Search

Computational models of human glucose homeostasis can provide insight into the physiological processes underlying the observed inter-individual variability in glucose regulation. Modelling approaches ranging from "bottom-up" mechanistic models to "top-down" data-driven techniques have been applied to untangle the complex interactions underlying progressive disturbances in glucose homeostasis. While both approaches offer distinct benefits, a combined approach taking the best of both worlds has yet to be explored. Here, we propose a sequential combination of a mechanistic and a data-driven modeling approach to quantify individuals' glucose and insulin responses to an oral glucose tolerance test, using cross sectional data from 2968 individuals from a large observational prospective population-based cohort, the Maastricht Study. The best predictive performance, measured by R2 and mean squared error of prediction, was achieved with personalized mechanistic models alone. The addition of a data-driven model did not improve predictive performance. The personalized mechanistic models consistently outperformed the data-driven and the combined model approaches, demonstrating the strength and suitability of bottom-up mechanistic models in describing the dynamic glucose and insulin response to oral glucose tolerance tests., Competing Interests: This study received funding from Janssen-Cilag, Novo Nordisk Farma., Sanofi-Aventis Netherlands, DSM Nutritional products, FrieslandCampina, and Danone Nutricia Research. There are no patents, products in development or marketed products to declare. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials., (Copyright: © 2023 Erdős et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Background: Detrimental associations of sedentary behaviour (time spent sitting) with musculoskeletal pain (MSP) conditions have been observed. However, findings on those with, or at risk of, type 2 diabetes (T2D) have not been reported. We examined the linear and non-linear associations of device-measured daily sitting time with MSP outcomes according to glucose metabolism status (GMS)., Methods: Cross-sectional data from 2827 participants aged 40-75 years in the Maastricht Study (1728 with normal glucose metabolism (NGM); 441 with prediabetes; 658 with T2D), for whom valid data were available on activPAL-derived daily sitting time, MSP [neck, shoulder, low back, and knee pain], and GMS. Associations were examined by logistic regression analyses, adjusted serially for relevant confounders, including moderate-to-vigorous intensity physical activity (MVPA) and body mass index (BMI). Restricted cubic splines were used to further examine non-linear relationships., Results: The fully adjusted model (including BMI, MVPA, and history of cardiovascular disease) showed daily sitting time to be significantly associated with knee pain in the overall sample (OR = 1.07, 95%CI: 1.01-1.12) and in those with T2D (OR = 1.11, 95%CI: 1.00-1.22); this was not statistically significant in those with prediabetes (OR = 1.04, 95%CI: 0.91-1.18) or NGM (OR = 1.05, 95%CI: 0.98-1.13). There were no statistically significant associations between daily sitting time and neck, shoulder, or low back pain in any of the models. Furthermore, the non-linear relationships were statistically non-significant., Conclusion: Among middle-aged and older adults with T2D, daily sitting time was significantly associated with higher odds of knee pain, but not with neck, shoulder, or low back pain. No significant association was observed in those without T2D for neck, shoulder, low back, or knee pain. Future studies, preferably those utilising prospective designs, could examine additional attributes of daily sitting (e.g., sitting bouts and domain-specific sitting time) and the potential relationships of knee pain with mobility limitations., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2023 Dzakpasu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Background: Closed-loop insulin delivery systems, which integrate continuous glucose monitoring (CGM) and algorithms that continuously guide insulin dosing, have been shown to improve glycaemic control. The ability to predict future glucose values can further optimize such devices. In this study, we used machine learning to train models in predicting future glucose levels based on prior CGM and accelerometry data., Methods: We used data from The Maastricht Study, an observational population-based cohort that comprises individuals with normal glucose metabolism, prediabetes, or type 2 diabetes. We included individuals who underwent >48h of CGM (n = 851), most of whom (n = 540) simultaneously wore an accelerometer to assess physical activity. A random subset of individuals was used to train models in predicting glucose levels at 15- and 60-minute intervals based on either CGM data or both CGM and accelerometer data. In the remaining individuals, model performance was evaluated with root-mean-square error (RMSE), Spearman's correlation coefficient (rho) and surveillance error grid. For a proof-of-concept translation, CGM-based prediction models were optimized and validated with the use of data from individuals with type 1 diabetes (OhioT1DM Dataset, n = 6)., Results: Models trained with CGM data were able to accurately predict glucose values at 15 (RMSE: 0.19mmol/L; rho: 0.96) and 60 minutes (RMSE: 0.59mmol/L, rho: 0.72). Model performance was comparable in individuals with type 2 diabetes. Incorporation of accelerometer data only slightly improved prediction. The error grid results indicated that model predictions were clinically safe (15 min: >99%, 60 min >98%). Our prediction models translated well to individuals with type 1 diabetes, which is reflected by high accuracy (RMSEs for 15 and 60 minutes of 0.43 and 1.73 mmol/L, respectively) and clinical safety (15 min: >99%, 60 min: >91%)., Conclusions: Machine learning-based models are able to accurately and safely predict glucose values at 15- and 60-minute intervals based on CGM data only. Future research should further optimize the models for implementation in closed-loop insulin delivery systems., Competing Interests: This study received funding from Janssen-Cilag B.V, Novo Nordisk Farma B.V., Sanofi-Aventis Netherlands B.V and Medtronic. There are no patents, products in development or marketed products to declare. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors. N.C. Schaper, R.M.A. Henry, and M.C.G.J. Brouwers were supported by Medtronic (External Research Program). Medtronic did not direct design, conduct, or outcomes of this study. The authors declare that there are no other relationships or activities that might bias, or be perceived to bias, their work.

