A continuous recording of the ECG obtained with a Holter monitor over a 24-hour period is well established for the detection and quantification of intermittent arrhythmias and monitoring the response to antiarrhythmic medications in dogs.1–4 Holter recordings can also provide information about heart rate variability, a theoretical concept first introduced in 1905 by Henri Poincaré through the use of mathematical and physical postulates supporting nonlinear dynamics5 and expanded by Edward Lorenz in 1963.6 Lorenz or Poincaré plots are one of several ways to graphically depict successive R-R intervals by plotting one R-R interval against the next R-R interval, thereby creating a scatterplot. These R-R interval scatterplots are useful to examine the relationship between consecutive heartbeats and provide information about the influences of sympathetic tone and parasympathetic tone on the cardiac rhythm.7–9 The visual patterns of these scatterplots have also been shown to be predictive of rhythm instability in some situations.10,11 Quantitative analysis of portions of LPs associated with a sinus rhythm has been shown to be reflective of sympathovagal input, and these portions are useful as a time-domain measure of heart rate variability.9,12 Both short-term (3-minute to 1-hour) and long-term (24-hour) ECG recordings of cardiac rhythms in people have been used to generate LPs for automated arrhythmia detection and classification.9,12–16 Applicability of LPs in veterinary medicine for these purposes remains to be fully explored.
The visual depiction of cardiac rhythms with LPs allows for a unique assessment of a composite of R-R intervals over a specific period. Patterns in LPs have been associated with various rhythm disturbances in people.14,15 However, only a few reports8,12,13,17,18 involving dogs and horses have mentioned the usefulness of LPs to identify patterns associated with cardiac rhythms, and patterns associated with various arrhythmias have not yet been described for these species. One example of the usefulness of LP pattern identification in human medicine is the observation of a double-sector LP (LP with 2 offset fan patterns) for some people with AF, which contributed to the understanding of the mechanism of dual AV nodal pathways.19 This specific feature of LPs found for some people with AF, suggesting 2 pathways, provides information about the properties of the AV node's functional refractory period, which may be modifiable by medications.20 Identifiable patterns, suggesting nonrandom causes, may contribute to understanding the mechanisms by which naturally occurring arrhythmias develop in dogs, a species that is commonly affected by clinically important and translatable cardiac diseases. The aim of the study reported here was to characterize the patterns identified in LPs derived from the Holter recordings of dogs with and without common arrhythmias. We speculated that LP patterns associated with common cardiac rhythms for dogs would be similar to those found with common cardiac rhythms for people.
Materials and Methods
The medical records of dogs with 24-hour Holter recordings performed between 2012 and 2018 through the Veterinary Teaching Hospital of the College of Veterinary Medicine at North Carolina State University were searched for specific rhythm diagnoses to obtain target samples of 20 dogs with apparently normal cardiac rhythms (arrhythmia-free dogs) and 60 dogs with arrhythmias. Among the latter group of dogs, 6 groups with a target of 10 dogs each with 1 of 6 arrhythmias as the predominant arrhythmia on the recording were sought. Signalment, treatment with antiarrhythmic medications, and arrhythmia frequency during a selected 1-hour period and over the 24-hour period for each Holter recording were noted for each dog.
Holter recording analysis
All Holter recordings were acquired and analyzed with a commercially available system.a This system was developed to enhance the interpretation of Holter recordings obtained from dogs, for which their Holter recordings can be challenging to interpret because of pronounced sinus arrhythmias and a wide range of heart rates. Recordings were of good quality (ie, > 90% interpretable by visual inspection), and only 1 Holter recording/dog was included. For dogs with abnormal cardiac rhythms, Holter recordings were excluded if a predominant type of arrhythmia could not be identified, often because many types of arrhythmias occurred in a given dog. Low numbers of concurrent nonpredominant arrhythmias, however, were acceptable for all dogs. Holter recordings were reviewed for accuracy of rhythm diagnosis by a single cardiology technician and one of the authors (DBA). Beat coding by the system was corrected if necessary. Activity diaries for the dogs were not available for review.
Cardiac rhythm categorization
Rhythm diagnoses were categorized as arrhythmia free, single SVPCs, CSVE, single VPCs, CVE, AF, and high-grade second-degree AVB. Holter recordings were categorized into one of these groups on the basis of the presence or absence of a predominant arrhythmia and other specific criteria. Holter recordings from dogs that were not receiving antiarrhythmic medications were defined as arrhythmia free when a sinus rhythm or sinus arrhythmia was present with < 25 premature complexes and without pauses > 3 seconds during the 24-hour recording period.21,22 Holter recordings of dogs with single SVPCs and VPCs were included in those respective groups if > 50 premature complexes were identified in the 24-hour recording period. Holter recordings of dogs with CSVE and CVE were included in those respective groups when any complex ectopy was present. Holter recordings of dogs with AF and without a sinus rhythm identified in the 24-hour recording period were included in the AF group. Holter recordings of dogs with high-grade second-degree AVB were included in that group when an AV nodal conduction ratio of > 2:1 was present.
LP generation and pattern classification
The Holter recording systema generated LPs for a 1-hour segment of the 24-hour Holter recording with the tachogram strip function. This system cannot generate LPs in segments longer than 1 hour. Lorenz plots for Holter recordings obtained from arrhythmia-free dogs were generated at the 4-hour time point to correspond with daytime hours (Holter monitors were placed on dogs during the hospital's normal business hours), which also simultaneously allowed dogs to acclimate to wearing the monitor, and at the 12-hour time point to correspond with nighttime hours. For LP generation, abnormal Holter recordings were searched for the 1 hour of the recorded 24 hours that best represented the predominant arrhythmia type.
