No afib since on flec -- 24/7 consumer ecg device?

I used to be able to tell when I was in afib, presumably all afib episodes. The 'symptoms' were more a feeling than anything overt (unless exercising). Our kardia 6L report would always confirm my feeling.

I haven't had any such feelings since being on flec. About a month now. Used to flec+card, but now just flec. I used our 6L a dozen times since starting the RXs, but it has never reported 'poss afib'. I haven't felt any episodes, so I just stopped checking to see what the 6L said.

I'm wondering if I'm having any unnoticeable afib episodes, and am looking for member experiences with a 24/7 consumer device that I can wear to monitor it, so see if I am having any unnoticeable episodes. I don't need the watch to report afib, just need it to alert me to check for afib with my 6L. If it is full of false positives, or if it never detects a positive, I'll probably stop wearing it after a week.

One website lists few watches ...

• Fitbit Sense Advanced
• Withings Scanwatch
• TicWatch Pro 3 Ultra
• Veepoo Health
• Garmin 010-02173-11 Venu,

I won't be going with an apple or amazon watch, or anything over, say, $250, since I expect it never go into alert mode. I just want to check for any unnoticeble episodes. Or should I ask my EP for a 7-day Holter?

I know this doesn't work for many afibbers, but feeling my carotid pulse for 3-5 seconds easily identified my afib. When I went into an episode, I could always feel it, then confirmed with the VERY irregular carotid pulse.

My partner can feel my pulse, I can't, but like you I always knew when I went into an episode. So those were the noticeable episodes. Maybe I had other episodes that flew by under my radar, so to say. Now I am wondering if the flec has truly eliminated all episodes, or if some are still occurring, but are diminished enough (in duration or severity or ...) by the flec to the point of being unnoticeable to me, but would be picked up on an ecg, say a Kardia 6L.

This guy posted about a consumer ECG device, don't know the cost: [www.afibbers.org]

I have used Polar H10 straps with an R to R (beat to beat) recording app that was built for researchers. This is a techy solution as you need to download the data through DropBox to a Windows PC (or virtual Windows PC on a Mac). Then do a visual examination of the data using no longer supported Polar software (I have an install file). Not a popular solution, though I've taught a few that were tech adept and interested here and they've told me they've gotten good benefit. You really want beat to beat data as a hallmark of afib is very high the beat to beat variability. This is masked even with readings every second. I have a ring that records every 4 seconds and it really mutes the variability. With the Polar, I find I get the best data, sans artefact if I'm sleeping to be sleeping on my back in a recliner. The total cost of this solution is under $100.

If your afib heart rate is high, you can see that in the app, though there can be other reasons for a high heart rate.

This is about a four hour recording with 1.5 hour afib in the early morning.


From the app:


Start of the afib, image from zoomed data on Mac virtual PC.


Full 4 hour recording with afib in the middle, image from zoomed data on Mac virtual PC. Afib stands out like a sore thumb.

celia thaxter poems

Another thought. Does your afib convert on its own, or with meds or ECV? If not on its own, then even once a day sampling would tell you that you likely haven't had afib during that day.

If they do self-convert, how short a run are you trying to catch? The methods in my above post should catch just about every episode, but that might not be necessary. Say if you sampled with your 6L three times/day (waking, middle of day, bedtime), you'd know you didn't have episodes longer than 6 ish hours during the day and 8 at night and likely much shorter episodes, if any.

You can have PACs and not AF. I can feel PACs when I'm in bed and trying to fall asleep. Thankfully, they wane in aggressiveness as I calm and relax, so they don't interfere with my sleep.

Most smart watches will record one's HR rather accurately, but they may show an HR of 67 instead of your average of 60. In those cases, you probably have PACs, and they can also give one a sensation that something unwelcome is going on. More forceful beats now and then, or what I term 'thumpiness'. You are not likely to be in AF until your HR climbs closer to 85 and on up to about 120---ish. Above that, you are just as likely to be in flutter as AF.

Unless you have real-time and constant monitoring with at least two electrode sensors, you are not likely to get a very accurate reading compared to a full ECG. A hospital ECG will record all electric events, including sensing their location because of time delays in pickup, sort of like triangulation.

Why would you expect that flecainide would make your usually identifiable episodes suddenly stealthy? It doesn't do that. If you were symptomatic before flecainide, you'll be symptomatic after flecainide. I'd say the flec works for you, be grateful for it, and leave it at that.

