I like to keep up on the goings-on in dialysis even though my daughter has a transplant - looked into dialysis eleven years ago, went into shock, and never quite came out of it.

Anyway, I was reading Dr. Belding Scribner’s and Dr. Dimitrios Oreopoulos’s article on HDP:

It looks like a very fine formula they came up with, but have to wonder why they don’t account for a person’s weight? Just off the top of my head, I would think dialysis would be different for a 100 pound person vs. a 200 pound person.

I was also wondering if anyplace eventually used HDP measurements that you know of?

I know the Haemodialysis Product (HDP) paper of Scribner and Oreopoulos very, very well. And, by the way, it is an easy read - so if you want to have a go at reading it, see ‘Pluggers’ link.

I loved it when it came out, and love it still - though it was published in Dialysis and Transplantation, a relatively low impact factor journal, and, perhaps as a result, it never quite got the ‘airplay’ it otherwise might have and which it certainly deserved. However, I think the paper needs to be read for its’ intent, and not as a literal application.

In fact the ‘formula’ for the HDP was … ‘a figment of imagination, sired by reality and born of desperation’ … thoroughbred terms for a thoroughbred idea, but an idea meant for breeding rather than for racing.

Those three words:

Imagination … sadly, too little of this surfaces in our trials-driven medical landscape now-a-days

Desperation … the ‘cry-in-the-wilderness’ by those who still possess imagination and ache to re-ignite it in others before it disappears from our minds altogether.

Reality … well, what I mean by the use of the word ‘reality’ is to emphasize that the ‘reality’ in vogue at the time the HDP was proposed … and, mind you, a ‘reality’ that (sadly) still is dominant in the thinking of most … was that the accepted measure of ‘dialysis adequacy’ was (and still is) Kt/V.

What I think ‘Scrib’ and Dimitri (correctly, in my view) wanted to point out by proposing the HDP was the total inadequacy of ‘adequacy’ as that ‘reality’ as it was (and is) measured by Kt/V.

Simultaneously, they ached to emphasize (and bring thinking back to) the importance of dialysis time (sessional duration) and dialysis frequency (sessions/week).

This is a view I personally have also long believed and have also ached to emphasize. It is a view for which I have also ‘beaten the drum’, both before and since the publication of the HDP.

Yes … some of us do truly believe in ‘t’ and ‘f’ - and fervently disbelieve in Kt/V! And yes … I am one such!

But, the problems with the HDP were several.

  1. It was ‘hidden away’ out of view of many/most nephrologists

  2. It truly was a formula ‘plucked’ from the air without validation or testing … as I said, a figment’ of imagination - despite being, in my view, a pretty good figment.

  3. The squaring of frequency - again, in my view - maybe overplayed frequency and underplayed time … though the two have never really been adequately trialled, one against the other. The recent FHN trial did not answer this point - if it answered any.

  4. It never ‘took off’ … and though I think it had many messages (all good) to trumpet, it never made the impact - if not in actual application but in thought-process - that it deserved to make.

As a Scrib-o-phile, I - along with Chris Blagg and others - loved the intent of the HDP. The ‘intent’ was to say to the dialysis community … 'Hey, out there, what matters is membrane-contact-time. What matters is the weekly dose - not ‘dose’ as in the fictitious mathematics of Kt/V, but dose as in hours done per week. What matters is longer, slower, more frequent dialysis".

Did we listen … oh, yes, some of us did. But not enough of us. That remains the sad conclusion.

The HDP was more of a call back to common-sense than a real marker of ‘good dialysis’. And, I think they would probably agree with that assessment. But, were they on the right tram? … you betcha!

I have come at the same emphasis too … differently, but along the same lines … with my proposal for a 'Good Dialysis Index" - which can be seen at my website http://www.nocturnaldialysis.org (click on the GDI tab) and which I have presented (by request) at the last two ADC’s in Seattle and Phoenix.

And, as a Scrib-o-phile, long live his memory - and the memory and intent of the Scribner/Oreopoulos HDP idea. As were many of Scrib and Dimitri’s ideas, it was an idea, spot-on-track but ahead of its time.

Thanks for your reply! Though it is sad to see HDP never really got out of the chute, I hope your ‘Good Dialysis Index’ gets much further!

I did have another question. What would be a good lab(s?) indicator that longer, more frequent dialysis - and therefore middle molecule clearance - was occurring?

Middle molecule clearance is hugely complex and cannot be distilled into a single answer. But, I’ll try to give a flavour of the issues:

To put it as simply as I can, some salts, ions and molecules are small, and will easily and rapidly pass through cell membranes.

For other, more complex substances, membrane transport is much more problematic.

The bigger, the more gnarly, the more electrically charged or water coated - the more complex the substance - the more difficulties that substance will have to pass across membranes and … simply put … be removable (or replaceable).

As important an issue - or even more important - is the structure and characteristics of the cell membrane (or, at the blood dialysate interface, the dialyzer membrane) through which these substances must pass.

