Interventions for dysphagia and nutritional support in acute and subacute stroke (Review)

Transcription

1 Interventions for dysphagia and nutritional support in acute and subacute stroke (Review) Geeganage C, Beavan J, Ellender S, Bath PMW This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2012, Issue 10

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY BACKGROUND OBJECTIVES METHODS RESULTS Figure Figure DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES Analysis 1.1. Comparison 1 Swallowing therapy, Outcome 1 Case fatality at end of trial Analysis 1.2. Comparison 1 Swallowing therapy, Outcome 2 Death or dependency at end of trial Analysis 1.3. Comparison 1 Swallowing therapy, Outcome 3 Institutionalisation Analysis 1.4. Comparison 1 Swallowing therapy, Outcome 4 Length of stay (days) Analysis 1.5. Comparison 1 Swallowing therapy, Outcome 5 Chest infection or pneumonia Analysis 1.6. Comparison 1 Swallowing therapy, Outcome 6 Dysphagia at end of trial Analysis 1.7. Comparison 1 Swallowing therapy, Outcome 7 Pharyngeal transit time (seconds) Analysis 1.8. Comparison 1 Swallowing therapy, Outcome 8 Swallow score Analysis 1.9. Comparison 1 Swallowing therapy, Outcome 9 Nutritional (albumin) Analysis 2.1. Comparison 2 Route of feeding, Outcome 1 Case fatality at end of trial Analysis 2.2. Comparison 2 Route of feeding, Outcome 2 Death or dependency at end of trial Analysis 2.3. Comparison 2 Route of feeding, Outcome 3 Institutionalisation Analysis 2.4. Comparison 2 Route of feeding, Outcome 4 Length of stay in hospital (days) Analysis 2.5. Comparison 2 Route of feeding, Outcome 5 Pressure sores Analysis 2.6. Comparison 2 Route of feeding, Outcome 6 Chest infection or pneumonia Analysis 2.7. Comparison 2 Route of feeding, Outcome 7 Dysphagia at end of trial Analysis 2.8. Comparison 2 Route of feeding, Outcome 8 Treatment failure Analysis 2.9. Comparison 2 Route of feeding, Outcome 9 Gastrointestinal bleeding Analysis Comparison 2 Route of feeding, Outcome 10 Feed delivery (%) Analysis Comparison 2 Route of feeding, Outcome 11 Weight at end of trial (last value carried forward) (kg) Analysis Comparison 2 Route of feeding, Outcome 12 Mid-arm circumference (last value carried forward) (cm). 116 Analysis Comparison 2 Route of feeding, Outcome 13 Albumin (last value carried forward) (g/l) Analysis 3.1. Comparison 3 Timing of feeding, Outcome 1 Case fatality at end of trial Analysis 3.2. Comparison 3 Timing of feeding, Outcome 2 Death or disabled at end of trial Analysis 3.3. Comparison 3 Timing of feeding, Outcome 3 Institutionalisation Analysis 4.1. Comparison 4 Fluid supplementation, Outcome 1 Time to resolution of dysphagia (days) Analysis 5.1. Comparison 5 Nutritional supplementation, Outcome 1 Case fatality at end of trial Analysis 5.2. Comparison 5 Nutritional supplementation, Outcome 2 Death or dependency at end of trial Analysis 5.3. Comparison 5 Nutritional supplementation, Outcome 3 Institutionalisation Analysis 5.4. Comparison 5 Nutritional supplementation, Outcome 4 Length of stay in hospital (days) Analysis 5.5. Comparison 5 Nutritional supplementation, Outcome 5 Pressure sores Analysis 5.6. Comparison 5 Nutritional supplementation, Outcome 6 Energy intake (kcal/day) Analysis 5.7. Comparison 5 Nutritional supplementation, Outcome 7 Protein intake (g/day) Analysis 5.8. Comparison 5 Nutritional supplementation, Outcome 8 Albumin (last value carried forward) APPENDICES WHAT S NEW i

3 HISTORY CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST SOURCES OF SUPPORT DIFFERENCES BETWEEN PROTOCOL AND REVIEW INDEX TERMS ii

4 [Intervention Review] Interventions for dysphagia and nutritional support in acute and subacute stroke Chamila Geeganage 1, Jessica Beavan 2, Sharon Ellender 3, Philip MW Bath 3 1 Clinical Pharmacology and Pharmacy, Faculty of Medicine, University of Colombo, Colombo, Sri Lanka. 2 Department of Stroke Medicine, Royal Derby Hospital, Derby, UK. 3 Division of Stroke Medicine, University of Nottingham, Nottingham, UK Contact address: Philip MW Bath, Division of Stroke Medicine, University of Nottingham, Nottingham, NG5 1PB, UK. philip.bath@nottingham.ac.uk. Editorial group: Cochrane Stroke Group. Publication status and date: New search for studies and content updated (no change to conclusions), published in Issue 10, Review content assessed as up-to-date: 14 March Citation: Geeganage C, Beavan J, Ellender S, Bath PMW. Interventions for dysphagia and nutritional support in acute and subacute stroke. Cochrane Database of Systematic Reviews 2012, Issue 10. Art. No.: CD DOI: / CD pub2. Background A B S T R A C T Dysphagia (swallowing problems) are common after stroke and can cause chest infection and malnutrition. Dysphagic, and malnourished, stroke patients have a poorer outcome. Objectives To assess the effectiveness of interventions for the treatment of dysphagia (swallowing therapy), and nutritional and fluid supplementation, in patients with acute and subacute (within six months from onset) stroke. Search methods We searched the Cochrane Stroke Group Trials Register (February 2012), MEDLINE (1966 to July 2011), EMBASE (1980 to July 2011), CINAHL (1982 to July 2011) and Conference Proceedings Citation Index- Science (CPCI-S) (1990 to July 2011). We also searched the reference lists of relevant trials and review articles, searched Current Controlled Trials and contacted researchers (July 2011). For the previous version of this review we contacted the Royal College of Speech and Language Therapists and equipment manufacturers. Selection criteria Randomised controlled trials (RCTs) in dysphagic stroke patients, and nutritional supplementation in all stroke patients, where the stroke occurred within six months of enrolment. Data collection and analysis Two review authors independently applied the inclusion criteria, assessed trial quality, and extracted data, and resolved any disagreements through discussion with a third review author. We used random-effects models to calculate odds ratios (OR), 95% confidence intervals (95% CI), and mean differences (MD). The primary outcome was functional outcome (death or dependency, or death or disability) at the end of the trial. 1