I found a different between tambocor and flecainide generic. Day and night difference. I got it online at a fantastic pharmacy.

Quote
GeorgeN
Another thought. Does your afib convert on its own, or with meds or ECV? If not on its own, then even once a day sampling would tell you that you likely haven't had afib during that day.

If they do self-convert, how short a run are you trying to catch? The methods in my above post should catch just about every episode, but that might not be necessary. Say if you sampled with your 6L three times/day (waking, middle of day, bedtime), you'd know you didn't have episodes longer than 6 ish hours during the day and 8 at night and likely much shorter episodes, if any.

Your posts are always useful, GeorgeN, thanks!

Self-convert, historically.

My beat intervals were highly variable back when I was having noticeable episodes. Flec may have altered that as well.



If those are 60 sec +- afib flase starts, then I'd want the device to alert me to those within 20 seconds of starting -- almost time enough to get a 6L recording. In a more perfect scenario, I'd learn of even shorter episodes, say 20 second duration. For example, situations where the electrophysiology is going through the initiation phase, but the flec dosage is weighting/dampening it back onto the NSR limit cycle before it can escape. Basically, at this early point in the flec (a few weeks), my heart electrophysiology and tissue has not changed a bit. Pathological remodeling of the LA was triggered by my previous episodes over the, likely, years, according to the literature (have not had an echocardiogram or anything yet). Now, with zero episodes of the noticeable type and possiblly no other episode types, that pathological remodeling is starting to reverse, according to the literature. But a month with zero episodes is not long enough (my speculation) for noticeable reversal. I ought to look more into the mechanisms of flec.



Edited 1 time(s). Last edit at 10/04/2022 10:01AM by Kwilk.

Quote
susan.d
I found a different between tambocor and flecainide generic. Day and night difference. I got it online at a fantastic pharmacy.

Good to know in case any subsequent monitoring reveals that some 'under the radar' episodes are occuring.

Quote
Kwilk

My beat intervals were highly variable back when I was having noticeable episodes. Flec may have altered that as well.

If those are 60 sec +- afib flase starts, then I'd want the device to alert me to those within 20 seconds of starting -- almost time enough to get a 6L recording. In a more perfect scenario, I'd learn of even shorter episodes, say 20 second duration.

For the questions you are asking, a holter with the ability to go look a specific traces is likely what you need. I say this not having used a Holter since I borrowed one from a friend in the late 2000's. The key is what the software will give you access to and if it is a Holter supplied by a doc, whether you'd even get to see those data. I haven't looked recently, but there are Chinese made Holters available on eBay. Again the software would be the issue.

What is the resolution of the data on your heart rate vs time graph. If beat to beat, you can tell a lot. Can you zoom in? If beat to beat, to my eyes it looks like some kind of heart rate increase associated with sleep with PAC's along the way and a run of 3 or 4 at the peak. Other than the PAC's, beat to beat variability looks like NSR. The troughs after the PAC peak are a compensatory beat. This means that the beat after the compensatory beat would start at the same time as it would have if there had not been a PAC/compensatory beat couplet. PAC's can occur with or without the compensatory beat.

My experience is that flec won't change the beat to beat interval during afib, until conversion happens. In the example I posted above, I took a loading dose (300 mg, in my case) probably 5 or so minutes after the episode started. I chew the flec so it gets into my system faster, but it will still take a while. In any case the variability did not change till I converted. This has always been my experience.

Quote
GeorgeN
For the questions you are asking, a holter with the ability to go look a specific traces is likely what you need.

Thanks, I'll look into that.

Quote
GeorgeN
What is the resolution of the data on your heart rate vs time graph.

Sorry, I should have mentioned that it was a crop of your graph.

Quote
GeorgeN
My experience is that flec won't change the beat to beat interval during afib

Let me understand, I know that your statement about flec during afib. Also your middle graph in your post seems to show that pre-afib hr variability is quite similar to the post-flec variability. Would you say that when not in afib, which is all the time now for me, my hr variability now (under 2x daily flec) is the same as before I started the RX? I ought to look more into the mechanisms of flec before I take up anymore of your time.

Quote
Kwilk
Would you say that when not in afib, which is all the time now for me, my hr variability now (under 2x daily flec) is the same as before I started the RX?.

