Ozempic, Heart Rate Variability, and What the Trials Actually Measured

What happens to heart rate variability after starting a GLP-1 receptor agonist, which trials measured HRV directly, and which populations they were run in. It does not cover whether the change matters for long term cardiovascular outcomes, and it does not cover tirzepatide, which none of the trials here tested.

Heart rate variability falls after starting a GLP-1 receptor agonist, and resting heart rate rises at the same time. The drop holds against placebo, and it holds even while the weight is coming off. A study that tracked people on their own wearables found the heart rate rise was statistically mediated by the drop in HRV, so the two numbers on a device are one effect rather than two. Liraglutide cut SDNN by 33.9 ms and lifted mean heart rate by 8.1 beats per minute in the trial that measured both, though those figures come from people with type 2 diabetes, not from people taking these drugs to lose weight.

The number on the ring changed before anything else did

You started a GLP-1 drug, and within a few weeks your tracker began showing a lower recovery or readiness score. Resting heart rate crept up. HRV came down. Nothing about your training or your sleep changed enough to explain it, and the weight is coming off, which is supposed to move these numbers in the other direction.

That last part is the confusing bit. Weight loss and better metabolic health usually pull HRV up. So when the drug delivers the weight loss and the HRV slides anyway, the obvious conclusion is that the device has lost the plot. I went looking for that explanation and did not find it.

The heart rate rise and the HRV drop are one finding, not two

The wearable study is the one that connects them. It did not simply note that both numbers moved, the authors tested whether the heart rate increase actually ran through the HRV change. It did. That gap between two things moving together and one thing driving the other is the whole question here.

What sits underneath is a shift in how much the parasympathetic side of the nervous system is modulating the heart. A mouse study measured this directly and found that GLP-1 receptor stimulation diminished both high and low frequency HRV power and reduced neurotransmission to the cardiac vagal neurons. But that is an animal result, and it cannot be read as a finding about people. What it does is explain why the human numbers move the way they do.

A randomised trial with both an acute exenatide arm and a 12 week liraglutide arm found resting heart rate rose by 7.5 and 6.6 beats per minute respectively, while sympathetic nervous system activity derived from heart rate variability remained unchanged. Which sounds like a contradiction until you set it beside the crossover trial, where the LF to HF ratio also held steady while SDNN and RMSSD fell. Overall HRV magnitude goes down. The sympathovagal balance index does not. Those are different measurements, and the index a device reports decides which one you ever see.

What this does not settle about a single reading

A 33.9 ms drop in SDNN is a group average from a crossover trial. It is not what any one person should expect to see on their own wrist. The observational work is blunt about this too, since the authors of a 12 week study with 24 hour monitoring before and after described their own findings as hypothesis generating and asked for confirmation in larger prospective studies.

It also says nothing about duration. Every trial here ran for 12 weeks or less, which leaves no evidence on whether HRV keeps falling, settles at a new baseline, or recovers as weight stabilises. A number lower than it was in the spring could be any of those three, and nothing published tells them apart. Not yet, anyway.

A lower HRV reading on a GLP-1 is also not the same signal as a sudden overnight drop, which usually has a different and more transient explanation. One is a sustained shift in baseline. The other is a single night, which is a different question entirely.

Common questions