Background Cardiac arrhythmias in horses are diagnosed by auscultation or electrocardiogram (ECG), which results in a low sensitivity for detecting arrhythmias that occur sporadically. Implantable loop recorders (ILRs) are small ECG devices placed subcutaneously, to automatically detect arrhythmias in human patients. Objectives To test ILRs ability to detect atrial fibrillation (AF) in horses. Furthermore, we hypothesised that anatomical location of the implant site might influence signal quality. Signal quality was evaluated both during exercise and over time. Study design Experimental study. Methods In five Standardbred mares, eleven ILRs were implanted subcutaneously in up to three different positions (Front: pectoral region, Left-6: sixth left intercostal space and Ventral: xiphoid region) and AF induced. The R- and T-wave amplitudes were measured in all positions over time during AF. AF burden automatically registered by the ILRs over a 2-month period was compared with selected Holter ECG recordings. Results All three positions had stable R- and T-wave amplitudes during the study period and were of sufficient quality to allow AF detection at rest. The position Left-6 showed significantly higher R- and T-wave amplitudes compared with the other positions. During submaximal exercise only the Left-6 position was able to record ECG signals of diagnostic quality. No position yielded diagnostic signals at maximum exercise due to artefacts. Main limitations Few horses and ILRs included and no spontaneous AF episodes were studied. Conclusions This preliminary study indicates that ILRs can be used for AF detection in horses, but the anatomical location is important for optimal ECG quality. Despite insufficient quality during exercise, ILRs were suitable for AF detection at rest. Therefore, the ILR may be a valuable diagnostic tool for detecting paroxysmal AF in horses.
Detection of atrial fibrillation with implantable loop recorders in horses
Review badges
Web of Science (Free Access)
Abstract
- Contribute
- pre-publication peer review (FINAL ROUND)
Decision Letter
2020/05/1818-May-2020
Dear Prof. Buhl:
I am pleased to inform you that your manuscript entitled "Detection of atrial fibrillation with implantable loop recorders in horses" is now acceptable for publication in the Equine Veterinary Journal.
Accepted articles are pre-published on Wiley Online Library within around 14 days, provided that we have received the appropriate license agreement. Your article cannot be published until the publisher has received the appropriate signed license agreement. Within the next few days the corresponding author will receive an email from Wiley’s Author Services system which will ask them to log in and will present them with the appropriate license for completion.
Equine Veterinary Journal's preprint policy requires that once a final version of a preprint paper is peer-reviewed, published, and becomes part of the version of record, the authors should link the preprint to the published paper. Please ensure that you do this if you have used a pre-print service.
You will receive proofs in due course and, once proofs are agreed upon, your paper will be placed on our Early View system prior to publication in the printed issue. PLEASE NOTE: THE CORRESPONDING AUTHOR IS SOLELY RESPONSIBLE FOR CHECKING THE PROOFS CAREFULLY WHEN THEY ARE RETURNED FROM WILEY.
Wiley-Blackwell also offers 'Online Open' allowing authors the opportunity to make their article available to non-subscribers. For more details, please see https://authorservices.wiley.com/bauthor/onlineopen_order.asp
How we will promote your article:
EVJ promotes articles by producing written, audio or other content to distribute via social media and this can include using images from published papers. Please let us know as soon as possible if you have any concerns about this. Authors are increasingly using Twitter to raise awareness about their research. If you do decide to tweet about your article, please include the link to your article on EVJ's Wiley Online Library page and direct a copy of your tweet to @EVJltd which we will be happy to retweet to our followers. It is most effective to use the link ending in '/abstract' - this link is set up when your article is posted online in our Accepted Article section and is maintained as it moves through our Early View and ultimately into a specific issue.Finally, in conjunction with Wiley, we are pleased to inform you that the Kudos service is now freely available for all EVJ authors. This web-based service helps authors explain, enrich, and share their published work for greater readership and impact. It also provides direct access to a publication dashboard so authors can measure the effect of their actions across a wide range of metrics. Authors who have registered with Wiley Author Services and opted into the mailing list will receive a notification email containing a direct link to claim their article in Kudos. More information regarding Kudos can be found in our Author Guidelines or via http://www.tamba.co.uk/dev/files/kudos/ANI_KUDOS_VID1_SPONSORS_V13_.mp4
Thank you for submitting this paper to the Equine Veterinary Journal. We look forward to your future contributions to the Journal.
Sincerely,
Celia Marr
Editor, Equine Veterinary Journal
cmm@evj.co.ukCelia Marr
Associate Editor, Equine Veterinary Journal
cmm@evj.co.ukP.S. – You can help your research get the attention it deserves! Check out Wiley's free Promotion Guide for best-practice recommendations for promoting your work at www.wileyauthors.com/eeo/guide. And learn more about Wiley Editing Services which offers professional video, design, and writing services to create shareable video abstracts, infographics, conference posters, lay summaries, and research news stories for your research at www.wileyauthors.com/eeo/promotion.
Editor Comments to the Author:
I have made some minor edits, including deleting the owner consent section as we do not usually include this for experimental studiesPlease let me know if you have any concerns.
Decision letter by
Cite this decision letter
Author Response
2020/04/16Dear Editor
We thank you for the fast and thorough handling of this manuscript. Please find our response to the reviewers comments below.
Sincerely,The Authors
Peer Reviewer: 2
Comments to the Author
The authors have revised the manuscript according to my comments and suggestions, and have adequately responded to my questions. I only have minor comments left:
We thank you for your reply and have responded to your comments below.Line 75: Recent study have shown should be has shown
Thank you. This has been corrected in the text.Line 116-118: where the performance instead of were, horses were part of instead of where part of
Thank you. This has been corrected in the text.Line 213-214: how do you explain that all ILRs showed correct R wave sensing at walk at the start of the exercise test, while the quality was only sufficient for Left-6 at walk after the exercise test?
We agree that this is imprecise described. Horses showed correct detection in all positions during baseline and walking. Immediately after the exercise test quality remained reduced for all but Left-6 for about 1-2 minutes, most likely due to increased body movement at the end of the exercise. We tried to express this clearer in the text. The sentence now read L214-215:
“When the horses were walking at the treadmill immediately after the test, the quality was sufficient for the Left-6 position and regained sufficient quality for the other positions one to two minutes after the test was stopped.”Line 223-224: no false positive detection of AF due to artefacts. Please specify whether the horses performed any type of exercise (walking, pasture, …) during this period or if they had box rest. Artefacts would be predominantly expected during exercise.
We agree that this has not been written clearly. The horses were kept at rest for two weeks and thereafter taken to the paddock on daily basis. We have included this, so it now read L225-226:
“However, horses were kept in box rest the first two weeks and allowed on the paddock on week three.”Line 228-229: it is still unclear what is meant by “intermittent beats with lower R wave amplitude”. Are these ventricular premature beats?
We agree with the reviewer that this is formulated a bit vague. This is due to the fact, that this is not clear how these beats should be classified. Even when analysing a three lead Holter ECG these beats remain difficult to classify and even more difficult with one lead as the ILR. In humans, it is known that broad, bundle branch block like deformed QRS complexes occur as aberrant conducted beats. Whether these beats seen in the horses are aberrant conducted AF beats or ventricular extra beats cannot be determined with the ECG recordings we have available. We have therefore included an extra sentence specifying that we cannot diagnose these beats further L231-232:
“Further classification of these beats were not possible from the recorded ECG traces”Line 230: was misclassification of T waves as R waves affected by the sign of the T wave (negative or positive?)
