Associations between physical activity and CVD-related metabolomic and proteomic biomarkers

Örjan Ekblom; Harry Björkbacka; Mats Börjesson; Elin Ekblom-Bak; Anders Blomberg; Kenneth Caidahl; Ewa Ehrenborg; Gunnar Engström; Jan Engvall; David Erlinge; Tove Fall; Bruna Gigante; Anders Gummesson; Tomas Jernberg; Lars Lind; David Molnar; Jonas Oldgren; Aidin Rawshani; Johan Sundström; Stefan Söderberg; Patrik Wennberg; Carl Johan Östgren

doi:10.1371/journal.pone.0325720

. 2025 Jun 11;20(6):e0325720. doi: 10.1371/journal.pone.0325720

Associations between physical activity and CVD-related metabolomic and proteomic biomarkers

Örjan Ekblom ^1,^2,^*, Harry Björkbacka ³, Mats Börjesson ^4,⁵, Elin Ekblom-Bak ¹, Anders Blomberg ⁶, Kenneth Caidahl ^7,⁸, Ewa Ehrenborg ⁹, Gunnar Engström ³, Jan Engvall ^10,¹¹, David Erlinge ¹², Tove Fall ¹³, Bruna Gigante ^14,¹⁵, Anders Gummesson ^16,¹⁷, Tomas Jernberg ¹⁸, Lars Lind ¹⁹, David Molnar ^17,²⁰, Jonas Oldgren ^21,²², Aidin Rawshani ^17,²³, Johan Sundström ^24,²⁵, Stefan Söderberg ⁶, Patrik Wennberg ²⁶, Carl Johan Östgren ^10,²⁷

Editor: Hidetaka Hamasaki²⁸

¹Department of Physical Activity and Health, The Swedish School of Sport and Health Sciences (GIH), Stockholm, Sweden

²Department of Neurobiology, Care Sciences and Society, Division of Nursing, Karolinska Institutet, Stockholm, Sweden

³Department of Clinical Sciences in Malmö, Lund University, Malmö, Sweden

⁴Center for Lifestyle Intervention, Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

⁵Sahlgrenska University Hospital, Region Västra Götaland, Gothenburg, Sweden

⁶Department of Public Health and Clinical Medicine, Umeå University, Umeå, Sweden

⁷Department of Clinical Physiology, Karolinska University Hospital, and Karolinska Institutet, Stockholm, Sweden

⁸Department of Clinical Physiology, Sahlgrenska University Hospital, and Sahlgrenska Academy, Gothenburg, Sweden

⁹Cardiovascular Medicine Unit, Department of Medicine, Center for Molecular Medicine, Karolinska Institute and Karolinska University Hospital, Stockholm, Sweden

¹⁰CMIV, Centre of Medical Image Science and Visualization, Linköping University, Linköping, Sweden

¹¹Department of Clinical Physiology, and Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden

¹²Department of Clinical Sciences Lund, Cardiology, Lund University, and Skåne University Hospital, Lund, Sweden

¹³Department of Medical Sciences, Molecular Epidemiology, Uppsala University, Uppsala, Sweden

¹⁴Division of Cardiovascular Medicine, Department of Medicine, Karolinska Institutet, Stockholm, Sweden

¹⁵Department of Cardiology, Danderyds Hospital, Stockholm, Sweden

¹⁶Department of Clinical Genetics and Genomics, Region Västra Götaland, Sahlgrenska University Hospital, Gothenburg, Sweden

¹⁷Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

¹⁸Department of Clinical Sciences, Danderyd University Hospital, Karolinska Institutet, Stockholm, Sweden

¹⁹Department of Medical Sciences, Clinical Epidemiology, University of Uppsala, Uppsala, Sweden

²⁰Department of Radiology, Region Västra Götaland, Sahlgrenska University Hospital, Gothenburg, Sweden

²¹Department of Medical Sciences, Cardiology, Uppsala University, Uppsala, Sweden

²²Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden

²³Wallenberg Laboratory for Cardiovascular and Metabolic Research, Institute of Medicine, University of Gothenburg, Gothenburg, Sweden

²⁴Department of Medical Sciences, Uppsala University, Uppsala, Sweden

²⁵The George Institute for Global Health, University of New South Wales, Sydney, Australia

²⁶Department of Public Health and Clinical Medicine, Family Medicine, Umeå University, Umeå, Sweden

²⁷Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden

²⁸Japan Medical Exercise Association, JAPAN

Competing Interests: J. S. reports direct or indirect stock ownership in companies (Anagram kommunikation AB, Sence Research AB, Symptoms Europe AB, MinForskning AB) providing services to companies and authorities in the health sector including Amgen, AstraZeneca, Bayer, Boehringer, Eli Lilly, Gilead, GSK, University of Gothenburg, Itrim, Ipsen, Janssen, Karolinska Institutet, LIF, Linköping University, Novo Nordisk, Parexel, Pfizer, Region Stockholm, Region Uppsala, Sanofi, STRAMA, Takeda, TLV, Uppsala University, Vifor Pharma, WeMind. S.S. has received compensation for consultancy or speaking for Johnson & Johnson and Merck. The rest of the authors declared that no competing interests exist.

^✉

* E-mail: orjan.ekblom@gih.se

Roles

Örjan Ekblom: Conceptualization, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing

Harry Björkbacka: Conceptualization, Methodology, Writing – original draft, Writing – review & editing

Mats Börjesson: Conceptualization, Methodology, Writing – original draft, Writing – review & editing

Elin Ekblom-Bak: Conceptualization, Methodology, Writing – original draft, Writing – review & editing

Anders Blomberg: Methodology, Writing – review & editing

Kenneth Caidahl: Methodology, Writing – review & editing

Ewa Ehrenborg: Methodology, Writing – review & editing

Gunnar Engström: Methodology, Writing – review & editing

Jan Engvall: Methodology, Writing – review & editing

David Erlinge: Methodology, Writing – review & editing

Tove Fall: Methodology, Writing – review & editing

Bruna Gigante: Methodology, Writing – review & editing

Anders Gummesson: Methodology, Writing – review & editing

Tomas Jernberg: Methodology, Writing – review & editing

Lars Lind: Methodology, Writing – review & editing

David Molnar: Methodology, Writing – review & editing

Jonas Oldgren: Methodology, Writing – review & editing

Aidin Rawshani: Methodology, Writing – review & editing

Johan Sundström: Methodology, Writing – review & editing

Stefan Söderberg: Methodology, Writing – review & editing

Patrik Wennberg: Methodology, Writing – review & editing

Carl Johan Östgren: Conceptualization, Investigation, Methodology, Writing – original draft, Writing – review & editing

Hidetaka Hamasaki: Editor

PMCID: PMC12157240 PMID: 40498722

Abstract

Aim

Habitual physical activity (PA) affects metabolism and homeostasis in various tissues and organs. However, detailed knowledge of associations between PA and cardiovascular disease (CVD) risk markers is limited. We sought to identify associations between accelerometer-assessed PA classes and 183 proteomic and 154 metabolomic CVD-related biomarkers.

Method

We utilized cross-sectional data from the main SCAPIS cohort (n = 4647, median age: 57.5 yrs, 50.5% female) as a discovery sample and the SCAPIS pilot cohort (n = 910, median age: 57.5 yrs, 50.3% female) as a validation sample. PA was assessed via hip-worn accelerometers, while plasma concentrations of proteomic biomarkers were measured using Olink CVD II and III panels. Metabolomic markers were assessed using the Nightingale NMR platform. We evaluated associations between four PA classes (moderate-to-vigorous PA [MVPA], low-intensity PA [LIPA], sedentary [SED], and prolonged SED [prolSED]) and biomarkers, controlling for potential confounders and applying a false discovery rate of 5% using multiple linear regressions.

Results

A total of eighty-five metabolomic markers and forty-three proteomic markers were validated and found to be significantly associated with one or more PA classes. LIPA and SED markers demonstrated significant mirroring or opposing relations to biomarkers, while prolSED mainly shared relations with SED. Notably, HDL species were predominantly negatively associated with SED, whereas LDL species were positively associated with SED and negatively associated with MVPA. Among the proteomic markers, eighteen were uniquely associated with MVPA (among those Interleukin – 6 [IL6] and Growth/differentiation factor 15 [GDF15] both negatively related), seven with SED (among those Metalloproteinase inhibitor 4 [TIMP4] and Tumor necrosis factor receptor 2 [TNFR2], both positively related), and eight were related to both SED/prolSED (among those Lipoprotein lipase [LPL] negatively related to SED and leptin [LEP] positively related to SED) and MVPA (with LPL positively related to MVPA and LEP negatively related to MVPA).

Conclusion

Our findings suggest the existence of specific associations between PA classes and metabolomic and cardiovascular protein biomarkers in a middle-aged population. Beyond validation of previous results, we identified new associations. This multitude of connections between PA and CVD-related markers may help elucidate the previously observed relationship between PA and CVD. The identified cross-sectional associations could inform the design of future experimental studies, serving as important outcome measures.

Introduction

Analyses of single plasma or serum biomarkers have increased our understanding of how PA may relate to, for example, CVD risk [1]. Still, analysing single or a few biomarkers is probably not enough to capture the full width of the processes associated with PA [2]. Simultaneous assessments of many hundred CVD-related biomarkers have been made possible using proteomics and metabolomics. Only a limited number of previous studies have been published using these techniques, but results can potentially shed light on the variation between individual biomarkers in cardiometabolic responses to PA. Stattin et al. [3] reported a relation between self-reported PA and 28 cardiovascular-specific proteins related to several atherosclerotic processes such as low-density lipoprotein oxidation, protein degradation, and immune cell adhesion and migration. However, self-reported PA is often less temporally granulated compared to accelerometer-based data, which hinders comparisons between time spent in various intensities of PA (i.e., moderate or intense PA vs sedentary behaviours or low-intensity PA). Further, a low-precision PA measurement with larger random and systematic errors, such as self-reports compared to sensor-based measures used as an exposure variable [4,5], will generate regression dilution [6] and thereby under-estimate associations. Accelerometer-based PA metrics have been shown to possess a stronger relation to CVD-related biomarkers, compared to that of self-reports [7]. Also, accelerometer-based variables have been shown to be highly predictive for all-cause and CVD mortality [8,9].

Using accelerometer-based PA data and NMR lipidomic, Henson et al. reported unfavourable relations between time spent sedentary and several markers (including VLDL, HDL, and apolipoprotein A1) in 509 middle-aged participants at high risk of type II diabetes. In younger men, Vaara et al. [10] reported only a few associations between accelerometer-based PA and IDL triglycerides 3triglycerides 3-hydroxybutyrate, indicating that associations are weaker in younger individuals. To what extent these associations are present in a mixed, middle-aged population is unknown.

Time spent sedentary (SED) and in moderate to vigorous physical activity (MVPA) is typically not strongly related, but SED and low-intensity physical activity (LIPA) are often strongly, negatively correlated [10]. As there is an interrelation between the different PA classes, such that for example time spent in MVPA may alter the relation between time spent in SED and CVD risk [11], it is important to be able to adjust analyses accordingly. There are several studies on proteomic and metabolomic biomarkers in relation to PA, based on either clinical populations [12], effects from short-term high-intensity interventions [13–15],or based on self-reported physical activity or fitness [16]. These studies generally report beneficial relations to PA and limited, but detrimental, relations to time spent sedentary. Studies on PA in relation to proteomics are less well understood, as many proteins are involved in multiple pathways, but data indicate a beneficial relation between PA and inflammatory markers and markers for lipid metabolism. However, there is a paucity of studies reporting relations between both proteomic and metabolomic biomarkers and sensor-based PA, in the same cohorts and applying a cross-validation design [17].

Regarding sedentary time, there is evidence to suggest that longer periods of uninterrupted sitting have a separate association with CVD health compared to short bouts of sitting or other physically inactive events [18]. Therefore, there may be unknown differences in associations between CVD-related biomarkers and different PA classes. This lack of data on which biomarkers or groups of biomarkers are related to PA classes limits the understanding of potential genetic or behavioural contributions. As previous studies have indicated causal relationships between PA to sarcopenia [19] and metabolic disorders [20], there is a need for such data in designing individualized preventative efforts including PA.

The aim of the present study was therefore to explore and validate associations between proteomic (Olink) and nuclear magnetic resonance metabolomic (NMR) CVD-related biomarkers and four different accelerometer-based PA classes in two population-based samples of middle-aged individuals in Sweden. We hypothesize that relations between PA and biomarkers are different across PA classes.

Materials and methods

This study was based on cross-sectional data from the Swedish CArdioPulmonary bioImage Study (SCAPIS) and included data from the main cohort as well as the local Gothenburg pilot study. We used data from the main SCAPIS cohort (n = 5557) as a discovery sample and data from the SCAPIS pilot cohort (n = 910) as a validation sample. Bergström et al have described the SCAPIS cohorts and data collection in detail [21].

SCAPIS main cohort (discovery sample)

The SCAPIS main cohort is a Swedish cohort of randomly selected participants from the general population (age 50–64 yrs) in six regions in Sweden. The SCAPIS data collection was conducted between February 14th, 2013, and October 30th, 2018, as a multicentre study at Swedish university hospitals.

SCAPIS pilot cohort (validation sample)

The Swedish CArdioPulmonary bioImage Study (SCAPIS) pilot study, which was conducted at the Sahlgrenska University Hospital in Gothenburg, Sweden, between March 12th, 2012, and October 26th, 2012, was used as a validation sample.

The data collection for the pilot study and the main study were approved by the ethics boards at Gothenburg University (No 638−16) and a Umeå University (No 2010–228-31M and 2017–183-31M), respectively. All data collection adheres to the Declaration of Helsinki and all participants gave written consent.

Assessment of the PA classes

Accelerometer-based sedentary behaviours and PA patterns were derived from tri-axial accelerometers (ActiGraph model GT3X and GT3X + , ActiGraph LCCo, Pensacola, FL, USA). The participants wore the accelerometers for 7 days. Participants were instructed to wear the accelerometer in an elastic belt over the right hip during all waking hours, except during water-based activities, and to return after the wearing period. ActiLife v.6.13.3 software was used to initialize the accelerometers and to download and process the collected data. The accelerometer recorded raw data (sample rate set to 30 Hz) from three axes, which were combined into a resulting vector, expressed in counts per minute (cpm). In the main SCAPIS cohort data were down-sampled and extracted as 60-second epochs using a low-frequency extension filter. The use of a low-frequency extension filter may affect the time spent in lower intensities, but the relations to biomarkers are influenced to a minimal degree. For the pilot SCAPIS sample, the data available were extracted as 15-second epochs.

In both samples, non-wear time was defined as sixty or more consecutive minutes with no movement (0 cpm), with allowance for a maximum of 2 minutes of counts between 0 and 199 cpm. Wear time was calculated as 24 hours minus non-wear time. Participants with a minimum of 600 minutes of valid daily wear time for at least 4 days were included [22].

Wear time was calculated as 24 hours minus non-wear time. Sedentary time was defined as <200 cpm [23], low-intensity PA (LIPA) as 200–2689 cpm, and moderate-to-vigorous intensity PA (MVPA) as ≥2690 [24]. A prolonged sedentary (prolSED) bout was defined as ≥20 min of cpm below 200 [25], with no allowance for interruption above the threshold. As daily wear time varied between study participants above the minimum of 600 min per day, the PA pattern is presented as a percentage of wear time spent in different intensity-specific categories.

Blood sampling and analysis

In the discovery sample, all sites contributed 850 samples each using random numbers, except for Stockholm, which contributed 825 samples, giving a total of 5075 samples. Discovery data initially included 30,154 subjects, and after the exclusion of subjects with completely missing biomarker data (n = 25079) and completely missing PA variables (n = 188), a total of 4647 subjects remained. Validation data initially included 1111 subjects, and after the exclusion of subjects with completely missing biomarker data (n = 43) and completely missing PA variables (n = 147), a total of n = 921 remained. In total 240 subjects in the discovery data had more than 5% missing proteins, and 11 subjects in validation data had more than 5% missing proteins. These subjects were removed from the data, leaving 4647 and 910 subjects in the two samples, respectively.

Samples were included consecutively starting from the earliest date possible for each site to optimize the potential of prospective studies. Blood samples were bio-banked before analyses.

Samples subsequently underwent proteomic profiling using the Olink Proseek Multiplex Cardiovascular II and III (CVD II and CVD III) panels (Olink Proteomics, Uppsala, Sweden), analysing a total of 183 protein biomarkers. Samples further underwent metabolomic profiling using a high-throughput NMR metabolomics platform (Nightingale) analysing 154 markers. Markers identified in proteomic and metabolomic profiling are reported separately, due to their differing methodology and units of expression (mmol/L for NMR and relative normalized protein expression [NPX] quantification for Olink) that limit comparisons.

Names of the metabolites and proteins are given in Tables 2–4.

Table 2. Markers significantly related to SED. Values are beta coefficients of the unadjusted model, model 1 and model 2.

