Preoperative nutrition therapy in people undergoing gastrointestinal surgery

Abstract

Background

Poor preoperative nutritional status has been consistently linked to an increase in postoperative complications and worse surgical outcomes. We updated a review first published in 2012.

Objectives

To assess the effects of preoperative nutritional therapy compared to usual care in people undergoing gastrointestinal surgery.

Search methods

We searched the Cochrane Central Register of Controlled Trials, MEDLINE, Embase, three other databases and two trial registries on 28 March 2023. We searched reference lists of included studies.

Selection criteria

We included randomised controlled trials (RCTs) of people undergoing gastrointestinal surgery and receiving preoperative nutritional therapy, including parenteral nutrition, enteral nutrition or oral nutrition supplements, compared to usual care. We only included nutritional therapy that contained macronutrients (protein, carbohydrate and fat) and micronutrients, and excluded studies that evaluated single nutrients. We included studies regardless of the nutritional status of participants, that is, well‐nourished participants, participants at risk of malnutrition, or mixed populations. We excluded studies in people undergoing pancreatic and liver surgery.

Our primary outcomes were non‐infectious complications, infectious complications and length of hospital stay. Our secondary outcomes were nutritional aspects, quality of life, change in macronutrient intake, biochemical parameters, 30‐day perioperative mortality and adverse effects.

Data collection and analysis

We used standard Cochrane methodology. We assessed risk of bias using the RoB 1 tool and applied the GRADE criteria to assess the certainty of evidence.

Main results

We included 16 RCTs reporting 19 comparisons (2164 participants). Seven studies were new for this update. Participants' ages ranged from 21 to 79 years, and 62% were men. Three RCTs used parenteral nutrition, two used enteral nutrition, eight used immune‐enhancing nutrition and six used standard oral nutrition supplements. All studies included mixed groups of well‐nourished and malnourished participants; they used different methods to identify malnutrition and reported this in different ways. Not all the included studies were conducted within an Enhanced Recovery After Surgery (ERAS) programme, which is now current clinical practice in most hospitals undertaking GI surgery.

We were concerned about risk of bias in all the studies and 14 studies were at high risk of bias due to lack of blinding.

We are uncertain if parenteral nutrition has any effect on the number of participants who had a non‐infectious complication (risk ratio (RR) 0.61, 95% confidence interval (CI) 0.36 to 1.02; 3 RCTs, 260 participants; very low‐certainty evidence); infectious complication (RR 0.98, 95% CI 0.53 to 1.80; 3 RCTs, 260 participants; very low‐certainty evidence) or length of hospital stay (mean difference (MD) 5.49 days, 95% CI 0.02 to 10.96; 2 RCTs, 135 participants; very low‐certainty evidence).

None of the enteral nutrition studies reported non‐infectious complications as an outcome. The evidence is very uncertain about the effect of enteral nutrition on the number of participants with infectious complications after surgery (RR 0.90, 95% CI 0.59 to 1.38; 2 RCTs, 126 participants; very low‐certainty evidence) or length of hospital stay (MD 5.10 days, 95% CI −1.03 to 11.23; 2 RCTs, 126 participants; very low‐certainty evidence).

Immune‐enhancing nutrition compared to controls may result in little to no effect on the number of participants experiencing a non‐infectious complication (RR 0.79, 95% CI 0.62 to 1.00; 8 RCTs, 1020 participants; low‐certainty evidence), infectious complications (RR 0.74, 95% CI 0.53 to 1.04; 7 RCTs, 925 participants; low‐certainty evidence) or length of hospital stay (MD −1.22 days, 95% CI −2.80 to 0.35; 6 RCTs, 688 participants; low‐certainty evidence).

Standard oral nutrition supplements may result in little to no effect on number of participants with a non‐infectious complication (RR 0.90, 95% CI 0.67 to 1.20; 5 RCTs, 473 participants; low‐certainty evidence) or the length of hospital stay (MD −0.65 days, 95% CI −2.33 to 1.03; 3 RCTs, 299 participants; low‐certainty evidence). The evidence is very uncertain about the effect of oral nutrition supplements on the number of participants with an infectious complication (RR 0.88, 95% CI 0.60 to 1.27; 5 RCTs, 473 participants; very low‐certainty evidence). Sensitivity analysis based on malnourished and weight‐losing participants found oral nutrition supplements may result in a slight reduction in infections (RR 0.58, 95% CI 0.40 to 0.85; 2 RCTs, 184 participants).

Studies reported some secondary outcomes, but not consistently.

Complications associated with central venous catheters occurred in RCTs involving parenteral nutrition. Adverse events in the enteral nutrition, immune‐enhancing nutrition and standard oral nutrition supplements RCTs included nausea, vomiting, diarrhoea and abdominal pain.

Authors' conclusions

We were unable to determine if parenteral nutrition, enteral nutrition, immune‐enhancing nutrition or standard oral nutrition supplements have any effect on the clinical outcomes due to very low‐certainty evidence. There is some evidence that standard oral nutrition supplements may have no effect on complications. Sensitivity analysis showed standard oral nutrition supplements probably reduced infections in weight‐losing or malnourished participants. Further high‐quality multicentre research considering the ERAS programme is required and further research in low‐ and middle‐income countries is needed.

Plain language summary

Is supplemental nutrition before surgery better than a regular diet for people having surgery on the digestive system?

Key messages

– We do not know if nutritional supplements in addition to normal diet are better for people having surgery on the digestive system due to uncertainty in the evidence.

– There is some evidence that oral supplements reduce infections in people who have lost weight or are malnourished.

– More research is required in this area.

Background

People who will be undergoing surgery on the digestive tract often have malnutrition before the surgery (that is, they have a poor level of nourishment, possibly because their diet does not contain the right amount of nutrients or their bodies do not absorb nutrients well). This can increase the possibility of death, other illnesses or problems, and result in longer hospital stays. Research shows that malnutrition and weight loss are linked to infections and other complications after surgery on the digestive system, such as poor healing at the site of surgery, the heart not pumping blood around the body properly, blood clots or bleeding. We looked at the effects of providing extra nourishment to people before an operation.

What did we want to find out?

We wanted to know if giving people nutritional supplements orally, via a tube into the stomach (enterally) or via a tube into the veins (parenterally) in addition to normal food before an operation on the digestive tract would lead to fewer non‐infectious complications (that is, complications other than an infection), infectious complications (that is, complications related to infections at the site of surgery or elsewhere in the body) and shorter stays in hospital. For the oral supplements, we looked at standard supplements and immune‐enhancing supplements (which are enriched with ingredients to boost the body's immune system to help fight infections).

What did we do?

We searched scientific databases for relevant studies that compared additional nutritional therapy versus usual care in people undergoing surgery on the digestive tract.

Main findings

We identified 16 studies including 2164 participants. For parenteral and enteral feeding, we are uncertain if there was any effect on the number of people with a non‐infectious complication, an infectious complication or length of hospital stay. For oral supplements with immune‐enhancing ingredients, we were uncertain if there was any effect on the number of people with a non‐infectious complication, an infectious complication or length of hospital stay. For standard oral supplements, we found little or no effect on the number of people with a non‐infectious or infectious complication. Although in two studies that looked at people losing weight or those who were malnourished, standard oral supplements probably reduced infections. Standard oral supplements may reduce length of hospital stay compared to usual care.

What are the limitations of this evidence?

We had limited confidence in the results as, for example, some studies were old, methods and results varied between studies, and in some studies, participants and researchers knew which treatment they received so may have favoured one treatment over another.

How up to date is this evidence?

This was an update of a review first published in 2012. The latest search was 28 March 2023.

Summary of findings

Summary of findings 1. Preoperative parenteral nutrition compared to usual care or control in people undergoing gastrointestinal surgery.

Preoperative parenteral nutrition compared to usual care/control in people undergoing gastrointestinal surgery
Patient or population: people undergoing gastrointestinal surgery Setting: hospital inpatients Intervention: preoperative parenteral nutrition Comparison: usualcare/control
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with control	Risk with preoperative parenteral nutrition compared usual care/control
Non‐infectious complications	262 per 1000	149 per 1000 (89 to 244)	RR 0.61 (0.36 to 1.02)	260 (3 RCTs)	⊕⊝⊝⊝ Very lowa,b	Time point outcome measured and not reported.
Infectious complications	587 per 1000	534 per 1000 (200 to 1000)	RR 0.98 (0.53 to 1.80)	260 (3 RCTs)	⊕⊝⊝⊝ Very lowa,b,c	Time point outcome measured and not reported.
Length of hospital stay (days)	The mean length of stay was 34.5 days	MD 5.49 higher (0.02 higher to 10.96 higher)	Not estimable	135 (2 RCTs)	⊕⊝⊝⊝ Very lowa,b	—
Adverse effects	—	—	—	260 (3 RCTs)	—	1 trial reported 7 instances of elevated liver function tests and 4 catheter‐related adverse events. 1 trial reported 2 febrile episodes. 1 trial reported 3 cases of diarrhoea in the enteral group, 3 cases of vomiting, 2 cases of gastric retention, 1 arterial puncture, 1 pneumothorax and 4 cases of catheter‐related sepsis.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded two levels as the studies were at high risk of bias. Studies had unclear or low risk of bias for selection, attrition and reporting bias. However, studies had high risk of bias for performance and detection and other bias.
^bDowngraded one level due to indirectness. These studies were old and clinical practice has changed substantially in relation to parenteral nutrition.
^cDowngraded one level due to inconsistency (high level of statistical heterogeneity, I² = 79%)

Summary of findings 2. Preoperative enteral nutrition (standard formula) compared to usual care or control in people undergoing gastrointestinal surgery.

Preoperative enteral nutrition (standard formula) compared to usual care/control in people undergoing gastrointestinal surgery
Patient or population: people undergoing gastrointestinal surgery Setting: hospital inpatients Intervention: preoperative enteral nutrition (standard formula) Comparison: usualcare/control
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with control	Risk with preoperative enteral nutrition (standard formula) compared to usual care/control
Non‐infectious complications	—	—	—	—	—	Not reported
Infectious complications	475 per 1000	456 per 1000 (313 to 669)	RR 0.90 (0.59 to 1.38)	126 (2 RCTs)	⊕⊝⊝⊝ Very lowa,b,c	Time point outcome measured and not reported.
Length of hospital stay (days)	The mean length of hospital stay was 18.4 days	MD 5.10 higher (1.03 lower to 11.23 higher)	Not estimable	126 (2 RCTs)	⊕⊝⊝⊝ Very lowa,b	Time point outcome measured and not reported.
Adverse effects	—	—	—	26 (1 RCT)	—	No adverse effects reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded two levels due to high risk of bias. Studies had an unclear risk of bias for selective reporting, random sequence generation and allocation concealment. There was a high risk of bias for lack of blinding and other bias. 
^bDowngraded one level for imprecision due to wide confidence intervals that overlapped the line of no effect.
^cDowngraded one level for inconsistency between studies.

Summary of findings 3. Immune‐enhancing supplements (oral or enteral) compared to usual care or control in people undergoing gastrointestinal surgery.

Immune‐enhancing supplements (oral or enteral) compared to usual care/control in people undergoing gastrointestinal surgery
Patient or population: people undergoing gastrointestinal surgery Setting: hospital inpatients Intervention: immune‐enhancing supplements (oral or enteral) Comparison: usualcare/control
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with control	Risk with immune‐enhancing supplements (oral or enteral) compared to usual care/control
Non‐ infectious complications recorded up to 30 days postdischarge	263 per 1000	216 per 1000 (163 to 282)	RR 0.79 (0.62 to 1.00)	1020 (8 RCTs)	⊕⊕⊝⊝ Lowa,b	—
Infectious complications recorded up to 30 days postdischarge	233 per 1000	173 per 1000 (124 to 245)	RR 0.74 (0.53 to 1.04)	925 (7 RCTs)	⊕⊕⊝⊝ Lowa,b	—
Length of hospital stay (days)	The mean length of hospital stay was 11.25 days	MD 1.22 lower (2.80 lower to 0.35 higher)	Not estimable	688 (6 RCTs)	⊕⊕⊝⊝ Lowa,c	—
Adverse effects	—	—	—	399 (3 RCTs)	—	1 trial reported 1 case of diarrhoea like symptoms, 1 spike in blood sugar. 1 trial reported 11 cases of abdominal cramping or bloating, 8 cases of diarrhoea, 3 cases of postoperative vomiting and the nasojejunal tube blocked in 5 participants. 1 trial reported cramping and distention, diarrhoea and vomiting.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded one level for risk of bias. Studies had an unclear or high risk of bias for blinding and an unclear bias in some studies for attrition and reporting bias.
^bDowngraded one level due to imprecision due to wide confidence intervals that overlapped the line of no effect. 
^cDowngraded one level for inconsistencies between studies (high level of statistical heterogeneity, I² = 76%).

Summary of findings 4. Oral nutritional supplements compared to usual care or control in people undergoing gastrointestinal surgery.

Oral nutritional supplement compared to usual care/control in people undergoing gastrointestinal surgery
Patient or population: people undergoing gastrointestinal surgery Setting: hospital inpatients Intervention: preoperative standard oral supplement compared to usual care/control Comparison: usualcare/control
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with control	Risk with preoperative standard oral supplement compared to usual care/control
Non‐infectious complications recorded up to 30 days postdischarge	341 per 1000	324 per 1000 (252 to 412)	RR 0.90 (0.67 to 1.20)	473 (5 RCTs)	⊕⊕⊝⊝ Lowa,b	—
Infectious complications recorded up to 30 days postdischarge	389 per 1000	343 per 1000 (238 to 502)	RR 0.88 (0.60 to 1.27)	473 (5 RCTs)	⊕⊝⊝⊝ Very lowa,b,c	Sensitivity analysis on 2 studies (184 participants) with data available on malnourished or weight‐losing participants found that standard oral supplements probably decreased infections (RR 0.53, 95% CI 0.37 to 0.77; I2 = 0%).
Length of hospital stay (days)	The mean length of hospital stay was 13.3 days	MD 0.65 lower (2.33 lower to 1.03 higher)	—	299 (3 RCTs)	⊕⊕⊝⊝ Lowa,b	—
Adverse effects	—	—	—	373 (4 RCTs)	—	1 trial reported 4 cases of nausea and vomiting, 2 cases of diarrhoea. 1 trial reported 4 cases of nausea, 3 cases of abdominal discomfort and 2 cases of diarrhoea. 2 trials reported no cases of adverse events.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded one level due to risk of bias. There was high or unclear risk of bias for blinding in most studies for this intervention. Also, a few studies were at high or unclear risk of bias for attrition and reporting bias. 
^bDowngraded one level due to imprecision due to wide confidence intervals that overlapped the line of no effect.
^cDowngraded one level due to inconsistencies between studies.

Background

Description of the condition

It is well documented that people undergoing gastrointestinal surgery are at risk of malnutrition preoperatively (Burden 2012; Shpata 2014). Those who are malnourished have increased rates of mortality and morbidity, and longer lengths of hospital stay, which have been shown to increase healthcare costs (Bozzetti 2007a; Thomas 2016). In older people, a low nutrition risk score preoperatively was independently associated with postoperative complications and poorer activities of daily living scores seven days after surgery (Hanada 2019). Data collected by the American College of Surgeons in the National Surgical Quality Improvement Program found that weight loss greater than 10% in the previous six months was one of 16 prognostic indicators of mortality in a sample of 222,741 people undergoing an operation (Vaid 2012). However, weight loss was one of the few variables considered modifiable (Vaid 2012).

Prevalence data vary in preoperative gastrointestinal surgery. Unintentional weight loss has been reported to affect between 54% and 67% of people prior to gastrointestinal surgery, with malnutrition occurring in 17% to 20% (Burden 2012; Fettes 2002; Garth 2010). Weight loss in people undergoing gastrointestinal surgery has been directly associated with functional impairment, decreased immune defences, delayed wound healing and organ dysfunction (Gillis 2018).

European surveys have reported the risk of malnutrition in people undergoing major abdominal surgery to be 44% using the Nutrition Risk Score 2002 (Sorensen 2008). Using the Malnutrition Universal Screening Tool (MUST), 26% of people on surgical wards were at risk of malnutrition or malnourished (Russell 2008).

The variation in reported data about malnutrition on surgical wards is possibly due to the different techniques used to measure nutritional status or screen for risk of malnutrition. There are many ways to measure malnutrition in inpatients, and this variation often makes comparisons difficult. In acknowledgement of this problem, the Global Leadership Initiative for Malnutrition has devised some standard criteria to be incorporated into nutritional assessments that aim to standardise measurements and the criteria required to identify malnutrition, both in clinical practice and in future research (Cederholm 2019).

The assessment of nutritional status in people during the preoperative period has been recommended in clinical guidelines developed by the European Society for Parenteral and Enteral Nutrition (ESPEN) (Weimann 2017). These guidelines state that perioperative nutrition can fulfil two purposes: initially fulfilling an energy deficit and secondly improving nutritional status in cases where there is an inadequate energy provision (Weimann 2017). However, once the risk of malnutrition has been identified, there clearly needs to be a nutritional intervention instigated to increase the provision of energy and protein to correct nutritional deficits, using a care plan following screening, food fortification, oral supplementation, or enteral or parenteral nutrition (Cooper 2013; Gillis 2018).

There is a lack of consensus relating to strategies to incorporate nutritional interventions into prehabilitation for people prior to elective gastrointestinal surgery, although some have proposed clinical pathways including assessment, immune‐enhancing nutrition, and high energy and protein supplements for those identified as malnourished (Williams 2019). Prehabilitation programmes for people awaiting surgery are starting to incorporate nutritional assessment into their programmes using body mass index, the MUST or the Patient‐Generated Subjective Global Assessment (Gillis 2018; Minnella 2017). Nevertheless, there remains doubt regarding the most appropriate nutritional intervention strategies suitable for routine implementation in clinical practice for preoperative feeding in people undergoing gastrointestinal surgery. However, there is some evidence of benefit from oral nutritional supplements for people losing weight and also some evidence of benefit of increasing protein intake with whey protein supplementation for people who are not taking 1.5 g/kg of protein prior to colorectal surgery (Burden 2017; Knight 2022; Minnella 2017).

The effects of preoperative weight loss and malnutrition are well documented in relation to clinical and economic outcomes (Williams 2019). However, people have reported that weight loss and malnutrition detrimentally affect their overall well‐being after surgery; poor food and nutritional intake affect their energy levels and recovery after surgery, have psychosocial consequences and impact negatively on quality of life (Burden 2016a; Stamataki 2011).