Each dot of an LP represented the R-R intervals surrounding a single heartbeat, such that the R-R interval (R-Rn; x-axis) was plotted against the next R-R interval (R-Rn+1; y-axis). The location of each LP dot was determined by the R-R interval immediately before and after a beat. Dots were color coded to reflect the origin of the electrical impulse (ie, sinoatrial node or elsewhere). With this system, the dots for normal beats were coded blue, those for SVPCs were coded green, and those for VPCs were coded pink. Beats associated with AF and high-grade second degree AVB were also coded blue. The scale for the x- and y-axes for all LPs was in increments of 0.5 seconds, to a limit of 3 seconds.
Complex supraventricular ectopy included supraventricular couplets and triplets and paroxysms of supraventricular tachycardia, and CVE included ventricular couplets and triplets and paroxysms of ventricular tachycardia. Because each LP dot represented 2 R-R intervals, premature beats associated with complex arrhythmias were uniquely located on the LP, compared with single premature beats (ie, lower left corner for complex ectopy vs a VSL for single premature beats). However, the type of CSVE defined by the number of ectopic beats (couplets, triplets, and paroxysms of supraventricular tachycardia) and the type of CVE defined by the number of ectopic beats (couplets, triplets, and paroxysms of ventricular tachycardia) could not be determined from the LPs because only 2 R-R intervals are plotted for each dot. Therefore, the Holter recordings of dogs with any type of CSVE were included in a single group for analysis, and the Holter recordings of dogs with any type of CVE were included in a single group for analysis. Patterns of LPs were determined on the basis of their overall geometric shape, as agreed by both authors, with the use of previously reported8,9,14,16 descriptors, when possible (Appendix).
Results
The 24-hour Holter recordings obtained from 20 arrhythmia-free dogs and 57 dogs with arrhythmias were retrieved from the medical records. Specifically, the Holter recordings from 10 dogs were identified for each arrhythmia of interest, with the exception of high-grade second-degree AVB, for which recordings from only 7 dogs were identified. The frequencies of each predominant arrhythmia varied among dogs (Table 1).
Median (95% CI) number of arrhythmias categorized by type of cardiac rhythm identified from 24-hour Holter recordings of 77 dogs (1 recording/dog) performed between 2012 and 2018.
| Cardiac rhythm group | No. of dogs | No. of single SVPCs/24 h | No. of single SVPCs/1 h (LP) | No. of CSVE events/24 h | No. of CSVE events/1 h (LP) | No. of single VPCs/24 h | No. of single VPCs/1 h (LP) | No. of CVE events/24 h | No. of CVE events/1 h (LP) |
|---|---|---|---|---|---|---|---|---|---|
| Arrhythmia free | 20 | 0 (0–1) | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0 (0–1) | 0 (0–0) | 0 (0–0) | 0 (0–0) |
| SVPC | 10 | 1,582 (92–16,868) | 146 (9–123) | 64 (8–3,809) | 7 (1–301) | 38 (0–93) | 1 (0–4) | 0 (0–19) | 0 (0–0) |
| CSVE | 10 | 969 (12–1,640) | 31 (11–299) | 686 (13–235) | 49 (11–614) | 179 (29–462) | 6 (0–75) | 3 (0–15) | 0 (0–0) |
| VPC | 10 | 2 (0–157) | 0 (0–0) | 0 (0–1) | 0 (0–0) | 4,843 (1,189–9,822) | 270 (65–618) | 0 (0–15) | 0 (0–1) |
| CVE | 10 | 1 (0–22) | 0 (0–1) | 0 (0–12) | 0 (0–1) | 4,298 (1,216–9,183) | 185 (74–558) | 304 (65–716) | 29 (3–392) |
| AF | 10 | 0 (0–0) | 0 (0–0) | 0 (0–48) | 0 (0–0) | 780 (4–6,040) | 2 (0–60) | 8 (0–912) | 0 (0–0) |
| High-grade second-degree AVB | 7 | 3 (0–15) | 0 (0–9) | 0 (0–38) | 0 (0–24) | 5 (0–689) | 0 (0–3) | 0 (0–8) | 0 (0–0) |
Numbers of arrhythmias are for the entire 24-hour period and the selected 1-hour period used for LP generation.
A Holter recording was categorized as arrhythmia free when a sinus rhythm or sinus arrhythmia was present with < 25 premature complexes and without pauses > 3 seconds during the 24-hour recording period.
LPs generated from Holter recordings of arrhythmia-free dogs
Holter recordings in the arrhythmia-free group were from dogs of various breeds, including Doberman Pinschers (n = 7), Boxers (7), Rhodesian Ridge-backs (2), and others (Shih Tzu, Cocker Spaniel, Norwegian Elkhound, and West Highland White Terrier [1 each]). Median age of the dogs was 5 years (range, 1 to 15 years). Among the 20 dogs, 14 were female (sexually intact, n = 11; spayed, 3) and 6 were male (sexually intact, 4; castrated, 2).
In this group, each Holter recording had a sinus rhythm or sinus arrhythmia that yielded LPs with blue-coded dots that formed a Y-shaped pattern similar to that previously reported8 for clinically normal dogs (Figure 1). Each LP generated from all of the Holter recordings had blue dots in a Y-shaped pattern at 12 hours; however, only 16 of the 20 LPs had a Y-shaped pattern at 4 hours. Instead, the other 4 recordings yielded LPs with blue dots in a comet-shaped pattern.
