Hemodialysis prescription

2.3.1 Dialysate composition and temperature

The concentrations of components of the dialysate, e.g. sodium, chloride, bicarbonate, and calcium, are an essential part of the prescription, but do not affect the removal of uremic toxins. Temperature likewise has nothing to do with the dose, but lowering it may improve the hemodynamic stability. The positive effects of HDF have sometimes been explained by that mechanism [Daugirdas 2015].

2.3.2 Filter and convection technique

Selection of the dialyzer determines the filtering characteristics and maximum instantaneous clearance (mass transfer area coefficient K0A). Many modern dialyzers combine biocompatibility, high efficiency (high K0A), high flux (high permeability to water), and large pore size (high permeability to large molecules permitting high clearance, especially with enhanced convection techniques). HDF combines high clearance of small molecules by diffusion and high clearance of big

molecules by convection. There are also differences in the adsorptive capacity of the membranes [Aucella et al. 2013, Perego 2013].

In the HEMO trial [Eknoyan et al. 2002], high flux correlated with better outcome in patients with long dialysis history, with probably negligible RRF. Later observational studies ([Canaud et al. 2006a, Canaud et al. 2015, Davenport et al.

2015]) and RCTs ([Santoro et al. 2008, Schiffl 2007], MPO [Locatelli et al. 2009], CONTRAST [Grooteman et al. 2012], Turkish [Ok et al. 2013] and ESHOL [Maduell et al. 2013a]) have also reported positive effects of convection. In MPO, only patients with S-Alb < 40 g/L benefited significantly from high-flux dialysis [Locatelli et al. 2009]. Mortality reduction was significant only in high-volume HDF (filtration >23 L, ESHOL trial) and HF, but not in high-flux HD and low-volume HDF. In high-flux HD, the convective transport is <10 L/session [Mostovaya et al. 2014].

In one heavily criticized meta-analysis [Rabindranath et al. 2005] survival was better with HD than with HDF. In one of the recent four meta-analyses HDF was significantly better [Mostovaya et al. 2014], in three the benefit remained unproven [Nistor et al. 2014, Susantitaphong et al. 2013, Wang et al. 2014]. [Locatelli et al.

2015] interpret these results as inconclusive. [Mostovaya et al. 2015] come to a different conclusion. In their opinion, high-volume (>20 L) online post-dilution HDF is an effective therapy and the convection volume is the most practical measure of the dose of HDF. EBPG have recently been updated to favor high-flux membranes [Tattersall et al. 2010], but probably their use is beneficial only in high-volume HDF. Obviously the added value of increased convection is rather small because it has been so difficult to demonstrate.

2.3.3 Blood, dialysate, and filtrate flow

Michaels’ equation [Daugirdas and Van Stone 2001, Ward et al. 2011] (36 on page 93) describes the dependence of dialyzer diffusive clearance (Kd) on blood and dialysate flow (Qb and Qd) and dialyzer K0A. With low permeability (low K0A, middle molecules) the effect of flows on clearance is small. Convective transport does not depend on dialysate flow. In RCTs a significant reduction in mortality has been achieved by HDF only with high filtration volumes (>23 L/session) [Maduell et al. 2013a]. Dialysate and replacement fluid cost, blood is free, but the access may restrict its flow. For optimal urea removal blood and dialysate flows should be in balance (1:1.5-2). In lengthy treatment sessions lower flows may be sufficient, but

dialysate consumption is still substantial. With modern dialyzers the effect of Qd on Kdurea may differ from that calculated by Michaels’ equation [Hauk et al. 2000, Leypoldt et al. 1997, Ward et al. 2011]. Ultrapure water is a prerequisite for convective techniques. In post-dilution HDF, high filtrate flow also requires high blood flow (Qb).

2.3.4 Duration and frequency

Treatment duration, frequency, and symmetry of the schedule are essential elements of the prescription and have a decisive effect on solute removal, survival, and HRQOL (sections 2.6.2 and 2.6.3). The interval between treatments is also important. The odd number of days in a week causes difficulties for hemodialysis patients in the conventional 3 x/week schedule. Mortality is highest on Mondays and Tuesdays [Bleyer et al. 1999, Foley et al. 2011, Fotheringham et al. 2015].

Splitting the weekly treatment time into smaller fractions corrects this problem and lowers concentrations, but may increase blood access complications [Jun et al.

2013, The FHN Trial Group 2010, Weinhandl et al. 2015].

2.3.5 Water removal

Fluid accumulation is a common problem in dialysis patients. Technically its removal is simple, but the patients do not always tolerate it. Antihypertensive medication may exacerbate blood pressure drops during dialysis and hamper ultrafiltration leading to volume load, blood pressure rise, and a vicious circle.

Water removal is not addressed in this thesis.

2.3.6 Prescribed and delivered dose

Dialyzer clearance multiplied by treatment time (Kt) is based on dialyzer characteristics, Qb, Qd, and td. It is a patient-independent, external measure of delivered dialysis dose and often scaled to patient urea distribution volume V (Kt/V). Traditionally the delivered dose has been estimated from its effects on the patient’s blood urea concentrations because measurement of true Kd was difficult before the advent of ionic dialysance monitoring (IDM) devices. Modeled Kt/V is

insensitive to errors of Kd, but V is not an ideal scaling factor (sections 2.6.4, 6.1 and 6.3).

In the USA the single pool Kt/Vurea is the primary measure of dialysis dose [National Kidney Foundation 2015], obviously because it can be easily prescribed.

Even some observational registry studies are based on the prescribed dose reflecting the intention-to-treat concept.

The KDOQI guidelines recommend a higher prescribed target Kt/V (1.4) to guarantee a minimum delivered Kt/V (1.2) for all [National Kidney Foundation 2015]. In Europe, Qb, Qd, and td are often corrected arbitrarily if the delivered dose target has not been achieved. The resulting new prescribed Kt/V is ignored and forgotten and the focus is in the delivered dose.

If the Kt/V target is A (e.g. 1.2), then the required dialyzer clearance is

K = A * V / t (3)

and the required treatment time

t = A * V / K, (4)

where K is dialyzer clearance in mL/min, V is in mL and t in min.

V can be estimated from anthropometric equations [Hume and Weyers 1971, Watson et al. 1980] and K from Michaels’ equation or nomograms published by the dialyzer manufacturer.

We can choose the dialyzer, set K, t, and Qd and determine the required Qb from a nomogram. Or we can set Qb and Qd and calculate Kd from Michaels’

equation and td from Equation 4.

Online monitors based on ionic dialysance or UV light absorption help in delivering exactly the prescribed dose.

Taking into consideration the compartment effects and prescribing the required Qb, Qd, and td to achieve a continuous-equivalent clearance target is more complex and among the topics addressed in Study IV.