| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
REVIEW ARTICLE |
1 Division of Human Nutrition, Wageningen University, Wageningen, The Netherlands.
2 Numico Research, Wageningen, The Netherlands.
3 Nursing Home De Landrijt, Eindhoven, The Netherlands.
4 Nursing Home De Riethorst, Geertruidenberg, and Faculty of Social and Behavioural Sciences, Tilburg University, The Netherlands.
5 Department of Geriatric Medicine, University Medical Centre Nijmegen, The Netherlands.
Address correspondence to Marleen Manders, Division of Human Nutrition, Wageningen University, P.O. Box 8129, 6700 EV Wageningen, The Netherlands. E-mail: Marleen.Manders{at}wur.nl
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Methods. The authors systematically reviewed randomized controlled trials that compared nutritional supplementation with a placebo treatment. Trials were identified from a MEDLINE search and from reference lists of identified studies and review articles. From each trial, information was gathered on the number and age of persons studied; the type, dosage, and duration of the intervention; and the assessed outcome measures.
Results. From 1086 titles, 571 articles were excluded based on their titles. Of the remaining 467 articles, the abstracts were read and 422 articles were excluded based on information found there. The remaining articles were screened for quality aspects of the study design, leaving 21 proper randomized, controlled trials. These trials are discussed in three groups according to the type of supplementation: multinutrient intervention or single components with or without a putative mechanism. Twelve studies, which were evenly distributed among the three supplement groups, found significantly positive effects of nutritional intervention on cognitive functioning, whereas nine studies did not. None of the studies found a significantly negative effect of nutritional intervention.
Conclusions. Shortcomings in methodology varying from the duration of intervention to outcome measures partly explain discrepancies in findings. Despite the heterogeneity in trial design, the results of this review suggest that nutritional supplements may improve the cognitive functioning of elderly persons and do no harm. Further well-designed studies are needed to support these findings.
The relation between malnutrition and cognitive functioning is complicated. Besides being a cause of cognitive impairment, malnutrition might also be a consequence of it. Cross-sectional studies have shown that low blood levels of certain nutrients are associated with impaired cognitive functioning in elderly persons (1–4). Longitudinal studies have shown that persons with suboptimal nutritional status seem to be at a higher risk for development of cognitive impairment (5–7). For this reason, it is important to determine whether cognitive functioning of elderly people would be improved by consumption of a nutritional supplement.
According to the literature, a few factors influence the effectiveness of a nutritional intervention on cognitive functioning: the stage of cognitive decline or nutritional deficiency of the investigated population, the applied nutrient and its dose, and the duration of intervention.
Most likely, benefits of nutritional supplementation can be expected mainly early after the onset of cognitive impairment. In later stages, the process of demyelination becomes irreversible and existing neuronal damage is not likely to be reversed (4,8). It is also hypothesized that the effects will be smaller in apparently healthy persons or those with a more advanced impairment of cognitive status. Therefore, the selection of the target group is important in intervention trials. Based on findings from the literature, it is not clear in which groups of elderly people improvement by nutritional intervention is most effective. Therefore, for this review, we wanted to evaluate studies in healthy elderly volunteers and in elderly patients. In our literature search, we did not select studies for the stage of cognitive decline of the target group they investigated.
In addition, we wanted to determine whether a single nutrient can improve cognitive functioning or whether a combination of more nutrients is needed. Table 1 presents single and mixtures of nutrients with their possible mechanisms in affecting cognition. Jorissen and Riedel (9) and Gonzalez-Gross and colleagues (4) have reviewed nutrition-related risk factors for cognitive impairment. Nutritional factors such as amino acids, antioxidants, lipids, and vitamins seem to have a relation to cognitive decline. The applied doses of the supplemented nutrients also might influence the effectiveness. For this review, we defined nutritional supplements as any nutritional intervention aimed at improving the dietary intake of participants. Therefore, we did not include trials that provided their participants with medication.
|
Regarding the fact that cognitive problems are diverse, it is also interesting to know which domain of cognitive functioning is impaired and in which domain improvement by nutrition is possible. Outcome measurements should be selected with care in studies relating diet to cognitive functioning, because of the complexity and multiplicity of cognitive functioning and the impossibility of measuring all cognitive functions using a single study design.