5 Main results We included 33 studies involving 6779 participants. Swallowing therapy: acupuncture, drug therapy, neuromuscular electrical stimulation, pharyngeal electrical stimulation, physical stimulation (thermal, tactile), transcranial direct current stimulation, and transcranial magnetic stimulation each had no significant effect on case fatality or combined death or dependency. Dysphagia at end-of-trial was reduced by acupuncture (number of studies (t) = 4, numbers of participants (n) = 256; OR 0.24; 95% CI 0.13 to 0.46; P < ; I 2 = 0%) and behavioural interventions (t = 5; n = 423; OR 0.52; 95% CI 0.30 to 0.88; P = 0.01; I 2 = 22%). Route of feeding: percutaneous endoscopic gastrostomy (PEG) and nasogastric tube (NGT) feeding did not differ for case fatality or the composite outcome of death or dependency, but PEG was associated with fewer treatment failures (t = 3; n = 72; OR 0.09; 95% CI 0.01 to 0.51; P = 0.007; I 2 = 0%) and gastrointestinal bleeding (t = 1; n = 321; OR 0.25; 95% CI 0.09 to 0.69; P = 0.007), and higher feed delivery (t = 1; n = 30; MD 22.00; 95% CI to 27.85; P < ) and albumin concentration (t = 3; n = 63; MD 4.92 g/l; 95% CI 0.19 to 9.65; P = 0.04; I 2 = 58%). Although looped NGT versus conventional NGT feeding did not differ for end-of-trial case fatality or death or dependency, feed delivery was higher with looped NGT (t = 1; n = 104; MD 18.00%; 95% CI 6.66 to 29.34; P = 0.002). Timing of feeding: there was no difference for case fatality, or death or dependency, with early feeding as compared to late feeding. Fluid supplementation: there was no difference for case fatality, or death or dependency, with fluid supplementation. Nutritional supplementation: there was no difference for case fatality, or death or dependency, with nutritional supplementation. However, nutritional supplementation was associated with reduced pressure sores (t = 2; n = 4125; OR 0.56; 95% CI 0.32 to 0.96; P = 0.03; I 2 = 0%), and, by definition, increased energy intake (t = 3; n = 174; MD kcal/day; 95% CI to ; P = 0.003; I 2 = 91%) and protein intake (t = 3; n = 174; MD g/day; 95% CI 1.99 to 32.56; P = 0.03; I 2 = 92%). Authors conclusions There remains insufficient data on the effect of swallowing therapy, feeding, and nutritional and fluid supplementation on functional outcome and death in dysphagic patients with acute or subacute stroke. Behavioural interventions and acupuncture reduced dysphagia, and pharyngeal electrical stimulation reduced pharyngeal transit time. Compared with NGT feeding, PEG reduced treatment failures and gastrointestinal bleeding, and had higher feed delivery and albumin concentration. Nutritional supplementation was associated with reduced pressure sores, and increased energy and protein intake. P L A I N L A N G U A G E S U M M A R Y Interventions for problems with swallowing and poor nutrition in patients who have had a recent stroke Stroke is often complicated by problems with swallowing (dysphagia) and poor nutrition. Normal oral feeding in those with swallowing problems may lead to pneumonia and an increased risk of death. Therapies to improve swallowing are designed to accelerate recovery of swallowing function and reduce the risk of developing pneumonia. We reviewed 33 studies involving 6779 patients (the average age of patients across the studies was 71 years). There was some evidence that acupuncture and behavioural interventions may reduce dysphagia but the roles of drug therapy, neuromuscular electrical stimulation, pharyngeal electrical stimulation, physical stimulation, transcranial direct current stimulation, and transcranial magnetic stimulation remain unclear. Liquid food may be given directly into the stomach through feeding tubes, either via the gullet, using a nasogastric tube (NGT), or directly into the stomach via a percutaneous endoscopic gastrostomy (PEG) tube. Starting tube feeding (with either NGT or PEG) early after stroke may reduce death although the information available remains inconclusive. If longer-term feeding is required PEG feeding provides better nutrition and is more secure than a NG tube. The available trial evidence does not support the routine use of protein and energy supplements in acute stroke patients who are able to take food by mouth; supplements may show benefit in those who have signs of malnutrition, for example through reducing pressure sores. 2

6 B A C K G R O U N D Description of the condition Dysphagia after stroke is common, affecting 27% to 64% of patients (Gordon 1987; Wolfe 1993; Odderson 1995; Smithard 1996; Mann 2000; Singh 2006a). Half of these patients will recover within two weeks; some will die and others will require longterm feeding with significant impairment of function, recovery, and quality of life (Barer 1989; Smithard 1997; Mann 1999; Perry 2004). Complications of dysphagia include aspiration leading to chest infection and pneumonia, malnutrition, inability to rehabilitate, increased risk of infection, prolonged length of stay in hospital, and an increased risk of death (Smithard 1993; Odderson 1995; Finestone 1996; Smithard 1996; Sharma 2001; Martino 2005). Dysphagia improves spontaneously in many stroke patients although at one month after stroke 15% of patients still have swallowing problems (Smithard 1993). The early identification and management of dysphagia has been shown to reduce pneumonia rates (Odderson 1995; Ramsey 2003; Hinchey 2005). Under-nutrition is common in stroke patients at the time of their admission (8% to 28%) and worsens during their hospital stay (Axelsson 1989; Gariballa 1998; FOOD 2003; Crary 2006). Malnutrition is associated with increased mortality, length of hospital stay (thereby increasing costs), inability to rehabilitate, and poor functional status (Smithard 1993; Finestone 1996; Gariballa 1998; Correia 2003; FOOD 2003). Whether and how food should be supplemented in stroke patients remains unclear. Under-nutrition and dysphagia after a stroke are risk factors for poor clinical outcomes. Therefore, treatment of these conditions may be beneficial. Description of the intervention Interventions for treating dysphagia are often administered by speech and language therapists (SLTs). These interventions involve the modification of fluid and food consistencies, postural techniques, swallowing exercises, and stimulation of oral and pharyngeal structures (Lazarra 1986; Logemann 1991; Logemann 1993). Local stimulation techniques include thermal and electrical stimulation. Transcranial direct current stimulation (TDCS) and transcranial magnetic stimulation (TMS) are also under investigation (Power 2004; Hamdy 2006). Acupuncture techniques are used routinely in some countries. A number of types of pharmacological agents (capsaicin, black pepper oil, cabergoline, angiotensin-converting enzyme (ACE) inhibitors, and nifedipine) have also been studied in patients (Ebihara 1993; Arai 1998; Arai 2003; Ebihira 2004; Ebihira 2005; Ebihara 2006), mostly with chronic or mixed aetiology dysphagia. The nutritional intake of patients with dysphagia may be managed by using modified diet consistencies or tube feedings. For the latter, tube feedings can be inserted in the nose and positioned in the stomach (nasogastric tube (NGT)), jejunum (nasojejunal tube (NJT)) or surgically placed in the stomach (percutaneous endoscopic gastrostomy (PEG)), radiologically inserted gastrostomy (RIG) tube feeding, or parenteral (intravenous (iv)) feeding. Insertion of an NGT is relatively easy, but requires training and awareness of the risks, which, although low (O Mahony 1995), have been highlighted by safety alerts from the UK National Patient Safety Agency (NPSA 2005). Many patients find NGTs uncomfortable, have poor understanding secondary to their stroke, and repeatedly pull out the tube resulting in interruption of feeding and subsequent worsening of their nutritional state. PEG insertion is an invasive procedure and can be complicated by bleeding, local infection, peritonitis, perforation, and aspiration leading to pneumonia and increased mortality in older stroke patients (Wanklyn 1995; NCEPOD 2004). PEG is more acceptable and less irritating to patients and is superior in delivery of feed and maintaining nutritional status in long-term dysphagic patients with traumatic brain damage and stroke (Wicks 1992; Norton 1996; Erdil 2005). Radiologically inserted equivalents such as RIG are available, but are less commonly used. It remains unclear whether PEG is superior to NGT in patients with acute stroke, the ideal time to start feeding following stroke onset, and after what time period PEG tubes are best inserted. The role of methods such as mittens for restraint, nasal bridles for holding NGT in place, and NJTs remains unclear. Intravenous feeding of dysphagic patients is generally not used unless there is enteral failure because of high complication rates through infection and thromboembolism. How the intervention might work These physical and pharmacological techniques may help recovery of dysphagia following stroke. In addition, they might hasten the natural recovery process. However, the improvement in swallowing and other measures could simply be because of the natural recovery process. Similarly, feeding, fluid, or nutritional supplementation may enhance stroke recovery or may accelerate the natural recovery process or improvements may only be because of natural recovery. Why it is important to do this review It remains unclear whether patients managed by these physical and pharmacological techniques fare better than those receiving no therapy. Improvements may be because of the natural recovery of swallowing function, acute stroke treatment, and stroke unit rehabilitation rather than just because of the dysphagia-targeted therapy. This review aimed to assess the effectiveness of interventions for the treatment of dysphagia (swallowing therapy), and nutritional and fluid supplementation, in patients with acute or subacute stroke. 3

