Monophasic action potential amplitude for substrate mapping

Abstract

Although radiofrequency ablation has revolutionized the management of tachyarrhythmias, the rate of arrhythmia recurrence is a large drawback. Successful substrate identification is paramount to abolishing arrhythmia, and bipolar voltage electrogram’s narrow field of view can be further reduced for increased sensitivity. In this report, we perform cardiac mapping with monophasic action potential (MAP) amplitude. We hypothesize that MAP amplitude (MAPA) will provide more accurate infarct sizes than other mapping modalities via increased sensitivity to distinguish healthy myocardium from scar tissue. Using the left coronary artery ligation Sprague-Dawley rat model of ischemic heart failure, we investigate the accuracy of in vivo ventricular epicardial maps derived from MAPA, MAP duration to 90% repolarization (MAPD₉₀), unipolar voltage amplitude (UVA), and bipolar voltage amplitude (BVA) compared with gold standard histopathological measurement of infarct size. Numerical analysis reveals discrimination of healthy myocardium versus scar tissue using MAPD₉₀ (P = 0.0158) and UVA (P < 0.001, n = 21). MAPA and BVA decreased between healthy and border tissue (P = 0.0218 and 0.0015, respectively) and border and scar tissue (P = 0.0037 and 0.0094, respectively). Contrary to our hypothesis, BVA mapping performed most accurately regarding quantifying infarct size. MAPA mapping may have high spatial resolution for myocardial tissue characterization but was quantitatively less accurate than other mapping methods at determining infarct size. BVA mapping’s superior utility has been reinforced, supporting its use in translational research and clinical electrophysiology laboratories. MAPA may hold potential value for precisely distinguishing healthy myocardium, border zone, and scar tissue in diseases of disseminated fibrosis such as atrial fibrillation.

NEW & NOTEWORTHY Monophasic action potential mapping in a clinically relevant model of heart failure with potential implications for atrial fibrillation management.

INTRODUCTION

Catheter-based radiofrequency ablation premiered in the late 1980s, resulting in new management options for complex arrhythmias such as atrioventricular nodal reentrant tachycardia, atrial fibrillation, and ventricular tachycardia. Whereas the value of radiofrequency ablation is evolving (4), shortcomings do exist. Of great importance is the arrhythmia recurrence rate, which repeatedly exposes patients to risks such as tamponade and permanent pacemaker placement during or after repeat ablation procedures.

Bipolar voltage electrograms are optimal in the context of ablation in that the field of view is sufficiently narrow to eliminate far-field noise and provide an accurate depiction of the local electric potential. However, in unique diseases such as atrial fibrillation that contain diffuse interstitial fibrosis (3), bipolar voltage electrograms may not be appropriate. Atrial fibrillation is a formidable disease because of the major quality-of-life challenges of thrombosis and stroke. These challenges are not always adequately mitigated with current radiofrequency ablation techniques. Reducing the recurrence rate for atrial fibrillation patients from 40% (9) to a more acceptable value will likely require techniques capable of highly sensitive identification of myocardial areas with electrical instability, which includes but is not limited to a decrease in tissue voltage.

We hypothesized that monophasic action potential (MAP) amplitude (MAPA) mapping would provide more accurate infarct sizes, compared with other mapping modalities, via increased sensitivity to distinguish healthy myocardium from scar tissue. Advantages of MAPA mapping include its inherent qualities relating to uniformly small field-of-view (2, 8), electrode-orientation insensitivity (12), and high-fidelity correlation to intracellular transmembrane action potentials. In this study, we use an in vivo animal model to demonstrate clinical feasibility.

METHODS

Ischemic heart failure model.

All rats in this study received humane care in compliance with Institutional Animal Care and Use Committee-approved protocols at the University of Arizona and in compliance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals. The male Sprague-Dawley rat (Envigo, Indianapolis, IN) permanent left coronary artery ligation model was used as previously described (10, 16, 19).

Rats are inducted with 3% isoflurane; anesthetized with a ketamine, xylazine, acepromazine, atropine, and saline cocktail; volume loaded intraperitoneally; undergo a medial sternotomy; and receive a permanent proximal left coronary artery ligature. Sham-operated surgical control rats receive the same procedure, excluding coronary ligation. Postsurgery, rats are maintained on standard light-dark cycles, rat chow, and water ad libitum for 6 wk while ischemic heart failure (HF) develops.

