An analysis of the impact of fluid overload and fluid depletion for all-cause and cardiovascular mortality

Abstract

Abstract Background Both baseline fluid overload (FO) and fluid depletion are associated with increased mortality risk and cardiovascular complications in haemodialysis patients. Fluid status may vary substantially over time, and this variability could also be associated with poor outcomes. Methods In our retrospective cohort study, including 4114 haemodialysis patients from 34 Romanian dialysis units, we investigated both all-cause and cardiovascular mortality risk according to baseline pre- and post-dialysis volume status, changes in pre- and post-dialysis fluid status during follow-up (time-varying survival analysis), pre–post changes in volume status during dialysis and pre-dialysis fluid status variability during the first 6 months of evaluation. Results According to their pre-dialysis fluid status, patients were stratified in the following groups: normovolaemic with an absolute FO (AFO) compartment between −1.1 and 1.1 L, fluid depletion with an AFO below −1.1 L, moderate FO with an AFO compartment >1.1 but <2.5 L and severe FO with the AFO compartment >2.5 L. Baseline pre-dialysis FO and fluid depletion patients had a significantly elevated risk of all-cause mortality risk {hazard ratio [HR] 1.53 [95% confidence interval (CI) 1.22–1.93], HR 2.04 (95% CI 1.59–2.60) and HR 1.88 (95% CI 1.07–3.39) for moderate FO, severe FO and fluid depletion, respectively}. In contrast, post-dialysis fluid depletion was associated with better survival [HR 0.71 (95% CI 0.57–0.89)]. Similar results were found when using changes in pre- or post-dialysis fluid status during follow-up (time-varying values): FO patients had an increased risk of all-cause [moderate FO: HR 1.39 (95% CI 1.11–1.75); severe FO: HR 2.29 (95% CI 2.01–3.31] and cardiovascular (CV) mortality [moderate FO: HR 1.34 (95% CI 1.05–1.70); severe FO: HR 2.34 (95% CI 1.67–3.28)] as compared with normohydrated patients. Using pre–post changes in volume status during dialysis, we categorized the patients into six groups: Group 1, AFO <−1.1 L pre- and post-dialysis; Group 2, AFO between −1.1 and 1.1 L pre-dialysis and <−1.1 L post-dialysis (the reference group); Group 3, AFO between −1.1 and 1.1 L pre- and post-dialysis; Group 4, AFO >1.1 L pre-dialysis and <−1.1 L post-dialysis; Group 5, AFO >1.1 L pre-dialysis and between −1.1 and 1.1 L post-dialysis; Group 6, AFO >1.1 L pre- and post-dialysis. Using the baseline values, only patients in Groups 1, 5 and 6 maintained an increased risk for all-cause mortality as compared with the reference group. Additionally, CV mortality risk was significantly higher for patients in Groups 5 and 6. When we applied the time-varying analysis, patients in Groups 1, 5 and 6 had a significantly higher risk for both all-cause and CV mortality risk. In the last approach, the highest risk for the all-cause mortality outcome was observed for patients with high-amplitude fluctuation during the first 6 months of evaluation [HR 2.75 (95% CI 1.29–5.84)]. Conclusion We reconfirm the association between baseline pre- and post-dialysis volume status and mortality in dialysis patients; additionally, we showed that greater fluid status variability is independently associated with higher mortality.