Postgraduate Medical Journal (November 1984) 60, 779-790

Enteral nutrition

D. B. A. SILKM.D., F.R.C.P.

Department of Gastroenterology and Nutrition, Central Middlesex Hospital, London, NWJO 7NS

This issue attests to the fact that Sir Francis AveryJones has had wide based interests during his 50years in clinical medicine. In recent years he hasbecome one of the major forces behind the promo-tion of clinical nutrition. He has directed much effortto bridging the gap that has for so long existedbetween the non-medical academic centres of re-search in nutrition and the practising clinician. It isno coincidence that several major recommendationsmade by the British Nutrition Foundation TaskForce (Avery Jones, 1983a) on which he was a veryactive participant, reflect completely his longstandingviews about what needs to be done to improve thenutritional management of patients.

Sir Francis has always believed that there has beena need for medically qualified personnel to partici-pate actively in research on clinical nutrition, andover the years research groups based in his Unit atCentral Middlesex Hospital have worked in this area.He has promoted the teaching of nutrition in bothhealth and disease at both the undergraduate andpostgraduate stages of medical training. It was clearwhen I arrived at Central Middlesex Hospital in1978, that my own personal research interests in noway conflicted with those existing within the Unit,and with Sir Francis Avery Jones' encouragement,we have merely enlarged an existing basic physio-logical and clinical research programme. Moreover,it was Sir Francis who persuaded Dr J. J. Misiewiczand myself to alter the name of our department fromthat of Gastroenterology alone to Gastroenterologyand Nutrition.One of Sir Francis' beliefs (Avery Jones, 1983b)

has been the need in district hospitals for onedepartment to take the lead in organizing a 'nutritionteam' to provide a nutrition advisory service for thehospital and to co-ordinate with the district medicalservices, thereby emphasizing the importance ofnutrition in the community and in postgraduateeducation-this I hope we have been able to achieve.

Sir Francis has provided encouragement for all thework we have done in the last 5 years, and especiallyfor our efforts in the field of enteral nutrition, I have

therefore chosen this subject for his Festschrift. Theensuing discussion will include not only an update onpractical techniques involved in feeding patientsenterally but also reference will be given to some ofthe problems that still exist in defining what consti-tutes 'malnutrition' in the hospitalized patient as wellas to the problems that exist in defining the nutri-tional requirements of the individual patient.

Incidence and diagnosis of protein calorie malnutri-tion

Recent studies have suggested that 20-50% ofhospitalized patients have some evidence of proteincalorie malnutrition (Bristian et al., 1974; Hill et al.,1977). These findings in turn indicate that awarenessof the nutritional status of patients is a neglected areaof clinical management and it follows that nutritionalassessment is the first essential step in the nutritionalcare of any patient. Few problems arise in recognis-ing the grossly malnourished patient in whom weightloss is marked, hypoproteinaemic oedema is present,muscle wasting extreme, and serum albumin levelsare lower than 25 g/l. Problems may arise, however,in reaching the diagnosis in the fitter looking patientwho at first sight may even look obese (Fig. 1).During the last few years a battery of clinical,biochemical, haematological and immunologicaltests have been described as aids to reaching adiagnosis of protein calorie malnutirition (Table 1).

Clinical historyA routine clinical history usually neglects the

nutritional intake. We have been surprised to findthat patients randomized to controlled clinical trialsof enteral nutrition have had a markedly reducednutritional intake for as long as 3 weeks beforenutritional support was instituted (Jones et al., 1983;Keohane et al., 1983).

Clinical examination

Many nutritional deficiencies are unaccompanied

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

D. B. A. Silk

TABLE 1. Diagnosis of protein calorie malnutrition (after Silk 1983)

Values suggestive ofprotein calorie

Measurement malnutrition

Clinical and dietary history seClinical examinationBody weight Loss> 10% normal

weightAnthropometric measurementMid-arm circumference < 23 cm c < 22 cm 9Mid-arm muscle circumference < 19 cm d < 17 cm 9Skinfold thickness see Silk (1983)

Measurements of protein stores24 hr urinary creatinineCreatinine height index see text3-methyl histidine excretionCirculating hepatic proteinsAlbumin <35 g/lTransferrin <2 g/lPrealbumin < 200 mg/lRetinol-binding protein < 100 mg/l

Immunological testingLymphocyte count <200/mm3Delayed hypersensitivity skin see Silk (1983)testing

by physical signs, but clinical examination shouldalways be carried out. Angular stomatitis, increasedcapillary fragility, anaemia, muscle wasting andoedema signify severe nutritional deficiency. Loss ofbody weight may provide the first clue that a patientis malnourished and we attach importance to a loss ofmore than 10%o of normal body weight. We havefound ideal weight for height and body mass index tobe of more limited value.

Anthropometric measurements

Anthropometric measurements, such as arm mus-cle circumference and triceps skinfold thickness(TSF) and derivatives of these arm measurementshave gained increasing importance in establishing theprotein and fat reserves of hospitalized patients. Anassessment ofmuscle mass can be gained by means ofthe midarm muscle circumference (MMC). This iscalculated from the midarm circumference andtriceps skinfold thickness, from the formula:MMC = midarm circumference minus (7r x tricepsskinfold thickness). Erroneous measurements can beobtained in oedematous patients in whom correla-tions between changes in body nitrogen and midarmcircumference and midarm muscle circumference arenot exact (Collins, McCarthy and Hull, 1979).Although of limited value in obese patients and thosewith peripheral oedema, body fat content can beestimated simply with Harpenden skin callipers tomeasure skinfold thickness at four sites (James,1982).

