Odor Identification Fell Over 10 Years Whether Older Adults Trained or Not

A five-year randomized exercise trial in Norwegian adults aged 70 to 77 tracked odor identification for a full decade. Scores fell steadily in every group, and neither moderate continuous training nor high-intensity intervals changed that trajectory.

Odor identification is among the first functions to fade in several age-related brain conditions, which is why a research group in Trondheim spent ten years measuring it inside a supervised exercise trial. Their analysis, published in Frontiers in Physiology on 6 November 2026, reports that five years of structured training did not slow the decline at all.

The work comes from the Generation 100 Brain Study, nested inside the larger Generation 100 trial registered as NCT01666340. The parent study enrolled 1,567 older adults living in Trondheim, Norway, and randomized them on a 2:1:1 basis to a control condition, moderate-intensity continuous training, or high-intensity interval training.

The moderate arm was asked to complete 50 minutes of continuous exercise at roughly 70 percent of peak heart rate, twice a week. The interval arm did four four-minute efforts at 85 to 95 percent of peak heart rate, also twice weekly. Controls were pointed at national physical activity guidelines and left unsupervised.

Smell testing covered two overlapping samples. A longitudinal cohort of 106 participants at baseline shrank to 62 by the ten-year mark. A separate cross-sectional cohort of 388 people was assessed once, at that same ten-year follow-up.

What Ten Years of Smell Testing Actually Showed

The decline arrived early and never reversed. By the first annual visit, odor identification had already fallen significantly against baseline, with a coefficient of -1.71 points and a confidence interval running from -2.29 to -1.13. Later time points showed further significant drops.

What the trial could not find was any separation between arms. The exercise-by-time interaction terms reported in the main table carried p-values between 0.10 and 0.75, a spread that offers no support for a training effect on odor identification.

The single-visit comparison of odor identification at ten years told the same story in a different format. Against controls, the moderate training group sat at a coefficient of -0.01 (95 percent CI -0.67 to 0.65, p = 0.99). The interval group sat at -0.22 (95 percent CI -0.88 to 0.45, p = 0.52). Both intervals straddle zero comfortably.

Why Neither Training Dose Moved the Needle

This was not a short or token intervention. Five years of twice-weekly supervised sessions, in a trial large enough to randomize more than 1,500 people, is about as serious a test of chronic exercise as the literature offers. The interval arm in particular trained at 85 to 95 percent of peak heart rate, the kind of intensity that reliably shifts aerobic capacity.

That makes the null result informative rather than merely disappointing. If a decade of follow-up on a five-year program cannot bend the odor identification curve, the hypothesis that structured training protects odor identification in healthy older adults becomes harder to sustain in its simple form.

How Odor Identification Is Measured, and Why Researchers Watch It

Smell testing in aging research is more structured than it sounds. Participants are presented with a series of standardized scents and asked to name each one from a short list of options, producing a score out of a fixed total. That scored task is what odor identification refers to here, and it is deliberately separate from how strong a smell seems or whether a person can detect it at all.

Researchers watch it because the olfactory system sits anatomically close to brain regions affected early in several age-related conditions, and because the test is cheap, quick and repeatable across a decade of visits. Those practical virtues are exactly what made a ten-year trajectory possible in a trial of this length.

Fitness Tracked Odor Identification Even When Training Did Not

The more interesting signal came from fitness itself rather than group assignment. Peak oxygen uptake measured at the ten-year visit was associated with better odor identification, at 0.06 points per unit (95 percent CI 0.01 to 0.11, p = 0.02).

Baseline fitness mattered too. Participants whose starting peak oxygen uptake sat above the national reference value scored higher at ten years, with a coefficient of 0.61 (95 percent CI 0.06 to 1.16, p = 0.04). In the longitudinal model, the fitness-by-time term at one year reached 0.07 (95 percent CI 0.01 to 0.14, p = 0.03), though other time points did not.

The authors are careful about what this pattern means. They state that the findings do not directly support the cardiorespiratory fitness hypothesis as a mechanism linking exercise to brain health, while still supporting a direct association between peak oxygen uptake and odor identification. Fitness and training are not interchangeable, a distinction that also surfaced in work on how differently older adults respond to the same prescribed program.

The Hippocampus Enters the Picture

Brain imaging added a third strand. In the cross-sectional cohort at ten years, hippocampal volume was positively associated with odor identification, at roughly 0.6 points of smell score per millilitre of volume (95 percent CI 0.20 to 1.01).

That association fits a well-established anatomy. Olfactory pathways run close to medial temporal structures, which is part of why smell testing has drawn interest as an early marker. It does not establish direction, however: a single imaging timepoint cannot say whether volume preserved smell or whether both reflect a shared underlying process. Related questions appear in research on physical activity and memory in older adults.

Why an Active Control Group Complicates a Null Result

One feature of the design deserves emphasis. The control arm was not sedentary. Participants were directed to national physical activity guidelines, and the authors note that their exercise durations ended up similar to the intervention groups.

That narrows the contrast the trial was able to test. A null finding between three groups that all exercised is a weaker statement than a null finding between training and genuine inactivity. The question of how much activity is enough recurs across the literature, including reviews of leisure-time exercise and dementia risk.

What Older Readers Can Reasonably Take From This

The finding is narrow, and it is worth stating plainly what it does not say. It does not say exercise fails to benefit the aging brain. It says that in this trial, in this sample, five years of supervised training did not change how odor identification scores fell.

Nor does it say that fitness is irrelevant. The peak oxygen uptake associations point the other way, and they held at both baseline and ten years. What the data cannot establish is causation, because the randomized comparison and the fitness association point in different directions.

General, non-individualized reading of the evidence:

• Structured aerobic training remains well supported for cardiovascular and functional outcomes, which this trial did not test and does not contradict.
• A declining sense of smell in later life is common and is not by itself a diagnosis; persistent or sudden changes are a conversation for a clinician, not a training plan.
• Higher measured fitness tracked better odor identification here, which is a reason to care about aerobic capacity rather than a promise that training will protect smell.

Readers following the broader fitness and longevity thread may also want the context in cardiorespiratory fitness and health outcomes across three cohorts.