- Email: [email protected]
Relationship between different blood pressure measurements and left ventricular mass by cardiac magnetic resonance imaging in end–stage renal disease
Journal of the American Society of Hypertension 9(4) (2015) 275–284
Research Article
Relationship between different blood pressure measurements and left ventricular mass by cardiac magnetic resonance imaging in end–stage renal disease Asad Merchant, MDa,b, Ron Wald, MD, MPHa,b,c,*, Marc B. Goldstein, MCDM, MDb,c, Darren Yuen, MD, PhDb,c, Anish Kirpalani, MDa,d, Niki Dacouris, BScb, Joel G. Ray, MD, MSca,e, Mercedeh Kiaii, MDf, Jonathan Leipsic, MDg, Vamshi Kotha, MDa,d, Djeven Deva, MDa,d, and Andrew T. Yan, MDa,c,h,** a
University of Toronto, Toronto, ON, Canada; Division of Nephrology, St. Michael’s Hospital, Toronto, ON, Canada; c Keenan Research Centre, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada; d Department of Medical Imaging, St. Michael’s Hospital, Toronto, ON, Canada; e Department of Obstetrics and Gynecology, and Division of Endocrinology and Metabolism, St. Michael’s Hospital, Toronto, ON, Canada; f Division of Nephrology, St Paul’s Hospital, University of British Columbia, Vancouver, BC, Canada; g Department of Radiology and Medicine, St. Paul’s Hospital, University of British Columbia, Vancouver, BC, Canada; and h Terrence Donnelly Heart Centre, St. Michael’s Hospital, Toronto, ON, Canada Manuscript received November 7, 2014 and accepted January 19, 2015 b
Abstract Hypertension is prevalent in patients with end–stage renal disease and is strongly associated with left ventricular hypertrophy (LVH), an independent predictor of cardiovascular mortality. Blood pressure (BP) monitoring in hemodialysis patients may be unreliable because of its lability and variability. We compared different methods of BP measurement and their relationship with LVH on cardiac magnetic resonance imaging. Sixty patients undergoing chronic hemodialysis at a single dialysis center had BP recorded at each dialysis session over 12 weeks: pre–dialysis, initial dialysis, nadir during dialysis, and post–dialysis. Forty–five of these patients also underwent 44–hour inter–dialytic ambulatory BP monitoring. Left ventricular mass index (LVMI) was measured using cardiac magnetic resonance imaging and the presence of LVH was ascertained. Receiver operator characteristic curves were generated for each BP measurement for predicting LVH. The mean LVMI was 68 g/m2 (SD ¼ 15 g/m2); 13/60 patients (22%) had LVH. Mean arterial pressure measured shortly after initiation of dialysis session was most strongly correlated with LVMI (Pearson correlation coefficient r ¼ 0.59, P < .0001). LVH was best predicted by post–dialysis systolic BP (area under the curve, 0.83; 95% confidence interval, 0.72–0.94) and initial dialysis systolic BP (area under the curve, 0.81; 95% confidence interval, 0.70–0.92). Forty–four–hour ambulatory BP and BP variability did not significantly predict LVH. Initial dialysis mean arterial pressure and systolic BP and post–dialysis systolic BP are the strongest predictors of LVH, and may represent the potentially best treatment targets in hemodialysis patients to prevent end–organ damage. Further studies are needed to confirm whether treatment targeting these BP measurements can optimize cardiovascular outcomes. J Am Soc Hypertens 2015;9(4):275–284. Ó 2015 American Society of Hypertension. All rights reserved. Keywords: Hemodialysis; hypertension; left ventricular hypertrophy; mean arterial pressure.
This study was supported by an operating grant from the Canadian Institutes of Health Research (MOP 89982). Conflict of interest: none. *Corresponding author: Dr. Ron Wald, MD, MPH, St. Michael’s Hospital, Division of Nephrology, 30 Bond Street, Toronto, Ontario, Canada M5B 1W8. E-mail: [email protected]
**Corresponding author: Dr Andrew T. Yan, MD, Division of Cardiology, St Michael’s Hospital, 30 Bond St, Room 6-030 Donnelly, Toronto, Ontario, Canada M5B 1W8. Tel: 416-8645465; Fax 416-864-5159. E-mail: [email protected]
1933-1711/$ - see front matter Ó 2015 American Society of Hypertension. All rights reserved. http://dx.doi.org/10.1016/j.jash.2015.01.011
276
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Introduction Cardiovascular disease is the leading cause of mortality and hospitalization in patients undergoing chronic hemodialysis (HD).1–5 Cardiovascular risk factors continue to be relatively under–treated,6 and hypertension is common in the end–stage renal disease (ESRD) population with an estimated prevalence of 50%–80%.6 In HD patients, blood pressure (BP) is partly related to extracellular volume status, which is largely dependent on the timing, frequency, and adequacy of dialysis.7 Since 2009, the Canadian Hypertension Education Program (CHEP) has recommended targeting a BP of less than 140/908; however, there is no consensus on when to measure BP—before, during, or after dialysis—and more importantly, which reading best correlates with cardiovascular outcomes. Current practice guidelines recommend using pre–dialysis systolic BP (SBP) to guide treatment of hypertension based on Grade C level evidence.9 Several studies have shown a close relation between BP and left ventricular mass (LVM)10,11 and left ventricular hypertrophy (LVH), which is strongly associated with cardiovascular events in patients with and without kidney disease.12,13 Furthermore, change in LVM has been used as a primary surrogate endpoint for therapeutic interventions in landmark trials of renal replacement therapies.14,15 Conversely, LVM regression is associated with favorable outcomes, including decreased likelihood of developing heart failure and lower all–cause mortality.16,17 Currently, cardiac magnetic resonance imaging (CMR) provides the most accurate and reproducible measurement of LVM.18,19 Although BP is an important and modifiable risk factor for cardiovascular disease in dialysis recipients, it is not clear which BP measurements should be used for therapeutic guidance. Given the clinical significance of LVH as a marker of end–organ damage and predictor of adverse cardiovascular events, the objective of this study was to evaluate the relationship between different BP measurements (pre–dialysis, initial dialysis, nadir dialysis, post–dialysis, inter–dialytic ambulatory) and LVM, as measured by CMR, in recipients of conventional HD.
Methods Study Design This was a cross–sectional study based on a single tertiary–center cohort of 60 prevalent in–center HD patients at St Michael’s Hospital in Toronto, Canada. All patients were receiving conventional HD (4 hours per session, 3–4 times weekly) at the time of assessment. Participants were recruited for an observational study comparing the cardiovascular impact of conversion to in-center nocturnal HD versus continuation of conventional dialysis. The data included herein reflect the baseline data for all study
participants who were receiving conventional HD. Adults 18 years or older with ESRD who were receiving conventional HD for at least 3 months were eligible for this study. Exclusion criteria for the study were any serious co– morbidity with a life expectancy of less than 1 year, a planned live donor kidney transplant within the next 12 months, contraindications to CMR, pregnancy, or inability to provide informed consent. BP data were collected during all dialysis sessions during a 12–week period.
Patient Demographic Data A chart review was performed for each patient via the St Michael’s Hospital electronic patient record system. We collected demographic and clinical data, which included age, gender, cause of ESRD, dialysis vintage, type of vascular access, history of coronary artery disease (defined as previous myocardial infarction or revascularization), cerebrovascular disease (history of documented stroke), diabetes, and peripheral vascular disease (amputation or peripheral revascularization). Relevant medications were also recorded including beta blockers, renin–angiotensin– aldosterone system (RAAS) blocking agents, cholesterol lowering agents, antiplatelet agents, anticoagulants, erythropoietin stimulating agents, and phosphate binders.
In–center Blood Pressure and Weight Measurements BP readings were collected retrospectively from dialysis treatment records over a 12–week period prior to CMR. These were routine blood pressure measurements taken by the dialysis nurses as part of standard dialysis care, using automated BP monitors built into the dialysis machines (Phoenix, Gambro, Richmond Hill, ON). A pre–dialysis BP was measured upon patient arrival in the dialysis unit but prior to the start of dialysis. Upon initiation of dialysis, the first recorded BP reading taken was termed the ‘‘initial dialysis BP’’; this was taken within the first 15 minutes after the HD circuit was initiated; subsequent BP measurements were taken every hour and at the nurse’s discretion. Post–dialysis BP was measured after cessation of dialysis within 15 minutes. The lowest BP reading measured at any point during the dialysis session was considered the ‘‘nadir BP.’’ Corresponding pulse rates were recorded with each BP reading. The dialysis records also contained information on pre–dialysis and post–dialysis weight, as well as volume of ultra–filtration achieved in each session. Inter– and intra–dialytic weight changes were calculated from these data.
Ambulatory Blood Pressure Monitoring Ambulatory blood pressure monitoring (ABPM) data were acquired over a 44–hour period. The duration of
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
monitoring reflected the time interval between the two consecutive HD sessions (first or midweek). BPs were recorded every 30 minutes using an ABPM device (TM– 2430; A&D Company Limited) in the non–access arm. A nurse in the BP lab programmed the hours of sleep, and recorded the mean total, awake, and nocturnal BP data accordingly. Patients who had a limited number of BP readings were not excluded from the analysis, since even these have been demonstrated to have prognostic significance.20
CMR All patients underwent CMR with a 1.5 Tesla whole– body magnetic resonance imaging scanner (Intera; Philips Medical Systems, Best, The Netherlands) using a phased– array cardiac coil and retrospective vectorocardiographic gating. Whenever possible, CMR was performed post–dialysis. A standardized CMR protocol was used as previously published.19 Images were acquired during breath–holds in end–expiration with the patient in supine position; 8–12 contiguous short–axis cine images were required to cover the entire left ventricle. Segmented, balanced steady–state free–precession imaging sequence was used, and typical imaging parameters were repetition time (TR), 4 ms; echo time (TE), 2 ms; slice thickness, 8 mm; field of view, 30–34 cm 30–34 cm; matrix size, 256 196, temporal resolution of <40 ms (depending on heart rate); flip angle, 50 . A blinded experienced cardiac imager (AY) reviewed all the CMR studies and performed the image post–processing offline using commercial software (ViewForum R 4.2; Philips Medical Systems). Manual tracing of endocardial borders at end–diastole and at end–systole using short–axis cine images was performed. Endocardial and epicardial borders in contiguous short–axis slices at end–diastole were also traced, and the difference in area was multiplied by the slice thickness. The sum of these differences throughout the entire left ventricle was multiplied by the myocardial specific density (1.05 g/cm3) to calculate the LVM. Papillary muscles were excluded from the LVM measurement. LVM values were then divided by body surface area (BSA; height in cm weight in kg/3600) to yield LVM index (LVMI). Left ventricular hypertrophy (LVH) was defined as LVMI >81 g/m2 for men and LVMI >79 g/m2 for women.19,21 To account for the fluctuation in intravascular volume status in HD patients, we also examined LVM/LV end–diastolic volume (LVEDV), which is an established prognostic marker for cardiovascular events.22
Laboratory Measurements Results of routine laboratory tests during the 12–week period were extracted retrospectively from electronic medical record and averaged. These included plasma
277
concentrations of calcium (corrected for albumin), phosphate, parathyroid hormone, alkaline phosphatase, ferritin, and hemoglobin.
