Abstract
Background and objectives: Although ambulatory BP recordings are found to be superior to dialysis unit recordings in predicting outcomes, ambulatory BP are difficult to obtain in the day-to-day treatment of hemodialysis patients. Home BP agree well with ambulatory BP, but the prognostic significance of home BP recordings is unknown in hemodialysis patients. This study ascertained the role of home BP in predicting all-cause and cardiovascular mortality.
Design, setting, participants, & measurements: A prospective cohort study was conducted in 150 patients who were on chronic hemodialysis dialyzing at four university-affiliated units. BP was self-measured at home for 1 wk, for an interdialytic interval by ambulatory recording, and by “routine” and standardized methods in the dialysis unit for 2 wk. Patients were followed for a median of 24 mo to assess the end points of all-cause and cardiovascular mortality.
Results: Cardiovascular death occurred in 26 (17%) patients and death in 46 (31%) patients. A 1-SD increase in systolic BP increased the risk for death by 1.35 (95% CI 0.99 to 1.84) and in diastolic BP by 1.40 (95% CI 1.03 to 1.93) for home BP and between 0.97 to 1.19 (P > 0.20) for all-cause mortality for dialysis unit BP recording. A dose–response relationship between increasing quartiles of home BP and all-cause mortality and cardiovascular mortality was seen.
Conclusions: Self-measured systolic BP of 125 to 145 mmHg and of 115 to 125 mmHg by ambulatory BP is associated with the best prognosis in hemodialysis patients.
It has long been recognized that hypertension occurs commonly in patients with chronic kidney disease and is often poorly controlled in hemodialysis (HD) patients (1). Routine BP agree poorly with ambulatory BP recordings that are obtained during an interdialytic interval (2,3). A meta-analysis comparing the diagnostic performance of routine BP for diagnosing hypertension found poor agreement between the techniques and questioned the use of routine BP in managing hypertension (4). Large cohort studies have found that lower systolic BP (SBP) that are obtained before or after dialysis are determinants of mortality (5,6). These BP, often obtained in the dialysis unit over several weeks, are obtained without attention to details of the measurement technique (7). BP that are obtained in the dialysis unit by technicians and nurses without attention to detail (routine BP) differ strikingly from BP that are obtained using standard methods. Nearly half of the SBP are >10 mmHg different from routine BP when standard methods of measurements are used (8). Higher ambulatory BP recordings are associated with increased mortality in HD patients (9). yet almost all hypertension management in dialysis units uses dialysis unit BP. For treating hypertension, national guidelines recommend the use of BP that are obtained in the dialysis unit (10).
The best BP measurement technique remains uncertain in HD patients. Although the diagnostic performance of BP that are obtained using standard measurement method in the dialysis unit to predict hypertension (using ambulatory BP as the gold standard) is reasonable, we previously reported (11) that it cannot be used to predict the presence or absence of left ventricular hypertrophy (LVH). In contrast, BP that are obtained outside the dialysis unit, whether obtained by interdialytic automatic BP measurements or self-measured BP at home, are useful in diagnosing LVH. Thus, out-of-dialysis BP measurement seems to be a promising technique to diagnose hypertension and target organ damage (12). In fact, numerous investigators have emphasized the superiority of ambulatory BP measurements compared with dialysis unit measurements in HD patients, yet no data exist on the prognostic significance of self-measured home BP compared with those obtained in the dialysis unit (13–16). If home BP are of prognostic value, then the level of BP associated with better outcomes is also undefined.
We hypothesized that home BP measurements will be of greater prognostic significance compared with routine or standardized BP recordings made in the dialysis unit. The purpose of this report was to evaluate the presence, strength, and shape of the relationship among BP that are measured using various modalities (home, ambulatory, and dialysis unit) and cardiovascular and all-cause mortality in HD patients.
Concise Methods
Participants
The cross-sectional data on this cohort have previously been reported (3,11). Enrolled patients who were ≥18 yr of age; had been on long-term HD for >3 mo; were free of vascular, infectious, or bleeding complications within 1 mo of recruitment; and were undergoing dialysis three times a week at one of the four dialysis units in Indianapolis that were affiliated with Indiana University. Those who missed two HD treatments or more during 1 mo, abused drugs, had chronic atrial fibrillation, or had body mass index of ≥40 kg/m2 were excluded. Patients who had a change in dry weight or antihypertensive drugs within 2 wk were also excluded. The study was approved by the institutional review board of Indiana University and Research and Development Committee of the Roudebush VA Medical Center, Indianapolis, and all patients gave written informed consent.
