Blood Flow Rate, Fistula Integrity, and Optimal Clearance

This blog post was made by Dr. John Agar on October 26th, 2017.
Blood Flow Rate, Fistula Integrity, and Optimal Clearance

A few topics seem to come up, over, and over, and over again at the Home Dialysis Central FaceBook discussion group. These are

1. Blood flow rates (read “blood pump speed”)

2. Fistula integrity relative to blood flow rate

3. The effect of increasing blood flow on clearance

Let me place my “take” on fistulas, blood flow rates, optimal clearance of small and middle molecules, and fistula integrity on record. I do this in response to—again and again—posts that advocate blood flow rates of 400, 450, and (ye gods!) even 520 ml/min! Often, it seems, these pump speeds have been “prescribed” by professionals = even more scary! So, here goes…

First and foremost, a fistula is a surgically formed “bridge” between a high flow, high pressure, rigid artery, and a low-flow, low-pressure, stretchy vein. A fistula can be made pretty much anywhere where an artery and vein can be physically joined. For dialysis purposes, most are made at the wrist or the elbow, with a less common site being in the upper thigh.

Veins are not really built to enjoy the high pressure flows that result from this surgical linkage. One of the surgical skills in fistula construction is to form a point of linkage of the right size, angle, and curvature to ensure there is as little turbulence as possible at the join, and on up into the first part of the now high-pressure-exposed vein. Turbulence disturbs the normal flow of fluid within a tube (here, the blood vessels) and creates forces (sheer stress) that deform and irritate the wall of the vein.

Think for a moment of a stream, and its bank. The two are usually calmly in harmony, each with the other. Water flows by, nurturing the plants that grow along the bank, the bank itself stable and undamaged. Suddenly, there is a storm. The stream floods, the water rages, the plants along the bank are pulled and battered, the turbulent waters dislodge the ordered bank and parts break away. Turbulence creates trouble at the river bank: turbulence is a damaging force.

So, back to the dialysis fistula. Just as turbulence damages the bank of a stream with high flow rates, so too does turbulence irritate, aggravate, and stimulate the endothelial cells that line the walls of the now high-flow-exposed vein. Their response is to proliferate—to increase in size and number. They begin to grow out into the lumen (flow channel) of the vein. In time, this causes a narrowing (stenosis) as the multiplying cells crowd into the flow channel. A nice review of this topic (by Mary Hammes) can be found here.

A word of warning: the paper is not for the faint-hearted, and it is in complex language. But, it does offer a wealth of research data and links around the topic of turbulence and the endothelial anger that it stimulates, promoting endothelial growth (intimal hyperplasia is the medical term) with the result being stenosis.

Here I am finally getting closer to addressing the FaceBook arguments: endothelial cell growth and proliferation is the vein wall response to turbulence, wherever it occurs. This will not only be at the anastomosis, or surgical join between the artery and vein but any other point, too. In dialysis fistulas, the classical and most common site for a venous stenosis is in the few centimetres that are downstream (in the direction of blood flow) from the venous needle return site. But, flow-related trouble can occur at several sites. So, let us think about the mechanical “hotspots” of dialysis for a moment.

Arterial Suction

During dialysis, the blood pump on the dialysis machine “sucks” blood out of the fistula through the arterial needle. The higher the pump speed, the greater the degree of “suck.” Veins are living, moving, delicate structures, and sucking blood from them too viciously can lead to vein deformity, vibratory harmonics, and vein wall damage at the arterial site. This is often shown as a falling arterial pressure. Note that arterial pressure is always a negative “suck” pressure. If it gets to be too negative, an alarm will sound to alert that the fistula cannot supply enough blood at the pump speed chosen. It is saying, “Slack off Buddy, you are pulling blood out too fast.” So, if the pump speed is too fast, damage can be done to the vessel lining (the endothelium) at the arterial end, causing suction-related injury.

Pump Trauma

Blood is drawn from the arterial needle by a blood pump. This is commonly a rotary (roller) pump—though the new Quanta SC+ uses an innovative hydraulic system. Watch your roller pump. Watch it squeeze the blood in the lumen of the plastic pump insert forward. Watch as it goes round and round, squeezing and “massaging” the blood in the lines forward, in a pulsatile flow, towards the dialyser. The lines jiggle and wiggle as it does.

