Nutrition after stroke - Semantic Scholar · Stroke is the better validated prognostic indicator. Based on these studies, there appears to be an association between malnutrition and

Nutrition after stroke

Martin DennisDepartment of Clinical Neurosctences, University of Edinburgh, Edinburgh, UK

Decisions about feeding are amongst the most difficult to face those managingstroke patients. About a fifth of patients with acute stroke are malnourished onadmission to hospital. Moreover, patients' nutritional status often deterioratesthereafter because of increased metabolic demands which cannot be met due tofeeding difficulties. Poor nutritional intake may result from, (i) reducedconscious level; (ii) an unsafe swallow (lii) arm or facial weakness; (iv) poormobility; or (v) ill fitting dentures. Malnutrition is associated with poorersurvival and functional outcomes, although these associations may not becausal. Patients often receive support with oral supplements or enteral tubefeeding via nasogastric or percutaneous endoscopic gastrostomy. Althoughthese probably improve nutritional parameters, it is unclear whether theyimprove patients' outcomes. Also the optimal timing, type and method ofenteral feeding is uncertain. Large randomised trials are now in progress toidentify the optimum feeding policies for stroke patients.

The influence of nutrition on the risk of stroke has been the subject ofmuch research. In contrast, few studies of nutrition after stroke,especially that during the first few days and weeks, have been reported.This is surprising because feeding problems are amongst the mostcommon and difficult management issues which confront the cliniciancaring for stroke patients.

This review aims to define what is known, what is not known andareas where more research is needed. It will not address the issue ofdietary modification in secondary prevention nor provide a detailedaccount of swallowing problems and their assessment in stroke patientswhich have been covered elsewhere1. This review will focus on severalclinically important questions including:

1 How can malnutrition be identified?

correspondence to 2 How common is malnutrition after a stroke?Dr Martin Dennis,

Department of clinical 3 Which patients are likely to have malnutrition?Neurosciences, Western

General Hospital, 4 Does malnutrition matter after a stroke?Crewe Road,

Edinburgh EH4 2xu, UK 5 Will nutritional support improve the patient's outcome?

British Medical Bulletin 2000, 56 (No 2) 466-475 C The British Courral 2000

Nutrition after stroke

6 How should stroke patients be fed?

7 When should tube feeding start after stroke?

8 Is feeding via a percutaneous endoscopic gastrostomy (PEG) better thanthat via a nasogastnc (NG) tube?

How can malnutrition be identified?

In routine clinical practice there are practical difficulties in assessingstroke patients' nutritional status. A dietary and weight history may notbe available because of patients' communication problems and analternative source of this information may not be available if, as iscommon, the patient lives alone. Simple assessments of weight and heightto estimate the body mass index (BMI) pose problems in immobile strokepatients. Specialised equipment, of limited availability, such as weighingbeds or scales which accommodate wheelchairs, may be required andheight may need to be estimated from the patient's demi-span or heel-kneelength. More complex anthropometric measures, e.g. mid-armcircumference (MAC) and triceps skin-fold thickness (TFT) which allowthe mid-arm muscle circumference (MAMC) to be calculated, require notonly a tape measure and skin-fold callipers but training for the assessor toobtain reproducible measures. Anthropometric measures may also changebecause of paralysis of the arm after stroke. Laboratory parameters suchas haemoglobin, serum protein, albumen and transferrin, are readilyavailable but low levels occur in many conditions and do not necessarilyreflect nutritional status. Indeed, in any acute illness, the serum albumentends to fall due to increased catabolism and preferential production ofacute phase proteins. More specialised measures such as vitaminestimations, antigen skin testing and bioelectric impedance (the latter toestimate body fat mass, body lean mass, body cell mass and total bodywater) are used in research, but are not widely available and are not suitedto routine clinical practice. An awareness of the possibility of malnutritionis a key factor in identifying malnourished patients. A simple end of thebed assessment reliably identifies most stroke patients with low BMI andabnormal anthropometry2. Estimation of the BMI, serial weights toidentify weight loss and monitoring of dietary intake could be used toscreen patients on admission and monitor patients on stroke unitsalthough no specific assessment tool has been developed and tested for usein stroke patients3. Some assessment of patients' nutritional status shouldbe routinely applied on admission to the stroke unit and periodicallythereafter.

