An overview of the modalities of renal replacement therapy

This chapter is vaguely related to the aims of Section H2(i) from the 2023 CICM Primary Syllabus, which expects the exam candidate to "describe the principles of dialysis and filtration". Among the things which could be listed under the heading of "principles of dialysis", one could certainly find space for the modalities of renal replacement therapy. 

The purpose and position of this summary is questionable. The primary exam candidate may safely overlook this material, because the college has never asked about it in the historic Part I papers and therefore anybody concerned with actually passing that exam would be well advised to ignore all this superfluous unexaminable material. The Fellowship exam candidate may be more interested in modalities of dialysis because the college frequently asks about this sort of stuff. A comparison of CRRT modalities was asked about in Question 21 from the second paper of 2015, Question 10 from the first paper of 2011, Question 19 from the second paper of 2008 and Question 2c from the second paper of 2001. However, the Part II candidates are likely looking for a shortform tabulated revision resource such as the comparison of renal replacement therapies available in the Required Reading section. Ergo, neither the CICM Part I nor Part II exam candidates would be served by a long confusing digression on RRT, particularly one which strays so far into historical and theoretic territories. It would perhaps be simpler to acknowledge that this chapter is the entirely self-indulgent product of an author who is not accountable to his readers.

Among authors who do have some sort of obligation to meet learning objectives or fulfill the demands of financial institutional affiliations, there are several excellent resources. Perhaps the best, though many years out of date, is the 1996 paper by Bellomo and Ronco, who attempted to introduce some standards to at least the field of continuous RRT, where in the nineties terminology proliferated with uncontrolled enthusiasm. Around that time, lots of papers were being published which referred to the same dialysis techniques by totally different names. The descriptive inconsistencies in the contemporary literature were bringing discredit to the field of critical care nephrology, it being "scientifically undesirable" to be "bewildered by the everchanging array of names and abbreviations". To address this lunacy, at the 1995 International Conference on RRT in San Diego a bunch of nephrologists took to the stage and wrestled the audience of intensivists until a consensus formed for CRRT nomenclature. It did not stop people from inventing random names for their modalities, but it did help make them look more silly. Intermittently, experts still meet to refine the definitions, but now the pursuit of standardised terminology has led to the next extreme, one of absurd granularity. The 2016 consensus statement published by Villa et al includes strict prescriptive recommendations for colour-coding of circuit tubes and naming of every little detail of machine operation (eg. collecting the equipment required to start dialsys from the store room is officially called the preparation phase).

In brief, modalities of RRT can be classified into groups on the basis of the manner of access, the duration of therapy, and the dominant mechanism of solute clearance. 

  • Duration: intermittent, continuous or "hybrid"
  • Access: arterio-venous, veno-venous
  • Mechanism: diffusional dialysis, ultrafiltration, or both (this is also occasionally referred to as the "mode" of RRT)

The combination of these three parameters describe the numerous permutations of RRT. Excluded from this list is peritoneal dialysis and blood purification therapies such as adsoprtive haemoperfusion and plasma exchange. They are excluded for a number of purely arbitrary reasons, including that some (PD) are not usually commenced in the ICU, whereas others are not "renal replacement" by the definition that they do not perform a function which the kidneys would have otherwise performed, had they not failed. ("renal replacement therapy" being defined as "any extracorporeal blood purification therapy intended to substitute for impaired renal function"). 

Continuous, intermittent and hybrid options

The CRRT, IHD and SLED descriptors refer specifically to the duration of the renal replacement treatment. These modalities differ in their efficiency of solute and water removal; the more efficient therapies move more molecules but also result in fluid and solute shifts which may rapidly empty the patient's circulating compartment, resulting in haemodynamic instability. 

