Aims Pancreatic leaks (PLs) are common complications following pancreatic surgery and
acute pancreatitis. Advancements in interventional endoscopy have considerably changed
their management, reducing the need for surgical intervention. The most widely used
classification, the International Study Group of Pancreatic Fistulas (ISGPF) Classification,
focuses on the clinical severity; in 2017 our group proposed an endoscopy-oriented
classification that integrates anatomical and functional criteria, aiming to guide
endoscopic treatment selection more effectively. We aimed to assess the clinical validity
and utility of an endoscopy-oriented classification for pancreatic leaks.
Methods We retrospectively reviewed all patients who underwent endoscopic retrograde cholangiopancreatography
(ERCP) for PLs between 1st February 2012 and 31st July 2024. Data collected included patient demographics, etiology of the PL, type
of endoscopic treatment, need for second-line endoscopic interventions, technical
and clinical success, early and late complications. PLs were classified as follows:
type I for side branch leaks (subgroups IH: head, IB: body, IT: tail); type II for
main pancreatic duct leaks, with subtypes II O (“open”) if no stricture or complete
disruption is present, and II C (“closed”), otherwise. Type III leaks occur in surgically
altered anatomy, with subtype IIID after pancreatoduodenectomy and subtype IIIP after
distal pancreatectomy.
Results A total of 109 patients treated endoscopically for PLs or fistulas were included.
Causes of pancreatic duct leaks included pancreatic or abdominal surgery (n=76), acute
or chronic pancreatitis (n=22), trauma (n=8), and post-ERCP complications (n=3). The
median age of patients was 61 years [IQR 46.0 – 72.0], with a male-to-female ratio
of 82:27. At presentation, 31 patients (28.4%) had sepsis. In our population, PLs
were distributed as follows: type I (n=16; subcategories IH n=5, IB n=2, IT n=9),
type II (n=39; subcategories IIO n=28, IIC n=11), and type III (n=54; subcategories
IIID n=34, IIIP n=20). This classification allowed targeted treatment selection. Technical
success was achieved in 100% of cases, while clinical success was 84.4%, with 7.3%
unevaluable due to loss to follow-up or death from other causes. Clinical success
required a mean of 2 procedures for patient (IQR 1 – 3). Five out of 92 patients (5.4%)
required an endoscopic ultrasound (EUS)-guided rendezvous procedure, and 4/109 (3.7%)
underwent a combined ERCP-EUS treatment. Transitioning to an EUS-guided approach was
necessary in 12/109 patients (11%) to achieve clinical success (Table 1).
Conclusions Our single-center experience shows high success rates achieved by tailoring treatment
on the anatomic and functional features of the leak according to the above-mentioned
classification. An endoscopy-oriented classification could lead to a better standardization
of treatment, guide physician in the clinical setting and improve outcomes.
