continuous-capture threshold was more reproducible than the sustained-VT/VF threshold for both unipolar AC (P<0.005) and bipolar AC (P<0.001).
In the final 20 patients, there were no significant differences among the capture threshold for 0.5 second of AC (68±18 μA), the capture threshold for 1 second of AC (67±18 μA), and the continuous-capture threshold for 5 seconds of AC (66±17 μA). In contrast, we induced sustained VT/VF in none of the tested patients when we gave 0.5 second of AC, in 10 patients when we gave 1 second of AC (50%), and in 16 patients when we gave 5 seconds of AC (80%) (P<0.001). In those 10 patients in whom the sustained-VT/VF threshold for 1 second of AC could be measured, the value was 474±255 versus 217±254 μA for 5 seconds (P<0.001). Figure 5⇓ shows an example. The lowest VF threshold for 1 second of AC was 140 μA.
The present study is the first report of the effects of AC applied directly to the heart in closed-chest humans. We found that continuous capture by AC causes cardiovascular collapse over a wide range of AC at levels that are below the VF threshold and in some patients, below the present national standard for a single-fault condition of 50 μA. At this 50 μA-level, AC causes no interference on the ECG, and continuous capture is thus indistinguishable from hypotensive VT. Leakage current ≤50 μA may be a cause of VT or VF in patients with intracardiac electrodes.
Cardiovascular Collapse Caused by Continuous Capture
Green et al5 first described a sequence of 3 responses to increasing AC that corresponds to the sequence of intermittent capture, continuous capture, and electrically induced VF observed in the present study. They and other early investigators were unable to record the ECG during AC because of inadequate filtering. They recognized continuous capture by its hemodynamic consequences.5 6 8 9 Subsequent investigators who used filtered ECGs described continuous capture as “rapid, ineffectual VT”6 and “runs of ectopic beats.”18 In canines, continuous capture for 3 to 5 minutes always resulted in death.8
The continuous-capture threshold for AC is substantially below the capture threshold for a single pacing stimulus. This observation suggests that continuous capture at low levels of AC requires a cumulative or summation effect of subthreshold stimuli.19 20 21
As a basis for safety standards, the continuous-capture threshold is superior to the VF threshold for the following reasons: (1) it defines the minimum unsafe level of AC lasting ≥5 seconds; (2) continuous capture results in cardiovascular collapse over a wide range of AC below the VF threshold; (3) the continuous capture threshold can be determined without induction of VF in the vast majority of patients; (4) the continuous-capture threshold is independent of the duration of AC, whereas the VF threshold is strongly dependent on duration; and (5) the continuous-capture threshold is more reproducible and behaves more like a step function than the VF threshold.
Factors That Influence Thresholds
Electrode size22 23 and location9 23 have been reported to influence the VF threshold in animal studies. To hold these factors constant, we placed a temporary electrode near the right ventricular apex, the most common clinical location. However, some clinically used electrodes have smaller surface areas than the electrodes used in the present study. They would be expected to have correspondingly lower continuous-capture and VF thresholds.22 23
Our study is the first to compare unipolar and bipolar thresholds for AC. We found that the group mean thresholds for continuous capture and VF were lower for bipolar AC than for unipolar AC. Bipolar AC may produce higher intramyocardial current density8 24 or cause unipolar stimulation from the ring electrode, which has a smaller surface area than the tip electrode. The common clinical path for leakage current is unipolar.
Our study is also the first to examine the strength-duration relationship for AC in humans. For AC durations between 0.5 and 5 seconds, we found that the VF threshold decreases but the capture threshold remains constant. The duration of leakage current may depend on the cause: short for a power-line surge, intermediate for current induced by electromagnetic interference from nearby equipment, or long for a broken ground wire.
We determined the minimum unsafe value of AC, not the maximum safe value. We studied patients under propofol anesthesia; thresholds might differ in conscious patients. Most patients in this study had ventricular arrhythmias and structural heart disease. However, patients with cardiac disease are most likely to undergo invasive cardiac procedures and thus to be exposed to intracardiac AC.
The principal clinical implication of our study is that AC causes cardiovascular collapse in closed-chest humans at levels substantially below the VF threshold.
A second implication is that the physician cannot rely on the ECG to distinguish continuous capture by AC from spontaneous VT. American Heart Association guidelines require 60-Hz notch filters on ECG equipment to suppress AC interference,4 and our data show that intracardiac AC at 50 μA causes no such interference. Thus, the clinical presentation of continuous capture by AC ≤50 μA is electrocardiographically silent and indistinguishable from hypotensive VT. Transient continuous capture may be misdiagnosed as nonsustained VT. This spurious diagnosis may lead to unnecessary diagnostic procedures, including costly and invasive electrophysiological studies. Sustained flow of intracardiac leakage current could present as VT refractory to cardioversion. In this situation, the patient’s survival depends on rapid interruption of the leakage-current circuit. However, a responsible physician would probably treat the patient unsuccessfully with the sequence of electrical cardioversions and antiarrhythmic drugs recommended for VT. In a patient with an intracardiac catheter, leakage current should be considered a new mechanism in the differential diagnosis of VT.
