A 2015 epidemiological study raised eyebrows when it claimed that people with the lowest body weight ran the highest risk of Alzheimer’s disease. The observation seemed to fly in the face of multiple independent reports linking midlife obesity to a greater risk of developing the disease later on. The study did not address cause and effect, leading critics to point out that the low body weight might indeed have been caused by the disease, not the other way around. A new study in the May 9 Journal of Clinical Endocrinology and Metabolism lends credence to this interpretation. Researchers led by Ruth Frikke-Schmidt at Copenhagen University Hospital, Denmark, looked for a link between Alzheimer’s disease and genetically determined body weight in cohorts totaling about 400,000 people. They found no relationship, suggesting that body weight by itself does not affect AD risk. Weight loss during early stages of this slowly developing disease may explain the apparent higher AD incidence among thinner people, Frikke-Schmidt suggested. “Testing causality of risk factors is pivotal before considering changing public health recommendations based on observational data alone,” she told Alzforum.

The size of the earlier study had sparked debate about public health guidelines. Its analysis of two million older adults in the U.K. suggested that skinnier people ran twice the AD risk of their heavier compatriots (see Apr 2015 news). Muddying the issue, previous data had been mixed. Some reported a protective effect from obesity (see Stewart et al., 2005; Fitzpatrick et al., 2009Strand et al., 2013). However, numerous other epidemiological studies claimed that obesity in midlife correlated with a higher AD risk in old age (see Sep 2015 news; Kivipelto et al., 2005; Xu et al., 2011).

To try to clarify the role of body weight in AD, first author Liv Tybjærg Nordestgaard analyzed 95,578 participants in the Copenhagen General Population Study. Over a period of 37 years, 645 of them were diagnosed with AD. When Nordestgaard and colleagues stratified this population by body mass index (BMI), they saw the same association found in the U.K. study—those in the lowest-weight quartile ran about twofold the risk of AD as the heaviest participants.

Did low weight bring on the disease? The authors picked five genes that have been shown to have the strongest effects on body mass: FTO, MC4R, TMEM18, BDNF, and GNPDA2. They reasoned that any linkage between variants that keep people slim and Alzheimer’s disease would support a causal effect of low BMI. Such Mendelian randomization studies approximate a randomized clinical trial, Frikke-Schmidt noted. “It’s a very clean measure, because the randomization is not confounded by other factors,” she told Alzforum.

In contrast to the observational data, common genetic variants that lead to low, or even high, BMI did not budge the risk of AD. To confirm BMI has no effect, the authors also analyzed data from two large genetic consortia, the Genetic Investigation of Anthropometric Traits (GIANT) and the International Genomics of Alzheimer’s Project (IGAP), which together comprise 303,958 participants. In these cohorts, the researchers looked at 32 genetic variants that influence body mass, and again saw no relationship to Alzheimer’s.

As a positive control, the researchers looked for a link between BMI and the risk of Type 2 diabetes. Obesity is a known risk factor for this disorder. Supporting this, genetic variants that lower BMI protected against diabetes in the Copenhagen cohort.

If lifelong body mass does not affect AD, what explains the higher rates of diagnosis observed among those with low BMI? Frikke-Schmidt noted two main possibilities. Either a confounding factor—something associated with low BMI—is the true source of AD risk, or reverse causality is at play, i.e. the AD itself causes the weight loss. Based on other evidence, Frikke-Schmidt considers the second idea most likely. “We know changes happen in the brain before an overt diagnosis of AD, and these may affect appetite,” she said.

Supporting this, an April 27 Scientific Reports paper detailed how a person’s weight tends to fall in preclinical stages of AD. Researchers led by Christoph Laske at the German Center for Neurodegenerative Diseases analyzed cross-sectional data from 120 cognitively healthy people who carry familial AD mutations and 110 non-carriers in the Dominantly Inherited Alzheimer Network (DIAN). They saw lower BMIs among carriers than non-carriers as early as 18 years before expected symptom onset. The difference was greater the closer people approached expected symptom onset, becoming significant 11 years prior. “Taking these findings together, it seems most likely that body mass decline is not a risk factor, but an early symptom in the preclinical stage of AD,” Laske wrote to Alzforum. He said BMI seems to decline even earlier than cognition in familial AD.

