Hostname: page-component-6898cfcb55-w5hgh Total loading time: 0 Render date: 2026-11-11T23:59:29.042Z Has data issue: false hasContentIssue false
  • English
  • Français

Relative associations of behavioral and physiological risks for cardiometabolic disease with cognition in bipolar disorder during mid and later-life: findings from the UK biobank

Published online by Cambridge University Press:  02 April 2024

Elysha RinginOpen the ORCID record for Elysha Ringin[Opens in a new window],
David W. Dunstan,
Denny Meyer,
Roger S. McIntyre,
Neville Owen,
Michael Berk,
Mats Hallgren,
Susan L. Rossell and
Tamsyn E. Van Rheenen
Elysha Ringin*
Affiliation:
Melbourne Neuropsychiatry Centre, Department of Psychiatry, University of Melbourne and Melbourne Health, Melbourne, Australia
David W. Dunstan
Affiliation:
Physical Activity Laboratory, Baker Heart & Diabetes Institute, Melbourne, VIC, AustraliaDeakin University, Institute for Physical Activity and Nutrition (IPAN), School of Exercise and Nutrition Sciences, Geelong, Australia
Denny Meyer
Affiliation:
Centre for Mental Health, School of Health Sciences, Swinburne University, Melbourne, Australia
Roger S. McIntyre
Affiliation:
Department of Psychiatry and Pharmacology, University of Toronto, Toronto, Canada
Neville Owen
Affiliation:
Physical Activity Laboratory, Baker Heart & Diabetes Institute, Melbourne, VIC, AustraliaCentre for Urban Transitions, Swinburne University of Technology, Melbourne, VIC, Australia
Michael Berk
Affiliation:
Deakin University, The Institute for Mental and Physical Health and Clinical Translation, Barwon Health, Geelong, VIC, AustraliaDepartment of Psychiatry, University of Melbourne, Melbourne, AustraliaOrygen Youth Health, Melbourne, Australia
Mats Hallgren
Affiliation:
Epidemiology of Psychiatric Conditions, Substance Use and Social Environment (EPiCSS), Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden
Susan L. Rossell
Affiliation:
Centre for Mental Health, School of Health Sciences, Swinburne University, Melbourne, AustraliaSt Vincent's Mental Health, St Vincent's Hospital, Melbourne, VIC, Australia
Tamsyn E. Van Rheenen*
Affiliation:
Melbourne Neuropsychiatry Centre, Department of Psychiatry, University of Melbourne and Melbourne Health, Melbourne, AustraliaCentre for Mental Health, School of Health Sciences, Swinburne University, Melbourne, Australia
*
Corresponding author: Elysha Ringin; Email: eringin@student.unimelb.edu.au; Tamsyn E. Van Rheenen; Email: tamsyn.van@unimelb.edu.au
Corresponding author: Elysha Ringin; Email: eringin@student.unimelb.edu.au; Tamsyn E. Van Rheenen; Email: tamsyn.van@unimelb.edu.au
Rights & Permissions [Opens in a new window]

Abstract

Background

Cardiometabolic disease risk factors are disproportionately prevalent in bipolar disorder (BD) and are associated with cognitive impairment. It is, however, unknown which health risk factors for cardiometabolic disease are relevant to cognition in BD. This study aimed to identify the cardiometabolic disease risk factors that are the most important correlates of cognitive impairment in BD; and to examine whether the nature of the relationships vary between mid and later life.

Methods

Data from the UK Biobank were available for 966 participants with BD, aged between 40 and 69 years. Individual cardiometabolic disease risk factors were initially regressed onto a global cognition score in separate models for the following risk factor domains; (1) health risk behaviors (physical activity, sedentary behavior, smoking, and sleep) and (2) physiological risk factors, stratified into (2a) anthropometric and clinical risk (handgrip strength, body composition, and blood pressure), and (2b) cardiometabolic disease risk biomarkers (CRP, lipid profile, and HbA1c). A final combined multivariate regression model for global cognition was then fitted, including only the predictor variables that were significantly associated with cognition in the previous models.

Results

In the final combined model, lower mentally active and higher passive sedentary behavior, higher levels of physical activity, inadequate sleep duration, higher systolic and lower diastolic blood pressure, and lower handgrip strength were associated with worse global cognition.

Conclusions

Health risk behaviors, as well as blood pressure and muscular strength, are associated with cognitive function in BD, whereas other traditional physiological cardiometabolic disease risk factors are not.

Keywords

bipolar disordercardiometabolic diseasecognitiondepressionhealth risk behaviorslifestylemaniamental healthphysiological riskspsychiatry

Information

Type
Original Article
Information
Psychological Medicine , Volume 54 , Issue 10 , July 2024, pp. 2612 - 2622
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
Copyright © The Author(s), 2024. Published by Cambridge University Press

Introduction

Cognitive deficits represent a significant and sustained challenge for over 50% of people with bipolar disorder (BD). Since these cognitive deficits contribute significantly to the psychosocial burden of BD (Van Rheenen et al., Reference Van Rheenen, Lewandowski, Bauer, Kapczinski, Miskowiak, Burdick and Balanzá-Martínez2020), identifying the factors they are associated with, and potentially driven by, is crucial. Physical health is one potential factor, given that people with BD are disproportionately affected by cardiometabolic diseases (e.g. heart disease and diabetes) and experience higher rates of conventional risk factors for these diseases compared to the general population (Coello et al., Reference Coello, Kjærstad, Stanislaus, Melbye, Faurholt-Jepsen, Miskowiak and Munkholm2019; Vancampfort et al., Reference Vancampfort, Ph, Vansteelandt, Ph, Correll, Mitchell and Ph2013; Weiner, Warren, & Fiedorowicz, Reference Weiner, Warren and Fiedorowicz2011). These risk factors include metabolic abnormalities such as obesity, dyslipidemia, hypertension, and the metabolic syndrome (Vancampfort et al., Reference Vancampfort, Ph, Vansteelandt, Ph, Correll, Mitchell and Ph2013), as well as the lifestyle/behavioral health risk factors that might drive such metabolic abnormalities, such as low physical activity, sedentary behavior, poor diet, and smoking (Vancampfort et al., Reference Vancampfort, Firth, Schuch, Rosenbaum, Mugisha, Hallgren and Stubbs2017).

General population studies have shown age-dependent associations of cardiometabolic diseases and their risk factors with both cognitive impairment and incident dementia (Qiu & Fratiglioni, Reference Qiu and Fratiglioni2015). Indeed, hemodynamic and serum markers of cardiac function have been associated with structural brain changes and accelerated cognitive decline and dementia (Hosoki et al., Reference Hosoki, Hansra, Jayasena, Poljak, Mather, Catts and Sachdev2023; Jensen, Zeller, Twerenbold, & Thomalla, Reference Jensen, Zeller, Twerenbold and Thomalla2023; van der Velpen, Feleus, Bertens, & Sabayan, Reference van der Velpen, Feleus, Bertens and Sabayan2017). Cognitive impairments associated with these markers, as well as with the presence of cardiometabolic diseases and their risk factors, are evident across a number of cognitive domains including executive function, processing speed, memory, attention, fluency, and global cognition (Qiu & Fratiglioni, Reference Qiu and Fratiglioni2015; Tahmi, Palta, & Luchsinger, Reference Tahmi, Palta and Luchsinger2021; van der Velpen et al., Reference van der Velpen, Feleus, Bertens and Sabayan2017; Waldstein & Wendell, Reference Waldstein and Wendell2010; Zonneveld et al., Reference Zonneveld, Abbel, le Cessie, Jukema, Noordam and Trompet2023). As such, cardiometabolic health appears to be broadly relevant to cognitive function.

Poor cardiometabolic health has been linked to an acceleration in the biological aging process in persons with BD (Rizzo et al., Reference Rizzo, Costa, Mansur, Swardfager, Belangero, Grassi-Oliveira and Brietzke2014), with cardiometabolic diseases manifesting during early mid-life and up to 17 years earlier than in the general population (Goldstein, Schaffer, Wang, & Blanco, Reference Goldstein, Schaffer, Wang and Blanco2015). In line with this, we recently reported that cardiometabolic disease (type 2 diabetes) in BD during mid to later-life is associated with more severe processing speed and memory deficits and putatively with premature age-related cognitive decline (Ringin et al., Reference Ringin, Dunstan, McIntyre, Berk, Owen, Rossell and Van Rheenen2022). Another recent large-scale study also found more pronounced adverse age-related changes in common risk factors for cardiometabolic disease, including blood pressure, pulse rate, body composition, and hand-grip strength, in BD patients compared to controls (Mutz, Young, & Lewis, Reference Mutz, Young and Lewis2022). However, in general, the association of cognition with these risk factors has received relatively sparse empirical attention in BD research to date. Indeed, a systematic review published in 2019 showed that only 11 neurocognitive studies of BD had examined physiological correlates of cardiometabolic disease, such as obesity, hypertension, dyslipidemia, diabetes mellitus, and the metabolic syndrome. The review reported that a heightened physiological risk for cardiometabolic disease was associated with more severe cognitive abnormalities in BD, namely in domains of executive function, processing speed, attention, reasoning, and global cognition, although the sample sizes of included studies were mostly modest (Bora, McIntyre, & Ozerdem, Reference Bora, McIntyre and Ozerdem2019).

