How Does CRRT Work: Circuit, Modes, and Waste Removal

Continuous renal replacement therapy, or CRRT, works by slowly pumping a patient’s blood through an external circuit containing a semipermeable filter, where waste products and excess fluid are removed around the clock rather than in a few concentrated hours. The process borrows the same basic physics that healthy kidneys use, diffusion and fluid pressure, but applies them mechanically through tubing, a blood pump, and a synthetic membrane called a hemofilter. What makes CRRT distinctive is its gentleness: running 24 hours a day at relatively low flow rates, it avoids the rapid fluid shifts that can destabilize critically ill patients.

The Extracorporeal Circuit

The word “extracorporeal” just means outside the body. In CRRT, blood leaves the patient through a large-bore double-lumen catheter, usually placed in the internal jugular or femoral vein. One lumen draws blood out; the other returns it. The catheter’s position and length matter more than you might expect. If the tip sits in the wrong spot, or if the ports press against the vessel wall, blood can recirculate, meaning it loops back into the circuit without ever reaching the patient’s bloodstream. Catheter-related problems like thrombosis, poor positioning, or external compression are among the most common causes of this recirculation.

From the catheter, blood travels through flexible tubing to a roller or peristaltic pump, which drives it at a controlled rate, typically somewhere between 100 and 250 mL per minute in adults. The blood then enters the hemofilter, a cartridge packed with thousands of hollow fibers made of a synthetic membrane. These fibers have pores small enough to let water, electrolytes, and waste molecules pass through while holding back blood cells and most proteins. After filtration, the cleaned blood exits the hemofilter and returns to the patient through the return lumen of the catheter.

Circuit lifespan varies. Without any anticoagulation, circuits may last only around 10 to 30 hours depending on conditions. With regional citrate anticoagulation, median lifespans can reach 42 to 72 hours. Outflow pressure problems, specifically when blood has trouble leaving the access catheter, independently shorten circuit life and degrade waste removal efficiency, with measurable declines in solute clearance appearing after about 12 hours of use and becoming pronounced at 24 hours when transmembrane pressure climbs above 150 mmHg.²

Three Ways the Filter Removes Waste

The hemofilter clears harmful substances from blood through three distinct physical processes, and different CRRT modes lean on each one to varying degrees.

Diffusion

Diffusion is the same phenomenon that lets a drop of food coloring spread through a glass of water. In CRRT, a sterile solution called dialysate flows on the outside of the hollow fibers while blood flows on the inside. Small molecules like urea and creatinine are concentrated in the blood but absent from the dialysate, so they naturally migrate across the membrane from the high-concentration side to the low-concentration side. The smaller the molecule, the faster it diffuses, which makes diffusion especially effective for clearing small waste products.³

Convection

Convection works more like a drain. A pressure gradient pushes water from the blood side of the membrane to the other side, and dissolved solutes get swept along with it, a process called solvent drag. This mechanism does not depend on a concentration difference; it physically pulls molecules through the membrane pores along with the water. Convection handles mid-sized molecules more effectively than diffusion alone. Interestingly, lab testing has shown that in larger hemofilters, convective and diffusive prescriptions clear middle molecules almost equally well across a wide weight range. In smaller filters, however, high ultrafiltration rates can cause a buildup of proteins against the membrane surface that acts as a secondary filter, paradoxically reducing clearance of those mid-sized molecules.⁴

Adsorption

The third mechanism is adsorption, where molecules stick to the membrane surface itself rather than passing through it. This is particularly relevant for inflammatory mediators like cytokines, which can be too large to filter efficiently by convection but can bind to certain membrane materials. The membrane chemistry matters here. Some membranes, like the AN69ST copolymer, carry an electrical charge that attracts positively charged inflammatory molecules. In trials comparing membrane types, AN69ST showed significantly better adsorption of specific inflammatory mediators compared to other common membrane materials.⁵ Adsorption is not something clinicians prescribe directly; it happens as a byproduct of the membrane chosen, and it diminishes over time as binding sites fill up.

CRRT Modes

The basic circuit stays the same, but how you configure the fluid flows determines the mode of therapy. Each mode emphasizes different combinations of diffusion, convection, and fluid removal. Clinicians choose among them based on what the patient needs most: pure fluid removal, solute clearance, or both.