At 6 wk, echocardiography is used to evaluate left ventricular (LV) function (ejection fraction, end-diastolic pressure, peak developed pressure, ±dP/dt, and time constant of relaxation), and a terminal invasive hemodynamic and electrophysiology (EP) study is performed. During the terminal study, the rat is anesthetized, intubated, and mechanically ventilated on a rectal probe-gated heated platform with electrocardiogram monitoring (10, 19). A 3-Fr solid-state micromanometer-tipped catheter is equilibrated and inserted in the carotid artery and advanced through the aortic valve to collect left ventricular hemodynamic parameters. Mapping is then performed using a custom software system by collecting data from roughly equidistant points in matrix format from the epicardial surface and interpolating the numerical data into color maps.

MAP acquisition.

In vivo MAPs were obtained using a tungsten concentric bipolar microelectrode (instrument no. TM53CCINS; World Precision Instruments, Sarasota, FL). This superficially penetrating microelectrode has a cone-shaped tip with a tip diameter of 3 μm, a tip electrode length of 250 μm, an electrode spacing of 300 μm, a ring electrode length of 200 μm, and an impedance of 15 kΩ. All EP signals were amplified with a BIOPAC MP150 data acquisition system and MCE100C amplifier modules (BIOPAC Systems, Goleta, CA). Filtering settings for MAPs were as follows: 0.05-Hz high pass, 95-Hz low pass, and 60-Hz notch filter.

One-second intramural recordings (sedated rat heart rate is ~400 beats/min) were obtained from 24 epicardial points within a matrix of 6 columns, each containing 4 rows; the matrix dimensions approximated 10 mm for each row in the y-axis and 15 mm for each column in the x-axis. Analysis included selecting three representative ventricular MAPs that contained a stable baseline consistent with the two adjacent action potentials and followed the general principles describing MAP phases (14). Amplitude is defined as the millivoltage difference between phase 4 and the peak of phase 0. Ninety percent repolarization was selected for analysis, as opposed to a different percentage, to capture the well-described phenomenon of impaired late ventricular cardiomyocyte repolarization in ischemic HF (7, 13).

Voltage electrogram acquisition.

One-second voltage electrograms were obtained at the same ventricular epicardial MAP points with a clinical 4-Fr quadripolar catheter (Bard Electrophysiology, Covington, GA). The four electrodes are 1-mm in length, each separated by 1 mm. Only two electrodes were used for data collection to ensure good contact with the ventricular epicardium. Analysis included filtering (0.05-Hz high pass, 240-Hz low pass, and 60-Hz notch filter) and selecting three representative tracings that contained a stable baseline consistent in the two adjacent tracings and followed the general principles in describing unipolar and bipolar electrogram waves. Amplitude is defined as peak-to-peak voltage, more specifically either Q-R, R-S, isoelectric line-R, or isoelectric line-QS.

Electrophysiological mapping.

Whereas the primary area of interest is the ventricular scar and border region, the healthy ventricular myocardium is equally important, particularly the nonlinear interface of healthy and scar tissue. In this left coronary ligation model, gross scar boundaries can be distinguished on the anterior left ventricular myocardium with the naked eye; this helped ensure that healthy myocardium, adjacent border tissue, and scar tissue were mapped. MAPA, MAP duration to 90% repolarization (MAPD₉₀), unipolar voltage electrogram amplitude (UVA), and bipolar voltage electrogram amplitude (BVA) values were assigned a color using a clinically relevant color bar scheme.

The true distinction between normal and abnormal EP values for all mapping modalities was determined using a normal distribution cutoff, specifically the ±2SD. The sign of 2SD deviation depended on the parameter in analysis. For example, with chronic ischemia, compromised cardiomyocyte action potential durations are expected to increase (+2SD; see Refs. 7 and 13), whereas compromised cardiomyocyte action potential amplitudes are expected to decrease (−2SD). Border values were calculated to be 10% of the abnormal values adjacent to the normal-abnormal cutoff. All matrix values were then interpolated into a color map representing the epicardial surface spanning heathy myocardium to scar tissue.

