visceral fat loss improves cognitive function

 this is the second blog on the clinical effects of decreasing visceral fat, this one finding improved brain architecture and cognitive function (see visceral fat loss inc cognitive function NatCommunic2026 or doi.org/10.1038/s41467-026-71141-4

    --a link to the last blog on cardiometabolic outcomes: https://gmodestmedblogs.blogspot.com/2026/07/visceral-fat-loss-is-sustained-long-term.html

Details:

-- this Follow-Interventions-Trials (FIT) project recruited 533 adults from four prior 18-24-month randomized clinical-trials assessing the effects of lifestyle changes on fat depots in the body

    -- DIRECT trial: 322 moderately obese participants randomly assigned to a diet of a low-fat restricted-calorie, a Mediterranean diet, a restricted-calorie diet, or a low carbohydrate but non-restricted-calorie diet, finding that the Mediterranean diet group had improved lipids, and those with diabetes had improved plasma glucose and insulin: https://www.nejm.org/doi/full/10.1056/NEJMoa0708681

    -- CASCADE trial: 48 participants randomized to Mediterranean diet and drinking 150ml of either mineral water, white wine, or red wine; all had similar weight loss, and there was not much difference in visceral adipose tissue, or superficial or deep subcutaneous adipose tissue: https://pubmed.ncbi.nlm.nih.gov/27692023/

    -- CENTRAL trial: 278 participants, finding that those on a Mediterranean diet and exercise program did better than those on a low-fat diet alone in decreasing intrahepatic, intrapericardial, and pancreatic fats and as well as having improved lipid profile and insulin sensitivity:  https://www.ahajournals.org/doi/10.1161/circulationaha.117.030501

    -- DIRECT-PLUS trial: 294 participants, finding that a green-Mediterranean diet (with 3–4 cups/day of green tea and Wolffia globosa (duckweed strain) plant green shakes (100 g frozen cubes/day) + 800mg/day polyphenols and reduced red meat intake doubled their VAT loss vs a regular Mediterranean diet

        -- the latter 2 studies were used in the last blog

-- these individuals had repeated visceral adipose tissue (VAT) assessments and underwent abdominal/brain magnetic-resonance-imaging (MRIs) and Montreal-Cognitive-Assessments (MoCA), with subsequent testing 5–16y after the interventions.

-- 647 individuals were identified for this FIT study, of whom 533 (89%) had brain MRI structural analysis

-- dietary follow-up: mean Mediterranean Diet adherence score overall was 7.0 (scores range from 0-14, 7 representing moderate adherence)

-- overall 86% male; rest of data in the Results section below

-- main outcome: the relationship between changes in abdominal adipose tissue depots with both structural changes in brain architecture per MRI scans as well as association with cognitive ability per MoCA evaluation

-- secondary outcomes included demographic parameters, anthropometric and blood biomarker measurements

Results:

-- the following are the baseline (preintervention) characteristics, and the 5-16 year follow-up assessments (lots of data, interpretation below)

    -- mean age:

        -- baseline: 52yo

        -- follow-up: 61yo

    -- whole brain, % of intracranial volume:

        -- baseline: 74.24%

        -- follow-up: 72.16%

    -- gray matter, % of intracranial volume:

        -- baseline: 30.61%

        -- follow-up: 29.62%

    -- white matter, % of intracranial volume:

        -- baseline: 30.82%

        -- follow-up: 30.04%

    -- total ventricles, % of intracranial volume:

        -- baseline: 1.73%

        -- follow-up: 2.46%

    -- hippocampal occupancy, % of intracranial volume:

        -- baseline: 0.86%

        -- follow-up: 0.79%

-- MoCA (Montreal Cognitive Assessment):

    -- baseline not done

    -- follow-up 23.56 (mild cognitive impairment)

-- VAT (visceral adipose tissue) area:

    -- baseline: 139.32 cm²

    -- follow-up: 130.96 cm²

-- Deep SAT (subcutaneous adipose tissue) area:

    -- baseline: 227.12 cm²

    -- follow-up: 196.92 cm²

-- superficial SAT area:

    -- baseline: 125.13 cm²

    -- follow-up: 114.77 cm²

-- VAT portion percent:

