The Guardian: an implantable system for chronic ambulatory monitoring of acute myocardial infarction
Article Outline
- Abstract
- Introduction
- Materials and methods
- Clinical history
- Results
- Summary
- Acknowledgments
- References
- Copyright
Abstract
The AngelMed Guardian is an implantable medical device that records cardiac data and detects ischemic events using a standard pacemaker intracardiac lead positioned in the right ventricular apex. The Guardian has been implanted in 55 people in the United States and Brazil and is currently undergoing a Food and Drug Administration phase 2 pivotal trial in the United States. The Guardian detects acute ischemic events by analyzing ST-segment shifts. The ST-segment shifts are calculated as the difference between the ST deviation of a current 10-second electrogram window and a baseline ST deviation value. If the ST-segment shift is greater than a heart rate–dependent programmable threshold, then the device generates an emergency alert signal. Results thus far have demonstrated that (i) the intracardiac electrogram is relatively noise-free and (ii) the ST-shift technique used by the Guardian is effective for detecting acute ischemic events.
Keywords: ST deviation, Acute ischemia, Myocardial infarction, STEMI, Implantable device, Chronic monitoring
Introduction
Acute myocardial infarction (MI) remains one of the leading causes of mortality in the Western world.1, 2 In the United States, approximately 1.3 million cases of MI occur each year, with 65% new and 35% reoccurring cases.3 Prompt treatment of a coronary artery occlusion can either prevent MIs or at least lessen their severity, thereby substantially improving clinical outcomes.3, 4, 5 Decreasing the time between occlusion and treatment has proven difficult for a number of reasons. In particular, some occlusions may result in silent ischemia, so that the affected population does not seek treatment. In addition, people often ignore symptoms or delay-seeking treatment because they are unsure whether they are really experiencing a severe medical problem.5 To detect the occurrence of an acute coronary occlusion and then quickly issue appropriate warnings to patients to seek medical attention requires chronic monitoring of patients' hearts. This type of long-term monitoring requires an implantable medical device (IMD) and cannot realistically be achieved using Holter-type technology. Long-term monitoring with an implantable device permits the use of self-normative–based, rather than population normative–based, detection of cardiac features that are determined to be abnormal for that patient.
The AngelMed Guardian device performs these chronic monitoring and warning functions. As will be described more fully below, the Guardian system includes an IMD implanted like a single-chamber pacemaker within the upper-left pectoral region. The IMD analyzes the electrical signals associated with the intracardiac electrogram (ICEG) that are recorded from a standard pacemaker lead within the right ventricular (RV) apex, with the IMD housing serving as a reference in a can-to-tip lead orientation. Clinical data support the use of an ICEG to detect acute ischemic events.6, 7 The ICEG has proven to be relatively free of various noise sources that tend to confound body surface recordings.8 In particular, because the recording electrode is within the heart itself and maintains a fixed position within the ventricle as the heart moves within the chest, the ICEG tends to not suffer from muscle noise, axis shifts, or motion artifact. The Guardian continuously monitors the ICEG. If an acute ischemic event occurs, the implanted device will alert the patient with vibration alerts and will issue RF signals to an external pager-sized device that provides audible and visual alerting signals. The goal is to enable patients to obtain prompt medical treatment upon the generation of such alerts.
The Guardian System has the risk profile for lead and pocket complications after surgical implantation similar to that of a single-chamber pacemaker (∼1%). The system must also be replaced when the battery runs out (4-8 years) as an additional surgery. The Guardian is also not usable in patients who have or need an implantable cardioverter defibrillator (ICD) or pacemaker, although the Guardian algorithm is already available in the AnalyST ICD from St Jude Medical (St. Paul, Minnesota) and is under development for use in pacemakers. The Guardian is most suitable for individuals who are at relatively high risk for a heart attack, including people with a family history of cardiac pathology or who have experienced a recent cardiac event such as a first heart attack.
Over the last 2 years, the Guardian has been implanted in 20 high-risk patients in Brazil and 17 patients in the United States in connection with a Food and Drug Administration (FDA) study showing safety and feasibility. In 2008, the FDA granted approval for a large randomized prospective pivotal study (AngelMed Early Recognition and Treatment of STEMI), which has enrolled and implanted the Guardian IMD in 18 patients to date.