British Medical Bulletin 2000;56 (No 2) 467

Stroke

How common is malnutrition after a stroke?

Malnutrition is a common and often unrecognised problem in patients,especially the elderly, admitted to hospital. Those who remain in hospitalfor prolonged periods are also at risk4. Inevitably, the reported frequencyof malnutrition after stroke has varied dependmg on patient selection, thedefinitions of malnutrition and the method and timing of assessments.Table 1 shows the various estimates of the frequency of malnutrition onadmission to hospital after an acute stroke. Studies have varied in thenumber of nutritional parameters measured, their reference ranges andthe number of abnormal results "required to categorise patients asmalnourished. Most have focused on undernutntion, but overnutritionwith obesity is probably more common in Western countries and posespractical difficulties for patients9. Transfers, walking, continence and skinhygiene may all be compromised by obesity.

Several studies have shown that stroke patients' nutritional status mayworsen during hospital admission5"8-11. However, these rely on groupeddata where estimates of nutritional status later in the admission excludethose who have died or have already been discharged. Few studies haveprovided serial measurements in surviving patients, but those that haveinevitably show that some patients improve, some deteriorate and someremain stable with respect to nutritional indices5.

Which patients are likely to have malnutrition?The factors which have been associated with malnutrition on admissionto hospital are shown in Table 1, but few conclusions can be drawn.

Table 1 Estimates of the frequency of malnutrition in various studies

Study

Axelsson5

Unosson'Davalos7

Ganballa1

Choi-Kwon»

Finestone10

n atbaseline

100

50104

20188

49

Type of patients

Acute admissions

Acute admissionsAcute admissions

All acute admissionsAcute females only(highly selected)Rehabilitation only

n (%) withlow albumen

23 (23%)

31 (62%) < 36g/l8(8%)

38 (19%) < 35g/1

n/a

n (%) classifiedas malnourishedand criteria

16 (16%) > 2 lowvalues

4 (8%) > 2 low values17 (16%) either lowalbumen or TSF/MAMCn/a30 (34%) > 2 low

Mild 7 (14%)Moderate 9 (19%)Severe 8 (16%)> 1 low level

Factors associatedwith malnutritionon admission

Increased age*Females*Prior peptic ulcer*Atnal f ibri l lation**n/an/a

n/a

Haemorrhagic stroke

Dysphagia*Diabetes**Previous stroke**

468 British Medical Bulletin 20OO;56 (No 2)

Nutrition after stroke

Drawing on the non-stroke literature, one would expect malnutrition tobe more frequent in older patients, those living in institutions and poorsocial circumstances, those with prior cognitive impairment, physicaldisability or gastrointestinal disease. Stroke, like any acute illness, maylead to a negative energy balance and greater nutritional demands, butstroke patients may be less able to meet these increased demands12.Davalos et aV showed that patients with severe strokes have a greaterstress response (based on cortisol levels) than those with milder strokesand that this was associated with a more marked deterioration innutritional status. Complications such as infections which increase thepatients' metabolic demands are associated with deteriorating nutritionalstatus5. To compound the general problem of malnutrition, it has beenestimated that up to 50% of hospitalised stroke patients are unable toswallow safely, although again the reported frequency depends on theselection of cases, the timing of assessments and the sensitivity of themethod used to detect swallowing problems11'13'14. In most studies,deterioration in nutrition occurred more often in dysphagic patients7-11

or in those who need help with feeding6. Even patients who are capableof swallowing liquids and food may have a poor appetite because of theeffects of intercurrent illness or medication and they may eat moreslowly or be less keen to eat because of facial weakness, lack of denturesor poor arm function15.

Does malnutrition matter after stroke?