    -- baseline: 28.67%

    -- follow-up: 29.86%

-- BMI:

    -- baseline: 30.36

    -- follow-up: 30.10

-- weight:

    -- baseline: 89.90 kg

    -- follow-up: 87.97 kg

-- waist circumference:

    -- baseline: 106.43 cm

    -- follow-up: 102.48 cm

-- systolic blood pressure:

    -- baseline: 129.16 mmHg

    -- follow-up: 133.98 mmHg

-- diastolic blood pressure:

    -- baseline: 79.65 mmHg

    -- follow-up: 77.65 mmHg

-- physical activity:

    -- baseline: 29.73 MET-h/week

    -- follow-up: 41.61 MET-h/week

-- fasting glucose:

    -- baseline: 108.87 mg/dL

    -- follow-up: 107.98 mg/dL

-- fasting insulin:

    -- baseline: 14.76 mcg/dL

    -- follow-up: 15.07 mcg/dL

-- HbA1c:

    -- baseline: 5.82%

    -- follow-up: 6.04%

-- triglycerides:

    -- baseline: 153.54 mg/dL

    -- follow-up: 149.85 mg/dL

-- HDL:

    -- baseline: 43.73 mg/dL

    -- follow-up: 46.15 mg/dL

-- LDL:

    -- baseline: 117.25mg/dL

    -- follow-up: 108.57 mg/dL

-- total cholesterol:

    -- baseline: 153.54 mg/dL

    -- follow-up: 149.85 mg/dL

-- triglycerides/HDL ratio:

    -- baseline: 3.99

    -- follow-up: 3.66

-- hsCRP:

    -- baseline: 3.83 mg/dL

    -- follow-up: 3.31 mg/dL

-- APOE ε4 alleles: at baseline 235 patients tested, 40 positive (17.02%)

-- Cross-sectional Analysis of the Interaction Between VAT and Brain Volumes in Association with Cognitive Function (MoCA) at Follow-Up:

    -- Linear regression models revealed significant interactions between VAT and several brain volume parameters in relation to MoCA scores. Significant interaction terms were:

        -- between VAT and Total Brain (VAT × TB; β = –0.30, p = 0.032)

        -- between Hippocampal Occupancy and VAT (VAT × HOC; β = −0.29, p = 0.048),

        -- between VAT and white matter (VAT × WM; β = –0.3, p = 0.028).

--interactions involving deep subcutaneous adipose tissue (DSAT) and superficial subcutaneous adipose tissue (SSAT) levels were not significantly associated with MoCA scores.

--participants with mid or high brain volumes exhibited a more pronounced inverse association between VAT andMoCA scores

    -- In contrast, among individuals with low brain volumes, MoCA scores remained relatively stable across the range of VAT values.

--adjusted for age, sex, BMI, and MED score, participants with MoCA scores of 18-25 had significantly higher VAT compared with those with a normal cognition score (β=0.21, p < 0.01).

    -- In contrast, differences in SSAT and DSAT across MoCA categories were not statistically significant (18-25 vs >=26 MoCA score: β = 0.07 for SSAT and β = 0.11 for DSAT). 

-- none of the adipose tissue depots showed significant differences between individuals with dementia-range MoCA scores (<18) and those with normal cognition

    -- in parallel models in which BMI, rather than VAT, was specified as the main adiposity exposure, BMI did not show significant interactions with brain volumes in relation to cognitive performance, either in models without VAT or in models additionally adjusted for VAT

    -- The BMI-brain interaction terms were consistently non-significant, in contrast to the significant VAT-brain interactions observed in our primary analyses

    -- In sensitivity analyses adjusted for incretin-based anti-obesity medication use (ie GLP-1 type meds), no significant associations were observed, and the inclusion of this covariate did not change the MoCA-related findings

        -- i am not sure how to interpret this, given my comments in the last visceral fat blog on cardiometabolic issues: https://gmodestmedblogs.blogspot.com/2026/07/visceral-fat-loss-is-sustained-long-term.html

        -- this current study assessed the long-term effects of the GLP-1 type meds, at a time when some of the most potent incretin meds were not available