Measurements of protein stores

Malnutrition is associated with a reduction in totalbody protein content and, in clinical practice, mea-surements of muscle mass have been regarded asrelatively sensitive measures of total body protein.Midarm muscle circumference, as described above, isthe simplest method available. Other methods forassessing protein stores, including measurements of24 hr urinary creatinine excretion, the creatinineheight index, neutron activation analysis for wholebody nitrogen, whole body counts of potassium andisotope techniques for measuring total exchangeablepotassium and water content have all been described(for review, see Silk, 1983) but have not found a placein our routine clinical practice.A great deal of interest has surrounded the

measurement of 3-methylhistidine to assess the de-gree of muscle breakdown. When released 3-methyl-histidine is not re-utilised in protein synthesis and isexcreted unchanged in the urine. The excretion of 4.2,umol of 3-methylhistidine represents the breakdownof 1 g of mixed muscle protein (Munro, 1978).Indirect measurements ofprotein synthesis rates

Numerous claims have been made that measure-ment of circulating hepatic proteins provides usefulinformation about nutritional status, and response tonutritional support. In the absence of parenchymalliver disease and proteinuria, the serum albuminconcentration provides useful information about thenutritional state, values below 30 g/l indicatingmoderate-to-severe malnutrition. Although claimshave been made to its value in monitoring theresponse to treatment, albumin has a long half-life,and in our experience (Jones et al., 1983), as well asthat of others, shows a poor response to short-termnutritional support (Shetty et al., 1979).

Transferrin has a half-life of 8 days, shorter thanalbumin, and plasma levels are reduced in proteincalorie malnutrition (Ingenbleek et al., 1975). Itsusefulness in clinical practice is limited by co-existingiron deficiency states which promote its synthesis,and in patients who are infected and undergoingstress as part of the 'acute phase' protein response.

Thyroxine-binding prealbumin (TBPA) and reti-nol binding protein (RBP) are two hepatic proteinswhose plasma levels are also reduced in obviousprotein calorie malnutrition (Ingenbleek et al., 1975).TBPA has a half-life of 2 days, and RBP of 12 hrand as such both would be expected to respond tochanges in nitrogen and energy input as occur duringthe nutritional therapy of malnourished patients.Studies indicate, though, that TBPA levels are moresensitive to changes in energy rather than nitrogenintake. Plasma levels ofRBP on the other hand alterrapidly in response to changes in both energy and

780

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

Enteral nutrition

nitrogen intake (Shetty et al., 1979) so that if closemonitoring of nutritional support is required in theclinical setting, plasma RBP, when measured regu-larly, should provide valuable information.

Immunological testingProtein calorie malnutrition has now frequently

been shown to be associated with impaired immuno-competence (Law, Dudrick and Abdon, 1979) and areduction in total peripheral lymphocyte count tobelow 1200/mm3 is seen in obvious protein caloriemalnutrition.

Negative skin testing with ubiquitous antigens hasbeen frequently reported in patients with proteincalorie malnutrition and skin testing has been pro-posed as a practical aid to the diagnosis of proteincalorie malnutrition (Pietsch, Meakins and Maclean,1977).We have confirmed that patients with obvious

protein calorie malnutrition have negative skin tests(i.e. are anergic), just as have patients with lessobvious clinical malnutrition. We have not, however,been able to confirm the observations of others(Pietsch et al., 1977) that effective nutritional supportis associated with reversal of the anergic state. In thisrespect it is important for the reader to be aware ofother data showing that ageing, cancer, sarcoidosis,infection, shock, acute trauma and zinc deficiencyhave all been shown to be associated with anergy perse, irrespective of nutritional status (see Silk, 1983).

Expectations of the nutritional measurements

Ideally, the nutritional measurements discussedabove should provide not only information about thediagnosis of protein calorie malnutrition, but serialmeasurements should also provide information aboutthe effectiveness of nutritional support. Finally, itwould be of immense value if clinically usefulmeasures of malnutrition could be established thatwould reliably predict the likely occurrence of post-operative complications. It is our feeling however,that an unreasonable burden of expectation has beenplaced on the use of these nutritional measurements.

1. Aids to diagnosis ofprotein calorie malnutritionAt the outset of our nutritional programme we

assiduously measured all, or most, of the nutritionalmeasurements, and decisions as to whether to insti-tute nutritional support were based on these results.We began to place less reliance on these tests when itwas shown that there was close agreement betweenassessments of nutritional status based solely on theexperience of an experienced clinician as comparedwith the use of a large battery of nutritional

FIG. 1. Severe protein calorie malnutrition, but less obviousclinically (female aged 65 years). Wound dehiscence and enterocuta-neous fistula following second laparotomy for recurrent small bowelobstruction. Note apparent obesity. Serum albumin 24 g/l. Patient is

receiving nutritional support via the enteral route.

measurements (Baker et al, 1982). Moreover, as wegained more experience, we began to realise that oneof the major roles of nutritional support is to preventthe development of protein calorie malnutrition.Consequently, our decisions as to whether to institutenutritional support are now influenced more byprevious dietary history and the natural history of theprimary disease process in respect of future nutri-tional intake than by nutritional measurements. Inpractical terms, therefore, we currently place verylittle emphasis on nutritional measurement.

2. Monitoring treatment

Many authors have advocated serial nutritionalmeasurements as a means of monitoring nutritionaltherapy (Law et al., 1973; Spanier et al., 1976;Meakins et al., 1977). We have not been impressedhowever, with the clinical significance of changesthat occur during treatment, and do not advocate theroutine measurement of all those listed in Table 1.

781

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

D. B. A. Silk

We place more emphasis on monitoring nitrogenbalance during treatment and the early detection ofclinical metabolic and haematological side effects.

3. Prediction ofpostoperative morbidityIn surgical circles, there is growing enthusiasm for

vigorous nutritional support of malnourished pa-tients about to undergo major surgery (Hill, 1979).This argument is based on the premise that malnour-ished patients have more postoperative complicationsand even an increased risk of dying.

