VO is the most important mechanism leading to decompensated HF [162]. When kidney function is impaired or in chronic HF associated with renal hypo-perfusion, the kidney ability to excrete sodium is exceeded, salt sensitivity increases, the sodium-effect of aldosterone is reduced and other conditions, such as renal resistance to natriuretic peptides and non-osmotic release of ADH, appear [1, 68, 163]. In the initial phases, the etiologic factors induced by HF are represented by arterial under-filling and venous congestion; these are further translated into compensatory mechanisms [1]. First, the decreased distension of arterial baroreceptors during arterial under-filling determines the activation of SNS, RAAS and non-osmotic release of vasopressin, leading to a decreased water and Na excretion [1, 163]. The response is represented by vasodilatation with natriuretic peptide release, activation of the kinin-kalikrein system and expression of endothelial relaxation factor, for an increased water and Na excretion [1]. These compensatory adaptive mechanisms eventually become maladaptive, further contributing to VO [162, 163]. There is still a paucity of data assessing the importance of VO as a risk factor for poor clinical outcomes in patients with CKD and CVD, as renal patients are an underrepresented subgroup population in CHF randomized controlled trials (RCT).

Recently, a direct and strong association between plasma Na and LA volume (LAV) was described in asymptomatic patients with 3–5 stages CKD [164]. The plasma Na had an important impact on the variability of LAV, only second to LVM, LV volume and age. Plasma Na was found to be a predictor of approx. 16 % of the total variance in LAV, independently of LVM and LV volume. Also, the study showed a very high prevalence (41.8 %) of subclinical LA enlargement that was not associated with a worsening GFR [164, 165]. Mallamaci et al. performed a subsequent analysis on this study and found that a-2 mmol/L increase in plasma Na is responsible for a 1 ml/m²] increase in LAV [165].

The relationship between VO and BP was evaluated in several studies [9, 110, 166]. In predialysis patients, Verdalles et al. used bioimpedance to assess fluid status and to guide diuretic therapy for treating hypertension in these patients. Thirty patients with extracellular volume (ECV) expansion and a diuretic were compared to 20 patients without ECV expansion who instead received another additional antihypertensive medication. At 6 months of follow-up, SBP decreased by 21 mmHg in patients with ECV expansion compared to 9 mmHg in patients with normal ECW (P < 0.01). In addition, more patients achieved the target BP of less than 140/90 mmHg at 6 months in the group with ECV expansion (nine of 30 patients with ECV expansion cf. two of 20 without ECV expansion).

Based on bioimpedance and cuff BP measurements, Wabel et al. described four distinct categories of individuals in dialysis: (i) normotensive – normovolemics; (ii) hypertensive – normovolemics; (iii) hypertensive – hypervolemics; (iv) normotensive –hypervolemics [110]. It is clear that BP management by different classes of drugs could be tailored much easier and related to prevailing underlying pathophysiological mechanisms. Similarly, our group found in 160 HD patients, using multi-frequency body impedance spectroscopy (MF-BIS) measurements, a large group of hypertensive patients with VO patients, despite an apparently achieved dry weight (DW) on clinical evaluation. Therefore, body composition monitoring (BCM) could represent an accurate measurement tool for the adequate management of BP and risk stratification in HD patients, providing accurate information of more abnormal cardiac and vascular profiles [9]. Similarly, in CKD predialysis and in dialysis patients, the inadequate evaluation of OH and its association with LVH have been established for a long time [9, 13, 109]. Essig et al. investigated 104 patients with early CKD followed-up for 5 years, in order to evaluate body composition, cardiac alterations and their association with GFR. The study showed that ECV was an independent predictor of LVH [7]. Moreover, the correction of the FO could be associated with an improvement of hypertension and left ventricular hypertrophy. Hur et al. in a recent RCT perform bioimpedance spectroscopy in 156 patients; at half of them, fluid removal during dialysis was adjusted based on bioimpedance measurement. After 1 year of follow-up, a substantial regression of the left ventricular mass index (mean difference between groups −10.2 g/m [2]; P < 0.001) and blood pressure was noted in the interventional group but not in the control group [111]. Also in PD patients, FO was associated with hypertension and LVH. Cader et al. aimed to assess the prevalence and LVH and its determinants in a cross-sectional observational study that enrolled 31 stable asymptomatic PD patients. OH was identified as a predictor of higher LV mass index (LVMI) and, on multivariate analysis, was the main predictor of LVH in these patients [167].