7 O B J E C T I V E S To determine: 1. if swallowing therapy improves clinical outcome; 2. the optimal administration (route, timing) of feeding and fluid administration; 3. if food supplementation improves clinical outcome. M E T H O D S Criteria for considering studies for this review Fluid supplementation Subcutaneous (sc) versus iv. Thickened versus non-thickened fluids. In participants without dysphagia Nutritional supplementation Supplementation versus no supplementation in nondysphagic patients. We excluded trials if they used a cross-over design, recruited patients after six months of stroke onset, or if they involved a large proportion of patients with non-stroke causes of dysphagia. Types of studies Randomised controlled trials involving patients with acute or subacute stroke comparing the following. Types of participants Definitions In participants with dysphagia Swallowing therapy Acupuncture versus no acupuncture or routine acupuncture or sham acupuncture. Behavioural interventions: swallowing exercises, dietary modification, positioning versus limited or usual or no treatment. Drug intervention versus none or placebo. Neuromuscular electrical stimulation (NMES) versus none or sham stimulation. Pharyngeal electrical stimulation (PES) versus none or sham stimulation. Physical stimulation: thermal, tactile versus limited, or usual or no treatment. TDCS versus none or sham stimulation. TMS versus none or sham stimulation. Comparisons of different strategies: NMES versus behavioural interventions. Route of feeding Parenteral versus enteral feeding. PEG versus NGT. NJT versus NGT. NGT with loop versus NGT. Timing of feeding Early versus late. Acute or subacute stroke recruited with a clinical diagnosis of stroke within six months of onset. Stroke type Ischaemic or haemorrhagic. Early feeding Within seven days of stroke onset. Dysphagia Diagnosed clinically (water swallow tests, modified diet and fluid assessments, swallowing test scores) by a range of clinicians, or using videofluoroscopy, or using flexible endoscopic evaluation of swallowing (FEES). Malnutrition or under-nutrition Subjective assessment based on body mass index (BMI), Demiquet index (a ratio to determine body mass in relation to skeletal size; used as an alternative to BMI, where measurement of height is difficult, and also in older people, where BMI is less reliable), nutritional risk score, anthropometric measures, and biochemical measures. 4

8 Types of interventions Swallowing therapy for dysphagia Acupuncture. Behavioural interventions: swallowing exercises/therapy and dietary modification. Drug therapy. NMES. PES. Physical stimulation (thermal, tactile). TDCS. TMS. Feeding and fluids Route of feeding: NGT, NJT, PEG, RIG, iv, sc. Timing of feeding. Fluid supplementation. Nutritional supplementation: providing protein and calorie supplements. Types of outcome measures Where available we obtained Information on the following outcome measures for each trial. Primary outcomes Functional outcome: death or dependency, or death or disability, at the end of the trial (we defined disability and dependency as a Barthel Index of 0 to 55 or Rankin score of 3 to 5). Secondary outcomes 1. Case fatality at the end of the trial. 2. Neurological deterioration as measured by a stroke impairment scale (e.g. National Institutes of Health Stroke Scale, Scandinavian Stroke Scale) within four weeks. 3. Late disability or dependency at the end of the trial. 4. Proportion with dysphagia at the end of the trial. 5. Improvement in dysphagia: videofluoroscopy, pharyngeal transit time, swallowing time, normal water swallow test, improvement in swallow function scales, functional oral intake scale (FOIS), Watian swallow scale, return to normal diet and fluids. 6. Aspiration: clinical, videofluoroscopy. 7. Pneumonia: clinical, radiologically. 8. Gastrointestinal bleeding. 9. Feeding tube failures: withdrawal of tube feeding. 10. Nutritional measures: weight, albumin, mid-arm circumference (MAC). 11. Length of hospital stay. 12. Pressure sores. 13. Institutionalisation: discharge destination, residential or nursing home or extended care facility. 14. Quality of life: for example Short Form-36 (SF-36), EuroQol. 15. Food intake: calories or volume of feed. Search methods for identification of studies See the Specialized register section in the Cochrane Stroke Group module. We searched for trials in all languages and arranged translation of relevant trials published in languages other than English. Electronic searches We searched the Cochrane Stroke Group Trials Register (last searched in February 2012), MEDLINE (1966 to July 2011) (Appendix 1), EMBASE (1980 to July 2011) (Appendix 2), CINAHL (1982 to July 2011) (Appendix 3), and Conference Proceedings Citation Index-Science (CPCI-S) (1990 to July 2011). Searching other resources In an effort to identify further published, unpublished, and ongoing trials, we: 1. searched the reference lists of relevant trials, review articles, and our own reference lists; 2. contacted researchers; 3. searched the ongoing trials register Current Controlled Trials ( (July 2011). For the previous version of this review we contacted the Royal College of Speech and Language Therapists Special Interest Group for adult-acquired dysphagia, and companies who manufacture PEG- or NGT-related equipment. Data collection and analysis Selection of studies For this update two review authors (CG and JB) scanned the titles and abstracts of the records identified from the searches of the electronic bibliographic databases and excluded obviously irrelevant articles. We then obtained the full text of the remaining studies and the same two review authors selected relevant trials based on the review inclusion criteria. These two review authors resolved any disagreements through discussion and consultation with a third review author (PB) if necessary. Randomised controlled trials in acute or subacute (less than six months) stroke of: 1. interventions for dysphagia; 2. feeding strategies and timing; 3. fluid supplementation; and 5