The association of cognition and lifestyle/health behaviors relevant to cardiometabolic disease, henceforth termed health risk behaviors, has received even less attention in BD (Van Rheenen et al., Reference Van Rheenen, Lewandowski, Bauer, Kapczinski, Miskowiak, Burdick and Balanzá-Martínez2020; Van Rheenen & Neil, Reference Van Rheenen and Neil2022). However, early studies on the topic generally indicate more detrimental cardiometabolic-disease-relevant health risk behaviors in those with cognitive impairment (Aas et al., Reference Aas, Djurovic, Ueland, Mørch, Fjæra, Reponen and Melle2019; Balanzá-Martínez, Crespo-Facorro, González-Pinto, & Vieta, Reference Balanzá-Martínez, Crespo-Facorro, González-Pinto and Vieta2015; Bradley, Anderson, Gallagher, & McAllister-Williams, Reference Bradley, Anderson, Gallagher and McAllister-Williams2020; Burgess, Bradley, Anderson, Gallagher, & McAllister-Williams, Reference Burgess, Bradley, Anderson, Gallagher and McAllister-Williams2022; Cardoso et al., Reference Cardoso, Bauer, Jansen, Suchting, Zunta-Soares, Quevedo and Soares2016; Fellendorf et al., Reference Fellendorf, Kainzbauer, Platzer, Dalkner, Bengesser, Birner and Reininghaus2017; Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023). Indeed, preliminary work on physical activity in BD suggests a positive association with executive function, attention and memory (Aas et al., Reference Aas, Djurovic, Ueland, Mørch, Fjæra, Reponen and Melle2019; Fellendorf et al., Reference Fellendorf, Kainzbauer, Platzer, Dalkner, Bengesser, Birner and Reininghaus2017), although our group recently observed worse global cognition in BD patients and controls with high self-reported physical activity levels (Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023). Other research indicates that cognition (specifically executive function, memory, attention, processing speed, and global cognition) is worse in BD patients with comorbid alcohol use disorders, sleep abnormalities, and for those who engage in more mentally passive sedentary behaviors (e.g. watching TV) (Balanzá-Martínez et al., Reference Balanzá-Martínez, Crespo-Facorro, González-Pinto and Vieta2015; Bradley et al., Reference Bradley, Anderson, Gallagher and McAllister-Williams2020; Burgess et al., Reference Burgess, Bradley, Anderson, Gallagher and McAllister-Williams2022; Cardoso et al., Reference Cardoso, Bauer, Jansen, Suchting, Zunta-Soares, Quevedo and Soares2016; Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023). In contrast, global cognitive performance appears to be better in people with BD engaging in more mentally active sedentary behaviors (e.g. reading, using the computer), with mentally active sedentary behaviors appearing to protect against age-related cognitive decline (Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023).

While these preliminary studies indicate the importance of cardiometabolic disease risk factors in the context of cognitive impairment in BD, the extent to which health-risk behaviors compared to physiological risk factors, are more, less, or equally relevant to cognition, and whether this differs by age, remains unknown. Thus, it is not clear which aspects of cardiometabolic disease risk could most usefully be addressed in cognitive BD research or in preventative cognitive interventions, nor whether different aspects of risk should be targeted at different stages of life. Risk assessment for cognitive impairment/decline in BD in clinical settings is also hampered by this absence of knowledge regarding the relative importance of different cardiometabolic disease risk factors at different life stages, particularly in the context of primary care.

Here, we: (i) identify which cardiometabolic disease risk factors are the more important correlates of cognitive impairment in BD, utilizing a large cross-sectional dataset to study a range of these factors simultaneously; and (ii) determine whether the nature of such relationships varies between mid and later life. In doing so, we aim to inform the strategic direction of resources in cognitive BD research, as well as the development of cognition-related risk assessment tools for use in clinical settings.

Materials and methods

Participants were drawn from the UK Biobank, a prospective dataset of 502 649 individuals aged 40–69. Baseline assessments were completed across 22 centers throughout the UK between 2005 and 2010, providing a range of lifestyle, health, demographic, cognitive, and biological data. Full details of the data collection procedures are provided elsewhere (UK Biobank: Protocol for a large-scale prospective epidemiological resource, 2007). All participants provided written informed consent. The UK Biobank has approval from the Northwest Multi-Centre Research Ethics Committee (reference 16/NW/0274 and 11/NW/0382).

BD diagnostic criteria

UK Biobank participants categorized as having BD were included in this analysis. No healthy controls were included. A detailed description of the methodology for categorizing participants as having BD has been provided previously (Smith et al., Reference Smith, Nicholl, Cullen, Martin, Ul-Haq, Evans and Pell2013). In brief, a touchscreen questionnaire, based on symptoms within the Structured Clinical Interview for DSM-IV axis I disorders (introduced in the final two years of recruitment), was utilized to identify participants with probable BD, major depressive disorder, or no indicated mental disorder. These classifications were validated against demographic and clinical information available in the dataset. In our analyses, participants who were categorized as having major depressive disorder or no indicated mental disorder were excluded from the primary analyses. We also excluded those who were pregnant, as well as those with neurological conditions known to affect cognitive functioning (see online Supplementary Material for details).

Cardiometabolic disease risk factors

We classified 16 individual cardiometabolic disease risk factors obtained at the baseline assessment into two key domains of interest, namely (1) health risk behaviors; including time spent in physical activity and mentally active and passive sedentary behavior, smoking status, and sleep duration, and (2) physiological risk factors; further broken down by (2a) anthropometric and clinical risk including systolic and diastolic resting blood pressure, waist circumference, handgrip strength, fat mass index, and fat free mass index, and (2b) cardiometabolic disease risk biomarkers including C-reactive protein (CRP), hemoglobin A1c (HbA1c), LDL cholesterol, HDL-L cholesterol, and triglycerides. We also considered the following variables as covariates; sex, socio-economic status (SES) – as measured by the Townsend Deprivation Index, educational level, and psychotropic medication use, as well as hypertension and cholesterol lowering medication use and BD subtype (BD I v. BD II). Brief descriptions of the health risk behaviors and physiological risk factors can be found below, with more extensive information provided in the online Supplementary Material. It should be noted that some of the risk factors of interest have previously been studied in relation to cognition in BD using UK Biobank participants. Further details of these studies can be found elsewhere (Firth et al., Reference Firth, Firth, Stubbs, Vancampfort, Schuch, Hallgren and Sarris2018a; Milton et al., Reference Milton, Ward, Ward, Lyall, Strawbridge, Smith and Cullen2021; Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023).

Health risk behaviors

Participants were categorized into low-to-moderate physical activity (=0) and high physical activity (=1) groups based on the self-report International Physical Activity Questionnaire data processing guidelines (International Physical Activity Questionnaire, 2005). Physical activity was categorized in this way given the previous observation in this sample that cognitive functioning did not significantly differ between the low and moderate IPAQ groups (Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023). Detailed information on this categorization is provided in the online Supplementary Material. Mentally passive (TV viewing) and mentally active (non-occupational computer use) sedentary behavior were self-reported as average hours per day of each respective behavior. Participants who reported greater than 16 h per day (indicating potentially implausible levels) of total sedentary behavior (TV viewing, computer use, and driving) or physical activity were excluded. Smoking status was defined based on participants reporting being either current smokers (=1) or non-smokers (=0). Non-smokers included participants with a history of smoking who did not currently smoke. Sleep duration was self-reported and dichotomized into inadequate (<6 and >9 h) (=1) and adequate (6–9 h) (=0) average nightly sleep duration. Sleep duration was analyzed categorically given consistent evidence of a u-shaped relationship with cognitive function (Wennberg, Wu, Rosenberg, & Spira, Reference Wennberg, Wu, Rosenberg and Spira2017; Yaffe, Falvey, & Hoang, Reference Yaffe, Falvey and Hoang2014).

Physiological risk factors

Anthropometric and clinical risk factors

Seated systolic and diastolic resting blood pressure (mmHg) were measured twice using an Omron 705 IT digital monitor, and the average of the two measurements was calculated. Hand-grip strength (a measure of muscular strength and physical fitness) was measured once on each hand using a Jamar J00105 hydraulic hand dynamometer, in line with standard procedures (Roberts et al., Reference Roberts, Denison, Martin, Patel, Syddall, Cooper and Sayer2011). The score from the participant's self-reported dominant hand was used for the analyses. When handedness was not reported, the highest scoring hand was used. Body composition measures were obtained with a Tanita BC-418 MA body composition analyzer. Fat Mass Index (FMI) and Fat Free Mass Index (FFMI) were calculated by dividing whole-body fat mass, and whole-body fat free mass, by height in meters squared. Waist circumference was measured by trained nurses.