Representative LPs derived from 24-hour Holter recordings of arrhythmia-free dogs. The recording was considered normal when a sinus rhythm or sinus arrhythmia was present with < 25 premature complexes and without pauses > 3 seconds during the 24-hour recording period. Each dot represents 2 R-R intervals associated with 1 beat. Normal beats were coded blue. A—An LP (left) with a Y-shaped pattern and a central silent zone was derived from 1 hour of a 24-hour Holter recording of a dog with a sinus arrhythmia. Also shown are the 24-hour (top right) tachogram and selected 1-hour (middle right) tachogram from the 24-hour tachogram (denoted by the 2 vertical lines in the 24-hour tachogram), from which the LP was generated, and an ECG trace of a portion of the selected 1 hour obtained from the Holter recording. B—The recording of a dog with an infrequent sinus arrhythmia (and without any other arrhythmia), compared with the relative higher frequency of sinus arrhythmia in the dog in panel A, yielded an LP with a comet-shaped pattern. C—An LP with a Y-shaped pattern and a silent zone slightly shifted to the upper left of center derived from the recording of a dog with a sinus arrhythmia. D—An LP with a torpedo-shaped pattern from the recording of a dog with sinus rhythm and sinus tachycardia. The small VSL, denoted by the pink dots, indicates concurrence of few VPCs.
Citation: American Journal of Veterinary Research 81, 9; 10.2460/ajvr.81.9.720
Silent zones, indicated by an absolute or relative absence of blue dots that contributed to the Y-shaped pattern, were present in 19 of the 20 LPs generated from the recordings at 12 hours and in 15 of the 20 LPs generated from the recordings at 4 hours. The silent zones for 12 dogs at 12 hours were centrally located in the LPs (ie, paucity of dots for which R-Rn and R-Rn+1 were both approx 1.2 seconds; Figure 1). A second location for the silent zone was noted to be shifted slightly to the upper left of the center of the LPs (ie, R-Rn < R-Rn+1) for 7 dogs at 12 hours.
LPs generated from Holter recordings of dogs with single SVPCs
Holter recordings in the single SVPC group were obtained from various breeds including Doberman Pinschers (n = 2), Boxers (2), and others (Labrador Retriever, Rhodesian Ridgeback, Irish Wolfhound, Whippet, Yorkshire Terrier, and Chihuahua [1 each]). Median age of the dogs was 8 years (range, 2 to 17 years). Among the 10 dogs, 2 were female (sexually intact, n = 1; spayed, 1) and 8 were male (sexually intact, 2; castrated, 6). Two dogs had been receiving atenolol (n = 1) or digoxin (1) at the time of the recordings.
In this group, each LP consistently had a VSL of green dots parallel to the y-axis (ie, R-Rn < R-Rn+1) and separate from the main cluster of blue dots, indicative of SVPCs (Figure 2). A complementary HSL of blue dots parallel to the x-axis (ie, R-Rn > R-Rn+1) was also noted, representing the varying R-R intervals immediately prior to the premature beat with a consistent coupling interval. Both lobes created a DSL pattern. The intensity of the VSL was subjectively greater for recordings with high numbers of SVPCs, compared with that for recordings with low numbers of SVPCs, but the perceived difference in intensities could not be quantified. The HSL was more intense for recordings with supraventricular bigeminy because the HSL, indicative of couplets of long-short R-R intervals, mirrored the couplets of short-long R-R intervals of the VSL. A sinus rhythm or sinus arrhythmia that was concurrently recorded resulted in blue dots in Y- (n = 6), comet- (2), and torpedo-shaped (2) patterns (Figure 1).
Representative LPs derived from 24-hour Holter recordings of dogs with SVPCs. A—An LP with an intense VSL, created by green dots that are representative of SVPCs, derived from a 1-hour period of the 24-hour recording of a dog with 475 SVPCs in the hour shown. B—An LP with a less intense VSL than that depicted in panel A corresponds to the 1-hour period of the 24-hour recording of a dog with 11 SVPCs in the hour shown.
Citation: American Journal of Veterinary Research 81, 9; 10.2460/ajvr.81.9.720
LPs generated from Holter recordings of dogs with CSVE
Holter recordings in the CSVE group were obtained from the following types of dogs: Labrador Retrievers (n = 2); Shetland Sheepdog, English Bulldog, Rottweiler, Irish Setter, Cocker Spaniel, Bernese Mountain Dog, and Golden Retriever (1 each); and mixed-breed dog (1). Median age of the dogs was 8.5 years (range, 3 to 14 years). Among the 10 dogs, 5 were female (spayed) and 5 were male (sexually intact, n = 2; castrated, 3). Six dogs had been receiving diltiazem (n = 5) or a combination of atenolol and sotalol (1) at the time of the recordings.
In this group, each LP had a small cluster of green dots in the lower left corner, representing the short R-R intervals associated with CSVE (Figure 3). Blue dots, consistent with a normal sinus rhythm or sinus arrhythmia, yielded undiscernible (n = 4), Y- (3), comet- (2), and torpedo-shaped (1) patterns. The lack of a clear pattern was due to undercoded SVPCs, which were too numerous to manually correct.
