Abstract
Whole blood viscosity (WBV) plays a role in hemorheology and is determined by many factors such as red blood cell factors, plasma protein and blood volume. As WBV changes during hemodialysis, mortality may be due to changes in WBV in patients on hemodialysis. However, there are few prospective data on the relationship between changes in WBV and overall mortality in dialysis patients. We tried to investigate the correlations between values of WBV at variable shear rates before and after hemodialysis and overall or atherosclerosis-related mortality in patients with end-stage kidney disease.
Forty-three patients receiving hemodialysis were enrolled in this study. In this 5.8-year prospective observational study, analyses of the effects of WBV at shear rates of 300 s−1 (systolic WBV; SBV), 5 s−1 (diastolic WBV5; DBV5), and 1 s−1 (diastolic WBV1; DBV1) during dialysis on all-cause and atherosclerotic mortality was performed.
Among a total of 43 patients, 27 (62.7%) died over the course of the study. Thirteen deaths were caused by atherosclerotic events. A high degree of change in WBV at shear rates of 300 s−1 and 5 s−1 during hemodialysis (ΔSBV, ΔDBV5) was positively correlated with overall mortality (HR = 4.688, 95% confidence interval [CI], 1.269–17.319, p = 0.020; HR = 3.941, 95% CI, 1.057–14.701, p = 0.041, respectively). A high degree of change in diastolic blood pressure (ΔDBP) during hemodialysis was also positively correlated with overall mortality (HR = 3.035, 95% CI, 1.039–8.867, p = 0.042). However, comparative analysis between WBV at shear rates of 300 s−1, 5 s−1, and 1 s−1 and overall mortality did not reveal any significant relationships. Kaplan-Meier analysis revealed that the all-cause mortality was significantly higher in patients from a high degree of change of WBV at shear rates of 300 s−1, compared to those from the moderate or low degree of changes of WBV at shear rates of 300 s−1 (p = 0.020, log-rank test). Survival rate in high ΔDBP was lower than that of moderate or low ΔDBP group in Kaplan-Meier survival analysis (p = 0.004, log-rank test).
Our data showed that a high degree of change in WBV at variable shear rates during hemodialysis might impact overall survival in patients with end-stage kidney disease. However, large-scale studies to evaluate the relationship of WBV with overall mortality and atherosclerotic mortality will be needed.
Introduction
Whole blood viscosity (WBV) has a critical role in flows in both large arteries and micro vessels [4]. When blood flows fast, the blood viscosity becomes relatively small, whereas when blood flows slowly, the blood viscosity greatly increases, a phenomenon that is called a Non-Newtonian shear-thinning viscosity behavior [12, 25]. Thus, the WBV is often described as a function of shear rate, which is defined as the ratio of blood velocity to the lumen diameter. Since the shear stress is mathematically given as the product of viscosity and shear rate (defined as the ratio of flow velocity to lumen diameter), this WBV is also associated with fluid shear stress within blood vessels and other characteristics, such as deformability and aggregation of erythrocyte, plasma protein concentration [6, 33]. Fluid shear stress within blood vessels can be correlated with growth and new formation of blood vessels [2, 37].
Shear stress within the vascular endothelium is also associated with localized inflammation and triggers atherosclerotic changes in human vasculature [2, 33]. The increased WBV theoretically makes endothelial shear stress more oscillatory, creating endothelial injury in human vasculature, and particularly at the bifurcation of large arteries, leading to the progression of atherosclerotic changes [37]. According to a previous study, blood viscosity is recognized as an atherothrombotic factor because of the causal relationship between WBV and cardiovascular events [12, 17]. Furthermore, the increased WBV increases the flow resistance in peripheral vessels and microvasculature, resulting in microvascular flow impairment [7].
Elevated WBV increases the wall shear stress and induces an increased release of NO, thereby strongly influencing blood cells and the vascular wall [35]. The elevated WBV can make endothelial shear stress more injurious, which consequently increases intravascular inflammation and atherothrombotic risk [23, 35]. Lee et al. [24] have demonstrated that blood viscosity is well correlated with carotid intima-media thickness. In patients with end-stage renal disease (ESRD) undergoing hemodialysis, changes in WBV during hemodialysis are developed easily due to hemoconcentration and are more pronounced at lower shear rates [22].