Statistical Analysis Data are presented as means standard deviation (SD) or median with inter–quartile ranges for continuous variables, and frequency (percent) for categorical data. BP variability was measured by calculating the standard deviation of the all SBPs, DBPs, and the mean arterial pressures (MAP) over the 12–week period. MAP were calculated as (SBP þ 2 DBP)/3. Pulse pressure was calculated as (SBPDBP). Fisher’s exact test was used to assess the relation between patient demographics and LVH. Student t–test was used to compare the difference in SBP, DBP, and MAP in the groups with and without LVH. Pearson product moment correlation coefficients are reported for each method of blood pressure measurement against LVMI as well as LVM/LVEDV. Pearson correlation coefficient of initial dialytic MAP was compared in pairwise fashion with pre–dialysis MAP, SBP, and DBP (Hochberg’s step– up procedure for multiple pairwise comparisons). LVH was analyzed as a dichotomous variable based on the normal gender–specific reference values for CMR–derived LVM,19 and receiver operating characteristic (ROC) curves were generated and thus areas under the curves (AUC) were calculated for the different BP measurements. Multivariable linear regression analyses were performed with LVMI as the dependent variable and each BP parameter as the independent variable, adjusting for age, gender, dialysis vintage, and number of anti–hypertensive medications. We examined the model R2 and verified model assumptions. Analyses were performed using SPSS version 22 (IBM).
Results Baseline Characteristics Table 1 shows the baseline characteristics of 60 patients with complete CMR data. The median age of the patients was 54 years (interquartile range, 44–60 years), and 44% were women. Diabetic nephropathy was the most common cause of ESRD, followed by glomerulonephritis. Vascular co–morbidities were common in the study population, including coronary artery disease and diabetes mellitus. The majority of patients were using anti–platelet agents and antihypertensive agents, primarily RAAS blockers. The mean LVMI was 68 g/m2 (SD ¼ 15 g/m2). Overall, 13 of 60 patients (22%) had LVH. Baseline characteristics of the patients were similar in patients with and without LVH (Table 1). There were no significant differences in causes of ESRD, co–morbidities, medications used, or levels of calcium, phosphate, or hemoglobin. Patients
278
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Table 1 Baseline patient characteristics Characteristics Age, y* Female Comorbid conditions Coronary artery disease Cerebrovascular disease Diabetes mellitus Peripheral vascular disease Ethnicity Caucasian Black Other AV fistula Duration of dialysis, mo* Cause of ESRD Diabetes mellitus Glomerulonephritis Polycystic kidney disease Hypertension Focal sclerosing glomerulosclerosis Other Medication use Number of antihypertensives* Antiplatelet agent Warfarin Beta–blocker ACE inhibitor or angiotensin II receptor blocker Lipid lowering agent ESA Calcium–based phosphate binder Non–calcium based phosphate binder Serum calcium corrected, mmol/L* Serum phosphate, mmol/L* Serum iPTH, pmol/L* Hemoglobin concentration, g/L*
No LVH (n ¼ 47) 53 (44–61) 22 (47) 14 3 24 6 14 12 21 22 40.3 14 14 1 2 3 11 2 27 7 23 36 25 44 36 18 2.26 1.77 40.3 112
(30) (6) (51) (13) (30) (26) (45) (48) (11.0–52.1) (30) (30) (2) (4) (7) (24) (1-2) (59) (15) (49) (76) (53) (94) (78) (39) (2.15–2.4) (1.46–2.05) (20.5–63.4) (106–120)
LVH (n ¼ 13) 58 (47–60) 4 (31) 2 1 6 1 5 4 4 7 49.2 5 2 1 3 0 4 3 8 0 9 13 4 15 14 7 2.36 1.5 49.5 110
(15) (7) (46) (8) (39) (31) (31) (47) (20.0–62.5) (38) (13) (7) (20) (0) (27) (2-3) (53) (0) (60) (87) (27) (100) (93) (47) (2.21–2.49) (1.24–1.66) (37.2–98.5) (107–115)
P–value .55 .36 .31 1.00 .56 .67
.66 1.00 .82
.26
.03 .77 .18 .77 .71 .07 1.00 .26 .76 .59 .10 .34 .73
ACE, Angiotensin–converting enzyme; AV, arteriovenous; ESA, erythropoeitin stimulating agent; ESRD, end–stage renal disease; iPTH, intact parathyroid hormone level; LVH, left ventricular hypertrophy. All data are presented as n (%) unless otherwise specified. * Median (interquartile range).
with LVH were taking more anti–hypertensive medications than patients without LVH (Table 1).
dialysis MAP, post–dialysis MAP, and mean ambulatory MAP (Table 3). In contrast, pulse pressure measurements did not differ significantly between the two groups (Table 3).
Blood Pressure Measurements The average BPs are shown in Table 2. ABPM studies were completed for 45 of 60 patients. SBP readings were significantly higher in patients with LVH as compared with patients without LVH (Table 2). There was no statistically significant difference in DBPs between patients with and without LVH except initial dialysis DBP (P ¼ .047). There were no significant differences in the two groups with respect to the amount of volume ultra–filtered during dialysis. Compared with the group without LVH, patients with LVH had higher pre–dialysis MAP, initial dialysis MAP, nadir
Relationship between BP and LVMI Table 4 demonstrates the associations of the various MAPs, SBPs, and DBPs with LVMI. Initial–dialysis MAP best correlated with LVMI with a Pearson correlation of 0.59 (P < .0001), followed by initial dialysis SBP (r ¼ 0.55; P < .0001; Figure 1). Compared with the corresponding pre–dialysis BP measurements (pairwise comparisons), initial dialysis MAP (P ¼ .009), SBP (P ¼ .01), and DBP (P ¼ .025) all demonstrated significantly stronger correlations with LVMI.
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
279
Table 2 Left ventricular mass by magnetic resonance imaging and blood pressure data for patients with and without left ventricular hypertrophy (LVH) Characteristics
No LVH (n ¼ 47)
LVH (n ¼ 13)
P–value
LV mass, g LV mass index, g/m2 LVEF, % Intra–dialytic weight loss, kg Blood pressure measurements, mm Hg Pre–dialysis SBP Pre–dialysis DBP Initial dialysis SBP Initial dialysis DBP Nadir dialysis SBP Nadir dialysis DBP Post–dialysis SBP Post–dialysis DBP Ambulatory SBP* Ambulatory DBP*
116 (93.5–133.2) 63 (56.5–68) 61 (57–64) 2.68 0.89
165 (153–190) 91 (88–93) 63 (57–65) 2.58 0.83
.38 .71
140 78 136 76 112 67 127 73 132 77
15 8 15 9 13 9 12 8 17 9
152 83 151 83 126 72 140 77 142 83
8 10 10 10 20 13 8 13 13 14
.013 .058 .0076 .047 .026 .14 .018 .20 .011 .058
DBP, Diastolic blood pressure; LV, left ventricular; LVEF, left ventricular ejection fraction; SBP, systolic blood pressure. Data presented as mean standard deviation or median (interquartile range). * Total sample size for ambulatory blood pressure measurement was 45 patients. Thirty–six patients did not have LVH, and nine patients had LVH.
All SBPs had better correlation with LVMI than DBPs during dialysis. Mean ambulatory DBP were better correlated with LVMI than mean ambulatory SBP (Table 4), but both showed only weak correlations with LVMI. There was no correlation between LVMI and intra–dialytic BP lability for SBP (P ¼ .52), DBP (P ¼ .90), or MAP (P ¼ .76; Table 4). Intra–dialytic BP variability also did not significantly predict LVH (Table 5). Table 3 Average mean arterial pressures (MAP) and pulse pressure for patients with and without left ventricular hypertrophy (LVH) by magnetic resonance imaging Blood Pressure Measurement MAP, mm Hg Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory* Pulse pressure, mm Hg Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory*
No LVH (n ¼ 47)
LVH (n ¼ 13)
P–value
99 96 82 91 95
9 9 10 8 11
106 105 90 98 104
8 9 14 10 13
.008 .0007 .018 .004 .041
63 60 46 54 55
13 13 9 11 11
69 68 55 63 59
12 12 17 11 9
.38 .13 .09 .19 .33
Data presented as mean standard deviation. * Total sample size for ambulatory blood pressure measurement was 45 patients. Thirty–six patients did not have LVH, and nine patients had LVH.
In an ancillary analysis, BP was correlated with LVM adjusted for left ventricular end diastolic volume (LVM/ LVEDV). Only initial dialysis MAP and initial dialysis SBP significantly correlated with LVM/LVEDV (r ¼ 0.26, P ¼ .045 and r ¼ 0.261, P ¼ .044 respectively). The areas under the ROC curves are summarized in Table 5. Post–dialysis and initial dialysis SBP performed best in predicting LVH (AUCs 0.83 [95% CI, 0.72–0.94] and 0.81 [95% CI, 0.70–0.92]), respectively. Initial dialysis MAP also performed well in its ability to predict LVH with an AUC of 0.78 (95% CI, 0.64–0.92). Apart from the initial dialysis DBP, none of the other DBP measures appeared to predict LVH. None of the ambulatory BP metrics was a significant predictor of LVH. For the in–center BP readings, compared with MAP, SBP was a better predictor of LVH. There remained independent associations between these BP measurements and LVMI after adjusting for age, gender, duration of dialysis and number of anti–hypertensive medications.
