Figures
Abstract
Endothelial dysfunction is a key initiating event in vascular disease in chronic kidney disease (CKD) patients and haemodialysis (HD) patients exhibit significant vascular abnormalities. To understand this further, we examined how ex vivo intrinsic function in isolated arteries correlates with in vivo assessments of cardiovascular status in HD patients. Abdominal fat biopsies were obtained from 11 HD patients and 26 non-uremic controls. Subcutaneous arteries were dissected and mounted on a wire myograph, and cumulative concentration-response curves to noradrenalin, endothelin-1, a thromboxane A2 agonist (U46619), angiotensin II, vasopressin, bradykinin (BK), acetylcholine (ACh) and sodium nitroprusside (SNP) were constructed. Pulse wave velocity and blood pressure were measured in HD patients. Enhanced (P<0.05−0.0001) maximal contractile responses (Rmax) to all spasmogens (particularly vasopressin) were observed in arteries from HD patients compared to controls, and this effect was more pronounced in arteries with an internal diameter>600 µm. The potency (pEC50) of U46619 (P<0.01) and vasopressin (P<0.001) was also increased in arteries>600 µm of HD patients. The maximal relaxant response to the endothelium-dependent dilators ACh and BK were lower in HD patients (P<0.01-P<0.0001) (worse for ACh than BK); however the endothelium-independent dilator SNP was similar in both groups. PWV was significantly correlated with the vasoconstrictor response to vasopressin (P = 0.042) in HD patients. HD patients are primed for hypertension and end organ demand ischaemia by a highly sensitised pressor response. The failure of arterial relaxation is mediated by endothelial dysfunction. Intrinsic vascular abnormalities may be important in sensitising HD patients to recurrent cumulative ischaemic end organ injury.
Citation: Abushufa AM, Eldehni MT, Odudu A, Evans PD, O′Sullivan SE, McIntyre CW (2014) Defining Uremic Arterial Functional Abnormalities in Patients Recently Started on Haemodialysis: Combined In Vivo and Ex Vivo Assessment. PLoS ONE 9(12): e113462. https://doi.org/10.1371/journal.pone.0113462
Editor: Christopher Torrens, University of Southampton, United Kingdom
Received: May 27, 2014; Accepted: October 26, 2014; Published: December 29, 2014
Copyright: © 2014 Abushufa et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper.
Funding: The authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Cardiovascular disease is the major cause of mortality in CKD patients. These patients are characterised by a wide range of both structural and functional abnormalities of the vessels. There is strong emerging evidence that the hemodialysis (HD) procedure causes significant systemic circulatory stress [1]. This circulatory stress interacts with complex haemodynamic factors causing perfusion anomalies that accelerate end organ damage in a wide range of vulnerable vascular beds [2]. Initially this work focused on the heart, but emerging evidence strongly suggests that a much wider range of organs are affected. This includes the gut, kidney, skin and critically, the brain [3]. Fixed structural abnormalities of the vascular tree develop during the lifetime of a progressive CKD patient. These include coronary calcification and other drivers of reduced arterial compliance. These changes are the result of fixed vascular structural change but importantly additional functional microvascular abnormalities are potentially superimposed onto these processes [4]. These may be more important in determining the propensity to multi organ based demand ischaemia.
It is unclear to what degree the observed abnormalities of microvascular function are related to the uremic milieu and exposure of relatively normal vasculature to a variety of functionally modifying factors, and to what degree there are intrinsic defects in pressor response. Vascular endothelium plays an important role in the regulation of vascular function and tone through production of a wide range of vasoactive substances including endothelin, nitric oxide (NO), and thromboxane A2 [5]. Different mechanisms have been proposed to cause endothelial dysfunction in HD patients, which is largely due to a defect in the endothelial NO bioavailability [6]. Several uraemia-associated factors including oxidative stress and reactive oxygen species [7], [8], homocysteine [9] and plasma nitric oxide synthase inhibitors (asymmetric dimethylarginine) [10] are found in high levels in the plasma of HD patients and may contribute to endothelial dysfunction. Impaired endothelium-dependent vasodilatation has been observed in several in vivo studies involving CKD [11], [12] and HD patients [13], [14]. A limited number of ex vivo studies [15]–[17] have examined small isolated arteries removed from uremic patients selected to be undergoing a surgical procedure have revealed similar vascular dysfunction ex vivo to a limited number of vasoactive agents.