Hemodialysis itself in ESRD patients also results in unfavorable hemodynamic outcomes that increase morbidity regardless of well-known cardiovascular risk factors, such as diabetes mellitus (DM), hypertension, and obesity. Recently, it has been reported that overhydration of patients with hemodialysis may be related with the mortality [20]. Since changes in WBV can directly influence endothelial shear stress in both large arteries and micro vessels, the intradialytic WBV changes can be clinically important factors related with morbidity in patients undergoing hemodialysis [4, 36]. Canaud et al. [6] demonstrated that higher WBV in a cardiovascular event group was found in hemodialysis patients, but the result was not statistically significant. Koenig et al. [22] reported that elevated blood viscosity is an independent risk factor for predicting all-cause mortality in apparently healthy men. However, there are few report about the relationship between WBV and mortality in patients undergoing hemodialysis.
We hypothesize that both all-cause and atherosclerosis-related mortalities are related to changes in WBV in patients undergoing hemodialysis. The objective of the present study was to investigate the correlation of the changes in WBV at shear rates of 1, 5, and 300 s−1 with all-cause and atherosclerosis-related mortality during hemodialysis in patients with ESRD. To accomplish this objective, we prospectively monitored patients for approximately 5.8 years.
Material and methods
Patients
This is a single-center, observational, prospective study carried out at Chonbuk National University Hospital from 2009 to 2015 to determine the impacts of changes in WBV during hemodialysis on all-cause and atherosclerosis-related mortality in patients with ESRD. This study was performed with approval from the Ethical Committee of Chonbuk National University Hospital, and all participating patients gave written informed consent prior to onset of the study. We previously reported that changes in WBV at low shear rates are related to changes in intravascular volume during hemodialysis [21]. Previous WBV data were used to evaluate the relationship of changes in WBV at variable shear rates with mortality (Table 1).
Enrolled patients underwent a regular hemodialysis two to three times a week. Hollow-fiber dialyzers (composed of Polyflux 14L® (Gambro, Lund, Sweden) and BLS512SD® (Fresenius Medical Care, Bad Homburg, Germany)) were used for this study. Polyflux 14L® was used for 39 patients, and BLS512SD® was used for 4 patients.
Measurement of WBV
This study included 43 patients who were enrolled in the previous study [21]. We measured the blood viscosity of a native hematocrit at the core body temperature of 37C. We use both systolic blood viscosity (SBV) and diastolic blood viscosity (DBV) in the present study as two representative characteristics of patient’s viscosity profile in clinical utility. We define that the SBV is blood viscosity at a shear rate of 300 s−1 and the DBV is blood viscosity at low shear rates, such as 5 s−1 and 1 s−1. Conceptually, the SBV provides useful information on the blood viscosity at relatively large vessels of lumen diameter of 3 mm or greater at the peak systole. On the other hand, DBV provides useful information on the blood viscosity in vessels where flow velocity is very small, i.e., less than 0.1 cm/s and thus the risk of erythrocyte aggregation is relatively high.
Pre- and post-dialysis WBV (Scanning capillary tube viscometer, Bio-Visco Inc., Republic of Korea) were measured at shear rates of 300 s−1, 5 s−1, and 1 s−1 before hemodialysis (pre-SBV, pre-DBV5, and pre-DBV1, respectively) and after hemodialysis (post-SBV, post-DBV5, and post-DBV1, respectively) [18]. From these pre- and post-values, the changes in WBV before and after hemodialysis, i.e., ΔSBV and ΔDBV, were determined. The whole blood viscosity measured at a shear rate of 300 s−1 is described as systolic blood viscosity (SBV). In addition, whole blood viscosities measured at shear rates of 5 and 1 s−1 are described as diastolic blood viscosities (DBV), which are identified as DBV5 and DBV1, respectively.
Outcome measurement
The primary outcome variables were all-cause and cerebro- or cardiovascular mortality. Deaths from cardiovascular diseases were defined as deaths associated with coronary artery disease (myocardial infraction, unstable angina, and stable angina which are diagnosed with the cardiac enzymes or echocardiogram or coronary angiography), congestive heart failure diagnosed with echocardiography and clinical symptoms or signs, sudden cardiac arrest, and acute pulmonary thromboembolism. Deaths from cerebrovascular diseases were defined as deaths caused by acute cerebral infarction or acute cerebral hemorrhage. Acute cerebral infarction or cerebral hemorrhage was recognized with brain computed tomography.