Discussion In this cross–sectional study of 60 patients receiving conventional HD, initial dialysis SBP and MAP and post–dialysis SBP and were most closely correlated with LVMI and the presence of LVH, as assessed by CMR. Our findings suggest these BP measurements may identify patients at risk of adverse cardiovascular outcomes and may represent useful therapeutic targets. Our study was prompted by a lack of consensus regarding which BP measurement best predicts clinical
280
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Table 4 Pearson’s product correlation coefficients for in–center dialysis (n ¼ 60) and ambulatory (n ¼ 45) blood pressure (BP) measurements, and BP lability with left ventricular mass index Blood Pressure Measurement
Mean Arterial Pressure Pearson’s Correlation Coefficient (95% CI)
P–value
Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory blood pressure* Lability
0.51 0.59 0.45 0.47 0.37
<.0001 <.0001 .0003 .0002 .013
(0.29–0.67) (0.40–0.74) (0.23–0.64) (0.24–0.65) (0.08–0.60)
0.04 (0.22–0.29)
Systolic BP
.76
Pearson’s Correlation Coefficient (95% CI) 0.47 0.55 0.47 0.53 0.33
(0.25–0.65) (0.34–0.70) (0.25–0.65) (0.32–0.69) (0.04–0.57)
0.085 (–0.17–0.33)
Diastolic BP P–value <.0001 <.0001 .001 <.001 .029 .52
Pearson’s Correlation Coefficient (95% CI) 0.41 0.50 0.37 0.32 0.36
(0.18–0.60) (0.28–0.67) (0.13–0.57) (0.07–0.53) (0.07–0.59)
0.017 (0.24–0.27)
P–value .001 .0001 .004 .013 .015 .90
CI, Confidence interval. * Total sample size for ambulatory blood pressure measurement was 45 patients. Thirty–six patients did not have left ventricular hypertrophy, and nine patients had left ventricular hypertrophy.
outcomes in dialysis recipients. This information is vital as it would serve to guide therapeutic targets in this population. Current guidelines recommend targeting treatment to pre–dialysis SBP, although this recommendation is based on rather limited data (grade C level evidence).9 Several studies have sought to determine the optimal BP measurement strategy, with conflicting results. Conlon et al concluded that pre–dialysis SBP best correlated with LVM (r ¼ 0.35, P ¼ .03) and should be the target of hypertension control.23 Their study was conducted over 12 dialysis sessions among a predominantly African American population on stable antihypertensive regimens (unchanged in the 3 months of the study),23 which may have limited the generalizability of their results. The Cardiovascular Risk Extended Evaluation investigators also found a significant association between LVMI and pre–dialysis SBP.24 Additionally, pre–dialysis DBP and pulse pressure were associated with LVMI.24 This study excluded diabetic patients and those with heart failure. Furthermore, the main anti– hypertensives used by patients in this population were beta–blockers, which does not reflect current practice.9 Foley et al examined the United States Renal Data System (USRDS) database of 11,142 subjects on HD, and found that higher post–dialysis SBP, but not pre–dialysis SBP, was associated with all–cause mortality. In this large study, wide post–dialysis pulse pressure was also associated with higher mortality. However, they only studied pre–dialysis and post–dialysis BP and used an average of only the last three BP values recorded prior to the study. They did not investigate if there was an association with cardiovascular mortality or events, and did not focus on patho–physiologic mechanisms such as end–organ damage.9,25 Cao et al found that both pre–dialysis SBP and DBP were significantly associated with LVH in 164 Chinese patients.26 In contrast, Koc et al found that in a sample population of 74 patients, both post–dialysis SBP (r ¼ 0.471, P < .001) and pre–dialysis SBP (r ¼ 0.461, P < .001) correlated with increased
LVMI as determined by echocardiography.27 They did not evaluate other in–center dialysis BP measurements and did not test which measurement best predicted the presence of LVH. In a study of 150 predominantly African American patients from four dialysis units in Cleveland, Agarwal et al showed a stronger association between LVH and SBP measured at home (using self–monitoring and ABPM) than with pre– or post–dialysis SBP.28 All peri–dialytic BPs had similar correlations with LVMI. The same group also examined pre–dialysis, post–dialysis, and intra– dialysis BP monitoring using 44–hour ABPM as a standard to diagnose hypertension. They showed that both mean intra–dialytic SBP and median intra–dialytic SBP had comparable diagnostic performance metrics.29 In this study, they did not correlate their findings with any marker of end–organ damage. More recently in 2010, Agarwal et al reported that out–of–dialysis unit BP had greater prognostic significance compared to pre– and post–dialysis BP, and higher ambulatory BP was associated with increased mortality.30 Although these studies provide useful mechanistic insights, they utilized echocardiography to measure LVM. Echocardiographic assessment of LVM relies on geometric assumptions that may not be valid in chronic dialysis patients due to cardiac remodeling from pressure and volume overload. Another limitation is the use of a formula in echocardiography that relies on the internal diameter of the left ventricular chamber; intra-cardiac volume may fluctuate greatly between pre–dialytic and post–dialytic periods, and may therefore cause considerable variation in the LVM calculation.31 Two studies have employed CMR to study the relation of BP measurements to LVMI, but each studied a limited set of BP parameters, and neither of them studied ABPM. Patel et al were the first investigators to use CMR to study the relation between BP and LVH in a cohort of 246 hemodialysis patients.32 MRI was performed 24 hours after the last dialysis session. They found pre–dialysis SBP to be a
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Figure 1. Scatterplots (n¼60) demonstrating the relationships of left ventricular mass index and (A) initial dialysis mean arterial pressure (MAP); (B) initial dialysis systolic blood pressure (SBP); and (C) post–dialysis SBP.
281
robust independent predictor of LVH. However, they did not compare initial dialysis, or nadir dialysis BP. In contrast, Khangura et al performed a study using CMR in a smaller sample size (n ¼ 39), and only studied pre–, post–, and intra–dialytic (average of all BP readings during a single dialysis session only) SBPs as well as a standardized pre–dialysis SBP.33 Their results supported a single baseline intra–dialytic SBP (AUC, 0.80; 95% CI, 0.64– 0.96) or post–dialysis SBP (AUC, 0.79; 95% CI, 0.64– 0.93) as being the best predictors of LVH. Pre–dialysis SBP poorly correlated with LVMI (r ¼ 0.3; P ¼ .068).33 However, they did not study MAP or DBP. A major strength of the current study was the use of CMR instead of echocardiography to assess LVM. CMR is now recognized as the gold standard for LVM measurement. It is more accurate and precise than echocardiography.18,19 In a comparison between CMR and echocardiography, echocardiography overestimated LV mass in the presence of LVH and dilatation, especially in the HD population.18 A recent study performed within the current cohort also showed a significant degree of overestimation of LVM and LVH prevalence by echocardiography in patients with ESRD.19 Another strength and novel aspect of our study was the comprehensive assessment of different in–center BP measurements (pre–dialysis, intra–dialytic, nadir dialysis, post–dialysis) and ABPM, and their association with LVH using CMR. We were able to collect a large amount of data over 3 months, with multiple BP measurements; this minimized effects of outliers and spurious results. Our study also reflected actual dialysis unit nursing practices and limitations, and therefore may be more representative of the information available to clinicians. We found that mean 44–hour ambulatory MAPs did positively correlate with LVMI, but the association was not as strong compared with other methods (Table 4). Interestingly, we found that ambulatory SBP, DBP, or MAP did not significantly predict LVH (Table 5). This result is quite different from what was reported by Agarwal et al.28 Our results suggest that BP monitoring in the dialysis unit over a longer period of time may be even more useful than ambulatory BP monitoring, which provides much fewer readings over a relatively shorter time period. Average BPs calculated from many readings may better predict end–organ damage, especially in dialysis patients, in whom wide fluctuations in BP are common. Of note, BP variability was not predictive of LVH in this study, which to the best of our knowledge, is the first to address this question. BP variability may result from changes in filling pressures in the setting of LVH and diastolic dysfunction, or heart failure with preserved ejection fraction. This is also a common cause of intra–dialytic hypotension, which is associated with increased cardiovascular mortality.34 LV systolic function was preserved in our study cohort, but diastolic dysfunction was not specifically
282
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Table 5 Receiver operating characteristics (ROC) statistics for left ventricular hypertrophy (LVH) for in–center dialysis (n ¼ 60) and ambulatory (n ¼ 45) blood pressure (BP) measurements Blood Pressure Measurement
Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory blood pressure* Lability
Mean Arterial Pressure
Systolic BP
Diastolic BP
ROC AUC (95% CI)
ROC AUC (95% CI)
ROC AUC (95% CI)
0.74 0.78 0.67 0.70 0.68 0.49
0.76 0.81 0.73 0.83 0.67 0.51
0.63 0.69 0.57 0.58 0.62 0.50
(0.59, (0.64, (0.48, (0.53, (0.47, (0.28,
0.88) 0.92) 0.85) 0.87) 0.88) 0.70)
(0.64, (0.70, (0.55, (0.72, (0.49, (0.30,
0.88) 0.92) 0.90) 0.94) 0.85) 0.72)
(0.44, (0.51, (0.37, (0.38, (0.38, (0.31,
0.82) 0.88) 0.78) 0.79) 0.86) 0.69)
AUC, Area under the curve; CI, confidence interval. * Total sample size for ambulatory blood pressure measurement was 45 patients, of whom nine had LVH.
assessed. Similar to BP variability, nadir BP was not significantly associated with LVH or LVMI. Most studies have investigated SBP and DBP, but there are a paucity of data on the association between LVH and MAPs or pulse pressures in the dialysis population. Increase in MAP has been shown to be associated with progressive increase in LVH and development of heart failure and ischemic heart disease,35 as well as higher cardiovascular mortality.20,36 We demonstrated that MAP correlated well with LVMI, but it did not predict LVH as well as SBP. In our study, pulse pressures did not correlate significantly with LVMI. Pulse pressure has been shown to be an independent predictor of coronary heart disease and all–cause mortality.25,37 This may be a phenomenon only applicable to central pulse pressure, as one study found carotid pulse pressure to be associated with all–cause and cardiovascular mortality, but not brachial pulse pressure.37 Furthermore, the association of pulse pressure with cardiovascular outcomes and mortality has been found to be independent of LVH and systolic dysfunction.38 Frequent changes in intravascular volume status may pose challenges to LVM measurements when geometric assumptions are made, for example, in echocardiography. Because CMR applies a truly volumetric approach to LVM measurements (ie, it directly measures the volume of the myocardium, rather than using a formula related to the LV cavity volume), it is relatively insensitive to changes in volume status. Moreover, we performed an additional analysis by normalizing LVM to LVEDV (LVM/LVEDV), thereby allowing for volume changes with HD. The left ventricular mass to volume ratio has been previously shown in the large Multi–Ethnic Study of Atherosclerosis to independently predict incident coronary heart disease and stroke.22 Thus, the significant correlations between initial dialysis MAP and SBP and LVM/LVEDV lend further credence to the prognostic significance of these BP measurements. There were several limitations in our study. This was a single center study with a limited number of patients.
