YCLNU3682_proof ■ 8 December 2018 ■ 1/7

Clinical Nutrition xxx (xxxx) xxx

55
Contents lists available at ScienceDirect 56
57
58
Clinical Nutrition 59
60
journal homepage: http://www.elsevier.com/locate/clnu 61
62
63
Original article 64
65
1 Gastrointestinal effects of extra-virgin olive oil associated with lower 66
2 67
3 Q4 postprandial glycemia in type 1 diabetes 68
4 69
5 Q3 Lutgarda Bozzetto a, Antonio Alderisio a, Gennaro Clemente b, Marisa Giorgini a, 70
6
Francesca Barone a, Ettore Griffo a, Giuseppina Costabile a, Claudia Vetrani a, 71
7
8 Paola Cipriano a, Angela Giacco a, Gabriele Riccardi a, Angela Albarosa Rivellese a, 72
73
9 Giovanni Annuzzi a, * 74
10 a
Department of Clinical Medicine and Surgery, Federico II University, Naples, Italy 75
11 b
Institute for Research on Population and Social Policies (IRPPS), National Research Council, Fisciano, SA, Italy 76
12 77
13 78
14 a r t i c l e i n f o s u m m a r y 79
15 80
16 Article history: Objective: To explore the possible mechanisms behind the lower glycemic response observed when 81
17 Received 2 April 2018 extra-virgin olive oil (EVOO) is added to a high-glycemic index meal in patients with type 1 diabetes 82
Accepted 25 November 2018
18 (T1D). 83
19 Research design and methods: According to a randomized cross-over design, eleven T1D patients (6
Keywords:
84
Q5 women, 5 men) on insulin pump consumed in the metabolic ward, one week apart, three high-glycemic
20 Type 1 diabetes 85
21 index meals differing only for amount and quality of fat: high-monounsaturated fat (EVOO), high-
Extra-virgin olive oil
saturated fat (Butter), and low-fat (LF). Before and after the meals, blood glucose (continuous glucose
86
22 Dietary fat 87
Postprandial glycemia
monitoring), gastric emptying rate (ultrasound technique), and plasma concentrations of glucagon-like
23 88
Incretins peptide 1 (GLP-1) and glucose-dependent insulinotropic polypeptide GIP (ELISA), glucagon (RIA), and
24 lipids (colorimetric assays) were evaluated. 89
Gastric emptying
25 Postprandial triglycerides Results: Blood glucose iAUC (mmol/lx360 min) was lower after the EVOO (690 ± 431) than after the 90
26 Butter (1320 ± 600) and LF meals (1007 ± 990) (M ± SD, p ¼ 0.041 by repeated measures ANOVA). Gastric 91
27 antrum volume was significantly larger in the early (60e90 min) postprandial phase (106 ± 21 vs. 92
28 90 ± 16 ml, p ¼ 0.048) and significantly smaller in the late phase (330e360 min) (46 ± 10 vs. 57 ± 22 ml, 93
29 p ¼ 0.045) after the EVOO than after Butter meal. EVOO significantly increased postprandial GLP-1 iAUC 94
30 (261 ± 311) compared to Butter (189 ± 349) (pmol/Lx180 min, p ¼ 0.009). Postprandial GIP and glucagon
95
31 responses were not significantly different between EVOO and Butter. Postprandial triglyceride iAUC was
96
32 significantly higher after EVOO (100 ± 53) than after Butter (65 ± 60) (mmol/l  360 min, p ¼ 0.048).
Conclusions: Changes in gastric emptying and GLP-1 secretion and reduced glucose absorption through
97
33 98
glucose-lipid competition may contribute to lower glycemia after a high-glycemic index meal with EVOO
34 99
in T1D patients.
35 Clinical trials number: NCT02330939. 100
36 © 2018 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved. 101
37 102
38 103
39 104
40 105
41 106
42 1. Introduction 107
43 108
44 Adding extra-virgin olive oil (EVOO) to a high glycemic index 109
45 meal attenuates the postprandial glucose response observed when 110
46 the meal is consumed with butter or very small amounts of fat, as 111
47 Abbreviations: CGM, continuous glucose monitoring; EVOO, extra-virgin olive previously shown in a controlled study in real life conditions, in 112
oil; GE, gastric emptying; GIP, glucose-dependent insulinotropic polypeptide; GLP- patients with type 1 diabetes on insulin pump [1]. How EVOO in-
48 113
1, glucagon-like peptide 1; LF, low-fat; MUFA, monounsaturated fatty acids; PUFA,
49 polyunsaturated fatty acids; SFA, saturated fatty acids; T1D, type 1 diabetes.
fluences postprandial glycemia is unknown. The main pathophys- 114
50 * Corresponding author. Department of Clinical Medicine and Surgery, Federico II iological regulators of postprandial glucose response may be 115
51 University, Via Pansini, 5, 80131 , Naples, Italy. Fax number: þ39 0817462311.
116
E-mail address: annuzzi@unina.it (G. Annuzzi).
52 117
53 118
https://doi.org/10.1016/j.clnu.2018.11.015
54 0261-5614/© 2018 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved. 119

Please cite this article as: Bozzetto L et al., Gastrointestinal effects of extra-virgin olive oil associated with lower postprandial glycemia in type 1
diabetes, Clinical Nutrition, https://doi.org/10.1016/j.clnu.2018.11.015
 YCLNU3682_proof ■ 8 December 2018 ■ 2/7

