Association of intradialytic blood pressure variability with increased all-cause and cardiovascular mortality in patients treated with long-term hemodialysis

Abstract

Background: Blood pressure is known to fluctuate widely during hemodialysis; however, little is known about the association between intradialytic blood pressure variability and outcomes.

Study design: Retrospective observational cohort.

Setting & participants: A random sample of 6,393 adult, thrice-weekly, in-center, maintenance hemodialysis patients dialyzing at 1,026 dialysis units within a single large dialysis organization.

Predictor: Intradialytic systolic blood pressure (SBP) variability. This was calculated using a mixed linear effects model. Peridialytic SBP phenomena were defined as starting SBP (regression intercept), systematic change in SBP over the course of dialysis (2 regression slopes), and random intradialytic SBP variability (absolute regression residual).

Outcomes: All-cause and cardiovascular mortality.

Measurements: SBPs (n = 631,922) measured during hemodialysis treatments (n = 78,961) during the first 30 days in the study. Outcome data were obtained from the dialysis unit electronic medical record and were considered beginning on day 31.

Results: High (ie, greater than the median) versus low SBP variability was associated with greater risk of all-cause mortality (adjusted HR, 1.26; 95% CI, 1.08-1.47). The association between high SBP variability and cardiovascular mortality was even more potent (adjusted HR, 1.32; 95% CI, 1.01-1.72). A dose-response trend was observed across quartiles of SBP variability for both all-cause (P = 0.001) and cardiovascular (P = 0.04) mortality.

Limitations: Inclusion of patients from a single large dialysis organization, over-representation of African Americans and patients with diabetes and heart failure, and lack of standardized SBP measurements.

Conclusions: Greater intradialytic SBP variability is associated independently with increased all-cause and cardiovascular mortality. Further prospective studies are needed to confirm findings and identify means of reducing SBP variability to facilitate randomized study.

Figure 1. Mixed linear systolic blood pressure metric by phase of treatment session

Systolic blood pressure (SBP) measurements during one hypothetical dialysis treatment in one individual (circles). The solid line represents the expected pattern of SBP during the course of a single dialysis treatment and is estimated from a mixed-effects model with main effect terms for the proportion of dialysis elapsed (considered as a 2-slope linear spline with knot at 0.25), random effects intercept terms for subject and treatment, and random effects slope terms for subject. Starting SBP is described as the intercept of the regression line; early slope as the slope of the predicted SBP line over the first 25% of dialysis; late slope as the slope of the predicted SBP line over the latter 75% of dialysis; absolute SBP residual is defined in terms of the absolute value of the vertical distance (dashed lines; d_i) between each SBP measurement and the predicted value for the corresponding time.

Figure 2. Adjusted associations between systolic blood pressure (SBP) metrics and all-cause and cardiovascular mortality

Panel A shows all-cause mortality, and Panel B shows cardiovascular mortality. Estimates for early and late slope are presented per 10 mmHg/treatment increment; all other estimates are presented as per category shown. Estimates in each panel are adjusted for one another as well as for age, sex, race (black, non-black), dialysis vintage (<3, ≥3 y), hypertension, heart failure, coronary artery disease, diabetes, vascular access type (fistula, graft, catheter), delivered dialysis session length, post-dialysis weight, interdialytic weight gain (≤1.5, >1.5–3, >3–4, >4 kg, missing), creatinine (quartiles, missing), equilibrated Kt/V (<1.2, ≤1.2, missing), albumin (≤3.5, >3.5–4, >4 g/dL, missing), calcium (<8.4, 8.4–<9.5, 9.5–<10, ≥10 mg/dL, missing), hemoglobin (<9, 9–<10, 10–<12, ≥12 g/dL, missing).

Figure 2. Adjusted associations between systolic blood pressure (SBP) metrics and all-cause and cardiovascular mortality

Panel A shows all-cause mortality, and Panel B shows cardiovascular mortality. Estimates for early and late slope are presented per 10 mmHg/treatment increment; all other estimates are presented as per category shown. Estimates in each panel are adjusted for one another as well as for age, sex, race (black, non-black), dialysis vintage (<3, ≥3 y), hypertension, heart failure, coronary artery disease, diabetes, vascular access type (fistula, graft, catheter), delivered dialysis session length, post-dialysis weight, interdialytic weight gain (≤1.5, >1.5–3, >3–4, >4 kg, missing), creatinine (quartiles, missing), equilibrated Kt/V (<1.2, ≤1.2, missing), albumin (≤3.5, >3.5–4, >4 g/dL, missing), calcium (<8.4, 8.4–<9.5, 9.5–<10, ≥10 mg/dL, missing), hemoglobin (<9, 9–<10, 10–<12, ≥12 g/dL, missing).