Classification of patients
To determine the effect of WBV on cardiovascular and cerebrovascular mortality, patients were classified according to the presence of cardiovascular diseases (Table 2) or cerebrovascular diseases (Table 3). Magnitudes of WBV at three different shear rates and systolic/diastolic blood pressures before and after hemodialysis are given in percentiles in Table 4. Patients were divided to high, moderate, and low groups accordance with percentiles in pre- and post-dialysis WBVs and blood pressures in Table 5. For example, if changes in SBV (i.e., ΔSBV) resulted in values higher than the 75th percentile, patients were included in the high blood viscosity group. The moderate blood viscosity group included WBV change between the 25th and 75th percentile, and the low blood viscosity group, less than the 25th percentile (Table 5). Table 5 shows the hazard ratios for the moderate and high groups relative to the low group in each factor.
Statistical analyses
Baseline characteristics of patients were described as median values and ranges, numbers, and percentages according to parameter. A Mann-Whitney test was also used to compare influences of values of blood viscosity according to presence of causes of death such cardiovascular and cerebrovascular disorders. Continuous variables and percentages of basic data in this study were categorized as median (25–75th percentile) values. Therefore, patients were divided into three groups according to blood viscosity (high, moderate, and low). All continuous variables were subjected to the Shapiro-Wilk test. Differences in continuous variables were determined by Wilcoxon rank-sum test. Survival times were calculated from the beginning date of this study. Patients who did not have to maintain hemodialysis due to kidney transplantation were considered censored events for survival analysis. Survival curves were generated by the Kaplan-Meier product-limit method and were compared using the log-rank test according to blood viscosity group. Univariate and multivariate analyses were based on the Cox proportional hazards model. For Cox univariate analysis, the relative importance of prognostic factors was determined according to Chi-squared values. And covariates, including DM, cardiovascular death, and cerebrovascular death were considered in Cox multivariate analysis. For Cox multivariate analysis, the relative importance of prognostic factors was also determined according to Chi-squared values. A p value < 0.05 was considered statistically significant. All data were statistically analyzed with SAS software, version 9.1.3 (SAS Institute Inc., Cary, NC).
Results
Baseline characteristics and laboratory findings of patients
The baseline characteristics and laboratory findings of all 27 patients who died during maintenance hemodialysis were summarized. The overall survival rate of patients with ESRD during approximately 70 months was about 37.2%. Causes of deaths and cardiovascular risk factors were also examined, including diabetes mellitus (Table 1).
Changes in WBV and systolic blood pressure correlated with overall survival
We evaluated the effects of both pre- and post-WBVs on the overall survival. In the beginning, it was not clear which parameter (i.e., pre- and post-WBVs or ΔWBV) is more useful indicator for predicting survival during hemodialysis. Hence, the effect of intradialytic changes in WBV (ΔWBV) at shear rates of 1 s−1 (ΔDBV1), 5 s−1 (ΔDBV5), and 300 s−1 (ΔSBV) on the overall survival in patients with ESRD were also evaluated.
The results of the univariate Cox proportional analysis are shown in Table 5, where, HRs for overall survival in patients with ESRD according to Pre-SBV, Pre-DBV5, and Pre-DBV1 showed a tendency toward higher overall mortality for moderate and high groups, although these analyses did not show statistical significances. There was also no positive relationship between post-SBV, post-DBV5, and post-DBV1 and overall mortality in univariate Cox proportional analysis (Table 5). Covariates were considered in the multivariate Cox analysis, including DM, cardiovascular death, and cerebrovascular death. However, even after the HRs of all SBV, DBV5, and DBV1 before and after hemodialysis were adjusted for covariates through multivariate Cox hazard analyses, there was still no statistical significance.
