4 Renal therapy techniques

Conventional haemodialysis requires a pressurised, purified water supply, and a greater risk of haemodynamic instability due to rapid fluid and osmotic shifts. Haemo(dia)filtration can be arterio‐venous, using the patient's BP to drive blood through the haemofilter, or pumped veno‐venous. The latter is preferable as it does not depend on the patient's BP, and the pump system incorporates alarms and safety features. Continuous veno‐venous haemo(dia)filtration (CVVH or CVVHD) is increasingly the technique of choice. Blood is usually drawn and returned via a 10–12Fr double‐lumen, central venous catheter (see figure 4.1).

CVVH relies on convection (bulk transfer of solute and water) to clear solute. In CVVHD, dialysate flows countercurrent to the blood, allowing small molecules to diffuse according to their concentration gradients.

Membranes are usually hollow fibre polyacrylonitrile, polyamide, or polysulphone with a surface area of 0.6–1m².

Both CVVH and CVVHD are effective for small molecule clearance (e.g. urea). CVVH is better at larger molecule clearance and can remove substances up to the membrane pore size cut‐off (usually 30–35kD). Filtrate is usually removed at 20–35mL/kg/h; fluid balance is adjusted by varying the rate of fluid replacement. High volume haemofiltration involves much higher ultrafiltration rates (e.g. 50–100mL/kg/h, usually for short periods, e.g. 4h) in an effort to remove inflammatory mediators. Variable outcomes are reported in studies.

Creatinine and K⁺ clearances are higher with CVVHD, but filtration alone is usually sufficient if ultrafiltrate volume is adequate. (1000mL/h approximates to a creatinine clearance of 16mL/min). CVVHD is preferred for pharmacologically‐resistant hyperkalaemia.

A balanced electrolyte solution buffers acidaemia and is titrated to desired fluid and electrolyte balance. Buffers include lactate (liver metabolised to bicarbonate) and bicarbonate. Acetate (metabolised by muscle) causes most haemodynamic instability and is now rarely used. Bicarbonate may be more efficient than lactate at reversing severe acidosis, but no outcome benefit has been shown. Care is needed when giving Ca²⁺ since calcium bicarbonate may crystallise. In hypoperfused liver, lactate may be inadequately metabolised.

Increasing metabolic alkalosis may be due to excessive buffer so use low buffer (30mmol/L lactate) replacement fluid. K⁺ can be added to maintain normokalaemia. 20mmol KCl in a 4.5L bag provides a concentration of 4.44mmol/L. K⁺ clearance is increased by reducing the concentration within replacement fluid or dialysate.

Haemo(dia)filtration (2), p110; Oliguria, p398; Acute renal failure—diagnosis, p400; Acute renal failure—management, p402.

The circuit can be anticoagulated with unfractionated heparin (200–2000IU/h), a prostanoid (prostacyclin or PGE₁) at 2–10ng/kg/min, or a combination of the two. Other alternatives include regional citrate anticoagulation, the low molecular weight heparinoid danaparoid, or direct thrombin inhibitors (e.g. lepirudin, argatroban). These agents are, as yet, under‐evaluated for renal replacement therapy and they risk important complications such as hypocalcaemia (with citrate) and prolonged bleeding with the long half‐lives of the thrombin inhibitors. Bivalirudin is a new agent related to hirudin and lepirudin but is reversible with a short half‐life of only 25min. No anticoagulant may be needed if the patient is auto‐anticoagulated.

Premature clotting may be due to mechanical kinking/obstruction of the circuit, insufficient anticoagulation, inadequate blood flow rates, or lack of endogenous anticoagulants such as antithrombin III.

Usual filter lifespan should be at least two days, but is often decreased in septic patients due to decreased endogenous anticoagulant levels. In this situation, consider use of fresh frozen plasma, a synthetic protease inhibitor such as aprotinin, or antithrombin III replacement (costly).

Flow through the filter is usually 100–200mL/min. Too slow a flow rate promotes clotting. Too high a flow rate will increase transmembrane pressures and decrease filter lifespan without significant improvement in clearance of ‘middle molecules’ (e.g. urea).

Haemo(dia)filtration (1), p108; Coagulation monitoring, p222; Anticoagulants, p318.

A slow form of dialysis, utilising the peritoneum as the dialysis membrane. It is now rarely used in UK Critical Care Units, having been largely superseded by continuous haemofiltration. The technique does not require complex equipment although continuous flow techniques do require continuous generation of dialysate. Treatment is labour‐intensive and there is considerable risk of peritoneal infection.

For acute peritoneal dialysis, a trochar and cannula are inserted through a small skin incision under local anaesthetic. The skin is prepared and draped as for any sterile procedure. The commonest approach is through a small midline incision 1cm below the umbilicus. The subcutaneous tissues and peritoneum are punctured by the trocar which is withdrawn slightly before the cannula is advanced towards the pouch of Douglas. In order to avoid damage to intra‐abdominal structures, 1–2L warmed peritoneal dialysate may be infused into the peritoneum by a standard, short intravascular cannula prior to placement of the trocar and cannula system. If the midline access site is not available, an alternative is to use a lateral approach, lateral to a line joining the umbilicus and the anterior superior iliac spine (avoiding the inferior epigastric vessels).

Warmed peritoneal dialysate is infused into the peritoneum in a volume of 1–2L at a time. During the acute phase, fluid is flushed in and drained continuously (i.e. with no dwell time). Once biochemical control is achieved, it is usual to leave fluid in the peritoneal cavity for 4–6h before draining. Heparin (500IU/L) may be added to the first six cycles to prevent fibrin catheter blockage. Thereafter, it is only necessary if there is blood or cloudiness in the drainage fluid.

The dialysate is a sterile balanced electrolyte solution with glucose at 75mmol/L for a standard fluid or 311mmol/L for a hypertonic fluid (used for greater fluid removal). The fluid is usually potassium‐free since potassium exchanges slowly in peritoneal dialysis although potassium may be added if necessary.

It is possible to sterilise the peritoneum and catheter by adding appropriate antibiotics to the dialysate. Suitable regimens include:

Oliguria, p398; Acute renal failure—diagnosis, p400; Acute renal failure—management, p402.

Plasma exchange may be used to remove circulating toxins or to replace missing plasma factors. It may be used in sepsis (e.g. meningococcaemia). In patients with immune mediated disease (e.g. Guillain–Barré syndrome, thrombotic thrombocytopaenic purpura), plasma exchange is usually a temporary measure while systemic immunosuppression takes effect. Most diseases require a daily 3–4L plasma exchange repeated for at least four further occasions over 5–10 days.

Blood is separated into components in a centrifuge. Plasma (or other specific blood components) is discarded and a plasma replacement fluid is infused in equal volume. Centrifugation may be continuous (where blood is withdrawn and returned by separate needles) or intermittent (where blood is withdrawn and separated, then returned via the same needle).

Plasma is continuously filtered through a large pore filter (molecular weight cut‐off typically 1,000,000D). Plasma is discarded and replaced by infusion of an equal volume of replacement fluid. The technique is similar to haemofiltration and uses the same equipment.

Most patients will tolerate replacement with a plasma substitute. Some fresh frozen plasma will be necessary after the exchange to replace coagulation factors. The only indication to replace plasma loss with all fresh frozen plasma is where plasma exchange is being performed to replace missing plasma factors.

Guillain Barré syndrome, p456; Myasthenia gravis, p458; Platelet disorders, p478; Vasculitides, p574.