9 4. effects of nutritional supplementation. Data extraction and management For this updated review, two review authors (CG and JB) assessed new trials, extracted data using a predefined proforma, and resolved disagreements through discussion and consultation with a third review author (PB). We sought additional information, where necessary, from the principal investigators of trials that appeared to meet the inclusion criteria. at the forest plots to see how CIs overlapped (non-overlapping studies are likely to exhibit statistical heterogeneity) and by the I 2 statistic (Higgins 2011). Assessment of reporting biases We assessed trials for selective outcome reporting and the assessment of each trial is reported in the Risk of bias table (Characteristics of included studies). Assessment of risk of bias in included studies We assessed risk of bias in the included trials using the Risk of bias tool as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The assessment included: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other issues. Measures of treatment effect We calculated weighted estimate of the typical treatment effect across trials using the odds ratio (OR) and 95% confidence intervals (CIs) for binary data and mean difference (MD) and 95% CIs for continuous data and used Review Manager 5.1 (RevMan 2011). We calculated ORs using the Mantel-Haenszel method and MDs using the inverse variance method. Data synthesis We obtained data on randomisation, blinding, the number of patients randomised, time of treatment from stroke, type of dysphagia therapy, patient withdrawals and losses to follow-up, and relevant outcomes (Types of outcome measures). Subgroup analysis and investigation of heterogeneity For each outcome we analysed different swallowing and nutritional interventions as different subgroups. We assessed heterogeneity by looking at the forest plots to see how well the CIs of trials in each subgroup overlapped (if studies did not overlap at all then it was likely to have more variation between the study results than expected by chance) and by the I 2 statistic. Unit of analysis issues Where outcome measures included different scores we converted these to grades in the same direction of desirability and analysed them using MDs. Three studies compared graduations of therapy (Yuan 2003; Carnaby 2006; Jing 2007). In these cases we divided the middle intensity group in two, and analysed the study data by comparing high intensity with medium intensity, and medium intensity with low intensity or no treatment. When a trial compared more than one active treatment with a common control group, we divided the control group patients equally between treatment groups to prevent control patients being counted more than once and thereby artificially narrowing the CIs. Dealing with missing data If trial publication did not provide relevant data we contacted the principle investigator in an effort to obtain the missing data. If they did not respond, then we excluded the trial from the analyses. Assessment of heterogeneity We used random-effects models as we anticipated that the trials would be heterogeneous in design, including different types of patients and interventions. We assessed heterogeneity by looking Sensitivity analysis We did not perform any sensitivity analyses because of the small number of trials. R E S U L T S Description of studies See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification; Characteristics of ongoing studies. We identified 195 studies. Of these, 15 studies are ongoing and we excluded 108 studies mainly because they have compared two active treatments without a control, the trials were not randomised, or no relevant outcome data were present (Characteristics of excluded studies). A further 38 studies are awaiting assessment because we are in the process of retrieving full-text articles of these publications (Studies awaiting classification). 6

10 Results of the search The database searches identified 975 references. A further 21 references were identified through other sources (Figure 1). We assessed 156 full-text articles for eligibility and we are in the process of retrieving a further 38 full-text articles (Figure 1). The present analyses included 33 studies involving 6779 patients (Included studies). The mean age across the included studies was 71 years. 7

11 Figure 1. Study flow diagram.* Further 38 studies are awaiting assessment. 8

12 Included studies Transcranial magnetic stimulation One study assessed TMS in 26 patients (Khedr 2009). Swallowing therapy for dysphagia We included 18 studies involving 967 patients. The trials looked at various forms of swallowing therapy after stroke. Acupuncture We included five acupuncture studies involving 321 patients (Liu 2000; Wei 2005; Bai 2007a; Bai 2007b; Huang 2010). Behavioural interventions Five studies tested behavioural interventions in 423 patients (Yuan 2003a; Yuan 2003b; Song 2004; Carnaby 2006a; Carnaby 2006b). Behavioural interventions consisted of swallowing exercises, environmental modifications such as upright positioning for feeding, safe swallowing advice, and appropriate dietary modifications. Drug therapy We included two studies with a total of 75 patients (Perez 1997; Gosney 2006). Drug interventions included nifedipine (17 patients) (Perez 1997) and an antibacterial oral gel in 58 dysphagic stroke patients (from a larger sample of 203 stroke patients) (Gosney 2006). Neuromuscular electrical stimulation One study assessed NMES in 22 patients (Lim 2009). Pharyngeal electrical stimulation PES was assessed in one study involving 28 patients (Jayasekeran 2010). Physical stimulation (thermal, tactile) Two studies assessed physical stimulation (thermal or tactile) in 35 patients with dysphagia (Bath 1997; Power 2006). Transcranial direct current stimulation One study involving 14 patients assessed TDCS (Kumar 2011). Feeding and fluids Route of feeding Five studies (455 patients) compared PEG with NGT feeding (Norton 1996; Bath 1997; PEGASUS 2004; FOOD ; Hamidon 2006). One study (104 patients) compared looped NGT versus conventional NGT (Beavan 2010). The studies ranged in size from 19 patients (single site) (Bath 1997) to 321 patients (47 sites) (FOOD ). Patients were recruited at between four and 30 days post stroke (Characteristics of included studies). We excluded several trials because of study design: chronic stroke, method of randomisation, low proportion of stroke patients, or lack of data (Characteristics of excluded studies). There were no trials on the use of RIG tubes or parenteral nutrition in stroke alone. Timing of feeding A second component within the FOOD (Feed Or Ordinary Diet) family of trials compared earlier (within seven days) versus later feeding in 859 patients from 83 sites (FOOD ). We found no other RCTs assessing timing of feeding in acute stroke. Fluid supplementation One study (20 participants) compared administering free water and thickened fluids with thickened fluids alone in patients known to aspirate thin fluids (Garon 1997). Nutritional supplementation Eight studies involving 4391 non-dysphagic patients assessed the effect of nutritional supplementation (Gariballa 1998; FOOD ; Aquilani 2008; Rabadi 2008; Nutristroke 2009a; Nutristroke 2009b; Nutristroke 2009c; Ha 2010). One study included 42 patients with impaired nutritional status (Gariballa 1998). The third component within the FOOD family of trials assessed protein-calorie supplementation in 4023 patients from 125 centres (FOOD ). Three studies assessed antioxidants and n3-fatty-acid supplementation in 52 post-stroke patients (Nutristroke 2009a; Nutristroke 2009b; Nutristroke 2009c). Another study assessed intensive nutritional supplementation in 102 under-nourished post-stroke patients (Rabadi 2008). The seventh study assessed protein-calorie supplementation in 48 post-stroke 9

13 patients (Aquilani 2008). The final study assessed the effects of individualised nutritional supplementation in 124 post-stroke patients aged over 65 years (Ha 2010). We excluded a further 12 studies (Characteristics of excluded studies). Excluded studies We excluded a further 108 studies, mainly because there was no control group, the trial was not randomised, or no relevant outcome data were available (Characteristics of excluded studies). Risk of bias in included studies We assessed key sources of bias as follows. Risk of bias across studies is summarised in Figure 2. Figure 2. Risk of bias graph: review authors judgements about each Risk of bias item presented as percentages across all included studies. Allocation Thirteen studies involved randomisation by computer, thereby ensuring concealment of allocation (Bath 1997; Garon 1997; Perez 1997; FOOD ; FOOD ; FOOD ; Carnaby 2006; Gosney 2006; Hamidon 2006; Aquilani 2008; Beavan 2010; Ha 2010; Jayasekeran 2010). Randomisation occurred using random numbers tables in two studies (Song 2004; Jing 2007); block randomisation by telephone in one study (Gariballa 1998); sealed opaque envelope containing block randomisation of 10 patients in one study (Rabadi 2008); and using a specific list in three studies (Nutristroke 2009a; Nutristroke 2009b; Nutristroke 2009c). Randomisation procedures were unclear in 10 studies (Norton 1996; Liu 2000; Yuan 2003; PEGASUS 2004; Wei 2005; Power 2006; Khedr 2009; Lim 2009; Huang 2010; Kumar 2011). Baseline prognostic factors were similar between intervention and control groups in 17 studies (Garon 1997; Perez 1997; Gariballa 1998; PEGASUS 2004; FOOD ; FOOD ; FOOD ; Carnaby 2006a; Carnaby 2006b; Aquilani 2008; Rabadi 2008; Khedr 2009; Nutristroke 2009a; Nutristroke 2009b; Nutristroke 2009c; Beavan 2010; Ha 2010); matching in the other 16 studies was unclear (Norton 1996; Bath 1997; Liu 2000; Yuan 2003a; Yuan 2003b; Song 2004; Wei 2005; Gosney 2006; 10