Marker	Name	Beta unadjusted (SE)	Beta model 1 (SE)	Beta model 2 (SE)
	Amino acids, fatty acids and ketones
Ala	Alanin	0,000757398 (6,72444E-05)	0,000683323 (6,9491E-05)	0,000342272 (8,26063E-05)
Crea	Creatinine	0,000206588 (1,48557E-05)	0,000153736 (1,0683E-05)	4,73282E-05 (8,96818E-06)
FAw6FA	Ratio of omega-6 fatty acids to total fatty acids	−0,035089296 (0,002763002)	−0,024238701(0,002738367)	−0,008927226 (0,003082267)
Gp	Glycoprotein acetyls, mainly a1-acid glycoprotein	0,002082828 (0,000185562)	0,001652642 (0,00018933)	0,000458226 (0,000207641)
Ile	Iso-leucin	0,000257519 (1,63193E-05)	0,000149325 (1,49111E-05)	6,04017E-05 (1,67903E-05)
Lac	Lactate	0,00431094 (0,000380532)	0,003250504 (0,000386103)	0,001642492 (0,000463092)
Leu	Leucine	0,000252418 (1,86517E-05	0,000120051 (1,65722E-05	4,36894E-05 (1,9182E-05)
MUFAFA	Ratio of monounsaturated fatty acids to total fatty acids	0,035991143 (0,002801885)	0,025614896 (0,002781239)	0,010656412(0,003087197)
Phe	Phenylalanine	7,02567E-05 (7,4994E-06)	5,10294E-05 (7,7009E-06)	2,02786E-05 (8,93771E-06)
Val	Valine	0,000591608 (4,26133E-05)	0,000331844 (3,97014E-05)	0,000203482 (4,61558E-05)
PUFAFA	Ratio of polyunsaturated fatty acids to total fatty acids	−0,036229424 (0,003004184)	−0,023644925 (0,00181495)	−0,007859597 (0,003311037)
TGPG	Ratio of triglycerides to phosphoglycerides	0,003697449 (0,000269492)	0,002577079 (0,000269207)	0,00112821(0,00030185)
UnSat	Estimated degree of unsaturation	−0,000598165 (5,92103E-05)	−0,000392529 (5,92365E-05)	−0,000160135 (6,83341E-05)
	Apolipoproteins
ApoA1	Apolipoprotein A-I	−0,003205023 (0,000223272)	−0,00178288 (0,001972295)	−0,001396951 (0,000244295)
ApoBApoA1	ApoB-to-ApoA1 ratio	0,002147062 (0,000173095)	0,001544738 (0,000174665)	0,000851516 (0,000204121)
	Triglycerides in:
IDLTG	IDL	0,000146643 (2,91207E-05)	0,000170783 (3,04712E-05)	7,76484E-05 (3,6111E-05
LVLDLTG	large VLDL	0,001485796 (0,00012601)	0,001059108 (0,000127491)	0,00035735 (0,000144793)
MVLDLTG	medium VLDL	0,002201863 (0,000175888)	0,00155024 (0,000177458)	0,000602896 (0,000200989)
SHDLTG	small HDL	0,000197954 (1,52525E-05)	0,000134137 (1,51856E-05)	5,47849E-05 (1,71227E-05)
SVLDLTG	small VLDL	0,001309077 (0,000104074)	0,000944677 (0,000105604)	0,000408588 (0,000119909)
VLDLTG	VLDL	0,005910212 (0,000484311)	0,004250725 (0,000490816)	0,001631061 (0,000557047)
XLVLDLTG	very large VLDL	0,000440238 (4,01758E-05)	0,000320525 (4,08346E-05)	0,000100027 (4,68506E-05)
XSVLDLTG	very small VLDL	0,000349484 (3,4231E-05)	0,000285587 (3,55203E-05)	0,000129653 (4,10133E-05)
XXLVLDLTG	chylomicrons and extremely large VLDL	0,000170209 (1,68287E-05)	0,000132453 (1,72914E-05)	4,06475E-05 (2,00668E-05)
SerumTG	Total serum TG	0,006288518 (0,000557214)	0,0047758 (0,0005718)	0,001795364 (0,000653029
	Cholesterols
HDL2C	Total cholesterol in HDL2	−0,00595068 (0,000373095	−0,003552335 (0,00034401	−0,002380541 (0,00039767
HDLC	Total cholesterol in HDL	−0,006319701 (0,000398226)	−0,003748321 (0,00036541	−0,002518947 (0,00042423
LHDLC	Total cholesterol in large HDL	−0,0033848 (0,000221823	−0,001936516 (0,0002028)	−0,001168106 (0,00023176)
LHDLCE	Cholesterol esters in large HDL	−0,002575691 (0,000169224)	−0,001473052 (0,000154688	−0,000895955 (0,000177116)
LHDLFC	Free cholesterol in large HDL	−0,000809016 (5,27052E-05)	−0,000463365 (4,81974E-05)	−0,000272034 (5,4795E-05)
LVLDLC	Total cholesterol in large VLDL	0,0005744 (4,99384E-05)	0,000425083 (5,09591E-05	0,000162113 (5,85212E-05)
LVLDLCE	Cholesterol esters in large VLDL	0,000306376 (2,61113E-05)	0,000229598 (2,66584E-05)	9,80094E-05 (3,07937E-05)
LVLDLFC	Cholesterol esters in large VLDL	0,000267983 (2,42264E-05)	0,000195454 (2,46738E-05)	6,40552E-05 (2,82497E-05)
MHDLC	Total cholesterol in medium HDL	−0,001237926 (9,41202E-05)	−0,000881011 (9,38655E-05)	−0,000821985 (0,000113623)
MHDLCE	Cholesterol esters in medium HDL	−0,000962191 (7,41617E-05)	−0,000692668(7,42398E-05)	−0,000648274 (8,9955E-05)
MHDLFC	Free cholesterol in medium HDL	−0,000275742 (2,04341E-05)	−0,000188349 (2,00818E-05)	−0,000173792 (2,42711E-05)
MVLDLC	Total cholesterol in medium VLDL	0,000926256 (8,52754E-05)	0,000725348 (8,77637E-05)	0,000313733 (0,000102088)
MVLDLCE	Cholesterol esters in medium VLDL	0,000459677 (4,73061E-05)	0,000381408 (4,89358E-05)	0,000180353 (5,77761E-05)
MVLDLFC	Free cholesterol in medium VLDL	0,000466661 (3,97015E-05)	0,000344025 (4,05208E-05)	0,00013344 (4,63021E-05)
SVLDLC	Total cholesterol in small VLDL	0,000658636 (8,02733E-05)	0,000562756 (8,34074E-05)	0,000283848 (9,89433E-05)
SVLDLCE	Cholesterol esters in small VLDL	0,000364147 (5,27335E-05)	0,000328136 (5,48881E-05)	0,000181352 (6,57135E-05)
SVLDLFC	Free cholesterol in small VLDL	0,00029447 (2,99862E-05)	0,00023458 (3,09752E-05)	0,000102452 (3,60542E-05)
VLDLC	Total cholesterol in VLDL	0,002411375 (0,000269418)	0,002065524 (0,000279525)	0,000945382 (0,000330269)
XLHDLFC	Free cholesterol in very large HDL	−0,000489711 (3,66678E-05)	−0,000237434 (3,28481E-05)	−0,000107204 (3,76589E-05)
XLVLDLCE	Cholesterol esters in very large VLDL	7,88053E-05 (7,0647E-06)	5,97828E-05 (7,23193E-06)	2,4636E-05 (8,43465E-06)
XLVLDLFC	Free cholesterol in very large VLDL	5,95526E-05 (5,84525E-06)	4,56085E-05 (5,99609E-06)	1,51758E-05 (6,97888E-06)
XLVLDLC	Total cholesterol in very large VLDL	0,000138378 (1,28571E-05)	0,000105409 (1,31899E-05)	3,98348E-05 (1,53459E-05)
XXLVLDLC	Total cholesterol in chylomicrons and extremely large VLDL	4,41523E-05 (4,38242E-06)	3,56332E-05 (4,5269E-06)	1,35055E-05 (5,30692E-06)
XXLVLDLCE	Cholesterol esters in chylomicrons and extremely large VLDL	2,4846E-05 (2,56015E-06)	2,09162E-05 (2,65561E-06)	9,19626E-06 (3,14581E-06)
	Phospholipids in:
LHDLPL	large HDL	−0,002800827 (0,000179897)	−0,001627117 (0,00016450)	−0,00106597 (0,00018976)
LVLDLPL	large VLDL	0,000452934 (3,90739E-05	0,000328975 (3,96998E-05)	0,000113296 (4,52355E-05)
MHDLPL	medium HDL	−0,000912957 (7,70612E-05	−0,000613855 (7,6003E-05)	−0,000623675 (9,21013E-05)
MVLDLPL	medium VLDL	0,00074992 (6,23607E-05)	0,000547846 (6,34244E-05)	0,000219399 (7,24465E-05)
SVLDLPL	small VLDL	0,000493055 (4,73479E-05)	0,000379489 (4,8694E-05)	0,000161985 (5,61787E-05)
XLHDLPL	very large HDL	−0,002054681 (0,000146845)	−0,001028559(0,00013092)	−0,000492023 (0,000149221)
XLVLDLPL	very large VLDL	0,000111759 (1,07427E-05)	8,40237E-05 (1,09902E-05)	2,62132E-05 (1,26947E-05)
	Total lipids in:
LHDLL	large HDL	−0,006342295 (0,000412701)	−0,003626729 (0,00037691)	−0,002280043 (0,00043196)
LVLDLL	large VLDL	0,002513186 (0,000214448)	0,001813235 (0,000217609)	0,000632821 (0,000247801)
MHDLL	medium HDL	−0,002036793 (0,000171113)	−0,001406397 (0,00017000)	−0,001427366 (0,000206204)
MVLDLL	medium VLDL	0,003878215 (0,000320133)	0,002823594 (0,000325469)	0,001136061 (0,000371424
SVLDLL	small VLDL	0,002460763 (0,000222008)	0,001886902 (0,000228097)	0,000854478 (0,000263458)
XLHDLL	very large HDL	−0,003635884 (0,000270317)	−0,001764195 (0,00024160)	−0,00082681 (0,000276791)
XLVLDLL	very large VLDL	0,000690364 (6,34996E-05)	0,000509953 (6,47649E-05)	0,000165993 (7,45401E-05)
XXLVLDLL	chylomicrons and extremely large VLDL	0,000244571 (2,41465E-05)	0,000191598 (2,48328E-05)	6,06016E-05 (2,8848E-05)
	Particle sizes:
HDLD	Mean diameter for HDL particles	−0,005039151 (0,00035135)	−0,002550716 (0,00031293)	−0,001244384 (0,000354833)
LHDLP	Concentration of large HDL particles	−9,87364E-09 (6,43413E-10)	−5,635E-09 (5,86774E-1	−3,56704E-09 (6,73749E-10)
LVLDLP	Concentration of large VLDL particles	4,34405E-11 (3,69497E-12)	3,12621E-11 (3,7477E-12)	1,08648E-11 (4,26369E-12)
MHDLP	Concentration of medium HDL particles	−4,52376E-09 (3,93504E-10)	−3,11244E-09 (3,9139E-10)	−3,23756E-09 (4,74873E-10)
MVLDLP	Concentration of medium VLDL particles	1,17455E-10 (9,61577E-12)	8,49667E-11 (9,76432E-12)	3,40594E-11 (1,11231E-11)
SVLDLP	Concentration of small VLDL particles	1,31344E-10 (1,15039E-11)	9,95346E-11 (1,18104E-11)	4,47564E-11 (1,35769E-11)
VLDLD	Mean diameter for VLDL particles	0,021243423 (0,001507626)	0,014063394 (0,001475238)	0,006080771 (0,001662287)
XLHDLP	Concentration of very large HDL particles	−3,57524E-09 (2,65935E-10)	−1,73154E-09 (2,3756E-10)	−8,16207E-10 (2,72231E-10)
XLVLDLP	Concentration of very large VLDL particles	7,0986E-12 (6,51041E-13)	5,22434E-12 (6,63501E-13)	1,68662E-12 (7,62916E-
XSVLDLP	Concentration of very small VLDL particles	4,40861E-11 (1,0309E-11)	5,3469E-11 (1,08305E-11)	2,656E-11 (1,29434E-11)
XXLVLDLP	Concentration of chylomicrons and extremely large VLDL particles	1,14038E-12 (1,12585E-13)	8,92743E-13 (1,15748E-13)	2,8134E-13 (1,34462E-13)
	Proteins
ADM	Pro-adrenomedullin	0,007446295 (0,000534891)	0,007359451 (0,000551924)	0,003457726 (0,000596551)
CA5A	Carbonic anhydrase 5A	0,011168224 (0,001126814)	0,007207642 (0,001156878)	0,002721892 (0,00136308)
CD4	T-cell surface glycoprotein CD4	0,002691825 (0,000389497)	0,00284425 (0,000415379)	0,001185998 (0,000496927)
CEACAM8	Carcinoembryonic antigen-related cell adhesion molecule 8	0,005953472 (0,000607435)	0,0046655 (0,00063822)	0,001833103 (0,000762638)
CTSZ	Cathepsin Z	0,005115518 (0,000535522)	0,003615147 (0,000554227)	0,001567717 (0,000649252)
EGFR	Epidermal growth factor receptor	−0,000891032 (0,00027802)	−0,000733592(0,00031096)	−0,001065195 (0,000362402)
FABP4	Fatty acid-binding protein, adipocyte	0,00756275 (0,000987665)	0,01119042 (0,000908023)	0,00357126 (0,000939422)
FGF21	Fibroblast growth factor 21	0,019073166 (0,001881712)	0,017273155 (0,001953605)	0,006661382 (0,002235862)
FGF23	Fibroblast growth factor 23	0,004162749 (0,000605142)	0,004631573 (0,000642545)	0,002362016 (0,000765342)
Gal9	Galectin-9	0,003435648 (0,000448696)	0,003281429 (0,000461537)	0,001481964 (0,000531889)
HAOX1	Hydroxyacid oxidase 1	0,015888091 (0,001733985)	0,009674681 (0,001759726)	0,004553619 (0,002101628)
IL1ra	Interleukin-1 receptor antagonist protein	0,010546778 (0,000967533)	0,0097498 (0,000950495)	0,003236482 (0,001030126)
ITGB2	Integrin beta-2	−0,003947141 (0,000491539)	−0,00358516 (0,000514364)	−0,003177349 (0,000619599)
LEP	Leptin	0,012177215 (0,001604368)	0,02013132 (0,001254941)	0,007419023 (0,001241799)
LPL	Lipoprotein lipase	−0,010698792 (0,000630343)	−0,007096573 (0,00057841)	−0,0055346 (0,000703881)
PI3	Elafin	0,007259485 (0,00069993)	0,005636906 (0,000724181)	0,002456884 (0,000860147)
RETN	Resistin	0,004342793 (0,000590814)	0,004665533 (0,000630287)	0,002395942(0,000750952)
SPON2	Spondin-2	0,001584103 (0,000240082)	0,001654546 (0,000253891)	0,000671623 (0,000301951)
TIMP4	Metalloproteinase inhibitor 4	0,001086715 (0,00056827)	0,002920281 (0,000582809)	0,001466391 (0,000693856)
TNFR1	Tumor necrosis factor receptor 1	0,005565843 (0,000452946)	0,004834209 (0,000471281)	0,002065514 (0,000538235)
TNFR2	Tumor necrosis factor receptor 2	0,004747785 (0,000510376)	0,004020958 (0,000527799)	0,001919902 (0,000625728)
TNFRSF11A	Tumor necrosis factor receptor superfamily member 11A	0,005346519 (0,000513226)	0,005409847 (0,000541223)	0,002900348 (0,000620613)
XCL1	Lymphotactin	0,002613806 (0,000766483)	0,003301193 (0,000827666)	0,001935281 (0,000985466)

Open in a new tab

Table 4. Markers significantly related to MVPA. Values are beta coefficients from unadjusted model, model 1, and model 2.

