2002/3 EDTNA/ERCA Journal Club Discussion Summary 

The May 2002 EDTNA/ERCA Renal Care Journal Club discussed the paper entitled 'Dissociation between dialysis adequacy and Kt/V' by Prof. Dr. Raymond VANHOLDER et al. (Belgium).

This paper was discussed and circulated in a group of 482 JC Members within the world.
Anyone interested in Renal Care can simply become a JC Member by subscribing for free at the Journal Club section; or by sending me your email-address and full name at devosjy@skynet.be

The paper has been actively discussed between André STRAGIER (Belgium), Elizabeth LINDLEY (UK), Hans POLASCHEGG (Austria), Jean-Yves DE VOS (Belgium) and Prof.Dr.Ray VANHOLDER (Belgium).

DISCUSSION SUMMARY

André STRAGIER: I want to make a comment on the general accepted idea that in vivo dialysis efficacy, generally expressed as Kt/V, is a simple notion. My point is that some units deduct Kt/V from the urea clearance – in one way or another – whereas others do this from the PRU (Percentual Reduction of Urea). This is clearly not the same. We also see different formulas being used. All methods give a Kt/V result that is understood as a universal standardised unit like a centimetre or a gram. Unfortunately this is far from being the case. Today it is perfectly possible that a Kt/V of 1.2 in unit A represents a higher efficacy than a Kt/V of 1.4 of unit B ! In an in vitro Kt/V assessment the urea reduction ratio (URR) or the PRU is for 100% predictable. However when we apply this same model in vivo the results become amazingly less reliable. This reveals that a whole series of factors interfere in vivo that do not in vitro. For example four hours in vitro dialysis with a urea clearance of 150 ml/min is identical to three hours dialysis with a urea clearance of 200 ml/min. This has “contaminated” the opinion that a three hours in vivo dialysis with a Kt/V of 1.2 has the same quality for the patient than a four hours dialysis with an identical Kt/V of 1.2 ! The K of Kt/V in vivo can be directly assessed either on the blood side or on the dialysate side, what is quite different. Sometimes the dialyser prospectus urea clearance or the mass transfer coefficient (KoA) is used instead. In practice units use different simplified models to deduct Kt/V from the Pre- and Post urea values, what again is different. There exist a considerable method variation in the world what is troublesome. If all patients would have the same dialysis frequency and duration, the same Kt/V formula being used, the same dialysis membranes, etc… only than I would agree that Kt/V is a simple index of adequacy. Perhaps lack of Kt/V standardisation has in some way contributed to the dissociation between dialysis adequacy and Kt/V ?

Ray VANHOLDER: Available methods to appreciate Kt/V in a surrogate way, remain a surrogate, which means that correlation regression analysis gives a decent r-value, but in individual cases differences might be as much as 100%. So this looks correct statistically, but on individual basis there might be tremendous differences. Because of this reason, in the upcoming European Best Practice Guideline (EBPG) from ERA/EDTA, we give preference to aim at Kt/V or equivalent of 1.4 to be sure ( or almost ) that everybody gets more or less above 1.2 in reality.

Jean-Yves DE VOS: In medicine we generally evaluate the “therapy efficiency” by the clinical outcome. In dialysis we strive however to achieve a good Kt/V value, NOT the normalisation of the biochemistry ! I agree with André that when somebody measures Kt/V or creatinin clearance or URR or PRU, they measure dialyser performances BUT not necessarily reflection of good dialysis in the patient ! Distribution volumes are not that simple to be general for all males and females; duration is no problem as long as correctly registered; blood samples pre- and post dialysis taken may vary again from centre to centre policies, from nurse to nurse, from day to day, from sample processing to processing, etc… However, Kt/V is often classically linked to mortality, but there are some pitfalls. There is surely an influence of uremia status and/or Kt/V on mortality causes, but who can calculate this on a scientific basis? Mortality is as high during the weekend as during the five other days of the week together ! … so something is still wrong, even with high Kt/V values. Patients often die acutely from overhydration and/or hyperkalemia; not because of a “low” Kt/V. Kt/V has been used to provide a minimum of therapy in a minimum of time while patients – without exclusions of being old or young; Black or White or Asian or …, having only Renal Failure or Multiple Organ Failure; etc… - still stay often chronically undertreated suffering from lots of co-morbidity’s: anaemia, polyneuritis, restless legs, osteoporosis, pruritis, hyper-/hypotension, … a too long list ! Kt/V, a diagnostic tool, became a therapy evaluation tool and became wrongly the “Golden Standard”.