14 Hamidon 2006; Power 2006; Bai 2007a; Bai 2007b; Lim 2009; Huang 2010; Jayasekeran 2010; Kumar 2011). Blinding Seven studies were double blind (Perez 1997; Aquilani 2008; Rabadi 2008; Nutristroke 2009a; Nutristroke 2009b; Nutristroke 2009c; Kumar 2011). One study was single blind (Power 2006). were assessed in a blinded fashion in eight studies ( Perez 1997; FOOD ; FOOD ; FOOD ; Wei 2005; Carnaby 2006; Khedr 2009; Jayasekeran 2010) and unblinded in eight studies (Norton 1996; Bath 1997; Garon 1997; Gosney 2006; Hamidon 2006; Bai 2007a; Bai 2007b; Beavan 2010); outcome blinding was unclear in nine studies (Gariballa 1998; Liu 2000; Yuan 2003a; Yuan 2003b; PEGASUS 2004; Song 2004; Lim 2009; Ha 2010; Huang 2010). Swallow score There was no difference in swallow scores between treatment and control groups. However, significant heterogeneity was noted (t = 3; n = 256; MD -0.41; 95% CI to 0.72; I 2 = 91%; P < ) for swallow scores in acupuncture studies (Analysis 1.8). Data on other outcomes were not available. Behavioural interventions Dysphagia at the end of the trial Behavioural interventions significantly reduced dysphagia by end of trial (t = 5; n = 423; OR 0.52; 95% CI 0.30 to 0.88; I 2 = 22%; P = 0.01) (Analysis 1.6). Incomplete outcome data Two studies reported no loss of patients during follow-up, 15 studies reported loss of patients during follow-up, and loss of patients during follow up was unclear in the remaining studies (Included studies). Length of stay A non-significant reduction in length of stay was noted (t = 4; n = 370; MD -2.70; 95% CI to 0.28; I 2 = 19%; P = 0.08) (Analysis 1.4). Selective reporting Twenty-three studies reported complete data; in another 10 studies it was unclear whether reported data were complete. Chest infection or pneumonia A non-significant reduction in chest infection/pneumonia was noted (t = 5; n = 423; OR 0.50; 95% CI 0.24 to 1.04; I 2 = 34%; P = 0.06) (Analysis 1.5). Other potential sources of bias The three acupuncture studies (Liu 2000; Wei 2005; Jing 2007) and two of the swallowing studies (Yuan 2003; Song 2004) were assessed from translations of the original text. Translations from Chinese to English were performed by native Chinese speakers. Effects of interventions Swallowing therapy for dysphagia Acupuncture Case fatality at the end of trial No effects were apparent on case fatality (t = 2; n = 306; OR 0.83; 95% CI 0.46 to 1.51) (Analysis 1.1). Death or dependency at the end of trial No effects were apparent on death or dependency (t = 2; n = 306; OR 1.05; 95% CI 0.63 to 1.75) (Analysis 1.2). Institutionalisation No effects were apparent on institutionalisation (t = 2; n = 306; OR 0.76; 95% CI 0.39 to 1.48) (Analysis 1.3). Dysphagia at the end of the trial Data from four studies showed a reduction in dysphagia by end of trial (t = 4; n = 256; OR 0.24; 95% CI 0.13 to 0.46; P < ) (Analysis 1.6). Nutrition (albumin) No effects were apparent on blood albumin concentration (t = 2; n = 64; MD 0.20; 95% CI to 5.17) (Analysis 1.9). 11

15 Drug therapy Physical stimulation (thermal, tactile) Case fatality at the end of trial Drug therapy was not associated with differences in case fatality (t = 1; n = 17; OR 1.14; 95% CI 0.06 to 21.87) (Analysis 1.1). Chest infection or pneumonia No effect on chest infection or pneumonia (t = 1; n = 58; OR 0.19; 95% CI 0.02 to 1.67) (Analysis 1.5). Dysphagia at the end of the trial No effect on dysphagia at end of trial (t = 1; n = 17; OR 0.48; 95% CI 0.07 to 3.35) (Analysis 1.6). Pharyngeal transit time (seconds) No effect on pharyngeal transit time (t = 1; n = 17; MD -0.21; 95% CI to 0.49) (Analysis 1.7). Neuromuscular electrical stimulation Dysphagia at the end of the trial NES did not alter dysphagia at the end of one small trial (t = 1; n = 22; OR 0.43; 95% CI 0.07 to 2.50) (Analysis 1.6). Data on other outcomes were not available. Pharyngeal electrical stimulation Case fatality at the end of trial In one small trial, physical stimulation had no effect on case fatality (t = 1; n = 19; OR 1.05; 95% CI 0.16, to 6.92) (Analysis 1.1). Dysphagia at the end of the trial No effect on dysphagia at the end of trial (t = 1; n = 7; OR 0.33; 95% CI 0.01 to 11.34) (Analysis 1.6). Pharyngeal transit time (seconds) One small study reduced pharyngeal transit time (t = 1; n = 28; MD -0.19; 95% CI to -0.04) (Analysis 1.7). Swallow score No effect on swallow score (t = 1; n = 28; MD 1.40; 95% CI to 5.38) (Analysis 1.8). Data on other outcomes were not available. Transcranial direct current stimulation Dysphagia at the end of the trial TDCS did not alter dysphagia at the end of one small trial (t = 1; n = 14; OR 0.29; 95% CI 0.01 to 8.39) (Analysis 1.6). Swallow score No effect on swallow score (t = 1; n = 14; MD 1.00; 95% CI to 2.50) (Analysis 1.8). Data on other outcomes were not available. Transcranial magnetic stimulation Pharyngeal transit time (seconds) PES significantly reduced pharyngeal transit time (t = 1; n = 28; MD -0.15; 95% CI to 0.20) (Analysis 1.7). Case fatality at the end of trial PES did not alter case fatality at end of trial (t = 1; n = 18; OR 4.31; 95% CI 0.19 to 98.51) (Analysis 1.1). Case fatality at the end of trial TMS did not alter case fatality at the end of one small trial (t = 1; n = 26; OR 0.26; 95% CI 0.01 to 7.12) (Analysis 1.1). Data on other outcomes were not available. Feeding and fluids Chest infection or pneumonia No effect on chest infection or pneumonia (t = 1; n = 28; OR 0.43; 95% CI 0.06 to 3.09) (Analysis 1.5). Percutaneous endoscopic gastrostomy versus nasogastric tube feeding Data were available for five studies (Norton 1996; Bath 1997; PEGASUS 2004; FOOD ; Hamidon 2006). 12