However, a high degree of intradialytic change in WBV at shear rates of 300 s−1 and 5 s−1 during hemodialysis (ΔSBV, ΔDBV5) was an independent predictor of all-cause mortality in patients with ESRD (HR = 4.688, 95% CI, 1.269–17.319, p = 0.020; HR = 3.941, 95% CI, 1.057–14.701, p = 0.041, respectively). In addition, a high degree of intradialytic change in diastolic blood pressure during hemodialysis (ΔDBP) was positively correlated with all-cause mortality (HR = 3.035, 95% CI, 1.039–8.867, p = 0.042). A moderate degree of intradialytic change in hematocrit after hemodialysis (ΔHct) also showed a positive correlation with all-cause mortality (HR = 4.508, 95% Cl, 1.019–19.952, p = 0.047).
WBV at variable shear rates during hemodialysis and cardiovascular mortality
A total of eight patients died due to cardiovascular events during this study, including two patients who died of sudden cardiac arrest, one of acute myocardial infarction, one of acute pulmonary thromboembolism, and four of aggravation of congestive heart failure and hypotension. The influence of several parameters on cardiovascular mortality was investigated, including values of ΔWBV at shear rates of 300, 5, and 1 s−1 before and after hemodialysis. In a comparative analysis of cardiovascular death and non-cardiovascular death, there were no statistically significant differences. The influence of the intradialytic change of WBV during hemodialysis at variable shear rates on cardiovascular mortality was also not statistically significant (Table 2).
WBV at variable shear rates during hemodialysis and cerebrovascular mortality
The influence of WBV on cerebrovascular mortality was also examined. Both WBV at variable shear rates during hemodialysis and the intradialytic change in WBV were not significantly higher in the cerebrovascular death group (Table 3).
ΔSBP and ΔDBP during hemodialysis and cardiovascular/cerebrovascular mortality
High blood pressure has been found to be related to cardiovascular mortality [1, 32]. Thus, we investigated the impact of systolic and diastolic blood pressures on both cardiovascular and cerebrovascular mortality in patients with ESRD undergoing hemodialysis. There were no statistically significant relationships between systolic or diastolic blood pressure or changes in these parameters during dialysis, and cardiovascular/cerebrovascular mortality (Tables 2 and 3).
Overall survival is lower in high ΔSBV or ΔDBP groups
Kaplan–Meier survival curves for intradialytic changes in WBV (i.e., ΔSBV, ΔDBV5, and ΔDBV1), blood pressure (i.e., ΔSBP and ΔDBP) and all-cause mortality are shown in Figs. 1 and 2. The median survival time of all patients enrolled in this study was 48 months. The overall survival rate was significantly lower in the high ΔSBV group than it was in the moderate or low ΔSBV groups (p = 0.020; Fig. 1). Patients in the high ΔDBP group had a lower survival rate than those in the moderate or low ΔDBP groups (p = 0.004; Fig. 2).
Discussion
In general, cardiovascular diseases in patients with ESRD occur much more often than in the general population. Cardiovascular diseases, including congestive heart failure, myocardial infarction, and sudden cardiac death, are important causes of an increased mortality rate in patients with ESRD undergoing hemodialysis [31, 38]. Cerebrovascular diseases are also important causes of death in patients with ESRD [28]. According to previous epidemiologic studies, cardiovascular mortality in patients with ESRD undergoing hemodialysis was more than 10 times higher than in the general population [3, 13].
The atherosclerotic vascular diseases, such as cardiovascular and cerebrovascular disease can be accelerated by increased blood viscosity. In addition, increased WBV was well correlated with increased carotid intima media, hypertension, and lipid profiles, which are recognized as major cardiovascular risk factors according to previous reports [2, 34]. The increased SBV and DBV definitely increase vascular flow resistance, which may result in atherosclerotic vascular complications [2, 33]. For example, the intima at bifurcations at both coronary and carotid arteries can be adversely affected as the flow separation and recirculation become more pathological (i.e., the condition that can cause intimal injury) with increased SBV, which alters the responsiveness of endothelial vaso-reactive materials including nitric oxide, nitric oxide synthase, thus accelerating the progression of atherosclerosis. On the other hand, the increased DBV can cause a microvascular perfusion problem. For example, the perfusion in small vessels where flow velocity is less than 0.1 cm/s will experience impaired flow and eventually the loss of capillary vessels with increasedDBV.
