Lifestyle,sleep quality,and cognitive function in resistant hypertension: One-year follow-up from the TRIUMPH trial

Abstract

Background

Treatment resistant hypertension (TRH) is associated with increased risk of cognitive decline, which may be reduced by healthy lifestyle changes.

Objective

To examine the effects of a comprehensive, rehabilitation-based lifestyle hprogram on cognitive function during a one-year follow-up of participants from the TRIUMPH clinical trial.

Methods

Among the 140 TRIUMPH participants originally randomized, 91 (65%) were available for one-year assessments prior to the COVID-19 lockdown. Participants were originally randomized to a Cardiac rehabilitation-based LIFEstyle program (C-LIFE) or to a Standardized Education and Physician Advice (SEPA) condition for 4-months. During their one-year follow-up, participants underwent assessments of sleep quality, body mass index, actigraphy-assessed physical activity levels, and cerebrovascular reactivity using functional near infrared spectroscopy (fNIRS). Cognitive function was assessed using a 45-min test battery incorporating tests of Executive Function/Learning, Memory, and Processing Speed. Regression-based models incorporating reliable change indices were used to assess cognitive change.

Results

Participants included 91 individuals (mean age = 63.6 [SD = 8.6]), evenly distributed in biological sex and race/ethnicity, and tended to be college-educated. The C-LIFE group had more preserved cognitive functioning compared to SEPA (C-LIFE: z = −0.26 [−0.40, −0.12] versus SEPA: −0.60 [−0.81, −0.39]; d = 0.44, p = 0.008), with reduced PSQI sleep symptoms associating with more preserved cognitive function (B = −0.18, p = 0.050 per 3-points). Treatment did not improve fNIRS markers, although changes in weight and physical activity associated with fNIRS outcomes.

Conclusions

Lifestyle modification may help preserve cognitive functioning among individuals with resistant hypertension.

Keywords

Alzheimer's disease blood pressure cognitive function executive function hypertension lifestyle memory sleep quality

Introduction

Hypertension is common and increasingly prevalent among United States adults.^1–3 Recent data suggest that as many as 75% of older adults are treated with an antihypertensive medication and this is projected to increase in parallel with rising obesity levels.³ Among hypertensive adults, those with resistant hypertension, characterized by taking at least three antihypertensives without an optimal blood pressure reduction, have been suggested to have the highest risk of stroke, cognitive decline, and Alzheimer's disease and related dementias (ADRD).⁴ Strategies to improve both blood pressure and cognition among individuals with treatment-resistant hypertension (TRH) are therefore critical to mitigate risk of ADRD in TRH.

Over the past decade, a growing number of clinical trials have demonstrated that reductions in blood pressure associates with decreased risk of cognitive impairment and ADRD.^1,2,5 Targeted blood pressure reductions using pharmacotherapy have been shown to reduce the risk of mild cognitive impairment (MCI)⁶ and a large body of evidence from multi-component randomized trials suggests that adoption of healthier lifestyle habits helps preserve cognition in adults at risk for ADRD.^7,8 Lifestyle modification may help lower ADRD risk, in part, through reducing hypertension and improving vascular function, tau hyperphosphorylation, and beta amyloid accumulation,⁹ even among individuals without measurable cognitive decline.¹⁰ The durability of these effects, however, remains in question, as many individuals are unable to sustain healthy lifestyle habits after the structure and support from the intervention team ends.^11,12 Follow-up analyses from randomized trials are therefore needed to better understand the longer-term effects of lifestyle modification on cognitive function. In addition, understanding how different aspects of lifestyle change differentially influence sustained cognitive function could have important implications for prioritizing treatment targets, such as choosing to target physical activity instead of weight loss. Finally, studies incorporating additional, subclinical markers of cerebral functions are needed to enhance the understanding of inconsistent cognitive changes, as measures of cerebral efficiency may provide insight into subtle brain changes that may not necessarily be reflected on clinical neuropsychological measures.^13,14

The specific lifestyle modification components that correlate with improved cognitive outcomes remain to be adequately characterized from clinical trial follow-up assessments. Although improvements in aerobic fitness,¹⁵ weight loss,^11,16–18 and reduced blood pressure¹⁹ have all been suggested as key correlates, the relationships between changes in these factors and cognitive outcomes has been inconsistent.²⁰ For example, although most aerobic trial data suggest improvements in cognitive outcomes,⁸ most data attempting to link changes in fitness to cognitive improvements have been equivocal.²⁰ In contrast, intentional weight loss appears to associate with improved cognition,¹⁶ regardless of whether it was achieved using physical activity, weight loss surgery, or through intensive caloric restriction.¹⁸ Recent data have extended the examination biobehavioral mechanisms to include improvements in sleep quality, which commonly parallel improvements in physical activity, diet, and weight loss.^21–23 Poor sleep quality is common among individuals with hypertension and robustly associated with worse cognitive functioning, even among individuals who do not have obstructive sleep apnea.²⁴ In order to examine the longer-term benefit of lifestyle modification on cognitive functioning and associated biobehavioral correlates, we examined longitudinal changes in cognition and associations with changes in physical activity, dietary quality, weight maintenance, sleep quality, and cerebrovascular functioning during a one-year follow-up from the TRIUMPH trial. The present analyses focus on cognitive and cerebrovascular outcomes, whereas comprehensive findings from the overall lifestyle intervention will be reported separately (Hinderliter et al., unpublished results). For cognitive and cerebrovascular outcomes, we hypothesized that the C-LIFE treatment group would show better cognitive function and greater cerebrovascular reactivity compared to SEPA at one-year and that these effects would be associated with higher levels of physical activity and DASH diet scores, better weigh maintenance, and better sleep quality.

Methods

Sample

The protocol was approved by the Duke University School of Medicine institutional review board following the Helsinki protocol and written informed consent was obtained from all participants. Adults with overweight / obesity and resistant hypertension were eligible to participate.^4,25 Treatment resistant hypertension (TRH) was defined as inadequate blood pressure response despite being optimally treated with at least three antihypertensives (including a diuretic). Specifically, as we previously reported,^4,25,26 TRH was defined as treatment for at least 6 weeks with 3 or more antihypertensive medications of different classes, including a diuretic, with clinic SBP ≥130 mm Hg or DBP ≥80 mm Hg, or the need for 4 or more drugs to achieve SBP ≤130 mm Hg and DBP ≤80 mm Hg, with SBP ≥120 mm Hg, were eligible. Additional inclusion criteria included body mass index ≥25 kg/m², lack of regular moderate or vigorous physical activity, and age 35 to 80 years. Exclusion criteria included known secondary hypertension, estimated glomerular filtration rate <40 mL/min/1.73 m2, moderate-severe ischemic heart disease, diabetes requiring insulin, and major psychiatric disorder or substance dependence, including alcohol consumption >14 drinks/week. Participants were enrolled beginning in June 2015 and ending in October 2019. Of note, because data for the present study were obtained from two separate grants with different hypotheses, the present analyses will focus on cognitive and cerebrovascular outcomes (HL130237), whereas complete data on blood pressure, physical activity, nutrition, and weight outcomes (HL122836) will be reported separately (Hinderliter et al., unpublished results).

Randomization

Participants were randomized 2:1 to a 4-month, cardiac rehabilitation-based rehabilitation program (C-LIFE) or to a standardized education and physician advice condition (SEPA). This randomization scheme was selected to leverage treatment heterogeneity in the C-LIFE condition, providing a more powerful test of efficacy.⁴ Assessments of cognitive function, vascular and cerebrovascular functioning, physical activity, blood pressure, and sleep quality were collected at baseline, 4 months, and one-year after randomization.