Cell membranes are highly complex structures but one of the chief ‘building blocks’ of the cell membrane is a complex interwoven chain of lipid and phosphate entities … known as phospholipid. Even phosphpolipids vary - in structure and complexity - from cell type to cell type. They can be thin or thick, sparse or dense, can carry varying degrees of electrical charge - positive and negative- and can be variably water resistant or water attracting.

There is a simple explanation of cell membranes at http://micro.magnet.fsu.edu/cells/plasmamembrane/plasmamembrane.html

To extract a small portion of this, it reads as follows:

" … the plasma membrane is composed of a double layer (bilayer) of lipids, oily substances found in all cells (NB: there is a nice picture at the website that shows this well). Most of the lipids in the bilayer can be more precisely described as phospholipids, that is, lipids that feature a phosphate group at one end of each molecule.

Phospholipids are characteristically hydrophilic (“water-loving”) at their phosphate ends and hydrophobic (“water-fearing”) along their lipid tail regions. In each layer of a plasma membrane, the hydrophobic lipid tails are oriented inwards and the hydrophilic phosphate groups are aligned so they face outwards, either toward the aqueous cytosol of the cell (the inward, water-repelling tails) or toward the outside-of-the-cell environment (the extracellular fluid). Phospholipids tend to spontaneously aggregate by this mechanism whenever they are exposed to water."

So … there is the substance itself - and then there is the structure through which it must pass.

Sodium, potassium, and chloride are examples of small, easily transported, uncomplicated substances. So, too, are some of the smaller metabolic wastes … like urea. Others, however, are a little more complex in their size and/or charge, and while they pass passably well - they don’t do so quite as quickly. Creatinine might be a fair example.

Middle molecules are bigger in molecular weight, they are gnarlier in outline, with twists and turns in their structure and with odd stickie-out-bits that ‘catch’ or cause hold-up during passage or which bear electrical charge (+ve or -ve) which can attract and bind or repel and prevent passage … in short, the more complex the molecule, the more tortuous and difficult the extrusion or inclusion across the cell barrier.

Nature has clever ways to import and export ‘stuff’ across cell membranes. Many cell surfaces have evolved specific transport systems - often involving ‘receptors’ on the cell membrane which are a bit like those transfer guys at airports, holding up a name when the plane arrives to signal 'here I am, I am your portal to the city … jump aboard" - while the rest of us take the bus or the train or wait in line for a cab!

But, many metabolic wastes don’t have specific transport systems 'flagging them down and making life easy" … so they wiggle and squeeze to work their way through.

Now … to answer your question: What would be a good lab(s?) indicator that longer, more frequent dialysis - and therefore middle molecule clearance - was occurring?

The answer is … phosphate! … but why? Isn’t phosphate a small molecule. It’s not dissimilar to sodium, or potassium. Why then is it my pick as an indicator of middle molecule clearance?

Why? … because it behaves as if it were a middle molecule!

And that is because phosphate has an odd characteristic … it is surrounded by a water ‘shell’, a bit like a walnut is surrounded by its seed casing. Phosphate, itself is small, but phosphate in living structure is not … and what’s more that shell is water.

Go back to your cell membrane for a moment. The phospholipid has a water-repelling (water resistant) ‘end’ to its chains … all pointing inwards into the cell. Phosphate is kind-of ‘trapped’ there and thus finds it hard to be ‘cleared’. It does escape … but it is slow to do so and it takes time!

Now this is a hopelessly simplified and a ‘not quite right’ explanation … you will get a flavour of the complexity at the Wikipedia site for phosphate @ http://en.wikipedia.org/wiki/Phosphate_homeostasis … but it gives the sense of why some substances move differently to the way you might expect them to.

Phosphate is an excellent ‘surrogate’ marker of middle molecular clearance and you will already know that one of the clear differentiating advantages for NHHD (or long, slow, extended hour and higher frequency dialysis) is the way better phosphate clearance.

Now - and this is a late addendum (see your follow-up question and my answer lower in the following posts of this thread) - this is predicated on the supposition that the patient is not on phosphate binbers. Clearly, these will muddy the interpretive waters considerably. We are used to our many long, slow, frequent dialysis patients not being on binders … so I think of that almost as a ‘given’ and I should not … especially when the vast majority of patients are still on hefty doses of binder medication.

OK … I agree, phosphate transport doesn’t represent the transport of all middle molecules. Nor does it reflect the transport of protein-bound groups. But it is a surrogate. It does loosely parallel the passage of other, far larger molecules in the 500 - 60,000 Dalton range (the European Uremic Toxin Work Group has defined the term middle molecule to be between 500 Daltons - 60 kD) … the so-called middle molecule range.

We also measure B2 microglobulin … and homocysteine too … regularly in our patients but these are more expensive ‘tools’ as indicators of MM clearance and also take a while to get laboratory results for. Phosphate, on the other hand, is cheap to measure and offers a fast turn-around on results. We measure pre- and post- phosphates in all our patients (conventional and nocturnal) and, someday when I get the time, it’d be cool to analyze the two groups.

So … in a sentence …phosphate clearance is this a useful laboratory tool to ‘speak for’ middle molecules. It is a pity we use it less often than we measure urea transport through Kt/V!