Ambulatory BP Monitoring.
Ambulatory BP monitoring was performed after the midweek HD session for 44 h. Ambulatory BP were recorded every 20 min during the day (6 a.m. to 10 p.m.) and every 30 min during the night (10 p.m. to 6 a.m.) using a Spacelab 90207 ambulatory BP monitor (SpaceLabs Medical, Redmond, WA) in the nonaccess arm, as reported previously (17). Awake and asleep readings were calculated for each patient by self-reported sleep and wake times by means of a diary. Even a limited number of ambulatory BP are useful for prognostic purposes in the general population; therefore, we did not exclude any patients on the basis of the number of ambulatory BP recordings (18).
Dialysis Unit BP.
Dialysis unit BP recordings as measured by the dialysis unit staff before and after dialysis were collected prospectively. These BP recordings were obtained using the sphygmomanometer equipped with HD machines without a specified technique and were averaged over 2 wk. Thus, each patient had six predialysis and six postdialysis BP recordings to provide routine dialysis unit BP.
BP was also recorded using a standard technique after at least a 5-min rest and using a validated oscillometric device (HEM 907; Omron Healthcare, Vernon Hills, IL) by research nurses who were trained in this technique. Three readings at each visit were averaged to provide one recording. Predialysis and postdialysis recordings were averaged separately over 2 wk to provide a standardized dialysis unit BP.
Home BP Monitoring.
Home BP monitoring was performed over 1 wk using a validated self-inflating automatic oscillometric device (HEM 705 CP; Omron Healthcare). The protocol specified home BP monitoring in the first week. Patients were instructed in the use of this monitor and asked not to share this monitor with others. Patients were asked to record their BP three times daily—on waking up, between noon and 6 p.m., and at bedtime—and log this on a chart provided for this purpose. Because this monitor is equipped with memory and a printer, we used only recordings that were recorded in the memory of the monitor.
Outcomes
All-cause mortality was the primary focus of our study because this outcome was available for every patient. Two physicians evaluated the patients’ medical charts to evaluate the cause of death. Cardiovascular death (defined as a composite end point of stroke, pulmonary edema, myocardial infarction, and sudden death) was also analyzed. Patients who received a transplant or transferred out of our unit were censored on their last dialysis date.
Statistical Analysis
Cox proportional hazards regression was used to determine the significance and strength of association of factors associated with mortality outcomes. The proportionality assumption was tested by evaluating the log minus log plot. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated from the proportional hazards model regression coefficients and their SE for a 1-SD change in each BP. Because all BP were obtained in the same population, the HR were not adjusted for the comparisons. Other investigators have taken a similar approach when comparing the relationship of various BP and mortality in the general population (19). To compare directly the BP on their impact on all-cause and cardiovascular mortality, we calculated the HR for a 1-SD change in SBP or diastolic BP (DBP). To ascertain which BP level is associated with the best survival, we divided our data into quartiles and tested the association of these quartiles with mortality using the Cox model.
All analyses were conducted using SPSS 14.0 (SPSS, Chicago, IL). The P values reported are two-sided and taken to be significant at <0.05.
Results
Between September 2003 and February 2005, we recruited 150 patients from four dialysis units staffed by the nephrology faculty of Indiana University, Indianapolis. The diagnostic performance of various BP from this cohort has been previously published (3).
The clinical characteristics of patients are shown in Table 1. The population was predominantly black with an average age of 56 yr. All patients were on thrice-weekly dialysis and were prescribed a dialysis time of 238 ± 25 min, blood flow rate of 405 ± 32 ml/min, and dialysate flow rate of 754 ± 88 ml/min. Average urea reduction ratio, serum albumin, and hemoglobin reflect the general HD population. Cardiovascular disease, defined as history of myocardial infarction, coronary or vascular bypass surgery or angioplasty, stroke, aortic aneurysm, or nontraumatic amputation, was present in 47% patients. The majority of patients received antihypertensive drugs, with an average of 2.3 drugs (Table 2). β Blockers were the most commonly used antihypertensive agent.
Clinical characteristics of the study population
Antihypertensive drug use and BP measurementsa
Median follow-up was 24 mo (interquartile range 16 to 32 mo). During this follow-up period, 46 (31%) patients died and 26 (17%) had cardiovascular death. The crude mortality rate was 157 per 1000 patient-years and cardiovascular mortality rate of 89 per 1000 patient-years. Causes of cardiovascular death were sudden cardiac death (n = 15), myocardial infarction (n = 6), pulmonary edema (n = 2), and stroke (n = 3).