In order to demonstrate significant clinical benefitsof pre-operative nutritional support, it would benecessary to identify those patients who would be atrisk of developing postoperative complications beforerandomization into controlled clinical trials. Severalgroups are therefore currently assessing the degree ofreliance that can be placed on the use of nutritionalmeasurements as a means of identifying 'at risk'patients.

Claims have been made that degree of weight loss,negative delayed hypersensitivity skin testing, re-

duced anthropometric indices, low transferrin levelsand reduced grip strength dynamometry values areall reliable predictors of postoperative morbidity.The problem has been however, that no singlemeasurement can be reliably used as a predictive testin the individual patient. It remains to be seenwhether the various prognostic nutritional indicesbased on the use of several tests will prove more

successful at reliably predicting whether the indivi-dual patient is at risk of developing postoperativecomplications.

Indications for providing nutritional supportThe indications for instituting nutritional support

have not been properly elucidated. One of the majorreasons for this has been the lack of properlyperformed clinical trials that include the withholdingof specific nutritional therapy from one group ofpatients. The difficulties in identifying 'at risk'patients with, for example, a high risk of developingpostoperative complications has often led to theinclusion of large numbers of normally, or nearlynormally nourished, rather than malnourished pa-tients in clinical nutrition studies. This has made theinterpretation of the benefits of nutritional supportvirtually impossible. Finally, when differences arenoted betweeen treatment and control groups it is amatter of opinion whether they are clinically signifi-cant; for example, reducing hospital in-patient stayafter surgery by a short period can hardly be cited as

being a clinically significant benefit of nutritionalsupport.

Guidelines for assessing the indications for provid-

ing nutritional support used at Central Middlesex aresummarised in Table 2.

TABLE 2. Indications for nutritional support

1. Obvious gross malnutrition: albumin<30 g/l; marked weightloss; muscle wasting; oedema

2. Nutritional measurements suggestive of protein calorie malnu-trition-dietary history shows impaired nutrient intake for oneweek or more

3. Medical and surgical disorders likely to result in protein caloriemalnutrition if nutritional support withheld

In many previous reviews it has of course beentraditional to categorise those clinical conditions thatmay be frequently associated with protein caloriemalnutrition and to imply that patients with theseconditions are those likely to require nutritionalsupport. Such an approach however, is a gross over-simplification of the problems that arise when decid-ing whether or not to institute nutritional supportbecause, as one's experience of clinical nutritionalbroadens, one realises that almost all medical andsurgical conditions can be accompanied by proteincalorie malnutrition. Despite the difficulties that existin defining the indications for nutritional support,clinical trials and clinical experience have high-lighted a number of areas where clear benefits havebeen obtained (Table 3).

TABLE 3. Specific indications for nutritional support

Definite benefitManagement of short bowel syndromeIntestinal fistulas-without distal obstructionIntra abdominal sepsisMajor burnsMajor sepsisMultiple injuries involving visceraPancreatic abscess/pseudocyst/trauma/fistula

Possible benefitAcute pancreatitisInflammatory bowel diseasePre-operatively*-major surgery including cancer surgeryPost-operatively-major surgery and complications

*See text-value of nutritional measurements in predictingpost-operative morbidity.

Assessment of nutritional requirements (Table 4)Any attempt to correct the nutritional deficits in

under-nourished or injured patients must include theprovision of a readily utilised energy substrate,nitrogen for protein synthesis, fluid, electrolytes,vitamins and haematinics.

782

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

Enteral nutrition

TABLE 4. A guide to the daily nutritionalrequirements for enteral and parenteral nu-

trition

NitrogenEnergy

Electrolytes andtrace elementsSodiumChloridePotassiumCalciumMagnesiumPhosphateFluorideManganeseCopperZincIronSeleniumIodineChromium

8-20g6270-16 720 kJ(1500-4000 cal)

70-150 mmol70-220 mmol50-100 mmol5-10 mmol5-20 mmol30-50 mmol

50 ,umol2-4 mg1.5-3.0 mg10-20 mg10-15 mg100-200 /lg10-200 jug5-10 ,lg

Energy and nitrogen (Table 5)There is a close relationship between energy and

nitrogen balances so that the requirements of the twocan be considered together. In a normal active man,positive nitrogen balance is achieved with a nitrogenintake of 8-9 g/24 hr. Non-protein energy require-ments are in the region of 1338 kJ/g (320 cal/g) ofnitrogen. As the metabolic expenditure of a starved,re-feeding patient is low, nitrogen equilibrium can beattained with a nitrogen intake of 7 g/day, with anon-protein requirement of 1200 kJ/g (250 cal/g) ofnitrogen. When the metabolic rate is increased afterinjury or during a febrile illness, relatively morenitrogen is needed to produce positive nitrogenbalance. Ideally nitrogen requirements should beestimated from direct measurements of nitrogenlosses in urine, stool and ostomy or fistula effluents.

TABLE 5. Approximate requirements for energy and nitrogen*

Non-catabolic Intermediate Catabolic

Nitrogen (g/day) 7.5 14 25Energy kJ/day(cal/day) 9600(2000) 14,400(3000) 19,200(4000)Non-protein energyin kJ/g 1200(250)(cal/g) of nitrogen*After Woolfson (1978).

In practice the methods are cumbersome, timeconsuming and costly. With the exception of patientswith protein-losing enteropathies and fistulae, reli-able estimations of nitrogen losses (NL) in g/24 hrcan be calculated from the following formula NL =(mmol urinary urea/24 hrx0-028)+2, where thefactor of 2 is added to take into account constant non-

urea urinary nitrogen as well as faecal and skinlosses. Simple corrections for proteinuria and risingblood urea values in renal failure patients have beendescribed.

