Incremental dialysis (iHD) is a term that is beginning to appear more frequently in the haemodialysis literature and, while I am not a particular fan of the practice—not for theoretical reasons, but for practical ones—it has been the subject of some recent papers from some recognised international dialysis heavyweights. These have included Andrew Davenport and Ken Farrington from the UK, and Tom Golper from the US, and whatever they say deserves due attention and consideration.
iHD has been promoted most aggressively by a Southern Californian team led by Kaymar Kalanter-Zadeh. While iHD does have some sound dialysis theory on its side, I still cannot bring myself to agree with it, or with those who proselytize it. Though I will come to the reasons why, first let me examine the underpinning theory behind iHD. As a key academic dialysis group from the UK including Andrew Davenport and Ken Farrington correctly note in their 2015 NDT paper “1”: “…the measurement of residual renal function is an accepted practice for peritoneal dialysis patients and the importance of residual renal function in determining technique success is well established …but few centres routinely assess residual renal function in haemodialysis patients”.
The UK team go on to correctly note that,“intra-dialytic hypotension and episodes of acute kidney injury (i.e., episodic renal “stun,” similar to the now well-described myocardial “stun,” “may lead to an earlier loss of residual renal function,” and they conclude that: “an incremental approach to the initiation of dialysis may potentially provide some haemodialysis patients with an improved quality of life and greater preservation of residual renal function, whilst fewer dialysis sessions may reduce health care costs.”
But…the devil is always in the details! The authors also warn—wisely—that… “an incremental approach requires careful monitoring of residual renal function,” and further, that: “once residual renal function has been lost, dialysis prescriptions should-must be re-examined to consider the use of longer or more frequent treatment sessions, and switching from low to high-flux dialysis or haemodiafiltration to offset the retention of middle sized molecules and protein-bound solutes.”
Another dialysis luminary for whom I have the utmost respect is Tom Golper. He published a paper “2” noting that: “incremental hemodialysis is not widely used nor is it well understood”. He also correctly observed that, “incremental hemodialysis is critically dependent on the amount of residual kidney function as well as the individualized goals of end-stage renal disease management.” But, importantly, he observed that, “residual kidney function must be assessed frequently and dialysis adjusted accordingly”. While I particularly liked his next statement that, “home HD lends itself to an incremental approach more so than in-center HD”, and think that this message is well worth serious thought, especially from those well-attuned to home therapies, Tom goes on to say, “I have had a long and successful experience in performing incremental dialysis (both peritoneal and hemodialysis),” and in the paper “shares practice strategies and approaches” to incremental therapies.”
Implicit in this last thought thread is that the application and implementation of iHD is for experienced and successful home therapies exponents. So, why do I hesitate when respected dialysis clinicians like Davenport, Farrington, Golper, and others support the concept? Well, while I understand and fully concur with the theoretical ideas behind iHD, I believe the wide adoption of iHD across an uncontrolled patient population in a wide cross-section of dialysis services/clinics would prove a dialysis disaster. There are several reasons why I say this:
- Central to the concept of iHD must be the regular, repetitive (and accurate) monitoring of residual renal function (RRF), and an accurate and reproducible measurement of RRF is NOT easy, especially in patients who are in a non-steady state, simply because of the intermittent dialytic component super-added by their treatment. Measurement of GFR at very low levels of renal function is inexact at the best of times, but add the near impossibility of careful measurement of native GFR in a patient also receiving concurrent dialysis and is thus in a permanent state of volume and biochemical flux, and a very complex assessment results! The summation of native GFR in dialysis-induced flux—and thus fluctuating—and dialysis-generated solute and volume clearance generates a nigh on impossible mathematical conundrum. Yet, this composite would need to be (a) assessed (and I can't see how), and (b) regularly re-measured to plot the inevitable background decline of native GFR and corresponding in-parallel rate at which to upscale weekly total membrane contact time as the native clearance component declined…While this may be an assessment within the capacity of some academic services, to think that this could be done widely across all services as each new patient entered his/her dialysis journey is, I am afraid, not likely to happen. Regarding RRF, current evidence suggests that the rate of fluid removal during dialysis is a key determinant (indeed, perhaps the key determinant) of the rate of RRF loss. Each episode of dialysis that imposes an excessive rate of volume removal…ergo drops circulating volume…ergo drops organ perfusion (myocardial stun is well understood but so, now, too, is there evidence for reduced renal perfusion with each episode of excessive UFR)…ergo inducing recurrent episodes of mini-AKI…contributes to an accelerated decline in residual GFR. This is the key reason for the more rapid loss of RRF seen in HD relative to the gentler decline seen in steady-state PD. For those on HD lucky enough to have a zero or near-zero sum volume status, then the HD regimen—whatever it may be—is unlikely to accelerate the loss of RRF at a rate greater than the rate of loss dictated by the ongoing, background native disease. A residual urine output does help—particularly regarding volume—and those with residual volume rarely need more than minimal amounts of volume removed by the dialytic process. But, the maintenance of a urine volume does not necessarily secure adequate solute clearance, and urine volume alone is not sufficient as a measure of RRF.