2 L. Bozzetto et al. / Clinical Nutrition xxx (xxxx) xxx

1 involved, including changes in gastric emptying rate and gastro- one-week intervals, in a random sequence determined by card 66
2 intestinal hormones [2]. drawing. The three meals had a similar carbohydrate content but 67
3 Gastric emptying (GE) is a major determinant of postprandial different amounts and type of fat: 1) high-monounsaturated fat 68
4 glycemia and is finely regulated by the characteristics of the meal. (EVOO), 37 g, 2) high-saturated fat (Butter), 43 g, and 3) low-fat 69
5 In particular, the amount of fat in a meal is able to slow down (LF), 8 g EVOO. 70
6 gastric emptying rate in healthy individuals [3] and in people with The participants came to the metabolic ward in the late morning 71
7 type 1 diabetes [4]. As to the quality of fat, monounsaturated fatty and consumed the test meals at lunchtime under the supervision of 72
8 acids (MUFA) have been shown to slow down GE more than n-3 an expert dietitian. The test meals had been previously prepared in 73
9 polyunsaturated (PUFA), n-6 PUFA, and saturated fatty acids (SFA) a metabolic kitchen and kept frozen until the test day, when they 74
10 in healthy people [5]. were defrosted and reheated on the cooker. The patients consumed 75
11 The quality of meal fat could influence gastric emptying through the meals and drank 250 ml of water within 15 min. 76
12 changes in postprandial secretion of glucagon-like peptide 1 (GLP- In case of pre-meal blood glucose levels outside the 5e8 mmol/l 77
13 1) and glucose-dependent insulinotropic polypeptide (GIP). In range or a rapid decrease/increase of glucose levels (±3.3 mmol/l) 78
14 particular, MUFA seem to enhance GLP-1 secretion more than SFA during the previous 60 min according to CGM measurement, the 79
15 in both healthy people [6] and in patients with type 2 diabetes test meal was postponed to the next week. On the three experi- 80
16 [7e9], while GIP secretion seems to be increased by SFA and mental days, the participants were asked to consume at home, no 81
17 reduced by fish oil compared to other types of fat in healthy people later than 7.00, the same light breakfast. They were asked to avoid 82
18 [10,11]. GLP-1 could influence postprandial glucose response in strenuous physical activity on the day before and on the morning of 83
19 people with type 1 diabetes also independently of its effects on the test meal, and to refrain from snacking or making insulin bo- 84
20 gastric emptying, by exerting its glucagonostatic effects [12]. luses or changes in insulin basal rate over the 6 h prior to the test 85
21 Another mechanism through which different types of fat could meals. In female participants, the experiments were not performed 86
22 influence postprandial glycemia is by eliciting a different post- during menses. After meals, the subjects were asked to remain in 87
23 prandial lipemic response, since postprandial lipid and glucose the upright position or sitting over 6 h, either reading, chatting, or 88
24 metabolism are reciprocally influenced and regulated [13]. watching TV. Pre-meal insulin doses, injected just before eating, 89
25 Therefore, the aim of the present study was to explore the were based on individual insulin/glycemic load ratio determined 90
26 possible mechanisms for the reduced postprandial glucose during the patients’ educational sessions with the study team [14]. 91
27 response to a high glycemic index meal enriched with EVOO For each subject, the same insulin dose was administered before 92
28 observed in our previous home-based study. For this purpose, we each of the three experimental meals. No changes from habitual 93
29 performed a new controlled study in the metabolic ward of our insulin basal rate were made during the 6-h postprandial phase and 94
30 Unit, involving patients with type 1 diabetes on insulin pump, to between the three different test meals. 95
31 evaluate the effects of different dietary fats on postprandial blood Over the whole experimental period, participants underwent 96
32 glucose, gastric emptying and plasma levels of gastrointestinal CGM. Moreover, on the test days, before and over the 6 h following 97
33 hormones and lipids. the meals, participants underwent gastric ultrasound for the eval- 98
34 uation of gastric emptying rate and venous blood sampling for the 99
35 2. Material and methods measurement of GLP-1, GIP, glucagon, and concentrations of plasma 100
36 lipids. 101
37 Fourteen patients with type 1 diabetes (8 women and 6 men) 102
38 were recruited at the outpatient diabetes clinic of the Federico II 103
2.2. Test meals composition
39 University Teaching Hospital (Naples, Italy). They were enrolled in 104
40 the study after giving their written informed consent. Inclusion 105
The meals consisted of white rice (60 g), white bread (75 g),
41 criteria were treatment with continuous subcutaneous insulin 106
minced beef (90 g), and banana (180 g) prepared with EVOO (37 g)
42 infusion for at least 6 months and an HbA1c less than 8.0% 107
in the EVOO meal, butter (43 g) in the Butter meal, and EVOO (8 g)
43 (64 mmol/mol). Exclusion criteria were pregnancy, celiac disease, 108
in the Low-fat meal. The whole content of Butter and EVOO was
44 severe microvascular and macrovascular diabetes complications 109
added to the meals before freezing. Energy content and total
45 including autonomic neuropathy possibly influencing gastric 110
amount of fat were similar for the EVOO (988 kcal; 40.5 g) and
46 emptying, and any other chronic or acute disease seriously affecting 111
Butter meals (982 kcal; 39.4 g); conversely, the Low-fat meal had a
47 health status. 112
lower energy content (721 kcal), due exclusively to the lower
48 Patients meeting the inclusion criteria were asked to participate 113
content of fat (10.6 g). While macronutrient composition was
49 in the study at their regular outpatient follow-up visits. One subject 114
similar for amount of carbohydrate (130 g) and protein (34 g) in all
50 was excluded from the final analysis because of a hypoglycemic 115
test meals, there was a substantial difference for saturated and
51 event during a test meal, and two subjects because of poor 116
monounsaturated fat between the EVOO (6.5 g and 27.9 g), Butter
52 compliance to the study design. Therefore, the analyses were per- 117
(22.1 g and 11.1 g), and Low-fat (2.2 g and 6.6 g) meals, respectively.
53 formed on 11 participants. The study protocol was approved by the 118
The small amount of fat in the Low-fat meal was necessary to allow
54 Federico II University Ethics Committee and registered at Clinical 119
the preparation of a meal with a similar food composition as the
55 Trials.gov, registration number: NCT02330939. 120
two fat-rich meals.
56 121
57 2.1. Study design 122
58 2.3. Continuous glucose monitoring 123
59 Before the intervention, patients underwent a one-week run-in 124
60 period on continuous glucose monitoring (CGM), and were asked to Glucose monitoring was performed using the Enlite™ Sensor 125
61 complete a seven-day dietary record. The run-in aimed at opti- Medtronic (n ¼ 7 participants) or the Dexcom G4® (n ¼ 4 partici- 126
62 mizing basal infusion rate and insulin prandial dose calculation by pants) systems. At the end of the experimental period, data from 127
63 insulin/glycemic load ratio, determined as previously described CGM and insulin pump were downloaded by dedicated informatics 128
64 [14]. After run-in, according to a triple crossover design, partici- platforms. Participants used the CGM system integrated with their 129
65 pants consumed one of three experimental meals, alternating, at insulin pump, as it was the one they were accustomed to. 130

