Abstract Background and Aims Cardiopulmonary monitoring during hemodialysis has the potential to improve dialysis adequacy, prevent adverse events, improve quality of dialysis patients’ lives and reduce mortality risks. Under stable conditions the knowledge of upper body oxygen consumption, the difference between arterial and venous oxygen saturation (ScvO2) and hemoglobin concentrations allow the calculation of estimated UBBF (eUBBF; Rosales 2019). We studied the association between eUBBF and cardiac output during hemodialysis. Method In patients with a central venous catheter as the vascular access, we simultaneously measured intradialytic central-venous oxygen saturation and hemoglobin levels once per minute using the Crit-Line Monitor (Fresenius Medical Care North America, Waltham, USA). Concomitantly we measured cardiac output (CO) using the Task Force Monitor (CNSystems, Graz, Austria). We depicted CO and eUBBF as box-whiskers in 5 minute periods from 20 to 180 minutes, tested the minute-level time series for stationarity using the Dickey-Fuller test, employed differencing to make the time series stationary, and tested for linear associations using cross-correlations and determined if the CO time series was useful to predict the eUBBF time series using the Granger Causality test. Results We studied 13 patients (59±14 years, 5 (38%) male, 93±22 kg pre-dialysis weight 170±7 cm tall) during 34 hemodialyses. CO and UBBF was 4.7±1.0 and 1.3±0.4 L/min, and showed a small trend downward for CO during hemodialysis. Cross-correlations showed no remarkable correlations for all treatments, Granger causality index was less than 20% in 8 treatments however without clearly discernible patterns. Conclusion Despite variability in the CO the eUBBF remains stable, a dynamic which pathophysiologically appears plausible. While testing for linear associations yielded no significant relationships, based on the Granger causality test there nevertheless appears to be some form of a relationship which requires some further investigation which will need to also include aspects such as fluid overload, body temperature and total peripheral resistance. Based on this data, cardiopulmonary monitoring appears possible and yields a reliable continuous overview of the hemodynamic status of a patient.
Read full abstract