Cardiometabolic disease risk biomarkers

CRP, HbA1c, LDL cholesterol, HDL cholesterol, and triglycerides were measured in blood samples collected at baseline. Details on serum sample handling and protocol in the UK Biobank have been described previously (UK Biobank, 2019). Serum CRP (mg/l) and the following lipids were measured by immunoturbidimetric analysis on a Beckman automated hematology analyzer; LDL cholesterol (mmol/l), HDL cholesterol (mmol/l), and triglycerides (mmol/l). HbA1c was measured by high performance liquid chromatography analysis on a Bio-Rad VARIANT II Turbo. CRP was dichotomized into normal (<5 mg/l) (=0) and elevated (⩾5 mg/l) (=1) levels as per the standard reference range provided by the Royal College of Pathologists in Australasia (The Royal College of Pathologists of Australasia, 2019). HbA1c was dichotomized into normal (<39 mmol/mol) and elevated (⩾39 mmol/mol) (The elevated category includes pathological HbA1c levels (⩾48mmol/mol) in line with references ranges from the American Diabetes Association and the International Diabetes Federation (International Diabetes Federation, 2021). Lipids were dichotomized into normal (=0) and abnormal (=1) levels as per standard reference ranges provided by the Australian Institute for Health and Welfare (Australian Institute of Health & Welfare, 2017). Abnormal levels were considered as follows: LDL cholesterol ⩾ 3.5 mmol/l (Given inconsistency in suggested normal/abnormal thresholds, analyses were re-run with a threshold of 2 mmol/l. Results were unchanged, and thus are not presented for brevity).; HDL cholesterol <1.00 mmol/l for men, <1.3 mmol/l for women; triglycerides ⩾ 2 mmol/l.

Cognitive assessment

Cognitive functioning was assessed through a brief computerized battery obtained at the same time as the physical activity and sedentary behavior data collection. The battery, which took approximately 15 min to complete, was developed specifically for the UK Biobank and was designed to be completed electronically without examiner supervision. Assessments were completed at the UK Biobank assessment centers and included measurement of the following cognitive domains (A fifth cognitive domain, numeric memory, was tested at baseline. A recent publication has queried whether the numeric memory test designed for the UK Biobank accurately tests its intended cognitive domain, working memory (Fawns-Ritchie & Deary, Reference Fawns-Ritchie and Deary2020). Further, the test was removed during the early stages of testing due to time constraints, subsequently resulting in a very low number of participants with available data. For these reasons, we have decided not to utilize this component in the current study).; visuospatial memory (pairs matching), processing speed (reaction time), fluid intelligence (reasoning test), and prospective memory (prospective memory test). Detailed information regarding each cognitive test can be found in the online Supplementary Material, but it should be noted that the tests were not validated in BD samples prior to their inclusion in the UK Biobank protocol. In the current study, scores for all tests were coded so that higher scores equated to better performance. A global cognitive score was then created by; (1) calculating z scores for the continuous measures (visuospatial memory, processing speed, and fluid intelligence) based on means and standard deviations in the full sample, and (2) summing these z-scores with the raw prospective memory score (dichotomous variable equaling 0 or 1), as has been done previously (Anatürk, Suri, Smith, Ebmeier, & Sexton, Reference Anatürk, Suri, Smith, Ebmeier and Sexton2021). A global cognitive score was used rather than individual domain scores, as aggregate scores are known to have greater validity in sampling the construct of interest and are more reliable (Harvey, Reference Harvey2019). A higher global cognitive score equates to better cognitive function.

Statistical analysis

All analyses were completed using the Statistical Package for the Social Sciences (SPSS) Version 29 (IBM). All variables were visually checked for extreme outliers and relevant statistical test assumptions were checked using the appropriate methods and residual plots. Preliminary analyses were run to determine whether the covariates of interest (sex, SES, educational level, BD subtype, mood stabilizer use, antidepressant use, first-generation antipsychotic use, second-generation antipsychotic use, sedatives/hypnotics use, cholesterol-lowering medication use, hypertension medication use, and diabetes medication use) were associated with cognitive function. Psychotropic medication use (all types), hypertensive medication use, diabetes medication use, and BD subtype were not significantly associated with global cognition (See online Supplementary Table S1 for full results). and were thus not included in the analyses specified below.

To identify individual cardiometabolic disease risk factors associated with global cognitive function in BD, separate multivariable linear regression models were fitted for each risk factor within each risk factor domain (1, health-risk behaviors; 2a, physiological risk – anthropometric and clinical risk factors; and 2b, physiological risk-cardiometabolic disease risk biomarkers; domain-based models, n = 3). Global cognition was specified as the outcome, while sex, education, SES, and cholesterol-lowering medication use (Given its association with measures of cholesterol (LDL-cholesterol, HDL-cholesterol, triglycerides), cholesterol-lowering medication use was added as a covariate in the analysis of cardiometabolic disease biomarkers). were added in block 1 of each respective model, and the risk factors of interest (either 1, health-risk behaviors; 2a, physiological risk - anthropometric and clinical risk factors; or 2b, physiological – cardiometabolic disease risk biomarkers) added in block 2. The cardiometabolic risk biomarkers model was re-run with participants with comorbid diabetes removed (n = 31), to ensure that any findings related to HbA1c reflected its contribution as a risk factor for cardiometabolic outcomes, given that HbA1c is a biological marker directly indicative of diabetes status. Findings did not change and thus are not presented for brevity. A final combined multivariable regression model (combined model, n = 1) for global cognition was then fitted, including the aforementioned covariates in block 1 and only the predictor variables that were significantly associated with cognition in the previous models in block 2.

In cases in which specific risk factors showed associations with cognition in domain-specific models but not in the combined model, post-hoc mediation models were conducted to ascertain the extent to which the association of these specific risk factors with cognition was being mediated by the other risk factors that were significant in the combined model. This was done using model 4 of the Preacher and Hayes PROCESS plugin for SPSS (v4.0). Coefficients with 95% bias-corrected bootstrapped confidence intervals (CIs) were calculated for the indirect path (5000 bootstrap samples used), and mediation was considered significant if the range of the CI did not span zero.

To explore whether there were life stage differences in the association of specific cardiometabolic disease risk factors at mid or later-life, we stratified participants by age and re-ran the final model in the stratified groups (age-stratified models, n = 2). Participants aged 40–59 years were grouped into the mid-life category and participants aged 60–69 years into the later-life category, based on (a) existing approximations of the mid-life stage (Infurna, Gerstorf, & Lachman, Reference Infurna, Gerstorf and Lachman2020) and (b) the maximum age of participants in the sample. All other variables in the model remained the same. A false discovery rate of p < 0.05 was applied to all results to account for multiple comparisons using the Benjamini-Hochberg method.

Results

Participants included in the analyses

Demographic characteristics of the participants included in the analyses are displayed in Table 1. The final analysis included 996 participants, all of whom met UK Biobank criteria for BD. The mean age of the sample was 54.29 (s.d. = 8.01), and 45.7% were female.

Table 1.

Characteristics of participants (mean ± standard errors)

a HbA1c data were missing for n = 101 participants and cholesterol medication data were missing for n = 8 participants.

Domain-based models

Results of the three domain-based models are reported in online Supplementary Tables S2–S4. From the health-risk behavior domain, global cognitive performance was positively associated with mentally active sedentary behavior, negatively associated with passive sedentary behavior, physical activity, and sleep duration, but not associated with smoking status. From the anthropometric and clinical subdomain of physiological risk factors, global cognitive performance was positively associated with diastolic blood pressure and hand-grip strength, and negatively associated with systolic blood pressure and waist circumference. From the cardiometabolic disease risk biomarkers subdomain of physiological risk factors, it was negatively associated with CRP, but not with HbA1c, HDL cholesterol, non-HDL cholesterol, or triglycerides.

Combined model

Results of the combined model are reported in Table 2. The full model explained 14.7% of the total variance of global cognition, and the included cardiometabolic disease risk factors explained 9.5% of the total variation after controlling for sex, SES, and educational level. In this model, global cognition was associated with the following variables (when other variables were controlled): (i) mentally active and mentally passive sedentary behavior; cognition was lower by 0.11 points on average per one-hour decrease of mentally active sedentary behavior (CI 0.05–0.18) and by 0.07 points on average for every 1-h increase in mentally passive sedentary behavior (CI: −0.13 to −0.01), (ii) physical activity and sleep duration; cognition was lower by 0.40 points on average in the high compared to low-to-moderate physical activity group (CI −0.62 to −0.17), and by 0.48 points on average in the group that slept <6 or >9 h per night v. those that slept 6–9 h (CI −0.82 to −0.15); (iii) handgrip strength; cognition was lower by 0.04 on average points for every one-kg hand-grip strength decrease (CI 0.03–0.05); and (iv) blood pressure; cognition was lower by 0.02 on average points for every one mmHg increase in systolic blood pressure (CI −0.03 to −0.01) and by 0.03 on average points for every one mmHg decrease in diastolic blood pressure (CI 0.02–0.05).

Table 2.

Combined model of cardiometabolic disease risk factors and global cognition (all age groups)

A * indicates significance at p < 0.05 before Benjamini–Hochberg FDR correction for multiple comparisons, and bolded values indicate significance after Benjamini–Hochberg FDR correction for multiple comparisons. Sex, male = 0, female = 1; educational level, did not attend university = 0, attended university = 1; physical activity, low-moderate activity = 0, high activity = 1; sleep duration, <6 or >9 = 0, 6–9 = 1; CRP normal = 0, elevated = 1.

a Unstandardized regression coefficient.

b Standardized regression coefficient.

c 95% confidence interval lower limit.

d 95% confidence interval upper limit.

Post-hoc analyses

Since waist circumference and CRP were significantly associated with cognition in the domain-based model only, post-hoc mediation models (n = 2) were run to explore the extent to which associations of these risk factors with cognition was being mediated by the other risk factors that were significant in the combined model. Continuous measures of physical activity and inadequate sleep were included as mediators in this model, as PROCESS does not allow the use of categorical mediators. More details on these variables are provided in the online Supplementary Material. CRP was significantly and negatively associated with global cognition, and this association was fully mediated by handgrip strength (CI −0.16 to −0.03), inadequate sleep (CI −0.11 to −0.009) mentally passive sedentary behavior (CI −0.09 to −0.0002), systolic blood pressure (CI −0.19 to −0.02) and diastolic blood pressure (CI 0.02–0.17) (Fig. 1). There was no direct or total effect of waist circumference with global cognition (online Supplementary Fig. S1).