Please cite this article as: Bozzetto L et al., Gastrointestinal effects of extra-virgin olive oil associated with lower postprandial glycemia in type 1
diabetes, Clinical Nutrition, https://doi.org/10.1016/j.clnu.2018.11.015
 YCLNU3682_proof ■ 8 December 2018 ■ 3/7

L. Bozzetto et al. / Clinical Nutrition xxx (xxxx) xxx 3

1 2.4. Gastric emptying rate between EVOO and butter [1]. Being a mechanistic study, multiple 66
2 endpoints were selected, namely postprandial changes in gastric 67
3 Gastric emptying rate was evaluated using a MINDRAY M7 emptying rate, gastrointestinal hormones, and blood lipids. 68
4 digital echographer with a convex-array scan (range 2e6 MHz). Data are expressed as mean ± standard deviation (SD) unless 69
5 Ultrasonographic scanning was done at the final portion of the otherwise stated. Differences in postprandial glucose, hormones, 70
6 stomach (antrum and pylorus) where constant viewing makes and lipid profiles were evaluated by two-way repeated-measures 71
7 volume measurement simpler. Scanning was performed with the ANOVA. Postprandial concentrations were included as levels of the 72
8 subjects standing up to favor air flow towards the fundus of the within-subject factor time, and EVOO, Butter, and Low-fat were 73
9 stomach. In order to calculate antral volume, three sagittal scans included as levels of the within-subject factor test-meal. Post- 74
10 were performed to measure anteroposterior (a, c and e, respec- prandial incremental area (iAUC) was calculated by the trapezoidal 75
11 tively) and longitudinal (b, d and f, respectively) diameters. The first rule as the area under the curve above the baseline value. Differ- 76
12 scan was taken on the angulus (transitional region between the ences between the three experimental meals in iAUCs or single 77
13 body of the stomach and the pylorus); the second, at an interme- time-point values were evaluated by one-way repeated measures 78
14 diate level corresponding to the superior mesenteric vein, which is ANOVA or, when data were not normally distributed, by Friedman's 79
15 easily detectable; the third, at the level of the pyloric sphincter analysis of variance by ranks. A p-value <0.05 was considered 80
16 where the thickness of the muscle wall is visible. Finally, the antral significant. Statistical analysis was performed according to stan- 81
17 length from the angulus to the pylorus (h) was measured with a dard methods using the Statistical Package for Social Sciences 82
18 transversal scan. The volume of the antrum was calculated software 21.0 (SPSS/PC; SPSS, Chicago, IL, USA). 83
19 using the formula by Bolondi et al. [15]: 84
20 0.065  h  (2ab þ 2efþ4cd þ cb þ ad þ ed þ cf). This method has 3. Results 85
21 been previously validated against the scintigraphic method [16]. A 86
22 limitation of the ultrasound technique is that it is unable to 3.1. Participants’ characteristics 87
23 differentiate the relative emptying of solid, aqueous, and oil 88
24 components. The study participants (5 men and 6 women) were 41 ± 9 89
25 (mean ± SD) years old and had a BMI of 24.9 ± 2.2 kg/m2. Diabetes 90
26 2.5. Laboratory assessment duration was 25 ± 9 years, and blood glucose control was accept- 91
27 able (HbA1c 7.0 ± 0.6%; 53 ± 6 mmol/mol). Their total daily insulin 92
28 2.5.1. Plasma gastrointestinal hormones dose was 41.9 ± 10.9 IU. As determined based on the individual 93
29 Blood samples were drawn before and 15, 30, 60, 90, 120, and insulin/glycemic load ratios, insulin doses administered before the 94
30 180 min after meal for the evaluation of GLP-1, GIP, and plasma three experimental meals were the same for each subject, ranging 95
31 levels of glucagon. Blood in EDTA- or EDTA and aprotinin (for the between 5 and 16 IU. One participant had background retinopathy 96
32 GLP-1 assay) tubes was centrifuged and plasma stored at 80  C and two had both background retinopathy and peripheral 97
33 until measurement. Active GLP-1 was assayed by a nonradioactive, neuropathy. 98
34 highly specific sandwich ELISA (Merck- Millipore, Darmstadt, Ger- Gastric volume and plasma concentrations of glucose, gastro- 99
