Version July 2025
INTRODUCTION —
Among other vital functions, the healthy kidney excretes water, electrolytes, and organic solutes (such as urea, creatinine, and uric acid). A few simple calculations help illustrate the degree to which kidney replacement therapy (KRT) must assume this excretory role in patients with acute kidney injury. In patients with normal kidney function, the glomerular filtration rate is approximately 170 to 180 L/day, which is roughly the clearance of solutes like creatinine that are excreted primarily by glomerular filtration. Most patients with acute kidney injury can tolerate the reduced solute removal that accompanies a glomerular filtration rate of about 10 percent of normal (17 L/day or 12 mL/min) [1-3]. This is therefore a reasonable initial target for KRT among patients who are not hypercatabolic. In comparison, hypercatabolic patients may require more aggressive solute removal.
The mechanisms of solute removal by KRT, and the hemodynamic changes that occur, are reviewed here. The indications for and prescription of KRT in patients with acute kidney injury, concerns related to whether KRT delays the recovery of kidney function, and the effects of different KRT membranes are discussed separately. (See "Kidney replacement therapy (dialysis) in acute kidney injury in adults: Indications, timing, and dialysis dose" and "Dialysis-related factors that may influence recovery of kidney function in acute kidney injury (acute renal failure)".)
MECHANISMS OF SOLUTE REMOVAL —
For dialysis to remove a solute, it must be present in the circulation, small enough to pass through the pores of the dialysis membrane, and relatively unbound to plasma proteins. The process of solute removal during dialysis can occur by two different mechanisms: passive diffusion down a favorable concentration gradient from the plasma water into the dialysate, and convection during the ultrafiltration of plasma water across the membrane of the hemofilter into the ultrafiltrate. The frictional forces between the solvent (water) and solutes (called solvent drag) result in the convective transport of small- and middle-molecular-weight solutes (less than 5000 daltons) in the same direction.
The dialytic clearance of a solute is dependent in part upon size, with larger molecules (including those that are protein bound) being less efficiently removed. The sieving coefficient (SC) is a measure of a solute's filterability during ultrafiltration, being equal to the ratio of the solute concentration in the filtrate to that in the arterial plasma water [4]. The SC ranges from 0, for a solute that is completely rejected, to 1, for a solute that is freely filtered (such as urea and creatinine). The total clearance of a solute during ultrafiltration is equal to the product of the SC and the rate of fluid removal (the ultrafiltration rate). For a solute with an SC of 1, the concentration of the solute in the filtrate is roughly the same as that in the plasma water, and the solute clearance by filtration is equal to the net ultrafiltration rate. A more detailed discussion of these concepts can be found elsewhere. (See "Drug removal in continuous kidney replacement therapy".)
With some continuous kidney replacement therapies (CKRTs), such as continuous arteriovenous or venovenous hemofiltration, all solute removal occurs by convection since there is no exposure to dialysis fluid (dialysate) [5] (see "Continuous kidney replacement therapy in acute kidney injury", section on 'Definition of CKRT modality'). If, as noted above, the aim is a clearance rate of 17 L/day, then the ultrafiltration rate for freely filtered small solutes (eg, urea) must be at least equal to this level to achieve the desired rate of solute removal. Small solute clearance per unit time is much less with CKRT than with conventional hemodialysis, but this is counteracted by much longer durations of therapy with CKRT (eg, 24 hours versus 4 hours per day) [6]. Although loss of ultrafiltrate with the same concentration as the plasma will not directly lower the blood urea nitrogen (BUN) or plasma creatinine concentration, the administration of replacement fluid to prevent volume depletion will do so by dilution. Thus, during isolated ultrafiltration, there will not be a decrease in plasma concentrations of those solutes because the ultrafiltered volume is not replaced with a solute-free substitution solution.
In general, the clearance of a small solute with an SC of 1 cannot exceed the ultrafiltration rate with continuous arteriovenous or venovenous hemofiltration. This limitation can be modestly exceeded by approximately 15 percent if the replacement fluid is administered before the filter (called predilution). This will allow some of the intracellular urea in red cells to diffuse out into the plasma and be removed in the filter.
Solute removal is substantially faster (over 160 mL/min for urea) with standard hemodialysis, in which small solutes passively diffuse from the plasma into the dialysis fluid down a favorable concentration gradient. Blood and dialysate move at high rates in a countercurrent fashion through the dialyzer. The constant resupply of new blood with a high solute concentration and new dialysate with a low solute concentration maximizes the rate of diffusive loss by maintaining a high concentration gradient between the two compartments. The net effect is that the loss of small solutes is much greater than that of water.