It is noteworthy that the preoperative time point is only the beginning of the overall treatment journey for people undergoing surgery; it can potentially influence the whole perioperative period, other adjuvant treatment modalities and life after treatment has finished. Nutrition has been evaluated throughout the treatment trajectory, including nutritional interventions during pelvic radiotherapy (Lawrie 2018), and after all treatments have been completed (Burden 2019a). Perioperative nutrition has also been evaluated in other tumour groups. Overall, systematic reviews have identified a lack of evidence for perioperative nutritional interventions in ovarian cancer, bladder cancer and oesophageal cancer (Billson 2013; Burden 2019b; Cao 2022). Where evidence was available for nutritional interventions, it was assessed as being predominantly of a low quality.

Description of the intervention

Nutritional therapy interventions include nutritional formulations that are used for medical purposes, administered via the oral, enteral or parenteral route. For this review, nutritional therapy interventions refer to mixed formulas containing macro‐ and micronutrients, with or without immunomodulating components.

How the intervention might work

The presence of malnutrition can contribute to poor clinical outcomes by affecting body structure, function, and physical and psychological health (Stratton 2003; Williams 2019). Malnutrition has been shown to be a significant prognostic indicator for postoperative complications (Sorensen 2008; Sungurtekin 2004; Vaid 2012), which significantly increases length of hospital stay (Leung 2009). Correcting malnutrition preoperatively to decrease postoperative morbidity and mortality may, therefore, be beneficial. When nutritional therapy has been instigated in malnourished people, positive effects on anthropometry, clinical outcomes and cost‐effectiveness have been demonstrated (Beier‐Holgersen 1996; Burden 2017; Knight 2022; Smedley 2004). Benefits from the provision of nutritional interventions have also been demonstrated in well‐nourished cohorts, without any direct effect on nutritional status measurements. This implies that there are physiological benefits to nutritional interventions that are not necessarily associated directly with improvements in anthropometric measurements. Such benefits may include improved immune, respiratory and cardiac function, along with improved wound healing and mobility (Adiamah 2019; Akbarshahi 2008).

Why it is important to do this review

Enhanced recovery after surgery (ERAS) is becoming standard practice in the management of people undergoing gastrointestinal surgery. ERAS includes recommendations on postoperative nutritional management and feeding with preoperative carbohydrate loading (Lassen 2009). However, there is no consensus regarding nutritional intervention in the period preceding hospital admission for people admitted for elective gastrointestinal surgery. This is the first update of an earlier Cochrane review (Burden 2012), and the protocol has been previously published (Burden 2010).

Objectives

To assess the effects of preoperative nutritional therapy compared to usual care in people undergoing gastrointestinal surgery.

Methods

Criteria for considering studies for this review

Types of studies

We included published randomised controlled trials (RCTs) that compared a nutritional therapy intervention versus usual care. We also included conference abstracts of RCTs where we could obtain sufficient data.

Types of participants

We included people undergoing non‐emergency oesophageal, gastric, intestinal or rectal surgery for any condition. We excluded studies where more than 60% of the participants had liver or pancreatic disease, due to the physiological and surgical differences between the gastrointestinal tract and accessory organs of digestion. Furthermore, hepatic surgery is undertaken less frequently than gastrointestinal surgery. Therefore, to ensure a homogeneous sample, we excluded studies if more than 60% of the participants had liver or pancreatic disease.

Types of interventions

The intervention of interest was nutritional therapy by any route, using any nutritional formulation that contained both macro‐ and micronutrients. We included studies if the nutritional formulation had a carbohydrate, fat and nitrogen source with vitamins and minerals, administered from up to three months prior to surgery to 24 hours preoperatively. We included studies if they had manipulated dietary intake to increase energy and protein. We excluded studies that evaluated a single nutrient or immune‐enhancing agent, or any combination of nutritional components that did not meet the inclusion criteria. We also excluded comparisons where there was no control group and nutritional supplementation was compared to an alternative nutritional substrate or feeding regimen.

Types of outcome measures

We collected the primary and secondary outcomes listed below for up to three months postoperatively. If the time point of outcome measurements was not recorded, we assumed that outcomes were assessed until the participant was discharged and so we considered studies where length of hospital stay was less than three months.

Primary outcomes

• Non‐infectious complications: included any complication that was not an infection

• Infectious complications: included surgical site infections (any infection that occurred at the site of incision following surgery) and another infection during time in hospital after surgery

• Length of hospital stay (days)

Secondary outcomes

• Nutritional aspects, including weight, anthropometric measurements, hand grip strength and subjective global assessment

• Quality of life (including participant‐reported outcomes)

• Change in macronutrient (energy and protein/nitrogen) intake

• Biochemical parameters, including albumin, prealbumin and C‐reactive protein

• 30‐day perioperative mortality

• Adverse effects from feeding and route of feeding

Search methods for identification of studies

Electronic searches

We identified RCTs by searching the following databases from the date of the last search in the previous published version of the review (Burden 2012).

• Cochrane Central Register of Controlled Trials (CENTRAL; 2023, Issue 3) in the Cochrane Library

• MEDLINE via Ovid (February 2012 to 28 March 2023)

• Embase via Ovid (February 2012 to 2023 week 13)

• AMED (Allied and Complementary Medicine) via Ovid (February 2012 to 28 March 2023)

• British Nursing Index Archive Ovid (February 2012 to 28 March 2023)

• Evidence based Medicine Reviews (Cochrane Database of Systematic Reviews, American College of Physicians Journal Club, Database of Abstracts of Reviews of Effectiveness, Cochrane Central Register of Controlled Trials, Cochrane Methodology Register, Health Technology Assessments and NHS Economic Evaluation Database) (February 2012 to 28 March 2023)

• World Health Organization International Clinical Trials Registry (WHO ICTRP; trialsearch.who.int) (28 March 2023)

• ClinicalTrials.gov (clinicaltrials.gov; 28 March 2023)

Details of the search strategies are shown in Appendix 1.

Searching other resources

We handsearched the reference lists of the articles selected, and would have contacted authors of conference abstracts if we required further data.

Data collection and analysis

Selection of studies

Two review authors (from JS, AMS, CF, SRK) independently assessed the titles and abstracts to determine relevance and eligibility; they excluded studies that failed to meet the eligibility criteria. If there was insufficient information in the title and abstract, we obtained the full‐text article for clarification. Two review authors (from AMS, JS or CF) independently assessed the full text of all the studies that they assessed to be potentially relevant and extracted data from those studies meeting the inclusion criteria. We listed any studies excluded at the full‐text assessment stage in the Characteristics of excluded studies table. We planned to translate any non‐English articles before assessment, if needed. Any conflicts would have been resolved by discussion with a third review author.

Data extraction and management

We devised a data collection form that facilitated data collection from the articles. This form allowed eligibility to be assessed by linking the studies directly to the research question. Two review authors (AMS, JS) piloted the data collection form on three studies and modified it as required. Two review authors (AMS, CF or JS) independently undertook the process of data extraction and discussed any discrepancies. Any conflicts would have been resolved by discussion with a third review author. They recorded the following information for each trial.

• Year of publication, country of origin, source of funding and number of participants.

• Details of participants, including proportion of malnourished participants (defined by body mass index less than 20 kg/m², weight loss greater than 10% in the previous three to six months, subjective global assessment or nutrition risk derived from a validated tool such as the MUST or the Global Leadership Initiative on Malnutrition (GLIM) (Cederholm 2019)).

• Number of participants, age, type of surgery, perioperative management (ERAS or traditional), gender, diagnosis (noting proportion of cancer and non‐cancer diagnoses).

• Details of type of intervention (nutritional substrate with or without immune‐enhancing agents), route of intervention (oral, enteral or parenteral), length of time on intervention, daily volume of nutritional substrate delivered.

• Details of primary and secondary outcomes.

No transformation of data was preplanned.

Assessment of risk of bias in included studies

Two review authors (AMS, CF or JS) independently assessed each trial's risk of bias in the following areas: random sequence generation, allocation concealment, blinding, complete outcome data, selective outcome reporting and other sources of bias using the Cochrane RoB 1 tool (Higgins 2017). Given the difficulties associated with blinding nutritional interventions, we considered performance bias and detection bias as one domain. We resolved any disagreements between review authors through discussion and would have resolved any conflicts through discussion with a third review author.

Measures of treatment effect

For dichotomous outcomes, we expressed estimates of the effect of an intervention as risk ratios (RR) together with 95% confidence intervals (CIs). We expressed the estimates of effect for continuous variables as mean differences (MD) with 95% CI. Had studies used different scales to measure an outcome, we planned to use standardised mean differences.

Unit of analysis issues

Where there were multiple groups in a study, we only included the groups that met the inclusion criteria. If we had needed to use multiple groups in the same analysis, we would have divided the control group by two to avoid double counting. In this review, we were able to include multiple groups from the same study in separate analyses, so we did not need to divide the control group in this case.

Dealing with missing data

We planned to contact authors for additional data from studies reported as abstracts, if required and where we identified missing data.

Assessment of heterogeneity

We assessed clinical heterogeneity by examining the type of participants, interventions and outcomes in each study. To assess statistical heterogeneity, we used the I² statistic and used the interpretation thresholds stated in the Cochrane Handbook for Systematic Reviews of Interventions as a guide (Higgins 2017). We only conducted meta‐analyses if there were studies reporting similar comparisons for the outcome measures.

Assessment of reporting biases

We planned to use funnel plots to evaluate publication bias had analyses contained 10 or more studies.

Data synthesis

We performed data analysis using Review Manager (RevMan 2022). We used the random‐effects models with inverse variance weighting for meta‐analyses where we anticipated that there would be heterogeneity in the study populations and outcome measurements of the included studies. In meta‐analyses where there were only a few studies included, and we anticipated that the level of heterogeneity in the study population and outcomes would be low, we used a fixed‐effect model with inverse variance weighting for meta‐analysis. We performed a meta‐analysis of the results assuming that included studies were sufficiently similar for the findings to be clinically meaningful. For continuous variables, there were no transformations made for skewed data.

Subgroup analysis and investigation of heterogeneity

We planned to undertake subgroup analyses on studies including participants with cancer and without cancer, elective versus semi‐elective surgery, those that stated the use of an ERAS protocol and route of feeding if data allowed. However, there was insufficient information in the studies regarding these issues, so no subgroup analyses were conducted.

Sensitivity analysis

We planned sensitivity analyses to examine the difference in the quality of the studies and to examine the difference in studies conducted before and after 1990. We used this cut‐off date because there have been advances in artificial feeding since 1990, including changes in technology, line care, feeding tubes and monitoring. The type and amount of enteral and parenteral nutrition delivered has also changed considerably. However, there was insufficient information in the studies regarding these issues, so no sensitivity analyses were conducted. We also planned and performed (where data permitted) sensitivity analyses in participants who were malnourished or weight‐losing prior to surgery.

Summary of findings and assessment of the certainty of the evidence

We presented the overall certainty of evidence for the primary outcomes (non‐infectious complications, infectious complications and length of hospital stay (days)) according to the GRADE approach, which takes into account issues related to internal validity (risk of bias, inconsistency, imprecision, publication bias) and external validity (e.g. directness of results). We created summary of findings tables for each of the identified comparisons based on the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017), and used GRADEpro GDT. We used the GRADE checklist and GRADE Working Group certainty of evidence definitions (Meader 2014). We downgraded the evidence from 'high certainty' by one level for serious (or by two levels for very serious) concerns for each limitation. Using the GRADE approach, we classified each outcome as one of the following levels.

• High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.

• Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.

• Very low certainty: we have very low confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

Results

Description of studies

Results of the search

The study selection process is summarised in the PRISMA flow diagram (Figure 1). The search strategy identified 9547 citations through database searching and 17 records through other sources. After adjusting for duplicates, 8917 studies remained. Of these, 8843 were excluded after title and abstract reviewing as they did not meet the inclusion criteria. We examined the full text of the remaining 74 citations and discarded 56 that did not meet the inclusion criteria, with reasons. We included 16 studies and two studies are ongoing (KCT0003717; NCT04513418). The 16 RCTs reported data for 2164 participants. Nine RCTs were from the previous version of the review (Burden 2012), and seven were new for this update (Barker 2013; Burden 2017; Fujitani 2012; Lee 2023; Campillo 2017; Moriya 2015; Zhao 2018). See Characteristics of included studies and Characteristics of ongoing studies tables. We contacted trial authors for additional information or data (or both), but received no responses; therefore, we only used published data.

1.

Flow diagram

Due to the complexity of the review, we focused on comparing the intervention versus usual or standard care. To do this, we excluded four studies from the original review as they did not have a usual care group (Braga 2002a; McCarter 1998; Okamoto 2009; Xu 2006).

Included studies

All included studies were RCTs that randomised participants to an intervention group receiving a nutritional intervention and a control group following standard diet or usual care. There was only one report for each included study. See Table 5.

1. Characteristics of the trials included on preoperative feeding.

Study and country	Site of surgery	Feed type and volume	Duration and route	Proportion of participants undernourished	Notes
Parenteral nutritional support vs usual care or control
Muller 1982 Germany	Oesophagus, stomach, colon, rectum, pancreas	Amino acids 1.5 g/kg Glucose 11 g/kg	10 days, CVC	C: 62%, I: 59% (WL > 5% in previous 3 months or albumin < 35 d/L)	—
Smith 1988 Australia	Major upper GI, colorectal, multiple operations	Glucose 50–60 kcal/kg and amino acid 150 kcal/1 g nitrogen	10 days, CVA	All (PNI > 30%)	—
Von Meyenfeldt 1992 The Netherlands	Gastric, colorectal	150% basal energy expenditure calculated from Harris and Benedict equation	≥ 10 days, enteral (nasogastric or oral)	All depleted (Nutrition Index)	Study funded by Wander Research and Clintec
Preoperative enteral nutrition (standard formula) vs usual care or control
Gunerhan 2009 Turkey	GI	Not reported	7 days, enteral	All at risk SGA	—
Immune‐enhancing supplementation (oral or enteral) vs usual care or control
Barker 2013 Australia	Upper or lower GI surgery	3 × 237 mL packs per day (Impact Advanced Recovery, Nestlé) Supplement content: carbohydrate, fat, protein, EPA, DPA, arginine	5 days, oral supplement	Well nourished (SGA‐A) I: 35, C: 40, malnourished (SGA‐B/C) I: 11, C: 9	—
Braga 2002b Italy	Colorectal cancer	1000 mL of IE formula with food ad libitum 1000 mL standard formula preoperatively	5 days, oral supplement	I: 12% C: 8% (10% WL in previous 6 months)	Diet provided by Novartis Consumer Health, Bern, Switzerland
Campillo 2017 Spain	Colorectal cancer	3 × 237 mL/day of supplement drink (Impact Oral) Supplement content: carbohydrate, fat, protein, fibre, arginine, omega‐3 fatty acids and RNA	8 days, oral supplement	Malnutrition: I: 27 (64.3%), C: 24 (57.1%) based on percentage of WL, BMI	—
Fujitani 2012 Japan	Gastric cancer	1000 mL/day of Impact plus diet Supplement content: carbohydrate, fat, protein, EPA, DPA, arginine, RNA	5 days, oral supplement	Well nourished: median (range): I: 123 (96.9), C: 116 (99.1), malnourished: median (range): I: 4 (3.1), C: 1 (0.9) [as reported in the publication]	—
Gianotti 2002 Italy	Oesophageal, pancreas colorectal cancer	1000 mL IE formula	5 days, oral supplement	Excluded participants with WL	Diet provided by Novartis Consumer Health, Bern, Switzerland
Lee 2023 South Korea	Colorectal	400 mL immunonutrient‐enriched supplement (Newcare Omega, Daesong life Science)	7 days, oral supplement	At risk of malnutrition I: 7, C: 7 Chonnam National University Hwasun Hospital‐Nutritional Risk Screening Tool	—
Moriya 2015 Japan	Colorectal cancer	Either 250 mL or 750 mL (Impact Nestle) Supplement content: carbohydrate, fat, protein, EPA, DPA, arginine	5 days, oral supplement	BMI at baseline 22 kg/m2
Standard oral nutritional supplements vs usual care or control
Burden 2011 UK	Colorectal cancer	400 mL standard supplement between meals 72% managed 400 mL 16% managed 200 mL	Minimum of 10 days, oral supplement	46% at risk using SGA	Oral supplements provided by Nutricia Ltd
Burden 2017 UK	Colorectal cancer	250 mL daily Fortisip Compact (10.1 kJ, 0.096 g of protein per mL, Nutricia UK) Supplement content: carbohydrate, fat, protein, fibre	5 days, oral supplement	MUST score: 0 = I: 28 (51%) vs C: 13 (28%) 1 = I: 13 (24%) vs C: 16 (35%) 2 = I: 9 (16%) vs C: 14 (30%) 3 = I: 1 (2%) vs C: 1 (2%) 4 = I: 4 (7%) vs C: 0 (0%)	Supplements provided by Nutricia UK
MacFie 2000 UK	Colorectal hepatobiliary	Minimum of 2 supplements a day and normal diet	Variable up until day before surgery, oral supplement	17 participants > 10% WL in previous 6 months	—
Smedley 2004 UK	Lower GI surgery	Ad libitum standard supplement in between meals	Minimum 7 days, oral supplement	34% at risk (determined by BMI and WL)	Study funded by Numico Research, Wageningen, the Netherlands
Zhao 2018 China	Siewert II /III adenocarcinoma oesophagogastric junction.	500 mL of Nutrison Fiber 500 mL (energy 500 kcal, protein 20 g, fat 19.45 g, carbohydrate 61.5 g) plus diet (35 kcal/kg/day) according to dietary guidance	7 days, oral supplement	48 hours prior to surgery prealbumin: mean: I: 265.56 (SD 38.67), C: 258.27 (SD 46.40) total protein: mean: I: 69.5 (SD 3.7), C: 69.2 (SD 4.7) BMI I: 24.2 kg/m2,C: 23.5 kg/m2	—

BMI: body mass index; C: control; CVA: central venous catheter; CVC: central venous catheter; DPA: docosahexaenoic acid; EPA: eicosapentaenoic acid; GI: gastrointestinal; I: intervention; IE: immune‐enhancing; MUST: Malnutrition Universal Screening Tool; PNI: prognostic nutrition index; RNA: ribonucleic acid; SGA: Subjective Global Assessment; WL: weight loss.

