Rising energetic cost of walking predicts cognitive impairment

Abstract

BACKGROUND

The energetic cost of walking increases with age and is linked to physical function impairment, but its relation to cognitive impairment is unknown.

METHODS

A total of 687 initially cognitively normal older adults (mean age 74.0 ± 7.2 years, 52% women) underwent repeated walking energy expenditure assessments (V̇O₂) and adjudicated cognitive diagnoses over 7.6 ± 3.8 years. We examined (1) trajectories in the energetic cost of walking prior to any clinical diagnosis of cognitive impairment, comparing adults who later developed cognitive impairment versus those who did not, and (2) the baseline energetic cost and future risk of cognitive impairment using linear mixed‐effects and Cox regression models.

RESULTS

Ninety‐one participants (13%) progressed to cognitive impairment. Progressors exhibited a steeper increase in energetic cost than non‐progressors (B = 0.13; p = 0.003). Higher baseline cost predicted impairment among adults ≥75 years (hazard ratio [HR] = 1.1, 95% confidence interval [CI] = 1.00 to 1.20, p = 0.039), but not those aged 65 to 74 (HR = 0.91, 95% CI = 0.81 to 1.01, p = 0.089).

CONCLUSION

Walking efficiency provides a physiological link between mobility and cognitive health; preserving efficiency may reduce risk of Alzheimer's disease and related dementias.

Highlights

• In initially cognitively normal adults aged 65 years and older, steeper increases in the energetic cost of walking, independent of walking speed, predicted progression to cognitive impairment.

• A single measurement of the energetic cost of walking in older adulthood (≥75 years) was associated with an increased risk of future cognitive impairment.

• The physiological mechanisms underlying mobility may offer novel insights into ADRD risk, beyond those offered by traditional physical function measures.

1. BACKGROUND

Alzheimer's disease (AD) and related dementias (ADRD) affect more than seven million older adults in the United States, ¹ underscoring the pressing need to identify early markers and potential remedial factors. Mobility measures, which reflect an individual's ability to move throughout their environment, have long been recognized as indicators of future morbidity, dementia, and mortality. ² , ³ , ⁴ , ⁵ , ⁶ A hallmark of mobility decline is the age‐related slowing of walking speed. Declines in usual walking speed become noticeable in older adulthood and are accompanied by increased energy costs, as documented by studies of oxygen consumption during normal walking. ⁷ Research has shown that a rising energetic cost of walking precedes mobility decline, ⁸ providing support for the hypothesis that older adults slow their walking speed to achieve an optimal level of energy expenditure. ⁹ , ¹⁰ , ¹¹

While prior research demonstrated that the level of energy required to walk at self‐selected speed provided invaluable insights about physical function, much less is known about whether this physiologic measure provides insights into cognitive function. Recent research from our group has shown that the energetic cost of walking is associated with several neuroimaging markers of age‐ and disease‐related brain pathology in cognitively normal adults, including amyloid beta (Aβ) and brain structure. ¹² , ¹³ , ¹⁴ These studies provide preliminary evidence that the energetic cost of walking tracks with the pathophysiological processes of ADRD during the asymptomatic stage, which may initiate decades before cognitive impairment becomes evident. ¹⁵ , ¹⁶ Notably, in these studies, the energetic cost of walking outperformed usual walking speed, suggesting that the underlying physiological processes associated with mobility decline may be more sensitive indicators of ADRD risk.

RESEARCH IN CONTEXT

1. Systematic review: The authors performed a traditional literature review related to energetic cost of walking, walking speed, and cognitive impairment. There is evidence that walking speed is associated with cognitive impairment. However, whether changes in the energetic cost of walking, which has been shown to precede gait speed decline, is linked to cognitive impairment is unknown.

