Академический Документы
Профессиональный Документы
Культура Документы
doi: 10.1093/ndt/gfz020
Downloaded from https://academic.oup.com/ndt/advance-article-abstract/doi/10.1093/ndt/gfz020/5364308 by Washington University, Law School Library user on 27 February 2019
Extracellular fluid volume expansion, arterial stiffness
ORIGINAL ARTICLE
and uncontrolled hypertension in patients with chronic
kidney disease
Branko Braam1,2, Chung Foon Lai1, Joseph Abinader2 and Aminu K. Bello2
1
Department of Medicine, Division of Nephrology, University of Alberta, Edmonton, Alberta, Canada and 2Department of Physiology,
University of Alberta, Edmonton, Alberta, Canada
ABSTRACT with the stage of CKD [1]. Similarly, ECFV expansion is com-
Background. Hypertension is prevalent in patients with mon in CKD and also increases with declining kidney function
chronic kidney disease (CKD) and is related to extracellular [2]. Not surprisingly, diuretics reduce ECFV and decrease blood
fluid volume (ECFV) expansion. Arterial stiffening is another pressure (BP) [1, 3]. However, a significant number of CKD
implication of CKD that can be caused by ECFV expansion. In patients have uncontrolled BP despite treatment [4, 5]. This
this study, we hypothesized that CKD patients with uncon- raises the question whether CKD patients with uncontrolled BP
trolled hypertension are more likely to be fluid volume ex- are not simply fluid volume expanded. If so, then this would
panded than normotensive patients, which in turn is associated need careful consideration to mitigate the risk of adverse conse-
with increased arterial stiffness. quences (morbidity and mortality) across the spectrum of CKD
Methods. Adult hypertensive patients with mild–severe CKD [6, 7].
(n ¼ 82) were recruited. ECFV was assessed using multifre- Arterial stiffening is another implication of CKD [8, 9],
quency bioimpedance and arterial stiffness by applanation which can be characterized by an increased augmentation index
tonometry and oscillometry. (AIx) or pulse wave velocity (PWV) [10]. Arterial stiffening has
Results. Patients with uncontrolled hypertension had fluid vol- been positively associated with the rate of decline of kidney
ume expansion compared with controls (1.0 6 1.5 versus function in patients with CKD [11, 12]. In addition, ECFV ex-
0.0 6 1.6 L, P < 0.001), and had a higher augmentation index pansion and hypertension can increase arterial wall stress,
(AIx) and pulse wave velocity. Fluid volume expansion was which might lead to structural (arterial wall thickness) and
more prevalent in patients with uncontrolled hypertension functional (distensibility) changes of the arterial wall and result
(58%) than patients who were at target (27%). Fluid volume ex- in arterial stiffening [13]. Taken together, our current study fo-
pansion was correlated with age, AIx and systolic blood pres- cuses on the effects of ECFV expansion in patients with CKD
sure. In a binary logistic regression analysis, AIx, age and fluid on BP control and arterial stiffness.
volume status were independent predictors of uncontrolled hy- We hypothesized that CKD patients with uncontrolled BP
pertension in both univariate and multivariate models. are more likely to have fluid volume expansion, which in turn is
Discussion. In summary, uncontrolled hypertension among then associated with increased arterial stiffness. We therefore
hypertensive CKD patients is associated with ECFV expansion. aimed to determine the association between ECFV expansion,
Our data suggest a relationship between ECFV expansion, in- uncontrolled BP and the relationship with arterial stiffness in a
creased arterial stiffness and uncontrolled hypertension. cohort of CKD patients in Canada.