Participants

The included studies randomised 2164 participants (791 (38%) women and 1276 (62%) men; gender was not reported for 97 participants). The age of participants ranged from 21 to 79 years. Most studies included people with mixed gastrointestinal cancers. Six studies specifically targeted colorectal cancer (Braga 2002b; Burden 2011; Burden 2017; Campillo 2017; Lee 2023; Moriya 2015); two studies gastric cancer (Fujitani 2012; Von Meyenfeldt 1992); one study oesophageal cancer (Zhao 2018); five studies mixed cancer populations, including oesophagus, stomach, colon, rectum, pancreas and liver (Gianotti 2002; Gunerhan 2009; MacFie 2000; Muller 1982; Smedley 2004); and two studies included people having upper and lower gastrointestinal surgery (Barker 2013; Smith 1988). Four studies were conducted in the UK (Burden 2011; Burden 2017; MacFie 2000; Smedley 2004); two in Australia (Barker 2013; Smith 1988); two in Italy (Braga 2002b; Gianotti 2002); two in Japan (Fujitani 2012; Moriya 2015); one in Germany (Muller 1982); one in the Netherlands (Von Meyenfeldt 1992); one in Spain (Campillo 2017); one in Turkey (Gunerhan 2009); and one in China (Zhao 2018).

Thirteen trials reported the nutritional status of participants. Two trials measured some nutritional status outcomes but did not report on prevalence of malnutrition (MacFie 2000; Zhao 2018). Except for the study by Gianotti 2002, which excluded malnourished people, all the studies included malnourished participants. Nutritional status was reported based on weight loss (Braga 2002b; Burden 2011; Burden 2017; Fujitani 2012; Gianotti 2002; MacFie 2000; Campillo 2017; Smedley 2004); body mass index (Lee 2023; Smedley 2004; Moriya 2015); subjective global assessment (Barker 2013; Burden 2011; Gunerhan 2009); MUST (Burden 2017); the Prognostic Nutritional Index (Smith 1988; Von Meyenfeldt 1992); or level of blood markers including prealbumin and albumin (Muller 1982; Zhao 2018). Between 34% and 100% of participants were malnourished in six trials (Burden 2011; Campillo 2017; Muller 1982; Smedley 2004; Smith 1988; Von Meyenfeldt 1992), and less than 25% of participants were malnourished in five trials (Barker 2013; Braga 2002b; Burden 2017; Fujitani 2012; MacFie 2000).

Interventions

The interventions varied in relation to route of feeding (oral supplementation, enteral and parenteral nutrition), length of supplementation prior to surgery, and frequency and volume of the nutritional intervention. Furthermore, nutritional composition of feeds and nutritional status of participants varied between studies. For each study, these details have been outlined in Table 5 and the Characteristics of included studies table.

The 16 trials allowed 19 comparisons, as three studies had two groups of interventions that met the inclusion criteria (Braga 2002b; Gunerhan 2009; Von Meyenfeldt 1992). Three studies evaluated preoperative parenteral nutrition (Muller 1982; Smith 1988; Von Meyenfeldt 1992), and two evaluated preoperative enteral nutrition (Gunerhan 2009; Von Meyenfeldt 1992). Eight studies evaluated preoperative immune‐enhancing nutrition (Barker 2013; Braga 2002b; Campillo 2017; Fujitani 2012; Gianotti 2002; Gunerhan 2009; Lee 2023; Moriya 2015). Six studies evaluated preoperative oral nutritional supplements (Braga 2002b; Burden 2011; Burden 2017; MacFie 2000; Smedley 2004; Zhao 2018).

All the parenteral nutrition trials planned to administer nutrition for 10 days preoperatively, and volumes administered exceeded current recommendations for macronutrients (Muller 1982; Smith 1988; Von Meyenfeldt 1992). Oral nutrition supplements were administered for a mean of five days in five studies (Barker 2013; Braga 2002b; Fujitani 2012; Gianotti 2002; Moriya 2015), seven days in two studies (Gunerhan 2009; Zhao 2018), eight days in one study (Campillo 2017), and in four studies the intervention was provided at the point from which surgery was a treatment option up to the day prior to the operation (Burden 2011; Burden 2017; MacFie 2000; Smedley 2004). Oral supplements were administered once a day or at multiple times. The volume consumed daily ranged from 250 mL in two studies (Burden 2017; Moriya 2015), 400 mL in two studies (Burden 2011; MacFie 2000), 500 mL in one study (Zhao 2018), 711 mL to 750 mL in three studies (Barker 2013; Campillo 2017; Moriya 2015), and 1000 mL in three studies (Braga 2002b; Fujitani 2012; Gianotti 2002). Participants consumed supplements ad libitum in one study (Smedley 2004), and one study did not report the volume of supplement consumed (Gunerhan 2009).

Eight trials used immune‐enhancing nutritional supplements, with a variety of additional immune‐enhancing components including eicosapentaenoic acid (EPA), docosahexaenoic acid (DPA), arginine and ribonucleic acid (RNA) (Barker 2013; Braga 2002b; Campillo 2017; Fujitani 2012; Gianotti 2002; Gunerhan 2009; Lee 2023; Moriya 2015). The trials of standard oral supplements used supplements from the same manufacturer, and all studies were conducted in the UK (Burden 2011; Burden 2017; MacFie 2000; Smedley 2004). One study did not report the supplement used (Zhao 2018).

Most studies monitored compliance with the supplementation regimen by using a supplement diary, but only four studies reported on the difference in energy consumption between intervention and control groups (Burden 2011; Burden 2017; MacFie 2000; Smedley 2004). Assessment methods included food diaries (Smedley 2004), and 24‐hour dietary recall (Burden 2011; Burden 2017; MacFie 2000).

Outcomes

All trials included postoperative complications as an outcome, although the definition applied to complications varied considerably. In some trials, postoperative complications were defined as infectious and non‐infectious complications (Braga 2002b; Burden 2011; Burden 2017; Fujitani 2012; Gianotti 2002; Lee 2023; Smedley 2004). Other trials outlined definitions used for complications in the publication (Barker 2013; Campillo 2017; MacFie 2000; Muller 1982; Smith 1988; Von Meyenfeldt 1992). The remaining trials did not give definitions for complications (Gunerhan 2009; Zhao 2018). Eight trials recorded biochemistry measures (albumin, transferrin or prealbumin) (Fujitani 2012; Gunerhan 2009; MacFie 2000; Moriya 2015; Muller 1982; Smith 1988; Von Meyenfeldt 1992; Zhao 2018). Other outcomes included are listed in Table 6. All adverse events reported in the trials are described in Table 1; Table 2; Table 3; and Table 4.

2. Outcomes and postoperative management.

Study	Complications definition	Mortality (n)	Anthropometry	Biochemistry	Oral nutritional intake	Postoperative management	Length of hospital stay (days)
Parenteral nutritional support vs usual care or control
Muller 1982	Major Infections Standardised observation forms used	C: 11 I: 3	Weight	Albumin Prealbumin	Not reported	Postoperative infusion regimen identical for groups	Not reported
Smith 1988	Major Minor predefined classification	C: 3 I: 1	Weight TSF	Albumin Transferrin	Not reported	Received postoperative nutritional support if surgeon deemed necessary	I: 44 (SD 13) C: 38 (SD 10)
Von Meyenfeldt 1992	Infectious Non‐infectious Defined in manuscript	PN: 2 EN: 4 C: 2	Not reported	Albumin baseline only	Not reported	PN continued until oral intake postoperatively Other groups increased food and fluid as tolerated	I: 36 (SD 17) C:31.7 (SD 22.1)
Preoperative enteral nutrition (standard formula) vs usual care or control
Gunerhan 2009	Infectious Non‐infectious	Not reported	Reported at baseline	Prealbumin	Not reported	Preoperative IE preoperative std nutrition C: no nutrition	IE: 16.5 (SD 14.8), Std: 14.2 (SD 9.1), C: 12 (SD 3.7)
Immune‐enhancing supplementation (oral or enteral) vs usual care or control
Barker 2013	Diagnostic criteria defined in manuscript for respiratory, wound, urinary infection, abdominal abscess, infective diarrhoea, septicaemia, anastomotic leak, deep venous thrombosis and pulmonary embolus	C: 1 I: 0	Not reported	Not reported	No details on other food consumed, no details on control group	Not reported	I: 7.1 (SD 4.1), C: 8.8 (SD 6.5)
Braga 2002b	Infections, non‐infectious complications, anastomotic leak, antibiotic therapy defined in manuscript	IE: 0 Std: 0 C: 1	Not reported	Not reported	Not reported	4 different groups: IN preoperative only, IE perioperative, standard nutrition perioperative, control	I: IE: 9.5 (SD 2.9), I: std 12 (SD 4.5), C: 12.2 (SD 4.5)
Campillo 2017	Minor (SSI, phlebitis, bacteraemia, urinary tract infection) and older (anastomotic leakage) defined in manuscript	Not reported	Not reported	Not reported	Not reported	Not reported	I: 12.7 (SD 8.3), C: 13.1 (SD 10.8)
Fujitani 2012	SSI definition Horan 1992, abdominal abscess, pancreatic fistula, anastomotic leakage, wound infections, drain infections, pneumonia, venous catheter infection, bleeding, respiratory failure, pleural effusion, heart failure, ileus	C: 0 I: 0	Not reported for full cohort	Not reported	Not reported	Not reported	Median (range) I: 18 (9–85), C: 17 (10–88)
Gianotti 2002	Complications and infections defined by Bozzetti 2001	I: 1 C: 1	Not reported	Albumin, prealbumin	Not reported	3 different groups: IN preoperative only, IN perioperative, C: no nutrition	I: 11.6 (SD 4.7) C: 14.0 (SD 7.7)
Lee 2023	Infectious complications – SSI and other infections, non‐infectious complications, length of hospital stay	Not reported	BMI at baseline Weight before surgery, day of discharge, 1 month after surgery	Albumin at baseline only	Not reported	Oral intake initiated the day after surgery	I: 7.6 (SD 2.5), C: 7.4 (SD 2.3)
Moriya 2015	SSI (Horan 1992), non‐infectious complications	No deaths	Not reported	Albumin, total protein	Not reported	Not reported	I high: 16, I low: 16.5, C: 16
Standard oral nutritional supplements vs usual care or control
Burden 2011	Infectious, non‐infectious Buzby 1988 and Ayliffe 1993, SSI, chest and urinary infection	3 deaths in cohort	Not reported	Not reported	Energy intake (kcal/day): I: 1722 (SD 488), C: 745 (SD 366.0), protein (g): I: 51.8 (SD 33.6), C: 33.0 (SD 16.0)	Sip feeds given up until surgery	Median: I: 13.5, C: 14
Burden 2017	SSI (Horan 1992), chest and urinary infections, complications (Dindo 2004)	I: 1 C: 4	Weight loss 5–7 days postsurgery	Not reported	I: 5302 (3973–7173) kJ, C: 4499 (3218–6416) kJ	Not reported	I: 7.13 (SD 1.9), C: 7.0 (SD 1.7)
MacFie 2000	Total complications, septic complications	I: 1 C: 1	Bodyweight grip strength	Albumin	Energy intake (kcal): I: 1198 (644–2022), C: 1315 (606–2215)	4 groups. We only compared data for preoperative nutrition group vs C	Mean: I: 12, C: 13
Smedley 2004	Minor, major, Buzby 1988	Not reported	Reported at baseline no differences between groups after intervention preoperatively	No difference between groups reported	Energy intake (kcal): I: 2528 (SD 606), C: 1923 (SD 475)	4 groups in trial: sip feeds preoperative compared with C	I: 12.8 (SD 4.5), C: 14.1 (SD 6.6)
Zhao 2018	Not reported	Not reported	BMI at 8 days postintervention	Total protein prealbumin albumin	C: routine diet (35 kcal/kg/day) according to dietary guidance	Both groups accepted fibre within 48 hours postoperative by nasojejunal tube	I: 7 (SD 1) C: 8 (SD 2)

BMI: body mass index; C: control; EN: enteral nutrition; I: intervention; IE: immune‐enhancing; IQR: interquartile range; n: number of participants; PN: parenteral nutrition; SD: standard deviation; SSI: surgical site infection; std: standard; TSF: triceps skinfold.

Excluded studies

We excluded 56 full‐text articles. Details of these are given in the Characteristics of excluded studies table. Reasons for exclusion included study design, postoperative nutritional support being given in conjunction with preoperative nutrition, no usual care group or use of an alternate feeding substrate as the comparator.

Ongoing studies

Two studies were ongoing (KCT0003717; NCT04513418). For further details, see Characteristics of ongoing studies table.

Risk of bias in included studies

We have given the risk of bias summary in Figure 2, and Figure 3 shows the judgements about each risk of bias item presented as percentages across all included studies.

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

Nine studies described sequence generation well, so were at low risk of bias (Barker 2013; Braga 2002b; Burden 2011; Burden 2017; Campillo 2017; Gianotti 2002; Lee 2023; Smedley 2004; Zhao 2018). Sequence generation was unclear in seven studies (Fujitani 2012; Gunerhan 2009; MacFie 2000; Moriya 2015; Muller 1982; Smith 1988; Von Meyenfeldt 1992).

Eight studies described allocation concealment well (Barker 2013; Burden 2011; Burden 2017; Fujitani 2012; Lee 2023; Moriya 2015; Smedley 2004; Smith 1988). Allocation concealment was unclear in eight studies (Braga 2002b; Campillo 2017; Gianotti 2002; Gunerhan 2009; MacFie 2000; Muller 1982; Von Meyenfeldt 1992; Zhao 2018).

Blinding

Given the difficulties associated with blinding nutritional interventions, we considered performance bias and detection bias as one domain.

Risk of bias for blinding was unclear in three studies (Barker 2013; Campillo 2017; Lee 2023). One study blinded hospital staff, but researchers and participants were unblinded (Barker 2013). One study blinded investigators except for the researchers prescribing the supplement; however, the participants were not blinded (Lee 2023). One study may have blinded study observers or researchers (Campillo 2017). Researchers and study participants were not blinded in the remaining 13 studies, so we assessed these at high risk of bias.

Incomplete outcome data

Several trials had a high risk or unclear risk of attrition bias (high: Burden 2011; Gunerhan 2009; Lee 2023; Smedley 2004; unclear: Campillo 2017; MacFie 2000; Muller 1982; Zhao 2018). This was primarily due to participants being recruited who did not have elective surgery or who had surgery at another hospital and were, therefore, not included in the analysis. Postrandomisation exclusions occurred for this reason in five trials (Burden 2011; Lee 2023; MacFie 2000; Muller 1982; Smedley 2004). Four trials excluded participants for other reasons, including gastrointestinal bleeding, emergency surgery to relieve obstruction, uncontrolled blood sugar levels, minimum oral intake of the intervention, no diagnosis of a malignancy, and some participants were excluded if they received postoperative enteral or parenteral nutrition (Campillo 2017; Gunerhan 2009; Muller 1982; Smedley 2004).

Eight studies had a low risk of attrition bias (Barker 2013; Braga 2002b; Burden 2017; Fujitani 2012; Gianotti 2002; Moriya 2015; Smith 1988; Von Meyenfeldt 1992).

Selective reporting

Eight studies had a low risk of reporting bias (Barker 2013; Braga 2002b; Burden 2017; Gianotti 2002; Lee 2023; MacFie 2000; Smedley 2004; Smith 1988). Seven studies had an unclear risk of reporting bias (Burden 2011; Campillo 2017; Fujitani 2012; Gunerhan 2009; Moriya 2015; Muller 1982; Von Meyenfeldt 1992). One study had a high risk of reporting bias due to discrepancies between outcomes reported in the publication and those stated in the trial register (Zhao 2018).

Other potential sources of bias

Seven studies had a low risk of other bias (Barker 2013; Braga 2002b; Burden 2017; Campillo 2017; Lee 2023; Moriya 2015; Zhao 2018), six studies had an unclear risk of other bias (Burden 2011; Fujitani 2012; Gianotti 2002; Gunerhan 2009; MacFie 2000; Smedley 2004), and three studies had a high risk of other bias (Muller 1982; Smith 1988; Von Meyenfeldt 1992).

Three trials predominantly included participants with malignant pathology, and they excluded people who had received preoperative chemotherapy, radiotherapy or immune suppressive treatment (Fujitani 2012; Gianotti 2002; Gunerhan 2009). This will introduce external bias and thus affect the generalisability of the results.

One study incorrectly randomised three participants with more than 10% weight loss into the intervention group, so the authors excluded them from the analysis (Fujitani 2012).

Only three trials specifically enroled malnourished people (Gunerhan 2009; Smith 1988; Von Meyenfeldt 1992), and most participants included in the trials were well nourished. Thus, participants who would be most likely to benefit from nutritional support were not specifically selected for inclusion in the RCTs. Perioperative surgical management has changed over the last decades with the advent of ERAS along with technological advances in the delivery, assessment and formulation of nutritional substrates, all of which may introduce temporal bias into the body of evidence.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

See: Table 1; Table 2; Table 3; and Table 4.

Comparison 1. Parenteral nutritional support versus usual care or control

Three trials with 260 participants compared parenteral nutritional support versus usual care or control (Muller 1982;Smith 1988; Von Meyenfeldt 1992). Primary outcomes are presented in Table 1. One study reported on infections, but the data were not in a format that could be used in the synthesis (Smith 1988). Both Smith 1988 and Muller 1982 reported on biochemistry, but the data were not in a suitable format to be included in the meta‐analyses.

Non‐infectious complications

The evidence is very uncertain about the effect of parenteral nutrition on the number of participants who had a non‐infectious complication after surgery (RR 0.61, 95% CI 0.36 to 1.02; I² = 0%; 3 RCTs, 260 participants; very low‐certainty evidence; Analysis 1.1). We downgraded the certainty of the evidence two levels due to high risk of bias and one level due to indirectness.

1.1. Analysis.

Comparison 1: Preoperative parenteral nutrition compared usual care/control, Outcome 1: Non‐infectious complications

Infectious complications

The evidence is very uncertain about the effect of parenteral nutrition on the number of participants who had an infection after surgery (RR 0.98, 95% CI 0.53 to 1.80; I² = 79%; 3 RCTs, 260 participants; very low‐certainty evidence; Analysis 1.2). We downgraded the certainty of the evidence two levels due to high risk of bias, one level for inconsistency due to the high level of statistical heterogeneity and one level due to indirectness.

1.2. Analysis.

Comparison 1: Preoperative parenteral nutrition compared usual care/control, Outcome 2: Infectious complications

Length of hospital stay (days)

The evidence is very uncertain about the effect of parenteral nutrition on length of hospital stay (MD 5.49 days, 95% CI 0.02 to 10.96; I² = 0%; 2 RCTs, 135 participants; very low‐certainty evidence; Analysis 1.3). We downgraded the certainty of the evidence two levels due to high risk of bias and one level due to indirectness.