Figure 1.

Effects (bootstrapped standard error in parenthesis) for mediation examining how elevated CRP was associated with global cognition after controlling for sex, educational level, and SES. *p < 0.05. Bolded (red) lines indicate significant mediation pathways (range of CI did not span 0).

Age-stratified models

Midlife

Results of the midlife model are reported in Table 3. In participants below 60 years of age, cardiometabolic disease risk factors explained 6.9% of the total variance in global cognition after controlling for sex, SES, and educational level. For these participants, global cognition was lower in those with high physical activity (by 0.38 points on average v. those with low-to-moderate physical activity, CI −0.65 to −0.11), those with a nightly sleep duration of <6 or >9 h (by 0.56 points on average v. those sleeping 6–9 h per night, CI −0.96 to −0.16), and those with lower hand-grip strength (by 0.04 points on average per one kg decrease, CI 0.02–0.05).

Table 3.

Age-stratified model of cardiometabolic disease risk factors and global cognition; midlife patients

A * indicates significance at p < 0.05 before Benjamini–Hochberg FDR correction for multiple comparisons, and bolded values indicate significance after Benjamini–Hochberg FDR correction for multiple comparisons. Sex, male = 0, female = 1; educational level, did not attend university = 0, attended university = 1; physical activity, low-moderate activity = 0, high activity = 1; sleep duration, <6 or >9 = 0, 6–9 = 1; CRP normal = 0, elevated = 1.

a Unstandardized regression coefficient.

b Standardized regression coefficient.

c 95% confidence interval lower limit.

d 95% confidence interval upper limit.

Later-life

Results of the later-life model are reported in Table 4. In participants 60 years of age or above, cardiometabolic disease risk factors of interest explained 10.9% of the total variance in global cognition after controlling for sex, SES, and educational level. For these older participants, global cognition was lower in those with high physical activity (by 0.51 points on average v. those with low-to-moderate physical activity, CI: −0.93 to −0.09), those with lower muscular strength (by 0.04 per one kg decrease, CI 0.01–0.06), those with less mentally active sedentary behavior (by 0.17 points on average per one hour decrease, CI 0.04–0.30) and those with low diastolic and high systolic blood pressure (by 0.04 on average for every one mmHg decrease, CI 0.01–0.07 and by 0.02 on average for every one mmHg increase, CI −0.04 to −0.004, respectively).

Table 4.

Age-stratified model of cardiometabolic disease risk factors and global cognition; later-life patients

An * indicates significance at p < 0.05 before Benjamini–Hochberg FDR correction for multiple comparisons, and bolded values indicate significance after Benjamini–Hochberg FDR correction for multiple comparisons. Sex, male = 0, female = 1; educational level, did not attend university = 0, attended university = 1; physical activity, low-moderate activity = 0, high activity = 1; sleep duration, <6 or >9 = 0, 6–9 = 1; CRP normal = 0, elevated = 1.

a Unstandardized regression coefficient.

b Standardized regression coefficient.

c 95% confidence interval lower limit.

d 95% confidence interval upper limit.

Discussion

Our study leveraged a large publicly available dataset to investigate correlates of global cognitive functioning in the context of cardiometabolic disease risk during mid and later life in BD. We specifically focused on understanding the relative importance of the health-risk behaviors and physiological risk factor domains, the latter including anthropometric, clinical, and cardiometabolic disease risk biomarkers. We found worse global cognition scores (encompassing an aggregate of visuospatial and prospective memory, processing speed, and fluid intelligence domain scores) albeit with small effect sizes, in BD patients with lower levels of mentally active sedentary behavior, hand-grip strength and diastolic blood pressure, and in those with higher levels of mentally passive sedentary behavior, systolic blood pressure, physical activity, and an average sleep duration of <6 or over 9 h per night.

These findings are broadly concordant with the literature on cognitive risk factors in the context of dementia (Baumgart et al., Reference Baumgart, Snyder, Carrillo, Fazio, Kim and Johns2015; Dintica & Yaffe, Reference Dintica and Yaffe2022). They also add to preliminary evidence linking health risk behaviors with executive, memory, processing speed, attentional, and global cognitive impairments in BD (Aas et al., Reference Aas, Djurovic, Ueland, Mørch, Fjæra, Reponen and Melle2019; Balanzá-Martínez et al., Reference Balanzá-Martínez, Crespo-Facorro, González-Pinto and Vieta2015; Bradley et al., Reference Bradley, Anderson, Gallagher and McAllister-Williams2020; Burgess et al., Reference Burgess, Bradley, Anderson, Gallagher and McAllister-Williams2022; Cardoso et al., Reference Cardoso, Bauer, Jansen, Suchting, Zunta-Soares, Quevedo and Soares2016; Fellendorf et al., Reference Fellendorf, Kainzbauer, Platzer, Dalkner, Bengesser, Birner and Reininghaus2017; Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023), as well as to a considerable literature indicating blood pressure as a key marker of cognitive health (in domains of executive function, processing speed, attention, reasoning, and global cognition) both in BD and in the general population (Bora et al., Reference Bora, McIntyre and Ozerdem2019; Ou et al., Reference Ou, Tan, Shen, Xu, Hou, Dong and Yu2020).

Further, hand-grip strength, a marker of muscular strength and physical fitness, has also been recently linked to memory, processing speed, reasoning, and global cognition in BD cohorts and elsewhere, as well as cognitive decline and dementia more generally (Aliño-Dies et al., Reference Aliño-Dies, Sánchez-Ortí, Correa-Ghisays, Balanzá-Martínez, Vila-Francés, Selva-Vera and Tabarés-Seisdedos2020; Cui, Zhang, Liu, Gang, & Wang, Reference Cui, Zhang, Liu, Gang and Wang2021; Firth et al., Reference Firth, Firth, Stubbs, Vancampfort, Schuch, Hallgren and Sarris2018a, Reference Firth, Stubbs, Vancampfort, Firth, Large, Rosenbaum and Yung2018b). Here, and in our recent UK Biobank study we also observed that physical activity, a correlate of physical fitness, was also associated with global cognition in BD patients and controls (Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023), although this association was inverse in direction. While this unexpected effect may be explained by the theorized negative effect on cognition of sustained activity (i.e. occupational activity, which may reflect more-manual job types) (see page 6 of Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023 for further explanation), perhaps more weighting should be given to the handgrip strength findings in terms of validity. This is because handgrip strength is a readily repeatable physical fitness measure that likely reflects longer-term physical activity, whereas in this study physical activity itself was measured with a subjective measure that relies on accurate recall over a 7-day period. Notably, handgrip strength (standardized β = 0.23) had the greatest association with global cognition of all continuously-measured variables, regardless of age, followed by systolic (standardized β = −0.19) and diastolic (standardized β = 0.17) blood pressure, mentally active sedentary behavior (standardized β = 0.10), and mentally passive sedentary behavior (standardized β = −0.07). Average sleep duration was also more strongly associated with cognitive function than physical activity levels.

Associations of poorer global cognition with high physical activity and low hand-grip strength were evident in BD patients in our sample irrespective of age. However, global cognition was associated with an average nightly sleep duration of <6 or >9 h (compared to 6–9 h) in only midlife patients, and with higher systolic blood pressure, lower diastolic blood pressure, and less mentally active sedentary behavior in only later-life patients. The blood pressure associations in later-life are consistent with known increases in systolic and decrease in diastolic blood pressure with age, which in turn increase pulse pressure (Wells & Townsend, Reference Wells and Townsend2019) and predict cognitive impairment in middle aged and older people (Sha, Cheng, & Yan, Reference Sha, Cheng and Yan2018). Indeed, high systolic blood pressure and low diastolic blood pressure are recognized as risk factors for dementia in late-life (>65 years) (Forte, De Pascalis, Favieri, & Casagrande, Reference Forte, De Pascalis, Favieri and Casagrande2020; Ou et al., Reference Ou, Tan, Shen, Xu, Hou, Dong and Yu2020).

The inverse association of mentally active sedentary behavior, a type of intellectual stimulation, and cognition in later-life participants is also in line with our earlier UK Biobank study in which we found that BD and control participants with less mentally active sedentary behavior were less protected against putative cognitive decline (Ringin et al., Reference Ringin, Dunstan, McIntyre, Owen, Berk, Rossell and Van Rheenen2023). Regarding sleep, preliminary BD research has linked sleep abnormalities with attention and processing speed impairments, although none have examined this association as a function of age (Bradley et al., Reference Bradley, Anderson, Gallagher and McAllister-Williams2020; Burgess et al., Reference Burgess, Bradley, Anderson, Gallagher and McAllister-Williams2022; Kanady, Soehner, Klein, & Harvey, Reference Kanady, Soehner, Klein and Harvey2017; Laskemoen et al., Reference Laskemoen, Büchmann, Barrett, Collier-Høegh, Haatveit, Vedal and Simonsen2020; Menkes et al., Reference Menkes, Andrews, Burgess, Carley, Marshall, Langenecker and Ryan2021; Russo et al., Reference Russo, Mahon, Shanahan, Ramjas, Solon, Purcell and Burdick2015). Nonetheless, meta-analytic data from the general population has shown a negative association of short and long sleep duration with multi-domain cognitive performance, executive functions, verbal, and working memory during later-life (Lo, Groeger, Cheng, Dijk, & Chee, Reference Lo, Groeger, Cheng, Dijk and Chee2016). Thus, it is possible that the relatively low number of late-life BD participants reporting a sleep duration outside of the range of hours recognized as reflecting good sleep (6–9 h) may explain the association between sleep duration and cognition in only midlife participants.

