- The 1992 Atlanta classification of acute pancreatitis defined mild and severe categories based on organ failure, but recent evidence shows this is too simplistic. - A proposed revision suggests classifying patients based on transient versus persistent organ failure, with persistent organ failure defining severe acute pancreatitis. - There is also a need to distinguish patients with local complications from those with both local and systemic complications, as mortality is much higher in the latter group. This suggests acute pancreatitis should be classified into mild, moderate, and severe/critical categories.

1. The American Journal of GASTROENTEROLOGY VOLUME 105 | JANUARY 2010 www.amjgastro.com
nature publishing group74
REVIEW
CLINICAL REVIEWS
“Crude classifications and false generalizations are the
curse of organized life.”
—George Bernard Shaw
The Atlanta classification, the most widely used classification
of acute pancreatitis, was introduced in 1992 and defined mild
and severe categories of acute pancreatitis (1). However, its
limitations have been highlighted in a number of publications
over the past 5 years and a call for its revision has been made
(2–4). The major impetus to revision has been the recent signi-
ficant advances in understanding the pathophysiology of acute
pancreatitis and especially the role of systemic complications.
Whereas the original Atlanta classification assessed only the
presence or absence of organ failure (OF), it is now recognized
that the number of organs that fail, the timing of onset, the
change in OF in response to initial treatment, and the duration
of OF, all contribute to severity (5–7). Given that OF lasting
for >48h is associated with significantly increased mortality in
patients with acute pancreatitis (8,9), the proposed revision of
the Atlanta classification suggests that patients with persistent
OF should be defined as having severe acute pancreatitis (10).
This means that patients with transient OF are considered to
have mild acute pancreatitis, and that local (peri)pancreatic
complications are not considered to contribute to the defini-
tion of severity in patients with acute pancreatitis.
Since the first attempt to classify the severity of acute pan-
creatitis by Fitz in 1889 and until the most recent Atlanta sym-
posium in 1992, a morphological component has always been
included (11). Whereas Fitz believed that the morphological
features of severe disease were evidence of pancreatic hemor-
rhage and disseminated fat necrosis, the morphological features
of severe disease in the original Atlanta classification were pan-
creatic necrosis, abscess, and pseudocyst. Since then, a number
of studies have demonstrated that infectious (peri)pancreatic
complications (IPCs), rather than the presence of necrosis
per se, are a key determinant of the high morbidity and morta-
lity in patients with acute pancreatitis (12–15). It therefore
seems reasonable to consider local complications in classifying
the severity of acute pancreatitis.
A retrospective study from the Mayo Clinic (16) showed that
patients with local pancreatic complications (as defined by the
1992 Atlanta classification) and no systemic complications at any
time during hospitalization had an almost negligible mortality
but an appreciable morbidity. Only 2 (2%) patients died among
99 patients with local complications and no OF, and this was
similar to those with mild acute pancreatitis. At the same time, it
was shown that these patients required an average stay in the ICU
of 5 days and a total hospital stay of 28 days, both of which are
more than expected for patients with mild acute pancreatitis. This
was recently confirmed by the same research group in a prospec-
tive study of 82 patients (17), as well as in a prospective study of
135 patients from Spain (18). On the basis of these findings, the
revision to the Atlanta classification should include a third cat-
egory, those with “moderate” acute pancreatitis, and these would
be those with local (peri)pancreatic complications but no persist-
ent systemic complications. These patients would have previously
been classified as having severe acute pancreatitis.
There is another subgroup of patients among those who
would have previously been classified as having severe acute
pancreatitis. This proposed category is at the severe end of
the spectrum and these patients have both local and systemic
complications during the course of acute pancreatitis. The rea-
son for defining this subgroup of patients as having extremely
severe (or “critical”) acute pancreatitis stems from findings of
several studies that demonstrated a marked difference in the
mortality rate of patients with OF depending on whether IPCs
are present or not. This was shown in a study from Switzerland
that prospectively enrolled 204 patients with acute pancreatitis,
Classification of the Severity of Acute Pancreatitis:
How Many Categories Make Sense?
Maxim S. Petrov, MD, MPH1
and John A. Windsor, MBChB, MD, FRACS1
There is an ongoing effort to revise the 1992 Atlanta classification of acute pancreatitis in the light of emerging
evidence. The categorization of the severity of acute pancreatitis is one of the key elements of the classification.
This paper aims to define the optimal number of categories and provide their definitions on sound clinical
grounds.