This is a very important question to raise. In our recordings both, positive and negative T waves were misclassified. We have added this fact to the manuscript L233-234:
“This was typically the case for T waves with high amplitude that were misclassified as R waves, regardless whether the T wave was positive or negative.”Line 261-262: Up to 2.7% misclassification of AF – it is unclear whether this consists of false positive or false negative AF detection, or both.
We agree that this is unclear. The authors of the cited study report false positive detection of AF. We have added this to the text. The sentence now reads L265-266:
“…up to 2.7% misclassification false positive AF episodes due to artefacts”Line 267: IRL should be ILR
Thank you. This has been corrected in the text.Figure 2D would be more informative if the timeline would be restricted to the study period, and if the study period would be indicated on the timeline.
We agree and have now changed the time axis to show the time relevant for the study only and indicated the study-days.Cite this author response
- pre-publication peer review (ROUND 2)
Decision Letter
2020/04/1414-Apr-2020
Dear Prof. Buhl:
Your manuscript ID EVJ-TN-19-302.R1 entitled "Detection of atrial fibrillation with implantable loop recorders in horses" which you submitted to the Equine Veterinary Journal, has been reviewed.
Decision
We require minor revisions to your manuscript before it can be reconsidered for publication. Therefore, I invite you to respond to the comments included at the bottom of this letter and revise your manuscript accordingly.Revision process
To revise your manuscript, log into https://mc.manuscriptcentral.com/evj and enter your Author Center, where you will find your manuscript title listed under "Manuscripts with Decisions." Under "Actions" click on "Create a Revision." Your manuscript number has been appended to denote a revision. Please DO NOT upload your revised manuscripts as a new submission.All revised articles must be uploaded with
(i) a main document containing a version which does not have a title page and is blinded and in which all changes are highlighted using 'track changes' mode. This version will be sent out to reviewers;
(ii) a document containing a version which includes the title page and in which blinding details have been re-inserted and changes are not marked. This should be uploaded with the file designation 'Clean Non-Blinded Copy'. If your manuscript is accepted, this version will be uploaded to our Accepted Articles section.All revised manuscripts must be accompanied by a point-by-point response, addressing every comment made by the Editors and Reviewers. When submitting your revised manuscript, you must respond to these comments in the space provided.
Your original files are available to you when you upload your revised manuscript. Please delete any redundant files before completing the submission.
Deadline
Your revised manuscript should be uploaded as soon as possible and within 12 weeks.Thank you for submitting your manuscript to the Equine Veterinary Journal.
Sincerely,
Celia Marr
Editor, Equine Veterinary Journal
cmm@evj.co.ukPeer Reviewer(s)' Comments to Author:
Peer Reviewer: 2
Comments to the Author
The authors have revised the manuscript according to my comments and suggestions, and have adequately responded to my questions. I only have minor comments left:
Line 75: Recent study have shown should be has shown
Line 116-118: where the performance instead of were, horses were part of instead of where part of
Line 213-214: how do you explain that all ILRs showed correct R wave sensing at walk at the start of the exercise test, while the quality was only sufficient for Left-6 at walk after the exercise test?
Line 223-224: no false positive detection of AF due to artefacts. Please specify whether the horses performed any type of exercise (walking, pasture, …) during this period or if they had box rest. Artefacts would be predominantly expected during exercise.
Line 228-229: it is still unclear what is meant by “intermittent beats with lower R wave amplitude”. Are these ventricular premature beats?
Line 230: was misclassification of T waves as R waves affected by the sign of the T wave (negative or positive?)
Line 261-262: Up to 2.7% misclassification of AF – it is unclear whether this consists of false positive or false negative AF detection, or both.
Line 267: IRL should be ILR
Figure 2D would be more informative if the timeline would be restricted to the study period, and if the study period would be indicated on the timeline.Decision letter by
Cite this decision letter
Reviewer report
2020/04/07The authors have revised the manuscript according to my comments and suggestions, and have adequately responded to my questions. I only have minor comments left:
Line 75: Recent study have shown should be has shown
Line 116-118: where the performance instead of were, horses were part of instead of where part of
Line 213-214: how do you explain that all ILRs showed correct R wave sensing at walk at the start of the exercise test, while the quality was only sufficient for Left-6 at walk after the exercise test?
Line 223-224: no false positive detection of AF due to artefacts. Please specify whether the horses performed any type of exercise (walking, pasture, …) during this period or if they had box rest. Artefacts would be predominantly expected during exercise.
Line 228-229: it is still unclear what is meant by “intermittent beats with lower R wave amplitude”. Are these ventricular premature beats?
Line 230: was misclassification of T waves as R waves affected by the sign of the T wave (negative or positive?)
Line 261-262: Up to 2.7% misclassification of AF – it is unclear whether this consists of false positive or false negative AF detection, or both.
Line 267: IRL should be ILR
Figure 2D would be more informative if the timeline would be restricted to the study period, and if the study period would be indicated on the timeline.Reviewed by
Cite this review
Author Response
2020/03/09Dear Editor,
Thank you very much for the comprehensive respond from the two reviewers. This is highly appreciated, and we believe their comments have strengthen the manuscript. We have responded point-by-point below.
On behalf of the authors,Sincerely
The Authors
Peer Reviewer(s)' Comments to Author:
Peer Reviewer: 1
Comments to the Author
General comments
This technical note describes the use of implantable loop recorders (ILRs) to detect atrial fibrillation in horses. It compares the signal quality obtained from 3 different anatomical sites both at rest and exercise. It also compares automatic detection of atrial fibrillation by the ILRs with manual evaluation of a Holter ECG recording. There is very little published information on the use of ILRs in horses -previous reports have only evaluated signal quality from one anatomical site (the left pectoral area) and have not evaluated a trace at exercise. In addition, these previous reports have not evaluated automatic detection of arrhythmias by the ILR but relied on manual activation of the ILR to save a recording after abnormal clinical signs were observed. This technical note therefore provides useful new information on the use of ILRs to detect arrhythmias in horses. This is clinically relevant as current diagnostic techniques are limited in their ability to detect sporadic arrhythmias.
Answer: We thank the reviewer for acknowledge the relevance of the study and will in the following respond to each of the comments/questions raised.Specific points
The terms P wave, R wave and f waves are variably hyphenated throughout the manuscript. These hyphens should be removed for consistency.
Answer: This has been corrected throughout the manuscriptMaterials and Methods
I found the timeline of the study difficult to follow – probably because this was a subset of horses used for another study. I think the “Materials and Methods” section could be re-ordered to make the timeline clearer. For example I think the section entitled “Implantable loop recorder” (lines 116-138) in “Materials and methods” would follow better after line 108.
Answer: We have included a timeline in the figure 1 and also included this in the text (L96). We have reordered the section as suggested (L127-143). Also, we included the sentence L116-118:
“Following implantation of pacemaker and ILRs, the horses had a resting period of three weeks where the performance of ILRs in sinus rhythm could be evaluated. Hereafter the horses were paced into sustained AF”.109-112 – It is unclear how this cardioversion relates to this particular study and presumably is because these horses were part of another study. Please could you just clarify this.
