Staying active has many health benefits, including, perhaps, to help preserve memory. Alas, is this true in some more than others? Scientists led by Kumar Rajan at Rush University, Chicago, suggest as much in the August 11 JAMA Network Open. They report that among old people whose plasma total tau levels are high, cognition slipped half as much in exercisers as in sedentary folks. Exercisers whose plasma tau was low benefitted less, seeing at best a 27 percent slower decline. Measuring blood biomarkers could flag those who stand to benefit most from early intervention, the authors conclude.

  • Cognition declined most quickly in older people whose plasma tau is high.
  • The most active of them saw less decline.
  • Highly active people with low tau declined most slowly.

"The big picture is that modifiable lifestyle factors help even people at high risk," Kristine Yaffe, University of California, San Francisco, told Alzforum.

Researchers ask new questions about biomarkers by digging out blood samples from completed studies and analyzing them using recently developed assays, such as those for plasma total tau. In this way, first author Pankaja Desai correlated total plasma tau, lifestyle history, and cognitive test scores collected during the Chicago Health and Aging Project. From 1993 to 2011, CHAP followed 10,800 cognitively normal people over 65. Every three years, researchers checked in with participants, and recruited new people who met the age requirement. At in-home visits, participants completed surveys that assessed their global cognition, episodic memory, and how fast their visual perception was. About 5,700 volunteers also provided a blood sample at each visit; of these, 3,000 were tested for tau.

Desai identified a subset of 1,159 participants for whom tau data and at least two timepoints for cognitive assessments were available. Their average age was 77; 63 percent were women, 60 percent African American. At baseline, the volunteers estimated how often and for how long they had been physically active in the prior two weeks, using the U.S. Health Interview Survey. Activities spanned from hobbies such as gardening and bowling, to calisthenics and sports. Desai calculated each person’s weekly activity level and grouped them by low, medium, and high activity. The low-activity group got no exercise, the medium group exercised less than 150 minutes per week, and the high activity group exercised for longer. About one-third of the participants fell into each group.

The scientists measured levels of plasma total tau using a Quanterix SIMOA assay, and split participants into low and high baseline tau groups as per a previously identified threshold for accelerated cognitive decline of 0.40 pg/mL (Rajan et al., 2020). 

Both plasma tau and exercise correlated with baseline cognition. Volunteers with low tau who were highly or moderately active had 47 and 41 percent higher cognitive scores, respectively, than their inactive counterparts. This exercise buffer was weaker in high-tau people, with both active groups scoring 8 percent higher than the sedentary group. The authors corrected for age, race/ethnicity, sex, years of education, APOE4 status, and chronic medical conditions.

Slow the Slide. Over 10 years, global cognition slipped faster in people with high plasma total tau who were inactive (dotted line) or only moderately active (gray). People with high tau who were highly active (gold) declined at about the same rate as people with low tau who were inactive (light blue), moderately active (blue), or highly active (black). [Courtesy of Desai et al., JAMA Neurology, 2021.]

Did tau or exercise influence the speed of decline? During the triennial visits, CHAP researchers assessed the volunteers’ cognition using the Mini-Mental State Exam, East Boston Immediate Memory and Delayed Recall, and Symbol Digit Modalities Tests. Averaging the z scores for these three tests gave a global cognitive score.

In people with high tau, high or moderate activity tracked with 41 or 58 percent less global decline, respectively. The low tau group saw a smaller benefit; high activity correlated with 27 percent less slippage, medium activity with 2 percent less. “Perhaps people with high tau loads had more room to benefit from lifestyle interventions because they are at higher risk of dementia,” Desai told Alzforum. Michelle Mielke, Mayo Clinic, Rochester, Minnesota, agreed. “Someone cognitively unimpaired without much neurodegeneration will not have much change, so it will be harder to see an effect,” she said.

In this paper, the authors only provided standard deviation units, not raw test scores. In addition, Desai acknowledged that physical activity was queried only once, using a single baseline survey. “We should probably look at activity over time as an outcome to track their exercise behavior,” she told Alzforum. Yaffe would also have liked to see changes in Aβ and phospho-tau species measured, and how exercise altered biomarkers based on sex, race, and APOE status.

To see how exercise may relate to other markers of neurodegeneration, Desai is comparing it to plasma neurofilament light and glial fibrillary acidic protein concentrations in this cohort. The neuroinflammation marker GFAP appears to track with brain amyloid and rises early in the course of AD (Aug 2021 news). Mielke is curious if people with high levels of one, two, or all three of these markers may benefit the most from exercise.

The blossoming plasma biomarker field will allow retrospective analysis of many cohort studies and clinical trials so long as samples were collected along the way and properly stored. Ongoing intervention trials that incorporate an exercise element, such as the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER), may be able to correlate plasma biomarker changes to physical activity levels (Jul 2014 news). 

“We plan to analyze plasma tau, among other markers, from baseline through the upcoming 11-year follow-up,” Miia Kivipelto, Karolinska Institutet, Sweden, who leads FINGER, wrote to Alzforum (full comment below). Other intervention trials are happening around the globe, including World-Wide FINGERS in 40 countries and POINTER at six U.S. sites. Ultimately, the authors suggest, plasma tests may be used as selection criteria for clinical intervention.—Chelsea Weidman Burke


  1. This is a very interesting study on AD biomarkers and physical activity (PA). It is probably one of the first studies to longitudinally evaluate PA in the association between total tau in serum and cognitive function.

    Those persons with high levels of tau had a slower cognitive decline if they had high levels of PA. People with both low and high levels of tau had a slower cognitive decline if they engaged in medium or high levels of PA.

    The results support the possibility of using blood biomarkers in measuring the benefits associated with PA, and eventually other health behaviors, and they may help to specify target populations or to tailor interventions for trials.

    There are some limitations: This is a population-based observational cohort, so the direction of causation is unclear. Also, PA was self-reported and the intensity is not known.

    In future, other plasma AD biomarkers could be analyzed. We aim to study this in the FINGER trial and within WW-FINGERS network that evaluate multi-domain lifestyle interventions.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Mirror, Mirror on the Wall, Who’s the Earliest of Them All?
  2. Healthy Lives, Healthy Minds: Is it Really True?

Paper Citations

  1. . Remote Blood Biomarkers of Longitudinal Cognitive Outcomes in a Population Study. Ann Neurol. 2020 Dec;88(6):1065-1076. Epub 2020 Sep 29 PubMed.

External Citations

  1. Chicago Health and Aging Project
  2. World-Wide FINGERS

Further Reading


  1. . Moderate intensity physical activity associates with CSF biomarkers in a cohort at risk for Alzheimer's disease. Alzheimers Dement (Amst). 2018;10:188-195. Epub 2018 Feb 6 PubMed.
  2. . Exploring the relationship between physical activity, beta-amyloid and tau: A narrative review. Ageing Res Rev. 2019 Mar;50:9-18. Epub 2019 Jan 4 PubMed.

Primary Papers

  1. . Longitudinal Association of Total Tau Concentrations and Physical Activity With Cognitive Decline in a Population Sample. JAMA Netw Open. 2021 Aug 2;4(8):e2120398. PubMed.