Large vessel compliance (arterial stiffness), as measured by increased PWV, is higher in patients on dialysis (and with non-dialysis dependant CKD) compared with the general population [18]. Increased PWV is associated with elevated cardiovascular morbidity and mortality risks in patients with CKD stage 5 [19]. A strong association has been observed between PWV and abdominal aorta calcification in HD patients [20]. This change in compliance is partially due to material alterations in the conduit arteries, however it is also blood pressure (BP) and endothelial dysfunction related.
This current study was focused on an investigation of the vascular function in isolated resistance arteries obtained from subcutaneous fat of HD patients, with comparison to appropriate matched non uremic control. We aimed to establish whether HD patients show enhanced responses to a wider range of different vasoconstrictors and impaired endothelial-dependent vasodilatation, as previously suggested [15], [16] and how that ex vivo intrinsic function correlates to in vivo assessments of cardiovascular status. Results from smaller (internal diameter less than 600 µM) were also compared with larger arteries (internal diameter greater than 600 µM) to establish whether different calibre arteries are affected differentially by HD.
Materials and Methods
Subcutaneous fat samples were obtained from 11 HD patients who were within the first 90 days of starting dialysis. Control fat samples were obtained from 26 appropriately consented healthy volunteers without documented renal disease, who were undergoing elective hernia repair. The ethical approval was granted by the Derbyshire Research Ethics Committee. Informed written consent was obtained in accordance with Good Clinical Practice guidelines.
Subcutaneous biopsies were obtained from HD patients using local anaesthetic (1% lignocaine) through a small (1–2 cm) lower lateral abdominal incision. Subcutaneous adipose tissue was dissected and separated from the adherent skin, and then a piece of fat was removed using scalpel and surgical scissor. Samples were placed immediately in chilled physiological salt solution (PSS) and experimented on the same day of biopsy.
In vitro assessments of vascular function
Arteries of varying diameters were dissected from each sample per patient and mounted as a ring preparation on two wire myographs (ie. Up to 8 segments of different sized arteries per patient)(model M620, Aarhus, Denmark) in PSS and continuously bubbled with 5% CO2 and 95% O2. PSS was composed of (in mmol/L) NaCl 119, KCL 4.7, CaCl2.2H2O 2.5, NaHCO3 25, MgSO4.7H2O 1.17, KH2PO4 1.18, D-glucose 5.5 and EDTA 0.026. Following an acclimatisation period, arteries were subjected to a standard normalization process [21] to determine the internal circumference of arteries. The integrity of arteries was tested by their ability to contract by at least 5 mN (0.5 g) tension to a high potassium PSS (KPSS, high potassium chloride concentration (124 mmol/L) substituted for sodium chloride). Arteries unable to contract to KPSS were discarded. Viable arteries were then contracted using 50 nM of the thromboxane mimetic U46619 and once a plateau contraction has reached, 10 µM of bradykinin was added to confirm an intact endothelium. All arteries were washed-out with PSS until re-establishment of the base line. Cumulative-concentration response curves for noradrenaline (NA, 100 pM to 100 µM), endothelin-1 (ET-1, 1 pM to 1 µM), U46619 (1 pM to 3 µM), angiotensin II (Ang II, 1 pM to 1 µM), and vasopressin (1 pM to 1 µM) were constructed in separate segments of artery. These started with the lowest concentration of the drug and allowed 3–5 minutes per stimulation.
To assess vasodilator function, separate segments of arteries were first constricted using a combination of 100 nM U46619 and 1 nM ET-1. Once the peak steady contraction had obtained, cumulative-concentration response curves for bradykinin (BK, 100 pM to 100 µM), acetylcholine (ACh, 100 pM to 100 µM) and sodium nitroprusside (SNP, 100 pM to 100 µM) were constructed.
In vivo haemodynamic measurement
Brachial artery BP was measured in the non-fistula arm using an oscillometric device. Carotid-to-femoral PWV was measured using a Vicorder device as previously described [22]. A carotid pressure cuff is applied over the neck to detect the right carotid artery and a pressure cuff is placed around the right upper thigh to detect the femoral artery. The distance from the suprasternal notch to the middle of the thigh cuff was measured. The signal from each cuff is analysed to derive arterial transit time. PWV is calculated using the software, dividing arterial transit time by measured distance.
Statistical analysis
GraphPad Prism (San Diego, CA, USA) was used to plot the data as mean percentage relaxation, with error bars representing standard error of the mean (SEM), and n being the number of arteries from different patients. The software was used to fit sigmoidal concentration-response curve to the mean data using a logistic equation. Percentage maximal vasorelaxant responses (Rmax) and potency (EC50) were calculated from this mean data were compared control and HD patients using unpaired Students' t-test, with p<0.05 taken as significant.