Background: This study aimed to determine the within-person and between-persons associations of low-grade inflammation (LGI) and endothelial dysfunction (ED) with echocardiographic measures related to diastolic dysfunction (DD) in two general populations and whether these associations differed by sex., Methods: Biomarkers and echocardiographic measures were measured at both baseline and follow-up in the Hoorn Study (n = 383) and FLEMENGHO (n = 491). Individual biomarker levels were combined into either a Z-score of LGI (CRP, SAA, IL-6, IL-8, TNF-α and sICAM-1) or ED (sICAM-1, sVCAM-1, sE-selectin and sTM). Mixed models were used to determine within-person and between-persons associations of biomarker Z-scores with left ventricular ejection fraction (LVEF), left ventricular mass index (LVMI) and left atrial volume index (LAVI). These associations were adjusted for a-priori selected confounders., Results: Overall Z-scores for LGI or ED were not associated with echocardiographic measures. Effect modification by sex was apparent for ED with LVEF in both cohorts (P-for interaction = 0.08 and 0.06), but stratified results were not consistent. Effect modification by sex was apparent for TNF-α in the Hoorn Study and E-selectin in FLEMENGHO with LVEF (P-for interaction≤0.05). In the Hoorn Study, women whose TNF-α levels increased with 1-SD over time had a decrease in LVEF of 2.2 (-4.5;0.01) %. In FLEMENGHO, men whose E-selectin levels increased with 1-SD over time had a decrease in LVEF of 1.6 (-2.7;-0.5) %., Conclusion: Our study did not show consistent associations of LGI and ED with echocardiographic measures. Some evidence of effect modification by sex was present for ED and specific biomarkers., Competing Interests: The authors have declared that no competing interests exist.

