An Electrocardiographic and Acoustic Cardiographic Study of Acute Coronary Occlusion During Percutaneous Coronary Intervention

Eunyoung Lee

Doctor of Philosophy

University of California, San Francisco, 2008

Ventricular dysfunction, as measured by a rise in ventricular filling pressure or abnormal wall motion on echocardiogram, has often been observed without or prior to ST changes during ischemia. However, these conventional measurement methods for ventricular function are either invasive or not readily available at the bedside.

The overall aim of this dissertation has been to investigate whether, using acoustic cardiography, the appearance or increased intensity of diastolic heart sounds, a non-invasive measurement of ventricular function, may (a) augment the detection of ischemia when combined with ST changes and (b) occur earlier than ST changes or angina during ischemia induced by coronary occlusion during PCI.

A prospective observational study was performed with a convenience sample of 80 patients referred for non-urgent cardiac catheterization with possible PCI. Acoustic and ECG data were gathered simultaneously and continuously before and during cardiac catheterization and PCI.

Among the 80 patients, 24 patients received PCI and developed ST changes > 2 SD in ≥ 2 contiguous leads during coronary occlusion. Diastolic heart sounds occurred in all patients who met the standard ST-T wave criteria for ischemia; in two-third of patients without ST changes; and in about 90% of patients without angina during coronary occlusion. All patients who did not have diastolic heart sounds during coronary occlusion in PCI did not have ST changes that met the standard ST-T wave criteria. Our study showed that the combined use of diastolic heart sounds and ST-T wave criteria improved the sensitivity to predict myocardial ischemia by 31.6% but reduced the specificity by 22.2%. The absence of diastolic heart sounds was shown to be helpful to rule out the presence of myocardial ischemia.

Electrocardiographic ST-T changes were the earliest sign of ischemia during coronary occlusion and they were also the first to resolve after balloon deflation. Anginal symptoms were either the last event in the ischemic cascade or, in 90% of patients, did not occur during coronary occlusion. The monitoring of changes in diastolic heart sounds along with ECG monitoring at the bedside may improve the early detection of myocardial ischemia.

Approved:

_______________________________________

Barbara J. Drew, RN, FANN, PhD

Dissertation Chairperson

Introduction: Physiologic monitors play a vital role in saving patients' lives but expose clinicians to an overwhelming number of alarms, many of which are false. Objective: Our study aims were; a) to determine whether the mean hourly oxygen saturation (SpO2) low-limit alarm rates could be reduced by modifying the default alarm setting and b) to determine whether there would be a reduction in mean hourly technical and critical arrhythmia alarm rates and the mean percentage of false-positive critical arrhythmia alarms following daily skin preparation and the application of high-quality ECG electrodes. Methods: We conducted a prospective randomized clinical trial in two neuroscience intensive care units, collecting data during two assessment periods. Each patient's alarm rate was calculated as the number of unique alarms divided by monitoring time. Critical arrhythmia alarms were determined (true vs. false positive) using a protocol. Means and standard deviations of the hourly alarm rates and the mean percentage of false-positive critical arrhythmia alarms were determined during both assessments. A negative binomial regression was performed to test the main effect of unit, the main effect of assessment, and the unit by assessment interaction. Results: The combined use of a lower SpO2 low-limit threshold and increased alarm delay resulted in a significant unit-by-assessment interaction (p <. 001). During Assessment 2, the experimental unit had a lower mean hourly SpO2 alarm rate while in the control unit, the rate increased. No significant unit-by-assessment interactions were observed for the mean hourly technical alarm rates; during Assessment 2, both units experienced an increase in ECG lead fail alarm rates and although both units had a reduction in mean hourly artifact alarms, it was insignificant. Similarly, no significant unit-by-assessment interaction in the mean hourly critical arrhythmia alarm rate was observed; no reduction in critical arrhythmia alarm rates was found in the experimental unit. Likewise, the intervention did not reduce the mean percentage of false-positive critical arrhythmia alarms. Conclusion: A lower SpO2 alarm limit and increased alarm delay safely reduces non-actionable alarms. However, our novel electrode regimen does not reduce critical arrhythmia and technical alarm rates or false-positive arrhythmia alarms.