Answer: We agree with you that it is unclear. The previous studies tested two drugs for cardioversion of AF, and therefore the horses had a few time periods in SR. To make this clearer to the reader, we have rewritten the sentence L118-121:
“As the horses were part of two pharmacological studies, cardioversion was attempted on day 1, 3, 9, and 27 following AF induction with intravenous infusion of 2 mg/kg flecainide (Tambocor®)i or 3 mg/kg Acetylcholine-Activated Inward-Rectifier K+ Current blocker (XAF-1407)j”113 – “Four horses in persistent AF” – It sounds like only 4 horses were in persistent AF. Were all 5 in persistent AF but only 4 went on the treadmill?
Answer: We fully agree that this is misleading. Therefore, we have rewritten the sentence L122-123:
“Four horses underwent treadmill testing when in persistent AF as previously described”113-115 – Please could you describe what tack the horses were wearing as this may be relevant to artefact generation.
Answer: The horses were not equipped with specific tack during the test, but just a headcollar and an elastic girth to fixate a surface ECG. L125-126:
“The horses were equipped with a headcollar and an elastic girth to fixate the surface ECG”.131 – Change “SR” to “sinus rhythm (SR)” as this abbreviation has not been used previously
Answer: This has been corrected L146.134 – Please add a definition for “blanking after sense” as you have done for the “sensing threshold”
Answer: We added the following description to the text L150-151:
“….blanking after sense (defines the minimal time before a new R-wave can be detected)”147-148 – I don’t understand these values – e.g. if a Holter ECG was in place for 14.4 days then doesn’t that generate (14.4 x 24 hrs) 345.6 hours of ECG recording?
Answer: We agree, this is an incorrect description. The horses were equipped with Holter ECGs on 14.4 days. However, not all of the recordings were of sufficient quality (electrodes detachment or poor contact in general), which resulted in a mean duration of only 145 hours readable ECG recordings. We have rewritten the sentence where we inform that we had 145 hours of diagnostic ECGs L161-164:
“Holter ECG recordingsh were obtained before AF induction and various times during AF and analysed manually to verify correct AF detection of the ILR. Holter ECGs were recorded on average for 14.4 days (range 13 – 16 days), resulting in ECGs of diagnostic quality from the Holter recordings of a mean duration of 145 hours (range 106 – 288 hours) for each horse.”151- AF burden histogram – is it possible to add one of these as another panel in figure 2 to allow the reader to visualize the data outputs better?
Answer: We have added an example of an AF-burden histogram as it is generated by the ILR as figure 2D to visualize the data.155-159 – I found this paragraph a bit confusing to read. Perhaps you could specify that R and T waves amplitude was compared between the 2 interrogations. I wasn’t sure what you meant by “taken form the same period” on line 157-158.
Answer: We fully agree that this might be unclear. We have now rewritten to make it clear that the R and T waves are measured at two timepoints for all three locations and compared. L178-182:
“As the ILR algorithm relies on R wave detection, the amplitude of the R and T waves were measured from 10 consecutive heart beats from registered AF episodes obtained approximately one and two months after ILR implantation. The average of R and T wave amplitude as well as the presence of f waves were compared for all the three locations. To evaluate the ECG over time, the R and T wave amplitude was compared between first and second interrogation.”Results
180 – There is no mention in the “Materials and Methods” that the horses were subject to euthanasia and post mortem examination – I think this should be included
Answer: We agree and have included this: L95-98:
“This study used a subset of horses that were used to evaluate the longitudinal effect of AF in horses (timeline, Fig 1) (reference “masked for peer review”) and the horses were euthanized and subjected to post mortem examination at the end of the study.”185 – “At baseline” presumably means “at rest” – perhaps reword?
Answer: We agree and has reworded L209-210:
“In the four horses tested on the treadmill, the ILRs sensed R waves correctly in the triggered ECG recordings for all positions at rest and during walking”186 – At what stage after cessation of maximal exercise did the traces became diagnostic again? I think this is relevant as presumably it could still help diagnose some cases of paroxysmal AF only present at exercise (and therefore identify a cause for poor performance) even if the trace is not diagnostic at maximal exercise.
Answer: The trace became diagnostic when the horses were walking for the Left-6 position but it was less clear for the two other positions. We have written that explicit in L214-215:
“When the horses were walking at the treadmill after the test, the quality was sufficient for the Left-6 position.”191 – Please clarify what you mean by “kinetics”.
Answer: The expression “similar duration and kinetics” is imprecise. We wanted to emphasize the all ILRs were capable of detecting AF episodes of comparable length, especially during episodes where cardioversion was successful. The “kinetics” of this is implicitly given by the time axis of the AF burden. We therefore removed the word “kinetics” from the manuscript.201- Insert “amplitude “so it reads “R/T wave amplitude ratio”
Answer: We agree that this is a better way of describing this ratio, and this has been inserted. L230-231
“Oversensing occurred when the R/T wave amplitude ratio was low”Discussion
213 -Insert “amplitude” so it reads “high R/T amplitude ratio”
Answer: This has been inserted: L243-244:
“….a high R/T amplitude ratio are essential to correctly identify arrhythmias.”218-219 – “These authors were not able to detect ECGs during collapses due to excessive artefacts in the acquired signals and therefore no arrhythmia diagnosis could be obtained”. This statement is incorrect and does not reflect what was written in that study. For example, in one horse the paper states that recordings were obtained during 2 subsequent collapse episodes and did not detect any arrhythmias or abnormal complexes.
Examples of traces from horses with ILRs in the pectoral region are published in another paper by some of the same authors - Lyle and Keen (2010) Episodic collapse in the horses Equine Veterinary Education 22 (11) 583-584. Even if artefacts are generated during movement associated with collapse this may not preclude diagnosis of an arrhythmia as the cause of the collapse as the arrhythmia would occur prior to the collapse. However, I agree that based on information from your study that the pectoral position is suboptimal compared to the left intercostal space.
Answer: We agree that we have not cited this study correct. Therefore we have rewritten the sentence L 248-249: “The ILR used in the previous study did not detect arrhythmias during the two recorded episodes of collapse.”222-224 – The range of ILR implantation time was 52-85 days so I don’t think it is possible to extrapolate to definitively say that signal quality would be good for years, e.g. migration of the ILR over time may occur.
Answer: We agree that this is not a correct way of writing this. We have racehorses at the racetracks with ILRs implanted for two years and they still work perfect (but of course we can’t use this argument here, as this is unpublished data). We have therefore rewritten the sentence. L254-255:
“….that these recorders are capable of recording sufficient quality ECGs over longer time in horses as it is seen in humans [19]”225-226 – Specify that manual analysis of a Holter ECG trace seems to be superior to use of an ILR to detect AF. Use of Holter software algorithms to analyse the trace is likely to misinterpret artefacts in the same way as the ILR algorithms so may also lead to false-positive diagnoses of arrhythmias.
Answer. We agree and this has been specified. L256-259:
“ILR can therefore not replace the manual analysis of a Holter ECG, which is the Golden Standard for AF detection, but the ILR could be useful in suspected cases of intermittent AF as they provide continuous arrhythmia detection over long time.“236-242 – The limited interrogation frequency was not a problem in this study (i.e. no episodes had been deleted from the ILR prior to interrogation) but I wonder if this could be due to the fact that the horses were mainly in AF? For example, I wonder if a horse was in sinus rhythm but regularly had episodes of physiological second-degree AV block that these would fill up the ILR quite quickly and then it might be possible that the pathological arrhythmia of interest would be missed? I think it would be worth saying that it would be good to investigate the use of ILR in horses in sinus rhythm to determine how quickly the episodes might fill up and therefore help determine how often interrogation might be required.
