Studies of extended duration of haemodialysis (beyond the conventional 3 to 5 hours three times a week) generally support improved BP control and or reduced BP medication burden, with inconsistent findings on left ventricular mass measurements (67-70).
More recently, however, extended duration HD has been shown to be associated with lower mortality. This is true whether the haemodialysis session is extended, or the frequency of haemodialysis is increased (71, 72). The Frequent Hemodialysis Network (FHN) Daily Trial reported that 2 months of a frequent haemodialysis regimen lowered pre-dialysis SBP by 7.7 mm Hg (95% CI: −11.9 to −3.5) and DBP by 3.9 mm Hg (95% CI −6.5 to −1.3) (73) whilst the FHN Nocturnal Trial reported a reduction in systolic SBP from baseline of 7.9 ± 18.4 mm Hg in the nocturnal cohort at 12 months (74). Short daily and nocturnal schedules also reduce the per-session probability of IDH by between 20 and 68%, while wellbeing and shorter recovery times are observed in those having extended hours HD, possibly mediated through a combination of optimizing sodium and volume status or enhanced solute clearance (70, 74).
Whilst there is good evidence to support extending haemodialysis for improved BP control, either by longer sessional hours or increased frequency of sessions, universal adoption of practice is unlikely due to the high cost of providing this service. The NightLife study is ongoing and will evaluate the cost and clinical effectiveness, including effect on BP, of thrice-weekly extended in-centre nocturnal haemodialysis versus daytime haemodialysis (75).
It has also been reported that HDF, compared to haemodialysis, improves cardiovascular stability during treatment sessions and that it reduces the frequency of IDH (76-78). These findings, however, are not consistent across all studies. Two meta-analyses on this topic have reached differing conclusions on the intervention effect on IDH, but neither study reports a significant effect on BP outcomes (79, 80). These inconsistent findings may be explained by failure to achieve sufficient convective volumes, different haemodialysis modalities (low-flux vs high-flux) and HDF techniques, participant demographics and access type. With the view that there was still clinical equipoise as to the potential benefits of high dose HDF compared to conventional high-flux haemodialysis, the CONVINCE and H4RT trials were conceived to address the issue. CONVINCE has recently reported on the primary outcome; a reduction in death from any cause in the HDF group (hazard ratio, 0.77; 95% CI, 0.65 to 0.93), although risk of death from CVD was similar in HDF and high-flux haemodialysis treatment arms (81). At present, there remains insufficient evidence to recommend the widespread implementation of HDF on the grounds of cardiovascular benefit. H4RT trial outcome data is awaited (82).
Active BP lowering in haemodialysis patients can lead to increased frequency of IDH. Large observational studies have demonstrated an association between frequency of IDH and mortality (83-85). It is therefore prudent that steps are taken to minimize IDH especially in IDH prone patients. Comparative studies have shown that haemodialysis tolerance, which includes IDH, is better when using bicarbonate rather than acetate dialysate (70, 71). However, we note that in an RCT (85) there appeared to be a dissociation between IDH reduction (reduced with temperature lowering) and mortality (no effect).
Different strategies for changing dialysate temperature have been studied including biofeedback and different degrees of temperature lowering. A systematic review (86) of 11 RCTs concluded that lowering dialysate temperature reduced the rate of IDH by 70% (95% CI 49% to 89%). However, more recent studies including a multi-centre RCT enrolling 73 patients for 12 months (87) and a cluster randomised study (88) did not demonstrate an advantage of setting dialysate temperature -0.5oC below body temperature vs. 37oC. On other hand, a recent large observational study using the Fresenius NephroCare Eclid database (85) did show that in case-mix, facility-level adjusted incident haemodialysis patients, a 0.5oC reduction in dialysate temperature was associated with a 33% risk reduction of IDH. It is possible that patient selection (patients at high risk of IDH (85) vs. patients at lower risk of IDH (87, 88)) accounts for the different outcomes.
Results of studies examining effect of changing dialysate sodium concentration on BP and IDH have been conflicting and covered in recent systematic reviews (89, 90). Whilst studies have shown that lowering dialysate sodium reduces pre-dialysis mean arterial pressure (MAP) by 3.6 mmHg (95% CI -5.7 to -1.7) and post-dialysis MAP by 3.3 mmHg (95% CI -1.7 to -4.8) (91-94), there is also weak evidence that lowering dialysate sodium may increase rate of IDH in IDH-prone patients (91, 92). The effect of varying dialysate sodium (sodium profiling) during a dialysis session has also been examined. Different profiles of sodium reduction were the subject of a meta-analysis (95). Stepwise rather than linear reduction of sodium concentrations was found to reduce IDH. A single centre RCT (96) also confirmed that profiling reduced IDH although, contrary to the meta-analysis, the authors reduced dialysate sodium concentrations linearly. This study demonstrated similar benefit on IDH by dialysate cooling but the benefits were not additive. Given the theoretical risk of “sodium loading” leading to increased dialytic weight gain, we do not feel there is enough evidence to recommend sodium profiling. This was decision has been supported by DOPPS cohort study showing that routine use of sodium profiling was associated with higher all-cause mortality (31). Sodium profiling may have a role in IDH prone patients but there is insufficient evidence to recommend as routine care.
The correction of chronic metabolic acidosis is an important goal of dialysis and dialysate bicarbonate should be optimised for mineral bone disorder and nutrition. However, acid-base balance might have an acute effect on BP and therefore IDH. We acknowledge that DOPPS (97) have suggested a reduction of the dialysate bicarbonate concentration can be considered for patients with significant peri- and post-dialysis alkalaemia with frequent IDH unresponsive to classical management protocols. However, we considered that the evidence based on three studies was weak (98-100).
There are few studies examining the effect of adjusting dialysate magnesium (101, 102) on BP and IDH. A single study of 14 patients (101) suggests that low dialysate concentration of magnesium (0.25mmol/L) be avoided with dialysate calcium concentration of 1.25mmol/L due to increased risk of IDH. We considered that the evidence base too weak to make a recommendation. A study has examined the effect of dialysate calcium concentration on IDH risk (99) but we suggest that dialysate calcium should be adjusted in accordance for optimal management of mineral bone disorder.
There is evidence that long intraperitoneal dwells of 7.5% icodextrin can achieve greater ultrafiltration than 2.27% glucose dwells (103). Sodium removal in peritoneal dialysis is dependent on convection so it is to be expected that using icodextrin can improve fluid status (104) and hence help to attain dry weight in people on peritoneal dialysis.