Materials and methods
System components
This device is implanted in a manner identical to a single-chamber pacemaker. The Guardian system includes a pacemaker-sized IMD that is inserted subcutaneously in the upper-left pectoral region and receives data from a standard pacemaker lead (eg, 1488T/TC, 1688TC Tendril SDX Steroid-Eluting Screw-In Lead; St Jude Medical, St. Paul, Minnesota) that is fed through a cardiac vein and into the apex of the right ventricle. The IMD connects to its lead using a lead adaptor that contains a long-range telemetry antenna. The antenna allows communication within a 6-ft radius between the IMD and external components such as a pager-type external alerting device (EXD) and a physician programmer. The programmer is used to set operational parameters of the IMD, such as ischemia-detection thresholds, and to retrieve cardiac waveforms, detected cardiac events, and summary data from the IMD. The programmer can display the ICEG waveforms and trends of various cardiac features that have been collected across days, weeks, or months. The components of the system are shown in Fig. 1.
Fig. 1.
Components of the implantable ischemia-detection system. Implantable components: the IMD, resembling a Ventricular Demand Inhibited Pacemaker (VVI) pacemaker, is shown (A), attached to the lead adapter (B) and a standard bipolar, screw-in, RV lead (C). External components: the external pager-sized alerting device (E) is attached using a cable (F) to the programmer (D) in order to permit IMD programming and data retrieval, via telemetry. In addition to sound cues, the EXD has 2 light-emitting diodes: red light for emergency alerts related to acute ischemia and severe cardiac event detection, and yellow light for “see your physician” alerts that are issued for lower priority events.
Normal operation
The IMD collects and processes a 10-second ICEG data window every 90 seconds for normal sinus beats within a defined resting heart rate range (eg, 50-90 beats/min). Data are amplified using a gain factor of between 62.5× and 625× and then band-pass filtered with low- and high-frequency cutoffs of 0.25 and 48 Hz, respectively. Analog-to-digital conversion then occurs at 200 Hz. If the ICEG data window has any potential deviation from this normal beat type, then the 90-second repetition rate is shortened to 30 seconds. The IMD analyzes the ICEG signal to determine whether an abrupt ST-segment shift has occurred by comparing the ST deviation (ST-voltage difference from isoelectric) of the current ICEG data window with the patient's normal (baseline) ST deviation. The ST-segment deviation for each beat is measured by subtracting the average PQ-segment voltage from the average ST-segment voltage. The ST- and PQ-segment voltages are calculated as the average of the potentials measured across an interval, having a start time (relative to the R wave peak) and duration that are defined in a heart rate–related manner, as will be described.
The baseline ST deviation is determined as a rolling 24-hour average of the ST deviations of baseline segments, which are acquired hourly. Specifically, once an hour, the IMD will attempt to find an “acceptable baseline” data window that meets 2 criteria: it must be characterized by a resting heart rate range, and it must have an average ST deviation that is not substantially shifted compared with the 24-hour average of the previous baselines. If the baseline data window meets these criteria, the average ST deviation, calculated across its individual heartbeats, is stored as 1 of 24 baseline values. In addition to this ST deviation data, the average R wave height is also stored and is used to normalize the ST deviation values according to the amplitude of the overall ICEG signal. This normalization is necessary because of the changes in the ICEG signal that may occur over time due to either changes that are physiological or which are likely related to the lead-tissue interface. This slowly adapting baseline reference allows both a person's ST deviation and signal strength to gradually drift without triggering detection of acute ischemia.
The ST and PQ start times and durations used by the IMD are manually selected, in a heart rate–dependent fashion, using the physician programmer (see Fig. 1). For each particular patient, heartbeats for 5 different heart rate ranges (“bins”) are displayed by the programmer. The heartbeats for these different heart rate ranges were acquired during a stress test that occurred shortly after device implantation. For each heart rate range, the medical personnel who programmed the IMD used a cursor to select the appropriate ST and PQ start times and durations by measuring these on a sample of displayed heartbeats. The programmer used these selections to calculate the start times and durations of the ST and PQ components relative to the R wave peak. This information was then uploaded into the IMD and used for monitoring the patient. During monitoring, to determine start times and durations of the ST and PQ intervals for particular beats, the IMD first determines the R wave peak and the heart rate range for the beats being measured, and then measures the ST and PQ features according to the start times and durations defined for that range.