Am J Gastroenterol 2010; 105:74–76; doi:10.1038/ajg.2009.597; published online 20 October 2009
1
Department of Surgery, The University of Auckland, Auckland, New Zealand. Correspondence: Maxim S. Petrov, MD, MPH, Department of Surgery, The
University of Auckland, Private Bag 92019, Auckalnd 1142, New Zealand. E-mail: max.petrov@gmail.com
Received 13 May 2009; accepted 11 September 2009

2. © 2010 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY
75
REVIEW
Classification of the Severity of Acute Pancreatitis
regardless of its severity (19). There were 74 patients who devel-
oped OF, of whom 9 (12%) died. Of these, 47 patients had OF
and no evidence of IPC, of whom 2 (4%) died. This is in con-
trast to the 27 patients who had both OF and IPC, of whom 7
(26%) died. In a further prospective study of 64 patients with
predicted severe acute pancreatitis in Greece, 33 developed OF,
of whom 8 (24%) died (20). There were 12 patients with both
OF and IPC, of whom 7 (58%) died. There were 21 patients
with OF and no evidence of IPC, of whom only 1 (5%) died.
In a prospective study from France, there were 30 patients with
multiple OF, of whom 10 (33%) died (21). When OF and IPC
were both present, 9 of 23 (39%) patients died, compared with
1 of 7 (14%) who died with OF and no IPC. These studies dem-
onstrate that the presence of OF and IPC significantly increases
mortality compared with OF alone. Although patients with
both OF and IPC are not common in routine clinical practice,
they are an important subgroup because of excessive mortality.
This provides a sound rationale for the introduction of a fourth
category, namely, “critical” acute pancreatitis.
It could be argued that patients with the most severe acute
pancreatitis are those with persistent OF in the early phase of
the disease, which is a time when IPCs are rare. This argument is
primarily based on two studies that demonstrated significantly
worse outcomes, in terms of OF and mortality, for patients with
so-called “early severe acute pancreatitis,” defined as a failure of
at least one organ system at admission or within 72h after onset
of symptoms (22,23). But this seems to be a circular argument,
as the diagnosis of “early severe acute pancreatitis” required the
presence of OF, resulting in a higher rate of OF. This is a clini-
cal illustration of the phenomenon widely referred to in social
sciences as “the Oedipus Effect.” In relation to mortality, one
might suggest that early persistent OF has greater impact on the
mortality of acute pancreatitis than late IPCs, but such a sug-
gestion cannot be proven in the absence of treatments that have
been studied to the same extent and are equally effective for
both IPCs and OF in acute pancreatitis. Whereas the preven-
tion and treatment of IPCs by means of enteral tube feeding,
prophylactic antibiotics, and delayed necrosectomy have been
extensively studied in clinical studies, including several high-
quality randomized controlled trials (24–26), there is relatively
poor evidence relating to the prevention and treatment of early
OF in patients with acute pancreatitis. There have been only
some “negative” studies examining antiproteases and lexipafant
(27,28), whereas other strategies, such as goal-directed resusci-
tation, inhibition of cytokines, calcium antagonists, and decom-
pression for abdominal compartment syndrome, have not been
studied in prospective clinical studies, let alone in randomized
controlled trials. Thus, the less-established treatment strategies
for the early phase of acute pancreatitis, compared with those
for the late phase, may account for the higher mortality ascribed
to “early” severe acute pancreatitis and persistent OF.
The proposed revision of the Atlanta classification is based on
the concept of a biphasic natural course of acute pancreatitis and
recommendsthatcliniciansuseadifferentmethodofclassification
for the early phase and the late phase of acute pancreatitis (10). In
the early phase of the disease, the classification of severity is to be
based on the presence or absence of persistent OF and/or death.
In the late phase, the classification of severity is to be based on the
needfor“activeintervention(operative,endoscopic,laparoscopic,
or percutaneous) or other supportive measures (such as need for
respiratory ventilation, renal dialysis, or nasojejunal feeding),” as
well as on the presence or absence of persistent OF and/or death.