When I look at HRV averages overnight from my Oura ring data, I see no material difference in Sept & Oct 2021 and Nov & Dec 2021. I took 50 mg flec once a day before bed during Nov & Dec.

Thanks.

As a side, in some other long lost thread here you mentioned nasal breathing and Zone 2 in the same thread. Do you use nasal breathing to keep your workouts below Zone 3 (when that is that is your goal). I was watching an Attia Q&A on Zone 2 last might, and did some browsing. His working def if I understand correctly is that Zone 2 is below lactate threshold, which is about 2mmol iirc, and that whatever watts/kilo effort keeps you under 2mmol, well, that's Zone 2 for you. Diabetetics might be 1watt/kilo, and TourDeFrance might be 5 or 6 watts/kilo, uping another 1-1.5 with EPO. Anyway, I was wondering what is a good proxy for lactate threshold, it seems nasal breathing will keep one below Zone 3.

Quote
Kwilk
Thanks.

As a side, in some other long lost thread here you mentioned nasal breathing and Zone 2 in the same thread. Do you use nasal breathing to keep your workouts below Zone 3 (when that is that is your goal). I was watching an Attia Q&A on Zone 2 last might, and did some browsing. His working def if I understand correctly is that Zone 2 is below lactate threshold, which is about 2mmol iirc, and that whatever watts/kilo effort keeps you under 2mmol, well, that's Zone 2 for you. Diabetetics might be 1watt/kilo, and TourDeFrance might be 5 or 6 watts/kilo, uping another 1-1.5 with EPO. Anyway, I was wondering what is a good proxy for lactate threshold, it seems nasal breathing will keep one below Zone 3.

Yes, short of measuring lactate (which I've not done, though may get a meter when I return from the travel I'm on), nasal breathing is a good proxy as is Phil Maffetone's MAF heart rate. It is nominally 180-age, but Phil has + & - from that depending on your situation. Attia is obsessive, so always does his Z2 on a trainer so he can make sure he is just on that edge. I'm not an avid cyclist, so only learned of Z2 in the more recent past. However John Douillard promoted nasal breathing in the 90's in his book. (and in seminars prior to the book). Effectively doing getting to the same place that Maffetone does. Early on in my afib journey, I learned empirically that I could do long duration activity with nasal breathing and not have that activity be a delayed vagal trigger. I get to the same place with a MAF heart rate if I wear a monitor, which many times I do not. More recently on the nasal breathing, instructor Patrick McKeown also promotes it in his books The Oxygen Advantage and The Breathing Cure available from Patrick or on Amazon. If I had followed Douillard's advice when I learned of it, I likely would have never gotten afib. I didn't have the patience then.

Unlike Attia, I'll do many things as Z2/MAF/Nasal Breathing: Heavy Hands walking, rucking with a 60# pack, Schwinn Airdyne fan bike, fan rower, 90's Nordic Track skier (& will do this with Heavy Hands also), TRX, bodyweight training, jumping rope, alpine skiing at 12-13,000', rock climbing, bear crawling, isometric training with 90 second holds and so on. At 67, I almost never open my mouth to breathe, even hiking in to climb up steep trails with the 60# pack at 7,500' elev. I also do 1-2x/week short duration high intensity intervals on the fan bike. I can do 13 seconds all out, 17 second easy intervals nasal breathing. My objective is to push it to 20:10 nasal breathing. It is even more fun to add KAATSU BFR bands while doing the intervals as they restrict blood flow to the muscles creating a mild hypoxic environment. And for even more fun, use the BFR while doing any of the exercise with an exhaled breath hold.

Thanks. I need to read up more about increasing mitochondria density. Do you know if BLASA does continuous lactate monitoring during their over-the-top physicals?

I know you don’t want an IWatch but if you can find an older cheaper version, you can set the afib alert settings and you will get a warning after 20-30 minutes of continuous afib. It isn’t 24/7 continuous but somewhat worth it if you get episodes lasting more than 30 minutes, especially if you are asleep. I have a graveyard of portable ekg monitors bought in 2004? 2005? 2007?. One is an ActiveEKG. You attach electrodes to your chest and just let it run. It was insanely expensive at $799 and used the Apple Pilot and original Trio phone which I was able to get on eBay. Unfortunately the company went bust (didn’t upgrade to newer phones) and they sold their inventory for $299. I have two in a closet along with other relics. Unfortunately I have a pacemaker which is the most accurate ekg running 24/7 which I download every 3 months. Otherwise I was a fan of the Kardia 6L- the battery is easy to replace.