Marker	Name	Beta unadjusted (SE)	Beta model 1 (SE)	Beta model 2 (SE)
	Amino acids, fatty acids and ketones
Ala	Alanine	−0,001997658 (0,000204386)	−0,001811386 (0,00020704)	−0,00085764 (0,000243055)
Gp	Glycoprotein acetyls, mainly a1-acid glycoprotein	−0,006675625 (0,000562613)	−0,006158627 (0,00057301)	−0,002791355 (0,000611732)
Ile	Isoleucine	−0,000312563 (5,00785E-05)	−0,00037182 (4,44112E-05)	−0,000119901 (4,93956E-05)
bOHBut	3-hydroxybutyrate	0,001344012 (0,000506077)	0,001222472 (0,00050869)	0,001257048 (0,000627057)
FAw6FA	Ratio of omega-6 fatty acids to total fatty acids	0,091508589 (0,008403092)	0,085137226 (0,008236262)	0,030175549 (0,009073206)
LAFA	Ratio of 18:2 linoleic acid to total fatty acids	0,082977986 (0,008521324)	0,072476319 (0,008283828)	0,027468151 (0,009518422)
TGPG	Ratio of triglycerides to phosphoglycerides	−0,008027147 (0,000821807)	−0,008207828 (0,00081024)	−0,003131338 (0,00088834)
PUFAFA	Ratio of polyunsaturated fatty acids to total fatty acids	0,093283636 (0,009135865)	0,088563088 (0,008891729)	0,029655329 (0,009747205)
MUFA	Monounsaturated fatty acids	−0,020985929 (0,002461154)	−0,0193762 (0,002534124)	−0,007706282 (0,002885652)
MUFAFA	Ratio of monounsaturated fatty acids to total fatty acids	−0,093028207 (0,00852279)	−0,087726216 (0,00840271)	−0,031206272 (0,009086148)
	Triglycerides in:
HDLTG	HDL	−0,000848775 (0,000113039)	−0,000780756 (0,00011654)	−0,000477207 (0,000137427)
MHDLTG	medium HDL	−0,0003581 (4,49834E-05)	−0,000341671 (4,57326E-05)	−0,000135641 (5,20453E-05)
SHDLTG	small HDL	−0,000419474 (4,64978E-05)	−0,000424427 (4,55902E-05)	−0,000139327 (5,03875E-05)
SLDLTG	small LDL	−0,000185789 (2,70266E-05)	−0,000171566 (2,80061E-05)	−6,62381E-05 (3,25961E-05)
LVLDLTG	large VLDL	−0,003653189 (0,00038336)	−0,00368966 (0,000384635)	−0,001498335 (0,000426318)
MVLDLTG	medium VLDL	−0,004935679 (0,000535763)	−0,005102922 (0,00053405)	−0,00194138 (0,000591614)
SVLDLTG	very small VLDL	−0,002874237 (0,000317085)	−0,00296683 (0,00031701)	−0,001052928 (0,000352857)
VLDLTG	medium VLDL	−0,013780726 (0,001474422)	−0,014057152 (0,00147675)	−0,00537104 (0,001639716)
XLVLDLTG	chylomicrons and extremely large VLDL	−0,001137596 (0,000122125)	−0,001139697 (0,00012276)	−0,000487657 (0,000137958)
XSVLDLTG	x-small VLDL	−0,000872787 (0,000104058)	−0,000852277 (0,00010626)	−0,000272007 (0,000120673)
XXLVLDLTG	chylomicrons and extremely large VLDL	−0,000464492 (5,11182E-05)	−0,000465627 (5,17935E-05)	−0,000218052 (5,90954E-05)
SerumTG	Serum total triglycerides	−0,015931527 (0,001694315)	−0,015926478 (0,00171542)	−0,006221542 (0,001922339)
	Cholesterol and cholesterol esters:
LVLDLC	Total cholesterol in large VLDL	−0,001370512 (0,000151945)	−0,001409557 (0,00015267)	−0,000569627 (0,000172274)
LVLDLCE	Cholesterol esters in large VLDL	−0,000689417 (7,94966E-05)	−0,000725759 (7,97544E-05)	−0,000286853 (9,06312E-05)
LVLDLFC	Free cholesterol in large VLDL	−0,000681152 (7,36567E-05)	−0,000683847 (7,40348E-05)	−0,000282875 (8,31775E-05)
MVLDLC	Total cholesterol in medium VLDL	−0,0021311 (0,000259457)	−0,002239775 (0,00026243)	−0,000880053 (0,00030045)
MVLDLCE	Cholesterol esters in medium VLDL	−0,001029626 (0,00014388)	−0,001108994 (0,00014683)	−0,000435955 (0,000170013)
MVLDLFC	Free cholesterol in medium VLDL	−0,001101776 (0,000120824)	−0,001131103 (0,00012161)	−0,000444413 (0,000136296)
SVLDLFC	Free cholesterol in small VLDL	−0,000668989 (9,1197E-05)	−0,000681991 (9,30445E-05)	−0,000208491 (0,000106087)
VLDLC	Total cholesterol in large VLDL	−0,005571649 (0,000818992)	−0,005808698 (0,00083929)	−0,002029188 (0,000971815)
XLVLDLC	Total cholesterol in very large VLDL	−0,000338969 (3,90982E-05)	−0,000349556 (3,94117E-05)	−0,00015193 (4,51787E-05)
XLVLDLCE	Cholesterol esters in very large VLDL	−0,000183233 (2,14963E-05)	−0,000192334 (2,1626E-05)	−8,22527E-05 (2,48286E-05)
XLVLDLFC	Free cholesterol in very large VLDL	−0,000155752 (1,77621E-05)	−0,000157235 (1,79563E-05)	−6,9732E-05 (2,05488E-05)
XXLVLDLC	Total cholesterol in chylomicrons and extremely large VLDL	−0,0001157 (1,33163E-05)	−0,000118391 (1,35317E-05)	−5,48797E-05 (1,5625E-05)
XXLVLDLCE	Cholesterol esters in chylomicrons and extremely large VLDL	−6,18548E-05 (7,78131E-06)	−6,52311E-05 (7,94284E-06)	−3,05962E-05 (9,2601E-06)
XXLVLDLFC	Free cholesterol in chylomicrons and extremely large VLDL	−5,38781E-05 (5,81818E-06)	−5,31954E-05 (5,88644E-06)	−2,43219E-05 (6,69699E-06)
	Phospholipids in:
LVLDLPL	large VLDL	−0,001115886 (0,000118858)	−0,001127558 (0,00011949)	−0,000455741 (0,000133182)
MVLDLPL	medium VLDL	−0,001687538 (0,000189891)	−0,001753362 (0,00018996)	−0,000663592 (0,000213231)
SVLDLPL	small VLDL	−0,00109916 (0,000144058)	−0,001118583 (0,00014611)	−0,000324217 (0,000165299)
XLVLDLPL	very large VLDL	−0,000294114 (3,26445E-05)	−0,000294821 (3,20144E-05)	−0,00012784 (3,73803E-05)
XXLVLDLPL	chylomicrons and extremely large VLDL	−8,62525E-05 (9,37464E-06)	−8,49089E-05 (9,50438E-06)	−3,89754E-05 (1,08257E-05)
	Total lipids in:
LVLDLL	large VLDL	−0,006139584 (0,000652414)	−0,006226773 (0,00065472)	−0,002523624 (0,000729571)
MVLDLL	medium VLDL	−0,008754832 (0,000974844)	−0,009096542(0,000977114)	−0,003485526 (0,001093228)
SVLDLL	small VLDL	−0,005274101 (0,000675893)	−0,005491108 (0,00068294)	−0,001754483 (0,000775147)
XLVLDLL	very large VLDL	−0,001770861 (0,000193032)	−0,001784267 (0,00019406)	−0,000767744 (0,000219485)
XXLVLDLL	chylomicrons and extremely large VLDL	−0,000666425 (7,33475E-05)	−0,000668905 (7,4405E-05)	−0,000311885 (8,49529E-05)
	Particle size:
LVLDLP	Concentration of large VLDL particles	−1,06107E-10 (1,12414E-11)	−1,07565E-10 (1,12881E-11)	−4,35723E-11 (1,25532E-11)
MVLDLP	Concentration of medium VLDL particles	−2,64211E-10 (2,92838E-11)	−2,74424E-10 (2,93226E-11)	−1,04815E-10 (3,27391E-11)
SVLDLP	Concentration of small VLDL particles	−2,82217E-10 (3,50296E-11)	−2,93511E-10 (3,52522E-11)	−9,5601E-11 (3,99468E-11)
VLDLD	Mean diameter for VLDL particles	−0,04492516 (0,004599996)	−0,04635674 (0,004445012)	−0,018100394 (0,004892523)
XLVLDLP	Concentration of very large VLDL particles	−1,82256E-11 (1,9791E-12)	−1,83399E-11 (1,98932E-12)	−7,86986E-12 (2,24644E-12)
XXLVLDLP	Concentration of chylomicrons and extremely large VLDL particles	−3,10609E-12 (3,4199E-13)	−3,11753E-12 (3,46775E-13)	−1,45393E-12 (3,95971E-13)
	Proteins
CEACAM8	Carcinoembryonic antigen-related cell adhesion molecule 8	−0,013298882 (0,001847549)	−0,014556686 (0,00191405)	−0,009631684 (0,002241865)
CHI3L1	Chitinase-3-like protein 1	−0,022971991 (0,003384647)	−0,01859018 (0,003508857)	−0,008647311 (0,004045689)
FGF21	Fibroblast growth factor 21	−0,070053334 (0,005694919)	−0,061600316 (0,00590028)	−0,028249103 (0,006571312)
GDF15	Growth/differentiation factor 15	−0,017464513 (0,002007169)	−0,013460106 (0,00194742)	−0,007778152 (0,002181139)
IL6	Interleukin-6	−0,033176711 (0,003011543)	−0,02826306 (0,003056445)	−0,016723585 (0,003485255)
ITGB2	Integrin beta-2	0,010231473 (0,001492903)	0,013163923 (0,001536924)	0,008061274 (0,00181918)
LEP	Leptin	−0,071021878 (0,00482695)	−0,052075226 (0,00368386)	−0,019940734 (0,003646182)
LOX1	Lipoxygenase-1	−0,018245904 (0,001960759)	−0,017345066 (0,00200973)	−0,012978217 (0,00237881)
LPL	Lipoproteinlinase	0,012131768 (0,001938325)	0,014898467 (0,001785016)	0,004199866 (0,00206001)
MMP7	Matrilysin	−0,013531344 (0,001502548)	−0,011039544 (0,00155021)	−0,007326018 (0,00180792)
MMP9	Matrix metalloproteinase-9	−0,022860288 (0,003071323)	−0,023206698 (0,00316118)	−0,0165899 (0,003705674)
MMP12	Matrix metalloproteinase-12	−0,018847613 (0,002510292)	−0,014898427 (0,00250909)	−0,009747345 (0,003043283)
PGF	Placenta growth factor	−0,006094464 (0,001096434)	−0,006663989 (0,00110869)	−0,003465778 (0,00126799)
PGLYRP1	Peptidoglycan recognition protein 1	−0,010924482 (0,001902916)	−0,011527805 (0,00200254)	−0,005172609 (0,002333144)
PI3	Elafin	−0,015131332 (0,002144302)	−0,018128302 (0,00218204)	−0,009887833 (0,002540709)
PON3	Paraoxonase	0,021626673 (0,002758333)	0,018190876 (0,00261885)	0,006281143 (0,002927796)
PRSS8	Prostasin	−0,013653318 (0,001724929)	−0,012535991 (0,0016866)	−0,006068555 (0,001883755)
PRTN3	Myeloblastin	−0,009462432 (0,001856551)	−0,009624614 (0,00197347)	−0,004607934 (0,002316056)
RARRES2	Retinoic acid receptor responder protein 2	−0,015754495 (0,001468753)	−0,013622475 (0,00149527)	−0,005501454 (0,001632155)
REN	Renin	−0,018404205 (0,003362319)	−0,018367703 (0,00334094)	−0,011326584 (0,003892206)
RETN	Resistin	−0,01228823 (0,001793415)	−0,01243717 (0,001894795)	−0,006682177 (0,002205628)
TFF3	Trefoil factor 3	−0,011514194 (0,001611293)	−0,010814729 (0,00165611)	−0,004751389 (0,001933227)
TNFSF13B	Tumor necrosis factor ligand superfamily member 13B	−0,010879679 (0,001430291)	−0,008632136 (0,00142081)	−0,004814957 (0,00173084)
tPA	Tissue-type plasminogen activator	−0,024025346 (0,003029865)	−0,021732233 (0,00301566)	−0,010129752 (0,003515359)
TRAILR2	TNF-related apoptosis-inducing ligand receptor 2	−0,012419908 (0,001450857)	−0,010752076 (0,00147532)	−0,004320619 (0,001617029)
UPAR	Urokinase plasminogen activator surface receptor	−0,012919005 (0,001428267)	−0,010474066 (0,00140669)	−0,006006859 (0,001647695)

Open in a new tab

Confounders

Sex was recorded binary as male and female using self-report. Age at assessment was recorded in years with one decimal. Kidney function was assessed as estimated glomerular filtration rate (eGFR) was calculated according to the Cockcroft-Gault equation and based on plasma creatinine. Diabetes was classified as No diabetes (fasting p-glucose <6,1 mmol/l and HbA1c < 42 mmol/mol), IFG (fasting p-glucose ≥6,1 and < 7,0 mmol/l and/or HbA1c > 42 and < 48 mmol/mol), or Diabetes (Diabetes stated by study participant at medical history interview or in subject questionnaire, or fasting p-glucose =>7,0 mmol/l or HbA1c =>48 mmol/mol).

Self-reported data on education (university degree vs no university degree) and smoking exposure (as pack-years) were collected. Diet was not standardized or controlled, but blood samples were drawn in a fasted state and intakes of carbohydrates, fats, and proteins (in grams per day, with calculations made from the MiniMeal-Q questionnaire) were used as potential confounders, together with measured body mass (in kilograms), and waist circumference (in centimetres).

Statistics

Subject characteristics are reported as median (IQR) and proportions (IQR) for continuous and categorical variables, respectively.

The aim was to explore any associations between biomarkers and each of the four PA classes. The analyses were divided into three steps.

In the first step, unadjusted (crude) analyses were performed, with the biomarker as the dependent variable and each PA class as independent. Biomarker concentrations were transformed to the log2 scale.

In the second step, the potential confounders sex, age, eGFR, and Diabetes were added to the model. For proteomics, CVD II, and CVD III biomarkers plate number was also included as a confounder. Normal linear regression models were estimated and each p-value for the PA variables was stored. The p-values are then sorted and following the FDR procedure (Benjamini-Hochberg correction) converted into q-values. Q-values can in principle be interpreted similarly to p-values. No imputation was performed for missing data or outliers. This procedure was performed on the discovery dataset. A cut-off of 5% for the q-values was used and for the biomarkers where the q-values fell below 5%, the model was re-estimated on the validation dataset. For the validation dataset, the p-values were used to establish significance (5% level used)

The biomarkers that were significant for the validation set in the second step were included in the third step.

In the third step, the same model as in Step 2 was used, but more potential confounders were included (education, smoking status, body mass, waist circumference, intake of carbohydrates, fats, and proteins). Also, for SED, prolSED, and LIPA, the variable MVPA was included, and for MVPA, the variable SED was included. The model for step 2 was estimated for both discovery and validation datasets and used p-values to establish significance. Beta coefficients from the linear regressions are presented when associations were significant at step 1 for both validation and discovery datasets and in model 2 in the discovery dataset.

The included variables all have their theoretical significance, but unadjusted beta coefficients along with those from the two regression models were applied for transparency.

We explored interactions regarding sex and waist circumference (dichotomized at 88 cm and 102 cm for females and males, respectively). Directions of associations are given as (+) for positive and (-) for negative associations, respectively. Venn diagrams were generated according to Hulsen et al [26].

To separate biomarkers with a more solid (“robust”) and more discrete (“separate”) relation to PA beyond the application of two-sample validation and two different regression models, “robust relations” were considered for markers inversely and significantly related to both MVPA and SED/prolSED. “Separate relations” were considered when a marker was significantly related to MVPA or SED/prolSED and with some marginal not related to another (p-value>0.1). In Figs 1 and 2, biomarkers with separate but not robust relations are in parentheses. Potential non-linearity for the association between significant biomarkers and PA classes MVPA and SED was explored using a second-degree term to model 2.

Results

Participant characteristics are shown in Table 1. Participants in the discovery and validation samples differed in some respects. Time spent in SED and in prolSED were higher, and LIPA lower, in the validation sample. A lower proportion of participants in the validation sample had a university degree, while sex distribution and age did not differ.

Table 1. Participant characteristics in the two samples.

Variable	Total (N = 5557)	Discovery (N = 4647)	Validation (N = 910)
	Median (IQR) or n (%)	Median (IQR) or n (%)	Median (IQR) or n (%)
SED (% of wear time)	56 (48-64)	54 (46-61)	69.7 (64.6-74.4)
prolSED (% of wear time)	186 (118-277)	164 (107.5-232)	362.1 (283.2-453.6)
LIPA (% of wear time)	38 (30-45)	40 (33-46)	25.1 (20.9-29.4)
MVPA (% of wear time)	6 (4-8)	6 (4-8)	4.8 (3.3-6.6)
Daily wear time (min)	875 (824-924.6)	876 (825-925)	869 (819-921.9)
Sex
Female, n (%)	2807 (50.5%)	2349 (50.5%)	458 (50.3%)
Male, n (%)	2750 (49.5%)	2298 (49.5%)	452 (49.7%)
Age (yrs)	57.5 (53.9-61.3)	57.5 (53.9-61.3)	57.5 (53.8-61.7)
Education
University degree, n (%)	2478 (44.8%)	2132 (46.0%)	346 (38.2%)
eGFR	118.1 (100.9-138.3)	116.9 (100-137)	123.8 (105.3-144.8)
Diabetes
Normoglycemia, n (%)	4371 (78.7%)	3654 (78.6%)	717 (79.2%)
IFG, n (%)	813 (14.6%)	685 (14.7%)	128 (14.1%)
Diabetes, n (%)	367 (6.6%)	307 (6.6%)	60 (6.6%)
Smoking
Pack-year	12.8 (5.3-23)	12.3 (5.3-22.5)	15 (6-25.2)
Body mass (kg)	65 (56-74)	65 (56-74)	64 (55-73.2)
Waist circumference (cm)	94 (85-102)	93 (85-102)	95 (86-102)
CHO intake (g/d)	173.6 (128.7-229.4)	170.1 (126.2-223)	194.6 (140.1-263.7)
Fat intake (g/d)	64 (47.9-84.3)	63.7 (47.8-83.7)	65.2 (49.5-88)
Protein intake (g/d)	65.4 (50.9-82.9)	64.6 (50.4-81.9)	70.2 (54.3-88.9)

Open in a new tab

Of all included markers, 193 were significantly related to SED in model 1 in the validation sample. Corresponding numbers were 176, 155, and 215 for prolSED, LIPA, and MVPA, respectively. A total of eighty-five metabolomic and forty-three proteomic biomarkers were significantly related to one or more PA classes in model 1 in both validation and discovery samples and in model 2 for the discovery sample (Figs 1 and 2). LIPA and SED almost completely shared related metabolomic and proteomic markers. LIPA shared all of its related markers with SED, and beta coefficients were close to opposite each other, which may be explained by the strong negative interrelation (rho = −0.94) between SED and LIPA. (S1 and S2 Figs) Arbitrarily, LIPA was excluded from the analyses for clarity.

Metabolomic markers

Of the eighty-three metabolic markers, eight were related to solely to MVPA, and thirteen to SED, with the remaining being related to two or three PA classes (Fig 2). MVPA and SED, the two most differing PA classes intensity-wise, shared forty-three metabolomic markers, mainly VLDLs. ProlSED and SED shared19 metabolomic markers, primarily HDLs. Of all eighty-three markers, only eight were not related to SED. Beta coefficients were nominally low (S3 Fig, Tables 2–4]. We found the nominally strongest associations between MVPA and LAFA (+), FAw6FA (+), MUFAFA (-), and PUFAFA (+). Robust relations (markers related to both MVPA and SED/prolSED) were found for a variety of VLDLs, serum TG, isoleucine, alanine, Gp, and FAw6FA (negative relations to MVPA and positive relations to SED/prolSED). Most separate relations were between HDLs and SED/prolSED, with a mix of markers separately related to MVPA.

Table 3. Markers significantly related to prolSED. Values are beta coefficients from unadjusted model, model 1, and model 2.

