Skin autofluorescence in the general population: associations and prediction
van Waateringe, Robert Paul
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
van Waateringe, R. P. (2019). Skin autofluorescence in the general population: associations and prediction. Rijksuniversiteit Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Serum free thyroxine has opposite effects on incident
type 2 diabetes and cardiovascular disease in the general
population
R.P. van Waateringe, S.N. Slagter, J.V. van Vliet-Ostaptchouk, A.C. Muller Kobold, H.J.C.M. Wouters, R. Graaff, A.D. Paterson, A.J. Smit, H.L. Lutgers, M.M. van der Klauw,
B.H.R. Wolffenbuttel Submitted
Abstract
Background Skin autofluorescence (SAF), a non-invasive measure of tissue advanced
glycation endproducts, predicts cardiovascular disease (CVD) and type 2 diabetes (T2D). Thyroid hormone (TH) levels may affect this association. In the present study, we examined whether TH levels predict 4-years risk of T2D, CVD and mortality in the general population independently of SAF.
Methods We included 27,086 participants (age 44±12yrs (mean±SD), BMI 25.9±4.2 kg/
m2) from the Dutch Lifelines Cohort Study, who had both SAF and TH levels (TSH, FT4 and
FT3) measured at baseline, and were not known to have diabetes or CVD at baseline, or using medication influencing TH levels (including levothyroxine). Diagnosis of incident T2D was by self-report, fasting blood glucose ≥7.0 mmol/l or HbA1c ≥6.5% at follow-up; diagnosis of incident CVD was by self-report. Metabolic syndrome (MetS) was defined by NCEP-ATPIII criteria. Mortality was ascertained with the Municipal Personal Records Database. The influence of TH levels on the composite outcome of incident T2D, CVD and mortality, and these outcomes separately, was evaluated with logistic regression.
Results FT4 and FT3 levels were positively associated with SAF. These associations,
however, became non-significant when adjusted for other factors influencing SAF, including age, glycaemic variables, coffee consumption, smoking and renal function. After a median follow-up of 3.9 (range 0.5-9) years, 372 participants had died, 496 developed CVD, and 381 developed T2D. TH levels were not significantly associated with the composite outcome. Subjects with incident T2D were significantly older (54 vs 44 yrs), had lower FT4 compared to those without diabetes (15.2 (IQR 13.8-16.6) vs 15.6 (IQR 14.3-17.0) pmol/l, p=9.3x10-7), but similar FT3 and TSH, while those with new-onset
CVD were older (57 vs 44 yrs) and had borderline significantly higher FT4 levels (15.8 (IQR14.4-17.3), p=0.06 vs those remaining without event). Logistic regression showed that each 1.0 pmol/l lower level of FT4 was associated with a 9-11 % higher T2D risk, adjusted for SAF, age, gender and BMI. After further adjustment for glucose, blood pressure, cholesterol and current smoking, FT4 retained its association (OR=0.90, p=2.0x10-4) with
T2D. In contrast, each 1.0 pmol/l higher FT4 was associated with a 5% increased CVD risk, adjusted for SAF, age, gender and BMI, but this association became non-significant after adjustment for blood pressure, cholesterol and smoking. In the multivariate models, higher FT3 was associated with increased risk of both T2D and CVD, when adjusted for SAF, age, gender and BMI, and with CVD after additional adjustment for actual BP levels, cholesterol and smoking.
Conclusions Serum FT4 levels have opposite effects on the risk of development of T2D
and CVD in the general population, independent of SAF, and the presence of metabolic syndrome. Higher FT3 is independently associated with CVD risk. Next steps will be to elucidate underlying mechanism(s).
Introduction
The worldwide prevalence of type 2 diabetes is increasing rapidly and is estimated to be close to 650 million in 2040 (1). Cardiovascular complications are the main drivers of increased morbidity and premature mortality in diabetes (2-4). The role of advanced glycation end products (AGEs) in ageing and in the pathophysiology of diabetes-related complications has been confirmed in several studies (5-7). AGEs may form cross-links between tissue proteins in the vascular wall, causing increased vascular stiffness and elevated blood pressure (8, 9). Moreover, binding of circulating AGEs to its receptor (RAGE) and their uptake into the vessel wall may accelerate the progression of atherosclerosis (10, 11). The accumulation of AGEs can be assessed non-invasively by measuring skin autofluorescence (SAF) (12). SAF increases with ageing, and is elevated in non-diabetic individuals with the metabolic syndrome (13). SAF is strongly associated with long-term cardiovascular complications and mortality in type 2 diabetes (14-17). Earlier, we have shown that SAF was significantly associated with the risk of incident type 2 diabetes, cardiovascular disease and mortality, independently of several conventional risk factors (18) (v Waateringe et al, Diabetologia).
Thyroid hormone plays a pivotal role in human metabolism. Population-based studies have assessed the effects of thyroid hormone levels on the risk of type 2 diabetes and cardiovascular disease. In the Rotterdam study, both low and low-normal thyroid function were a risk factor for incident diabetes, especially in people with prediabetes (19). This is in agreement with our earlier report on the association between low FT4 levels and metabolic syndrome (20). On the other hand, higher FT4 levels have been associated with higher coronary artery calcification scores, and higher incidence of atherosclerotic events (21). Thyroid hormone levels may influence tissue turnover. Experimental studies have suggested that thyroid hormone imbalance influences structural wall changes in the aorta of rats (22). In addition, T3 may stimulate arterial smooth muscle cell proliferation (23), and directly regulates collagen synthesis and collagen cross-linking and thereby influences bone quality (24). Similarly, lowered thyroid hormone levels were associated with higher fructosamine levels, and thereby may -indirectly- influence skin autofluorescence (25).
Therefore, the goal of this study was to assess whether thyroid hormone levels are associated with skin autofluorescence measurements and whether they modulate the association between SAF and the development of type 2 diabetes, cardiovascular disease and mortality in the general population.
Materials and methods
Participants
Subjects included were participants from the Lifelines Cohort Study, a large population-based cohort study in the northern region of The Netherlands. Lifelines is a multi-disciplinary prospective population-based cohort study examining in a unique three-generation design the health and health-related behaviours of 167,729 persons living in the North of The Netherlands. It employs a broad range of investigative procedures in assessing the biomedical, socio-demographic, behavioural, physical and psychological factors which contribute to the health and disease of the general population, with a special focus on multi-morbidity and complex genetics (20, 26, 27). All participants have provided written informed consent before participating in the study, which has been approved by the Medical Ethics Review Committee of the University Medical Center Groningen.
For the present study, we included all subjects of Western European descent between 18 and 90 years of age in whom both a validated SAF measurement and thyroid hormone measurements were available at baseline and prospective follow-up was performed between January 2014 and January 2018 (Figure 1). There was no relevant
difference in sex distribution, age, body mass index, and glycaemic variables between those with and without SAF measurements and those with or without thyroid hormone levels (data not shown). As Lifelines decided to stop the measurements of thyroid hormones in the beginning of 2012, only a subset of its participants has thyroid hormone measurements and follow-up data available (n=31,760, Figure 1). We excluded subjects who at baseline had clinical cardiovascular disease, type 1 diabetes, type 2 diabetes, or reported a history of gestational diabetes, and those with suppressed TSH (<0.4 mU/l) or severely elevated TSH (>10 mU/l). We also excluded participants who were treated because of a thyroid disorder, or who were current users of drugs known to influence thyroid hormone parameters, like lithium and amiodarone. Furthermore, individuals with new type 1 diabetes and a new history of gestational diabetes at follow-up were removed from the dataset, as these are disorders with a different pathophysiology than type 2 diabetes. In addition, we excluded those without documented follow-up (n=1511). This resulted in 27,086 individuals available for analyses (Figure 1). Of these, 22,282 had filled in all follow-up questionnaires and underwent a follow-up examination with detailed blood pressure measurement and laboratory examinations, while in 4,804 only interim follow-up questionnaire data were available. Median follow-up was 3.9 years, with a range of 0.5 to 9 years, for a total of 104,936 participant years.
Figure 1. Flow chart of the study population.
