A 4-week high-AGE diet does not impair glucose metabolism and vascular function in obese individuals

BACKGROUND Accumulation of advanced glycation endproducts (AGEs) may contribute to the pathophysiology of type 2 diabetes and its vascular complications. AGEs are widely present in food, but whether restricting AGE intake improves risk factors for type 2 diabetes and vascular dysfunction is controversial. METHODS Abdominally obese but otherwise healthy individuals were randomly assigned to a specifically designed 4-week diet low or high in AGEs in a double-blind, parallel design. Insulin sensitivity, secretion, and clearance were assessed by a combined hyperinsulinemic-euglycemic and hyperglycemic clamp. Micro- and macrovascular function, inflammation, and lipid profiles were assessed by state-of-the-art in vivo measurements and biomarkers. Specific urinary and plasma AGEs Nε-(carboxymethyl)lysine (CML), Nε-(1-carboxyethyl)lysine (CEL), and Nδ-(5-hydro-5-methyl-4-imidazolon-2-yl)-ornithine (MG-H1) were assessed by mass spectrometry. RESULTS In 73 individuals (22 males, mean ± SD age and BMI 52 ± 14 years, 30.6 ± 4.0 kg/m2), intake of CML, CEL, and MG-H1 differed 2.7-, 5.3-, and 3.7-fold between the low- and high-AGE diets, leading to corresponding changes of these AGEs in urine and plasma. Despite this, there was no difference in insulin sensitivity, secretion, or clearance; micro- and macrovascular function; overall inflammation; or lipid profile between the low and high dietary AGE groups (for all treatment effects, P > 0.05). CONCLUSION This comprehensive RCT demonstrates very limited biological consequences of a 4-week diet low or high in AGEs in abdominally obese individuals. TRIAL REGISTRATION Clinicaltrials.gov, NCT03866343; trialregister.nl, NTR7594. FUNDING Diabetesfonds and ZonMw.

where D is the arterial diameter; ΔD is the distension; IMT the intima-media thickness; and PP the pulse pressure. Local carotid PP was estimated according to the calibration method described by Kelly and Fitchett (5), with the use of carotid tonometry waveforms as adapted by van Bortel et al. (6). This method assumes a constant difference between MAP and diastolic pressure along the arterial tree. PP can then be calculated at a carotid artery (PPcar) from the uncalibrated carotid pressure waveform

Radial Pulse Wave Analysis
Radial artery pulse wave analysis was measured in triplicate at the wrist of the right arm using tonometry (SphygmoCor v9; AtCor Medical), as described previously (7). In short, the central arterial waveform was derived from the peripheral arterial waveform using a validated transfer function. The augmentation index was defined as the difference between the first and second peak of the central arterial waveform, expressed as a percentage of the pulse pressure and corrected for heart rate. We used the median of 3 consecutive measurements. As in some participants more than 3 measurements were performed, we chose the measurements with the highest quality based on four criteria (8): (1) average pulse height above 100 units, (2) pulse height variation < 5%, (3) diastolic variation < 5%, and (4) systolic peak between 80 and 150 ms from the start of the wave. Measurements were scored on a scale from 0 to 4 based on the number of criteria met. Measurements with low quality scores (0 or 1) were excluded from the analysis. For one participant, this resulted in 2 measurements instead of 3, and these 2 measurements were averaged instead of using the median.

Flow-mediated dilation
Flow-mediated dilation (FMD) of the brachial artery was assessed by ultrasound echography in dual mode (MyLab70, Esaote) and recording of echo images on DVD, as described previously (7). These images were analyzed offline using a customwritten Matlab program (MyFMD; AP Hoeks, Department of Biomedical Engineering, Maastricht University Medical Center, Maastricht, the Netherlands). After a 5-minute reference period, a pneumatic cuff placed around the participant's right forearm was inflated to 200 mmHg for 5 minutes to ensure arterial occlusion. After 5 minutes of arterial occlusion, the cuff was deflated and images were obtained for an additional 5 minutes. The FMD response was quantified as the maximal percentage change in post occlusion arterial diameter relative to the baseline diameter.