Haemolysis—red blood cell destruction—has been quite frequently described with high pump speeds. This is true not only in dialysis, but also with high pump speeds used by heart bypass systems during modern cardiac surgery. See Twardowski et al. The higher the pump speed, the greater is the threat of red cell trauma. The red cells can be literally pummelled till they burst, worsening anaemia and releasing free haemoglobin into the bloodstream. Not nice! High pump speeds can thus cause issues at the pump as well.

Venous return

The main mechanical issue that results from high pump speeds is at the venous return site. Not only is there (unavoidable) vein wall trauma from the insertion of the venous return needle but, the returning blood creates turbulence as it “squirts” back into the vein from the venous return needle. Simply—and it is simple—the higher the return “squirt,” the greater the turbulence caused, the greater the non-harmonic vibration against the vein wall, and the greater the disturbance of normal “laminar” blood flow. (NOTE: laminar flow is described by Wikipedia as… “fluid flowing in parallel layers, with no disruption between the layers. At low velocities, the fluid tends to flow without lateral mixing, and adjacent layers slide past one another like playing cards.”

*I will purposefully avoid discussion of the pros and cons of ladder vs buttonhole needling—that is another discussion altogether…

Back to our example of a stream, but now think of two streams joining to form a larger river. If the flow is gentle, imperceptible mixing of the two bodies of water occurs, and all is quiet. But, if one stream is in raging flood, it will cause flow disruption, whirlpools may form, and eddies will ripple the surface. The higher the entry speed of the laterally joining stream, the greater the turbulence, and the further down the river it will extend, until all becomes quiet again and normal flow patterns are restored. So it is with a high venous return. The higher the flow, the more deforming will be the flow effect. Again, intimal anger results! The endothelium proliferates, narrows the vessel, and stenosis results. Our road authorities often—and quite rightly—use the catch phrase “speed kills” in an attempt to minimise the road toll. In dialysis, we could use a similar warning: “speed kills,” as we attempt to minimise fistula damage.

OK, I hear some who use 400, 450 (or, ye gods! 520) ml/min pump speeds saying, “But, my fistula is OK,” or, “I don't have stenosis,” or, “I don’t get arterial alarms.” Well may that be true—we are all built and react differently. But DOPPS published their conclusions a few years back, and here is an abbreviated version (by me) of their abstract.

FISTULA PROCEDURE RATES AND PRESCRIBED DIALYSIS BLOOD FLOW RATE: FINDINGS FROM THE DIALYSIS OUTCOMES AND PRACTICE PATTERNS STUDY (DOPPS)

Introduction and Aims: We postulated that dialysis facilities that routinely prescribe lower blood flow rates would have lower fistula procedure rates.

Methods: Fistula procedure rates were calculated as total numbers of fistula procedures reported during follow-up divided by fistula follow-up time during the study. Procedures included angioplasty, stent, surgical revision, banding, thrombectomy, thrombolysis, and other. Blood flow rates were divided into low and high within each region based on the regional median: 200 in Japan, 300 in Europe/Australia/New Zealand, and 400 in the US and Canada.

Conclusions: Fistula procedure rates were higher among patients in facilities with higher blood flow rates in North America, Japan, and E/A/NZ.

For those who use high flow rates and have thus far escaped trouble…good for you. But, perhaps, for you the bell is tolling! Just pause for a minute to look at the flow rates (above) in Japan, ANZ, Europe, and ponder the very different blood flow rates across the world. Dialysis survival in the US is worse than in any other western nation. It is far too long a bow to blame blood flow rates as the only reason for this, but there are data that link this as a contributing factor.

Next, high(er) blood flow rates are encouraged for those who use the NxStage system. Indeed, it is actually essential to run high(er) blood flows if using the NxStage to compensate for the low dialysate flows this system uses. I have discussed the concept of the ratio of blood flow to dialysate flow elsewhere. But, 400? 450? Or (ye gods) 520 ml/min? Why? The usual conventional dialysis systems—Fresenius, Baxter/Gambro, Braun, Belco, Nikisso, and other systems—all depend on a blood flow to dialysate flow ratio of around 1:2 That is, a blood flow of 300 ml/min is paired with a dialysate flow of about 600 ml/min. Certainly, the blood flow rate is always significantly less than the dialysate flow rate.

When NxStage came along, they strove for a small, portable system! To do this, they had to reduce the volume of dialysate. They came up with the idea of reversing the blood flow to dialysate ratio. To downsize the dialysate flow, they had to increase the blood flow rate to compensate. They created the concept of the “Filtration Fraction” (again, see the blog reference just above) where the blood flow rate needs to be 2-3 times the dialysate flow rate to compensate for the less efficient low flow rate of the dialysate. Typically, this means attaining a blood flows rate of up to 400+ ml/min against dialysate flow rates of maybe 120 ml/min.