Results
There was no difference in the characteristics of HD and non-HD patients other than uremic status (see Table 1). Haemodialysis is generally commenced at an eGFR below 10 mL/min/1.73 m2, so patients will have a lower eGFR than the controls by virtue of being on dialysis.
Effects of HD on contractile responses
In vasoconstrictor studies, the average internal diameter (ID) of arteries used was similar between HD and control groups for smaller (HD; 473±42 µm (n = 30 individual segments) cf control: 449±79 µm (n = 45)) or larger arteries (HD; 738±96 µm (n = 27) cf control: 717±89 µm (n = 40)).
In arteries less than 600 µm in diameter, the maximum KPSS contraction was greater in the HD group; 9.8±1.2 mN in HD (n = 11) and 6.99±0.32 mN in controls (n = 26), P<0.05, Students t test with Welch's correction. There was no significant difference in the mean KPSS response in arteries with an ID greater than 600 µm; 14.7±1.2 mN in HD (n = 11) and 13.0±1.2 mN in controls (n = 26).
In arteries less than 600 µm in diameter, the potency (EC50) of vasoconstrictor agents were similar between groups, with the exception that the potency of NA was reduced in HD patients (P<0.05, Fig. 1A). However, the maximal contractile response to ET-1 (P<0.01, Fig. 1A), U46619 (P<0.0001, Fig. 2A), Ang II (P<0.01, Fig. 2C) and vasopressin (P<0.0001, Fig. 2E) were all significantly enhanced in arteries from HD patients.
Data are expressed as mean ± SEM and the comparison is by students t-test; *P<0.05, **P<0.01, ***P<0.001. HD, haemodialysis; n, number of arteries; M, molar; mN, milliNewtons.
Data are expressed as mean ± SEM and the comparison is by students t-test; **P<0.001, ***P<0.001, ****P<0.0001. HD, haemodialysis; U4, U46619; n, number of arteries; M, molar; mN, milliNewtons.
In arteries greater than 600 µm in diameter, the enhanced vasocontractility in HD patients was even more pronounced. The Rmax to NA (P<0.001, Fig. 1B), ET-1 (P<0.05, Fig. 1D), U46619 (P<0.01, Fig. 2B), Ang II (P<0.0001, Fig. 2D) and vasopressin (P<0.0001, Fig. 2F) were all enhanced in HD patients. In these larger arteries, the potency of U46619 (P<0.05, Fig. 2B) and vasopressin (P<0.001, Fig. 2F) was also enhanced in subcutaneous arteries obtained from HD patients.
Effects of HD on endothelium-dependent vasorelaxant responses
In a separate set of studies, vasodilator responses were measured in fresh arterial segments. The average ID of arteries less than 600 µm was 478±81 µm in HD (n = 18) and 512±44 µm in controls (n = 27). In arteries greater than 600 µm, it was 761±62 µm (n = 15) and 744±91 µm (n = 25) in HD and controls respectively.
In arteries less than 600 µm, the potency of BK was reduced (P<0.05) in arteries from HD patients. The maximal vasorelaxant response to BK (P<0.001) and ACh (P<0.0001) was also significantly lower in the HD group compared to controls (Figs. 3A,B). Similarly, in arteries greater than 600 µm, significantly lower vasorelaxation to BK and ACh were obtained in HD patients compared to controls (Figs. 3D,E). The effect of HD was most pronounced on the ACh response in arteries greater than 600 µm, as evidenced by the greatest reduction in the overall response (area under the curve, AUC) (see Fig. 3C,F).
Data are expressed as mean ± SEM and the comparison is by students t-test; *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. HD, haemodialysis; n, number of arteries; M, molar.
Effects of HD on endothelium-independent vasorelaxant responses
The vasorelaxant response to SNP was similar between groups in subcutaneous arteries in arteries of either sized diameter (Fig. 4).
Data are expressed as mean ± SEM and the comparison is by student's t-test. HD, haemodialysis; n, number of arteries; M, molar.