Lifelines participation and baseline measurement n=167729
Exclusion for baseline: Type 1 diabetes, n=72 Type 2 diabetes, n=1112 Gestational diabetes, n=42 Cardiovasc disease, n=672 Medication use, n=837 TSH<0.4 or > 10, n=386 Prospective follow-up data available: n=27086 By questionnaire only: n=4804 Questionnaire & measurements: n=22282
Validated baseline SAF and thyroid hormone measurements available
n=31760
Exclusion for follow-up: Type 1 diabetes, n=16 Gestational diabetes, n=26
Clinical examination
The specific methodology employed by Lifelines has been described previously (27-29). At both baseline and follow-up examination, participants completed a self-administered questionnaire on medical history, past and current diseases, and health behaviour. Medication use was verified -only at baseline- by a certified research assistant, and scored by the Anatomical Therapeutic Chemical (ATC) Classification System, a system developed for the classification of active ingredients of drugs. Information regarding smoking behaviour (never, former and current smoking), and quantity smoked, as well as coffee consumption (cups per day) was collected from the questionnaires as described previously (28). A standardised protocol was used to obtain blood pressure and anthropometric measurements: height, weight, and waist circumference. Weight was measured to the nearest 0.1 kg and height and waist circumference to the nearest 0.5 cm by the research assistant using calibrated measuring equipment, with participants wearing light clothing and no shoes. Waist circumference was measured with a tape around the body between the lower rib margin and the iliac crest. Body mass index (BMI) was calculated as weight divided by height squared (kg/m²). Systolic and diastolic blood pressure (BP) and heart rate were measured every minute during 10 minutes in the supine position using an automated Dinamap Monitor (GE Healthcare, Freiburg, Germany). The average of the last three readings was recorded for each blood pressure parameter and heart rate.
Skin autofluorescence (SAF)
SAF was measured non-invasively using the AGE Reader (Diagnoptics Technologies, Groningen, The Netherlands), as described previously (12, 30). The AGE Reader illuminates a skin surface of approximately 4 cm², guarded against surrounding light, with an excitation light source with wavelengths between 300 and 420nm (peak intensity at ~ 370nm). Emission light and reflected excitation light from the skin are measured with an internal spectrometer in the range of 300 to 600nm. Measurements were performed three times on the volar side of the forearm, 10cm below the elbow. SAF was based on the ratio of the average emitted light intensity per nanometer in the range of 420-600 nm and the average reflected light intensity per nanometer in the range 300-420 nm, multiplied by 100, and is expressed in arbitrary units (AU), taking skin color into account (31). Previous studies have shown an error percentage of 5% when repeated SAF measurements were taken over a single day in control subjects and individuals with diabetes (12). More details about the number of machines in the Lifelines study and validation of measurements have been described previously (van Waateringe et
al, Diabetologia). Age-adjusted SAF levels (Z-scores) were calculated based on the total population, separately for men and women.
Biochemical measurements
Blood samples were taken in the fasting state between 8:00 and 10:00 a.m. and transported to the laboratory facility at room temperature or at 4 °C, depending on the sample requirements. On the same day, glycated haemoglobin (HbA1c, EDTA-anticoagulated) was analysed using a NGSP-certified turbidimetric inhibition immunoassay on a Cobas Integra 800 CTS analyser (Roche Diagnostics Nederland BV, Almere, The Netherlands). Serum creatinine was measured on a Roche Modular P chemistry analyser (Roche, Basel, Switzerland), and renal function was calculated as estimated glomerular filtration rate (eGFR) with the formula developed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) (32). Total and high density lipoprotein (HDL) cholesterol were measured using an enzymatic colorimetric method, triglycerides using a colorimetric UV method, and low density lipoprotein (LDL) cholesterol using an enzymatic method, on a Roche Modular P chemistry analyser (Roche, Basel, Switzerland). Fasting blood glucose was measured using a hexokinase method. Levels of thyroid stimulating hormone (TSH), as well as free thyroxine (FT4) and free triiodothyronine (FT3) were assayed by electrochemiluminescent immunoassay (Roche Modular E170, Roche, Switzerland). Normal values according to the package insert for TSH are 0.4-4.5 mU/l, for FT4 11-22 pmol/l, and for FT3 4.4-6.7 pmol/l. The general Dutch population is iodine sufficient. Anti-thyroid peroxidase antibody levels were not available.
Calculations, definitions and statistical analyses
Diagnosis of metabolic syndrome was established if a subject at baseline satisfied at least three out of five criteria according to the modified guidelines of the National Cholesterol Education Programs Adults Treatment Panel III (NCEP-ATPIII criteria): (1) systolic BP ≥ 130 mmHg and/or diastolic BP ≥ 85 mmHg and/or use of antihypertensive medication; (2) HDL-cholesterol levels < 1.03 mmol/l in men and < 1.30 mmol/l in women and/or use of lipid-lowering medication influencing HDL-cholesterol levels; (3) triglyceride levels ≥ 1.70 mmol/l and/or use of triglyceride-lowering medication; (4) waist circumference ≥ 102 cm in men and ≥ 88 cm in women; (5) fasting glucose level ≥ 5.6 mmol/l and/or use of blood glucose-lowering medication and/or diagnosis of type 2 diabetes (33). Incident type 2 diabetes was based either on self-report or on the finding of a fasting blood glucose ≥ 7.0 mmol/l, and/or HbA1c ≥ 6.5% (48 mmol/mol) at follow-up. Incident cardiovascular disease was defined when subjects reported either one of the following diseases: myocardial
infarction, percutaneous transluminal coronary angioplasty (PTCA), stent positioning, coronary artery bypass grafting (CABG), transient ischaemic attack (TIA), cerebrovascular accident (CVA), intermittent claudication and peripheral artery vascular surgery. Vital status was ascertained with the Municipal Personal Records Database (GBA). Data on causes of death were not available. The incidences of type 2 diabetes, cardiovascular disease and mortality were considered as a composite outcome as well as separate outcomes.
All analyses were conducted using SPSS Statistics (Version 22, IBM, Armonk, NY, USA). Data are presented as mean ± SD, or median and interquartile range when not normally distributed. Means were compared between groups with analysis of variance. When variables were not normally distributed, medians were compared with the nonparametric Mann-Whitney U test. Chi-square test was used to analyse categorical variables. Uni- and multivariate logistic regression analyses were performed to examine the association between SAF and thyroid hormone levels and the composite outcome of incident type 2 diabetes, cardiovascular disease and mortality, as well as these outcomes separately, while adjusting for relevant clinical, biochemical and lifestyle risk factors. We compared different models in which we cumulatively adjusted for the most important determinants of SAF and thyroid hormone levels, i.e. age (model 1, used as a basic model), SAF (model 2), SAF and age (model 3), and additionally adjusted for presence of metabolic syndrome (model 4), confounding non-biochemical factors like gender and BMI (model 5), and all relevant factors including classic cardiovascular risk factors as blood pressure, cholesterol, fasting blood glucose, smoking and eGFR (model 6). To allow for comparison with data from the Rotterdam study (19), we also performed multivariable logistic regression with thyroid hormone levels and all relevant risk factors including the use of cholesterol- and blood pressure lowering medication. Finally, we performed Cox proportional hazard analysis to assess the effects of thyroid hormone levels on mortality. P-values < 0.05 were considered statistically significant.
Results
Association between thyroid hormone levels and skin autofluorescence
As thyroid hormone levels change significantly during ageing, we assessed their influence on SAF with adjustment for age. There was a significant association between FT3 and SAF (Beta = 0.032, p = 1.1 x 10-13) and between FT4 and SAF (Beta = 0.028, p = 1.8 x
10-8, Table 1). However, after adjustment for all known factors influencing SAF (age, sex,
waist circumference, BMI, glycaemic parameters, coffee consumption, smoking behaviour and cigarette pack-years, both FT4 and FT3 were no longer significantly associated with SAF (Table 1). In supplemental Figure1, we depict the relationship between age, thyroid hormone levels FT4 and FT3, and skin autofluorescence.