Background: Beyond antihyperglycemic effects, metformin may improve cardiovascular outcomes. Patients with type 2 diabetes often have an elevated plasma level of N-terminal pro B-type as a marker of (sub) clinical cardiovascular disease. We studied whether metformin was associated with a reduction in the serum level of N-terminal pro B-type natriuretic peptide (NT-proBNP) in these patients., Methods: In the HOME trial 390 insulin-treated patients with type 2 diabetes were randomized to 850 mg metformin or placebo three times daily. Plasma samples were drawn at baseline, 4, 17, 30, 43 and 52 months. In a post-hoc analysis we analyzed the change in NT-proBNP in both groups. We used a longitudinal mixed model analysis adjusting for age, sex and prior cardiovascular disease. In a secondary analysis we assessed a possible immediate treatment effect post baseline., Results: Metformin did not affect NT-proBNP levels over time in the primary analysis (-1% [95%CI -4;3, p = 0.62]). In the secondary analysis there was also no sustained time independent immediate treatment effect (initial increase of 17% [95%CI 4;30, p = 0.006] followed by yearly decrease of -4% [95%CI -7;0, p = 0.07])., Conclusions: Metformin as compared to placebo did not affect NT-proBNP plasma levels in this 4.3-year placebo-controlled trial. Potential cardioprotective effects of metformin cannot be explained by changes in cardiac pressures or volumes to the extent reflected by NT-proBNP., Competing Interests: The authors have read the journal’s policy and have the following competing interests: the HOME trial was supported by grants from Altana, Lifescan, Merck Santé, Merck Sharp & Dohme, and Novo Nordisk awarded to Bethesda General Hospital, Hoogeveen, The Netherlands (1996- 2000), guided by a registered accountant, and raised by AK, initiator and principal investigator of the HOME study. This does not alter our adherence to PLOS ONE policies on sharing data and materials. There are no patents, products in development or marketed products associated with this research to declare.

In response to a study previously published in PLOS Biology, this Formal Comment thoroughly examines the concept of 'glucotypes' with regard to its generalisability, interpretability and relationship to more traditional measures used to describe data from continuous glucose monitoring., Competing Interests: The authors have declared that no competing interest exists.

Objective: Low individual socioeconomic status (SES) is known to be associated with a higher risk of type 2 diabetes mellitus (T2DM), but the extent to which the local context in which people live may influence T2DM rates remains unclear. This study examines whether living in a low property value neighbourhood is associated with higher rates of T2DM independently of individual SES., Research Design and Methods: Using cross-sectional data from the Maastricht Study (2010-2013) and geographical data from Statistics Netherlands, multilevel logistic regression was used to assess the association between neighbourhood property value and T2DM. Individual SES was based on education, occupation and income. Of the 2,056 participants (aged 40-75 years), 494 (24%) were diagnosed with T2DM., Results: Individual SES was strongly associated with T2DM, but a significant proportion of the variance in T2DM was found at the neighbourhood level (VPC = 9.2%; 95% CI = 5.0%-16%). Participants living in the poorest neighbourhoods had a 2.38 times higher odds ratio of T2DM compared to those living in the richest areas (95% CI = 1.58-3.58), independently of individual SES., Conclusions: Neighbourhood property value showed a significant association with T2DM, suggesting the usefulness of area-based programmes aimed at improving neighbourhood characteristics in order to tackle inequalities in T2DM., Competing Interests: This study received funding from Janssen-Cilag B.V, Novo Nordisk Farma B.V. and Sanofi-Aventis Netherlands B.V. There are no patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors. Hans Bosma serves as guest editor for the Health Inequities and Disparities Call for papers. This does not alter our adherence to PLOS ONE policies on sharing data and materials.

Background: Small-molecules that disrupt the binding between glucokinase and glucokinase regulatory protein (GKRP) in the liver represent a potential new class of glucose-lowering drugs. It will, however, take years before their effects on clinically relevant cardiovascular endpoints are known. The purpose of this study was to estimate the effects of these drugs on cardiorenal outcomes by studying variants in the GKRP gene (GCKR) that mimic glucokinase-GKRP disruptors., Methods: The MEDLINE and EMBASE databases were searched for studies reporting on the association between GCKR variants (rs1260326, rs780094, and rs780093) and coronary artery disease (CAD), estimated glomerular filtration rate (eGFR), and chronic kidney disease (CKD)., Results: In total 5 CAD studies (n = 274,625 individuals), 7 eGFR studies (n = 195,195 individuals), and 4 CKD studies (n = 31,642 cases and n = 408,432 controls) were included. Meta-analysis revealed a significant association between GCKR variants and CAD (OR:1.02 per risk allele, 95%CI:1.00-1.04, p = 0.01). Sensitivity analyses showed that replacement of one large, influential CAD study by two other, partly overlapping studies resulted in similar point estimates, albeit less precise (OR:1.02; 95%CI:0.98-1.06 and OR: 1.02; 95%CI: 0.99-1.04). GCKR was associated with an improved eGFR (+0.49 ml/min, 95%CI:0.10-0.89, p = 0.01) and a trend towards protection from CKD (OR:0.98, 95%CI:0.95-1.01, p = 0.13)., Conclusion: This study suggests that increased glucokinase-GKRP disruption has beneficial effects on eGFR, but these may be offset by a disadvantageous effect on coronary artery disease risk. Further studies are warranted to elucidate the mechanistic link between hepatic glucose metabolism and eGFR., Competing Interests: The authors have declared that no competing interests exist.