Laser Doppler Flowmetry
Skin blood flow was measured both in the basal state, during acute hyperinsulinemia, and during acute local heating, as described previously, by means of a laser-Doppler system (Periflux 5000; Perimed, Järfalla, Sweden) equipped with two thermostatic laser-Doppler probes (PF457; Perimed) at the dorsal side of the wrist of the left hand (9). The laser-Doppler output was recorded for 35 minutes with a sample rate of 32 Hz, which gives semi quantitative assessment of skin blood flow expressed in arbitrary perfusion units.

Flowmotion
Since skin microvascular flowmotion (SMF) has predominantly been observed in participants with a skin temperature above 29.3°C (10), the laser-Doppler probe was set at 30°C. The skin blood flow signal was transformed into five different SMF components by means of a Fast-Fourier transform algorithm using dedicated custom build software (FlowPSD; AP Hoeks, Department of Biomedical Engineering, Maastricht University Medical Center, Maastricht, the Netherlands ). The frequency spectrum between 0.01 and 1.6 Hz was divided into five components: (1) endothelial, 0.01-0.02 Hz, (2) neurogenic, 0.02-0.06 Hz, (3) myogenic, 0.06-0.15 Hz, (4) respiratory, 0.15-0.40 Hz, and (5) heartbeat, 0.40-1.60 Hz (11). Additionally, total SMF energy was obtained by the sum of the power density values of the total frequency spectrum.

Heat-induced skin hyperemic response
With the second probe, skin blood flow was first recorded unheated for 2 minutes to serve as a baseline. After the 2 minutes of baseline, the temperature of the probe was rapidly and locally increased to 44°C and was then kept constant until the end of the registration. The heat-induced skin hyperemic response was expressed as the percentage increase in average perfusion units during the 33-minute heating phase over the average baseline perfusion units.

Retinal imaging
Fundus photographs were obtained to assess static retinal microvascular diameters with a non-mydriatic manual-focus fundus camera (Canon). To this end, three opticdisc centered photographs of the right eye were taken. The detailed procedure has been explained elsewhere (12). In short, retinal arteriolar and venular diameters were measured at an area 0.5-1.0 disc diameter away from the optic disc margin with semiautomatic analyzing software (Vesselmap 3.0, Visualis, Imedos Systems UG).
Arteriolar and venular diameters were averaged to central retinal arteriolar (CRAE) and venular (CRVE) equivalents using the Parr-Hubbard formula (13). Vessel diameters are presented in µm, as one measuring unit of the imaging device relates to 1 µm in the model of Gullstrand's normal eye. The same researcher took all images, and all images were analysed by the same independent researcher, unaware of a participant's treatment allocation. Participants with retinal pathologies that influence microvascular calibers (e.g. macular degeneration, n = 1) were excluded from the analyses.

Habitual food intake and dietary advanced glycation endproducts
We assessed habitual dietary intake by a validated 253-item food frequency questionnaire (FFQ) (14). This FFQ contains 101 questions on consumption with a reference period of one year. The FFQ collected information on the intake of major food groups. Food intake was determined by the combination of frequency questions with quantity questions. For the frequency questions, 11 options were available ranging from "not used" to 7 days/week. For the quantity questions, variable options were available based on fourteen standard household servings, ranging from < 1/day to > 12/day. Average daily consumption of food items was then calculated by multiplying the frequency and amount. Energy and nutrient intakes were subsequently determined by transcribing food items into food codes embedded in the Dutch Food Composition Table 2011 (15). Additionally, we determined the Dutch Healthy Diet (DHD) index based on this food intake data. The DHD-index is a measure of diet quality as it assesses adherence to the Dutch dietary guidelines (16). A higher index has been associated with more nutrient-dense diets and lower risk of mortality (17,18).
Dietary AGE intake was determined by coupling the consumption of food items within the FFQ to our dietary AGE database (19). In this database, three major AGEs, CML, CEL, and MG-H1, were quantified in protein fractions of food products using highly specific ultra performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS). In total, this database includes over 200 food products commonly consumed in a Western diet. For each participant, AGE intake was estimated as described previously (20). Some of the food products in the FFQ were not analyzed for AGEs content. AGE content of these specific products were estimated by matching them to other products that were comparable in macronutrient profile and preparation method. For example, for several fresh vegetables boiled in water, such as endive, beets, leek, and spinach, the same AGE content was used. By comparison, jarred peas and carrots were measured separately from fresh peas and carrots, as AGEs in jarred peas and carrots are higher as they contain added sugar and are heated to prolong shelf life (19).