Table 1. Linear regression analysis for the predictive value of thyroid hormone levels on SAF
adjusted for important confounders, in participants with normal and slightly elevated TSH (range 0.4 - 10 mU/l)
Multivariate model of SAF adjusted for age
Beta SE P value TSH (mU/l) -0.015 0.002 0.004 FT4 (mU/l) 0.028 0.001 1.8 x 10-8 FT3 (mU/l) 0.032 0.003 1.1 x 10-13 FT4 (pmol/l) 0.001 0.001 0.775 Age (years) 0.485 0.001 <1.0 x 10-200
Male gender (y/n) -0.011 0.005 0.059 Body mass index (kg/m2) -0.020 0.001 0.038
Waist circumference (cm) 0.046 0.001 1.0 x 10-5
Glucose (mmol/l) 0.003 0.005 0.643 HbA1c (mmol/mol) 0.008 0.001 0.145 Coffee (cups per day) 0.141 0.001 4.2 x 10-146
Current smoking (y/n) 0.078 0.006 2.3 x 10-42
Pack years (years) 0.112 0.001 2.6 x 10-79
eGFR (ml/min) -0.031 0.001 6.6 x 10-7
FT3 (pmol/l) -0.002 0.003 0.702 Age (years) 0.485 0.001 <1.0 x 10-200
Male gender (y/n) 0.012 0.005 0.057 Body mass index (kg/m2) -0.020 0.001 0.038
Table 1. Continued
Beta SE P value Glucose (mmol/l) 0.003 0.005 0.637 HbA1c (mmol/mol) 0.008 0.001 0.148 Coffee (cups per day) 0.141 0.001 4.5 x 10-146
Current smoking (y/n) 0.078 0.006 2.9 x 10-42
Pack years (years) 0.112 0.001 2.6 x 10-79
eGFR (ml/min) -0.031 0.001 8.0 x 10-7
Incidence of type 2 diabetes, cardiovascular disease and mortality
Table 2 depicts the incidence of all outcomes. An individual may have had more than one outcome. After a median follow-up of 47 months, 381 individuals had developed type 2 diabetes (1.4%), and 496 individuals (1.8%) had developed cardiovascular disease. Death was reported in 372 (1.4%) individuals. Neither event was observed in 25,861 participants. As depicted in Table 3, participants with an event were more frequently males and current and former smokers, and had higher BMI, blood pressure, glucose and HbA1c, dyslipidaemia, SAF, and lower eGFR. Also, they more frequently used cholesterol- and blood-pressure-lowering drugs. Overall, there were no significant differences in thyroid hormone levels when compared between individuals who developed an event or who remained free of events. However, when all three outcomes were assessed separately, subjects with incident type 2 diabetes were significantly older (54 vs 44 yrs), had lower FT4 compared to those without diabetes, cardiovascular disease or mortality (FT4 15.2 (IQR 13.8-16.6) vs 15.6 (IQR 14.3-17.0) pmol/l, p = 9.3 x 10-7), but similar FT3 and
TSH. Those with new-onset CVD were older (57 vs 44 yrs) and had borderline significantly higher FT4 levels (15.8 (IQR 14.4-17.3) vs. 15.6 (14.3-17.0) pmol/l, p = 0.06, Table 4). Thyroid hormone levels did not differ between participants who died, and those who remained free of any event.
Table 2. Clinical endpoints as defined in the study (n=27086).
No type 2 diabetes, CVD or death n = 25861 Either type 2 diabetes, CVD or death n = 1225 Incident type 2 diabetes only n = 358 Incident CVD only n = 468
Death * n = 364
Incident type 2 diabetes and CVD n=15 Type 2 diabetes and death n=7
CVD and death n=12
Type 2 diabetes and CVD and death n=1 * without / prior to ascertainment of diabetes or CVD status
Table 3. Clinical characteristics of the study population at baseline in relation to outcome status No events Composite outcome P value
N (males/females) 10480 / 15381 623 / 602 % males 40.5 50.9 <0.0001 Age (yrs) 43.6 ± 12.0 54.4 ± 12.1 <0.0001 BMI (kg/m2) 25.8 ± 4.1 27.8 ± 4.5 <0.0001 Waist (cm) 90 ± 12 97 ± 12 <0.0001 Systolic BP (mmHg) 125 ± 15 134 ± 17 <0.0001 Diastolic BP (mmHg) 74 ± 9 78 ± 10 <0.001
Heart rate (/min) 71 ± 11 72 ± 11 0.110 Creatinine (µmol/l) 74 ± 12 76 ± 19 <0.0001
eGFR ( ml/min) 96 ± 14 89 ± 15 <0.0001
Total cholesterol (mmol/l) 5.1 ± 1.0 5.4 ± 1.1 <0.0001
HDL-cholesterol (mmol/l) 1.49 ± 0.39 1.40 ± 0.40 <0.0001 LDL-cholesterol (mmol/l) 3.2 ± 0.9 3.5 ± 0.9 <0.0001 Triglycerides (mmol/l) 0.95 (0.70-1.35) 1.21 (0.87-1.74) <0.0001 Glucose (mmol/l) 4.9 ± 0.5 5.3 ± 0.6 <0.0001 HbA1c (%) 5.6 ± 0.3 5.9 ± 0.4 <0.0001 HbA1c (mmol/mol) 37 ± 3 39 ± 4 <0.0001 TSH (mU/l) 2.13 (1.54-2.94) 2.13 (1.48-2.98) 0.709 FT4 (pmol/l) 15.7 ± 2.0 15.6 ± 2.2 0.076 FT3 (pmol/l) 5.25 ± 0.70 5.24 ± 0.65 0.670 Current smokers (%) 21.21 25.5 <0.0001 Former smokers (%) 30.5 40.9 <0.0001
Table 3. Continued
No events Composite outcome P value
Metabolic syndrome (%) 13.2 34.9 <0.0001 BP-lowering therapy (%) 8.6 26.4 <0.0001 Cholesterol-lowering
therapy (%) 3.0 11.3 <0.0001 Skin autofluorescence (AU) 1.93 ± 0.42 2.24 ± 0.49 < 0.0001 SAF Z-score 0.06 ± 0.82 0.28 ± 1.03 < 0.0001 Data are presented as mean ± standard deviation, median (IQR), or number (%). Composite outcome refers to the combination of incident type 2 diabetes, cardiovascular disease and mortality.
Abbreviations: AU, arbitrary units; BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate; FT3, free tri-iodothyronine; FT4, free thyroxine; HbA1c, glycated haemoglobin; HDL, high density lipoprotein; LDL, low density lipoprotein; SAF, skin autofluorescence; TSH, thyroid-stimulating hormone.
Table 4. Baseline thyroid hormone levels according to outcome
No event New T2D New CVD Mortality
TSH (mU/l) 2.13 (1.54-2.94) 2.12 (1.52-2.92) 2.19 (1.52-3.13) 2.07 (1.40-2.97) FT4 (pmol/l) 15.6 (14.3-17.0) 15.2 (13.8-16.6) $ 15.8 (14.4-17.3) 15.5 (14.3-17.1) FT3 (pmol/l) 5.2 (4.8-5.6) 5.3 (4.8-5.7) 5.2 (4.9-5.6) 5.2 (4.8-5.5) Values are median (IQR)
T2D, type 2 diabetes; CVD, cardiovascular disease. $ p = 9.3 x 10-7 vs. no event, all other comparisons p>0.05
Association with separate outcomes
Next, we analysed whether thyroid hormone levels were independently associated with type 2 diabetes, cardiovascular disease and mortality. Table 5 first shows the univariate association between all relevant parameters, with type 2 diabetes and cardiovascular disease separately. Lower FT4 was significantly associated with a higher risk of incident 2 diabetes (OR 0.88, p=8.0 x 10-7), while higher FT4 levels were associated with an increased risk of cardiovascular disease (OR 1.04, p = 0.056). Low levels of FT4 remained significantly associated with increased risk type 2 diabetes, when adjusted for age (model 1), SAF (model 2), SAF and age (model 3), SAF, age and presence of metabolic syndrome (model 4), SAF, age, male gender and BMI (model 5), and all relevant parameters including blood pressure, cholesterol, smoking and glucose (model 6).
