I learned a lot about the heart over the last 2 weeks, and discovered an amazing similarity between our heart and distributed computer systems. Let me show you how our heart deals with the CAP theorem. Here's my story: Starting in August, I increased my exercise volume to try to make up for some fitness I lost during the summer. For the 6 weeks that followed I averaged 60min of cardio per day, mostly cycling and running. My average before that was ~30min/day. About 70% was low intensity (Z2-Z3), and 30% high (Z4-Z5). In mid September, I felt something strange when I was in bed at night. My heart rate felt slower, and it felt kind of relaxing. At the time, I thought this was a good thing. I also noticed that the Heart Variability Rate on my Apple Watch shot up by 150% from Sep 10. I didn't make the link with the above though. I had opted into the latest watchOS beta and I suspected it was due to a bug or an algorithm change. I felt the same thing in bed for the rest of the week. Then it disappeared for a week. Then it came back the week after. But this time I was feeling it more frequently. I would be quiet at my desk, and I'd occasionally feel a 2 second delay between beats, followed by a big pounding beat. In the meantime, my fitness was improving. I was doing personal best times in 3 mile runs and 6 mile bike rides. I felt very good when exercising and I didn't notice anything wrong with my heart rhythm when it was elevated during exercise. But towards the start of October, these abnormal beats became too frequent to ignore. They were even starting to affect my sleep. On Oct 1, I did an ECG Afib test on my Apple Watch and it came "Inconclusive". I had never seen that outcome before. I repeated it a dozen times, and it kept arriving to the same result. The next day, I did a brief workout and after my heart rate was down to ~80bpm I repeated the Afib test. It came positive: The next day I went to my doctor and get hooked up on a 12-lead EKG. Within just one 10 second sample, we caught one of abnormal beats. It wasn't Afib, luckily. The diagnosis from the EKG monitor was "Premature Ventricular Contraction", which was confirmed by a radiologist a day later: My doctor told me this is usually a benign thing... unless it was "too frequent", which would be above 10,000 PVCs per day. I did some quick math in my head, and I was sure I was at that threshold or thereabouts. Doctor still tried to reassure me, but when I went home I did some research and found several recent studies which linked a PVC burden of just 70 events per day to significant increases in mortality, cardiovascular events, sudden cardiac arrests, heart attacks, dangerous arrhythmias, stokes, and may other bad things. One study that followed up with 7,500 participants with no known cardiovascular disease found that 75% of those who had shown shown one PVC event on an EKG (like me) were dead after 18 years, while only 20% from the control group with no PVC had died. The studies on athletes with PVC were even more concerning, with some strong links to sudden death. My doctor gave me a 14 day heart rate monitor, which was going to count all PVC events, and the circumstances in which they happen. As a precaution, I stopped all my exercise. Yet, the frequency of my PVCs keep increasing until Oct 11, where they were happening every 4 or 5 beats. Because of the very evident high frequency, I got to do a bunch of other tests to try to rule out any damage in the heart that might be causing them. First, was blood work. It looked for electrolyte imbalance, inflammation markers, thyroid issues, and a few other things. All came normal. Then I did an echocardiogram (ultrasound) of the heart and main arteries. It found everything normal. The left atrium had slightly higher volume than normal, but all the wall thicknesses were normal. There were no signs of inflammation (one of the covid risks that could last several months), the valves were working well, and my ejection fraction (the amount of blood the heart pumps each time it beats) was calculated to 60-65%, which is good. No plaque or narrowing of the arteries. No heart visible damage. Then I also did a CT coronary angiogram. My LDL cholesterol level has always been borderline or slightly elevated, and this test would look for any blockages in the vessels that supply oxygen to the heart. The test produces a 3D imagine of the heart and the blood flow gets enhanced with a contrast dye that looks bright on the CT scan. This also went well. No narrowing was found, and my calcium score was zero. My risk of heart attack within the next 5yrs was zero according to the radiologist. And finally I got another EKG which shed some light on what was actually happening. This time the EKG finding was "Incomplete Right Bundle Branch Block", which in the original EKG was either missed, or the signal was disrupted by the PVC event. This is when there's a reduction of electrical conductivity in the wall that separates the left