Marker	Name	Beta unadjusted (SE)	Beta model 1 (SE)	Beta model 2 (SE)
	Amino acids, fatty acids and ketones
Ala	Alanin	5,75413E-05 (7,13547E-06)	4,72956E-05 (7,61497E-06)	1,70196E-05 (7,97096E-06)
Crea	Creatinine	1,62948E-05 (1,57815E-06)	1,2152E-05 (1,16368E-06)	3,29834E-06 (8,65408E-07)
Lac	Lactate	0,000416154 (4,02922E-05)	0,000304941 (4,23551E-05)	0,000187495 (4,46365E-05)
Phe	Phenylalanine	6,80737E-06 (7,93749E-07)	4,71088E-06 (8,46153E-07)	2,38666E-06 (8,61623E-07)
	Apolipoprotein:
ApoA1	Apolipoprotein A-I	−0,000248806 (2,37354E-05)	−0,000129877 (2,24717E-05)	−6,98032E-05 (2,35894E-05)
	Cholesterol and cholesterol esters:
HDL2C	Total cholesterol in HDL2	−0,000453316 (3,97217E-05)	−0,000249333 (3,51116E-05)	−0,000106809 (3,84093E-05)
HDLC	Total cholesterol in HDL	−0,00048184 (4,23939E-05)	−0,000263807 (3,99499E-05)	−0,000113594 (4,09741E-05)
LHDLC	Total cholesterol in large HDL	−0,000256243 (2,3605E-05)	−0,000130745 (2,22266E-05)	−4,54449E-05 (2,23754E-05)
LHDLCE	Cholesterol esters in large HDL	−0,000195458 (1,80066E-05)	−0,000100041 (1,58769E-05)	−3,5466E-05 (1,70997E-05)
MHDLC	Total cholesterol in medium HDL	−0,000103954 (9,98879E-06)	−7,56782E-05 (1,02903E-05)	−5,46403E-05 (1,0974E-05)
MHDLCE	Cholesterol esters in medium HDL	−8,15207E-05 (7,86909E-06)	−6,01198E-05 (8,15666E-06)	−4,38739E-05 (8,68748E-06)
MHDLFC	Free cholesterol in medium HDL	−2,24368E-05 (2,16985E-06)	−1,55625E-05 (2,202E-06)	−1,07764E-05 (2,34454E-06)
SVLDLCE	Cholesterol esters in small VLDL	3,03649E-05 (5,58258E-06)	2,66092E-05 (6,05147E-06)	1,28808E-05 (6,33734E-06)
	Phospholipids in:
LHDLPL	large HDL	−0,000213246 (1,91468E-05)	−0,000112972 (1,80184E-05)	−4,53817E-05 (1,83251E-05)
MHDLPL	medium HDL	−7,70391E-05 (8,1726E-06)	−5,42839E-05 (7,80536E-06)	−4,16509E-05 (8,89338E-06)
	Total lipids in:
LHDLL	large HDL	−0,000481475 (4,39193E-05)	−0,00024765 (4,12621E-05)	−9,21692E-05 (4,17084E-05)
MHDLL	medium HDL	−0,000174058 (1,81454E-05)	−0,000125655 (1,87233E-05)	−9,7655E-05 (1,99112E-05)
	Particle sizes:
LHDLP	Concentration of large HDL particles	−7,50161E-10 (6,84696E-11)	−3,85555E-10 (6,42494E-11)	−1,45017E-10 (6,50543E-11)
MHDLP	Concentration of medium HDL particles	−3,89794E-10 (4,17176E-11)	−2,82222E-10 (4,29098E-11)	−2,24501E-10 (4,58501E-11)
	Proteins:
ADM	Pro-adrenomedullin	0,000647433 (5,67638E-05)	0,000593559 (6,08141E-05)	0,000291703 (5,85873E-05)
CA5A	Carbonic anhydrase 5A, mitochondrial	0,000957928 (0,00011939)	0,00061161 (0,00013099)	0,000289831 (0,000133788)
CTSL1	Cathepsin L1	0,000253652 (4,62612E-05)	0,000160632 (4,65883E-05)	0,000116998 (5,24017E-05)
EGFR	Epidermal growth factor receptor	−8,85146E-05 (2,94056E-05)	−6,90074E-05 (3,4407E-05)	−8,98081E-05 (3,55784E-05)
FABP4	Fatty acid-binding protein, adipocyte	0,000660856 (0,000104561)	0,000896701 (9,49827E-05)	0,00035056 (9,22136E-05)
FGF23	Fibroblast growth factor 23	0,000328836 (6,40726E-05)	0,000365183 (7,24377E-05)	0,000190799 (7,51372E-05)
Gal9	Galectin-9	0,000349071 (4,74661E-05)	0,000296887 (5,07129E-05)	0,000171524 (5,12729E-05)
LEP	Leptin	0,000733019 (0,000170036)	0,001361011 (0,000132829)	0,000313281 (0,000122112)
PI3	Elafin	0,000635571 (7,60459E-05)	0,000473655 (8,26773E-05)	0,000237512 (8,65987E-05)
SPON2	Spondin-2	0,000173148 (2,53873E-05)	0,000170664 (2,83772E-05)	0,00010115 (2,96257E-05)
TNFR1	Tumor necrosis factor receptor 1	0,00050181 (4,80113E-05)	0,000406556 (5,12016E-05)	0,000189236 (5,19002E-05)
TNFRSF11A	Tumor necrosis factor receptor superfamily member 11A	0,000550731 (5,42958E-05)	0,000540855 (5,624E-05)	0,000325564 (6,0891E-05)
TRAILR2	TNF-related apoptosis-inducing ligand receptor 2	0,000336565 (5,06483E-05)	0,00026247 (5,47382E-05)	0,000125594 (5,29428E-05)

Open in a new tab

ProlSED was related only to cholesterols in HDL (-) and MVPA was only related to cholesterols in VLDL (-). SED (+) and MVPA (-) but not prolSED were related to triglycerides in lipoproteins. SED and prolSED were negatively related to phospholipids in HDLs and positively related to phospholipids in VLDL. MVPA was only related to phospholipids in VLDLs (-). SED was related to total lipid content in HDL (-) and VLDL (+), with reversed associations for MVPA. MVPA was only related to total lipid content in VLDLs (-). Regarding particle size and concentration, SED was positively related to the size and concentration of VLDL but negatively related to HDLs. LIPA and MVPA were exclusively related to the size and concentration of VLDLs (-).

Proteomic markers

Of the 43 related proteomic markers, a considerable proportion (18 markers) were uniquely related to MVPA. Only eight markers were jointly related to MVPA and SED and/or prolSED which is significantly less than for the metabolic markers (χ2 = 22.5.0, df = 1, p < 0.001). ProlSED shared ten of thirteen markers with SED. Relations were nominally stronger for proteomic markers, compared to metabolomic markers (S4 Fig, Tables 2–4].

We found the nominally strongest associations for LEP (-), FGF21 (-), IL6 (-), MMP9 (-), and LOX1 (-), all associations with MVPA. (Tables 1 and 2) Proteomic markers significantly related to the PA classes were only weakly inter-related. Robust negative relations to MVPA (and positive to SED/prolSED) were found for TRAILR2, LEP, PI3, CEACAM8, FGF21, and RETN. Opposite relations were found for ITGB2 and LPL, with positive relations to MVPA and negative to SED. Separate relations were noted for most significant markers, with exceptions for Il1ra, ADM, SPON2, TNFR1, MMP7, PON3, and tPA.

Non-linear associations.

Potential non-linearity associations between significant biomarkers and PA classes were explored using a second-degree term. For MVPA, only three of the 103 significant markers showed a significant second-degree term, which is close to the expected number of findings. For SED, forty-seven of the 127 significant biomarkers showed a significant second-degree association, which is more than expected. However, explanatory values of the non-linear associations were not significantly higher than linear, and visual inspection did not reveal any meaningful cut-off points for the associations.

Interaction between PA class and sex.

The interaction between PA class and sex in the significant associations was further investigated.

For SED, associations were stronger in men for the metabolomic markers FAw6FA, Ile, and Leu and a group of VLDL-related markers, while HDL-related markers generally showed a stronger relation to SED in women. Associations between SED and proteomic markers RETN and SPON2 were both strongly related in men. For SED, relations were in the same direction for both sexes. For MVPA, five HDL-related markers and proteomic markers ITGB2 were all strongly related in females. Interestingly, for MVPA, relations were opposite for men and women except for ITGB2. (S2 and S3 Tables).

Interaction between PA class and waist circumference.

For SED, associations to VLDL-related markers and UnSat were significantly stronger among participants with low waist circumference, while creatinine, lactate, phenylalanine, and proteomic markers FAPB4, HAOX1, PARP1, SPON2, TIMP4, and TNFRSF11 showed stronger associations to SED in participants with high waist circumference. All relations between SED and biomarkers were in the same direction for Low and High WC.

ApoB-to-ApoA1 and three VLDL-related markers were strongly related to MVPA in participants with low waist circumference, while UnSat was strongly related to MVPA among individuals with high waist circumference. Two proteomic markers (CTSZ and HAOX1) showed significantly stronger associations with MVPA in participants with high waist circumference, while LEP was strongly related to MVPA in participants with low waist circumference. Except for UnSat and CTSZ, relations between biomarkers and MVPA were in the opposite direction for Low and High WC (S4 and S5 Tables).

Discussion

To our knowledge, this is the first time that the association between accelerometer-based measures of PA at different intensities and plasma metabolomic and proteomic biomarkers has been investigated in two population-based samples. Using two independent samples for discovery and validation, we identified significant associations between several metabolomic and proteomic markers and the investigated PA classes.

We found associations between PA and eighty-five metabolomic species identified by metabolomic profiling and forty-three cardiovascular biomarkers from proteomic profiling in these two independent population-based cohorts. Metabolic markers were to a larger degree related to several PA classes while proteomic markers were significantly more often uniquely related to one PA class. MVPA, i.e., PA of higher intensity showed the strongest association with markers, both metabolic and proteomic. Although speculative, this may indicate if repeated in experimental trials, that this PA class may have a stronger association with CVD risk compared to SED or prolSED.

Prolonged SED was only marginally different from SED for metabolic markers but exhibited unique associations with several proteomic biomarkers. LIPA almost totally shared markers with SED. This makes it questionable if one of these two PA classes adds any meaningful information in risk assessment in middle-aged adults in the presence of the other.

It may be that SED lacks additional information to LIPA and that the relations between PA and biomarkers may be intensity dependent. In such interpretation, PA at higher intensities (e.g., MVPA) has the strongest relationship to biomarkers, and LIPA has a weaker relationship. This is supported by the findings that the strongest relationships in both metabolic and proteomic markers were found for MVPA.

As proteomic markers, rather than metabolic markers, distinguished the SED and MVPA PA classes, they may be central to explaining varying associations between different PA classes and CVD health. In longitudinal epidemiological studies, SED has been related to CVD mortality primarily in individuals lacking MVPA [27]. Proteomic markers uniquely related to MVPA may therefore be of interest when identifying the reason behind this. In a recent paper by Siegbahn et al [28], a novel CVD risk panel consisting of twenty-one biomarkers was validated. Among the most prognostic markers were MMP-12, UPAR, REN, and TFF3 (all related to MVPA in the present data) as well as ADM and FGF23 (both related to SED). This could indicate that these markers are of special importance to explain the above findings.

Metabolomic profiling

LAFA (or linoleic fatty acids) [29], FAw6FA [30]and PUFAFA [31] are related to cardio-protective effects via different pathways and were found to be related to MVPA. Their positive associations with MVPA may be linked to the observed associations between MVPA and CVD. MUFAFA has been linked to an increased risk for CVD [32].

As significantly more metabolomic markers were jointly related to SED and prolSED, the proposed unique association between prolSED and vascular health [33], may be related to metabolic factors, rather than proteomic. Experimental work could be applied to verify this and examine any causal links.

In early work, Kujala et al [34] showed similar associations to ours between PA and metabolomic CVD markers. The analyses were partly conducted in twin pairs and indicate that levels of these biomarkers are, to a limited degree, genetically dependent since they differ between active and inactive twins. Interestingly, this included a1-acid glycoprotein, an inflammatory marker related to CVD risk. In the present study, we observed associations between glycoprotein acetyls and SED/LIPA and MVPA. Glycoprotein acetyls are not related to SED in young individuals [35], which could play a role in the differing associations between time spent on SED and metabolic risk markers among young age and adults.

SED/LIPA and prolSED, alone and in combination were negatively related to medium, large, and very large HDLs. MVPA was related to a limited number of HDL types. The majority of the metabolomic markers were jointly related to both MVPA and SED/LIPA. Those markers were predominantly LDLs or VLDLs. If causal, these findings may be of importance when prescribing PA to patients with blood lipid disorders.

In a sample of adults at high risk for type 2 diabetes, Henson et al. [36] showed negative associations between medium and small HDLs and PA corresponding to SED and positive associations between PA corresponding to MVPA and large, medium, and small-sized HDL. Further, they showed positive associations between LDLs and PA corresponding to SED, but no association to higher intensities. In their study, the intensity corresponding to SED was negatively related to ApoA1 but unrelated to the ApoB-to-ApoA1 ratio. We could not fully confirm these findings, as we found associations between SED and both ApoA1 and the ApoB-to-ApoA1 ratio in this mixed population. Based on data from younger men (27 yrs.), Vaara et al [37] reported strong, significant associations between ApoB-to-ApoB1-ratio and both SED and MVPA, indicating that these associations may be stronger in younger or more healthy populations. Differences between studies may be dependent on the PA definition and partly dependent on the different models used and the populations studied.

Proteomic profiling

We found the strongest robust associations between MVPA and LEP (-), FGF21 (-), IL6 (-), MMP9 (-) and LOX1 (-). Leptin is related to cardiovascular dysfunction [38,39] and previous research has shown that LEP may decrease after exercise training [40]. IL6, MMP9 [41], and LOX1 [42] have been linked to inflammation, insulin resistance, or arteriosclerosis. Both MMP9 [43] and LOX1 [44] are further linked to a higher risk of ruptures of atherosclerotic plaques. Previous research has linked LOX1 to endothelial oxidated-LDL uptake, and thereby endothelial dysfunction.

A group of robustly related markers has been linked to inflammation or immune response markers, including TNFRSF11A [45], and Gal9 [46]. Inflammatory markers separately related to MVPA were IL-6 [47], MMP9 [48], MMP12 [49], CHI3L1 [50], TNFSF13B [50], TFF3 [51], and PGF [52]. Further, FGF23 [53], and GDF-15 [54], were robustly related to PA and have been linked to metabolic disorders such as diabetes or obesity. Renin, GDF15 [55], PRSS8 [56], RARRES2 [57], PGLYRP1 [58], LOX1 [59], and UPAR [60], all robustly related to PA, have been related to cardiovascular or kidney disease via different pathways.

In a recent paper, Lind et al. [61] found forty-two proteomic markers linked to incident CVD, of which eleven were related to MVPA in the present study and nine others to SED, with only two (FGF-21 and RETN) shared across PA classes. In the paper by Lind et al., forty-nine proteins were linked to MI, of which ten were related to SED and ten to MVPA, with only FGF-21 being related to both PA classes. Further, in a study of over 800 Olink proteins [62], 19 different markers were found to be predictive of future myocardial infarction. Among these were TRAILR2. MMP-12, GDF-15, CHI3L1, and IL-6 all were related to MVPA in the present analyses, and TRAILR2 was also related to prolSED.

However, there may be site-specific associations between PA classes and arteriosclerosis, with different markers indicating specific mechanisms. In previous work on SCAPIS data, time spent in SED was found to be related to a lower prevalence of carotid atherosclerosis but a risk factor for coronary stenosis [63]. Based on results from the present investigation, biomarkers separately related to SED may therefore be candidates for explaining these opposite associations. For example, TNFR2 which is positively related to SED in the present paper, is considered a risk factor for myocardial infarction [64], but earlier studies have identified TNFR2 signaling as protective after ischemic strokes [65,66]. Another example is LPL, which is negatively related to SED and considered to be protective for coronary artery disease [67], but some genetic variants of the LPL gene have been related to a higher risk of carotid stenosis [68].

While relations to PA classes might share physiological mechanisms, there is most probably a large number of different pathways involved in the identified relationships between PA classes and biomarkers. For example, regarding metabolomic markers, PA is known to increase lipid metabolism and could thereby affect lipid profile via LPL activation or (up-)regulation of LDL receptors. Another example regarding proteomic markers related to inflammation (such as IL-6) may be affected by PA by alterations to the autonomic system, AMPK activation or changes in for example visceral fat deposits.

Interactions

Associations between SED or MVPA and HDL-related markers were stronger in females, while associations with VLDL markers were stronger in males. Sex differences in associations between PA and some proteomic markers were also noted. Several of these proteomic markers are related to metabolic dysregulation and inflammation (e.g., RETN, ITGB2, LEP). Previous investigations [69,70] have noted sex-specific associations between exercise and vascular health, which may, to some degree, be explained by the different associations between markers and activity in the present study. Also, based on SCAPIS data, we have previously shown a borderline interaction between SED/LIPA and sex for coronary atherosclerosis, and the identified markers may be key candidates to explain these differences. Interactions were also noted between activity and waist circumference. For the proteomic markers, all associations were stronger in participants with higher waist circumference (>102 cm for males and >88 cm for females). Limited data is available on the interaction between obesity and PA for the development of cardiometabolic disease, but the present data indicates that associations may be stronger in obese than in normal-weight individuals.

Strengths and limitations

Strengths of our study include the objective measurement of PA and the examination of lipids and cardiovascular plasma protein biomarkers in association with various characteristics across a range of PA intensities. Moreover, the study includes the discovery and replication of associations between PA and metabolic and cardiovascular protein biomarkers in two separate large, well-characterized population-based cohorts. Further, the study is based on a substantial number of biomarkers analyzed simultaneously using two sensitive, highly specific methods without cross-reactivity and with good agreement compared to other assays. Several limitations balance the strengths of the study. Due to the cross-sectional design, it is not possible to determine the causal direction of the associations. Having or being at high risk for cardiovascular disease could influence the amount of PA. [71] Although accelerometers provide a more robust assessment of PA compared to self-report measures, they are not without limitations. They rely on categorizing the strength of movement (acceleration) rather than directly distinguishing between postures, domains or types of PA. Further, intensities used in this study, and most other studies based on accelerometers, are absolute in the sense that they are the same for individuals with differing work tolerance, or fitness level. As most physiological responses to PA are at least partly related to relative intensity (as a proportion of the individual fitness level), this may constitute a source of misclassification, potentially leading to regression dilution and unidentified associations dependent on fitness levels. Another limitation is the lack of information on ongoing therapies, such as statins or lipid uptake inhibitors. Many of the included variables in the models are subjected to potential residual confounding, i.e.,that a confounding variable is not fully adjusted for, due to imperfect measurement or unwarranted categorization. We have not been able to estimate the potential impact of residual confounding. As in all larger data collections, not all invited subjects choose to participate, leading to potential selection biases. This can be of major importance when assessing prevalence or estimating reference values. While it still may constitute a potential limitation, it is most probably to a lesser degree when assessing relations within a population such as in these analyses.

The SCAPIS cohort is based on middle-aged Swedish men and women. Although this population is heterogenous in several aspects, the generalizability to other populations is unknown.

Conclusions

In conclusion, our data suggest differences in the associations between physical activity of various intensity and metabolomic or proteomic biomarkers, in a mixed middle-aged population. These results confirm and expand on previous knowledge, by showing that MVPA and lower-intensity activity have (partly) different relations to CVD risk.

Supporting information

S1 Table. Lists of non-significant biomarkers.

(DOCX)

pone.0325720.s001.docx^{(36.8KB, docx)}

S2 Table. Interactions between SED and sex, using model 2 plus an interaction term.

(DOCX)

pone.0325720.s002.docx^{(17.6KB, docx)}

S3 Table. Interactions between MVPA and sex, using model 2 plus an interaction term.

(DOCX)

pone.0325720.s003.docx^{(16KB, docx)}

S4 Table. Interactions between SED and waist circumference (dichotomized at 102 cm and 88 cm for male and females, respectively) using model 2 plus an interaction term.

(DOCX)

pone.0325720.s004.docx^{(17.5KB, docx)}

S5 Table. Interactions between MVPA and waist circumference (dichotomized at 102 cm and 88 cm for male and females, respectively) using model 2 plus an interaction term.

(DOCX)

pone.0325720.s005.docx^{(16.1KB, docx)}

S1 Fig. Wenn-diagram showing number of metabolomic biomarkers related to SED, prolSED, LIPA and MVPA.

SED encapsuled all biomarkers related to LIPA.

(TIF)

pone.0325720.s006.tif^{(52.2KB, tif)}

S2 Fig. Wenn-diagram showing number of proteomic biomarkers related to SED, prolSED, LIPA and MVPA.

SED encapsuled all biomarkers related to LIPA.

(TIF)

pone.0325720.s007.tif^{(65.2KB, tif)}

S3 Fig. Heat map of relations between metabolomic markers and the PA classes.

(TIF)

pone.0325720.s008.tif^{(172.1KB, tif)}

S4 Fig. Heat map of relations between proteomic markers and the PA classes.

