Could Afib rhythm on Kardia be misinterpreted

[yzmd]

At the end of April I received my 14 day Zio report that showed very frequent rhythms of supraventricular ectopies and no AFib. Most heart rates averaged below 100. A few weeks later I used my Kardia on 1 lead which showed sinus rhythm with supraventricular ectopy at 64 BPM. 1 minute later I used Kardia 6 lead which showed AFib at 74 BPM. I went on chatgpt and asked whether supraventricular ectopy can be misinterperted as AFib and vice versa and answer was yes. I wonder how common a possible misinterpretation of premature atria contractions/AFib is.

[Hidden]

store.alivecor.co.uk/produc...

You need a Kardia Care Membership for the algorithm to report sinus rhythm with supraventricular ectopics.

I don’t know whether that explains part of the problem here.

[mjames1]

Kardia's AI determinations are very accurate, but not perfect. Multiple ectopics could throw it off. So best to have the EKG read by an EP to be sure. You also may be entitled to a complimentary EKG review by one of Kardia's cardiologists. BTW kardia only makes its determinations on Lead One even if you're using it in six lead mode.

Jim

[Buffafly]

Don’t know about common but it is has been reported several times. It is also quite difficult to tell the difference manually if you have numerous ectopics.

[AAJJTt]

Agree with previous comments. I rate my Kardiamobile 6L but it is not infallible. For me, it has been very accurate reporting Afib.

However, I have noticed, 3 or 4 ectopics within a recording can produce a determination of ‘unclassified’ or ‘possible atrial fibrillation’.

I usually see the ectopics on leads II & III and the resemblance of a p wave on lead I, and along with absence of other Afib symptoms, I draw my conclusions.

[CDreamer]

Had this earlier today, my Apple Watch showed AF, then as episode ended showed ‘Undetermined’ so out came the Kardia - AF then SVT ectopy from the 6L. P wave came and went so it kept flipping and out of AF, then SVT ectopy - but that’s not unusual for me and I take it as a sign I’m, on the way back to NSR and indeed I reverted.

[Singwell]

Not expert on different readings but I do know that PACs are precursors of AFib. My EP explained it as AF trying to kick off. So if you got one short reading of AF amidst lots of PACs maybe that's not so surprising. And re chatgpt - very useful for gathering info but be careful about which ones you use. I gather there are more specialist search engines for academic and medical queries. And of course, chatgpt cannot diagnose.

[Ppiman]

It seems to me that all AI based ECG readings are prone to errors. Both Wellue and Kardia have been shown in recent posts to suggest even the worrying VT as a determination when the cause was, in both times, atrial. I think the heart varies in some people to render AI less useful. In my case, my cardiologist says that my bundle block was the cause of mine mis-reporting.

Interestingly, the Apple Watch I use has never once misreported - but it only shows AF, of course. Last night it proved very useful to me.

Steve

[Buffafly in reply to Ppiman]

Hi, which BBB do you have? I went into AF yesterday and was a bit alarmed by the appearances of my Kardia trace as all my R waves had double peaks or wide flat tops, no Q waves but I never do. A bit of Duckduckgoing revealed I seem to have LeftBBB which apparently needs investigation but as I have had various investigations before I’m assuming it’s known about but not mentioned. I thought of calling 111 but they were bound to want me to go to A&E so I decided to see how I am today - things have quietened down but funny R waves persist.

[Ppiman in reply to Buffafly]

I have LBBB but I gather most people with it have RBBB. I think a twelve lead ECG is needed to work out which. The R wave is broad on mine occasionally with some double peaks. It comes and goes as yours likely does. It’s said not to be of great concern but, unlike what I’ve read, I wouldn’t call it symptomless. There’s another member here who described it well recently.

I had a long episode of AF last night. I hope you felt okay during yours. Mine was the usual (and they are becoming much more usual this year…): heavy palpitations. We were on the last day of our holiday in Suffolk, too, and so this morning is the long drive home! I managed to sleep with the help of a tablet, despite the AF.

Steve

[Ennasti in reply to Buffafly]

Do you have an image of your trace? My LBBB is very distinct. It doesn’t matter if it’s single lead or 6 - it’s clearly LBBB with the double peaked wide QRS.

[Buffafly in reply to Ennasti]

I have just checked, gone back to NSR and result is ‘SR with wide QRS’ for the first time. It’s a wide double peaked to flat top R wave with slight dip and then very low rise up followed by a wavy line - I’ll try to take a screenshot. This is new so obviously something happened last night, should have called my helpline who would have sent an ambulance but too late now 😕 Feel OK but a bit breathless.

[Buffafly in reply to Buffafly]

Rhythm strip

NSR

[Ennasti in reply to Buffafly]

Yep, looks like mine. I’ve never called an ambulance as for me it’s not an urgent thing - not sure what else you’re affected by though. Lots of us live with an lbbb. Here’s mine. In this strip it’s flicking back and forth between normal and lbbb.

Ecg strip showing lbbb

[Buffafly in reply to Ennasti]

Yes, I think mine used to do that but now constant. My concern was the sudden change combined with a very rough episode of AF including throat pain which made me worry that ‘something happened’ but now just an extra to add to the mix - leaky valve, artery spasm, asthma, PAF…..

