Drugs That Distort Heart Rate Variability (HRV): What Skews Your Results

Heart rate variability (HRV) is the beat-to-beat variation in the time between heartbeats, measured from an ECG or a photoplethysmography-based wearable and reported in metrics such as SDNN, RMSSD, and the LF/HF ratio. It is not a brand name or a single test; it is a family of time-domain and frequency-domain calculations derived from R-R interval data, and different devices (clinical Holter monitors versus consumer rings and watches) use different windows and algorithms that are not interchangeable.

The useful question is not "is my HRV normal" but "does this HRV number reflect my autonomic nervous system or my medication list." Several common drug classes change HRV readings by acting directly on the sinoatrial node or on autonomic neurotransmission, independent of whether a person's underlying vagal and sympathetic function has changed at all. Beta-blockers and the sinoatrial-node blocker ivabradine tend to raise time-domain HRV metrics; anticholinergic drugs, opioids, and stimulants tend to lower them; SSRIs and GLP-1 receptor agonists have mixed or agent-specific effects. A person who starts or stops one of these drugs and sees their wearable HRV jump or drop is very often watching pharmacology, not fitness.

What is established, what is plausible, and what is not established

This is the boundary the rest of the article works within, and it matters more here than a list of numbers does.

Established: Drugs that block sympathetic input to the heart (beta-blockers) or block vagal muscarinic receptors (anticholinergics, atropine) predictably move HRV in the expected direction, because the mechanism is direct autonomic pharmacology, not an indirect or speculative pathway. This is consistent with basic cardiac electrophysiology and is reflected in HRV methodology guidance that instructs researchers to record and control for concurrent medications (Quintana et al., 2019).

Plausible but not precisely quantified in this draft: Many of the specific effect sizes circulated for individual drugs (for example, an exact percentage increase in SDNN on a named beta-blocker dose, or an exact percentage drop in HF power on a named opioid) come from a scattered and sometimes small clinical literature. This draft does not carry forward exact percentages tied to sources that could not be verified against the specific claim, because a mismatched citation is worse than an honest range. Where a number is retained, it is described as "reported in small studies" rather than as an established constant, and a qualified reviewer should confirm the figure against the primary paper before publication.

Not established: There is no validated formula for "subtracting out" a drug's effect to recover a person's true unmedicated HRV. Wearable-reported HRV scores also have not been validated against clinical HRV standards, so a wearable trend and a clinical Holter trend should not be treated as the same measurement even before medications are considered.

Beta-blockers: the most common HRV inflator

Beta-adrenergic blockers (metoprolol, atenolol, bisoprolol, propranolol, carvedilol) reduce sympathetic drive to the sinoatrial node. That mechanism lengthens and diversifies the R-R interval, which raises time-domain HRV metrics like SDNN and RMSSD and typically raises high-frequency (HF) power as well. Small clinical studies have reported meaningful increases in SDNN and RMSSD within days to weeks of starting a beta-blocker, and non-selective agents (propranolol, carvedilol) appear to raise HF power more than cardioselective agents, plausibly because beta-2 blockade also deepens respiratory sinus arrhythmia. Exact percentage figures for specific drugs and doses require verification against the primary literature before being stated as fixed numbers.

The clinical point does not depend on the exact percentage: if a wearable RMSSD rises sharply after starting a beta-blocker, that rise is the drug acting on the sinoatrial node, not a sign that the autonomic nervous system itself became healthier. The 1996 Task Force standards for HRV measurement, still the reference document for the field, note that agents affecting autonomic tone must be accounted for when interpreting HRV data rather than compared directly to unmedicated population norms.

Anticholinergic medications: stripping away the parasympathetic signal

Anticholinergic drugs block muscarinic acetylcholine receptors at the sinoatrial node, which suppresses the vagal input that generates the high-frequency HRV component. This class includes medications many patients do not associate with the heart: oxybutynin for overactive bladder, diphenhydramine, tricyclic antidepressants such as amitriptyline, some older antipsychotics, and inhaled agents like ipratropium and tiotropium. Cumulative anticholinergic burden across several of these drugs at once (a common pattern in older adults on multiple prescriptions) is plausibly associated with lower RMSSD and HF power in observational data, though the exact magnitude reported in any single study should be checked against the source before being used as a clinical benchmark.