A few thoughts:

Most modern smart-watches have the capability to measure both BP and HR. To get real-time AFib notifications, you will have to enter the appropriate app's settings and turn the monitoring from manual, or every 10 minutes, to 'continuous', or it will be something like that.

A continuous monitoring setting will drain your watch's batter much more quickly, probably inside 8 hours from a full charge. So, if you really do want the real-time information, you should probably only enable it when the stakes are high for you. For example, when you don't feel right, when you want to know what happened while you slept...that kind of thing. Otherwise, a simple monitoring of HR every few minutes will give you a rough approximation of what your heart is doing.

Lastly, there can be a problem with knowing too much. These devices, always using the haptic buzz to alert you, or beeping, or texting you, will quite probably become intrusive and raise your stress levels. That, alone, can be bad for your health and well-being.

Just some caveats to consider with 'wearable's.

“ These devices, always using the haptic buzz to alert you, or beeping, or texting you, will quite probably become intrusive and raise your stress levels. That, alone, can be bad for your health and well-being.”

I take meds from 5am to 1am -20 hours of buzzing alerts daily until I take it. It wakes people up. I’m use to it but others in this family said their BP is higher from the 15 minutes buzzing every hour or two all day and night long. I space my drugs thus the many alerts. I could just turn it off but I then forget to take them…Cardizem 3/day buzzing every 8 hours, sotalol every 8 hours, Eliquis and 8 others etc. I think life was calmer before smart phones.

Quote
susan.d
20 hours of buzzing alerts daily until I take it. It wakes people up.

The Medisafe App lets you choose from many different alert sounds--you want something that registers but doesn't annoy.

...just not when resting/sleeping. Unless one's situation is quite dire and there is danger to the bearer of the alerting device if a parameter the device is measuring is of critical importance, alerts of any kind are best turned off. Interference with rest is dangerous because it disrupts sleep patterns, usually causing arousals, which in turn lead to increased serum cortisol, high BP, metabolic syndrome and obesity, etc, etc.

I don't think it is likely I have episodes lasting more than, say, an hour because I've 30-second sampled myself with our 6L about a dozen times since being on flec, with the 6L never reporting afib-anything. So in some back-of-the-envelope statistical sense, it's not likely I have hour long episodes on a daily basis, or, less assuredly, on weekly basis.

I'd speculate that the mostly likely episodes are just bumps off the NSR limit cycle which get weighted back onto it rather quickly by the steady-state flec effects. My afib now is at best, latent. All the makings of it are still there, it is just jailed in a cement cell and hard to hear without special attention. So for that, I'd want to see 10 second episodes, for example.

Quote
susan.d
<big snip>on eBay <big snip>

Thanks, I had not thought of that. If my EP has my back, then this would be strictly optional, more a matter of curiosity than something my EP would find actionable if it did indeed turn up some unnoticeable-type episodes.

Quote
Kwilk
So for that, I'd want to see 10 second episodes, for example.

Maybe with a lack of p wave, but I'm guessing the single lead consumer devices won't be able to tell on something that short. Note your Kardia wants 30 seconds. Below is a zoom of one of the heart rate elevations of mine that you zoomed in on. Note there are 7 beats in a PAC run. Maybe I could tell the difference with a full ECG with multiple leads, but it'd be a bit tough.

The consumer device I imagined in the top post would pick up highly variable beat-to-beat intervals after 5-10 beats, then the device would beep. Upon hearing the beep, I'd get a 6L recording, some times missing the episode entirely, sometimes recording 10 seconds of highly variable beat-to-beat, and sometimes recording 30 seconds of it.

Your plot is average BPM vs time. Do you have the corresponding beat-to-beat-variability vs time?

Quote
Kwilk
Your plot is average BPM vs time. Do you have the corresponding beat-to-beat-variability vs time?

My plot is ACTUAL beat to beat heart rate vs time NOT average. Hence, the 7 beats that are likely PAC's near time 2:30:25 visually show the beat-to-beat-variability. I have the actual beat length for each beat in milliseconds in my data file and I can configure the plot to show the data in these units. I'd need to go back to the data file to dig out each actual data point value.