Correlation of ex vivo data with in vivo measurements in HD patients
PWV was significantly correlated with the maximum contractile response of arteries greater than 600 µm to vasopressin (r = 0.829, P = 0.042) (Table 2). When contractile responses were expressed as a percentage of maximum KPSS response, a positive correlation was also seen between PWV and percentage of maximum response of arteries less than 600 µm to vasopressin (r = 0.886, P = 0.019), see table 3. There was a trend for contractile responses to U46619 (P = 0.058) and Ang II (P = 0.083) (Table 3) in arteries greater than 600 µm to also be correlated with PWV. A negative correlation between DBP and the normalised vasopressin responses in arteries greater than 600 µm (r = −0.829, P = 0.042, Table 3). A negative correlation was observed between the DBP of HD patients and the response of large arteries to SNP (r = −0.974, P = 0.016) (Table 2).
Discussion
Endothelial dysfunction is a crucial element in the pathophysiology of increased cardiovascular risk amongst CKD patients, including those on HD. Limited human studies reported to date [15]–[17] have examined the effect of uraemia on isolated vascular function. These studies were conducted on small arteries isolated from ESRD patients (with a very wide range of dialysis vintages), investigating a limited suit of vasoactive agents. The current study was therefore conducted on subcutaneous arteries isolated from HD patients investigating different vasoconstrictors and vasodilators. The present study provides a number of key findings. Firstly there is a markedly increased responsiveness of isolated arteries from HD patients to a range of vasoconstrictors. Secondly, impaired endothelium-dependent vasodilatation, but preserved endothelium-independent vasodilatation, is observed in arteries of HD patients. Thirdly, ex vivo responses of small isolated arteries were correlated to clinically relevant in vivo assessment of cardiovascular system function.
The data published to date on enhanced vasoconstrictor responses of isolated arteries have been somewhat contradictory. Increased contractile responses of isolated arteries have been reported before in CKD patients (not HD), characterised by a greater response to NA and ET-1 [15]. However, in a later study, similar contractile response of small arteries to NA, ET-1, and Ang II was observed between uremic and control groups [16]. This is in contrast to our study with consistently observed significant increments in the arterial contraction in isolated different-sized arteries of HD patients in response to NA, ET-1, U46619, Ang II, and vasopressin in arteries of 2 calibres. The effects of HD were most pronounced on the contractile response to U46619, angiotensin II and vasopressin, and the effect of HD was greater in the arteries of diameter greater than 600 µm, where an increased sensitivity (pEC50) to U46619 and vasopressin was also observed. The data derived from our study are consistent with a previous report on isolated subcutaneous arteries where increased contractions to NA and ET-1 were observed in patients with ESRD [15], but is in contrast with Aalkjaer et al, in which a similar response to NA and Ang II was observed between controls and patients with chronic intermittent dialysis [23]. Our population was unselected (no survivor or transplant bias) and all had been exposed to a very similar and short period of HD (<90 days), furthermore a larger number of arterial segments were studied with the benefits of robust normal volunteer comparators.
Interestingly, our data showed differential effects of HD on contractile responses to the various agents with some ligands (U46619, angiotensin II and vasopressin) being more affected than others (noradrenalin and endothelin-1). This suggests that there is not a global upregulation of contractile function intrinsic to smooth muscle, and must relate to either the expression or function of the receptors involved, or the intracellular mechanisms to which the receptors are coupled. Since all the receptors activated by these ligands are Gq- coupled (bringing about contraction via increased IP3 and DAG levels is to increase intracellular Ca2+), changes in the expression of the receptors seems a more likely explanation.
In addition to excessive pressor response of isolated arteries from HD patients, all showed significant decreased endothelium-dependent vasodilatation. This defect was observed in both artery sizes in HD patients, although the blunted response to acetylcholine was bigger in the arteries with a larger diameter. By contrast, the vasorelaxant response to SNP (endothelium-independent vasodilator) was preserved in the HD group. Previous studies on isolated subcutaneous arteries of uremic ESRD patients have similarly demonstrated impaired vasodilatory function in response to ACh but not SNP [15], [16]. Luksha et al. also found that the percentage relaxation and the sensitivity (EC50) to both BK and ACh were reduced in isolated small arteries of patients commencing PD [17]. Similar results have been reported previously in vivo using non-invasive techniques such as Doppler ultrasound [13], and forearm plethysmography [11]. Together this suggests that endothelial-dependent dysfunction is widespread in CKD, observed in conduit arteries, as well as the microcirculation.