Arterial stiffness is best described by the viscoelastic property of the arterial wall that represents the relationship between pressure response and changes in volume [168]. The principal factors that alter the arterial wall are classified as stable factors (e.g. atherosclerosis) and dynamic hemodynamic changes (e.g. BP) [168].

In ESRD patients, the aorta, the common carotid artery and large arteries are enlarged in comparison to age-, sex- and BP-matched control subjects [169]. GusbethTatomir – Covic et al. found an association between VO (a common status in ESRD patients) and arterial stiffness [170], results that were confirmed by similar findings [65, 171]. Furthermore, an improvement in arterial stiffness was associated with a better survival in dialysis patients [172]. In these patients, chronic VO produces the proper conditions for arterial remodeling as a response to an increased wall stress, determining a thickening of the arterial intima-media [173, 174]. The increase in the arterial intima-media thickness (IMT) will further lead to a decreased arterial distension, an increased pulse wave velocity (PWV) and an early return of wave reflection [173, 175]. LVH in ESRD patients develops as a combined consequence of both non-hemodynamic uremic factors and of hemodynamic overload, the latter being the effect of flow and pressure overload [173]. Flow overload determines the enlargement of LV and it is caused by anemia, arteriovenous fistulas and chronic VO; pressure overload is associated with structural and functional modifications of large arteries including stiffening of the arterial tree. At the same time, the arterial tree is also suffering remodeling from the flow overload burden. It is evident that cardiac and arterial changes are the consequence of the same hemodynamic abnormalities [173].

Regarding the decrease/amelioration of arterial stiffness by improving VO with HD, the literature provides conflicting data. Ie et al. combined carotid pulse contour analysis with aortic outflow measurement and found that aortic compliance was improved by HD fluid removal in patients with already achieved DW and steady BP although PWV did not differ after the HD session, suggesting that arterial stiffness in HD patients could be explained by a reversal reduction of aortic compliance due to volume expansion [176]. Another study measured PWV in 19 HD patients before and 24 h after they had achieved post-HD DW. The results failed to associate an improvement in PWV after volume reduction; this correlation was highlighted only after adding angiotensin converting enzyme inhibitors (ACEI) treatment for 1 week in a subgroup of ten patients, suggesting that the contribution of angiotensin II to arterial stiffness [177]. A more recent study demonstrated, in the same clinical settings, that HD significantly reduced PWV from pre-HD to post-HD [178]. These later findings were in addition supported by a study conducted by the same group that evaluated the long-term association of VO with arterial stiffness. PWV was measured at baseline, at the end of the intervention period (2.5 years) and at the end of the study (3.5 years) and was analyzed in correlation with VO status assessed by clinical methods and BIA. The results showed significant difference in PWV between the groups in relationship to VO, with PWV decrease in the BIA-managed group and a significant increase in the clinical-managed group, underlying the advantage of an adequate fluid status management in arterial stiffness improvement [109].

The newest information is provided by a very recent study that investigated for the first time the effect of different HD techniques on arterial stiffness. The authors aimed for a comparative evaluation between hemodialysis and hemodiafiltration on PWV, wave reflections, and central hemodynamic parameters. The results showed a decrease in the augmentation index (AIx) by both methods. Also, the aortic and brachial PWV were not modified by any of the two HD methods [179].

The presence of OSA in ESRD patients was associated with hypertension, LVH and increased mortality [180–182]. OSA presents a higher prevalence in patients that have VO status, such as CHF and ESRD in comparison to the general population [183–185]. The hypothesized mechanism is that the fluid overload status, so common in these patients, could increase the amount of fluid displaced from the legs into the neck during nighttime, leading to a compression of the upper airways. Indeed, it has been demonstrated that OSA in ESRD is ameliorated with conversion from conventional to nocturnal HD or conversion from conventional ambulatory PD to nocturnal- cycler-assisted PD [186, 187].

The proposed mechanism is supported by data from several studies. In a cohort of 26 ESRD patients studied throughout polysomnography, with BIA and neck circumference measurements it was found that nocturnal rostral fluid shift is indeed associated with OSA [188]. Another cohort of 20 patients was evaluated with polysomnography, leg fluid volume measurement and magnetic resonance imaging (MRI) of the upper airway. It was found that VO, through increased internal jugular vein volumes and increased amount of mucosal water is a contributing factor in the pathogenesis of OSA in ESRD patients [189].