2. Interpretation: Within an initially cognitively normal cohort of community‐dwelling older adults, an accelerated increase in the energetic cost of walking over time, independent of walking speed, distinguished those who later developed cognitive impairment from those who remained cognitively normal. Furthermore, a single measurement of the energetic cost of walking in older adults was associated with an increased risk of future cognitive impairment. These data suggest the physiological mechanisms underlying mobility offer novel insights into ADRD risk and may serve as a potential target for therapeutic intervention.

3. Future directions: Studies that assess changes in walking energetics, ADRD biomarkers, and cognitive function are needed to better elucidate causality and directionality of the observed associations.

This study expands on this emerging area of research in examining the longitudinal associations between the energetic cost of walking and incident cognitive impairment, including mild cognitive impairment (MCI), the earliest clinical stage of cognitive impairment, ¹⁷ and dementia in community‐dwelling participants of the Baltimore Longitudinal Study of Aging (BLSA). We hypothesized that (1) adults who progress to cognitive impairment will display an accelerated increase in the energetic cost of walking preceding the impairment compared to those who remain cognitively normal and (2) a higher baseline energetic cost of walking is associated with an earlier onset of cognitive impairment.

2. METHODS

2.1. Participants

At enrollment, all BLSA participants are community‐dwelling adults free of major chronic conditions and cognitive and functional impairment. Once enrolled, participants are followed for life and undergo comprehensive health, cognitive, neuroimaging, and functional assessments every 1 to 4 years, depending on age (<60: every 4 years, 60 to 79: every 2 years, ≥80: every year). Trained staff administer all assessments following standardized protocols. Additional study enrollment and design details were previously described. ¹⁸ The sample for the present study consists of initially cognitively normal participants aged ≥65 years free of Parkinson's disease or history of stroke who underwent at least two energetic cost of walking assessments no more than 10 years apart, physical examinations, health history assessments, functional testing, and clinical cognitive testing between 2007 and 2022. The Institutional Review Board of the Intramural Research Program of the National Institutes of Health approved the study protocol. At each examination, participants provided written informed consent.

2.2. Energetic cost of walking

The energetic cost of walking was assessed as the energy expended during an overground customary‐paced walking test. Participants were instructed to walk at their “normal comfortable pace” for 2.5 min in a continuous loop around a 20‐m course laid out in an uncarpeted corridor marked by traffic cones. At the start, participants stood with their feet behind a taped starting line. After a command of “go,” timing was initiated with the first foot‐fall over the line and stopped after 2.5 min of walking. Oxygen consumption (V̇O₂), carbon dioxide production (V̇CO₂), minute ventilation (V̇E), respiratory exchange ratio (RER), and metabolic work rate were obtained during the walking test using a portable indirect calorimeter (Cosmed K4b²/K5, Cosmed) and two‐way non‐rebreathing mask. The Cosmed unit continuously collects and analyzes V̇O₂ and V̇CO₂ using breath‐by‐breath measurement and averages these measures over 30‐s intervals to reduce variability. The Cosmed unit was calibrated using standard procedures prior to each test to ensure accuracy (i.e., volumetric/gas). To calculate average customary‐paced walking V̇O₂ (mL/kg/min), readings from the first 1.5 m of testing were discarded to allow the participant to adjust to the workload. The average V̇O₂ (mL/kg/min) recorded during the final minute was expressed per meter walked (V̇O₂, mL/kg/m) to standardize walking speed and derive a single energetic cost of walking measure.