Continuous renal replacement therapy (CRRT) replace renal function at a relatively slow pace. The rate of solute and fluid removal is sufficiently slow that even haemodynamically unstable patients can tolerate the transcompartment water shifts. The therapy is relatively ineffective at replacing renal function (i.e. nothing is as good as a real kidney) but because it runs continuously it is capable of doing enough kidney-like work to mitigate the complications of renal failure. Crudely, 24 hours of CRRT can be said to replace 24 hours of normal renal function. Apart from the slower rates of flow, CRRT is usually dependent on the use of premixed sterile dialysate and replacement fluid.

Intermittent haemodialysis (IHD) is the much faster alternative. It does 72 hours worth of kidney work in 3-4 hours. As the result, fluid and solute shifts between compartments are quite vigorous. The haemodynamic consequences of IHD often make it unsuitable for critically ill patients, and poor cardiovascular fitness is a strong argument for not offering such dialysis support in the long term.

Apart from faster flow rates, a key feature distinguishing IHD from CRRT is the origin of the dialysate fluid. Unlike CRRT which usually relies on pre-mixed bags of sterile dialysate and replacement fluid, IHD typically uses "online" water, which is ultrapure but not sterile. This water comes from a reverse osmosis purification plant which separates water from its solutes by forcing it through a semipermeable membrane. As it is administered into the circuit, the water is mixed with some solutes (sodium chloride, bicarbonate, potassium etc) to more closely approximate plasma. This is probably the only way of producing sufficient volumes of dialysate for IHD, because massive numbers of people require intermittent dialysis for ends stage renal disease and it would be totally impractical to store thousands of litres of sterile dialysate in bags (and thousands is a conservative estimate, each four-hour session of IHD requiring about 120L of dialysate). Strictly speaking, the dialysate never comes in direct contact with the blood (they are separated by a membrane) and therefore does not need to be sterile - "ultrapure" is good enough. Practically, the use of online dialysate is not without adverse consequence (Tetta et al, 2006), and the piped systems become contaminated often enough to occasionally take out a whole unit for weeks or months. The practice and theory of online fluid production is extensively described elsewhere, and it will suffice here to offer this excellent 2002 article by Ingrid Ledebo as a reference.

Sustained low-efficiency dialysis (SLED) is a "hybrid" therapy. "Hybrid" in this context refers purely to the blending of timing characteristics between IHD and CRRT. SLED is not continuous, but a standard SLED session runs for longer than an IHD session. It is usually performed on a device which is identical to normal IHD apparatus, but using slower dialysate and blood flow rates. The technique is less efficient than IHD in terms of solute removal and ultrafiltration, in terms of millimoles and litres of water removed; however it is better tolerated than IHD precisely for this reason. However, it still moves more solute than CRRT and the more haemodynamically fragile patients will not tolerate it. Because this technique uses the same machinery and relatively high dialysate flow rates, use of premixed fluid is also impractical, and therefore SLED is reliant upon the availability of online dialysate.

SLED seems to have a massive number of subgenres. Villa et al lists "sustained low-efficiency dialysis (SLED), slow low-efficiency extended daily dialysis (SLEDD), prolonged intermittent RRT (PIRRT), extended daily dialysis (EDD), extended daily dialysis with filtration (EDDf), extended dialysis (ED), “go slow dialysis”, and accelerated veno-venous hemofiltration (AVVH)". These are all essentially describing the same thing, and no further effort will be wasted on discussing them.

Arterio-venous versus veno-venous renal replacement modalities

The "VV" in "CVVHDF" describes the source and destination for the filtered blood. I.e "VV" is veno-venous; the blood comes out of a vein and returns into a vein after it has passed through the filter. For continuous RRT, the veno-venous method is nowadays the standard. For many intermittent haemodialysis patients, the access point is an arteriovenous fistula, and therefore the modality is AV.