Answer: The diagnostic memory of an ILR is divided into two parts. The first one is a simple list containing the overall count/number of arrhythmic episodes (i.e. AF 20 episodes, pauses 300 episodes and so on). The second one is a list that can only contain a limited number of selected episodes that provides further detail (i.e. mean/min/max heart rate and ECG examples). While no information of the total number of episodes is lost, a more in depth analysis of episodes is only available for a selected number. We extended the description of the ILR to the manuscript L132-143:
“A list containing all types of arrhythmia and its respective number of occurrence is stored automatically for the total life-time of the device. Furthermore, the reveal LINQ can store a list containing up to 30 episodes in greater detail, including date and time of onset, duration as well as mean and maximal heart rate. This list is complemented with up to 14 ECG traces (examples Fig 2). The ECGs carry annotations under each detected R wave and can be used to retrace the ILRs diagnose. If more episodes occur, the ECG of the oldest episodes will be overwritten, but the list containing the total count will not be affected by that. Total duration of AF episodes is stored as AF burden (percentage in AF per day). Recorded episodes can be downloaded using a programming device or transferred automatically to Medtronic’s care link server every 24 hours using a home monitoring device. Battery life is typically three years depending on the model. For further details on the functionality of the ILR the manufacturer’s manual can be consulted.”
Furthermore we added to the text L224-225:
“All horses had been in sinus rhythm for 3 weeks with ILRs implanted (day -21 to 0). None of the ILRs detected a false positive AF episode during this period.”Peer Reviewer: 2
Comments to the Author
The manuscript describes the use of ILRs for detection of AF in horses. This technical note clearly describes the Left-6 position as the best position for implantation of an ILR, which is interesting information for future studies using ILRs. However, based on the described methods and results, it is unclear how the AF burden was measured using the ILRs and how the comparison with Holter ECG was made. A timeline is missing and it is unclear how many hours/days of AF were included in the study period. This is essential information to draw conclusions about the usefulness of ILRs.
Answer: We thank the reviewer for these important comments, which has stimulated us to rewrite part of the manuscript. We have decided to exclude the Pearsons correlation comparing the Holter recordings and ILRs AF detection. The ILR used in our study has an automatic artifact rejection. Unfortunately the system does not inform how much of the ECG recording that is rejected. In other words, the clinician don’t know if the ILR has discarded parts of the automatic recordings due to artefacts. When we look at the AF burden over the specific days for each individual horse in AF (see example in figure 2) we can see that almost 24 hours a day is recorded as AF and thereby indicates that the artefacts rejection is low. But as we don’t have the exact AF time registered each day and the full 24 hours Holter recordings for each day a quantitative analysis such as Pearsons correlation is therefore not possible and we decided to down-tone this part. Instead, we now use the Holter recordings to ensure correct AF detection of the ILR. We have addressed this further below. Furthermore a timeline was added to figure 1.The authors also describe the reduced diagnostic quality due to movement artefacts. Is this also a problem in human patients or other species, and how is it tackled in these species? This would be interesting information for improving the diagnostic value of ILRs, as these probably have most potential for use in exercising horses.
Answer: Surprisingly little is reported about the function of the ILRs during exercise in the human literature, but there are studies reporting the reasons for AF misdetection. In one study 20% of the AF misdetections registered was caused by artefacts (Eitel et al 2011) but whether these occurred due to exercise is not reported. Malmo et al 2016 reported up to 2.7% misclassification due to sinus arrhythmia, ectopic beats or noise during activity, but the exact percentage caused by exercise was not reported.
The ILR used in our study has an automatic artifact rejection: ECGs with poor diagnostic value (as defined by the internal algorithm) are not used to diagnose arrhythmia. Unfortunately neither whether this function was active nor the percentage of poor ECG that is rejected are presented to the clinician by the ILR setting. In other words, the clinician don’t know if the ILR has discarded parts of the automatic recordings due to artefacts. When we look at the AF burden over the specific days where the horses were in AF (see example in figure 2) we can see that almost 24 hours a day are recorded as AF and thereby indicate that the artefacts rejection is low. We have added this in the discussion by rewriting the following paragraph,
L257-271:
“The performance of ILRs during exercise showed artefacts due to myopotential interference resulting in poor quality ECGs for most positions in the present study. Only the Left-6 position showed acceptable ECGs during trot at 5 m/sec. Studies in humans using ILRs in a setting where exercise has been performed have functioned sufficiently [19-21], although it is not described in details. One study that investigated effect of training on AF burden reported up to 2.7% misclassification of AF due to artefacts during activity, ectopic beats or sinus arrhythmia [20]. These results are in accordance with our findings where none of the detected AF episodes documented with ECG by the ILR were false positive due to artefacts. The Reveal LINQ is equipped with a built-in algorithm that refrains noisy signals from being used in arrhythmia detection. The intervention of this algorithm or the amount of rejected signals are not presented to the clinician and a quantitative analysis is therefore not possible. However, the IRL rejected ECG episodes superimposed with artefacts (e.g. recordings during maximal exercise) and did not use them to detect an arrhythmia like AF or tachycardia. Despite this limitation, the AF burden histogram for the specific days showed that AF occurred for almost 24 hours/day for the horses, which indicate that the automatic artefact reduction is low in horses at rest.”The title is a bit misleading as the potential diagnostic implication does not seem the main focus of the manuscript.
Answer: The title has been changed to L2-3: “Detection of atrial fibrillation with implantable loop recorders in horses”Some sentences in the abstract need to be clarified:
Line 50-51: R and T wave amplitudes measured, why not P wave or S wave amplitude?
Answer: As we have mainly analysed ECG examples in AF and P waves were not present in these recordings. The internal detection of the ILR only uses R-waves. We therefore focused on R waves. We have added that measurements were performed when the horses were in AF in the abstract L50-51:
“The R and T wave amplitudes were measured in all positions over time during AF”Line 51: AF burden registered by ILRs, automatically detected or based on manual analysis?
Answer: With ILRs, only automatically detected AF burden can be registered as ILRs do not provide/store continues ECG recordings. An example of AF burden registration is given in figure 2D. We have rewritten the sentence in the abstract L51-52:
“AF burden automatically registered by the ILRs over a 2-month period were compared with selected Holter ECG recordings”Line 52: What is meant by ‘ECG signals stable over time’ or ‘suitable to detect AF’?
Answer: We agree that this imprecise and have rewritten the sentence: L52-54:
“All three positions had stable R and T wave amplitudes during the study period and were of sufficient quality to allow AF detection at rest”Manuscript:
Line 71-73: “short lasting episodes may not be diagnosed, and early recognition of the disease is essential for prompt treatment”. One would think that the main advantage of ILRs would be to detect paroxysmal AF, however, which implications does this have for treatment as cardioversion occurs spontaneously?