So if the goal is to find out if a patient is doing longer dialysis by checking the patient’s phosphate levels, could the person doing the checking be misled if the patient is using phosphate binders?

And if a person is being misled by just looking at phosphate levels, is there another lab value or something that could be looked at to tell if a patient is taking phosphate binders?

My error: A correction

You are absolutely right. In my previous answer, I neglected to say what to me is daily fact of life (and ‘a given’), as I am so used to our patients having normal phosphate levels pre-dialysis and sub-normal phosphates post-dialysis with their interdialytic mean being the mid point of normal and yet they do not take binders.

But - and here is my error - I assumed that that would be understood … as you clearly have, but others may not have (especially those who are not so used to long, slow, frequent dialysis patients) that the surrogacy of phosphate as a marker is dependent on the patient not having adjunct binder therapy. So, yes … the comments I made regarding phosphate as a surrogate marker for middle molecular clearance due to its’ behaviour as a middle molecule despite that it appears to be a small molecule were predicated on the supposition that the patients are not on binders.

I should have said that - and I didnt!

With your permission, I will go back into the previous post and ammend it in case other read it - but not this correction.

As far as the second part of your question … “is there another lab value that can point to when a patient isn’t taking binder therapy” … sadly, I don’t know of one. It’d be nice, wouldn’t it, if that were the case : it would remove much of the angst and friction that occurs between staff and patient when accusations of ‘non-compliance’ fill the space between. However - no measure fills that void , at least to my knowledge.

Ok, I’m going to lay my cards on the table. The reason I’m asking all this is because we are pushing my State Representative to think about standards of care for dialysis patients here in Colorado.

At first I had hoped HDP had become tried and true, somewhere, somehow. Then I hoped a lab value or two might suffice - looks like phosphate levels ALMOST made it. Now we are eyeballing your GDI real hard.

I am glad you are moving ahead

But … remember that:

[B][I]1. the GDI is a figment of my imagination … just as HDP was a figment of Scrib and Dimitri’s imagination

  1. it has only been tested in our small dialysis service

  2. questions 7, 8 and 20 bias (unashamedly, but heavily) toward long and frequent dialysis and should be excluded if assessing programs that do not offer these options

  3. the GDI was but a 1st attempt. It needs tweeking, improving, localizing (ie: the laboratory units) to local practice and laboratory normal values

  4. there may be questions I should have added, some that might be removed, some that could do with altering

  5. It has not been validated … and should be (or needs to be), before it is used as a replacement for you current practices[/I][/B]

It is a concept to be played with, to be improved … but, a few things about the GDI, to me, are clear …

[B][I]1. it attempts to put the patient back into the assessment where the patient belongs

  1. It allows the patient to speak to their own aspirations, their own view of ‘how they are doing’

  2. by spreading the questions over 5 key areas: the patient, the process, the laboratory, the chart, the medications … each gets ‘a say’ in the progress of the individual

  3. and … it is for and of the individual

If you look at it - or something like it - play with it, shape it, make it fit your needs … it is not immutable or fixed in stone! It is but one idealistic (?) nephrologists’ view of what should constitute better dialysis. If you like, it’s a patient KPI list!

Good luck … and if you use it (or a variant of it), please, one day, let me know how you think it fits. Is it better than simply dialysing to …

eKt/V = (- Ln (R - 0.008 x t) + (4 - (3.5 x R)) x UFR/W) – (0.6) x ((- Ln (R - 0.008 x t) + (4 - (3.5 x R)) x UFR/W)/t ) + 0.03

… and then turning off the machine. I don’t accept that that equation can wrap up, ‘bundle’ and express good dialysis. We are not mathematical equations … we are humans, with aspirations and frailties. Maths will never, in my view, equate to the beauty of a human spirit.

Good luck

Just all pieces of puzzle the way I look at it. Anyway, thought I would continue our discussion with a PM I sent you.

As long as we are talking about ideals, I’d want a measure that is intuitive to understand, that had meaning without being explained. I think one potential measure that meets that bar is blood cell life span. If you could say to someone that your BCLS has gone from 100 days to 90 days, I think you would have their attention as you explain that it likely due to their shortening of treatments or skipping. And it wouldn’t have to be blood cells I would theorize that all the cells in the body have a shortened life span when the body is under dialyzed. It could be a definitional measure - you are under dialyzed when your average cell life span is X.

The most powerful outcome we can measure, the outcome of greatest importance to dialyzors is mortality but of course for an individual a binary mortality measure does not make a good CPM but using cells as a surrogate for the whole would increase the N=1 to N=billions.

Now we just need an inexpensive and highly accurate way to measure average cell life span.

Possibly, a composite marker

As ever, Bill, from you, a thoughtful post – and one I doubt I can come up with an answer for – but, ‘oh, that it could be so’.

As Hamlet opines in his ‘To be or not to be’ soliloquy … ‘T’is a consummation devoutly to be wished’

Your search for a cell whose lifespan (a) is easily accessible to repeated analysis,(b) is affected by CKD and/or uraemia, and © … and here you make a leap of faith … is so sensitively affected by the level of metabolic ‘derangement’ that even small differences in the ‘goodness’ or ‘badness’ of dialysis might make it into a measurable indicator of dialysis efficacy … is going to be an unrequited search!