Poor nutrition, although not specifically in stroke patients, has beenassociated with reduced muscle strength, reduced resistance to infectionand impaired wound healing1617. Among patients with stroke, most ofwhom are elderly, muscle weakness, infections and pressure sores arecommon and account for significant mortality and morbidity18. It isplausible that malnutrition could increase the frequency of theseproblems and result in poorer outcomes. Davalos et aPshowed thatmalnutrition after the first week of admission was associated with anincreased risk of a poor outcome (dead or Barthel index <50) at 1month, death, a greater frequency of infections and pressure sores, andlonger length of stay. However, these associations were not statisticallysignificant after adequate adjustment for stroke severity, using theCanadian Stroke Scale. More recently, Gariballa et aP showed thatserum albumen, which may reflect nutritional status, predicted post-stroke survival, but did not adequately adjust for baseline stroke severityin their Cox proportional hazards' model. Although the ModifiedRankin and Orpington Prognostic score at baseline was collected, theauthors only included the former in their model, even though the latter

British Medical BulleVn 2000.56 (No 2) 469

Stroke

is the better validated prognostic indicator. Based on these studies, thereappears to be an association between malnutrition and poor outcome,but this is not necessarily independent of other prognostic factors andmay not be causal.

Will nutritional support improve the patient's outcome?

Evidence of a causal relationship between malnutrition and poor outcomecould come from intervention studies if improving nutrition resulted inbetter outcomes. There have been a large number of randomised trialstesting the effects of providing protein calorie supplementation to diversegroups of patients. These studies have been individually too small toreliably demonstrate an effect on their own, but a recent systematic reviewof all of the available trials suggested that oral or enteral (i.e. via a feedingtube) nutritional supplementation definitely improves nutritionalparameters and may reduce the odds of death (odds ratio = 0.66; 95% CI0.48-0.91 )19. However, this review included trials of differingmethodological quality and when only more rigorous studies wereincluded in the analysis the effect was statistically non significant (oddsratio = 0.81; 95% CI 0.44-1.50). None of these studies were specificallyfor stroke patients and few stroke patients were included. One smallrandomised trial (n = 42) has suggested that oral supplementation afterstroke improves nutritional parameters20 and a retrospective non-randomised study showed that early enteral nutrition after stroke reducedlength of stay in hospital21. Thus, nutritional support probably improvesnutritional parameters, but it is unclear whether this leads to improvedclinical outcomes.

How should stroke patients be fed?

Even if there is little evidence to support nutritional supplementation, it isobvious that stroke patients require feeding. Ideally, patients would eatnormally, but this is not possible for the important minority of patientswho cannot swallow safely. Patients with swallowing difficulties areusually put 'nil by mouth' and given parenteral fluids believing (but withlittle scientific justification) that this will reduce the risk of aspirationpneumonia. Of course, these patients must still cope with their saliva.Patients' swallowing usually recovers over the first few days or weeks toan extent which allows most patients to safely take fluids and food, ifnecessary with a modified consistency13-14'22^3. Indeed, many patients canswallow safely if carefully positioned, given food and fluids of appropriateconsistency and using a variety of compensatory strategies.

470 British Medical Bulletin 2000;56 (No 2)

Nutrition after stroke

It is unclear how we should best support patients' nutritional statusduring the period when their oral intake is inadequate. How long is itreasonable to wait before starting feeding and what is the best route?Intravenous feeding can be used, but in practice is rarely justified in strokepatients who are able to absorb nutrients from their gut. Peripheral totalparenteral nutrition (TPN) offers a less invasive option than that deliveredvia central venous line and may become more widely used where enteralroutes are impractical. In most settings, the choice of support lies betweenenteral feeding via a nasogastric (NG) or percutaneous endoscopicgastrostomy (PEG) tube (Fig. 1) or one of the closely related alternatives,e.g. a radiologically-guided gastrostomy or a jejunostomy.

When should tube feeding start after stroke?

Whilst the patient cannot swallow adequate food, their nutritional statuswill inevitably deteriorate unless supported. If tube feeding was welltolerated and carried no hazard, then one would lose nothing by startingearly. However, patients find tube feeding uncomfortable and it carries arisk of complications which have to be set against the benefits. By fillingthe patient's stomach, enteral feeding inevitably increases the risk ofaspiration in patients who do not adequately protect their airway. Intheory, early feeding might be associated with metabolic changes {e.g.hyperglycaemia) which could be detrimental to the ischaemic penumbra24.The balance of risk and benefit will vary depending on the nutritionalstatus of the patient and whether they are taking any food orally. The riskassociated with tube feeding, in turn, will depend on the method used, itsduration and local factors {e.g. complication rates associated with PEGinsertion). Some clinicians prefer to introduce tube feeding very soon afterthe stroke, others delay for days and sometimes weeks. There are nocompleted large randomised trials to guide our use of enteral feeding.