(TIF)

pone.0325720.s009.tif^{(123.5KB, tif)}

S1 Checklist. STROBE-checklist-v4_B4_241209.

(DOCX)

pone.0325720.s010.docx^{(51.2KB, docx)}

Acknowledgments

We would like to acknowledge the help of Biobank Sweden and the local biobank facilities for their services in handling of biological samples and biobanking.

Data Availability

The General Data Protection Regulation (EU 2016/679) classifies de-identified versions of sensitive data that are sufficiently detailed to allow for re-identification as sensitive personal information. According to Swedish law (Law 2003:460 for ethical review of research involving humans), ethical permission is required to process such data. In accordance with Swedish legislation, the data can and will be made available to researchers who meet the criteria for access to confidential data, which includes obtaining their own ethics approval from the Swedish Ethical Review Authority (email: registrator@etikprovning.se; website: https://etikprovningsmyndigheten.se). Data applications can then be made by contacting SCAPIS (email: scapis@scapis.org; website: https://www.scapis.org/data-access/).

Funding Statement

ÖE was funded by Livförsäkringsbolaget Skandia, Risk&Hälsa in the form of salary. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1.Wood G, Taylor E, Ng V, Murrell A, Patil A, van der Touw T, et al. Estimating the Effect of Aerobic Exercise Training on Novel Lipid Biomarkers: A Systematic Review and Multivariate Meta-Analysis of Randomized Controlled Trials. Sports Med. 2023;53(4):871–86. doi: 10.1007/s40279-023-01817-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Kelly RS, Kelly MP, Kelly P. Metabolomics, physical activity, exercise and health: A review of the current evidence. Biochim Biophys Acta Mol Basis Dis. 2020;1866(12):165936. doi: 10.1016/j.bbadis.2020.165936 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Stattin K, Lind L, Elmståhl S, Wolk A, Lemming EW, Melhus H, et al. Physical activity is associated with a large number of cardiovascular-specific proteins: Cross-sectional analyses in two independent cohorts. Eur J Prev Cardiol. 2019;26(17):1865–73. doi: 10.1177/2047487319868033 [DOI] [PubMed] [Google Scholar]
4.Sharifzadeh M, Bagheri M, Speakman JR, Djafarian K. Comparison of total and activity energy expenditure estimates from physical activity questionnaires and doubly labelled water: a systematic review and meta-analysis. Br J Nutr. 2021;125(9):983–97. doi: 10.1017/S0007114520003049 [DOI] [PubMed] [Google Scholar]
5.White T, Westgate K, Hollidge S, Venables M, Olivier P, Wareham N, et al. Estimating energy expenditure from wrist and thigh accelerometry in free-living adults: a doubly labelled water study. Int J Obes (Lond). 2019;43(11):2333–42. doi: 10.1038/s41366-019-0352-x [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Aadland E, Nilsen AKO, Ylvisåker E, Johannessen K, Anderssen SA. Reproducibility of objectively measured physical activity: Reconsideration needed. J Sports Sci. 2020;38(10):1132–9. doi: 10.1080/02640414.2020.1743054 [DOI] [PubMed] [Google Scholar]
7.Atienza AA, Moser RP, Perna F, Dodd K, Ballard-Barbash R, Troiano RP, et al. Self-reported and objectively measured activity related to biomarkers using NHANES. Med Sci Sports Exerc. 2011;43(5):815–21. doi: 10.1249/MSS.0b013e3181fdfc32 [DOI] [PubMed] [Google Scholar]
8.Ledbetter MK, Tabacu L, Leroux A, Crainiceanu CM, Smirnova E. Cardiovascular mortality risk prediction using objectively measured physical activity phenotypes in NHANES 2003-2006. Prev Med. 2022;164:107303. doi: 10.1016/j.ypmed.2022.107303 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Leroux A, Xu S, Kundu P, Muschelli J, Smirnova E, Chatterjee N, et al. Quantifying the Predictive Performance of Objectively Measured Physical Activity on Mortality in the UK Biobank. J Gerontol A Biol Sci Med Sci. 2021;76(8):1486–94. doi: 10.1093/gerona/glaa250 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Ekblom-Bak E, Börjesson M, Bergman F, Bergström G, Dahlin-Almevall A, Drake I, et al. Accelerometer derived physical activity patterns in 27.890 middle-aged adults: The SCAPIS cohort study. Scand J Med Sci Sports. 2022;32(5):866–80. doi: 10.1111/sms.14131 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Ekelund U, Tarp J, Fagerland MW, Johannessen JS, Hansen BH, Jefferis BJ, et al. Joint associations of accelero-meter measured physical activity and sedentary time with all-cause mortality: a harmonised meta-analysis in more than 44 000 middle-aged and older individuals. Br J Sports Med. 2020;54(24):1499–506. doi: 10.1136/bjsports-2020-103270 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Pang Y, Kartsonaki C, Du H, Millwood IY, Guo Y, Chen Y, et al. Physical Activity, Sedentary Leisure Time, Circulating Metabolic Markers, and Risk of Major Vascular Diseases. Circ Genom Precis Med. 2019;12(9):386–96. doi: 10.1161/CIRCGEN.118.002527 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Sakaguchi CA, Nieman DC, Signini EF, Abreu RM, Catai AM. Metabolomics-Based Studies Assessing Exercise-Induced Alterations of the Human Metabolome: A Systematic Review. Metabolites. 2019;9(8):164. doi: 10.3390/metabo9080164 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Schranner D, Kastenmüller G, Schönfelder M, Römisch-Margl W, Wackerhage H. Metabolite Concentration Changes in Humans After a Bout of Exercise: a Systematic Review of Exercise Metabolomics Studies. Sports Med Open. 2020;6(1):11. doi: 10.1186/s40798-020-0238-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Latino F, Cataldi S, Carvutto R, De Candia M, D’Elia F, Patti A, et al. The Importance of Lipidomic Approach for Mapping and Exploring the Molecular Networks Underlying Physical Exercise: A Systematic Review. Int J Mol Sci. 2021;22(16):8734. doi: 10.3390/ijms22168734 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Tian Q, Corkum AE, Moaddel R, Ferrucci L. Metabolomic profiles of being physically active and less sedentary: a critical review. Metabolomics. 2021;17(7):68. doi: 10.1007/s11306-021-01818-y [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Khoramipour K, Sandbakk Ø, Keshteli AH, Gaeini AA, Wishart DS, Chamari K. Metabolomics in Exercise and Sports: A Systematic Review. Sports Med. 2022;52(3):547–83. doi: 10.1007/s40279-021-01582-y [DOI] [PubMed] [Google Scholar]
18.Chang Y-J, Tuz-Zahra F, Godbole S, Avitia Y, Bellettiere J, Rock CL, et al. Endothelial-derived cardiovascular disease-related microRNAs elevated with prolonged sitting pattern among postmenopausal women. Sci Rep. 2021;11(1):11766. doi: 10.1038/s41598-021-90154-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Park S. Association between polygenetic risk scores related to sarcopenia risk and their interactions with regular exercise in a large cohort of Korean adults. Clin Nutr. 2021;40(10):5355–64. doi: 10.1016/j.clnu.2021.09.003 [DOI] [PubMed] [Google Scholar]
20.Zhou J-Y, Park S. Regular exercise, alcohol consumption, and smoking interact with the polygenetic risk scores involved in insulin sensitivity and secretion for the risk of concurrent hyperglycemia, hypertension, and dyslipidemia. Nutrition. 2021;91–92:111422. doi: 10.1016/j.nut.2021.111422 [DOI] [PubMed] [Google Scholar]
21.Bergström G, Berglund G, Blomberg A, Brandberg J, Engström G, Engvall J, et al. The Swedish CArdioPulmonary BioImage Study: objectives and design. J Intern Med. 2015;278(6):645–59. doi: 10.1111/joim.12384 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Migueles JH, Cadenas-Sanchez C, Ekelund U, Delisle Nyström C, Mora-Gonzalez J, Löf M, et al. Accelerometer Data Collection and Processing Criteria to Assess Physical Activity and Other Outcomes: A Systematic Review and Practical Considerations. Sports Med. 2017;47(9):1821–45. doi: 10.1007/s40279-017-0716-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Aguilar-Farías N, Brown WJ, Peeters GMEEG. ActiGraph GT3X+ cut-points for identifying sedentary behaviour in older adults in free-living environments. J Sci Med Sport. 2014;17(3):293–9. doi: 10.1016/j.jsams.2013.07.002 [DOI] [PubMed] [Google Scholar]
24.Sasaki JE, John D, Freedson PS. Validation and comparison of ActiGraph activity monitors. J Sci Med Sport. 2011;14(5):411–6. doi: 10.1016/j.jsams.2011.04.003 [DOI] [PubMed] [Google Scholar]
25.Dunstan DW, Kingwell BA, Larsen R, Healy GN, Cerin E, Hamilton MT, et al. Breaking Up Prolonged Sitting Reduces Postprandial Glucose and Insulin Responses. Diabetes Care. 2012;35(5):976–83. doi: 10.2337/dc11-1931 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Hulsen T, de Vlieg J, Alkema W. BioVenn - a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams. BMC Genomics. 2008;9:488. doi: 10.1186/1471-2164-9-488 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Sagelv EH, Hopstock LA, Morseth B, Hansen BH, Steene-Johannessen J, Johansson J, et al. Device-measured physical activity, sedentary time, and risk of all-cause mortality: an individual participant data analysis of four prospective cohort studies. Br J Sports Med. 2023;57(22):1457–63. doi: 10.1136/bjsports-2022-106568 [DOI] [PubMed] [Google Scholar]
28.Siegbahn A, Eriksson N, Assarsson E, Lundberg M, Ballagi A, Held C, et al. Development and validation of a quantitative Proximity Extension Assay instrument with 21 proteins associated with cardiovascular risk (CVD-21). PLoS One. 2023;18(11):e0293465. doi: 10.1371/journal.pone.0293465 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Yang T, Yi J, He Y, Zhang J, Li X, Ke S, et al. Associations of Dietary Fats with All-Cause Mortality and Cardiovascular Disease Mortality among Patients with Cardiometabolic Disease. Nutrients. 2022;14(17):3608. doi: 10.3390/nu14173608 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Zhang Y, Sun Y, Yu Q, Song S, Brenna JT, Shen Y, et al. Higher ratio of plasma omega-6/omega-3 fatty acids is associated with greater risk of all-cause, cancer, and cardiovascular mortality: a population-based cohort study in UK Biobank. medRxiv. 2024. doi: 10.1101/2023.01.16.23284631 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Kaikkonen JE, Kresanov P, Ahotupa M, Jula A, Mikkilä V, Viikari JSA, et al. High serum n6 fatty acid proportion is associated with lowered LDL oxidation and inflammation: the Cardiovascular Risk in Young Finns Study. Free Radic Res. 2014;48(4):420–6. doi: 10.3109/10715762.2014.883071 [DOI] [PubMed] [Google Scholar]
32.Bajerska J, Skoczek-Rubińska A, Małczak L, Vucic V, Arsic A, Kojadinovic M, et al. Plasma fatty acid composition and some markers of dietary habits are associated with cardiovascular disease risk determined by an atherogenic plasma index in postmenopausal women. Nutr Res. 2023;115:47–60. doi: 10.1016/j.nutres.2023.05.008 [DOI] [PubMed] [Google Scholar]
33.Climie RE, Wheeler MJ, Grace M, Lambert EA, Cohen N, Owen N, et al. Simple intermittent resistance activity mitigates the detrimental effect of prolonged unbroken sitting on arterial function in overweight and obese adults. J Appl Physiol (1985). 2018;125(6):1787–94. doi: 10.1152/japplphysiol.00544.2018 [DOI] [PubMed] [Google Scholar]
34.Kujala UM, Mäkinen V-P, Heinonen I, Soininen P, Kangas AJ, Leskinen TH, et al. Long-term leisure-time physical activity and serum metabolome. Circulation. 2013;127(3):340–8. doi: 10.1161/CIRCULATIONAHA.112.105551 [DOI] [PubMed] [Google Scholar]
35.Bell JA, Hamer M, Richmond RC, Timpson NJ, Carslake D, Davey Smith G. Associations of device-measured physical activity across adolescence with metabolic traits: Prospective cohort study. PLoS Med. 2018;15(9):e1002649. doi: 10.1371/journal.pmed.1002649 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Henson J, Edwardson CL, Davies MJ, Gill JMR, Heaney LM, Khunti K, et al. Physical activity and lipidomics in a population at high risk of type 2 diabetes mellitus. J Sports Sci. 2020;38(10):1150–60. doi: 10.1080/02640414.2020.1744836 [DOI] [PubMed] [Google Scholar]
37.Vaara JP, Kyröläinen H, Vasankari T, Kainulainen H, Raitanen J, Kujala UM. Accelerometer-Based Sedentary Time, Physical Activity, and Serum Metabolome in Young Men. Metabolites. 2022;12(8):700. doi: 10.3390/metabo12080700 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Zhao S, Kusminski CM, Scherer PE. Adiponectin, Leptin and Cardiovascular Disorders. Circ Res. 2021;128(1):136–49. doi: 10.1161/circresaha.120.314458 [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Söderberg S, Stegmayr B, Stenlund H, Sjöström L-G, Agren A, Johansson L, et al. Leptin, but not adiponectin, predicts stroke in males. J Intern Med. 2004;256(2):128–36. doi: 10.1111/j.1365-2796.2004.01351.x [DOI] [PubMed] [Google Scholar]
40.Eriksson M, Johnson O, Boman K, Hallmans G, Hellsten G, Nilsson TK, et al. Improved fibrinolytic activity during exercise may be an effect of the adipocyte-derived hormones leptin and adiponectin. Thromb Res. 2008;122(5):701–8. doi: 10.1016/j.thromres.2008.01.010 [DOI] [PubMed] [Google Scholar]
41.Kostov K, Blazhev A. Use of Glycated Hemoglobin (A1c) as a Biomarker for Vascular Risk in Type 2 Diabetes: Its Relationship with Matrix Metalloproteinases-2, -9 and the Metabolism of Collagen IV and Elastin. Medicina (Kaunas). 2020;56(5):231. doi: 10.3390/medicina56050231 [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Barreto J, Karathanasis SK, Remaley A, Sposito AC. Role of LOX-1 (Lectin-Like Oxidized Low-Density Lipoprotein Receptor 1) as a Cardiovascular Risk Predictor: Mechanistic Insight and Potential Clinical Use. Arterioscler Thromb Vasc Biol. 2021;41(1):153–66. doi: 10.1161/ATVBAHA.120.315421 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Chen X, Wang S, Xu W, Zhao M, Zhang Y, Xiao H. Metformin Directly Binds to MMP-9 to Improve Plaque Stability. J Cardiovasc Dev Dis. 2023;10(2):54. doi: 10.3390/jcdd10020054 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Jin P, Cong S. LOX-1 and atherosclerotic-related diseases. Clin Chim Acta. 2019;491:24–9. doi: 10.1016/j.cca.2019.01.006 [DOI] [PubMed] [Google Scholar]
45.Fellström B, Helmersson-Karlqvist J, Lind L, Soveri I, Wu P-H, Thulin M, et al. Associations Between Apolipoprotein A1, High-Density Lipoprotein Cholesterol, and Urinary Cytokine Levels in Elderly Males and Females. J Interferon Cytokine Res. 2020;40(2):71–4. doi: 10.1089/jir.2019.0074 [DOI] [PubMed] [Google Scholar]
46.Patel H, Ashton NJ, Dobson RJB, Andersson L-M, Yilmaz A, Blennow K, et al. Proteomic blood profiling in mild, severe and critical COVID-19 patients. Sci Rep. 2021;11(1):6357. doi: 10.1038/s41598-021-85877-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Hejazi K, Mohammad Rahimi GR, Rosenkranz SK. Effects of Exercise Training on Inflammatory and Cardiometabolic Risk Biomarkers in Patients With Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. Biol Res Nurs. 2023;25(2):250–66. doi: 10.1177/10998004221132841 [DOI] [PubMed] [Google Scholar]
48.Hamirani YS, Pandey S, Rivera JJ, Ndumele C, Budoff MJ, Blumenthal RS, Nasir K. Markers of inflammation and coronary artery calcification: a systematic review. Atherosclerosis. 2008;201(1):1–7. doi: 10.1016/j.atherosclerosis.2008.04.045 [DOI] [PubMed] [Google Scholar]
49.Karason K, Girerd N, Andersson-Asssarsson J, Duarte K, Taube M, Svensson P-A, et al. Heart failure in obesity: insights from proteomics in patients treated with or without weight-loss surgery. Int J Obes (Lond). 2022;46(12):2088–94. doi: 10.1038/s41366-022-01194-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Ponce-de-Leon M, Hannemann A, Linseisen J, Nauck M, Lerch MM, Bülow R, et al. Links between ectopic and abdominal fat and systemic inflammation: New insights from the SHIP-Trend study. Digestive and Liver Disease. 2022;54(8):1030–7. doi: 10.1016/j.dld.2022.02.003 [DOI] [PubMed] [Google Scholar]
51.Hasan MM, Gazi MA, Das S, Fahim SM, Hossaini F, Khan A-R, et al. Gut biomolecules (I-FABP, TFF3 and lipocalin-2) are associated with linear growth and biomarkers of environmental enteric dysfunction (EED) in Bangladeshi children. Sci Rep. 2022;12(1):13905. doi: 10.1038/s41598-022-18141-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Peiris HN, Romero R, Vaswani K, Reed S, Gomez-Lopez N, Tarca AL, et al. Preterm labor is characterized by a high abundance of amniotic fluid prostaglandins in patients with intra-amniotic infection or sterile intra-amniotic inflammation. J Matern Fetal Neonatal Med. 2021;34(24):4009–24. doi: 10.1080/14767058.2019.1702953 [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Hocher C-F, Chen X, Zuo J, Horvathova K, Hocher B, Krämer BK, et al. Fibroblast growth factor 23 is associated with the development of gestational diabetes mellitus. Diabetes Metab Res Rev. 2023;39(8):e3704. doi: 10.1002/dmrr.3704 [DOI] [PubMed] [Google Scholar]
54.Sanders-van Wijk S, Tromp J, Beussink-Nelson L, Hage C, Svedlund S, Saraste A, et al. Proteomic Evaluation of the Comorbidity-Inflammation Paradigm in Heart Failure With Preserved Ejection Fraction: Results From the PROMIS-HFpEF Study. Circulation. 2020;142(21):2029–44. doi: 10.1161/CIRCULATIONAHA.120.045810 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Royston L, Isnard S, Perrin N, Sinyavskaya L, Berini C, Lin J, et al. Growth differentiation factor-15 as a biomarker of atherosclerotic coronary plaque: Value in people living with and without HIV. Front Cardiovasc Med. 2022;9:964650. doi: 10.3389/fcvm.2022.964650 [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Planès C, Randrianarison NH, Charles R-P, Frateschi S, Cluzeaud F, Vuagniaux G, et al. ENaC-mediated alveolar fluid clearance and lung fluid balance depend on the channel-activating protease 1. EMBO Mol Med. 2010;2(1):26–37. doi: 10.1002/emmm.200900050 [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Batista AP, Barbosa KF, de Azevedo RJ, Vianna VN, de Queiroz EM, Marinho CC, et al. Hypertension is associated with a variant in the RARRES2 gene in populations of Ouro Preto, Minas Gerais, Brazil: a cross-sectional study. Int J Mol Epidemiol Genet. 2021;12(3):40–51. [PMC free article] [PubMed] [Google Scholar]
58.Han Y, Hua S, Chen Y, Yang W, Zhao W, Huang F, et al. Circulating PGLYRP1 Levels as a Potential Biomarker for Coronary Artery Disease and Heart Failure. J Cardiovasc Pharmacol. 2021;77(5):578–85. doi: 10.1097/FJC.0000000000000996 [DOI] [PubMed] [Google Scholar]
59.Lankin VZ, Tikhaze AK, Melkumyants AM. Dicarbonyl-Dependent Modification of LDL as a Key Factor of Endothelial Dysfunction and Atherosclerotic Vascular Wall Damage. Antioxidants. 2022;11(8):1565. doi: 10.3390/antiox11081565 [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Kanno Y. The uPA/uPAR System Orchestrates the Inflammatory Response, Vascular Homeostasis, and Immune System in Fibrosis Progression. IJMS. 2023;24(2):1796. doi: 10.3390/ijms24021796 [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Lind L, Titova O, Zeng R, Zanetti D, Ingelsson M, Gustafsson S, et al. Plasma Protein Profiling of Incident Cardiovascular Diseases: A Multisample Evaluation. Circ Genom Precis Med. 2023;16(6):e004233. doi: 10.1161/CIRCGEN.123.004233 [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Gustafsson S, Lampa E, Jensevik Eriksson K, Butterworth AS, Elmståhl S, Engström G, et al. Markers of imminent myocardial infarction. Nat Cardiovasc Res. 2024;3(2):130–9. doi: 10.1038/s44161-024-00422-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Ekblom-Bak E, Börjesson M, Ekblom Ö, Angerås O, Bergman F, Berntsson C, et al. Accelerometer derived physical activity and subclinical coronary and carotid atherosclerosis: cross-sectional analyses in 22 703 middle-aged men and women in the SCAPIS study. BMJ Open. 2023;13(11):e073380. doi: 10.1136/bmjopen-2023-073380 [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Carlsson AC, Ruge T, Kjøller E, Hilden J, Kolmos HJ, Sajadieh A, et al. 10-Year Associations Between Tumor Necrosis Factor Receptors 1 and 2 and Cardiovascular Events in Patients With Stable Coronary Heart Disease: A CLARICOR (Effect of Clarithromycin on Mortality and Morbidity in Patients With Ischemic Heart Disease) Trial Substudy. J Am Heart Assoc. 2018;7(9):e008299. doi: 10.1161/JAHA.117.008299 [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Probert L. TNF and its receptors in the CNS: The essential, the desirable and the deleterious effects. Neuroscience. 2015;302:2–22. doi: 10.1016/j.neuroscience.2015.06.038 [DOI] [PubMed] [Google Scholar]
66.Thougaard E, Nielsen PV, Forsberg A, Phuong V, Velasco AM, Wlodarczyk A, et al. Systemic treatment with a selective TNFR2 agonist alters the central and peripheral immune responses and transiently improves functional outcome after experimental ischemic stroke. J Neuroimmunol. 2023;385:578246. doi: 10.1016/j.jneuroim.2023.578246 [DOI] [PubMed] [Google Scholar]
67.Rip J, Nierman MC, Wareham NJ, Luben R, Bingham SA, Day NE, et al. Serum lipoprotein lipase concentration and risk for future coronary artery disease: the EPIC-Norfolk prospective population study. Arterioscler Thromb Vasc Biol. 2006;26(3):637–42. doi: 10.1161/01.ATV.0000201038.47949.56 [DOI] [PubMed] [Google Scholar]
68.Elosua R, Cupples LA, Fox CS, Polak JF, D’Agostino RA Sr, Wolf PA, et al. Association between well-characterized lipoprotein-related genetic variants and carotid intimal medial thickness and stenosis: The Framingham Heart Study. Atherosclerosis. 2006;189(1):222–8. doi: 10.1016/j.atherosclerosis.2005.12.005 [DOI] [PubMed] [Google Scholar]
69.Black MA, Cable NT, Thijssen DH, Green DJ. Impact of age, sex, and exercise on brachial artery flow-mediated dilatation. Am J Physiol Heart Circ Physiol. 2009;297(3):H1109–16. doi: 10.1152/ajpheart.00226.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Pierce GL, Eskurza I, Walker AE, Fay TN, Seals DR. Sex-specific effects of habitual aerobic exercise on brachial artery flow-mediated dilation in middle-aged and older adults. Clin Sci (Lond). 2011;120(1):13–23. doi: 10.1042/CS20100174 [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Barker J, Smith Byrne K, Doherty A, Foster C, Rahimi K, Ramakrishnan R, et al. Physical activity of UK adults with chronic disease: cross sectional analysis of accelerometer measured physical activity in 96 706 UK Biobank participants. Int J Epidemiol. 2019;48(4):1386–1386. doi: 10.1093/ije/dyz148 [DOI] [PMC free article] [PubMed] [Google Scholar]

