Early postoperative feeding in resectional gastrointestinal surgical cancer patients

Table 1 Randomised controlled trials investigating early feeding published since 2005

Study	Year	Types ofGastrointestinal Surgery	n(Trad/Early)	Earlyfeeding protocol	Outcomes
Lucha et al[4]	2005	Open colorectal surgery	25/26	Regular diet from 8 hr following surgery	No difference in post operative complications between groups (1 d vs 1 d) or LOS 6.6 d vs 6.3 d
Zhou et al[5]	2006	Excision and anastomosis for colorectal tumour	155/161	Liquid fibreless diet D1-3 post op	Statistically significant benefits of early feeding Flatus 3.0 ± 0.9 d vs 3.6 ± 1.2 d, P = 0.000 Stool 4.1 ± 1.1 d vs 4.8 ± 1.4 d, P = 0.000 LOS 8.4 ± 3.4 d vs 9.6 ± 5.0 d, P = 0.016 Reduced complications with early feeding Reduced febrile illness: 3 vs 15, P = 0.042 Pulmonary infection: 1 vs 7, P = 0.034 Pharyngolaryngitis: 5 vs 36, P = 0.000 No differences in wound complications 4 vs 3, P = 1.0 No differences in anastomotic leakage 2 vs 4, P = 0.441
Han-Geurts et al[6]	2007	Open colorectal surgery	50/46	Regular diet from D1 post op	No statistically significant differences in outcomes between groups in any in-hospital complication, including mortality. No statistically significant differences between return of bowel function and length of hospital stay between groups
Lassen et al[7]	2008	Hepatic, pancreatic, oesophageal, gastric resections, bilioenteric and gastroenteric bypass procedures, unspecified procedures in which traditional NBM management would be indicated	227/220	Early oral feeding provided with ordinary hospital diet from D1 post op NB control group received enteral nutrition via a jejunostomy tube from D1 post op	No differences between number of patients major complications between groups (33% in jejunum fed vs 28% early oral, P = 0.26); less overall complications in early oral feeding group (100 vs 165, P = 0.012) No differences in mortality between groups within the trial period (8.4% early jejunum feeding vs 5.9% early oral, P = 0.36) Increased likelihood of intra-abdominal abscesses in gastrectomy patients with early jejunum feeding vs early oral intake (6 vs 0, P = 0.012) Shorter duration to passage of flatus early oral feeding group (2.6 vs 3.0 d, P = 0.01); no difference for duration to first bowel motion (4.3 vs 4.0 d, P = 0.112) Longer length of stay with jejunum fed patients (16.7 vs 13.5 d, P = 0.046)

Table 2 Comparison of outcomes and characteristics of published meta-analyses on early feeding