Downloaded from https://academic.oup.com/ndt/advance-article-abstract/doi/10.1093/ndt/gfz020/5364308 by Washington University, Law School Library user on 27 February 2019
analyses on arterial stiffness assessment, by year of enrolment.
of fluid volume status (VS). Hypertension was defined as BP Differences between groups were tested for continuous varia-
140/90 mmHg (130/80 mmHg for diabetic patients) or the bles by Student’s t-test and for categorical variables by Fisher’s
use of antihypertensive medications as per standard guideline exact test. All values were expressed as mean 6 standard deviation
recommendation [14]. The study was approved by the and P < 0.05 was taken to be significant. Univariate relationships
University of Alberta Human Research Ethics Board and writ- between ECFV expansion and related variables were assessed by
ten consent was obtained from each participant in the study. Pearson’s correlation analyses. In a binary logistic regression
model, we examined the associations of key predictor variables
(demographic, clinical and laboratory) with uncontrolled BP us-
Biochemical data collection ing ECFV expansion as the primary predictor variable. Linear re-
Biochemical parameters, such as serum creatinine and albu- gression analyses were also conducted to investigate, if present,
min, were obtained from patient records at the clinic. eGFRs any statistical differences between the results obtained from the
were calculated using the CKD Epidemiology Collaboration SphygmoCor arterial tonometer (2014 patient cohort) and those
equation [15]. from the Arteriograph (2016 patient cohort).
We also conducted a series of sensitivity analyses. First, we
conducted a series of correlation analyses to examine the effect
Measurement of ECFV of baseline proteinuria and proteinuria changes (baseline to fol-
To determine fluid VS, we used an approach previously used low-up) on salt retention (fluid VS) and vascular stiffness (AIx).
in hemodialysis patients by our group [16]. Using the Body Second, we investigated the impact of BP control, vascular stiff-
Composition Monitor (BCM; Fresenius Medical Care, Bad ness and fluid volume expansion on changes in kidney function
Homburg, Germany), a multifrequency bioimpedance device from baseline to follow-up (DeGFR) in a multivariate liner re-
that has been validated previously [17], fluid volume expansion gression model adjusting in a stepwise fashion for the key pre-
is calculated by comparison of the measured ECFV with the dictor variables, with DeGFR as the dependent variable.
predicted ECFV using a standard physiological model [18]. Student’s t-test and correlations were performed using
This model uses adjustments for subjects with high body mass GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA). All
index [19]. For each patient in the study, measurements were regression analyses were performed using Statistical Package
performed in triplicate on one occasion. Electrodes were ap- for the Social Sciences, version 25.0 (IBM, Armonk, NY, USA).
plied on an ipsilateral arm and foot. The BCM calculates VS,
which is expressed as volume excess or depletion in liters com- RESULTS
pared with the estimated normal ECFV. Normovolemia was de-
fined as any measurement between 1.1 and 1.1 L relative to Patient characteristics
normal ECFV, as 90% of fluid volumes in the healthy popula- Of the total 82 patients (52 male and 30 female) enrolled in
tion are within the range of 1.1 to 1.1 L [20]. Fluid volume the study, 72 patients (88%) were on BP-lowering medications
overload was considered >1.1 L above normal ECFV. and 40 patients (49%) were receiving diuretic treatment; 38
patients (46%) displayed uncontrolled BP. Overall, patients had
mild–severe CKD, with a mean eGFR value of 46 6 26 mL/
Arterial stiffness measurement min/1.73 m2. In total, 28 patients (34%) were diabetic and 35
Arterial stiffness was defined and assessed using standard patients (43%) had signs of edema (Table 1). Regarding fluid
methods [21]. This was evaluated in two ways during the study VS, 34 patients (41%) were found to be fluid volume overloaded
period (for technical reasons): the SphygmoCor (AtCor (1.1 L above normal ECFV). The patients displayed substan-
Medical, Sydney, NSW, Australia) arterial tonometer was used tial variation of fluid VS, ranging from 3.5 to 4.5 L below and
during the first phase of data collection in 2014 and the above the estimated normal ECFV (Figure 1A).