Let me see if I understand. There is one point plotted per beat, where y=60000/L with L being the number of milleseconds since the previous beat? If so, there is no need on my account to dig up the original L data. I can get it from the plot, or just write over the y axis scale, the new values being 60000/y

Yes, you are correct. I still had the graph up in that software, so just changed the scale to ms.



Here are the data in the graph, date, time & beat ms:

2022-06-02 02:29:12 +0000 1161
2022-06-02 02:29:15 +0000 1124
2022-06-02 02:29:16 +0000 1085
2022-06-02 02:29:17 +0000 1124
2022-06-02 02:29:18 +0000 1172
2022-06-02 02:29:19 +0000 1234
2022-06-02 02:29:20 +0000 1224
2022-06-02 02:29:21 +0000 1264
2022-06-02 02:29:22 +0000 1307
2022-06-02 02:29:23 +0000 1269
2022-06-02 02:29:24 +0000 1181
2022-06-02 02:29:27 +0000 1128
2022-06-02 02:29:28 +0000 1091
2022-06-02 02:29:29 +0000 1119
2022-06-02 02:29:30 +0000 1121
2022-06-02 02:29:31 +0000 1114
2022-06-02 02:29:32 +0000 1138
2022-06-02 02:29:32 +0000 1123
2022-06-02 02:29:33 +0000 1092
2022-06-02 02:29:34 +0000 1095
2022-06-02 02:29:36 +0000 1106
2022-06-02 02:29:38 +0000 1108
2022-06-02 02:29:39 +0000 1088
2022-06-02 02:29:40 +0000 1083
2022-06-02 02:29:41 +0000 1059
2022-06-02 02:29:42 +0000 1046
2022-06-02 02:29:43 +0000 1051
2022-06-02 02:29:44 +0000 1051
2022-06-02 02:29:44 +0000 1045
2022-06-02 02:29:45 +0000 1044
2022-06-02 02:29:46 +0000 1044
2022-06-02 02:29:47 +0000 1038
2022-06-02 02:29:48 +0000 1024
2022-06-02 02:29:50 +0000 1004
2022-06-02 02:29:51 +0000 995
2022-06-02 02:29:52 +0000 1000
2022-06-02 02:29:53 +0000 1000
2022-06-02 02:29:54 +0000 980
2022-06-02 02:29:55 +0000 966
2022-06-02 02:29:56 +0000 945
2022-06-02 02:29:56 +0000 931
2022-06-02 02:29:57 +0000 928
2022-06-02 02:29:58 +0000 923
2022-06-02 02:29:58 +0000 911
2022-06-02 02:29:59 +0000 899
2022-06-02 02:30:00 +0000 882
2022-06-02 02:30:02 +0000 878
2022-06-02 02:30:03 +0000 873
2022-06-02 02:30:04 +0000 868
2022-06-02 02:30:05 +0000 863
2022-06-02 02:30:05 +0000 833
2022-06-02 02:30:06 +0000 890
2022-06-02 02:30:07 +0000 841
2022-06-02 02:30:08 +0000 833
2022-06-02 02:30:08 +0000 814
2022-06-02 02:30:09 +0000 816
2022-06-02 02:30:09 +0000 814
2022-06-02 02:30:10 +0000 785
2022-06-02 02:30:11 +0000 668
2022-06-02 02:30:12 +0000 896
2022-06-02 02:30:12 +0000 813
2022-06-02 02:30:14 +0000 776
2022-06-02 02:30:15 +0000 775
2022-06-02 02:30:16 +0000 775
2022-06-02 02:30:17 +0000 772
2022-06-02 02:30:17 +0000 765
2022-06-02 02:30:18 +0000 737
2022-06-02 02:30:19 +0000 630
2022-06-02 02:30:20 +0000 889
2022-06-02 02:30:20 +0000 737
2022-06-02 02:30:20 +0000 759
2022-06-02 02:30:21 +0000 724
2022-06-02 02:30:21 +0000 724
2022-06-02 02:30:22 +0000 733
2022-06-02 02:30:23 +0000 734
2022-06-02 02:30:24 +0000 732
2022-06-02 02:30:24 +0000 736
2022-06-02 02:30:26 +0000 727
2022-06-02 02:30:27 +0000 729
2022-06-02 02:30:27 +0000 722
2022-06-02 02:30:28 +0000 585
2022-06-02 02:30:29 +0000 488
2022-06-02 02:30:29 +0000 609
2022-06-02 02:30:30 +0000 842
2022-06-02 02:30:31 +0000 655
2022-06-02 02:30:31 +0000 709
2022-06-02 02:30:32 +0000 807
2022-06-02 02:30:32 +0000 619
2022-06-02 02:30:33 +0000 987
2022-06-02 02:30:34 +0000 584
2022-06-02 02:30:36 +0000 1036
2022-06-02 02:30:36 +0000 1036
2022-06-02 02:30:38 +0000 1089
2022-06-02 02:30:40 +0000 1135
2022-06-02 02:30:41 +0000 1178
2022-06-02 02:30:42 +0000 1248
2022-06-02 02:30:44 +0000 1307
2022-06-02 02:30:44 +0000 1332
2022-06-02 02:30:46 +0000 1239
2022-06-02 02:30:46 +0000 1254
2022-06-02 02:30:47 +0000 1264
2022-06-02 02:30:48 +0000 1099
2022-06-02 02:30:50 +0000 1116
2022-06-02 02:30:52 +0000 1102
2022-06-02 02:30:53 +0000 1062
2022-06-02 02:30:54 +0000 1010
2022-06-02 02:30:55 +0000 1081
2022-06-02 02:30:56 +0000 1112
2022-06-02 02:30:56 +0000 1077
2022-06-02 02:30:57 +0000 1059
2022-06-02 02:30:58 +0000 1229
2022-06-02 02:31:00 +0000 1250
2022-06-02 02:31:02 +0000 1153
2022-06-02 02:31:03 +0000 1125
2022-06-02 02:31:04 +0000 1170
2022-06-02 02:31:05 +0000 1234
2022-06-02 02:31:06 +0000 1119
2022-06-02 02:31:08 +0000 678
2022-06-02 02:31:08 +0000 1496
2022-06-02 02:31:09 +0000 1312
2022-06-02 02:31:10 +0000 1219
2022-06-02 02:31:11 +0000 1091
2022-06-02 02:31:12 +0000 1164
2022-06-02 02:31:15 +0000 1203