1.3. Analysis.

Comparison 1: Preoperative parenteral nutrition compared usual care/control, Outcome 3: Length of hospital stay (days)

Nutritional aspects

One study reported weight gain as an increase of 1.98 kg in the parenteral nutrition group compared to a 1.04 kg loss in the control group (Muller 1982). One study reported a difference in triceps skinfold thickness of 0.6 mm (standard deviation (SD) 1.2; P = 0.05) and a difference in the prognostic nutrition index of 5.5 (SD 10.1; P = 0.05) (Smith 1988).

Quality of life

None of the studies for this comparison reported this outcome.

Change in macronutrient intake

None of the studies for this comparison reported this outcome.

Biochemical parameters

One study reported that total protein was the same in each group and albumin decreased in both groups (Muller 1982). Another study reported transferrin, with similar levels in each group (Smith 1988).

Thirty‐day perioperative mortality

Three studies reported 30‐day perioperative mortality data. Muller 1982 reported 11 deaths in the control group and three deaths in the intervention group. Smith 1988 reported three deaths in the control group and one in the intervention group. Von Meyenfeldt 1992 reported two deaths in each group. Due to the low number of deaths, we did not combine the data for this outcome.

Adverse effects

One study reported complications relating to the central venous catheter occurred in four people (Muller 1982). One study reported two febrile episodes within the cohort (Smith 1988). One study reported six catheter‐related adverse events (Von Meyenfeldt 1992).

Comparison 2. Preoperative enteral nutrition (standard formula) versus usual care or control

Two trials with 120 participants compared preoperative enteral nutrition (standard formula) versus usual care or control (Gunerhan 2009; Von Meyenfeldt 1992). Primary outcomes are presented in Table 2.

Non‐infectious complications

Neither study reported the number of participants with a non‐infectious complication.

Infectious complications

The evidence is very uncertain if enteral nutrition has any effect on the number of participants who had an infection after surgery (RR 0.90, 95% CI 0.59 to 1.38; I² = 11%; 2 RCTs, 126 participants; very low‐certainty evidence; Analysis 2.1). We downgraded the certainty of the evidence two levels due to a high risk of bias, one level due to wide CIs (imprecision) and one level for inconsistency.

2.1. Analysis.

Comparison 2: Preoperative enteral nutrition (standard formula) compared to usual care/control, Outcome 1: Infectious complications

Length of hospital stay (days)

The evidence is very uncertain if enteral nutrition has any effect on length of hospital stay (MD 5.10 days, 95% CI −1.03 to 11.23; I² = 54%; 2 RCTs, 126 participants; very low‐certainty evidence; Analysis 2.2). We downgraded the certainty of the evidence two levels for high risk of bias and one level due to wide CIs (imprecision).

2.2. Analysis.

Comparison 2: Preoperative enteral nutrition (standard formula) compared to usual care/control, Outcome 2: Length of hospital stay (days)

Nutritional aspects

Neither study reported nutritional aspects.

Quality of life

Neither study reported quality of life.

Change in macronutrient intake

Neither study reported change in macronutrient intake.

Biochemical parameters

For prealbumin, Gunerhan 2009 reported no difference between the group that received enteral nutrition compared to the control group (16.41 d/L (SD 7.81) with enteral nutrition versus 18.13 d/L (SD 6.76) with control; P = 0.75).

Thirty‐day perioperative mortality

One study reported four deaths in the intervention group and two in the control group (Von Meyenfeldt 1992).

Adverse effects

One study reported three adverse reactions in the enteral nutrition group (Von Meyenfeldt 1992). One study did not report adverse effects (Gunerhan 2009).

Comparison 3. Immune‐enhancing supplementation (oral or enteral) versus usual care or control

Eight studies with 1007 participants compared immune‐enhancing nutrition supplementation to a control group(Barker 2013; Braga 2002b; Campillo 2017; Fujitani 2012; Gianotti 2002; Gunerhan 2009; Lee 2023; Moriya 2015).Primary outcomes are presented in Table 3.

Non‐infectious complications

Immune‐enhancing nutrition compared to control may result in little to no difference in the number of participants who had a non‐infectious complication after surgery (RR 0.79, 95% CI 0.62 to 1.00; I² = 0%; 8 RCTs, 1020 participants; low‐certainty evidence; Analysis 3.1). We downgraded the certainty of the evidence one level for risk of bias and one level for imprecision.

3.1. Analysis.

Comparison 3: Immune‐enhancing nutrition (oral or enteral) compared to usual care/control, Outcome 1: Non‐infectious complications

Infectious complications

Immune‐enhancing nutrition may result in little to no difference in the number of participants who had an infection after surgery (RR 0.74, 95% CI 0.53 to 1.04; I² = 44%; 7 RCTs, 925 participants; low‐certainty evidence; Analysis 3.2). We downgraded the certainty of the evidence one level for risk of bias and one level for imprecision.

3.2. Analysis.

Comparison 3: Immune‐enhancing nutrition (oral or enteral) compared to usual care/control, Outcome 2: Infectious complications

Length of hospital stay (days)

Immune‐enhancing nutrition may result in little to no difference in length of hospital stay (MD −1.22 days, 95% CI −2.80 to 0.35; I² = 76%; 6 RCTs, 688 participants; low‐certainty evidence; Analysis 3.3). We downgraded the certainty of the evidence one level for risk of bias and one level for inconsistency as there was a high level of statistical heterogeneity.

3.3. Analysis.

Comparison 3: Immune‐enhancing nutrition (oral or enteral) compared to usual care/control, Outcome 3: Length of hospital stay (days)

Nutritional aspects

One study reported nutritional status. Bodyweight decreased in both groups. However, the immune‐enhancing group recovered significantly more weight at 30 days compared to the control group (0.4 (SD 2.1) kg with IE nutrition versus −0.7 (SD 2.3) kg with control) (Lee 2023).

Quality of life

None of the studies reported quality of life.

Change in macronutrient intake

None of the studies reported change in macronutrient intake.

Biochemical parameters

Three studies reported biochemical parameters but each reported different serum proteins and there were insufficient data to undertake meta‐analyses.

One study reported prealbumin levels (Gunerhan 2009). The group given immune‐enhancing nutrition showed an increase compared to baseline (15.98 (SD 8.66) mg/dL; P = 0.037). Moriya 2015 reported albumin (4.1 (SD 0.1) g/dL with immune‐enhancing nutrition versus 4.2 (SD 0.1) g/dL with control) and total protein (6.9 (SD 0.1) g/dL with immune‐enhancing nutrition versus 6.9 (SD 0.1) g/dL with control). Gianotti 2002 reported prealbumin (0.25 (SD 0.03) g/L with immune‐enhancing nutrition versus 0.26 (SD 0.04) g/L with control) and albumin (41.6 (SD 4.9) g/L with immune‐enhancing nutrition versus 42.1 (SD 4.5) g/L with control).

Thirty‐day perioperative mortality

Five studies reported 30‐day perioperative mortality data. Two studies reported one death in the intervention group and one death in the control group (Barker 2013; Gianotti 2002), two studies reported no deaths (Fujitani 2012; Moriya 2015), and one study reported no deaths in the intervention group and one death in the control group (Braga 2002b).

Adverse effects

Two studies reported adverse effects including abdominal discomfort, vomiting and diarrhoea (Braga 2002b; Gianotti 2002). One study reported diarrhoea and increased blood sugars (Barker 2013).

Comparison 4. Standard oral nutritional supplements compared to usual care or control

Six studies compared standard oral nutritional supplements to controls (Braga 2002b; Burden 2011; Burden 2017; MacFie 2000; Smedley 2004; Zhao 2018). Primary outcomes are presented in Table 4. Four studies reported data on hospital stay, dietary intake, nutritional status and quality of life (Braga 2002b; Burden 2011; Burden 2017; Smedley 2004). These data were not in a format that was suitable for inclusion in meta‐analyses.

Non‐infectious complications

Standard oral supplements may result in little to no difference in the number of participants with non‐infectious complications after surgery (RR 0.90, 95% CI 0.67 to 1.20; I² = 0%; 5 RCTs, 473 participants; low‐certainty evidence; Analysis 4.1). We downgraded the certainty of the evidence one level due to risk of bias and one level for imprecision.

4.1. Analysis.

Comparison 4: Standard oral supplement compared to usual care/control, Outcome 1: Non‐infectious complications

Infectious complications

The evidence is very uncertain about the effects of standard oral supplements on the number of participants who had an infection after surgery (RR 0.88, 95% CI 0.60 to 1.27; I² = 50%; 5 RCTs, 473 participants; very low‐certainty evidence; Analysis 4.2). We downgraded the certainty of the evidence one level due to risk of bias, one level for imprecision and one level for inconsistency between studies.

4.2. Analysis.

Comparison 4: Standard oral supplement compared to usual care/control, Outcome 2: Infectious complications

Length of hospital stay (days)

Standard oral supplements may result in little to no difference in the length of hospital stay compared to usual care (MD −0.65 days, 95% CI −2.33 to 1.03; I² = 1%; 3 RCTs, 299 participants; low‐certainty evidence; Analysis 4.3). We downgraded the certainty of the evidence one level due to risk of bias and one level for imprecision. Findings from Zhao 2018 could not be incorporated into the meta‐analysis but they found that preoperative oral nutritional supplements decreased median length of hospital stay compared to participants in the control group (median: 7 (quartile range 6 to 8) days with intervention versus 8 (quartile range 6 to 10) days with control; P = 0.004).

4.3. Analysis.

Comparison 4: Standard oral supplement compared to usual care/control, Outcome 3: Length of hospital stay (days)

Nutritional aspects

None of the studies reported nutritional aspects in a format suitable for inclusion in meta‐analysis. We summarised the data in Table 6.

Quality of life

None of the studies reported quality of life in a format suitable for inclusion in meta‐analysis. We summarised the data in Table 6.

Change in macronutrient intake

None of the studies reported change in macronutrient intake in a format suitable for inclusion in meta‐analysis. We summarised the data in Table 6.

Biochemical parameters

None of the studies reported biochemical parameters in a format suitable for inclusion in meta‐analysis. We summarised the data in Table 6.

Thirty‐day perioperative mortality

Four studies reported 30‐day perioperative mortality data. MacFie 2000 reported one death each in the intervention and control groups. Burden 2011 reported four deaths in the cohort. Burden 2017 reported four deaths in the control and one death in the intervention group. Braga 2002b reported one death in the control group and none in the intervention group.

Adverse effects

Burden 2011 reported four episodes of nausea and vomiting, two episodes of diarrhoea and 20 infections in the intervention group. They reported 27 people with an infection in the control group measured using the Centers for Disease Control and Prevention definitions of infections (Ayliffe 1993). Burden 2017 reported three episodes of abdominal pain, two episodes of diarrhoea and 20 infections in the intervention group and 26 infections in the control group. Braga 2002b reported five participants with abdominal cramping or bloating, four with diarrhoea, 18 with infectious complications and five with non‐infectious complications in the intervention group and seven with abdominal cramping or bloating, three with diarrhoea, 17 with infectious complications and five with non‐infectious complications in the control group.

Subgroup analysis

No subgroup analysis was undertaken as there were insufficient data to conduct them.

Sensitivity analysis

We undertook a sensitivity analysis based on two studies that had data available to evaluate malnourished or weight‐losing participants and found that standard oral supplements probably decreased infectious complications (RR 0.58, 95% CI 0.40 to 0.85; I² = 0%; 2 RCTs, 184 participants; Analysis 5.1). This indicates that standard oral supplements seem to have a greater effect at reducing the number of participants with an infection after surgery in those who are malnourished or weight‐losing.

5.1. Analysis.

Comparison 5: Sensitivity analysis, Outcome 1: Oral standard nutritional supplements compared to usual care or control including malnourished and weight‐losing participants for infections

For non‐infectious complications and length of hospital stay (days) there were limited data, so we did not undertake further sensitivity analyses.

Discussion

Summary of main results

We included 16 studies in this review, which recruited 2164 participants. We compared preoperative nutritional interventions, including parenteral, enteral, immune‐enhancing supplements and standard oral supplements, to usual care. Most participants were men and had gastrointestinal surgery for colorectal cancer. The amount of evidence included varied for the different interventions. We found few studies for parenteral and enteral nutrition, and there was more evidence for immune‐enhancing supplements (eight studies) and standard oral supplements (six studies). We summarise the main findings below.

Parenteral nutrition

We found very low‐certainty evidence for preoperative parenteral nutrition for all the primary outcomes, so we are uncertain if preoperative parenteral nutrition has any effect on non‐infectious complications, infectious complications or length of hospital stay after surgery.

Enteral nutrition

We found very low‐certainty evidence for preoperative parenteral nutrition for infectious complications or length of hospital stay, so we are uncertain if there is any effect on these outcomes. None of the studies reported non‐infectious complications.

Immune‐enhancing nutrition

We identified low‐certainty evidence that immune‐enhancing nutrition probably had little or no effect on non‐infectious complications, infectious complications and length of hospital stay.

Oral nutritional supplements

We found low‐certainty evidence that standard oral nutritional supplements may result in little to no difference on non‐infectious complications. The evidence is very uncertain about the effects of standard oral supplements on the number of participants who had an infection after surgery. The evidence suggests that standard oral supplements may result in little to no effect in length of hospital stay compared to usual care with low‐certainty evidence.

However, in a sensitivity analysis of weight‐losing and malnourished participants, we found that standard oral supplements probably reduced infectious complications.

Overall completeness and applicability of evidence

The evidence we present here is applicable for the current management of people undergoing gastrointestinal surgery with regard to immune‐enhancing nutrition and standard supplements. The evaluation of parenteral nutrition is only of academic relevance and included for completeness due to temporal modifications in indications, assessment, prescribing and monitoring of intravenous nutritional support.

There was a lack of data on some important outcomes within the review, including longer‐term follow‐up. There were few trials on preoperative enteral nutrition and the use of oral nutritional supplements in people who are malnourished preoperatively. More studies are required in the era of ERAS programmes and their clinical implementation of prehabilitation in oncology surgery.

Quality of the evidence

According to GRADE assessments, the certainty of the evidence for enteral and parenteral nutrition was very low, downgraded due to risk of bias, imprecision and indirectness. For immune‐enhancing and standard supplements, the certainty of the evidence was low due to risk of bias, imprecision and inconsistency.

Potential biases in the review process

Limitations of the review were that databases searched were limited to those commonly used in Western countries and excluded Asian and other databases. Therefore, some literature could have been missed.

The inclusion criteria excluded liver and pancreatic cancer surgery, which will have affected the review outcomes, and the inclusion criteria limited the studies by only including mixed nutrient substrates.

We recorded the number of complications and the number of infections. It was not always possible from the data presented within the studies to determine if the results were for events within the group or the number of participants with one or more events. This limitation could potentially influence the effect sizes reported. Where studies reported both the total number of events in the cohort and the number of events per participant, we included the number of participants with one or more events either for complications and infections.

Agreements and disagreements with other studies or reviews

One review incorporating immune‐enhancing nutrition in the preoperative period included studies that recruited participants with hepatobiliary and liver cancers (Adiamah 2019). This review showed a reduction of infectious complications in the immune‐enhancing group compared to the control group. Similar results were also seen in one study on people without malnutrition (Moya 2016). The Adiamah 2019 review differed from our findings as we did not demonstrate a reduction in length of hospital stay with immune‐enhancing nutritional support preoperatively. However, one review that only included studies of people with oesophageal cancer reported no effect on clinical complications postoperatively or hospital length of stay (Li 2020a). In contrast, one review investigating the effect of preoperative oral nutrition prior to gastrointestinal cancer surgery in low‐ and middle‐income countries found a reduction in surgical complications, infectious complications and mortality 30 days postoperatively (Knight 2022). However, there was evidence of methodological bias in how the studies were conducted. One additional review addressed a very similar question to our review with a focus on standard ONS, and found low‐ or very low‐certainty evidence for the use of oral nutritional supplements in relation to influencing weight, and energy and protein intake in the preoperative period and immediate postoperative period (Reece 2020).

The earlier studies included in this review agree with some later research indicating parenteral nutrition may have some benefit for non‐infectious complications but not for infectious complications or length of hospital stay (Bozzetti 2000; VA TPN study group 1991).

Authors' conclusions

Implications for practice.

We were unable to determine if parenteral nutrition, enteral nutrition, immune‐enhancing nutrition or standard oral nutritional supplements have any effect on the clinical outcomes due to very low‐certainty evidence. Preoperative immune‐enhancing nutrition probably results in little to no difference in length of hospital stay. However, studies were not conducted within an enhanced recovery after surgery (ERAS) programme, which is now current clinical practice in most hospitals undertaking gastrointestinal surgery. Standard oral nutritional supplements probably result in little to no difference on complications; however, preoperative standard oral supplements showed some benefit in reducing infection rates after surgery in weight‐losing and malnourished participants. This review only included studies on preoperative nutritional support.

Implications for research.

Immune‐enhancing nutritional supplementation needs to be evaluated in the context of ERAS programmes and in multicentre studies in people undergoing surgery of the lower gastrointestinal tract to improve generalisability.

Multicentre trials are needed on immune‐enhancing nutrition in all gastrointestinal patients. For standard oral nutritional supplements, multicentre trials are required that include participants who are losing weight or are malnourished.

Preoperative enteral and parenteral nutrition need further evaluation in the context of current clinical practice.

Cost‐effectiveness studies need to be undertaken in preoperative nutritional interventions.

It would be preferable to use standard criteria for nutritional assessment and adhere to the recommendations from the Global Leadership Initiative on Malnutrition (GLIM).

Future trials should include an evaluation of nutritional screening and assessment to identify people at risk of malnutrition prior to surgery.

While not within the scope of the current review, it would be interesting to conduct a network meta‐analysis to compare the impact of oral/enteral immune‐enhancing nutrition with a standard oral/enteral supplement. Future work should focus on preoperative standard nutritional supplements compared to usual care in the context of prehabilitation.

What's new

Date	Event	Description
8 April 2024	New citation required and conclusions have changed	New authors added (Anne Marie Sowerbutts, Jana Sremanakova, Chloe French, Stephen R Knight, Ewen M Harrison). Seven new studies added since the original review in 2012 comprising 854 participants. GRADE has been applied to the primary outcomes.
8 April 2024	New search has been performed	New search performed

History

Protocol first published: Issue 12, 2010
Review first published: Issue 11, 2012

Appendices

Appendix 1. Search strategy

MEDLINE

1. exp Preoperative Period/ or exp Preoperative Care/ or preoperative.mp.

2. exp Perioperative Nursing/ or exp Perioperative Care/ or perioperative.mp.

