Recommendations for Safe Current Limits for Electrocardiographs

The release of a revised American National Standard¹ that permits an increase in current limits purported to be safe makes it urgent that the American Heart Association review its recommendations regarding safety and electrical shock hazards for electrocardiographs. These recommendations cover two aspects of electrical safety. The first is the level of current allowable in any patient-connected lead that may flow through the myocardium without inducing ventricular fibrillation. The second is the allowable chassis leakage current that may flow from the electrocardiograph to ground, passing through patient or operator. These recommendations for safe current limits reaffirm the levels of allowable current established for patient-connected leads and revise the amount of allowable chassis leakage current established in previous AHA recommendations.

Current in Patient-Connected Leads

Recommendation: The electrocardiographic (ECG) or vectorcardiographic apparatus shall be designed so that no more than 10 μA root mean square, from direct current to the tenth harmonic of power line frequency, shall flow through any patient-connected lead under either normal or single-fault conditions.

Justification: In 1972 the AHA, in an amendment to its 1967 report,² recommended an upper limit of 10 μA for current between any patient electrode and either power line ground or the accessible part of the electrocardiograph. The concept of a single fault was introduced to define the occasional failure of one component of the ECG equipment, an error in power distribution connections, or a wiring error on the part of the operator. The AHA recommended that the 10 μA limit not be exceeded, even in the presence of a single fault, to address occurrences such as an insulation failure in a line-operated component, an incorrectly wired power line receptacle, a single failure in an electronic circuit, or a disconnected power line ground. The AHA based its 1972 recommendation on the fact that the 10 μA limit had been previously specified by the 1971 National Electrical Code,³ the 1971 National Fire Protection Association,⁴ the 1972 Underwriters’ Laboratory,⁵ and the 1971 report of the Subcommittee on Electrical Safety of the Association for the Advancement of Medical Instrumentation (AAMI).⁶ In addition, in one set of experiments in dogs⁷ ventricular fibrillation was produced with a 20-μA current.

Watson et al,⁸ Starmer and Whalen,⁹ and Raftery et al¹⁰ have described electrical threshold experiments in humans. Watson et al studied 15 patients with an endocardial electrode; the smallest current that produced ventricular fibrillation in humans was 15 μA. These studies were another basis for the AHA’s reaffirmation of the 10 μA limit in 1975.¹¹

Watson et al⁸ raised the argument that little or no current is required to fibrillate a sufficiently irritable heart, and mechanical stimulation itself may be enough. Nevertheless, in 1972 the AHA decided that it was unrealistic, from the engineering design point of view, to establish a limit below 10 μA. The reasons for this are related to power transformer design as well as the lengths of conducting leads and power line cables and their capacitive coupling to ground. Another important consideration is that immunity to electromagnetic interference is more difficult to achieve as lower limits are placed on current-through-patient leads and chassis leakage current. These reasons remain relevant in 1996.

Since 1975 no human studies on electric current levels and ventricular fibrillation have been published. Thus, the AHA reaffirms its 1975 recommendations limiting unintended currents in patient-connected leads to 10 μA under no-fault or single-fault conditions.

The most recent American National Standard¹ specifies 10 μA as the maximum risk current in patient leads under no-fault conditions. The national standard relaxes this limit to 50 μA under single-fault conditions, based on the judgment that a single fault is a low-probability event. The AHA does not agree with that judgment. The likelihood of a single fault arising from the electrocardiograph itself, miswiring of a power line receptacle, or connection of an indwelling line to power line voltage because of human error does not differ fundamentally today from 20 years ago. Moreover, as noted in a recent editorial,¹² manufacturers of electromedical equipment continue to meet the 1975 AHA risk current limits, ie, 10 μA under either single-fault or no-fault conditions.

Chassis Leakage Current

Recommendation: Chassis leakage current shall be limited to 100 μA at line frequency under either no-fault or single-fault conditions. Ideally, 10 μA should be the goal.

Justification: Unintentional currents that flow from the electrocardiograph to ground are usually referred to as leakage currents. Pathways for these leakage currents include stray capacitances within the device and capacitances from patient leads to the ECG enclosure or chassis. The ground wire of the electrocardiograph or other electromedical equipment should carry leakage currents to ground. In certain circumstances, such as when the operator or patient contacts the conducting surface of the equipment while touching another grounded, conducting surface, leakage currents may pass through the operator or patient. Thus, limits were set in previous versions of the American National Standard.¹³ For devices such as electrocardiographs, which may be used with an indwelling lead, the limit was specified as 100 μA under either no-fault or single-fault conditions. The 1993 revision,¹ however, relaxes the limit to 300 μA under single-fault conditions.

The AHA Committee on Electrocardiography previously addressed chassis leakage current in both 1972 and 1975. In both reports the 10 μA limit is recommended for current flowing between any two of the following:

• Electrodes

• Case (chassis)

• Other accessible part normally or potentially in contact with the patient

• Power line ground at the receptacle from which the apparatus is energized

The AHA’s justification for including the case or chassis was the potential for the patient touching or contacting the case (either directly or indirectly through an operator in contact with the case) while at the same time being grounded through another pathway, such as an indwelling catheter (in a blood vessel or the heart chamber) connected to an instrument other than the electrocardiograph. In such an instance the ECG chassis leakage current will have a low resistance pathway to ground through the heart. Thus, the 10 μA limit appears to be as applicable to chassis leakage current as it is to any ECG patient lead. No reliable data indicate the relative probabilities of hazards from chassis leakage current vis-a-vis current-through-patient leads.