The purpose of our systematic review was to evaluate the existing evidence based on studies on the effectiveness of nutritional supplementation in improving cognitive functioning in elderly persons. We selected only randomized controlled trials of high quality. Therefore, we used a priori defined criteria for selection. Our hypothesis was that cognitive functioning would be improved in elderly people by a complete supplement, as combined factors in the diet cause cognitive decline. The factors we have noted already may influence the effect. We performed a MEDLINE search for articles describing intervention studies on this subject.
|
METHODS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Inclusion Criteria
We identified 1038 potentially eligible trials. If, according to the title, articles did not report on human elderly subjects, an intervention trial, nutritional supplementation, or cognitive function or nutritional status as primary outcomes, we did not consider the studies for our review. We excluded 571 studies accordingly. We further explored the 467 remaining articles by reviewing the abstract. We used the same criteria as for the title screen to exclude abstracts from our review, which ultimately resulted in 47 potentially eligible trials. We scored these articles on quality aspects using a checklist for randomized controlled trials of the Dutch Cochrane Centre, extended with items from the Delphi list (10). Considering the quality aspects, we excluded 26 articles from the review: 3 were not an intervention study, 14 did not use a placebo treatment, 4 were not double blind, and 4 did not describe a randomization procedure. Later, in Results, we evaluate only the 21 best-conducted trials (randomized, double-blind, placebo-controlled studies).
Data Extraction
From each trial we gathered information on the number and age of the persons examined; the description of the investigated population; the type, dosage, and duration of the intervention; and the assessed outcome measures. Our tables summarize this information.
|
RESULTS |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
In each group of studies, we first searched for an overall explanation for the lack of effect. If such an explanation was missing, we further explored the characteristics of the studies without positive effects.
The first group of studies consisted of four trials that addressed the effect of a multinutrient supplement (Table 2). Except for the study of Hogarth and colleagues (11), all participants in these four studies were community-dwelling elderly people. In none of the studies were participants screened for existing vitamin deficiencies or impaired cognitive functioning before the intervention. The interventions in the study by Hogarth and colleagues (11) and the study by Bryan and colleagues (12) were short and involved a mega-dose of vitamins. Conversely, de Jong and colleagues (13) and Chandra (14) investigated a multivitamin supplement during 17 weeks and 1 year, respectively, and did not use high doses. Differences in population, doses, or duration could not explain the difference in effectiveness of the supplements. Typically, the studies that did not show a significant effect used simple, general tests for cognitive functioning. In contrast, positive effects were noted in the studies in which different domains of cognitive functioning were specifically measured. This suggests that the difference in result could be explained in part by the choice of outcome measures. None of the studies in Table 2 reported adverse effects of supplementation.
Table 3 lists studies that evaluated the effect of supplying a single nutrient rather than multinutrients. In total, 7 studies of the 21 randomized, double-blind, placebo-controlled trials investigated 1 active component. In four intervention studies, B vitamins (B1, B6, and B12) were supplied. In one study, guarana (a Brazilian medicinal plant) was given, in one alpha-tocopherol, and in one inositol. In three of the four positive studies, patients with Alzheimer's disease were examined, and in another study, apparently healthy elderly persons were enrolled. In all studies, participants were on average older than 70 years, except for the study of Galduroz and Carlini (15) in which they were younger. One of the significantly positive studies (16) was a long-term investigation with a supplementation period of 2 years. The period of supplementation in the other three positive studies was 4 or 12 weeks. In the vitamin studies (B vitamins and alpha-tocopherol), a high dose was used (4 to 3000 times the recommended daily allowance). Sano and colleagues (16) used no neuropsychological test battery, in contrast to the outcome measures in the other significantly positive studies. The nutrients that showed a positive effect were alpha-tocopherol, inositol, thiamine, and pyridoxine. However, the study by Nolan and colleagues (17) did not show a positive effect of thiamine. Overall, no clear pattern explains the difference in results between the studies.
In the study of Seal and colleagues (18), which was a small study population, the short period of intervention or the use of the Mini-Mental State Examination as primary outcome might explain the lack of effect. Galduroz and Carlini (15) investigated apparently healthy, elderly volunteers who were not screened for vitamin deficiencies. This might explain why they did not find any significant alteration in cognitive functioning. The only factor that might explain the lack of effect in the study by Nolan and colleagues (17) is that the study was insufficiently powered, because they included only 10 patients. None of the suggested factors completely explains the difference in effectiveness of the studies.
Galduroz and Carlini (15) used guarana, which was not completely free of adverse effects, because four volunteers showed some discomfort. Overall, in the study by Sano and colleagues (16), no statistically significant differences occurred in adverse event categories among the four treatment groups. Barak and colleagues (19) reported adverse effects of inositol such as mild insomnia and flatus during the treatment period. Meador and colleagues (20) noted no adverse systemic effects of thiamine. Seal and colleagues (18), Nolan and colleagues (17), and Deijen and colleagues (21) reported no data concerning adverse effects.
Table 4 summarizes the findings of the 10 studies performed to investigate the effect of acetyl-L-carnitine (ALCAR). All studies addressed the effect of ALCAR supplementation in persons with mild or moderate cognitive impairment. Here we evaluate the most remarkable characteristics of those studies. One of the six studies with positive findings included participants who were aged 40 years or older. The most recent significantly positive studies were long-term ones that had an intervention period between 6 months and 1 year. The three earlier studies had a much shorter intervention period (4, 6, or 12 weeks). The supplemented dose in all studies ranged from 1 to 3 g/day. Extensive cognitive tests were used as outcome measures in all studies, except for the study by Herrmann and colleagues (22). For the difference in results in the 10 ALCAR studies, no unambiguous reason can be given.
Thal and colleagues (23) reported 20 different treatment-related adverse effects, but they noted no difference in the number of serious adverse effects between the treatment groups. Thal and colleagues (24) noted no clinically significant supplement-related adverse effects in their earlier study. Herrmann and colleagues (22) found that some central nervous symptoms, such as headache and dizziness, occurred slightly more frequently under the supplement condition than under the placebo condition. Sano and colleagues (25) reported a few adverse effects possibly related to ALCAR treatment (nausea, vomiting, and abdominal discomfort). In their study, Spagnoli and colleagues (26) observed no significant difference either in the incidence rate or severity of adverse effects (agitation was the most common adverse effect) between the experimental and control groups. Rai and colleagues (27) reported nausea and vomiting as adverse effects that may have been related to ALCAR therapy. None of the patients in Arrigo's study reported adverse effects (28). In the earliest study, adverse effects (insomnia, epigastralgia, and pruritus) were limited in both incidence and intensity and did not cause discontinuation of treatment (29).
|
DISCUSSION |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
A weakness of all the reviews concerning the effectiveness of treatment is publication bias as a result of the selective nonpublication of null or negative trials. For this systematic review, we performed a MEDLINE search of the literature. We used carefully determined key words and selection criteria to define our topic. We used the criteria at different stages in the selection procedure of articles for our review. In addition to the MEDLINE search, we manually checked the references of selected articles and some reviews on nutritional supplements and cognitive functioning to yield the best possible literature search.