Follow-up assessments

Following the 4-month trial, we assessed lifestyle habits, blood pressure, and cognitive function during a one-year naturalistic follow-up. Following their 4-month assessment, participants received no further intervention but were periodically contacted to maintain study contact (Figure 1). During the active intervention period, four individuals (all in SEPA) dropped out, possibly due to their dissatisfaction with treatment group assignment. The remainder of individuals with missing data were impacted by the initial COVID-19 lockdown and associated contact precautions. Individuals lost to follow-up had higher BMI and worse insulin sensitivity levels (as indicated by higher HOMA-IR [Homeostatic Model Assessment for Insulin Resistance] levels), as well as lower levels of intima medial thickness and kidney disease, but otherwise did not differ from completers. Specifically, participants lost to follow-up had higher BMI levels (37.3 [SD = 5.7] versus 35.3 [SD = 5.6]), higher HOMA-IR levels (5.5 [SD = 4.0] versus 4.4 [SD = 2.2]), lower IMT levels (0.73 [SD = 0.14] versus 0.78 [SD = 0.16]), and lower rates of kidney disease (14% versus 30%) compared to participants who were available for follow-up assessments.

Figure 1.

Consort chart showing the flow of participants through screening, randomization, post-treatment assessments and one-year follow-up. Participants were randomized to either a Cardiac rehabilitation-based LIFEstyle program (C-LIFE; n = 90) or to a Standardized Education and Physician Advice (SEPA; n = 50) for four months.

Cognitive performance

We adopted a battery similar to that proposed by the Neuropsychological Working Group for vascular cognitive disorders.²⁷ As detailed below, we used reliable change indices to quantify cognitive changes using regression-based techniques, the sources of which are cited for individual subtests. Assessments included tests of Executive Function/Learning (Trail Making Test Part B [lower scores = better performance],²⁸ Stroop Color-Word Section,²⁹ Animal Naming,²⁸ COWA,³⁰ CVLT-II Discrimination Index,³¹ BVMT-R Learning),³⁰ Processing Speed (Trail Making Test Part A [lower scores = better performance],²⁸ Stroop Word Section, Stroop Color Section, Digit Symbol Substitution Test, Ruff 2&7 Test),³² and Memory (CVLT-II Total Learning [sum of trials 1–5], CVLT-II Immediate Recall,³¹ and CVLT Delayed Recall).^23,26,33 On the CVLT-II, we incorporated several indices that measure efficiency of encoding information, including semantic clustering and learning slope efficiency. Semantic clustering represents the degree to which participants recognized and grouped words into categories to enhance their ability to learn and recall words on the CVLT-II.³⁴ Total learning slope efficiency represents how quickly and efficiently individuals were able to learn CVLT-II word items, with higher score suggesting more effective learning over trials.³⁵

As previously reported, cognitive testing followed a fixed order of administration and utilized alternative forms for all tests where available (e.g., CVLT-II, BVMT-R, COWA, Trail Making Test, Digit Symbol Substitution Test), utilizing an alternating schedule of administration between assessment time periods (e.g., A-B-A and B-A-B). Assessments were performed by a data technician who was not involved in the participants’ treatment regimen and participants were asked not to disclose their treatment assignment during assessments.

Cerebrovascular functioning

We utilized functional near infrared spectroscopy (fNIRS) to non-invasively cerebrovascular reserve and vasomotor function in parallel with cognitive function measures. When assessed by fNIRS, the increase in cerebral blood flow in the surface layers of the cortex during localized neural activity is seen as an increase in the total concentration of hemoglobin (total-Hb) and comparative decrease in deoxy-Hb, with both variables corresponding strongly with the functional MRI BOLD signal.³⁶

Cerebrovascular reserve (CVR)

CVR was measured as the change in tissue oxygen saturation from baseline during a standardized breath holding index (BHI).^37,38 The BHI has been used extensively as a marker of CVR, has been shown to decrease with age,³⁹ is impaired among individuals with cerebrovascular disease^40,41 and cognitive impairment,⁴² and is predictive of subsequent stroke.⁴³ Consistent with standard methodology, changes in TOS during BH was measured using two, near-infrared optodes of the cerebral oximetry, each consisting of two light-source fibers and one light-collecting fiber. Sensors was placed in two symmetrical points on the forehead in symmetric areas of the frontal lobe of the two hemispheres (2 cm beside the midline and about 3 cm above the supraorbital ridge) and fixed with a headband. CVR was quantified as changes in TOS from baseline to BHI, divided by baseline TOS and total BHI time (BHI = [TOSBH − TOSBAS] × 100/TOSBAS/DBH).⁴⁴ Voluntary breath holding was used to induce hypercapnia, with a duration ranging from 20 to 30 s. Assessments lasting less than 20 s were repeated.

Cerebral hemodynamic changes

We also assessed changes in oxygenated hemoglobin (HbO₂) while completing cognitive challenges for selected subset of three cognitive tasks amenable to efficient assessment in an ambulatory setting. We utilized cognitive assessments known to elicit cortical activation and that have been shown to have prognostic benefit in older adults at risk for ADRD.⁴⁵ The cognitive challenge tasks therefore included semantic fluency (articles of clothing, items in a grocery store, and fruits), mental arithmetic (subtraction using serial sevens with varying start numbers; max score = 12), and phonemic fluency (letters B, F, and W; selected in order not to conflict with the COWA administration in our main test battery). As recently reviewed,⁴⁶ contemporary analyses of fNIRS data during cognitive assessments has primary utilized measures of changes in HBO₂ while performing cognitive assessments, which is thought to reflect hemodynamic efficiency.⁴⁷ Cognitive tasks were administered in a fixed order (1: semantic fluency, 2: mental arithmetic, 3: phonemic fluency), counterbalanced across participants in order to minimize practice effects. Each task was conducted for one-minute, with changes in oxygenated hemoglobin (maximum – minimum HbO₂ levels during one-minute assessment) used as the hemodynamic measure of interest.

Physical activity levels

Physical activity levels were assessed using 24-h actigraphy with an accelerometer (ActiGraph GT9X Link Monitor, Pensacola, FL). Wrist-worn actigraphy monitors were worn for seven consecutive days both before and after the intervention in order to assess measures of sleep quality and markers of weekly physical activity. We focused on total step count as the metric of interest, which was found to have been improved in the C-LIFE group relative to SEPA controls, as previously published.

Sleep quality

Sleep quality was assessed using Pittsburgh Sleep Quality Inventory (PSQI). The PSQI provides a subjective measure of sleep quality and is composed of nineteen individual items that generate seven “component” scores (i.e., subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction).^23,48

Blood pressure

Both clinic and ambulatory blood pressures (BPs) were obtained. Clinic-measured BP was determined according to JNC-7 guidelines. Ambulatory blood pressure monitoring (ABPM) was also assessed during a typical day, in which participants wore an Oscar 2 (Suntech Medical Inc, Raleigh, North Carolina) ABPM.⁴⁹ The Oscar 2 was programmed to record BP measurements 3 times per hour during waking hours and 2 times per hour during sleep, with included assessments requiring at least 70% of planned readings.⁵⁰ Total number of blood pressure medications and BP medication burden were also obtained using total count and mean Daily Defined Dose (DDD) levels.⁵¹ The DDD was developed by the World Health Organization to systematize the quantification of medication dosing across potentially heterogeneous treatment regimens.

Nutritional and weight assessment

Dietary and nutritional content was obtained both using a retrospective food frequency questionnaire (FFQ) requiring participants to recall typical consumption during a 4-week period⁵² and a 2-day (weekend/weekday) food diary.⁵³ The FFQ was analyzed by NutritionQuest (Berkeley, CA), and the diary data were analyzed using the Automated Self-Administered 24-h (ASA 24^®) Dietary Assessment Tool (https://epi.grants.cancer.gov/asa24/).

Data analysis

Analyses were conducted in SAS 9.4 (Cary, NC) and R 4.5.0 (https://cran.r-project.org/). In order to provide a parsimonious analytical approach across cognitive outcomes, we elected to use a global score approach, which combined reliable change indices for all subtests into a unified z-score outcome. This approach parallels the “gatekeeper” methodology we utilized in the primary trial analyses for cognitive and biomarker outcomes,^25,26 and is an approach we have used extensively in other clinical trials.^18,^54–56 Our analyses followed an intention-to-treat approach, including all 140 participants regardless of their follow-up status. Following our primary analysis, we conducted explanatory, secondary analyses of individual changes in each subtest with accompanying effect sizes to provide a characterization of where group differences were arising, if observed.^57–60 This type of explanatory analytical step has been advocated as appropriate within the “gatekeeper” framework to account for multiplicity in clinical trials settings because of its control for type-I error. However, this approach also has the limitation that small-to-moderate patterns of change, if consistently observed, will yield a significant composite outcome despite relatively small changes on individual variables when examined in isolation.