16 Treatment failures PEG was associated with fewer treatment failures (t = 3; n = 72; OR 0.09; 95% CI 0.01 to 0.51; P = 0.007; I 2 = 0%) (Analysis 2.8). Gastrointestinal bleeding PEG was associated with fewer gastrointestinal bleeding events (t = 1; n = 321; OR 0.25; 95% CI 0.09 to 0.69; P = 0.007) (Analysis 2.9). Feed delivery (%) PEG was associated with higher feed delivery (t = 1; n = 30; MD 22.00; 95% CI to 27.85; P < ) (Analysis 2.10). Albumin (g/l) PEG was associated with higher albumin (t = 3; n = 63; MD 4.92; 95% CI 0.19 to 9.65; P = 0.04; I 2 = 58%) (Analysis 2.13). Mid-arm circumference (cm) PEG was also associated with a trend to a higher MAC (t = 3; n = 58; MD 2.29; 95% CI to 4.89; P = 0.08; I 2 = 0%) (Analysis 2.12). Pressure sores PEG was associated with fewer pressure sores (t = 1; n = 321; OR 3.10; 95% CI 0.98 to 9.83; P = 0.05) (Analysis 2.5). Chest infection or pneumonia No effect on chest infection/pneumonia rates (t = 2; n = 93; OR 0.65; 95% CI 0.23 to 1.86) (Analysis 2.6). Dysphagia at end of trial No effect on dysphagia at end of trial (t = 2; n = 66; OR 0.76; 95% CI 0.05 to 11.77) (Analysis 2.7). Weight at end of trial (kg) No effect on weight at end of trial (t = 2; n = 34; MD 4.08; 95% CI to 12.48) (Analysis 2.11). Looped nasogastric tube versus conventional nasogastric tube One small study compared looped NGT with conventional NGT feeding (Beavan 2010). Feed delivery (%) Feed delivery was significantly higher in the looped NGT group than conventional NGT group (t = 1; n = 104; MD 18.00; 95% CI 6.66 to 29.34; P = 0.002) (Analysis 2.10). Case fatality at the end of trial Looped NGT versus conventional NGT feeding did not differ for end-of-trial case fatality (t = 1; n = 104; OR 0.60; 95% CI 0.27 to 1.33) (Analysis 2.1). Case fatality at the end of trial PEG and NGT feeding did not differ for end-of-trial case fatality (t = 5; n = 455; OR 0.81; 95% CI 0.42 to 1.56) (Analysis 2.1). Death or dependency at the end of trial No effect on death or dependency (t = 1; n = 104; OR 0.52; 95% CI 0.18 to 1.57) (Analysis 2.2). Death or dependency at the end of trial No effect on death or dependency (t = 3; n = 400; OR 0.80; 95% CI 0.12 to 5.55) (Analysis 2.2). Institutionalisation No effect on institutionalisation (t = 1; n = 104; OR 1.73; 95% CI 0.78 to 3.81) (Analysis 2.3). Institutionalisation No effect on institutionalisation (t = 2; n = 364; OR 0.62; 95% CI 0.15 to 2.57) (Analysis 2.3). Length of stay in hospital (days) No effect on length of stay in hospital (t = 1; n = 104; MD 7.00; 95% CI to 22.48) (Analysis 2.4). Length of stay in hospital (days) No effect on length of stay in hospital (t = 2; n = 384; MD 14.32; 95% CI to 40.68) (Analysis 2.4). Pressure sores No effect on pressure sores (t = 1; n = 104; OR 1.04; 95% CI 0.28 to 3.84) (Analysis 2.5). 13

17 Chest infection or pneumonia No effect on chest infection or pneumonia (t = 1; n = 104; OR 0.84; 95% CI 0.39 to 1.84) (Analysis 2.6). Protein intake (g/day) Protein intake was increased (t = 3; n = 174; MD 17.28; 95% CI 1.99 to 32.56; P = 0.03; I 2 = 92%) (Analysis 5.7). Treatment failure No effect on treatment failure (t = 1; n = 104; OR 1.67; 95% CI 0.64 to 4.34) (Analysis 2.8). Case fatality at the end of trial A non-significant reduction in case fatality was noted (t = 7; n = 4343; OR 0.58; 95% CI 0.28 to 1.21; P = 0.14; I 2 = 38%) (Analysis 5.1). Gastrointestinal bleeding No effect on gastrointestinal bleeding (t = 1; n = 104; OR 1.63; 95% CI 0.43 to 6.17) (Analysis 2.9). Institutionalisation A non-significant reduction in institutionalisation was noted (t = 1; n = 102; OR 0.48; 95% CI 0.22 to 1.07) (Analysis 5.3). Timing of feeding One medium-sized study compared starting feeding earlier (less than one week) or later (FOOD ). Feeding commenced earlier rather than later was associated with a tendency to a lower end-of trial case-fatality (t = 1; n = 859; OR 0.79; 95% CI 0.61 to 1.04; P = 0.09) (Analysis 3.1). The timing of feeding did not differ for death or disability (t = 1; n = 859; OR 0.94; 95% CI 0.68 to 1.31) (Analysis 3.2) or rate of institutionalisation (t = 1; n = 859; OR 1.12; 95% CI 0.81 to 1.56) (Analysis 3.3). Fluid supplementation Data were only available from one small study (Garon 1997). Fluid supplementation did not alter the time to resolution of dysphagia (t = 1; n = 20; MD -8.10; 95% CI to 4.64) (Analysis 4.1). No episodes of pneumonia were reported. Nutritional supplementation Data were available for eight studies involving 4391 patients ( Gariballa 1998; FOOD ; Aquilani 2008; Rabadi 2008; Nutristroke 2009a; Nutristroke 2009b; Nutristroke 2009c; Ha 2010). Pressure sores Nutritional supplementation was associated with reduced pressure sores (t = 2; n = 4125; OR 0.56; 95% CI 0.32 to 0.96; P = 0.03; I 2 = 0%) (Analysis 5.5). Energy intake (kcal/day) Energy intake was increased (t = 3; n = 174; MD ; 95% CI to ; P = 0.003; I 2 = 91%) (Analysis 5.6). Death or dependency at the end of trial No effect on death or dependency (t = 1; n = 4023; OR 1.06; 95% CI 0.94 to 1.20) (Analysis 5.2). Length of stay in hospital (days) No effect on length of stay in hospital (t = 2; n = 4114; MD 1.40; 95% CI to 3.60) (Analysis 5.4). Albumin (g/l) No effect on albumin concentration (t = 2; n = 144; MD 0.29; 95% CI to 1.24) (Analysis 5.8). D I S C U S S I O N We identified 33 studies that assessed feeding and swallowing treatment strategies in stroke patients. A further 14 studies are ongoing (Characteristics of ongoing studies). Eighteen completed studies assessed the effect of swallowing therapy in patients with post-stroke dysphagia. A variety of stimulatory techniques have been tested - acupuncture, behavioural therapy, drug therapy, NMES, PES, physical stimulation, TDCS, and TMS. None of the techniques showed, individually, significant effects on functional outcome (primary outcome) or case fatality, albeit each based on limited data. Both acupuncture and behavioural interventions significantly reduced dysphagia at the end of trial. In the absence of significant effects on the primary outcome, significant findings in secondary and explanatory outcomes may reflect chance (e.g. owing to multiple comparisons) and further trials are needed to test these observations. Limited evidence from five studies suggested that there might be a trend towards a lower death rate with PEG as compared with NGT feeding, although the results were heterogeneous and largely 14

