. A 32-year prospective study of change in body weight and incident dementia: the Honolulu-Asia Aging Study. Arch Neurol. 2005 Jan;62(1):55-60. PubMed.


Please login to recommend the paper.


  1. Observations reported from the Honolulu-Asia Aging Study (HAAS) illustrate the relationship between dementia and weight history, change in weight and body mass index (BMI), and how dementia risk factors are related to weight changes. In general, the HAAS shows that weight loss precedes a clinical diagnosis of dementia, perhaps due to the underlying disease. These data are new from the standpoint that the report is based on 34 years of follow-up, with frequent measurements of weight, BMI, and other risk factors during mid- to late life. The HAAS also addresses the role of mid- versus late-life risk factors for dementia, and finds that they have different relationships to degree of change in body weight.

    Recent previous studies have reported that a high BMI may increase risk for dementia and age-related brain changes (Gustafson et al., 2003). There are a number of reasons that can account for why some studies show a high BMI to be a risk factor for dementia and others, such as the HAAS report, may not. As indicated in the summary below, a number of aspects differ between the Swedish study and the HAAS. Perhaps the biggest reasons for differences between the results from the two studies are sex differences in absolute BMI and consideration of changes in BMI over time (in addition to other potential biological differences, e.g., estrogen). HAAS men were at least two units lighter at baseline than our sample of Swedish women, were generally not "overweight" at any measurement, and experienced significant declines in weight over the 34-year follow-up, which began in mid-life and continued through late life. In our sample of Swedish women, who were 70 years old at baseline, similar measures of mid-life to late-life weight history are not available; our women were, on average, overweight at each measurement, which is typical of this age group of white women, and there were no significant declines, on average, over the 10-year period preceding dementia onset. Another difference between the studies is potential genetic differences due to race and ethnicity.

    Taken together, these studies show that a higher dementia risk may be related to a higher BMI measured during mid-life to age 70 years, a sufficient length of time, e.g., 10 years, prior to a dementia diagnosis, before the (separate) decline in body weight that often precipitates dementia.

    Stewart et al.
    Japanese American men
    34-year follow-up
    Born 1900-1919
    46-68 years at baseline, 77-98 at follow-up
    Screening process to select those who receive complete dementia evaluation
    Results applicable to Japanese-American men
    Baseline exams in 1965-68 during midlife
    Last follow-up in 1999
    Mean BMI is at baseline is about 23.8-24 kg/m2
    Mean BMI is at last follow-up is about 21.5-22.5 kg/m2
    Primary analysis is change in weight or BMI over time related to dementia

    Gustafson et al.
    Swedish women and men
    18-year follow-up
    Born 1901-02
    70 years at baseline, and 88 at last follow-up
    All participants completed neuropsychiatric examinations, and are thus eligible for a dementia diagnosis
    Results applicable to northern European women; not enough men for evaluation
    Baseline exams in 1970 at late life
    Last follow-up in 1988
    Mean BMI is at baseline is about 26 kg/m2
    Mean BMI at last follow-up is about 25-30 kg/m2
    Primary analysis is related to the role of BMI as a risk factor for dementia at least 10 years preceding a dementia diagnosis; there was no difference in BMI in cross-sectional analyses between those who were and were not demented in the Swedish sample


    . An 18-year follow-up of overweight and risk of Alzheimer disease. Arch Intern Med. 2003 Jul 14;163(13):1524-8. PubMed.

    . Body mass index, other cardiovascular risk factors, and hospitalization for dementia. Arch Intern Med. 2005 Feb 14;165(3):321-6. PubMed.

    . A 24-year follow-up of body mass index and cerebral atrophy. Neurology. 2004 Nov 23;63(10):1876-81. PubMed.

    . Body mass index and white matter lesions in elderly women. An 18-year longitudinal study. Int Psychogeriatr. 2004 Sep;16(3):327-36. PubMed.