Percentage of infarcted ventricular myocardium was extrapolated from color maps by comparing the two-dimensional surface area of scar tissue with the total area of the color map, which included healthy myocardium and border zone.

The same ventricular epicardial area was mapped in every rat, within a negligible error rate, because a single surgeon performed all median sternotomies, creating a surgical window of identical size in every rat. Furthermore, as mentioned before, the border of scar tissue is visible to the naked eye and was used to guide the placement of graded surgical marker boundaries.

Ex vivo histopathological infarct sizing.

After hemodynamic and EP evaluation, all rats were euthanized via a 3-ml potassium chloride intracavitary injection. The heart was immediately excised from the mediastinum, flushed intracavitary with heparinized saline, formalin fixed at 100 mmHg in a Langendorff preparation, and stored for 24 h in a glass jar. The fixed heart was then primed by removing nonleft ventricular tissue, and the remaining left ventricle was sent to an institutional core laboratory (Tissue Acquisition and Cellular/Molecular Analysis Shared Resource, University of Arizona Cancer Center) for staining and embedding. Left ventricle samples were sectioned by applying two equidistant cuts creating three portions (apex, mid, and base) and then stained with Masson’s trichrome to distinguish between healthy myocardium and scar tissue for histopathological circumferential infarct sizing (15). Only the midsection was used for infarct sizing.

This circumference-based method was used, as opposed to an area-based method, to best approximate the two-dimensional surface area evaluated by the EP mapping modalities. An area-based method to size an infarct includes not only the epicardial scar tissue but also intramyocardial scar tissue and endocardial scar tissue for a three-dimensional volume value (example: arbitrary units³). The circumference-based method used in this study employed an average of epicardial and endocardial scar length based on previously published techniques (15) to yield a two-dimensional surface area value (example: arbitrary units²).

Statistical analysis.

Data are presented as group means ± SE. Statistical significance is defined as a P < 0.05. Unpaired two-tailed Student’s t-tests were used to compare the location means (healthy, border, scar) within the two groups (Sham and HF) and between groups.

In assessing the accuracy between EP color map-derived infarct size and histopathological infarct size, average percent difference was used. Percent difference is defined as the absolute value of the difference between the experimental color map value and the histology value, divided by the average of the two numbers, all multiplied by 100. Related-Samples Friedman’s Two-Way ANOVA was performed on the percent difference values to determine statistical significance in mapping accuracy.

RESULTS

Ischemic HF model.

Induction of HF in infarcted rats was confirmed by echocardiography and invasive hemodynamic measurements at 6 wk post-MI (n = 6 Sham; n = 11 HF). The LV ejection fraction decreased from 74 ± 3 to 39 ± 4% (P < 0.001), LV peak-developed pressure fell from 189 ± 4 to 140 ± 12 mmHg (P = 0.0104), LV +dP/dt decreased from 7,853 ± 417 to 5,360 ± 393 mmHg/s (P = 0.0011), and time constant of LV relaxation increased from 20 ± 1 to 28 ± 3 ms (P = 0.0752). LV end-diastolic pressure increased from 7 ± 1 to 19 ± 4 mmHg (P = 0.0470) and LV −dP/dt increased from −6,864 ± 343 to −3,981 ± 490 mmHg/s (P = 0.0011).

MAP acquisition.

MAPs obtained (Fig. 1) exhibited expected morphology for rat ventricular cardiomyocytes (14). Approximately 7% of MAPs obtained from dense scar were unusable, the highest percentage from any tissue type. For Sham rats (n = 6), MAPA and MAPD₉₀ did not differ in any of the three zones, namely healthy, border, and scar (S) (Fig. 2). However, in HF rats (n = 21), MAPA decreased between healthy and border (P = 0.0015) and between border and scar (P = 0.0094), whereas MAPD₉₀ only increased between the healthy and border (P = 0.0158).

Fig. 1.

Monophasic action potential (MAP) tracings. All MAP tracings are from a single heart failure rat’s ventricular epicardium. The x-axis is time, in s; the y-axis is amplitude, in mV. The highest amplitude (gray) tracing represents healthy myocardium (H). The intermediate-amplitude (yellow) tracing represents tissue border (B) to scar tissue (S), depicted in red.