A third implication is that under certain circumstances, routine methods would fail to detect leakage-current–induced VF, resulting in sudden cardiac death.14 AC at 50 μA is below the threshold of cutaneous sensation25 and could thus be conducted through an unsuspecting device operator to an unsuspecting patient.
A fourth implication is that safety standards may consider the duration of AC and, by implication, its cause. Under transient conditions ≤1 second, such as power-line surges, the 50-μA standard may be safe.
A final implication is that the 50-μA standard is insufficient to protect patients with intracardiac electrodes from VT or VF caused by leakage current lasting ≥5 seconds. Our results indicate that the safety standard for AC lasting ≥5 seconds must be based on the continuous-capture threshold. In the present study, the maximum value that did not cause cardiovascular collapse in any patient was 20 μA.
This work was done during the tenure of a national Grant-in-Aid from the American Heart Association and the Council on Clinical Cardiology to Dr Swerdlow.
- Received October 26, 1998.
- Revision received February 8, 1999.
- Accepted February 12, 1999.
- Copyright © 1999 by American Heart Association
Association for the Advancement of Medical Instrumentation. American National Standard, Safe Current Limits for Electromedical Apparatus. Arlington, Va: Association for the Advancement of Medical Instrumentation; 1993. ANSI/AAMI ES1–1993.
International Electrotechnical Commission. Medical Electrical Equipment, Part 1. General Requirement for Safety. Geneva, Switzerland: International Electrotechnical Commission; 1988.
Pipberger HV, Arzbaecher RC, Berson AS, Briller SA, Geselowitz DB, Horan LG, Rautaharju P, Schmitt OH. Amendment of recommendations for standardization of specifications for instruments in electrocardiography and vectorcardiography concerning safety and electrical shock hazards: report of the Committee of Electrocardiography, American Heart Association. Circulation. 1972;46:1–2.
Pipberger HV, Arzbaecher RC, Berson AS, Briller SA, Brody DA, Flowers NC, Geselowitz DB, Lepeschkin E, Oliver GC, Schmitt OH, Spach M. Recommendations for standardization of leads and of specifications for instruments in electrocardiography and vectorcardiography: report of the Committee on Electrocardiography, American Heart Association. Circulation. 1975;52:11–31.
Roy OZ, Scott JR, Park GC. 60 Hz ventricular fibrillation and pump failure thresholds versus electrode area. IEEE Trans Biomed Eng. 1976;BME-23:45–48.
Graystone P, Ledsome J. Microshock hazards in hospital: fibrillation thresholds: the wrong parameter. In: Digest of the 10th International Conference on Medical and Biologic Engineering. Dresden, Germany, 1973:159.
Weinberg DI, Artley JL, Whalen RE, McIntosh HD. Electric shock hazards in cardiac catheterization. Circ Res. 1962;11:1004–1009.
Whalen RE, Starmer CF, McIntosh HD. Electrical hazards associated with cardiac pacemaking. Ann N Y Acad Sci. 1964;111:922–931.
Bruner J, Leonard PF. Codes and standards: who makes the rules? In: Electricity, Safety, and the Patient. Chicago, Ill: Year Book; 1989:240–279.
Laks M, Arzbaecher RC, Bailey J, Berson A, Briller S, Geselowitz D. Will relaxing safe current limits for electromedical equipment increase hazards to patients? Circulation. 1994;89:909–910.
Laks MM, Arzbaecher RC, Bailey JJ, Geselowitz DB, Berson AS. Recommendations for safe current limits for electrocardiographs: a statement for healthcare professionals from the Committee of Electrocardiography, American Heart Association. Circulation. 1996;93:837–839.
Laks MM, Arzbaecher RC, Bailey JJ, Geselowitz DB, Berson AS. Comments on “Special report: recommendations for safe current limits for electrocardiographs.” Circulation. 1997;95:277–278. Letter.
Fisch C, Greenspan K. Wedensky’s observations. Circulation. 1967;35:819–820.
Antzelevitch C, Moe GK. Electrotonic inhibition and summation of impulse conduction in mammalian Purkinje fibers. Am J Physiol. 1983;245:H42–H53.
Swerdlow CD, Liem LB, Franz MR. Summation and inhibition by ultrarapid train pacing in the human ventricle. Circulation. 1987;76:1101–1109.
Geddes L, Cabler P, Moore A, Rosborough J, Tacker W. Threshold 60-Hz current required for ventricular fibrillation in subjects of various body weights. IEEE Trans Biomed Eng. 1973;465–468.
Starmer CF, Whalen RE. Current density and electrically induced ventricular fibrillation. Med Instrum. 1973;7:3–6.