I’m afraid it just ain’t gonna happen.

While a great thought … and I mean that … and while the lifespan of an RBC might seem a potential candidate, the means by which we can measure sensitive and small decrements in RBC lifespan just aren’t there.

First, they are not accurate enough to measure the small changes (and they would be small) that might accurately grade levels of uraemic impact.

Second, too many extraneous factors also impact on RBC lifespan: iron status is one … but so, too, are ESA responsiveness and dose, folate levels - and many more.

To fulfill the ‘easily accessible to repeated measurement’ criteria, any measurable ‘factor’ would have to be a blood element. No dialysis patient is going to accept the regular and repeated taking of an ‘attached tissue’ sample … nor can I envisage one that would possibly fulfill the 3rd criteria of being so sensitively affected by uraemia as to be of use, even if sampling were possible.

So, no, I don’t think that would work. Nor would the white cell or platelet population be useful either – for similar as well as other ‘won’t work’ reasons. Much as it seems a nice idea, I just don’t see a cell, or the life of one - attractive though the idea may seem - working. Until such time as far more accurate and sensitive estimates of red cell survival that are also separable from other significant ‘uraemic influences’ - like EPO, ESAs and iron - it will always come back to a biochemical measure.

And, if biochemical, then what measure? And, here, I am again tempted to stick my neck out (badly) … why do I always end up doing this? …

I have long been tempted to consider the possibility of a ‘composite’ marker. So much has this been a temptation, that I have been patiently accumulating post-dialysis data in the hope that, one day, I might get the chance to potter around with it with this very thought in mind: a better biochemical aggregate by which to measure and guide ‘good dialysis’.

While you know my views about Kt/V, so I won’t expand here other than to say that its’ use as a ‘marker’ of effective dialysis is – in my view – seriously flawed … it is (likely) a fair marker of urea clearance (and some of the other smaller metabolic beasties) removed by dialysis. It just doesn’t represent all the other stuff that dialysis must remove (clear) – especially the middle molecule ‘group’.

But, as phosphate moves, acts, and behaves as a middle molecule even though it seems small (see my earlier post in this thread), then might not its’ clearance be used in conjugation with the urea clearance to give a combined or composite measure of small and middle molecule removal.

To this end, we have amassed at my unit over the past 6+ years a data-set of pre- and post-dialysis biochemical markers by routinely drawing not only pre- and post-dialysis urea and creatinine but also pre- and post- phosphate and calcium.

In Australia, few (if any) use Kt/V … but we all routinely measure the URR (ie: the urea reduction ratio … or, if expressed as a %, the PRU = the percentage reduction in urea … as was proposed by John Daugirdas as a ‘poor man’s Kt/V’).

The URR (or PRU) is a national measure in our ANZDATA registry – we all collect it.

The PRU is the pre-dialysis urea minus the post-dialysis urea, divided by the pre-dialysis urea, and expressed as a percentage – and is a measure of the degree (the %) by which the urea has been lowered by that dialysis treatment. While >65% is regarded as ‘adequate’, we shoot for >70% in all patients. Pump speed, dialyser surface area, dialysate flow rate and time (ie: dialysis sessional duration) can all be up-sized (surprise, surprise that time yields the biggest impact) to achieve a ‘better’ PRU.

Urea reflects small molecules (well, at least some of them) while phosphate is a ‘surrogate’ marker for middle (or larger) sized molecules.

Sure, other 'middle molecules exist: vitamin B12, B2-microglobulin, homocysteine are three that come easily to mind … but none of these is cheap to measure and none have the instantaneous ‘get a result in minutes’ value that phosphate offers.

We (or, at least, some of us) seem to ‘accept’ the PRU as a marker … why not, then, a “PPO4” equivalent … ie: a phosphate-based, PRU look-alike … the percentage reduction in phosphate.

We have >6 years of accumulated pre-and post- data in both conventional and home NHHD patient populations - all sequential and in all patients (with outcome data too) - waiting for me to potter with – if and when I get the time.

Then, maybe, some combination like the average of (PRU + PPO4) might emerge to better reflect both small and middle molecular clearance.

All that may be pure rubbish … but I haven’t yet put it out of mind to try.

We have this data on our NHHD’s going back >10 years. Our conventional dialysis group data goes back 6+ years … we only began to collect the conventional group data as a comparator group when it occurred that it might be useful. Now, I need the time …

Of course, PO4 may be be-devilled (as a marker) by the potential effect(s) of binder agents and their impact on the serum phosphate level (both pre- and post-dialysis) … and would this matter? To tell the truth, I am not sure how to rationalise this … though when eyeballing our pre- and post-dialysis data-sets for long, slow extended hour and high frequency dialysis for binder-free patients versus the data-sets for our ‘bound’ conventional dialysis patients, I don’t think it would make the numbers unusable.

The temptation is to think that if we (or, at least, some) put so much store by the PRU (or Kt/V), then it is not too far a stretch to think that similar devotion might accord to a composite measure which in-builds that pesky set of bigger beasts we call middle molecules. If the clearance of PO4 is a ‘fair’ marker of the clearance of these (and PO4 is certainly easy, quick and cheap to measure), then why have we left it - for so long - out in the cold.

