MIAMI BEACH* – There appears to be no difference in the rate of anastomotic leak in pancreaticogastrostomy compared with pancreaticojejunostomy reconstruction after a Whipple procedure, but pancreaticogastrostomy does appear to increase the risk of some serious postoperative complications, a study found.
"We didn’t show any significant differences in terms of leaks, severity of leaks, or even overall complications," Dr. Jan Grendar said at the annual meeting of the Americas Hepato-Pancreato-Biliary Association. "But we did show a significant difference in complications of Clavien grades III-V severity. This anastomosis just did a little bit worse in terms of severe complications."
Grade III Clavien complications are those that require a second invasive procedure, like reoperation or drainage, with local or general anesthesia. Grade IV complications are more serious – they include organ failures such as renal failure requiring hemodialysis, heart failure, and liver failure, and intubation. The Grade V complication is patient death.
Dr. Grendar, a surgical resident at the University of Calgary (Alta.), and his colleagues randomized 98 patients with benign or malignant pancreatic disease to either of the two reconstruction techniques after a Whipple procedure. Pancreaticojejunostomy (PJ) is the procedure typically performed in North America. Pancreaticogastrostomy (PG) is typically performed in Europe. The hepatobiliary pancreatic surgeons and residents at the university, however, have become adept at this second procedure and employ it as an alternative reconstruction, depending on the characteristics of the pancreas discovered during surgery. They most often use it on patients who have a soft pancreas or small pancreatic duct.
All of the patients in the study had a pancreatic or periampullary neoplasm that appeared resectable on preoperative imaging. There were no differences in baseline patient characteristics. The mean age was 64 in the PG group and 68 in the PJ group. The median Charlson comorbidity index score was 2.
The intraoperative outcomes were mostly similar. Gland mobilization occurred significantly more often in the PG group (31 vs. 18 mm). Significantly fewer PG patients had a two-layer anastomosis (77% vs. 94%). The mean pancreatic duct size was smaller in the PG patients (3.8 vs. 4.3 mm), but this wasn’t statistically significant. Significantly fewer PG patients had a pancreatic duct stent placed (22% vs. 83%).
Postoperative outcomes included pancreatic fistula, any postoperative complications, severe complications, and length of stay.
There was no difference between the groups in Clavien grade complications separately, but significantly more PG patients had severe, Clavien grade III-V complications (31% vs. 12%). But these didn’t affect the mean length of stay, which was similar between the groups (17 vs. 14 days).
In a multivariate analysis, only having a soft pancreas significantly increased the risk of developing a pancreatic fistula (odds ratio, 5.87). Female gender significantly decreased the risk (OR, 0.31). This most likely represents a correlation between pancreatic texture and patient sex rather than a true protective property of being female. Having a soft pancreas also significantly increased the risk of any postoperative complications (OR, 3.08).
Several factors significantly increased the risk of a severe complication. These included a baseline American Society of Anesthesiologists grade of 2 or 3 (OR, 12.75 and 29.56, respectively). Being randomized to the PJ procedure significantly decreased the risk of a severe complication (OR, 0.11).
"Unlike the European centers that report different outcomes, these results are coming from a center that wasn’t very enthusiastic about this type of reconstruction prior to this study," Dr. Grendar said in an interview. "Despite the initial preference, high-risk patients in Calgary, those with soft pancreatic glands and small pancreatic ducts, are now likely to be offered a pancreaticogastrostomy"
Dr. Grendar said he had no relevant financial disclosures.
*Correction, 3/11/2014: An earlier version of the article misstated the name of the city where the AHPBA meeting took place.