Liz LINDLEY: Imagine you have an ultra-efficient dialyser and a blood flow of five litres per minute. If you were to dialyse a person with this imaginary system, the urea level in the blood might be reduced to one third of the pre-dialysis level in a few minutes. This would give a PRU of over 65% but it would not mean that the patient has been well dialysed as very little urea would have been removed from the tissues and the blood level would soon rebound almost up to the pre-dialysis level. Theoretically a PRU of 65% can be achieved with any dialysis time from 5 minutes to 5 hours, but the amount of urea actually removed will be much higher if the time is longer. If you take urea generation into account, this effect is even greater. The more sophisticated urea kinetic models ( e.g. TATTERSALL, SMYE and DAUGIRDAS ) do make a correction for time. They do require a computer program and more information than the PRU – i.e. the dialysis time and pre- and post weights – but this information is routinely collected so there is no reason why these models should not be used. Meta-analysis of Kt/V data is impossible if the methods used to calculate Kt/V are different, so it is very important that centres providing this kind of data give full details of the model they use. Urea is not a good marker to use as it is removed just as easily with low flux or high flux dialysers. So it gives no indication of the membrane quality. The substances that are more clinically relevant are harder to measure or have distribution volumes that are more complex. Urea is okay for routine monitoring but ideally a periodic check should be made on the more relevant substances. I would be interested to know if we would choose for a four hour high convection HDF treatment three times a week OR two hours of standard HD each day ?

Jean-Yves DE VOS: We are in favour of installing more frequent dialysis instead of three times a week "High Performance” (?) treatments. Not how much therapy do you deliver is important but HOW do you deliver therapy !

Ray VANHOLDER: It is as if you make me choose between a nice meal and a good glass of wine; both are good things. Daily dialysis cuts off important shifts in toxin concentration, prevents rebound phenomenon, and is probably better for removal of protein bound compounds. HDF removes middle molecules, is biocompatible, and probably improves morbidity and mortality. So, why not daily HDF or slow long nocturnal HDF ?

André STRAGIER: I was reading the report on the HEMO study ( JAMA Vol 287, May 22-29, 2002 or http://jama.ama- assn.org/issues/current/full/jmn0522-3.html ) . The purpose of the HEMO study in the US was to determine whether the type of dialysis membrane used or the amount of dialysis (Kt/V) can influence well-being and survival. The results showed that mortality rates were basically the same for patients regardless of combinations of standard or high doses of haemodialysis and low and high dialysis membrane flux. My problem is that these results do not fit well with our general opinion.

Ray VANHOLDER: This is at least surprising but also intriguing news. If we are supposed now no more to take into account Kt/V, this really is shocking, but it might be conform with my hypothesis that it has something to do with potassium. May be the quality of dialysis today is so good that all levels of Kt/V imposed in the HEMO study corresponded to enough potassium removal so that the change between the highest and the lowest level had no impact. Perhaps we have here a bias due to body weight where heavy-set patients with low Kt/V due to high V survive long because of good nutritional status and vice versa for the thin malnourished ones. The HEMO study is on patients already present in the dialysis unit irrespective of whether they were treated since three months or ten years. We should look therefor at the full paper.

Hans POLASCHEGG: I completely agree with the title of the discussion. The preliminary data of the HEMO study shows the limitations of Kt/V. it is a single dimension equivalent to describing the three dimensional world by a single parameter: length. Obviously, to reach a point in space you have to move certain distances in three dimensions. What are the other dimensions in dialysis? Clinical research has focused on urea because of the requirement to define an easy to measure quality parameter for reimbursement in the US. Now Europe and Japan is following. Reimbursement changes in Germany and Japan require implementation of quality control. Prof. Vanholder mentioned one important parameter: potassium. While sodium variation or feedback control is widely popular and although it is claimed to have beneficial effects – a claim that I oppose – the few tests with variable dialysate potassium have shown more convincing and remarkable results. These effects are not simple water shift effects but seems to have a more fundamental basis. There is no evidence that the enhanced removal of water soluble middle molecules has any beneficial effect. Protein bound molecules seem plausible but the evidence is missing. We have to wait for suitable adsorbers and long term studies. Science is important for our understanding and future improvements but one must stop wasting time and money for the improvement of methods that have shown to be of limited use, such as urea kinetics. I suggest a paradigm change. After having met with “daily dialysis patients” I strongly advocate the freedom of choice for patients to switch to daily dialysis. There is a cost problem, of course, but for daily dialysis we can forget all gimmicks the modern dialysis machines offer at high costs. We can return to central dialysate delivery systems. Together with the cost savings provided by the reduction of morbidity, daily dialysis will become affordable.

CLOSING WORDS by Prof.Dr.Ray VANHOLDER.

I have been asked to write down a number of final conclusions regarding this Journal Club discussion. First of all I would like to thank everybody who contributed for the interesting suggestions and the intense discussion. The remarks were correct and outstanding. This is an interesting initiative.
We might distillate out of the discussions the following concerns:

The old concept of Kt/V might be an interesting “ a minima “ principle for dialysing uremics, but there is more happening to dialysis patients than urea retention only.
There is a substantial concern about the flaws that may occur during the calculation of Kt/V. These are mainly attributed to the multicompartmental distribution of urea, and the unpredictability of this distribution ( also related to hemodynamics and the shut-down of parts of the compartments if hypotension occurs ).
There is a major concern about the fact that other than small water soluble compounds ( e.g. middle molecules ) might be responsible for biological/clinical defects in uremia.
A bit as a surprise, and not directly allowing clear-cut conclusions, was the recent publication of the HEMO-study from the US, where neither Kt/V nor application of large pores seems to bring in an extra impact on survival.

Jean-Yves DE VOS (BELGIUM)
EDTNA/ERCA JC Manager

A big thanks to all JC members who makes this forum a living tool, sharing ideas and practical experiences.