18 reflected the results from one study (Norton 1996) where timings for NG feeding were much later than current practice. PEG feeding appeared to improve overall delivery of feed. The data related to timing of feeding suggested that enteral nutrition should be commenced earlier (within seven days) rather than later (FOOD ). Nutritional supplementation involving seven studies was associated with a non-significant reduction in case fatality although, again, considerable heterogeneity existed. Pressure sores were significantly lower with nutritional supplementation. However there was no effect on death or dependency, length of stay in hospital, or albumin concentration. Albumin is a poor marker of nutritional status and more closely relates to sepsis, severe illness, and inflammatory conditions. Several studies assessing nutritional supplementation have provided data on albumin levels and it was the main reason for adding this biochemical indicator into the present version of this review. Studies included post-stroke patients irrespective of their swallowing status and had variable baseline nutritional status. Of four studies, the first recruited underweight stroke patients (Rabadi 2008), the second recruited older (over 65 years) stroke patients (Ha 2010), the third trial recruited a small number of under-nourished patients (FOOD ), and the fourth recruited patients only with under-nutrition (Gariballa 1998). No studies reported data on food intake using calories or volume of feed. This measurement would be useful concordance with the reported energy intake (kcals/day) consumption measures that are already reported in this review, especially during transition from non-oral feeding routes to oral dietary consumption (liquids or thickened liquids) of high-calorie supplements. Results of the present analysis were subject to several caveats. First, we excluded 108 studies from the analysis. One common reason for exclusion was that studies compared two active treatments without having a control or placebo. Therefore, we would encourage trialists to design a control or placebo group for future trials. Lack of uniformity in outcome measures and lack of data on clinical outcomes, such as dependency, mortality, institutionalisation, and chest infections, has led us to exclude many trials. These trials have used various swallowing assessment techniques, cortical excitability techniques, and videofluoroscopic measurements. In future, trialists should be encouraged to report clinical outcomes. Second, a further 38 studies are awaiting assessment and we acknowledge that this is a significant number of publications that may have the potential to affect the results of the review. We will seek full-text articles for these studies and we will add them to the review as soon as possible. Third, with regard to acupuncture, data from the three studies may have been confounded owing to the use of routine acupuncture or a different type of acupuncture as control, variation in the delivery of therapy, and the risk of language bias since the majority of acupuncture literature is only available in full in Chinese language journals. Fourth, the risk of bias assessments were completed for newly added studies in this update; however, these were incomplete for some studies that were already in the review, mainly because the publications did not provide the relevant information. In addition, the present analysis included studies up to six months from stroke onset. However, in future, it may be useful to analyse acute and rehabilitation studies separately. Ongoing trials should add substantially to the existing data (Characteristics of ongoing studies). A U T H O R S C O N C L U S I O N S Implications for practice Acupuncture and behavioural therapy (as provided by SLTs) may reduce dysphagia, although the components of each that are effective remain unclear. In the short term, the available evidence suggests that survival may be better if feeding is started earlier, and there is no clear advantage of PEG over NG feeding. For those patients who require long-term nutritional support (feeding beyond six months) PEG feeding results in fewer treatment failures and gastrointestinal bleeding and better feed delivery. Finally, nutritional supplements do not appear to be of value to the majority of patients except for those who are admitted malnourished or possibly in those who are at particular risk of malnourishment. Implications for research Further research is needed to discover which components of swallowing therapy, including acupuncture, are beneficial. Research studies into dysphagia and under-nutrition need to ensure that standardised outcome measures are used to allow comparison of trials. However, measuring nutritional status is difficult and there are no indicators validated in the stroke population. Reporting of proportions of patients who develop pneumonia or have signs of aspiration should be an important outcome measure in all dysphagia and feeding-related trials. Few studies (FOOD ; FOOD ; FOOD ) assessed quality of life, which has relevance when balancing the risks and benefits of interventions in severely disabled stroke survivors. In addition, several studies have compared active treatments without a control group and were excluded from this review. For future studies we recommend trialists include a control or placebo group. A number of studies assessing interventions for dysphagia and nutritional support are ongoing and these will add further information on this important research question (Characteristics of ongoing studies). A number of studies of mixed groups of chronic dysphagia have been done or are ongoing: a systematic review of these studies may inform the management of acute and subacute dysphagia post stroke. 15

19 A C K N O W L E D G E M E N T S We thank Ms Jean Kerr and Ms Morwenna Collins (SLTs) for their help with the early stages of the first version of the review, and Cameron Sellars and David Smithard for their involvement in the completion of the first version (through searches, interpretation of data, and writing the review). We thank the Cochrane Stroke Group for helping identify trials, and their editors and external assessor for comments on the review. Several trialists and other interested healthcare staff reviewed the draft of the first version and made comments - we thank each of them: CGMI Baeten (Netherlands), MS Dennis (UK), BR Garon (USA), GJ Hankey (Australia), GKT Holmes (UK), PR Mills (UK), B Norton (UK), C Ormiston (USA), J Rosenbek (USA), and G Vanhooren (Belgium). We also thank D Luo and G Lan who translated five of the papers from Chinese into English. Finally, we are grateful to the funding bodies that supported this research. Naturally any mistakes are our own. We would be very grateful to be informed of any other related completed or ongoing trials that are not listed in the review. R E F E R E N C E S References to studies included in this review Aquilani 2008 {published data only} Aquilani R, Scocchi M, Boschi F, Viglio S, Iadarola P, Pastoris O, et al.effect of calorie-protein supplementation on the cognitive recovery of patients with subacute stroke. Nutritional Neuroscience 2008;11(5): Bai 2007a {published data only} Bai J, Li B, Wang Z, Gao W, Wang L. The role of different needling manipulation in adjusting swallow period obstacle of dysphagia after stroke. Zhongguo Zhenjiu 2007;27(1): Bai 2007b {published data only} Bai J, Li B, Wang Z, Gao W, Wang L. The role of different needling manipulation in adjusting swallow period obstacle of dysphagia after stroke. Zhongguo Zhenjui 2007;27(1): Bath 1997 {unpublished data only} Bath PMW, Kerr J, Collins M. Factorial trial of swallowing versus conventional therapy, and PEG versus nasogastric tube feeding, in dysphagic patients with recent stroke. Unpublished Beavan 2010 {published data only} Beavan J, Conroy SP, Harwood R, Gladman JR, Leonardi- Bee J, Sach T, et al.does looped nasogastric tube feeding improve nutritional delivery for patients with dysphagia after acute stroke? A randomised controlled trial. Age and Ageing 2010;39(5): Carnaby 2006a {published and unpublished data} Carnaby G, Hankey GJ, Pizzi J. Behavioural intervention for dysphagia in acute stroke: a randomised controlled trial. Lancet Neurology 2006;5:31 7. Mann G, Baxter K, Hankey G, Davis B, Stewart-Wynne E. Treatment for swallowing disorders following acute stroke: a randomised controlled trial. Stroke Society of Australia Annual Scientific Meeting Mann G, Hankey G, Davis B, Stewart-Wynne E. Swallowing therapy after acute stroke study (STAASS): where are we now?. Journal of Clinical Neuroscience 1999;6 (3):281. Carnaby 2006b {published data only} Carnaby G, Hankey GJ, Pizzi J. Behavioural interventions for dysphagia in acute stroke: a randomised controlled trial. Lancet Neurology 2006;5:31 7. FOOD {published and unpublished data} Dennis M. FOOD trial (Feed Or Ordinary Diet): a multicentre trial to evaluate various feeding policies in patients admitted to hospital with a recent stroke. Stroke 1998;29:551. Hankey GJ, Dennis MS. Food (Feed Or Ordinary Diet): a family of three randomised trials evaluating feeding policies for patients admitted to hospital with a recent stroke. Journal of Clinical Neuroscience 2002;9(4):483. Ricci S. International Stroke Trials Collaboration: FOOD Trial (Feed Or Ordinary Diet). Revista Medica 1999;5(4): Signorini DF, on behalf of the International Stroke Trials Collaboration - FOOD. Advantages of an inclusive trial: the FOOD pilot experience. Cerebrovascular Diseases 1998; 8 Suppl 4:83. The FOOD trial collaboration. Routine oral nutritional supplementation for stroke patients in hospital. Lancet 2005;365: The International Stroke Trials Collaboration. FOOD Trial (Feed Or Ordinary Diet). Protocol. FOOD {published and unpublished data} Dennis M. FOOD trial (Feed Or Ordinary Diet): a multicentre trial to evaluate various feeding policies in patients admitted to hospital with a recent stroke. Stroke 1998;29:551. Hankey GJ, Dennis MS. Food (Feed Or Ordinary Diet): a family of three randomised trials evaluating feeding policies for patients admitted to hospital with a recent stroke. Journal of Clinical Neuroscience 2002;9(4):483. Ricci S. International Stroke Trials Collaboration: FOOD Trial (Feed Or Ordinary Diet). Revista Medica 1999;5(4): Signorini DF, on behalf of the International Stroke Trials Collaboration - FOOD. Advantages of an inclusive trial: the FOOD pilot experience. Cerebrovascular Diseases 1998; 16