Standard hemodialysis also can be used to remove fluid by increasing the transmembrane pressure gradient. This is essentially performed by pumping dialysate out of the dialyzer.
In stable patients with severe acute kidney injury, conventional hemodialysis is typically the KRT modality of choice. However, hemodialysis (for either acute kidney injury or chronic kidney disease) is frequently associated with episodes of hypotension, making it less useful in hypotensive or hemodynamically unstable patients. It is in this setting that CKRT may have an advantage since the risk of hypotension is much lower.
MECHANISMS OF HEMODIALYSIS-INDUCED HYPOTENSION —
An understanding of how hemodialysis causes hypotension allows us to appreciate the potential advantages of continuous kidney replacement therapy (CKRT). The major determinants of hemodynamic stability are the rate of net fluid removal and the rate of solute removal. It had been assumed that the primary cause of intradialytic hypotension was induction or exacerbation of intravascular volume depletion by rapid ultrafiltration. This simplistic hypothesis was disproved by the observation that patients who became hypotensive during hemodialysis tolerated the same degree and rate of fluid removal with pure hemofiltration, in which there was no diffusive component of solute loss [7-9].
The proposed explanation for this phenomenon is illustrated in the figure (figure 1). During conventional hemodialysis, the rapid diffusive removal of urea (and other small osmotically active solutes) results in a reduction in the plasma osmolality, which is now less than that in the cells. As a result, water moves osmotically into the cells, further depleting the extracellular volume, which has also been diminished by the extracorporeal ultrafiltration. It is also possible that the rapid fall in plasma osmolality itself contributes to the hemodynamic instability, perhaps by interfering with sympathetic responsiveness to volume depletion (see below). This mechanism may partially explain why decreasing the extracorporeal blood flow rate during hemodialysis improves blood pressure stability.
Urea has been a biomarker solute whose removal has been a part of the definition of dialysis efficiency [10]. Although urea is generally considered to be an "ineffective osmole" because it can cross cell membranes, this process requires several hours to reach equilibrium. Thus, urea acts as an effective osmole that can induce water shifts in the setting of acute dialysis. In addition to promoting the development of hypotension, these water shifts can also induce cerebral edema and the neurologic symptoms of the dialysis disequilibrium syndrome. (See "Dialysis disequilibrium syndrome".)
In comparison to hemodialysis, the fluid removed via isolated ultrafiltration has the same concentration of small solutes as the plasma water. As a result, no osmotic gradient is created between the cells and the extracellular fluid, and water shifting into the cells does not occur. To the contrary, the fall in intravascular volume will raise the plasma protein concentration and therefore the plasma oncotic pressure. This will draw water from the extravascular space and the cells into the vascular space, thereby leading to the relative preservation of intravascular volume (figure 2).
The etiologic importance of osmolar changes in dialysis-related hypotension has been suggested by additional studies [11,12] and the following observations:
●The intracellular fluid volume increases during conventional hemodialysis [13].
●Water moves from the extracellular fluid into the cells with hemodialysis but in the opposite direction during hemofiltration [14].
●Hemodynamic instability is improved with hemodiafiltration compared with hemodialysis [8,9].
Even when diffusive solute loss is allowed to occur during CKRT (by running dialysate on the other side of the dialysis membrane), hypotension is less likely because of the much slower rate of urea removal. The urea clearance during continuous arteriovenous hemodialysis (CAVHD), for example, is approximately 17 mL/min, depending on the operating conditions, as compared with values above 160 mL/min with conventional hemodialysis.
Hemodynamics of hemofiltration and hemodialysis — The net effect is that arterial blood pressure and systemic vascular resistance are usually maintained or even increased during hemofiltration or hemodiafiltration [8,9,15-18]. In comparison, both of these parameters tend to fall when a similar volume of fluid is removed by hemodialysis. In addition to the osmotic shifts, patients treated with hemodialysis tend to have more severe derangements in autonomic function [16,17,19]. How this occurs is not known, but venous vasoconstriction (which increases venous return to the heart) is an important mechanism to support cardiovascular stability during hemofiltration and may be impaired during hemodialysis [18,20,21]. (See "Intradialytic hypotension in an otherwise stable patient", section on 'Autonomic dysfunction'.)
The role of myocardial stunning in chronic hemodialysis is common but poorly understood [22]. It appears that neither dialysate cooling nor hemofiltration are particularly beneficial in reducing myocardial stunning [23].
