9.1. Introduction
For the purposes of these Guidelines, enteral tube feeding (ETF) refers to the delivery of a nutritionally complete feed (containing protein or amino acids, carbohydrate +/− fibre, fat, water, minerals and vitamins) directly into the gut via a tube. The tube is usually placed into the stomach, duodenum or jejunum via either the nose, mouth or the direct percutaneous route37. ETF is not exclusive and can be used in combination with oral and/or parenteral nutrition. Patients receiving ETF should be reviewed regularly to enable re-instigation of oral nutrition when appropriate. Most enteral feeding tubes are introduced at the bedside but some are placed surgically, at endoscopy or using radiological techniques, and some are inserted in the community. Whenever possible the patient should be aware of why this form of nutrition support is necessary, how it will be given, for how long, and the potential risks involved. There may be considerable ethical difficulties in deciding if it is in a patient’s best interests to start a tube feed.
Innumerable questions regarding best ETF practice could be asked but for these guidelines, reviews were restricted to studies providing potential guidance on the indications for ETF, studies on the benefits of ETF compared to oral or parenteral nutrition, and studies on some technical aspects of delivering enteral feeds. No studies on different types of enteral feed were reviewed.
9.4. Enteral tube feeding versus standard care
9.4.1. Introduction
Some patients are put at potential risk of malnutrition (or worsening of pre-existing malnutrition) through a limitation of oral intake or absorptive capabilities from effects of their disease or direct and indirect consequences of surgery (e.g. nausea or ileus and/or clinical practice of restricting post surgical oral intake). If this limitation is severe and long-lasting, nutrition support using ETF or PN will be needed but ETF could also be beneficial for patients who are likely to have limited intake for only a few days (as in most post-operative patients), especially if they already malnourished. However, the benefits from using ETF in this elective, supplementary role is uncertain and it is possible that the risks might outweigh any clinical benefits. Two reviews were therefore conducted to identify:
RCTs comparing patients who received ETF (with or without oral intake) vs. patients receiving standard care (e.g. normal hospital diet and/or oral nutrition supplements) and
RCTs that included patients receiving elective early post-operative ETF vs. no early postoperative nutrition (i.e. nil by mouth post-surgical dietary care with simple IV fluids until clinical signs of returning GI function).
9.4.2. Studies of ETF vs. standard care
The review conducted identified 10 RCTs20,52,143,180,228,315,318,339,340,361 (Table 41). Four of these compared the effect of patients receiving 12 to 24 hours of nasogastric tube feeding plus continued normal hospital diet with patients receiving a standard hospital diet only 143,228,339,340. Two studies compared nasogastric/nasoduodenal feeding with standard hospital diet 52,180. One study compared nasogastric feeding with standard hospital diet plus ad lib snacks 20, while another had two intervention arms in which patients received a nasogastric feed with amino acids alone or a nasogastric feed containing amino acids plus carbohydrates 218. The control group continued on a normal hospital diet. A further study compared oesophagostomy tube feeding with a clear liquid diet, advancing to a normal diet as tolerated 315 and one investigated the benefits of pre-operative ETF (nasogastric tube feeding) compared with routine hospital diet 318. The final study examined the effect of perioperative nutrition in malnourished head and neck cancer patients 361 using three intervention arms: one group received no preoperative and standard postoperative ETF; another group received standard preoperative and postoperative ETF; and the third group received arginine supplemented preoperative and postoperative tube feeding.
The patients included in the studies were orthopaedic hip fracture patients (four studies covering 337 patients)20,143,339,340, people who were generally malnourished (one study covering 86 patients)228, malnourished surgical patients (one study covering 110 patients) 318, total laryngectomy patients (one study covering 67 patients)315, malnourished patients undergoing surgery because of a head and neck malignancy (one study covering 49 patients) 361 and patients with alcoholic liver disease (two studies covering 66 patients)52,180.
9.4.3. Clinical evidence ETF vs. standard care
The main outcomes reported were nutritional intake achieved, changes in nutritional status, mortality, length of stay and complications associated with tube feeding (e.g. tolerance of the feeding tube).
The difference in nutritional intake (usually reported as energy and/or protein intake) between the enterally tube fed patients and those receiving standard care was reported in six studies52,143,180,228,339,340. In all six studies, the enterally fed group achieved a significantly greater nutritional intake (range p<0.0001 to 0.012).
Five studies reported changes in measures of nutritional status 20,180,228,318,361 with three showing improvement20,228,318 (range p=0.001 to p=0.05) while two showed no differences 180,361.