35 many), which has a 100% cross-reactivity with the active isoforms intestinal hormones, and lipids did not differ before the EVOO, 100
36 of GLP-1 (7e36 amide and 7e37 glycine extended) but no reactivity Butter, and Low-fat meals. Data are reported in Supplemental 101
37 with inactive isoforms (9e36 amide and 9e37 glycine extended), Table S1. 102
38 GLP-2 or glucagon [17]. Total GIP was assayed by a nonradioactive, 103
39 highly specific sandwich ELISA (Merck-Millipore, Darmstadt, Ger- 3.2. Postprandial glycemia 104
40 many) with 100% cross reactivity with human GIP (1e42) and GIP 105
41 (3e42). The intra- and interassay coefficients of variation of the Postprandial blood glucose response to the EVOO meal was 106
42 GLP-1 and GIP assays were <5% and <10%, respectively. Plasma blunted compared with the Butter and Low-fat meals (p < 0.037 for 107
43 glucagon was assayed by competitive RIA using a rabbit antiserum time  meal interaction by two-way repeated measures ANOVA) 108
44 against a glucagonealbumin conjugate (Euria-Glucagon, Euro- (Fig. 1). Blood glucose progressively increased in the later time of 109
45 Diagnostica, Malmo €, Sweden). Glucagon in standards and samples observation (3e6 h) after the two fat meals, at variance with the 110
46 compete with 125I labeled glucagon in binding to the antibodies in Low-fat meal, which peaked at 2 h. Differences between the EVOO 111
47 a two-step incubation. Antibody-bound 125I-glucagon is separated and Butter meals were also evident in the later postprandial phase. 112
48 from the unbound fraction using double antibody solid phase and Blood glucose iAUC was significantly lower after the meal with 113
49 measured in a gamma counter. The limit of sensitivity for the EVOO (690 ± 431) than with Butter (1320 ± 600) or Low-fat 114
50 glucagon assays was 3 pmol/l. The intra- and interassay coefficients (1007 ± 990) (mmol/l  360 min; p ¼ 0.041, repeated measures 115
51 of variation were <10%. ANOVA; EVOO vs. Butter, p ¼ 0.004). 116
52 117
53 2.5.2. Other measurements 3.3. Gastric emptying 118
54 Plasma triglyceride (Roche Molecular Biochemicals, Mannheim, 119
55 Germany) and FFA (Wako Chemicals GmbH, Neuss, Germany) were Gastric antrum volume in the early postprandial phase was 120
56 measured before and 60, 120, 180, 240, 300 and 360 min after meals significantly larger after the EVOO than after the Butter meal (mean 121
57 by a Roche Cobas Mira autoanalyzer (ABX Diagnostics, Montpellier, values 60e90 min: 106 ± 21 vs. 90 ± 16 ml, p ¼ 0.048), indicating a 122
58 France) using a colorimetric assay. All laboratory analyses were slower gastric emptying after EVOO (Fig. 2). Conversely, at the end 123
59 performed blinded to the assigned treatment. of the observation, antrum volume was significantly smaller after 124
60 the EVOO than after the Butter meal (mean values 330e360 min: 125
61 2.6. Statistical analysis 46 ± 10 vs. 57 ± 22 ml, p ¼ 0.045). 126
62 Combining all meals, in the late postprandial phase, blood 127
63 A sample size of 13 participants with a randomized crossover glucose levels (mean values 300e330 min) were significantly and 128
64 design was determined. In our previous study, this size had been positively correlated with gastric antrum volumes at 330 min 129
65 sufficient to discriminate the postprandial blood glucose patterns (r ¼ 0.528, p ¼ 0.002) and 360 min (r ¼ 0.493, p ¼ 0.004). 130

Please cite this article as: Bozzetto L et al., Gastrointestinal effects of extra-virgin olive oil associated with lower postprandial glycemia in type 1
diabetes, Clinical Nutrition, https://doi.org/10.1016/j.clnu.2018.11.015
 YCLNU3682_proof ■ 8 December 2018 ■ 4/7

4 L. Bozzetto et al. / Clinical Nutrition xxx (xxxx) xxx

1 8 66
2 67
3 68
4 6
69
5 70
6 71
7 72
Blood glucose

4
8 73
mmol/l

9 74
10 75
11 2 76
12 77
13 78
14 0 79
15 80
16 81
17 82
-2
18 0 30 60 90 120 150 180 210 240 270 300 330 360 83
19 84
Minutes after meal
20 85
21 Fig. 1. Absolute changes in blood glucose concentrations after EVOO (full squares), Butter (empty circles), and Low-fat (empty squares) meals (p ¼ 0.037 for time  meal interaction 86
22 by two-way repeated measures ANOVA; EVOO vs. Butter, p ¼ 0.004). 87
23 88
24 89
25 90
26 90 91
27
28
* *
92
93
Gastric antrum volume

