Acute kidney injury (AKI) is a common complication of cardiovascular surgery affecting >20% of patients. The aetiology is complex and includes ischaemia-reperfusion injury, pro-inflammatory processes following prolonged cardiopulmonary bypass, haemodynamic instability, haemolysis, nephrotoxins and radiocontrast dye. It commonly develops on the background of pre-existing comorbidities such as vascular disease, older age and chronic kidney disease (CKD). Following cardiac surgery, patients are particularly vulnerable to the consequences of AKI, especially electrolyte derangement, metabolic acidosis and fluid accumulation. Between 1 and 5% of patients require renal replacement therapy (RRT). Several risk prediction models for RRT have been developed.
removal of excess fluid;
correction of electrolyte abnormalities (most importantly hyperkalaemia);
removal of metabolic byproducts (e.g. urea, ammonia, sulphates);
correction of metabolic acidosis;
removal of toxins, including drugs;
removal of inflammatory mediators.
RRT is considered a supportive therapy during a period where metabolic and fluid demands exceed native kidney function. It can also help to limit non-renal organ dysfunction that may be exacerbated by the consequences o f AKI. Although it achieves solute clearance, acid-base homeostasis and fluid removal, it does not fully replace all kidney functions, such as reabsorption of amino acids, activation of vitamin D and erythropoietin production.
There are several techniques of administering RRT, which can be classified according to the technique used (modalities) and the duration of therapy (modes).
Renal replacement therapy can be administered either by utilising an extracorporeal circuit containing an artificial membrane (haemodialysis or haemofiltration) or by utilising the abdominal cavity and the patient’s own peritoneal membrane (peritoneal dialysis).
Haemodialysis (HD) refers to the removal of solute and water across a semipermeable membrane by means of diffusion across a concentration gradient. Blood is pumped through a filter whilst electrolyte containing dialysate fluid flows in the reverse direction on the opposite (non-blood) side of the filter membrane (see Figure 18.1).
Haemofiltration (HF) relies on the principle of convection. A pressure gradient is applied across a semipermeable membrane resulting in removal of solute and water. A balanced replacement fluid is added pre-filter and/or post-filter to maintain electrolyte concentrations and fluid balance (see Figure 18.2). It should be remembered that the addition of pre-filter replacement fluid reduces the delivered RRT dose.
Haemodiafiltration (HDF) incorporates the removal of solute and water across a semipermeable membrane by means of the application of both a concentration gradient and a transmembrane pressure (see Figure 18.3).
Peritoneal dialysis (PD) uses the patient’s peritoneum as a membrane. Dialysate fluid is instilled into the abdominal cavity and drained out several hours later. Fluids, electrolytesand small molecules are exchanged across the peritoneal membrane by diffusion.
HD, HF and HDF should be viewed as equivalent therapies for AKI after cardiac surgery. Middle molecular weight molecules may be removed more effectively by convection than by diffusion but there is no evidence to suggest that this leads to clinically important differences in outcome.
PD has theoretical advantages in that haemodynamic disturbance is rare and anticoagulation is not necessary. However, the efficiency and control of fluid and solute balance is inferior to HD, HF and HDF. In chronic PD patients, PD has a role for maintenance RRT during the recovery period post surgery.
Figure 18.1 Haemodialysis.
Figure 18.2 Haemofiltration.
Figure 18.3 Haemodiafiltration.
RRT can be provided intermittently (≈4 hours daily or alternate days) or continuously (24 hours/day). During continuous renal replacement therapy (CRRT), solute clearance and fluid removal occur over a longer time resulting in less fluctuation in the concentrations of osmotically active molecules, such as urea and ammonia. Intermittent RRT (IRRT) is more effective at clearing small solutes and fluid rapidly.
CRRT is recommended for patients who are haemodynamically unstable and/or do not tolerate rapid fluid removal. It is also superior to intermittent RRT in patients with cerebral injury or cerebral oedema who may not tolerate fluctuations in osmotically active molecules. However, the potential advantages of CRRT are offset by the need for immobilisation, longer exposure to anticoagulation and increased health care costs (Table 18.1). Intermittent RRT has a role in situations where patients are haemodynamically more stable and where the focus of care has shifted to rehabilitation.
|Continuous RRT||Intermittent RRT|
|Continuous removal of toxins and electrolytes||Rapid clearance of toxins and electrolytes|
|Slower removal of excess fluid (i.e. better haemodynamic tolerability)||More time for diagnostic and therapeutic procedures|
|Less fluid and metabolic fluctuations||More time for physiotherapy/rehabilitation|
|Reduced exposure to anticoagulation|
|Lower financial costs|
Abbreviations: RRT renal replacement therapy.
Haemodialysis is usually IRRT but can be provided as CRRT. Haemofiltration is usually CRRT, and haemodiafiltration can be performed as CRRT or IRRT.