Table 3 shows the HR for death associated with a 1-SD change in respective BP. By standardizing the scale, the HR can be compared directly across the BP measurement methods, because the measurements are made in the same population of patients. Measurements in the dialysis unit for SBP and DBP were not predictive of total mortality; however, postdialysis SBP was associated with an increase in cardiovascular mortality. In contrast, ambulatory BP was associated with a nearly 50% higher death rate for a 1-SD or 22.3/13.8-mmHg increment in SBP/DBP. In comparison, home BP was associated with a 35 to 40% elevation in risk for death.
HR for mortality associated with 1 SD of BPa
To explore better the relationship between SBP and outcomes, we calculated the HR of the progressive quartiles compared with the first. Table 4 demonstrates the relationship between quartiles of SBP obtained by various methods and the HR for all-cause mortality and cardiovascular death. The second quartile was associated with a better survival compared with others. The fourth quartile of ambulatory SBP (>145 mmHg) was significantly associated with higher mortality compared with the first quartile. Figure 1 shows HR depicting the relationship between all-cause mortality and quartiles of BP measured using ambulatory, home, or routine dialysis unit BP measurements. A dose-response relationship between increasing levels of SBP and all-cause mortality was seen with home and ambulatory BP; dialysis unit BP recordings were of no prognostic importance.
Hazard ratios for all-cause mortality for quartiles of systolic BP. Both home BP and ambulatory BP had prognostic information. Conversely, dialysis unit BP recordings did not achieve statistical significance in terms of prognostic value. P values are those reported for linear trend. HD, hemodialysis; Q, quartile.
HR for mortality in hypertensive patients associated with SBP quartile
Discussion
The results of our study demonstrate that prognostic information is contained in BP that are obtained outside the dialysis unit. Although ambulatory BP measurements contained the greatest prognostic information, home BP recordings served as a useful surrogate for ambulatory BP recordings and could be used to provide prognostic information. In contrast, dialysis unit–recorded BP were generally not of prognostic significance, with the exception of the association of postdialysis SBP and cardiovascular mortality. The relationship of BP recordings and mortality followed a U-shaped curve, as has been previously reported.
The intent of this study was to describe the prognostic value of BP in the sample of patients in whom clinicians may be interested in managing hypertension. The large cohort studies report a consistently higher mortality for lower BP (5,6). Examination of Table 4 shows that the mortality of the lowest quartile indeed seems to be higher compared with the second quartile, which confirms previous observations made in larger cohorts; however, using home BP recordings, we saw a clear trend of increasing all-cause and cardiovascular mortality from quartile 2 to quartile 4. A similar trend was seen for ambulatory BP monitoring, suggesting that BP that are recorded outside the dialysis unit may contain greater prognostic information compared with BP that are measured in the dialysis unit at two extremes of volume state. Thus, a U-shaped relationship may link hypertension to cardiovascular mortality or all-cause mortality. The strong relationship between postdialysis routine SBP and cardiovascular mortality is notable and is consistent with previous reports (20,21).
Given that all of the analyses in this article are unadjusted, we cannot conclude that any of these modalities for measuring BP are independently associated with mortality. Covariates such as age, ethnicity, intradialytic weight gain, smoking, diabetes, and cardiovascular disease presumably can influence mortality as well as BP. At the bedside, if one modality of measuring BP is a stronger predictor of mortality, then it stands to reason that that measure contains greater prognostic power compared with the comparator. Because all measurements were implemented in the same population, we did not make any adjustments for the patient characteristics. Sega et al. (19) took a similar approach when conducting these analyses of home, ambulatory, and office BP and mortality outcomes in the general population.
Although a similar study that links ESRD mortality to home or ambulatory BP monitoring is not available for comparison, the results of our study are consistent with previously published cohort studies that examined the influence of home BP with clinic BP. For example, the prognostic importance of home BP monitoring was evaluated in a large, rural, community-based study in Japan (22). The investigators reported that among 1789 individuals who were aged ≥40 yr, home BP measurement had a stronger predictive power for mortality than clinic BP. In another study, 4939 treated elderly hypertensive patients who were a mean age of 70 yr had clinic and home BP recorded and were followed for 3.2 yr (23). Whereas each 10/5-mmHg increase in SBP and DBP measured at home increased the cardiovascular mortality by 17.2/11.7%, no similar increase was associated with clinic BP recordings. In patients who had chronic kidney disease and were not on dialysis, 1-SD increase in SBP increased the risk for renal end point by 1.27 (95% CI 1.01 to 1.60) for routine clinic measurement, by 1.69 (95% CI 1.32 to 2.17) for standardized clinic measurement, and by 1.84 (95% CI 1.46 to 2.32) for home BP recording (24).
