Preventing organ dysfunction due to hypoperfusion is the goal of most therapies addressing circulatory failure. In shock states, such as those caused by sepsis and acute pancreatitis, the intravascular volume may be reduced. This may be the result of several factors, such as dehydration from vomiting, reduced fluid intake, and fever. The negative effect of the reduced volume is compounded by interstitial fluid sequestration and vasoplegia with venous pooling of blood leading to a reduced preload of the heart. However effective in re-establishing normal blood pressure and cardiac output, the administration of large volumes of intravenous fluids comes at a price. There is ample evidence from observational trials in the critically ill that overzealous volume administration may negatively affect outcomes leading to increased mortality, morbidity, and length of stay1. In conditions of high systemic inflammation, the infusion of crystalloids has the potential to harm endothelial integrity leading to a vicious circle of increased intravascular fluid loss and oedema, further compromising organ function in critically ill patients.
In acute pancreatitis, the presence of microcirculatory dysfunction caused by local inflammation and the release of inflammatory mediators leads to a systemic inflammatory response that reduces pancreatic perfusion and aggravates injury. This local and systemic inflammation may lead to intestinal paralysis and retroperitoneal oedema. This state may lead to intra-abdominal hypertension and eventually development of abdominal compartment syndrome as the most severe form. Such severe conditions can further compromise a patient’s haemodynamic status, lead to further organ failure, and are associated with high mortality.
Fluid resuscitation has been the mainstay of initial therapy in acute pancreatitis. Even in the absence of obvious hypovolaemia and shock, early fluid resuscitation has been propagated based on evidence from experimental studies showing a reduction in pancreatic injury and necrosis and thereby preventing the progression to more severe forms2. However, there has been a lack of data from high-quality RCTs to guide use of fluid administration. Enter the WATERFALL study published in 20223. This study was the first large RCT comparing different fluid administration strategies in acute pancreatitis. The WATERFALL study compared an initial crystalloid bolus of 20 ml/kg with a bolus of 10 ml/kg followed by an infusion of 3 ml/kg/h and 1.5 ml/kg/h respectively. Notably, patients presenting with shock and patients with moderately severe or severe acute pancreatitis were excluded. Initially planned to include 744 patients, the trial ended up including 249 patients after it was halted due to an interim analysis showing a significantly higher incidence of fluid overload in the aggressive group (20.5 compared with 6.9 per cent) without significant differences in the incidence of the primary outcome of development of moderately severe or severe acute pancreatitis. Although not statistically significant, the incidences of important secondary outcomes, such as pancreatic necrosis, organ failure, ICU admission, and death, were all numerically higher in the aggressive group. The median amounts of fluid administered were 7.8 l in the aggressive group and 5.5 l in the moderate group.
The WATERFALL study provides evidence that an aggressive approach to fluid administration does not confer any benefits on patient-centred outcomes in mild acute pancreatitis. Whether this can be extrapolated to patients with more severe disease is unknown, but the totality of evidence at this time does not indicate that aggressive fluid administration is warranted. Having this in mind, it is important that patients presenting with haemodynamic compromise due to hypovolaemia if indicated by history, clinical examination, or a high haematocrit are resuscitated with intravenous fluids, as simply leaving organs underperfused is not an option that is in line with basic principles of resuscitation (Fig. 1). The question that remains unanswered is how to optimize fluid administration in a way that reaps the benefits of increased perfusion while not exacerbating capillary leak and oedema.
Proposed algorithm for appropriate fluid management in acute pancreatitis
Hct: hematocrit; MAP: mean arterial pressure; UO: urine output; CRT: capillary refill time; ACS: abdominal compartment syndrome.
Finding the sweet spot in fluid resuscitation is notoriously hard in the surgical ward, and even in a high-tech environment, such as the ICU. Traditionally, administering fluids while a patient is on the ascending slope of the Frank–Starling curve has been the goal in patients in shock4. While administering fluid in non-fluid-tolerant patients is certainly detrimental due to the increased stress on vessel walls and the risk of reduced organ perfusion from high venous pressures, there is an understanding that even fluid-responsive patients may experience serious side effects, as these fluids accumulate over time outside the circulatory system. In a patient with a protracted course of instability, a strategy of filling the patient to the point of loss of fluid responsiveness is likely to lead to gross oedema and the risk of organ dysfunction.
So, where do we stand regarding fluid resuscitation at this point? Patients with acute pancreatitis who are well perfused should at most receive fluids as prescribed in the moderate arm of the WATERFALL study3. Frequent assessments of fluid overload should be performed in all patients (Fig. 1). Special caution is warranted in patients with known heart failure who may be highly sensitive to volume overload.
When a patient presents in shock, the picture becomes murkier5. In the absence of high-quality evidence in acute pancreatitis, we must probably seek guidance in the evidence from fluid treatment in septic shock. In the CLASSIC trial6, there was no difference in patient-centred outcomes using a very restrictive fluid protocol in septic shock compared with usual practice. In this trial, all patients received at least 1 l of fluid before ICU admission and additional fluids in the restrictive group were only administered when there was refractory hypotension, lactate greater than 4 mmol/l, or skin mottling.
When combined with evidence from small trials in severe acute pancreatitis showing increased mortality in aggressive fluid therapy7, the safety of a restrictive therapy in septic-shock patients should temper the inclination of some to treat these patients with large volumes of fluid to obtain transient haemodynamic or biochemical improvements. However, there are caveats to extrapolating data from septic shock to acute pancreatitis, as there are obvious differences in the presentation, time course, and pathophysiological basis of these two diseases.
The WATERFALL study has provided us with a valuable insight into the effect of fluid administration in mild acute pancreatitis, but there is a great need for well-designed studies in the treatment of severe acute pancreatitis. It is far from satisfactory that we have such a small evidence base to guide us in the most common treatment of a disease with a mortality reaching 20–40 per cent in the severe categories.
Funding
The authors have no funding to declare.
Disclosure
The authors declare no conflict of interest.