Answer: We agree that we at present will not change treatment of horses with PAF. However, PAF episodes might be preceding sustained AF episodes. As our knowledge is very limited within the field, extended knowledge about horses with PAF may result in change of management of these horses and potentially medical treatment (betablockers and sotalol as examples) may be initiated to prevent the disease to become chronic. A recent study in humans has shown changed treatment strategy in people with PAF that has been diagnosed by ILRs (Verma et al. Changes in management following detection of previously unknown atrial fibrillation by an insertable cardiac monitor. Am J Cardiol 2019). This has been added L75-77:
“Interestingly, a recent study have shown changed treatment strategy in people with PAF diagnosed with ILRs to prevent chronic AF and co-morbidities (Verma et al 2019).”To detect short-lasting persistent AF other approaches have been described such as the use of heart rate variability measurements from heart rate monitors. No reference to this work is made.
Answer: We have included the study from Broux et al 2018 and rewritten the sentence L70-73:
“Also heart rate monitors are able to distinguish AF from sinus rhythm (SR) (Broux et al. 2018, BMC Vet Res). However, for practical reasons, the duration of heart rate monitor recordings or Holter ECGs is limited to a few days and electrodes can be cumbersome to keep in place.”Line 80-81 about the use in human asymptomatic patients: these are probably patients with paroxysmal AF?
Answer: We agree that these patients most likely are PAF patients. We have changed AF to PAF and the sentence now reads L84-85:
“In humans, ILRs are used to detect PAF or AF recurrence in asymptomatic patients, with greater diagnostic yield than conventional Holter ECG [9-13].”Line 114-115: how many 10 minute ECGs over the entire test? A little bit extra information about the exercise test would be interesting for the reader, e.g. was this a submaximal or maximal exercise test.
Answer: We agree and has now included more information about the test L122-126:
“Four of the horses underwent an incremental treadmill testing to fatigue when in persistent AF as previously described [6]. During exercise, ILRs were manually activated (only 10-minutes can be recorded with manual trigger) to cover the latest part of the warm up period, the exercise test to fatigue and immediate post exercise period. The horses were equipped with a headcollar and an elastic girth to fixate the surface ECG.”Line 117-118: This information should be in the introduction. How is the R-R interval detected? Based on a positive peak of the QRS complex, a negative peak or any deflection? Was the largest deflection of the QRS complex positive in all ILR anatomical locations? And the T wave?
Answer: Medtronic does not provide the details on the underlying detection algorithm. The manual gives the impression that the signal is positively rectified and only mentions R-wave detection on filtered ECGs. The manual reads:
“The automatic detection of arrhythmias by the Reveal LINQ ICM is based on R-wave sensing. (…). The Reveal LINQ ICM filters the ECG signal to reduce noise and to reduce the number of sensed P-waves and T-waves. The filtered ECG signal is compared against the sensing threshold.”
We have added that the readers can refer to manufacturer’s manual for further details on the functionality, L142-143:
“For further details on the functionality of the ILR the manufacturer’s manual can be consulted.”Further to your question, the largest deflection of the QRS complex in all three positions was positive. The T wave was positive in Left, negative in Front and Ventral that varies between the three anatomical locations. We have added the description to the results, L195-199:
“The R waves were positive in all three positions and significantly higher in the Left-6 position (0.52 ± 0.03 mV; example see Fig 1), compared to the Front and Ventral positions (0.12 ± 0.01 and 0.13 ± 0.01 mV, respectively; P < 0.0001). The T wave was positive in Left-6 and negative in Front and Ventral position. Its amplitude was significantly higher for the Left-6 (0.15 ± 0.01 mV) compared to the Front and Ventral positions (0.03 ± 0.01 and 0.06 ± 0.02 mV, respectively; P < 0.0001).”
Furthermore the section about the implantable loop recorder has been inserted earlier according the whishes from Reviewer 1, please see L127-143.Line 120 should read “AF detection”. How does the ILR define AF? As more than one P wave between two R waves or absent P wave? Or based on heart rate variability parameters?
Answer: “AF detection” has now been corrected (L132). We have added further information on the detection algorithm – at least as far as it is made public by the manufacturer, L129-132:
“The AF detection algorithm uses a 2 minute ECG interval to detect variation in the RR intervals. The variation is measured by plotting the current RR interval versus the RR interval of the preceding heart beat in a Lorenz-Plot, Fig 2. To refine AF detection, additionally the ILR searches for a P wave between two R waves.”
Furthermore the reader is referred to the manufactures manual for greater detail L142-143:
“For further details on the functionality of the ILR the manufacturer’s manual can be consulted”Line 131: how much time in SR?
Answer: The horses were in SR for about three weeks before the pacemakers to induce AF were turned on. To make this clearer we have included a sentence about this L116-118:
“Following implantation of pacemaker and ILRs, the horses had a resting period of three weeks were the performance of ILRs in sinus rhythm could be evaluated. Hereafter the horses were paced into sustained AF”Line 134: why were these settings chosen?
Answer: These are the predefined standard settings the ILR is shipped with and has been included in the sentence, L148-149:
“When initialising the ILR, the following were chosen by the predefined standard settings….”Line 140-141: the timeline is unclear. How was this related to the experimental AF period?
Answer: We added a timeline in figure 1. The large variation in interrogation intervals was due to the limited availability of the programming device to our facility.Line 145-148: similar to my previous comments, how was this related to the entire duration of ILR recording / AF period?
Answer: Holter recordings were performed on selected time points during the experiments. These recordings were mainly used to perform analysis on AF fibrillatory rate as previously published. Furthermore, Holter recordings was used to ensure correct AF classification by the ILR. The paragraph therefore now reads L161-164:
“Holter ECG recordingsh were obtained before AF induction and various times during AF and analysed manually to verify correct AF detection of the ILR. Holter ECGs were recorded on average for 14.4 days (range 13 – 16 days), resulting in ECGs of diagnostic quality from the Holter recordings of a mean duration of 145 hours (range 106 – 288 hours) for each horse”Line 157-159: why were R and T wave measurements only performed during AF? Isn’t it more difficult to measure the amplitude during AF as the baseline is undulating?
Answer: The ILRs only record the ECGs when there is an arrhythmic event (unless you activate the recording manually) and therefore the ECGs recorded in the present study were mainly AF recordings. The baseline was extremely stable (see Figure 2 for examples), so this was not difficult to measure.Line 162: From this description “Pearson’s correlation to compare AF burden” it seems like only 5 data points were included for this analysis (5 horses) and perhaps a different statistical analysis would be more accurate eg. Bland Altman analysis. In the results section, it becomes clear that the correlation was made for each horse individually, again, it is unclear how many time points were included per horse, and how many of those time points were in SR or in AF. This is crucial to conclude whether the automatic detection of AF works or not.
Answer: During the entire study period Holter ECG at various length (dependent on electrode stability and signal quality) were performed. Furthermore, the presence of AF was evaluated clinically every day. We initially aimed to correlate the hours in AF as seen in Holter with the ILR. However, as the ILR only provides as roughly daily burden and no detailed time information we think it is not correct to perform correlation tests as they are presented here. (comparing hours of AF/day rather than the actual occurrence). We therefore removed the Pearsons correlation test. Instead we now use the Holter ECGs confirmed AF and to compare this with the AF burden histogram provided by the ILRs to confirm that AF is registered on this specific day by the ILR. We have included this in the discussion of the study, L265-273:
“The Reveal LINQ is equipped with a built-in algorithm that refrains noisy signals from being used in arrhythmia detection. The intervention of this algorithm or the amount of rejected signals is not presented to the clinician and a quantitative analysis is therefore not possible. However, the IRL rejected ECG episodes superimposed with artefacts (e.g. recordings during maximal exercise) and did not use them to detect an arrhythmia like AF. Despite this limitation, the AF burden histogram for the specific days showed that AF occurred for almost 24 hours/day for the horses, which indicate that the automatic artefact reduction is low in horses at rest. Further, we did not observe false positive AF episodes during sinus rhythm, before AF induction pacing was turned on.”Line 168: Does this mean that only 10 ILRs were included in the study? This could be more clearly stated in the manuscript. The low number of horses and ILRs is indeed a major limitation, as stated in the discussion.