  2. Weight loss is common among persons with dementia and Alzheimer disease (AD). Nevertheless, it has remained unclear why and when AD patients lose weight. Some recent studies have suggested that weight loss may start already during the preclinical phase of AD. Previous studies had relatively short follow-up times and only a limited number of body weight assessments, making it difficult to clarify more specifically the time course of dementia-associated weight loss.

    Changes in Body Weight in the HAAS
    Steward and colleagues studied this issue using the well-described cohort of the Honolulu-Asia Aging Study (HAAS). The study has several strengths, including a large population sample, extensive long-term follow-up (32 years), and multiple body weight measurements and cognitive evaluations. Thus, it was possible to analyze mid-life and late-life body mass trends among those who developed dementia at old age vs. the non-demented. The study indicated that men who developed dementia started losing weight several years prior to the clinical diagnosis. The weight loss was independent of a large number of potential confounding factors. A more extreme weight loss (5 kg or more) during the six years prior to diagnosis was much more common among those who developed dementia compared to the non-demented men (57 vs. 35 percent). Further, weight loss accelerated in the three-year period prior to diagnosis.

    Possible Mechanisms—Role of ApoE?
    The exact mechanisms responsible for these findings remain unclear. It is possible that weight loss is a consequence of underlying neurodegenerative processes and changes in cognition, behavior, and appetite. We have previously reported that women with AD carrying the ApoE4 allele were more likely to lose weight than were the non-carriers (Vanhanen et al., 2001). AD pathology is known to impair olfactory function early in AD, particularly among the ApoE4 carriers (e.g., Bacon et al., 1998; Wang et al., 2002), which could explain this association. However, Steward and colleagues found that the rate of weight change was, if anything, weaker in AD cases with the ApoE4 allele. This difference between the HAAS study and our results may be partly due to the differences in genetic, demographic, and lifestyle characteristics of the populations.

    First, the HAAS study included only men, and in our study, the ApoE-weight loss association was detected only among women. Body fat distribution and metabolism differ between men and women, and there may be gender-specific differences in pathophysiological mechanisms involved in dementia-related weight loss. It is also possible that changes in BMI may measure different things and have a different meaning in men and women.

    Second, levels of obesity were low in the HAAS, and the mean BMI (23.9 kg/m2) was much lower than in many Western countries. Also the ApoE4 allele frequency was relatively low, and actually did not differ much among the demented vs. non-demented persons (21 vs. 18 percent). The contribution of the ApoE4 allele for dementia may decline among the elderly cohorts that the HAAS represents (mean age at the time of the dementia diagnoses was 83 years), and thus, finally, selective survival may affect the results, also. In this context, it is important to bear in mind that our Finnish cohort was younger (mean age 73 years), and the follow-up time was much shorter (mean 3.5 years).

    Mid-life BMI and Risk of Dementia
    Steward and colleagues did not find any difference in BMI in middle age between men who later developed vs. did not develop dementia in the current study. However, an earlier report from the HAAS indicated that a higher cardiovascular metabolic risk factor burden, including high BMI in midlife, increased the risk of dementia 25 years later (Kalmijn et al., 2000). These differences may be related to the previously described factors (age and BMI distribution, survival, etc.). Also, in the Finnish cohort study where data about mid-life risk factor was available, high BMI at mid-life was associated with an increased risk of dementia later in life (mean follow-up 21 years) (Kivipelto et al., 2004).

    Implications and Future Directions
    The processes leading to weight loss in AD are likely to be complex and merit further investigation. It is important to study these issues in different populations, including both men and women, and having wider BMI distributions, and also in different age groups.