29 70 94
30 95
31 96
32 50 97
ml

33 98
34 99
30
35 100
36 101
37 102
10
38 103
39
* * 104
40 -10 105
41 0 30 60 90 120 150 180 210 240 270 300 330 360 106
42 Minutes after meal
107
43 108
44 Fig. 2. A. Absolute changes in gastric antrum volume after EVOO (full squares), Butter (empty circles), and Low-fat (empty squares) meals. Data in the figure are means and SEM. 109
45 *p < 0.05 EVOO vs. Butter. 110
46 111
47 3.4. Postprandial hormones postprandial changes could explain some differences between the 112
48 statistical tests. The postprandial concentration of glucagon 113
49 GLP-1 plasma concentrations at 15, 30, and 60 min increased increased significantly in EVOO at 15, 60, and 120 min when 114
50 significantly more after the EVOO than after the Butter and LF meals compared to LF, and at 15 and 120 min when compared to the 115
51 (Fig. 3A). Postprandial GLP-1 iAUC was significantly higher after the butter meal. The mean concentration at all time-points 116
52 EVOO (261 ± 311) than after the Butter (189 ± 349) and LF meals (15e180 min) was significantly higher after both the EVOO and 117
53 (180 ± 337) (pmol/l  180 min; p ¼ 0.009 by repeated measures Butter (46.3 ± 8.0 and 47.0 ± 8.7, respectively) than after the LF meal 118
54 ANOVA; p ¼ 0.028 EVOO vs. Butter, p ¼ 0.004 EVOO vs. LF) (Fig. 3B). (42.5 ± 8.0) (pmol/l, p ¼ 0.013 by repeated measures ANOVA) 119
55 There was a two-fold increase in plasma GIP concentrations (Fig. 3E). However, glucagon postprandial iAUCs were not signifi- 120
56 after the EVOO and Butter meals than after LF meal, with significant cantly different between the three meals (p ¼ 0.503) (Fig. 3F). 121
57 differences between the two fat-rich meals and the low fat meal at 122
58 all postprandial time-points (Fig. 3C). Postprandial GIP iAUC was 3.5. Postprandial lipids 123
59 significantly higher after the EVOO (18873 ± 7250) and Butter 124
60 (17217 ± 5344) than after LF meals (10634 ± 3680) (pmol/ Plasma triglyceride concentrations increased significantly more 125
61 l  180 min; p < 0.001 by repeated measures ANOVA; p ¼ 0.001 after the two fat-rich meals than after the low-fat meal, with a 126
62 EVOO vs. LF, p ¼ 0.001 Butter vs. LF) (Fig. 3D). significantly higher increase after EVOO than after Butter at 60, 120, 127
63 Plasma glucagon concentrations increased only slightly after the and 360 min (p ¼ 0.005 for time  meal interaction by two-way 128
64 three meals (on average less than 10%) with a polyphasic pattern repeated measures ANOVA) (Fig. 4A). Postprandial triglyceride 129
65 and without an evident peak. The high variability due to the small iAUC was significantly higher after EVOO (100 ± 53) than after the 130

Please cite this article as: Bozzetto L et al., Gastrointestinal effects of extra-virgin olive oil associated with lower postprandial glycemia in type 1
diabetes, Clinical Nutrition, https://doi.org/10.1016/j.clnu.2018.11.015
 YCLNU3682_proof ■ 8 December 2018 ■ 5/7

L. Bozzetto et al. / Clinical Nutrition xxx (xxxx) xxx 5

1 3 66
2 67
3 68
4 2 69
400
5 *† 70
6 71
300
7 1 72
8 73
9 200 74
10 0 75
11 100 76
12 77
13 -1 0 Categoria 1 78
0 15 30 45 60 75 90 105 120 135 150 165 180 EVOO Butter LF
14 79
15 80
16 81
17
120
D. 82
18 83
19 84
20 90 20000
85
†
21 86
22
† 87
†
23 60 † 88
†
24 10000
89
25 90
26 30 91
27 92
28 93
0 0 Categoria 1

29 0 15 30 45 60 75 90 105 120 135 150 165 180 EVOO Butter LF 94

30 Minutes after meal 95
31 96
32 97
33 E. 10 F. 98
34 † 99
*†
35 8 100
*†
36 101
*
37 6 1500 102
†
38 103
39 4 104
1000
40 105
41 2 106
42 500 107
0
43 108
44 109
-2 0
45 0 15 30 45 60 75 90 105 120 135 150 165 180 EVOO Butter
Categoria 1

LF 110
46 Minutes after meal 111
47 112
48 Fig. 3. Absolute changes in plasma concentrations of GLP-1 (A), GIP (C), and glucagon (E) after EVOO (full squares), Butter (empty circles), and Low-fat (empty squares) meals. GLP-1 113
(B), GIP (D), and glucagon (F) postprandial iAUCs after the three test meals. Data in the figure are means and SEM. *p < 0.05 vs. Butter; yp < 0.05 vs. Low-fat.
49 114
50 115
51 Butter (65 ± 60) and LF meals (23.9 ± 23.7) (mmol/l  360 min; differential changes in gastric emptying, increased GLP-1 secre- 116
52 p < 0.001 by repeated measures ANOVA; p ¼ 0.048 EVOO vs. Butter, tion, a more pronounced postprandial triglyceride response and 117
53 p < 0.001 EVOO vs. LF, p < 0.014 Butter vs. LF) (Fig. 4B). lower postprandial free fatty acids inhibition may contribute to 118
54 Postprandial FFA concentrations were less suppressed after the this effect. 119
55 EVOO meal than after the Butter and LF meals, with significant In our study, EVOO slowed down early gastric emptying, as 120
56 differences at 60 and 120 min (Fig. 4C). The difference in post- shown by significant differences in early postprandial gastric vol- 121
57 prandial FFA iAUCs after the EVOO and Butter meals tended to be umes between EVOO and butter intake. This suggests that only 122
58 statistically significant (80 ± 91 vs.203 ± 174 mEq/L  360 min, early postprandial changes in gastric emptying contributed to 123
59 respectively; p ¼ 0.066) (Fig. 4D). reducing postprandial glycemia. During the late postprandial 124
60 phase, in line with a bidirectional relationship between gastric 125
61 4. Discussion emptying and hyperglycemia [18], a more relevant factor slowing 126
62 down gastric emptying may became the higher blood glucose 127
63 The present study confirms, in a controlled setup, our previous levels. This is also suggested, in our study, by the significant and 128
64 home-based findings that the addition of EVOO to a high-glycemic positive correlations between blood glucose concentrations and 129
65 index meal reduces postprandial glycemia, and shows that gastric volumes observed 6 h after the meals. 130