Answer: We added this clearly to the manuscript L190-192:
“In total, 10 ILRs could be used for analysis and mean duration of ILR time per horse from insertion until the last interrogation of the ILR was 70 days (range 52– 85 days).”Line 169: “interrogation with the ILR” seems like an odd phrase?
Answer: Interrogation is the technical term used to describe the process of receiving diagnostic date and perform programming of the device. We corrected the phrase to: “interrogation of the ILR” (L191).Line 200: aberrant conducted beats with lower R wave amplitude: were these present at rest or during exercise? Aberrant conduction has been described in horses with AF during exercise (Verheyen et al. 2010).
Answer: We believe that the word aberrant may be misleading and therefore removed this from the text. We have reworded the sentence so it reads as follow, L228-229:
“Undersensing was typically observed when intermittent beats with lower R wave amplitude were present.”Line 232-233: no false positive detection of AF due to artefacts: was the exercise test also performed with the horses in SR?
Answer: The exercise testing was only performed in AF. Unfortunately, no treadmill testing during SR was performed.Table 1: the percentage of QRS complexes with preceding f waves in the table is difficult to interpret without the manuscript – how was this calculated?
Answer: We added the following sentence to the table’s footnote L318-319: “f waves: percentage of 10 consecutive QRS complexes that showed preceding f waves.”Line 235-240: This information is difficult to understand without a clear timeline, especially line 239-240. I assume that the horses were in AF for a relatively long period, but the ILR did not record more than 14 AF episodes – is this because persistent AF is only recorded as one episode by the ILR?
Answer: Horses were in sinus rhythm for 3 weeks, followed by approximately 6 weeks of AF (55 days minus the time it took to induce AF). We did not state that only 14 episodes were recorded, but we agree that it was not clearly explained how the memory of the ILR is build up. We have added this now to the description of the Implantable loop recorder, L132-136:
“A list containing all types of arrhythmia and its respective number of occurrence is stored automatically for the total life-time of the device. Furthermore, the reveal LINQ can store a list containing up to 30 episodes in greater detail, including date and time of onset, duration as well as mean and maximal heart rate. This list is complemented with up to 14 ECG traces (examples Fig 2).”Cite this author response
- pre-publication peer review (ROUND 1)
Decision Letter
2020/01/1010-Jan-2020
Dear Prof. Buhl:
Your manuscript ID EVJ-TN-19-302 entitled "Detection of atrial fibrillation with implantable loop recorders – preliminary results and potential diagnostic implication" which you submitted to the Equine Veterinary Journal, has been reviewed.
Decision
We require major revisions to your manuscript before it can be reconsidered for publication. Therefore, I invite you to respond to the comments included at the bottom of this letter and revise your manuscript accordingly.Revision Process
To revise your manuscript, log into https://mc.manuscriptcentral.com/evj and enter your Author Center, where you will find your manuscript title listed under "Manuscripts with Decisions." Under "Actions," click on "Create a Revision." Your manuscript number has been appended to denote a revision. Please DO NOT upload your revised manuscripts as a new submission.All revised articles must be uploaded with
(i) a main document containing a version which does not have a title page and is blinded and in which all changes are highlighted using 'track changes' mode. This version will be sent out to reviewers;
(ii) a document containing a version which includes the title page and in which blinding details have been re-inserted and changes are not marked. This should be uploaded with the file designation 'Clean Non-Blinded Copy'. If your manuscript is accepted, this version will be uploaded to our Accepted Articles section.All revised manuscripts must be accompanied by a point-by-point response, addressing every comment made by the Editors and Reviewers. When submitting your revised manuscript, you must respond to these comments in the space provided.
Your original files are available to you when you upload your revised manuscript. Please delete any redundant files before completing the submission.
Deadline
Your revised manuscript should be uploaded within 12 weeks. Please contact the Editorial Office if you need extra time to make your revisions.Thank you for submitting your manuscript to the Equine Veterinary Journal and I look forward to receiving your revision.
Sincerely,
Celia Marr
Editor, Equine Veterinary Journal
cmm@evj.co.ukPeer Reviewer(s)' Comments to Author:
Peer Reviewer: 1
Comments to the Author
General comments
This technical note describes the use of implantable loop recorders (ILRs) to detect atrial fibrillation in horses. It compares the signal quality obtained from 3 different anatomical sites both at rest and exercise. It also compares automatic detection of atrial fibrillation by the ILRs with manual evaluation of a Holter ECG recording. There is very little published information on the use of ILRs in horses -previous reports have only evaluated signal quality from one anatomical site (the left pectoral area) and have not evaluated a trace at exercise. In addition, these previous reports have not evaluated automatic detection of arrhythmias by the ILR but relied on manual activation of the ILR to save a recording after abnormal clinical signs were observed. This technical note therefore provides useful new information on the use of ILRs to detect arrhythmias in horses. This is clinically relevant as current diagnostic techniques are limited in their ability to detect sporadic arrhythmias.Specific points
The terms P wave, R wave and f waves are variably hyphenated throughout the manuscript. These hyphens should be removed for consistency.
Materials and Methods
I found the timeline of the study difficult to follow – probably because this was a subset of horses used for another study. I think the “Materials and Methods” section could be re-ordered to make the timeline clearer. For example I think the section entitled “Implantable loop recorder” (lines 116-138) in “Materials and methods” would follow better after line 108.
109-112 – It is unclear how this cardioversion relates to this particular study and presumably is because these horses were part of another study. Please could you just clarify this.
113 – “Four horses in persistent AF” – It sounds like only 4 horses were in persistent AF. Were all 5 in persistent AF but only 4 went on the treadmill?
113-115 – Please could you describe what tack the horses were wearing as this may be relevant to artefact generation.
131 – Change “SR” to “sinus rhythm (SR)” as this abbreviation has not been used previously
134 – Please add a definition for “blanking after sense” as you have done for the “sensing threshold”
147-148 – I don’t understand these values – e.g. if a Holter ECG was in place for 14.4 days then doesn’t that generate (14.4 x 24 hrs) 345.6 hours of ECG recording?
151- AF burden histogram – is it possible to add one of these as another panel in figure 2 to allow the reader to visualize the data outputs better?
155-159 – I found this paragraph a bit confusing to read. Perhaps you could specify that R and T waves amplitude was compared between the 2 interrogations. I wasn’t sure what you meant by “taken form the same period” on line 157-158.
Results
180 – There is no mention in the “Materials and Methods” that the horses were subject to euthanasia and post mortem examination – I think this should be included
185 – “At baseline” presumably means “at rest” – perhaps reword?