2.3. Diagnoses of cognitive impairment and dementia

Clinical diagnoses are based on consensus diagnostic procedures implemented for many years at the BLSA, comparable to the National Institute on Aging–Alzheimer's Association criteria. ¹⁷ , ¹⁹ Clinical and selected neuropsychological data from BLSA participants were reviewed at a consensus conference if participants screened positive on the Blessed Information‐Memory‐Concentration Test score (i.e., score ≥4) if their Clinical Dementia Rating (CDR) score was ≥0.5 using subject or informant report or if concerns were raised about their cognitive status. The clinical and cognitive data are first reviewed to determine a syndromic diagnosis (i.e., cognitively normal, MCI, impaired not MCI, or dementia). MCI was determined using the Petersen criteria and diagnosed when (1) cognitive impairment was evident for a single domain or (2) cognitive impairment in multiple domains occurred without significant functional loss in activities of daily living. ²⁰ Both cross‐sectional and longitudinal neuropsychological performance (e.g., Trail Making Test, letter and category fluency), as well as the CDR, are used to determine cognitive impairment. Those judged to be cognitively impaired (e.g., MCI, dementia) are then further classified by presumed etiology (e.g., AD, frontotemporal dementia, dementia with Lewy bodies). More than one etiology can be endorsed (e.g., AD and vascular disease). Additional diagnostic procedures were previously described. ²¹

2.4. Covariate measures

All participants completed a variety of health‐related questionnaires and measurements at each study visit. Variables investigated as potential confounders included age, sex, race, apolipoprotein E epsilon 4 (APOE ε4) carrier status, years of education, body composition, and comorbid conditions. Age, sex, race, and years of education were determined by self‐report. Presence of the APOE ε4 allele was determined through restriction isotyping using standard procedures. ²² Body composition, specifically lean mass (kg) and fat mass (kg), was estimated using a dual‐energy X‐ray absorptiometry scan (Prodigy Scanner, GE). Comorbid conditions were defined as a history of two or more of the following: cardiovascular disease, lung disease, liver disease, kidney disease, peripheral neuropathy, hypertension, diabetes, cancer, and lower extremity arthritis pain.

2.5. Walking Speed

Usual walking speed was measured over a 6‐m course in an uncarpeted corridor. Participants stood with their feet behind a taped starting line and were asked to walk at a “usual comfortable pace.” After a command of “go,” timing was initiated with the first foot‐fall over the starting line and stopped after the first foot‐fall over the finish line. Two timed trials were conducted to derive the usual walking speed in meters/second; the faster of the two trials was used for analyses.

2.6. Statistical Analysis

Baseline and follow‐up participant characteristics were summarized with means (standard deviations) and frequencies (percentages). The association between the energetic cost of walking and the progression to cognitive impairment was examined from two different perspectives. First, a linear mixed‐effects model was employed to examine how the longitudinal trajectories of the energetic cost of walking over time (i.e., progression with age) differed by subjects’ final cognitive status (cognitively normal vs cognitively impaired). For participants who progressed to cognitive impairment, only energetic cost of walking measures collected at earlier research visits prior to the cognitive impairment diagnosis were included in the analyses. This model included interaction terms between final cognitive status and age, adjusting for sex, race, APOE ε4 carrier status, education, body composition, and medical comorbidities (Model 1). The final cognitive status × age (i.e., time) interaction was the term of interest, as it indicated whether longitudinal changes in the energetic cost of walking differed as a function of progressing status (i.e., remaining cognitively normal or developing cognitive impairment).

Second, Cox proportional hazard regression models were employed to examine how baseline energetic cost of walking levels associated with age at the onset of cognitive impairment. To test the hypothesis that the effect of baseline energetic cost of walking differed at an older age, the interaction term between baseline energetic cost of walking and the indicator variable for baseline age (≥75 years) was included in the model. Additionally, this model was adjusted for age, sex, race, APOE ε4 carrier status, education, and body composition (Model 2). The baseline energetic cost of walking × indicator for baseline age 65 to 74.9 versus ≥75 years interaction was the term of interest, as it indicated whether the association between baseline energetic cost of walking and risk of cognitive impairment differed between younger and older participants. Hazard ratios and Kaplan Meier analysis with log‐rank tests were used to examine the association between baseline energetic cost of walking and age of onset of cognitive impairment (Model 2). Schoenfeld residuals were used to evaluate whether the proportional hazards assumption was violated. The competing risk of death was considered by treating the non‐progressor who passed away before becoming cognitively impaired as being censored at the time of the occurrence of death, using the cause‐specific hazard models. ²³

To verify the robustness of the results, we performed a series of sensitivity analyses by (1) controlling for usual walking speed in the primary analysis (Model 1), (2) using multiple imputations to account for missing covariate values (Models 1 and 2), and (3) linear mixed‐effects models with energetic cost of walking as the outcome, number of total visits as independent variable, adjusting for baseline age and other covariates. These were conducted to confirm results were (1) independent of usual walking speed, (2) not influenced by the small percentage (< 8%) of missing covariate data, and (3) not biased by dropouts, and non‐participation or dropout during follow‐up were missing at random (MAR). To improve the interpretability, both the energetic cost of walking (mL/kg/m) and walking speed (m/s) were scaled to 100 units in the regression models. Statistical significance was set at an alpha level of 0.05. All analyses were conducted using Stata (version 18.0, StataCorp LLC, College Station, TX, USA).

3. RESULTS

We examined 687 participants (mean age 74.0 ± 7.2 years, 52.4% women, 29.5% non‐White) with longitudinal energetic cost of walking and clinical cognitive diagnosis data over 7.6 ± 3.8 years. A flowchart of participant selection is provided in Figure S1. Among 687 participants, 91 progressed to cognitive impairment, including dementia (13%). Compared to participants who remained cognitively normal, those who progressed to cognitive impairment were more likely to be older, have lower body fat, a higher energetic cost of walking, and slower walking speed (all p < 0.05, Table 1). Results from the linear fixed‐effects model revealed that participants who progressed to cognitive impairment had a steeper increase in the energetic cost of walking over time than non‐progressors (Model 1 interaction term; B = 0.13; p = 0.003; Table 2, Figure 1). The Cox proportional hazard regression model revealed the association of baseline energetic cost of walking and progression to cognitive impairment differed at older age (Model 2 interaction term: hazard ratio [HR] = 1.2, 95% CI = 1.05 to 1.39, p = 0.008). Baseline energetic cost of walking was not significantly associated with progression to cognitive impairment among participants aged 65 to 74.9 (Model 2; HR = 0.91, to CI = 0.81 to 1.01, p = 0.089); however, among those aged ≥75, a higher energetic cost of walking was significantly associated with increased risk of progressing to cognitive impairment (Model 2; HR = 1.1, to CI = 1.00 to 1.20, p = 0.039, Table 3; log‐rank p < 0.001, Figure 2). Visual inspection of the energetic slopes in Figure 1 revealed a clear divergence between individuals who remained cognitively intact and those who became impaired beginning around age 75, potentially explaining the observed associations at this time point. In sensitivity analyses, results remained consistent after accounting for walking speed, missing values in covariates, and attrition bias (Tables S1–S4).

TABLE 1.

Baselines characteristics of Baltimore Longitudinal Study of Aging study participants.

N (%) or mean (SD)	All N = 687	Non‐progressor N = 596	Progressor N = 91	P value
Age (years)	74.00 (7.19)	73.41 (7.02)	77.89 (7.10)	<0.001
Baseline age category				<0.001
65 to 74.9	254 (37.0 %)	239 (40.1 %)	15 (16.5 %)
≥75	433 (63.0 %)	357 (59.9 %)	76 (83.5 %)
Age at last visit or progressed visit	80.88 (7.05)	80.43 (7.03)	83.83 (6.46)	<0.001
Number of visits, median (range)	4 (2 to 13)	4 (2 to 13)	3 (2 to 10)	0.33
Non‐White	203 (29.5 %)	172 (28.9 %)	31 (34.1 %)	0.28
Women	360 (52.4 %)	318 (53.4 %)	42 (46.2 %)	0.20
APOE ε4 carriers	171 (25.4 %)	145 (24.9 %)	26 (29.2 %)	0.38
Education (years)	17.83 (2.71)	17.76 (2.62)	18.34 (3.20)	0.06
Fat mass (kg)	26.62 (9.69)	26.93 (9.69)	24.67 (9.57)	0.04
Lean mass (kg)	46.31 (9.47)	46.32 (9.65)	46.22 (8.35)	0.92
Height (cm)	166.75 (8.49)	166.90 (8.39)	164.25 (10.05)	0.27
Comorbid conditions (≥2)	413 (60.6 %)	353 (59.7 %)	60 (65.9 %)	0.26
Energetic cost of walking (mL/kg/m)	16.64 (3.24)	16.53 (3.18)	17.33 (3.57)	0.03
Usual walking speed (m/s)	1.15 (0.21)	1.17 (0.20)	1.06 (0.23)	<0.001

Note: Values indicate mean and standard deviation unless indicated otherwise.

Abbreviation: APOE ε4, ε4 allele of the apolipoprotein E gene.

TABLE 2.

Results from linear mixed‐effects model to examine how longitudinal trajectories of energetic cost of walking over time differ by subjects’ final cognitive status prior to cognitive impairment diagnosis.

Predictors	Coefficient	95% CI	P value
Progressors versus non‐progressors	−0.75	−1.896, 0.46	0.224
Age
Age: Non‐progressors	0.04	0.013, 0.08	0.005
*  Progressors versus non‐progressors	0.13	0.04, 0.22	0.003
Woman versus Man	0.80	0.04, 1.55	0.038
Race
Black versus White	−0.99	−1.45, −0.53	<0.001
Other versus White	0.01	−0.80, 0.78	0.981
APOE ε4	−0.25	−0.69, 0.18	0.256
Education	−0.07	−0.14, 0.004	0.062
Lean mass	0.04	−0.001, 0.08	0.046
Fat mass	−0.04	−0.06, −0.02	< 0.001
Comorbid conditions	0.15	−0.17, 0.48	0.357

Abbreviations: APOE ε4, ε4 allele of the apolipoprotein E gene; CI, confidence interval.

^*Interaction term of interest = Final cognitive status × Age (i.e., time). Non‐progressors = remained cognitively normal; progressors = progressed to cognitive impairment.

FIGURE 1.

Energetic cost of walking trajectories among older adults who progressed to cognitive impairment versus those who did not.

TABLE 3.

Results from Cox regression model to examine how baseline energetic cost of walking levels associate with age at onset of cognitive impairment.

Predictors	HR	95% CI	P value
Baseline age	1.13	1.09, 1.18	<0.001
Baseline energetic cost of walking
Baseline energetic cost of walking: <75 years	0.91	0.81, 1.01	0.089
*  <75 years versus ≥75 years	1.21	1.05, 1.39	0.008
Women versus men	1.53	0.66, 3.52	0.319
APOE ε4 carrier versus non‐carrier	1.45	0.89, 2.36	0.137
Education	1.02	0.93, 1.11	0.658
White versus non‐White	0.67	0.42, 1.09	0.105
Lean mass	0.99	0.95, 1.04	0.687

Abbreviations: HR, hazard ratio; CI, confidence interval.

^*Interaction term of interest = Baseline energetic cost of walking × Baseline indicator for age ≥75 years.

FIGURE 2.

Kaplan Meier analysis with log‐rank tests investigated baseline energetic cost of walking and progression to cognitive impairment.

4. DISCUSSION

In initially cognitively normal men and women aged 65 years and older, an accelerated increase in the energetic cost of walking distinguished those who later developed cognitive impairment from those who remained cognitively normal. Notably, these findings were independent of usual walking speed. Furthermore, we found that a single measurement of the energetic cost of walking in older adulthood (≥75 years) is associated with an increased risk of future cognitive impairment. These data add to a growing body of evidence indicating that the physiological mechanisms underlying mobility may offer novel insights into ADRD risk, beyond those offered by traditional physical function measures.

Assessing mobility in older adulthood holds clear value. Usual walking speed has historically served as a marker of physical function due to its ease of measurement and prognostic value across a range of health outcomes. ² , ³ , ⁴ , ⁵ , ⁶ Research has shown that walking speed tends to slow in later adulthood and may coincide with declines in cognitive function. ⁴ Although walking speed has been the focus of considerable research, far less is known about the relationship of walking‐related energetic efficiency to higher‐order cognitive function and capacities. Previous research from our group found the energetic cost of walking to increase with age and be an important indicator of future physical function health ⁸ ; in that study, adults who displayed an accelerated increase in the energetic cost of walking from midlife to late life were more likely to develop slow walking speed in older adulthood. ⁸ Consistent with this line of research, our prior studies also showed that the energetic cost of walking could provide meaningful insights into cognitive and brain health, beyond the traditional physical function tests. ¹² , ¹³ , ¹⁴

Specifically, we found that a higher energetic cost of walking was cross‐sectionally associated with lower gray and white matter volumes in frontal, parietal, and temporal lobes, as well as smaller hippocampal and larger ventricular volumes. ¹³ Longitudinal analyses revealed that higher baseline levels of the energetic cost of walking predicted accelerated hippocampal atrophy and ventricular enlargement over time and that participants with the highest energy cost (∼0.25 mL/kg/m) displayed approximately twice the annual rate of volume change in these regions compared to more efficient walkers (∼0.10 mL/kg/m). ¹³ Notably, hippocampal and ventricular volume trajectories have been shown to differentiate cognitively normal individuals from those with ADRD ²⁴ and are strongly linked to disease severity. ²⁵ Building on these findings, we examined Aβ deposition, a hallmark of AD pathology, in a subset of cognitively normal BLSA participants who underwent positron emission tomography imaging and observed that each 0.01 mL/kg/m greater energy cost was associated with 18% higher odds of amyloid positivity. ¹² Collectively, these studies suggest that the energy needed for walking is linked to age‐ and disease‐related processes and may offer novel prognostic value for identifying individuals at elevated risk for ADRD.

Walking, among the most common forms of physical activity in older adulthood, has long been recognized as a potential modifier of cognitive decline and ADRD risk. Early studies showed that adults who self‐reported higher levels of walking tended to experience slower rates of age‐related cognitive decline. ²⁶ , ²⁷ These findings are supported by recent research using objectively assessed walking using accelerometers, which further demonstrate that greater daily walking is associated with a reduced risk of developing ADRD. ²⁸ , ²⁹ However, very little is known about how the level of energy required for walking relates to cognitive health. In this study, we found that a rising energetic cost of walking throughout older adulthood was associated with a higher likelihood of being diagnosed with cognitive impairment over an average 8‐year follow‐up period. In the linear mixed‐effects model (Model 1), we observed a distinctly higher energetic cost of walking trajectories among participants who progressed to cognitive impairment versus those who did not. Interestingly, these trajectories appeared to diverge around age 75, suggesting a potential inflection point at which the energetic cost of walking became higher among individuals who later developed cognitive impairment compared to those who did not (Figure 1). These results were further supported by the Cox regression model (Model 2), which revealed a significant interaction between baseline energetic cost of walking × age group (<75 years vs ≥75 years), demonstrating that higher baseline energetic cost of walking in older adulthood was associated with an increased risk of progressing to cognitive impairment (Figure 2). Notably, those who progressed to cognitive impairment still had, on average, usual walking speeds in the normal range (e.g., >1.0 m/s). These findings expand upon and complement our previous research findings that adults with a higher energetic cost of walking experienced greater rates of neurodegeneration and Aβ deposition and demonstrate that the energetic cost of walking is also associated with a clinical cognitive diagnosis.

