Ambulatory blood pressure monitoring in dialysis patients and estimation of mean interdialytic blood pressure

Abstract

To define blood pressure (BP) patterns and control in dialysis patients, 48-hour ambulatory BP monitoring was performed in 36 hemodialysis and 18 peritoneal dialysis patients. Monitoring began during a dialysis session for hemodialysis patients. Data revealed significantly lower diastolic BP (DBP) and lower diastolic load (percentage of diastolic values > 90 mm Hg) in hemodialysis patients compared with peritoneal dialysis patients (80.6 mm Hg v 88.8 mm Hg, respectively, [P < 0.03] and 26% v 45%, respectively [P < 0.03]) for the 48-hour period. When the 2 days were analyzed separately, the difference in diastolic pressures and loads was significant only for the first (dialysis) day. Similarly, trends toward lower systolic BP (SBP) and systolic load in hemodialysis patients existed throughout monitoring and were greater in magnitude during the first day. BP data were fit to a random-coefficient growth curve model to detect periodicity. This sensitive model did not detect diurnal variation of BP in either group. The incidence of hypotension did not differ between the two groups (2.0% v 1.0% of total observations, hemodialysis v peritoneal dialysis). In the hemodialysis group, the proportion of hypotensive observations was significantly greater during the 4 hours postdialysis compared with other periods (5.6% v 1.6%; P < 0.02), a finding that likely reflects the practice of holding antihypertensives until after hemodialysis. However, patient diaries did not reflect hypotensive symptoms during this time. In the hemodialysis group, mean BP and predialysis BP did not correlate with interdialytic sodium load or weight gain. Predialysis and postdialysis BP (recorded by dialysis nurses) correlated significantly with mean BP. Predialysis SBP overestimated mean SBP by an average of 10 mm Hg, while postdialysis SBP underestimated mean SBP by an average of 7 mm Hg. To create formulas to estimate mean SBP and DBP in hemodialysis patients, multiple linear regression was used to model these variables against age, sex, race, and average prehemodialysis/posthemodialysis BP. The model achieved a high degree of fit (r2 = 0.72 for SBP; r2 = 0.65 for DBP), demonstrating that prehemodialysis and posthemodialysis BP can be used to predict mean BP in hemodialysis patients. In summary, our data show the absence of a diurnal variation of BP in dialysis patients and lower BP in hemodialysis patients compared with peritoneal dialysis patients. Among hemodialysis patients, more hypotension occurred after dialysis compared with other periods, and predialysis and postdialysis BP can be used to model mean BP levels.