186 – At what stage after cessation of maximal exercise did the traces became diagnostic again? I think this is relevant as presumably it could still help diagnose some cases of paroxysmal AF only present at exercise (and therefore identify a cause for poor performance) even if the trace is not diagnostic at maximal exercise.
191 – Please clarify what you mean by “kinetics”
201- Insert “amplitude “so it reads “R/T wave amplitude ratio”
Discussion
213 -Insert “amplitude” so it reads “high R/T amplitude ratio”
218-219 – “These authors were not able to detect ECGs during collapses due to excessive artefacts in the acquired signals and therefore no arrhythmia diagnosis could be obtained”. This statement is incorrect and does not reflect what was written in that study. For example, in one horse the paper states that recordings were obtained during 2 subsequent collapse episodes and did not detect any arrhythmias or abnormal complexes.
Examples of traces from horses with ILRs in the pectoral region are published in another paper by some of the same authors - Lyle and Keen (2010) Episodic collapse in the horses Equine Veterinary Education 22 (11) 583-584. Even if artefacts are generated during movement associated with collapse this may not preclude diagnosis of an arrhythmia as the cause of the collapse as the arrhythmia would occur prior to the collapse. However, I agree that based on information from your study that the pectoral position is suboptimal compared to the left intercostal space.
222-224 – The range of ILR implantation time was 52-85 days so I don’t think it is possible to extrapolate to definitively say that signal quality would be good for years, e.g. migration of the ILR over time may occur.
225-226 – Specify that manual analysis of a Holter ECG trace seems to be superior to use of an ILR to detect AF. Use of Holter software algorithms to analyse the trace is likely to misinterpret artefacts in the same way as the ILR algorithms so may also lead to false-positive diagnoses of arrhythmias.
236-242 – The limited interrogation frequency was not a problem in this study (i.e. no episodes had been deleted from the ILR prior to interrogation) but I wonder if this could be due to the fact that the horses were mainly in AF? For example, I wonder if a horse was in sinus rhythm but regularly had episodes of physiological second-degree AV block that these would fill up the ILR quite quickly and then it might be possible that the pathological arrhythmia of interest would be missed? I think it would be worth saying that it would be good to investigate the use of ILR in horses in sinus rhythm to determine how quickly the episodes might fill up and therefore help determine how often interrogation might be required.Peer Reviewer: 2
Comments to the Author
The manuscript describes the use of ILRs for detection of AF in horses. This technical note clearly describes the Left-6 position as the best position for implantation of an ILR, which is interesting information for future studies using ILRs. However, based on the described methods and results, it is unclear how the AF burden was measured using the ILRs and how the comparison with Holter ECG was made. A timeline is missing and it is unclear how many hours/days of AF were included in the study period. This is essential information to draw conclusions about the usefulness of ILRs.
The authors also describe the reduced diagnostic quality due to movement artefacts. Is this also a problem in human patients or other species, and how is it tackled in these species? This would be interesting information for improving the diagnostic value of ILRs, as these probably have most potential for use in exercising horses.
The title is a bit misleading as the potential diagnostic implication does not seem the main focus of the manuscript.
Some sentences in the abstract need to be clarified:
Line 50-51: R and T wave amplitudes measured, why not P wave or S wave amplitude?
Line 51: AF burden registered by ILRs, automatically detected or based on manual analysis?
Line 52: What is meant by ‘ECG signals stable over time’ or ‘suitable to detect AF’?
Manuscript:
Line 71-73: “short lasting episodes may not be diagnosed, and early recognition of the disease is essential for prompt treatment”. One would think that the main advantage of ILRs would be to detect paroxysmal AF, however, which implications does this have for treatment as cardioversion occurs spontaneously? To detect short-lasting persistent AF other approaches have been described such as the use of heart rate variability measurements from heart rate monitors. No reference to this work is made.
Line 80-81 about the use in human asymptomatic patients: these are probably patients with paroxysmal AF?
Line 114-115: how many 10 minute ECGs over the entire test? A little bit extra information about the exercise test would be interesting for the reader, e.g. was this a submaximal or maximal exercise test.
Line 117-118: This information should be in the introduction. How is the R-R interval detected? Based on a positive peak of the QRS complex, a negative peak or any deflection? Was the largest deflection of the QRS complex positive in all ILR anatomical locations? And the T wave?
Line 120 should read “AF detection”. How does the ILR define AF? As more than one P wave between two R waves or absent P wave? Or based on heart rate variability parameters?
Line 131: how much time in SR?
Line 134: why were these settings chosen?
Line 140-141: the timeline is unclear. How was this related to the experimental AF period?
Line 145-148: similar to my previous comments, how was this related to the entire duration of ILR recording / AF period?
Line 157-159: why were R and T wave measurements only performed during AF? Isn’t it more difficult to measure the amplitude during AF as the baseline is undulating?
Line 162: From this description “Pearson’s correlation to compare AF burden” it seems like only 5 data points were included for this analysis (5 horses) and perhaps a different statistical analysis would be more accurate eg. Bland Altman analysis. In the results section, it becomes clear that the correlation was made for each horse individually, again, it is unclear how many time points were included per horse, and how many of those time points were in SR or in AF. This is crucial to conclude whether the automatic detection of AF works or not.
Line 168: Does this mean that only 10 ILRs were included in the study? This could be more clearly stated in the manuscript. The low number of horses and ILRs is indeed a major limitation, as stated in the discussion.
Line 169: “interrogation with the ILR” seems like an odd phrase?
Line 200: aberrant conducted beats with lower R wave amplitude: were these present at rest or during exercise? Aberrant conduction has been described in horses with AF during exercise (Verheyen et al. 2010).
Line 232-233: no false positive detection of AF due to artefacts: was the exercise test also performed with the horses in SR?
Table 1: the percentage of QRS complexes with preceding f waves in the table is difficult to interpret without the manuscript – how was this calculated?
Line 235-240: This information is difficult to understand without a clear timeline, especially line 239-240. I assume that the horses were in AF for a relatively long period, but the ILR did not record more than 14 AF episodes – is this because persistent AF is only recorded as one episode by the ILR?Decision letter by
Cite this decision letter
Reviewer report
2019/12/26The manuscript describes the use of ILRs for detection of AF in horses. This technical note clearly describes the Left-6 position as the best position for implantation of an ILR, which is interesting information for future studies using ILRs. However, based on the described methods and results, it is unclear how the AF burden was measured using the ILRs and how the comparison with Holter ECG was made. A timeline is missing and it is unclear how many hours/days of AF were included in the study period. This is essential information to draw conclusions about the usefulness of ILRs.
The authors also describe the reduced diagnostic quality due to movement artefacts. Is this also a problem in human patients or other species, and how is it tackled in these species? This would be interesting information for improving the diagnostic value of ILRs, as these probably have most potential for use in exercising horses.
The title is a bit misleading as the potential diagnostic implication does not seem the main focus of the manuscript.
Some sentences in the abstract need to be clarified:
Line 50-51: R and T wave amplitudes measured, why not P wave or S wave amplitude?
Line 51: AF burden registered by ILRs, automatically detected or based on manual analysis?
Line 52: What is meant by ‘ECG signals stable over time’ or ‘suitable to detect AF’?