I won’t go over this again here, but this is a classic case of “robbing Peter to pay Paul,” or “giving with one hand, but taking away with the other.” To understand this better, I encourage you to read my previous blog on dialysate flow and study the graphs that come later in this blog. Dialysis duration and frequency are inextricably linked. In the end, you can’t have it both ways unless UNLESS you dialyse longer. I prefer, advise, and prescribe longer, slower, dialysis with fistula-gentle blood flow rates of 325 ml/min or less in our facility-based patients and, in our home nocturnals, a mean of 225 ml/min. But, that’s just me! To be fair, it is now possible to do longer treatments using NxStage with blood flow rates in the mid 300s—acceptable, though not ideal—as the dialysate flow rate remains limited.

Clearance

There is an erroneous notion out there in FaceBook-land and elsewhere in U.S. nephrology practice that by simply speeding up the blood flow rate, better clearance will result. Well, I am sorry to disappoint, but…no! Or, so minimally and uselessly so that making that choice will certainly risk damage to the fistula (see above), yet only marginally (if at all) improve clearance. I have copied the following graph from a paper by Tom Golper et al in AJKD. I have been naughty, Tom, and have not asked your permission, but I don't think you will mind, and the graphs are quite generic, with similar versions having been out for years. Take a moment to study this with me. It is a graph that shows the clearance from blood of various “nasties,” as measured against the blood flow (read pump speed).

Look first at the black line extending upwards to the urea curve from the horizontal axis (the blood flow) at a blood flow rate of 300 ml/min. Where it hits the urea curve, you can draw a red line across to the vertical axis (clearance). It hits the axis at a clearance of about 200 ml/min. Then increase the blood flow to 500 ml/min and redraw the black line (2) and its corresponding red line. There is a big jump in blood flow (on the horizontal axis) yet a minimal improvement in clearance (on the vertical axis). Considering the potential fistula damage from turbulence by increasing the blood flow, yet the tiny increase in clearance, is the trade-off worth it? I think not!

Now, look at the curves for clearance by dialysate compared to dialysate flow.

Clearance (the red line) at 120 ml/min dialysate flow rate (similar to NxStage) is quite low. Clearance is significantly improved if the dialysate flow rate is increased to 300 ml/min, and even more (>50% more than those of the low flow state) if the dialysate flow rate is increased to 500 ml/min. This shows why conventional systems still prefer to offer a high(er) dialysate flow rate coupled with a low(er) blood flow rate. Importantly, as a lower blood flow rate is also sympathetic to the fistula, one might reasonably argue that this is better for the fistula, and equal for clearance. So, NxStage has brought a smaller, more mobile system to the market = good. However, this has been at some cost! The cost is in clearance, and potentially in fistula preservation.

Small Solutes vs. Middle Molecules

I have dealt with this topic extensively in a number of past blogs at KidneyViews. It is easier – as I tire – to refer you there, rather than reiterate here. I suggest you read:

Summary

Life is full of trade-offs, and this is yet another. At the end, blood flow rate is a choice you must make with your dialysis team. But, as is sometimes noted at the HDC FaceBook page, dialysis education even for professionals, especially perhaps in the US, remains less than ideal. This means: YOU must learn. Learn ALL you can. Learn without passion or emotion. In dialysis, chose your path with care and with knowledge, and not with emotion. I hope this has helped fuel your learning.

References

For the sleuths among you … here are some references to aid a deeper understanding … though for most, I suspect, they will simply make your eyes glaze over!

Comments

  • Cheok PP

    Oct 30, 7:34 AM

    Thanks , Dr. Agar for giving more information on fistula and high blood flow effect. I learned something new from you.

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
  • Jermaine Mintuck

    Oct 29, 11:17 PM

    IDK how I managed to dodge the warning good, but my fistula works BEAUTIFULLY for over a year so far and I run at a European 300, not a crazy high speed, like I have seen others. I am doing great. I was surprised that 300 seemed to be fair. I am in Canada, and their usual seems to be 400+ and all the way up to 500+ .

    Scared of situations like stuff happening to fistulas and risk of hemolysis. None of this seems to happen to me. Glad to find that speed of 300 is very good, not crazily slow or insanely fast. Wouldn't want to speed up. Nope. Just where I am at now.

