INTRODUCTION
Over the past decade, the use of levetiracetam in critical care settings both in the treatment and prevention of various seizure disorders has significantly increased [1-2] . This increase is owed in large part to several studies that have showed it to be a safe, broad-spectrum and highly effective anti-epileptic drug with minimal drug interactions and a wide therapeutic index [3-5] . There is no official requirement for therapeutic drug monitoring and, as such, no agreed target range for levetiracetam exists. However, a large scale dose-ranging study conducted among epileptic patients identified a desired therapeutic target of 12 – 46 mcg/mL at trough level[6] .
Levetiracetam is a small, hydrophilic molecule that is weakly bound to proteins (<10%) with linear kinetics [5,7-9] . In a healthy adult, the total clearance of levetiracetam is approximately 4.03L/hr, the half-life is 6 – 8 hours and the volume of distribution is 0.5 – 0.7L/kg [10] . It is almost exclusively (>90%) excreted via the kidneys and so dose adjustments are recommended in patients with impaired renal function[6,8,11,12] . Even mild-to-moderate renal impairment has been shown to double plasma concentration and significantly increase drug half-life [13] .
Renal impairment is common in critical care settings with around half of people admitted to intensive care developing acute kidney injury[14-15] . Continuous renal replacement therapy (CRRT) is used in severe acute kidney injury to support renal function and minimise the risk of multi-organ failure and fluid overload. Additionally, people with end-stage renal disease who require intermittent haemodialysis often benefit from a temporary transition to continuous renal replacement therapy in the event of critical illness due to the enhanced haemodynamic and metabolic control it provides. The introduction of renal replacement therapy creates challenges for pharmacological management as pharmacokinetics are altered through both extrinsic (e.g. filtration mode, filter type, flow rate) and intrinsic factors (e.g. residual renal function, fluid volume status, protein binding). For medications that are renally excreted, such as levetiracetam, the effects on clearance can be significant[16] .
Intermittent haemodialysis (IHD) has already been shown to eliminate 50% of levetiracetam within four hours [13] . As a result, the current dosing recommendation for patients receiving IHD is a 750mg loading dose followed by 500 – 1,000mg once daily dosing. However, in patients on CRRT the current recommended dosing is 250 – 750mg twice daily without loading – equivalent to a patient with CKD Stage 3b[17] . While there have been a multitude of studies investigating the pharmacokinetics of levetiracetam in haemodialysis to support the current dosing recommendation in IHD [18-21] , the evidence base for patients undergoing CRRT remains limited. Indeed, the current dosing recommendations are based on three case reports alone. In the last few years there has been a renewed focus on broadening the evidence base [22] .
In this systematic review, we evaluate the latest available evidence on the pharmacokinetics of levetiracetam in critically ill patients undergoing CRRT. This review was performed in light of several case reports and case studies suggesting that the clearance of levetiracetamin vivo may be significantly higher than previously thought – raising the risk of sub-therapeusis in such patients[22-25] . We undertake a quality assessment of the current evidence in addition to meta-analysis and computational modelling to understand the factors that affect clearance and to simulate the effect of different dosing strategies on plasma concentration.