Edited 1 time(s). Last edit at 10/08/2022 11:02AM by GeorgeN.

Thanks so much. The HUGE benefit of your patience with me is that it has now occurred to me that don't really need to know when/if I am having an arrhythmia episode. I can just use a log of R-R intervals. Granted it won't catch all types of arrhythmia. If I find episodes of high variability in R-R, then at that point, my EP will might order a Holter, or I'd have good footing to request one. Then the establishment can then use their state of the art software to score the recording for all kinds of things. Thanks for hanging in there!

So now all I need is an device that will log time stamps of every beat and allow a download in a non-proprietary format.

My current setup is posted above here:

RR data are prone to artifact. I initially used a Polar S810 watch with a chest strap 18 years ago. The watch would store 8 hours of data and you would then download the data to the computer. The data files are ASCII so you can pull then into anything you want. The data I posted is from the app I use. The app exports four different files, one is heart rate, which is on 1 second intervals. This is not good for our purposes. Then there is the RR file which is shown above, it also has a cumulative time in ms which I didn't show. Then there is a file they call Kubios, which is for HRV software of that name. It is just the RR intervals in ms. My graphs are from the ancient Polar software. I that the RR intervals from the Kubios file and paste them into a skeleton Polar data file and then pull that file into their software. Though you can look at and graph the data in Excel, Their program allows me to look at say 8 hours of data quickly. Twenty minute windows show most things I care about and then I can zoom in to look at individual beats, as shown above. A

t the beginning of my afib journey, in 2004 & 2005, I had a two pronged approach. One was balancing electrolytes so afib didn't happen (along with proper training intensity). The second was to create a software/hardware solution that would analyze ECG's in real time looking for a signal to predict afib. My hypothesis was that If I could predict 5 minutes in advance, a dose of sildenafil citrate would increase adrenergic tone enough to ward off the impending afib (my then GP was on board with giving me a script for sildenafil. The prediction piece was very difficult and I never solved it. What I can say is that looking at 8 hours of heart rate data is much easier than looking at 8 hours of ECG's. I'm fortunate that the electrolyte/training approach worked well enough to not need to worry about the prediction.

Back to artifact, movement can cause artefact with the strap. Polar advises not using RR recording during exercise. Their instantaneous heart rate is always a moving average, as is the heart rate shown by the app I use. I mitigated this issue by recording either during meditation, or recording while sleeping on my back in a recliner as I don't move much in this position. ECG gel on the strap instead of water also helps, but I normally just record with water sleeping on my back.