Higher FT4 levels were significantly associated with a higher risk of incident cardiovascular disease when adjusted for SAF and age, metabolic syndrome, gender and BMI, but became non-significant (p = 0.069) when additional adjustment for all relevant parameters was performed (model 6).
Table 5. Univariate and multivariate logistic regression analyses for incident type 2 diabetes and
cardiovascular disease, at a median of 4-years follow-up
New type 2 diabetes New cardiovascular disease Univariate OR 95% CI P value OR 95% CI P value
SAF (AU) 2.64 2.20-3.18 2.1 x 10-25 3.01 2.56-3.52 6.1 x 10-42 Age (years) 1.05 0.91-1.07 2.0 x 10-36 1.07 1.06-1.07 4.9 x 10-65 TSH (mU/1) 0.99 0.91-1.07 0.761 1.05 0.98-1.12 0.190 FT4 (pmol/l) 0.88 0.83-0.92 8.0 x 10-7 1.04 1.00-1.09 0.056 FT3 (pmol/l) 1.05 0.94-1.18 0.400 1.03 0.92-1.16 0.595 Male gender 1.47 1.20-1.80 1.9 x 10-5 1.52 1.28-1.82 4.0 x 10-6 BMI (kg/m2) 1.15 1.13-1.17 1.0 x 10-58 1.07 1.05-1.09 6.9 x 10-13 Waist circumference (cm) 1.07 1.06-1.07 1.6 x 10-67 1.03 1.03-1.04 6.9 x 10-22 Syst. BP (mmHg) 1.03 1.03-1.04 5.4 x 10-33 1.03 1.02-1.03 8.5 x 10-27 Diast. BP (mmHg) 1.04 1.03-1.05 2.0 x 10-15 1.04 1.03-1.05 7.0 x 10-19 Cholesterol (mmol/1 1.24 1.13-1.37 1.4 x 10-5 1.36 1.25-1.48 1.3 x 10-12 Triacylglycerol (mmol/l) 1.46 1.37-1.56 2.0 x 10-29 1.22 1.13-1.31 3.7 x 10-7
Blood glucose (mmol/1) 12.1 10.2-14.4 1.0 x 10-250 1.85 1.57-2.18 3.2 x 10-13
HbA1c 1.41 1.36-1.45 8.6 x 10-103 1.14 1.11-1.17 1.11-1.17
eGFR (ml/min) 0.98 0.97-0.99 1.1 x 10-8 0.97 0.97-0.98 7.5 x 10-22
Former smoking (y/n) 1.59 1.30-1.96 9.0 x 10-6 1.45 1.21-1.74 6.8 x 10-6
Current smoking (y/n) 1.05 0.82-1.34 0.710 1.45 1.19-1.76 2.6 x 10-4
MetS 7.8 6.3-9.5 9.5 x 10-86 2.17 1.77-2.66 1.0 x 10-13
Multivariate Model 1 OR 95% CI P value OR 95% CI P value
Age (years) 1.05 1.05-1.06 1.4 x 10-35 1.06 1.06-1.07 1.4 x 10-65 FT4 (pmol/l) 0.89 0.84-0.93 7.0 x 10-6 1.06 1.01-1.10 0.013 Multivariate Model 2 SAF (AU) 2.64 2.20-3.17 3.0 x 10-25 3.01 2.56-3.52 6.0 x 10-42 FT4 (pmol/l) 0.88 0.84-0.93 1.0 x 10- 1.04 1.00-1.09 0.051 Multivariate Model 3 SAF (AU) 1.59 1.25-2.02 1.3 x 10-5 1.61 1.31-1.98 8.0 x 10-6 Age (years) 1.04 1.03-1.051 1.8 x 10-18 1.06 1.05-1.06 5.9 x 10-36
Table 5. Continued
New type 2 diabetes New cardiovascular disease Univariate OR 95% CI P value OR 95% CI P value FT4 (pmol/l) 0.89 0.84-0.93 4.0 x 10-6 1.05 1.01-1.10 0.019 Multivariate Model 4 SAF (AU) 1.40 1.09-1.78 0.008 1.55 1.26-1.91 4.1 x 10-5 Age (years) 1.04 1.02-1.04 5.6 x 10-12 1.05 1.04-1.06 2.5 x 10-33 FT4 (pmol/l) 0.91 0.87-0.96 0.001 1.06 1.01-1.10 0.010 MetS (y/n) 6.10 4.90-7.51 8.2 x 10-64 1.60 1.30-1.97 1.1 x 10-5 Multivariate Model 5 SAF (AU) 1.41 1.10-1.80 0.007 1.53 1.24-1.89 8.2 x 10-5 Age (years) 1.04 1.03-1.05 1.4 x 10-17 1.05 1.04-1.06 2.3 x 10-34 FT4 (pmol/l) 0.91 0.86-0.95 2.0 x 10-4 1.05 1.01-1.10 0.019 Male gender 1.53 1.24-1.88 6.8 x 10-5 1.38 1.15-1.66 4.6 x 10-4 BMI (kg/m2) 1.15 1.13-1.17 1.7 x 10-47 1.06 1.03-1.08 4.3 x 10-7 Multivariate Model 6 SAF (AU) 1.26 0.96-1.64 0.093 1.33 1.07-1.66 0.011 Age (years) 1.03 1.01-1.04 8.8 x 10-5 1.06 1.05-1.071 4.5 x 10-24 FT4 (pmol/l) 0.90 0.86-0.95 2.0 x 10-4 1.04 1.00-1.09 0.069 Male gender 0.93 0.74-1.16 0.520 1.30 1.07-1.56 0.007 BMI (kg/m2) 1.07 1.05-1.10 6.8 x 10-10 1.05 1.02-1.07 3.0 x 10-5 Syst. BP (mmHg) 1.01 1.0-1.01 0.56 1.01 1.00-1.02 0.004 Cholesterol (mmol/l) 0.93 0.83-1.04 0.201 1.03 0.94-1.14 0.487 Glucose (mmol/l) 9.01 7.41-10.83 9.2 x 10-111 1.01 0.83-1.22 0.932 eGFR (ml/min) 1.01 1.00-1.02 0.143 1.00 0.99-1.01 0.599 Current smoking (y/n 1.28 0.98-1.67 0.076 1.76 1.42-2.17 2.0 x 10-7
Data are presented as OR, odds ratios (95% confidence interval). Baseline risk factors were used to predict the median 4 years risk of the outcomes of type 2 diabetes and cardiovascular disease. FT4, FT3, SAF, age, BMI, blood pressure, cholesterol, glucose were defined as
continuous variables. Male gender, current smoking, and metabolic syndrome were defined as categorical variables. Abbreviations: see Table 2.
Figure 2. Crude incidence of type 2 diabetes and cardiovascular disease according to FT4 levels.
N of participants: FT4 10-12: n=1997; FT4 13-14: n=7839; FT4 15-16: n=10301; FT4 17-18: n=5293; FT4 19-21: n=1656.
Figure 2 shows the incidence of both type 2 diabetes and cardiovascular disease according to FT4 levels. There is a significant increase of type 2 diabetes incidence with lower levels of FT4, and similarly an increase of incident cardiovascular disease with higher FT4, especially with FT4 levels of 17 pmol/l and higher.
Thus, each 1.0 pmol/l lower level of FT4 was associated with a 9-11% higher type 2 diabetes risk, adjusted for all relevant parameters, SAF, age, gender and BMI. In contrast, each 1.0 pmol/l higher FT4 was associated with a 5% increased cardiovascular disease risk, when adjusted for SAF, age, gender and BMI, but this association lost its significance when adjusting further for blood pressure, cholesterol and smoking.
Similarly, we analysed the association of FT3 with type 2 diabetes and cardiovascular disease. Higher FT3 levels were associated with an increased risk of type 2 diabetes (Table 6), when adjusted for age, age and SAF, and metabolic syndrome, but lost its significance when further adjustment was performed for classic risk factors like blood pressure, cholesterol, glucose and smoking (model 6). Higher FT3 was associated with increased risk of CVD, even when corrected for all classic risk factors in model 6.