Background. Little is known about the prevalence and prognostic significance of long QT interval (QTi) among patients with chest pain during the acute phase of suspected cardiovascular injury. Previous attempts to research similar questions resulted in conflicting conclusions. This is primarily because of the inappropriate expression of the QT/RR relation when calculating the QTi corrected for heart rate (HR).

Aims. The author aims to (1) determine the prevalence and time duration of long QTi , (2) investigate whether a long QTi is associated with mortality and other adverse cardiovascular events at 1 year follow-up, and (3) determine whether the QT/RR slope is different among different HR zones (bradycardia, normal, and tachycardia).

Methods. This is a secondary analysis of data obtained from the IMMEDIATE AIM trial. Data included 24-hour 12-lead Holter electrocardiographic recordings. The sample of the present analysis consisted of patients who presented to the emergency department with chest pain (N, 145). The QTi was measured automatically and rate corrected using seven QTc formulas including subject specific correction. The formula with the closer to zero absolute mean QTc/RR correlation was considered the most accurate. The QT/RR slope was compared among different HR zones applying steady (with minimal fluctuation) beat selection technique.

Results. Subject specific QT rate correction outperformed other QTc formulas and resulted in the closest to zero absolute mean QTc/RR correlation. Using linear subject specific correction, the prevalence of long QTc interval was 14.5%. The QTc interval did not predict mortality or hospital admission at short and long term follow-up. The QT/RR slope was significantly less under the bradycardia zone (0.074 ± 0.07) compared to the normal (0.114 ± 0.1, p < 0.01) or the tachycardia zones (0.147 ± 0.1, p < 0.001). There was no difference between the normal and the tachycardia zone.

Conclusion. Adequate QT rate correction can only be performed using subject specific correction. Long QTi is not uncommon among patients presenting to the emergency department with chest pain. The QT/RR slope is different among different HR zones. These differences can be used in resetting the limits defining different HR zones boundaries.

Background

Methods

Results

Conclusions

Autonomic nervous system (ANS) dysfunction has been associated with cardiac arrhythmias and death. The aim of this research was to determine whether electrocardiographic (ECG) measures of ANS function, heart rate variability (HRV) and heart rate turbulence (HRT), have prognostic value in patients with acute coronary syndrome (ACS). A secondary analysis was performed using 24-hour ECG recordings from patients enrolled in the IMMEDIATE AIM Trial, a prospective NIH-funded study conducted from 2002 - 2005. The ECG recordings were begun within minutes (median, 45 minutes) from patients' presentation to the UCSF Emergency Department with ACS symptoms. HRV and HRT were computed using research software (HRV Laboratory, St. Louis, MO). A total of 193 patients had recordings with sufficient data for analysis.

Multivariate analyses demonstrated that 5 HRV measurements were significantly associated with all-cause 1-year rehospitalization (p-values<.05). HRT was associated with cardiac-related 30-day and 1-year ED readmission and/or death. Patients with disrupted HRT markers had greater risk of adverse 30-day outcomes compared to patients with healthy markers (hazard ratio=3.6, p=.01), and greater risk of adverse 1-year outcomes (hazard ratio=3.2, p<.001). Comparison of HRT and routinely-measured TIMI risk scores showed similar sensitivity and specificity for 30-day (areas under the curve: HRT=0.647; TIMI=0.687) and 1-year outcomes (HRT=0.657; TIMI=0.605).

In conclusion, HRT provides a viable alternative to TIMI scores for risk assessment in patients with ACS. With appropriate software, HRT and HRV could be obtained from the ECG monitoring nurses routinely perform. Future studies are needed to determine whether ANS ECG markers identify patients who warrant closer follow-up.