ElectrolytesAttention should be paid to correcting electrolyte

imbalance in all patients receiving nutritional ther-apy; the values shown in Table 4 can be consideredonly as a guide to requirements. Potassium lossesoccur in a fixed ratio to nitrogen losses, so at least 5mmol potassium should be administered with eachgram of nitrogen.As a general principle, biological fluids should be

collected from appropriate orifices of sick patients(e.g. gastric aspirates, fistula and drain effluents,urine, diarrhoea, faeces) and analysed for electrolytecontent. Hypophosphataemia is a common complica-tion of parenteral (intravenous) nutrition; and there-fore at least 30 mmol phosphate must be adminis-tered daily.

Minerals

Mineral deficiencies in patients receiving long-term nutritional support have been well documented.Unfortunately the minimal daily requirements ofmost minerals are not known with any certainty, sothat the values for manganese, copper, zinc andselenium given in Table 4 are, at best, approximate,and further research will almost certainly indicatethat additional trace metals should be added to thislist.

Vitamins

Vitamins are necessary for the utilisation ofnutritional components and vitamin deficiencies canoccur rapidly in the debilitated state. Vitamin supple-ments should be given as soon as nutritional supportis started.

Methods ofproviding nutritional support (Table 6, 7)It is important to emphasise that when gastrointes-

tinal function is normal attempts should be made toinstitute enteral rather than parenteral nutrition.Ideally the patient should be fed orally, either withwhole or liquidised food, or with the addition ofpalatable supplements. If this is not possible liquidfeeds can be administered via a naso-enteric tube.Some patients requiring nutritional therapy have agastrostomy or jejunostomy and nutrition can beadministered via these routes. Only if these methodsfail should nutrients be provided via the parenteralroute.

783

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

D. B. A. Silk

TABLE 6. Factors limiting achievement of nutritional objectives

Most commonSepsisTrauma

Severe fracturesMajor visceral injuryMajor burnsMajor operations

ImmobilityBedboundSplinting of fracturesNeurological causesMuscle relaxant drugs in intensive carePain

CommonTechnique

Insufficient energy supplyInsufficient amino-nitrogen supplyInsufficient co-factors for nitrogen utilisation

Fluid balanceInability to tolerate volume required for nutritional requirements

UnusualMetabolic

Hyperglycaemia unresponsive to insulinHyponatraemia due to inappropriate anti-diuretic hormone secre-

tionOccult losses of proteinLow ambient temperature

Enteral nutrition

Once the decision is made to provide nutritionalsupport via the enteral route, the best formulation tomeet the nutritional needs of the individual patientmust be selected. There are now a good manycommercially available enteric diets in the U.K. andimportant differences exist in respect of their nutri-tional components.

Diet formulation

1. Diets for patients with normal gastrointestinalfunction

Patients with normal gastrointestinal function arecapable of assimilating whole protein and unhydro-lysed triglyceride. Enteric diets for these patientsshould therefore contain whole protein as the nitro-gen source, and a mixture of triglycerides andcarbohydrate as the energy source. Our recentphysiological studies indicate that high molecularweight glucose polymers are the most suitable car-bohydrate energy source (Jones et al., 1981). Contro-versy exists as to what constitutes the optimum non-protein energy to nitrogen ratio of these diets. Sincethe nitrogen and energy requirements of differentpatients varies (Table 5), a ratio of 150 kcal/g Nwould appear a reasonable compromise. Enteric dietsfor those patients with normal gastrointestinal func-tion should contain sufficient electrolytes, traceelements and vitamins to satisfy nutritional require-ments.

TABLE 7. Methods of providing nutritional support

Route Nutritional source

Oral feeding FoodLiquidised foodPalatable enteral diet

Tube feeding Enteral dietsNasogastricNasoduodenalNasojejunal

Gastrostomy Liquidised foodEnteral diet

Jejunostomy Enteral dietParenteral nutrition

2. Diets for patients with impaired gastrointestinalfunction

In a small group of patients, nutrient assimilationmay be impaired by insufficient luminal nutrienthydrolysis or because the functional absorptivecapacity of the intestine is so reduced as not to beable to cope with the quantities of nutrients presentedto it for absorption. In these circumstances it followsthat nutrients should be presented to the gut in apredigested form, and indeed to ensure that themaximal possible absorption is achieved the predi-gested nutrients should ideally be presented in theform in which they are absorbed fastest in the normalclinical setting.

Conditions in which luminal nutrient hydrolysis isseverely impaired include severe exocrine pancreaticinsufficiency, obstructive jaundice and the shortbowel syndrome. The functional absorptive capacityof the intestine may be severely reduced in the shortbowel syndrome and in clinical conditions character-ized by a severe and extensive mucosal lesion (coeliacdisease, severe Crohn's disease). These then becomethe indications for the use of predigested diets. Theold term 'elemental diet' is a misnomer. Used initiallyto describe the early free amino-acid and glucose-containing diets, it has later been used to describe allthe other diets containing predigested nutrients thatwere far from 'elemental' in design.The use of 'elemental' diets has been advocated in

a variety of clinical conditions (Russell, 1975) includ-ing the management of gastrointestinal fistulae,inflammatory bowel disease, the nutritional manage-ment of patients with gastrointestinal cancer andthose with severe maldigestion and malabsorptionand, finally, their use as a means of pre-operativebowel preparation has also been suggested. In abrilliant expose of the myths surrounding the indica-tions and uses of 'elemental' diets, Koretz and Meyer(1980) point out that there is very little controlleddata to support any of these claims. In our unit, wehave just shown (Jones et al., 1983) that there is noevidence for the superiority of 'elemental' over

784

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

Enteral nutrition

polymeric diets containing whole protein as a nitro-gen source in the management of malnourishedpatients with normal gastrointestinal function, anobservation that was also noted during an investiga-tion of the efficacy of 'elemental' and polymeric dietsin unconscious patients with head injury (Jones et al.,1980).The nitrogen source of these predigested diets

should, in the light of our current knowledge, consistof oligopeptides rather than free amino acids. Partialenzymic hydrolysates of whole protein constitute themost suitable peptide-based nitrogen source, al-though the optimum starter protein, hydrolysis pro-cedure and peptide chain length have yet to bedefined. Because our intestinal perfusion studiesindicate the purified high molecular weight glucosepolymers are assimilated in the absence of luminalamylase activity (Jones et al., 1981) these shouldprobably constitute the carbohydrate fraction of theenergy source.

