• 1.

  To assess the effect of Qd (500, 600, and 700mL/min) in Kt and VI in OL-HDF.

• 2.

  To quantify the amount of water that can be saved.

This is a prospective cross-over study conducted in a single dialysis facility. Subjects had to be older than 18 years and on OL-HDF for more than 3 months.

Demographics data included: sex, age, time in HD, and aetiology of renal disease. The Kt objective (Ktobj) was individually adjusted for BSA.

Thirty-seven patients were enrolled (16 women and 21 men). Monitors and dialysers were the same throughout the study. A total of 16 patients were dialysed with AK 200® and 21 with Fresenius 5008®. The membranes used were distributed as follows: 20 FX800® and 17 Polyflux 210H®. All patients underwent OL-HDF as shown below:

• -

  6 sessions at Qd of 500mL/min.

• -

  6 sessions at Qd of 600mL/min.

• -

  6 sessions at Qd of 700mL/min.

Dialysis time, anticoagulation, and blood flow rate were unchanged throughout the entire study period. Ultrafiltration volumes were adjusted according to individual clinical requirements. When recirculation is measured by 5008®, there is a temporal increase of Qd to 800mL/min.; recirculation was routinely measured once in the session.

The following parameters were directly recorded from the monitor display: effective blood flow rate (Qbe), Qd, effective time in dialysis, final Kt (Ktf), and final VI.

Based on the collected data the following parameters were calculated:

• 1.

  Mean individual values of Qbe, effective time in dialysis, VI, and Ktf at each Qd.

• 2.

  Kt differences: Ktobj−Ktf.

• 3.

  VI differences (target=24L): VIobj−final VI.

Qualitative variables are expressed as percentages and quantitative variables as means (standard deviation) or medians (minimum−maximum). Quantitative variables were compared by Paired t tests and ANOVA while. Chi-square tests were used to compare qualitative variables. A p value below 0.05 was considered statistically significant.

Analyses were conducted using the software SPSS version 15.0.

The median age of the 37 patients enrolled was 67.4 years (36–92). The aetiology of CKD was: diabetes mellitus in 13, glomerular disease in 7, unknown in 6, vascular origin in 5, interstitial disease in 4, and polycystic kidney disease in 2. Dialysis was administered 3 times a week in 36 patients and twice a week in one patient because this patient had good residual renal function: mean 24h urea and creatinine clearance, greater that 5mL/min. Dialysis duration was programmed at 240min in 7 patients, 255min in 25 patients, 270min in 4 patients and 300min in one patient. A fistula was used for dialysis in thirty patients, while the remaining patients used a tunnelled catheter.

Data from a total of 565 sessions were collected: 192 at 500mL/min, 194 at 600mL/min and 179 at 700mL/min. Sessions with a deviation in prescribed treatment time, Qb or sessions with no measurements of K available due to technical problems were excluded. All patients had undergone at least 3 sessions with each Qd.

The main finding of our study is the slight improvement in Kt when increasing Qd in OL-HDF, with practically no effect in VI. Even though this difference is statistically significant, the clinical relevance of these differences is questionable, even more when considering that amount of water consumption that is necessary to achieve this modest result.

There is almost no information available on the use of a specific Qd in OL-HDF. A Qd of 700mL/min is used in most facilities for increasing the efficacy of diffusive transport, but there is no rationale for its use. In fact, only one publication has addressed the effect of Qd in OL-HDF, comparing HD to OL-HDF by using AutoFlow (AF), a built-in system in the 5008® monitors where Qd is adjusted to Qb. The authors conclude that OL-HDF may obtain a higher Kt/V using less dialysate than in HD.20 But we have not found reports comparing the effect of the various Qds on efficacy or VI in OL-HDF, which makes our study original, and practical. Just as in HD, our results prove that increasing Qd has a minimal effect on the efficacy of HD, which is even more relevant considering that it is not based on Kt/V calculated from pre- and post-dialysis urea levels or the V entered to the monitor. Our measure of dialysis adecuacy is based on the BSA-adjusted Kt, which is a rigorous assesssement.21 The target was achieved in all except one patient. In this patient which had vascular access problem the increase in Qd may not seem to be a useful strategy for improvement of dialysis adecuacy. This result shows that classical indications of elevated Qd in OL-HDF to increase efficacy should be disregarded, even more considering that decreased mortality resulting from this technique is associated with the amount of convective volume administered. It could be argued that the effective time was lower with a Qb=700mL/min, and the achieved Kt would be greater. As extending time is known to be associated with improved survival irrespective of dialysis dose,22 the effect of time loss potentially due to internal machine controls vs. improved Kt should be balanced, as both differences are not significant and possibly irrelevant, and that is why we believe that using a Qd above 500mL/min is not an effective strategy.