Myocardial ischemia is dynamic and unpredictable which, if not identified and treated promptly, can lead to devastating outcomes. Pre-hospital electrocardiography may facilitate early intervention by identifying patients with acute coronary syndrome (ACS) promptly. The goal of this dissertation research was to determine the prevalence and clinical significance of electrocardiographic (ECG) abnormalities of ischemia that occur in the pre-hospital setting in patients who contact `911' for chest pain and other anginal equivalent symptoms. While previous investigations have examined the impact of a snapshot 12-lead pre-hospital ECG (PH ECG), this study is the first to analyze ST-segment monitoring triggered 12-lead ECGs that have been acquired in the pre-hospital setting. The diagnostic and prognostic capabilities of PH ECG in the management of ACS have been examined and described.

The first paper is a critical literature review of the scientific literature about pre-hospital electrocardiography. Several themes that emerged from the literature and form the bases of this paper include: clinical effects of PH ECG (timing, roles, effects on patient outcome and characteristics of patients with PH ECG), diagnostic performance, and prognostic/predictive value of PH ECG.

The second paper is a secondary analysis of data from patients (n=620) who activated `911' for chest pain symptoms to determine the sensitivity and specificity of 12-lead ECG ST-segment monitoring in the pre-hospital period for diagnosing acute ACS. Both the PH ECG and the initial hospital ECG data for each patient were included in this study. The sensitivity of the ECG for diagnosing ACS was significantly improved when used in conjunction with the initial hospital ECG.

The third paper is a secondary analysis of data from the same cohort as previously described above, but includes only the PH ECG data for each patient (n=630). The purpose of the study was to determine whether manifestations of acute myocardial ischemia on the PH ECG are predictive of adverse hospital outcomes. Patients with evidence of PH ECG ischemia had a significantly higher proportion of adverse hospital outcomes than those without PH ECG ischemia. Furthermore, PH ECG ischemia was found to be an independent predictor of adverse hospital outcomes.

Abstract

Determining the Ideal Electrode Configuration for Continuous In-Hospital Electrocardiographic (ECG) Monitoring

Richard L. Fidler, PhD(c), MSN, MBA, CRNA, ANP

Significance: Hospital ECG-monitoring is done using the Mason-Likar electrode configuration with chest-mounted and newer technology allows the addition of precordial electrodes to the bedside monitor to acquire a 12-lead ECG. Mason-Likar limb electrodes need to move to the limbs for a standard 12-lead ECG; however, if this step is missed nonstandard and nonequivalent ECG is obtained. The Lund electrode configuration, with more distal limb electrodes was proposed as a solution, but it is unknown how Lund and Mason-Likar compare in signal quality, false lethal arrhythmia alarms, and patient comfort.

Methods: One hundred patients from ICU and PCU were enrolled, and in addition to standard hospital monitoring equipment, each subject wore two Holter monitors, one in the Mason-Likar and the other in the Lund electrode configurations for a 24-hour period. Randomization to abrasive skin prep was conducted. ECG signals were sent for blinded analysis for signal quality using the Hook-Up Advisor® and arrhythmia analysis using EK-Pro®. Signal quality was rated as "green-yellow-red", and lethal arrhythmia alarms were categorized as true or false by clinicians. Qualitative patient data regarding the monitoring experience was also gathered.

Results: Subjects each provided a mean of 23.8-hours of data in both electrode configurations, and 45 subjects received abrasive skin preparation. Signal quality was compared between configurations using a paired t-test showing that the Mason-Likar configuration spent 8.2% more time in "green". There was no differences between electrode configurations in the numbers of false lethal arrhythmia alarms. Abrasive skin preparation did not confer a benefit in signal quality or false lethal arrhythmia alarms. Patients prefer options to carry monitoring equipment. Hairy patients prefer to be shaved to reduce pain at electrode removal.

Implications: There is a difference favoring the Mason-Likar configuration over Lund for mean ECG signal quality, and there is no difference in false lethal arrhythmia alarms. Mason-Likar should remain the choice for continuous in-hospital ECG monitoring. Skin preparation conferred no benefit in signal quality or false lethal arrhythmia detection.