Thyroid hormone levels were not associated with mortality, neither in a univariate model, nor in any of the multivariate models (data not shown). Figure 3 shows the survival analysis according to tertiles of FT4 and FT3, and according to TSH group. No significant effect of FT4- and FT3-tertiles on mortality were observed. In univariate analysis, higher SAF, age, BMI, waist circumference, cholesterol, and blood glucose and HbA1c, as well as male gender, were associated with higher mortality. In the final model, only SAF, age, male gender, BMI and current smoking were independently associated with mortality (data not shown).
Table 6. Univariate and multivariate logistic regression analyses for incident type 2 diabetes and
cardiovascular disease with FT3 levels at a median of 4-years follow-up
New type 2 diabetes New cardiovascular disease Multivariate Model 1 OR 95% CI P value OR 95% CI P value
Age (years) 1.06 1.05-1.06 3.6 x 10-38 1.07 1.06-1.07 1.5 x 10-67 FT3 (pmol/l) 1.16 1.07-1.26 0.001 1.08 1.17-1.28 1.3 x 10-4 Multivariate Model 2 SAF (AU) 2.67 2.23-3.21 7.2 x 10-26 3.04 2.59-3.56 1.7 x 10-42 FT3 (pmol/l) 1.10 0.99-1.21 0.068 1.09 0.99-1.19 0.76 Multivariate Model 3 SAF (AU) 1.56 1.23-1.99 2.6 x 10-4 1.61 1.31-1.98 7.0 x 10-6 Age (years) 1.05 1.04-1.06 3.0 x 10-20 1.06 1.05-1.07 2.0 x 10-37 FT3 (pmol/l) 1.16 1.06-1.26 0.001 1.08 1.07-1.28 1.4 x 10-4
6
Table 6. Continued
New type 2 diabetes New cardiovascular disease Multivariate Model 4 SAF (AU) 1.37 1.08-1.76 0.011 1.56 1.26-1.92 3.2 x 10-5 Age (years) 1.04 1.03-1.05 7.8 x 10-13 1.06 1.05-1.07 1.0 x 10-34 FT3 (pmol/l) 1.11 0.99-1.23 0.065 1.16 1.07-1.26 2.5 x 10-4 MetS (y/n) 6.20 5.03-7.02 1.1 x 10-64 1.54 1.25-1.90 5.2 x 10-5 Multivariate Model 5 SAF (AU) 1.40 1.09-1.79 0.009 1.54 1.24-1.90 6.7 x 10-5 Age (years) 1.05 1.04-1.06 6.7 x 10-5 1.06 1.05-1.06 1.8 x 10-35 FT3 (pmol/l) 1.14 1.03-1.26 0.012 1.15 1.05-1.25 0.002 Male gender 1.39 1.12-1.71 0.002 1.35 1.13-1.62 0.001 BMI (kg/m2) 1.15 1.13-1.18 1.3 x 10-50 1.05 1.03-1.07 2.0 x 10-6 Multivariate Model 6 SAF (AU) 1.24 0.95-1.62 0.109 1.34 1.07-1.66 0.009 Age (years) 1.03 1.01-1.04 6.8 x 10-5 1.06 1.05-1.07 1.4 x 10-24 FT3 (pmol/l) 1.00 0.84-1.19 0.974 1.12 1.01-1.25 0.026 Male gender 0.90 0.71-1.12 0.342 1.27 1.06-1.54 0.012 BMI (kg/m2) 1.08 1.06-1.10 4.2 x 10-11 1.05 1.02-1.07 6.1 x 10-5 Syst. BP (mmHg) 1.01 1.00-1.01 0.085 1.01 1.00-1.02 0.004 Cholesterol (mmol/l) 0.94 0.84-1.05 0.273 1.04 0.94-1.14 0.470 Glucose (mmol/l) 9.02 7.41-10.33 3.0 x 10-111 0.99 0.82-1.20 0.950 eGFR (ml/min) 1.01 1.00-1.02 0.108 1.00 0.99-1.01 0.688 Current smoking (y/n) 1.22 0.93-1.60 0.149 1.75 1.42-2.17 2.3 x 10-7
Data are presented as OR, odds ratios (95% confidence interval). Baseline risk factors were used to predict the median 4 years risk of the outcomes of type 2 diabetes and cardiovascular disease. FT3, SAF, age, BMI, blood pressure, cholesterol, glucose were defined as continuous variables. Male gender, current smoking, and metabolic syndrome were defined as categorical variables.
Figure 3. Association between thyroid hormone levels and mortality.
A: FT4 tertiles; B: FT3 tertiles; C: TSH groups
A.
B.
C.
In order to compare our data with results by others, we also performed the multivariate logistic regression analyses without the inclusion of SAF (Supplemental Tables 2 and 3). From these results, similar associations emerge: for the development of type 2 diabetes, fasting blood glucose, BMI and age were the most significant predictors, and again FT4 was an independent additional predictor (OR 0.90, p = 2.0 x 10-4). For the development of cardiovascular disease, age, male gender, BMI, current smoking, and use of statins and blood-pressure-lowering agents were the strongest predictors, while FT4 was not significant when adjusted for by these factors (OR 1.04, p = 0.104). In a similar multivariate model, FT3 was not independently associated with incident type 2 diabetes or mortality, and only marginally associated with incident CVD.
Discussion
When adjusted for age, there was a significant and positive association between both FT4 and FT3 levels and skin autofluorescence. However, this association was no longer significant when adjustments were made for other variables known to be associated with SAF. Both SAF and FT4/FT3 were significantly associated with incident type 2 diabetes and CVD, but FT4/FT3 was not associated with mortality. Serum FT4 levels have opposite effects on risk of development of type 2 diabetes and cardiovascular disease in the general population, independent of SAF, and presence of metabolic syndrome. Similarly, higher FT3 was associated with increased risk of type 2 diabetes and cardiovascular disease, when adjusted for SAF, age, gender and BMI, but not after additional adjustment for blood pressure, cholesterol and smoking.
Only a limited number of earlier studies have assessed the effects of thyroid hormone levels on future disease events. In the Rotterdam study, it was shown that low and low-normal thyroid function are risk factors for the development of type 2 diabetes, especially in people who already have slightly increased blood glucose levels (19). A decrease of FT4 levels from 22 to 12 pmol/l was associated with an increase of absolute diabetes risk, from approximately 17 to 33%. In the Lifelines population, we calculated a 9-11% increase of type 2 diabetes risk for every 1.0 pmol/l lower level of FT4. The average age of the participants in the Rotterdam study, however, was significantly higher than in Lifelines (respectively 64 vs 44 years), resulting in a considerable difference in absolute incidence of type 2 diabetes (9.5% after 7.9 years follow-up vs. 1.4% in 4 years follow-up, respectively). Nevertheless, these data are in agreement with our earlier finding that low FT4 levels are strongly associated with the presence of metabolic syndrome, a cluster of risk factors, which is a recognized predictor for the development of type 2 diabetes (20).
Also in the Rotterdam study, higher FT4 levels were associated with incident atherosclerotic cardiovascular events and with cardiovascular mortality (21). The data of our study confirm these findings, be it in a younger population. However, we were not able to assess the effects of FT4 levels on cardiovascular mortality, because Lifelines has at the time of writing not yet cause-specific death in its participants. Again, because of the age difference, the absolute incidence of cardiovascular events in the Rotterdam study was significantly higher: incidence of atherosclerotic cardiovascular events was 10%, and cardiovascular mortality was 6.5%, over a follow-up of 8.8 years. Incident (atherosclerotic) cardiovascular events in Lifelines was 1.8% over 4 years. Importantly, in addition to the difference in age, our definition of cardiovascular disease was broader, and also included cardiovascular procedures like percutaneous transluminal coronary
angioplasty (PTCA), stent positioning, or coronary artery bypass grafting (CABG), as well as intermittent claudication and peripheral artery vascular surgery. Also, 9.8% of the participants in the Rotterdam study had a history of atherosclerotic cardiovascular disease, and 11.4% had diabetes mellitus at baseline (21), while we included only participants without baseline cardiovascular disease or diabetes.