To examine longitudinal associations between putative mechanistic markers and cognitive outcomes, we conducted a parallel regression in which changes in each physical activity, weight, DASH diet scores, and sleep quality were modeled together. Our primary analyses were therefore conducted using PROC REG and PROC GLM. Within our modeling approach, the one-year global RCI z-score served as the outcome variable with age, education (years), sex, baseline creatinine, baseline stroke risk (using the Framingham Stroke Risk Profile), and baseline cognitive function (mean rank across all cognitive subtests) as adjustment variables, with group assignment as the variable of interest. Similarly, to examine changes in task-based hemodynamic efficiency, we created a mean-rank, global score across all three cognitive challenge tasks, residualized on cognitive performance levels.

Changes in physical activity, weight, DASH diet quality, and sleep quality were conducted using a parallel approach, with baseline levels of the respective outcome as an additional control variable. Missing data was accounted for using multiple imputation (PROC MI within SAS 9.4), in which 200 imputations were utilized. Model diagnostics were assessed through visual inspection of residuals in comparison with observed values, as well as assessing imputations for potential influential outliers. Within our imputation models, we incorporated 1) all variables that were subsequently included in our full models (e.g., cognitive functioning, age, education, sex, creatinine, Framingham stroke risk profile, and treatment group), 2) variables predictive of missingness within our dataset (e.g., intima medial thickness, body mass index, additional markers of kidney function [e.g., eGFR], HOMA-IR), and 3) variables assessing putative mechanistic factors, such as sleep quality and blood pressure.

Follow-up, explanatory analyses of individual cognitive changes were conducted using repeated measures, linear mixed effect models through PROC MIXED. Within these models, cognitive function for individual subtests at 4-months and one-year served as the repeated measures outcome, controlling for the same adjustment variables as in our primary analyses above. Our a priori hypotheses were that C-LIFE would improve cognitive functioning across domains in comparison with controls at one-year. We further hypothesized that higher levels of physical activity, DASH diet scores, ambulatory systolic blood pressure, weight loss, and sleep quality would associate with better cognitive and cerebrovascular outcomes. Consistent with contemporary modeling approaches, we examined associations between putative mechanistic markers and outcomes regardless of whether a main effect was observed.^61,62 Model assumptions regarding independence, additivity, and linearity of residuals were evaluated and found to be acceptable prior to analysis.

Results

Background and clinical characteristics from baseline assessments are presented in Table 1. From the original 140 participants, 91 were available for inclusion during the one-year follow-up (Figure 1). Analyses of one-year lifestyle habits will be presented in detail elsewhere (Hinderliter et al., unpublished results) and are therefore described only briefly here. Results indicated that the C-LIFE group exhibited lower ambulatory SBP (133 mm Hg [129, 137] versus 140 mm Hg [135, 144], p = 0.044) and ambulatory DBP (71 mm Hg [69, 73] versus 74 mm Hg [72, 77], p = 0.046) compared with SEPA). In terms of other lifestyle habits, we found no group differences in dietary quality (DASH diet scores: C-LIFE = 3.8 [3.5, 4.0] versus SEPA = 3.5 [3.0, 3.9]. p = 0.171), objectively assessed physical activity (Daily Actigraphy Steps: C-LIFE = 9844 [9360, 1–320] versus SEPA = 9667 [8949, 10385], p = 0.653), or weight (C-LIFE: 219.6 [217, 222] lbs versus SEPA: 220.7 [217, 225] lbs, p = 0.282). In exploratory analyses of sleep quality changes on the PSQI, we found that the C-LIFE group continued to demonstrate improved sleep quality compared with SEPA (6.7 [6.2, 7.3] versus 7.7 [6.9, 8.6], d = 0.45, p = 0.049) (Table 2), corresponding to % changes from baseline of +5.6 [−3.7, 14.9] for SEPA versus −14.1 [−27.9, −0.4] in C-LIFE) (Supplemental Figure 1).

Table 1.

Background and demographic characteristics of the sample.

Variable	C-LIFE(n = 90)	SEPA(n = 50)	Full Cohort(n = 140)
Age, Years	62.2 (8.7)	63.2 (9.3)	62.5 (8.9)
Biological Sex, Female, n (%)	43 (48%)	24 (48%)	67 (48%)
Race / Ethnicity, n (%)
African-American	47 (52%)	35 (70%)	82 (59%)
White	39 (43%)	11 (22%)	50 (36%)
Asian	1 (1%)	1 (2%)	2 (1%)
Hispanic	1 (1%)	1 (2%)	2 (1%)
Other	2 (2%)	0 (0%)	2 (1%)
Education, Years	16.1 (2.3)	15.7 (2.4)	15.9 (2.4)
Clinic SBP (mm Hg)	139 (10)	140 (10)	139 (10)
Clinic DBP (mm Hg)	79 (9)	80 (8)	79 (9)
Body Mass Index (kg/m²)	36 (6)	36 (5)	36 (6)
Creatinine, µmol/L	81 (33)	90 (45)	84 (38)
Diabetes, n (%)	33 (37%)	11 (22%)	44 (31%)
Current smoker, n (%)	1 (1%)	5 (10%)	6 (4%)
Antihypertensive Medication Daily Defined Dose Sum	5.1 (2)	4.9 (2)	5.0 (2)
Sleep Characteristics
Pittsburgh Sleep Quality Inventory Score, m	7.3 (3.2)	7.3 (3.4)	7.3 (3.2)
Actigraphy Sleep Efficiency, %	77.2 (9.1)	77.5 (9.8)	77.3 (9.3)
Actigraphy Sleep Fragmentation, m	38.2 (11.6)	37.2 (12.6)	37.8 (11.9)
Actigraphy Awakenings Per Night, m	20.2 (6.9)	18.5 (7.0)	19.6 (7.0)
Actigraphy Length of Awakenings, mins	5.3 (2.4)	5.4 (2.7)	5.4 (2.5)
Reported Diagnosis of Sleep Apnea, n (%)	40 (44%)	23 (46%)	63 (45%)
STOP-BANG Score, m	5.1 (1.7)	5.1 (1.8)	5.1 (1.7)

Values are presented as mean (SD) unless otherwise indicated.

STOP-BANG is a commonly used screening tool for obstructive sleep apnea. Symptoms included Snoring, Tiredness, Observed choking or gasping while sleeping, high blood Pressure, Body mass index, Age, Neck size, and Gender. For each item, score responses are Yes = 1 and No = 0, with a resulting score range of 0–8. Scores of 0–2 suggest low OSA risk, scores of 3–4 suggest intermediate OSA risk, and scores of 5–8 suggest high OSA risk.

Table 2.

Changes in physical activity and sleep quality metrics from self-report and actigraphy assessments.

	Sleep Quality Indices at One-Year Follow-up(95% confidence interval)
Variable	C-LIFE	SEPA	Cohen's D
Pittsburgh Sleep Quality Index	6.7 (6.2, 7.3)	7.7 (6.9, 8.6)	0.45
Actigraphy Sleep Fragmentation	37.4 (35.2, 39.6)	38.8 (35.4, 42.3)	0.15
Actigraphy Sleep Efficiency	76.9 (75.3, 78.6)	75.6 (73.2, 78.0)	0.17
Sleep Time, Hours	6.0 (5.9, 6.2)	5.9 (5.6, 6.1)	0.21
Nighttime Awakenings	20.0 (18.9, 21.0)	18.7 (17.1, 20.3)	0.18
Length of Nighttime Awakenings	5.2 (4.7, 5.7)	5.9 (5.2, 6.6)	0.25

	Physical Activity Indices at One-Year Follow-up(95% confidence interval)
	C-LIFE	SEPA	Cohen's D
Moderate-Vigorous Physical Activity	712 (650, 774)	704 (606, 802)	0.02
Sedentary Activity, hours per day	5.4 (5.0, 5.8)	5.6 (5.0, 6.2)	−0.08
Light Activity per day	6.8 (6.4, 7.2)	7.3 (6.9, 7.7)	−0.24
K-Cals per day	2221 (2094, 2348)	2194 (2007, 2380)	0.04
Average Steps per day	9886 (9394, 10379)	9639 (8902, 10376)	0.09

One-year cognitive findings

Examination of cognitive changes from baseline to one-year follow-up were accomplished using reliable change indices. Across the cohort, most individuals showed small-to-moderate cognitive changes (mean global RCI = −0.40 [−0.52, −0.29], p < 0.001), with the largest changes in Psychomotor Speed (mean RCI = −0.63 [−0.89, −0.37], p < 0.001) and Executive Function/Learning (mean RCI = −0.50 [−0.61, −0.39], p < 0.001), without substantial Memory changes (mean RCI = 0.03 [−0.14, 0.21], p = 0.702). The majority of participants available for one-year assessments (73%) showed clinically meaningful declines on at least one cognitive subtest (i.e., < −1.96 RCI for subtest change from baseline to one-year), with 24% exhibiting one impairment, 18% exhibiting two impairments, and 31% impairments on ≥ three subtests. Examination of individual cognitive subtests revealed that the highest frequency of worsening of performance was observed on the Ruff 2&7 Test (17% and 18% for Speed and Accuracy, respectively), followed by the Stroop Color-Word Test (12%).

Examination of treatment group differences revealed main effect of treatment on our global RCI z-score, such that the C-LIFE group showed a lesser decline compared to SEPA (C-LIFE: z = −0.26 [−0.40, −0.12] versus SEPA: −0.60 [−0.81, −0.39]; d = 0.44, p = 0.008) (Figure 2). Examination of individual subtests (Table 3) demonstrated that the largest group differences were in Memory subtests, as observed on the CVLT-II Learning (d = 0.62) and CVLT-II Recall subtests (d = 0.36). Explanatory analyses of individual subtest changes in their raw form similarly revealed the largest group differences for CVLT-II Learning (C-LIFE = 49.1 [47.4, 50.9] versus SEPA = 43.8 [41.3, 46.3], p = 0.001), Immediate Recall (C-LIFE = 9.9 [9.3, 10.5] versus 8.8 [7.9, 9.7], p = 0.048), and Delayed Recall (C-LIFE = 10.4 [9.8, 11.0] versus 9.5 [8.7, 10.4], p = 0.112) (Supplemental Figure 2). Moreover, these improvements appeared to be paralleled by greater Learning Slope Efficiency, as evidenced by improvements in Semantic Clustering (C-LIFE = 1.2 [0.8, 1.5] versus SEPA = 0.5 [0, 0.9], p = 0.016) and Learning Slope (C-LIFE = 1.4 [1.3, 1.6] versus 1.2 [1.0, 1.4], p = 0.080). In contrast, we found no consistent improvements in measures of Visual Memory performance, such as the BVMT-R (BVMT-R Total Learning: d = 0.10, p = 0.272; BVMT-R Total Recall: d = 0.06, p = 0.650). Explanatory analyses of other cognitive outcomes revealed smaller and less consistent group differences in Executive Function / Learning and Psychomotor Speed subtests. For example, we observed modest changes on the Stroop Word subtest (d = 0.43, p = 0.002), Ruff 2&7 Test (d = 0.29, p = 0.039), and trends towards group differences on the Trail Making Test Part B (d = 0.35, p = 0.094) and Animal Naming (d = 0.39, p = 0.066) (Supplemental Figure 3).

Figure 2.

Treatment-related changes in our global, composite measure of cognitive change from reliable change indices. Participants in the Cardiac rehabilitation-based LIFEstyle (C-LIFE) and Standardized Education and Physician Advice (SEPA) groups were compared after adjusting for age, education (years), sex, baseline creatinine, baseline stroke risk (using the Framingham Stroke Risk Profile), and baseline cognitive function.

Table 3.

Changes in individual cognitive subtests using reliable change indices.

	One-Year Reliable Change, z-score (95% confidence interval)
Cognitive Function Subtest	C-LIFE	SEPA	Cohen's D
CVLT-II Total Learning	0.14 (−0.12, 0.39)	−0.63 (−1.01, −0.25)	0.62
CVLT-II Recall	−0.19 (−0.45, 0.06)	−0.61 (−1.01, −0.20)	0.36
Digit Span Forwards	0.11 (−0.36, 0.58)	−0.64 (−1.34, 0.05)	0.35
Ruff 2&7 Controlled Speed	−0.41 (−0.86, 0.04)	−1.14 (−1.82, −0.47)	0.31
Ruff 2&7 Automatic Speed	−0.56 (−1.13, 0.00)	−1.33 (−2.18, −0.49)	0.26
Ruff 2&7 Test Controlled Accuracy	−0.94 (−1.29, −0.58)	−1.37 (−1.89, −0.86)	0.25
Animal Fluency	−0.04 (−0.44, 0.35)	−0.42 (−1.08, 0.24)	0.22
Trail Making Part B	−0.05 (−0.45, 0.35)	−0.30 (−0.95, 0.35)	0.14
Digit Span Backwards	−0.05 (−0.39, 0.30)	−0.23 (−0.74, 0.27)	0.12
BVMT-R Total Learning	0.51 (0.15, 0.87)	0.33 (−0.16, 0.82)	0.11
Controlled Oral Word Association Test	0.06 (−0.27, 0.39)	−0.05 (−0.61, 0.52)	0.07
Ruff 2&7 Test Automatic Accuracy	−1.36 (−1.74, −0.97)	−1.49 (−2.04, −0.93)	0.07
BVMT-R Recall	0.34 (0.00, 0.68)	0.39 (−0.11, 0.90)	−0.03
Trail Making Test Part A	−0.30 (−0.60, 0.00)	−0.18 (−0.69, 0.34)	−0.07
Stroop Test, Color-Word Section	−1.16 (−1.43, −0.90)	−1.03 (−1.52, −0.55)	−0.11

Models are from multiple imputation analyses, adjusted for age, education, sex, creatinine, and Framingham Stroke Risk Profile.

One-year cerebrovascular findings

Examination of cerebrovascular reactivity and task-related hemodynamic outcomes are shown in Table 4. As shown, we found no consistent group differences at one-year across any of our fNIRS outcomes, including changes in TOI (p = 0.489) and oxygenated hemoglobin (p = 0.964) during a breath-holding challenge. Similarly, analyses of changes in our global score of task-based oxygenated hemoglobin efficiency showed no group differences at one-year (B = −0.01 [−0.02, 0.01], p = 0.126).

Table 4.

Changes in cognitive and hemodynamic responses during cognitive challenge tasks.

Domain	C-LIFE	SEPA	Cohen's D, Contrast p-value
Cerebrovascular Reactivity: Breath Holding Task
Δ Oxygenated Hemoglobin (HbO₂)	36.3 (29.7, 43.0)	40.5 (30.8, 50.3)	Cohen's D = 0.05, p= 0.489
Δ Tissue Oxygenation Index (TOI)	36.2 (29.5, 42.9)	35.9 (26.3, 45.5)	Cohen's D = 0.05, p = 0.964
Cognitive Challenge Task
Semantic Fluency	14.7 (13.7, 15.8)	15.2 (13.3, 17.0)	Cohen's D = 0.10, p = 0.770
Mental Arithmetic (# correct [max = 12])	8.9 (8.2, 9.5)	8.3 (7.0, 9.6)	Cohen's D = 0.17, p = 0.493
Phonemic Fluency	15.5 (14.2, 16.7)	16.9 (14.9, 19.0)	Cohen's D = 0.11, p = 0.191
Hemodynamic Changes
Semantic Fluency, HbO₂	35.9 (29.1, 42.8)	44.9 (32.7, 57.1)	Cohen's D = 0.33, p = 0.173
Mental Arithmetic, HbO₂	39.0 (33.0, 45.0)	45.5 (33.5, 57.6)	Cohen's D = 0.17, p = 0.437
Phonemic Fluency, HbO₂	40.4 (33.2, 47.5)	37.3 (27.7, 47.0)	Cohen's D = 0.20, p = 0.619
Hemodynamic Efficiency (HbO₂ Change / Cognitive Performance)
Semantic Fluency, HbO₂ / Performance	34.4 (27.5, 41.2)	49.7 (38.4, 61.1)	Cohen's D = 0.42, p = 0.025
Mental Arithmetic, HbO₂ / Performance	37.5 (31.7, 43.3)	45.2 (36.1, 54.3)	Cohen's D = 0.22, p = 0.168
Phonemic Fluency, HbO₂ / Performance	43.3 (35.9, 50.6)	35.7 (23.9, 47.5)	Cohen's D = 0.21, p = 0.280
Global Score, HbO₂ / Performance (mean rank)	68.1 (60.8, 75.4)	75.6 (64.6, 86.6)	Cohen's D = 0.20, p = 0.126

Outcomes of primary interest are italicized and included changes in oxygenated hemoglobin and tissue oxygenation index levels during a breath holding task. We also explored changes in task-based hemodynamic recruitment while participants performed three different cognitive challenge tasks.

Longitudinal associations with one-year cognitive function and cerebrovascular outcomes

Examination of correlates of one-year cognitive changes were also conducted. In an examination of changes in putative mechanistic markers and cognitive outcomes, we found that improvements in (reductions in) PSQI scores associated with global cognitive changes (B = −0.17 every 3-point PSQI reduction, p = 0.038) (Figure 3 ), whereas increased steps (B = −0.01 for every 2500 steps, p = 0.822), DASH diet scores (B = −0.04 for every 2-point change, p = 0.502), weight loss (B = 0.08 for every 15-lbs of weight loss, p = 0.270), and ambulatory SBP reductions (B = 0.01 for every 15-mmHg reduction, p = 0.939) were not associated with cognitive changes. Exploratory analysis of persistent treatment group effects demonstrated that treatment group differences remained significant in a final model incorporating the group contrast and putative mechanistic markers (B = 0.12, p = 0.005).

Figure 3.

Changes in the Pittsburgh Sleep Quality Index and global cognitive scores from reliable change indices after adjusting age, education (years), sex, baseline creatinine, baseline stroke risk (using the Framingham Stroke Risk Profile), and baseline cognitive function. RCI: reliable change indices.

Figure 4.

Changes in the weight from baseline to one-year follow-up and cerebrovascular reactivity (CVR) changes from functional near infrared spectroscopy (fNIRS) assessments. CVR was assessed using tissue oxygenation index changes during a breath holding provocative task for up to 30 s. CVR was residualized for age, education (years), sex, baseline creatinine, baseline stroke risk (using the Framingham Stroke Risk Profile), and baseline CVR levels.

In analyses examining correlates of cerebrovascular outcomes we found that greater amounts of weight loss (ΔTOI Saturation: 0.38 [0.06, 0.71] per 15 lbs of weight loss from baseline; p = 0.023) (Figure 4), whereas we found no such relationships with ambulatory SBP (0.06 [−0.29, 0.41], p = 0.757), DASH diet scores (−0.02 [−0.4, 0.3], p = 0.903), step count (0.01 [−0.3, 0.3], p = 0.960), or PSQI scores (0.01 [−0.4, 0.4], p = 0.980). Similarly, we found no associations between putative correlates of cerebrovascular outcomes and oxygenated hemoglobin changes: ambulatory SBP (p = 0.881), DASH diet scores (p = 0.993), step count (p = 0.371), or PSQI scores (p = 0.763).

In parallel analyses of task-based oxygenated hemoglobin changes, we found that greater levels of actigraphy-assessed physical activity associated with greater cognitive efficiency (as indicated by lower oxygenation per cognitive performance unit: B = −9.3 [−18.4, −0.3], p = 0.044). In contrast, changes in sleep (p = 0.177), ambulatory SBP (p = 0.981), DASH diet (p = 0.782), and weight (p = 0.659) were not associated with changes in cognitive efficiency levels.

Discussion

Our finding demonstrated that a cardiac rehabilitation-based lifestyle program among individuals with TRH, targeting physical activity and behavioral weight loss, associated with modestly preserved cognitive function over a one-year follow-up compared to individuals receiving an enhanced standard of care. In contrast, we did not observe parallel associations with cerebrovascular function or in several biobehavioral outcomes. These findings extend prior evidence linking lifestyle modification to improved cognitive outcomes among individuals with TRH by suggesting that sustained lifestyle adherence may help to preserve cognitive and cerebrovascular function among some individuals with TRH. Surprisingly, our findings suggested that sleep quality improvements following lifestyle modification may associate with greater preservation of cognitive function, although this association was not hypothesized and requires replication. In addition, it is notable that treatment group differences were strongest for some aspects of memory performance, whereas we observed less consistent group differences in other cognitive domains. In parallel, we found that greater weight loss and physical activity levels associated with different cerebrovascular outcomes. Finally, our findings were somewhat surprising in failing to show an association between ambulatory blood pressure, physical activity, dietary quality, and weight loss with preserved cognitive function.

Numerous prior studies have demonstrated that lifestyle modification could help sustain cognitive functioning among older adults and those with chronic medical conditions. Multimodal interventions targeting vascular risk factors using exercise and diet have provided the strongest evidence of cognitive benefits, such as from the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) trial,^63,64 the PreDIVA trial,⁶⁵ MAPT⁶⁶ and most recently the U.S. Pointer trial,⁶⁷ among others.^68–70 Moreover, recent data from the SPRINT MIND trial⁶ suggest that intensive blood pressure lowering results in reduced risk of MCI and dementia, as well as reducing white matter lesion burden on magnetic resonance brain imaging.⁷¹ Finally, prior meta-analyses have demonstrated that programs increasing physical activity or implementing behavioral weight loss appear to improve cognition across multiple domains of function.^72,73

The lack of improvement in cerebrovascular function, as well as the lack of association between individual biobehavioral markers (e.g., physical activity and weight loss) with cognitive function in the present study is notable and warrants additional exploration. There are several potential explanations for this finding, in addition to the possibility that our analyses were underpowered to detect associations due to missing data associated with the COVID-19 lockdown. First, as we previously reported from the primary treatment findings in the TRIUMPH trial,²⁵ individuals in our enhanced standard of care control condition (SEPA) showed noteworthy, within-group changes in weight (–8.5 lbs), urinary sodium [–187 mg/day) and potassium levels (292 mg/day), in addition to reducing their blood pressure levels. Although the C-LIFE treatment group nevertheless showed greater improvements in most of these outcomes, the lifestyle changes made in the SEPA group were potentially impactful and may have made it more difficult to detect associations between lifestyle factors and cognitive changes. It is further possible that the intensive lifestyle intervention we provided was harder for individuals in the C-LIFE group to maintain even before the unprecedented complexities introduced by the COVID-19 lockdown. For example, preliminary estimates from one-year follow-up outcomes suggested comparable weight loss maintenance between groups (C-LIFE: −10.6 [−13.4, −7.9] lbs versus SEPA: −9.6 [−13.7, −5.6] lbs from baseline to one-year). These findings suggest that, although the C-LIFE group showed greater weight loss after 4-months of treatment, these improvements declined over follow-up. It is possible that individuals in the SEPA group, although showing lesser lifestyle changes and blood pressure benefits immediately following trial completion, may have been better able to sustain their health habits over time. Similarly, our failure to find an association between blood pressure reductions and cognitive functions could also have been influenced by the substantive reductions in the SEPA group, making it harder to discern group differences. Finally, differential dropout in our sample could plausibly have influenced the pattern of observed findings, as BMI, metabolic function, and baseline atherosclerosis all predicted drop-out status.

One notable nuance in the present study was the differential changes in cognition observed across domains when examined using reliable change indices. The majority of TRIUMPH participants appeared to show some level of cognitive decline, albeit subtle in many cases. Cognitive changes were most commonly observed in Executive Function/Learning and Psychomotor Speed domains, whereas Memory function was more stable. Observing changes in these domains among hypertensive patients is consistent with prior literature demonstrating preferential impairments on these tasks due to their greater vulnerability to microvascular ischemic changes.^1,74–78 Executive Functions, in particular, are conventionally thought of as heavily reliant on integrity within frontal-subcortical brain regions, which are sensitive to the effects of chronic hypertension.^8,79,80 Our finding that Memory performance was most improved relative to other cognitive domains may partly reflect this secular decline in other domains across both groups. It is also possible that our selection criteria, which would have eliminated individuals with observable cognitive impairments on the Montreal Cognitive Assessment Battery,^4,26 could have biased our sample to over-enroll individuals with cognitive impairments due to vascular risk factors and exclude individuals showing the memory changes that typify ADRD. In the present study, although treatment improvements were most notable on tasks of memory, explanatory analyses indicated improvements were likely driven by improved learning efficiency and complex information sequencing, which are conventionally associated with the dorsal attention network within the prefrontal cortex.⁸¹ Memory is closely associated with attention and concentration abilities for many individuals and our findings seem consistent with the notation that improvements in some aspects of selective attention, a component of executive function that has an important influence on memory abilities, may have played an important role.⁸²

Recent studies have reported similar findings, demonstrating that physical activity and sleep quality associate independently with cognitive functioning.⁸³ Interestingly, an emerging body of data have extended these findings^84,85 to suggest that physical activity and sleep may have moderating effects on cognitive and brain health outcomes, with greater physical activity levels protecting against the untoward effects of poor sleep.^86,87 These findings underscore the need for future trials to explore the potentially moderating effects of different lifestyle components on cognitive outcomes. This is particularly important as existing evidence not only suggests that aerobic exercise training improves sleep outcomes,⁸⁸ but that treating sleep dysfunction alone does not appear sufficient to improve cognitive function.⁸⁹ Moreover, improvements in cognitive function following aerobic exercise training among individuals with MCI do not appear to be paralleled by improved sleep,⁹⁰ suggesting that both lifestyle factors interact in their impact on brain outcomes.

Although we failed to find treatment group differences on any fNIRS indices, observed associations between biobehavioral factors and fNIRS outcomes are potentially interesting and warrant further study. Our finding that changes in cerebrovascular reactivity, assessed by fNIRS, associated with changes in actigraphy-assessed is consistent with prior trial data suggesting that physical activity improves cerebrovascular functions.⁹¹ Similarly, prior studies examining task-based changes in hemodynamic recruitment have suggested that physical activity improves hemodynamic efficiency during cognitive task performance. For example, in a one-year study examining fNIRS correlates of physical activity among individuals with MCI, MCI individuals engaging in physical activity showed decreased task-related hemodynamic responses despite stable cognitive performance.⁹² Similarly, our finding that greater weight loss associated with greater CVR changes during a breath holding task parallels prior work suggesting that improved cardiorespiratory fitness (which overlaps substantially with weight) associates with increased cerebrovascular functions.¹⁰

The present study should be viewed with several important limitations in mind. First, these findings were observed in a small and select cohort of individuals with resistant hypertension and require replication in a larger cohort. We were further limited by loss of data due to the COVID-19 pandemic lockdown, which coincided with the post-treatment assessments for final participants in the active treatment portion of the TRIUMPH trial. Second, we are limited in our ability to follow up with all participants due to distancing requirements that occurred due to the COVID-19 lockdown. Our findings are therefore encouraging in that we found improvements in memory despite our small sample size and underpowered statistical analyses. Third, our study sample was highly selected, such that more highly educated individuals with access to cardiac rehabilitation were preferentially included due to fidelity considerations. Fourth, our use of cognitive testing could have been enhanced by the inclusion of ancillary neuroimaging outcomes or ADRD biomarkers, which would have provided a more comprehensive assessment of the role lifestyle modification has on systemic markers of brain health.

In conclusion, our findings suggest that a comprehensive lifestyle modification program can improve cognitive function among individuals with resistant hypertension. These findings further emphasize the importance of comprehensive lifestyle changes as a means of protecting against the deleterious effects of hypertension among aging adults. Our findings also suggest that, even in a sample of individuals who have a high burden of cardiovascular risk factors and are refractory to pharmacological therapy for hypertension, changing lifestyle habits may nevertheless result in improved cognitive functioning and help to reduce the risk of ADRD.

Supplemental Material

sj-docx-1-alz-10.1177_13872877251409716 - Supplemental material for Lifestyle, sleep quality, and cognitive function in resistant hypertension: One-year follow-up from the TRIUMPH trial

Supplemental material, sj-docx-1-alz-10.1177_13872877251409716 for Lifestyle, sleep quality, and cognitive function in resistant hypertension: One-year follow-up from the TRIUMPH trial by Patrick J. Smith, James A. Blumenthal, Krista Ingle, Lana L. Watkins, Forgive Avorgbedor, Stephanie K. Mabe, William Kraus, Crystal Tyson, Alan Hinderliter and Andrew Sherwood in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

Competing interests and sources of financial support: this work was supported by Grants HL122836 and HL130237 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD. All authors declare no relevant financial interests.

ORCID iDs

Patrick J. Smith

Stephanie K. Mabe

Andrew Sherwood

Ethical considerations

Participants in the current study were monitored by the Duke University Institutional Review Board to ensure ethical treatment and clinical research practices.

Consent to participate

All participants provided informed consent to participate.

Consent for publication

Not applicable

Author contribution(s)

Patrick J. Smith: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Writing – original draft; Writing – review & editing.

James A. Blumenthal: Conceptualization; Investigation; Writing – original draft; Writing – review & editing.

Krista Ingle: Investigation; Writing – original draft; Writing – review & editing.

Lana L. Watkins: Conceptualization; Investigation; Writing – original draft.

Forgive Avorgbedor: Investigation; Writing – original draft; Writing – review & editing.

Stephanie K. Mabe: Investigation; Methodology; Writing – original draft.

William Kraus: Conceptualization; Funding acquisition; Investigation; Writing – original draft.

Crystal Tyson: Investigation; Writing – original draft.

Alan Hinderliter: Investigation; Methodology; Writing – original draft.

Andrew Sherwood: Conceptualization; Funding acquisition; Investigation; Methodology; Writing – original draft.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Grants HL122836 and HL130237 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD. 1R61AG080615-01 from the National Institute of Aging.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

Select data from the present analyses may be available upon request from the corresponding author.

Supplemental material

Supplemental material for this article is available online.

References

Cheon

. Hypertension and cognitive dysfunction: a narrative review. J Yeungnam Med Sci 2023; 40: 225–232.

Canavan

O'Donnell

. Hypertension and cognitive impairment: a review of mechanisms and key concepts. Front Neurol 2022; 13: 821135.

Muli

Meisinger

Heier

, et al. Prevalence, awareness, treatment, and control of hypertension in older people: results from the population-based KORA-age 1 study. BMC Public Health 2020; 20: 1049.

Blumenthal

Sherwood

Smith

, et al. Lifestyle modification for resistant hypertension: the TRIUMPH randomized clinical trial. Am Heart J 2015; 170: 986–994 e985.

Badji

Youwakim

Cooper

, et al. Vascular cognitive impairment - Past, present, and future challenges. Ageing Res Rev 2023; 90: 102042.

Kjeldsen

Narkiewicz

Burnier

, et al. Intensive blood pressure lowering prevents mild cognitive impairment and possible dementia and slows development of white matter lesions in brain: the SPRINT Memory and Cognition IN Decreased Hypertension (SPRINT MIND) study. Blood Press 2018; 27: 247–248.

Yang

Moore

Mpofu

, et al. Effectiveness of combined cognitive and physical interventions to enhance functioning in older adults with mild cognitive impairment: a systematic review of randomized controlled trials. Gerontologist 2020; 60: 633–642.

Sanders

LMJ

Hortobagyi

la Bastide-van Gemert

, et al. Dose-response relationship between exercise and cognitive function in older adults with and without cognitive impairment: a systematic review and meta-analysis. PLoS One 2019; 14: e0210036.

Pacholko

Iadecola

. Hypertension, neurodegeneration, and cognitive decline. Hypertension 2024; 81: 991–1007.

10.

DuBose

Weng

Pierce

, et al. Association between cardiorespiratory fitness and cerebrovascular reactivity to a breath-hold stimulus in older adults: influence of aerobic exercise training. J Appl Physiol 2022; 132: 1468–1479.

11.

Smith

Whitson

Merwin

, et al. Engineering virtuous health habits using emotion and neurocognition: flexibility for lifestyle optimization and weight management (EVEN FLOW). Front Aging Neurosci 2023; 15: 1256430.

12.

Fianu

Bourse

Naty

, et al. Long-term effectiveness of a lifestyle intervention for the primary prevention of type 2 diabetes in a low socio-economic community–an intervention follow-up study on Reunion Island. PLoS One 2016; 11: e0146095.

13.

Matton

Stephen

Daniilidou

, et al. Mechanisms of interventions targeting modifiable factors for dementia risk reduction. Mol Neurodegener 2025; 20: 75.

14.

Mielke

Evans

Neiberg

, et al. Alzheimer disease blood biomarkers and cognition among individuals with diabetes and overweight or obesity. JAMA Netw Open 2025; 8: e2458149.

15.

Voss

Heo

Prakash

, et al. The influence of aerobic fitness on cerebral white matter integrity and cognitive function in older adults: results of a one-year exercise intervention. Hum Brain Mapp 2013; 34: 2972–2985.

16.

Veronese

Facchini

Stubbs

, et al. Weight loss is associated with improvements in cognitive function among overweight and obese people: a systematic review and meta-analysis. Neurosci Biobehav Rev 2017; 72: 87–94.

17.

Joseph

Alonso-Alonso

Bond

, et al. The neurocognitive connection between physical activity and eating behaviour. Obes Rev 2011; 12: 800–812.

18.

Smith

Blumenthal

Babyak

, et al. Effects of the dietary approaches to stop hypertension diet, exercise, and caloric restriction on neurocognition in overweight adults with high blood pressure. Hypertension 2010; 55: 1331–1338.

19.

Barnes

Corkery

. Exercise improves vascular function, but does this translate to the brain? Brain Plast 2018; 4: 65–79.

20.

Etnier

Nowell

Landers

, et al. A meta-regression to examine the relationship between aerobic fitness and cognitive performance. Brain Res Rev 2006; 52: 119–130.

21.

Sewell

Rainey-Smith

Villemagne

, et al. The interaction between physical activity and sleep on cognitive function and brain beta-amyloid in older adults. Behav Brain Res 2023; 437: 114108.

22.

Sewell

Erickson

Rainey-Smith

, et al. Relationships between physical activity, sleep and cognitive function: a narrative review. Neurosci Biobehav Rev 2021; 130: 369–378.

23.

Smith

Sherwood

Avorgbedor

, et al. Sleep quality, metabolic function, physical activity, and neurocognition among individuals with resistant hypertension. J Alzheimers Dis 2023; 93: 995–1006.

24.

Geer

Jeon

O'Connell

, et al. Correlates of cognition among people with chronic heart failure and insomnia. Sleep Breath 2023; 27: 1287–1296.

25.

Blumenthal

Hinderliter

Smith

, et al. Effects of lifestyle modification on patients with resistant hypertension: results of the TRIUMPH randomized clinical trial. Circulation 2021; 144: 1212–1226.

26.

Smith

Sherwood

Hinderliter

, et al. Lifestyle modification and cognitive function among individuals with resistant hypertension: cognitive outcomes from the TRIUMPH trial. J Hypertens 2022; 40: 1359–1368.

27.

Hachinski

Iadecola

Petersen

, et al. National Institute of Neurological Disorders and Stroke-Canadian Stroke Network vascular cognitive impairment harmonization standards. Stroke 2006; 37: 2220–2241.

28.

Gavett

Ashendorf

Gurnani

. Reliable change on neuropsychological tests in the Uniform Data Set. J Int Neuropsychol Soc 2015; 21: 558–567.

29.

Attix

Story

Chelune

, et al. The prediction of change: normative neuropsychological trajectories. Clin Neuropsychol 2009; 23: 21–38.

30.

Hammers

Porter

Dixon

, et al. Validating 1-year reliable change methods. Arch Clin Neuropsychol 2021; 36: 87–98.

31.

Woods

Delis

Scott

, et al. The California Verbal Learning Test–second edition: test-retest reliability, practice effects, and reliable change indices for the standard and alternate forms. Arch Clin Neuropsychol 2006; 21: 413–420.

32.

Knight

McMahon

Skeaff

, et al. Reliable change indices for the ruff 2 and 7 selective attention test in older adults. Appl Neuropsychol 2010; 17: 239–245.

33.

Smith

Sherwood

Hinderliter

, et al. Cerebrovascular function, vascular risk, and lifestyle patterns in resistant hypertension. J Alzheimers Dis 2022; 87: 345–357.

34.

Stricker

Brown

Wixted

, et al. New semantic and serial clustering indices for the California Verbal Learning Test-Second Edition: background, rationale, and formulae. J Int Neuropsychol Soc 2002; 8: 425–435.

35.

Stepanov

Abramson

Hoogs

, et al. Overall memory impairment identification with mathematical modeling of the CVLT-II learning curve in multiple sclerosis. Mult Scler Int 2012; 2012: 312503.

36.

Huppert

Hoge

Diamond

, et al. A temporal comparison of BOLD, ASL, and NIRS hemodynamic responses to motor stimuli in adult humans. Neuroimage 2006; 29: 368–382.

37.

Kastrup

Glover

, et al. Cerebral blood flow-related signal changes during breath-holding. AJNR Am J Neuroradiol 1999; 20: 1233–1238.

38.

Kastrup

Takahashi

, et al. Functional magnetic resonance imaging of regional cerebral blood oxygenation changes during breath holding. Stroke 1998; 29: 2641–2645.

39.

Zavoreo

Demarin

. Breath holding index and arterial stiffness as markers of vascular aging. Curr Aging Sci 2010; 3: 67–70.

40.

Apruzzese

Silvestrini

Floris

, et al. Cerebral hemodynamics in asymptomatic patients with internal carotid artery occlusion: a dynamic susceptibility contrast MR and transcranial Doppler study. AJNR Am J Neuroradiol 2001; 22: 1062–1067.

41.

Cupini

Diomedi

Placidi

, et al. Cerebrovascular reactivity and subcortical infarctions. Arch Neurol 2001; 58: 577–581.

42.

Zavoreo

Kes

Morovic

, et al. Breath holding index in detection of early cognitive decline. J Neurol Sci 2010; 299: 116–119.

43.

Reinhard

Schwarzer

Briel

, et al. Cerebrovascular reactivity predicts stroke in high-grade carotid artery disease. Neurology 2014; 83: 1424–1431.

44.

Vasdekis

Tsivgoulis

Athanasiadis

, et al. Cerebrovascular reacivity assessment in patients with carotid artery disease: a combined TCD and NIRS study. J Neuroimaging 2012; 22: 261–265.

45.

Yucebas

Fox-Fuller

Badillo Cabrera

, et al. Associations of category fluency clustering performance with in vivo brain pathology in autosomal dominant Alzheimer's disease. J Int Neuropsychol Soc 2024; 30: 77–83.

46.

Udina

Avtzi

Durduran

, et al. Functional near-infrared spectroscopy to study cerebral hemodynamics in older adults during cognitive and motor tasks: a review. Front Aging Neurosci 2019; 11: 367.

47.

Lingling

Huaqing

Xuan

, et al. Acute effect of low-intensity aerobic exercise on eliciting enhanced parietal activation and promoting executive function performance more than moderate-intensity exercise. Front Physiol 2025; 16: 1581481.

48.

Buysse

Hall

Strollo

, et al. Relationships between the Pittsburgh Sleep Quality Index (PSQI), Epworth Sleepiness Scale (ESS), and clinical/polysomnographic measures in a community sample. J Clin Sleep Med 2008; 4: 563–571.

49.

Goodwin

Bilous

Winship

, et al. Validation of the Oscar 2 oscillometric 24-h ambulatory blood pressure monitor according to the British Hypertension Society protocol. Blood Press Monit 2007; 12: 113–117.

50.

O'Brien

Parati

Stergiou

. Ambulatory blood pressure measurement: what is the international consensus? Hypertension 2013; 62: 988–994.

51.