and right ventricles. It can lead to all sorts of bad things, but mine is considered a benign variant (based on other signs on the EKG), and something that 35-50% of endurance athletes have developed. There are no known harmful outcomes for the variant I have (good news), and it usually disappears when people stop exercising for a couple of months. So here's what's likely happening: The contractions of our heart are not controlled by the central nervous system, like most of our other muscles. Instead, the heart is able to generate its own spontaneous electrical activity to activate its muscles in a fully autonomous and decentralized manner. The master pacemaker for the electrical activity to start and generate the heartbeat contractions is called the SA node, which sits at the top of the heart. When it generates activity, the top 2 chambers of the heart (atria) contract to push blood to the lower ventricles where the main pumping function happens. But the lower ventricles should only contract after a small delay, or else the blood wouldn't have enough time to make it from the atria to the ventricles. So the SA node only activates the atrial muscles (not the lower ventricles), and the electrical activity gets absorbed by a secondary pacemaker structure called the AV node. The AV node delays the conduction for about a tenth of a second (enough time for blood to move from the top to bottom chambers), and then it forwards the electrical activity through the middle wall and down to the ventricles. The ventricles then contract which push blood to the body and lungs, and the cycle repeats itself. (I'm simplifying A LOT of course.) What's interesting is that the AV node can also act like a backup pacemaker if the SA node doesn't do its job. If the SA node doesn't initiate a heart beat for some reason, the AV node will initiate its own electrical activity, contract the atria, and send the electrical activity down to the ventricles for them to contract. What's even more fascinating is that this system is completely decentralized. There is no "brain" watching for missed SA node activity to failover to pacemaking function to the AV node. Instead, there's a hierarchy of several fallbacks (not just SA and AV nodes), and they all watch each other. If they never see activity from the higher up node, they start their own heart beat. The way it works is quite simple. The master SA node is configured to beat at, say, 1000ms intervals. The AV node is configured to beat at 1200ms intervals. When the AV node receives electricity from the SA node, it resets its timer. So basically, if the AV node fails to hear anything from the SA node in 1200ms, it would start its own heart beat. But the redundancy doesn't stop with the SA node and AV node. There are actually 3 other parts of the heart that have the automaticity to trigger a heart beat on their own if none of the other nodes do their job. Just like the AV node runs at a slower frequency than the SA node, all the other fallback structures run at increasingly slower rates. Then when they receive electricity from up the hierarchy, they all reset their timer at the same time without triggering a beat. Here's an illustration of all the conductivity structures and their location. All of these except 2 and 7 have the capability to trigger a heart beat, but in normal conditions 1 is the true pacemaker: So here's what seems to have happened to me: Because of my increased exercise load, my heart did some cardiac remodeling to adapt to the increased demand. The left atrium increased in volume (confirmed by the ultrasound), which compressed the cells of structure 5 (the bundle branches). This compression led to slightly slower conduction velocity, which means that structure 6 was receiving the electrical activity a bit later than normal. And when this delay happened to be a bit too late, it thinks that the heart beat didn't happen and spontaneously triggers its own beat. This is the PVC deflection shown in my original EKG. When this structure triggers its premature beat, there would be almost no blood in the heart (because a regular beat would have just happened a few milliseconds earlier). But since this generates new electrical activity, all the other pacemaking structures reset their timers, and that's how I end up with a skipped heart beat. Then the beat that comes after the skipped beat finds the heart with significantly more blood than usual, which causes the pounding sensation that was making it hard for me to sleep at night. Basically, in terms of the CAP theorm, my heart experienced a partitioning problem. One part of the heart wasn't aware of the real state of the system. But the heart is designed for max availability over consistency, so when in doubt, it triggers a beat. For a full cardiac arrest to happen, all 5 pacemakers or conduction pathways need to fail. The heart is an AP system (highly available and partition loss tolerant, over perfect consistency). Pretty remarkable.