From Blog to Dialysis Chair: Implementing Volume 101

This blog post was made by Dr. John Agar on September 23, 2015.
From Blog to Dialysis Chair: Implementing Volume 101

A few weeks ago, I sent Dori Schatell an email for a blog post that I initially called Volume 101.1 Please read that post first, as doing so will give you my rationale for a maximum limit on the ultrafiltration rate (UFR) of <10 ml/kg/hour, and thus help you understand this current, second post of the series: how we implemented volume 101 (rate-limited UFR) in our dialysis clinic.

Even knowing why I believe UFR should be limited, many may still harbour understandable scepticism about how this recommendation could translate into practice at the dialysis chair or bedside. We have begun that journey here in Australia.2 And, so too have John Burkart and Jimmy Pirkle at Wake Forest University, in North Carolina. Their work at Emory has already been published in abstract form.3 They have already shown how a rational attack on volume through the limitation of UFR can be made to work in the US context—and they have found it to be a whole lot easier than durational sceptics might think.

The steps we have taken here in Australia are as follows:

  1. We first determined the UFR patterns of our facility-based patients. We have now analysed 20 months of prospectively collected data, but for purposes of clarity, I will only include data for the calendar year – January 1st to December 31st 2014. In that time:

    • 141 conventional thrice-weekly HD patients completed the full 12 months.
    • Over this period, this group completed 14,701 dialysis sessions.
    • “Acute” (<3 months duration) and short daily HD patients were excluded.

    The UFR was calculated from weight (read fluid) removed during each HD treatment: Pre-dialysis weight (always known) minus post dialysis weight.*

    As the post dialysis weight is also always known, the UFR in ml/kg/hr can be calculated, once the duration of the HD session is known. And, sessional duration is known as it is a prescribed treatment variable. However, I use the word “variable” here in hope rather than reality, as the treatment time (other than at Emory) is rarely varied in the light of the volume to be removed—as it should be.

    Why Sessional Time Matters

    Remembering that 1 litre in volume = 1 Kg in weight, this example will show you why time is the key factor in whether a treatment is or is not safe. If pre-dialysis weight is 100 Kg and post dialysis target weight is 96 Kg, then 4 litres (4,000mL) will be removed:

    Shorter Treatment: If the intended HD time is 3 hours, the fluid to remove each hour is 4000 ÷ 3 = 1333 mL. This translates to 1333 ÷ 96 = 13.9 ml/kg/hour; a fluid removal rate in the maximum risk range. (See Flythe’s graph here.)


    Longer Treatment: If the intended treatment time is 5 hours, the fluid to remove each hour is 4000 ÷ 5 = 800 mL. This translates to 800 ÷ 96 = 8.3 ml/kg/hour; a fluid removal rate in the minimal risk range.

    It came as no surprise that the distribution of ultrafiltration rates for all 14,701 treatments in our clinic followed a normal distribution, or a bell-shaped curve. The UFR was low in some patients, high in others, but grouped towards the middle for most.

    Ultrafiltration Rate by Dialysis Session

    Of the 14,701 dialysis sessions

    • The mean UFR was 7.95 ml/kg/hour.
    • 3314 sessions (23%) had a UFR >10 ml/kg/hour.
    • 17% of all sessions had a UFR between 10 and 13 ml/kg/hour.
    • 835 (6%) had UFR’s in excess of 13 ml/kg/hour.

    On a pie chart, it looked like this … the ‘at risk’ UFRs depicted being orange and red.

    Ultrafiltration Rate Quartiles by Dialysis Session

    As the data may be better expressed as a median, rather than as a mean:

    • The median 12 month average UFR was 7.73ml/kg/hour.
    • 20 of the 141 patients (14%) had a 12-month median UFR >10ml/kg/hour.
    • 6 of the 141 patients (4%) had a 12-month median UFR >13ml/kg/hour.
  2. Identify patients with the greatest UFR numbers. Although the UFR was predictably greatest for dialysis treatments that followed the long break,** importantly, those with the greatest UFR numbers were easily identifiable: They were the 20 (14%) and 6 (4%) of patients with an annual average UFR in excess of 10-13 and 13 ml/kg/hour respectively.