The IMD detects acute ischemia when 3 consecutive data windows have an ST-segment shift% (ie, ST deviation of current window minus the ST deviation of 24-hour average baseline as a function of R amplitude) that exceeds a programmable, heart rate–dependent, ischemia-detection threshold. To determine whether a particular ST-segment shift is indicative of acute ischemia, the IMD applies heart rate–dependent thresholds. In clinical exercise stress testing, the standard 12-lead electrocardiogram (ECG) will register ST-segment shifts as heart rates increase. This is true for intracardiac data as well. As will be shown in “Results,” the ICEG tends to register ST-segment depression at higher heart rates. These heart rate shifts, which are normal and which vary from person to person, should not trigger an acute ischemia alarm. To avoid an alarm, the ischemia-detection threshold is set at a different level for different heart rate ranges, in a patient-specific manner. For example, if a negative shift of −10% (compared with baseline) is normal for a particular individual at a high heart rate, the detection threshold is set lower than −10% (eg, −25%). Accordingly, even if a patient normally shows ST-segment changes during heart rate increases in the absence of ischemia (in the case of either precordial or intracardiac leads), these changes will be larger when ischemia is present and will surpass this “normal” range for the patient.
The IMD allows the detection thresholds for a positive ST shift% to be set at different levels than the thresholds for a negative ST shift%.
Patient alerting
Upon detecting an acute ischemic event, the IMD generates a vibratory alert and communicates with the EXD so that it generates visual and auditory alerts. The patient is then expected to immediately seek treatment. To aid medical personnel in the diagnosis of acute ischemia, the IMD saves raw ICEG data pertaining to periods both before (up to 24 hours) and after (8 hours) an alert, which can be downloaded and reviewed upon meeting with the patient.
To assist doctors in the diagnosis of chronic ischemia, the IMD also stores ST deviation trend data, organized as a function of several defined heart rate ranges, during periods of weeks or months. These data may be downloaded from the IMD using the physician programmer and analyzed by a physician to help determine whether the patient may require additional diagnostic tests, such as stress tests, to help diagnose chronic ischemia.
The IMD can also be set to detect conditions other than ischemia such as tachycardia, bradycardia, irregular rhythms, and conditions relevant to the device's function (eg, “low battery”). These events may be programmed to generate a “See Doctor” alert that has different characteristics than the emergency alert associated with an acute ischemic event.
Clinical history
Early concept testing6 was conducted by monitoring ICEG's using a temporary pacing lead pushed against the RV apex during percutaneous coronary intervention (PCI). The results showed significant ST shifts, typically 20% to 50% of the QRS amplitude, after 2 minutes of arterial occlusion. These data were used to develop the Guardian detection algorithm.
The Guardian has undergone both animal-model and human-clinical testing. The preclinical Good Laboratory Practices animal study involved ambulatory pigs that were implanted with copper stents in selected coronary arteries to provoke acute thrombotic events.7 In an early study, 20 high-risk patients indicated for PCI with identified coronary stenoses and failed stress tests received Guardian IMDs in Brazil. After the implant, data were recorded by the IMDs during repeated stress testing and subsequent PCI. In the United States, a phase I feasibility study was carried out with 17 high-risk patients recruited by 3 centers in the United States (Borgess Research Institute, Kalamazoo, MI; Virtua Hospital, Cherry Hill, NJ; Baptist Hospital West, Knoxville, TN). The phase I study was completed in August 2008. In December 2008, the USFDA granted full approval for a phase II randomized prospective study, which started enrolling patients later that month. In connection with this study, 18 patients have been implanted as of the date of this writing.
Results
Pig study
Detailed results for the porcine study have been described previously.7 Briefly, the Guardian generated ischemia-detection ST-shift alert signals for all the pigs that received copper stents. Pathology analysis revealed that all of the stented animals experienced vessel occlusions that led to MIs.
Human implants (balloon angioplasty)
A number of patients who have been implanted with Guardians also underwent balloon angioplasty. The Guardian generally recorded negative shifts for left anterior descending artery (LAD) angioplasty and positive shifts for right coronary artery (RCA) and left circumflex artery (LCX). Fig. 2 shows sequences of 10-second electrogram windows recorded by the Guardian during LAD, LCX, and RCA PCI angioplasties.