3. 1 or 2

4. exp Food, Formulated/ or exp Nutritional Status/ or exp Dietary Supplements/ or

exp Malnutrition/ or exp Enteral Nutrition/ or sip feeds.mp. or exp Nutritional Requirements/

5. Oral supplements.mp.

6. parenteral nutrition.mp. or exp Parenteral Nutrition/

7. exp Parenteral Nutrition/ or exp Enteral Nutrition/ or enteral.mp.

8. jejunostomy.mp. or exp Jejunostomy/

9. exp Enteral Nutrition/ or nasogastric.mp.

10. gastrostomy.mp. or exp Gastrostomy/

11. 4 or 5 or 6 or 7 or 8 or 9 or 10

12. exp Glutamine/ or exp Food, Formulated/ or exp Arginine/ or exp Fatty Acids,

Omega‐3/ or immunonutrition.mp.

13. Novel substrates.mp.

14. exp Carbohydrates/ or carbohydrate.mp.

15. glucose.mp. or exp Glucose/

16. protein.mp. or exp Proteins/

17. Amino acids.mp. or exp Amino Acids/

18. 12 or 13 or 14 or 15 or 16 or 17

19. exp Esophagectomy/ or gastrointestinal surgery.mp. or exp Gastrectomy/

20. colorectal surgery.mp. or Colorectal Surgery/

21. exp Gastrectomy/ or gastric cancer surgery.mp.

22. exp Esophagectomy/ or oesophageal cancer surgery.mp.

23. pancreatic cancer surgery.mp. or exp Digestive System Surgical Procedures/ or

exp Pancreatectomy/ or exp Pancreaticoduodenectomy/

24. 19 or 20 or 21 or 22 or 23

25. 3 and 11 and 18 and 24

Embase

1 exp Preoperative Period/ or exp Pre‐operative Period/ or exp Preoperative Care/ or exp Pre‐operative Care/ or preoperative.mp. or pre‐operative.mp.

2 exp Perioperative Nursing/ or exp Peri‐operative Nursing/ or exp Perioperative Care/ or exp Peri‐operative Care/ or perioperative.mp. or peri‐operative.mp.

3 1 or 2

4 Exp elemental diet/ or exp Nutritional Status/ or exp Dietary Supplements/ or exp Malnutrition/ or Exp enteric feeding/ or sip feeds.mp. or exp Nutritional Requirements/

5 Oral supplements.mp.

6 parenteral nutrition.mp. or exp Parenteral Nutrition/

7 exp Parenteral Nutrition/ or Exp enteric feeding/ or enteral.mp.

8 jejunostomy.mp. or exp Jejunostomy/

9 exp Exp enteric feeding/ or nasogastric.mp.

10 gastrostomy.mp. or exp Gastrostomy/

11 4 or 5 or 6 or 7 or 8 or 9 or 10 410148

12 exp Glutamine/ or exp Food, Formulated/ or exp Arginine/ or exp Fatty Acids, Omega‐3/ or immunonutrition.mp.

13 Novel substrates.mp.

14 exp Carbohydrates/ or carbohydrate.mp.

15 glucose.mp. or exp Glucose/

16 protein.mp. or exp Proteins/

17 Amino acids.mp. or exp Amino Acids/

18 12 or 13 or 14 or 15 or 16 or 17

19 exp Esophagectomy/ or gastrointestinal surgery.mp. or exp Gastrectomy/

20 colorectal surgery.mp. or Colorectal Surgery/

21 exp Gastrectomy/ or gastric cancer surgery.mp.

22 exp Esophagectomy/ or oesophageal cancer surgery.mp.

23 pancreatic cancer surgery.mp. or exp Digestive System Surgical Procedures/ or exp Pancreatectomy/ or exp Pancreaticoduodenectomy/

24 19 or 20 or 21 or 22 or 23

25 3 and 11 and 18 and 24

Cochrane Central Register of Controlled Trials

#1 MeSH descriptor: [Preoperative Care] explode all trees

#2 MeSH descriptor: [Perioperative Nursing] explode all trees

#3 MeSH descriptor: [Preoperative Period] explode all trees

#4 MeSH descriptor: [Perioperative Care] explode all trees

#5 Pre‐operative Period

#6 Preoperative Care

#7 Pre‐operative Care

#8 preoperative

#9 pre‐operative

#10 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9

#11 Perioperative Nursing

#12 Peri‐operative Nursing

#13 Perioperative Care

#14 Peri‐operative Care

#15 perioperative

#16 peri‐operative

#17 #11 or #12 or #13 or #14 or #15 or #16

#18 #10 or #17

#19 MeSH descriptor: [Dietary Supplements] explode all trees

#20 MeSH descriptor: [Nutritional Status] explode all trees

#21 MeSH descriptor: [Malnutrition] explode all trees

#22 MeSH descriptor: [Enteral Nutrition] explode all trees

#23 MeSH descriptor: [Nutritional Requirements] explode all trees

#24 MeSH descriptor: [Parenteral Nutrition] explode all trees

#25 MeSH descriptor: [Jejunostomy] explode all trees

#26 MeSH descriptor: [Gastrostomy] explode all trees

#27 Oral supplement?

#28 sip feed?

#29 parenteral nutrition

#30 enteral

#31 jejunostomy

#32 nasogastric

#33 gastrostomy

#34 #19 or #20 or #21 or #22 or #23 or #24 or #25 or #26 or #27 or #28 or #29 or #30 or #31 or #32 or #33

#35 MeSH descriptor: [Glutamine] explode all trees

#36 MeSH descriptor: [Arginine] explode all trees

#37 MeSH descriptor: [Fatty Acids, Omega‐3] explode all trees

#38 immunonutrition

#39 Novel substrates

#40 MeSH descriptor: [Carbohydrates] explode all trees

#41 carbohydrate

#42 MeSH descriptor: [Glucose] explode all trees

#43 glucose

#44 MeSH descriptor: [Proteins] explode all trees

#45 protein

#46 Amino acids

#47 MeSH descriptor: [Amino Acids] explode all trees

#48 #35 or #36 or #37 or #38 or #39 or #40 or #41 or #42 or #43 or #44 or #45 or #46 or #47

#49 #18 and #34 and #48

#50 MeSH descriptor: [Esophagectomy] explode all trees

#51 MeSH descriptor: [Gastrectomy] explode all trees

#52 gastrointestinal surgery

#53 colorectal surgery

#54 MeSH descriptor: [Colorectal Surgery] explode all trees

#55 gastric cancer surgery

#56 oesophageal cancer surgery

#57 pancreatic cancer surgery

#58 MeSH descriptor: [Digestive System Surgical Procedures] explode all trees

#59 MeSH descriptor: [Pancreatectomy] explode all trees

#60 MeSH descriptor: [Pancreaticoduodenectomy] explode all trees

#61 #50 or #51 or #52 or #53 or #54 or #55 or #56 or #57 or #58 or #59 or #60

#62 #49 and #61

EBM reviews including ACP Journal Club, Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, Cochrane Methodology Register, Health Technology Assessment, NHS Economic Evaluation Database

Preoperative Period or Preoperative Care or preoperative or Pre‐operative Period or Pre‐operative Care or pre‐operative or Perioperative Nursing or Perioperative Care or peri‐operative or Peri‐operative Nursing or Peri‐operative Care or peri‐operative

AND

Food, Formulated or Nutritional Status or Dietary Supplement* or Malnutrition or Enteral Nutrition or sip feed* or Nutritional Requirement* or Oral supplement* or parenteral nutrition or jejunostomy or nasogastric or gastrostomy

AND

Glutamine or Arginine or Fatty Acids, Omega‐3 or immunonutrition or Novel substrate* or Carbohydrate* or Glucose or protein* or Amino acid*

AND

Esophagectomy or gastrointestinal surgery or Gastrectomy or colorectal surgery or gastric cancer surgery or oesophageal cancer surgery or pancreatic cancer surgery or Digestive System Surgical Procedure* or Pancreatectomy or Pancreaticoduodenectomy

British Nursing Index Archive

Preoperative Period or Preoperative Care or preoperative or Pre‐operative Period or Pre‐operative Care or pre‐operative or Perioperative Nursing or Perioperative Care or peri‐operative or Peri‐operative Nursing or Peri‐operative Care or peri‐operative

AND

Food, Formulated or Nutritional Status or Dietary Supplement* or Malnutrition or Enteral Nutrition or sip feed* or Nutritional Requirement* or Oral supplement* or parenteral nutrition or jejunostomy or nasogastric or gastrostomy

AND

Glutamine or Arginine or Fatty Acids, Omega‐3 or immunonutrition or Novel substrate* or Carbohydrate* or Glucose or protein* or Amino acid*

AND

Esophagectomy or gastrointestinal surgery or Gastrectomy or colorectal surgery or gastric cancer surgery or oesophageal cancer surgery or pancreatic cancer surgery or Digestive System Surgical Procedures or Pancreatectomy or Pancreaticoduodenectomy

AMED (Allied and Complementary Medicine)

1 exp Preoperative Care/ or preoperative.mp. or pre‐operative.mp.

2 (perioperative or peri‐operative).mp.

3 1 or 2

4 exp Nutritional Status/ or exp Dietary Supplements/ or exp Malnutrition/ or exp Enteral Nutrition/ or sip feeds.mp.

5 Oral supplements.mp.

6 parenteral nutrition.mp. or exp Parenteral feeding/

7 exp Parenteral feeding/ or exp Enteral Nutrition/ or enteral.mp.

8 jejunostomy.mp.

9 exp Enteral Nutrition/ or nasogastric.mp.

10 gastrostomy.mp. or exp Gastrostomy/

11 4 or 5 or 6 or 7 or 8 or 9 or 10

12 (glutamine or Arginine).mp. or exp Fatty Acids, Omega‐3/ or immunonutrition.mp.

13 Novel substrates.mp.

14 exp Carbohydrates/ or carbohydrate.mp.

15 glucose.mp. or exp Glucose/

16 protein.mp. or exp Proteins/

17 Amino acids.mp. or exp Amino Acids/

18 12 or 13 or 14 or 15 or 16 or 17

19 (Esophagectomy or gastrointestinal surgery).mp. or exp Gastrectomy/

20 colorectal surgery.mp. or Colorectal Surgery/

21 exp Gastrectomy/ or gastric cancer surgery.mp.

22 exp Esophageal neoplasms/ or oesophageal cancer surgery.mp.

23 (pancreatic cancer surgery or Digestive System Surgical Procedures or Pancreatectomy or Pancreaticoduodenectomy).mp.

24 19 or 20 or 21 or 22 or 23

25 3 and 11 and 18 and 24

ClinicalTrials.Gov

gastrointestinal, nutrition, preoperative

Limiting to intervention

WHO International Clinical Trials Registry

gastrointestinal, nutrition, preoperative

Data and analyses

Comparison 1. Preoperative parenteral nutrition compared usual care/control.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Non‐infectious complications	3	260	Risk Ratio (M‐H, Fixed, 95% CI)	0.61 [0.36, 1.02]
1.2 Infectious complications	3	260	Risk Ratio (M‐H, Random, 95% CI)	0.98 [0.53, 1.80]
1.3 Length of hospital stay (days)	2	135	Mean Difference (IV, Random, 95% CI)	5.49 [0.02, 10.96]

Comparison 2. Preoperative enteral nutrition (standard formula) compared to usual care/control.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Infectious complications	2	126	Risk Ratio (IV, Random, 95% CI)	0.90 [0.59, 1.38]
2.2 Length of hospital stay (days)	2	126	Mean Difference (IV, Random, 95% CI)	5.10 [‐1.03, 11.23]

Comparison 3. Immune‐enhancing nutrition (oral or enteral) compared to usual care/control.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Non‐infectious complications	8	1020	Risk Ratio (M‐H, Random, 95% CI)	0.79 [0.62, 1.00]
3.2 Infectious complications	7	925	Risk Ratio (M‐H, Random, 95% CI)	0.74 [0.53, 1.04]
3.3 Length of hospital stay (days)	6	688	Mean Difference (IV, Random, 95% CI)	‐1.22 [‐2.80, 0.35]

Comparison 4. Standard oral supplement compared to usual care/control.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Non‐infectious complications	5	473	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.67, 1.20]
4.2 Infectious complications	5	473	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.60, 1.27]
4.3 Length of hospital stay (days)	3	299	Mean Difference (IV, Random, 95% CI)	‐0.65 [‐2.33, 1.03]

Comparison 5. Sensitivity analysis.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Oral standard nutritional supplements compared to usual care or control including malnourished and weight‐losing participants for infections	2	184	Risk Ratio (M‐H, Random, 95% CI)	0.58 [0.40, 0.85]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Barker 2013.

Study characteristics
Methods	Design: single‐centre, prospective parallel randomised controlled trial Country: Australia
Participants	Inclusion criteria: people undergoing upper or lower gastrointestinal surgery Exclusion criteria: people who were immunosuppressed or had chronic infections, people with chronic cardiac or respiratory dysfunction, renal failure or hepatic dysfunction (excluding jaundice) and who were unable to provide consent Dates of study: not reported Number randomised Eligible: 103, 8 excluded Randomised: 95 Allocation to control group: 49 Allocation to intervention group: 46 Age (years): mean:intervention: 61.2 (SD 13.3), control: 64.5 (SD 15.3) Gender (male/female): intervention: 17/29, control: 31/18 Condition: well‐nourished and malnourished people undergoing upper or lower gastrointestinal surgery Weight (kg): not reported Body mass index (kg/m2): intervention: 26.8 (SD 6.2), control: 26.3 (SD 4.4) Weight loss (%): not reported Nutritional status Well‐nourished (SGA‐A) intervention: 35, control: 40 Malnourished (SGA‐B/C) intervention: 11, control: 9
Interventions	Comparison Intervention: supplementdrink for the 5 days before their surgery Control: no supplementation, which was in line with current hospital practice Nutritional support 3 × 237 mL packs per day (Impact Advanced Recovery) Supplement content: carbohydrate, fat, protein, eicosapentaenoic acid, docosahexaenoic acid arginine
Outcomes	Length of hospital stay (excluding rehabilitation stay) determined by treating physicians (who were blinded to the treatment allocation) based on participants' clinical progress and standard hospital procedures. Infectious and non‐infectious complications as inpatients. Defined as: Respiratory: abnormal chest radiograph with fever (temperature > 37.5 °C) treated with chest physiotherapy with/without antibiotics Wound infection: any redness, tenderness or purulent discharge of the wound requiring drainage or antibiotic therapy Intra‐abdominal abscess: intra‐abdominal collection treated with antibiotic therapy or drainage Urinary tract infection: > 107 micro‐organisms per 1 mL of urine Infective diarrhoea (Clostridium): positive Clostridium difficile toxin on stool culture Septicaemia: positive blood cultures treated with antibiotic therapy Anastomotic leak: any dehiscence with clinical or radiological evidence Deep venous thrombosis: thrombosis of the pelvic or deep veins of the thigh confirmed on radiological imaging Pulmonary embolus: thrombus in the pulmonary arteries confirmed on radiological imaging Antibiotic use: for > 24 hours Unplanned intensive care unit admissions Mortality Surgical site infection or other complications: 30 days after discharge, self‐report by telephone
Notes	Declaration of interests: nil declared Funding: no funding sought We contacted the author for more data in order to calculate risk ratios and mean differences for outcome variables. The authors provided data to complete the analyses.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation undertaken by researcher.
Allocation concealment (selection bias)	Low risk	Undertaken by a pre‐admissions clinic nurse who had no involvement with the study.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Hospital staff were blinded to allocation, researcher and participants were unblinded to treatment allocation though the outcome measures were considered objective.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 participant in each group was lost to 30‐day follow‐up.
Selective reporting (reporting bias)	Low risk	Reported on all outcomes stated in registration.
Other bias	Low risk	No other bias detected.

Braga 2002b.

Study characteristics
Methods	Design: multigroup parallel randomised controlled study Country: Italy
Participants	Inclusion criteria: people with histologically confirmed colorectal cancer who were candidates for elective curative surgery Exclusion criteria: aged < 18 years, pregnant, respiratory dysfunction, cardiac dysfunction, hepatic dysfunction, renal failure Dates of study: not reported Number randomised Eligible: 233 Randomised: 200 Allocation to group 1: 50 Allocation to group 2: 50 Allocation to group 3: 50 Allocation to group 4: 50 Age (years): mean: group 1:60.5 (SD 11.5), group 2: 63.0 (SD 8.1), group 3: 61.8 (SD 9.9), group 4: 62.2 (SD 10.4) Gender (male/female): group 1: 28/22, group 2: 30/20, group 3: 31/19, group 4: 29/21 Condition: colorectal cancer Weight (kg): not reported Body mass index (kg/m2): not reported Weight loss > 10%: group 1:5, group 2: 6, group 3: 4, group 4: 5 Nutritional status Albumin (g/L): mean: group 1: 41.5 (SD 4.1), group 2: 41.8 (SD 5.0), group 3: 42.2 (SD 4.8), group 4: 40.9 (SD 4.3) Prealbumin (g/L): mean: group 1: 0.23 (SD 0.06), group 2: 0.24 (SD 0.05), group 3: 0.23 (SD 0.10), group 4: 0.22 (SD 0.07)
Interventions	Comparison Group 1: 1 L of IE nutrition 5 days before surgery and continued after surgery with jejunal feeding Group 2: 1 L of IE orally 5 days before surgery Group 3: 1 L of isonitrogenous and isoenergetic specially formulated diet preoperatively Group 4: conventional diet, did not receive any artificial diet before or after surgery Nutritional support IE nutrition with arginine and omega 3 fatty acids (Oral Impact, Novartis. Berne, Switzerland)
Outcomes	Postoperative complications were determined by a surgical team not involved with the study. Participants were followed up for complications 30 days postdischarge. Infectious complications confirmed by microbiological analysis Number of days of antibiotic therapy in participants with postoperative infections Non‐infectious complications Anastomotic leak: clinically evident and radiologically confirmed Postoperative gastrointestinal symptoms such as abdominal cramping, bloating, diarrhoea (defined as > 3 liquid stools per day) and vomiting Length of stay: participants were discharged after meeting the following: occurrence of bowel movement, full recovery of deambulation and oral food intake > 1000 kcal/day. Mortality
Notes	10% of participants had a weight loss > 10% in the previous 6 months. Outcomes were defined (Bozzetti 2001), and were recorded by a member of surgical staff independent of the study. Declaration of interests: nil declared Funding: products were provided by Novartis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by a computer‐generated list.
Allocation concealment (selection bias)	Unclear risk	No mention of concealment of randomisation sequencing.
Blinding (performance bias and detection bias) All outcomes	High risk	No details of blinding reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants included in the study were included in the intention‐to‐treat analysis.
Selective reporting (reporting bias)	Low risk	All outcomes measured were reported in the results.
Other bias	Low risk	No other bias detected.