Arteriograph (TensioMed, Budapest, Hungary) was used dur-
ing the second phase in 2016. The SphygmoCor arterial tonom- Differences between patients with controlled versus
eter applies applanation tonometry to measure waveforms at uncontrolled BP
the radial and carotid arteries three times at each site while the The 82 hypertensive CKD patients were categorized into two
patient is in a supine position. Through a computerized process, groups based on whether they had controlled BP or not (uncon-
the aortic pressure waveform was generated [22] and the pulse trolled: systolic BP 140 mmHg or diastolic BP 90 mmHg for
wave analysis of the patient obtained [10]. The Arteriograph is nondiabetic patients and 130 mmHg or 80 mmHg for dia-
an oscillometric device that applies occlusion on the brachial ar- betic patients). The two groups did not differ in BMI, diastolic
tery of the patient to record pulse waves. By calculating the time BP, serum creatinine level, eGFR, presence of diabetes and use
difference between the beginning of the first wave and the sec- of antihypertensive medication. However, patients with uncon-
ond wave (reflective wave), the AIx and PWV of the patient trolled BP had a higher average fluid volume than patients with
were determined [23]. controlled BP (P ¼ 0.0052) (Table 1). Furthermore, 58% of
2 B. Braam et al.
Table 1. Patient characteristics
Downloaded from https://academic.oup.com/ndt/advance-article-abstract/doi/10.1093/ndt/gfz020/5364308 by Washington University, Law School Library user on 27 February 2019
Fluid VS (L) 0.5 6 1.62 0.0 6 1.57 1.0 6 1.52 0.0052
Age (years) 63 6 14.4 57.9 6 15.0 68.8 6 11.3 0.0005
BMI (kg/m2) 29.2 6 7.00 29.0 6 6.62 29.6 6 7.50 NS
Systolic BP (mmHg) 135 6 20.5 121 6 12.4 152 6 13.8 <0.0001
Diastolic BP (mmHg) 74.6 6 12.8 71.7 6 11.7 78.0 6 13.4 NS
PP (mmHg) 60.8 6 18.0 49.5 6 11.2 74.5 6 15.9 <0.0001
MAP (mmHg) 94.9 6 13.2 88.1 6 10.9 103 6 11.3 <0.0001
Serum creatinine (lmol/L) 157 6 76.4 143 6 59.1 177 6 85.8 NS
Baseline eGFR (mL/min/1.73 m2) 46 6 26 53 6 27 39 6 23 NS
Diabetic (yes), n (%) 28 (34.1) 11 (25.0) 17 (44.7) NS
Signs of edema (yes), n (%) 35 (42.7) 14 (31.8) 21 (55.3) 0.0443
Use of antihypertensive medications (yes), n (%) 72 (87.8) 39 (88.6) 33 (86.8) NS
Use of diuretics (yes), n (%) 40 (48.8) 14 (31.8) 26 (68.4) 0.0017
Values are expressed as mean 6 SD unless stated otherwise. P-values: unpaired Student’s t-test between patients with fluid volume overload and normal fluid VS. PP, pulse pressure;
MAP, mean arterial pressure; NS, not significant; M, male; F, female.
4 B. Braam et al.
Several predictors for uncontrolled BP were identified in bi- using different baseline BP measures, which indicated the im-
nary logistic regression analysis. First, fluid VS was strongly as- pact of ECFV expansion on GFR decline. We were also in-
sociated with the outcome variable in both univariate and trigued that a third of the patients with uncontrolled BP were
multivariate models, indicating a higher chance for patients to not receiving diuretics, and this could be partly explained by the
Downloaded from https://academic.oup.com/ndt/advance-article-abstract/doi/10.1093/ndt/gfz020/5364308 by Washington University, Law School Library user on 27 February 2019
have uncontrolled BP if they have increased ECFVs. This was inability of practitioners to detect volume overload using clini-
not investigated in previous clinical studies in the CKD popula- cal examination only. This might reflect practice variations
tion. Another identified predictor was age, which showed across settings and within practitioners. This is an important
strong associations in both univariate and multivariate models, practice implication of our findings. This is an additional impe-
indicating that older CKD patients are less likely to be uncon- tus for the use of bio-impedance (even considering its limita-
trolled. With its negative correlation with fluid VS, aging plays tions) for volume assessments, as demonstrated in our previous
a critical role in contributing to ECFV expansion and hyperten- work [16].
sion, as aging is a significant modifier of hypertension in CKD Our study has several limitations. First, this was a snapshot
patients [4, 26]. Interestingly, arterial stiffness, measured by AIx study, as all data were collected only once for this study; longi-
and PWV, was associated with uncontrolled hypertension. In tudinal assessments and interventions were not applied.