Please cite this article as: Bozzetto L et al., Gastrointestinal effects of extra-virgin olive oil associated with lower postprandial glycemia in type 1
diabetes, Clinical Nutrition, https://doi.org/10.1016/j.clnu.2018.11.015
 YCLNU3682_proof ■ 8 December 2018 ■ 6/7

6 L. Bozzetto et al. / Clinical Nutrition xxx (xxxx) xxx

1 66
2
A. B. 67
0.60
3 68
4 Plasma triglyceride *† † 69
0.40 150
5 70
6 *† † *† *† 71

Triglyceride iAUC
mmol/lx 360 min
†
mmol/l

7 0.20 100 72
8 † † 73
9 † 74
10 0.00 50 75
11 76
12 -0.20 0 Categoria 1
77
13 0 60 120 180 240 300 360 EVOO Butter LF 78
14 Minutes after meal 79
15 80
16 81
17 C. D. 82
18 0 83
19 EVOO Butter LF 84
* *† 0 Categoria 1

20 85
-0.3
21 86
†
Plasma FFA

22 87
mEq/l

-100

mEq/lx360 min
23 -0.6 88

FFA iAUC
24 89
25 -200
90
-0.9
26 91
27 92
28 -1.2 -300
93
29 0 60 120 180 240 300 360 94
30 Minutes after meal 95
31 96
Fig. 4. Absolute changes in plasma concentrations of triglyceride (A) and FFA (C) after the EVOO (full squares), Butter (empty circles), and Low-fat (empty squares) meals. Tri-
32 97
glyceride (B) and FFA (D) postprandial iAUCs after the three test meals. Data in the figure are means and SEM. *p < 0.05 vs. Butter; yp < 0.05 vs. Low-fat.
33 98
34 99
35 In our cohort of patients with type 1 diabetes, EVOO induced a As another potential mechanism, we investigated the glucago- 100
36 higher postprandial increase in plasma GLP-1 concentrations than nostatic effect of GLP-1 shown with exogenous GLP-1 in patients 101
37 did the butter or low fat meals. Although GLP-1 is mainly a glucose- with type 1 diabetes [24]. In our study, in line with Kielgast et al. 102
38 dependent insulin secretagogue, its postprandial secretion is pre- [12], we observed a slight postprandial increase in glucagon levels, 103
39 served in people with type 1 diabetes [12,19], and the extent of which tended to be even higher after EVOO. Consequently, it is 104
40 postprandial GLP-1 response is influenced by meal size and fat unlikely that the higher GLP-1 concentrations due to EVOO acted 105
41 content meal [4,9,19,20]. There is also evidence that the quality of fat through glucagon suppression. Possibly, the postprandial glucagon 106
42 may influence postprandial GLP-1 secretion. In particular, mono- increase in our patients was, instead, a consequence of the post- 107
43 unsaturated fatty acids given either as fat emulsion in healthy people prandial GIP increase, which was significantly higher after both 108
44 [21] or as olive oil in healthy people [6] and in patients with type 2 high-fat meals than after the low-fat meal, as previously observed 109
45 diabetes [7] increased postprandial GLP-1 secretion compared to by Lund et al. [25] in patients with type 2 diabetes. 110
46 butter. Conversely, EVOO increased postprandial GLP-1 secretion to EVOO may have exerted its postprandial hypoglycemic effects 111
47 the same extent as butter compared with carbohydrate alone in in- also by influencing carbohydrate absorption at the intestinal level. 112
48 sulin resistant individuals [22]. Our study shows that the quality of The consistent increases in GIP, GLP-1, and glucagon in response to 113
49 fat added to a high glycemic index meal modulates the postprandial the fat-meals suggests that dietary fats stimulate the regulation of 114
50 increase in GLP-1 concentrations also in people with type 1 diabetes, postprandial hormonal response more than carbohydrates. In the 115
51 possibly contributing to the lower postprandial blood glucose intestine, the absorption of fat competes with the absorption of 116
52 response observed after EVOO consumption. Considering that GLP-1 carbohydrates, thus blunt the latter [13]. This may hold particularly 117
53 increased in the very early postprandial phase, we can speculate that true for monounsaturated fatty acids, which are preferentially 118
54 this was due to indirect stimulation, likely through a neuro/endo- absorbed compared with other types of fat [26]. Consistent with 119
55 crine pathway. While in healthy people and in patients with type 2 this, we observed a higher triglyceride response and a lower FFA 120
56 diabetes blood glucose lowering by endogenous or exogenous GLP-1 suppression after EVOO meals. A higher postprandial increase in 121
57 is mainly related to the stimulation of insulin secretion, this mech- plasma triglyceride and free fatty acids with EVOO than with butter 122
58 anism of action unlikely played a role in our patients with type 1 has been observed in studies on healthy people [27]. Likewise, in 123
59 diabetes of long duration. The concomitant increases in plasma GLP- patients with type 2 diabetes, despite its beneficial chronic effects 124
60 1 and gastric volume during the early postprandial phase in our on fasting parameters, EVOO determined a higher postprandial 125
61 study support the possibility that EVOO may have reduced post- triglyceride response mainly related to an increase in lipoproteins 126
62 prandial glycemia by slowing down gastric emptying through the of intestinal origin and, therefore, to a direct effect of dietary fatty 127
63 stimulation of GLP-1 secretion [23]. acids [28,29]. 128
64 129
65 130