Mortality was reported in 8 studies20,52,143,180,318, 339,340,361. Four showed no differences between groups 20,339,340,361 but one 52 did show significantly lower mortality in the ETF group (p=0.02) and two further studies reported lower mortalities but with no significance values given 180,318. One study, 143 noted a higher mortality rate for the patients who were tube fed but again no p-value was reported.
There were no significant differences in postoperative complications reported in four studies315,339,340,361; nor in the incidence of pressure sores in one study143; diarrhoea in one study 180, or infection rates in one study 52. In one study 318 the incidence of wound infection, nausea and vomiting were lower in the ETF group although no p-value was reported.
Five studies reported that ETF had no influence on length of hospital stay 180,315,339,340,361; although in one study 20, median time to independent mobility was lower in the ETF group (p 0.02 –0.04).
Three studies20,228,318 provided information on patient’s tolerance of ETF but no p-values were reported. In two studies 22%20 and 30%228 of study participants experienced problems tolerating the nasogastric tube. In the third study 318 7 out of 67 patients receiving ETF (10.5%) needed it to be discontinued due to uncontrollable diarrhoea, vomiting or severe aversion to the smell and taste of the feed.
9.4.4. Cost-effectiveness evidence ETF vs. standard care
Four studies were found that reported a cost comparison104,210,234,256: two RCTs, one retrospective cohort study and a study that constructed a simple model on the basis of two small trials (Table 56).
One RCT210 evaluated insertion of double-lumen gastrojejunostomy tube compared with routine care by the surgeon after pancreatico-duodenectomy. Half the patients in the routine care arm received PN; and the other group probably received NG feeding (but the route of feeding was unclear). The study found significant reductions in gastro-paresis and in costs. The second RCT104 compared early nasogastric enteral feeding with early oral feeding after colorectal resection in cancer patients. They found that early oral intervention was safe but there were no cost savings or improvements in clinical outcomes.
The aim of the retrospective study 234 was to test whether there were cost savings in using tube-feeding rather than a carer manually feeding the patient (which requires expensive staff time and risks causing aspiration) for patients with advanced dementia. The results showed that the total costs were higher for the patients with feeding tubes compared with those without tubes (£5,600 vs. £3,100, p=0.04). The difference was due to tube feeding placement cost and hospital costs arising from complications directly related to tube feeding. However, the sample size of this study was small (11 patients in each group) and potentially biased since it was a convenience sample. Costing was also made using Medicaid and Medicare reimbursement rates, which may not be applicable to the UK NHS setting.
The fourth cost-effectiveness study evaluated the cost of preoperative enteral nutrition 256. ETF (10–21 days) was compared with no ETF. The study was a sensitivity analysis based on the two small trials with the largest reduction in complication rate. Incremental cost per complication averted was between £9,000 and £94,500 with hospital preoperative ETF, depending on the assumptions made. However, they found that home preoperative ETF is more likely to be cost saving.
There were no economic studies evaluating pre and post-operative ETF.
9.4.5. Studies of early post-operative ETF vs. no early post-operative nutrition
We identified one systematic review206 that looked at early post-operative feeding (oral or enteral) versus post-operative ‘nil by mouth’. There were 11 studies included in this review: 6 on early post-operative enteral feeding versus no early post-operative nutrition 27,58,148,301,310,367 and 5 on early post-operative oral feeding versus postoperative ‘nil by mouth’ 32,144,258,291,334 (included in the oral chapter 8). In this section we have included the six studies from the systematic review that looked at the effect of early postoperative ETF. In addition to the studies from this systematic review, we identified 17 further studies that looked at the effect of early post-operative ETF versus no early post-operative nutrition. The RCTs were analysed according to the type of surgical patients included in the studies.
Five studies included patients undergoing upper GI surgery 45,148,263,341,367 (Table 51). Three studies included patients undergoing lower GI surgery 215,301,310 (Table 51). Six studies included both upper and lower surgery 27,58,160,298,321,328 (Table 53). Three studies included patients undergoing hepatobiliary surgery 117,145,164 (Table 54). Six studies included acute trauma patients 65,98,175,216,238,281 (Table 45).
We extracted data on seven outcomes: vomiting, anastomotic dehiscence, pneumonia, death, intra-abdominal abscess, wound infection and hospital length of stay (LOS) where available. Where appropriate we pooled the data for these outcomes. We were unable to pool the data for LOS as the studies reported the data in different units and information needed to convert these units was not available.
9.4.6. Clinical evidence: early post-operative ETF vs. no early post-operative nutrition
Analyses for each of the surgical subgroups showed no statistically significant differences in any of the outcomes extracted. The P value from tests for heterogeneity was greater than 0.1 for all outcomes in all the groups.