binary logistic regression analyses in both the 2014 and 2016 Second, the sample size was small and could possibly not repre-
cohorts, AIx and PWV showed strong positive associations in sent the general CKD population across the full spectrum.
the univariate model, indicating that CKD patients with en- However, the finding of a high prevalence of fluid volume over-
hanced arterial stiffness are less likely to be controlled. load is in keeping with existing literature in the field studies [16,
Interestingly, being diabetic was not a predictor for fluid over- 17, 20]. Third, the assessment of ECFV using multifrequency
load as observed previously by others [27]. bioimpedance spectroscopy comes with some error. Yet the
A major implication of this study is on understanding the method seems comparable in performance to radioisotope
pathophysiology as well as management approach to resistant methods [33]. Lastly, the two cohorts of patients in this study
hypertension [28]. A previous study demonstrated that patients had their arterial stiffness assessed by different methods due to
with resistant hypertension on a low-salt diet showed a reduc- technical issues with the equipment initially used in measure-
tion of mean office systolic and diastolic BP compared with ments. The results obtained by both methods were comparable
patients on a regular, high-salt diet [29]. In a double-blind and did not impact on our conclusions and implications of this
study, arterial stiffness (baseline PWV) was determined as a sig- study, but they still share slight discrepancies that make merg-
nificant independent predictor of achieving a reduction in sys- ing the data not justifiable [34]. In our study, AIx was signifi-
tolic BP after 12 months of treatment [30]. Although patients cantly different between the patients in 2014 and 2016, which
enrolled in our study had CKD and were not categorized on the was likely due to the differences in age and renal function.
basis of resistant hypertension, similar predictors—fluid vol- Recognizing these limitations, our conclusions still hold with
ume and arterial stiffness—were identified in our binary logistic respect to ECFV expansion being a determinant of uncontrolled
regression analyses. Therefore ECFV expansion could well be a BP and a likely determinant of arterial stiffness. Other limita-
factor in resistant hypertension without CKD as well. tions with potential implications for further work in this area
Fluid volume overload was significantly and positively corre- included a lack of data in this analysis on other comorbid con-
lated with systolic BP, age and AIx. Therefore a relationship be- ditions and hospitalization rates, as well as laboratory markers
tween ECFV expansion and arterial stiffness is suggested. such as serum B-type natriuretic peptide and C-reactive
Although the PWV data collected in our study were not ade- protein.
quate for correlation analysis, PWV has been associated with An increase in total body sodium is the basis for ECFV ex-
fluid volume overload among peritoneal dialysis patients in a pansion. Several studies have indicated that high-sodium states
small study (n ¼ 122) [31]. Also, correction of ECFV expansion could lead to increased vascular stiffness [35–37]. In salt-resis-
was shown to improve AIx in hemodialysis patients [32]. tant healthy humans, a high-sodium diet was associated with
Therefore we propose that ECFV expansion is associated with decreased endothelium-dependent vasodilation [35]. There is
arterial stiffness in our cohort of CKD patients. accumulating evidence that the mechanism for impaired vascu-
To enhance the robustness of our findings, we have con- lar function is related to the composition and function of the
ducted a series of sensitivity analyses in order to determine if endothelial glycocalyx [36, 37]. While plasma sodium levels
there are any interesting associations between salt retention were not significantly different between controlled and uncon-
(fluid VS) and vascular stiffness (AIx) with proteinuria in the trolled subjects and the study was not designed to investigate
subjects with and without diabetes, and there were no signifi- endothelial function and the glycocalyx, it is possible that the
cant findings in both study groups. Furthermore, we also exam- differences in vascular stiffness and in outcome in diabetic
ined the impact of BP control (using different measures of BP patients were related to endothelial glycocalyx alterations.
