When "mild" isn't mild: Champalimaud Foundation surgeons propose a new way to grade bleeding after pancreatic surgery

Surgery is often the main treatment for pancreatic cancer and for several other pancreatic conditions. It is also one of the most complex surgical operations in medicine. A particularly feared complication afterwards is internal bleeding, known as postpancreatectomy haemorrhage. Since 2007, surgeons worldwide have graded these bleeds with an international system. It combines three factors: when the bleeding starts, where it happens and how severe it seems.

The team, led by first author Filipe C. Borges and senior author Markus W. Büchler, set out to test whether this system still reflects reality. They analysed 763 patients who had pancreatic surgery at the BCPCC between July 2020 and September 2025. About one in eight of them (91 patients) had bleeding after the operation. That bleeding mattered as these patients were far more likely to have serious complications and to stay longer in hospital.

The results were striking. Neither the timing nor the location of the bleeding significantly affected how patients did. Bleeds labelled "mild" under the 2007 grading system were often far from mild. Around 7 in 10 of these patients still developed serious complications, and almost 6% died.

"A grading system should tell doctors how worried they need to be, and the current one often doesn't," says Filipe Borges. "We wanted something simpler that reflects what actually happens to the patient."

The team's answer is a classification built on one question: what are the consequences for the patient? Bleeding that has no consequences is classed as grade A. These patients do as well as those who don't bleed at all, so it no longer needs to be reported as a complication. When the bleeding needs action, such as a blood transfusion, an endoscopic procedure, a minimally invasive procedure guided by imaging or another operation, it becomes grade B. Grade C is kept for life-threatening bleeding that leads to organ failure or death.

When the team applied the new system to their patients, it separated risk clearly. Fewer than 2% of grade B patients died, compared with 30% in grade C. In other words, the most serious grade was reserved for the patients in real danger.

These observations may also be relevant to other hospitals. Shared definitions are how surgeons around the world compare results, judge new techniques and improve care. A system that is easier to use and better matches what happens to patients could make those comparisons more reliable. "This is only the beginning," says Filipe Borges. "We looked back at the records of a single centre, so this potentially new classification now needs to be tested at other hospitals before it can be widely adopted."
 

Text by Teresa Fernandes, Co-Coodinator of the Champalimaud Foundation's Communication, Events & Outreach Team