Somewhat surprisingly, we found no associations between global cognition and most of the cardiometabolic disease risk biomarkers or body composition measures of interest. This contrasts with findings from extant research. For example, two previous, albeit small, BD studies, linked elevated triglycerides to poor executive function (Naiberg et al., Reference Naiberg, Newton, Collins, Dickstein, Bowie and Goldstein2016; Van Rheenen, McIntyre, Balanzá-Martínez, Berk, & Rossell, Reference Van Rheenen, McIntyre, Balanzá-Martínez, Berk and Rossell2021a). Another meta-analysis of BD also recently linked obesity to executive function and processing speed (Bora et al., Reference Bora, McIntyre and Ozerdem2019). However, there was no association between cognition and memory (verbal, visual, or working) or attention. This may help to explain the absence of adiposity-cognition associations in the current work, given two memory tests were included in the global cognitive score and potentially weighted it more heavily toward a cognitive domain that may not be relevant to measures of body composition. Recent work from our group found domain-specific associations of type 2 diabetes with processing speed and visuospatial memory in a similar UKB sample (Ringin et al., Reference Ringin, Dunstan, McIntyre, Berk, Owen, Rossell and Van Rheenen2022), and as such, the lack of an association with elevated, but not necessarily pathological HbA1c here suggests that cognition in BD is more likely to be affected at pathological HbA1c levels. In support of this, two studies BD studies found no associations between glucose levels and cognition are primarily normoglycemic samples (Hubenak, Tuma, & Bazant, Reference Hubenak, Tuma and Bazant2015; Naiberg et al., Reference Naiberg, Newton, Collins, Dickstein, Bowie and Goldstein2016), whereas diagnosed diabetes has been linked with cognition in another BD cohort (Tsai, Lee, Chen, & Huang, Reference Tsai, Lee, Chen and Huang2007).

Only elevated CRP was associated with worse global cognitive function in the domain-model, and in subsequent analyses it was revealed that handgrip strength, mentally passive sedentary behavior, inadequate sleep, systolic blood pressure, and diastolic blood pressure were mediators of its association with global cognition (Fig. 1). All of these risk factors have been previously associated with CRP (Chuang et al., Reference Chuang, Hsu, Chang, Bai, Yeh and Pan2013; Irwin, Olmstead, & Carroll, Reference Irwin, Olmstead and Carroll2016; Lakoski et al., Reference Lakoski, Cushman, Palmas, Blumenthal, D'Agostino and Herrington2005; Tuttle, Thang, & Maier, Reference Tuttle, Thang and Maier2020; Wirth et al., Reference Wirth, Klenk, Brefka, Dallmeier, Faehling, Roqué i Figuls and Stubbs2017). Relevantly, although immune dysregulation and inflammation, as indexed by elevated CRP, has been associated with worse cognitive performance (namely in executive function, processing speed, attention, verbal fluency and memory domains) in several BD studies (Congio, Urbano, Soares, & Nunes, Reference Congio, Urbano, Soares and Nunes2022; Dickerson et al., Reference Dickerson, Stallings, Origoni, Vaughan, Khushalani and Yolken2013; Millett et al., Reference Millett, Perez-Rodriguez, Shanahan, Larsen, Yamamoto, Bukowski and Burdick2019; Milton et al., Reference Milton, Ward, Ward, Lyall, Strawbridge, Smith and Cullen2021), a recent meta-analysis of severe mental illness showed only weak associations between inflammation and global cognition (Morrens et al., Reference Morrens, Overloop, Coppens, Loots, Van Den Noortgate, Vandenameele and De Picker2022). In light of our findings, it is possible that this weak association may be explained as a function of CRP acting indirectly on cognition via a range of other risk factors from the health risk behavior and anthropometric risk factor domains that are not typically considered in inflammation-cognition studies.

Indeed, although previous BD cognition research has focused on some of the variables we included in our analysis independently, to our knowledge, this is the first BD study to have examined associations of cognitive function with a broad range of health behavior-related and physiological cardiometabolic disease risk factors concurrently. Although this was a strength of the study, other limitations should be considered when interpreting our findings. First, several important cardiometabolic disease risk factors were not measured, or did not have detailed measures, in the UK Biobank dataset (i.e. diet, alcohol use, other inflammatory markers), and as such were not included in our analyses. Moreover, detailed data describing the clinical characteristics of the sample that may impact cardiometabolic outcomes, such as illness duration or episode history, was not available. Second, physical activity, sedentary behavior, and sleep duration were self-reported by participants, and thus may be subject to measurement error. Third, the study was cross-sectional which precludes inferences regarding causality. Fourth, the cognitive data did not come from validated cognitive assessments, and only a few domains were measured. Although there is evidence to show that the available cognitive tests are valid measures of general cognitive functioning (Fawns-Ritchie & Deary, Reference Fawns-Ritchie and Deary2020), they have not been validated in BD and may not be particularly sensitive to the deficits common to this population. To offset this, we used an aggregate of the individual cognitive test scores to improve validity and reliability in the measurement of cognition, meaning that domain-specific associations may have been missed. Nonetheless, readers are also cautioned to keep in mind that the global cognition score encompassed only four underpinning domains (visuospatial and prospective memory, processing speed, and fluid intelligence), and this score was also not an ideal global cognitive measure. The UK Biobank sample was limited to participants aged between 40 and 69 years of age living in the UK, and as such it is unclear whether these associations would be evident in younger or older cohorts, or in wider geographical regions. Finally, participants with mental health disorders in the UK Biobank have been shown to be generally higher functioning than those with disorders in the general population (Kendall et al., Reference Kendall, Rees, Escott-Price, Einon, Thomas, Hewitt and Kirov2017), which limits generalizability and suggests a potential underestimation of our findings.

Overall, our findings appear to suggest that behavioral cardiometabolic disease risk is particularly associated global cognitive function in BD, potentially more-so than physiological cardiometabolic disease risk factors given the predominant associations of health risk behaviors and cognition over and above that of physiological risk factors. Considering this, our findings provide preliminary evidence that health behavior-related cardiometabolic disease risk may be considered a marker of cognitive function in BD. This has implications for patient phenotyping and clinical care given that health-behaviors can be readily assessed and thus measured in multiple clinical settings (e.g. general practice, psychology) without the need for specialized equipment or knowledge. Given the clear importance of cognitive dysfunction in the clinical management of BD (Van Rheenen, Miskowiak, & Burdick, Reference Van Rheenen, Miskowiak and Burdick2021b), health behavior risk presence as a proxy for poor cognition in BD may be particularly helpful in clinical practice to facilitate; (a) identification of those in need of more complex cognitive assessment, (b) identification of which patients may benefit from additional support, and (c) insight into patient behaviors, thus fostering empathy and understanding in the treating clinician which may lead to improved clinical outcomes. The overlap between measures of cardiometabolic risk and cognitive risk supports the notion that the drivers of neuroprogression, the processes driving progression of psychiatric disorders overlap substantially with those for somatoprogression, the drivers of progression of physical health disorders (Morris et al., Reference Morris, Puri, Walker, Maes, Carvalho, Bortolasci and Berk2019). This suggests the necessity for common approaches to prevention and management of these shared pathways for diverse non communicable disease endpoints (O'Neil et al., Reference O'Neil, Jacka, Quirk, Cocker, Taylor, Oldenburg and Berk2015).

For future research, our findings suggest that cognitive studies of BD with a cardiometabolic disease risk focus should consider concentrating efforts on better understanding the role of health risk behaviors, as well as measures of blood pressure and muscular strength, beyond other physiological risk factors such as lipids and inflammatory markers. It would be of particular benefit to explore how health-risk behaviors may aggregate or interact with one another using latent profile analysis, to better understand whether they collectively compound, or further explain, cognitive impairment in BD. Given recent evidence linking poor cognition to a high immune dysregulation subgroup of psychiatric patients (Sæther et al., Reference Sæther, Ueland, Haatveit, Maglanoc, Szabo, Djurovic and Ueland2022), exploring physiological risk factor profiles in relation to cognition in BD in that context may reveal more robust or stronger associations than those seen here. Future research would also do well to explore cardiac biomarkers and their relation to cognition and cognitive markers in BD. An association between these markers and cognition has already been established in the general population (Hosoki et al., Reference Hosoki, Hansra, Jayasena, Poljak, Mather, Catts and Sachdev2023; Jensen et al., Reference Jensen, Zeller, Twerenbold and Thomalla2023; van der Velpen et al., Reference van der Velpen, Feleus, Bertens and Sabayan2017), where relevantly, troponin – a widely recognized indicator of heart muscle damage – has been linked to individual variation in the factors associated with cognition in BD in this study; blood pressure, physical activity, and sedentary behavior (Aakre & Omland, Reference Aakre and Omland2019; Xue, Iqbal, Chan, & Maisel, Reference Xue, Iqbal, Chan and Maisel2014). Exploring the association of cognition and cardiac biomarkers like troponin could, therefore, help to delineate the mechanistic pathways involved in cognitive dysfunction within the disorder.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/S0033291724000722

Acknowledgements

This research was conducted using the UK Biobank Resource under application number 60698. We acknowledge the efforts of the UK Biobank research team and thank all UK Biobank participants for agreeing to volunteer for this research.

