Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Estimating the Direct Effect between Dietary Macronutrients and Cardiometabolic Disease, Accounting for Mediation by Adiposity and Physical Activity

Pomares-Millan, Hugo LU orcid ; Atabaki-Pasdar, Naeimeh LU orcid ; Coral, Daniel LU orcid ; Johansson, Ingegerd ; Giordano, Giuseppe N. LU and Franks, Paul W. LU (2022) In Nutrients 14(6).
Abstract

Assessing the causal effects of individual dietary macronutrients and cardiometabolic disease is challenging owing to the complexity to distinguish direct effects from those mediated or confounded by other factors. To estimate these effects, intake of protein, carbohydrate, sugar, fat, and its subtypes were obtained using food frequency data derived from a Swedish population-based cohort (n~60,000). Data on clinical outcomes (i.e., type 2 diabetes (T2D) and cardiovascular disease (CVD) incidence) were obtained by linking health registry data. We assessed the magnitude of direct and mediated effects of diet, adiposity and physical activity on T2D and CVD using structural equation modelling (SEM). To strengthen causal inference, we used... (More)

Assessing the causal effects of individual dietary macronutrients and cardiometabolic disease is challenging owing to the complexity to distinguish direct effects from those mediated or confounded by other factors. To estimate these effects, intake of protein, carbohydrate, sugar, fat, and its subtypes were obtained using food frequency data derived from a Swedish population-based cohort (n~60,000). Data on clinical outcomes (i.e., type 2 diabetes (T2D) and cardiovascular disease (CVD) incidence) were obtained by linking health registry data. We assessed the magnitude of direct and mediated effects of diet, adiposity and physical activity on T2D and CVD using structural equation modelling (SEM). To strengthen causal inference, we used Mendelian randomization (MR) to model macronutrient intake exposures against clinical outcomes. We identified likely causal effects of genetically predicted carbohydrate intake (including sugar intake) and T2D, independent of adiposity and physical activity. Pairwise, serial-and parallel-mediational configurations yielded similar results. In the integrative genomic analyses, the candidate causal variant localized to the established type 2 diabetes gene TCF7L2. These findings may be informative when considering which dietary modifications included in nutritional guidelines are most likely to elicit health-promoting effects.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Adiposity, Cardiometabolic risk, Cardiovascular disease, Causal inference, Macronutrient intake, Mediation, Physical activity
in
Nutrients
volume
14
issue
6
article number
1218
publisher
MDPI AG
external identifiers
  • scopus:85126465144
  • pmid:35334875
ISSN
2072-6643
DOI
10.3390/nu14061218
language
English
LU publication?
yes
id
266b179f-12e9-4208-bc07-a0d336dc2983
date added to LUP
2022-05-20 16:38:09
date last changed
2024-06-13 12:14:10
@article{266b179f-12e9-4208-bc07-a0d336dc2983,
  abstract     = {{<p>Assessing the causal effects of individual dietary macronutrients and cardiometabolic disease is challenging owing to the complexity to distinguish direct effects from those mediated or confounded by other factors. To estimate these effects, intake of protein, carbohydrate, sugar, fat, and its subtypes were obtained using food frequency data derived from a Swedish population-based cohort (n~60,000). Data on clinical outcomes (i.e., type 2 diabetes (T2D) and cardiovascular disease (CVD) incidence) were obtained by linking health registry data. We assessed the magnitude of direct and mediated effects of diet, adiposity and physical activity on T2D and CVD using structural equation modelling (SEM). To strengthen causal inference, we used Mendelian randomization (MR) to model macronutrient intake exposures against clinical outcomes. We identified likely causal effects of genetically predicted carbohydrate intake (including sugar intake) and T2D, independent of adiposity and physical activity. Pairwise, serial-and parallel-mediational configurations yielded similar results. In the integrative genomic analyses, the candidate causal variant localized to the established type 2 diabetes gene TCF7L2. These findings may be informative when considering which dietary modifications included in nutritional guidelines are most likely to elicit health-promoting effects.</p>}},
  author       = {{Pomares-Millan, Hugo and Atabaki-Pasdar, Naeimeh and Coral, Daniel and Johansson, Ingegerd and Giordano, Giuseppe N. and Franks, Paul W.}},
  issn         = {{2072-6643}},
  keywords     = {{Adiposity; Cardiometabolic risk; Cardiovascular disease; Causal inference; Macronutrient intake; Mediation; Physical activity}},
  language     = {{eng}},
  month        = {{03}},
  number       = {{6}},
  publisher    = {{MDPI AG}},
  series       = {{Nutrients}},
  title        = {{Estimating the Direct Effect between Dietary Macronutrients and Cardiometabolic Disease, Accounting for Mediation by Adiposity and Physical Activity}},
  url          = {{http://dx.doi.org/10.3390/nu14061218}},
  doi          = {{10.3390/nu14061218}},
  volume       = {{14}},
  year         = {{2022}},
}