Representative LP derived from a 1-hour period of a 24-hour Holter recording of a dog with CSVE. Also shown are the 24-hour (top right) tachogram and selected |-hour (middle right) tachogram from the 24-hour tachogram (denoted by 2 vertical lines in the 24-hour tachogram), from which the LP was generated and which best represented the predominant arrhythmia (CSVE), and an ECG trace of a portion of the selected 1 hour obtained from the Holter recording. The green dots in the lower left corner (e) of the LP indicate CSVE and correspond to consecutive couplets of short-short R-R intervals noted in the ECG trace. The dog also had supraventricular bigeminy, with a DSL (VSL [green dots {a}] and HSL [blue dots {b}]), single VPCs (few pink dots in the lower left corner of the LP), and a sinus rhythm (d). N = Sinus beat. S = SVPC.
Citation: American Journal of Veterinary Research 81, 9; 10.2460/ajvr.81.9.720
LPs generated from Holter recordings of dogs with single VPCs
Holter recordings in the VPC group were obtained from 1 mixed-breed dog and dogs of various breeds including Boxers (n = 3), Labrador Retrievers (2), and Airedale Terrier, Plott Hound, Yorkshire Terrier, Weimaraner (1 each). Median age of the dogs was 9 years (range, 4 to 12 years). Among the 10 dogs, 9 were female (sexually intact, n = 1; spayed, 8) and 1 was male (castrated). Four dogs had been receiving sotalol (n = 2), mexiletine (1), or a combination of sotalol and mexiletine (1) at the time of the recordings.
In this group, each LP had a VSL of pink dots, indicative of VPCs (Figure 4). A complementary HSL of varying length of blue dots and the VSL produced a DSL pattern, which corresponded to long-short R-R intervals consistent with ventricular bigeminy (6/10 recordings) or VPCs and pronounced sinus arrhythmia (4/10). The Ashman phenomenon was not evident. Three LPs each had 2 VSLs consistent with 2 distinct coupling intervals for VPCs, and 3 other LPs each had a TSL pattern consistent with 3 distinct interectopic intervals (normal R-R interval preceding a premature beat, single coupling interval associated with a premature beat, and postectopic pause after a premature beat). The intensity of the VSL was subjectively greater in the LPs derived from recordings with high numbers of VPCs, compared with that in the LPs derived from recordings with low numbers of VPCs, and the intensity of the HSL was subjectively greater with the presence of ventricular bigeminy. However, the perceived differences in intensities could not be quantified. The main cluster of blue dots was in a Y- (n = 5) or comet-shaped (5) pattern.
Representative LPs derived from 24-hour Holter recordings of dogs with single VPCs. Also shown for each dog are the 24-hour (top right) tachogram and the selected 1-hour (middle right) tachogram from the 24-hour tachogram (denoted by 2 vertical lines in the 24-hour tachogram), from which the LP was generated and which best represented the predominant arrhythmia (single VPCs), and an ECG trace of a portion of the selected 1 hour obtained from the Holter recording. A—An LP derived from a 1-hour period of the 24-hour recording of 1 dog has a VSL of pink dots (a), which corresponds to the short coupling intervals associated with VPCs, and a complementary HSL of blue dots (b), illustrating a DSL pattern. Side lobes are similar in intensity because of the presence of ventricular bigeminy, in which the VSL reflects short-long R-R intervals and the HSL reflects complementary long-short R-R intervals. The small group of pink dots in the lower left corner of the LP represents the short-short R-R intervals consistent with CVE. A Y-shaped pattern of blue dots is also present, consistent with variable R-R intervals of sinus rhythm and sinus arrhythmia (d). B—Three of the LPs had TSL patterns. The HSL (blue dots [b]) is less dense than the VSL (pink dots [a]) because ventricular bigeminy was uncommon. The third side lobe (blue dots [c]) reflects a constant compensatory pause after an ectopic beat, which was then followed by 2 sinus beats, yielding another group of blue dots (d). N = Sinus beat. V = VPC.
Citation: American Journal of Veterinary Research 81, 9; 10.2460/ajvr.81.9.720
LPs generated from Holter recordings of dogs with CVE
Holter recordings in the CVE group were obtained from 1 mixed-breed dog and dogs of various breeds including Boxers (n = 4) and Australian Shepherd, German Shepherd, Doberman Pinscher, Poodle, and English Bulldog (1 each). Median age of the dogs was 9 years (range, 6 to 15 years). Among the 10 dogs, 3 were female (sexually intact, n = 1; spayed, 2) and 7 were male (sexually intact, 2; castrated, 5). Six dogs had been receiving sotalol (n = 4), mexiletine (1), or atenolol (1) at the time of the recordings.
In this group, each LP had a small cluster of pink dots in the lower left corner, representative of the short-short R-R intervals of CVE (Figure 4). Occasionally, a DSL pattern that supported the concurrence of ventricular bigeminy was seen. Two LPs showed 2 VSLs consistent with 2 distinct coupling intervals for coexisting VPCs. The blue dots yielded Y- (n = 7), comet- (1), and torpedo-shaped (1) patterns, consistent with a normal sinus rhythm and sinus arrhythmia, and a fan (1) pattern, consistent with AF.14,15,19 The recording from the dog with AF was included in the CVE group because the presence of any CVE qualified a recording to be included in the CVE group (Figure 5).
An LP derived from a 1-hour period of a 24-hour Holter recording of a dog with CVE and AF. The small cluster of pink dots in the lower left corner indicates CVE. The fan pattern of blue dots indicates AF. Also shown are the 24-hour (top right) tachogram, the selected 1-hour (middle right) tachogram from the 24-hour tachogram from which the LP was generated, and an ECG trace of a portion of the selected 1 hour obtained from the Holter recording. See Figure 4 for remainder of key.