  3. Focus on the patients with the highest average UFR numbers. These 26 patients (20% of the total patient pool) are the ones we have need to focus on—not the 115 patients with an occasionally excessive UFR >10, but with mean annual UFR targets within the “safe range.” Notably, 23% of sessions—but only 14% of patients—were “at risk”; the difference being largely accounted for by the intermittent occurrence of a high long-break UFR in a mostly-safe patient.

    We need to learn to walk before we can run!

    While clearly this needs far more research and quantification, we are able to put names to our 26 patients whose UFR exceeded 10 ml/kg/hour, and especially to the 6 who regularly exceeded 13 ml/kg/hour.


    If we were to extend the duration of dialysis in only the 6/141 patients (though, ideally of course, we should do so in all of the 26/141 who persistently breached the 10 ml/kg/hour safety limit), we would bring the UFR for all patients down into an overall “safe” range. This does not appear to be as disruptive as it we first feared it might. Not only do 26 have identifiably high UFRs but, relating back to the distribution graph, a similar number also have very low UFR targets.

  4. Consider changing session length. If we adjust session length to pair those whose treatment time might be safely made shorter (i.e., those who already have a very low mean UFR), providing, of course, that their Kt/Vurea (or PRU) permits this (and our data say that in the main it does) with those whose volume safety demands a longer treatment (i.e., those with a high mean UFR), the total daily service duration for the provider should not change.

    We just need to be smarter and more flexible.

    We need to move away from the one-size-fits-all approach and realise that all patients differ, and all need different schedules tailored to individual needs. Yet, we do not commonly do this most obvious of all things.

    As has happened in John Burkart’s service in North Carolina, using session length as the service variable rather than fomenting and perpetuating staff vs. patient argument and angst through accusations of non-compliance is a far more effective way to ensure (1) greater cardiovascular safety of our patients, and (2) a smoother, less symptomatic, and less nurse-interventional delivery of their dialysis.


Having said all that, there are traps for the unwary in the application of a flexible approach. I am, at heart, a believer in longer duration and higher frequency dialysis schedules. That is, after all, what we do here in Australia and New Zealand. So, the “rob Peter to pay Paul” concept has its risks; the main one being that those with the lowest UFRs may also be those with a residual urine output (see my recent blog here, “How much can I drink, Doc?”.

To risk another idiomatic expression, we must also avoid “throwing the baby out with the bathwater”! Before further abbreviating the duration of treatments for those with the lowest UFRs, we need to be sure that they are not then placed at risk for being “fluid safe”. There may also be unintended consequences for removal of slow-moving molecules—like phosphate—where dialysis time is crucial to ensure linear trans-membrane transit.

I may extend this series of “volume” posts to explore these issues, so rather than offering a detailed explanation now, I will address some of the risk-benefits of flexible dialysis scheduling in a future KidneyViews blog!

While clearly far more research is needed to validate the risk thresholds and adverse event profiles within individual and collective dialysis services, the very first step is to start collecting the data! Only then will we be able to tease out the risks and benefits around optimal fluid management and dialysis duration.

To enable this, UFR should—no, must—become a required key performance indicator (KPI). It should become a mandated responsibility for every dialysis service to collect these data. The data are not complex and they are right in front of us. But, we haven’t yet troubled ourselves to look.

Who knows…we might be surprised how simple it is to achieve volume safety in tandem with solute adequacy, if we but gave it our best shot!


  1. Agar JWM and Schatell D.
  2. Agar JWM. Personal Viewpoint: Limiting maximum ultrafiltration rate as a potential new measure of dialysis adequacy. Hemodialysis International. Published as a ‘early view’ at
  3. Pirkle JL, Hawfield AT, Russell GB, Burkart JM. The effects of limiting maximum ultrafiltration rate in an in-center hemodialysis population. JASN. 2012: 23: (Abstract) 6A:TH-OR026.

