Background: The opioid epidemic highlights the need for objective biomarkers that identify high-risk chronic pain patients before overdose events occur. Existing prescription drug monitoring programs such as the NARX score are primarily reactive, underscoring the potential value of physiologic and behavioral risk markers.

Methods: In a retrospective cohort of 98 adults with chronic pain (2019–2023), heart rate variability (HRV) was measured using the Max Pulse system, ultra-processed food addiction (UPFA) symptoms were assessed with an 8-item modified Yale Food Addiction Scale, and prescription patterns were obtained from state PDMP data. High-risk (NARX ≥200; n=74) and control (n=24) groups were compared using standard univariable tests, while parsimonious multivariable logistic regression, correlation analyses, and receiver operating characteristic (ROC) modeling with bootstrap internal validation evaluated associations and predictive performance.  

Results: High-risk patients showed more frequent pathologic HRV (HRV <40 ms) and substantially lower mean HRV than controls (both p<0.001). Each 10 ms decrement in HRV was associated with roughly a twofold higher odds of high-risk status in adjusted models, with a plausible 95% confidence interval replacing the previously inconsistent estimate. UPFA measures were higher in high-risk patients, and a strong but more conservative positive correlation between sugar cravings and opioid dose escalation was observed in the high-risk group, with a weak, non-significant correlation in controls. A composite NARX-HRV-UPFA model achieved higher apparent and optimism-adjusted AUC and sensitivity than NARX alone, though performance remained exploratory given internal validation only.  

Conclusions: Combined autonomic and behavioral profiling via HRV and UPFA screening is associated with elevated prescription-based overdose risk and improves risk discrimination beyond NARX alone in this cohort. HRV and UPFA emerge as promising, hypothesis-generating biomarkers that require prospective validation and external replication before informing routine clinical or policy decisions.