In this study, we recruited total 120 patients on maintenance HD as a retrospective cohort, and calculated the intra-dialytic as well as visit-to-visit BPV metrics using peri-dialytic SBP during the 3 months exposure period. We found both intra-dialytic and visit-to-visit BPV were associated with CVD morbidity, while only intra-dialytic SD was associated with all-cause mortality. Overall, intra-dialytic BPV outperformed visit-to-visit BPV in prediction of CVD event. Although without reaching significance, intra-dialytic BPV also showed better performance in prediction of all-cause mortality than visit-to-visit BPV.
There are already multiple studies that proved both short-term BPV [8–10] and long-term BPV [11–14] are associated with CVD morbidity, CVD mortality and all-cause mortality. But none of them compared the prognostic ability of adverse outcomes between short-term BPV and long-term BPV. Hence, we covered all the commonly used metrics for intra-dialytic and visit-to-visit BPV, including SD, CV, VIM, ARV and residual, in order to obtain a comprehensive comparison. In the prediction of CVD event, intra-dialytic BPV metrics are better predictors with higher AUC than visit-to-visit BPV. Across intra-dialytic BPV metrics, residual is inferior to SD (P = 0.029) while no difference was discovered in the rest of metrics. Novel metrics such as VIM, ARV and residual didn’t show a significantly better performance than conventional metrics. With additional consideration of independency from SBP and simplicity of calculation, we recommended intra-dialytic CV as a favorable target. Moreover, intra-dialytic ARV is the only metric showing correlation with inter-dialytic weight gain and ultrafiltration volume in our study, implying its unique characteristic and potential application value. Similar association between intra-dialytic ARV and fluid removal was also discovered in other study [24]. Intra-dialytic BPV also had better performance in prediction of all-cause mortality than visit-to-visit BPV. But the limitation is that BPV does not significantly associate with all-cause mortality in our study.
Short-term BPV reflects the various physiologic mechanisms, such as autonomic modulation, arterial stiffness, humoral response, blood viscosity, behavior, emotional and environmental factors, whereas long-term BPV additionally involves compliance with prescribed therapeutic regimen and seasonal climate changes [7]. In HD patients, the dialysis-related and interdialytic factors should also be considered, including UF-driven fluid shifts, serum osmolality changes, dialytic removal of antihypertensives and vasoactive substances [25], and interdialytic fluid accumulation [9]. Although it is pointless to identify precise contribution of distinct mechanisms in clinical setting, it helps us to better understand the BPV.
Greater prognostic ability of intra-dialytic BPV may suggest a major role of dialysis-related factors causing adverse outcomes. Dialysis could be considered as a shock due to the rapid removal of fluids and electrolytes, which may better reflect the body’s ability to regulate BP under stress, like a cardiac stress test for heart evaluation. The limitation of intra-dialytic BPV is that it can’t reflect influence out of dialysis unit such as fluid accumulation, circadian fluctuation and behavior changes. Moreover, peri-dialytic SBP is not the only choice to evaluate BPV. Ambulatory SBP and home SBP measurement have shown superior risk prediction for adverse outcomes [26–28]. And BPV derived from ambulatory and home SBP measurements manifested association with CVD morbidity and all-cause mortality as well [28]. The BPV based on 44-h ambulatory monitoring during interdialytic interval has been proved to predict outcomes [9]. Here, we didn’t include ambulatory and home SBP, because these two measurements have not been routinely monitored in clinic yet. Further investigation comparing between BPV derived from various measurements is required. Perhaps metrics, obtained by combination of intra-dialytic and inter-dialytic BPV, can be explored in the future.
Our study comprehensively compared the prognostic ability of intra-dialytic and visit-to-visit BPV metrics in predicting CVD events and all-cause mortality, which provides evidence for selecting the most efficient BPV metric. Determining a widely accepted and efficient BPV metric is the first step to build a BPV-based guideline for HD population. Non-pharmacologic treatments, such as restricting sodium intake and dialysis prescription, and antihypertensive medications are main approaches to manage BP and volume in HD patients [18]. Nevertheless, studies investigating the influence of these interventions on BPV are scarce. Our work can also provide reference for reliable BPV evaluation in the further research. However, there is also some limitations in our study. First of all, this is a single center retrospective study and the cohort size is limited, which can cause bias. Second, we didn’t include anti-hypertensive medication as a covariate due to the lack of an organized record. Finally, our conclusion for BPV was based on an exposure period for 3 months. Different durations may also affect the results.