|Home | About | Journals | Submit | Contact Us | Français|
Most patients with dementia develop feeding problems, leading physicians and families to consider tube feeding or oral feeding options. Tube feeding offers limited benefit, but current decision-making includes limited information on other options.
To review the evidence for oral feeding options in dementia.
PubMed/MEDLINE, EMBASE, the Cochrane Library, CINAHL and PsychINFO literature indices between January 1990 and October 2009.
Clinical trials with random or non-random control groups were included if they reported on clinical outcomes of oral feeding interventions for patients with dementia.
Systematic literature search with review of potentially eligible studies by two independent investigators. Investigators abstracted data from included studies using a structured instrument. Studies were graded on quality and potential bias, and overall strength of evidence summarized.
Thirteen controlled trials provide data on use of supplements for patients with dementia, and twelve controlled trials test assisted feeding or other interventions. Studies provide moderate strength evidence for high calorie supplements, and low strength evidence for appetite stimulants, assisted feeding and modified foods to promote weight gain in dementia. The few studies measuring function or survival showed no difference.
High calorie supplements and other oral feeding options can help dementia patients with feeding problems to gain weight; they are unlikely to improve other outcomes. These treatments can be offered alone or in combination as an alternative to tube feeding.
Dementia is a syndrome of decline in cognitive domains causing functional impairment. In early dementia taste and smell dysfunction, medications or depression may reduce intake. 1 In advanced dementia apraxia and attention deficits interfere with self-feeding, and dysphagia causes choking or food avoidance.2 Feeding problems cause important health effects such as weight loss, dehydration, poor wound healing, and pneumonia. In the CASCADE Study 86% of persons with advanced dementia developed a feeding problem, and onset was associated with 39% mortality at 6 months.3
Treatments include medical feeding through a feeding tube, or modifications of oral feeding including high calorie supplements, appetite stimulants, modified foods, enhanced dining environments or personal assistance. The use of feeding tubes has increased for patients with serious illness, particularly dementia and other neurologic diseases.4,5,6,7 Controlled observational studies of persons with dementia provide evidence tube feeding does not prolong life or promote wound healing.8,9,10,11
Physicians and families make choices about feeding in dementia. Interview studies suggest they expect benefits from tube feeding that exceed actual outcomes.12,13 Consent for feeding tubes usually focuses on procedural risks, with limited information on outcomes and alternatives.14,15 If the choice is framed as opting for or against tube feeding, families may fear starvation without understanding other options.
To enhance evidence-based decision-making, we conducted a systematic review of oral feeding options in dementia. We aimed to answer two key questions for dementia care: 1) does the addition of high calorie supplements improve clinical outcomes including weight gain, function or survival; and 2) do other oral feeding interventions such as appetite stimulants, assisted feeding or modified diets improve outcomes?
Assisted by an experienced health services librarian, investigators searched PubMed/MEDLINE, EMBASE, the Cochrane Library, CINAHL and PsychINFO literature indices between January 1990 and October 2009. Searches used a combination of medical subheadings, keywords and text words. We used search terms for dementia and long-term care in combination with terms to identify interventions for nutrition or feeding problems. (Table 1) We added evaluation terms to ensure we did not miss smaller studies. Reference lists of selected studies were hand searched for additional clinical trials.
Investigators defined eligibility prior to searches using the PICOT framework.16 This framework designates 1) population, 2) intervention, 3) comparison groups, 4) outcomes, and 5) time frame for outcomes. (Table 2) Eligible studies were in English. Study participants were aged 50 or older, with dementia of any stage or etiology, and evidence of a feeding problem. Initially, we planned to include only studies of dementia. Because some study populations varied in cognitive status, we accepted those that combined frail older persons with and without dementia. Participants could be in any setting – community, long-term care or hospital. Eligible interventions included 1) prescribed use of high calorie or protein supplements, or 2) other interventions such as appetite stimulants, modified diets, assistance in feeding, or modified dining environments. Supplements without caloric value, such as vitamins, were excluded. We required a comparator group, but accepted randomized or non-randomized controls including pre-post design. Studies had to address at least one major clinical outcome, including survival, hospitalization, pneumonia, aspiration, function, quality of life, weight change, or wound healing, with follow-up of at least 1 month. Intermediate outcomes such as serum proteins or amount of intake were excluded. Two investigators (RG, LCH) reviewed all titles and abstracts, and excluded duplicates, studies that were not clinical trials, studies with ineligible interventions, and studies excluding dementia.