Finally, while on the topic of markers - and composite markers at that … wouldn’t it be good to in-build into our ‘composite’ a marker which might ‘mark’ volume. That would yield a marker reflecting small molecule clearance (urea), bigger molecule clearance (phosphate) and the dimension of intravascular volume contraction during dialysis. What might work, here, is the serum albumin.

I have published a few peer-review papers and abstracts about this … see my last effort (this one with Andreas Pierratos) which deal with this area: Agar JWM and Pierratos A. The serum albumin and hemoglobin: Behavior patterns in nocturnal hemodialysis. Hemodialysis International 11(3): 303-308, June 2007.

In simplistic terms, the serum albumin rises in “wham, bam, thank-you ma’am” dialysis as albumin stays within the intravascular volume and, as it is contracted by rapid fluid removal during conventional dialysis, it becomes increasingly concentrated.

In other words, the measured concentration (of albumin) will rise if the rate of fluid removal from the intravascular volume constantly exceeds the rate at which fluid can flow from tissue and cellular fluid stores to replace that which is being removed by dialysis. The more rapid the dialysis, the greater the impact on blood volume and the greater the hyperconcentration of albumin.

But, as the dialysis treatment lengthens and slows, the rate of replacement from tissue and cellular fluid stores can keep pace with the rate of fluid removal, and the serum albumin concentration will change less, and less, until - at the point of equality where the rate of fluid removal by dialysis is no greater than the rate capacity of the tissues and cells to replace that which is being removed - the serum albumin will no longer be concentrated during dialysis and remain unaltered. The longer and slower (ie: the more physiological) the dialysis, the less the pre- to post- change in the serum albumin … until there is no change at all.

Measuring the pre- and post-dialysis albumin should, therefore, give a measure of the impact of the dialysis process on the blood volume. the less the change, the less the impact and (arguably) the ‘better’ the dialysis - from a volume aspect.

That’s complex - so if you don’t follow, say so … though you may need to re-read a few times.

In the end, a composite small molecule + large molecule + volume ‘measure’ might be possible … where each is accorded a measure and the summation equates to a biochemical and volume yardstick of the ‘goodness’ - or ‘badness’ - of the treatment process.

I still think more than just biochemistry and volume matter. The ‘good dialysis index’ … see http://www.nocturnaldialysis.org … assesses far more than just biochemistry and volume … but as a measure of these parameters, urea, phosphate and volume in a tripartite composite might be better than Kt/V (or PRU) alone.

Food for thought …

I needed to sleep, dialyze, drink a cup a coffee and then reread it (:

First, the thing that I like the most about being able to report cell life span is that it is intuitive but that is a moot point if as you say it wouldn’t be sensitive enough to provide what we want. However, to use a golf analogy, if I make a blind shot to a green and on approaching the green I don’t see my ball, the first place to look is in the cup even if the most likely explanation is that it rolled off the back. Even though the ball has never been in the cup - it’s the first place I look. Why not check where you would like the ball to be?

I’d like to think that even skin cells would reflect the cumulative impact of dialysis, so maybe it isn’t that cell life span would respond to day to day renal replacement events but maybe cell life span could be more like an end of the year report card letting everyone know how we’ve done over the last year. Ah well, one can wish. What a piece of work is man … perhaps there is a solute that is the quintessence of dialysis health.

This is the first time I’ve heard the idea that pre and post Albumin could be a surrogate for volume control. It can say that one hasn’t exceeded the UFR speed limit but can it say that the person finished dialysis at their proper hydration? I suppose using it in conjunction with BP - does good volume control = good dialysis? If so can we just use water as our surrogate solute?

Hmmm thinking out loud. Maybe the issue is that we’re too close to the trees to see the forest. Maybe a treatment to treatment measure wouldn’t be as useful as a “dialysis year” measure. For people like me, last month I marked 23 years of renal replacement therapy, the day to day seems less consequential than the year to year. If we want to talk about how I did last year replacing my renal function does that lead to a different measure? What would be a CPM to describe my treatment regime generally? Are the parts - each treatment - a good surrogate for the whole? Or is the whole (the dialysis year) greater than the sum of its parts (each treatment)?

A couple of quick thoughts:

You commented … “I’d like to think that even skin cells would reflect the cumulative impact of dialysis”

I suspect skin cells do take a battering in the uraemic state - and maybe that would be an interesting cell line to follow - but, again, while I understand your golf shot analogy (btw: mine would always be over the back of the green) I just can’t get my head around other than the RBC as the only cell-line likely to be a meaningful marker. And, sadly, our means of assessing the life span of the RBC, as a population, are so blunt, slow or expensive, that it doesn’t seem likely - at least in the near future.

Of course, we do use the RBC as the yardstick of ‘goodness’ of diabetic control … the HbA1c! Is there a time-related, easily measurable, quantitative ‘change’ that might represent a summation ‘beat up and battery’ of the red cell and its’ haemoglobin by uraemia (just as the percentage of haemoglobin glycation leads to a measurable yardstick of red cell exposure to glucose through its life span) that might be used in a similar fashion? … not one I know. But that doesn’t mean at all that there isn’t one. That is maybe one search, or one bit of thinking, that still needs to be done!