        -- no comment/granular information on how many people were on GLP-1 type meds or what the meds/doses were, or the trend in staying on the meds (did initial use of the meds continue throughout the study, an issue since the studies relied on just the initial use and did not tract changes over the course of the studies

        -- in the supplementary materials, they did note that, adjusted for age, sex, BMI, and Mediterranean diet adherence score:

            -- Hippocampal Occupancy Score: P-value 0.048, which is statistically significant (though the confidence intervals for the statistics were wide and did cross into the nonsignficant range)

                -- the hippocampus role: memory consolidation by transferring short-term memories to long-term memory storage, ability o recall facts and events, and spatial navigation (ie, acting like the brain's internal GPS)

            -- Cerebral White Matter: P-value 0.029, which is statistically significant (though the confidence intervals for the statistics were wide and did cross into the nonsignficant range)

            -- Whole Brain: P-value 0.032, which is statistically significant (though the confidence intervals for the statistics in this analysis also were wide and did cross into the nonsignificant range)

-- Longitudinal analysis of visceral adipose tissue (VAT) change during the intervention and brain structural outcomes at follow-up:

    -- among participants from the CENTRAL and DIRECT-PLUS trials (n = 297), the mean VAT change during the 18-month intervention was after adjustment for age, baseline fat depot levels, BMI changes, MED diet adherence score, sex, and diet group:

        -- the mean VAT change during the 18-month intervention was −24% ±23.4%

        -- changes in abdominal adipose depots during the intervention period and brain structure and cognitive function at 5 and 10 years post-intervention found:

            -- greater reduction in VAT area during the intervention was associated with better preservation of brain structure at follow-up, with higher total brain volume (β = −0.13, p = 0.05), higher gray matter (β = –0.19, p = 0.01) and higher HOC score (β = −0.10, p = 0.028) 

                -- these associations remained robust after adjustment for age, baseline fat depot levels, BMI changes during intervention, Mediterranean diet adherence score, sex, and diet group.

                -- however, reductions in deep and superficial SAT during the intervention were not associated with brain structure outcomes

                -- relative change in BMI during the intervention was not significantly associated with brain structural measures at follow-up, though BMI change was negatively associated with MoCA score at follow-up (β = –0.153, p = 0.014)

                -- sex-by-intervention interaction did not reach statistical significance in any of the models

-- Interaction between longitudinal visceral adipose tissue (VAT) exposure and brain atrophy rate:

    -- for participants from the DIRECT-PLUS trial who had repeated brain MRI measurements at baseline, 18 months, and 5 years (n = 188):

        -- the association between cumulative VAT exposure and the rate of brain structural change over time, dividing participants into 3 groups (those with slow, moderate, and fast declines in brain atrophy rate, all adjusted for age, sex, BMI, and MED diet adherence score):

                -- VAT exposure and brain volume was found to be independent of the rate of brain volume decline, though

                    -- a significant effect of time was observed across all other brain volume structures:

                        -- total brain volume: β = −0.02, p < 0.001

                        -- gray matter: β = −0.01, p < 0.001

                        -- white matter: β = −0.005, p < 0.001

                        -- hippocampal occupancy: β = −0.003, p < 0.001

                        -- total ventricles: β = 0.003, p < 0.001 (larger ventricles are expected with brain shrinkage)

        -- APOE ε4 carrier status showed no significant main effect and no significant interactions with time

        -- sensitivity analyses in which baseline glycemic, lipid, and inflammatory biomarkers were individually entered as interaction terms, and adjusting for visceral adipose tissue accumulation, age, sex, and MED score, found that only fasting glucose and HbA1c remained significantly associated with longitudinal brain change. All other baseline glycemic, lipid, and inflammatory biomarkers showed no significant interactions

  

Commentary:

-- several large observational studies have also found associations between the amount of visceral adipose tissue (VAT) and adverse brain outcomes including greater brain atrophy; lower hippocampal size (crucial for generating memories), gray matter and white matter volumes, larger ventricular volume and cognitive decline

-- this current FIT study was done to assess whether long-term exposure to lower VAT, independent of weight loss, could attenuate brain atrophy and protect cognitive function

-- ectopic fat in organs including liver, pancreas, skeletal muscle, kidney, and heart were not measured; a meta-analysis of 21 studies found ectopic fat is associated with decreased brain volume and cognitive decline: see ectopic fat effect on brain and cognition FrontNeuroendo2023 in dropbox, or DOI: 10.1016/j.yfrne.2023.101082.  this all means that there is an adipose-brain axis.