Three investigators reviewed all full text articles (RG, LCH, ME). Two investigators reached consensus on abstracted data on methods, results, study quality and potential sources of bias. A third investigator settled cases of disagreement. The structured abstraction tool recorded study size, dementia diagnosis and severity, type of feeding problems, intervention and control conditions, type of controls, outcome definitions and results, time to follow-up, study setting, as well as methods indicative of study quality.
Two investigators graded each included study on: 1) the strength of evidence and 2) risk of bias using the Cochrane rating approach.17 Strength of evidence was graded A if randomized, placebo-controlled, with concealed allocation; B if randomized without clearly defined concealment; and C if not randomized and concealment inadequate. Risk of bias was rated on type of controls, methods of double or single blinding, well-specified outcomes, well-specified inclusion criteria, >75% complete outcome assessment, risk of confounding bias, and intention to treat analysis. Studies meeting 5–7 of these criteria were judged to have low risk, those meeting 3–4 had medium risk, and those meeting 0–2 had high risk of bias. Data tables were created for summative assessment of evidence for each key question. Investigators used the PRISMA Statement to guide reporting of evidence, and AHRQ comparative effectiveness review guidelines and GRADE criteria for overall strength of evidence.18,19,20
Databases identified 1147 potential studies, and hand searches 65, representing 912 unduplicated publications. Abstract and title review excluded 802, leaving 110 articles for full-text review. Eighty-five studies failed to meet PICOT criteria (n=69) or did not test an intervention (n=16), leaving 25 studies for review. (Figure)
We identified 13 original studies that examined the effect of high calorie supplements on clinical outcomes for feeding problems in dementia. 21,22,23,24,25,26,27,28,29,30,31,32,33 Studies examined a variety of outcomes meeting our criteria – change in weight or body mass index, mortality, morbidity, wound healing, physical or cognitive function. (Table 3) Follow-up varied from 1 month to 1 year. Ten studies found evidence for benefit using at least one outcome. Nine of twelve studies examining weight or body mass index found improvement. One study found improved pressure ulcer healing with supplementation.23 Four studies measuring function, three measuring cognition, and one measuring mortality found no differences with supplementation. One study reported reduced infections at 1 year (47% vs 66%, p=0.05).25
Between 1990 and 2009, two high quality randomized clinical trials have addressed the effectiveness of high calorie supplements for dementia patients. (Table 4) Wouters-Wessling randomized 35 nursing home residents with moderate to severe dementia and low body mass index to a micronutrient-enriched supplement compared to placebo.31 Controls lost an average of 0.8 kg over 12 weeks, while intervention subjects gained 1.4 kg. Cereda enrolled 28 nursing home residents with Stage II–IV pressure ulcers to receive a high calorie, high protein supplement with micronutrients compared to usual diets.23 Although this study met inclusion criteria, two-thirds of participants used tube feeding and only one-third used oral nutrition for prescribed supplements. Wound healing was more rapid in the intervention group over 12 weeks, with mean reduction of wound size of 75% vs 45% (p<0.005).