Citation: American Journal of Veterinary Research 81, 9; 10.2460/ajvr.81.9.720
LPs generated from Holter recordings of dogs with AF
Holter recordings in the AF group were from dogs of various breeds including Boxers (n = 2) and others (Irish Setter, Jack Russell Terrier, Great Dane, Whippet, Cavalier King Charles Spaniel, English Bulldog, German Shorthaired Pointer, and Rottweiler [1 each]). Median age of the dogs was 10 years (range, 5 to 13 years). Among the 10 dogs, 4 were female (sexually intact, n = 1; spayed, 3) and 6 were male (sexually intact, 2; castrated, 4). Dogs had been receiving diltiazem (n = 2), sotalol (1), or combinations of diltiazem and digoxin (5), sotalol and digoxin (1), and sotalol, diltiazem, and digoxin (1). Among the 10 dogs, the mean ± SD heart rate was 123 ± 27 beats/min (range, 71 to 163 beats/min).
In this group, each LP had a fan pattern (Figure 6). Heart rate was visually assessed by the location of the fan and distribution of the dots. High heart rates resulted in clustering of the dots in the lower left corner, and lower heart rates resulted in relative dispersion of the dots toward the upper right corner. A DSL was evident in 5 of 10 recordings, which indicated the concurrence of VPCs. No silent zones or double sectors were noted.
Representative LPs derived from 24-hour Holter recordings of dogs with AF and few VPCs (pink dots). A—For 1 dog with a mean heart rate of 123 beats/min, the fan pattern of blue dots is clustered in the lower left corner, indicative of variable but relatively short R-R intervals. B—For another dog with a mean heart rate of 71 beats/min, the blue dots of the fan pattern are more distributed, indicative of variable but relatively longer R-R intervals.
Citation: American Journal of Veterinary Research 81, 9; 10.2460/ajvr.81.9.720
LPs generated from Holter recordings of dogs with high-grade second-degree AVB
Holter recordings in the high-grade second-degree AVB group were obtained from various breeds including Beagle, Labrador Retriever, Australian Shepherd, Boxer, Havanese, Chow Chow, and Basenji (1 each). Median age of the dogs was 11 years (range, 5 to 13 years). Among the 7 dogs, 4 were female (spayed) and 3 were male (castrated). Two dogs had been receiving diltiazem (n = 1) and a combination of propantheline, hyoscyamine, or theophylline (1) at the time of the recordings.
In this group, each LP had an island pattern, 5 of which were round (Figure 7) and 2 of which were elliptical. The AVB was intermittent, and therefore, patterns consistent with a sinus rhythm or sinus arrhythmia (Y- [n = 3], comet- [2], and torpedo-shaped [2] patterns) were also evident.
Representative LP derived from a 1-hour period of a 24-hour Holter recording of a dog with high-grade second-degree AVB and few VPCs (pink dots). Also shown are the 24-hour (top right) tachogram and selected 1-hour (middle right) tachogram from the 24-hour tachogram (denoted by 2 vertical lines in the 24-hour tachogram), from which the LP was generated and which best represented the predominant arrhythmia (high-grade second-degree AVB), and an ECG trace of a portion of the selected 1 hour obtained from the Holter recording. Each cluster (island) of blue dots represents different AV conduction ratios (eg, 2:1 and 3:1). In the LP, atrial-to-ventricular conduction ratios are shown for each island, with the conduction ratio for the R-Rn interval shown first (top) and the conduction ratio for the R-Rn+1 interval shown second (bottom). N = Sinus beat.
Citation: American Journal of Veterinary Research 81, 9; 10.2460/ajvr.81.9.720
Discussion
In the present study, specific, consistent LP patterns were identified for common cardiac rhythms in dogs. The LP patterns identified for apparently normal rhythms were similar to those previously reported8 for dogs but different from those for people.14 The LP patterns identified for common arrhythmias were similar to those previously reported14–16 for people and those described for dogs in a few reports.12,13 The data were obtained with a Holter recording system, including software, specifically designed for dogs and that had been used for assessment of a Boxer dog with multiple episodes of syncope.13
Lorenz plots generated from most Holter recordings of arrhythmia-free dogs (sinus rhythm or sinus arrhythmia with < 25 premature complexes and without pauses > 3 seconds) showed a Y-shaped pattern, as has been previously reported8 for dogs. However, the LPs derived from Holter recordings of clinically normal people do not have Y-shaped patterns but rather have comet- and torpedo-shaped patterns.9,14,16 The Y-shaped pattern is a result of variable R-R intervals associated with sinus arrhythmia and relatively shorter R-R intervals associated with sinus rhythm and sinus tachycardia. This finding corroborated the strong influence of vagal tone on the heart rate and rhythm of dogs.8,23 Interestingly, the LPs derived from the Holter recordings of several arrhythmia-free dogs yielded Y-shaped patterns at the 12-hour time point and comet-shaped patterns at the 4-hour time point. The comet-shaped pattern indicated that a sinus arrhythmia infrequently occurred at the 4-hour time point, compared with its frequency at the 12-hour time point. These findings are consistent with sympathetic tone input and parasympathetic tone withdrawal during daytime hours (the 4-hour time point), when the dogs were more likely to be awake and active and, therefore, have regular sinus rhythms. Parasympathetic tone likely dominated during nighttime hours (the 12-hour time point), when the dogs were more likely to be sleeping and, therefore, have sinus arrhythmias. However, activity diaries were not available for review; thus, we could not confirm the relationship between the cardiac rhythm and activity.