As far as VI, Qd has almost no effect as would have been anticipated. Automated OL-HDF techniques currently look for a high-performance convective transport, but each monitor has different control systems and dissimilar results. In fact, VI is “regulated” based on methods that nephrologists may often be unaware of. In any case, elevated Qd has clearly no effect or utility to change VI in OL-HDF. After evaluation of the predetermined objectives it was observed that only one patient was far from reaching the target 24L (achieving 20L) because of problem a poor Qb when undergoing dialysis using a tunnelled catheter. Interestingly, this patient had a bilateral supracondylar amputation. At present no BSA-adjusted VI can be prescribed as in Kt. There is a need to determine what is the VI target for each patient and to develop a fully individualised OL-HDF. For patients who were close but did not achieve the target, this was not dependant of the Qd used. These were patients whose VI was very close to 24L, and therefore the target could possibly be achieved irrespective of Qd by increasing Qb or time, or if other methods were used for enhanced performance. Lastly, many papers include VI and ultrafiltration volume in convective volume. We have not used the ultrafiltration volume, which would have helped achieve the target of 24L.

We did not compare Kt or VI resulting from the use of different monitors, as this was not the purpose of our study and it was not designed for this. We believe that the differences listed in Table 2 in terms of Kt are consistent with the results reported by Maduell et al., who showed that ionic dyalisance calculated by Fresenius monitors was higher as compared to the AK200® monitor.23 Yet, the higher VI achieved by the Ultracontrol® method is consistent with our results previously published.24

Drinking water for human use is not suitable for production of dialysate it has to be purified. For OL-HDF, the water and the dialysate have to be ultrapure independently of the amount required. Treating a patient for several hours 3 times a week requires large amounts of water plus consumption of energy and the generate undesired throwaway medical products. Mean environmental balance per HD session is estimated at 400–500L of water, 10kW/h of electricity, and up to 3kg of clinical disposables.25 As a result, in terms of dialysate manufacturing ultrapure dialysate results in economical and ecological problems. We have addressed the impact of large amounts of dialysate; but, particular attention should be paid to the fact that in order to produce 1L of dialysate another litre was lost in the treatment process; i.e., 2L of water are required for the generation of 1L of DF by reverse osmosis, thereby the consumption is double. Water is essential for life and its management is completely necessary as part of good use of natural resources; awareness of environmental issues should be raised among personnel in dialysis facilities. In a 255min session, decreasing the Qd from 700 to 500mL/min and saving 51L of dialysate in each patient may appear irrelevant however more than 500,000L of dialysate are required during one year in a dialysis facility with 75 patients or, as explained before, this equal to 1,000,000L of drinking water. In addition, by maintaining the Qd there is a significant amount of acid is saved with the consequent financial and ecological advantage. IN the present study we have analysed the cost, of 1L of dialysate including water and acid, ignoring bicarbonate and assuming that one cartridge is used per session with independently of the Qd used Even though the cost of one litre seems unimportant at first glance, we stress the relevance of n: annual consumption. Also, in our calculations water used for preparation or disinfestation has not been included in our calculations (this is why savings could be increased with optimal use), or other maintenance expenses from the water treatment plant, which may partially be affected by increased water consumption.

Our study's main limitation is that the sample size is small, but the number of dialysis sessions is enough and the cross-over design allows for comparing each patient with themselves, which increases the power of results. Our study had been conducted in a single dialysis facility, but our results may be extrapolated to all facilities working under similar conditions.

Our data show that increasing Qd above 500mL/min in OL-HDF provides a limited advantage. Improving the efficacy of dialysis by saving water is necessary for our environment and to meet the current and future human demand for water, thereby achieving a much more effective dialysis. Further studies should be conducted to determine whether or not using lower Qds is feasible.

Dr Pérez García, Dr de Sequera, and Dr M. Albalate have been involved in meetings with Fresenius and Gambro.