measurements), ECFV expansion and vascular stiffness on re- In summary, our study demonstrated associations of ECFV
nal function decline, and this also showed no significant rela- expansion with uncontrolled BP among hypertensive CKD
tionship between baseline eGFR, change in renal function patients. Also, causes of uncontrolled BP are multifactorial, in-
(DeGFR) and BP control and our key outcomes. However, in- volving patient age, fluid volume overload and likely arterial
triguingly, fluid VS showed a strong association with the out- stiffening. Therefore we suggest that better fluid volume control
come variable among diabetic patients in the regression models would help improve BP control and preserve arterial function,
Downloaded from https://academic.oup.com/ndt/advance-article-abstract/doi/10.1093/ndt/gfz020/5364308 by Washington University, Law School Library user on 27 February 2019
via body composition spectroscopy in health and disease. Physiol Meas
and would lead to insight into whether prevention of fluid over- 2006; 27: 921–933
load will also prevent adverse cardiovascular outcomes. 18. Wang Z, Deurenberg P, Wang W et al. Hydration of fat-free body mass:
new physiological modeling approach. Am J Physiol 1999; 276: E995–E1003
19. Chamney PW, Wabel P, Moissl UM et al. A whole-body model to distin-
SUPPLEMENTARY DATA guish excess fluid from the hydration of major body tissues. Am J Clin Nutr
Supplementary data are available at ndt online. 2007; 85: 80–89
20. Wabel P, Moissl U, Chamney P et al. Towards improved cardiovascular
management: the necessity of combining blood pressure and fluid overload.
FUNDING Nephrol Dial Transplant 2008; 23: 2965–2971
B.B. was supported by the Kidney Health Translational 21. O’Rourke MF, Staessen JA, Vlachopoulos C et al. Clinical applications of ar-
terial stiffness; definitions and reference values. Am J Hypertens 2002; 15:
Research Chair by the Faculty of Medicine and Dentistry,
426–444
Department of Medicine and Division of Nephrology. 22. O’Rourke MF, Pauca A, Jiang XJ. Pulse wave analysis. Br J Clin Pharmacol
2001; 51: 507–522
CONFLICT OF INTEREST STATEMENT 23. Trachet B, Reymond P, Kips J et al. Numerical validation of a new
method to assess aortic pulse wave velocity from a single recording of a bra-
Fresenius Canada made the BCM available for the present chial artery waveform with an occluding cuff. Ann Biomed Eng 2010; 38:
study. 876–888
24. Hung SC, Kuo KL, Peng CH et al. Volume overload correlates with cardio-
REFERENCES vascular risk factors in patients with chronic kidney disease. Kidney Int
2014; 85: 703–709
1. Locatelli F, Pozzoni P, Tentori F. Epidemiology of cardiovascular risk in 25. Tsai YC, Chiu YW, Tsai JC et al. Association of fluid overload with cardio-
patients with chronic kidney disease. Nephrol Dial Transplant 2003; 18: vascular morbidity and all-cause mortality in stages 4 and 5 CKD. Clin J Am
vii2–vi9 Soc Nephrol 2015; 10: 39–46
2. Vasavada N, Agarwal R. Role of excess volume in the pathophysiology of 26. Coresh J, Selvin E, Stevens LA et al. Prevalence of chronic kidney disease in
hypertension in chronic kidney disease. Kidney Int 2003; 64: 1772–1779 the United States. JAMA 2007; 298: 2038–2047
3. Dussol B, Moussi-Frances J, Morange S et al. A pilot study comparing furo- 27. Tsai YC, Chiu YW, Kuo HT et al. Fluid overload, pulse wave velocity, and
semide and hydrochlorothiazide in patients with hypertension and stage 4 ratio of brachial pre-ejection period to ejection time in diabetic and non-
or 5 chronic kidney disease. J Clin Hypertens (Greenwich) 2012; 14: 32–37 diabetic chronic kidney disease. PLoS One 2014; 9: e111000
4. Buckalew VM, Berg RL, Porush JG et al. Prevalence of hypertension in 28. Braam B, Taler SJ, Rahman M et al. Recognition and management of resis-