A silent zone (area of absolute or relative absence of dots), which has been previously reported,8 was also noted in many LPs from the Holter recordings of arrhythmia-free dogs in the present study; a few LPs also had a second silent zone. The silent zone is the space between the arms of the “Y” of the Y-shaped pattern visualized in the LPs. The Y-shaped pattern indicated that successive long-short, short-long, short-short, and long-long R-R intervals produced dots that bordered but did not often enter the silent zone. The paucity of successive R-R intervals of 1.0 to 1.3 seconds, corresponding to the silent zone and an instantaneous heart rate of 40 to 60 beats/min, supported the generation of a nonrandom rhythm.13 The infrequency of data points within this central area does not necessarily mean that R-R intervals do not occur at these instantaneous rates, just that successive R-R intervals at these specific rates are uncommon. In dogs, fluctuations in sympathetic tone and parasympathetic tone that change the impulse origination from distinct areas of the sinoatrial node have been postulated to explain the observed clustering of dots surrounding these LP silent zones.8,13 Silent zones have not been noted in the LPs generated from the Holter recordings of clinically normal people but have been identified in the LPs generated from the Holter recordings of people with AF.14,20 The proposed mechanisms of silent zones in people with AF are increased parasympathetic tone, nonlinear conduction dynamics, dual AV nodal physiology, and the effects of negative chronotropic medications on the functional refractory period of the AV node.14,20 The Holter recordings of the arrhythmia-free dogs in our study, however, were not affected by negative chronotropic medications. The cause of the 2 distinct silent zones in the LPs generated from the Holter recordings of 7 of the 20 arrhythmia-free dogs is not known.
Color coding of beats with the software used in the present study aided visual recognition of the beat origin, supraventricular or ventricular, and facilitated the rapid identification of patterns consistent with various arrhythmias. As with any system, however, inaccurate coding occasionally occurred; therefore, manual inspection of the Holter recordings was necessary to ensure accuracy. Patterns in LPs that could not be categorized had been inaccurately coded in some areas.
The dots' locations in the LP, especially when coupled with color coding, can also provide information on the cause of an arrhythmia. For example, pink dots corresponding to VPCs are plotted along the y-axis as a VSL, whereas pink dots corresponding to ventricular escape rhythms are plotted in the upper right corner because of their timing relative to the previous beat.24 The DSL pattern often indicated bigeminy but also occurred as a result of premature beats preceded by variably long R-R intervals associated with sinus arrhythmia. This pattern was seen because the VSL of the DSL pattern represented the short R-R interval before the premature beat and the long R-R interval after the premature beat, and the HSL of the DSL pattern represented the long R-R interval after the premature beat and the short R-R interval before the next premature beat.
Some LPs showed multiple parallel VSLs consistent with the multiple coupling R-R intervals of premature beats; multiple coupling intervals may not be apparent by evaluating an ECG trace but may be noticeable by evaluating the LP because LPs display numerous, superimposed R-R intervals.25 A TSL pattern was variably identified in the LPs of recordings with VPCs as a result of postectopic compensatory pauses that produced distinct R-R intervals not seen in LPs of recordings with SVPCs.14 The TSL pattern was not evident in the LPs derived from all recordings with VPCs, possibly because interpolated VPCs did not have compensatory pauses or the third side lobe blended with the Y-shaped pattern, but the TSL pattern was only noted in the LPs from the recordings with VPCs. Therefore, identification of a TSL pattern in an LP supports the usefulness of an LP for confirming the presence of VPCs. Double side lobe and TSL patterns in LPs derived from the Holter recordings of people have also been reported,14 and the densities of the side lobes were positively correlated to the frequency of premature complexes. Although the intensity of the side lobes was subjectively related to the number of VPCs in the present study, the relationship between the intensities of the side lobes and the number of VPCs was not evaluated systematically.
The LPs of the recordings with CSVE and CVE showed green and pink dots, respectively, concentrated in the lower left corner. The location of the dots was consistent with repetitive, short R-R intervals associated with couplets, triplets, and paroxysmal tachycardia. The degree of dot concentration in the lower left corner indicated the rapidity of the consecutive premature beats; however, the number of consecutive premature beats in the CVE or CSVE could not be determined from the LPs.
A fan pattern was consistently identified in the LPs from the Holter recordings with AF. Results of studies14,19,26 of people with AF indicate that although the LPs derived from the Holter recordings of some people have homogenous fan patterns similar to that in the LPs of the dogs with AF of the present report, others have patterns of multimodal interval distributions (eg, double sector fan or clusters), which support nonlinear dynamics or dual AV nodal conduction.14,19,26 No dogs with AF in the present study had fan patterns with double sectors or clusters. The fan pattern supports a nonlinear relationship between the atrial electrical impulses and the AV node that is associated with the irregularly irregular rhythm of AF.14
In the present study, AF associated with fast heart rates resulted in a fan pattern with dots concentrated in the lower left corner of the LPs, and AF associated with lower heart rates resulted in a fan pattern with dots that had spread toward the right and upper parts of the LPs. On the basis of these observations, LPs of recordings with AF may allow for rapid assessment of the overall heart rate and, therefore, aid in the decision to begin or adjust the dosage of negative chronotropic medications.