You also wrote re the “pre and post albumin (as) a surrogate for volume control” and rightly wondered … “can it say that the person finished dialysis at their proper hydration?”

No … it won’t say that … true. It will only say something about the volume impact of that treatment relative to the blood volume pre- and post-dialysis. It won’t tell you if that blood volume is right. A BVM (blood volume monitor) may help there … or, maybe better, a BCM (body composition monitor) … but both of these also have their limits and befuddlements.

As for long range measures … there are still things like echocardiography - to assess the impact of dialysis (over time) on the LV: or nerve conduction - to follow the impact of amyloid deposition in peripheral nerves (but its way too blunt, after the fact, and out of favour) …

You pose an interesting question to which there are no answers, other than the musings of an aging brain!

One day, we may discover the obvious, like the ball in the hole in your golf analogy. Till then, all one can do is offer (likely) off-beam thoughts. This, of course, is the danger of Internet question sites … where things can go off-piste, rather quickly, like this discussion … yet, they are fun to have, as long as the reader(s) don’t take the musings too seriously.

Back to the central theme then: how to measure good dialysis?

I again emphasise (at least to my mind) that the patient is and has to be at the centre of that measurement. I don’t believe he/she currently is - or not enough - and we need to re-insert the ‘you’ back into our thinking, putting ‘you’ at the apex of the search for a better way to measure and express the ‘goodness’ (or otherwise) of the therapy.

Hence, my naive attempt at a “Good Dialysis Index”. For better, for worse, it’s there, to be considered, on its’ merits or its’ failings, among the mathematical formulae that others use to tell ‘you’ how well ‘you’ are going!

Talking about what would be good CPMs forces us to define what we mean by good dialysis and that’s enough of a reason in itself to make the effort worthwhile.

The good dialysis index does a much better job than any urea based measure could but the way it is today everyone in healthcare, from the dialyzor to the docs to the payers, has a fetishistic reliance on laboratory numbers. Blood measures are healthcare’s umpires and guidebook all in one. A lot of good comes from that but I think we’re missing a lot too.

Maybe what has to happen is that laboratories need to start reporting GDIs (a person’s Good Dialysis Index), then getting docs and dialyzors to base treatment decisions on the GDI will be a snap.

No-one knows the GDI exists

Sadly, but truthfully, I doubt anyone has ever seen the GDI except for those who may have stumbled across it at my www.nocturnaldialysis.org website, or who have attended the last 2 x pre-ADC Home HD Symposia, or who may have bumped into it here, or who are on my side of the ‘big pond’ we call the Pacific.

No journal would ever consider it worth publishing - after all, where is the validation, the RCT, the trial, or the proof that it might just have some value?

Yet, I really believe it has something to say - some merit for you, the dialyzors. Maybe it needs work - I don’t deny that - but its’ heart is in the right spot, I think.

In a way, and to me, that is a core sadness in the current state of medical literature … that unless there is trial evidence, a concept, an ‘idea’ cannot be published. There is no place for thought bubbles, provocation, discourse, the floating of thought - even if that thought be shown, in due course, to fail.

Peer review would never allow a concept as ‘waffly’ as the GDI to gain journal space. I know it. Yet, it is sad that that is so … for contained within ideas ‘like’ the GDI are kernels of thought that just might ‘lead us somewhere’.

I can present it at meetings - I have - here and, when brave and forward-thinking meeting conveners (like Chris Blagg) can take a chance on an idea and allow it airplay, in the US. But in a journal … sadly, no.

And, if and idea, a thought-bubble, a concept - even if a good one - is not published, then it must perish.

That is the way of modern medicine - though it is sad that it is so.

Anyway, thought we could continue our discussion about Kt/V here on the main board since pm’s seem to be acting a little squirrelly.

First off, I wanted to make sure I understand Kt/V properly. To boil it down, isn’t it: (K: dialyzer specifications and pump speed) times (t: treatment time) divided by (V: patient’s total body water)?

If what I am assuming is correct, and the goal is to increase t (treatment time) - while keeping Kt/V the same - would it be best to limit pump speed or K? (wondering if by limiting K instead that would add some flexibility - which I’m not sure is good or bad)

Plugger I think the discussion needs to clearly differentiate between Kt/V - the formula you asked about and standardized Kt/V. They are not the same thing - it is confusing because they sound more similar than they are. I would think that if you’re considering what makes a good CPM than you would be looking at std Kt/V and not the original Kt/V.

I think this is an important point of confusion that needs to be discussed. Here is a starting point, the Wikipedia article on std Kt/V.

It is a math rich environment but lets not be intimidated. I’d really like us to work through it - and doing it here (v private message) would help a lot of people understand what is meant.

In simple terms, the answer is ‘yes’ to the question posed by Plugger …

“K” (dialyzer specifications and pump speed) multiplied by “t” (treatment time) and divided by “V” (patient’s total body water)… is the ‘boil-down’ nub of Kt/V.