This does not explain why other studies have found a link between midlife obesity and dementia. Frikke-Schmidt noted that many of those studies looked at all-cause dementia, rather than AD specifically. She speculated that obesity may contribute to vascular dementia, mixed dementia, and AD through vascular mechanisms (see Jul 2016 news). These likely include diabetes and hypertension, both of which are known risk factors for Alzheimer’s. Frikke-Schmidt is now examining potential links between vascular risk factors and AD in the Copenhagen cohort.—Madolyn Bowman Rogers


  1. We were also puzzled about the apparent increased dementia risk for those with low BMI, and have written about this in a correspondence letter in Lancet Diabetes & Endocrinology in 2015 (see Strand et al., 2015). 

    We followed this up in a full paper, recently published in the Journals of Gerontology, where we found that BMI loss in midlife was predictive for dementia-related mortality, while maintaining a low (or high) BMI was not related to any increased risk (Strand et al., 2017). 

    The findings of no causal relationship between stable low BMI and Alzheimer´s disease in this paper by Frikke-Schmidt et al., wherein they apply Mendelian randomization to avoid confounding and reverse causality, is intriguing, and fits with our findings described above. Thus, the increased AD risk for those with low BMI in several observational studies might be due to reverse causality. Their results fit with ours; those with stable low BMI through midlife (35-62 years) were not at increased risk of dementia-related mortality, while only those with decreased BMI in midlife were at increased risk, up to four decades later. Gaining weight reduced the risk in our analyses.  Why weight loss is related to dementia is poorly understood, and there is an ongoing discussion on this issue. Reverse causality has been suggested, especially in studies where BMI is measured in old age. Our study members were measured at baseline when they were 35-49 years and then followed up five-13 years later for a second BMI measurement at ages 40-62 years. We argued our results were less prone to reverse causality due to young age and very long follow-up time (as long as 35 years follow-up time after the second BMI measure), but we cannot rule it out even in this fairly young population. This new paper, using Mendelian randomization, strongly supports reverse causality as the main driver linking low weight with AD.


    . Does midlife obesity really lower dementia risk?. Lancet Diabetes Endocrinol. 2015 Jul;3(7):498-9. PubMed.

    . Weight Change in Midlife and Risk of Mortality From Dementia up to 35 Years Later. J Gerontol A Biol Sci Med Sci. 2016 Aug 10; PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Pound-for-Pound Protection: Does Obesity Guard Against Dementia?
  2. Extra Weight in Midlife Hastens Onset of Alzheimer’s Disease
  3. LOAD of Data Place Vascular Malfunction as Earliest Event in Alzheimer’s

Paper Citations

  1. . 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.
  2. . Midlife and late-life obesity and the risk of dementia: cardiovascular health study. Arch Neurol. 2009 Mar;66(3):336-42. PubMed.
  3. . Midlife vascular risk factors and their association with dementia deaths: results from a Norwegian prospective study followed up for 35 years. J Neurol Sci. 2013 Jan 15;324(1-2):124-30. PubMed.
  4. . Obesity and vascular risk factors at midlife and the risk of dementia and Alzheimer disease. Arch Neurol. 2005 Oct;62(10):1556-60. PubMed.
  5. . Midlife overweight and obesity increase late-life dementia risk: a population-based twin study. Neurology. 2011 May 3;76(18):1568-74. PubMed.

External Citations

  1. obesity
  2. diabetes
  3. hypertension

Further Reading

Primary Papers

  1. . Body Mass Index and Risk of Alzheimer's Disease: A Mendelian Randomization Study of 399,536 Individuals. J Clin Endocrinol Metab. 2017 Jul 1;102(7):2310-2320. PubMed.
  2. . Decreased body mass index in the preclinical stage of autosomal dominant Alzheimer's disease. Sci Rep. 2017 Apr 27;7(1):1225. PubMed.