[This corrects the article DOI: 10.1371/journal.pone.0191172.].

Aims: Carotid intima media thickness (CIMT) predicts cardiovascular (CVD) events, but the predictive value of CIMT change is debated. We assessed the relation between CIMT change and events in individuals at high cardiovascular risk., Methods and Results: From 31 cohorts with two CIMT scans (total n = 89070) on average 3.6 years apart and clinical follow-up, subcohorts were drawn: (A) individuals with at least 3 cardiovascular risk factors without previous CVD events, (B) individuals with carotid plaques without previous CVD events, and (C) individuals with previous CVD events. Cox regression models were fit to estimate the hazard ratio (HR) of the combined endpoint (myocardial infarction, stroke or vascular death) per standard deviation (SD) of CIMT change, adjusted for CVD risk factors. These HRs were pooled across studies. In groups A, B and C we observed 3483, 2845 and 1165 endpoint events, respectively. Average common CIMT was 0.79mm (SD 0.16mm), and annual common CIMT change was 0.01mm (SD 0.07mm), both in group A. The pooled HR per SD of annual common CIMT change (0.02 to 0.43mm) was 0.99 (95% confidence interval: 0.95-1.02) in group A, 0.98 (0.93-1.04) in group B, and 0.95 (0.89-1.04) in group C. The HR per SD of common CIMT (average of the first and the second CIMT scan, 0.09 to 0.75mm) was 1.15 (1.07-1.23) in group A, 1.13 (1.05-1.22) in group B, and 1.12 (1.05-1.20) in group C., Conclusions: We confirm that common CIMT is associated with future CVD events in individuals at high risk. CIMT change does not relate to future event risk in high-risk individuals.

Background: Chronic kidney disease, which is defined as having a reduced kidney function (estimated glomerular filtration rate (eGFR)) and/or signs of kidney damage (albuminuria), is highly prevalent in Western society and is associated with adverse health outcomes, such as cardiovascular disease. This warrants a search for risk factors of lower eGFR and higher albuminuria. Physical activity and sedentary behavior may be such risk factors., Objective: To examine associations of physical activity (total, high, low), sedentary time and sedentary behavior patterns (breaks, prolonged bouts, average bout duration) with eGFR and albuminuria., Methods: We examined these associations in 2,258 participants of the Maastricht Study (average age 60.1±8.1 years; 51.3% men), who wore an accelerometer 24h/day on 7 consecutive days. Associations with continuous eGFR and categories of urinary albumin excretion (UAE; <15 [reference category], 15-<30, ≥30 mg/24h) were evaluated with linear regression analyses and multinomial logistic regression analyses, respectively., Results: After adjustment for potential confounders, each extra hour of total physical activity was associated with a more favorable kidney function (betaeGFR = 2.30 (95%CI = 1.46; 3.14)), whereas each extra hour of sedentary behavior was associated with a more adverse kidney function (betaeGFR = -0.71 (-1.08; -0.35)). Also, compared to individuals with the lowest levels of total physical activity, individuals with the highest levels had less kidney damage (OR15-<30mg/24h = 0.63 (0.41; 0.96), OR≥30mg/24h = 0.84 (0.53; 1.35). An extra hour of sedentary behavior was associated with more kidney damage (OR15-<30 mg/24h = 1.11 (1.01; 1.22), OR≥30 mg/24h = 1.10 (0.99; 1.22)). Further, a highly sedentary pattern was associated with a more adverse kidney function, but no association was seen with kidney damage., Conclusions: Physical activity and sedentary behavior were associated with kidney function and kidney damage. Additionally, sedentary behavior patterns were associated with kidney function. Causal studies are required to examine whether this indeed implicates that prevention strategies should focus not only on increasing physical activity, but on reducing sedentary behavior as well.