Skin autofluoresence
Skin autofluoresence (SAF) was measured with the AGE Reader (DiagnOptics Technologies BV, Groningen, The Netherlands). The AGE reader is a desktop device that uses the characteristic fluorescent properties of certain AGEs to estimate the level of AGE accumulation in the skin. Technical details of this noninvasive method have been described more extensively elsewhere (21). In short, the AGE Reader illuminates a skin surface of 4 cm2 guarded against surrounding light, with an excitation wavelength range of 300 to 420 nm, with a peak excitation of 370 nm. SAF was calculated as the ratio between the emission light from the skin in the wavelength range of 420 to 600 nm (fluorescence) and excitation light that is reflected by the skin (300-420 nm), multiplied by 100 and expressed in arbitrary units. Participants were asked not to use any sunscreen or self-browning creams on their lower arms within 2 days before the measurement. SAF was measured at room temperature in a semidark environment, where participants were at rest in a seated position. The inner side of the forearm ≈4 cm below the elbow fold of a participant was positioned on top of the device, as described by the manufacturer. The mean of 3 consecutive measurements was used in the analyses. Percentage difference in intake of AGEs between the habitual diet (assessed with food frequency questionnaires) and during the intervention (assessed as the average daily intake from two five-day dietary logs). Bar plots indicate mean ± SD. Black circles indicate individuals that showed no increase or decrease in respective AGE intake during their intervention. n=34 for the low-AGE group, n=38 for the high-AGE group.

Supplemental Tables
Supplemental Table 1 Comparison of baseline characteristics from participants included in the complete case analysis to those excluded resulting from missing the primary outcome.   were assessed from two 24-hour recalls in week 3 and week 4 of the intervention. Differences between intervention groups were tested by a one-factor ANCOVA with energy intake, sex, and age as covariates. 1 Energy intake was not included as a covariate. ) of micronutrients were assessed from two five-day dietary logs at week 1 and week 4 of the intervention. Differences between intervention groups were tested by a one-factor ANCOVA with energy intake, sex, and age as covariates. 1 Dietary logs were not returned by one participant in the low AGE group.   .32 Values are presented as means ± SD. Within-group changes were evaluated with a paired-samples t test. Overall differences after the low compared to high AGE diet were evaluated with a one-way ANCOVA with adjustment for age, sex, and the baseline variable of interest. Abbrevations: Β-GS: beta-cell glucose sensitivity. ISR: Insulin secretion rates. M/I: insulin sensitivity adjusted for plasma insulin. 1 n=35 for low AGE, n=38 for high AGE. 2 n=34 for low AGE, n=38 for high AGE. 3 n=34 for low AGE, n=36 for high AGE. Table 6 Effects of a 4-week low-and high AGE diet on micro and macrovascular function of abdominally obese individuals .46 Values are presented as means ± SD. Within-group changes were evaluated with a paired-samples t test. Overall differences after the low compared to high AGE diet were evaluated with a one-way ANCOVA with adjustment for age, sex, and the baseline variable of interest. Abbreviations: Gamma-GT: Gamma-glutamyltransferase.

Supplemental
Supplemental Table 9 Multivariate-adjusted associations between indices of AGE intake and outcomes in 72 abdominally obese individuals after a low and high AGE diet.
AGEs SD and 95% CIs indicate the difference in outcome per unit change in determinant. Please note that AGEs in urine and plasma were standardized. Associations were adjusted for age, sex, and intake of carbohydrates, fat, and protein as energy-percentages. Statistically significant associations are shown bold.