This impaired relaxation response could be due to NO dysfunction or decreased ability of vascular endothelium to produce and release vasodilator substances such as endothelium-derived hyperpolarizing factor (EDHF). These vasodilators act by stimulating release of both NO and NO-independent factor (possibly EDHF), from human vascular endothelium [24]. Circulating uremic factors in HD patients may responsible for these changes. These toxins have being widely considered and include elevated plasma homocysteine level [25], increase reactive oxygen species [7], [8], and elevated plasma levels of nitric oxide synthase inhibitors such as ADMA in dialysis patients [10], [26]. In our study, arteries were examined after being removed from exposure to humeral consequences of CKD, suggesting adaptive responses have occurred within the arteries, and the blunted responses are not just a consequence of circulating factors.
Interestingly, the effects of HD were more pronounced when looking at the relaxant response to ACh than BK, and in the larger than smaller calibre arteries. Endothelium-dependent relaxations are achieved by a combination of endothelium-derived prostaglandins, NO, and EDHF. At least for ACh, the role for NO is thought to be more important in conduit arteries and the role for EDHF more important in distal resistance vessels. Thus the greater blunting of ACh responses in arteries greater than 600 µm in our study could reflect a greater contribution of NO production in these arteries, and point to defective NO production as a mechanism of action for the blunted endothelium-dependent responses in the HD patients.
This study represents the first attempt to examine any association between in vivo cardiovascular performance and ex vivo arterial function. PWV positively correlated with a greater response to vasopressin, and there was a trend for an association with the enhanced responses to U46619 and angiotensin II also. It is of note, that these are also the ligands whose responses were enhanced more by HD, and suggests the enhanced responses to these ligands play contribute to the increased PWV seen in HD patients. In addition, the present study found that the DBP was inversely correlated with the contractile response of small arteries to vasopressin, and negatively correlated with the vasorelaxation response of large arteries to SNP. The existence of overhydration in those dialysis patients which have been identified as hypertensive is far from universal. Wabel and co-workers measured pre-dialysis SBP and fluid status in 500 HD patients by bioimpedance monitoring (compared to a matched healthy population). Only 15% of patients fitted the stereotype of fluid overload with hypertension. 13% of patients had hypertension despite dehydration and 10% had fluid overload despite normal BP or hypotension [27]. Similar findings were seen in a another study using the same methodology in 639 patients using PD [28]. These studies highlight that although physicians often estimate hydration status by BP, the two factors are often dissociated in the setting of the profound physiological derangement characteristic of dialysis dependant CKD. A wide variety of potential other pathophysiological processes characteristic of dialysis patients may contribute to generating hypertension without recourse to simple fluid overload, and enhanced pressor response of abnormal uremic resistance vessels appears important.
In conclusion, this study provides additional insights into the effect of HD on vascular function. Exaggerated pressor response and endothelium based failure of relaxation are associated with in vivo measurements of markers of cardiovascular performance we already know to be important in determining HD patient survival. This defective vascular response may be important in sensitising HD patients to recurrent cumulative ischaemic end organ injury driven by the circulatory stress of HD.
Author Contributions
Conceived and designed the experiments: SOS CmcI. Performed the experiments: AA. Analyzed the data: AA SOS. Wrote the paper: AA SOS CmcI. Subject recruitment, subject consenting, taking biopsies, ex vivo cardiovascular measurements: ME AO PE.
References
- 1. Burton JO, Jefferies HJ, Selby NM, McIntyre CW (2009) Hemodialysis-induced cardiac injury: determinants and associated outcomes. Clin J Am Soc Nephrol 4:914–920.
- 2. McIntyre CW (2010) Recurrent circulatory stress: the dark side of dialysis. Semin Dial 23:449–451.
- 3. Eldehni MT, McIntyre CW (2012) Are there neurological consequences of recurrent intradialytic hypotension? Semin Dial 25:253–256.
- 4. Sigrist MK, McIntyre CW (2008) Vascular calcification is associated with impaired microcirculatory function in chronic haemodialysis patients. Nephron Clin Pract 108:c121–126.
- 5. Stankevicius E, Kevelaitis E, Vainorius E, Simonsen U (2003) Role of nitric oxide and other endothelium-derived factors. Medicina (Kaunas) 39:333–341.
- 6. Baylis C (2008) Nitric oxide deficiency in chronic kidney disease. Am J Physiol Renal Physiol 294:F1–9.
- 7. Ferraro B, Galli F, Frei B, Kingdon E, Canestrari F, et al. (2003) Peroxynitrite-induced oxidation of plasma lipids is enhanced in stable hemodialysis patients. Kidney Int 63:2207–2213.