Vasoconstriction due to lower temperatures with hemofiltration also may contribute to the difference in vascular reactivity between hemodialysis and hemofiltration. The role of temperature was addressed in a crossover study in which hemodynamic measurements were determined in 11 patients who underwent cold hemofiltration, warm hemofiltration, and combined ultrafiltration and hemodialysis [24]. When care was taken to insure that hemofiltration and combined therapy were performed at equivalent warm temperatures, no difference in blood pressure or reactivity was observed between the two modalities. By comparison, cold hemofiltration was associated with significant increases in blood pressure and vascular resistance when compared with warm hemofiltration and combined therapy.
CKRT may also protect systemic hemodynamics by a similar mechanism. The high rate of fluid removal (which can exceed 500 mL/hour) may require the administration of a substitution fluid. When given unwarmed (ie, at room temperature), the ensuing cooling of the patient can lead to cutaneous vasoconstriction and preservation of the central blood volume and blood pressure. (See "Intradialytic hypotension in an otherwise stable patient", section on 'Second-line approach'.)
It must be emphasized, however, that the protection afforded by hemofiltration is relative, not absolute. Hypotension can still occur if too much fluid is removed or if fluid is removed too rapidly.
Other factors — A variety of other factors can also contribute to hemodialysis-induced hypotension in selected patients. These include bleeding induced by the heparin given to prevent clotting in the dialyzer, the use of acetate rather than bicarbonate as the dialysate buffer (now a rare problem), certain medications, underlying heart disease, and, in relatively healthy patients, food ingestion during dialysis. The association between acetate and cardiovascular instability led KDIGO guidelines to specify the use of bicarbonate as the dialysate buffer in almost every circumstance [25]. (See "Intradialytic hypotension in an otherwise stable patient".)
Type of dialysis membrane — The type of dialysis membrane used may be an additional factor that can affect hemodynamic stability. There is evidence in some, but not all, studies that hemofiltration and/or continuous dialysis with very porous membranes may be beneficial by removing middle to large-middle molecules with cardiodepressant, vasodilatory, or immunomodulatory properties that cannot easily cross conventional dialysis membranes [26-30]. Complement-activating membranes can initiate an inflammatory process that contributes to vasodilation (see "Continuous kidney replacement therapy in acute kidney injury"). An additional hemodynamic benefit of porous biocompatible membranes may be to enhance the rate of recovery of kidney function and to improve patient outcomes (see "Dialysis-related factors that may influence recovery of kidney function in acute kidney injury (acute renal failure)"). Thus, the minimal extra cost of these membranes appears to be justified by the potential benefits. This conclusion is supported by the 2012 KDIGO guidelines that recommend the use of biocompatible membranes for kidney replacement therapy in acute kidney injury [25].
SOCIETY GUIDELINE LINKS —
Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Fluid and electrolyte disorders in adults".)
SUMMARY
●Mechanisms of solute removal – The process of solute removal during hemodialysis can occur by two different mechanisms: passive diffusion down a favorable concentration gradient from the plasma into the dialysis fluid and during the ultrafiltration (or convection) of plasma water across the membrane of the hemofilter. With some continuous kidney replacement therapies (CKRTs), such as continuous arteriovenous or venovenous hemofiltration, all solute removal occurs by convection since there is no exposure to dialysate. (See 'Mechanisms of solute removal' above.)
●Mechanisms of hemodialysis-induced hypotension – The major determinants of hemodynamic stability during kidney replacement therapy are the rates of fluid and solute removal. During conventional hemodialysis, the rapid diffusive removal of solutes results in a reduction in the plasma osmolality, thereby causing water to move osmotically into cells. In combination with ultrafiltration, this further depletes the extracellular volume and, together, lowers the systemic blood pressure. By comparison, the fluid removed via isolated ultrafiltration has the same concentration of osmotically active solutes as plasma. Thus, there is no osmotic gradient to draw water movement into cells and intravascular volume is relatively preserved. With CKRT, hypotension is also less likely because of the much lower rates of solute and net fluid removal. (See 'Mechanisms of hemodialysis-induced hypotension' above.)
●Hemodynamics of hemofiltration and hemodialysis – Given solute and fluid removal considerations, the arterial blood pressure and systemic vascular resistance are usually maintained or even increased during hemofiltration. By comparison, both the blood pressure and vascular resistance tend to fall when a similar volume of fluid is removed by hemodialysis. However, the protection afforded by hemofiltration is only relative since hypotension still occurs if too much fluid is removed or if fluid is removed too rapidly. (See 'Hemodynamics of hemofiltration and hemodialysis' above.)