Hybrid therapies such as slow low efficiency dialysis (SLED) or prolonged intermittent renal replacement therapy (PIRRT) are variations of RRT, which are usually provided daily for 6–12 hours. They offer the advantages of both intermittent and continuous RRT and are often employed during the transition period from CRRT to IRRT.
In patients with life threatening complications of AKI such as severe hyperkalaemia, marked metabolic acidosis or fluid overload, the decision to urgently start RRT is generally unequivocal (Table 18.2).
Abbreviations: AKI acute kidney injury; RRT renal replacement therapy.
In the absence of overt or impending life threatening complications, the optimal threshold for starting RRT after cardiac surgery remains uncertain. Important considerations include the severity of illness, non-renal organ dysfunction, degree of fluid overload, severity of metabolic acidosis, clinical reserve to tolerate fluid overload and metabolic disturbances, anticipated fluid administration and likelihood of spontaneous recovery of renal function.
Potential advantages of earlier RRT initiation may be offset by the risks of catheter insertion, bleeding, treatment related haemodynamic instability, infectious complications and unwanted losses of nutrients and drugs (Table 18.3). In contrast, delays in initiating RRT may put patients at risk of serious fluid overload and other life threatening complications.
|Avoidance and/or early control of fluid accumulation and overload||Risk of complications associated with dialysis catheter insertion|
|Avoidance and/or earlier control of complications of uraemia||Risk of complications from anticoagulation|
|Avoidance and/or earlier control of electrolyte/metabolic derangement||Risk of haemodynamic instability|
|Avoidance and/or earlier control of acid-base derangement||Clearance of micronutrients and trace elements|
|Avoidance of unnecessary diuretic exposure||Excess clearance of dialysable medications (i.e. antimicrobials, antiepileptics)|
|Clearance of inflammatory mediators||Workload for providers|
Abbreviations: AKI acute kidney injury; RRT renal replacement therapy.
A recent meta-analysis of two randomised controlled trials (RCTs) and nine retrospective cohort studies which included a total of 841 patients following cardiac surgery concluded that the early initiation of RRT was associated with a lower 28-day mortality and shorter stay in ICU. However, there was marked heterogeneity between the studies including the definitions of ‘early’ and ‘late’ initiation.
Clinical practice is often variable. Ideally, the decision to start RRT should be individualised and based on the dynamic context and trajectory of the patient, illness severity, non-renal organ dysfunction, along with physiological and laboratory data, rather than relying on absolute laboratory values. Figure 18.4 shows an algorithm which incorporates these principles.
Figure 18.4 Algorithm for deciding whether to begin RRT.
The dose of RRT is a measure of the quantity of a solute that is removed from the patient during extracorporeal treatment. In patients on intermittent haemodialysis, dose of treatment is expressed as urea reduction ratio (URR) or Kt/V (K is dialyser clearance of urea, t is dialysis time and V is the volume of distribution of urea). Both parameters have important limitations in critically ill patients with AKI where neither urea generation rate nor volume of distribution can be clearly defined. In patients receiving CRRT, the dose may be estimated considering the effluent flow rate indexed by the patient’s body weight. It is usually described as ml/kg/hour.
Based on two large RCTs, the current recommendation is to deliver a target dose of 20–25 ml/kg/hour. However, unintended interruptions in treatment often occur which will reduce the effective dose. Therefore, doses higher than 20–25 ml/kg/hour may have to be prescribed but there is no role for high volume RRT.
Although the dose of RRT classically refers to solute clearance, ultrafiltration rate and target fluid balance should also be considered as important components of the prescription. Cardiac surgery results in an inflammatory state and capillary leak, which may result in fluid accumulation. Progressive tissue and pulmonary oedema can be attenuated by close attention to fluid balance and judicious use of ultrafiltration.
In patients with AKI requiring RRT, contact of blood with the extracorporeal circuit results in activation of the coagulation cascade. Additionally, the haematocrit within the filter increases as a result of fluid removal, which adds to the risk of the filter clotting. Premature clotting of the circuit reduces effective clearance, leads to blood loss and increases the workload and financial costs. The goal of anticoagulation is to maintain filter patency, and thus avoid these complications.
The most commonly used anticoagulation strategies include unfractionated heparin delivered systemically or via the circuit, low molecular weight heparin (LMWH), systemic epoprostenol, regional anticoagulation with citrate and non-pharmacological measures. Other options are thrombin inhibitors such as argatroban, factor Xa inhibitors such as danaparoid or fondaparinux, or serine proteinase inhibitors, for example nafamostat mesylate. The choice should be individualised and based on characteristics of the patient, potential risks and benefits of the anticoagulant, availability and local expertise (Table 18.4).