Because these diets are probably indicated inconditions in which luminal fat digestion is impaired,they should not contain large amounts of triglycer-ide-based energy source. It is not completely clear asto whether large quantities of medium-chain trigly-cerides (MCT) are absorbed in the absence ofluminal hydrolysis, so medium-chain triglyceridesshould probably not be included in these diets. Thediets should contain at least 4% of their total energycontent as linoleic acid, however, to prevent thedevelopment of essential fatty acid deficiency, as wellas sufficient electrolytes, trace elements and vitaminsto satisfy the requirements of the average malnour-ished patient.

Enteral diets currently available

Before the recent resurgence of interest in enteralnutrition, those patients with normal gastrointestinalfunction who did receive nutritional support were fedtube feeds prepared in the hospital dietetic depart-ment. Although these tube feeds are cheap, thepreparation of large quantities places a significantburden on the workload of the dietetic department,which is the major reason why they are not routinelyused in our hospital. Care should be taken to ensurethat these diets are prepared under reasonably sterileconditions, as problems with infection have been welldocumented (Casewell, 1979) and there is nowcontrolled data to show that the incidence of diar-rhoea is higher when 'home brew', rather thancommercial diets, are used for enteral feeding(Keighley et al., 1982).Some of the more widely used polymeric and

predigested 'elemental diets' are listed in Table8.

Routes of administration

OralfeedingWhen the patient is able to swallow, but still

unable to eat normal food (e.g. because of anoesophageal stricture), a liquid diet can be providedfrom any hospital diet kitchen. The large volume offluid required to liquidize whole food may make thistype of supplement unacceptable to some patients.The palatable standard enteric diets and the

elemental diets listed in Table 8 are suitable for oralfeeding, both as a means of supplementing aninadequate intake of normal food and as the solemeans of nutritional support.

If it is planned to use the proprietary enteric dietsas the sole means of nutritional support, we prefer toadminister the diets via a nasoenteric tube, becausewe find it easier to document and control intakeprecisely. Another reason is that the patients find thatmost of the diets are not as palatable as themanufacters claim and excessive nursing time istaken up cajoling patients to ingest the desiredquantities.

Nasogastric tubefeedingPatients who cannot swallow, or who will not

tolerate oral feeding, can be fed via a nasogastrictube. The time-honoured method of tube feeding hasbeen to place a large-bore Ryle's tube into thestomach and intermittently to instil up to 200 mlliquid feed, having first aspirated the gastric residue.This method is probably responsible for the poorreputation of enteral feeding, as the incidence ofside effects, namely diarrhoea, aspiration pneumoniaand oesophageal ulceration, has been unacceptablyhigh.Major new developments have taken place in this

field and most of these problems have now beencircumvented. Enteric feeds (Fig. 2) can be adminis-tered from containers by gravity infusion via narrowbore feeding tubes, most of which are inserted, withthe aid of an introducer, into the stomach. These fine-bore tubes are more comfortable than the older widerbore tubes and their use has not so far beenassociated with oesophageal erosions, ulcers or stric-tures. Although easy to pass, the final position shouldbe checked radiologically or by insufflating 5 ml airand auscultating over the epigastrium to ensure thatthe feeding tube has been positioned in the stomachand not the bronchial tree.As Fig. 3 shows, there are now a large number of

fine-bore feeding tubes on the market. Broadlyspeaking there are the simple unweighted tubes orfeeding tubes with weighted tips. In our experience,in routine clinical use, the more expensive weightedtubes offer few advantages over the simple un-

785

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

786 D. B. A. Silk

TABLE 8. Proprietary enteral diets

Non-proteinFlavouring Osmolality energy/Nitrogennecessary Presentation mosmol/kg ratio (kcal/g N)

Polymeric diet*Build up (Carnation) No Powder 575 100Clinifeed 400 (Roussel) No Liquid 426 142Clinifeed Favour (Roussel) No Liquid 403 142Clinifeed Iso (Roussel) No Liquid 321 200Ensure (Abbott) No Liquid 430 154Fortison standard No Liquid 300 131(Cow and Gate)

Fortison Energy-Plus No Liquid 410 167(Cow and Gate)

Isocal (Mead Johnson) Yes Liquid 350 170Nutrauxil (Kabivitrum) No Liquid 386 140Trisorbon (BDH) No Powder 400 130

Chemically defined(elemental diets)Flexical (Mead Johnson) Yest Powder$ 580 256Nutranel (Roussel) Yest Powder$ 550 131Vivonex HN (Eaton) Yest Powderx 830 121

*All have whole protein as nitrogen source.tin the opinion of the author.tOligopeptate nitrogen source.xFree amino acid nitrogen source.