An island pattern was consistently identified in the LPs from recordings with high-grade second-degree AVB. The island pattern represents the variability of AV nodal conduction, with conduction ratios of 2:1, 3:1, or 4:1.12 The island patterns were round in some LPs and elliptical in others; an elliptical island pattern may be explained by the autonomic influence on the rate of conducted beats.14
Although the visual inspection of an LP does not replace the visual inspection of a Holter recording, the visual inspection of an LP is valuable prior to the complete analysis of a recording. Instantaneous heart rates can be gleaned from all LPs, and the rapid assessment of instantaneous heart rates can be helpful for determining the need for and the effectiveness of antiarrhythmic medications for dogs with AF, CSVE, and CVE. Visual inspection of an LP aids in the detection of sinus, supraventricular, and ventricular beats through color coding of the plotted dots and the presence of multiple coupling intervals for premature beats through identification of > 1 VSLs.
The present study had several limitations. First, the number of Holter recordings for each arrhythmia type was low. Although the color coding was corrected for most of the LPs when necessary, some with frequent undercoded SVPCs were too laborious to manually correct. Second, identification of the LP patterns was subjective, which may have then led to their misclassification. Grouping of Holter recordings on the basis of the predominant arrhythmia type may have introduced error because of the concurrence of nonpredominant arrhythmias in some recordings; however, the unique color coding of beats by the Holter recording system allowed for visual recognition of the beats' origins. Third, cardiac and other medications and underlying structural cardiac and systemic diseases may have affected the LP patterns. The effect of cardiac medications was especially relevant for the recordings from dogs with AF because all of those dogs received negative chronotropic medications; without those medications, the fan pattern would have likely shifted closer to the lower left corner of the LP, reflecting a high heart rate. Yet, ineffective negative chronotropic treatment would have also similarly shifted the fan pattern. Fourth, the recording system used in this study was limited to the generation of an LP from a single hour; however, generation of a single LP from all 24 hours would have provided a more comprehensive assessment of the rhythms for each dog. Because the periods, ranging from 3 minutes to 24 hours, for which LPs are generated from the recordings of people14–16 and dogs8,13 are variable, results of the present study may not be directly comparable to those of other studies. Each LP should always be interpreted with knowledge of the represented time period and limitations of the recording system.
In the present study, we identified specific LP patterns for common cardiac rhythms of dogs, some of which were similar to those reported for people and some of which were similar to those reported for dogs of other studies. The patterns identified here implied nonrandomness to rhythm generation. Visual recognition of LP patterns associated with various cardiac rhythms in dogs may have value for understanding the origins of the rhythm and interpreting normal and abnormal rhythms in clinical and research settings, especially considering the ability to visually assess the rhythm over a large time period with a single LP. The similarities of LP patterns for dogs and people make feasible the translational efforts to understand mechanisms of arrhythmias and the effects of interventions. Future studies could explore possible automated identification of arrhythmias, which may reduce manual review of Holter recordings, allow rapid quantification of arrhythmia frequency, and facilitate assessment of pattern changes that might reflect autonomic vacillations or medical interventions.
Acknowledgments
The authors did not receive financial support for this study. The authors declare that there were no conflicts of interest.
This was a retrospective study of clinical animal patients; therefore, institutional approval was not required. No laboratory animals were used for this study.
ABBREVIATIONS
| AF | Atrial fibrillation |
| AV | Atrioventricular |
| AVB | Atrioventricular block |
| CSVE | Complex supraventricular ectopy |
| CVE | Complex ventricular ectopy |
| DSL | Double side lobe |
| HSL | Horizontal side lobe |
| LP | Lorenz plot |
| SVPC | Supraventricular premature complex |
| TSL | Triple side lobe |
| VPC | Ventricular premature complex |
| VSL | Vertical side lobe |
Footnotes
Trillium 5000, version 4.41, Forest Medical LLC, East Syracuse, NY.
References
1. Gelzer ARM, Kraus MS, Rishniw M, et al. Combination therapy with mexiletine and sotalol suppresses inherited ventricular arrhythmias in German Shepherd Dogs better than mexiletine or sotalol monotherapy: a randomized cross-over study. J Vet Cardiol 2010;12:93–106.
2. Spier AW, Meurs KM. Evaluation of spontaneous variability in the frequency of ventricular arrhythmias in Boxers with arrhythmogenic right ventricular cardiomyopathy. J Am Vet Med Assoc 2004;224:538–541.
3. Spier AW, Meurs KM. Assessment of heart rate variability in Boxers with arrhythmogenic right ventricular cardiomyopathy. J Am Vet Med Assoc 2004;224:534–537.
4. Miller RH, Lehmkuhl LB, Bonagura JD, et al. Retrospective analysis of the clinical utility of ambulatory electrocardiographic (Holter) recordings in syncopal dogs: 44 cases (1991–1995). J Vet Intern Med 1999;13:111–122.
5. Poincaré H. Calculus of probabilities and electrodynamics, In: Science and hypothesis. New York: The Science Press, 1905;218–224, 257–272.
6. Lorenz E. Deterministic nonperiodic flow. J Atmos Sci 1963;20:130–141.
7. Nayak SK, Bit A, Dey A, et al. A review on the nonlinear dynamical system analysis of electrocardiogram signal. J Healthc Eng [serial online]. 2018;2018:1–19. Available at: www.hindawi.com/journals/jhe/2018/6920420/ Accessed Nov 19, 2018.
