Anxiety Shrinks Heart Rate Variability — Genes and Wearables Confirm the Link
A large genomic-wearable study finds anxiety and antidepressants independently reduce pulse rate variability, with causal evidence from Mendelian randomization.
Summary
Researchers combined Fitbit wearable data, whole-genome sequencing, and electronic health records from 920 All of Us participants to study how anxiety affects pulse rate variability (PRV). A higher genetic risk score for anxiety was independently linked to lower SDANN — a key PRV metric reflecting autonomic nervous system control. Several antidepressants, including venlafaxine, bupropion, SSRIs, SNRIs, and tricyclics, also reduced PRV independently. Mendelian randomization analysis supported a causal effect of anxiety on reduced PRV. These findings matter because wearable-derived PRV is increasingly used for stress detection and mental health screening, and understanding what drives PRV changes is critical to interpreting those signals accurately.
Detailed Summary
Heart rate and pulse rate variability (HRV/PRV) are widely used markers of autonomic nervous system health, and reduced variability is linked to poorer cardiovascular and mental health outcomes. While ECG-derived HRV has been studied in the context of anxiety disorders, PRV — derived from photoplethysmography (PPG) sensors in consumer wearables like Fitbit — had not been systematically investigated in relation to anxiety. This gap matters because PRV and HRV, though correlated, respond differently to stress and movement in real-world conditions.
This study used data from the All of Us Research Program, selecting 920 individuals of European ancestry with whole-genome sequencing, Fitbit wearable data, and electronic health records (EHRs), yielding 61,333 daily data points. PRV was quantified as SDANN — the standard deviation of average five-minute pulse intervals across a full 24-hour period. Anxiety polygenic risk scores (PRS) were constructed using PRS-CS after meta-analyzing GWAS data from UK Biobank, FinnGen, and the Million Veterans Program (combined N=364,550), identifying 41 genome-wide significant loci and confirming robust polygenic architecture.
The anxiety PRS was independently associated with reduced SDANN (beta=−0.08; p=0.003), even after adjusting for age, sex, BMI, principal components, and antidepressant use. Lifetime clinical diagnosis of anxiety was also associated with lower SDANN. Several antidepressant classes and individual drugs showed significant negative associations with SDANN: SNRIs (beta=−0.16; p=2×10⁻⁶), TCAs (beta=−0.177; p=0.0008), SSRIs (beta=−0.069; p=0.0008), venlafaxine (beta=−0.12; p=0.002), and bupropion (beta=−0.071; p=0.01). Anxiety PRS and antidepressant effects were shown to be statistically independent of each other.
Critically, one-sample Mendelian randomization — using genetic variants as instrumental variables to test causality — found an inverse causal effect of anxiety on SDANN (beta=−2.22; p=0.03). This suggests the association is not merely due to confounding or reverse causation, but that anxiety itself mechanistically suppresses autonomic variability as captured by wearable PRV sensors.
These findings have direct implications for the growing use of consumer wearables in health monitoring. Clinicians and researchers interpreting PRV data from fitness trackers should account for anxiety burden and antidepressant exposure as independent modulators of PRV. The study also raises important questions about whether PRV-based algorithms for stress or depression screening will perform differently in anxious populations or those on antidepressants. Limitations include the restriction to European-ancestry participants, reliance on EHR-based diagnoses, inability to assess inter-beat intervals directly, and the cross-sectional nature of the wearable data collection.
Key Findings
- Anxiety polygenic risk score independently associated with reduced SDANN PRV (beta=−0.08; p=0.003).
- SNRIs showed the strongest antidepressant-related PRV reduction (beta=−0.16; p=2×10⁻⁶).
- TCAs, SSRIs, venlafaxine, and bupropion each independently lowered SDANN.
- Mendelian randomization supports a causal effect of anxiety on reduced PRV (beta=−2.22; p=0.03).
- Anxiety and antidepressant effects on PRV were statistically independent of each other.
Methodology
Retrospective study using 920 European-ancestry All of Us participants with whole-genome sequencing, Fitbit PPG-derived PRV (SDANN), and EHR data (61,333 data points). Anxiety PRS were built via PRS-CS from a meta-GWAS of UKB, FinnGen, and MVP (N=364,550). Associations were tested with generalized linear mixed models; causal inference used one-sample Mendelian randomization.
Study Limitations
Analyses were restricted to European-ancestry participants, limiting generalizability. EHR-based diagnoses may miss subthreshold or undiagnosed anxiety cases. Wearable Fitbit data lacks inter-beat interval access, preventing assessment of ectopic beats and limiting PRV metric options.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.