Thanks for reply and information

[mjames1 in reply to Ppiman]

Steve: Both Wellue and Kardia have been shown in recent posts to suggest even the worrying VT as a determination when the cause was, in both times, atrial.

Wellue, yes, but VT is not a Kardia determination.

Jim

[Ppiman in reply to mjames1]

You missed the post, I guess, in which it was mentioned? It was an advanced determination that Kardia apparently provided after the trace was sent to them, indeed it seems (from what was said) that twice they got it wrong, with one AI and one a technician / medic determination. It hadn't occurred to me that Kardia use AI, too, which, had I thought about it for even a moment, I would have realised.

I was toying with buying another Kardia last week, rather regretting my selling it, until that thread and then the realisation that the list of arrhythmias that they advertise even in the "Advanced" category is next to none if one excludes tachy- and bradycardia - which hardly needs a device to determine. Essentially all the Kardia will determine is AF, PVC, SVE and "wide QRS". As I sometimes have the latter thanks to intermittent LBBB, Kardia frequently determined only "NSR with wide QRS" to cover all and any ectopics that happened during the trace. That is a frustrating determination to pay for as it surely ought to separate PVCs from PACs (at a cost in the UK of around $13.00 / month).

I was taken aback, too, to read the large number of negative "Trust Pilot" reviews of the Kardia, since the Amazon reviews are generally all pretty good. Maybe have a look yourself and see what you think. I have to say, I found mine a very good device indeed, except for the failings I mention above.

I have come to the conclusion that the Wellue devices, and especially the small handheld one with touchscreen at about $80, despite the occasionally flawed AI, are something of a bargain. I would say that the VT determination needs to be sorted by them, most certainly, and I will be writing to them about this once I have had a further discussion with my EP, but, as the other thread shows, any "aberrancy" or "wide QRS" can fool both devices. What is outstanding with the Wellue is the trace can be done using gel electrodes, which is, naturally far less noisy and shows much more detail of use to an EP. Depending on the positioning of the two electrodes, also, different ECG "views" of the heart are possible - something I hadn't realised.

I had the chance to try a different make yesterday, too, by HealForce. It was an old-fashioned looking device and a bit large but very much like the Kardia in having three contact points but it could also be used "noise free" with up three gel electrodes, much like the Holter I have just used.

I wish there were a three electrode / six view home ECG device around that gave better AI determinations, but it seems not.

Steve

[mjames1 in reply to Ppiman]

So yes, I missed the post, but again VT is not one of Kardia's advanced determinations. And if a kardia tech said it looked like VT that is still not an advanced determination simply the tech's opinion.

I don't know if it's changed or not but at least in the United States we used to have a choice between a tech and a cardiologist to look at a tracing. The tech was less expensive than the Cardiologist. I think you get what you pay for.

Personally, I trust an electrophysiologist over both of them. But this is not a deficiency of the Kardia device just an added service that has its plusses and minuses.

My electrophysiologist's PA, probably more advanced than kardia's tech, also saw VT in one of my tracings. The EP explained that it was simply an artifact. That particular tracing was from my Apple Watch.

I have both the watch and the Kardia and I find them both invaluable, but neither is perfect

Jim

[Ppiman in reply to mjames1]

In the UK, Kardia Care includes “advanced determinations”, which included one reading that can be referred to them for professional analysis, I gather, every month or so. That was what produced the VT determination - not the app AI itself, of course, but their “advanced determinations” via email, as it were.

Clearly the device’s own AI is limited in what it can determine to just a very few arrhythmias indeed: sinus rhythm with … PVC, QRS or SVE, and, of course AF.

This is far fewer than Wellue determines (using an FDA approved AI software, assistant) and for free - and with a device that is half the price, which can be noise free and which analyses not just the first 30 seconds but the whole trace of up to five minutes. I think that is a bargain, frankly and very useful for anyone who has problems with ectopic beats as well as AF. As I say, they need to sort out the VT error but on an aberrant or wider QRS with LBBB trace even a trained eye struggles, I gather.

The Kardia is useful to determine QRS, for sure, but it cannot determine the type of block nor determine the nature of any ectopic beats that occur when the QRS is wide (I.e. during a block of any kind).

Steve

[SuziElley]

I was originally diagnosed with AFib but when seen by EP that was changed to atrial fibrillation and ventricular together with supraventricular ectopics, so yes to your question

[Snowgirl65]

I also questioned my Kardia readings a few months ago that indicated I was in a-fib. After my EP checked out the actual print-out, he said I wasn't in a-fib after all. I don't think the device can pinpoint with minute accuracy, but just give a generic indication.

[Ennasti]

My EP and cardiologist review my Kardia and watch readings to double check what it says. Most often it seems both are correct.

[yzmd]

The misinterpretation seemed to have been very plausable because the 30 second ecg by Kardia showed afib at 64 BPM but 2 weeks earlier t he 14 day continuous 24 hour/day Zio ECG showed no AFib but many pacs daily at below 100 BPM.