Objective: Microvascular dysfunction is an important underlying mechanism of microvascular diseases. Determinants (age, sex, hypertension, dyslipidemia, hyperglycemia, obesity, and smoking) of macrovascular diseases affect large-artery endothelial function. These risk factors also associate with microvascular diseases. We hypothesized that they are also determinants of microvascular (endothelial) function., Methods: In The Maastricht Study, a type 2 diabetes-enriched population-based cohort study (n = 1991, 51% men, aged 59.7±8.2 years), we determined microvascular function as flicker light-induced retinal arteriolar %-dilation and heat-induced skin %-hyperemia. Multiple linear regression analyses were used to assess the associations of cardiovascular risk factors (age, sex, waist circumference, total-to-high-density lipoprotein (HDL) cholesterol ratio, fasting plasma glucose (FPG), 24-h systolic blood pressure, and cigarette smoking) with retinal and skin microvascular function., Results: In multivariate analyses, age and FPG were inversely associated with retinal and skin microvascular function (regression coefficients per standard deviation (SD) were -0.11SD (95%CI: -0.15;-0.06) and -0.12SD (-0.17;-0.07) for retinal arteriolar %-dilation and -0.10SD (-0.16;-0.05) and -0.11SD (-0.17;-0.06) for skin %-hyperemia, respectively. Men and current smokers had -0.43SD (-0.58;-0.27) and -0.32SD (-0.49;-0.15) lower skin %-hyperemia, respectively. 24-h systolic blood pressure, waist circumference, and total-to-HDL cholesterol ratio were not statistically significantly associated with these microvascular functions., Conclusions: Associations between cardiovascular risk factors and retinal and skin microvascular function show a pattern that is partly similar to the associations between cardiovascular risk factors and macrovascular function. Impairment of microvascular function may constitute a pathway through which an adverse cardiovascular risk factor pattern may increase risk of diseases that are partly or wholly of microvascular origin.

Aims: Tailored, patient-centred innovations are needed in the care for persons with type 2 diabetes mellitus (T2DM), in particular those with insufficient glycaemic control. Therefore, this study sought to assess their biopsychosocial characteristics and explore whether distinct biopsychosocial profiles exist within this subpopulation, which differ in health-related quality of life (HRQoL)., Methods: Cross-sectional study based on data from The Maastricht Study, a population-based cohort study focused on the aetiology, pathophysiology, complications, and comorbidities of T2DM. We analysed associations and clustering of glycaemic control and HRQoL with 38 independent variables (i.e. biopsychosocial characteristics) in different subgroups and using descriptive analyses, latent class analysis (LCA), and logistic regressions., Results: Included were 840 persons with T2DM, mostly men (68.6%) and with a mean age of 62.6 (±7.7) years. Mean HbA1c was 7.1% (±3.2%); 308 patients (36.7%) had insufficient glycaemic control (HbA1c>7.0% [53 mmol/mol]). Compared to those with sufficient control, these patients had a significantly worse-off status on multiple biopsychosocial factors, including self-efficacy, income, education and several health-related characteristics. Two 'latent classes' were identified in the insufficient glycaemic control subgroup: with low respectively high HRQoL. Of the two, the low HRQoL class comprised about one-fourth of patients and had a significantly worse biopsychosocial profile., Conclusions: Insufficient glycaemic control, particularly in combination with low HRQoL, is associated with a generally worse biopsychosocial profile. Further research is needed into the complex and multidimensional causal pathways explored in this study, so as to increase our understanding of the heterogeneous care needs and preferences of persons with T2DM, and translate this knowledge into tailored care and support arrangements.