And… it’s a ‘yes’, too, to the comment from Bill that Kt/V and standardized Kt/V are not the same measures.

In the Wikipedia article on stdKt/V cited by Bill Peckham, my difficulties with Kt/V (in all or any of it’s formats … Kt/V, spKt/V, eKt/V, stdKt/V … and there are others) are covered under the ‘disadvantages’ heading.

Think of urea. Urea is a small, easily diffusible, nitrogenous waste that crosses most cell membranes and tissue densities, layers or interfaces reasonably rapidly. Despite this, it is known to diffuse only slowly across the blood:brain barrier such that, if the blood urea relative to the intracellular urea of the brain cells drops too quickly (ie: as a result of overly aggressive dialysis … another argument against ‘bazooka’, short duration, conventional dialysis schedules), the intracellular urea of the brain cells can act osmotically and ‘suck’ water into brain cells down a urea-driven concentration gradient.This causes them to swell, resulting in the dialysis syndrome known as the disequilibration syndrome … a syndrome characterized by the onset of headache towards the end of dialysis – though worse things can result than just headache.

Think of the things that contribute to the blood urea concentration. Urea is produced from protein metabolism. It is also absorbed across the gut from the intestines where protein, after being taken in as food, is broken down inside the gut by urea-splitting organisms … it is meant to be … and urea is thus absorbed directly into the blood stream. If there is any blood in the gut, this accentuates this as urea is released from the blood cells by urea-splitting organisms and the urea is absorbed. Furthermore, functioning kidneys also reabsorb urea – along with water - from the last part of the renal tubular system … though this is clearly going to be a less relevant component in dialysis-dependent CKD5 patients where, largely, kidney function has been already lost.

So, the blood urea is a composite of metabolic contribution, gut absorption, and tubular reabsorption (the latter not significant in established ESRD) as the major contributors.

  1. The blood urea is the pre-dialysis ‘marker’ used to set in train the measurement of any form of Kt/V (or URR, or PRU)

But, then, during dialysis, urea is progressively removed by the dialytic process. That’s clear.

So … it should be easy, shouldn’t it, to calculate how much has been removed, especially once knowing the blood urea at the end of the dialysis session.

But … it’s not quite so easy. Urea is still being made by metabolism during the dialysis. And, here, the nutritional status of the patient is important. Both in the inter-dialytic and intra-dialytic period, a malnourished patient will have a very different metabolic rate to a well nourished one, and will ‘create’ a very different urea load as a result.

Urea is also still being absorbed from the gut through the dialysis treatment – and this may vary depending upon (for example) what was eaten in the pre-dialysis meal and its’ associated ‘load’ of protein … was it a breakfast of cereal and a cup of tea or a breakfast of Canadian bacon, sausages and scrambles egg? … the two would clearly yield a very different urea absorption load to then remove, after absorption, during dialysis.

These issues, and others, ‘bedevil’ efforts to characterize the behaviour, the load, the making and the removal of urea.

  1. Then there is the conundrum of when to measure the urea at the end of dialysis.

This problem is created by the ‘rebound’ phenomenon. As urea is removed from blood, there remains a ‘lag’ period, where the concentration of urea is made lower in the blood (from direct dialysis removal) than it still is within the tissues and cells of the body. Even though we say that urea is a rapidly diffusing small molecule, even that doesn’t make its’ concentration either (a) instantaneous or (b) equal, tissue to tissue … remember, for example, the issue with the blood brain barrier.

So, after dialysis finishes, there is a ‘bounce-back’ (or ‘rebound’) in the blood concentration of urea until equilibrium is established between the concentration of urea in all tissues and in the blood. This takes a little time … and that time depends on differing urea diffusion times, tissue to tissue and tissue to blood.

Question: when is the right time to take the blood sample that is to be used to measure the post dialysis urea (and from which the calculation of Kt/V can be made).

Should it be taken, lightning quick, at the end of dialysis and thus represent a single urea ‘pool’ (the blood compartment) … the spKt/V … or should equilibration first be allowed … the eKt/V … but if the latter, how long should that equilibration be for, as minutes pass. How long should the ‘equilibration time’ be to allow full tissue/blood pool equilibration. And, to complicate the decision, remember that more urea is being created by post-dialysis metabolism and gut absorption … ie: the pool is changing. Should the sample be taken at 2 minutes, 5 minutes, 20 minute, ½ hour … and there are arguments for and against all.

The problem: humans before, during and after dialysis are not in a steady, unchanging state but are a dynamic, changing, many-pooled, complex living system.

  1. Then, there is the problem of body water … the notional fluid inside us in which urea is dissolved.

Body water varies by as much as 15-20% from body conformation to body conformation and from sex to sex. If the volume urea is dissolved in varies … ‘Houston, we have a problem!’

And … body water isn’t in one bucket of water where, like adding a dye, the water could be swished around and thoroughly and equally mixed. It is hidden in nooks and crannies, in different tissue densities in organs, muscles, bone … within cells, outside cells. Water is in accessible places, inaccessible spots and everywhere in between.