Burden 2011.

Study characteristics
Methods	Design: unblinded randomised trial using block randomisation Country: UK
Participants	Inclusion criteria: elective curative surgery for colorectal cancer, aged > 18 years and gave informed consent Exclusion criteria: pregnant, enroled in another trial, could not give informed consent, inoperable tumour Dates of study: not reported Number randomised Eligible: 226 Randomised: 125 Allocation to control group: 59 Allocation to intervention group: 66 Age (years): mean: intervention:64.5 (SD 13.9), control: 65.3 (SD 2.7) Gender (male/female): intervention: 34/20, control: 38/24 Condition: colorectal cancer Weight (kg): not reported Body mass index (kg/m2): mean: intervention: 25.0 (SD 4.8), control: 26.8 (SD 4.7) Weight loss (%): mean: intervention: 6.2 (SD 6.8), control: 3.9 (SD 4.8) Nutritional status SGA B and C, indicating moderate‐to‐high risk of malnutrition: 45% of participants
Interventions	Comparison Intervention group: 400 mL of oral supplement (Fortisip, Nutricia Clinical Care, Nutricia Ltd, Wiltshire, UK) for 10 days and dietary advice Control group: dietary advice
Outcomes	Infectious and non‐infectious complications: 2 sets of published definitions were applied to postoperative complications (Ayliffe 1993; Buzby 1988), recorded up to 3 months postsurgery. Antibiotics given postoperatively to treat infections Nutritional intake using 24‐hour unstructured dietary recall at baseline and preoperatively Hospital Anxiety and Depression score at baseline Karnofsky's performance index at baseline Hand grip strength: measured on the non‐dominant hand using a handgrip dynamometer at baseline Length of hospital stay: from date of hospital admission to date of discharge
Notes	Dietary advice was to increase energy and protein: participants were given written information. Declaration of interests: nil declared Funding: National Health Service grant and products were supplied by Nutricia Limited.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation.
Allocation concealment (selection bias)	Low risk	Sequentially numbered brown opaque envelopes.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding undertaken.
Incomplete outcome data (attrition bias) All outcomes	High risk	There were participants who did not go on to have surgery and these were not included in the analysis.
Selective reporting (reporting bias)	Unclear risk	Some outcomes were not reported, including anthropometry and quality of life measures.
Other bias	Unclear risk	Imbalance at baseline: more weight‐losing participants in intervention group. However, this did not affect outcomes on adjusted analysis.

Burden 2017.

Study characteristics
Methods	Design: single‐blind, multicentre randomised controlled trial, used block randomisation Country: UK
Participants	Inclusion criteria: hada primary colorectal tumour, aged > 18 years, listed for radical surgery, had capacity for informed consent and reported unintentional weight loss over previous 3–6 months (> 1 kg) Exclusion criteria: pregnant or had a pacemaker precluding the use of bioelectrical impedance analysis, already on a similar nutritional supplement or had insulin‐dependent diabetes Dates of study: November 2013 to February 2015 Number randomised Eligible: 453, 352 excluded Randomised: 101 Allocation to control group: 46 Allocation to intervention group: 55 Age (years): mean:intervention: 70.5 (SD 11.66), control: 68.9 (SD 11.49) Gender (male/female) (%): intervention: 35/20 (64%/36%), control: 32/14 (70%/30%) Condition: colorectal cancer Weight (kg): not reported Body mass index (kg/m2): mean: intervention: 25.9 (SD 4.8), control: 25.5 (SD 4.54) Weight loss (%): median: intervention: 4.90 (IQR 2.2–8.8), control: 6.8 (IQR 3.4–12.1) Nutritional status MUST score 0: intervention: 28 (51%), control: 13 (28%) 1: intervention: 13 (24%), control: 16 (35%) 2: intervention: 9 (16%), control: 14 (30%) 3: intervention: 1 (2%), control: 1 (2%) 4: intervention: 4 (7%), control: 0 (0%)
Interventions	Comparison Intervention group: oral supplement (Fortisip Compact Nutricia Clinical Care, Wiltshire, UK) a minimum of 5 days prior to surgery with dietary advice Control group: without oral supplement and dietary advice Nutritional support 250 mL/day of supplement drink (2 cartons of 125 mL) Supplement content: carbohydrate, fat, protein, micronutrients
Outcomes	Clinical outcomes were recorded from day 1 up to 30 days postoperatively. Infectious complications: chest or surgical site infection as defined by the US Centre for Disease Control and Prevention definitions (Horan 2008). Postoperative complications: from medical records using standard definitions (Dindo 2004). Length of hospital stay: recorded from the day of surgery until date of discharge Height measured: using a stadiometer to nearest 0.1 cm at baseline Weight measured on calibrated scales to nearest 0.1 kg at baseline, 24–48 hours preoperatively and 5–7 days postoperatively Patient‐Generated Subjective Global Assessment at baseline, 24–48 hours preoperatively and 5–7 days postoperatively Malnutrition Universal Screening Tool score at baseline visit Fat mass and fat‐free mass measured by bioelectrical impedance analysis with participant being adequately hydrated and in a supine position at baseline, 24–48 hours preoperatively and 5–7 days postoperatively.
Notes	Declaration of interests: travel grant received from Nutricia Limited Funding: Fellowship grant from Macmillan and products were supplied by Nutricia Limited.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation used, with random number tables.
Allocation concealment (selection bias)	Low risk	Used opaque sequentially number envelopes.
Blinding (performance bias and detection bias) All outcomes	High risk	Single‐blind only: researchers were blinded, participants were not.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 participant withdrew from the control group, so was not included in the analysis.
Selective reporting (reporting bias)	Low risk	Protocol available: all outcomes were reported on.
Other bias	Low risk	No other bias detected.

Campillo 2017.

Study characteristics
Methods	Design: parallel randomised controlled trial Country: Spain
Participants	Inclusion criteria: aged > 18 years diagnosed with colorectal cancer, proposals for programmed surgery with curative intent (including those with resectable metastatic disease), where a surgical technique with intestinal resection was performed, signed informed consent for performance of surgery and for inclusion in the study. Exclusion criteria: urgent surgery; pregnancy; preoperative evidence of unresectable metastatic disease; American Society of Anaesthesiologists Grade 5; respiratory, cardiac, hepatic or renal dysfunction; active preoperative infection; need for enteral or parenteral nutrition in the immediate preoperative (15 days) Dates of study: not reported Number randomised Eligible: not reported Randomised: 84 Allocation to control group: 42 Allocation to intervention group: 42 Age (years): mean:69.9 (SD 11) Gender (males): 69.05% Condition: colorectal cancer Weight (kg): not reported Body mass index (kg/m2): not reported Weight loss (%): not reported Nutritional status Protein malnutrition: intervention:12 (28.6%), control: 14 (33.3%), based on altered retinol binding protein, transferrin, albumin and prealbumin Caloric malnutrition: intervention: 27 (64.3%), control: 24 (57.1%), based on percentage of weight lost, body mass index and total cholesterol Nutritional Risk Screening (NRS): intervention:27 (64.3%), control: 27 (64.3%)
Interventions	Comparison Intervention: preoperative immunonutrition and normal diet for 8 days prior to surgery Control: who did not receive immunonutrients and followed usual preoperative management. Nutritional support 3 × 237 mL/day of supplement drink (Impact Oral) Supplement content: carbohydrate, fat, protein, fibre, arginine, omega‐3 fatty acids and RNA
Outcomes	Postoperative infectious complications: minor (surgical wound infection, phlebitis, bacteraemia, urinary infection) or major (anastomotic leak) up to 30 days after discharge from hospital Overall hospital stay: not further defined Average costs: calculated according to the diagnosis‐related group assigned by codification department. Participants with any complication – diagnosis‐related group 148: major small and large bowel procedures with complications. Participants with no complications = diagnosis‐related group 149: major small and large bowel procedures with no complications. Cost of the daily hospital stay = diagnosis‐related group × total days of hospital stay + cost of the preoperative oral supplements in the group following the 'immune diet'.
Notes	Declaration of interests: nil declared Funding: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Used random numbers to establish sequencing.
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Not reported.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No CONSORT diagram included.
Selective reporting (reporting bias)	Unclear risk	No protocol found.
Other bias	Low risk	No other bias detected.

Fujitani 2012.

Study characteristics
Methods	Design: prospective randomised trial stratified by institution Country: Japan
Participants	Inclusion criteria: confirmed resectable primary gastric, adenocarcinoma; fit for elective total gastrectomy with adequate bone marrow, hepatic and renal function; performance status 0 or 1 on the Eastern Cooperative Oncology Group scale; age ≤ 80 years; bodyweight loss ≤ 10% within 6 months before entry; tolerance of oral feeding; no other severe medical conditions including insulin‐dependent diabetes mellitus; no concurrent active infection; no known allergy to any of the ingredients of immunonutrition; no preoperative chemotherapy or radiotherapy Exclusion criteria: not reported Dates of study: February 2006 to December 2009 Number randomised Eligible: not reported Randomised: 244 Allocation to control group: 117 Allocation to intervention group: 127 Age (years): median:intervention:64 (range 26–78), control: 65 (range 30–79) Gender (male/female): intervention: 97/30, control: 84/33 Condition: gastric cancer Weight (kg): median: intervention:60.9 (range 38.0–97.0), control: 60.0 (range 40.1–92.2) Body mass index (kg/m2): median: intervention:22.8 (range 15.1–33.8), control: 22.6 (range 17.8–33.1) Weight loss (%): intervention:0 (range 0–16.9), control: 0 (range 0–10.0) Nutritional status Well‐nourished: intervention: 123 (96.9%), control: 116 (99.1%) Malnourished: intervention: 4 (3.1%), control: 1 (0.9%)
Interventions	Comparison Intervention: oral supplementation 5 days prior to surgery in intervention group Control: access to a regular diet without any nutritional supplementation in control group Nutritional support 1000 mL/day of Impact drink added to normal diet Supplement content: carbohydrate, fat, protein, eicosapentaenoic acid, docosahexaenoic acid, arginine, RNA
Outcomes	Surgical and non‐surgical complications from surgery to hospital discharge were documented prospectively. Surgical site infections: superficial incisional, deep incisional, and organ or space SSI, as defined in the Center for Disease Control and Prevention guidelines (Horan 1992). Other complications Abdominal abscess: collection of pus confirmed by percutaneous drainage Pancreatic fistula: drain output of any measurable volume of fluid on or after the third day after surgery, with an amylase content > 3 times the serum amylase level Anastomotic leakage: positive contrast swallow test Wound infection: purulent exudate in the wound with positive bacterial culture Drain infection: purulent exudate around a percutaneous drainage tube Pneumonia: clinical signs of pneumonia with radiographic evidence and positive sputum culture or bronchoalveolar lavage Venous catheter infection: local signs of inflammation or the isolation of pathogenic organisms in culture Bleeding: need for blood transfusion of ≥ 2 units Respiratory failure: presence of dyspnoea and respiratory rate > 35 breaths/minute or arterial partial pressure of oxygen < 70 mmHg, pleural effusion Heart failure: unstable blood pressure requiring use of additional intravenous fluids or cardiac stimulants Ileus Systemic inflammatory response syndrome: clinical manifestation of ≥ 2 of the following features in the first week after operation: temperature exceeding 38 °C or < 36 °C; heart rate > 90 beats/minute; respiratory rate > 20 breaths/minute or arterial partial pressure of carbon dioxide < 32 mmHg; white blood cell count > 12 000/mm3, < 4000/mm3 or > 10% immature (band) forms Prognostic Nutritional Index: 10 × albumin (g/dL) + 0.005 × lymphocyte counts (per mm3) based on albumin levels measured within 2 weeks before trial entry Serum levels of C‐reactive protein: day 3 or 4 after surgery Mortality
Notes	Declaration of interests: nil declared Funding: the immunonutrition (Impact) was purchased by Osaka Gastrointestinal Cancer Chemotherapy Study Group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Used minimisation method with an algorithm that stratified for institutions.
Allocation concealment (selection bias)	Low risk	Randomisation carried out by data centre staff.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding described.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up; those who did not receive a gastrectomy were not included in analysis and 3 participants with > 10% weight loss were excluded from analysis.
Selective reporting (reporting bias)	Unclear risk	Protocol not available.
Other bias	Unclear risk	Eligibility criteria extensive.

Gianotti 2002.

Study characteristics
Methods	Design: multigroup parallel randomised controlled trial Country: Italy
Participants	Inclusion criteria: histologically documented neoplasm of the gastrointestinal tract and planned major elective surgery Exclusion criteria: weight loss 10% (with respect to usual bodyweight) in the past 6 months, aged < 18 years, hepatic dysfunction (Child‐Pugh class B), respiratory dysfunction (arterial partial pressure of oxygen 70 torr), renal dysfunction (serum creatinine level 3 mg/dL, haemodialysis), cardiac dysfunction (New York Heart Class 3), Karnofsky Performance Score 60, pregnancy, ongoing infections, and immune disorder (neoadjuvant radiochemotherapy, circulating neutrophils 2.0 × 109/L) Dates of study: not reported Number randomised Eligible: 517 Randomised: 305 Allocation to control group: 102 Allocation to group 1: 102 Allocation to group 2: 101 Age (years): mean:control:63.4 (SD 11.9), group 1: 62.3 (SD 12.3), group 2: 65.6 (SD 11.5) Gender (male/female): control: 56/46, group 1: 50/52, group 2: 60/41 Condition: cancer of the gastrointestinal tract Weight (kg): mean: control: 68.1 (SD 11.7), group 1: 69.4 (SD 10.1), group 2: 69.0 (SD 13.3) Body mass index (kg/m2): mean:control:23.8 (SD 4.1), group 1: 24.5 (SD 4.9), group 2: 24.2 (SD 4.5) Weight loss (%): control:2.3 (SD 2.7), group 1: 2.4 (SD 2.6), group 2: 2.5 (SD 2.7) Nutritional status Albumin (g/L): mean: control: 40.3 (SD 6.5), group 1: 40.2 (SD 5.6), group 2: 39.9 (SD 5.6) Prealbumin (g/L): mean: control: 0.24 (SD 0.07), group 1: 0.26 (SD 0.08), group 3: 0.25 (SD 0.06)
Interventions	Comparison Group 1: IE nutrition 5 days preoperatively of a supplemented liquid diet Group 2: IE nutrition 5 days preoperatively of a supplemented liquid diet and postoperatively jejunal feeding with the same formula as before, starting within 12 hours after surgery Group 3: no artificial nutrition before or after surgery Nutritional support Immune‐enhancing nutrition with arginine and omega 3 fatty acids (Oral Impact, Novartis. Berne, Switzerland)
Outcomes	Postoperative complications recorded up to 30 days by a member of surgical staff not directly involved in the study Postoperative complications defined by Bozzetti 2001 Infections complications: confirmed by microbiological analysis and positive culture. Non‐infectious complications Wound dehiscence: any dehiscence of the fascia > 3 cm Bleeding: necessary blood transfusion ≥ 2 units Anastomotic leak: any dehiscence with clinical and radiological evidence Respiratory failure: presence of dyspnoea and respiratory rate > 35 breaths/minute or partial pressure of arterial oxygen < 70 mmHg Circulatory insufficiency: unstable blood pressure requiring use of extra fluids or cardiac stimulants (or both) Renal dysfunction: increased serum urea or creatinine concentration 50% above baseline (or both) Renal failure: necessary haemodialysis Hepatic dysfunction: increased serum bilirubin concentration 50% above baseline Pancreatic fistula: daily output of fluid > 10 mL from surgical drainage with amylase concentration 5 times higher than that in serum Delayed gastric emptying: necessity of nasogastric suction for > 8 days after surgery Multiple organ dysfunction syndrome: a state of physiological derangement in which organ function is not capable of maintaining homoeostasis Mortality Length of hospital stay: not further defined
Notes	Declaration of interests: nil declared Funding: products were supplied by Novartis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Individual random numbers.
Allocation concealment (selection bias)	Unclear risk	No details on the process reported.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding. However, surgical staff not involved in the trial applied the definitions for complications.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants enroled were included in the results.
Selective reporting (reporting bias)	Low risk	All outcomes reported.
Other bias	Unclear risk	Extensive exclusion criteria limit generalisability of the results.

Gunerhan 2009.

Study characteristics
Methods	Design: parallel randomised controlled trial Country: Turkey
Participants	Inclusion criteria: people with gastrointestinal tumour admitted to surgical unit Exclusion criteria: diabetes mellitus, renal or hepatic failure (or both), active infection, people with a history of immunosuppressive drug use or clinical signs of vitamin or trace element deficiency Dates of study: not reported Number randomised Eligible: not reported Randomised: 42 Allocation to group 1: 16 Allocation to group 2: 13 Allocation to group 3: 13 Age (years): mean:group 1:64.56 (SD 16.16), group 2: 64.38 (SD 11.67), group 3: 61.31 (SD 12.13) Gender (male/female): group 1: 8/8, group 2: 9/4, group 3: 8/5 Condition: gastrointestinal cancer Weight (kg): not reported Body mass index (kg/m2): mean: group 1:24.11 (SD 3.67), group 2: 23.02 (SD 5.66), group 3: 22.24 (SD 4.87) Weight loss (%): not reported Nutritional status Albumin (mg/dL): mean: group 1: 3.6 (SD 0.59), group 2: 3.46 (SD 0.57), group 3: 3.26 (SD 0.51) Prealbumin (mg/dL): mean: group 1: 13.64 (SD 8.83), group 2: 15.71 (SD 6.97), group 3: 17.72 (SD 8.39) Subjective global assessment (SGA) Moderate malnutrition: group 1: 9, group 2: 7, group 3: 6 Severe malnutrition: group 1: 7, group 2: 6, group 3: 7
Interventions	Comparison Group 1: immune‐enhancing nutrition Group 2: normal nutrition Group 3: standard enteral feed Nutritional support 7 days prior to surgery Immune‐enhancing nutrition: IMPACT Standard enteral feed: Fresubin
Outcomes	C‐reactive protein and prealbumin: days 4 and 7 following randomisation Lymphocyte, CD4+, CD8+, CD16+/56+, CD69+ counts: T‐lymphocyte subpopulations were determined with immunofluorescent stained mouse antihuman monoclonal antibodies – taken prior to nutrition and morning of day 8 following randomisation Non‐infectious complications: not further defined Infection rate: not further defined Duration of hospital stay: not further defined
Notes	Declaration of interests: nil declared Funding: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information reported.
Allocation concealment (selection bias)	Unclear risk	No information reported.
Blinding (performance bias and detection bias) All outcomes	High risk	Not blinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	14 participants excluded from the analysis due to gastrointestinal bleeding, emergency surgery to relieve an obstruction and uncontrolled blood sugars.
Selective reporting (reporting bias)	Unclear risk	All outcomes reported.
Other bias	Unclear risk	Extensive exclusion criteria.