-- the FIT findings in summary:

    -- MRI-assessed VAT and Brain measures over long-term interventions and follow-up found that sustained reduction in VAT, independent of weight loss, significantly attenuates brain atrophy and enhances cognitive function across a span of up to 10 years

        -- this MRI-quantification of VAT trajectories included adjustment for dietary adherence and body weight that render this work both robust and translational

            -- these findings address a critical gap by identifying visceral fat, not just obesity, as a modifiable risk factor for neurodegeneration, and they thereby present a compelling target for early intervention. In parallel analyses that substituted BMI for VAT as the primary adiposity measure, BMI showed no significant associations with brain structures, which may help explain why VAT, rather than BMI, emerges as a more specific adiposity-related risk factor in our cohort despite the expected correlation between these measures

                -- as found in the last blog, this reinforces the lack of utility of BMI as a predictor of clinical problems vs various measures that reflect visceral fat: https://gmodestmedblogs.blogspot.com/2024/09/body-roundness-index-is-better.html

                -- of course, there is a rough correlation (though erratic) between individuals with high BMI and visceral fat; but some people with high BMIs have preponderance of superficial fat depots which in some studies is actually beneficial (see last blog for details)

    -- long-term exposure to higher visceral adiposity is associated with both accelerated brain atrophy and lower cognitive performance

    -- but sustained VAT reduction, achieved through lifestyle interventions, predicted preserved brain structure and cognition in late midlife

        -- lower cumulative VAT exposure is associated with attenuation of brain atrophy, higher brain volumes, and better MoCA scores at follow-up

        -- there were also significant interactions between VAT and total brain, white matter, and hippocampal occupancy in relation to MoCA scores

            -- so, this all reinforces the importance of focusing on VAT as the modifiable cognitive risk factor

-- researcher postulate that glycemic control may represent a key pathway linking visceral adiposity to brain outcomes

    -- and, overall, their findings underscore the potential for long-lasting neuroprotective effects of targeted VAT loss

-- Likely mechanisms for the improvement in brain architecture and cognitive testing, per the FIT researchers:

    -- VAT is a metabolically active depot that secretes pro-inflammatory cytokines (e.g., IL-6, TNF-α, CRP), promoting chronic low-grade systemic inflammation that can induce neuroinflammation, disrupt blood-brain barrier integrity, and impair synaptic plasticity and neuronal survival.

    -- Sustained exposure to VAT-derived inflammatory mediators may also interfere with hippocampal neurogenesis and accelerate white matter degeneration

        -- both of these are hallmarks of age-related neurodegeneration

   -- in addition, VAT is closely linked to insulin resistance and vascular dysfunction, which contribute to impaired cerebral perfusion, microvascular damage, and reduced clearance of neurotoxic proteins such as β-amyloid

        -- the researchers found that among the examined biomarkers, fasting glucose and HbA1c were uniquely associated with longitudinal brain atrophy, whereas lipid-related and inflammatory markers showed no consistent associations over time

        -- and improvements in glycemic control are associated with attenuation of brain atrophy

-- longitudinal analyses further demonstrated that cumulative exposure to higher levels of abdominal adiposity over time is inversely associated with lower MoCA and MoCA-MIS (MoCA Memory Index Score) scores, independent of BMI, diet adherence, and brain volume

    -- participants with greater cumulative VAT levels exhibited faster rates of brain atrophy across multiple regions, including gray matter, and hippocampal occupancy: this is a cumulative process

        -- therefore, it is likely that the earlier visceral fat is detected would be associated with likely increased benefit from diet/exercise and potentially GLP-1 like meds; this result was consistent after adjusting for key confounders, supporting the notion that VAT is not only a correlate of cross-sectional brain structure but also a marker of longitudinal neurodegeneration

        -- the Health, Aging and Body Composition (ABC) study found that CT-measured VAT independently predicted global cognitive decline over several years in older adults, even after adjusting for BMI and baseline cognition: https://jamanetwork.com/journals/jamaneurology/fullarticle/1033004