To our knowledge, this is the first study to assess FT3 levels in relation to incidence of type 2 diabetes, cardiovascular disease and mortality within the same population. This is due to the fact that only few epidemiologic studies have measured FT3 levels. T3 in serum is in part directly secreted by the thyroid, and in part the result of conversion from T4 by deiodinase enzymes (34). Conversion of T4 to T3 is reduced in inflammation, caloric restriction, infection, and possibly in low-grade inflammation associated with vascular disease (35, 36). The latter makes a clear a priori hypothesis regarding the association between FT3 and CVD difficult, as FT3 may lose its association when CVD phenotype becomes more severe. We showed that higher FT3 levels were associated with an increased risk of type 2 diabetes as well as cardiovascular disease, at least in the initial models which we explored. This may in part be an independent effect, as FT3 levels remained significantly associated with cardiovascular disease when adjustment for the presence of metabolic syndrome and for gender and BMI was performed. In addition, FT3 was also significant when further adjusted for cardiovascular risk factor like blood pressure, cholesterol and smoking. There are only a few other studies which have evaluated the role of T3 on mortality and cardiovascular disease. Cappola et al did not find an association between total T3 and incidence of atrial fibrillation, coronary heart disease, heart failure and all-cause mortality in adults older than 65 years (mean age 74.5 years) (37). A study in Korea reported that FT3 levels within the normal range were inversely associated with all-cause mortality, but not with cardiovascular mortality in middle-aged people (mean age 40 years) (38). In this study, FT3 above the median was associated with a 30% increase of cardiovascular mortality, but with a large confidence interval due to a limited number of events (n=62 in over 400,000 person-years). In the Atherosclerosis Risk in Communities (ARIC) study, which comprised over 11,000 participants with a mean age of 57±6 years, T3 levels -adjusted for relevant factors like age, BMI, blood pressure, lipids, and lipid-lowering and blood-pressure-lowering medication- were not associated with incident cardiovascular events (39). Our group has reported in an earlier paper that presence of metabolic syndrome was associated with higher levels of FT3. We suggested that, on one hand, this may be due to the consequence of an excess of nutrition, especially when consuming a diet high in fat, in obese individuals, who more frequently are insulin-resistant (40, 41). On
the other hand, metabolic syndrome or insulin resistance itself may lead to a relatively higher intracellular production of T3 as a way to compensate -on a local tissue level- for the disturbances in the metabolic state, intended to limit nutrient overload by inducing tissue thermogenesis and simulating metabolic activity (20, 42, 43). Thus, from our study higher FT3 levels appear to be independently associated with a higher CVD risk.
In an earlier publication, we have shown that SAF was an independent and strong predictor of mortality. From the current study, we can conclude that thyroid hormone levels were not associated with increased mortality risk.
Strengths and limitations
Our study has several strengths but also some limitations. We have data on a prospective population-based study, which included over 27,000 participants within a broad range of age and cardiovascular risk. This is the first prospective study that examined SAF and thyroid hormone levels together as predictors for type 2 diabetes, cardiovascular disease and mortality in the general population. We included participants with normal or slightly elevated TSH levels, but still normal FT4 levels. Although extensive biomedical information was collected at Lifelines baseline evaluation, unfortunately the number of participants in whom thyroid hormone levels were measured was lower compared to those in whom SAF was measured (Waateringe et al, Diabetologia). Also no data were available on the use of new medications, changes in medications or development of new thyroid disorders, as this was not included in the follow-up questionnaires. Medication prescriptions, in particular oral blood-glucose-lowering agents and/or insulin, can be used to validate self-reported diagnosis of type 2 diabetes, or even ascertain presence of diabetes when a participant did not report presence of diabetes correctly in the questionnaire. Also, it has to be noted that some aspects of cardiovascular disease, like PTCA and CABG procedures, as well as carotid and peripheral artery surgery or interventions, were not systemically asked for in all questionnaires, which may influence allocation to the healthy or cardiovascular disease group. Also, data regarding exact time of diabetes diagnosis and cardiovascular events, as well as cause of death were not collected. As a consequence, we were not able to perform survival analyses for these outcomes.
Conclusions
Serum FT4 levels have opposite effects on the risk of development of type 2 diabetes and cardiovascular disease in the general population, and the association with type 2 diabetes is independent of SAF, glycaemic variables, and presence of metabolic syndrome, and other risk factors. Higher FT4 was no longer associated with cardiovascular disease when additional adjustment for all relevant cardiovascular risk parameters was performed. Higher FT3 levels were associated with risk of CVD, but not type 2 diabetes or death. Next steps will be to elucidate underlying mechanism(s), and possibly to set up an intervention study to assess the effects of thyroid hormone supplementation in people with low normal thyroid function on disease outcomes.
Acknowledgements
The authors wish to acknowledge the services of the Lifelines Cohort Study, the contributing research centres delivering data to Lifelines, and all the study participants.
Availability of data and material
The manuscript is based on data from the Lifelines Cohort Study. Lifelines adheres to standards for data availability. The data catalogue of Lifelines is publicly accessible on http://www.lifelines.net. All international researchers can obtain data at the Lifelines research office (research@lifelines.nl), for which a fee is required. The Lifelines system allows access for reproducibility of the study results.
Funding
Lifelines has been funded by a number of public sources, notably the Dutch Government, The Netherlands Organization of Scientific Research NWO [grant 175.010.2007.006], funds from FES (Fonds Economische Structuurversterking), SNN (Samenwerkingsverband Noord Nederland) and REP (Ruimtelijk Economisch Programma), Dutch Ministry of Economie Affairs, Pieken in de Delta, Provinces of Groningen and Drenthe, the Target project, BBMRI-NL, the University of Groningen, and the University Medical Center Groningen, The Netherlands. This work was supported by the National Consortium for Healthy Ageing, and funds from the European Union’s Seventh Framework program (FP7/2007-2013) through the BioSHaRE-EU (Biobank Standardisation and Harmonisation for Research Excellence in the European Union) project, grant agreement 261433. Lifelines (BRIF4568) is engaged in a Bioresource research impact factor (BRIF) policy pilot study, details of which can be found at: https://www.bioshare.eu/content/bioresource-impact-factor
Duality of interest
RG and AJS are founders and shareholders of DiagnOptics BV, Groningen, the Netherlands, manufacturer of the AGE Reader (http://www.diagnoptics.com) which has been used in the present study. All other authors have declared no conflict of interest.
Contribution statement
RPvW, SNS, MMvdK and BHRW contributed to the study design. RPvW, SNS and BHRW performed the statistical analyses. All authors contributed to analyses and interpretation of the data. RPvW and BHRW drafted the initial version of the manuscript. All authors read and approved the final manuscript.
References
1. Ogurtsova K, da Rocha Fernandes JD, Huang Y, Linnenkamp U, Guariguata L, Cho NH, et al. IDF Diabetes Atlas: Global estimates for the prevalence of diabetes for 2015 and 2040. Diabetes Res Clin Pract. 2017;128:40-50.
2. Wolffenbuttel BHR, Van Haeften TW. Prevention of complications in non-insulin-dependent diabetes mellitus (NIDDM). Drugs. 1995;50(2):263-88.
3. Brownlee M. The pathobiology of diabetic complications: a unifying mechanism. Diabetes. 2005;54(6):1615-25.
4. Ford ES. Trends in the risk for coronary heart disease among adults with diagnosed diabetes in the U.S.: findings from the National Health and Nutrition Examination Survey, 1999-2008. Diabetes Care. 2011;34(6):1337-43.
5. Monnier VM. Nonenzymatic glycosylation, the Maillard reaction and the aging process. J Gerontol. 1990;45(4):B105-11.
6. Brownlee M. Advanced protein glycosylation in diabetes and aging. AnnuRevMed. 1995;46:223-34.
7. Baynes JW, Thorpe SR. Glycoxidation and lipoxidation in atherogenesis. Free Radic Biol Med. 2000;28(12):1708-16.
8. Aronson D. Cross-linking of glycated collagen in the pathogenesis of arterial and myocardial stiffening of aging and diabetes. J Hypertens. 2003;21(1):3-12.