Fig. 2.
Guardian recordings from before (baseline) and 2 to 3 minutes after balloon inflation, which occurred during angioplasty (inflation). Data are shown from 3 patients (B7, B10, and B9, respectively) who underwent balloon occlusions of the LAD, LCX, and RCA, resulting in ST shifts of −42%, 34%, and 18%, respectively. In the LAD occlusion recording, the arrow points to the time at which the balloon was deflated. The negative ST shift begins to shift positive (toward the shape it demonstrated during the preocclusion baseline recording) within 2 beats.
Human implants (alerts in ambulatory patients)
During the feasibility study, 10 abnormal ischemia detections occurred across 6 patients and resulted in emergency alert signals being triggered. These alerts have led to medical interventions such as stenting of a narrowed coronary artery and have been supported by evidence from angiography, intravascular ultrasound, stress testing, and biomarker techniques. A study of representative clinical-case reports is being prepared for publication.
ST-segment stability
The average of ST-segment shift% in the normal/resting heart rate range, for both positive and negative shifts (ie, polarity ignored), across all patients is within the range of 5%, where ST-segment shift% is calculated as:
(1)where STdeviationC is the ST deviation of the current window, STdeviationB is the ST deviation of an average 24-hour baseline and RwaveH is the R wave height of the average 24-hour baseline. The corresponding average standard deviation (calculated for each patient and then averaged) is within the range of 3% of R wave height, with 90% of all values being less than 6%. In summary, the ICEG ST segments of our ambulatory patients have generally been quite stable.
Stress tests
All patients underwent an exercise treadmill or pharmacological (dobutamine) stress test to enable the ST-segment detection thresholds to be set in different heart range ranges (eg, normal resting range, elevated range number 1, etc.). This stress test data can be used to allow the programmer's software to automatically set ischemia-detection thresholds based on statistical assessment of the variance of the ST segments that occurs during the test. In almost all instances, ST-segment deviation became more negative at higher heart rates.
For the 17 patients involved in the US phase I clinical trial, the average heart rate increased from 57 to 124 beats/min. The average ST shift% for these patients was −40%. To enable a rough comparison between this −40% shift and the change that occurs on surface leads, assuming a surface R wave amplitude of 2 mV, a −40% shift is −0.8 mV, whereas (horizontal or downsloping) ST depression of at least 0.1 mV on the 12-lead ECG during stress tests is often considered indicative of coronary artery disease.9 Intracardiac potentials are far larger than body surface potentials because of the electrical loading effects of the tissue and body fluids outside the heart, which diminish current flow with increasing distance from the heart.
Of the previously mentioned 17 patients, 7 showed maximal ST depression simultaneous with maximum heart rate, whereas the remaining 10 patients showed maximal ST depression after the heart rate began decreasing from its maximum value. For these 10 patients, maximal ST depression occurred an average of 2.8 minutes after heart rate recovery began. Maximal ST depression during the recovery phase has also been recorded in various clinical studies that involve surface recordings.10, 11
The Guardian stress test recordings are generally free from artifact. A major source of noise in surface recordings—motion artifact—is typically absent in the Guardian recordings because of the secure attachment of the recording electrodes to areas inside the body.
Polarity of ST shifts
The polarity of ST changes associated with ischemia appears to depend on 2 factors: (i) whether the change occurs in connection with a heart rate increase and (ii) if the change occurs at a normal heart rate, as a result of a coronary artery blockage, the location of the culprit artery.
As mentioned previously, ST shifts associated with heart rate increases have been negative in almost all cases. This is consistent with at least some reports of surface recordings, which show heart rate–related ST depression at similar torso positions (eg, in the area around V5 of the standard 12-lead ECG) in patients with chronic ischemia, regardless of the culprit artery.12 Some modeling work13 suggests that these surface results may be explained by a sort of global endocardial “ischemia.” According to this theory, the endocardium as a whole tends to repolarize earlier than normal at high heart rates. This causes the entire endocardium to be electrically positive during the early ST segment. The Guardian IMD, which reverses the polarity of signals measured from the RV endocardium (by measuring potential in a “can-to-tip” direction), would thus register ST depression, as is observed (ie, as an increased negative shift). This is far from a definitive proof of the theory: more detailed modeling and experimental work, including simultaneous body surface mapping and Guardian-based IMD recordings, could help to shed light on the mechanism underlying ST shifts at high heart rates.