Other good reading

  1. Flythe JE, Curhan GC, Brunelli SM. Shorter length dialysis sessions are associated with increased mortality, independent of body weight. Kidney Int. 2013. 83(1): 104-13. doi: 10.1038/ki.2012.346. (Epub 2012 Sep 26, 2013).
  2. Flythe JE, Curhan GC, Brunelli SM Disentangling the ultrafiltration rate-mortality association: the respective roles of session length and weight gain. Clin J Am Soc Nephrol. 2013. 8(7): 1151-61. doi: 10.2215/CJN.09460912. (Epub 2013 Mar 14, 2013).

*Note that the post dialysis weight may or may not equate to the post-dialysis target weight, while target weight is only a clinical estimate of true ‘dry weight’. Furthermore, ‘dry weight’ remains a notional value that continues to elude established, accurate, reproducible, and all-service measurement, despite recent advances in body composition monitoring, bioimpedance spectroscopy and the like.

**NB: this culpable facility-based treatment anomaly could and should be abolished through a combination of political, administrative and funding courage, but this anomaly is for another fight, another day, and is not the focus of this current exercise.


  • Henning Sondergaard

    Sep 30, 2015 10:19 PM

    While I absolutely agree with the concept of fluid removal as a KPI, I do think that pushing longer, more frequent sessions will make the idea of slow UFR obsolete by its very nature.

    The true 'enemy' here is the practice of 3x weekly "death diysis" (the most apt term for the common in-center practice, IMHO)

    Now, how we change that is a great question. I think we need different approaches in various parts of the world. But the most important one as I see it, is to simply to shame practitioners for deliberately killing off their patients - because with the mounting evidence there is not much room for any other viewpoint.

    Is this a harsh way of seeing it? Indeed! But no harsher than knowing that my fellow patients are receiving sub-optimal and lethal care
    Reply to a Comment
    *All fields are required.
    Your email will not be displayed publicly
    • John Agar

      Oct 01, 2015 12:02 AM

      Dear Henning

      Of course you are right - and I think you will know well that I have campaigned ceaselessly for longer and more frequent dialysis for at least a couple of decades ... though most of that has fallen on deaf ears outside of ANZ. In addition, the weekend long break (the so-called 'killer weekend') is clearly and above all the beast we ought to slay ... though there has been no appetite shown to do this from providers of funders, both of whom would have to increase service availability, nurse hours, and funding envelopes were this to happen. We have tried - and failed - and while that doesn't mean we should stop explaining the sound reasoning for abolishing the killer weekend, you and I know in our heart of hearts that it is still but a distant hope.

      You will also know that here, we DO more than talk the talk - we walk the walk - with - in my own unit nearly 40% of all HD at home on long hour (8 hr) and high frequency (2 nights on, 1 night off) home overnight dialysis. And - across this antiopean pair of nations, ANZ has a mean centre-based duration of 275 minutes (in the US, I believe the mean is 211 minutes) with a national ethic for home where possible and the numbers at home to back up that ethic.

      Meantime, there is a problem. Meantime, we are desperately failing to understand and make inroads into the rate of volume removal on dialysis. We are ignoring it. We don't even collect the data that - for each and every tratment - is so easy to collect: (1) the pre-dialysis weight; (2) the post dialysis weight; (3) the duration of each dialysis ... data which allows the mean per-treatment UFR to be calculated. This shows - even in our longer-hour, facility-based program - that a significant number of patients exceed a UFR of 10 ml/kg/hr. We need to come up with (and quickly) a means to correct that. And, that IS possible ... with will, explanation, and logistically non-disruptive means ... while better longer-term solutions are found - such as rolling alternate day programs, and longer dialysis hours.

      I am not suggesting the argument I advanced in the blog as ideal. Not at all. It is simply a way to save the hearts and lives of fluid-depleted maxi-UFR dialysis patients while we continue to argue for rolling schedules and routine 4.5-5 hour treatments.
      Reply to a Comment
      *All fields are required.
      Your email will not be displayed publicly
Leave a New Comment
*All fields are required.
Your email will not be displayed publicly