PLoS One. 2025 Jun 11;20(6):e0325720. doi: 10.1371/journal.pone.0325720.r001

Author response to Decision Letter 0

16 Dec 2024

PLoS One. doi: 10.1371/journal.pone.0325720.r002

Decision Letter 0

Hidetaka Hamasaki

28 Jan 2025

Dear Dr. Ekblom,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

==============================

ACADEMIC EDITOR:

==============================

Please submit your revised manuscript by Mar 14 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Hidetaka Hamasaki

Academic Editor

PLOS ONE

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following financial disclosure: [ÖE was funded by Livförsäkringsbolaget Skandia, Risk&Hälsa.]. At this time, please address the following queries:

a) Please clarify the sources of funding (financial or material support) for your study. List the grants or organizations that supported your study, including funding received from your institution.

c) If any authors received a salary from any of your funders, please state which authors and which funders.

d) If you did not receive any funding for this study, please state: “The authors received no specific funding for this work.”

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

3. We note that you have indicated that there are restrictions to data sharing for this study. For studies involving human research participant data or other sensitive data, we encourage authors to share de-identified or anonymized data. However, when data cannot be publicly shared for ethical reasons, we allow authors to make their data sets available upon request. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

Before we proceed with your manuscript, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., a Research Ethics Committee or Institutional Review Board, etc.). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of recommended repositories, please see https://journals.plos.org/plosone/s/recommended-repositories. You also have the option of uploading the data as Supporting Information files, but we would recommend depositing data directly to a data repository if possible.

Please update your Data Availability statement in the submission form accordingly.

Additional Editor Comments (if provided):

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

Reviewer #4: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: N/A

Reviewer #4: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

Reviewer #4: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

Reviewer #1: I would like to congratulate the authors of developing this manuscript which is challenging due to the number of variables and sample size. It is an important information due to unclear association status between accelerometer-based measures of different physical activities and plasma metabolomic and proteomic biomarkers in a population-based study. Many relationship were anticipated, like the association of sedentary with HDL and higher activity level with VLDL and LDL.

It was not clear from the manuscript whether diet plans were selected or advised to the participants and whether their diet could have any differences on the outcomes.

There was two short titles for the manuscript, to review this.

Reviewer #2: The manuscript explores an important and timely topic, providing valuable insights into the associations between physical activity and cardiovascular biomarkers. While the study employs decent methodologies and presents significant findings, there are critical areas requiring attention. Addressing these issues will enhance the manuscript's impact. Below, I outline specific critiques for each section of the manuscript.

Title

Abstract

Line 71–98: The terminology used (e.g., "mirroring," "robust relations") might be too technical for a general audience. Brief definitions or rephrasing could help.

The inclusion of exact numerical values for validated biomarkers might overwhelm the abstract's narrative. Consider summarizing these findings qualitatively.

The abstract fails to explicitly state the study's limitations, which would provide a more balanced overview.

Introduction

The aim of the study (line 130–133) lacks a hypothesis-driven narrative. Stating a clear research question would enhance focus.

Methods

Line 136–243:

The description of accelerometer processing could be expanded to address potential biases introduced by the low-frequency extension filter.

No mention is made of how outliers or missing data were handled statistically (e.g., imputation methods).

Confounding variables (lines 203–214): While a comprehensive list is provided, there is no discussion on whether any were prioritized based on their theoretical significance.

Results

Line 254–305:

The authors mention "robust relations" but do not adequately define how this robustness was assessed beyond statistical significance.

Table 1: Some demographic variables (e.g., education) are reported without discussing their impact on the study findings.

Discussion

Line 307–319:

Potential limitations like the cross-sectional nature of the study are mentioned but downplayed. A more critical acknowledgment is warranted, particularly around confounding and selection biases.

No mention is made of generalizability. Since participants are middle-aged Swedes, findings may not apply to diverse populations.

Reviewer #3: The focus of this study by Ekblom et. al. is to find an association between various intensity PA classes and metabolomic and cardiovascular protein biomarkers in a middle-aged population.

The main drawback of this study is that it is not giving any new information except expanding on the previous correlation between these variables. Metabolomic and Proteomic Profiling are done in different conditions and cohort which doesn't give much emphasis to the manuscript.

Reviewer #4: I am grateful for the opportunity to review this manuscript, which examined the associations between accelerometer-assessed physical activity behaviours and CVD-related metabolomic and proteomic biomarkers using two cross-sectional data from the main SCAPIS cohort (discovery data) and SCAPIS pilot cohort (validation data). Though the manuscript is well-written and contains some interesting findings, a few issues may have impacted the scientific quality of the findings.

Below are some suggestions/comments/questions the authors may consider if they find them intuitive to enhance the quality of the paper.

Abstract:

1. The authors could briefly describe the demographic characteristics of the studied participants, both the discovery and validated samples.

2. The number of the main SCAPIS cohort (discovery = 5557) does not match the description in the main text – discovery = 4647)

3. The authors can consider mentioning the analytical method used to examine the associations in the abstract.

4. The description of the observed associations is unclear, and the authors can help the readers by clarifying the metabolic and proteomic biomarkers and the direction of their associations with the physical activity behaviour classes, especially for the proteomic markers (lines 90 – 92).

Introduction:

1. The authors made several factual statements in the first four opening sentences (page 5, lines 100 – 104) and it would be more appropriate for them to provide some supporting citations (references).

2. Similarly, the sentences on page 5 lines 110 – 114 may need some supporting references.

3. Though the introduction looks well-written, it lacks some relevant and specific details that justify the authors’ decision to examine the relationships of physical activity behaviour classes with metabolic and proteomic biomarkers. There is extensive evidence of the associations between the physical activity behaviour spectrum and biomarkers. A brief contextual review of the evidence on moderate-to-vigorous intensity physical activity (MVPA), light-intensity physical activity (LIPA) and sedentary behaviour (SB) (including SB bouts), and the interdependence of these behaviours, would have been appropriate to build their study rationale.

Methods:

1. The description on page 9, lines 181 – 184 is a little confusing. The authors may consider clarifying the total number of discovery samples from the regions/sites (5075) and the initial number of included subjects (30154) and after excluding those with missing data the total number of remaining subjects (4887).

2. It is not clear what “FA variables” means. It may presumably be a typo (PA – physical activity); if it isn’t, it would be appropriate to define abbreviations when used for the first time in the document.

3. It would be appropriate to define the abbreviation “E-GFR”.

4. Some readers may also be interested in how the biomarkers used in the Cockcroft-Gault Equation to measure the estimated glomerular filtration rate (eGFR) were assessed.

Results:

1. The analysis and results are complex, and the descriptions are comprehensive. There are a few instances, though, where the authors would have been more specific about the direction of the associations (beneficial ‘+’ or detrimental ‘-’) of the physical activity behaviours with the biomarkers. For instance, the description of the associations on page 27, line 298.

2. The sentence “Opposite relations were found for ITGB2 and LPL” on page 27, line 316 needs clarity.

3. It is impressive that the authors also explored the potential non-linear associations; however, describing the non-linear analytic approach utilised in the statistical analysis section would be more appropriate.

4. Did the authors also examine the non-linear associations with the metabolic/proteomic biomarkers of LIPA and prolonged sedentary behaviour?

5. There seem to be gender differences in the associations, what is not clear, though, is the direction of the associations. The authors may consider describing the direction of the physical activity behaviours and sex interaction associations with the biomarkers.

6. Similarly, the authors could describe the direction of the physical activity behaviours and waist circumference interaction associations with the biomarkers.

7. Did the authors also examine the interaction of LIPA with sex or waist circumference?

8. Association tables: the authors may have a specific reason to only report coefficients. However, reporting coefficients with confidence intervals or standard errors would be more informative in understanding the statistical significance of the findings.

Discussion:

1. The findings described in the results mainly focus on the directions rather than the magnitudes/strengths of the associations, however, the authors seem to be highlighting the strengths of the associations in their discussion of the findings.

2. There are several factual statements in the discussion that may need appropriate supporting evidence (references). For example, the sentence on page 34, lines 454 – 457.

Figures:

1. The authors did not adequately justify why LIPA was excluded from the diagram. As the authors noted in the results section LIPA and sedentary behaviour have a strong negative interrelation, some readers may find it informative if all the physical activity behaviours, including LIPA, are shown in the vain diagram with the intersections.

2. The authors could consider a brief description of the vain diagram, highlighting the intersections in the legend.

**********

what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

Reviewer #4: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/ . PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org . Please note that Supporting Information files do not need this step.

PLoS One. 2025 Jun 11;20(6):e0325720. doi: 10.1371/journal.pone.0325720.r003

Author response to Decision Letter 1

12 Mar 2025

Journal requirements

Comment Author response Alterations

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at We have compared requirements including naming.

2. Thank you for stating the following financial disclosure: [ÖE was funded by Livförsäkringsbolaget Skandia, Risk&Hälsa.]. At this time, please address the following queries:

a) Please clarify the sources of funding (financial or material support) for your study. List the grants or organizations that supported your study, including funding received from your institution. Thanks for pointing this out. This information has been added to the submission system. In the manuscript, this information has been moved from Acknowledgements to Funding. Text updated on p 39, under "Acknowledgements" and "Funding"

b) State what role the funders took in the study. If the funders had no role in your study, please state: “The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.” As the funding bodies did not have any such role, the statement has been added to the submission system and the manuscript. During the work on this paper, the SCAPIS office launched an altered standard wording for Acknowledgements, which has now been added Text updated on p 39, under "Acknowledgements" and "Funding"

c) If any authors received a salary from any of your funders, please state which authors and which funders. The main author was funded by a grant from Livförsäkringsbolaget Skandia in the form of salary. This has been added to the manuscript Added text on p 39 under "Funding"

d) If you did not receive any funding for this study, please state: “The authors received no specific funding for this work.”

We note that you have indicated that there are restrictions to data sharing for this study. For studies involving human research participant data or other sensitive data, we encourage authors to share de-identified or anonymized data. However, when data cannot be publicly shared for ethical reasons, we allow authors to make their data sets available upon request. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

Before we proceed with your manuscript, please address the following prompts:

Data used in this paper are a part of the SCAPIS database. Researcher aiming to obtain access to these data may apply for them to the data access committee at the SCAPIS office. The data access agreement between the authors and the SCAPIS hinders any direct sharing of data. Further, the ethical permissions for this paper do not allow for direct sharing of de-identified data. The data access committee can be reached at scapis@scapis.org, with more information at the website www.scapis.org.

Please update your Data Availability statement in the submission form accordingly. Data Availability statement updated aThe following text has been added: The General Data Protection Regulation (EU 2016/679) classifies de-identified versions of sensitive data that are sufficiently detailed to allow for re-identification as sensitive personal information. According to Swedish law (Law 2003:460 for ethical review of research involving humans), ethical permission is required to process such data. In accordance with Swedish legislation, the data can and will be made available to researchers who meet the criteria for access to confidential data, which includes obtaining their own ethics approval from the Swedish Ethical Review Authority (email: registrator@etikprovning.se; website: https://etikprovningsmyndigheten.se). Data applications can then be made by contacting SCAPIS (email: scapis@scapis.org; website: https://www.scapis.org/data-access/).

Reviewer #1

I would like to congratulate the authors of developing this manuscript which is challenging due to the number of variables and sample size. It is an important information due to unclear association status between accelerometer-based measures of different physical activities and plasma metabolomic and proteomic biomarkers in a population-based study. Many relationship were anticipated, like the association of sedentary with HDL and higher activity level with VLDL and LDL. We agree that the number of biomarkers in this paper is large, but also constitute a strength of the paper and that some of the relationships were expected. None.

It was not clear from the manuscript whether diet plans were selected or advised to the participants and whether their diet could have any differences on the outcomes. No diet plan was given, and diet was not controlled other than that subjects were in a fasted state during blood sampling. Food frequency questionnaires were used to survey the dietary habits of the participants. Data on macronutrient intake were included in the analyses. Altered text on p 11, rows, 247-248

It is worth to be mentioned by authors in the manuscript is the lack of detailed individuals health, medical conditions and therapies which might also affect outcomes like dyslipidemia and being on statin therapy as examples. We agree that there are conditions that may be of relevance that we could not include, such as those mentioned. However, we chose to include overweight status, kidney function, diabetes status, along with behaviours such as diet and smoking. We have included a paragraph on this in the limitations section, on p 37, rows 546-548

There was two short titles for the manuscript, to review this. We apologize, this has been adjusted. The latter short title has been removed.

Reviewer #2

The manuscript explores an important and timely topic, providing valuable insights into the associations between physical activity and cardiovascular biomarkers. While the study employs decent methodologies and presents significant findings, there are critical areas requiring attention. Addressing these issues will enhance the manuscript's impact. Below, I outline specific critiques for each section of the manuscript. We thank the reviewer for the time taken to review our manuscript and the valuable input.

The title could be more concise. Consider removing redundant phrases like "assessed physical activity variables" to streamline it without losing clarity (e.g., "Associations between Physical Activity and CVD-related Biomarkers"). Thank you for the input. We agree that the title can be perceived as bulky. Title shortened for clarity.

Line 71–98: The terminology used (e.g., "mirroring," "robust relations") might be too technical for a general audience. Brief definitions or rephrasing could help. We agree and have tried to clarify this further.