Funding statement

TVR was supported by a Dame Kate Campbell Fellowship from the University of Melbourne and an Al and Val Rosenstrauss Fellowship from the Rebecca Cooper Foundation. MB was supported by a NHMRC Senior Principal Research Fellowship and Leadership 3 Investigator grant (1156072 and 2017131). SLR was supported by a NHMRC Senior Research Fellowship (GNT1154651). DD (GNT1078360), and NO (GNT1003960) were supported by NHMRC Senior Research Fellowships and the Victorian Operational Infrastructure Support (OIS) Program.

Competing interests

RM has received research grant support from CIHR/GACD/Chinese National Natural Research Foundation; speaker/consultation fees from Lundbeck, Janssen, Purdue, Pfizer, Otsuka, Takeda, Neurocrine, Sunovion, Bausch Health, Novo Nordisk, Kris, Sanofi, Eisai, Intra-Cellular, NewBridge Pharmaceuticals, Abbvie. RM is a CEO of Braxia Scientific Corp. The remaining authors report no financial relationships with commercial interests.

References

Aakre, K. M., & Omland, T. (2019). Physical activity, exercise and cardiac troponins: Clinical implications. Progress in Cardiovascular Diseases, 62(2), 108–115. https://doi.org/10.1016/j.pcad.2019.02.005CrossRefGoogle ScholarPubMed
Aas, M., Djurovic, S., Ueland, T., Mørch, R. H., Fjæra, J., Reponen, E. J., … Melle, I. (2019). The relationship between physical activity, clinical and cognitive characteristics and BDNF mRNA levels in patients with severe mental disorders. The World Journal of Biological Psychiatry, 20(7), 567–576. https://doi.org/10.1080/15622975.2018.1557345CrossRefGoogle ScholarPubMed
Aliño-Dies, M., Sánchez-Ortí, J. V., Correa-Ghisays, P., Balanzá-Martínez, V., Vila-Francés, J., Selva-Vera, G., … Tabarés-Seisdedos, R. (2020). Grip strength, neurocognition, and social functioning in people with type-2 diabetes mellitus, major depressive disorder, bipolar disorder, and schizophrenia. Frontiers in Psychology, 11, 1–12. https://doi.org/10.3389/fpsyg.2020.525231CrossRefGoogle Scholar
Anatürk, M., Suri, S., Smith, S. M., Ebmeier, K. P., & Sexton, C. E. (2021). Leisure activities and their relationship with MRI measures of brain structure, functional connectivity, and cognition in the UK biobank cohort. Frontiers in Aging Neuroscience, 13, 1–11. https://doi.org/10.3389/fnagi.2021.734866CrossRefGoogle ScholarPubMed
Australian Institute of Health and Welfare. (2017). Abnormal blood lipids (dyslipidaemia). Retrieved from Risk Factors to Health website. https://www.aihw.gov.au/reports/risk-factors/risk-factors-to-health/contents/abnormal-blood-lipids-dyslipidaemiaGoogle Scholar
Balanzá-Martínez, V., Crespo-Facorro, B., González-Pinto, A., & Vieta, E. (2015). Bipolar disorder comorbid with alcohol use disorder: Focus on neurocognitive correlates. Frontiers in Physiology, 6(APR), 1–9. https://doi.org/10.3389/fphys.2015.00108Google ScholarPubMed
Baumgart, M., Snyder, H. M., Carrillo, M. C., Fazio, S., Kim, H., & Johns, H. (2015). Summary of the evidence on modifiable risk factors for cognitive decline and dementia: A population-based perspective. Alzheimer's and Dementia, 11(6), 718–726. https://doi.org/10.1016/j.jalz.2015.05.016CrossRefGoogle ScholarPubMed
Bora, E., McIntyre, R. S., & Ozerdem, A. (2019). Neurococognitive and neuroimaging correlates of obesity and components of metabolic syndrome in bipolar disorder: A systematic review. Psychological Medicine, 49(5), 738–749. https://doi.org/10.1017/S0033291718003008CrossRefGoogle ScholarPubMed
Bradley, A. J., Anderson, K. N., Gallagher, P., & McAllister-Williams, R. H. (2020). The association between sleep and cognitive abnormalities in bipolar disorder. Psychological Medicine, 50(1), 125–132. https://doi.org/10.1017/S0033291718004038CrossRefGoogle ScholarPubMed
Burgess, J. L., Bradley, A. J., Anderson, K. N., Gallagher, P., & McAllister-Williams, R. H. (2022). The relationship between physical activity, BMI, circadian rhythm, and sleep with cognition in bipolar disorder. Psychological Medicine, 52(3), 467–475. https://doi.org/10.1017/S003329172000210XCrossRefGoogle ScholarPubMed
Cardoso, T. d. A., Bauer, I. E., Jansen, K., Suchting, R., Zunta-Soares, G., Quevedo, J., … Soares, J. C. (2016). Effect of alcohol and illicit substance use on verbal memory among individuals with bipolar disorder. Psychiatry Research, 243(3), 225–231. https://doi.org/10.1016/j.psychres.2016.06.044CrossRefGoogle ScholarPubMed
Chuang, S. Y., Hsu, P. F., Chang, H. Y., Bai, C. H., Yeh, W. T., & Pan, H. W. (2013). C-reactive protein predicts systolic blood pressure and pulse pressure but not diastolic blood pressure: The cardiovascular disease risk factors two-township study. American Journal of Hypertension, 26(5), 657–664. https://doi.org/10.1093/ajh/hps095CrossRefGoogle Scholar
Coello, K., Kjærstad, H. L., Stanislaus, S., Melbye, S., Faurholt-Jepsen, M., Miskowiak, K. W., … Munkholm, K. (2019). Thirty-year cardiovascular risk score in patients with newly diagnosed bipolar disorder and their unaffected first-degree relatives. Australian and New Zealand Journal of Psychiatry, 53(7), 651–662. https://doi.org/10.1177/0004867418815987CrossRefGoogle ScholarPubMed
Congio, A. C., Urbano, M. R., Soares, M. R. Z., & Nunes, S. O. V. (2022). Cognitive impairment, childhood trauma, sedentary behaviour, and elevated C-reactive protein levels in major affective disorders. Journal of Psychiatric Research, 155(July), 1–9. https://doi.org/10.1016/j.jpsychires.2022.07.052CrossRefGoogle Scholar
Cui, M., Zhang, S., Liu, Y., Gang, X., & Wang, G. (2021). Grip strength and the risk of cognitive decline and dementia: A systematic review and meta-analysis of longitudinal cohort studies. Frontiers in Aging Neuroscience, 13, 1–11. https://doi.org/10.3389/fnagi.2021.625551CrossRefGoogle ScholarPubMed
Dickerson, F., Stallings, C., Origoni, A., Vaughan, C., Khushalani, S., & Yolken, R. (2013). Elevated C-reactive protein and cognitive deficits in individuals with bipolar disorder. Journal of Affective Disorders, 150(2), 456–459. https://doi.org/10.1016/j.jad.2013.04.039CrossRefGoogle ScholarPubMed
Dintica, C. S., & Yaffe, K. (2022). Epidemiology and risk factors for dementia. Psychiatric Clinics of North America, 45(4), 677–689. https://doi.org/10.1016/j.psc.2022.07.011CrossRefGoogle ScholarPubMed
Fawns-Ritchie, C., & Deary, I. J. (2020). Reliability and validity of the UK biobank cognitive tests. PLoS ONE, 15(4), 1–24. https://doi.org/10.1371/journal.pone.0231627CrossRefGoogle ScholarPubMed
Fellendorf, F. T., Kainzbauer, N., Platzer, M., Dalkner, N., Bengesser, S. A., Birner, A., … Reininghaus, E. Z. (2017). Gender differences in the association between physical activity and cognitive function in individuals with bipolar disorder. Journal of Affective Disorders, 221, 232–237. https://doi.org/10.1016/j.jad.2017.06.048CrossRefGoogle ScholarPubMed
Firth, J., Firth, J. A., Stubbs, B., Vancampfort, D., Schuch, F. B., Hallgren, M., … Sarris, J. (2018a). Association between muscular strength and cognition in people with major depression or bipolar disorder and healthy controls. JAMA Psychiatry, 75(7), 740–746. https://doi.org/10.1001/jamapsychiatry.2018.0503CrossRefGoogle ScholarPubMed
Firth, J., Stubbs, B., Vancampfort, D., Firth, J. A., Large, M., Rosenbaum, S., … Yung, A. R. (2018b). Grip strength is associated with cognitive performance in schizophrenia and the general population: A UK biobank study of 476559 participants. Schizophrenia Bulletin, 44(4), 728–736. https://doi.org/10.1093/schbul/sby034CrossRefGoogle ScholarPubMed
Forte, G., De Pascalis, V., Favieri, F., & Casagrande, M. (2020). Effects of blood pressure on cognitive performance: A systematic review. Journal of Clinical Medicine, 9(1), 1–22. https://doi.org/10.3390/jcm9010034Google Scholar
