Objectives: This study investigated whether frequency-domain heart rate variability (HRV) analysis could detect autonomic dysfunction linked to incipient renal impairment.
Methods: Cross-sectional study of 82 adults with type 2 diabetes. RR intervals were captured using the Polar H10 sensor and processed via Kubios HRV software. LF power (0.04-0.15 Hz) was log-transformed (lnLF). Participants were stratified by uACR ≥ 30 mg/g.
Results: The 82 patients were divided into two groups (38 patients normoalbuminuria, 44 patients microalbuminuria) were comparable in age (p = 0.588), BMI (p = 0.535), and HbA1c (p = 0.198). eGFR was lower in the microalbuminuria group (median 40.5 vs. 52.5 mL/min/1.73 m², p = 0.028). Time-domain (SDNN, RMSSD) and nonlinear indices (SD1, SD2) did not differ significantly between groups. In contrast, lnLF was significantly lower in the microalbuminuria group (3.24 ± 1.78 vs. 2.23 ± 2.02, p = 0.019); the LF/HF ratio was also significantly lower (p = 0.042). Spearman correlation confirmed an inverse association between lnLF and uACR (ρ = -0.249, p = 0.024).
Conclusion: Frequency-domain HRV, specifically lnLF, was the only HRV domain significantly associated with albuminuria. Wearable sensor-derived LF-HRV may serve as a non-invasive biomarker warranting validation in longitudinal studies.