The dialysate Na (DNa) prescription is a modifiable, but underused parameter for achieving Na balance in HD patients. In HD, Na removal is achieved by convection or diffusion; under the current practice, approx. 80 % of Na is removed by convection and 20 % by diffusion [236, 237]. Hypothetically, a regular removal of 1 L of plasma water by UF, when a theoretically isotonic Na concentration of 140 mmol/L is considered in the ultrafiltrate, could remove 140 mmol of Na (8 g NaCl consumption for each interdialytic day) [236]. Diffusive Na losses take place across the dialyzer membrane in line with the diffusion gradient between plasma and dialysate, whereas the convective Na losses represents the quantity of Na removed by UF and it depends on the prescribed UFR [237]. Currently, the regular dialysate prescription is based on an isotonic Na level of 135–145 mmol/L [236]. Experiments for a better fluid control in HD patients, by altering Na prescription to a higher value of DNa concentration (>141 mmol/L) or to a lower value of DNa concentration (<137 mmol/L) have been conducted in various forms and protocols. Higher DNa prescription determines a rapid UF and the removal of fluid excess from the interdialyitic period, but, as highlighted by Davenport et al., this is not as adequate method to achieve Na balance, as is leading to hypernatremia, increased thirst and higher IDWG, higher pre- and post-dialysis SBP and an intensified antihypertensive medication [237, 238]. On the other hand, using a low DNa concentration will lead to a rapid decrease in plasma osmolality and a hemodynamic instability and intradialytic hypotension [239, 240]. Various studies advocate for the benefits of using a lower DNa prescription, whereas others emphasize the ‘dark sides’ of the complications that arise with this method.

Na profiling is a HD method developed to avoid hypotension-related discomfort during HD (IHD) by modulating DNa prescription according to pre-established profiles; it was designed to preserve the advantages of a high DNa prescription without its intradialyitic complications [241]. It usually represents the prescription of a high DNa at the beginning of the HD session in order to determinate a hyponatremic status that will facilitate the water shift from the ICV to ECV compartment; during the HD session, DNa will be progressively decreased to avoid VO [242]. Time-averaged mean of DNa is associated with intradialytic diffusion Na overload and is generally higher with Na profiling than with a fixed DNa concentration [241, 243].

UF profiling is another way to prevent IHD and it represents the intermittent interruption or progressive reduction of UFR in order to promote plasma refilling. Therefore, during the high DNa period, a high UFR could provide an increase in plasma tonicity and a higher plasma refilling; during the low DNa period, a low UFR could avoid hypovolemia [241].

Song et al. conducted a prospective study to determine the optimal Na balance by Na profiling for prevention of IHD and to investigate if such an optimal Na profiling could result in HD sessions without Na gain-related complications. In 11 HD patients, 8 treatment modalities were evaluated: conventional HD (CHD) – control group, Na balance-positive step-down profiling HD (PS), Na balance-neutral step-down profiling HD (NS), Na balance-neutral alternating Na profiling HD (NA), UF only and PS + UF, NS + UF, NA + UF. Only PS, PS + UF, NS + UF and NA + UF had a significant positive impact; these were further analyzed in a phase II, randomized controlled, 6 weeks, crossover study [241]. The results were: (a) diffusive Na gain was significantly higher with both PS and PS + UF; (b) PS and PS + UF increased IDWG; (c) PS and PS + UF increased pre-dialysis weight and the amount of UF; (d) %UF achieved targetingDW was higher with NS + UF and NA + UF; (e) post-dialysis weight as closest to DW was achieved with NS + UF and NA + UF; (f) incidence of excessive weight gain and IHD increased significantly with PS. The investigators concluded that NS + UF and NA + UF are associated with less Na and weight gain, better UF performance, fewer IHD, a hemodynamic benefit and post-dialysis weight closest to DW [241]. DOPPS, an international database of dialysis patients was analyzed in an observational study by Hecking et al. respective to SNa and DNa [244]. The study found: (a) a higher DNa is associated with lower risk of hospitalization, even though it was, also, associated with higher IDWG; (b) DNa was not associated with mortality, but neither with lower SBP values [244]. The important disadvantages of inadequate Na profiling and the unconfirmed ‘benefits’ of this method are detailed elsewhere [245, 246].