20 8 Suppl 4:83. The FOOD Trial Collaboration. Effect of timing and method of enteral tube feeding for dysphagic stroke patients (FOOD): a multicentre randomised controlled trial. Lancet 2005;365: The International Stroke Trials Collaboration. FOOD Trial (Feed Or Ordinary Diet). Protocol. FOOD {published and unpublished data} Dennis M. FOOD trial (Feed Or Ordinary Diet): a multicentre trial to evaluate various feeding policies in patients admitted to hospital with a recent stroke. Stroke 1998;29:551. Hankey GJ, Dennis MS. Food (Feed Or Ordinary Diet): a family of three randomised trials evaluating feeding policies for patients admitted to hospital with a recent stroke. Journal of Clinical Neuroscience 2002;9(4):483. Ricci S. International Stroke Trials Collaboration: FOOD Trial (Feed Or Ordinary Diet). Revista Medica 1999;5(4): Signorini DF, on behalf of the International Stroke Trials Collaboration - FOOD. Advantages of an inclusive trial: the FOOD pilot experience. Cerebrovascular Diseases 1998; 8 Suppl 4:83. The FOOD Trial Collaboration. Effect of timing and method of enteral tube feeding for dysphagic stroke patients (FOOD): a multicentre randomised controlled trial. Lancet 2005;365: The International Stroke Trials Collaboration. FOOD Trial (Feed Or Ordinary Diet). Protocol. Gariballa 1998 {published data only} Gariballa SE, Parker SG, Castledon CM. A randomised controlled trial of nutritional supplementation after stroke. Age and Ageing 1998;27 Suppl I:66. Gariballa SE, Parker SG, Taub N, Castleden M. A randomized, controlled, single blind trial of nutritional supplementation after acute stroke. Journal of Parenteral and Enteral Nutrition 1998;22(5): Garon 1997 {published and unpublished data} Garon BR, Engle M, Ormiston C. A randomized control study to determine the effects of unlimited oral intake of water in patients with identified aspiration. Journal of Neurological Rehabilitation 1997;11: Gosney 2006 {published data only} Gosney M, Martin MV, Wright AE. The role of the selective decontamination of the digestive tract in acute stroke. Age and Ageing 2006;35:42 7. Ha 2010 {published data only} Ha L, Hauge T, Spenning AB, Iversen PO. Individual, nutritional support prevents undernutrition, increases muscle strength and improves QoL among elderly at nutritional risk hospitalized for acute stroke: a randomized, controlled trial. Clinical Nutrition 2010;29(5): Hamidon 2006 {published data only} Hamidon BB, Abdullah SA, Zawawi MF, Sukumar N, Raymond AA. A prospective comparison of percutaneous endoscopic gastrostomy and nasogastric tube feeding in patients with acute dysphagic stroke. Medical Journal of Malaysia 2006;61(1): Huang 2010 {published data only} Huang Z, Huang F, Yan HX, Min Y, Gao Y, Tan BD, et al.dysphagia after stroke treated with acupuncture or electric stimulation: a randomized controlled trial. Zhongguo Zhen Jiu 2010;30(12): Jayasekeran 2010 {published data only} Jayasekeran V, Singh S, Tyrrell P, Michou E, Jefferson S, Mistry S, et al.adjunctive functional pharyngeal electrical stimulation reverses swallowing disability after brain lesions. Gastroenterology 2010;138(5): Khedr 2009 {published data only} Khedr EM, Abo-Elfetoh N, Rothwell JC. Treatment of post-stroke dysphagia with repetitive transcranial magnetic stimulation. Acta Neurologica Scandinavica 2009;119(3): Kumar 2011 {published data only} Kumar S, Wagner CW, Frayne C, Zhu L, Selim M, Feng W, et al.noninvasive brain stimulation may improve stroke-related dysphagia: a pilot study. Stroke 2011;42(4): Lim 2009 {published data only} Lim KB, Lee HJ, Lim SS, Choi YI. Neuromuscular electrical and thermal-tactile stimulation for dysphagia caused by stroke: a randomized controlled trial. Journal of Rehabilitation Medicine 2009;41(3): Liu 2000 {published data only} Liu L. Acupuncture treatment of bulbar palsy - a report of 54 cases. Journal of Traditional Chinese Medicine 2000;20 (1):30 2. Norton 1996 {published data only} Norton B, Holmes GKT. Percutaneous endoscopic gastrostomy feeding after acute dysphagic stroke. BMJ 1996;312: Norton B, Homer-Ward M, Donnelly MT, Long RG, Holmes GKT. A randomised comparison of percutaneous endoscopic gastrostomy feeding and nasogastric tube feeding following acute dysphagic stroke. Gut 1994;35 Suppl 5:S6. Norton B, Homer-Ward M, Donnelly MT, Long RG, Homes GKT. A randomised prospective comparison of percutaneous endoscopic gastrostomy and nasogastric tube feeding after acute dysphagic stroke. BMJ 1996;312:13 6. Norton B, Long RG, Holmes GKT. Tube feedings and file drawers. Gastroenterology 1996;111: Sanders H, Newall S, Norton B, Holmes GTK. Gastrostomy feeding in the elderly after acute dysphagic stroke. Journal of Nutrition Health and Aging 2000;4(1): Nutristroke 2009a {published data only} Garbagnati F, Cairella G, De Martino A, Multari M, Scognamiglio U, Venturiero V, et al.is antioxidant and n- 3 supplementation able to improve functional status in poststroke patients? Results from the Nutristroke Trial. Cerebrovascular Diseases 2009;27(4):