Recent data from a study of 32 middle-aged (mean 51yo) with BMI of 32.28 and cognitively normal who had abdominal MRI, and amyloid and tau PET scans had their VAT/SAT ratio that was significantly associated with beta-amyloid plaques and cortical thinning in Alzheimer’s disease-related regions

but only in males: https://pmc.ncbi.nlm.nih.gov/articles/PMC11272197/

Limitations:

-- this was a predominantly male population, as noted in the prior blog, limiting generalizability to women

-- the participants were required to have an elevated BMI at study initiation, limiting generalizability to those with lower BMIs (with the caveat that BMI is not a reliable marker of overweight as per above, and about 21.7% of patients with a significant visceral fat depot have a normal BMI (18.5-24.9) globally, but still have more hypertension, diabetes, and elevated lipids see: visceral obesity with normal BMI outcomes JAMA2025in dropbox, or doi:10.1001/jamanetworkopen.2025.37942)

-- the MoCA value of 24 for people aged 61 is on the low side, also limiting generalization to the population in general; this lower MoCA value is perhaps related to the people studied in FIT, with high levels of long-standing abdominal obesity, dyslipidemia, and diabetes

-- one concern in a systematic review/meta-analysis such as this one is that it combined 4 studies, the largest having 294 patients and the smallest having 48 patients. therefore the combination will effectively undercut any real contribution of the smallest study

-- they did not find clear benefit from GLP-1 type meds (incretins); it is likely that many patients were on older GLP-1s and we do not know anything about which ones, what their doses were, how many people continued on them for the long-term (and several had many more adverse effects), limiting the legitimacy of this conclusion

-- the long-term 5-16 year follow-up was not randomized, only allowing us to have an association of the outcomes and not determining causality

-- only 2 of the 4 studies (CENTRAL and DIRECT-PLUS trials, the 2 largest ones) had baseline and end-of-intervention abdominal adipose tissue relations with brain MRIs, so outcomes were limited to the 5- and 10-year follow-ups and not longer-term assessments

-- MoCA cognitive function was not systematically assessed at the pre- or post-intervention times, limiting the cognitive assessment accuracy (?were MoCAs done predominantly in one group of patients over another, perhaps in those with more apparent cognitive issues?, was there a bias to which country the assessments were done and the associated different cultures/diets/psychosocial or psychiatric issues...). This could also limit interpretation and generalizability of the results

-- ectopic fat in organs including liver, pancreas, skeletal muscle, kidney, and heart were not measured, and a meta-analysis of 21 studies found ectopic fat is associated with decreased brain volume and cognitive decline: see ectopic fat effect on brain and cognition FrontNeuroendo2023 in dropbox, or DOI: 10.1016/j.yfrne.2023.101082. 

-- there are no specific Alzheimer's disease biomarkers in this study (eg PET scan or blood-based markers), limiting our ability to attribute cognitive decline to specific brain problems; however, of note, the APOE ε4 carrier status did not influence the results (though APOE ε4 were not systematically assessed, allowing for bias in interpretation here)

-- we have no information on the granular assessments of changes in diet and other lifestyle issues over time, in particular after the studies were completed. did some/many people continue with their diets from the studies? continue with their exercise? perhaps some of the decreased visceral fat long-term was because some/many individuals continued the intervention to maintain visceral fat loss??

so,

-- another impressive study of the relationship between visceral adipose tissue and bad outcomes, in this case both decreases in brain volume and cognition

-- a benefit of this study is that they assessed these results many years after the formal studies ended

-- this study and others mentioned above confirm that visceral fat represents a potentially modifiable risk factor for cognitive decline/neurodegeneration, an issue of increasingly importance given the increasing prevalence of cognitive decline globally in the aging population (currently listed as 23.7% of the global geriatric population) and the increasing levels of obesity and diabetes

-- and, i promise to have less complex studies in the near future... it just seemed that these last 2 studies broached really important and clinically relevant findings

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

to get access to the dropbox: https://www.dropbox.com/scl/fo/pcf4vymtpo77f49wd2ljr/ANyJqB40CuBd-DMEr67XDoI?rlkey=5dsalmuhf8ekd0svsgvn93xm1&dl=0

or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them