Na individualization was designed as an improved alternative for the adverse effects of Na profiling. It has been hypothesized that humans have an individual Na ‘set-point’, therefore, whenever there is a Na excess, higher than the fluid intake, SNa will increase above the set-point leading to a further fluid intake [69, 247], especially in HD patients where pre-dialysis SNa tends to remain stable over time [248]. Accordingly, if a patient has a positive Na balance during HD (dialysis against a higher DNa prescription) that individual’s thirst will be stimulated and he/she will drink the amount of water necessary to drive the osmolality back to its set-point [237]. As a general concern, lower DNa is considered a potential determinant of SNa [249, 250]. Data, associating a low DNa with mortality, was previously reported, so lowering DNa for an alignment of the DNa with pre-dialysis SNa has to be pursued with caution [249, 251–253]. Also, Na individualization was associated with better surrogate and hard end points, supporting the lowering/tailoring of DNa in line with the SNaset-point, rather than using a fixed DNa concentration or DNa modelling [103, 249].

Na individualization method is well described by Arramreddy et al.’s case series of 13 patients that were switched from CHD to DNa individualization prescription. The strategy began with a phase A-stepwise weekly reduction of the standard DNa (140 Meq/L) by 2–3 mEq/L until reaching a Na gradient of −2 mEq/L. The tapering phase lasted 4 weeks. The Na gradient was defined as difference between prescribed DNa and average SNa and average SNa was calculated as mean SNa over the preceding 3 months. During the intervention period, the DNa was not further readjusted in line to monthly SNa levels. Further, on the same day that patients reached a −2 mEq/L Na gradient, they entered a 4 week-phase B-period during which all patients were dialyzed with a Na gradient of -2 mEq/L. Baseline pre-HD SNa was 135.3 ± 3.7 mEq/L; in phase A, post-HD SNa increased by 1.5 ± 3.7 mEq/L compared to pre-HD SNa. HD with an individually reduced DNa, achieving a gradient of −2 mEq/L, in phase B did not modify pre-HD SNa. Compared to standard DNa, individualized DNa reduced IDWG without significant increase in the frequency of intradialytic hypotension or cramps [254].

Hecking et al. performed two separate investigations on DOPPS trial results [244, 255]. In the first one, the investigators found an association between SNa and DNa and that this correlation influenced the mortality: DNa >140 mEq/L was associated with a trend towards higher mortality in patients with SNa ≥140 mEq/L; higher DNa was associated with lower mortality in patients with SNa <137 mEq/L [249, 255]. In the second study, the analysis revealed: (a) a large variability of DNa prescription between centers: 55 % used fixed DNa and 44 % used DNa individualization; (b) in non-individualized DNa facilities, higher DNa was associated with higher SBP, but with a lower mortality; (c) in individualized DNa facilities, higher DNa was associated with lower SBP, but with higher mortality [247, 249]. Nevertheless, it should be noted that patients with higher non-individualized DNa were younger, lees likely to have DM and had lower serum albumin. Also, as in the non-individualized centers, DNa was not prescribed in line with patient’ characteristics, there is possibility for confounding by indication [249]. The authors concluded that there is rather limited support for lowering DNa to control HTN, especially in the view of this method’s association with hospitalization risk and mortality [244]. A higher DNa could be beneficial in some patients but detrimental in others [249]. De Paula et al.’s study, on 37 HD patients that received 3 weeks of HD with DNa at 138 mEq/L and 3 weeks of HD with DNa individualized according to patient’s average pre-dialysis SNa, found that DNa individualization lead to lower SBP in patients with uncontrolled HTN, but with no impact on BP in those with controlled BP (<150/85 mmHg) [256].

Sayarlioglu et al. showed that DNa individualization is associated with better outcomes in 18 HD patients that underwent HD with DNa of 135 mEq/L when pre-HD SNa was <137 mEq/L and HD with DNa of 137 mEq/l when pre-HD SNa was >137 mEq/L. After a follow-up period of 8 weeks, lowered DNa was associated with lower pre- and post-HD SBP, lower pre-HD DBP, lower IDWG and improved cardiovascular parameters (reduction in LV systolic diameter, tricuspid regurgitation, IVC diameter and pulmonary arterial pressure) [257]. Actually, this is the first of the only two studies that analyzed the relationship between a lower DNa and cardiovascular parameters. The second study was performed by Kutlugun et al., in 30 HD patients, and showed that endothelial dysfunction, measured by flow-mediated dilatation (FMD), was significantly higher after using low DNa prescription (137 mEq/L) than after assigned standard DNa (143 mEq/L). Lower DNa resulted in a better BP and IDWG control, but no improvement in the LVM [258].

