To determine the feasibility and safety of deresuscitation using slow continuous ultrafiltration with regional citrate anticoagulation and peripheral or standard central venous access.

Fluid deresuscitation has been shown to have favourable outcomes in the management of the critically ill patient. Whilst diuretic medications have previously been the mainstay of achieving this deresuscitation, the metabolic and biochemical complications associated with their use can limit their application. Mechanical ultrafiltration is another means of achieving fluid deresuscitation and has theoretical benefits over the use of intravenous diuretics. It allows a more titratable process, without the potential metabolic complications. The use of ultrafiltration has been limited by the requirement of wide bore central venous access. The SCUFFD study is a means of assessing whether ultrafiltration with regional anticoagulation can be achieved through standard central or peripheral venous access.

Inclusion Criteria:

• Adults (>18 years)
• Evidence of fluid overload (greater than trace amount of peripheral oedema in >1 site, pulmonary oedema on chest radiograph, or positive fluid balance equivalent to >5% body weight)
• Clinician intention to target a negative fluid balance
• Expected to remain in a critical care setting beyond the next calendar day

Exclusion Criteria:

Anticipated unavailability of suitable vascular access

• Lack of commitment to full support
• Receiving or imminently planned to receive renal replacement therapy
• Hyponatraemia (Sodium <130mmol/L)
• Hypernatraemia (Sodium >150mmol/L)
• Significant metabolic alkalosis (Bicarbonate>30 and pH>7.5)
• Significant metabolic acidosis (HCO3- <18 mmol/l and pH < 7.30)
• Uncorrected hypokalaemia (Potassium <3.0mmol/L)
• Liver failure (Child-Pugh Grade B or above)
• Shock (any of: lactate >3mmol/L, extensive skin mottling, central capillary refill time>3 seconds, more than 1 vasoactive agent in use, noradrenaline dose >0.2 mcg/kg/min, use of dobutamine, adrenaline, milrinone, enoximone or levosimendan)
• Receiving any systemic anticoagulation other than for routine venous thromboembolism prophylaxis, or dual antiplatelet therapy
• Prisoner
• Known pregnancy

No personalized IPD will be available to be share with other researchers. Anonymised data will be grouped into a database which may be shared with other researchers.