Please cite this article as: Bozzetto L et al., Gastrointestinal effects of extra-virgin olive oil associated with lower postprandial glycemia in type 1
diabetes, Clinical Nutrition, https://doi.org/10.1016/j.clnu.2018.11.015
 YCLNU3682_proof ■ 8 December 2018 ■ 7/7

L. Bozzetto et al. / Clinical Nutrition xxx (xxxx) xxx 7

1 5. Conclusions [3] Cunningham KM, Read NW. The effect of incorporating fat into different 66
components of a meal on gastric emptying and postprandial blood glucose
2 67
and insulin responses. Br J Nutr 1989;61:285e90.
3 The present study explored for the first time the possible [4] Lodefalk M, Aman J, Bang P. Effects of fat supplementation on glycaemic 68
4 mechanisms underlying the reduced postprandial glycemic response and gastric emptying in adolescents with Type 1 diabetes. Diabet 69
5 response determined by the addition of EVOO to a high glycemic Med 2008;25:1030e5. 70
[5] Robertson MD, Jackson KG, Fielding BA, et al. Acute ingestion of a meal rich in
6 index meal. These results indicate that EVOO improves post- n-3 polyunsaturated fatty acids results in rapid gastric emptying in humans. 71
7 prandial glucose response in patients with type 1 diabetes through Am J Clin Nutr 2002;76:232e8. 72
8 complex interactions between the macronutrient composition of [6] Thomsen C, Rasmussen O, Lousen T, et al. Differential effects of saturated and 73
monounsaturated fatty acids on postprandial lipemia and incretin responses
9 the meal and gastro-intestinal sensing that involve gastric in healthy subjects. Am J Clin Nutr 1999;69:1135e43.
74
10 emptying, incretins secretion, and lipid metabolism. Moreover, this [7] Thomsen C, Storm H, Holst JJ, Hermansen K. Differential effects of saturated 75
11 study confirms, in well-controlled experimental conditions, the and monounsaturated fats on postprandial lipemia and glucagon-like peptide 76
1 responses in patients with type 2 diabetes. Am J Clin Nutr 2003;77:605e11.
12 differences in postprandial blood glucose profiles observed in our [8] Gentilcore D, Chaikomin R, Jones KL, et al. Effects of fat on gastric emptying of
77
13 previous home-based study. This has relevant direct implications in and the glycemic, insulin, and incretin responses to a carbohydrate meal in 78
14 the treatment of patients with type 1 diabetes, as it shows that type 2 diabetes. J Clin Endocrinol Metab 2006;91:2062e7. 79
[9] Mansour A, Hosseini S, Larijani B, Pajouhi M, Mohajeri-Tehrani MR. Nutrients
15 insulin doses and time/duration of administration may differ 80
related to GLP1 secretory responses. Nutrition 2013;29:813e20.
16 consistently after meals with a similar amount of carbohydrates [10] Itoh K, Moriguchi R, Yamada Y, et al. High saturated fatty acid intake induces 81
17 and total fat. This information may be especially useful for patients insulin secretion by elevating gastric inhibitory polypeptide levels in healthy 82
individuals. Nutr Res 2014;34:653e60.
18 on insulin pump and can be of help in defining algorithms for the 83
[11] Lardinois CK, Starich GH, Mazzaferri EL. The postprandial response of gastric
19 determination of prandial insulin administration. inhibitory polypeptide to various dietary fats in man. J Am Coll Nutr 1988;7: 84
20 241e7. 85
21 [12] Kielgast U, Holst JJ, Madsbad S. Antidiabetic actions of endogenous and 86
Funding exogenous GLP-1 in type 1 diabetic patients with and without residual beta-
22 cell function. Diabetes 2011;60:1599e607. 87
23 Q1 The work presented here was supported by Ministero Istruzione [13] Emerson SR, Haub MD, Teeman CS, Kurti SP, Rosenkranz SK. Summation of 88
24  e Ricerca PRIN 2010JCWWKM, 2010. blood glucose and TAG to characterise the “metabolic load index”. Br J Nutr 89
Universita 2016;116:1553e63.
25 [14] Bozzetto L, Giorgini M, Alderisio A, et al. Glycaemic load versus carbohydrate
90
26 Declaration of interest counting for insulin bolus calculation in patients with type 1 diabetes on 91
27 insulin pump. Acta Diabetol 2015;52:865e71. 92
[15] Bolondi L, Bortolotti M, Santi V, et al. Measurement of gastric emptying time
28 None. by real-time ultrasonography. Gastroenterology 1985;89:752e9.
93
29 [16] Marzio L, Giacobbe A, Conoscitore P, et al. Evaluation of the use of ultraso- 94
30 nography in the study of liquid gastric emptying. Am J Gastroenterol 1989;84: 95
Author contributions 496e500.
31 96
[17] Di Marino L, Griffo E, Maione S, Mirabella M. Active glucagon-like peptide-1
32 (GLP-1): storage of human plasma and stability over time. Clin Chim Acta 97