- Even more than the issues around residual renal function and its measurement, the natural reluctance embodied in human nature will make most—if not all patients—argue, resist, and down-right refuse to agree or accept the timely incremental increase in dialysis time and frequency that would need to accompany the inevitable decline of native GFR. This, to me, is the true Achilles heel of the iHD argument. Not only will most patients strongly resist any sequential increase in their dialysis frequency and time, but there maybe a parallel reticence on the part of services to lengthen and increase dialysis time, too. The complex logistics of varying sessional hours and frequencies would prove a nightmare for services to embrace and implement. Many simply would not do it…or would implement far too late, allowing patients to drift ever further into under-dialysis…as if this isn’t already a huge underbelly problem, especially in the US. Moreover, in the US, where funding streams are difficult enough to follow already, the complexity that iHD may add to billing and charging may be yet another disincentive—for patient and centres alike—though I dare not even peek into that Pandora’s Box! For all theses reasons and more, I think iHD, if applied widely, would be a recipe for worse dialysis, and not better. Let us say that someone starts dialysis twice a week, for 2-3 hours per session to augment a native GFR of 8 or 9. After 2 or 3 months, just when that patient has adjusted to a regimen she or he hates, but has finally (and grudgingly) come to accept, the nephrologist waltzes up and says, “hmmm, my fuzzy formula (and one that I cannot rely on) tells me that you now need 3 treatments a week, not 2, and oh—by the way—your sessional length will need to increase to 3.5 hours.” I know exactly what would come back from the patient. “Why? I don't feel any different! Sod off! I am not going to do a moment longer than I already am!” Try, then, also saying that she/he really needs 4.0, 4.5 or 5.0+ hours, and the retort will inevitably be “No!” It just won't work! Maybe the occasional very disciplined patient will understand and/or accept, but for most it will be a flat no! Human nature is hard (if not impossible) to change once a precedent of lesser sessions or hours has been set.
- Finally, I (we) already do introduce dialysis incrementally, though not quite for the same reasons, and I (we) always have. While I admit that I do always start patients on a 3 x week regimen, the first few weeks are commonly iso-volaemic, while dialysis time is truncated at the start to 2.0-2.5 hours, then in the second week = 3.0-3.5 hours, etc, until a full 3 weekly sessions of 4.5 to 5 hours is achieved. OK - that is clearly not what is meant by iHD, but it is a form of mini-incremental start! While I know of no data to support my next suggestion—though perhaps there should be. Rather than fight the patient again and again over sessional frequency and duration, if an incremental start is planned, perhaps it should be governed by incremental settings, rather than by incremental duration and frequency. The patient starts at full frequency and duration from the get-go, but the dialysis itself is made subliminal by using lower pump speeds (150-200), iso-volaemic or ultra-low ultrafiltration with minimal volume removal, smaller and perhaps even low flux-dialysers, with the adjustments over time being setting changes, not time and frequency changes.
The comment from Tom Golper that home haemodialysis lends itself best to incremental dialysis is true—save for the training period. To train for home HD, a relatively intense and “full-on” training period lasting (in our country, at least) 6-8 weeks, or even significantly longer, could significantly impede any intent to offer less, rather than more, dialysis. Intense, frequent, careful, and as-long-as-is-needed training—our current model for home training—would diametrically oppose an incremental, less is better approach.
Should our training model change? Possibly, but we have come to believe that good training is long, slow, and often training. Training in the home might help serve both masters, but training 1-2 times per week for short periods, the introductory iHD model, would, in my view, be poor and set-to-fail home training. So, even initiating iHD at home presents its conundrums.
Though I will not dwell on peritoneal dialysis, iPD is far, far easier. iPD has long and commonly been used—slowly ramping up PD exchanges from 1 or 2 exchanges per each 24 hour cycle to a full overnight APD regimen as the far-more-easily performed Adequest assessment (i.e. Peritoneal Equilibration Test) of native clearance dwindles. So … while iPD is easy, iHD is not.
I could say a lot more, but that is likely more than enough.
So…back to the start! While there is a theoretical argument—though not a particularly strong one—for some devotee and/or largely academic services to “play with” the iHD models, in my view the practical application would be a “dogs breakfast,” and would lead to comprehensively worse dialysis for all but the occasional, highly disciplined, well-informed patient who understood the residual RRF/dialysis interface as well (or likely far better) than I do.