We also conducted a combined analysis which included all the surgical studies (Appendix Eight: Meta-Analyses Enteral versus Nil Post Operative Nutrition Support). This also identified no statistically significant differences in any of the outcomes extracted which included vomiting, anastomotic dehiscence, pneumonia, intra-abdominal abscess, wound infection and mortality. The data on lengths of hospital stay reported in fourteen studies 58,98,117,145,148,216,238,263,281,301,321,328,341,367 were not adequate to permit a combined analysis but statistically significant differences were only detected in two studies with one showing that early feeding led to fewer days in hospital (p< 0.05)301 whilst the other showed it extended length of stay (p< 0.01)328.
Table 19Outcomes reported in studies of early enteral tube feeding
View in own window
| No. patients (early feeding/late feeding) | RR (fixed) 95% CI |
---|
Vomiting (reported in four studies27,148,175,216 | 298/280 | 1.27 [0.92, 1.75] |
P value from test for heterogeneity | P= 0.21 | |
Anastomotic dehiscence (reported in 10 studies 27,148,215,263,301,310,321,328,341,367 | 257/264 | 0.60 [0.33, 1.10] |
P value from test for heterogeneity | P= 0.79 | |
Pneumonia (reported in 9 studies 27,98,148,216,238,263,310,321,328 | 355/361 | 0.76 [0.53, 1.08] |
P value from test for heterogeneity | P= 0.36 | |
Death (reported in 10 studies 27,98,148,216,263,281,301,310,321,328 | 368/375 | 0.72 [0.45, 1.15] |
P value from test for heterogeneity | P=0.37 | |
Intra-abdominal abscess (reported in eight studies 27,98,148,238,301,310,321,328 | 250/256 | 0.60 [0.32, 1.14] |
P value from test for heterogeneity | P=0.69 | |
Wound infection (reported in 12 studies 27,98,117,148,163,216,263,281,301,321,328,341 | 402/408 | 0.92 [0.68, 1.23] |
P value from test for heterogeneity | P= 0.26 | |
9.4.7. Cost effectiveness evidence: early post-operative ETF vs. no early post-operative nutrition
We identified three cost-effectiveness analyses for ETF compared to nil nutrition postoperatively27,145,147, although all three were small and potentially biased due to methodological weaknesses. Results were inconsistent although all reported a lower number of infections in the ETF groups compared to the nil groups. Estimated effects on cost were as follows:
A non-randomised prospective US study of patients undergoing bowel resection147 showed a cost saving (the magnitude and statistical significance is unclear due to poor reporting) with jejunal feeding tube placed during surgery and feeding initiated within 12 hours of surgery compared with usual care (which was not detailed). The cost savings were due to a reduction in infections.
A small Danish RCT27 reported a non-significant difference in (median) cost of about £1,500 for a 4 day nasoduodenal intervention compared with placebo after major abdominal surgery. Mean costs, which are more relevant than median costs, were not reported.
A small US RCT comparing nasojejunal tube feeding from 12 hours after surgery with maintenance iv fluid after liver transplantation 145 found a non-significant incremental cost of £1,200, despite a 50% reduction in infections. Control patients that were moved to tube feeding were excluded.
9.4.8. Conclusions
ETF in patients where there is some doubt about the adequacy of oral intake is effective in increasing nutritional intake over and above the intake observed with standard care and/or oral supplements and this usually leads to an improvement in nutritional status. However, this does not seem to produce consistent benefit in terms of length of stay or mortality rates and tube tolerance is sometimes a problem in these patients. The evidence of benefit related to complications, quality of life, costs and cost-effectiveness is very limited and ETF use in older people with dementia could be more expensive than oral feeding. The cost-effectiveness of preoperative enteral nutrition is unclear but might be improved if administered in the patients’ home. However, oral nutrition support is likely to be more cost-effective, when this can be tolerated by the patient.
The studies on early post-operative ETF compared to standard practice of nil by mouth until return of GI function, do not support the use of early ETF although most did not focus on very malnourished patients who might benefit from this approach. There may be cost benefits associated with the use of post-operative jejunostomy feeding in some circumstances but more research is needed.
The studies that examined elective ETF in malnourished patients prior to surgery suggest that they benefit in terms of nutritional status. However, much larger trials are needed to determine whether there are any benefits in lengths of hospital stay or mortality.
9.4.9. Rationale for recommendation(s)
Although ETF does increase nutritional intakes in patients the evidence that this benefits outcomes such as length of hospital stay or mortality is not clear.