8. Blake RR, Shaw DJ, Culshaw GJ. Poincaré plots as a measure of heart rate variability in healthy dogs. J Vet Cardiol 2018;20:20–32.
9. Esperer HD, Oehler M. Automatic quantification of the Poincaré plot asymmetry of NN-interval recordings. Physiol Meas 2010;31:395–413.
10. Thomsen MB, Volders PGA, Beekman JDM, et al. Beat-to-beat variability of repolarization determines proarrhythmic outcome in dogs susceptible to drug-induced torsades de pointes. J Am Coll Cardiol 2006;48:1268–1276.
11. Thomsen MB, Truin M, van Opstal JM, et al. Sudden cardiac death in dogs with remodeled hearts is associated with larger beat-to-beat variability of repolarization. Basic Res Cardiol 2005;100:279–287.
12. Gladuli A, Moise NS, Hemsley S, et al. Poincaré plots and tachograms reveal beat patterning in sick sinus sndrome with supraventricular tachycardia and varying AV nodal block. J Vet Cardiol 2011;13:63–70.
13. Moïse NS, Gladuli A, Hemsley SA, et al. “Zone of avoidance”: RR interval distribution in tachograms, histograms, and Poincaré plots of a Boxer dog. J Vet Cardiol 2010;12:191–196.
14. Esperer HD, Esperer C, Cohen RJ. Cardiac arrhythmias imprint specific signatures on Lorenz plots. Ann Noninvasive Electrocardiol 2008;13:44–60.
15. Zhang L, Guo T, Xi B, et al. Automatic recognition of cardiac arrhythmias based on the geometric patterns of Poincare plots. Physiol Meas 2015;36:283–301.
16. Borracci RA, Montoya Pulvet JD, Ingino CA, et al. Geometric patterns of time-delay plots from different cardiac rhythms and arrhythmias using short-term EKG signals. Clin Physiol Funct Imaging 2018;38:856–863.
17. Gelzer ARM, Moïse NS, Vaidya D, et al. Temporal organization of atrial activity and irregular ventricular rhythm during spontaneous atrial fibrillation: an in vivo study in the horse. J Cardiovasc Electrophysiol 2000;11:773–784.
18. Moïse NS, Dugger DA, Brittain D, et al. Relationship of ventricular tachycardia to sleep/wakefulness in a model of sudden cardiac death. Pediatr Res 1996;40:344–350.
19. Oka T, Nakatsu T, Kusachi S, et al. Double-sector Lorenz plot scattering in an R-R interval analysis of patients with chronic atrial fibrillation: incidence and characteristics of vertices of the double-sector scattering. J Electrocardiol 1998;31:227–235.
20. Pedich W, Popiolek J, Tobiaszewska D. The mechanism of the silent zone on Lorenz plots in atrial fibrillation. Rocz Akad Med Bialymst 1998;43:232–244.
21. Lamb AP, Meurs KM, Hamlin RL. Correlation of heart rate to body weight in apparently normal dogs. J Vet Cardiol 2010;12:107–110.
22. Meurs KM, Spier AW, Wright NA, et al. Use of ambulatory electrocardiography for detection of ventricular premature complexes in healthy dogs. J Am Vet Med Assoc 2001;218:1291–1292.
23. Moïse NS. From cell to cageside: autonomic influences on cardiac rhythms in the dog. J Small Anim Pract 1998;39:460–468.
24. Pan Y, Fanfang Z, Ru L, et al. The characteristics of RR-Lorenz plot in persistent atrial fibrillation patients complicating with escape beats and rhythm. Zhonghua Xin Xue Guan Bing Za Zhi 2014;42:481–483.
25. Katohno J, Nakai K, Itoh C, et al. Analysis of the preceding R-R interval and the coupling interval on premature ventricular contractions. Rinsho Byori 1994;42:419–424.
26. Climent AM, Guillem MDLS, Husser D, et al. Poincaré surface profiles of RR intervals: a novel noninvasive method for the evaluation of preferential AV nodal conduction during atrial fibrillation. IEEE Trans Biomed Eng 2009;56:433–442.
Appendix
Descriptions of patterns in LPs and, where applicable, the corresponding underlying cardiac rhythms responsible for pattern generation.
| LP pattern | Description of dot distribution | Cardiac rhythm |
|---|---|---|
| Torpedo shape | Elliptical appearance | Sinus rhythm without rate variability |
| Y shape | Elliptical appearance in the lower left corner of the LP with wide distribution of dots toward the upper right corner of the LP | Sinus rhythm with pronounced rate variability |
| Comet shape | Club-like shape with subtle distribution of dots to the upper right corner of the LP | Sinus rhythm with some rate variability |
| DSL | Elliptical clusters orientated parallel to the x- and y-axes of the LP | Premature beats |
| TSL | Elliptical clusters orientated parallel to the x- and y-axes with a third elliptical cluster above the cluster parallel to the x-axis of the LP | 3 dot locations based on consistent interectopic intervals associated with VPCs (dots associated with a sinus beat preceding a VPC, dots associated with VPCs of short coupling intervals, and dots associated with sinus beats after the compensatory pause of a VPC) |
| Fan | Triangular and diffuse appearance | AF |
| Island | Regularly spaced clusters | Second-degree AVB |
| Silent zone | Discrete area with relative absence of dots | NA |
| Cluster | Discrete area with predominance of dots | NA |
| Side lobes | Elongated cluster near the x- or y-axis of the LP | NA |
NA = Not applicable.