Slow Continuous Ultrafiltration (SCUF)

SCUF is the simplest mode. Blood passes through the hemofilter, and a pressure gradient pulls water across the membrane, but there is no dialysate and no replacement fluid. The sole purpose is to remove excess fluid, making it useful for patients with severe fluid overload who do not yet need significant waste clearance, such as those with decompensated heart failure who have stopped responding to diuretics. In a series of heart-failure patients treated with SCUF, clinicians removed roughly two liters of fluid per session at a gentle rate of about 5 to 6 mL per minute, with no significant drops in blood pressure or heart rate.⁶ Even brief runs can sometimes jump-start the kidneys: one case report described a patient whose urine output surged to 200 mL per hour after just 10 hours of SCUF and one kilogram of fluid removal, allowing therapy to be stopped early.⁷

Continuous Venovenous Hemofiltration (CVVH)

CVVH relies on convection. Large volumes of water are pushed through the membrane, dragging dissolved waste with it. Because so much fluid is removed, an equal or near-equal volume of sterile replacement fluid is infused back into the circuit to keep the patient’s blood volume stable. This replacement fluid can be added before the filter (pre-dilution) or after it (post-dilution), and the choice matters for circuit longevity and efficiency.

Pre-dilution thins the blood before it enters the filter, which reduces the chance of clotting and can substantially extend filter life. One study found median filter survival of about 46 hours with pre-dilution compared to 16 hours with post-dilution.⁸ The trade-off is that diluting the blood before it reaches the membrane lowers the concentration of waste at the membrane surface, reducing clearance per liter of filtrate. That same study measured a median creatinine clearance of 33 mL/min with pre-dilution versus 45 mL/min with post-dilution. In practice, clinicians often compensate by prescribing a higher total replacement fluid volume when using pre-dilution.

The picture gets muddier when anticoagulation is adequate. In heparinized patients, one trial found no significant difference in filter life between the two approaches, with median survival around 24 to 29 hours in both groups, and no meaningful difference in waste clearance either.⁹ So the benefit of pre-dilution is most apparent when anticoagulation is limited or absent.

Continuous Venovenous Hemodialysis (CVVHD)

CVVHD uses diffusion as its primary mechanism. Dialysate runs along the outside of the membrane fibers, typically in the opposite direction to blood flow (counter-current). There is minimal or no net fluid removal beyond what is needed for volume management. This mode excels at clearing small solutes like urea and creatinine. Counter-current flow keeps the concentration gradient steep along the entire length of the filter, and one clinical study showed it improved removal of these small solutes by roughly 20% compared to running the dialysate in the same direction as blood flow.¹⁰

Continuous Venovenous Hemodiafiltration (CVVHDF)

CVVHDF combines both mechanisms, running dialysate across the membrane while also generating ultrafiltrate and infusing replacement fluid. This hybrid approach offers both diffusive clearance of small solutes and convective clearance of mid-sized molecules simultaneously. It is the most commonly prescribed CRRT mode in many intensive care units, and research suggests that using CVVHDF independently prolongs circuit lifespan compared to purely convective modes.²

Keeping the Circuit From Clotting

Blood exposed to foreign surfaces wants to clot. The tubing, the pump head, and the hemofilter membrane all trigger the coagulation cascade, and without intervention, most circuits would clot within hours. Two main anticoagulation strategies dominate CRRT practice.

Systemic heparin is the traditional choice: an infusion of unfractionated heparin thins the blood throughout the entire body, keeping the circuit open but raising bleeding risk everywhere else. Regional citrate anticoagulation (RCA) takes a different approach. Citrate solution is infused into the circuit before the filter, where it binds calcium ions. Because calcium is essential for clot formation, removing it from the blood in the circuit effectively prevents coagulation there. The target citrate concentration in the circuit is about 4 to 6 mmol/L, which drops ionized calcium below the threshold needed for clotting. Citrate has a systemic half-life of only about five minutes; the liver, muscles, and kidneys metabolize it rapidly after the blood returns to the patient. A separate calcium infusion runs through a different line to replace the calcium lost in the circuit, keeping the patient’s systemic calcium normal.¹¹

A meta-analysis of randomized trials found that regional citrate significantly prolonged circuit life in CVVH and reduced bleeding risk by about 70% compared to systemic heparin.¹² A pediatric study echoed those findings, reporting median hemofilter survival of 51 hours with citrate versus about 30 hours with heparin, and a clotting-related filter failure rate roughly half as high.¹³ Regional citrate does have limits: patients with severe liver failure may not metabolize citrate fast enough, leading to citrate accumulation and a dangerous drop in ionized calcium. For these patients, heparin or even no anticoagulation with frequent circuit changes may be the safer option.

Monitoring Circuit Pressures

Modern CRRT machines continuously track several pressure readings that together tell clinicians how the circuit is performing. The two most important are the filter pressure drop (the difference in pressure across the hemofilter from the blood inlet to the outlet) and the transmembrane pressure (the pressure pushing fluid through the membrane). Rising values in either one signal that the filter is clogging, either from protein fouling or from early clot formation.