Is feeding via a PEG better than that via an NG tube?

Nasogastric (NG) tubes are often inserted to allow fluid and food to begiven to patients. However, in patients who are unable to swallow, they arenot always easy to insert and they are often pulled out by patients and haveto be replaced, because they are uncomfortable. This adds to patients'distress, may necessitate a further X-ray to check position and interruptsany feeding regimen. NG tubes, especially the fine bore variety, maybecome displaced and cause aspiration. If left in situ for prolonged periods,ulceration of the nostril, oesophageal strictures and oesophagotrachealfistulae have been described. Because of these problems, some workers

British Medical Bulletin 2000,56 Wo 2) 471

Stroke

Fig. 1 A diagramshowing a

percutaneousendoscopic

gastrostomy (PEG)tube in situ.

Y connector

advocate the increased and earlier use of PEG tubes. This technique, whichcan be performed with little or no sedation, provides an effective and quiteacceptable method of enteral feeding (Fig. 1). However, any advantages ofPEG over NG tubes, as far as improved delivery of nutrition is concerned,have to be carefully weighed against their relative complication rates.Unfortunately, despite the frequency of their use, there are few data con-cerning the complication rates associated with NG tubes in stroke patients.In contrast, literally hundreds of series have been published reporting theexperience with PEG tubes, although relatively few specifically in strokepatients. A systematic review of a large number (but not all) of thesestudies25 suggested that there was a 0.3% risk of death related to theprocedure itself and a 10% risk of major complications (Table 2).However, these figures are likely to underestimate the risks in strokepatients because: (i) specialist centres which achieve better results are morelikely to publish than those with less interest or higher complication rates;(ii) many studies are retrospective and rely on routine recording ofcomplications; and (iii) stroke patients who tend to be elderly and frail mayhave higher complication rates.

Table 2 Reported frequency of major complications afterpercutaneous endoscopic gastrostomy (from Wollman25)

Wound relatedAbscessSepticaemiaNecrotising faciitis

Aspiration pneumoniaPeritonitisOtherGl

PerforationGastro-colic fistulaHaemorrhage

Dislodged tube requiring repeat procedure

3 3%

2 1%0 5%2 4%

0 9 %

472 British Medical Bulletin 2000,56 (No 2)

Nutrition after stroke

Recently, several reports of the complication rates of PEG insertionamongst stroke patients have been published26"30. These included a total of310 patients of whom about 8% died within a week or two of theprocedure and 25% died during the hospital admission. The rates ofvarious complications during variable follow-up periods were: aspirationpneumonia 19%, tube blockage/breakage or removal 11%, woundinfection 8%, gastrointestinal haemorrhage 0.6% (one fatal), and fatalperforation 0.3%. Unfortunately, such studies are of limited value mguiding practice and indirect comparisons of complication rates betweenNG and PEG tubes are bound to be unreliable.

There have been only three small randomised comparisons of NG andPEG tube feeding. These suggested that the latter provided more effectivenutritional support with less interruption of feeding30"32. One of thesetrials30 was in severe stroke patients and showed that those fed by PEG hadan implausibly large (70% relative) reduction in case fatality compared withthose fed via NG tube. However, this trial only included 30 patients andlittle data were provided to allow an assessment of the effectiveness ofrandomisation in achieving balanced groups. It seems most likely that someimbalance in baseline factors accounted for much of the observed differencein outcome. Thus, the relative merits of the two types of tube are uncertain,at least in the first month or so after the stroke. There is little doubt that aPEG tube is a better option if feeding is to be prolonged. Also, in practice,there may be no alternative to a PEG tube if feeding is required andnasogastnc feeding has been unsuccessful.