Alterations in results section of abstract.

The inclusion of exact numerical values for validated biomarkers might overwhelm the abstract's narrative. Consider summarizing these findings qualitatively. We agree with the reviewer that this could be less detailed. However, we have discussed this and tried to express the results in overarching terms, without being too vague. We hope that this way of expressing the results might work in many readers minds. none

Lines 100–133: There is an overreliance on citing older studies without critical analysis of gaps they leave. For example, it would be helpful to contrast self-reported physical activity data with accelerometer-based data more thoroughly. We agree with the reviewer that this is most important. However, to our knowledge, there are no recent analysis that compare the predictive validity on biomarkers between self-reports and motion sensors. We therefore chose to base a discussion on this on an earlier paper and a few newer on similar topics. Addition of discussion and references on rows 127-130.

The aim of the study (line 130–133) lacks a hypothesis-driven narrative. Stating a clear research question would enhance focus. We agree with the reviewer that our aim is not hypothesis driven, but rather explorative. The large number of biomarkers studied makes it difficult to synthesise a single hypothesis, but based on studies presented in the background, it would be plausible to state that we believe that the relations would be different across PA classes. On rows 165-166, we state We hypothesize that relations between PA and biomarkers are different across PA classes.

The description of accelerometer processing could be expanded to address potential biases introduced by the low-frequency extension filter. The low-frequency extension (LFE) filter may to some extent affect time spent in SED or in prolSED, but not the other PA classes. The relations to the studied biomarkers are assumed to be more stable. We have added a short discussion on this on rows 200-201.

No mention is made of how outliers or missing data were handled statistically (e.g., imputation methods).

We agree that this could be described in more detail. No imputation was applied as stated on row 235, but the text is now slightly more detailed. The extent of missing data is given under “Blood sampling and analysis”. We state that “No imputation was applied for missing data.” on row 265-266.

Confounding variables (lines 203–214): While a comprehensive list is provided, there is no discussion on whether any were prioritized based on their theoretical significance. We acknowledge this shortcoming to the text. The included independent variables all have their theoretical significance, and it is difficult to rank them due to the varying nature of the different outcome variables. However, we considered it more transparent to apply two models. Age, sex, kidney function and diabetes function (and plate number) were thought as variables suitable to adjust for in the first model. In model 2, remaining variables were included, due to their potential influence on the relation between PA and biomarker level.

We have added a short part on model development.

Added sentence on row 282-283

Statistical findings are often presented without adequate contextualization. For instance, what are the implications of HDL species being negatively associated with SED? A brief interpretation in the results section could aid clarity. We agree that there are many relations worth discussing. The implications are many and we have tried to elaborate on this in the discussion. However, the cross-sectional design limits the ability to discuss implications. Added short sentence on row 453-454

The authors mention "robust relations" but do not adequately define how this robustness was assessed beyond statistical significance. Under “Statistics” we define “robust” relations as those with opposing relations to MVPA vs SED/prolSED and to separate those from biomarkers with relations to only one PA-class. The idea is to separate biomarkers with a more solid (“robust”) and more discrete (“separate”) relation to PA beyond the application of two-sample validation and two different regression models.

We agree that this could be further described, and we have altered the description in the text for clarity. Added sentence on rows 288-290

Table 1: Some demographic variables (e.g., education) are reported without discussing their impact on the study findings. We thank the reviewer for this comment. The results from the regression analyses including the other independent variables than PA could be presented and discussed. However, to keep the focus on the aim of the paper, we chose to report only the results for the PA classes.

The discussion is overly descriptive and does not sufficiently explore the causal implications of the findings. For instance, are the observed associations indicative of mechanistic pathways or simply correlations? We agree that the discussion refrains from discussing in causal terms, which is due to the study design. However, the reviewer is correct in that causality could be discussed to a certain degree, based on potential physiological mechanisms. For metabolomic markers, PA is known to increase lipid metabolism and thereby could affect lipid profile via LPL activation or (up-)regulation of LDL receptors. Proteomic markers related to inflammation (such as IL-6) may be affected by PA by alterations to the autonomic system, AMPK activation or changes in for example visceral fat deposits. We have added a few examples, without arguing that we have covered all such mechanisms.

We deemed the number of such mechanisms to be too high to be systematically covered in the discussion. Added examples of potential physiological pathways on rows 504-510.

Potential limitations like the cross-sectional nature of the study are mentioned but downplayed. A more critical acknowledgment is warranted, particularly around confounding and selection biases. We agree with the reviewer that the cross-sectional design prevents any causal inferences, and we have mentioned this throughout the manuscript. We have added a section on selection biases and possible residual confounding in the “Strength and limitations”-section. Added a paragraph on this on rows 546-555

No mention is made of generalizability. Since participants are middle-aged Swedes, findings may not apply to diverse populations. We agree with the reviewer and have added a text on this Added text on rows 556-558

Reviewer #3

The focus of this study by Ekblom et. al. is to find an association between various intensity PA classes and metabolomic and cardiovascular protein biomarkers in a middle-aged population.

We agree with the reviewer that previous investigations have described several of these relations. However, we believe that the study including sensor-based PA assessments and cross-validation in a relatively large and heterogenous population has a value in describing the wider patterns of relations. Also, the lack of convincing relationships between prolSED and biomarkers outside of those found for SED is of potential importance for the design of future intervention studies.

Reviewer #4

I am grateful for the opportunity to review this manuscript, which examined the associations between accelerometer-assessed physical activity behaviours and CVD-related metabolomic and proteomic biomarkers using two cross-sectional data from the main SCAPIS cohort (discovery data) and SCAPIS pilot cohort (validation data). Though the manuscript is well-written and contains some interesting findings, a few issues may have impacted the scientific quality of the findings.

Below are some suggestions/comments/questions the authors may consider if they find them intuitive to enhance the quality of the paper.

We thank the reviewer for critically assessing our manuscript and the possibility to revise in accordance

Attachment

Submitted filename: Ekblom_Response to reviewers.docx

pone.0325720.s012.docx^{(58.6KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0325720.r004

Decision Letter 1

Hidetaka Hamasaki

28 Mar 2025

Dear Dr. Ekblom,

Please submit your revised manuscript by May 12 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Hidetaka Hamasaki

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

Reviewer #4: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions??>

Reviewer #1: Partly

Reviewer #2: Yes

Reviewer #4: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #4: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #4: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #4: Yes

**********

Reviewer #1: The authors have adequately addressed my comments raised in the previous review and I don't have any additional comments.

Reviewer #2: I would like to congratulate and thank the author for carefully addressing my comments. I have no further concerns.

Reviewer #4: The authors did a great job satisfactorily addressing most of the concerns; however, their response to one of the comments raises a major concern about the credibility of the reported findings. The reported effect sizes look rather very small. The tables do not look that huge, and the authors may consider restructuring the tables. For them to only report the significant findings without confidence intervals or standard errors may indicate that the authors were selective in their reporting and compromised the validity of the findings. The question, though, is why the authors did not report non-significant results. In some points, the authors also highlight the strength of the associations in their discussion, which is not obvious in the reported coefficients without confidence intervals. The comment with the response is below:

Comment: Association tables: the authors may have a specific reason to only report coefficients. However, reporting coefficients with confidence intervals or standard errors would be more informative in understanding the statistical significance of the findings.

Response: We agree that adding CIs or measures of variance would give more information. However, the tables are rather large as is and contain only the significant relations. Confidence intervals would all exclude “zero”. The magnitude of for example standard deviations would be small and difficult to read. Therefore, we chose not to include these data.

**********

what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #1: No

Reviewer #2: No

Reviewer #4: No

**********

PLoS One. 2025 Jun 11;20(6):e0325720. doi: 10.1371/journal.pone.0325720.r005

Author response to Decision Letter 2

30 Apr 2025

Reviewer #4:

1. The authors did a great job satisfactorily addressing most of the concerns; however, their response to one of the comments raises a major concern about the credibility of the reported findings. The reported effect sizes look rather very small. The tables do not look that huge, and the authors may consider restructuring the tables. For them to only report the significant findings without confidence intervals or standard errors may indicate that the authors were selective in their reporting and compromised the validity of the findings.

2. The question, though, is why the authors did not report non-significant results.

3. In some points, the authors also highlight the strength of the associations in their discussion, which is not obvious in the reported coefficients without confidence intervals. The comment with the response is below:

[Citing earlier conversations]

Author response:

1. We agree with the reviewer that the coefficients are nominally low, which was mentioned in the manuscript. Based on the suggestion from the reviewer, we have added SE to tables 2, 3, and 4, which may help in interpreting the values.

2. We agree with the reviewer that it is essential to report null findings in scientific reports and are therefore listing biomarkers with non-significant relations to the different PA classes in supplementary table 5.

3. The addition of SEs to the tables may help interpretation in this instance. However, we regarded all significant relations as essential and chose to mention the significant biomarkers with the nominally largest coefficient instead of, for example, those with the lowest p-value. Thus note, all markers in tables 2-4 are significant.

Attachment

Submitted filename: Response to reviewer250430.docx

pone.0325720.s013.docx^{(17KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0325720.r006

Decision Letter 2

Hidetaka Hamasaki

18 May 2025

Associations between physical activity and CVD-related metabolomic and proteomic biomarkers.

PONE-D-24-56786R2

Dear Dr. Ekblom,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. If you have any questions relating to publication charges, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Hidetaka Hamasaki

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for resubmitting the revised manuscript.

On behalf of Reviewer 4, I have reviewed your responses and the corresponding revisions. Regarding the comment about the small effect size, you have appropriately provided the standard errors. As for the biomarkers that did not show a significant association, they are now presented in Supplementary Table 5. These revisions sufficiently address the concerns raised by Reviewer 4. Therefore, I am pleased to inform you that your manuscript is accepted for publication. Congratulations, and thank you for reporting such a valuable piece of research.

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0325720.r007

Acceptance letter

Hidetaka Hamasaki

PONE-D-24-56786R2

PLOS ONE

Dear Dr. Ekblom,

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now being handed over to our production team.

At this stage, our production department will prepare your paper for publication. This includes ensuring the following:

* All references, tables, and figures are properly cited

* All relevant supporting information is included in the manuscript submission,

* There are no issues that prevent the paper from being properly typeset

You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days to review your paper and let you know the next and final steps.

Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

If we can help with anything else, please email us at customercare@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Hidetaka Hamasaki

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Table. Lists of non-significant biomarkers.

(DOCX)

pone.0325720.s001.docx^{(36.8KB, docx)}

S2 Table. Interactions between SED and sex, using model 2 plus an interaction term.

(DOCX)

pone.0325720.s002.docx^{(17.6KB, docx)}

S3 Table. Interactions between MVPA and sex, using model 2 plus an interaction term.

(DOCX)

pone.0325720.s003.docx^{(16KB, docx)}

S4 Table. Interactions between SED and waist circumference (dichotomized at 102 cm and 88 cm for male and females, respectively) using model 2 plus an interaction term.

(DOCX)

pone.0325720.s004.docx^{(17.5KB, docx)}

S5 Table. Interactions between MVPA and waist circumference (dichotomized at 102 cm and 88 cm for male and females, respectively) using model 2 plus an interaction term.

(DOCX)

pone.0325720.s005.docx^{(16.1KB, docx)}

S1 Fig. Wenn-diagram showing number of metabolomic biomarkers related to SED, prolSED, LIPA and MVPA.

SED encapsuled all biomarkers related to LIPA.

(TIF)

pone.0325720.s006.tif^{(52.2KB, tif)}

S2 Fig. Wenn-diagram showing number of proteomic biomarkers related to SED, prolSED, LIPA and MVPA.

SED encapsuled all biomarkers related to LIPA.

(TIF)

pone.0325720.s007.tif^{(65.2KB, tif)}

S3 Fig. Heat map of relations between metabolomic markers and the PA classes.

(TIF)

pone.0325720.s008.tif^{(172.1KB, tif)}

S4 Fig. Heat map of relations between proteomic markers and the PA classes.

(TIF)

pone.0325720.s009.tif^{(123.5KB, tif)}

S1 Checklist. STROBE-checklist-v4_B4_241209.

(DOCX)