Min

Aubert

, et al. A method to quantify mean hypertension treatment daily dose intensity using health care system data. JAMA Netw Open 2021; 4: e2034059.

52.

Eck

Klesges

Hanson

, et al. Measuring short-term dietary intake: development and testing of a 1-week food frequency questionnaire. J Am Diet Assoc 1991; 91: 940–945.

53.

Bazzano

Ogden

, et al. Agreement on nutrient intake between the databases of the First National Health and Nutrition Examination Survey and the ESHA Food Processor. Am J Epidemiol 2002; 156: 78–85.

54.

Smith

Mabe

Sherwood

, et al. Metabolic and neurocognitive changes following lifestyle modification: examination of biomarkers from the ENLIGHTEN randomized clinical trial. J Alzheimers Dis 2020; 77: 1793–1803.

55.

Blumenthal

Sherwood

Smith

, et al. Enhancing cardiac rehabilitation with stress management training: a randomized, clinical efficacy trial. Circulation 2016; 133: 1341–1350.

56.

Sherwood

Blumenthal

Koch

, et al. Effects of coping skills training on quality of life, disease biomarkers, and clinical outcomes in patients with heart failure: a randomized clinical trial. Circ Heart Fail 2017; 10: e003410.

57.

O'Brien

. Procedures for comparing samples with multiple endpoints. Biometrics 1984; 40: 1079–1087.

58.

Koch

Gansky

. Statistical considerations for multiplicity in confirmatory protocols. Drug Inform J 1996; 30: 523–533.

59.

Dmitrienko

D'Agostino Sr

. Multiplicity considerations in clinical trials. N Engl J Med 2018; 378: 2115–2122.

60.

Dmitrienko

. Special issue: multiplicity issues in clinical trials. J Biopharm Stat 2018; 28: 1–2.

61.

O'Rourke

MacKinnon

. Reasons for testing mediation in the absence of an intervention effect: a research imperative in prevention and intervention research. J Stud Alcohol Drugs 2018; 79: 171–181.

62.

Whittle

Mansell

Jellema

, et al.

Applying causal mediation methods to clinical trial data: what can we learn about why our interventions (don't) work?

Eur J Pain 2017; 21: 614–622.

63.

Kivipelto

Solomon

Ahtiluoto

, et al. The Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER): study design and progress. Alzheimers Dement 2013; 9: 657–665.

64.

Ngandu

Lehtisalo

Solomon

, et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised controlled trial. Lancet 2015; 385: 2255–2263.

65.

van Charante

EPM

Richard

Eurelings

, et al. Effectiveness of a 6-year multidomain vascular care intervention to prevent dementia (preDIVA): a cluster-randomised controlled trial. Lancet 2016; 388: 797–805.

66.

Vellas

Carrie

Gillette-Guyonnet

, et al. MAPT Study: a multidomain approach for preventing Alzheimer's disease: design and baseline data. J Prev Alzheimers Dis 2014; 1: 13–22.

67.

Baker

Espeland

Whitmer

, et al. Structured vs self-guided multidomain lifestyle interventions for global cognitive function: the US POINTER randomized clinical trial. JAMA 2025; 334: 681–691.

68.

Coley

Ngandu

Lehtisalo

, et al. Adherence to multidomain interventions for dementia prevention: data from the FINGER and MAPT trials. Alzheimers Dement 2019; 15: 729–741.

69.

Kivipelto

Mangialasche

Snyder

, et al. World-Wide FINGERS Network: a global approach to risk reduction and prevention of dementia. Alzheimers Dement 2020; 16: 1078–1094.

70.

Pang

Zhou

, et al. The multi-domain lifestyle intervention for cognitive impairment in community-dwelling older adults in Hangzhou (The Heritage Study): study design and protocol. J Prev Alzheimers Dis 2024; 11: 601–611.

71.

SPRINT MIND Investigators for the SPRINT Research Group, Nasrallah

Pajewski

, et al. Association of intensive vs standard blood pressure control with cerebral white matter lesions. JAMA 2019; 322: 524–534.

72.

Ngandu

Lehtisalo

Korkki

, et al. The effect of adherence on cognition in a multidomain lifestyle intervention (FINGER). Alzheimers Dement 2022; 18: 1325–1334.

73.

Kivipelto

Mangialasche

Ngandu

. Lifestyle interventions to prevent cognitive impairment, dementia and Alzheimer disease. Nat Rev Neurol 2018; 14: 653–666.

74.

Joyce

McHugh

Mockler

, et al. Midlife hypertension is a risk factor for some, but not all, domains of cognitive decline in later life: a systematic review and meta-analysis. J Hypertens 2024; 42: 205–223.

75.

Hua

Yang

, et al. Different hypertension thresholds and cognitive decline: a pooled analysis of three ageing cohorts. BMC Med 2021; 19: 287.

76.

Moraes

Muela

HCS

MemOria

, et al. Systemic arterial hypertension and cognition in adults: effects on executive functioning. Arq Neuropsiquiatr 2020; 78: 412–418.

77.

Alvarez-Bueno

Cunha

Martinez-Vizcaino

, et al. Arterial stiffness and cognition among adults: a systematic review and meta-analysis of observational and longitudinal studies. J Am Heart Assoc 2020; 9: e014621.

78.

Hughes

Sink

Williamson

, et al. Relationships between cerebral structure and cognitive function in African Americans with type 2 diabetes. J Diabetes Complications 2018; 32: 916–921.

79.

Zhou

Wang

, et al. Effects of exercise interventions for specific cognitive domains in old adults with mild cognitive impairment: a protocol of subgroup meta-analysis of randomized controlled trials. Medicine (Baltimore) 2018; 97: e13244.

80.

Smith

Blumenthal

Hoffman

, et al. Aerobic exercise and neurocognitive performance: a meta-analytic review of randomized controlled trials. Psychosom Med 2010; 72: 239–252.

81.

Carnevale

Maffei

Landolfi

, et al. Brain functional magnetic resonance imaging highlights altered connections and functional networks in patients with hypertension. Hypertension 2020; 76: 1480–1490.

82.

deBettencourt

Williams

Vogel

, et al. Sustained attention and spatial attention distinctly influence long-term memory encoding. J Cogn Neurosci 2021; 33: 2132–2148.

83.

Falck

Best

Davis

, et al. Sleep and cognitive function in chronic stroke: a comparative cross-sectional study. Sleep 2019; 42: zsz040.

84.

Sandison

Callan

NGL

Rao

, et al. Observed improvement in cognition during a personalized lifestyle intervention in people with cognitive decline. J Alzheimers Dis 2023; 94: 993–1004.

85.

Mace

Lyons

Cohen

, et al. Optimizing the implementation of a lifestyle dementia prevention intervention for older patients in an academic healthcare system. J Alzheimers Dis 2024; 100: 1237–1259.

86.

Simon

Nadel

Payne

. The functions of sleep: a cognitive neuroscience perspective. Proc Natl Acad Sci U S A 2022; 119: e2201795119.

87.

Callow

Zipunnikov

Spira

, et al. Actigraphy estimated sleep moderates the relationship between physical activity and cognition in older adults. Ment Health Phys Act 2024; 26: 100573.

88.

Dolezal

Neufeld

Boland

, et al. Interrelationship between sleep and exercise: a systematic review. Adv Prev Med 2017; 2017: 1364387.

89.

Franks

Rowsthorn

Nicolazzo

, et al. The treatment of sleep dysfunction to improve cognitive function: a meta-analysis of randomized controlled trials. Sleep Med 2023; 101: 118–126.

90.

Ahn

Kim

. Effects of aerobic exercise on global cognitive function and sleep in older adults with mild cognitive impairment: a systematic review and meta-analysis. Geriatr Nurs 2023; 51: 9–16.

91.

Shen

Hou

Xia

, et al. Exercise promotes brain health: a systematic review of fNIRS studies. Front Psychol 2024; 15: 1327822.

92.

Talamonti

Gagnon

Vincent

, et al. Exploring cognitive and brain oxygenation changes over a 1-year period in physically active individuals with mild cognitive impairment: a longitudinal fNIRS pilot study. BMC Geriatr 2022; 22: 648.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.58 MB