Goldstein, B. I., Schaffer, A., Wang, S., & Blanco, C. (2015). Excessive and premature new-onset cardiovascular disease among adults with bipolar disorder in the US NESARC cohort. Journal of Clinical Psychiatry, 76(2), 163–169. https://doi.org/10.4088/JCP.14m09300CrossRefGoogle Scholar
Harvey, P. D. (2019). Domains of cognition and their assessment. Dialogues in Clinical Neuroscience, 21(3), 227–237. https://doi.org/10.31887/DCNS.2019.21.3/pharveyCrossRefGoogle ScholarPubMed
Hosoki, S., Hansra, G. K., Jayasena, T., Poljak, A., Mather, K. A., Catts, V. S., … Sachdev, P. S. (2023). Molecular biomarkers for vascular cognitive impairment and dementia. Nature Reviews Neurology, 19(12), 737–753. https://doi.org/10.1038/s41582-023-00884-1CrossRefGoogle ScholarPubMed
Hubenak, J., Tuma, I., & Bazant, J. (2015). Association of arterial hypertension and cognitive impairment in euthymic bipolar disorder. Neuroendocrinology Letters, 36(3), 294–300.Google ScholarPubMed
Infurna, F. J., Gerstorf, D., & Lachman, M. E. (2020). Midlife in the 2020s: Opportunities and challenges. American Psychologist, 75(4), 470–485. https://doi.org/10.1037/amp0000591CrossRefGoogle Scholar
International Diabetes Federation. (2021). IDF Diabetes altas (10th ed.). Retrieved from https://diabetesatlas.org/idfawp/resource-files/2021/07/IDF_Atlas_10th_Edition_2021.pdfGoogle Scholar
International Physical Activity Questionnaire. (2005). Guidelines for Data Processing and Analysis of the International Physical Activity Questionnaire (IPAQ) – Short and Long Forms, revised on November 2005, 1–15.Google Scholar
Irwin, M. R., Olmstead, R., & Carroll, J. E. (2016). Sleep disturbance, sleep duration, and inflammation: A systematic review and meta-analysis of cohort studies and experimental sleep deprivation. Biological Psychiatry, 80(1), 40–52. https://doi.org/10.1016/j.biopsych.2015.05.014CrossRefGoogle ScholarPubMed
Jensen, M., Zeller, T., Twerenbold, R., & Thomalla, G. (2023). Circulating cardiac biomarkers, structural brain changes, and dementia: Emerging insights and perspectives. Alzheimer's and Dementia, 19(4), 1529–1548. https://doi.org/10.1002/alz.12926CrossRefGoogle Scholar
Kanady, J. C., Soehner, A. M., Klein, A. B., & Harvey, A. G. (2017). The association between insomnia-related sleep disruptions and cognitive dysfunction during the inter-episode phase of bipolar disorder. Journal of Psychiatric Research, 88(1), 80–88. https://doi.org/10.1016/j.jpsychires.2017.01.001.TheCrossRefGoogle Scholar
Kendall, K. M., Rees, E., Escott-Price, V., Einon, M., Thomas, R., Hewitt, J., … Kirov, G. (2017). Cognitive performance among carriers of pathogenic copy number variants: Analysis of 152000 UK biobank subjects. Biological Psychiatry, 82(2), 103–110. https://doi.org/10.1016/j.biopsych.2016.08.014CrossRefGoogle Scholar
Lakoski, S. G., Cushman, M., Palmas, W., Blumenthal, R., D'Agostino, R. B., & Herrington, D. M. (2005). The relationship between blood pressure and C-reactive protein in the Multi-Ethnic Study of Atherosclerosis (MESA). Journal of the American College of Cardiology, 46(10), 1869–1874. https://doi.org/10.1016/j.jacc.2005.07.050CrossRefGoogle ScholarPubMed
Laskemoen, J. F., Büchmann, C., Barrett, E. A., Collier-Høegh, M., Haatveit, B., Vedal, T. J., … Simonsen, C. (2020). Do sleep disturbances contribute to cognitive impairments in schizophrenia spectrum and bipolar disorders? European Archives of Psychiatry and Clinical Neuroscience, 270(6), 749–759. https://doi.org/10.1007/s00406-019-01075-0CrossRefGoogle ScholarPubMed
Lo, J. C., Groeger, J. A., Cheng, G. H., Dijk, D. J., & Chee, M. W. L. (2016). Self-reported sleep duration and cognitive performance in older adults: A systematic review and meta-analysis. Sleep Medicine, 17, 87–98. https://doi.org/10.1016/j.sleep.2015.08.021CrossRefGoogle Scholar
Menkes, M. W., Andrews, C. M., Burgess, H. J., Carley, I., Marshall, D. F., Langenecker, S. A., … Ryan, K. A. (2021). Sleep quality and neuropsychological functioning in bipolar I disorder. Journal of Affective Disorders, 293(September 2020), 133–140. https://doi.org/10.1016/j.jad.2021.06.022CrossRefGoogle ScholarPubMed
Millett, C. E., Perez-Rodriguez, M., Shanahan, M., Larsen, E., Yamamoto, H. S., Bukowski, C., … Burdick, K. E. (2019). C-reactive protein is associated with cognitive performance in a large cohort of euthymic patients with bipolar disorder. Molecular Psychiatry, 26(8), 4096–4105. https://doi.org/10.1038/s41380-019-0591-1CrossRefGoogle Scholar
Milton, D. C., Ward, J., Ward, E., Lyall, D. M., Strawbridge, R. J., Smith, D. J., … Cullen, B. (2021). The association between C-reactive protein, mood disorder, and cognitive function in UK Biobank. European Psychiatry, 64(1), 1–10. https://doi.org/10.1192/j.eurpsy.2021.6CrossRefGoogle Scholar
Morrens, M., Overloop, C., Coppens, V., Loots, E., Van Den Noortgate, M., Vandenameele, S., … De Picker, L. (2022). The relationship between immune and cognitive dysfunction in mood and psychotic disorder: A systematic review and a meta-analysis. Molecular Psychiatry, 27(8), 3237–3246. https://doi.org/10.1038/s41380-022-01582-yCrossRefGoogle ScholarPubMed
Morris, G., Puri, B. K., Walker, A. J., Maes, M., Carvalho, A. F., Bortolasci, C. C., … Berk, M. (2019). Shared pathways for neuroprogression and somatoprogression in neuropsychiatric disorders. Neuroscience and Biobehavioral Reviews, 107(June), 862–882. https://doi.org/10.1016/j.neubiorev.2019.09.025CrossRefGoogle ScholarPubMed
Mutz, J., Young, A. H., & Lewis, C. M. (2022). Age-related changes in physiology in individuals with bipolar disorder. Journal of Affective Disorders, 296, 157–168. https://doi.org/10.1016/j.jad.2021.09.027CrossRefGoogle ScholarPubMed
Naiberg, M. R., Newton, D. F., Collins, J. E., Dickstein, D. P., Bowie, C. R., & Goldstein, B. I. (2016). Elevated triglycerides are associated with decreased executive function among adolescents with bipolar disorder. Acta Psychiatrica Scandinavica, 134(3), 241–248. https://doi.org/10.1111/acps.12603CrossRefGoogle ScholarPubMed
O'Neil, A., Jacka, F. N., Quirk, S. E., Cocker, F., Taylor, C. B., Oldenburg, B., & Berk, M. (2015). A shared framework for the common mental disorders and non-communicable disease: Key considerations for disease prevention and control. BMC Psychiatry, 15(1), 15. https://doi.org/10.1186/s12888-015-0394-0CrossRefGoogle Scholar
Ou, Y.-N., Tan, C.-C., Shen, X.-N., Xu, W., Hou, X.-H., Dong, Q., … Yu, J.-T. (2020). Blood pressure and risks of cognitive impairment and dementia. Hypertension, 76(1), 217–225. https://doi.org/10.1161/HYPERTENSIONAHA.120.14993CrossRefGoogle Scholar
Qiu, C., & Fratiglioni, L. (2015). A major role for cardiovascular burden in age-related cognitive decline. Nature Reviews Cardiology, 12(5), 267–277. https://doi.org/10.1038/nrcardio.2014.223CrossRefGoogle Scholar
Ringin, E., Dunstan, D. W., McIntyre, R. S., Berk, M., Owen, N., Rossell, S. L., … Van Rheenen, T. E. (2022). Interactive relationships of type 2 diabetes and bipolar disorder with cognition: Evidence of putative premature cognitive ageing in the UK Biobank Cohort. Neuropsychopharmacology, 48, 362–370. https://doi.org/10.1038/s41386-022-01471-6CrossRefGoogle ScholarPubMed
Ringin, E., Dunstan, D. W., McIntyre, R. S., Owen, N., Berk, M., Rossell, S. L., … Van Rheenen, T. E. (2023). Differential associations of mentally-active and passive sedentary behaviours and physical activity with putative cognitive decline in healthy individuals and those with bipolar disorder: Findings from the UK Biobank cohort. Mental Health and Physical Activity, 24, 100514. https://doi.org/10.1016/j.mhpa.2023.100514CrossRefGoogle Scholar
Rizzo, L. B., Costa, L. G., Mansur, R. B., Swardfager, W., Belangero, S. I., Grassi-Oliveira, R., … Brietzke, E. (2014). The theory of bipolar disorder as an illness of accelerated aging: Implications for clinical care and research. Neuroscience and Biobehavioral Reviews, 42, 157–169. https://doi.org/10.1016/j.neubiorev.2014.02.004CrossRefGoogle ScholarPubMed