Second, this was a cross–sectional observational study, and causality cannot be established between elevated BP and LVH. Therefore, caution must be exercised when interpreting these results. We used LVMI and the presence of LVH as our primary outcome of interest, and our study was not powered to examine cardiovascular outcomes. Our study population had a lower prevalence of LVH than most other studies; although this might be partly due to an over–estimation of LVH by echocardiography or different definition of LVH, this might also reflect the particular case mix at our single dialysis center that is different from other study populations. Another limitation to our study was that 44–hour ABPM was not available for all patients. This may have underpowered our study and led to an apparently weaker correlation between ambulatory blood pressure and left ventricular mass. It should also be noted that all other BP data were extracted retrospectively from routine dialysis records. These BP measurements were not done according to a standardized research protocol. We were therefore not able to compare HD unit BP data with inter–dialytic standardized BPs.
Conclusions In conclusion, initial dialysis MAP, SBP, and post–dialysis SBP both strongly correlate with LVMI and LVH, as measured by CMR. These findings suggest that therapies targeting these BP measurements may attenuate end–organ damage, thereby improving cardiovascular outcomes. These results need to be confirmed in larger multi–center clinical trials randomizing a broader spectrum of HD patients to various BP targets with long–term follow up. Acknowledgment The authors would like to thank all the patients and staff at the St Michael’s Hospital hemodialysis unit who contributed the data for this study.
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
References 1. Foley RN, Parfrey PS. Cardiovascular disease and mortality in ESRD. J Nephrol 1998;11:239–45. 2. Mittal SK, Kowalski E, Trenkle J, McDonough B, Halinski D, Devlin K, et al. Prevalence of hypertension in a hemodialysis population. Clin Nephrol 1999;51: 77–82. 3. Rostand SG, Brunzell JD, Cannon RO III, Victor RG. Cardiovascular complications in renal failure. J Am Soc Nephrol 1991;2:1053–62. 4. Salem MM. Hypertension in the hemodialysis population: a survey of 649 patients. Am J Kidney Dis 1995;26:461–8. 5. Tomita J, Kimura G, Inoue T, Inenaga T, Sanai T, Kawano Y, et al. Role of systolic blood pressure in determining prognosis of hemodialyzed patients. Am J Kidney Dis 1995;25:405–12. 6. Agarwal R, Nissenson AR, Batlle D, Coyne DW, Trout JR, Warnock DG. Prevalence, treatment, and control of hypertension in chronic hemodialysis patients in the United States. Am J Med 2003;115:291–7. 7. Horl MP, Horl WH. Hemodialysis-associated hypertension: pathophysiology and therapy. Am J Kidney Dis 2002;39:227–44. 8. Prasad GV, Ruzicka M, Burns KD, Tobe SW, Lebel M. Hypertension in dialysis and kidney transplant patients. Can J Cardiol 2009;25:309–14. 9. Jindal K, Chan CT, Deziel C, Hirsch D, Soroka SD, Tonelli M, et al. Hemodialysis clinical practice guidelines for the Canadian Society of Nephrology. J Am Soc Nephrol 2006;17(3 Suppl 1):S1–27. 10. Harnett JD, Kent GM, Barre PE, Taylor R, Parfrey PS. Risk factors for the development of left ventricular hypertrophy in a prospectively followed cohort of dialysis patients. J Am Soc Nephrol 1994;4:1486–90. 11. Devereux RB, de SG, Koren MJ, Roman MJ, Laragh JH. Left ventricular mass as a predictor of development of hypertension. Am J Hypertens 1991; 4:603S–7S. 12. Silberberg JS, Barre PE, Prichard SS, Sniderman AD. Impact of left ventricular hypertrophy on survival in end-stage renal disease. Kidney Int 1989;36:286–90. 13. Bouzas-Mosquera A, Broullon FJ, Alvarez-Garcia N, Peteiro J, Mosquera VX, Castro-Beiras A. Association of left ventricular mass with all-cause mortality, myocardial infarction and stroke. PLoS One 2012;7: e45570. 14. Kannel WB, Gordon T, Offutt D. Left ventricular hypertrophy by electrocardiogram. Prevalence, incidence, and mortality in the Framingham study. Ann Intern Med 1969;71:89–105. 15. Casale PN, Devereux RB, Milner M, Zullo G, Harshfield GA, Pickering TG, et al. Value of
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
283
echocardiographic measurement of left ventricular mass in predicting cardiovascular morbid events in hypertensive men. Ann Intern Med 1986;105:173–8. London GM, Pannier B, Guerin AP, Blacher J, Marchais SJ, Darne B, et al. Alterations of left ventricular hypertrophy in and survival of patients receiving hemodialysis: follow-up of an interventional study. J Am Soc Nephrol 2001;12:2759–67. Foley RN, Parfrey PS, Kent GM, Harnett JD, Murray DC, Barre PE. Serial change in echocardiographic parameters and cardiac failure in end-stage renal disease. J Am Soc Nephrol 2000;11:912–6. Stewart GA, Foster J, Cowan M, Rooney E, McDonagh T, Dargie HJ, et al. Echocardiography overestimates left ventricular mass in hemodialysis patients relative to magnetic resonance imaging. Kidney Int 1999;56:2248–53. Jakubovic BD, Wald R, Goldstein MB, Leong-Poi H, Yuen DA, Perl J, et al. Comparative assessment of 2dimensional echocardiography vs cardiac magnetic resonance imaging in measuring left ventricular mass in patients with and without end-stage renal disease. Can J Cardiol 2013;29:384–90. Gerin W, Schwartz JE, Devereux RB, Goyal T, Shimbo D, Ogedegbe G, et al. Superiority of ambulatory to physician blood pressure is not an artifact of differential measurement reliability. Blood Press Monit 2006;11:297–301. Hudsmith LE, Petersen SE, Francis JM, Robson MD, Neubauer S. Normal human left and right ventricular and left atrial dimensions using steady state free precession magnetic resonance imaging. J Cardiovasc Magn Reson 2005;7:775–82. Bluemke DA, Kronmal RA, Lima JA, Liu K, Olson J, Burke GL, et al. The relationship of left ventricular mass and geometry to incident cardiovascular events: the MESA (Multi-Ethnic Study of Atherosclerosis) study. J Am Coll Cardiol 2008;52:2148–55. Conlon PJ, Walshe JJ, Heinle SK, Minda S, Krucoff M, Schwab SJ. Predialysis systolic blood pressure correlates strongly with mean 24-hour systolic blood pressure and left ventricular mass in stable hemodialysis patients. J Am Soc Nephrol 1996;7:2658–63. Zoccali C, Mallamaci F, Tripepi G, Benedetto FA, Cottini E, Giacone G, et al. Prediction of left ventricular geometry by clinic, pre-dialysis and 24-h ambulatory BP monitoring in hemodialysis patients: CREED investigators. J Hypertens 1999;17(12 Pt 1):1751–8. Foley RN, Herzog CA, Collins AJ. Blood pressure and long-term mortality in United States hemodialysis patients: USRDS Waves 3 and 4 Study. Kidney Int 2002;62:1784–90. Cao X, Zou J, Teng J, Zhong Y, Ji J, Chen Z, et al. BMI, spKt/V, and SBP but not DBP are related to
284
27.
28.
29.
30. 31.
32.
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
LVH in Chinese maintenance hemodialysis patients. Ren Fail 2011;33:269–75. Koc Y, Unsal A, Kayabasi H, Oztekin E, Sakaci T, Ahbap E, et al. Impact of volume status on blood pressure and left ventricle structure in patients undergoing chronic hemodialysis. Ren Fail 2011;33:377–81. Agarwal R, Brim NJ, Mahenthiran J, Andersen MJ, Saha C. Out-of-hemodialysis-unit blood pressure is a superior determinant of left ventricular hypertrophy. Hypertension 2006;47:62–8. Agarwal R, Metiku T, Tegegne GG, Light RP, Bunaye Z, Bekele DM, et al. Diagnosing hypertension by intradialytic blood pressure recordings. Clin J Am Soc Nephrol 2008;3:1364–72. Agarwal R. Blood pressure and mortality among hemodialysis patients. Hypertension 2010;55:762–8. Kilickap M, Turhan S, Sayin T, Nergizoglu G, Kutlay S, Duman N, et al. Intravascular volume dependency of left ventricular mass calculation by twodimensional guided M-mode echocardiography. Can J Cardiol 2007;23:219–22. Patel RK, Oliver S, Mark PB, Powell JR, McQuarrie EP, Traynor JP, et al. Determinants of left ventricular mass and hypertrophy in hemodialysis patients assessed by cardiac magnetic resonance imaging. Clin J Am Soc Nephrol 2009;4:1477–83.
33. Khangura J, Culleton BF, Manns BJ, Zhang J, Barnieh L, Walsh M, et al. Association between routine and standardized blood pressure measurements and left ventricular hypertrophy among patients on hemodialysis. BMC Nephrol 2010;11:13. 34. Losi MA, Memoli B, Contaldi C, Barbati G, Del PM, Betocchi S, et al. Myocardial fibrosis and diastolic dysfunction in patients on chronic haemodialysis. Nephrol Dial Transplant 2010;25:1950–4. 35. Foley RN, Parfrey PS, Harnett JD, Kent GM, Murray DC, Barre PE. Impact of hypertension on cardiomyopathy, morbidity and mortality in end-stage renal disease. Kidney Int 1996;49:1379–85. 36. Charra B, Calemard E, Ruffet M, Chazot C, Terrat JC, Vanel T, et al. Survival as an index of adequacy of dialysis. Kidney Int 1992;41:1286–91. 37. Safar ME, Blacher J, Pannier B, Guerin AP, Marchais SJ, Guyonvarc’h PM, et al. Central pulse pressure and mortality in end-stage renal disease. Hypertension 2002;39:735–8. 38. Palmieri V, Devereux RB, Hollywood J, Bella JN, Liu JE, Lee ET, et al. Association of pulse pressure with cardiovascular outcome is independent of left ventricular hypertrophy and systolic dysfunction: the Strong Heart Study. Am J Hypertens 2006;19: 601–7.