pone.0325720.s010.docx^{(51.2KB, docx)}

Attachment

Submitted filename: Ekblom_Response to reviewers.docx

pone.0325720.s012.docx^{(58.6KB, docx)}

Attachment

Submitted filename: Response to reviewer250430.docx

pone.0325720.s013.docx^{(17KB, docx)}

Data Availability Statement

[pone.0325720.ref001] 1.Wood G, Taylor E, Ng V, Murrell A, Patil A, van der Touw T, et al. Estimating the Effect of Aerobic Exercise Training on Novel Lipid Biomarkers: A Systematic Review and Multivariate Meta-Analysis of Randomized Controlled Trials. Sports Med. 2023;53(4):871–86. doi: 10.1007/s40279-023-01817-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref002] 2.Kelly RS, Kelly MP, Kelly P. Metabolomics, physical activity, exercise and health: A review of the current evidence. Biochim Biophys Acta Mol Basis Dis. 2020;1866(12):165936. doi: 10.1016/j.bbadis.2020.165936 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref003] 3.Stattin K, Lind L, Elmståhl S, Wolk A, Lemming EW, Melhus H, et al. Physical activity is associated with a large number of cardiovascular-specific proteins: Cross-sectional analyses in two independent cohorts. Eur J Prev Cardiol. 2019;26(17):1865–73. doi: 10.1177/2047487319868033 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref004] 4.Sharifzadeh M, Bagheri M, Speakman JR, Djafarian K. Comparison of total and activity energy expenditure estimates from physical activity questionnaires and doubly labelled water: a systematic review and meta-analysis. Br J Nutr. 2021;125(9):983–97. doi: 10.1017/S0007114520003049 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref005] 5.White T, Westgate K, Hollidge S, Venables M, Olivier P, Wareham N, et al. Estimating energy expenditure from wrist and thigh accelerometry in free-living adults: a doubly labelled water study. Int J Obes (Lond). 2019;43(11):2333–42. doi: 10.1038/s41366-019-0352-x [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref006] 6.Aadland E, Nilsen AKO, Ylvisåker E, Johannessen K, Anderssen SA. Reproducibility of objectively measured physical activity: Reconsideration needed. J Sports Sci. 2020;38(10):1132–9. doi: 10.1080/02640414.2020.1743054 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref007] 7.Atienza AA, Moser RP, Perna F, Dodd K, Ballard-Barbash R, Troiano RP, et al. Self-reported and objectively measured activity related to biomarkers using NHANES. Med Sci Sports Exerc. 2011;43(5):815–21. doi: 10.1249/MSS.0b013e3181fdfc32 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref008] 8.Ledbetter MK, Tabacu L, Leroux A, Crainiceanu CM, Smirnova E. Cardiovascular mortality risk prediction using objectively measured physical activity phenotypes in NHANES 2003-2006. Prev Med. 2022;164:107303. doi: 10.1016/j.ypmed.2022.107303 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref009] 9.Leroux A, Xu S, Kundu P, Muschelli J, Smirnova E, Chatterjee N, et al. Quantifying the Predictive Performance of Objectively Measured Physical Activity on Mortality in the UK Biobank. J Gerontol A Biol Sci Med Sci. 2021;76(8):1486–94. doi: 10.1093/gerona/glaa250 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref010] 10.Ekblom-Bak E, Börjesson M, Bergman F, Bergström G, Dahlin-Almevall A, Drake I, et al. Accelerometer derived physical activity patterns in 27.890 middle-aged adults: The SCAPIS cohort study. Scand J Med Sci Sports. 2022;32(5):866–80. doi: 10.1111/sms.14131 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref011] 11.Ekelund U, Tarp J, Fagerland MW, Johannessen JS, Hansen BH, Jefferis BJ, et al. Joint associations of accelero-meter measured physical activity and sedentary time with all-cause mortality: a harmonised meta-analysis in more than 44 000 middle-aged and older individuals. Br J Sports Med. 2020;54(24):1499–506. doi: 10.1136/bjsports-2020-103270 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref012] 12.Pang Y, Kartsonaki C, Du H, Millwood IY, Guo Y, Chen Y, et al. Physical Activity, Sedentary Leisure Time, Circulating Metabolic Markers, and Risk of Major Vascular Diseases. Circ Genom Precis Med. 2019;12(9):386–96. doi: 10.1161/CIRCGEN.118.002527 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref013] 13.Sakaguchi CA, Nieman DC, Signini EF, Abreu RM, Catai AM. Metabolomics-Based Studies Assessing Exercise-Induced Alterations of the Human Metabolome: A Systematic Review. Metabolites. 2019;9(8):164. doi: 10.3390/metabo9080164 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref014] 14.Schranner D, Kastenmüller G, Schönfelder M, Römisch-Margl W, Wackerhage H. Metabolite Concentration Changes in Humans After a Bout of Exercise: a Systematic Review of Exercise Metabolomics Studies. Sports Med Open. 2020;6(1):11. doi: 10.1186/s40798-020-0238-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref015] 15.Latino F, Cataldi S, Carvutto R, De Candia M, D’Elia F, Patti A, et al. The Importance of Lipidomic Approach for Mapping and Exploring the Molecular Networks Underlying Physical Exercise: A Systematic Review. Int J Mol Sci. 2021;22(16):8734. doi: 10.3390/ijms22168734 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref016] 16.Tian Q, Corkum AE, Moaddel R, Ferrucci L. Metabolomic profiles of being physically active and less sedentary: a critical review. Metabolomics. 2021;17(7):68. doi: 10.1007/s11306-021-01818-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref017] 17.Khoramipour K, Sandbakk Ø, Keshteli AH, Gaeini AA, Wishart DS, Chamari K. Metabolomics in Exercise and Sports: A Systematic Review. Sports Med. 2022;52(3):547–83. doi: 10.1007/s40279-021-01582-y [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref018] 18.Chang Y-J, Tuz-Zahra F, Godbole S, Avitia Y, Bellettiere J, Rock CL, et al. Endothelial-derived cardiovascular disease-related microRNAs elevated with prolonged sitting pattern among postmenopausal women. Sci Rep. 2021;11(1):11766. doi: 10.1038/s41598-021-90154-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref019] 19.Park S. Association between polygenetic risk scores related to sarcopenia risk and their interactions with regular exercise in a large cohort of Korean adults. Clin Nutr. 2021;40(10):5355–64. doi: 10.1016/j.clnu.2021.09.003 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref020] 20.Zhou J-Y, Park S. Regular exercise, alcohol consumption, and smoking interact with the polygenetic risk scores involved in insulin sensitivity and secretion for the risk of concurrent hyperglycemia, hypertension, and dyslipidemia. Nutrition. 2021;91–92:111422. doi: 10.1016/j.nut.2021.111422 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref021] 21.Bergström G, Berglund G, Blomberg A, Brandberg J, Engström G, Engvall J, et al. The Swedish CArdioPulmonary BioImage Study: objectives and design. J Intern Med. 2015;278(6):645–59. doi: 10.1111/joim.12384 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref022] 22.Migueles JH, Cadenas-Sanchez C, Ekelund U, Delisle Nyström C, Mora-Gonzalez J, Löf M, et al. Accelerometer Data Collection and Processing Criteria to Assess Physical Activity and Other Outcomes: A Systematic Review and Practical Considerations. Sports Med. 2017;47(9):1821–45. doi: 10.1007/s40279-017-0716-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref023] 23.Aguilar-Farías N, Brown WJ, Peeters GMEEG. ActiGraph GT3X+ cut-points for identifying sedentary behaviour in older adults in free-living environments. J Sci Med Sport. 2014;17(3):293–9. doi: 10.1016/j.jsams.2013.07.002 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref024] 24.Sasaki JE, John D, Freedson PS. Validation and comparison of ActiGraph activity monitors. J Sci Med Sport. 2011;14(5):411–6. doi: 10.1016/j.jsams.2011.04.003 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref025] 25.Dunstan DW, Kingwell BA, Larsen R, Healy GN, Cerin E, Hamilton MT, et al. Breaking Up Prolonged Sitting Reduces Postprandial Glucose and Insulin Responses. Diabetes Care. 2012;35(5):976–83. doi: 10.2337/dc11-1931 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref026] 26.Hulsen T, de Vlieg J, Alkema W. BioVenn - a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams. BMC Genomics. 2008;9:488. doi: 10.1186/1471-2164-9-488 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref027] 27.Sagelv EH, Hopstock LA, Morseth B, Hansen BH, Steene-Johannessen J, Johansson J, et al. Device-measured physical activity, sedentary time, and risk of all-cause mortality: an individual participant data analysis of four prospective cohort studies. Br J Sports Med. 2023;57(22):1457–63. doi: 10.1136/bjsports-2022-106568 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref028] 28.Siegbahn A, Eriksson N, Assarsson E, Lundberg M, Ballagi A, Held C, et al. Development and validation of a quantitative Proximity Extension Assay instrument with 21 proteins associated with cardiovascular risk (CVD-21). PLoS One. 2023;18(11):e0293465. doi: 10.1371/journal.pone.0293465 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref029] 29.Yang T, Yi J, He Y, Zhang J, Li X, Ke S, et al. Associations of Dietary Fats with All-Cause Mortality and Cardiovascular Disease Mortality among Patients with Cardiometabolic Disease. Nutrients. 2022;14(17):3608. doi: 10.3390/nu14173608 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref030] 30.Zhang Y, Sun Y, Yu Q, Song S, Brenna JT, Shen Y, et al. Higher ratio of plasma omega-6/omega-3 fatty acids is associated with greater risk of all-cause, cancer, and cardiovascular mortality: a population-based cohort study in UK Biobank. medRxiv. 2024. doi: 10.1101/2023.01.16.23284631 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref031] 31.Kaikkonen JE, Kresanov P, Ahotupa M, Jula A, Mikkilä V, Viikari JSA, et al. High serum n6 fatty acid proportion is associated with lowered LDL oxidation and inflammation: the Cardiovascular Risk in Young Finns Study. Free Radic Res. 2014;48(4):420–6. doi: 10.3109/10715762.2014.883071 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref032] 32.Bajerska J, Skoczek-Rubińska A, Małczak L, Vucic V, Arsic A, Kojadinovic M, et al. Plasma fatty acid composition and some markers of dietary habits are associated with cardiovascular disease risk determined by an atherogenic plasma index in postmenopausal women. Nutr Res. 2023;115:47–60. doi: 10.1016/j.nutres.2023.05.008 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref033] 33.Climie RE, Wheeler MJ, Grace M, Lambert EA, Cohen N, Owen N, et al. Simple intermittent resistance activity mitigates the detrimental effect of prolonged unbroken sitting on arterial function in overweight and obese adults. J Appl Physiol (1985). 2018;125(6):1787–94. doi: 10.1152/japplphysiol.00544.2018 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref034] 34.Kujala UM, Mäkinen V-P, Heinonen I, Soininen P, Kangas AJ, Leskinen TH, et al. Long-term leisure-time physical activity and serum metabolome. Circulation. 2013;127(3):340–8. doi: 10.1161/CIRCULATIONAHA.112.105551 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref035] 35.Bell JA, Hamer M, Richmond RC, Timpson NJ, Carslake D, Davey Smith G. Associations of device-measured physical activity across adolescence with metabolic traits: Prospective cohort study. PLoS Med. 2018;15(9):e1002649. doi: 10.1371/journal.pmed.1002649 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref036] 36.Henson J, Edwardson CL, Davies MJ, Gill JMR, Heaney LM, Khunti K, et al. Physical activity and lipidomics in a population at high risk of type 2 diabetes mellitus. J Sports Sci. 2020;38(10):1150–60. doi: 10.1080/02640414.2020.1744836 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref037] 37.Vaara JP, Kyröläinen H, Vasankari T, Kainulainen H, Raitanen J, Kujala UM. Accelerometer-Based Sedentary Time, Physical Activity, and Serum Metabolome in Young Men. Metabolites. 2022;12(8):700. doi: 10.3390/metabo12080700 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref038] 38.Zhao S, Kusminski CM, Scherer PE. Adiponectin, Leptin and Cardiovascular Disorders. Circ Res. 2021;128(1):136–49. doi: 10.1161/circresaha.120.314458 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref039] 39.Söderberg S, Stegmayr B, Stenlund H, Sjöström L-G, Agren A, Johansson L, et al. Leptin, but not adiponectin, predicts stroke in males. J Intern Med. 2004;256(2):128–36. doi: 10.1111/j.1365-2796.2004.01351.x [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref040] 40.Eriksson M, Johnson O, Boman K, Hallmans G, Hellsten G, Nilsson TK, et al. Improved fibrinolytic activity during exercise may be an effect of the adipocyte-derived hormones leptin and adiponectin. Thromb Res. 2008;122(5):701–8. doi: 10.1016/j.thromres.2008.01.010 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref041] 41.Kostov K, Blazhev A. Use of Glycated Hemoglobin (A1c) as a Biomarker for Vascular Risk in Type 2 Diabetes: Its Relationship with Matrix Metalloproteinases-2, -9 and the Metabolism of Collagen IV and Elastin. Medicina (Kaunas). 2020;56(5):231. doi: 10.3390/medicina56050231 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref042] 42.Barreto J, Karathanasis SK, Remaley A, Sposito AC. Role of LOX-1 (Lectin-Like Oxidized Low-Density Lipoprotein Receptor 1) as a Cardiovascular Risk Predictor: Mechanistic Insight and Potential Clinical Use. Arterioscler Thromb Vasc Biol. 2021;41(1):153–66. doi: 10.1161/ATVBAHA.120.315421 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref043] 43.Chen X, Wang S, Xu W, Zhao M, Zhang Y, Xiao H. Metformin Directly Binds to MMP-9 to Improve Plaque Stability. J Cardiovasc Dev Dis. 2023;10(2):54. doi: 10.3390/jcdd10020054 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref044] 44.Jin P, Cong S. LOX-1 and atherosclerotic-related diseases. Clin Chim Acta. 2019;491:24–9. doi: 10.1016/j.cca.2019.01.006 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref045] 45.Fellström B, Helmersson-Karlqvist J, Lind L, Soveri I, Wu P-H, Thulin M, et al. Associations Between Apolipoprotein A1, High-Density Lipoprotein Cholesterol, and Urinary Cytokine Levels in Elderly Males and Females. J Interferon Cytokine Res. 2020;40(2):71–4. doi: 10.1089/jir.2019.0074 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref046] 46.Patel H, Ashton NJ, Dobson RJB, Andersson L-M, Yilmaz A, Blennow K, et al. Proteomic blood profiling in mild, severe and critical COVID-19 patients. Sci Rep. 2021;11(1):6357. doi: 10.1038/s41598-021-85877-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref047] 47.Hejazi K, Mohammad Rahimi GR, Rosenkranz SK. Effects of Exercise Training on Inflammatory and Cardiometabolic Risk Biomarkers in Patients With Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. Biol Res Nurs. 2023;25(2):250–66. doi: 10.1177/10998004221132841 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref048] 48.Hamirani YS, Pandey S, Rivera JJ, Ndumele C, Budoff MJ, Blumenthal RS, Nasir K. Markers of inflammation and coronary artery calcification: a systematic review. Atherosclerosis. 2008;201(1):1–7. doi: 10.1016/j.atherosclerosis.2008.04.045 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref049] 49.Karason K, Girerd N, Andersson-Asssarsson J, Duarte K, Taube M, Svensson P-A, et al. Heart failure in obesity: insights from proteomics in patients treated with or without weight-loss surgery. Int J Obes (Lond). 2022;46(12):2088–94. doi: 10.1038/s41366-022-01194-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref050] 50.Ponce-de-Leon M, Hannemann A, Linseisen J, Nauck M, Lerch MM, Bülow R, et al. Links between ectopic and abdominal fat and systemic inflammation: New insights from the SHIP-Trend study. Digestive and Liver Disease. 2022;54(8):1030–7. doi: 10.1016/j.dld.2022.02.003 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref051] 51.Hasan MM, Gazi MA, Das S, Fahim SM, Hossaini F, Khan A-R, et al. Gut biomolecules (I-FABP, TFF3 and lipocalin-2) are associated with linear growth and biomarkers of environmental enteric dysfunction (EED) in Bangladeshi children. Sci Rep. 2022;12(1):13905. doi: 10.1038/s41598-022-18141-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref052] 52.Peiris HN, Romero R, Vaswani K, Reed S, Gomez-Lopez N, Tarca AL, et al. Preterm labor is characterized by a high abundance of amniotic fluid prostaglandins in patients with intra-amniotic infection or sterile intra-amniotic inflammation. J Matern Fetal Neonatal Med. 2021;34(24):4009–24. doi: 10.1080/14767058.2019.1702953 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref053] 53.Hocher C-F, Chen X, Zuo J, Horvathova K, Hocher B, Krämer BK, et al. Fibroblast growth factor 23 is associated with the development of gestational diabetes mellitus. Diabetes Metab Res Rev. 2023;39(8):e3704. doi: 10.1002/dmrr.3704 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref054] 54.Sanders-van Wijk S, Tromp J, Beussink-Nelson L, Hage C, Svedlund S, Saraste A, et al. Proteomic Evaluation of the Comorbidity-Inflammation Paradigm in Heart Failure With Preserved Ejection Fraction: Results From the PROMIS-HFpEF Study. Circulation. 2020;142(21):2029–44. doi: 10.1161/CIRCULATIONAHA.120.045810 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref055] 55.Royston L, Isnard S, Perrin N, Sinyavskaya L, Berini C, Lin J, et al. Growth differentiation factor-15 as a biomarker of atherosclerotic coronary plaque: Value in people living with and without HIV. Front Cardiovasc Med. 2022;9:964650. doi: 10.3389/fcvm.2022.964650 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref056] 56.Planès C, Randrianarison NH, Charles R-P, Frateschi S, Cluzeaud F, Vuagniaux G, et al. ENaC-mediated alveolar fluid clearance and lung fluid balance depend on the channel-activating protease 1. EMBO Mol Med. 2010;2(1):26–37. doi: 10.1002/emmm.200900050 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref057] 57.Batista AP, Barbosa KF, de Azevedo RJ, Vianna VN, de Queiroz EM, Marinho CC, et al. Hypertension is associated with a variant in the RARRES2 gene in populations of Ouro Preto, Minas Gerais, Brazil: a cross-sectional study. Int J Mol Epidemiol Genet. 2021;12(3):40–51. [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref058] 58.Han Y, Hua S, Chen Y, Yang W, Zhao W, Huang F, et al. Circulating PGLYRP1 Levels as a Potential Biomarker for Coronary Artery Disease and Heart Failure. J Cardiovasc Pharmacol. 2021;77(5):578–85. doi: 10.1097/FJC.0000000000000996 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref059] 59.Lankin VZ, Tikhaze AK, Melkumyants AM. Dicarbonyl-Dependent Modification of LDL as a Key Factor of Endothelial Dysfunction and Atherosclerotic Vascular Wall Damage. Antioxidants. 2022;11(8):1565. doi: 10.3390/antiox11081565 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref060] 60.Kanno Y. The uPA/uPAR System Orchestrates the Inflammatory Response, Vascular Homeostasis, and Immune System in Fibrosis Progression. IJMS. 2023;24(2):1796. doi: 10.3390/ijms24021796 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref061] 61.Lind L, Titova O, Zeng R, Zanetti D, Ingelsson M, Gustafsson S, et al. Plasma Protein Profiling of Incident Cardiovascular Diseases: A Multisample Evaluation. Circ Genom Precis Med. 2023;16(6):e004233. doi: 10.1161/CIRCGEN.123.004233 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref062] 62.Gustafsson S, Lampa E, Jensevik Eriksson K, Butterworth AS, Elmståhl S, Engström G, et al. Markers of imminent myocardial infarction. Nat Cardiovasc Res. 2024;3(2):130–9. doi: 10.1038/s44161-024-00422-2 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref063] 63.Ekblom-Bak E, Börjesson M, Ekblom Ö, Angerås O, Bergman F, Berntsson C, et al. Accelerometer derived physical activity and subclinical coronary and carotid atherosclerosis: cross-sectional analyses in 22 703 middle-aged men and women in the SCAPIS study. BMJ Open. 2023;13(11):e073380. doi: 10.1136/bmjopen-2023-073380 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref064] 64.Carlsson AC, Ruge T, Kjøller E, Hilden J, Kolmos HJ, Sajadieh A, et al. 10-Year Associations Between Tumor Necrosis Factor Receptors 1 and 2 and Cardiovascular Events in Patients With Stable Coronary Heart Disease: A CLARICOR (Effect of Clarithromycin on Mortality and Morbidity in Patients With Ischemic Heart Disease) Trial Substudy. J Am Heart Assoc. 2018;7(9):e008299. doi: 10.1161/JAHA.117.008299 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref065] 65.Probert L. TNF and its receptors in the CNS: The essential, the desirable and the deleterious effects. Neuroscience. 2015;302:2–22. doi: 10.1016/j.neuroscience.2015.06.038 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref066] 66.Thougaard E, Nielsen PV, Forsberg A, Phuong V, Velasco AM, Wlodarczyk A, et al. Systemic treatment with a selective TNFR2 agonist alters the central and peripheral immune responses and transiently improves functional outcome after experimental ischemic stroke. J Neuroimmunol. 2023;385:578246. doi: 10.1016/j.jneuroim.2023.578246 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref067] 67.Rip J, Nierman MC, Wareham NJ, Luben R, Bingham SA, Day NE, et al. Serum lipoprotein lipase concentration and risk for future coronary artery disease: the EPIC-Norfolk prospective population study. Arterioscler Thromb Vasc Biol. 2006;26(3):637–42. doi: 10.1161/01.ATV.0000201038.47949.56 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref068] 68.Elosua R, Cupples LA, Fox CS, Polak JF, D’Agostino RA Sr, Wolf PA, et al. Association between well-characterized lipoprotein-related genetic variants and carotid intimal medial thickness and stenosis: The Framingham Heart Study. Atherosclerosis. 2006;189(1):222–8. doi: 10.1016/j.atherosclerosis.2005.12.005 [DOI] [PubMed] [Google Scholar]

[pone.0325720.ref069] 69.Black MA, Cable NT, Thijssen DH, Green DJ. Impact of age, sex, and exercise on brachial artery flow-mediated dilatation. Am J Physiol Heart Circ Physiol. 2009;297(3):H1109–16. doi: 10.1152/ajpheart.00226.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref070] 70.Pierce GL, Eskurza I, Walker AE, Fay TN, Seals DR. Sex-specific effects of habitual aerobic exercise on brachial artery flow-mediated dilation in middle-aged and older adults. Clin Sci (Lond). 2011;120(1):13–23. doi: 10.1042/CS20100174 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0325720.ref071] 71.Barker J, Smith Byrne K, Doherty A, Foster C, Rahimi K, Ramakrishnan R, et al. Physical activity of UK adults with chronic disease: cross sectional analysis of accelerometer measured physical activity in 96 706 UK Biobank participants. Int J Epidemiol. 2019;48(4):1386–1386. doi: 10.1093/ije/dyz148 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Associations between physical activity and CVD-related metabolomic and proteomic biomarkers

Örjan Ekblom

Harry Björkbacka

Mats Börjesson

Elin Ekblom-Bak

Anders Blomberg

Kenneth Caidahl

Ewa Ehrenborg

Gunnar Engström

Jan Engvall

David Erlinge

Tove Fall

Bruna Gigante

Anders Gummesson

Tomas Jernberg

Lars Lind

David Molnar

Jonas Oldgren

Aidin Rawshani

Johan Sundström

Stefan Söderberg

Patrik Wennberg

Carl Johan Östgren

Roles

Abstract

Aim

Method

Results

Conclusion

Introduction

Materials and methods

SCAPIS main cohort (discovery sample)

SCAPIS pilot cohort (validation sample)

Assessment of the PA classes

Blood sampling and analysis

Table 2. Markers significantly related to SED. Values are beta coefficients of the unadjusted model, model 1 and model 2.

Table 4. Markers significantly related to MVPA. Values are beta coefficients from unadjusted model, model 1, and model 2.

Confounders

Statistics

Fig 1. Proportional Wenn-diagram for metabolomic markers significantly related to the three PA classes, SED (light blue), prolSED (dark blue), and MVPA (yellow).

Fig 2. Proportional Wenn-diagram for proteomic markers significantly related to the three PA classes, SED (light blue), prolSED (dark blue), and MVPA (yellow).

Results

Table 1. Participant characteristics in the two samples.

Metabolomic markers

Table 3. Markers significantly related to prolSED. Values are beta coefficients from unadjusted model, model 1, and model 2.

Proteomic markers

Non-linear associations.

Interaction between PA class and sex.

Interaction between PA class and waist circumference.

Discussion

Metabolomic profiling

Proteomic profiling

Interactions

Strengths and limitations

Conclusions

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Author response to Decision Letter 0

Decision Letter 0

Hidetaka Hamasaki

Roles

Author response to Decision Letter 1

Decision Letter 1

Hidetaka Hamasaki

Roles

Author response to Decision Letter 2

Decision Letter 2

Hidetaka Hamasaki

Roles

Acceptance letter

Hidetaka Hamasaki

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS