点击显示 收起
1 From the Departments of Gerontology (OG, MM, RB, and PB) and Gastroenterology and Clinical Nutrition (SMS), University Hospital, Nice, France, and the Departments of Public Health (SA) and Internal Medicine and Clinical Gerontology (OG and BV), University Hospital, Institut National de la Sante et de la Recherche Medicale (INSERM) U558, Toulouse, France
2 Supported by a grant from the French Ministry of Health (Programme Hospitalier de Recherche Clinique). 3 Reprints not available. Address correspondence to B Vellas, Hôpital Casselardit, Pavillon Junod, 170, Avenue de Casselardit, 31059 Toulouse Cedex 9, France. E-mail: vellas.b{at}chu-toulouse.fr.
ABSTRACT
Background: Alzheimer disease is often accompanied and worsened by malnutrition. Patterns of weight loss can differ by the patients concerned and by the outcome and interventions required.
Objective: Our aim was to describe and analyze 2 modes of weight loss (progressive and severe) in the course of Alzheimer disease.
Design: This was a prospective study of 395 patients with Alzheimer disease, who had a mean age of 75.4 y. A standardized gerontologic evaluation was conducted at 6 mo and 1 y, including assessments of nutrition, neuropsychology, function, and caregiver burden.
Results: We investigated 2 modes of weight loss. The first, progressive loss (4% in 1 y), affected 33.4% of subjects. Disease severity was a risk factor [odds ratio (OR): 7.2; 95% CI: 1.4, 38.2 for a Reisberg score 5], whereas treatment with cholinesterase inhibitors at baseline decreased this risk (OR: 0.33; 95% CI: 0.14, 0.79). The second mode of weight loss, a severe loss of 5 kg in 6 mo, affected 10.2% of subjects. The existence of an acute phase reaction was a risk factor (OR: 2.4; 95% CI: 1.2, 4.8), as was an intercurrent event, such as hospitalization, acute disease, institutionalization, and change of living arrangements (OR: 6.8; 95% CI: 1.2, 39.9).
Conclusion: During the follow-up of patients with Alzheimer disease, risk factors for these 2 modes of weight loss should be sought to identify patients who would benefit from a nutritional intervention. Our findings lead us to advocate follow-up, which involves an assessment of functional, nutritional, and neuropsychologic status every 6 mo.
Key Words: Weight loss • malnutrition • dementia • Alzheimer disease • aging • elderly • prospective study • malnutrition • nutritional assessment
INTRODUCTION
Alzheimer disease (AD) is a neurodegenerative condition that is characterized by a progressive loss of cognitive function. It is the most frequent type of dementia, affecting 30-50% of persons >85 y (1). An interventional strategy to prevent or delay the complications of this disease would have an important public health effect in industrialized societies (2).
Weight loss is a frequent complication of AD and occurs in 40% of patients at all stages, even in the early stages before diagnosis is possible (3). Malnutrition (namely undernutrition) contributes to the alteration of general health status, to the frequency and gravity of complications, especially infections, and to a faster loss of independence. These states of malnutrition can be prevented or at least improved if an early intervention strategy is set up, but management must be rapid and appropriate (4). In a geriatric population, a loss of >4% of body weight in 1 y is known to be an independent factor of morbidity and mortality (5). In addition, weight loss is a phenomenon whose kinetics may vary; it can be a dramatic loss of several kilograms in a few months (severe weight loss) or a moderate but continuous loss as the disease progresses (progressive weight loss).
The aims of this prospective study were to identify patients with each of these 2 modes of weight loss in a large AD population with a standardized nutritional follow-up and to determine the risk factors for these patterns of weight loss.
SUBJECTS AND METHODS
Study design
This single-center, prospective cohort study was conducted in the hospitals of Toulouse, France, as part of the Etude Longitudinale de Suivi de la Maladie d'Alzheimer (ELSA) study. The ELSA study has followed patients with AD since 1994. Follow-up included a clinical evaluation that was completed every 6 mo in the daycare hospital. The collation of intercurrent events by telephone was every 3 mo.
The diagnosis of AD was confirmed by initial evaluation through a detailed personal and family history and the administration of psychometric tests to confirm cognitive impairment. Conventional tests could also be used to exclude other causes of dementia (standard biological tests, thyroid function, vitamin measurements, brain computed tomography scan, and, in some cases, brain perfusion radioisotope scan and apolipoprotein E genotyping). All patients fulfilled the criteria for dementia as described in the Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition (6) and had probable AD according to the criteria of the National Institute of Neurological and Communicative Disorders and Strokes–Alzheimer's Disease and Related Disorders Association (7).
The presence of a caregiver able to ensure the quality of follow-up was an inclusion criterion. Exclusion criteria were diagnosis of AD >5 y before the study, an activities of daily living (ADL) score < 3, uncontrolled heart failure, severe or unstable angina, uncontrolled arterial hypertension, severe orthostatic hypotension, severe renal or liver failure, severe anemia, vascular disorders, systemic disease, clinically relevant hyperthyroidism or hypothyroidism, clinically relevant vitamin deficiency, concomitant malignancy, severe or total blindness, or deafness. The study was approved by the local research ethics board, CCPPRB Haute-Garonne.
Evaluation
The prestudy and 6-mo evaluations were multidimensional and multidisciplinary. The patients underwent clinical, neuropsychological, and biological investigations. The 6-mo evaluations included the following. A cognitive evaluation was conducted by a neuropsychologist who used Folstein's Mini-Mental State Examination (8). Independence was evaluated by interviewing the family with the use of the ADL scale (9) and the Instrumental Activities of Daily Living (IADL) scale (10). Mood was scored with the use of Cornell's depression scale (11). Behavior was assessed with the Cohen-Mansfield scale (12). A full nutritional evaluation included clinical [weight and body mass index (BMI; in kg/m2)] and biological variables [albumin, prealbumin, orosomucoid, C-reactive protein (CRP), prognostic inflammatory and nutritional index (PINI); (CRP x orosomucoid/albumin x prealbumin) (13)]. Nutritional status was quantified with the Mini-Nutritional Assessment (MNA) (14, 15). A dietitian also administered a dietary questionnaire to assess qualitative and quantitative food intakes. The caregivers were interviewed for assessment of their material and psychological burden (burden interview, 22 items) and their reactions to the patients' behavioral problems and increased dependence (Memory and Behavior Problem Checklist, 17 items) or the Zarit scale (16).
The multiple components of the 6-mo evaluations were used to classify the patients according to the Reisberg scale (17), which indicates the stage of AD. All intercurrent events, whether medical (hospitalization, visit to the treating physician, hospital consultation, and the reason thereof) or nonmedical (social in particular), were carefully recorded. Enrollment was completed in 2000; 395 patients were included.
Classification of weight loss
Two modes of weight loss were examined: progressive weight loss and severe weight loss. A progressive weight loss was determined by using a cutoff of 4% weight loss in 1 y in accordance with the works of Wallace et al (3), who found weight loss of this magnitude to be an independent risk factor of increased mortality in an elderly population. Severe weight loss was determined by using a cutoff of 5-kg weight loss in 6 mo. This pattern of weight loss was selected in accordance with the works of Guyonnet et al (18), who determined that there are 2 types of weight loss in AD patients: a slowly progressive and a severe weight loss of >10%/y (which was >5 kg for most subjects in this study). This loss of 5 kg was routinely observed as an important intercurrent event between 2 visits by the physicians who followed patients in the ELSA cohort, and it was always reported.
Statistical analysis
We individualized 2 groups of subjects: those who had lost 4% of body weight during the first year, and those who had lost 5 kg during the first 6 mo of follow-up. We also determined the factors at inclusion that were associated with each of these 2 modes of weight loss in the course of AD.
Student's t tests were used to compare normally distributed quantitative variables across groups. When the hypothesis of normality was not met, nonparametric Kruskal-Wallis tests were used. For qualitative variables, chi-square or Fisher's exact tests were performed as required in each category. Multivariate logistic regression was used to identify the factors independently associated with weight loss. The selection criteria for the variables introduced in the abovementioned multivariate models were a critical probability of P < 0.20 after bivariate analysis. When several strongly linked variables met this criterion, a single variable was selected to avoid the risk of multicolinearity. Statistical analyses of the data were carried out with the SAS statistical software package (version 8.0.2; SAS Institute, Cary, NC).
RESULTS
Population characteristics at inclusion
Of the 434 patients who fulfilled the inclusion criteria, 395 agreed to participate. Of these 395 patients, the weight of 341 subjects was known at 6 mo and that of 308 subjects at 1 y. During follow-up, 28 subjects left the study at the request of their family and 19 died. The flow diagram of the study is represented in Figure 1 (19). The characteristics of the population at inclusion in the ELSA study are represented in Table 1. Patients were on average in their 70s and had a normal nutritional status according to clinical and biological markers and to the MNA. More than one-half had normal ADL scores, indicating full autonomy, and their depression and behavioral disorders scores were low. Most had a moderate cognitive impairment, and the severity of their AD was also moderate. One of 3 patients received cholinesterase inhibitors for their AD.
FIGURE 1.. Flow diagram of the study according to the Consolidated Standards of Reporting Trials (CONSORT) statement.
View this table:
TABLE 1. Characteristics of the population at inclusion1
Progressive weight loss
Progressive weight loss (4% weight loss in the first year of follow-up) was observed in 33.4% of subjects (n = 97) at 1 y.
Bivariate analysis
The variables that were different between this group of patients and the rest of the initial cohort are shown in Table 2. At baseline, patients with a progressive weight loss had a more severe AD (Reisberg scale) and a more pronounced cognitive impairment (Mini-Mental State Examination). The 1-point difference between groups in the mean IADL score for everyday activities is extremely relevant in clinical practice. The rate of admission to a hospital or institution was higher in these patients, as was the burden for their caregivers. Finally, treatment with cholinesterase inhibitors was inversely associated with progressive weight loss. The presence of an intercurrent event (besides hospitalization or institutionalization), depression, behavioral disorders, living arrangement, food intake, and nutritional status at inclusion was not different between patients with a progressive weight loss and the others.
View this table:
TABLE 2. Comparison of variables at inclusion between groups with and without a 4% weight loss in 1 y (bivariate analysis)1
Multivariate analysis
After multivariate analysis, 3 variables were independently associated with a progressive weight loss: a higher initial weight, the absence of a specific treatment of AD, and more severe AD (Reisberg score) (Table 3).
View this table:
TABLE 3. Comparison of variables at inclusion between groups with and without a 4% weight loss in 1 y
Severe weight loss
A loss of 5 kg body weight during the first 6 mo of follow-up occurred in 10.2% of the subjects (n = 32). Twenty (62.5%) of these patients were also part of the first group (progressive weight loss at 1 y).
Bivariate analysis
The variables that were different between this group of patients and the rest of the initial cohort are shown in Table 4. At baseline, patients with a severe weight loss had a higher initial weight and BMI and had lower prealbumin and higher CRP values. Their functional status was lower, and their educational level was higher. The rate of admission to a hospital or institution was higher in these patients, as was the burden for their caregivers. Living arrangements differed significantly between the 2 groups: patients who lost weight generally lived at home alone or in an institution. The severity of AD, cognitive status, depression, and food intake at inclusion were not different between patients with and without severe weight loss.
View this table:
TABLE 4. Comparison of variables at inclusion between groups with and without a 5-kg weight loss during the first 6 mo (bivariate analysis)1
Multivariate analysis
After multivariate analysis, 3 variables were independently associated with a severe weight loss at 6 mo (Table 5). These variables included a higher initial weight, a higher PINI score, and the presence of an intercurrent event (including hospitalization, acute disease, institutionalization, and change of living arrangements).
View this table:
TABLE 5. Comparison of variables at inclusion between groups with and without a 5-kg weight loss during the first 6 mo1
DISCUSSION
To our knowledge this was the first study to determine the effect of various nutritional variables (weight, MNA, albumin, prealbumin, orosomucoid, CRP, and PINI), as part of a standardized systematic follow-up procedure, on the progression of AD. Malnutrition is a feature of the original description of AD as reported by Aloïs Alzheimer in 1901. Afterward, in the early 1980s, Morgan and Hullin (20) reported that the weight of patients hospitalized for dementia was lower than that of healthy elderly subjects. White et al (21) explored the association between AD and weight changes in 362 subjects with AD and 317 control subjects for 2 y. The results showed that nearly twice as many subjects with AD experienced weight loss of 5% than did control subjects. Weight loss in AD can follow 2 patterns that are relevant to elderly people (22): progressive and severe.
In our cohort, more than one-third of subjects lost progressively (over 1 y) >4% of their body weight. The disease per se seemed to be the most important risk factor for this mode of weight loss, because patients with the most severe forms of AD were 6-7 times more likely to undergo a progressive weight loss. These results confirm the findings of White et al (23) who explored the association between progression of AD and weight loss in 666 subjects. They showed that changes in stage of AD correlated significantly with weight changes. Some patients with AD show a rapid cognitive decline (24) with risk factors (worse cognitive status, greater BMI, and less anxiety of burden) that are similar to risk factors for progressive weight loss (worse cognitive status and greater initial weight). Future studies are required to compare individuals with the different patterns of weight loss to identify common features. A higher initial body weight may confirm the relation that has been described between obesity and AD (25), maybe through the association of cardiovascular risk factors and AD, even though a small proportion of our patients were obese. A more surprising finding was the identification of treatment with cholinesterase inhibitor as a protective factor against progressive weight loss, even though we cannot attribute a cause-and-effect relation. Indeed, this study was not a controlled study of cholinesterase inhibitors, which have been proved effective against cognitive decline and behavioral disorders in patients with AD (26). Prevention of progressive weight loss may be part of the benefits of these drugs in AD. Unlike other investigators (27), we did not identify behavioral disorders (including anorexia) as risk factors for weight loss in AD. We observed no differences in food intake according to the pattern of weight loss. Similar findings have been reported, and those findings leave the question of anorexia and energy expenditure in AD intact (28). This disease-related weight loss can lead to cachexia at the end of the disease's evolution, and at that stage nutritional support has been much debated and is generally not advocated (29). However, an earlier intervention might allow improvement of the nutritional status, thus justifying an early diagnosis of this progressive weight loss (4).
Ten percent of the cohort underwent a severe weight loss of >5 kg in 6 mo. This was the first time that such a weight loss had been prospectively quantified. The independent effect of a higher initial body weight on severe weight loss may be explained because we chose 5 kg as an absolute value rather than a percentage of weight. This severe weight loss was likely due to a concomitant disease, as shown by the frequency of an inflammatory syndrome or intercurrent event (health-related or social) in these subjects. Stress, especially in its medical meaning, is reflected here by a higher number of intercurrent events and by an elevated PINI, which takes into account at the same time the acute phase protein reaction (with a switch in liver protein synthesis as a result of the action of proinflammatory cytokines such as tumor necrosis factor and interleukin 6) and the ensuing decrease in nutritional proteins. In elderly hospitalized patients, this prognostic index has proved to be helpful to predict both nearest lethality and chronic institutionalization (30). Cytokine-driven anorexia is a common symptom during an acute stress (such as infection, trauma, surgery) and is usually reversible within a few days after the course of the initial disease. However, this is only true for children and younger adults, and the concept of secondary anorexia is well known in the elderly who are unable to develop hyperphagia after a period of food restriction (31). The ensuing malnutrition-induced morbidity and mortality may superimpose those related to AD (32). Nutritional care is crucial in elderly patients with secondary anorexia, because it can result in a return to previous food intake and a subsequent regaining of the weight that was lost (33). Therefore, we believe that screening for this condition (acute weight loss) is important because it may lead to early and beneficial nutritional care. This needs to be confirmed by an interventional study in such patients.
Twenty patients (62.5%) with a severe weight loss were also part of the group with a progressive weight loss (n = 97). We did not exclude those patients from the analysis because all patients with AD can experience an acute stress: approximately one third will recover part or all of the weight lost, whereas approximately two thirds will evolve toward a disease-related progressive weight loss.
In conclusion, 2 distinct modes of weight loss with different risk factors were identified in patients with AD. The first mode, which occurs quite frequently, is a progressive loss, which is related to the evolution and worsening of AD. The second mode, which occurs less frequently, is a severe loss, which is related to intercurrent medical or social events. Because the prognosis of these 2 modes of weight loss differ, with a nutritional support potentially being useless if provided too late in severe weight loss, we believe that a rigorous, early, and regular follow-up of nutritional variables is mandatory in patients with AD.
ACKNOWLEDGMENTS
We thank Mélanie Koning and Christelle Cantet for assistance with the analysis and Sophie Gillette-Guyonnet and Ian Janssen for their review of the manuscript and much other advice.
SA and BV were responsible for the study concept and design. OG, SA, and BV were responsible for the analysis and interpretation of the data. OG, SA, SMS, and BV drafted the manuscript. All authors were responsible for the critical revision of the manuscript for important intellectual content. SA provided statistical expertise. BV obtained funding. MM, RB, and PB provided administrative, technical, or material support. BV supervised the conduct of the study. None of the authors had a conflict of interest.
REFERENCES