Monday, December 05, 2011

Fluid resuscitation in acute pancreatitis: is less actually more?

Traditional teaching and guidelines held that we should pour the fluids early on in the treatment of acute pancreatitis. Nobody would say just how much, but a lot. The problem was, these recommendations were not driven by high level data. We had expert opinion, animal data, pathophysiologic rationale and low level studies in patients but nothing more.

A new study, representing the best evidence we have to date, challenges that teaching:

RESULTS:
A total of 247 patients were analyzed. Administration of greater than 4.1 l during the initial 24 h was significantly and independently associated with persistent OF, acute collections, respiratory insufficiency, and renal insufficiency. Administration of less than 3.1 l during the initial 24 h was not associated with OF, local complications, or mortality. Patients who received between 3.1 and 4.1 l during the initial 24 h had an excellent outcome.


CONCLUSIONS:
In our study, administration of a small amount of fluid during the initial 24 h was not associated with a poor outcome.

I don't have access to the full text of the paper to know how severity adjustment was done. The authors, when interviewed by Medscape, spoke strongly against the practice of massive fluid resuscitation for all patients with pancreatitis.

How might this study change practice? To me it challenges the dogma that patients with acute pancreatitis should undergo massive initial fluid resuscitation just because they have acute pancreatitis. The game changer may be that we'll have to adopt a more individualized approach to fluid management based on volume assessment. For patients who seem well perfused at presentation, a strategy like that suggested by the authors, aiming for 3-4 liters or so over the first 24 hours, may be reasonable with the caveat that frequent clinical and laboratory assessment over that time may be necessary to identify those patients who need to be switched to a more aggressive volume loading strategy. We'll need further study of this question to refine the approach.