9. Liu CY, Huang QF, Cheng YB, Guo QH, Chen Q, Li Y, et al. A Comparative Study on Skin and Plasma Advanced Glycation End Products and Their Associations with Arterial Stiffness. Pulse (Basel). 2017;4(4):208-18.
10. Basta G, Schmidt AM, De Caterina R. Advanced glycation end products and vascular inflammation: implications for accelerated atherosclerosis in diabetes. Cardiovasc Res. 2004;63(4):582-92.
11. Isermann B, Bierhaus A, Humpert PM, Rudofsky G, Chavakis T, Ritzel R, et al. [AGE-RAGE: A Hypothesis or a Mechanism?]. Herz. 2004;29(5):504-9.
12. Meerwaldt R, Graaff R, Oomen PHN, Links TP, Jager JJ, Alderson NL, et al. Simple non-invasive assessment of advanced glycation endproduct accumulation. Diabetologia. 2004;47(7):1324-30.
13. van Waateringe RP, Slagter SN, van Beek AP, van der Klauw MM, van Vliet-Ostaptchouk JV, Graaff R, et al. Skin autofluorescence, a non-invasive biomarker for advanced glycation end products, is associated with the metabolic syndrome and its individual components. Diabetol Metab Syndr. 2017;9:42.
14. Lutgers HL, Graaff R, Links TP, Ubink-Veltmaat LJ, Bilo HJ, Gans RO, et al. Skin autofluorescence as a noninvasive marker of vascular damage in patients with type 2 diabetes. Diabetes Care. 2006;29(12):2654-9.
15. Gerrits EG, Lutgers HL, Kleefstra N, Graaff R, Groenier KH, Smit AJ, et al. Skin autofluorescence: a tool to identify type 2 diabetic patients at risk for developing microvascular complications. Diabetes Care. 2008;31(3):517-21.
16. Noordzij MJ, Mulder DJ, Oomen PH, Brouwer T, Jager J, Castro Cabezas M, et al. Skin autofluorescence and risk of micro- and macrovascular complications in patients with Type 2 diabetes mellitus-a multi-centre study. Diabet Med. 2012;29(12):1556-61.
17. Tanaka K, Tani Y, Asai J, Nemoto F, Kusano Y, Suzuki H, et al. Skin autofluorescence is associated with severity of vascular complications in Japanese patients with Type 2 diabetes. Diabet Med. 2012;29(4):492-500.
18. Fokkens BT, van Waateringe RP, Mulder DJ, Wolffenbuttel BHR, Smit AJ. Skin autofluorescence improves the Finnish Diabetes Risk Score in the detection of diabetes in a large population-based cohort: The LifeLines Cohort Study. Diabetes Metab. 2017.
19. Chaker L, Ligthart S, Korevaar TI, Hofman A, Franco OH, Peeters RP, et al. Thyroid function and risk of type 2 diabetes: a population-based prospective cohort study. BMC medicine. 2016;14(1):150.
20. Wolffenbuttel BHR, Wouters HJCM, Slagter SN, van Waateringe RP, Muller Kobold AC, van Vliet-Ostaptchouk JV, et al. Thyroid function and metabolic syndrome in the population-based LifeLines cohort study. BMC Endocr Disord. 2017;17(1):65.
21. Bano A, Chaker L, Mattace-Raso FUS, van der Lugt A, Ikram MA, Franco OH, et al. Thyroid Function and the Risk of Atherosclerotic Cardiovascular Morbidity and Mortality: The Rotterdam Study. Circ Res. 2017;121(12):1392-400.
22. Zaki SM, Youssef MF. Thyroid hormone dysfunctions affect the structure of rat thoracic aorta: a histological and morphometric study. Folia Morphol (Warsz). 2013;72(4):333-9.
23. Wang X, Sun Z. Thyroid hormone induces artery smooth muscle cell proliferation: discovery of a new TRalpha1-Nox1 pathway. J Cell Mol Med. 2010;14(1-2):368-80.
24. Varga F, Rumpler M, Zoehrer R, Turecek C, Spitzer S, Thaler R, et al. T3 affects expression of collagen I and collagen cross-linking in bone cell cultures. Biochem Biophys Res Commun. 2010;402(2):180-5.
25. Udupa SV, Manjrekar PA, Udupa VA, Vivian D. Altered fructosamine and lipid fractions in subclinical hypothyroidism. J Clin Diagn Res. 2013;7(1):18-22.
26. Stolk RP, Rosmalen JG, Postma DS, de Boer RA, Navis G, Slaets JP, et al. Universal risk factors for multifactorial diseases: LifeLines: a three-generation population-based study. EurJEpidemiol. 2008;23(1):67-74.
27. Scholtens S, Smidt N, Swertz MA, Bakker SJ, Dotinga A, Vonk JM, et al. Cohort Profile: LifeLines, a three-generation cohort study and biobank. Int J Epidemiol. 2015;44(4):1172-80.
28. van Waateringe RP, Slagter SN, van der Klauw MM, van Vliet-Ostaptchouk JV, Graaff R, Paterson AD, et al. Lifestyle and clinical determinants of skin autofluorescence in a population-based cohort study. Eur J Clin Invest. 2016;46(5):481-90.
29. Wouters HJ, van Loon HC, van der Klauw MM, Elderson MF, Slagter SN, Kobold AM, et al. No Effect of the Thr92Ala Polymorphism of Deiodinase-2 on Thyroid Hormone Parameters, Health-Related Quality of Life, and Cognitive Functioning in a Large Population-Based Cohort Study. Thyroid. 2017;27(2):147-55.
30. Koetsier M, Lutgers HL, de Jonge C, Links TP, Smit AJ, Graaff R. Reference values of skin autofluorescence. Diabetes Technol Ther. 2010;12(5):399-403.
31. Koetsier M, Nur E, Chunmao H, Lutgers HL, Links TP, Smit AJ, et al. Skin color independent assessment of aging using skin autofluorescence. Opt Express. 2010;18(14):14416-29. 32. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, 3rd, Feldman HI, et al. A new equation
to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604-12.
33. Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation. 2009;120(16):1640-5. 34. Bianco AC, Salvatore D, Gereben B, Berry MJ, Larsen PR. Biochemistry, cellular and molecular
biology, and physiological roles of the iodothyronine selenodeiodinases. Endocr Rev. 2002;23(1):38-89.
35. Boelen A, Kwakkel J, Alkemade A, Renckens R, Kaptein E, Kuiper G, et al. Induction of type 3 deiodinase activity in inflammatory cells of mice with chronic local inflammation. Endocrinology. 2005;146(12):5128-34.
36. Gereben B, McAninch EA, Ribeiro MO, Bianco AC. Scope and limitations of iodothyronine deiodinases in hypothyroidism. Nature reviews Endocrinology. 2015;11(11):642-52.
37. Cappola AR, Arnold AM, Wulczyn K, Carlson M, Robbins J, Psaty BM. Thyroid function in the euthyroid range and adverse outcomes in older adults. J Clin Endocrinol Metab. 2015;100(3):1088-96.
38. Zhang Y, Chang Y, Ryu S, Cho J, Lee WY, Rhee EJ, et al. Thyroid hormones and mortality risk in euthyroid individuals: the Kangbuk Samsung health study. J Clin Endocrinol Metab. 2014;99(7):2467-76.
39. Martin SS, Daya N, Lutsey PL, Matsushita K, Fretz A, McEvoy JW, et al. Thyroid Function, Cardiovascular Risk Factors, and Incident Atherosclerotic Cardiovascular Disease: The Atherosclerosis Risk in Communities (ARIC) Study. J Clin Endocrinol Metab. 2017;102(9):3306-15.
40. Mayer-Davis EJ, Monaco JH, Hoen HM, Carmichael S, Vitolins MZ, Rewers MJ, et al. Dietary fat and insulin sensitivity in a triethnic population: the role of obesity. The Insulin Resistance Atherosclerosis Study (IRAS). Am J Clin Nutr. 1997;65(1):79-87.
