Table 1.

Pancreatic duct occlusion

Ref	Study design	Population	Intervention and comparison	Co-interventions	Outcomes	Comment
	RCT, multicentre; stratified randomisation balancing every 4 people within each stratum (type of resection, pancreaticoduodenectomy, or distal pancreatectomy; pathology; tumour or chronic pancreatitis) and at each centre	182 people having pancreaticoduodenectomy (type of surgery not reported): 65% with malignant disease; mean age 56 years	Anastomosis plus duct occlusion v anastomosis alone (use of PJ v PG was similar between groups, 60%–70% having PJ)	Significantly more people having duct occlusion also received octreotide (53% with duct occlusion v 26% with anastomosis alone; P less than 0.001) and had reinforcement of the anastomosis with fibrin glue (59% with duct occlusion v 10% with anastomosis alone; P less than 0.001)	Perioperative mortality:No significant difference between adding duct occlusion and anastomosis alone in perioperative mortality at 1 month (9/102 [9%] with duct occlusion v 4/80 [6%] with anastomosis alone; reported as not significant; P value not reported)	Significantly more participants having occlusion had fibrotic pancreatic stumps (46% with duct occlusion v 30% with anastomosis alone; P = 0.02); this makes the results difficult to interpret. Multivariate analysis suggested that normal pancreatic parenchyma significantly influenced the onset of intra-abdominal collections (OR 3.23, 95% CI 1.30 to 8.17)
					Intra-abdominal collections:No significant difference between adding duct occlusion and anastomosis alone (15% with duct occlusion v 24% with anastomosis alone; P = 0.20)
					Pancreatic fistula:No significant difference between adding duct occlusion and anastomosis alone (9% with duct occlusion v 6% with anastomosis alone; reported as not significant; P value not reported)
					Blood loss:No significant difference between adding duct occlusion and anastomosis alone (7% with duct occlusion v 14% with anastomosis alone; reported as not significant; P value not reported)
	RCT, single-centre	169 people having pancreaticoduodenectomy: 62% pylorus-preserving; 59% with pancreatic adenocarcinoma; 27% with peri-ampullary cancer	Anastomosis plus duct occlusion v anastomosis alone (use of PJ v PG was similar between groups, 60%–70% having PJ)	Use of octreotide not reported	Survival at 1 year:Similar with duct occlusion and PJ (63% with duct occlusion v 69% with PJ; significance not assessed; absolute numbers not reported)	Survival was a secondary end point
					Subgroup analysis in participants with malignant disease (58% with duct occlusion v 66% with PJ; significance not assessed; absolute numbers not reported)
					Overall complications:No significant difference between duct occlusion and PJ (proportion with no complications: 54/86 [64%] with duct occlusion v 63/83 [76%] with PJ; P = 0.07)
					Pancreatic fistula:Significantly higher rates with duct occlusion compared with PJ (15/86 [17%] with duct occlusion v 4/83 [5%] with PJ; P = 0.02)
					Intra-abdominal abscesses:No significant difference between duct occlusion and PJ, although lower with PJ (13/86 [15%] with duct occlusion v 8/83 [10%] with PJ; P = 0.35)
					Postoperative blood loss:No significant difference between duct occlusion and PJ (7/86 [8%] with duct occlusion v 6/83 [7%] with PJ; P = 1.00)
					Need for enzyme replacement:Significantly higher with duct occlusion compared with PJ (77/86 [90%] with duct occlusion v 57/83 [67%] with PJ; P = 0.01) and at 3 months (75/86 [87%] with duct occlusion v 63/83 [76%] with PJ; P = 0.03) but similar at 12 months (51/86 [59%] with duct occlusion v 48/83 [58%] with PJ; P value not reported)
					Diabetes mellitus:Significantly increased at 3- and 12-month follow-up in the duct occlusion group (at 12 months: 34% with duct occlusion v 14% with PJ; P = 0.001; follow-up of 105 people; absolute numbers not reported)

PG, pancreaticogastrostomy; PJ, pancreaticojejunostomy; Ref, reference.