The SOLID trial is a multi-center, prospective, randomized, single-blind, controlled parallel 3 years clinical trial designed to investigate the impact of a low DNa on cardiovascular risk in HD patients. The study protocol was initiated after a previous investigation performed by the same group that showed a decrease by 4–5 mmHg in SBP and 2–3 mmHg in DBP with a 3 mmol decrease in DNa [259]. The present study intends to enroll 118 home HD patients from 6 sites for a follow-up period of 12 months; during this time the intervention and control groups will undergo HD with DNa of 135 and 140 mmol respectively. The primary outcome will be LVMI changes with other approximately 11 more secondary outcomes [240].

The virtues of more frequent dialysis on improved BP control, and even on survival are supported by results of small, observational cohort studies, as well as larger, RCTs. Prolonged dialysis schedules represent a more physiologic solute and fluid removal (a gradual UF) that can be translated into an improvement of dialysis adequacy. All of these HD methods can be performed in different locations: in-center, home-assisted, home and include: Short Daily HD (SDHD) – less than 3 h/session, 6 times/week; Long Daily HD (LDHD) – >5 h/session, 6 times/week; Nocturnal HD (NHD) – 6 to 8 h/session, 3–5 or 6 times/week.

In 1983, Charra and the Tassin group were the first to report the benefits of 3 times/week, 8 h/session, diurnal and nocturnal HD, along with reduction of salt intake on volume and BP control and survival [260–262]. Consistent with their results, similar findings were further provided by different studies using the same strategy. Beginning with 1998, studies performed in patients receiving SDHD or NHD reported improved BP, LVH, phosphorus and hemoglobin levels, decreased risk of hospitalization and even a better survival [250, 263–273].

A major concern with all of these techniques remains whether they are helpful for preservation of residual renal function (RRF). RRF is increasingly becoming an important parameter for nephrologists and one of the major goals in the management of RRT patients. Jansen et al. showed in 279 incident HD patients and 243 PD patients that predictors for RRF decline are elevated DBP and higher proteinuria, dialysis hypotension and dehydration [287]. It has been also demonstrated that PD patients present, in general, a slower decline of RRF than patients on HD, maybe due to a more ‘natural’ fluid removal and fewer hypotension episodes [287–290]. It was found, in PD patients, that for every 1 ml/min increase in RRF there is a 50 % decrease in the risk of death; also, an analysis of the CANUSA study showed that a 250 ml increase in daily diuresis in associated with a 36 % reduction in mortality, suggesting, by expressing the RRF as urinary output, that a fluid control is more important than solute clearance [291–294]. Preserving RRF is associated with better fluid control and a more liberal fluid intake, maintained endocrine function, secretion of organic acids and a better dialysis adequacy [287]. RRF loss was associated with mortality, LVH, arterial stiffness and other ESRD complications [291, 293, 295–299].

An optimal management for preservation of RRF has not been yet determined since renal function decline appears both in dehydration and overhydration. Thus, some nephrologists are in favor of maintaining patients ‘wet’, whereas others prefer the patients ‘dry’. Hypovolemia and hypotension are known factors to induce loss of RRF, especially in PD patients [295, 296, 300], but so are overhydration-associated HTN and LVH [287]. Gunal et al. showed that controlling fluid status, by salt restriction and increased UFR, led to 2.8 kg weight loss and a 28 % reduction in urine output [301]. McCafferty et al. failed to demonstrate in PD patients that a better preservation of RRF is achieved with overhydration status [293].

As mentioned before, PD patients seem to preserve RRF for a longer time than HD patients, implying therefore that loss of RRF could be also determined by HD technique-related factors. Daugirdas et al. conducted recently an analysis of FHN studies to investigate if frequent HD could result in a more rapid decline of RRF, compared to CHD. They found that in both of FHN trials, baseline RRF was inversely associated with dialysis vintage. Frequent HD participants from these trials, with a percentage of RRF at baseline, suffered a significant progressive loss of RRF at the end of follow-up period: at 12 months, in the Nocturnal FHN, 67 % of patients had zero urine output in comparison to 36 % in CHD group and in Daily FHN, the method of treatment did not had a significant impact. The authors concluded that frequent NHD appearsto promote, by undetermined mechanisms, a more rapid loss of RRF [302].