Nine additional randomized trials provided Grade B evidence for supplement use in dementia, and six demonstrated benefit. One additional trial with concerns for inadequate randomization and one study using pre-post design provided lower quality evidence. Lower and higher quality studies had similar outcomes. Study populations varied in the severity of cognitive impairment and nutritional problems, yet weight gains were consistently 0.5–2.0 kg. Studies rarely reported potential harms from supplements. In one study, 5 of 31 (16%) subjects, particularly those with lower baseline weight, reduced their lunch intake after the supplement.33
We identified 12 studies of diverse nutritional and environmental interventions to improve clinical outcomes for dementia patients.34,35,36,37,38,39,40,41,42, 43,44,45 Interventions included appetite stimulants (3); changes in environment with buffet-style dining (1) or music (1); thickened liquid or semi-solid food (2); or individualized nutritional care plans (4) including exercise (1). Ten studies used weight or body mass index as outcomes; six found positive effects. (Table 3) Follow-up ranged from 3–12 months. Function and mortality were rarely studied; two trials using these outcomes found no benefit.35,39 Three studies tested effects of feeding interventions on behaviors, with mixed results.36,43,45
Three studies tested appetite stimulants for dementia patients. Volicer found dronabinol improved weight and reduced negative affect with a trend toward improving disruptive behaviors.43, 43 Yeh randomized nursing home residents to megestrol acetate 800 mg or placebo. Analyzing subjects with complete data, 43% of the intervention vs. 18% of the placebo group gained > 1.82 kg over 12 weeks, but those with advanced dementia were less likely to respond.44 A pilot study of megestrol demonstrated no change in weight, but secondary analyses suggested benefit when combined with optimal assisted feeding.41
Four non-medication interventions showed positive effects on weight, including chocolate and exercise, enhanced dietician time, lyophilized foods with modified texture, and feeding assistance.34,37,40,42 One other study of enhanced dietician care had no effect, but enrolled less cognitively and nutritionally impaired patients.35 Buffet style dining at one meal a day had no effect on weight.38 A randomized trial of chin down posture compared to thickened liquids found no differential effect for aspiration; investigators commented a “usual care” control could not be used due to ethical concerns.39
The two high quality randomized trials in this group tested appetite stimulants and found beneficial effects on weight. (Table 4) Six additional randomized trials of non-medication interventions provided Grade B evidence with low to medium risk of bias. Four additional studies used non-random controls to test enhanced nutritional assessment, use of calming music, and megestrol. We detected no correlation between study quality and likelihood of positive results.
Four studies explicitly collected and reported data on harms. One randomized trial comparing chin down posture to thickened liquids as interventions to prevent aspiration pneumonia found that 23% of participants experienced an adverse event. Adverse events were equal between groups, but the combined outcome of fever or urinary infection or dehydration tended to be more common for the thickened liquid group (9% vs. 5%, p=0.055).39 In a randomized trial of lyophilized food compared to nutritional advice, nearly all deaths or hospitalizations involved patients with more advanced dementia, but rates did not differ between groups.40 In Simmons’ study of megestrol, nearly half of 17 participants reported new fatigue or loss of strength; 5 reported leg swelling but information on thrombotic disease was not provided.41 Finally, in a placebo-controlled study of dronabinol, one patient on treatment had a seizure, and the active treatment phase was marked by increases in somnolence and euphoria.43
The evidence base for oral feeding options in dementia includes an encouraging number of randomized trials with low to medium risk of bias. Examining both the quantity and quality of research, there is moderate evidence to support the use of high calorie supplements to improve weight for patients with dementia and feeding problems, and low evidence these supplements promote wound healing and reduce infection risk. For the outcome of weight gain, current research offers low evidence for appetite stimulants, assisted feeding and modified foods. Findings regarding weight gain are consistent among trials, nearly uniformly positive, and effect size for supplements can be estimated at 0.5–2.0 kg of weight gain over 1–6 months. Evidence is sparse but consistent in showing no effect of oral feeding options on function, cognition or mortality for patients with moderate to severe dementia.