9.13. Mode of delivering Enteral Tube Feeding
9.13.1. Bolus v continuous
Administering an enteral feed into the stomach rather than small intestine permits the use of hypertonic feeds, higher feeding rates and bolus feeding. Enteral feeding pumps are available to alter rates and in patients with doubtful GI motility, the stomach may be aspirated every 4 hours. If aspirates are high (e.g. exceed 200 – 300 mls depending upon local policy), the pump rate may be reduced and/or prokinetic drugs considered. This is usually recommended in the critical care setting though an aspirate of under 400 ml correlates poorly with the risk of aspiration or pneumonia225. Enteral feeding delivery is usually increased gradually over the first 24 hours (or slower in the very malnourished, see section 6.6).
When using NG feeding, enteral feeds can be delivered continuously over a variable number of hours or intermittently as boluses (or as a combination of both methods). There are potential advantages and disadvantages to both methods. We therefore identified studies that compared different modes of delivering enteral feeds. The RCTs found were categorised into continuous v bolus and continuous (24hr) v continuous (16–18hr). The rationale for non-continuous feeding is that it is more physiological and allows the stomach to completely empty and hence may reduce bacterial colonisation of the stomach which may be safer should an episode of aspiration occur.
9.13.2. Studies on bolus vs. continuous
Nine studies compared continuous v bolus regimens in neurological dysphagic patients, patients with injuries to the head, post-operative cancer patients, critically ill patients314, older patients and healthy adults24,35,66,154,188,201,276,332 (Table 44). Most regimens described in the studies compared 24 hourly continuous feeding with 3–6 hour bolus feeds (250 –500ml). The main outcomes reported were: abdominal discomfort, aspiration pneumonia, change in nutritional status, clogged tubes, nurse preference and biochemical changes.
9.13.3. Clinical evidence on bolus vs. continuous
For abdominal discomfort, aspiration pneumonia and nurse preference there was no evidence of benefit between the continuous and bolus fed group54,66,201,332. However, in one study276 the continuous group were found to have a significant improvement in nutritional status (body weight and arm circumference) compared to the bolus fed group (p<0.01), while in another 66 there was less clogging of nasogastric tubes with bolus feeding (p=0.01).
9.13.4. Continuous vs. cycled continuous
Five studies compared continuous ETF (24hours) v cycled continuous ETF (16–18hours) with daily breaks (2 – 4 hours) or even intermittent ETF (e.g. 4–6 hours feeding then 2 hours rest). Studies were undertaken in critically ill, ventilated patients and post surgical patients35,55,134,323,360. The main outcomes reported were; length of hospital stay, duration of enteral feeding, mortality, ventilator associated pneumonia, gastric pH and rate of gastric colonisation.
9.13.5. Clinical evidence: continuous vs. cycled continuous
There were no significant differences between the 24 hour continuous feeding groups and the 16–18 hour feeding groups in either mortality or ventilator associated pneumonia; 35,134,360, and rates of gastric colonisation and levels of gastric pH were also similar35,323. In one study however 360 there was a significant reduction in hospital stay for a 16 hour fed group compared to a 24 hour continuous group (p=0.04).
9.13.6. Cost-effectiveness
No study reporting cost or cost-effectiveness was found.
9.13.7. Conclusions
Bolus feeding is as effective as continuous (16–24 hours) feeding. Overall, however, the mode of feed delivery can be dictated by practical issues. For example, in patients who pull or dislodge nasogastric tubes regularly, bolus feeding can be used as a practical safe alternative to continuous feeding, while in intensive care the severity of illness and issues of gastric emptying, metabolic stability and control of glucose levels favour continuous feed administration.
9.19. Research recommendations
What are the benefits to Intensive care patients likely to stay for >5 days, who are offered ETF only compared to ETF and PN if they fail to tolerate >60% of their target nutritional needs 2 days after starting ETF in terms of survival, complications and hospital costs?
This is an area of common practice but where the benefits of these interventions are unclear and poorly reported.
What are the benefits to malnourished surgical patients who have indications for ETF being offered ETF only compared to ETF and PN if they fail to tolerate >60% of their target nutritional needs two days after starting ETF in terms of survival, complications and hospital costs?
Currently patients who present with the indications for enteral feeding are being given PN early when it seems that they are not tolerating enough enteral feed to meet requirements, however the benefits of fairly early intervention with PN are unclear.
What are the benefits of enteral tube feeding compared to no enteral tube feeding in people with dysphagia and early to mid stage dementia in terms of reduced complications associated with swallowing, improved nutritional status, delay onset of advanced stage dementia, hospital admissions, cost effectiveness and survival?
Much of the research tends to tends to focus or concentrate on tube feeding people with advanced dementia or those who may be in terminal stages of the disease. Depending upon the types of dementia a person has swallowing disorders may occur at an earlier stage in the disease, for example vascular dementia. The benefits and complications of tube feeding may be quite different in people in the earlier stages than those who are in the advanced stage of dementia.