These pressure trends are predictive. In pediatric circuits, each 1 mmHg increase in transmembrane or filter pressure was independently associated with about a 1.5% higher risk of circuit clotting.¹⁴ The challenge is that both measures are also affected by normal fluctuations in blood flow rate and ultrafiltration settings, which can trigger false alarms. Newer research has proposed adjusted parameters that account for flow velocity, aiming to separate genuine clotting signals from routine pressure changes.¹⁵

Electrolyte and Acid-Base Balance

CRRT does not just remove waste; it also strips electrolytes and buffer from the blood. Every liter of ultrafiltrate or dialysate effluent carries away sodium, potassium, bicarbonate, phosphate, magnesium, and calcium. To prevent dangerous imbalances, the replacement fluid and dialysate are carefully formulated to contain electrolytes at concentrations close to what clinicians want in the patient’s blood. The buffer in these fluids is usually bicarbonate or lactate (or, in citrate-based protocols, citrate itself, which the body converts to bicarbonate). Maintaining acid-base balance is one of the core reasons CRRT is prescribed in the first place, and the composition of these fluids is adjusted based on frequent blood gas measurements.¹⁶

Drug Dosing Complications

One of the less intuitive consequences of CRRT is that it removes medications from the bloodstream along with waste, and the degree to which it does so varies enormously from drug to drug. Only the unbound fraction of a drug, the portion not attached to blood proteins, can cross the hemofilter membrane. Whether that removal matters clinically depends on the drug’s characteristics: its molecular size, how tightly it binds to proteins, and how large a volume of body tissue it distributes into.¹⁷

The drug most susceptible to significant removal by CRRT is one with low protein binding, a small volume of distribution, and limited clearance by the liver or other non-kidney routes. Many antibiotics fit this profile. The practical risk is underdosing: a critically ill patient on CRRT for a life-threatening infection may be getting sub-therapeutic antibiotic levels without anyone realizing it, which not only risks treatment failure but also promotes antibiotic resistance. On the flip side, a drug cleared mainly by the liver will not be meaningfully affected by CRRT even if the filter pulls some of it out, because the liver is already removing it much faster.¹⁸ The CRRT mode matters too: in post-dilution hemofiltration, drug clearance essentially equals the ultrafiltration rate, whereas pre-dilution requires an additional correction because the blood is diluted before reaching the membrane.

Why CRRT Instead of Standard Dialysis

Conventional intermittent hemodialysis accomplishes the same waste removal goals in about three to four hours, so why run a machine around the clock? The answer lies almost entirely in hemodynamic stability. ICU patients are often on vasopressors with fragile blood pressures, and rapid fluid and solute removal hammers an already stressed cardiovascular system. In a direct comparison, patients treated with intermittent hemodialysis had substantially greater drops in mean arterial pressure and larger increases in heart rate after sessions than those treated with CRRT.¹⁹ CRRT’s slow, steady approach spreads the same total fluid removal across 24 hours, keeping blood pressure more stable and reducing the need to escalate vasopressor doses.

When CRRT Stops

Deciding when to disconnect a patient from CRRT involves watching for signs that the kidneys are waking up. Urine output is the most widely studied predictor, though the evidence is imperfect. A meta-analysis found that urine output had moderate accuracy for predicting successful liberation from CRRT, with a sensitivity of about 66% and specificity around 74%, though no single threshold volume emerged as the clear cutoff across studies.²⁰ A separate pooled analysis identified three factors that independently predicted successful weaning: higher urine output at the time CRRT stopped, shorter total duration of CRRT, and the absence of pre-existing chronic kidney disease.²¹ In practice, clinicians often trial a brief pause in CRRT while monitoring urine output and lab values over several hours to see whether the kidneys can sustain clearance on their own.

Adapting CRRT for the Smallest Patients

Standard CRRT machines were designed for adults, and using them on newborns or small infants creates real problems. The extracorporeal circuit volume in adult-sized tubing can represent a significant fraction of a tiny patient’s total blood volume, causing dangerous drops in blood pressure at the moment the circuit is connected. Neonatal-specific platforms have been engineered to address this. The CARPEDIEM device, designed for infants weighing 2.5 to 10 kg, uses a circuit priming volume under 30 mL, miniaturized roller pumps, and ultrafiltration scales precise to one gram.²² Retrospective data suggest that infants under 5 kg treated on the CARPEDIEM platform had better survival during therapy than those treated on standard machines, which required substantially more blood pressure support at the start of treatment.²³

From Arterial Shunts to Pump-Driven Circuits

The story of CRRT starts in 1977, when Peter Kramer in Göttingen, Germany, performed the first continuous arteriovenous hemofiltration (CAVH). That original technique relied on the patient’s own arterial blood pressure to push blood through the filter, which meant it required cannulation of an artery and worked poorly whenever blood pressure dropped.²⁴ The limitations were obvious: arterial access carried bleeding and ischemia risks, and the low flow rates limited waste clearance in highly catabolic patients. These problems drove the development of pump-driven venovenous circuits (CVVH, CVVHD, CVVHDF), which freed the therapy from dependence on the patient’s blood pressure and eliminated the need for arterial puncture.²⁵ By the 1990s, arteriovenous CRRT had largely been abandoned in favor of the venovenous platforms that remain standard today.