Historically, acute CRRT was also performed by the arteriovenous route. Because reliable roller pumps were not available, the patient became their own pump - arterial blood pressure was used to push blood through the circuit. That might sound barbaric, but it was in fact remarkably efficient. Kramer et al (1977) were able to remove 200-600ml by ultrafiltration, per hour - and that was in a bunch of heart failure patients, people whose circulation is poor by definition. Even when blood pumps became available arteriovenous dialysis remained popular because it was felt that the arteriovenous pressure difference was sufficient to push blood around in the circuit (ultrafiltration being produced by the suction of the effluent pump). Lauer et al (1983) is a good representative article on CAVH and CAVHD, demonstrating the currents of though in the early 1980s.

Obviously, access was a major issue. The very first dialysis treatments used improvised equipment. Haas (1924) used glass cannulae to access the radial artery, returning blood into the cubital vein. Kolff (1943) used standard venepuncture needles to access the femoral artery. For maintenance haemodialysis, Quinton et al (1960) developed an implantable arteriovenous Teflon shunt, which ultimately led to the development of modern AV fistula technology. For acute RRT, arterial access was usually achieved with 16G or 14G cannulae (similar to the access needles used for intermittent haemodialysis). These cannulae have theoretical maximum flow rates in the order of 220-300ml/min, but realistically the critically ill patient's arteriovenous pressure gradient never approached these values -  Lauer et al (1983) report blood flow rates in the order of 30-70ml/min. 

Haemodialysis, haemofiltration and haemodiafiltration

IHD, SLED and CRRT can be either purely reliant on diffusional clearance (haemodialysis), or on convection and ultrafiltration (haemofiltration), or on some combination of the two (haemodiafiltration). Most frequently, the latter is used. Haemodiafiltration is the best of both worlds, as there is both opportunity to efficiently move small molecules by dialysis and the option of removing middle molecules together with water. Strictly speaking, most intermittent "haemodialysis" is usually a combined treatment where some fluid is removed by ultrafiltration and some solute is therefore also removed convectively. 


Negi, Shigeo, et al. "Renal replacement therapy for acute kidney injury." Renal Replacement Therapy 2.1 (2016): 31.

Bellomo, Rinaldo, and Claudio Ronco. "Nomenclature for continuous renal replacement therapies." Critical Care Nephrology. Springer, Dordrecht, 1998. 1169-1176.

Villa, Gianluca, et al. "Nomenclature for renal replacement therapy and blood purification techniques in critically ill patients: practical applications." Critical Care 20.1 (2016): 283.

Burchardi, Hilmar. "History and development of continuous renal replacement techniques." Kidney international. Supplement 66 (1998): S120-4.

Kramer, Peter, et al. "Arteriovenous haemofiltration: a new and simple method for treatment of over-hydrated patients resistant to diuretics." Klinische Wochenschrift 55.22 (1977): 1121-1122.

Konner, Klaus. "History of vascular access for haemodialysis." Nephrology Dialysis Transplantation 20.12 (2005): 2629-2635.

Paskalev, Dobrin N. "Georg Haas (1886-1971): The forgotten hemodialysis pioneer." Dialysis and Transplantation 30.12 (2001): 828-832.

Kolff, W. J., et al. "The artificial kidney: a dialyser with a great area. 1944." Journal of the American Society of Nephrology8.12 (1997): 1959-1965.

Quinton, Wayne, David Dillard, and Belding H. Scribner. "Cannulation of blood vessels for prolonged hemodialysis." ASAIO Journal 6.1 (1960): 104-113.

Lauer, Allan, et al. "Continuous arteriovenous hemofiltration in the critically ill patient: clinical use and operational characteristics." Annals of Internal Medicine 99.4 (1983): 455-460.

Schiffl, H., and S. M. Lang. "Current practice of conventional intermittent hemodialysis for acute kidney injury." Indian journal of nephrology 23.6 (2013): 395.

Tetta, Ciro, et al. "Clinical effects of online dialysate and infusion fluids." Hemodialysis International 10.s1 (2006).

Ledebo, Ingrid. "On-line preparation of solutions for dialysis: practical aspects versus safety and regulations." Journal of the American Society of Nephrology 13.suppl 1 (2002): S78-S83.