Although the energy costs of human mobility have been the subject of scientific investigation for over a century, ³⁰ to date research in this area has mainly focused on aerobic fitness, which reflects an individual's capacity for exercise (i.e., V̇O₂max). Similar to the energetic cost of walking, aerobic fitness worsens with age, and low fitness levels are predictive of future mobility limitations. ³¹ , ³² , ³³ Further, when low fitness combines with a high energetic cost of walking, there may be only a negligible difference between peak capacity and the energy needed for mobility. ³⁴ Work from our group and others has identified fitness as a physiological attribute that may delay onset of the pathophysiological and clinical features of ADRD, ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ but relatively little research has focused on the links between the energy needed for mobility and brain health. The current findings contribute to the growing body of research on the energetic costs of mobility by showing that the energy required for customary‐paced walking may serve as a novel energetic marker of ADRD risk that is safer to measure than V̇O₂max (e.g., risk of arrythmia). These data further underscore the potential importance of preserving walking efficiency as a strategy to mitigate physical and ADRD risk and suggest the viability of clinical measurement of the energetic cost of walking as a screening tool for older adults.

Several pathways may explain the association between the energetic cost of walking and risk of cognitive impairment. One hypothesis is that maintaining a physically active lifestyle during midlife leads to more efficient walking in later life, with the energetic cost of walking serving as a marker of long‐term physical activity. A potential pathway through which this may manifest is skeletal muscle mitochondrial health. Higher levels of physical activity are associated with enhanced skeletal muscle mitochondrial function, which in turn is linked to more favorable energetic profiles in older adults. ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ Accordingly, a lower energetic cost of walking in later life may reflect the cumulative effects of sustained physical activity across the lifespan, in part through maintained skeletal muscle health. Indeed, higher levels of physical activity are associated with numerous indices of brain health in midlife, prior to cognitive impairment. ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ Taken together, these observations suggest that walking efficiency may serve as a valuable indicator of both lifelong physical activity and its downstream benefits for brain health and resilience to cognitive impairment. Alternatively, a rising energetic cost of walking may be a consequence of age‐ and disease‐related changes. It is plausible that changes in the central nervous system, such as the accumulation of white matter hyperintensities, Aβ plaques, and neurodegeneration, lead to increased energetic costs for mobility through shared neurological pathways. The aggregation of these pathologies has been shown to cause myelin and axonal loss, ⁴⁹ , ⁵⁰  which may disrupt corticospinal pathways critical for neuromuscular signaling and motor unit recruitment. These disruptions in the central nervous system may lead to inefficient motor programming and/or subtle motor coordination dysfunction that results in higher energy costs for mobility.

Future studies that assess changes in walking energetics, ADRD biomarkers, and cognitive function are needed to better elucidate causality and directionality of the observed associations. Additional research into the biological mechanisms connecting walking energetics with age‐ and disease‐related neuropathology is also warranted. The BLSA is a well‐characterized sample of highly educated healthy middle to older aged adults, making it uncertain whether our results generalize to more socioeconomically diverse populations. These limitations should be considered in the context of the study's strengths, which include the large sample size, clinically adjudicated cognitive diagnoses, and objectively measured customary‐paced walking energetics throughout mid to late life.

In summary, this study demonstrates that among initially cognitively normal, middle‐ to older‐aged community‐dwelling adults, divergent walking energetic trajectories were observed between those who did and did not develop cognitive impairment, with higher energetic costs of walking associated with greater risk. Notably, these findings persisted after accounting for walking speed. This work adds to a growing body of evidence suggesting that the physiological mechanisms underlying mobility may offer novel insights into ADRD risk, beyond those offered by traditional physical function measures. Preserving walking efficiency could be a key strategy for reducing the risk of ADRD, with the energetic cost of walking serving as a novel target for future interventions.

FUNDING SOURCES

Data used in the analyses were obtained from the Baltimore Longitudinal Study of Aging, a study performed by the National Institute on Aging Intramural Research Program (blsa.nih.gov). Ryan J. Dougherty was supported by grant K01AG080122.

CONFLICT OF INTEREST STATEMENT

All authors report no conflicts of interest. Author disclosures are available in the supporting information.

CONSENT STATEMENT

All participants provided written informed consent at each study visit.

Supporting information

Supporting Information

Supporting Information

Supporting Information