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
    • John Agar

      Oct 30, 4:30 PM

      Dear Jermaine ... nice to hear you are from Canada. There is clearly a difference between the Qb run for home overnight frequent nocturnal patients ... here, we run a mean somewhere between 225 and 250 (not greater) ... and in-facility 3 x week 'conventional' dialysis (here I think the sweet spot is around 300-325 but rarely more and not more than 350 - though conventional centre-based haemodialysis in Australia is most commonly 4.5-5 hours per session. I think it sounds like you are doing just fine.

      Reply to this Comment

      * All fields are required.

      1. Your email will not be displayed publicly
  • John Agar

    Oct 29, 7:47 AM

    Most of the interplay from this post seems to have been at the HomeDialysis Central FaceBook page. Many of the responses have been from NxStage users quoting their blood glow rates in excess of 400 ... and many higher.

    My response to that group is as follows ...

    “I wonder if some misunderstood the basic physiological reversal between blood flow and dialysate flow that I have tried to describe ... that conventional systems (ie: all EXCEPT the NxStage system) allow, permit, use, encourage, depend upon a dialysate flow rate of about twice that of the blood flow rate. That means a slow blood flow rate and a fast dialysate flow rate. That means conventional dialysis can accomodate slow(er) blood flow rates .. something I personally like and encourage.

    But ... because the NxStage system significantly limits the dialysate flow rate (by its bag system, its PureFlow, indeed both) it has compensated by encouraging short hour treatments at a much higher blood flow rate. They have encouraged this pattern in an effort to ensure sufficient clearance. They have reversed the conventional paradigm ... and have opted for a very high blood flow rate of 2-3 times the dialysate flow rate in order that the dialysate volume can be reduced. A low dialysate volume was seen by NxStage as the key to portability (and correctly so) but the cost has been the flow reversal known as the filtration fraction. This has driven up the blood flow rate, something that I have argued in this blog as being an at least ‘questionable’ trade off.

    So, for those who have quoted back to me at the FaceBook site blood flows of 400+ are ... yes ... doing what is needed to ensure adequate clearance. But I would argue that while this flow reversal practice is necessary for the short hour treatments so common with the NxStage prescription, that it isn’t necessarily an ideal outcome. Rather, it’s just what’s required.

    Some use the NxStage rather differently ... and use it for longer, quieter sessions = something I applaud. Some run near 8 hr runs at very low dialysate flow rates (but also at gentler blood flow rates) ... a way of using the NxStage that permits fistula-happier flows yet far better phosphate-clearing duration.

    If I were ever to use NxStage for our patients - which I don’t - that would be my prescription and my recommendation.

    To be honest, I abhor short hour treatment. Here in Oz, I much prefer to opt for and prescribe what I believe is better dialysis for our many home patients ... ie: 7-8 hrs x alternate nights at a minimum or, better, 4-5 nights a week for most ... even if it means using a more cumbersome system like a Fresenius 4008S = a system that permits unlimited duration with good dialysate flows coupled with low blood flows.

    OK ... I think differently to those in the US, who prescribe as they prescribe. I disagree with that dominant yet peculiarly US-centric prescription of short hour, high blood flow therapy ... but, hey, that’s just me ... and maybe I am wrong!

    Those who run those (in my view) ultra-high blood flows must do what you think right ... but, I ask that you do consider the consequences too. I can’t put it more plainly. I think that practice is wrong.

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
  • Chris Walton

    Oct 29, 3:22 AM

    Good you wrote this and very interesting. I was “told off” by the manager the other month for running at 420 when my prescription was 350. So for one month I ran at 350 and waited for blood results. My URR dropped from .78 to .66.
    My nephrologist increased my pump speed prescription to 420. I do 3x4hrs on a 5008 and I buttonhole 14g blunts, my fistula is elbow and very well developed over 18 yrs.
    I feel therefore this report does not apply to me but am intrigued why my clearance dropped so much for the month at 350 speed as according to the graph it should have been much better

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
    • John Agar

      Oct 30, 5:04 PM

      Thanks for the comment Chris ... and I share your perplexity at the rather large drop in your URR (or PRU) from 78% to 66% with a change from 420 to 350. I would have expected it to have fallen, but not into the <70% range. Other things beyond pump speed affect urea clearance ... for example, the inter-dialytic period of the week used to do the calculation, the urea load for that period = (in part) nutritionally dependent, the protein catabolic rate ... and other stuff. But, yes, the drop was more than I might have expected. Going back to the curve for blood flow alone (above), you are playing with a fairly flat part of the clearance advantage curve at these blood speeds, and while some reduction in clearance would be expected, not a whole 12 percentage points. So ... hmmm. And, do I have a ready answer? ... nope.