	Lewis, Egger, Sylvester & ThomasBMJ 2001[8]	Andersen, Lewis & ThomasCochrane Database Syst Rev 2006[9]	Lewis, Andersen & ThomasJ Gastrointest Surg 2009[10]
Inclusion criteria	Elective gastrointestinal surgery RCTs Enteral feeding within 24 h post op vs NBM/traditional management Included unpublished data	RCTs (un/published) Colorectal surgery Early feeding (within 24 h) vs NBM Malignant/benign disease incl. IBD Studies solely in paediatric population RCTs with no blinding If reported on outcomes including adverse outcomes, mortality	RCTs (unpublished/published) Colorectal surgery Early feeding (within 24 h) vs NBM Malignant/benign disease including inflammatory bowel diseases Studies solely in paediatric population RCTs with no blinding If reported on outcomes including adverse outcomes, mortality
Exclusion criteria	Not stated	PN Non-RCTs Unpublished abstracts with no correspondence data	PN Non-RCTs Unpublished abstracts with no correspondence data
Number of patients	929	1173	1173
Number of included studies	11	13	13
Publication dates	1979-1998	1979-2004	1979-2004
Gastrointestinal surgery types included	Colonic, ileal or colonic resection; oesophago-gastrectomy, gastrectomy, ileoanal J pouch, reanastomosis; esophagectomy, pancreatoduodenectomy; unspecified laparotomy	Colonic, ileal or colonic resection; oesophago-gastrectomy, gastrectomy, ileoanal J pouch, reanastomosis; esophagectomy, pancreatoduodenectomy; unspecified laparotomy	Colonic, ileal or colonic resection; oesophago-gastrectomy, gastrectomy, ileoanal J pouch, reanastomosis; esophagectomy, pancreatoduodenectomy; unspecified laparotomy
Outcomes
Wound infections	RR 0.71 (0.44-1.17) χ2 value not reported, P = 0.074	RR 0.77 (0.48-1.22) P = 0.3 (FEM) χ2 = 10.30 P = 0.26	RR 0.78 (0.38, 1.68) (REM) RR 0.77 ( 0.48-1.22) P = 0.3 (FEM) χ2 = 10.30 P = 0.26
Intra-abdominal abscesses	RR 0.87 (0.31-2.42) χ2 value not reported, P = 0.84	RR 0.87 (0.31–2.42) P = 0.8 χ2 = 1.45 P = 0.84	RR 0.94 (0.32, 2.77) (REM) RR 0.87 (0.31–2.42) P = 0.8 (FEM) χ2 = 1.45 P = 0.84
Pneumonia	RR 0.73 (0.33–1.59) χ2 value not reported, P = 0.85	RR 0.76 (0.36-1.58) P = 0.5 χ2 = 3.73 P = 0.81	RR 0.71 (0.32, 1.59) (REM) RR 0.76 (0.36-1.58) P = 0.5 (FEM) χ2 = 3.73 P = 0.81
Any infection	RR 0.72 (0.54-0.98) P = 0.036 χ2 = 10.7, P = 0.22	Not assessed	Not assessed
Mortality	RR 0.48 (0.18-1.29) P = 0.15 χ2 value not reported, P = 0.99	RR 0.41 (0.18-0.93) P = 0.03 χ2 = 0.6 P = 0.99	RR 0.42 (0.18, 0.96) (REM) RR 0.41 (0.18-0.93) P = 0.03 (FEM) χ2 = 0.6 P = 0.99
Anastomotic dehiscence	RR 0.53 (0.26-1.08) P = 0.08 χ2 = 2.1, P = 0.96 NB-little evidence that data from proximal vs distal feeding results differed P = 0.42	RR 0.69 (0.39-1.32) P = 0.3 χ2 = 4.89 P = 0.77	RR 0.62 (0.30, 1.28) (REM) RR 0.69 (0.39-1.32) P = 0.3 (FEM) χ2 = 4.89, P =0.77 for FEM. No χ2 reported for REM
Length of hospital stay	-0.84 d (-0.36-1.33) P = 0.001 χ2 = 16.2, P = 0.094	-0.60 d (-0.66, -0.54) χ2 = 18.86 P = 0.06	-0.89 d (-1.58, -0.20) (REM) -0.60 d (-0.66, -0.54) (FEM) χ2 = 18.86 P = 0.06
Vomiting	RR 1.27 (1.01-1.61) P = 0.045 χ2 value not reported, P = 0.52 NB-non-significant increase in N&V with early feeding where NGs were not placed at time of surgery RR 1.21 (0.73-1.99) P = 0.46	RR 1.27, (1.01-1.61) P = 0.04 χ2 = 4.21 P = 0.52	RR 1.23 (0.97, 1.55) (REM) RR 1.27 (1.01-1.61) (FEM) χ2 = 4.21 P = 0.52

95% Confidence intervals in closed brackets. RCT: Randomised controlled trial; PN: Parenteral nutrition; NBM: Nil by mouth; RR: Relative risk ratio; FEM: Fixed effects model (of meta-analysis); REM: Random effects model (of meta-analysis); N&V: Nausea and vomiting.