Two studies using ambulatory BP monitoring in HD patients support our observations of a greater strength of relationship with ambulatory BP and mortality. Amar et al. (9) reported 57 treated French hypertensive HD patients, 10 of whom died of cardiovascular causes. Although this study had a small number of patients with a follow-up of 34 ± 20 mo, it demonstrated that nocturnal SBP (relative risk 1.41; 95% CI 1.08 to 1.84) and 24-h pulse pressure (relative risk 1.85; 95% CI 1.28 to 2.65) were independent predictors of cardiovascular mortality in treated hypertensive HD patients. In the largest study using 24-h ambulatory BP, the relationship of hypertension and cardiovascular outcomes was prospectively analyzed in 168 nondiabetic HD patients without preexisting cardiovascular events (25). The ratio of the average SBP during the night and day was used to indicate the nocturnal fall or the dipping phenomenon. This ratio was the only BP indicator that was associated with all-cause and cardiovascular mortality on both univariate and multivariate analyses.
Why were BP recordings that were made outside the dialysis unit of greater prognostic significance? Although the exact answer to this question is not known, there are several possible reasons. First, timing of the administration of antihypertensive drugs can be variable, and multiple BP measurements during the course of the day, as done with home or ambulatory BP monitoring, can average out the troughs and peaks in BP swings that clinic BP recordings are unable to do (26). Second, dialysis unit BP are influenced by the white-coat effect—elevated BP only in the dialysis setting—which is less pronounced with home BP and eliminated by ambulatory BP monitoring (27). Also, masked hypertension—elevated BP at home but normal in the dialysis unit—is potentially detected with home BP monitoring and ambulatory BP monitoring and may be of prognostic significance (28). There is little relationship between BP that are obtained in the dialysis unit and LVH, but a significant relationship is seen between BP that are obtained outside the dialysis unit and LVH (11). Finally, BP that are sampled from a broader pool of situations may make them more representative of the person’s typical BP (18)
There are several strengths and limitations of our work. Our study was largely limited to blacks, and we excluded certain patients, such as those with morbid obesity and atrial fibrillation because of difficulties with accurate BP assessment in this group. Whether the same results would hold in people of other ethnicities and of broader clinical characteristics is not known and will require verification in future cohorts. The sample size of our study was relatively small, yet the statistical significance of the prognostic information contained in home BP recordings is the first reported in this population. By dividing the sample into quartiles, we could test the threshold of BP associated with a better prognosis, instead of using arbitrary definitions of normotension and hypertension, which are debatable in the HD population. The home BP monitor that we used was a validated device equipped with a memory device and printer, so there was a mechanism in place to confirm the authenticity of the patient reports (29). Whether home BP is independently related to total or cardiovascular mortality was not tested in this study and would require a larger cohort. Thus, further efforts are needed to establish the usefulness of BP measurements that are made outside the dialysis environment.
Conclusions
Dialysis unit BP predict neither target organ damage (11) nor hard outcomes. Thus, in dialysis patients more so than in the general population, BP measurement and treatment should occur with recordings that are made outside the clinic. This paradigm is similar to that of blood glucose monitoring in the management of diabetes. The results of the study support the view of a conventional risk factor being very important for prognostication of mortality and illustrate that inadequate measurements of a conventional risk factor may confer undue advantage to nonconventional risk factors. We cannot imply causality in a cohort study; however, this study may offer some guidelines with respect to BP goals. Self-measured SBP of approximately 125 to 145 mmHg (rounded) and of approximately 115 to 125 mmHg by ambulatory BP were associated with the best prognosis. These thresholds may be used to test the hypothesis of whether controlling hypertension in HD patients using out-of-dialysis unit BP recordings would make a difference to cardiovascular mortality.
Disclosures
None.
Acknowledgments
This study was supported by grants 5RO1-063020 and 5T32DK062711 from the National Institutes of Health.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
- Received May 28, 2007.
- Accepted July 17, 2007.
- Copyright © 2007 by the American Society of Nephrology