Fig. 2.

Monophasic action potential (MAP) data. A: an ex vivo heart failure (HF) heart with an example electrophysiology (EP) color map of 4 rows and 6 columns. Each black dot represents a mapped point, and each dot’s label is hypothetical. B: MAP data for Sham (n = 6) and HF (n = 21). The MAP parameters are amplitude, in mV, and duration at 90% repolarization, in ms. Values are reported as means + SE. Significance (t-test, P < 0.05) vs. scar (*) and border (#).

Voltage electrogram acquisition.

Voltage tracings were of high resolution (Fig. 3) and followed expected wave morphology for both unipolar and bipolar. Approximately 1% of MAPs obtained from dense scar were unusable, the highest percentage from any tissue type. For Sham rats (n = 6), UVA and BVA did not differ in any of the three zones (Fig. 4). However, in HF rats (n = 21), UVA and BVA both decreased in the infarcted myocardium. UVA decreased between healthy and scar (P < 0.001), whereas BVA decreased between healthy and border (P = 0.0218) as well as between border and scar (P = 0.0037).

Fig. 3.

Unipolar and bipolar voltage electrogram tracings. All unipolar (A) and bipolar (B) voltage electrogram tracings are from a single heart failure (HF) rat’s ventricular epicardium. The x-axis is time in s; the y-axis is amplitude in mV. The highest amplitude (gray) tracing represents healthy tissue (H). The intermediate-amplitude (yellow) tracing represents tissue border (B) to scar (S), depicted in red.

Fig. 4.

Voltage electrogram data. See Fig. 2A for reference. Unipolar and bipolar voltage electrogram data for Sham (n = 6) and heart failure (n = 21). The electrogram parameter is voltage amplitude, in mV. HF, heart failure. Values are reported as means + SE. Significance (t-test, P < 0.05) vs. scar (*) and border (#).

Electrophysiological mapping.

Of the total 21 HF rats, six representative rats were selected for color map infarct sizing comparison because of superior histopathology specimen quality (Table 1 and Fig. 5). Using Masson’s trichrome histopathology stain as the gold standard, BVA was found to be the most accurate method of approximating the size of the infarct region followed by UVA and then MAPA (Table 1). Statistical significance distinguished MAPD₉₀ from BVA (P = 0.010). Qualitative assessment supported the notion of MAPA’s superior spatial resolution (Figs. 5 and 6), despite MAPA not being the most accurate mapping method for determining infarct size.

Table 1.

Mapping vs. histopathology infarct sizing

HF Rat No.	MAPA, au2	MAPD90, au2	UVA, au2	BVA, au2	Histology, au2
1	0.29	0.06	0.61	0.62	0.54
2	0.39	0.05	0.49	0.47	0.28
3	0.32	0.30	0.49	0.50	0.47
4	0.27	0.12	0.65	0.63	0.50
5	0.31	0.19	0.52	0.46	0.38
6	0.42	0.03	0.56	0.55	0.39
Average difference, %	36	117*	27	24	0

Comparison of 6 rats’ infarct size derived from different methods of electrophysiology (EP) mapping, namely monophasic action potential (MAP) amplitude (MAPA), MAP duration to 90% repolarization (MAPD₉₀), unipolar voltage electrogram amplitude (UVA), and bipolar voltage electrogram amplitude (BVA) vs. the gold standard (histopathology). au, Arbitrary units; HF, heart failure. BVA is the most consistently accurate in approximating histopathological infarct size.

^*Significance vs. BVA (related-samples Friedman’s 2-way ANOVA, P = 0.010).

Fig. 5.

Color maps and histopathology sections. Images for the same 6 rats highlighted in Table 1. Color maps generated from the 4 different methods of electrophysiology (EP) mapping, namely monophasic action potential amplitude (MAPA), monophasic action potential duration to 90% repolarization (MAPD₉₀), unipolar voltage amplitude (UVA), and bipolar voltage amplitude (BVA), with a single corresponding histopathological section (mid). At the top of each column, the color bar for each mapping modality can be found, with black bars approximating the value threshold between each tissue type. For orientation, asterisks on histopathology slices denote where the right ventricle was removed. The broken lines on the histopathology sections approximate the mapped epicardium and the portion of histopathology that was included in the infarct sizing (black stars). Whereas most maps contain scar tissue on the right side, MAPD₉₀ maps have obvious flaws, suggesting an incapability to determine infarcted tissue in vivo.