1,795 subjects with chronic renal disease: the modification of diet in renal tant hypertension. Clin J Am Soc Nephrol 2017; 12: 524–535
disease study baseline cohort. Am J Kidney Dis 1996; 28: 811–821 29. Pimenta E, Gaddam KK, Oparil S et al. Effects of dietary sodium reduction
5. Tanner RM, Calhoun DA, Bell EK et al. Prevalence of apparent treatment- on blood pressure in subjects with resistant hypertension: results from a
resistant hypertension among individuals with CKD. Clin J Am Soc Nephrol randomized trial. Hypertension 2009; 54: 475–481
2013; 8: 1583–1590 30. Protogerou A, Blacher J, Stergiou GS et al. Blood pressure response under
6. Karaaslan DC, Afsar B, Kanbay M. Is relative overhydration measurement chronic antihypertensive drug therapy: the role of aortic stiffness in the
by bioimpedance spectroscopy useful in reducing morbidity and mortality REASON (preterax in regression of arterial stiffness in a controlled double-
in chronic kidney disease? Clin Kidney J 2018; 11: 370–371 blind) study. J Am Coll Cardiol 2009; 53: 445–451
7. Essig M, Escoubet B, de Zuttere D et al. Cardiovascular remodelling and ex- 31. Kwan BC, Szeto CC, Chow KM et al. Bioimpedance spectroscopy for the
tracellular fluid excess in early stages of chronic kidney disease. Nephrol Dial detection of fluid overload in Chinese peritoneal dialysis patients. Perit Dial
Transplant 2007; 23: 239–248 Int 2014; 34: 409–416
8. Kheder-Elfekih R, Yannoutsos A, Blacher J et al. Hypertension and chronic 32. Tycho Vuurmans JL, Boer WH, Bos WW et al. Contribution of volume
kidney disease: respective contribution of mean and pulse pressure and arte- overload and angiotensin II to the increased pulse wave velocity of hemodi-
rial stiffness. J Hypertens 2015; 33: 2010–2015 alysis patients. J Am Soc Nephrol 2002; 13: 177–183
9. London GM. Arterial stiffness in chronic kidney disease and end-stage renal 33. Raimann JG, Zhu F, Wang J et al. Comparison of fluid volume estimates in
disease. Blood Purif 2018; 45: 154–158 chronic hemodialysis patients by bioimpedance, direct isotopic, and dilution
10. Kim DH, Braam B. Assessment of arterial stiffness using applanation to- methods. Kidney Int 2014; 85: 898–908
nometry. Can J Physiol Pharmacol 2013; 91: 999–1008 34. Ring M, Eriksson MJ, Zierath JR et al. Arterial stiffness estimation in healthy
11. Briet M, Bozec E, Laurent S et al. Arterial stiffness and enlargement in mild- subjects: a validation of oscillometric (Arteriograph) and tonometric
to-moderate chronic kidney disease. Kidney Int 2006; 69: 350–357 (SphygmoCor) techniques. Hypertens Res 2014; 37: 999–1007
12. Ford ML, Tomlinson LA, Chapman TP et al. Aortic stiffness is indepen- 35. DuPont JJ, Greaney JL, Wenner MM et al. High dietary sodium intake
dently associated with rate of renal function decline in chronic kidney dis- impairs endothelium-dependent dilation in healthy salt-resistant humans.
ease stages 3 and 4. Hypertension 2010; 55: 1110–1115 J Hypertens 2013; 31: 530–536
13. Zieman SJ, Melenovsky V, Kass DA. Mechanisms, pathophysiology, and 36. Oberleithner H. A physiological concept unmasking vascular salt sensitivity
therapy of arterial stiffness. Arterioscler Thromb Vasc Biol 2005; 25: 932–943 in man. Pflugers Arch 2012; 464: 287–293
14. Levin A, Stevens PE. Summary of KDIGO 2012 CKD guideline: behind the 37. Oberleithner H. Vascular endothelium: a vulnerable transit zone for merci-
scenes, need for guidance, and a framework for moving forward. Kidney Int less sodium. Nephrol Dial Transplant 2014; 29: 240–246
2014; 85: 49–61
15. Levey AS, Stevens LA, Schmid CH et al. A new equation to estimate glomer-
Received: 26.8.2018; Editorial decision: 14.1.2019
ular filtration rate. Ann Intern Med 2009; 150: 604–612
6 B. Braam et al.