Lee 2023.

Study characteristics
Methods	Design: single‐centre, phase 3, parallel randomised controlled trial Country: South Korea
Participants	Inclusion criteria: primary colon cancer, aged 20–80 years and gave informed consent Exclusion criteria: needed emergency surgery, people with difficulty in oral intake, pregnant women, people planned to have ostomy surgery Dates of study: 2019–2020 Number randomised Eligible: 200, 24 excluded Randomised: 176 Allocation to control group: 88 Allocation to intervention group: 88 Age (years): mean: intervention: 65.3 (SD 9.2), control: 65.3 (SD 11.7) Gender (male/female) (%): intervention:56/23 (71%/29%), control: 50/32 (61%/39%) Condition: colorectal cancer Weight (kg): not reported Body mass index (kg/m2): mean: intervention: 24.4 (SD 3.5), control: 24.1 (SD 4.1) Weight loss (%): not reported Nutritional status Nutritional risk screening (NRS) using Chonnam National University Hwasun Hospital: Nutritional Risk Screening Tool: intervention: 7 (9%), control: 7 (8%) at risk of malnutrition
Interventions	Comparison Intervention: 400 mL/day of preoperative immunonutrient‐enriched oral supplementation (Newcare Omega) for 7 consecutive days prior to surgery Control group: normal diet
Outcomes	Infectious complications: incisional and organ/space surgical site infection and other infections such as respiratory and urinary tract up to 30 days postoperatively Postoperative complication rate up to 30 days postoperatively Change in bodyweight: weight recorded at the time of allocation, the day before surgery, 1–2 weeks after surgery and 1 month after surgery Length of hospital stay: not further defined
Notes	Declaration of interests: nil declared Funding: supported by Chonnam National University Hwasun Hospital, Institute for Biomedical Science (grant number HCRI19014).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed in a 1:1 ratio according to computer‐generated random numbers.
Allocation concealment (selection bias)	Low risk	After enrolment, the investigators contacted a central randomisation centre.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Investigators blinded except for 1 researcher who prescribed ONS after random allocation, participants unblinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	15 participants were lost to follow‐up. For 13 participants, this was due to surgery happening at another hospital, 1 participant refused surgery and 1 for another reason.
Selective reporting (reporting bias)	Low risk	Protocol was available and all the primary and secondary outcomes related to the review were reported in the prespecified way.
Other bias	Low risk	Study appeared free of other source of bias.

MacFie 2000.

Study characteristics
Methods	Design: multigroup parallel randomised controlled trial Country: UK
Participants	Inclusion criteria: people requiring elective major gastrointestinal surgery Exclusion criteria: dementia; major concurrent metabolic problems, such as uncontrolled diabetes, advanced liver disease or uraemia; and those requiring emergency surgery Dates of study: not reported Number randomised Eligible: not reported Randomised: 100 Allocation to group 1: 24 Allocation to group 2: 24 Allocation to group 3: 27 Allocation to group 4: 25 Age (years): mean: group 1: 63 (range 41–86), group 2: 68 (range 23–84), group 3: 66 (range 23–86), group 4: 64 (range 42–85) Gender (male/female): group 1: 11/13, group 2: 15/19, group 3: 8/19, group 4: 12/13 Condition: people undergoing colorectal, gastrointestinal, hepatobiliary surgery Weight (kg): not reported BMI (kg/m2): mean: group 1: 23 (range 17–30), group 2: 23 (range 15–31), group 3: 25 (range 17–34), group 4: 25 (range 21–35) Weight loss (number) < 5%: group 1: 10, group 2: 13, group 3: 19, group 4: 15 5–9%: group 1: 6, group 2: 9, group 3: 6, group 4: 5 ≥ 10%: group 1: 8, group 2: 2, group 3: 2, group 4: 5
Interventions	Comparison Group 1: pre‐ and postoperative supplements Group 2: preoperative supplements Group 3: postoperative supplements Group 4: no supplements Nutritional support Minimum of 2 × 200 mL cartons oral dietary supplement at the point the participant was informed about surgery (Fortisip, Nutricia Ltd, Trowbridge, Wiltshire, UK). Mean duration of taking oral supplements 15 days (range 5–59 days). Supplement content: carbohydrate, fat, protein
Outcomes	Nutritional status was assessed on entry into the study, the day before the operation, the day of discharge from hospital and approximately 4 weeks after discharge. Weight: Marsden electronic chair scales Weight change: recalled as a percentage of pre‐illness weight Height: wall‐mounted measure Body mass index: calculated from weight and height Handgrip strength: Duffield Medical dynamometer. 3 measurements recorded using non‐dominant hand with a minimum 30 seconds of rest between each. Highest value obtained recorded. Mid‐arm circumference (MAC): non‐stretch tape, having determined midpoint between tip of acromion and top of radius, of the non‐dominant arm. Triceps skinfold thickness (TSF): Harpenden skinfold callipers. Mid‐arm muscle circumference (cm) = MAC − TSF × 0.314 Albumin: from venous blood samples Voluntary food intake: after surgery was assessed prospectively for a random sample of participants in each group whilst inpatients. Length of hospital stay Total and septic complications: defined by Copeland 1991 up to 6 months postoperatively Mortality Hospital Anxiety and Depression Scale questionnaire: 4 weeks after discharge
Notes	Declaration of interests: nil declared Funding: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information reported.
Allocation concealment (selection bias)	Unclear risk	No information reported.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding of oral supplements.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	12 people excluded from the analysis as they did not go on to have surgery.
Selective reporting (reporting bias)	Low risk	All outcomes reported.
Other bias	Unclear risk	Some groups were merged after surgery allowing for some confusion in interpretation of results.

Moriya 2015.

Study characteristics
Methods	Design: multigroup parallel randomised controlled trial Country: Japan
Participants	Inclusion criteria: people undergoing colorectal surgery Exclusion criteria: aged < 18 years, active preoperative infections, administration of corticosteroids or immune suppressive agents, gastrointestinal obstruction, respiratory dysfunction, cardiac dysfunction, renal failure, hepatic failure or widespread metastatic disease Dates of study: not reported Number randomised Eligible: not reported Excluded: not reported Randomised: 88, analyses: 85 Allocation to control group: 31 Allocation to intervention high group: 26 Allocation to intervention low group: 31 Age (years): mean:intervention high: 64.7 (SD 2.3), intervention low: 64.8 (SD 2.3); control: 63.8 (SD 2.0) Gender (male/female): intervention high: 15/11, intervention low: 19/11; control: 18/11 Condition: colorectal surgery Weight (kg): mean:intervention high: 57.3 (SD 2.3), intervention low: 56.8 (SD 2.3); control: 57.1 (SD 2.2) Body mass index (kg/m2): mean:intervention high: 22.6 (SD 0.8), intervention low: 22.0 (SD 0.7); control: 22.2 (SD 0.6) Weight loss (%): not reported Nutritional status: not reported
Interventions	Comparison Intervention high or low: supplementdrink for the 5 days before surgery Control group: no supplementation, which was in line with current hospital practice Nutritional support Low 250 mL/day High 750 mL/day (Oral Impact) Supplement content: carbohydrate, fat, protein, eicosapentaenoic acid, docosahexaenoic acid, arginine
Outcomes	Length of hospital stay: number of days from the day of the operation until the date of discharge Total protein, albumin before and after administration of supplement White blood cell count and serum C‐reactive protein levels: preoperatively, and 1, 4 and 7 days after surgery Infections measured up to 30 days postsurgery: surgical site infections as defined by (Horan 1992) and non‐surgical site infection: infection in organ remote from the surgical site, e.g. urinary tract infection, pneumonia Non‐infectious complications Systemic inflammatory response syndrome: ≥ 2 of the following: body temperature > 38 °C or < 36 °C; heart rate > 90 beats/minute; respiratory rate > 20 breaths/minute or partial pressure of arterial carbon dioxide < 32 torr; and white blood cell count > 12,000/mm3 or < 4000 cells/mm3 or > 10% immature band forms at 1, 4 and 7 days after surgery Mortality
Notes	Declaration of interests: nil declared Funding: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Procedure not described.
Allocation concealment (selection bias)	Low risk	Secretary in the Surgical Department opened a sealed opaque envelope.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding of participants or researchers.
Incomplete outcome data (attrition bias) All outcomes	Low risk	88 participants randomised and 3 not included in the analysis.
Selective reporting (reporting bias)	Unclear risk	Protocol not obtained.
Other bias	Low risk	No other bias detected.

Muller 1982.

Study characteristics
Methods	Design: parallel randomised controlled trial Country: Germany
Participants	Inclusion criteria: people with carcinoma of the oesophagus, stomach, colon, rectum or pancreas admitted to the surgical department. Exclusion criteria: total obstruction of the gut Dates of study: December 1978 to July 1980 Number randomised Eligible: 160 Randomised: 125 Allocation to control group: 59 Allocation to intervention group: 66 Age (years): mean: intervention: 58.9 (SD 11.5), control: 59.4 (SD 12.6) Gender (male/female): intervention: 43/23, control: 34/25 Condition: carcinoma of the oesophagus, stomach, colon, rectum or pancreas Weight (kg): not reported Body mass index (kg/m2): not reported Weight loss (%): not reported Nutritional status: not reported
Interventions	Comparison Intervention: parenteral nutrition for 10 days Control: regular hospital diet Nutritional support Parenteral nutrition content: amino acids 1.5 g/kg bodyweight; glucose 11 g/kg bodyweight; electrolytes, trace elements, and vitamins by a central venous catheter. Controls received regular hospital diet of 2400 kcal/day
Outcomes	Infectious: wound infection and pneumonia (not further defined) Non‐infectious complications: intraabdominal abscess, peritonitis, anastomotic leakage, ileus (not further defined) The clinical progress of the participants was assessed daily by the same surgeon on standardised observation sheets. Mortality Serum protein levels: monitored preoperatively and postoperatively Immunological variables: IgA, IgG, IgM, C3A, C4, reactivity against 5 skin‐test antigens (mumps, Candida, purified protein derivative, trichophytin, streptokinase/streptodornase), phytohaemagglutin stimulation of T‐lymphocytes, nitroblue tetrazolium reduction of monocytes, and phagocytosis of erythrocytes induced by complement receptors monitored preoperatively and postoperatively
Notes	Participants were considered malnourished if they had incurred a weight loss of > 5 kg in the previous 3 months prior to admission, serum albumin < 3.5 g/dL and the responses to 5 skin tests were negative. Declaration of interests: nil declared Funding: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details reported.
Allocation concealment (selection bias)	Unclear risk	No details reported.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not all participants included in the trial were in the analysis; 2 participants did not have malignant disease and the remaining participants did not go on to have surgery.
Selective reporting (reporting bias)	Unclear risk	All outcomes were reported.
Other bias	High risk	Lacks clinical applicability as current recommendations outline that parenteral nutrition should only be administered in people who cannot meet their nutritional requirements via oral or enteral route. Amount of nutrition administered was quite high compared to current practices.

Smedley 2004.

Study characteristics
Methods	Design: multigroup parallel randomised controlled trial Country: UK
Participants	Inclusion criteria: people undergoing lower gastrointestinal tract surgery Exclusion criteria: aged < 18 years, pregnancy, overt dementia, emergency or laparoscopic surgery, receipt of other forms of preoperative nutritional support and inability to take ONS for a minimum of 7 days before operation Dates of study: October 1998 to March 2001 Number randomised Eligible: 532 Randomised: 179 Allocation to group 1: 42 Allocation to group 2: 48 Allocation to group 3: 39 Allocation to control group: 50 Age (years): mean:group 1:55 (range 26–81), group 2: 61 (range 23–84), group 3: 62 (range 22–83), control: 63 (range 25–88) Gender (male/female): group 1: 19/23, group 2: 33/15, group 3: 20/19, control: 28/22 Condition: people undergoing gastrointestinal tract surgery Weight (kg): not reported Body mass index (kg/m2): group 1:24.9 (SD 4.5), group 2: 26.9 (SD 4.9), group 3: 25.5 (SD 4.5), control: 27.8 (SD 5.6) Weight loss (%): not reported Nutritional status: not reported
Interventions	Comparison Group 1: supplements were given preoperatively for a minimum of 7 days Group 2: supplements were given pre‐ and postoperatively up to 4 weeks after discharge from hospital Group 3: supplements were given postoperatively up to 4 weeks after discharge from hospital Control: no artificial nutrition administered Nutritional support Encouraged to drink supplements ad libitum Supplement was Fortisip (Nutricia, Wageningen, the Netherlands).
Outcomes	Outcomes measured up to 4 weeks postoperatively Change in bodyweight Major and minor complications using definitions by Buzby 1988 Anthropometric measurements: bodyweight, body mass index, mid‐arm circumference, triceps skinfold thickness, mid‐arm muscle circumference, hand‐grip dynamometry using standard techniques at recruitment, on admission to hospital, on the day of resumption of free fluids or light diet, at hospital discharge, and 2 and 4 weeks after discharge by a single dietitian at each centre Nutritional intake: 4‐day food diaries for outpatients and food record charts for inpatients. The diaries were completed before admission and at 2 and 4 weeks after discharge. The inpatient charts were kept from the day of resumption of free fluids to day 7 or discharge, whichever occurred earlier. Quality of life: 36‐item Short Form, EuroQol instruments, a well‐being index for people undergoing surgery and a 10‐point visual analogue scale for fatigue completed by participants were possible on recruitment, admission, at the time of discharge and 4 weeks later. Length of hospital stay: not further defined Health service costs: all resource elements in both primary and secondary care from beginning of hospital stay to 28 days postdischarge.
Notes	Declaration of interests: nil declared Funding: Numico Research
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sealed envelopes stratified according to nutritional status.
Allocation concealment (selection bias)	Low risk	Used sealed envelopes.
Blinding (performance bias and detection bias) All outcomes	High risk	Unblinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	27 people were withdrawn from the study at the participant's request, surgery cancelled, or enteral or parenteral nutrition was started.
Selective reporting (reporting bias)	Low risk	Stated no difference in quality of life, but data not reported.
Other bias	Unclear risk	Dose of supplement was not specified as intervention given ad libitum.

Smith 1988.

Study characteristics
Methods	Design: parallel randomised controlled trial Country: Australia
Participants	Inclusion criteria: people referred for major upper gastrointestinal surgery and people with significant weight loss who were referred for colorectal surgery, having a prognostic nutritional index score > 30% Dates of study: not reported Number randomised Eligible: not reported Randomised: 34 Allocation to control group: 17 Allocation to intervention group: 17 Age (years): mean: intervention: 67 (SD 4), control: 68 (SD 3) Gender (male/female): intervention: 12/5, control: 15/5 Condition: people undergoing major gastrointestinal surgery including upper gastrointestinal surgery and colorectal surgery Weight (kg): intervention: 57 (SD 3), control: 65 (SD 3) Body mass index (kg/m2): not reported Weight loss (%): not reported Nutritional status: not reported
Interventions	Comparison Intervention: intravenous nutrition prior to surgery Control: standard care Nutritional support 10 days of preoperative intravenous nutrition Intravenous nutrition content: infusing 50–60 kcal/kg/day of glucose/amino acid intravenous nutrition mixture containing 150 kcal/1 g of nitrogen. Replacement of electrolytes, trace elements, vitamins, and essential fatty acids was also given.
Outcomes	Infectious and non‐infectious complications categorised as minor and major Respiratory infection: minor if treated with physiotherapy or oral antibiotics and major if oxygen therapy or ventilatory assistance required Other septic: major if resulted in hypotension, use of intravenous antibiotic or prolonged hospital stay Thromboembolic disease: major if resulted in clinical pulmonary emboli or gross swelling of the leg Mortality Hospital length of stay: not further defined Triceps skinfold thickness taken using Holtaine skinfold callipers, weight change, albumin measured once in the control group and before and after intravenous nutrition in treatment group
Notes	Declaration of interests: nil declared Funding: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomly ordered cards in sealed envelopes opened after the prognostic nutrition index was obtained. Did not describe an audit trail for sealed envelopes.
Allocation concealment (selection bias)	Low risk	Sealed envelopes.
Blinding (performance bias and detection bias) All outcomes	High risk	Not blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants included in the results.
Selective reporting (reporting bias)	Low risk	All outcomes reported.
Other bias	High risk	Lacked clinical applicability as current recommendations outline that parenteral nutrition should only be administered in people who cannot meet their nutritional requirements via oral or enteral routes. Amount of nutrition administered was high compared to current practices.

Von Meyenfeldt 1992.

Study characteristics
Methods	Design: multigroup parallel randomised controlled trial Country: the Netherlands
Participants	Inclusion criteria: people with newly detected, histologically confirmed gastric or colorectal carcinoma requiring surgical treatment, who had not undergone treatment for other malignant tumours. Exclusion criteria: aged > 80 years and people with a normal nutritional status Dates of study: not reported Number randomised Eligible: not reported Randomised: 200 Allocation to group 1: 51 Allocation to group 2: 50 Allocation to group 3: 50 Allocation to group 4: 49 Age (years): mean: group 1: 67.3 (SD 10.2), group 2: 65.7 (SD 9.3), group 3: 65.8 (SD 7.5), group 4: 61.3 (SD 9.5) Gender (male/female): group 1: 29/22, group 2: 32/18, group 3: 32/18, group 4: 33/16 Condition: gastric or colorectal carcinoma Weight (kg): not reported Body mass index (kg/m2): not reported Weight loss (%): not reported Nutritional status Albumin (g/L): mean: group 1: 33.5 (SD 3.8), group 2: 35.5 (SD 4.0), group 3: 33.7 (SD 4.3), group 4: 39.1 (SD 3.4)
Interventions	Comparison Group 1: preoperative parenteral nutrition for 10 days Group 2: preoperative enteral nutrition via nasogastric tube or by mouth for 10 days Group 3: no nutrition Group 4: non‐depleted group not randomised Nutritional support Enteral nutrition used was Precitene or Isotein.
Outcomes	Total complications monitored daily as inpatients Minor Wound infection: clinical aspect, positive culture or necessity for surgical drainage Urinary tract infection: positive culture > 107 micro‐organisms per mL of urine Pulmonary tract infection: positive sputum culture and abnormal chest X‐ray requiring treatment with antibiotics Major Respiratory insufficiency: ventilatory assistance for > 12 hours postoperative or renewed mechanical ventilation after detubation Cardiac insufficiency: clinical and radiological diagnosis requiring cardiotonics or diuretics (or both) or myocardial infarction confirmed on electrocardiography and by serum enzyme elevations Renal insufficiency: plasma urea levels elevated by 50% or a plasma creatinine elevation with 20% Intra‐abdominal abscess: purulent fluid collection, demonstrated by ultrasound or computer tomography scan‐guided drainage or at reoperation Fistula: diagnosis on clinical grounds and confirmed by fistulography Anastomotic leakage: after clinical or radiological diagnosis, necessitating reoperation or drainage Wound dehiscence: dehiscence of the wound requiring surgical closure Sepsis: positive blood culture persisting temperature > 39 °C associated with hypotension and hypoperfusion Disseminated intravascular coagulation syndrome: positive fibrinogen‐degradation product test, abnormal clotting tests and decreasing thrombocyte counts Mortality: any postoperative inhospital mortality Length of hospital stay: not further defined
Notes	Complications defined in the manuscript Declaration of interests: nil declared Funding: Wander Research
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information reported.
Allocation concealment (selection bias)	Unclear risk	No information reported.
Blinding (performance bias and detection bias) All outcomes	High risk	Not blinded; however, would have been difficult to blind the trial.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants randomised were included.
Selective reporting (reporting bias)	Unclear risk	Protocol not available.
Other bias	High risk	Lacked clinical applicability as current recommendations outline that parenteral nutrition should only be administered in people who cannot meet their nutritional requirements via oral or enteral route. Amount of nutrition administered was quite high compared to current practices.