TABLE 9. Complications of enteral nutrition

Tube insertionOesophagitisOesophageal erosionsOesophageal strictureTube misplacementTube withdrawal

Gastrointestinal side effectsDiarrhoeaAbdominal distensionAbdominal painIntussusception

Metabolic complicationsHyperglycaemiaHypokalaemiaHypomagnesaemiaHypocalcaemiaHypophosphataemiaLow zinc levelsLow red cel folate

Abnormalities of liver functionIntravenous administration of enteric feedsRegurgitation and aspiration

weighted and open-ended feeding tubes. There arethree clinical areas, however, where we have noted adistinct advantage (Keohane and Silk, 1983).1. Nasogastric intubation of patients already intu-

bated with an endotracheal tube.2. Intubation of patients with oesophageal strictures

in whom endoscopic intubation with unweightedtubes has failed.

3. Nasoenteric feeding (see below).

Endoscopic tube placementAlthough nasogastric tube placement usually

presents few problems, difficulties do arise whenattempts are made to perform nasogastric intubationin patients with oesophageal strictures, and nasoen-teric intubation of those with gastric atony. In thefirst group of patients, irrespective of whether thestricture is malignant or benign, intubation is usuallyrequired either during diagnostic endoscopy or im-mediately following endoscopic dilatation (Keohane,Attrill and Silk, 1982). In the second group, despiteclaims to the contrary (Metz, Dilawari and Kellock1978; Dobbie and Butterick, 1977), feeding tubesrarely pass through the pylorus spontaneously, andhave to be placed in the jejunum or duodenum atendoscopy.

Nasoenteric feedingIn certain patients with neurological disorders of

swallowing mechanism (e.g. motor neurone disease,pseudobulbar palsy) or gastric atony, regurgitation oraspiration of enteric feeds administered nasogastri-cally necessitates the cessation of nasogastric feeding.Theoretically, these side effects should be circum-vented by direct duodenal orjejunal feeding. In thesepatients it is our current policy to intubate theduodenum endoscopically with one of the longerweighted tubes.

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

Enteral nutrition 787

7i

FiG. 2. Enteral feeding of a patient following neurosurgery. Patientis intubated nasogastrically with a fine bore feeding tube, andreceiving an enteral diet by constant gravity infusion from a 500 ml

glass container via a standard giving set.

Early postoperative nasoenteralfeeding'The recognition that small bowel function returns

earlier on in the postoperative period than gastricfunction has lead to the concept of early post-operative enteral feeding via fine needle catheterjejunostomy or via nasoenteric tubes positioned atthe time -of laparotomy (Sagar, Harland and Shields,1979). On theoretical grounds, this would appear to.represent an ideal means of providing nutritionalsupport in the early postoperative period, althoughthe data is far from clear in respect of clinicaladvantages (Young et at., 1978). In most of thepublished studies 'elemental' rather than polymericdiets were used (for which there appears to be littlerationale) and in one controlled study (Sagar et at.,1979) which claims advantages, the only possiblebenefit (and a questionable one at that) that seemsactually to have accrued in the treatment group wasthat the patients were discharged from hospitalsooner than those not receiving nutritional support.

Administration techniques

Enteric diets are best administered from feedreservoirs via a giving set and feeding tube directlyinto the stomach, duodenum or jejunum (Fig. 2).There has been a recent proliferation of deliverysystems (Fig. 4) and debate exists as to the idealvolume of the food reservoirs. A recent study(Bastow, Allison and Greaves, 1981) indicates thatfeeds which are mixed in the diet kitchen are subjectto significant bacterial contamination (e.g. prepara-tion of standard 'home brew' tube feeds or energysupplementation of other feeds). During administra-tion, bacterial multiplication occurs, so those diets, ifused, should not be prepared in volumes of morethan 0.5 litres. No significant bacterial contamination

FIG. 3. A selection of commercially available nasogastric andnasoenteric feeding tubes. From inside to outside: (1) ClinifeedingSystem 1 (Roussel Laboratories: PVC, 70 cm. int. diam 1 mm); (2)Entri nasogastric feeding tube (Biosearch: Erythrothane, 81 cm. no.6 French); (3) Prima enteral feeding tube (Portex; PVC, 85 cm, 1.1mm); (4) Hydromer-Dobbhoff enteric feeding tube (Biosearch;Erythrothane, 109 cm. no. 8 French); (5) Entriflex nasogastricfeeding tube (Biosearch; Erythrothane, 91 cm. no. 8 French); (6)Enteral feeding tube (SHS; Silicone, 125 cm. no. 7 French); (7) DuoTube enteral feeding unit (Argyle; Silicone, 102 cm. no. 8 French).

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

788 D. B. A. Silk

FIG. 4. A selection of enteral feeding bags and reservoirs. From left to right: (1) Roussel Clinifeeding System 3 (1-5 litres); (2) Re-usuableDHSS Winchester container (0-5 litres); (3) Express enteral feeding bag (2 litres); (4) Viomedex enteral feeding bag (15 litres); (5) Boots Flow

Fuser enteral feeding bottle (0-5 litres).

of pre-sterilised enteral feeds appears to occur duringfilling of reservoirs, so if these diets are employed, upto 2 litre volume reservoirs can be used.

Enteralfeeding pumps

In our unit all diets are administered continuouslyover 24 hr, as a means of reducing the incidence ofgastrointestinal side effects (see below). In a prospec-tive study (Jones, Payne and Silk, 1980) over 85% of80 patients were fed successfully when the diets wereadministered by simple gravity infusion using thegiving-set clamp to control the infusion rate. Clearly,therefore, one does not have to advocate the routineuse of an enteral feeding pump. The use of a pumpwas, however, shown to be beneficial in patients withimpaired gastrointestinal function who developeddiarrhoea during enteral feeding and a saving of upto 30 min nursing time per patient per day can beachieved if diets are administered using an enteralfeeding pump.