So, the notion that urea, dissolved in those water ‘tables’ inside us is all equally accessible and diffusible just isn’t so.

To then think we can treat it like a one-size-fits-all ‘volume’ isn’t quite valid.

And … the issues go on … and on.

So …

This is why it is so hard to develop a mathematical model to cover and correct for all variabilities.

This is why I personally find it so hard to rely on any form of Kt/V – ‘original’, ‘sp’, ‘e-’, ‘standardized’ … take your pick. While I respect those who believe in Kt/V, I cannot bring myself to think that the wonder of a living organism can be modeled by mathematics.

And … as a measure of dialysis adequacy: one marker, yes … but the marker, no. It can be used as one of the markers of good dialysis. But it shouldn’t, in my humble view, be the bees knees. That is why I include it in my ‘20’ … but I give it no greater value than any of the other 19 measures.

However, Bill is right … standardized Kt/V should be the Kt/V ‘variant’ used for inclusion in any measure – particularly as it allows cross modality comparisons. I put Kt/V in to the model without stipulating ‘standardized’ – and I should have.

Kt/V, in my view has served a useful purpose: it has given us a minimum measure. But, as a measure of what is good, as a measure of excellence - even of ‘adequacy’ … and how I dislike that word and wish it had never been applied to dialysis … no. It doesn’t do it for me.

I don’t know if that’s been helpful.

I hope it’s correct.

Whatever the case, I will, I am sure, upset the mathematical among you and my colleagues by saying that I am not a believer in Kt/V (of any ilk) … full stop.

Ok, maybe I will forget the private messages since I might know enough now to be dangerous.

What I’m getting is HDP is something that didn’t make it out of the gate, GDI and cell life spans are works-in-progress, phosphate levels may not indicate middle molecule clearance if binders are used, and Kt/V is a poor indicator of good dialysis which also sounds like it can be gamed.

But let’s say the state of Colorado told the clinics they must turn down their pump speeds and must increase their dialysis times. Is there any way to check to see if they are complying? (maybe I should have asked this in the first place)

And thank you both for your help with all this!


Ah … now your real question, I think

  1. Can time (duration of treatment) be simply measured - for compliance purposes?


  1. If time is lengthened, can pump speed be reduced?

In answer to the second … yes.

Here in Australia, the mean ‘national’ pump speed runs between somewhere 300 and 325. This is +/- 100 ml/min less than many current units run in the US.

Here in Australia, the mean dialysis sessional duration is about 270-275 minutes while in the US - correct me if I am wrong, it lies somewhere between 210 and 220 minutes. The Australian sessions (t) are full hour more per session (or in a 3 x week program, 3 hours more dialysis a week).

Yet, we both generate the same weekly Kt/V (or roughly so) … if and/or when Kt/V is calculated on Australian patients (as it is not routinely recorded).

We, broadly, record the simpler URR (or PRU) in our national registry. And here, Kt/V is not a KPI measure - nor is it likely to become one - probably because most of us don’t care much for it. We do measure (in our unit … and most, if not all, others do too) the PRU on a 2nd monthly basis.

But, even then, the PRU is used an internal KPI and not as a reported, national one - except as an after-the-fact record to ANZDATA once yearly at the end of each year.

While we drive for (and achieve) a PRU > 70% in almost all patients - those less being those with access issues under direct and current review and (where possible) repair - it is but one of many measures of good dialysis … and, even then, is not the key or primary measure I use.

I like (as you know) a wider, more patient-inclusive view of dialysis … the Good Dialysis Index. While not validated (agreed), it makes more holistic sense to me, as a practicing nephrologist, than an isolated Kt/V (or PRU).

So … by turning down the pump speed, one lessens Kyes.

And … by turning up treatment duration, one increases tyes.

But, as a result, the Kt combination may, in effect, remain unaltered. So … while there are other factors which may exert an influence, in the simplest construct of your proposition, the answer is ‘yes’.

Your problem (and your question) - as I think I now understand it is … how could compliance with a lesser K and higher t be measured if the end calculation Kt/V remains essentially unaltered.

Kt/V won’t give it to you. Nor will it’s cheaper cousin, URR.

Thinking about it, its probably back to Scrib!.

The HDP certainly would, in this instance, be an answer. The HDP is simply a function of duration x frequency … nothing else. Forget the ‘squaring’ … in this case that’s immaterial, though I do have an issue with that in the HDP paper in D&T, as it was presented there.

If frequency remained unaltered in a program (3 x week), the only way to measure compliance using (t) that I can think of off the top of my head, would be to calculate both Kt/V and the HDP, and observe (as a compliance measure for t) that while the Kt/V remained unaltered, the HDP was seen to be lengthening.

That would confirm longer dialysis (t) and presume … most likely … that if KT/V was unaltered, the lesser value in Kt would have to be K. And, if the only measure of K that had altered was the Qb (read pump speed), then … ergo … you have your measure.

It’s a broad brush measure, to be sure, but it’s easily do-able.

The HDP would need to be converted from the Scribner/Oreopoulos ‘hours’ to a ‘minutes’ calculation … = simply done.

Then, yes, it would likely work.

Does that answer your question?