Thisapproachhassomeimportantlimitationsthatmakeitsubop-
timal. First, there is imprecision in the definition of the duration
of the early phase as “within the first 1–2 weeks of onset,” which
probably reflects the lack of consensus in the literature. Moreover,
a recent large population-based study of all deaths due to acute
pancreatitis in Scotland over a 6-year period does not support the
concept of a biphasic natural course of acute pancreatitis as it did
not reveal a bimodal distribution of mortality (6). Second, it is not
appropriate to use mortality as both an indicator of the natural
course of the disease and as a part of the definition of severity, the
latter of which lacks clinical utility postmortem. Third, significant
variance will result from a classification of severity that is based
on the need for an intervention or supportive care. This is because
there is a lack of international standardization of management,
including indications for endoscopic procedures, enteral nutri-
tion, and criteria for admission to intensive care units. Finally, the
prognosticandclinicalutilityofmanyofthesuggestednewradio-
logical terms (e.g., “acute peripancreatic fluid collection,” “acute
postnecrotic collection,” “walled-off necrosis”) has not been dem-
onstrated, and these may require further revision. Furthermore,
there is no consensus on the use of this terminology even among
radiologists, and an alternative image-based classification based
on retroperitoneal extension has been proposed (29). Therefore,
a suggestion that the radiologists should refine their imaging
criteria for the diagnosis of (peri)pancreatic complications and
that the clinicians should integrate them into a clinical classifica-
tion system seems to be reasonable and justified (30).
It is appreciated that an ideal classification of the severity of
acute pancreatitis would reflect, through its categories, clini-
cally relevant changes that occur in individual patients (31). The
proposed revision to the Atlanta classification (10) only reflects
one clinically relevant variable—persistent OF. But there is a
large body of evidence that demonstrates a broader spectrum of
clinically relevant changes in acute pancreatitis. These include
(peri)pancreatic complication (absence, sterile, infectious) and
OF (absence, transient, persistent). These clinically relevant
variables provide the basis for a classification of the severity of
acute pancreatitis comprising four categories (Table 1). This
classification of severity uses widely accepted and unambigu-
ous terms, can be applied in both early and late phases of acute
pancreatitis, and will prove to be useful in tracking individual
patients and comparing groups of patients.
We conclude that a classification of the severity of acute pan-
creatitis that includes just two categories does not reflect all the
clinically important changes in patients with acute pancreatitis.
There are sound clinical grounds for introducing two additional
categories, namely, “moderate” and “critical” acute pancreati-
tis. The four categories will better accomplish the main objec-

3. The American Journal of GASTROENTEROLOGY VOLUME 105 | JANUARY 2010 www.amjgastro.com
76
REVIEW
Petrov and Windsor
6. Mole DJ, Olabi B, Robinson V et al. Incidence of individual organ
dysfunction in fatal acute pancreatitis: analysis of 1024 death records.
HPB 2009;11:166–70.
7. Lankisch PG. Natural course of acute pancreatitis: what we know today and
what we ought to know for tomorrow. Pancreas 2009;38:494–8.
8. Buter A, Imrie CW, Carter CR et al. Dynamic nature of early organ dysfunction
determines outcome in acute pancreatitis. Br J Surg 2002;89:298–302.
9. Johnson CD, Abu-Hilal M. Persistent organ failure during the first week as
a marker of fatal outcome in acute pancreatitis. Gut 2004;53:1340–4.
10. Acute Pancreatitis Classification Working Group. Revision of the Atlanta
classification of acute pancreatitis (3rd revision). www. pancreasclub.com/
resources/AtlantaClassification.pdf Accessed 1 April 2009.
11. Pannala R, Kidd M, Modlin IM. Acute pancreatitis: a historical perspective.
Pancreas 2009;38:355–66.
12. Gloor B, Müller CA, Worni M et al. Late mortality in patients with severe
acute pancreatitis. Br J Surg 2001;88:975–9.
13. Garg PK, Madan K, Pande GK et al. Association of extent and infection
of pancreatic necrosis with organ failure and death in acute necrotizing
pancreatitis. Clin Gastroenterol Hepatol 2005;3:159–66.
14. Beger HG, Rau BM. Severe acute pancreatitis: clinical course and manage-
ment. World J Gastroenterol 2007;13:5043–51.
15. Xue P, Deng LH, Zhang ZD et al. Infectious complications in patients with
severe acute pancreatitis. Dig Dis Sci 2008; [e-pub ahead of print].
16. Vege SS, Gardner TB, Chari ST et al. Low mortality and high morbidity
in severe acute pancreatitis without organ failure: a case for revising the
Atlanta classification to include “moderately severe acute pancreatitis”.
Am J Gastroenterol 2009;104:710–5.
17. Talukdar R, Vege SS, Chari ST et al. Moderately severe acute pancreatitis:
a prospective validation study of this new subgroup of acute pancreatitis.
Pancreatology 2009;9:434.
18. De-Madaria E, Soler G, Martinez J et al. Update of the Atlanta classification
of severity of acute pancreatitis: should a moderate category be included?
Pancreatology 2009;9:433–4.