We prospectively analyzed the correlation between intradialytic changes in WBV during hemodialysis and mortality in patients with ESRD over approximately 70 months. In addition, several studies have demonstrated a relationship between blood pressure (systolic or diastolic) and cardiovascular mortality, so we also evaluated the impact of intradialytic changes in diastolic and systolic blood pressure during hemodialysis on atherosclerotic mortality in patients with ESRD [1, 32].
Firstly, the mortality rate of this study over approximately 70 months was 62.8%, which was higher than that of another study [14]. In epidemiologic studies of Korean patients with ESRD, the mortality rate of patients with ESRD undergoing hemodialysis was also reported to be approximately 60–70% like the present study [16]. This significant difference in mortality rate between the present study and Green et al. [14] might result from differences in baseline characteristics for the enrolled patients in the two studies. The present study generally enrolled patients with ESRD who had already begun hemodialysis about 2 years before the beginning of this study. Many patients had serious underlying conditions at enrollment, such as malignant neoplasms. These etiologies might have resulted in a much higher mortality rate in patients with ESRD receiving maintenance hemodialysis. As several studies have reported, WBV at a low shear rate showed a positive correlation with the serum concentration of several plasma contents, such as immunoglobulins and fibrinogens [5, 30]. In addition, as mentioned early, the low shear blood viscosity indirectly indicates the microvascular perfusion in small vessels where blood moves very slowly. In the science of hemodynamics, this is called as the low Reynolds number flow, where the viscous force represented by DBV is approximately 100–1,000 greater than the inertia force even in a normal physiology [11]. With the increased DBV, the vascular flow resistance further increases, forcing the microvascular flow to stop, a phenomenon that may result in atherosclerotic vascular complications as well as flow impairment [8]. This relationship reflects the impact of the blood viscosity in the flow of relatively small arterioles, strongly related to tissue hypoxia and subsequent organ damage [10, 37]. Finally, in this study, overall mortality in patients with ESRD undergoing hemodialysis was well correlated with ΔSBV, ΔDBV5, and ΔDBP. Among these values, both ΔSBV and ΔDBP showed a significant relationship with overall survival on Kaplan-Meier analysis. When intradialytic changes in blood viscosity and diastolic blood pressure were low during hemodialysis, the overall survival rate increased significantly in patients with ESRD. These results support evidence from previous studies suggesting that atherosclerosis-related mortality in patients with ESRD undergoing hemodialysis is more than 10 times higher than that of the general population with respect to intradialytic changes inWBV [3, 13].
This study has some limitations, including a small number enrolled patients, blood viscosity measurement error, difficulty of clinical application of blood viscosity, and difficulty in controlling variables that impact blood viscosity. However, we cautiously suggest that intradialytic changes in WBV and diastolic blood pressure during hemodialysis may be better indicators than WBV values for overall mortality in patients with ESRD. As mentioned above, this study was a prospective observational study that did not consider many variable factors that may impact blood viscosity in patients with ESRD undergoing hemodialysis. If the limitations associated with this study are accounted for, the intradialytic changes in WBV especially ΔSBV and ΔDBV5 values, will be useful markers for predicting overall mortality in patients with ESRD undergoing hemodialysis. We also found that the blood viscoisty values in this study are relative high in comparison with blood viscosity obtained with other viscometer [6]. This may be due to the difference of measurement methods.
All of our data suggest that changes in WBV and diastolic blood pressure during hemodialysis predict survival in patients with ESRD undergoing hemodialysis. Future prospective, large-scale, more sophisticated studies are needed to further evaluate this relationship.
Footnotes
Acknowledgments
This paper was supported by the Fund of the Biomedical Research Institute, Chonbuk National University Hospital (Park S.K.); a National Research Foundation of Korea (NRF) grant funded by the Korean Government (MISP) (No. 2008-0062279, Kim W); and a grant (CUHBRI-2012-02-003) of the CNUH-BRI (Kim W.). The biospecimens and data used in this study were provided by the Biobank of Chonbuk National University Hospital, a member of the Korea Biobank Network, which is supported by the Ministry of Health, Welfare and Family Affairs. All samples derived from the Korea Biobank Network were obtained with informed consent under institutional review board-approvedprotocols.
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (NRF-2014R1A2A2A01002545).