The AHA has been reluctant to raise the 10 μA limit for chassis leakage current recommended in 1975. However, two arguments have been persuasive in the decision to recommend a limit of 100 μA. From the technical/engineering perspective, electrocardiographs have generally become more complex, with many systems now capable of measuring amplitudes and durations, averaging, and implementing interpretive algorithms. This increased complexity increases the difficulty of designing an apparatus that limits chassis leakage current to 10 μA. From another viewpoint, over the past 15 to 20 years there have been no reported and documented instances of ventricular fibrillation resulting from ECG chassis leakage current. Yet most, if not all, electrocardiographs used in the United States have been designed to meet the American National Standard that specifies 100 μA as the limit for chassis leakage current. Therefore, the AHA judges 100 μA to be a more realizable limit for chassis leakage current at present, under either no-fault or single-fault conditions. The many possibilities for single faults that may lead to chassis leakage current flowing through the heart make it equally important to limit this current to 100 μA.

While the 100 μA limit is the recommendation for chassis leakage current, it is important to keep in mind that unintentional current through the heart, regardless of its source, is the important consideration. Thus, the AHA strongly encourages manufacturers of electromedical equipment to work toward achieving a 10 μA limit for chassis leakage current. Many modern electrocardiograph designs use a nonconductive (plastic) case, which is an effective way to limit chassis leakage current without requiring expensive circuit components. Other design strategies may also be useful for achieving the 10 μA limit. Additionally, it should be recognized that the absence of documented incidents of ventricular fibrillation resulting from chassis leakage current does not lead logically to the conclusion that 100 μA chassis leakage current provides adequate safety for the patient with an indwelling conductor. No controlled studies have been undertaken to search specifically for data to substantiate the role of chassis leakage current in instances of ventricular fibrillation. This gap in knowledge must be addressed.

Electric Current Perception

What about other effects of electric current? This report focuses on electrocardiographs and vectorcardiographs and unintentional current that may induce ventricular fibrillation. However, electric current generated by other electromedical equipment may flow through the skin and introduce nonlethal shock effects that should be avoided.

In 1990 Tan and Johnson¹⁴ performed experiments in which 60-Hz currents ranging up to 500 μA were applied through electrodes on the forearms of 95 adult volunteers. The subjects’ perceptions and reactions were recorded:

• At an average current of 83 μA some sensation was detected.

• At an average current level of 298 μA strong sensations were detected. The lower limit of these strong sensations was 130 μA for women and men.

• At more than 400 μA most “. . . described the sensation to be strong and painful.”

The Tan and Johnson study establishes a basis for setting a limit on current that may be delivered to the skin. Given the need to accommodate variation among individuals and provide a reasonable safety factor, there is ample justification for limiting this current to 100 μA. It is especially important to maintain this limit under single-fault conditions, because in the presence of a single fault, such as an open ground wire, this protection is most needed. The relaxation of chassis leakage current to 300 μA under single-fault conditions in the latest American National Standard¹ thus stands out in contradistinction to the study by Tan and Johnson.

In a recent editorial Laks et al¹² noted that currently available ECG systems meet the 10 μA allowable limit for current through patient-connected conductors and described the value of an appropriate reporting system. Some of their remarks, slightly modified, are repeated below to emphasize their importance:

Of significance, in 1996, manufacturers of electromedical equipment continued to meet the AHA’s 1975 safety recommendations limiting current in patient-connected leads to 10 μA. During this period there were no reports either in the literature or from the Food and Drug Administration of an instance of ventricular fibrillation or other harmful effects caused by alternating current leakage. This is not surprising. Reconstruction of equipment and connection conditions after an incident occurs creates a major problem. Unless a study is designed with the specific goal of assessing the effects of current flow through the body or the myocardium, it is unlikely that an association will be found. In the absence of a precise nationwide system for evaluating the incidence of adverse effects of unintentional electric current, the apparently benign nature of even the modest currents allowed by the AHA cannot be ensured. In the absence of such a reporting system and without new scientifically based data to supersede the old, manufacturers of electromedical equipment are strongly urged to continue to design systems that meet the AHA’s 1975 recommendations.

In the absence of credible data, the increases in risk current permitted by the American National Standard¹ constitute experimentation on humans without their consent to determine the safe range of such currents. There is a clear need for ethically approved clinical studies to resolve this issue.

Test methodologies are extremely important for validating the allowable current levels for patient-connected leads and chassis leakage. The AHA recommends the test methods described in the latest American National Standard.¹

In summary, the AHA recommends that under both normal and single-fault conditions, ECG risk currents be limited to 10 μA through patient-connected leads. This recommendation remains unchanged from previous AHA recommendations. In addition, the chassis leakage current must be limited to 100 μA under both normal and single-fault conditions. The AHA emphasizes that the important goal of reducing safety hazards will be significantly advanced by instrument improvements targeted to achieve a chassis leakage current limit of 10 μA. Clinical investigators and members of industry are strongly urged to collaborate in conducting experiments and gathering appropriate data to establish a modern scientific basis for safe current limits and to establish a reporting system for documentation of ventricular fibrillation induced by electric current.

“Recommendations for Safe Current Limits for Electrocardiographs” was approved by the American Heart Association SACC/Steering Committee on October 19, 1995.