- 8. Hasdan G, Benchetrit S, Rashid G, Green J, Bernheim J, et al. (2002) Endothelial dysfunction and hypertension in 5/6 nephrectomized rats are mediated by vascular superoxide. Kidney Int 61:586–590.
- 9. Bostom AG, Culleton BF (1999) Hyperhomocysteinemia in chronic renal disease. J Am Soc Nephrol 10:891–900.
- 10. Vallance P, Leone A, Calver A, Collier J, Moncada S (1992) Accumulation of an endogenous inhibitor of nitric oxide synthesis in chronic renal failure. Lancet 339:572–575.
- 11. Morris ST, McMurray JJ, Rodger RS, Jardine AG (2000) Impaired endothelium-dependent vasodilatation in uraemia. Nephrol Dial Transplant 15:1194–1200.
- 12. Annuk M, Lind L, Linde T, Fellstrom B (2001) Impaired endothelium-dependent vasodilatation in renal failure in humans. Nephrol Dial Transplant 16:302–306.
- 13. Hand MF, Haynes WG, Webb DJ (1998) Hemodialysis and L-arginine, but not D-arginine, correct renal failure-associated endothelial dysfunction. Kidney Int 53:1068–1077.
- 14. Oflaz H, Pusuroglu H, Genchallac H, Demirel S, Bugra Z, et al. (2003) Endothelial function is more impaired in hemodialysis patients than renal transplant recipients. Clin Transplant 17:528–533.
- 15. Morris ST, McMurray JJ, Spiers A, Jardine AG (2001) Impaired endothelial function in isolated human uremic resistance arteries. Kidney Int 60:1077–1082.
- 16. Luksha N, Luksha L, Carrero JJ, Hammarqvist F, Stenvinkel P, et al. (2011) Impaired resistance artery function in patients with end-stage renal disease. Clin Sci (Lond) 120:525–536.
- 17. Luksha L, Stenvinkel P, Hammarqvist F, Carrero JJ, Davidge ST, et al. (2012) Mechanisms of endothelial dysfunction in resistance arteries from patients with end-stage renal disease. PLoS One 7:e36056.
- 18. Shinohara K, Shoji T, Tsujimoto Y, Kimoto E, Tahara H, et al. (2004) Arterial stiffness in predialysis patients with uremia. Kidney Int 65:936–943.
- 19. Blacher J, Guerin AP, Pannier B, Marchais SJ, Safar ME, et al. (1999) Impact of aortic stiffness on survival in end-stage renal disease. Circulation 99:2434–2439.
- 20. Raggi P, Bellasi A (2007) Clinical assessment of vascular calcification. Adv Chronic Kidney Dis 14:37–43.
- 21. Mulvany MJ, Halpern W (1977) Contractile properties of small arterial resistance vessels in spontaneously hypertensive and normotensive rats. Circ Res 41:19–26.
- 22. Pucci G, Cheriyan J, Hubsch A, Hickson SS, Gajendragadkar PR, et al. (2012) Evaluation of the Vicorder, a novel cuff-based device for the noninvasive estimation of central blood pressure. J Hypertens
- 23. Aalkjaer C, Pedersen EB, Danielsen H, Fjeldborg O, Jespersen B, et al. (1986) Morphological and functional characteristics of isolated resistance vessels in advanced uraemia. Clin Sci (Lond) 71:657–663.
- 24. McIntyre CA, Williams BC, Lindsay RM, McKnight JA, Hadoke PW (1998) Preservation of vascular function in rat mesenteric resistance arteries following cold storage, studied by small vessel myography. Br J Pharmacol 123:1555–1560.
- 25. Mallamaci F, Zoccali C, Tripepi G, Fermo I, Benedetto FA, et al. (2002) Hyperhomocysteinemia predicts cardiovascular outcomes in hemodialysis patients. Kidney Int 61:609–614.
- 26. Kielstein JT, Boger RH, Bode-Boger SM, Frolich JC, Haller H, et al. (2002) Marked increase of asymmetric dimethylarginine in patients with incipient primary chronic renal disease. J Am Soc Nephrol 13:170–176.
- 27. Wabel P, Moissl U, Chamney P, Jirka T, Machek P, et al. (2008) Towards improved cardiovascular management: the necessity of combining blood pressure and fluid overload. Nephrology Dialysis Transplantation 23:2965–2971.
- 28. Van Biesen W, Williams JD, Covic AC, Fan S, Claes K, et al. (2011) Fluid Status in Peritoneal Dialysis Patients: The European Body Composition Monitoring (EuroBCM) Study Cohort. PLoS One 6.