Manuscript:
Line 71-73: “short lasting episodes may not be diagnosed, and early recognition of the disease is essential for prompt treatment”. One would think that the main advantage of ILRs would be to detect paroxysmal AF, however, which implications does this have for treatment as cardioversion occurs spontaneously? To detect short-lasting persistent AF other approaches have been described such as the use of heart rate variability measurements from heart rate monitors. No reference to this work is made.
Line 80-81 about the use in human asymptomatic patients: these are probably patients with paroxysmal AF?
Line 114-115: how many 10 minute ECGs over the entire test? A little bit extra information about the exercise test would be interesting for the reader, e.g. was this a submaximal or maximal exercise test.
Line 117-118: This information should be in the introduction. How is the R-R interval detected? Based on a positive peak of the QRS complex, a negative peak or any deflection? Was the largest deflection of the QRS complex positive in all ILR anatomical locations? And the T wave?
Line 120 should read “AF detection”. How does the ILR define AF? As more than one P wave between two R waves or absent P wave? Or based on heart rate variability parameters?
Line 131: how much time in SR?
Line 134: why were these settings chosen?
Line 140-141: the timeline is unclear. How was this related to the experimental AF period?
Line 145-148: similar to my previous comments, how was this related to the entire duration of ILR recording / AF period?
Line 157-159: why were R and T wave measurements only performed during AF? Isn’t it more difficult to measure the amplitude during AF as the baseline is undulating?
Line 162: From this description “Pearson’s correlation to compare AF burden” it seems like only 5 data points were included for this analysis (5 horses) and perhaps a different statistical analysis would be more accurate eg. Bland Altman analysis. In the results section, it becomes clear that the correlation was made for each horse individually, again, it is unclear how many time points were included per horse, and how many of those time points were in SR or in AF. This is crucial to conclude whether the automatic detection of AF works or not.
Line 168: Does this mean that only 10 ILRs were included in the study? This could be more clearly stated in the manuscript. The low number of horses and ILRs is indeed a major limitation, as stated in the discussion.
Line 169: “interrogation with the ILR” seems like an odd phrase?
Line 200: aberrant conducted beats with lower R wave amplitude: were these present at rest or during exercise? Aberrant conduction has been described in horses with AF during exercise (Verheyen et al. 2010).
Line 232-233: no false positive detection of AF due to artefacts: was the exercise test also performed with the horses in SR?
Table 1: the percentage of QRS complexes with preceding f waves in the table is difficult to interpret without the manuscript – how was this calculated?
Line 235-240: This information is difficult to understand without a clear timeline, especially line 239-240. I assume that the horses were in AF for a relatively long period, but the ILR did not record more than 14 AF episodes – is this because persistent AF is only recorded as one episode by the ILR?Reviewed by
Cite this review
Reviewer report
2019/12/19General comments
This technical note describes the use of implantable loop recorders (ILRs) to detect atrial fibrillation in horses. It compares the signal quality obtained from 3 different anatomical sites both at rest and exercise. It also compares automatic detection of atrial fibrillation by the ILRs with manual evaluation of a Holter ECG recording. There is very little published information on the use of ILRs in horses -previous reports have only evaluated signal quality from one anatomical site (the left pectoral area) and have not evaluated a trace at exercise. In addition, these previous reports have not evaluated automatic detection of arrhythmias by the ILR but relied on manual activation of the ILR to save a recording after abnormal clinical signs were observed. This technical note therefore provides useful new information on the use of ILRs to detect arrhythmias in horses. This is clinically relevant as current diagnostic techniques are limited in their ability to detect sporadic arrhythmias.Specific points
The terms P wave, R wave and f waves are variably hyphenated throughout the manuscript. These hyphens should be removed for consistency.
Materials and Methods
I found the timeline of the study difficult to follow – probably because this was a subset of horses used for another study. I think the “Materials and Methods” section could be re-ordered to make the timeline clearer. For example I think the section entitled “Implantable loop recorder” (lines 116-138) in “Materials and methods” would follow better after line 108.
109-112 – It is unclear how this cardioversion relates to this particular study and presumably is because these horses were part of another study. Please could you just clarify this.
113 – “Four horses in persistent AF” – It sounds like only 4 horses were in persistent AF. Were all 5 in persistent AF but only 4 went on the treadmill?
113-115 – Please could you describe what tack the horses were wearing as this may be relevant to artefact generation.
131 – Change “SR” to “sinus rhythm (SR)” as this abbreviation has not been used previously
134 – Please add a definition for “blanking after sense” as you have done for the “sensing threshold”
147-148 – I don’t understand these values – e.g. if a Holter ECG was in place for 14.4 days then doesn’t that generate (14.4 x 24 hrs) 345.6 hours of ECG recording?
151- AF burden histogram – is it possible to add one of these as another panel in figure 2 to allow the reader to visualize the data outputs better?
155-159 – I found this paragraph a bit confusing to read. Perhaps you could specify that R and T waves amplitude was compared between the 2 interrogations. I wasn’t sure what you meant by “taken form the same period” on line 157-158.
Results
180 – There is no mention in the “Materials and Methods” that the horses were subject to euthanasia and post mortem examination – I think this should be included
185 – “At baseline” presumably means “at rest” – perhaps reword?
186 – At what stage after cessation of maximal exercise did the traces became diagnostic again? I think this is relevant as presumably it could still help diagnose some cases of paroxysmal AF only present at exercise (and therefore identify a cause for poor performance) even if the trace is not diagnostic at maximal exercise.
191 – Please clarify what you mean by “kinetics”
201- Insert “amplitude “so it reads “R/T wave amplitude ratio”
Discussion
213 -Insert “amplitude” so it reads “high R/T amplitude ratio”
218-219 – “These authors were not able to detect ECGs during collapses due to excessive artefacts in the acquired signals and therefore no arrhythmia diagnosis could be obtained”. This statement is incorrect and does not reflect what was written in that study. For example, in one horse the paper states that recordings were obtained during 2 subsequent collapse episodes and did not detect any arrhythmias or abnormal complexes.
Examples of traces from horses with ILRs in the pectoral region are published in another paper by some of the same authors - Lyle and Keen (2010) Episodic collapse in the horses Equine Veterinary Education 22 (11) 583-584. Even if artefacts are generated during movement associated with collapse this may not preclude diagnosis of an arrhythmia as the cause of the collapse as the arrhythmia would occur prior to the collapse. However, I agree that based on information from your study that the pectoral position is suboptimal compared to the left intercostal space.
222-224 – The range of ILR implantation time was 52-85 days so I don’t think it is possible to extrapolate to definitively say that signal quality would be good for years, e.g. migration of the ILR over time may occur.
225-226 – Specify that manual analysis of a Holter ECG trace seems to be superior to use of an ILR to detect AF. Use of Holter software algorithms to analyse the trace is likely to misinterpret artefacts in the same way as the ILR algorithms so may also lead to false-positive diagnoses of arrhythmias.
236-242 – The limited interrogation frequency was not a problem in this study (i.e. no episodes had been deleted from the ILR prior to interrogation) but I wonder if this could be due to the fact that the horses were mainly in AF? For example, I wonder if a horse was in sinus rhythm but regularly had episodes of physiological second-degree AV block that these would fill up the ILR quite quickly and then it might be possible that the pathological arrhythmia of interest would be missed? I think it would be worth saying that it would be good to investigate the use of ILR in horses in sinus rhythm to determine how quickly the episodes might fill up and therefore help determine how often interrogation might be required.Reviewed by
Cite this review