The need for more research

A survey of almost 3000 physicians who manage stroke in the UKdemonstrated wide variation in the use of oral supplements and in thetiming and method of feeding in dysphagic stroke patients33. Suchvariation reflects the lack of clear evidence to guide practice. There isclearly a need for large randomised trials to establish how best to feedpatients after a stroke. Trials are needed to address several importantquestions including:

1 Should patients who can take adequate fluids orally routinely receivenutritional supplements to improve their outcome? If not are thereparticular groups of patients who should?

2 In patients who are unable to take adequate fluid and food orallyimmediately after the stroke, should we start tube feeding early or wait fora few days to allow their swallowing to improve?

3 If tube feeding is required, is feeding via a percutaneous endoscopic gastro-stomy (PEG) superior to that via the traditional nasogastnc tube (NG)?

British Medical Bulletin 20OO;56 (No 2) 4 7 3

Stroke

The FOOD trial (Feed Or Ordinary Diet; www.dcn.ed.ac.uk/food) isan on-going multicentre international randomised trial which aims toaddress these and other important questions relating to nutrition afterstroke.

Ethical considerations

References

Feeding is regarded by some as a basic component of care and by othersas a medical intervention. Decisions about whether to feed, when to feedand how to feed stroke patients are among the most difficult to confrontthe professions involved in their care34-35. Many patients who are unableto eat normally are likely to have a poor functional outcome and perhapsa quality of life which some would judge to be worse than death. It is farfrom clear whether judgements regarding the quality of life in a dependentstate made by the person before their stroke, or by their relatives orinvolved professionals are valid in making decisions about starting orcontinuing nutritional support. Unfortunately, most patients in whom thisissue arises are unable to communicate their own wishes. Relatives may ormay not be able to speak on their behalf. Nurses, doctors and therapistswill all have an opinion, but these may not converge and may differ fromthose of the family. Moreover, the lack of reliable information about thebenefits and risks of feeding techniques adds further uncertainty todecision making. Because of these difficulties, it is essential that issuessurrounding nutritional support are discussed openly with all concerned,and that lines of communication are kept open so that the best decisioncan be made to minimize potential conflict.

1 O'Neill PA. Swallowing and prevention of complications Br Med Bull 2000, 56 457-652 Mead GE, Donaldson L, North P, Dennis MS. An informal assessment of nutritional status in acute

stroke for use in an international multicentre trial of feeding regimens. Int ] Ckn Pract 1998; 52:316-8

3 Nightingale JMD, Walsh N, Bullock ME, Wicks AC. Three simple methods of detectingmalnutrition on medical wards / R Soc Med 1996; 89: 144-8

4 Sullivan DH, Sun S, Walls RC. Protein-energy undernutntion among elderly hospitalized patientsa prospective study. JAMA 1999; 281: 2013-9

5 Axelsson K, Asplund K, Norberg A, Alafuzoff I Nutritional status in patients with acute stroke.Ada Med Scand 1988; 224: 217-24

6 Unosson M, Ek AC, Bjurulf P, von Schenck H, Larsson J Feeding dependence and nutritionalstatus after acute stroke. Stroke 1994, 25 366-71

7 Davalos A, Ricart W, Gonzalez-Huix F et al Effect of malnutrition after acute stroke on clinicaloutcome Stroke 1996, 27: 1028-32

8 Ganballa SE, Parker SG, Taub N, Castleden CM Influence of nutritional status on clinicaloutcome after stroke Am ] Clm Nutr 1998; 68: 275-81

9 Qioi-Kwon S, Yang YH, Kim EK, Jeon MY, Kim JS Nutritional status in acute stroke undemutrraonversus overnutnnon in different stroke subtypes. Acta Neurol Scand 1998; 98: 187-92

474 British Medial Bulletin 2OOO;56 (No 2)

Nutrition after stroke

10 Finestone HM, Greene-Finestone LS, Wilson ES, Teasell RW. Malnutntion in stroke patients onthe rehabilitation service and at follow up. prevalence and predictors. Arch Phys Med Rehabil1995; 76: 310-6

11 Smithard DG, O'Neill PA, Park C et al. Complications and outcome after acute stroke. Doesdysphagia matter' Stroke 1996: 27: 1200-4

12 Klipstein-Grobusch K, Redly JJ, Potter J, Edwards CA, Roberts MA. Energy intake andexpenditure in elderly patients admitted to hospital witli acute illness BrJ Nutr 1995, 73. 323-34

13 Gordon C, Langton Hewer R, Wade DT. Dysphagia in acute stroke. BM] 1987; 295: 411-1414 Barer DH. The natural history and functional consequences of dysphagia after hemispheric stroke.