    The current findings highlight the fact that primary prevention and treatment strategies for AD may be quite different. Moreover, strategies for primary vs. secondary prevention and recommendations for various age groups (mid-life vs. late life) may be somewhat different. Avoiding being overweight prior to entering one’s later years might delay AD. In contrast, weight loss among the elderly may be a sign of impending dementia, and if it occurs concurrently with disease onset or later, it may accelerate disease progression. Accordingly, measurements of body weight and nutritional status can be considered to be important for evaluations not only of persons with manifest dementia, but also persons with mild cognitive impairment. It seems possible that dietary interventions may prevent weight loss and even delay cognitive decline and mortality (Riviere et al., 2001). However, to what extent weight loss may be prevented in dementia, and how much this would affect the disease course, need to be clarified further in controlled studies.

    See also:

    Kivipelto M. Body-mass index, clustering of vascular factors and the risk of dementia: A longitudinal, population-based study. The 9th International Conference on Alzheimer’s Disease and Related Disorders. Philadelphia 2004.


    . APOE-epsilon4 is associated with weight loss in women with AD: a population-based study. Neurology. 2001 Mar 13;56(5):655-9. PubMed.

    . Very early changes in olfactory functioning due to Alzheimer's disease and the role of apolipoprotein E in olfaction. Ann N Y Acad Sci. 1998 Nov 30;855:723-31. PubMed.

    . Olfactory identification and apolipoprotein E epsilon 4 allele in mild cognitive impairment. Brain Res. 2002 Sep 27;951(1):77-81. PubMed.

    . Metabolic cardiovascular syndrome and risk of dementia in Japanese-American elderly men. The Honolulu-Asia aging study. Arterioscler Thromb Vasc Biol. 2000 Oct;20(10):2255-60. PubMed.

    . A nutritional education program could prevent weight loss and slow cognitive decline in Alzheimer's disease. J Nutr Health Aging. 2001;5(4):295-9. PubMed.

  3. Several studies have found that in the elderly, low weight is associated with onset of dementia [1]. Other studies have shown that high body mass index in middle age [2] and in the elderly [3] are associated with a higher risk of dementia. These discrepancies may be explained by changes in weight that occur after middle age, accompanied by attrition and resulting biases in elderly cohorts. It can also be explained by weight loss associated with preclinical dementia.

    The study by Stewart et al. is an important contribution to the field. It shows that men who eventually developed dementia in the Honolulu Asia Aging Study were more likely to have weight loss in old age but not in middle age. What does this mean? It could represent an increase in metabolic rate associated with the onset of a neurodegenerative process, and thus a potential clinical predictor of dementia onset, but it could also represent a marker of a process that is a cause of dementia. One possible candidate is hyperinsulinemia. Hyperinsulinemia has been shown to be associated with a higher risk of dementia [4,5], and may precede weight loss [6], and it would be interesting for the authors to explore this association.

    The study also showed that weight in middle age was not different in men who developed dementia compared to men who did not develop dementia, contrary to what other studies have shown. However, it is important to point out, as the authors do in the manuscript, that the participants in the Honolulu Asia Aging study were men of Asian ancestry, and these results may not be generalizable to other populations.


    . Body mass index and incidence of dementia: the PAQUID study. Neurology. 2003 Jan 14;60(1):117-9. PubMed.

    . Midlife vascular risk factors and Alzheimer's disease in later life: longitudinal, population based study. BMJ. 2001 Jun 16;322(7300):1447-51. PubMed.

    . An 18-year follow-up of overweight and risk of Alzheimer disease. Arch Intern Med. 2003 Jul 14;163(13):1524-8. PubMed.

    . Fasting insulin and incident dementia in an elderly population of Japanese-American men. Neurology. 2004 Jul 27;63(2):228-33. PubMed.

    . Hyperinsulinemia and risk of Alzheimer disease. Neurology. 2004 Oct 12;63(7):1187-92. PubMed.

    . Insulin resistance precedes weight loss in adults without diabetes : the Rancho Bernardo Study. Am J Epidemiol. 2001 Jun 15;153(12):1199-205. PubMed.

Make a Comment

To make a comment you must login or register.