Use of starter regimes

Upper abdominal symptoms including distension,discomfort and colicky pains, as well as diarrhoea,are said to frequently occur if full-strength entericfeeding regimes are introduced too quickly (Silk,1980). A number of factors have been implicated inthe pathogenesis of these side effects. These include

intolerance to high osmotic loads of nutrients ad-ministered during enteral feeding, lactose intoler-ance, use of contaminated feeds and concomitantantibiotic therapy (Silk, 1983). It has been generallybelieved that the incidence of these side effects can beminimized by gradually introducing full-strengthenteric feeds over a 3-4 day period by means of'starter regimes'.Our most recent controlled studies have however,

shown that starter regimes in patients with normalgastrointestinal function fed with polymeric diets byconstant gravity infusion are unnecessary (Keohaneet al., 1984a). Full strength regimes can be prescribedfrom the outset, and diarrhoea, which did occur insome 10-15% of cases, appeared related solely toconcomitant oral or parenteral antibiotic therapy(Keohane et al., 1984).

Complications of enteral nutrition (Table 9)

Tube-related problems

Complications associated with larger-bore Ryle'stubes (oesophageal erosions, haemorrhage and stric-tures) have not been reported with the fine borenasogastric tubes. The fine bore tubes can be passedinto the trachea rather than into the oesophagus,especially in comatose patients, so care must be takento ensure correct positioning before enteric feeding isstarted. A common problem with the narrow bore

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

Enteral nutrition 789

feeding tube is the ease with which the tube rides upinto the oesophagus, or is removed completely by thepatient; a significant number of patients on entericfeeding regimens require more than one intubation(Keohane et al., 1983c). Intravenous administrationof an enteric feed has been reported but thiscomplication should not occur if feeding tubes andgiving sets are used with reversed Luers.

Gastrointestinal side effectsAs mentioned above, diarrhoea, abdominal disten-

tion and abdominal pain are common side effects,occurring at some stage during the treatment in up to25% of patients receiving enteral nutrition (Jones etal., 1983). As the most likely cause of these problemsis concomitant antibiotic therapy (Keohane et al.,1984), antibiotics should be stopped if clinicallypossible. If not, symptomatic treatment with loper-amide or codeine phosphate, if diarrhoea is thepredominant problem, is often effective, and enteralnutrition rarely has to be stopped on account of theseproblems.

Metabolic complications

Hyperglycaemia may occur during enteric feeding,caused by excessive sugar intake or insulin resistanceassociated with trauma and injury. Initially, frequenturine testing for sugar should be performed as well asblood glucose measurement. Electrolyte abnormali-ties, particularly hypokalaemia, commonly occurduring enteral nutrition and are related not only tothe feeding regimen, but also to the underlyingmedical or surgical disorder. Low blood levels ofcalcium, magnesium, zinc and phosphate can alsooccur, and supplements may be required.

Abnormalities of liverfunctionAbnormal liver function tests have been reported

in patients receiving enteral as well as parenteralnutrition, though earlier reports that abnormalities ofliver function inevitably occur have not been con-firmed. The aetiology of these changes is uncertain.Changes include elevations of alkaline phosphatase,y-glutamyl transpeptidase and the hepatocellularenzymes. It is common for minor and intermittentelevations of hepatocellular enzymes to occur andlittle significance need be attached to this. Occasion-ally a marked and persistent elevation of alkalinephosphatase has occurred. In our experience thehighest levels up to 48 KA units (1000 iu/litre) havebeen observed before the start of therapy, andhistology shows that this is due to marked fattyinfiltration, presumably related to associated protein-calorie malnutrition. All tests usually return tonormal when enteral nutrition is terminated and we

have never had to discontinue enteric feeding onaccount of any of the above changes.

ConclusionsA full clinical, biochemical, haematological and

immunological assessment of nutritional statusshould be performed before any form of nutritionalsupport is instituted and baseline values obtained forall other haematological and biochemical measure-ments. If electrolyte or other abnormalities occur,measurements should be made at frequent intervals.Careful monitoring is important to ensure earlyidentification of possible complications of enteral andparenteral nutrition.

ReferencesBAKER, J.P., DETSKY, A.S., WESSON, D.E., WOLMAN, S.L.,STEWART, S. WHITEWELL, J., LANGER, N. & JEEJEEBHOY, K.N.(1982). Nutritional assessment. A comparison of clinical judge-ment and objective measurements. New England Journal ofMedicine, 306, 169.

BASTOW, M.D., ALLISON, S.P. & GREAVES, P. (1981) Study ofmicrobial contamination of nasogastric feeds. In: Proceedings ofthe 3rd European Congress on Parenteral and Enteral Nutrition.p. 75.

BRISTIAN, B.R., BLACKBURN, G.L., HALLOWELL, E. & HEDDLE, R.(1974) Protein status of general surgical patients. Journal of theAmerican Medical Association, 230, 858.

CASEWELL, M.W. (1979) Nasogastric feeds as a source of klebsiellainfection for intensive care patients. Research and Clinical Forums1, 101.

COLLINS, J.P., MCCARTHY, I.D. & HULL, G.L. (1979) Assessment ofprotein nutrition in surgical patients-the value of anthropome-trics. American Journal of Clinical Nutrition 32, 1527.

DOBBIE, R.P. & BUTTERICK, O.D. (1977) Continuous pump/tubeenteric hyperalimentation-use in oesophageal disease. Journal ofParenteral and Enteral Nutrition, 1, 100.

HILL, G.L., BLACKETT, R.L., PICKFORT, I. & YOUNG, G.A. (1977)Malnutrition in surgical patients. An unrecognized problem.Lancet, i, 689.

HILL, G.L. (1979) Do malnourished patients need nutritionaltherapy before major surgery? Medical Journal ofAustralia, 2, 464.