/ Neurol Neurosurg Psychiatry 1989; 52: 236-4115 Carr EK, Hawthorn PJ. Lip function and eating after a stroke: a nursing perspective. / Adv Nurs

1988,13- 447-5116 Fiatarone MA, Evans WJ. The etiology and reversibility of muscle dysfunction in the aged. /

Gerontol 1993; 48: 77-8317 Chandra RK Graying of the immune system. Can nutrient supplements improve immunity in the

elderly' JAMA 1997; 277- 1398-918 Davenport RJ, Dennis MS, Wellwood I, Warlow CP. Complications following acute stroke. Stroke

1996; 27: 415-2019 Potter JN, Langhorne P, Roberts M. Routine protein energy supplementation in adults: systematic

review BM] 1998, 317. 495-50120 Ganballa SE, Parker SG, Castleden CM. A randomised controlled trial of nutritional

supplementation after stroke. Age Ageing 1998; 27 (Suppl 1): 6621 Nyswonger GD, Helmchen RH. Early enteral nutrition and length of stay in stroke patients. /

Neurosa Nurs 1992; 24: 220-322 Smithard D, O'Neill P, England RE et al. The natural history of dysphagia following a stroke.

Dysphagia 1997; 12: 188-9323 Mann G, Dip PG, Hankey GJ, Cameron D Swallowing function after stroke. Prognosis and

prognostic factors at 6 months. Stroke 1999, 30: 744-824 Scott JF, Gray CS, O'Connell JE, Albem KGMM. Glucose and insulin therapy in acute stroke; why

delay further? Q ] Med 1998; 91: 511-525 Wollman B, D'Agostino HB, Walus-Wigle JR, Easter DW, Beale A. Radiologic, endoscopic and

surgical gastrostomy: an institutional evaluation and a meta-analysis of the literature. Radiology1995; 197: 699-704

26 Wanklyn P, Cox N, Belheld P. Outcome in patients who require a gastrostomy after stroke AgeAgeing 1995; 24: 510-4

27 Wijdicks EFM, McMahon MM. Percutaneous endoscopic gastrostomy after acute stroke:complications and outcome. Cerebrovasc Dis 1999; 9: 109-11

28 James A, Kapur K, Hawthorne AB. Long-term outcome of percutaneous endoscopic gastrostomyfeeding m patients with dysphagic stroke. Age Ageing 1998; 27: 671-6

29 Skelly R, Terry H, Millar E, Cohen D. Outcomes of percutaneous endoscopic gastrostomy feeding.Age Ageing 1999; 28: 416

30 Norton B, Homer-Ward M, Donnelly MT, Long RG, Holmes GKT. A randomised prospectivecomparison of percutaneous endoscopic gastrostomy and nasogastnc tube feeding after acutedysphagic stroke. BM] 1996; 312. 13-6

31 Park RH, Allison MC, Lang J et al. Randomised comparison of percutaneous endoscopicgastrostomy and nasogastnc tube feeding in patients with persisting neurological dysphagia. BM]1992; 304: 1406-9

32 Baeten C, Hoefnagels J. Feeding via nasogastnc tube or percutaneous endoscopic gastrostomy.A companson. Scand] Gasterenterol 1992; 27 (Suppl 194). 95-8

33 Ebrahim S, Redfern J. Stroke Care - A matter of chance. A national survey of stroke services.London: Stroke Association, 1999

34 Rabeneck L, McCullough LB, Wray NP. Ethically justified, clinically comprehensive guidelinesfor percutaneous endoscopic tube placement. Lancet 1997; 349: 496-8

35 Macfie J. Ethics and nutritional support. Wien Klin Wochenschr 1997; 109- 850-7

British Medical Bulletin 2000;56 (No 2) 475