Research Article
Relationship between different blood pressure measurements and left ventricular mass by cardiac magnetic resonance imaging in end–stage renal disease Asad Merchant, MDa,b, Ron Wald, MD, MPHa,b,c,*, Marc B. Goldstein, MCDM, MDb,c, Darren Yuen, MD, PhDb,c, Anish Kirpalani, MDa,d, Niki Dacouris, BScb, Joel G. Ray, MD, MSca,e, Mercedeh Kiaii, MDf, Jonathan Leipsic, MDg, Vamshi Kotha, MDa,d, Djeven Deva, MDa,d, and Andrew T. Yan, MDa,c,h,** a
University of Toronto, Toronto, ON, Canada; Division of Nephrology, St. Michael’s Hospital, Toronto, ON, Canada; c Keenan Research Centre, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada; d Department of Medical Imaging, St. Michael’s Hospital, Toronto, ON, Canada; e Department of Obstetrics and Gynecology, and Division of Endocrinology and Metabolism, St. Michael’s Hospital, Toronto, ON, Canada; f Division of Nephrology, St Paul’s Hospital, University of British Columbia, Vancouver, BC, Canada; g Department of Radiology and Medicine, St. Paul’s Hospital, University of British Columbia, Vancouver, BC, Canada; and h Terrence Donnelly Heart Centre, St. Michael’s Hospital, Toronto, ON, Canada Manuscript received November 7, 2014 and accepted January 19, 2015 b
Abstract Hypertension is prevalent in patients with end–stage renal disease and is strongly associated with left ventricular hypertrophy (LVH), an independent predictor of cardiovascular mortality. Blood pressure (BP) monitoring in hemodialysis patients may be unreliable because of its lability and variability. We compared different methods of BP measurement and their relationship with LVH on cardiac magnetic resonance imaging. Sixty patients undergoing chronic hemodialysis at a single dialysis center had BP recorded at each dialysis session over 12 weeks: pre–dialysis, initial dialysis, nadir during dialysis, and post–dialysis. Forty–five of these patients also underwent 44–hour inter–dialytic ambulatory BP monitoring. Left ventricular mass index (LVMI) was measured using cardiac magnetic resonance imaging and the presence of LVH was ascertained. Receiver operator characteristic curves were generated for each BP measurement for predicting LVH. The mean LVMI was 68 g/m2 (SD ¼ 15 g/m2); 13/60 patients (22%) had LVH. Mean arterial pressure measured shortly after initiation of dialysis session was most strongly correlated with LVMI (Pearson correlation coefficient r ¼ 0.59, P < .0001). LVH was best predicted by post–dialysis systolic BP (area under the curve, 0.83; 95% confidence interval, 0.72–0.94) and initial dialysis systolic BP (area under the curve, 0.81; 95% confidence interval, 0.70–0.92). Forty–four–hour ambulatory BP and BP variability did not significantly predict LVH. Initial dialysis mean arterial pressure and systolic BP and post–dialysis systolic BP are the strongest predictors of LVH, and may represent the potentially best treatment targets in hemodialysis patients to prevent end–organ damage. Further studies are needed to confirm whether treatment targeting these BP measurements can optimize cardiovascular outcomes. J Am Soc Hypertens 2015;9(4):275–284. Ó 2015 American Society of Hypertension. All rights reserved. Keywords: Hemodialysis; hypertension; left ventricular hypertrophy; mean arterial pressure.
This study was supported by an operating grant from the Canadian Institutes of Health Research (MOP 89982). Conflict of interest: none. *Corresponding author: Dr. Ron Wald, MD, MPH, St. Michael’s Hospital, Division of Nephrology, 30 Bond Street, Toronto, Ontario, Canada M5B 1W8. E-mail: [email protected]
**Corresponding author: Dr Andrew T. Yan, MD, Division of Cardiology, St Michael’s Hospital, 30 Bond St, Room 6-030 Donnelly, Toronto, Ontario, Canada M5B 1W8. Tel: 416-8645465; Fax 416-864-5159. E-mail: [email protected]
1933-1711/$ - see front matter Ó 2015 American Society of Hypertension. All rights reserved. http://dx.doi.org/10.1016/j.jash.2015.01.011
276
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Introduction Cardiovascular disease is the leading cause of mortality and hospitalization in patients undergoing chronic hemodialysis (HD).1–5 Cardiovascular risk factors continue to be relatively under–treated,6 and hypertension is common in the end–stage renal disease (ESRD) population with an estimated prevalence of 50%–80%.6 In HD patients, blood pressure (BP) is partly related to extracellular volume status, which is largely dependent on the timing, frequency, and adequacy of dialysis.7 Since 2009, the Canadian Hypertension Education Program (CHEP) has recommended targeting a BP of less than 140/908; however, there is no consensus on when to measure BP—before, during, or after dialysis—and more importantly, which reading best correlates with cardiovascular outcomes. Current practice guidelines recommend using pre–dialysis systolic BP (SBP) to guide treatment of hypertension based on Grade C level evidence.9 Several studies have shown a close relation between BP and left ventricular mass (LVM)10,11 and left ventricular hypertrophy (LVH), which is strongly associated with cardiovascular events in patients with and without kidney disease.12,13 Furthermore, change in LVM has been used as a primary surrogate endpoint for therapeutic interventions in landmark trials of renal replacement therapies.14,15 Conversely, LVM regression is associated with favorable outcomes, including decreased likelihood of developing heart failure and lower all–cause mortality.16,17 Currently, cardiac magnetic resonance imaging (CMR) provides the most accurate and reproducible measurement of LVM.18,19 Although BP is an important and modifiable risk factor for cardiovascular disease in dialysis recipients, it is not clear which BP measurements should be used for therapeutic guidance. Given the clinical significance of LVH as a marker of end–organ damage and predictor of adverse cardiovascular events, the objective of this study was to evaluate the relationship between different BP measurements (pre–dialysis, initial dialysis, nadir dialysis, post–dialysis, inter–dialytic ambulatory) and LVM, as measured by CMR, in recipients of conventional HD.
Methods Study Design This was a cross–sectional study based on a single tertiary–center cohort of 60 prevalent in–center HD patients at St Michael’s Hospital in Toronto, Canada. All patients were receiving conventional HD (4 hours per session, 3–4 times weekly) at the time of assessment. Participants were recruited for an observational study comparing the cardiovascular impact of conversion to in-center nocturnal HD versus continuation of conventional dialysis. The data included herein reflect the baseline data for all study
participants who were receiving conventional HD. Adults 18 years or older with ESRD who were receiving conventional HD for at least 3 months were eligible for this study. Exclusion criteria for the study were any serious co– morbidity with a life expectancy of less than 1 year, a planned live donor kidney transplant within the next 12 months, contraindications to CMR, pregnancy, or inability to provide informed consent. BP data were collected during all dialysis sessions during a 12–week period.
Patient Demographic Data A chart review was performed for each patient via the St Michael’s Hospital electronic patient record system. We collected demographic and clinical data, which included age, gender, cause of ESRD, dialysis vintage, type of vascular access, history of coronary artery disease (defined as previous myocardial infarction or revascularization), cerebrovascular disease (history of documented stroke), diabetes, and peripheral vascular disease (amputation or peripheral revascularization). Relevant medications were also recorded including beta blockers, renin–angiotensin– aldosterone system (RAAS) blocking agents, cholesterol lowering agents, antiplatelet agents, anticoagulants, erythropoietin stimulating agents, and phosphate binders.
In–center Blood Pressure and Weight Measurements BP readings were collected retrospectively from dialysis treatment records over a 12–week period prior to CMR. These were routine blood pressure measurements taken by the dialysis nurses as part of standard dialysis care, using automated BP monitors built into the dialysis machines (Phoenix, Gambro, Richmond Hill, ON). A pre–dialysis BP was measured upon patient arrival in the dialysis unit but prior to the start of dialysis. Upon initiation of dialysis, the first recorded BP reading taken was termed the ‘‘initial dialysis BP’’; this was taken within the first 15 minutes after the HD circuit was initiated; subsequent BP measurements were taken every hour and at the nurse’s discretion. Post–dialysis BP was measured after cessation of dialysis within 15 minutes. The lowest BP reading measured at any point during the dialysis session was considered the ‘‘nadir BP.’’ Corresponding pulse rates were recorded with each BP reading. The dialysis records also contained information on pre–dialysis and post–dialysis weight, as well as volume of ultra–filtration achieved in each session. Inter– and intra–dialytic weight changes were calculated from these data.
Ambulatory Blood Pressure Monitoring Ambulatory blood pressure monitoring (ABPM) data were acquired over a 44–hour period. The duration of
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
monitoring reflected the time interval between the two consecutive HD sessions (first or midweek). BPs were recorded every 30 minutes using an ABPM device (TM– 2430; A&D Company Limited) in the non–access arm. A nurse in the BP lab programmed the hours of sleep, and recorded the mean total, awake, and nocturnal BP data accordingly. Patients who had a limited number of BP readings were not excluded from the analysis, since even these have been demonstrated to have prognostic significance.20
CMR All patients underwent CMR with a 1.5 Tesla whole– body magnetic resonance imaging scanner (Intera; Philips Medical Systems, Best, The Netherlands) using a phased– array cardiac coil and retrospective vectorocardiographic gating. Whenever possible, CMR was performed post–dialysis. A standardized CMR protocol was used as previously published.19 Images were acquired during breath–holds in end–expiration with the patient in supine position; 8–12 contiguous short–axis cine images were required to cover the entire left ventricle. Segmented, balanced steady–state free–precession imaging sequence was used, and typical imaging parameters were repetition time (TR), 4 ms; echo time (TE), 2 ms; slice thickness, 8 mm; field of view, 30–34 cm 30–34 cm; matrix size, 256 196, temporal resolution of <40 ms (depending on heart rate); flip angle, 50 . A blinded experienced cardiac imager (AY) reviewed all the CMR studies and performed the image post–processing offline using commercial software (ViewForum R 4.2; Philips Medical Systems). Manual tracing of endocardial borders at end–diastole and at end–systole using short–axis cine images was performed. Endocardial and epicardial borders in contiguous short–axis slices at end–diastole were also traced, and the difference in area was multiplied by the slice thickness. The sum of these differences throughout the entire left ventricle was multiplied by the myocardial specific density (1.05 g/cm3) to calculate the LVM. Papillary muscles were excluded from the LVM measurement. LVM values were then divided by body surface area (BSA; height in cm weight in kg/3600) to yield LVM index (LVMI). Left ventricular hypertrophy (LVH) was defined as LVMI >81 g/m2 for men and LVMI >79 g/m2 for women.19,21 To account for the fluctuation in intravascular volume status in HD patients, we also examined LVM/LV end–diastolic volume (LVEDV), which is an established prognostic marker for cardiovascular events.22
Laboratory Measurements Results of routine laboratory tests during the 12–week period were extracted retrospectively from electronic medical record and averaged. These included plasma
277
concentrations of calcium (corrected for albumin), phosphate, parathyroid hormone, alkaline phosphatase, ferritin, and hemoglobin.