INGENBLEEK, Y., VAN DEN SHRIEK H.G., DE NAYER P. & DEBISSCHER, M. (1975) Albumin, transferrin and thyroxine bindingprealbumin/retinol binding protein (TBPA/RBP) complex inassessment of malnutrition. Clinica Chimica Acta, 63, 61.

JAMES, W.P.T. (1982) Assessment of nutritional status. MedicineInternational, 15, 663.

JONES, F. AVERY (1983a) In: Nutrition in Medical Education Reportof the British Nutrition Foundation's Task Force on ClinicalNutrition. Ed. J. Gray. June 1983.

JONES, F. AVERY (1983b) In: Nutritional support in hospitalpractice.Blackwell Scientific Publications, Oxford, London etc. p vi.

JONES, B.J.M., PAYNE, S. & SILK, D.B.A. (1980) Indications forpump assisted enteral feeding. Lancet, i, 1057.

JONES, B.J.M., BROWN, B.E., GRIMBLE, G.K. & SILK, D.B.A. (1981)The formulation of energy dense enteral feeds-the use of highmolecular weight glucose polymers. In: Proceedings 3rd EuropeanCongress on Parenteral and Enteral Nutrition, p. 75.

JONES, B.J.M., LEES, R., ANDREWS, J., FROST, P. & SILK, D.B.A.(1983) Comparison of an elemental and polymeric enteral diet inpatients with normal gastrointestinal function. Gut, 24, 78.

JONES, D.C., RICH, A.J., WRIGHT, P.D. & JOHNSON, I.P.A. (1980)Comparison of proprietary elemental and whole-protein diets in

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from

790 D. B. A. Silk

unconscious patients with head injury. British Medical Journal, 1,1493.

KEIGHLEY, M.R.S., MOGG, B., BENTLEY, S. & ALLAN, C. (1983)'Home brew' compared with commercial preparations for enteralfeeding. British Medical Journal, 1, 163.

KEOHANE, P.P., ATTRILL, H., JONEs, B.J.M., BROWN, I., FROST, P. &SILK, D.B.A. (1983) Influence of lactose and Cl. difficile in thepathogeneisis of enteral feeding associated diarrhoea. ClinicalNutrition, 1, 259.

KEOHANE, P.P., ATTRILL, H. & SILK, D.B.A. (1982) Endoscopicplacement of fine bore nasogastric and nasoenteric feeding tubes.Clinical Nutrition, 1, 245.

KEOHANE, P.P., ATTRILL, H. & SILK, D.B.A. (1983) Clinicalindications for weighted enteral feeding tubes. Clinical Nutrition,2, 27.

KEOHANE, P.P., ATTRILL, H., JONES, B.J.M. & SILK, D.B.A. (1983)Limitations and drawbacks of 'fine bore' nasogastric feedingtubes. Clinical Nutrition, 2, 85.

KEOHANE, P.P., ATTRILL, H., LOVE, M., FROST, P. & SILK, D.B.A.(1984) Relation between osmolality of diet and gastrointestinalside effects in enteral nutrition. British Medical Journal, 288, 678.

KORETZ, R.L. & MEYER, J.H. (1980) Elemental diets-facts andfantasies. Gastroenterology, 78, 393.

LAW, D.K., DUDRICK, S.J. & ABDON, N.I. (1973) Immunocompet-ence of patients with protein calorie malnutrition. Effects ofnutritional repletion. Annals of Internal Medicine, 79, 545.

MEAKINS, J.L., PIETSCH, J.B., BUBENICK, O., KELLY, R., RODE, H.,GORDON, J. & MACLEAN, L.B. (1977) Delayed hypersensitivity:indicator of acquired failure of host defences in sepsis and trauma.Annals of Surgery, 186, 241.

METZ, G., DILAWARI, J. & KELLOCK, T.D. (1978) Simple techniquefor naso-enteric feeding. Lancet, ii, 454.

MUNRO, H. (1978) Biological limiting factors to parenteral aminoacid feeding. In: Advances in Parenteral Nutrition 1978 ed. I. D. A.Johnston, MTP Press. Lancaster p. 107.

PIETSCH, J.B., MEAKINS, J.L. & MACLEAN, L.D. (1977) The delayedhypersensitivity response: application in clinical surgery. Surgery,82, 349.

RUSSELL, R.I. (1975) Elemental diets. Gut, 16, 68.SAGAR, S., HARLAND, P. & SHIELDS, R. (1979) Early postoperative

feeding with elemental diet. British Medical Journal, 1, 293.SHETTY, P.S., WATRASIEWICZ, K.E., JUNG, R.T. & JAMES, W.P.T.

(1979) Rapid-turnover transport proteins: an index of subclinicalprotein energy malnutrition. Lancet, ii, 230.

SILK, D.B.A. (1980) Enteral nutrition. Hospital Update, 6, 761.SILK, D.B.A. (1983) Nutritional support in hospital practice.

Blackwell Scientific Publications, Oxford.SPANIER, A.H., PIETSCH, J.B., MEAKINS, J.L., MACLEAN, L.D. &SHIZGAL, H.M. (1976) Relationship between immune competenceand nutrition. Surgical Forum, 27, 332.

WOOLFSON, A.M.J. (1978) Metabolic considerations in nutritionalsupport. Proceedings of Symp. Roy Coll Phys. Research andClinical Forum. 1, 35.

YEUNG, C.K., YOUNG, G.A., HACKETT, A.F. & HILL, G.L. (1979)Fine needle catheter jejunostomy-an assessment of a newmethod of nutritional support after major gastrointestinal surgery.British Journal of Surgery, 66, 727.

by copyright. on D

ecember 6, 2020 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.60.709.779 on 1 Novem

ber 1984. Dow

nloaded from