Fig. 6.

Mapping percent scar versus histopathology percent scar. Monophasic action potential amplitude (MAPA, A), monophasic action potential duration to 90% repolarization (MAPD₉₀, B), unipolar voltage amplitude (UVA, C), and bipolar voltage amplitude (BVA, D) maps are depicted for a single heart failure (HF) rat. E: histopathology slides relative to the specific mapped epicardium of the same HF rat. For orientation, asterisks on histopathology sections denote where the right ventricle was removed. The ovals (A and E) suggest MAPA’s high spatial resolution, distinguishing adjacent healthy and scar tissue (×4 magnification). The red oval highlights epicardial collagen insulation; the black dotted oval highlights diffuse scar infiltration; the double yellow oval highlights normal myocardium.

DISCUSSION

This report defends the utility of MAPA as a mapping parameter and reinforces its robustness in in vivo cardiac EP applications. Although not as useful as BVA for quantifying scar tissue, these qualitative data (Figs. 5 and 6) may suggest that MAPA mapping attains a high degree of spatial resolution for distinguishing healthy, partially impaired, and grossly impaired cardiac tissue. We believe this may be because of the intrinsic qualities of MAPs relating to narrow field-of-view, electrode-orientation insensitivity, and ultrasensitivity to aberrations. These findings may hold potential value for clinical procedures relating to high-fidelity mapping (18) [ex.: accessory pathway ablation (1), rotor termination (11)].

Whereas MAPA’s ultrasensitivity to impaired depolarization asserts its utility in carefully dissecting substrate with respect to membrane potential, it also limits MAPA’s clinical value with respect to infarct sizing. The data suggest that MAPA attains high spatial resolution (Fig. 5), yet MAPA fails to most accurately depict infarct size (Table 1). This finding indicates that MAPA’s sensitivity to “electrical infarct” may render it relatively impervious to “histopathological infarct,” since there is work supporting the notion that the two are not one-in-the-same (20).

Perhaps BVA provides the most accurate depiction of infarct size by averaging the optimal number of cardiomyocytes’ electrical potentials to detect macroirregularities, which have a stronger correlation to histopathological infarct rather than the microirregularities of MAPA. While detecting ultrastructural subtleties with MAPA may be important in specific clinical cases such as atrial fibrillation, overall, it seems to provide too sensitive of data, and currently there are no approved clinical therapies that are evaluated via halting or reversing adverse remodeling at this microscopic tissue level.

In Table 1, including the border zone surface area in the calculation of infarct size decreases the average percent difference of MAPA to levels comparable to BVA, suggesting that any deviation from normal-phase zero-one activity has a stronger correlation to histopathological scar. While using the negative second deviation as a cutoff for MAPA seemed to underestimate the histopathological infarct size, the same cutoff value was used for BVA, which produced overestimates (Table 1). We suspect that this is because of the previously mentioned qualities of MAPA and BVA relating to field-of-view: BVA averages the electrical potential from a larger population of cardiomyocytes and can overlook general anatomic variants in voltage such as endocardial trabeculations while still achieving sufficient sensitivity to distinguish healthy myocardium from border zone and border zone from scar tissue (Fig. 4).

Action potential duration is a common parameter used to measure cardiomyocyte damage from ischemia (6). We have shown that MAPA has a greater ability than MAPD₉₀ to distinguish the infarcted myocardium tissue types (Fig. 2). The inferior ability of MAPD₉₀ to accurately quantify infarct size may relate to the excessive finesse of MAPs. These recordings are particularly susceptible to artifact given the small margin of repolarization voltages and the stepwise return to isoelectric potential that is easily affected by operator-mediated contact error or minor cell damage. Furthermore, improper electrode contact or epicardial collagen insulation may yield inappropriately small MAPs, which would affect the MAPD₉₀ to a greater extent than the MAPA by depicting nearly normal repolarization times even in the face of overt impaired repolarization. This theory may have been exhibited in this study, leading to MAPD₉₀ color maps depicting healthy myocardium where all other mapping modalities depict scar tissue (Figs. 5 and 6B).