Zhao 2018.

Study characteristics
Methods	Design: parallel randomised controlled trial Country: China
Participants	Inclusion criteria: people diagnosed with gastroscopy, pathological examination, and computer tomography as Siewert II or III adenocarcinoma oesophagogastric junction with the diameter of tumour 8 cm; confirmed at progressive stage (T3/4, N+, M0) but without remote metastasis, no participants accepted antitumour therapy before; no contraindications to chemotherapy and surgical operation; score of Karnofsky Performance Score > 60, score of Eastern Cooperative Oncology Group 0–2; aged 18–75 years, with informed consent; participants did not have any other systemic diseases Exclusion criteria: not reported Dates of study: not reported Number randomised Eligible: not reported Randomised: 66 Allocation to control group: 33 Allocation to intervention group: 33 Age (years): mean 62 (range 26–74) Gender (male/female): 57/9 Condition: Siewert II and III adenocarcinoma of oesophagogastric junction. All participants accepted chemotherapy for 2 cycles (35 days) before surgery Weight (kg): not reported Body mass index (kg/m2): mean: intervention:24.24 (SD 3.20), control: 23.53 (SD 3.03) (48 hours from admission to hospital) Weight loss (%): not reported Nutritional status: prealbumin (mg/L): mean: intervention:265.56 (SD 38.67), control: 258.27 (SD 46.40)
Interventions	Comparison: intervention group receiving enteral feeding (Nutrition Fibre and oral nutrition therapy) versuscontrol group who followed routine preoperative diet (35 kcal/kg/day) according to dietary guidance Nutritional support 500 mL oral nutrition supplementation 7 days prior to surgery. Both groups accepted Nutrison Fibre within 48 hours after surgery by nasojejunal tube inserted during the surgery Supplement content: protein, fat, carbohydrate, fibre
Outcomes	All outcomes measured within 48 hours of first hospitalisation, days 1 and 8 after surgery Body mass index Serum prealbumin, total protein, albumin, haemoglobin, total lymphocyte count Nutrition Risk Score 2002 and Patient‐Generated Subjective Global
Notes	Declaration of interests: nil declared Funding: LI Jieshou's Intestinal Barrier Association
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Used random numbers for sequence generation.
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding (performance bias and detection bias) All outcomes	High risk	No blinding.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Not clearly reported as no CONSORT diagram.
Selective reporting (reporting bias)	High risk	Discrepancies between outcomes measures in publication and the trial registry: NCT01962246.
Other bias	Low risk	No other bias detected.

IE: immune‐enhancing; Ig: immunoglobulin; IQR: interquartile range; MUST: Malnutrition Universal Screening Tool; ONS: oral nutrition supplements; RNA: ribonucleic acid; SGA: Subjective Global Assessment (categories A, B or C); SD: standard deviation.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abe 2019	Retrospective observational study
Ali 2019	Perioperative nutrition
Bozzetti 2001	Postoperative nutrition
Braga 1999	Perioperative nutrition
Braga 2002a	Not a usual care group for control
Burden 2016b	Abstract to included study
Caglayan 2012	Not an RCT
Ding 2015	Single nutrient supplementation
Finco 2007	Perioperative nutrition
Gade 2016	Ineligible population
Gianotti 1999	Perioperative nutrition
Giger 2007	Postoperative nutrition
Giger‐Pabst 2013	Ineligible comparator
Gillis 2016	Single nutrient supplementation
Healy 2017	Ineligible comparator
Heatley 1979	Not an RCT
Hendry 2008	Not an RCT
Hiki 2015	Perioperative nutrition
Horie 2006	Not an RCT
Hossain 2016	Perioperative nutrition
Huang 2014	Not an RCT
Hubner 2012	Ineligible comparator
Imamura 2016	Postoperative nutrition
Jie 2012	Not an RCT
Kabata 2015	Perioperative nutrition
Kanekiyo 2019	Ineligible comparator
Khomyakov 2019	Not an RCT
Klek 2011	Perioperative nutrition
Lee 2020	Protocol to included study
Li 2020b	Ineligible comparator
Lidder 2013	Single nutrient supplementation
Lim 1981	Perioperative nutrition
Lin 1997	Ineligible comparator
Liu 2015	Not an RCT
Lluch Taltavull 2014	Retrospective observational study
Ma 2017	Perioperative nutrition
McCarter 1998	Not a usual care group for control
Mudge 2018	Perioperative nutrition
Mueller 1982	Ineligible comparator
Okamoto 2009	Not a usual care group for control
Ozkan 2002	Perioperative nutrition
Rombeau 1982	Not an RCT
Ruiz‐Tovar 2016	Perioperative nutrition
Ryan 2009	Perioperative nutrition
Sakurai 2007	Perioperative nutrition
Sangster 2019	Perioperative nutrition
Senkal 1999	Perioperative nutrition
Senkal 2005	Incorrect comparison
Sodergren 2010	Postoperative nutrition
Sultan 2012	Ineligible comparator
Takeuchi 2007	Perioperative nutrition
Wierdak 2021	Ineligible comparator
Xu 2006	Not a usual care group for control
Yildiz 2016	Ineligible comparator
Yu 2021	Ineligible comparator
Zareba 2013	Ineligible aim

RCT: randomised controlled trial

Characteristics of ongoing studies [ordered by study ID]

KCT0003717.

Study name	Multicenter randomized controlled trial comparing intensive nutritional support versus conventional nutritional support for gastric cancer patients who underwent radical gastrectomy
Methods	Multicentre single‐blind RCT
Participants	Inclusion criteria Aged 20–80 years with stomach cancer diagnosed by biopsy Preoperative body mass index > 23.5 kg/m2 ECOG performance 0 or 1 American Society of Anesthesiologist score class I–III On preoperative computed tomography examinations, participants who were found to be able to undergo curative gastrectomy without any other organs except spleen Scheduled for total gastrectomy or subtotal gastrectomy Living in areas where home nursing is available Signed written consent form to participate in this study according to their free will Exclusion criteria Gastric outlet obstruction Scheduled for pylorus‐preserving gastrectomy or proximal gastrectomy Scheduled for resection of other organs (ovaries, adrenals, small intestine, etc.) Siewert type 1 oesophagogastric junctional cancer Other organ cancer history within 5 years before surgery, or with present cancers of other organ Treated with systemic inflammatory disease or who are currently being treated History of gastrectomy
Interventions	Control group No preoperative nutritional therapies are performed. If there is no postoperative complication, water intake begins 1–4 days after surgery, 2–5 days for semi‐fluid diet, and 3–6 days for soft‐blended diet. If participants do not have any problems after 1–2 days of soft‐blended diet, participant will be discharged. When taking hydration, the parenteral nutritional support (WINUF Peri) are administered. Administration will be stopped after the start of semi‐fluid diet. At discharge, prescribe oral nutritional supplements for up to 10 days after surgery. Intervention group Allow oral nutritional supplements to be administered 3–14 days before surgery. Intake of an oral carbohydrate drink (Fastop) is recommended until 2 hours before surgery. Progression of the diet is the same as for the conventional nutritional support group. Only allow ingestion of oral carbohydrate drinks if participant starts oral hydration. The initiation of parenteral nutritional support is the same as for the conventional nutritional support group. However, the parenteral nutritional support is administered with soft‐blended diet until the oral intake calorie exceeds 70% of the daily requirement. Insert a central catheter at discharge. After discharge, the home nurse will administer parenteral nutritional supplement 4 times a week via peripherally inserted central catheter. Nutritional counselling is performed at each outpatient clinic and according to the result, the parenteral nutritional supplement are administered 2–7 times per week. Oral nutritional supplements should be taken for 30 days after surgery.
Outcomes	Primary outcome Percentage of weight loss 6 months after surgery Secondary outcome Percentage of weight loss 1 and 3 months after surgery Change in body mass index 1 and 6 months after surgery Percentage of underweight participants (body mass index < 18.5 kg/m2) 1 and 6 months after surgery Changes in quality of life 1 and 6 months after surgery Compliance with postoperative chemotherapy in participants with advanced gastric cancer 1 and 6 months after surgery Complication rate, mortality rate 6 months after surgery Compliance with postoperative chemotherapy in participants with advanced gastric cancer after end of chemotherapy Complication rate, mortality rate 1 and 3 months after surgery
Starting date	Registered on World Health Organization International Clinical Trials Registry Platform in 2019 as not yet recruiting (KCT0003717).
Contact information	Youngsuk Park youngsukmd@gmail.com
Notes

NCT04513418.

Study name	Effects of preoperative enteral immunonutrition for esophageal cancer patients given neoadjuvant chemoradiotherapy (POINT)
Methods	Randomised controlled trial evaluating the effects of preoperative enteral immunonutrition in people with oesophageal cancer undergoing neoadjuvant therapy.
Participants	Inclusion criteria Histologically confirmed oesophageal cancer Staging as cT2‐T3, N0‐3, M0 with the need of neoadjuvant therapy before radical oesophagectomy Tolerance for oral intake (at least fluid diet) ECOG performance status ≤ 2 Body mass index ≥ 18.5 kg/m2 before recruitment Participants approve and sign the informed consent Exclusion criteria Expected survival time < 6 months Complete dysphagia Pregnant or breastfeeding women Unable to obey the interventions for any reason Serious comorbidities (cardiac, pulmonary, liver, kidney, brain, haematological, endocrine and other diseases) in people who cannot tolerate neoadjuvant therapy or surgery
Interventions	Intervention group: participants receive omega‐3 fatty‐acid enriched enteral nutritional emulsion from the start of neoadjuvant chemoradiotherapy until surgery. Participants are meanwhile encouraged to intake 25–30 kcal/kg through regular food. Preoperative immunonutrition includes an omega‐3 fatty‐acid‐enriched enteral nutritional emulsion given by oral intake, nasogastric feeding tube or jejunostomy 600 mL/day, lasting from the start of neoadjuvant chemoradiotherapy until surgery. Meanwhile, oral intake is encouraged to reach 25–30 kcal/kg through regular food. Control group: participants are encouraged to intake 25–30 kcal/kg through regular food without supplemental nutritional support before oesophagectomy.
Outcomes	Primary outcomes Rate of postoperative nutrition and immune‐related complications (up to 30 days after surgery) Rate of gastrointestinal complications (anastomotic leakage, gastrointestinal dysfunction), metabolic complications (electrolyte disturbances, liver or renal dysfunction) and infectious complications (wound infection, catheter‐related infection, pneumonia, sepsis, or other infections requiring antibiotics) Secondary outcomes Completion rate of neoadjuvant therapy and oesophagectomy (2–3 months before surgery) Adverse events during neoadjuvant chemoradiotherapy (2–3 months before surgery) Pathological complete response rate (2–3 months before surgery) Blood loss (at operation day) Duration of surgery (at operation day) Rate of surgery‐related complications (at operation day): conversion to open surgery, recurrent nerve injury, cardiac and cerebrovascular accident Length of hospital stay (time point: postoperative in‐hospital stay) Hospitalisation costs (time point: postoperative in‐hospital stay) 30‐day and 90‐day mortality (up to 90 days after surgery) Weight loss (from neoadjuvant chemoradiotherapy to 6 months after surgery) Perioperative change in Patient‐Generated Subjective Global Assessment score (from neoadjuvant chemoradiotherapy to 6 months after surgery) Perioperative change in blood pressure (from neoadjuvant chemoradiotherapy to 6 months after surgery) Perioperative change in biochemical indicators (from neoadjuvant chemoradiotherapy to 6 months after surgery): white blood cells; albumin; C‐ reactive protein; tumour necrosis factor‐alpha; interleukin 6; immunoglobulins A, G, M; insulin; C‐peptide; glucose; haemoglobin; vitamin A; vitamin D; serum ferritin; transferrin receptor Quality of life: European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire and European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire ‐ Oesophageal Cancer Module (1, 3 and 5 years after surgery) Long‐term survival (1, 3 and 5 years after surgery): overall survival and progression‐free survival Daily dietary intake (through study completion, a mean of 6 months): 24‐hour meal review every week
Starting date	August 2020
Contact information	Hecheng Li, PhD, MD, tel: 00862164370045 ext 664566, Shanghai Ninth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine Shanghai, Shanghai, China
Notes

ECOG: Eastern Cooperative Oncology Group; NYHA: New York Heart Association; PG‐SGA: Patient‐Generated Subjective Global Assessment.

Differences between protocol and review

We made the following changes from the protocol and first version of the review (Burden 2010; Burden 2012)

• Previously we did not specify the exclusion of specific populations of participants and diseases within the gastrointestinal tract. However, in this update, we found studies that included a mixture of people undergoing gastrointestinal surgery, so needed to decide on clearer inclusion and exclusion criteria. Due to the physiological and surgical differences between the gastrointestinal tract and accessory organs of digestion, such as the liver and pancreas, we decided to exclude studies if more than 60% of participants were having surgery for pancreatic or liver diseases. In the protocol, we also stated that we would perform sensitivity analyses on quality of studies, the difference in studies before and after 1990, and studies before and after the ERAS (enhanced recovery after surgery) programme. Due to the limited data extracted from the included studies, this was not undertaken.

• Due to substantial developments in nutritional formulations for oral, enteral and parenteral nutrition, further clarification of the comparison was necessary to meet the initial aims of the review, which were to determine if preoperative nutritional support affected postoperative outcomes. In the updated review, we have specified the comparisons as preoperative nutritional intervention versus control or usual care.

• In the first version of the review (Burden 2012), we expressed continuous outcomes using mean differences and standard deviations to summarise the data for each group. However, in this update, we reported 95% confidence intervals in the abstract and summary of findings tables. In the outcomes measured, we did not state that we would examine within‐group changes and planned to examine between‐group differences in macronutrient intake. We have changed calories to energy as energy can be measured using either kilojoules or kilocalories. In the sensitivity analyses, we included analyses on weight‐losing participants combined with those who were malnourished.

• Four studies included in the review did not record the time point of outcome measures (Gunerhan 2009; Muller 1982;Smith 1988; Von Meyenfeldt 1992). We included these studies in the review although in the protocol and first version of the review we had specified that we would assess outcomes at up to three months postoperatively. We included these studies as we have assumed complications were measured while the participants were in hospital and length of hospital stay (including variance) fell within the three‐month time frame.

• The protocol did not contain a strategy to handle multigroup studies; however, we decided to include groups that met the inclusion criteria to maximise the available data.

• We included additional information in this update regarding the assessment of heterogeneity, data synthesis, summary of findings and assessment of the certainty of evidence. This additional information was added in line with the Cochrane Handbook for Systematic Reviews of Interventions that was published after the original review (Higgins 2017).

• We have amended all occurrences of 'nutrition support' to 'nutrition therapy' (including in the review title) to reflect that this is now the preferred term (Cederholm 2017).

• In line with changes in Cochrane style, in the title of the review we have amended 'Pre‐operative' to 'Preoperative' and 'patients' to 'people'.

• In the first version of the review (Burden 2012), complications was considered a primary outcome, which considered both 'infective' and 'non‐infective' complications. In this update, we have considered these separately to have three primary outcomes: non‐infectious complications, infectious complications and length of hospital stay.

Methods not implemented

We planned to undertake subgroup analyses on studies including participants with cancer and without cancer, elective versus semi‐elective surgery, those that stated the use of an ERAS protocol and route of feeding if data allowed. However, there was insufficient information in the studies regarding these issues, so no subgroup analyses were conducted.

We planned sensitivity analyses to examine the difference in the quality of the studies and to examine the difference in studies conducted before and after 1990. We used this cut‐off date because there have been advances in artificial feeding since 1990, including changes in technology, line care, feeding tubes and monitoring. The type and amount of enteral and parenteral nutrition delivered has also changed considerably. However, there was insufficient information in the studies regarding these issues, so no sensitivity analyses were conducted. For complications and length of hospital stay there were limited data, so we did not undertake further sensitivity analyses on these outcomes.

Planned funnel plots were not undertaken as there were fewer than 10 studies in each analysis (Higgins 2017).

Contributions of authors

Conception and design of the review: SB

Selecting studies: JS, AMS, CF, SRK

Assessing risk of bias, extracting data: JS, AMS, CF

GRADE assessments: JS, AMS, CF

Performing meta‐analysis and writing draft of review: JS

Interpretation of data: JS, SB, AMS

Revision of draft review: AMS, SB, JS, SRK, EMH, CF

All review authors approved the final manuscript.

Sources of support

Internal sources

• University of Edinburgh, UK

  Staff time

• University of Manchester, UK

  Staff time

External sources

• No external sources of support, Other

  No external sources of support

Declarations of interest

AMS: none.

SB: received research grants from Baxter 2021, Nutricia 2022 (small award) and Takeda 2018, 2019. Received travel award from Nutricia 2022. SB is also an author on two of the included studies but was not involved in eligibility decisions, data extraction, risk of bias assessment or GRADE assessments of these studies.

JS: none.

CF: none.

SRK: none.

EMH: none.

New search for studies and content updated (conclusions changed)