      Reply to this Comment

      * All fields are required.

      1. Your email will not be displayed publicly
  • Terry Chamney

    Oct 27, 8:15 PM

    Great Article John. As a layman it made me understand the whole speed thing. My wife has always run at a speed of 300 with an upper arm graft as this was ( we were told ) her only option. The graft has since failed after about 3 interventions and they tried the Radiocephalic Avf which we have recently been told will not mature, so now are looking at creating a Brachiobasilic Avf I think it is. Thanks for the link from Mary Hammes, it was very interesting. I have one simple question, If my mathematics is correct, if I have 5 litres of blood in my body and run at 300 my complete blood supply will run through the dialysis system twice. Is that correct?

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
    • John Agar

      Oct 30, 4:45 PM

      Terry ... the 300 = 300 mls per minute ... so in pure maths terms, that's 300 x 60 ml = 18,000 ml = 18L per hour passes through the dialyser. That means [notionally] that all your blood volume (assuming it is 5L) traverses the dialyser 3.6 times each hour, or, for a 'standard' 4 hr dialysis, 14.4 times (say 15 in round figures) in each dialysis session. I say 'notionally' as of course not every drop from every nook and cranny will necessarily pass through the extra-corporeal circulation, but you get my drift.

      Reply to this Comment

      * All fields are required.

      1. Your email will not be displayed publicly
  • William Withers

    Oct 27, 7:17 PM

    Hi John, you mention a figure of 225ml/pm for nocturnal patients. Would there be anything wrong with 100/120 on single needle?

    My original rationale was that this is close to the body's own speed, so less trauma and my results remain good and unremarkable in the sense that they never change.

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
    • John Agar

      Oct 30, 4:56 PM

      Hmmmm ... William ... I don't think that's quite right, mathematically at least. Though a lower flow rate through a single needle should, in theory, reduce the risk of any re-circulation, it should only make a difference of a few percent (if that). Slowing the draw speed DOWN from 225 to 120 or less would be impacting the steepest slope of the clearance curve ... at least for urea - though much less so for the middle molecules (see the blood speed curves for urea and other substances above) ... and have a proportionately greater effect on clearance than increasing the blood speed UP, where the clearance rates have flattened right off and less and less relative clearance accrues to each hike up in blood speed.

      Reply to this Comment

      * All fields are required.

      1. Your email will not be displayed publicly
  • Robert Siebenaler

    Oct 27, 11:12 AM

    I've been doing this over 10 years in center. I stick myself I use a 14 gauge buttonhole needles and my flow is 500-550 for 3 hours. NEVER had a problem. Also I workout 3-5 times a week. I believe everyone is different and this works for me.

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
    • John Agar

      Oct 27, 5:41 PM

      As I said in the text of this blog, Robert .....

      OK, I hear some who use 400, 450 (or, ye gods! 520) ml/min pump speeds saying, “But, my fistula is OK,” or, “I don't have stenosis,” or, “I don’t get arterial alarms.” Well may that be true—we are all built and react differently.

      ...... but regardless that YOU are OK, the principle stands - and applies - for the majority. I am glad all is still OK, for you as an individual, but this blog raises principles that apply to the many ...

      Reply to this Comment

      * All fields are required.

      1. Your email will not be displayed publicly
  • Nieltje Gedney

    Oct 27, 11:01 AM

    Great, great analogy John about the river, whirlpools, etc...I have used Dori's analogy of the firehouse (running at 500) vs the water house on low! I, for one, cannot stress the importance of lower BFR's, and appreciate so much your summarizing this for everyone again. Thanks for all that you do to keep us educated!

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
  • Tami Ramsey

    Oct 27, 9:10 AM

    Thanks, Dr. Agar for such a fine description and explanation of blood flow rates. It's so critically important in the life of a fistula and I hope more people read this!

    Reply to this Comment

    * All fields are required.

    1. Your email will not be displayed publicly
    • John Agar

      Oct 27, 5:42 PM

      Thanks Tami

      Reply to this Comment

      * All fields are required.

      1. Your email will not be displayed publicly
Leave a New Comment

* All fields are required.

  1. Your email will not be displayed publicly