Atropine, the prototypical muscarinic antagonist, is used experimentally to abolish respiratory sinus arrhythmia almost completely, which is part of how researchers confirmed that HF power is vagally mediated in the first place. That research use is a helpful confirmation of mechanism; it is not a reason to expect the same magnitude of suppression from a single anticholinergic tablet at a normal clinical dose.

SSRIs and SNRIs: direction depends on the specific drug

Selective serotonin reuptake inhibitors do not have one uniform HRV effect, and this inconsistency is itself clinically relevant because untreated depression is independently associated with lower HRV. That creates two competing signals in a depressed patient starting treatment: the disease process (which may improve) and the drug's own autonomic pharmacology (which varies by agent).

Paroxetine has more anticholinergic activity than other SSRIs and has been reported to reduce HF power in small trials. Sertraline has shown a neutral-to-slightly-favorable HRV signal in cardiac populations studied for depression treatment safety, though differences from placebo in at least one such trial did not reach statistical significance. SNRIs (venlafaxine, duloxetine) add norepinephrine reuptake inhibition, which can raise resting heart rate by several beats per minute and lower overall HRV, with the effect appearing more pronounced at higher doses. None of the exact percentage changes attributed to these drugs in older summaries should be treated as confirmed without checking the primary source; the reliable takeaway is the direction (paroxetine and venlafaxine trending HRV down, sertraline trending closer to neutral) rather than a specific number.

A decision framework: is this HRV change the drug or the person?

Use this sequence before drawing any conclusion from an HRV reading, whether it comes from a clinical recording or a wearable export. It does not replace a clinician's evaluation, and it is not a substitute for cardiovascular autonomic reflex testing (CARTs) when a formal diagnosis of autonomic neuropathy is being considered.

Known direction of common drug classes (for step 2):

Two failure modes matter more than any single number. An artificially elevated HRV from a beta-blocker or ivabradine can mask a real, developing autonomic problem such as diabetic cardiac autonomic neuropathy, where a genuine decline in HRV is an early warning sign that a drug-inflated reading could hide. An artificially suppressed HRV from anticholinergics or opioids can trigger unnecessary alarm or testing when the nervous system itself may be functioning normally. Neither error is solved by picking a single "corrected" number; both are reduced by tracking trends on a stable regimen and by naming the drug explicitly whenever an HRV value is reported or discussed with a clinician.

Opioids, stimulants, and other agents

Opioids suppress both autonomic branches but disproportionately reduce vagally mediated HF power, an effect reported to appear within an hour of acute intravenous dosing in small studies. Chronic opioid therapy has been associated with lower SDNN and RMSSD in cross-sectional data on patients with long-standing pain, and methadone carries an additional confound because it can prolong the QT interval, which may distort automated HRV algorithms that assume a normal repolarization pattern. Buprenorphine, a partial agonist, appears to have a milder suppressive effect than full agonists in the available literature, so a switch from methadone to buprenorphine could raise HRV metrics without reflecting a true autonomic improvement.

Stimulants (amphetamine-based ADHD medications, methylphenidate, pseudoephedrine) increase sympathetic outflow and are reported to shift the LF/HF ratio toward sympathetic dominance while reducing RMSSD. Supraphysiologic thyroid hormone dosing, whether from over-replacement of levothyroxine or from endogenous hyperthyroidism, raises resting heart rate and has been associated with lower SDNN and HF power; even modest TSH suppression used intentionally in some thyroid cancer treatment protocols is a plausible confound for anyone tracking HRV for fitness or recovery purposes. Caffeine above roughly 300 mg (about three cups of brewed coffee) has been reported to reduce HF power and raise the LF/HF ratio for two to four hours after ingestion, which is a simple and often overlooked source of day-to-day HRV variability that has nothing to do with a prescription at all.