Statistical Analysis Data are presented as means standard deviation (SD) or median with inter–quartile ranges for continuous variables, and frequency (percent) for categorical data. BP variability was measured by calculating the standard deviation of the all SBPs, DBPs, and the mean arterial pressures (MAP) over the 12–week period. MAP were calculated as (SBP þ 2 DBP)/3. Pulse pressure was calculated as (SBPDBP). Fisher’s exact test was used to assess the relation between patient demographics and LVH. Student t–test was used to compare the difference in SBP, DBP, and MAP in the groups with and without LVH. Pearson product moment correlation coefficients are reported for each method of blood pressure measurement against LVMI as well as LVM/LVEDV. Pearson correlation coefficient of initial dialytic MAP was compared in pairwise fashion with pre–dialysis MAP, SBP, and DBP (Hochberg’s step– up procedure for multiple pairwise comparisons). LVH was analyzed as a dichotomous variable based on the normal gender–specific reference values for CMR–derived LVM,19 and receiver operating characteristic (ROC) curves were generated and thus areas under the curves (AUC) were calculated for the different BP measurements. Multivariable linear regression analyses were performed with LVMI as the dependent variable and each BP parameter as the independent variable, adjusting for age, gender, dialysis vintage, and number of anti–hypertensive medications. We examined the model R2 and verified model assumptions. Analyses were performed using SPSS version 22 (IBM).
Results Baseline Characteristics Table 1 shows the baseline characteristics of 60 patients with complete CMR data. The median age of the patients was 54 years (interquartile range, 44–60 years), and 44% were women. Diabetic nephropathy was the most common cause of ESRD, followed by glomerulonephritis. Vascular co–morbidities were common in the study population, including coronary artery disease and diabetes mellitus. The majority of patients were using anti–platelet agents and antihypertensive agents, primarily RAAS blockers. The mean LVMI was 68 g/m2 (SD ¼ 15 g/m2). Overall, 13 of 60 patients (22%) had LVH. Baseline characteristics of the patients were similar in patients with and without LVH (Table 1). There were no significant differences in causes of ESRD, co–morbidities, medications used, or levels of calcium, phosphate, or hemoglobin. Patients
278
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Table 1 Baseline patient characteristics Characteristics Age, y* Female Comorbid conditions Coronary artery disease Cerebrovascular disease Diabetes mellitus Peripheral vascular disease Ethnicity Caucasian Black Other AV fistula Duration of dialysis, mo* Cause of ESRD Diabetes mellitus Glomerulonephritis Polycystic kidney disease Hypertension Focal sclerosing glomerulosclerosis Other Medication use Number of antihypertensives* Antiplatelet agent Warfarin Beta–blocker ACE inhibitor or angiotensin II receptor blocker Lipid lowering agent ESA Calcium–based phosphate binder Non–calcium based phosphate binder Serum calcium corrected, mmol/L* Serum phosphate, mmol/L* Serum iPTH, pmol/L* Hemoglobin concentration, g/L*
No LVH (n ¼ 47) 53 (44–61) 22 (47) 14 3 24 6 14 12 21 22 40.3 14 14 1 2 3 11 2 27 7 23 36 25 44 36 18 2.26 1.77 40.3 112
(30) (6) (51) (13) (30) (26) (45) (48) (11.0–52.1) (30) (30) (2) (4) (7) (24) (1-2) (59) (15) (49) (76) (53) (94) (78) (39) (2.15–2.4) (1.46–2.05) (20.5–63.4) (106–120)
LVH (n ¼ 13) 58 (47–60) 4 (31) 2 1 6 1 5 4 4 7 49.2 5 2 1 3 0 4 3 8 0 9 13 4 15 14 7 2.36 1.5 49.5 110
(15) (7) (46) (8) (39) (31) (31) (47) (20.0–62.5) (38) (13) (7) (20) (0) (27) (2-3) (53) (0) (60) (87) (27) (100) (93) (47) (2.21–2.49) (1.24–1.66) (37.2–98.5) (107–115)
P–value .55 .36 .31 1.00 .56 .67
.66 1.00 .82
.26
.03 .77 .18 .77 .71 .07 1.00 .26 .76 .59 .10 .34 .73
ACE, Angiotensin–converting enzyme; AV, arteriovenous; ESA, erythropoeitin stimulating agent; ESRD, end–stage renal disease; iPTH, intact parathyroid hormone level; LVH, left ventricular hypertrophy. All data are presented as n (%) unless otherwise specified. * Median (interquartile range).
with LVH were taking more anti–hypertensive medications than patients without LVH (Table 1).
dialysis MAP, post–dialysis MAP, and mean ambulatory MAP (Table 3). In contrast, pulse pressure measurements did not differ significantly between the two groups (Table 3).
Blood Pressure Measurements The average BPs are shown in Table 2. ABPM studies were completed for 45 of 60 patients. SBP readings were significantly higher in patients with LVH as compared with patients without LVH (Table 2). There was no statistically significant difference in DBPs between patients with and without LVH except initial dialysis DBP (P ¼ .047). There were no significant differences in the two groups with respect to the amount of volume ultra–filtered during dialysis. Compared with the group without LVH, patients with LVH had higher pre–dialysis MAP, initial dialysis MAP, nadir
Relationship between BP and LVMI Table 4 demonstrates the associations of the various MAPs, SBPs, and DBPs with LVMI. Initial–dialysis MAP best correlated with LVMI with a Pearson correlation of 0.59 (P < .0001), followed by initial dialysis SBP (r ¼ 0.55; P < .0001; Figure 1). Compared with the corresponding pre–dialysis BP measurements (pairwise comparisons), initial dialysis MAP (P ¼ .009), SBP (P ¼ .01), and DBP (P ¼ .025) all demonstrated significantly stronger correlations with LVMI.
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
279
Table 2 Left ventricular mass by magnetic resonance imaging and blood pressure data for patients with and without left ventricular hypertrophy (LVH) Characteristics
No LVH (n ¼ 47)
LVH (n ¼ 13)
P–value
LV mass, g LV mass index, g/m2 LVEF, % Intra–dialytic weight loss, kg Blood pressure measurements, mm Hg Pre–dialysis SBP Pre–dialysis DBP Initial dialysis SBP Initial dialysis DBP Nadir dialysis SBP Nadir dialysis DBP Post–dialysis SBP Post–dialysis DBP Ambulatory SBP* Ambulatory DBP*
116 (93.5–133.2) 63 (56.5–68) 61 (57–64) 2.68 0.89
165 (153–190) 91 (88–93) 63 (57–65) 2.58 0.83
.38 .71
140 78 136 76 112 67 127 73 132 77
15 8 15 9 13 9 12 8 17 9
152 83 151 83 126 72 140 77 142 83
8 10 10 10 20 13 8 13 13 14
.013 .058 .0076 .047 .026 .14 .018 .20 .011 .058
DBP, Diastolic blood pressure; LV, left ventricular; LVEF, left ventricular ejection fraction; SBP, systolic blood pressure. Data presented as mean standard deviation or median (interquartile range). * Total sample size for ambulatory blood pressure measurement was 45 patients. Thirty–six patients did not have LVH, and nine patients had LVH.
All SBPs had better correlation with LVMI than DBPs during dialysis. Mean ambulatory DBP were better correlated with LVMI than mean ambulatory SBP (Table 4), but both showed only weak correlations with LVMI. There was no correlation between LVMI and intra–dialytic BP lability for SBP (P ¼ .52), DBP (P ¼ .90), or MAP (P ¼ .76; Table 4). Intra–dialytic BP variability also did not significantly predict LVH (Table 5). Table 3 Average mean arterial pressures (MAP) and pulse pressure for patients with and without left ventricular hypertrophy (LVH) by magnetic resonance imaging Blood Pressure Measurement MAP, mm Hg Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory* Pulse pressure, mm Hg Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory*
No LVH (n ¼ 47)
LVH (n ¼ 13)
P–value
99 96 82 91 95
9 9 10 8 11
106 105 90 98 104
8 9 14 10 13
.008 .0007 .018 .004 .041
63 60 46 54 55
13 13 9 11 11
69 68 55 63 59
12 12 17 11 9
.38 .13 .09 .19 .33
Data presented as mean standard deviation. * Total sample size for ambulatory blood pressure measurement was 45 patients. Thirty–six patients did not have LVH, and nine patients had LVH.
In an ancillary analysis, BP was correlated with LVM adjusted for left ventricular end diastolic volume (LVM/ LVEDV). Only initial dialysis MAP and initial dialysis SBP significantly correlated with LVM/LVEDV (r ¼ 0.26, P ¼ .045 and r ¼ 0.261, P ¼ .044 respectively). The areas under the ROC curves are summarized in Table 5. Post–dialysis and initial dialysis SBP performed best in predicting LVH (AUCs 0.83 [95% CI, 0.72–0.94] and 0.81 [95% CI, 0.70–0.92]), respectively. Initial dialysis MAP also performed well in its ability to predict LVH with an AUC of 0.78 (95% CI, 0.64–0.92). Apart from the initial dialysis DBP, none of the other DBP measures appeared to predict LVH. None of the ambulatory BP metrics was a significant predictor of LVH. For the in–center BP readings, compared with MAP, SBP was a better predictor of LVH. There remained independent associations between these BP measurements and LVMI after adjusting for age, gender, duration of dialysis and number of anti–hypertensive medications.
