Predictive value of heart rate variability on long-term mortality in end-stage kidney disease on hemodialysis

Autonomic disturbance is common in end-stage kidney disease (ESKD). Heart rate variability (HRV) is a useful tool to assess autonomic function. We aimed to evaluate the predictive value of HRV on all-cause mortality and explore the proper timing of HRV assessment. This prospective cohort study enrolled 163 ESKD on hemodialysis patients from April-December 2018. HRV measurements were recorded ten minutes before hemodialysis, four hours during hemodialysis, and ten minutes after hemodialysis. Clinical parameters and all-cause mortality were recorded. Cox-proportional hazard regression was used for statistical analysis. After a median follow up of 40 months, 37 (22.7%) patients died. Post-dialysis HRV parameters including higher very low frequency (VLF) (hazard ratio [HR], 0.881; 95%confidence interval [CI], 0.828-0.937; p<0.001), higher normalized low frequency (nLF) (HR, 0.950; 95%CI, 0.917-0.984; p = 0.005) and higher LF/HF ratio (HR, 0.232; 95%CI, 0.087-0.619; p = 0.004) were the independent predictors associated with lower risk for all-cause mortality. Higher post-dialysis normalized high frequency (nHF) increased risk of mortality (HR, 1.051; 95%CI, 1.015-1.089; p = 0.005). HRV parameters at pre-dialysis and during dialysis were not predictive for all-cause mortality. The area under receiver operating characteristic curve (AuROC) of VLF for survival was highest compared to other HRV parameters at post-dialysis period (AuROC 0.71; 95% CI; 0.62-0.79; p<0.001). In conclusion, post-dialysis HRV parameters predicted all-cause mortaliy in ESKD. VLF measured at post-dialysis exhibited best predictive value for survival in chronic hemodialysis patients.