Roberts, H. C., Denison, H. J., Martin, H. J., Patel, H. P., Syddall, H., Cooper, C., & Sayer, A. A. (2011). A review of the measurement of grip strength in clinical and epidemiological studies: Towards a standardised approach. Age and Ageing, 40(4), 423–429. https://doi.org/10.1093/ageing/afr051CrossRefGoogle ScholarPubMed
Russo, M., Mahon, K., Shanahan, M., Ramjas, E., Solon, C., Purcell, S. M., & Burdick, K. E. (2015). The relationship between sleep quality and neurocognition in bipolar disorder. Journal of Affective Disorders, 187(1), 156–162. https://doi.org/10.1016/j.jad.2015.08.009CrossRefGoogle ScholarPubMed
Sæther, L. S., Ueland, T., Haatveit, B., Maglanoc, L. A., Szabo, A., Djurovic, S., … Ueland, T. (2022). Inflammation and cognition in severe mental illness: Patterns of covariation and subgroups. Molecular Psychiatry, 28, 1284–1292. https://doi.org/10.1038/s41380-022-01924-wCrossRefGoogle ScholarPubMed
Sha, T., Cheng, W., & Yan, Y. (2018). Prospective associations between pulse pressure and cognitive performance in Chinese middle-aged and older population across a 5-year study period. Alzheimer's Research and Therapy, 10(1), 1–11. https://doi.org/10.1186/s13195-018-0355-1CrossRefGoogle ScholarPubMed
Smith, D. J., Nicholl, B. I., Cullen, B., Martin, D., Ul-Haq, Z., Evans, J., … Pell, J. P. (2013). Prevalence and characteristics of probable major depression and bipolar disorder within UK Biobank: Cross-sectional study of 172751 participants. PLoS ONE, 8(11), 1–7. https://doi.org/10.1371/journal.pone.0075362CrossRefGoogle Scholar
Tahmi, M., Palta, P., & Luchsinger, J. A. (2021). Metabolic syndrome and cognitive function. Current Cardiology Reports, 23(12), 1–20. https://doi.org/10.1007/s11886-021-01615-yCrossRefGoogle ScholarPubMed
The Royal College of Pathologists of Australasia. (2019). C-Reactive Protein. Retrieved from https://www.rcpa.edu.au/Manuals/RCPA-Manual/Pathology-Tests/C/C-Reactive-proteinGoogle Scholar
Tsai, S. Y., Lee, H. C., Chen, C. C., & Huang, Y. L. (2007). Cognitive impairment in later life in patients with early-onset bipolar disorder. Bipolar Disorders, 9(8), 868–875. https://doi.org/10.1111/j.1399-5618.2007.00498.xCrossRefGoogle ScholarPubMed
Tuttle, C. S. L., Thang, L. A. N., & Maier, A. B. (2020). Markers of inflammation and their association with muscle strength and mass: A systematic review and meta-analysis. Ageing Research Reviews, 64(September), 101185. https://doi.org/10.1016/j.arr.2020.101185CrossRefGoogle Scholar
UK Biobank. (2019). Biomarker assay quality procedures: approaches used to minimise systematic and random errors (and the wider epidemiological implications). Retrieved from https://biobank.ndph.ox.ac.uk/showcase/ukb/docs/biomarker_issues.pdfGoogle Scholar
UK Biobank: Protocol for a large-scale prospective epidemiological resource. (2007). Retrieved from ukbiobank.ac.uk/media/gnkeyh2q/study-rationale.pdfGoogle Scholar
van der Velpen, I. F., Feleus, S., Bertens, A. S., & Sabayan, B. (2017). Hemodynamic and serum cardiac markers and risk of cognitive impairment and dementia. Alzheimer's & Dementia, 13(4), 441–453. https://doi.org/10.1016/j.jalz.2016.09.004CrossRefGoogle ScholarPubMed
Van Rheenen, T. E., & Neil, A. O. (2022). What can lifestyle factors tell us about cognitive dysfunction in bipolar disorder? European Neuropsychopharmacology, 64, 1–3. https://doi.org/10.1016/j.euroneuro.2022.07.574CrossRefGoogle ScholarPubMed
Van Rheenen, T. E., Lewandowski, K. E., Bauer, I. E., Kapczinski, F., Miskowiak, K., Burdick, K. E., & Balanzá-Martínez, V. (2020). Current understandings of the trajectory and emerging correlates of cognitive impairment in bipolar disorder: An overview of evidence. Bipolar Disorders, 22(1), 13–27. https://doi.org/10.1111/bdi.12821CrossRefGoogle ScholarPubMed
Van Rheenen, T. E., McIntyre, R. S., Balanzá-Martínez, V., Berk, M., & Rossell, S. L. (2021a). Cumulative cardiovascular disease risk and triglycerides differentially relate to subdomains of executive function in bipolar disorder; preliminary findings. Journal of Affective Disorders, 278(September 2020), 556–562. https://doi.org/10.1016/j.jad.2020.09.104CrossRefGoogle ScholarPubMed
Van Rheenen, T. E., Miskowiak, K., & Burdick, K. E. (2021b). Recognising the relevance of cognitive dysfunction in the clinical management of bipolar disorder. Bipolar Disorders, 23(4), 414–415. https://doi.org/10.1111/bdi.13063CrossRefGoogle ScholarPubMed
Vancampfort, D., Firth, J., Schuch, F. B., Rosenbaum, S., Mugisha, J., Hallgren, M., … Stubbs, B. (2017). Sedentary behavior and physical activity levels in people with schizophrenia, bipolar disorder and major depressive disorder: A global systematic review and meta-analysis. World Psychiatry, 16(3), 308–315. https://doi.org/10.1002/wps.20458CrossRefGoogle Scholar
Vancampfort, D., Ph, D., Vansteelandt, K., Ph, D., Correll, C. U., Mitchell, A. J., … Ph, D. (2013). Metabolic Syndrome and Metabolic Abnormalities in Bipolar Disorder: A Meta-Analysis of Prevalence Rates and Moderators. (March), 265–274.CrossRefGoogle Scholar
Waldstein, S. R., & Wendell, C. R. (2010). Neurocognitive function and cardiovascular disease. Journal of Alzheimer's Disease, 20(3), 833–842. https://doi.org/10.3233/JAD-2010-091591CrossRefGoogle ScholarPubMed
Weiner, M., Warren, L., & Fiedorowicz, J. G. (2011). Cardiovascular morbidity and mortality in bipolar disorder. Annals of Clinical Psychiatry, 23(1), 40–47.Google ScholarPubMed
Wells, C. L., & Townsend, M. (2019). Management of the acutely ill and medically complex older patient. In Guccione's geriatric physical therapy (pp. 309–343). St. Louis, Missouri: Elsevier. https://doi.org/10.1016/B978-0-323-60912-8.00014-2Google Scholar
Wennberg, A., Wu, M., Rosenberg, P., & Spira, A. (2017). Sleep disturbance, cognitive decline, and dementia: A review. Seminars in Neurology, 37(04), 395–406. https://doi.org/10.1055/s-0037-1604351Google ScholarPubMed
Wirth, K., Klenk, J., Brefka, S., Dallmeier, D., Faehling, K., Roqué i Figuls, M., … Stubbs, B. (2017). Biomarkers associated with sedentary behaviour in older adults: A systematic review. Ageing Research Reviews, 35, 87–111. https://doi.org/10.1016/j.arr.2016.12.002CrossRefGoogle ScholarPubMed
Xue, Y., Iqbal, N., Chan, J., & Maisel, A. (2014). Biomarkers in hypertension and their relationship with myocardial target-organ damage. Current Hypertension Reports, 16(12), 1–7. https://doi.org/10.1007/s11906-014-0502-4CrossRefGoogle ScholarPubMed
Yaffe, K., Falvey, C. M., & Hoang, T. (2014). Connections between sleep and cognition in older adults. The Lancet Neurology, 13(10), 1017–1028. https://doi.org/10.1016/S1474-4422(14)70172-3CrossRefGoogle ScholarPubMed
Zonneveld, M. H., Abbel, D., le Cessie, S., Jukema, J. W., Noordam, R., & Trompet, S. (2023). Cardiac troponin, cognitive function, and dementia: A systematic review. Aging and Disease, 14(2), 386–397. https://doi.org/10.14336/AD.2022.0818Google ScholarPubMed
Figure 0

Table 1.Characteristics of participants (mean ± standard errors)

Figure 1

Table 2.Combined model of cardiometabolic disease risk factors and global cognition (all age groups)

Figure 2

Figure 1.Effects (bootstrapped standard error in parenthesis) for mediation examining how elevated CRP was associated with global cognition after controlling for sex, educational level, and SES. *p < 0.05. Bolded (red) lines indicate significant mediation pathways (range of CI did not span 0).

Figure 3

Table 3.Age-stratified model of cardiometabolic disease risk factors and global cognition; midlife patients

Figure 4

Table 4.Age-stratified model of cardiometabolic disease risk factors and global cognition; later-life patients

Supplementary material: File

Ringin et al. supplementary material

Ringin et al. supplementary material
Download Ringin et al. supplementary material(File)
File 222.6 KB