At normal heart rates, in all instances of LCX and RCA occlusions (whether as a result of angioplasty or spontaneously occurring), the ST shift has been positive, whereas in most instances, LAD ST shifts have been negative. These results accord with a rather simplified theory of ischemia, according to which ischemic tissue is electrically positive during the ST segment, and nonischemic tissue is reciprocally negative. If the RV apex—where the Guardian electrode records—is perfused by the LAD, occlusions of that artery would tend to cause ST depression (in the can-to-tip lead orientation) because the ischemic RV apex becomes electrically positive. An opposite result could occur if an LAD occlusion resulted in a more left-lateral ischemic region, resulting in ST elevation. In contrast, LCX and RCA occlusions would tend to cause the RV apex to become reciprocally electrically negative, which results in ST elevation in the Guardian's can-to-tip lead orientation.
Summary
We have provided an overview of an implantable ischemia-detection device that underwent 2 studies with a total of 37 subjects and is now in FDA pivotal trials. The Guardian system is based on an intracardiac lead, which has known drawbacks related to implantation, infection, and replacement that are similar to those encountered with ICDs and pacemakers. The Guardian device has already shown to be useful in inducing patients to rapidly seek care in order to deter or decrease damage related to the occurrence of an MI. Examination of the polarity of the ST shifts that have occurred supports a model of ischemia previously proposed by our group (17), whereby the polarity is related to the location of ischemia in relation to the position of the intracardiac electrode of the device. The measures of ST shift% variation computed on the ICEG data indicate that 90% of daily fluctuations are less than 6%, suggesting a relatively stable environment for detection of abnormal intervals of acute ischemia.
Acknowledgments
The authors would like to acknowledge the substantial contributions of Jonathan P. Harwood, Steven R. Johnson, Dave Keenan, Rich Bantel, Jill Schweiger, Santiago Marques, and Rosana Suemi Nakamura to this paper.
References
- . The world health Report. http://www.who.int/entity/whr/2004/en/report04_en.pdf2004;[Accessed on 17 April 2009]
- . Heart disease and stroke statistics—2009 update. http://www.americanheart.org/presenter.jhtml?identifier=3000090[Accessed on 17 April 2009]
- A comparison of immediate angioplasty with thrombolytic therapy for acute myocardial infarction. N Engl J Med. 1993;328:673
- Immediate angioplasty compared with the administration of a thrombolytic agent followed by conservative treatment for myocardial infarction. N Engl J Med. 1993;328:685
- . Timing is everything: motivating patients to call 9-1-1 at onset of acute myocardial infarction. Circulation. 2001;104:1210
- Potential of an intracardiac electrogram for the rapid detection of coronary artery occlusion. Cardiovasc Revasc Med. 2005;6:14
- . Real-time detection and alerting for acute ST-segment elevation myocardial ischemia using an implantable, high-fidelity, intracardiac electrogram monitoring system with long-range telemetry in an ambulatory porcine model. J Am Coll Cardiol. 2006;48:2306
- . Using simulated noise to define optimal QT intervals for computer analysis of ambulatory ECG. Med Eng Phys. 1999;21:15
- . Exercise Electrocardiogram Testing: Beyond the ST Segment. Circulation. 2006;114:2070
- Comparison of exercise-positive with recovery-positive treadmill graded exercise tests. Am J Cardiol. 1987;60:544
- . Diagnostic and prognostic value of ST segment depression limited to the recovery phase of exercise stress test. Heart. 2004;90:1417
- ST-T integral and T-wave amplitude in detection of exercise-induced myocardial ischemia evaluated with body surface potential mapping. J Electrocardiol. 2003;36:89
- . ST segment depression: the possible role of global repolarization dynamics. Biomed Eng Online. 2007;6:6
PII: S0022-0736(09)00263-5
doi:10.1016/j.jelectrocard.2009.06.017
© 2009 Elsevier Inc. All rights reserved.