Discussion In this cross–sectional study of 60 patients receiving conventional HD, initial dialysis SBP and MAP and post–dialysis SBP and were most closely correlated with LVMI and the presence of LVH, as assessed by CMR. Our findings suggest these BP measurements may identify patients at risk of adverse cardiovascular outcomes and may represent useful therapeutic targets. Our study was prompted by a lack of consensus regarding which BP measurement best predicts clinical
280
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Table 4 Pearson’s product correlation coefficients for in–center dialysis (n ¼ 60) and ambulatory (n ¼ 45) blood pressure (BP) measurements, and BP lability with left ventricular mass index Blood Pressure Measurement
Mean Arterial Pressure Pearson’s Correlation Coefficient (95% CI)
P–value
Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory blood pressure* Lability
0.51 0.59 0.45 0.47 0.37
<.0001 <.0001 .0003 .0002 .013
(0.29–0.67) (0.40–0.74) (0.23–0.64) (0.24–0.65) (0.08–0.60)
0.04 (0.22–0.29)
Systolic BP
.76
Pearson’s Correlation Coefficient (95% CI) 0.47 0.55 0.47 0.53 0.33
(0.25–0.65) (0.34–0.70) (0.25–0.65) (0.32–0.69) (0.04–0.57)
0.085 (–0.17–0.33)
Diastolic BP P–value <.0001 <.0001 .001 <.001 .029 .52
Pearson’s Correlation Coefficient (95% CI) 0.41 0.50 0.37 0.32 0.36
(0.18–0.60) (0.28–0.67) (0.13–0.57) (0.07–0.53) (0.07–0.59)
0.017 (0.24–0.27)
P–value .001 .0001 .004 .013 .015 .90
CI, Confidence interval. * Total sample size for ambulatory blood pressure measurement was 45 patients. Thirty–six patients did not have left ventricular hypertrophy, and nine patients had left ventricular hypertrophy.
outcomes in dialysis recipients. This information is vital as it would serve to guide therapeutic targets in this population. Current guidelines recommend targeting treatment to pre–dialysis SBP, although this recommendation is based on rather limited data (grade C level evidence).9 Several studies have sought to determine the optimal BP measurement strategy, with conflicting results. Conlon et al concluded that pre–dialysis SBP best correlated with LVM (r ¼ 0.35, P ¼ .03) and should be the target of hypertension control.23 Their study was conducted over 12 dialysis sessions among a predominantly African American population on stable antihypertensive regimens (unchanged in the 3 months of the study),23 which may have limited the generalizability of their results. The Cardiovascular Risk Extended Evaluation investigators also found a significant association between LVMI and pre–dialysis SBP.24 Additionally, pre–dialysis DBP and pulse pressure were associated with LVMI.24 This study excluded diabetic patients and those with heart failure. Furthermore, the main anti– hypertensives used by patients in this population were beta–blockers, which does not reflect current practice.9 Foley et al examined the United States Renal Data System (USRDS) database of 11,142 subjects on HD, and found that higher post–dialysis SBP, but not pre–dialysis SBP, was associated with all–cause mortality. In this large study, wide post–dialysis pulse pressure was also associated with higher mortality. However, they only studied pre–dialysis and post–dialysis BP and used an average of only the last three BP values recorded prior to the study. They did not investigate if there was an association with cardiovascular mortality or events, and did not focus on patho–physiologic mechanisms such as end–organ damage.9,25 Cao et al found that both pre–dialysis SBP and DBP were significantly associated with LVH in 164 Chinese patients.26 In contrast, Koc et al found that in a sample population of 74 patients, both post–dialysis SBP (r ¼ 0.471, P < .001) and pre–dialysis SBP (r ¼ 0.461, P < .001) correlated with increased
LVMI as determined by echocardiography.27 They did not evaluate other in–center dialysis BP measurements and did not test which measurement best predicted the presence of LVH. In a study of 150 predominantly African American patients from four dialysis units in Cleveland, Agarwal et al showed a stronger association between LVH and SBP measured at home (using self–monitoring and ABPM) than with pre– or post–dialysis SBP.28 All peri–dialytic BPs had similar correlations with LVMI. The same group also examined pre–dialysis, post–dialysis, and intra– dialysis BP monitoring using 44–hour ABPM as a standard to diagnose hypertension. They showed that both mean intra–dialytic SBP and median intra–dialytic SBP had comparable diagnostic performance metrics.29 In this study, they did not correlate their findings with any marker of end–organ damage. More recently in 2010, Agarwal et al reported that out–of–dialysis unit BP had greater prognostic significance compared to pre– and post–dialysis BP, and higher ambulatory BP was associated with increased mortality.30 Although these studies provide useful mechanistic insights, they utilized echocardiography to measure LVM. Echocardiographic assessment of LVM relies on geometric assumptions that may not be valid in chronic dialysis patients due to cardiac remodeling from pressure and volume overload. Another limitation is the use of a formula in echocardiography that relies on the internal diameter of the left ventricular chamber; intra-cardiac volume may fluctuate greatly between pre–dialytic and post–dialytic periods, and may therefore cause considerable variation in the LVM calculation.31 Two studies have employed CMR to study the relation of BP measurements to LVMI, but each studied a limited set of BP parameters, and neither of them studied ABPM. Patel et al were the first investigators to use CMR to study the relation between BP and LVH in a cohort of 246 hemodialysis patients.32 MRI was performed 24 hours after the last dialysis session. They found pre–dialysis SBP to be a
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Figure 1. Scatterplots (n¼60) demonstrating the relationships of left ventricular mass index and (A) initial dialysis mean arterial pressure (MAP); (B) initial dialysis systolic blood pressure (SBP); and (C) post–dialysis SBP.
281
robust independent predictor of LVH. However, they did not compare initial dialysis, or nadir dialysis BP. In contrast, Khangura et al performed a study using CMR in a smaller sample size (n ¼ 39), and only studied pre–, post–, and intra–dialytic (average of all BP readings during a single dialysis session only) SBPs as well as a standardized pre–dialysis SBP.33 Their results supported a single baseline intra–dialytic SBP (AUC, 0.80; 95% CI, 0.64– 0.96) or post–dialysis SBP (AUC, 0.79; 95% CI, 0.64– 0.93) as being the best predictors of LVH. Pre–dialysis SBP poorly correlated with LVMI (r ¼ 0.3; P ¼ .068).33 However, they did not study MAP or DBP. A major strength of the current study was the use of CMR instead of echocardiography to assess LVM. CMR is now recognized as the gold standard for LVM measurement. It is more accurate and precise than echocardiography.18,19 In a comparison between CMR and echocardiography, echocardiography overestimated LV mass in the presence of LVH and dilatation, especially in the HD population.18 A recent study performed within the current cohort also showed a significant degree of overestimation of LVM and LVH prevalence by echocardiography in patients with ESRD.19 Another strength and novel aspect of our study was the comprehensive assessment of different in–center BP measurements (pre–dialysis, intra–dialytic, nadir dialysis, post–dialysis) and ABPM, and their association with LVH using CMR. We were able to collect a large amount of data over 3 months, with multiple BP measurements; this minimized effects of outliers and spurious results. Our study also reflected actual dialysis unit nursing practices and limitations, and therefore may be more representative of the information available to clinicians. We found that mean 44–hour ambulatory MAPs did positively correlate with LVMI, but the association was not as strong compared with other methods (Table 4). Interestingly, we found that ambulatory SBP, DBP, or MAP did not significantly predict LVH (Table 5). This result is quite different from what was reported by Agarwal et al.28 Our results suggest that BP monitoring in the dialysis unit over a longer period of time may be even more useful than ambulatory BP monitoring, which provides much fewer readings over a relatively shorter time period. Average BPs calculated from many readings may better predict end–organ damage, especially in dialysis patients, in whom wide fluctuations in BP are common. Of note, BP variability was not predictive of LVH in this study, which to the best of our knowledge, is the first to address this question. BP variability may result from changes in filling pressures in the setting of LVH and diastolic dysfunction, or heart failure with preserved ejection fraction. This is also a common cause of intra–dialytic hypotension, which is associated with increased cardiovascular mortality.34 LV systolic function was preserved in our study cohort, but diastolic dysfunction was not specifically
282
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
Table 5 Receiver operating characteristics (ROC) statistics for left ventricular hypertrophy (LVH) for in–center dialysis (n ¼ 60) and ambulatory (n ¼ 45) blood pressure (BP) measurements Blood Pressure Measurement
Pre–dialysis Initial dialysis Nadir dialysis Post–dialysis Mean ambulatory blood pressure* Lability
Mean Arterial Pressure
Systolic BP
Diastolic BP
ROC AUC (95% CI)
ROC AUC (95% CI)
ROC AUC (95% CI)
0.74 0.78 0.67 0.70 0.68 0.49
0.76 0.81 0.73 0.83 0.67 0.51
0.63 0.69 0.57 0.58 0.62 0.50
(0.59, (0.64, (0.48, (0.53, (0.47, (0.28,
0.88) 0.92) 0.85) 0.87) 0.88) 0.70)
(0.64, (0.70, (0.55, (0.72, (0.49, (0.30,
0.88) 0.92) 0.90) 0.94) 0.85) 0.72)
(0.44, (0.51, (0.37, (0.38, (0.38, (0.31,
0.82) 0.88) 0.78) 0.79) 0.86) 0.69)
AUC, Area under the curve; CI, confidence interval. * Total sample size for ambulatory blood pressure measurement was 45 patients, of whom nine had LVH.
assessed. Similar to BP variability, nadir BP was not significantly associated with LVH or LVMI. Most studies have investigated SBP and DBP, but there are a paucity of data on the association between LVH and MAPs or pulse pressures in the dialysis population. Increase in MAP has been shown to be associated with progressive increase in LVH and development of heart failure and ischemic heart disease,35 as well as higher cardiovascular mortality.20,36 We demonstrated that MAP correlated well with LVMI, but it did not predict LVH as well as SBP. In our study, pulse pressures did not correlate significantly with LVMI. Pulse pressure has been shown to be an independent predictor of coronary heart disease and all–cause mortality.25,37 This may be a phenomenon only applicable to central pulse pressure, as one study found carotid pulse pressure to be associated with all–cause and cardiovascular mortality, but not brachial pulse pressure.37 Furthermore, the association of pulse pressure with cardiovascular outcomes and mortality has been found to be independent of LVH and systolic dysfunction.38 Frequent changes in intravascular volume status may pose challenges to LVM measurements when geometric assumptions are made, for example, in echocardiography. Because CMR applies a truly volumetric approach to LVM measurements (ie, it directly measures the volume of the myocardium, rather than using a formula related to the LV cavity volume), it is relatively insensitive to changes in volume status. Moreover, we performed an additional analysis by normalizing LVM to LVEDV (LVM/LVEDV), thereby allowing for volume changes with HD. The left ventricular mass to volume ratio has been previously shown in the large Multi–Ethnic Study of Atherosclerosis to independently predict incident coronary heart disease and stroke.22 Thus, the significant correlations between initial dialysis MAP and SBP and LVM/LVEDV lend further credence to the prognostic significance of these BP measurements. There were several limitations in our study. This was a single center study with a limited number of patients.