19. Büchler MW, Gloor B, Müller CA et al. Acute necrotizing pancreatitis: treat-
ment strategy according to the status of infection. Ann Surg 2000;232:619–26.
20. Lytras D, Manes K, Triantopoulou C et al. Persistent early organ failure:
defining the high-risk group of patients with severe acute pancreatitis?
Pancreas 2008;36:249–54.
21. Le Mée J, Paye F, Sauvanet A et al. Incidence and reversibility of organ fail-
ure in the course of sterile or infected necrotizing pancreatitis. Arch Surg
2001;136:1386–90.
22. Isenmann R, Rau B, Beger HG. Early severe acute pancreatitis: characteris-
tics of a new subgroup. Pancreas 2001;22:274–8.
23. Tao HQ, Zhang JX, Zou SC. Clinical characteristics and management of
patients with early acute severe pancreatitis: experience from a medical
center in China. World J Gastroenterol 2004;10:919–21.
24. Petrov MS, van Santvoort HC, Besselink MG et al. Enteral nutrition
and the risk of mortality and infectious complications in patients with
severe acute pancreatitis: a meta-analysis of randomized trials. Arch Surg
2008;143:1111–7.
25. Windsor JA. Minimally invasive pancreatic necrosectomy. Br J Surg
2007;94:132–3.
26. Petrov MS. Meta-analyses on the prophylactic use of antibiotics in acute
pancreatitis: many are called but few are chosen. Am J Gastroenterol
2008;103:1837–8.
27. Cavallini G, Frulloni L. Somatostatin and octreotide in acute pancreatitis:
the never-ending story. Dig Liver Dis 2001;33:192–201.
28. Abu-Zidan FM, Windsor JA. Lexipafant and acute pancreatitis: a critical
appraisal of the clinical trials. Eur J Surg 2002;168:215–9.
29. Ishikawa K, Idoguchi K, Tanaka H et al. Classification of acute pancreatitis
based on retroperitoneal extension: application of the concept of
interfascial planes. Eur J Radiol 2006;60:445–52.
30. Bradley EL III. Confusion in the imaging ranks: time for a change?
Pancreas 2006;33:321–2.
31. Frey CF. Classification of pancreatitis: state-of-the-art, 1986. Pancreas
1986;1:62–8.
tives of the revision of the Atlanta classification, which are to
improve clinical assessment, facilitate communication between
treating physicians and promote standardization for reporting
clinical studies.
ACKNOWLEDGMENTS
We are indebted to Professor Peter A. Banks (Brigham and
Women’s Hospital, Harvard Medical School, Boston, MA) for
helpful discussion. Dr. Maxim S. Petrov is supported by the
Kenneth Warren Foundation of the International
Hepato-Pancreato-Biliary Association.
CONFLICT OF INTEREST
Guarantor of the article: Maxim S. Petrov, MD, MPH.
Specific author contributions: Planning, conducting, and
drafting the manuscript: Maxim S. Petrov; drafting and critical
reviewing of the manuscript: John A. Windsor.
Financial support: None.
Potential competing interests: None.
REFERENCES
1. Bradley EL III. A clinically based classification system for acute pancrea-
titis. Summary of the International Symposium on Acute Pancreatitis,
Atlanta, Ga, September 11 through 13, 1992. Arch Surg 1993;128:586–90.
2. Vege SS, Chari ST. Organ failure as an indicator of severity of acute pancreatitis:
time to revisit the Atlanta classification. Gastroenterology 2005;128:1133–5.
3. Banks PA, Freeman ML. Practice guidelines in acute pancreatitis. Am J
Gastroenterol 2006;101:2379–400.
4. Pandol SJ, Saluja AK, Imrie CW et al. Acute pancreatitis: bench to the
bedside. Gastroenterology 2007;132:1127–51.
5. Flint R, Windsor JA. Early physiological response to intensive care as a
clinically relevant approach to predicting the outcome in severe acute
pancreatitis. Arch Surg 2004;139:438–43.
Table 1. Classification and definitions of four categories for
the severity of acute pancreatitis
Severity
category
Local complications Systemic
complications
Mild No (peri)pancreatic
complication
and No organ failure
Moderatea
Sterile (peri)pancreatic
complication
or Transient organ
failure
Severea
Infectious (peri)pancreatic
complication
or Persistent organ
failure
Critical Infectious (peri)pancreatic
complication
and Persistent organ
failure
a
Severity is graded on the basis of more severe local or systemic complication
(e.g., sterile pancreatic necrosis without organ failure has to be graded as
“moderate”; sterile pancreatic necrosis with persistent organ failure has to be
graded as “severe”).