GLP-1 receptor agonists

GLP-1 receptor agonists (semaglutide, tirzepatide, liraglutide) are associated with a modest increase in resting heart rate in cardiovascular outcome trials, an effect documented as of trial publication dates in the mid-2010s for semaglutide specifically. This heart rate increase mechanically lowers time-domain HRV metrics even when autonomic function is unchanged, and the leading proposed mechanism is direct GLP-1 receptor activity at the sinoatrial node rather than a general sympathetic surge, which means the usual "higher LF/HF equals more sympathetic drive" interpretation may not cleanly apply here. Patients starting semaglutide or tirzepatide who track wearable HRV should expect their baseline to drift lower for reasons that are not necessarily concerning, but any large or unexplained drop in HRV alongside other new symptoms (palpitations, dizziness, chest discomfort) still warrants a call to the prescribing clinician rather than self-reassurance based on this pattern.

How to interpret HRV when you cannot stop a medication

Most people taking cardiovascular, psychiatric, or pain medications cannot and should not stop them to get a "clean" HRV reading. A workable approach has three parts. First, document every medication and dose alongside any HRV report, clinical or consumer-grade, since interpretation without that context is not reliable. Second, favor within-person trends over comparisons to population reference ranges: a person on a stable metoprolol dose whose RMSSD falls steadily over months is showing a real decline even though both the starting and ending values might look "normal" against an unmedicated norm. Third, for research-grade or diagnostic-grade HRV assessment, a supervised medication washout of roughly two to three days is sometimes used to reduce confounding, but this is only feasible for some drugs (caffeine, short-acting stimulants) and is not appropriate or safe for others (beta-blockers in heart failure, opioids in chronic pain), and it should never be attempted without the prescriber's involvement. Methodological guidance for designing HRV studies explicitly recommends recording and, where possible, controlling for concurrent medication use as a standard part of study design (Quintana et al., 2019).

Normal HRV ranges and why they are only a starting point

Reported reference ranges for 24-hour clinical SDNN in healthy adults commonly fall in the roughly 100-180 ms range, with lower 24-hour SDNN linked to increased cardiovascular risk in post-myocardial-infarction populations. Short-term (5-minute) resting RMSSD in healthy adults is commonly reported in the 20-75 ms range, and age is the strongest non-pharmacological determinant, with RMSSD tending to decline across the decades of adult life. These ranges come from clinical recording protocols and are not directly interchangeable with a wearable's overnight "HRV score," which uses proprietary algorithms and a different recording window. A wearable reading of 45 ms while taking a beta-blocker could correspond to a meaningfully lower unmedicated baseline; without naming the drug, that gap is invisible.

Guideline bodies have moved diagnostic weight away from passive HRV monitoring for exactly this reason. Diabetes care guidelines recommend formal cardiovascular autonomic reflex testing (CARTs), rather than passive HRV recording alone, as the standard approach for evaluating cardiac autonomic neuropathy, in part because HRV is vulnerable to exactly the kind of pharmacological confounding described throughout this article. Anyone using HRV to screen for a specific autonomic condition, rather than to track personal wellness trends, should raise that goal directly with a clinician who can order the appropriate test rather than relying on a wearable trend line.

When to seek care rather than interpret the number yourself

An unexplained, large, or rapid drop in HRV accompanied by palpitations, fainting, chest pain, or shortness of breath is not something to resolve by adjusting for medications at home. HRV interpretation, medication-aware or not, is a supplement to clinical evaluation, not a replacement for it, and any new cardiac symptom deserves prompt medical attention regardless of what a wearable trend shows.

Reported effect sizes for beta-blockers, anticholinergics, SSRIs, opioids, and GLP-1 receptor agonists vary between studies and have not been independently confirmed here, so specific figures are not provided. Similarly, guideline language on cardiovascular autonomic reflex testing may be updated over time and should be checked against current diabetes care guidelines rather than assumed from this text.