Second, this was a cross–sectional observational study, and causality cannot be established between elevated BP and LVH. Therefore, caution must be exercised when interpreting these results. We used LVMI and the presence of LVH as our primary outcome of interest, and our study was not powered to examine cardiovascular outcomes. Our study population had a lower prevalence of LVH than most other studies; although this might be partly due to an over–estimation of LVH by echocardiography or different definition of LVH, this might also reflect the particular case mix at our single dialysis center that is different from other study populations. Another limitation to our study was that 44–hour ABPM was not available for all patients. This may have underpowered our study and led to an apparently weaker correlation between ambulatory blood pressure and left ventricular mass. It should also be noted that all other BP data were extracted retrospectively from routine dialysis records. These BP measurements were not done according to a standardized research protocol. We were therefore not able to compare HD unit BP data with inter–dialytic standardized BPs.
Conclusions In conclusion, initial dialysis MAP, SBP, and post–dialysis SBP both strongly correlate with LVMI and LVH, as measured by CMR. These findings suggest that therapies targeting these BP measurements may attenuate end–organ damage, thereby improving cardiovascular outcomes. These results need to be confirmed in larger multi–center clinical trials randomizing a broader spectrum of HD patients to various BP targets with long–term follow up. Acknowledgment The authors would like to thank all the patients and staff at the St Michael’s Hospital hemodialysis unit who contributed the data for this study.
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
References 1. Foley RN, Parfrey PS. Cardiovascular disease and mortality in ESRD. J Nephrol 1998;11:239–45. 2. Mittal SK, Kowalski E, Trenkle J, McDonough B, Halinski D, Devlin K, et al. Prevalence of hypertension in a hemodialysis population. Clin Nephrol 1999;51: 77–82. 3. Rostand SG, Brunzell JD, Cannon RO III, Victor RG. Cardiovascular complications in renal failure. J Am Soc Nephrol 1991;2:1053–62. 4. Salem MM. Hypertension in the hemodialysis population: a survey of 649 patients. Am J Kidney Dis 1995;26:461–8. 5. Tomita J, Kimura G, Inoue T, Inenaga T, Sanai T, Kawano Y, et al. Role of systolic blood pressure in determining prognosis of hemodialyzed patients. Am J Kidney Dis 1995;25:405–12. 6. Agarwal R, Nissenson AR, Batlle D, Coyne DW, Trout JR, Warnock DG. Prevalence, treatment, and control of hypertension in chronic hemodialysis patients in the United States. Am J Med 2003;115:291–7. 7. Horl MP, Horl WH. Hemodialysis-associated hypertension: pathophysiology and therapy. Am J Kidney Dis 2002;39:227–44. 8. Prasad GV, Ruzicka M, Burns KD, Tobe SW, Lebel M. Hypertension in dialysis and kidney transplant patients. Can J Cardiol 2009;25:309–14. 9. Jindal K, Chan CT, Deziel C, Hirsch D, Soroka SD, Tonelli M, et al. Hemodialysis clinical practice guidelines for the Canadian Society of Nephrology. J Am Soc Nephrol 2006;17(3 Suppl 1):S1–27. 10. Harnett JD, Kent GM, Barre PE, Taylor R, Parfrey PS. Risk factors for the development of left ventricular hypertrophy in a prospectively followed cohort of dialysis patients. J Am Soc Nephrol 1994;4:1486–90. 11. Devereux RB, de SG, Koren MJ, Roman MJ, Laragh JH. Left ventricular mass as a predictor of development of hypertension. Am J Hypertens 1991; 4:603S–7S. 12. Silberberg JS, Barre PE, Prichard SS, Sniderman AD. Impact of left ventricular hypertrophy on survival in end-stage renal disease. Kidney Int 1989;36:286–90. 13. Bouzas-Mosquera A, Broullon FJ, Alvarez-Garcia N, Peteiro J, Mosquera VX, Castro-Beiras A. Association of left ventricular mass with all-cause mortality, myocardial infarction and stroke. PLoS One 2012;7: e45570. 14. Kannel WB, Gordon T, Offutt D. Left ventricular hypertrophy by electrocardiogram. Prevalence, incidence, and mortality in the Framingham study. Ann Intern Med 1969;71:89–105. 15. Casale PN, Devereux RB, Milner M, Zullo G, Harshfield GA, Pickering TG, et al. Value of
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
283
echocardiographic measurement of left ventricular mass in predicting cardiovascular morbid events in hypertensive men. Ann Intern Med 1986;105:173–8. London GM, Pannier B, Guerin AP, Blacher J, Marchais SJ, Darne B, et al. Alterations of left ventricular hypertrophy in and survival of patients receiving hemodialysis: follow-up of an interventional study. J Am Soc Nephrol 2001;12:2759–67. Foley RN, Parfrey PS, Kent GM, Harnett JD, Murray DC, Barre PE. Serial change in echocardiographic parameters and cardiac failure in end-stage renal disease. J Am Soc Nephrol 2000;11:912–6. Stewart GA, Foster J, Cowan M, Rooney E, McDonagh T, Dargie HJ, et al. Echocardiography overestimates left ventricular mass in hemodialysis patients relative to magnetic resonance imaging. Kidney Int 1999;56:2248–53. Jakubovic BD, Wald R, Goldstein MB, Leong-Poi H, Yuen DA, Perl J, et al. Comparative assessment of 2dimensional echocardiography vs cardiac magnetic resonance imaging in measuring left ventricular mass in patients with and without end-stage renal disease. Can J Cardiol 2013;29:384–90. Gerin W, Schwartz JE, Devereux RB, Goyal T, Shimbo D, Ogedegbe G, et al. Superiority of ambulatory to physician blood pressure is not an artifact of differential measurement reliability. Blood Press Monit 2006;11:297–301. Hudsmith LE, Petersen SE, Francis JM, Robson MD, Neubauer S. Normal human left and right ventricular and left atrial dimensions using steady state free precession magnetic resonance imaging. J Cardiovasc Magn Reson 2005;7:775–82. Bluemke DA, Kronmal RA, Lima JA, Liu K, Olson J, Burke GL, et al. The relationship of left ventricular mass and geometry to incident cardiovascular events: the MESA (Multi-Ethnic Study of Atherosclerosis) study. J Am Coll Cardiol 2008;52:2148–55. Conlon PJ, Walshe JJ, Heinle SK, Minda S, Krucoff M, Schwab SJ. Predialysis systolic blood pressure correlates strongly with mean 24-hour systolic blood pressure and left ventricular mass in stable hemodialysis patients. J Am Soc Nephrol 1996;7:2658–63. Zoccali C, Mallamaci F, Tripepi G, Benedetto FA, Cottini E, Giacone G, et al. Prediction of left ventricular geometry by clinic, pre-dialysis and 24-h ambulatory BP monitoring in hemodialysis patients: CREED investigators. J Hypertens 1999;17(12 Pt 1):1751–8. Foley RN, Herzog CA, Collins AJ. Blood pressure and long-term mortality in United States hemodialysis patients: USRDS Waves 3 and 4 Study. Kidney Int 2002;62:1784–90. Cao X, Zou J, Teng J, Zhong Y, Ji J, Chen Z, et al. BMI, spKt/V, and SBP but not DBP are related to
284
27.
28.
29.
30. 31.
32.
A. Merchant et al. / Journal of the American Society of Hypertension 9(4) (2015) 275–284
LVH in Chinese maintenance hemodialysis patients. Ren Fail 2011;33:269–75. Koc Y, Unsal A, Kayabasi H, Oztekin E, Sakaci T, Ahbap E, et al. Impact of volume status on blood pressure and left ventricle structure in patients undergoing chronic hemodialysis. Ren Fail 2011;33:377–81. Agarwal R, Brim NJ, Mahenthiran J, Andersen MJ, Saha C. Out-of-hemodialysis-unit blood pressure is a superior determinant of left ventricular hypertrophy. Hypertension 2006;47:62–8. Agarwal R, Metiku T, Tegegne GG, Light RP, Bunaye Z, Bekele DM, et al. Diagnosing hypertension by intradialytic blood pressure recordings. Clin J Am Soc Nephrol 2008;3:1364–72. Agarwal R. Blood pressure and mortality among hemodialysis patients. Hypertension 2010;55:762–8. Kilickap M, Turhan S, Sayin T, Nergizoglu G, Kutlay S, Duman N, et al. Intravascular volume dependency of left ventricular mass calculation by twodimensional guided M-mode echocardiography. Can J Cardiol 2007;23:219–22. Patel RK, Oliver S, Mark PB, Powell JR, McQuarrie EP, Traynor JP, et al. Determinants of left ventricular mass and hypertrophy in hemodialysis patients assessed by cardiac magnetic resonance imaging. Clin J Am Soc Nephrol 2009;4:1477–83.
33. Khangura J, Culleton BF, Manns BJ, Zhang J, Barnieh L, Walsh M, et al. Association between routine and standardized blood pressure measurements and left ventricular hypertrophy among patients on hemodialysis. BMC Nephrol 2010;11:13. 34. Losi MA, Memoli B, Contaldi C, Barbati G, Del PM, Betocchi S, et al. Myocardial fibrosis and diastolic dysfunction in patients on chronic haemodialysis. Nephrol Dial Transplant 2010;25:1950–4. 35. Foley RN, Parfrey PS, Harnett JD, Kent GM, Murray DC, Barre PE. Impact of hypertension on cardiomyopathy, morbidity and mortality in end-stage renal disease. Kidney Int 1996;49:1379–85. 36. Charra B, Calemard E, Ruffet M, Chazot C, Terrat JC, Vanel T, et al. Survival as an index of adequacy of dialysis. Kidney Int 1992;41:1286–91. 37. Safar ME, Blacher J, Pannier B, Guerin AP, Marchais SJ, Guyonvarc’h PM, et al. Central pulse pressure and mortality in end-stage renal disease. Hypertension 2002;39:735–8. 38. Palmieri V, Devereux RB, Hollywood J, Bella JN, Liu JE, Lee ET, et al. Association of pulse pressure with cardiovascular outcome is independent of left ventricular hypertrophy and systolic dysfunction: the Strong Heart Study. Am J Hypertens 2006;19: 601–7.