There may be other RR recording solutions in the market that I'm not aware of. Any device that measures HRV has to have it, but actually getting access to the RR data it hard to come by, in my experience. Polar may have a watch that does it, I haven't looked recently. My old Polar S810 watches still work, but they require a Windows machine to transfer the data. I've never gotten the transfer to work on virtual Windows on my Mac. I have plenty of Windows machines, but it is more hassle to get them out and set up just for this one application.

Thanks. We've a couple Garmin Forerunner 305s that my partner uses for tracking open water distance swimming. They record heart rate via a wireless torso strap. But the HR data isn't embedded in the gpx file. I don't consider a strap option to be appealing.

It may just be terminology, but for my purposes the data file really only needs timestamps, doesn't need the calculate RR intervals. Movement could be tracked by smartphone accelerometers, and that data could be used to ignore periods of potential movement-artifact. Some artifact-free data would be discounted in the process of course. Not sure I want to go that deep into it. My use of the data is not for treatment, as your is/was. It's just to detect something that might be occurring, and it is only potentially actionable (not for sure), and only relevant for the next couple of months while waiting to get my heart zapped, if that's what I end up doing.

On other notes ...

Quote
GeorgeN
t the beginning of my afib journey, in 2004 & 2005, I had a two pronged approach. <snip> The second was to create a software/hardware solution that would analyze ECG's in real time looking for a signal to predict afib. My hypothesis was that If I could predict 5 minutes in advance, a dose of sildenafil citrate would increase adrenergic tone enough to ward off the impending afib (my then GP was on board with giving me a script for sildenafil. The prediction piece was very difficult and I never solved it.

Sorry if you've already thought of these things or read about them: Now a days, an easy to pick up language like R probably has some AI modules that could be trained on your personal data to predict onset 5 minutes ahead of time, then compiled for speed and run real time on your streaming ECG data. Or pair up with a research group:

Quote
Web
Advance knowledge of an imminent switch from sinus rhythm (SR) to AF could prompt patients to take preventive actions to avoid AF, like taking oral antiarrhythmic drugs. The question is whether there is information, even if subtle, in the minutes prior to AF to indicate an imminent switch from SR. This paper shows that, for the vast majority of patients, the answer is affirmative. On test data, our algorithm can predict the onset of AF on average 31 minutes before it appears, with an accuracy of 83\% and an F1-score of 85\%. The predictions were based on deep learning and data from 350 patients, plus an external validation of 48 patients. Overall, the proposed method has low computational complexity and can be embedded in common wearable devices for continuous heart monitoring and early warning of AF onset.

Real-time Monitoring and Early Warning of Atrial Fibrillation



Edited 1 time(s). Last edit at 10/08/2022 02:12PM by Kwilk.

Quote
Kwilk
Sorry if you've already thought of these things or read about them: Now a days, an easy to pick up language like R probably has some AI modules that could be trained on your personal data to predict onset 5 minutes ahead of time, then compiled for speed and run real time on your streaming ECG data.

Thanks! That is interesting using RRI data. Here is the full paper: [www.medrxiv.org]

At this point not really attractive to pursue for me. I'm currently at one 1.5 hour episode (converted with a loading dose of flec) in over 49 weeks. I have taken prophylactic flec for much of that time, but only two months were at 50 mg qd, four at 25 mg qd and the rest of the time at 10 mg qd. Per this paper, the flec should not be doing anything as they saw a minor impact from flec at 25 mg bid as well as 50 mg bid. When I return from my current travel, I will stop the 10 mg qd. I did not implement the iodine fully until after my episode on June 2. Per amiodarone data, it takes a while (many weeks) before a person is fully saturated with iodine. The mineral "annecdata" agree with the amio data.



I've been repleting iodine per a protocol by Dr. David Brownstein and my hypothesis is I can replicate the efficacy of organic iodine as amiodarone using mineral iodine (Lugol's solution: 60% iodide and 40% iodine) and paying careful attention to all the details in Brownstein and others' repletion protocols to avoid the negatives assumed by standard of care with high dose iodine (I'm taking 37.5 mg qd) or the negatives associated with amiodarone.. I'm not suggesting others play with iodine without studying the protocols in depth.