This systematic review summarizes research on oral feeding in dementia. High calorie supplements are an evidence-based option to promote weight gain for patients with dementia and feeding problems. Assisted feeding, appetite stimulants, and modified foods may also improve weight, and treatments can be used individually or in combination. Based on current evidence, specialized oral feeding interventions are unlikely to change how patients with dementia function or how long they live.
Our results are consistent with a meta-analysis which found protein energy supplementation improved nutrition and reduced morbidity and mortality for undernourished older hospitalized patients.46 Another systematic review found moderate evidence to support use of supplements for healing of pressure ulcers.47 The single study in this review demonstrating wound healing included patients who were both tube fed and orally fed, limiting clear conclusions about oral supplements and pressure ulcer healing. To be effective, prescribed supplements must be ingested. Incomplete administration of supplements occurs in practice, and is associated with weight loss among nursing home residents with dementia.48
This systematic review combines studies that are heterogeneous in the dementia status and feeding problems of enrolled participants, interventions, and outcome measures, precluding meta-analysis. Variation in baseline dementia severity and nutritional status of study subjects raises questions about optimal timing for nutritional interventions in the progression of dementia. Many studies target moderate to severe dementia, but are too small to stratify findings by stage. Some studies enrolled patients at risk for nutritional decline, while others enrolled patients with clear indications of nutritional insufficiency. Findings from individual studies suggest that interventions may be ineffective when initiated before nutrition is a major issue, or for very advanced dementia or very low body mass index, when interventions may be too late. This review focused on dementia, and results may not extend to other populations with nutritional problems.
Our review did identify several areas for improvement in this body of research. Future studies will be strengthened by careful definition of dementia stage and feeding problems of the enrolled subjects. Study of more complex, programmatic interventions make double blinding difficult; single-blinding of outcome assessment and concealment of randomization allocation can avoid important sources of bias. Several studies that demonstrated positive results were supported by manufacturers of nutritional supplements; it is unclear whether or not reporting bias (i.e., failure to publish studies with negative findings) affected the available published literature. Current evidence relies on numerous single-site small studies (average sample size n=73). Future interventions, and combinations of promising interventions in comprehensive nutritional programs, could be tested in multi-site randomized trials. Investigators should be encouraged to design trials that view intake and weight gain as intermediate outcomes, so as to provide stronger evidence about the effects on function, behavior, infection risk and wound healing. Ethical concerns about withholding feeding treatments may limit the range of possible control conditions, including a randomized comparison of oral assisted feeding to tube feeding.
Feeding treatments choices may cause great legal, ethical and clinical controversy, and remain emotionally difficult for family caregivers.49 State laws reflect this controversy, and many set stricter legal requirements, such as health status or explicit evidence of patient wishes, in order to withhold or withdraw medical forms of nutrition and hydration. Patients, and families of patients with dementia, may rely heavily on medical advice to understand other treatment options. No randomized trials of tube feeding compared to oral feeding exist, but observational studies of dementia patients indicate tube feeding is not superior for promoting survival, function or wound healing. In these observational studies, patients without tube feeding served as controls and many may have received oral feeding treatments included in this review.
Nearly all patients with advanced dementia develop feeding problems. Health care providers may confidently advise families that high calorie supplements, perhaps in combination with assisted feeding, foods modified in taste or texture, and appetite stimulants can promote weight gain for several months. Given the progressive nature of dementia, families should be counseled not to expect improvements in function or survival with any available form of feeding. In end-stage dementia, oral feeding may no longer be possible, and tastes and sips of food combined with mouth care may be used to promote comfort.50 Families and health care providers may improve the quality of informed decision-making using current evidence for oral feeding options in dementia.
This research was funded by the National Institute for Nursing Research, R01 NR009826
Sponsor’s Role: The sponsor had no role in the design, methods, subject recruitment, data collection, analysis or writing of this manuscript.
Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper.
Author Contributions: Study concept, design, data analysis and interpretation, preparation of manuscript: Hanson, Ersek, Gilliam, Carey