UVA and BVA are both highly used in the clinical EP laboratory, particularly bipolar voltage electrograms for defining substrate. We included these parameters in our study (Fig. 4) to assess the potential clinical application of creating myocardial maps with MAPA. Statistical analysis reveals that MAPA maps are inferior to both UVA and BVA maps when quantifying histopathological infarct size (Table 1), suggesting that quantifying scar tissue quantity and creating high-resolution epicardial maps are distinct tasks that warrant individualized techniques.

Limitations.

In this body of work, we use in vivo cardiac EP parameters to estimate the location and amount of histopathological scar. Due to the fact that perturbations in action potentials or electrograms do not unequivocally correspond with scar tissue, we acknowledge that an EP map-derived percentage scar value could be identical to a histopathological percentage scar value but be unrelated with respect to anatomic location. However, we contend that this phenomenon is not confounding the results of this study, particularly because of the normal distribution cutoff criteria.

MAPs, particularly the third phase, have high sensitivity to electrode motion and contact pressure artifacts (5) and therefore are more likely to exhibit atypical morphology that may obscure the landmarks necessary for analysis. Nonetheless, MAPs have a high-fidelity correlation to intracellular transmembrane action potentials (17), and the methods for training the surgeon-electrode handler, acquiring the data with immediate review, and selecting representative tracings during postacquisition processing in our study attempted to minimize artifacts by seeking consistent electrical activity. The fact that only 7% of MAPs from dense scar were unusable highlights our approach’s reproducibility.

Scar tissue is a nonuniform three-dimensional structure that may not be fully encapsulated by a two-dimensional histopathology approximation (example: arbitrary units²) such as the one employed in this study. Nonetheless, our work attempted to depict the epicardial layer of the healthy tissue-scar tissue interface in two dimensions and not the entire three-dimensional scar structure. We used a circumference-based histopathology infarct sizing methodology to simplify the scar to transmural thickness.

Although this study is posed as LV mapping, a portion of the color maps involved the right ventricle-septum border, which served as healthy myocardium. This was done to create a uniform area of mapping in all rats, since some infarcted rats have extensive anterior LV scar consequent to efficacious coronary ligation. We did not object to involving a portion of the right ventricle-septal border because the thickness difference between the right ventricle (RV) and LV in rats is less pronounced compared with humans’ RV-LV thickness difference, mainly for two reasons as follows: 1) substantially smaller body habitus and 2) quadrupedal stature, lessening the antagonism of gravity on systemic circulation.

Finally, a different method to define infarct size is cardiac magnetic resonance imaging. We acknowledge that cardiac magnetic resonance imaging-based computations would have served as a more clinically relevant gold standard method to quantify the infarct size in vivo, as opposed to the histopathology we used in this work.

Conclusion.

We describe the creation of electrophysiological color maps using the ventricular epicardial parameters of MAPA, MAPD₉₀, UVA, and BVA. To our knowledge, this is the first attempt to use MAPA for two-dimensional mapping in an in vivo animal model and is also the first to compare four different mapping modalities simultaneously to assess infarct sizing and spatial resolution.

We hypothesized that MAPA mapping would provide more accurate infarct sizes via increased sensitivity to distinguish between healthy myocardium and scar tissue compared with the other mapping modalities. Using a rat model of ischemic HF, we quantified the percentage of infarcted myocardium from each two-dimensional map and determined that BVA maps create the most accurate depiction of percentage of infarcted myocardium, relative to histopathology. Although MAPA is not of the same power with respect to depicting infarct size, it may be useful for characterizing convoluted ischemic myocardial pathways in atrial fibrillation.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant HL-007249-43, The WARMER Research Foundation, Sarver Heart Center, University of Arizona, and The Martin and Carol Reid Charitable Remainder Trust.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

I.R.C., J.L., S.G., and E.J. conceived and designed research; I.R.C. performed experiments; I.R.C. and M.H. analyzed data; I.R.C., M.H., T.M., and S.G. interpreted results of experiments; I.R.C. prepared figures; I.R.C. drafted manuscript; I.R.C., M.H., and S.G. edited and revised manuscript; I.R.C., T.M., J.L., S.G., and E.J. approved final version of manuscript.