L.B. designed the experiment, acquired, analyzed and inter-
33 2011;412:1693e4. 98
preted data, and wrote the manuscript. A.A., M.G., and F.B. acquired [18] Phillips LK, Deane AM, Jones KL, Rayner CK, Horowitz M. Gastric emptying and
34 99
and interpreted data and edited the manuscript. G.C., E.G., C.V., and glycaemia in health and diabetes mellitus. Nat Rev Endocrinol 2015;11:
35 112e28. 100
P.C. acquired data and edited the manuscript. A.G. interpreted data
36 [19] Vilsboll T, Krarup T, Sonne J, et al. Incretin secretion in relation to meal size 101
and contributed to discussion. G.R. contributed to the conception of and body weight in healthy subjects and people with type 1 and type 2
37 102
the study and reviewed and edited the manuscript. A.A.R. designed diabetes mellitus. J Clin Endocrinol Metab 2003;88:2706e13.
38 [20] Bozzetto L, Annuzzi G, Pacini G, et al. Polyphenol-rich diets improve glucose 103
the experiment, contributed to the discussion, and reviewed and
39 metabolism in people at high cardiometabolic risk: a controlled randomised 104
edited the manuscript. G.A. designed the experiment and wrote the intervention trial. Diabetologia 2015;58:1551e60.
40 105
manuscript. G.A. is the guarantor of this work and, as such, had full [21] Beysen C, Karpe F, Fielding BA, et al. Interaction between specific fatty acids,
41 106
access to all the data in the study and takes responsibility for the GLP-1 and insulin secretion in humans. Diabetologia 2002;45:1533e41.
42 [22] Paniagua JA, De La Sacristana AG, Sanchez E, et al. A MUFA-rich diet improves 107
integrity of the data and the accuracy of the data analysis. All au-
43 posprandial glucose, lipid and GLP-1 responses in insulin-resistant subjects. 108
thors have approved the final article. J Am Coll Nutr 2007;26:434e44.
44 [23] Dupre J. Glycaemic effects of incretins in Type 1 diabetes mellitus: a concise
109
45 review, with emphasis on studies in humans. Regul Pept 2005;128:149e57. 110
Acknowledgments
46 [24] Creutzfeldt WO, Kleine N, Willms B, et al. Glucagonostatic actions and 111
reduction of fasting hyperglycemia by exogenous glucagon-like peptide I(7-
47 112
The authors thank Dr Stefano Signorini (Laboratory Medicine, 36) amide in type I diabetic patients. Diabetes Care 1996;19:580e6.
48 [25] Lund A, Vilsboll T, Bagger JI, Holst JJ, Knop FK. The separate and combined 113
 Milano Bicocca, Milan, Italy) for the
Desio Hospital, Universita
49 impact of the intestinal hormones, GIP, GLP-1, and GLP-2, on glucagon 114
glucagon assay. secretion in type 2 diabetes. Am J Physiol Endocrinol Metab 2011;300:
50 115
E1038e46.
51 [26] Jackson KG, Robertson MD, Fielding BA, Frayn KN, Williams CM. Olive oil in- 116
52 Appendix A. Supplementary data creases the number of triacylglycerol-rich chylomicron particles compared 117
53 with other oils: an effect retained when a second standard meal is fed. Am J 118
Supplementary data to this article can be found online at Clin Nutr 2002;76:942e9.
54 [27] Mekki N, Charbonnier M, Borel P, et al. Butter differs from olive oil and 119
55 https://doi.org/10.1016/j.clnu.2018.11.015. sunflower oil in its effects on postprandial lipemia and triacylglycerol-rich 120
56 lipoproteins after single mixed meals in healthy young men. J Nutr 121
References 2002;132:3642e9.
57 [28] De Natale C, Annuzzi G, Bozzetto L, et al. Effects of a plant-based high-car-
122
58 bohydrate/high-fiber diet versus high-monounsaturated fat/low-carbohy- 123
[1] Bozzetto L, Alderisio A, Giorgini M, et al. Extra-virgin olive oil reduces gly-
59 drate diet on postprandial lipids in type 2 diabetic patients. Diabetes Care 124
cemic response to a high-glycemic index meal in patients with type 1 dia- 2009;32:2168e73.
60 Q2 betes: a randomized controlled ctTrial. Diabetes Care 2016;39:518e24.
[29] Bozzetto L, Annuzzi G, Costabile G, et al. A CHO/fibre diet reduces and a MUFA
125
61 [2] Marathe CS, Rayner CK, Jones KL, Horowitz M. Relationships between gastric diet increases postprandial lipaemia in type 2 diabetes: no supplementary 126
emptying, postprandial glycemia, and incretin hormones. Diabetes Care
62 2013;36:1396e405.
effects of low-volume physical training. Acta Diabetol 2014;51:385e93. 127
63 128
64 129
65 130

Please cite this article as: Bozzetto L et al., Gastrointestinal effects of extra-virgin olive oil associated with lower postprandial glycemia in type 1
diabetes, Clinical Nutrition, https://doi.org/10.1016/j.clnu.2018.11.015