41. Samuel VT, Petersen KF, Shulman GI. Lipid-induced insulin resistance: unravelling the mechanism. Lancet. 2010;375(9733):2267-77.
42. Bianco AC, Silva JE. Intracellular conversion of thyroxine to triiodothyronine is required for the optimal thermogenic function of brown adipose tissue. J Clin Invest. 1987;79(1):295-300.
43. Silva JE. The thermogenic effect of thyroid hormone and its clinical implications. Ann Intern Med. 2003;139(3):205-13.
Sup
pl
em
en
tal
da
ta
Su ppl em en ta l T abl e 1. Mul tiv ar iat e lo gi sti c re gr es sio n an al ys es fo r i nc id en t t yp e 2 diab et es , c ar di ov as cul ar di se as e an d m or tali ty , at a m edian o 4-ye ar s f ol lo w -u p, w ith e xc lu sio n o f S AF . Ne w t yp e 2 d ia be te s Ne w ca rdi ov as cu la r di se as e De at h M ul tiv ar ia te OR (9 5% C I) P v al ue OR (9 5% C I) P v al ue OR (9 5% C I) P v al ue FT 4 (p m ol /l) 0. 90 (0 .8 6-0. 95 ) 2. 3 x 1 0 -4 1. 04 (0 .9 9-1. 08 ) 0.1 04 0. 98 (0 .9 3-1. 04 ) 0. 507 Age (y ea rs ) 1. 02 (1 .0 1-1.0 3) 2. 0 x 1 0 -4 1. 05 (1. 05 -1. 06 ) 1. 3 x 1 0 -32 1.1 0 (1. 09 -1. 11 ) 7. 4 x 1 0 -7 8 M al e g en de r ( y/ n) 0. 94 (0 .7 5-1 .1 8) 0. 59 5 1. 36 (1. 12 -1. 64 ) 0. 002 1. 44 (1. 16 -1. 79 ) 0. 001 BM I ( kg /m 2) 1. 06 (1. 04 -1. 09 ) 2. 0 x 1 0 -6 1. 04 (1. 01 -1. 06 ) 0. 002 1. 04 (1 .0 1-1. 07) 0. 00 6 Sy st . B P ( m m Hg ) 1. 00 (1. 00 -1. 01 ) 0.1 60 1. 01 (1. 00 -1. 01 ) 0. 020 1. 01 (1. 00 -1. 02 ) 0. 002 Cho le st er ol (m m ol /l) 0. 95 (0 .8 5-1 .0 6) 0. 382 1.1 0 (1. 00 -1. 20 ) 0. 061 0. 98 (0 .8 7-1 .0 9) 0. 651 Gl uc os e ( m m ol /l) 8.1 0 (6 .6 -10 .0) 2. 3 x 1 0 -86 0. 95 (0 .7 8-1 .1 6) 0. 58 8 0. 84 (0 .6 7-1. 05) 0.1 25 M et S ( y/ n) 1. 36 (1. 04 -1. 79 ) 0. 026 1. 07 (0 .8 3-1 .3 7) 0. 61 4 1.1 9 (0 .9 0-1. 58 ) 0. 222 Cu rr en t s m ok in g ( y/ n) 1. 32 (1. 02 -1. 72 ) 0.0 38 1. 85 (1 .5 0-2. 28 ) 7. 2 x 1 0 -9 2. 04 (1 .6 0-2.6 1) 1. 5 x 1 0 -8 Ch ol -lo w er in g t he ra py ( y/ n) 1. 34 (0 .9 2-1. 94 ) 0.1 30 1. 73 (1 .2 6-2. 36 ) 0. 001 0. 70 (0. 46 -1 .0 6) 0. 09 4 BP -lo w er in g t he ra py ( y/ n) 1. 28 (0 .9 7-1. 68 ) 0. 08 3 1.6 6 (1 .3 0-2. 10) 3. 3 x 1 0 -5 1. 23 (0 .9 4-1. 61 ) 0.1 26Su ppl em en ta l T abl e 2. Mul tiv ar iat e lo gi sti c re gr es sio n an al ys es fo r i nc id en t t yp e 2 diab et es , c ar di ov as cul ar di se as e an d m or tali ty , at a m edian o 4-ye ar s f ol lo w -u p, w ith e xc lu sio n o f S AF . Ne w t yp e 2 d ia be te s Ne w ca rdi ov as cu la r De at h M ul tiv ar ia te OR (9 5% C I) P v al ue OR (9 5% C I) P v al ue OR (9 5% C I) P v al ue FT 4 (p m ol /l) 0. 99 (0 .8 3-1 .1 8) 0. 91 7 1.1 2 (1. 00 -1. 24 ) 0. 04 6 1.0 3 (0 .8 7-1 .2 2) 0. 73 4 Age (y ea rs ) 1. 02 (1 .0 1-1.0 3) 2. 4 x 1 0 -4 1. 06 (1. 05 -1. 06 ) 2. 9 x 1 0 -33 1.1 0 (1. 09 -1. 11 ) 4. 2 x 1 0 -7 6 M al e g en de r ( y/ n) 0. 91 (0 .7 2-1 .1 4) 0. 41 2 1. 34 (1. 10 -1. 62 ) 0.0 03 1. 42 (1. 14 -1. 77 ) 0. 002 BM I ( kg /m 2) 1. 06 (1. 04 -1. 09 ) 5. 2 x 1 0 -7 1. 04 (1. 01 -1. 06 ) 0.0 03 1. 04 (1 .0 1-1. 07) 0. 00 5 Sy st . B P ( m m Hg ) 1. 00 (1. 00 -1. 01 ) 0. 231 1. 01 (1. 00 -1. 01 ) 0. 02 2 1. 01 (1. 00 -1. 02 ) 0. 002 Cho le st er ol (m m ol /l) 0. 96 (0 .8 6-1. 07) 0. 46 6 1.1 0 (1. 00 -1. 21 ) 0. 05 7 0. 98 (0 .8 7-1 .0 9) 0. 67 5 Gl uc os e ( m m ol /l) 8. 01 (6 .5 -9 .9 ) 4. 8 x 1 0 -86 0. 94 (0 .7 7-1. 14 ) 0. 526 0. 84 (0 .6 7-1. 05) 0.1 23 M et S ( y/ n) 1. 42 (1. 08 -1. 87 ) 0. 01 1 1. 05 (0 .8 2-1 .3 5) 0.69 5 1. 20 (0 .9 0-1. 58 ) 0. 211 Cu rr en t s m ok in g ( y/ n) 1. 26 (0 .9 7-1. 64 ) 0. 08 4 1. 84 (1 .5 0-2. 27 ) 7. 9 x 1 0 -9 2. 01 (1 .5 7-2. 58 ) 3. 0 x 1 0 -8 Ch ol -lo w er in g t he ra py ( y/ n) 1. 36 (0 .9 3-1. 98 ) 0.1 10 1. 71 (1 .2 5-2. 34 ) 0. 001 0. 70 (0 .4 6-1. 07) 0. 09 9 BP -lo w er in g t he ra py ( y/ n) 1. 21 (0 .9 2-1. 60 ) 0. 17 0 1.6 8 (1 .3 2-2 .1 3) 1. 9 x 1 0 -5 1. 22 (0 .9 4-1. 60 ) 0.1 43 Da ta a re p re se nt ed a s O R, o dd s r ati os ( 95 % c on fid en ce i nt er va l). B as el in e r isk f ac to rs w er e u se d t o p re di ct t he m ed ia n 4 y ea rs r isk o f t he o ut co m es of t yp e 2 d ia be te s a nd c ar di ov as cu la r d ise as e. F T3 , a ge , B M I, b lo od p re ss ur e, c ho le st er ol , g lu co se w er e d efi ne d a s c on tin uo us v ar ia bl es . M al e g en de cu rr en t s m ok in g, a nd m et ab ol ic s yn dr om e w er e d efi ne d a s c at ego ric al v ar ia bl es . Ab br ev ia tio ns : s ee T ab le 2 .
6
Supplemental Figure 1. Heatmaps showing the relationship between FT4 and FT3 and age and
