Does antecolic reconstruction for duodenojejunostomy improve delayed gastric emptying after pylorus-preserving pancreaticoduodenectomy? A systematic review and meta-analysis

Table 4 Sensitivity analysis performed for studies comparing antecolic and retrocolic reconstruction for duodenojejunostomy

Outcome	Number of studies	Number of patients		OR/MD	95% CI	P value	Heterogeneity (I2)
		AG	RG
Randomized controlled trials
Delayed gastric emptying	2[17,19]	2/37	14/38	0.1	0.02, 0.47	0.004	0%
Mortality	2[17,19]	0/37	1/38	0.32	0.01, 8.26	0.49	_
Postoperative hospital stay (d)	2[17,19]	37	38	-7.4	-27.2, 12.40	0.46	79%
Non-randomized controlled trials
Morbidity	2[15,16]	26/178	49/154	0.33	0.19, 0.57	< 0.00 001	0%
Delayed gastric emptying	3[14,15,16]	15/203	82/173	0.05	0.01, 0.20	< 0.00 001	72%
Reconstruction with Billroth II in the two groups
Morbidity	2[16,19]	20/117	30/118	0.85	0.16, 4.69	0.86	79%
Delayed gastric emptying	3[16,17,19]	7/137	38/138	0.15	0.06, 0.34	< 0.00 001	0%
Mortality	3[16,17,19]	0/137	1/138	0.32	0.01, 8.26	0.49	_
Postoperative hospital stay (d)	2[17,19]	37	38	-7.4	-27.2, 12.40	0.46	79%
Reconstruction with Billroth II in AG and Billroth I in RG
Delayed gastric emptying	2[14,15]	10/103	58/73	0.03	0.01, 0.06	< 0.00 001	0%

AG: Antecolic group; RG: Retrocolic group; OR: Odds ratio; MD: Mean difference.