We reviewed all potentially eligible articles using a checklist for randomized controlled trials by the Dutch Cochrane Centre, extended with items from the Delphi list (10) to select only trials with high-quality designs. In most excluded articles, the methodologic aspects, such as randomization, blindness of treatment, and placebo treatment, were not clearly described. As a result of the selection, we found a wide variety in participants, intervention dose and duration, and applied outcome measures in the performed trials.
Because of the wide variety in the study designs, it is difficult to draw a firm conclusion about which nutrient or combination of nutrients supplied would have a positive effect on which domain of cognitive functioning. No clear pattern emerged to explain the observed positive effects.
The clearest finding was that in the multinutrient studies that did not show positive effects, cognitive functioning was tested using simple, general tests. Conversely, significantly positive effects were found in the multinutrient studies in which functioning was specifically measured on different aspects of cognitive functioning. Martin and colleagues (30) also noted this phenomenon in their review. A possible explanation is the difference between fluid and crystallized abilities. Cognition has two aspects: one is the capacity to apply information learned during the life span (crystallized abilities) and the other reflects information-processing capacities (fluid abilities). The latter are more vulnerable to changes in nutritional status and could be influenced more by B vitamin supplementation (31). Results of intervention studies, therefore, depend in part on which of the two aspects of cognition is tested. Another aspect influencing the effect is that changes in cognitive functioning by nutrition are subtle. The outcome measures included in intervention trials should be able to detect small changes over time. The use of several different outcome measures in the studies reviewed makes it difficult to interpret the results.
Another interesting aspect of the trials reviewed is the large number of articles on ALCAR, mostly funded by Sigma-Tau Pharmaceuticals (Gaithersburg, MD). A variety of nutritional components may positively influence cognitive functioning. Table 1 lists the putative mechanisms behind these results. In this review, alpha-tocopherol, inositol, and some B vitamins (thiamine, pyridoxine, folate, and B12) had a positive effect on cognitive functioning in some studies. In other studies, vitamin B12 and thiamine did not show this positive effect. In the group of ALCAR studies, six trials found significantly positive effects but the other four did not. Thal and colleagues (23,24) suggest a possible mechanism for the effects of ALCAR. ALCAR's most common function is to shuttle acetyl groups across the inner mitochondrial membrane to be available for oxidation, thereby participating in cellular energy production. In addition, ALCAR appears to play a role in the removal of toxic accumulations of fatty acids from mitochondria. Furthermore, ALCAR functions as a membrane-stabilizing agent. The best way to improve cognitive functioning in elderly persons might be to combine all positive effects of different nutrients, as we hypothesized in the introduction of this review. A proper supply of nutrients is probably more important for cognitive functioning than the availability of a single component.
Besides the selection of outcome measures, other factors that might explain the absence of effect are the size of the study population, the age of the participants, the presence of Alzheimer's disease, and the dose and duration of the intervention. Because of the heterogeneity of the trials and the lack of a clear pattern of explanatory factors, it is difficult to draw a conclusion about the effectiveness of single nutrients in improving cognitive functioning or preventing impairment in elderly persons. Therefore, we conclude that additional high-quality trials are needed to cluster the results in different components of the diet and to suggest firm conclusions about single nutrients. Future trials should be conducted in patients with mild cognitive impairment, because the window of opportunity for effective intervention may be as short as 1 year from the onset of medical symptoms, as Martin and colleagues (8) suggested. The duration of administration of the nutrient should be long enough to counteract reversible damage processes to the brain. In designing a trial, researchers should keep in mind which possible nutritional mechanism underlies their study. For the outcome measures, the most ideal situation would be that researchers in the field would use one standard neuropsychological battery of sensitive tests. Comparing studies would become easier and more accurate. Finally, dose–response relationships of nutrients should be investigated to determine an optimal dose for improving cognitive functioning.
Our results suggest that nutritional supplements may improve the cognitive functioning of elderly persons and do no harm. A great opportunity exists to conduct well-designed studies to support these findings.
Received January 16, 2004
Accepted January 22, 2004
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
