Effect of a Mediterranean diet on endothelial progenitor cells and carotid intima-media thickness in type 2 diabetes: Follow-up of a randomized trial

Abstract

Background

We assessed the long-term effects of a Mediterranean diet on circulating levels of endothelial progenitor cells (EPCs) and the carotid intima-media thickness (CIMT) in patients with type 2 diabetes.

Design

This was a parallel, two-arm, single-centre trial.

Methods

Two hundred and fifteen men and women with newly diagnosed type 2 diabetes were randomized to a Mediterranean diet (n = 108) or a low-fat diet (n = 107). The primary outcome measures were changes in the EPC count and the CIMT of the common carotid artery after the treatment period defined as the end of trial (EOT).

Results

At the EOT, both the CD34⁺KDR⁺ and CD34⁺KDR⁺CD133⁺ counts had increased with the Mediterranean diet compared with the low-fat diet (p < 0.05 for both). At the EOT evaluation, there was a significant (p = 0.024) difference of −0.025 mm in the CIMT favouring the Mediterranean diet. Compared with the low-fat diet, the rate of regression in the CIMT was higher in the Mediterranean diet group (51 vs. 26%), whereas the rate of progression was lower (25 vs. 50%) (p = 0.032 for both). Changes in the CIMT were inversely correlated with the changes in EPC levels (CD34⁺KDR⁺, r = −0.24, p = 0.020; CD34⁺KDR⁺CD133⁺, r = −0.28, p = 0.014). At the EOT, changes in levels of HbA1c, HOMA, total cholesterol, high-density lipoprotein cholesterol and systolic blood pressure were significantly greater with the Mediterranean diet than with the low-fat diet.

Conclusion

Compared with a low-fat diet, a long-term trial with Mediterranean diet was associated with an increase in circulating EPCs levels and prevention of the progression of subclinical atherosclerosis in patients with newly diagnosed type 2 diabetes.

Mediterranean diet, type 2 diabetes, endothelial progenitor cells, carotid intima-media thickness

Introduction

Current nutrition recommendations for patients with type 2 diabetes mellitus emphasize a greater intake of fruit and vegetables and a reduced intake of saturated fats and low-fat dairy products.¹ Strong evidence from prospective cohort studies and randomized trials shows reduced coronary heart disease with the consumption of Mediterranean diet.² On the basis of the current evidence, the traditional Mediterranean-type diet, including plant foods with emphasis on plant protein sources, provides a well-tested healthy dietary pattern to reduce cardiovascular disease, also in diabetic patients.³ It has been suggested that a Mediterranean diet down-regulates cellular and humoral inflammatory pathways related to atherosclerosis.^4,5

Vascular homeostasis results from the balance between endothelial damage and repair. Circulating endothelial progenitor cells (EPCs) are stem cells derived from bone marrow that are able to differentiate in mature endothelial cells.⁶ Once in the bloodstream, EPCs reach the target tissues where they promote endothelial repair by forming a patch at sites of endothelial injury. Diabetes mellitus is characterized by a defective mechanism of vascular repair caused by an impaired regenerative capacity of EPCs.⁶ A reduction in EPCs may represent one of the mechanisms linking increased vascular risk with diabetes mellitus.⁷

The effect of a Mediterranean diet on circulating EPCs has never been investigated in patients with type 2 diabetes free of cardiovascular disease (CVD). Therefore, using the data from a randomized trial (MÈDITA; MEditerranean DIet and Type 2 diAbetes),⁸ we investigated (a) whether a Mediterranean diet affected circulating levels of EPCs in patients newly diagnosed with type 2 diabetes, (b) whether changes in circulating levels of EPCs were associated with changes in the carotid intima-media thickness (CIMT), considered to be a surrogate marker for CVD in patients with type 2 diabetes⁹ and (c) whether the effect was sustained during long-term follow-up.

Methods

Study design and participants

The MÈDITA trial is a single-centre, randomized controlled trial that was designed to investigate whether a Mediterranean diet with a low carbohydrate content (<50% of total energy) reduced the need for drug treatment in patients newly diagnosed with type 2 diabetes. The two-arm trial design of the study has been described previously.⁸ The trial was conducted between January 2004 (the first patient enrolled) and September 2008 (the end of the 4-year follow-up of the last patient). We decided to continue to monitor the participants who did not reach the primary endpoint so that the trial ended in early autumn 2012 after a total follow-up of 8.1 years (i.e. at the time they reached the primary endpoint of HbA1c ≥ 7%).¹⁰ The follow-up visits included the same procedures as during the core intervention and were similar for all participants. The dates of our present analysis were 6 September 2013, to 31 December 2015. Middle-aged white patients with type 2 diabetes defined according to the American Diabetes Association criteria, who had never been treated with diabetes drugs, were eligible for the study. The trial was conducted in accordance with the Declaration of Helsinki and with the institutional review board’s approval.

After obtaining written informed consent, 215 men and women with newly diagnosed type 2 diabetes were randomized using a computer-generated random number sequence (simple randomization) at the baseline visit to one of two treatment modalities: a Mediterranean diet group or a low-fat diet group. The nurses who scheduled the study visits did not have access to the randomization list and the laboratory staff did not know the participants’ group assignments. The staff members involved in the intervention were aware of the group assignments, but those who assessed achievement of the primary outcome were blinded to the intervention.

Nutritional interventions and dietary assessment

The main goals of the dietary interventions were the restriction of energy intake to 1500 kcal/day for women and 1800 kcal/day for men in both groups. The Mediterranean diet was rich in vegetables and whole grains and low in red meat, which was replaced with poultry and fish, with the goal of no more than 50% of calories from carbohydrates and no less than 30% calories from fat, with the main source of added fat 30–50 g of olive oil. The low-fat diet was rich in whole grains and restricted additional fats, sweets and high-fat snacks, with the goal of no more than 30% of calories from fat and no more than 10% of calories from saturated fat. Participants were instructed on how to record their intake using food models and actual weights or amounts in terms of common measures. Adherence to the diets was assessed by session attendance and review of the diet diaries. Participants in both groups were also advised to increase their level of physical activity, with programmes tailored on the basis of the results of a baseline physical fitness test and safety concerns. For the ascertainment of physical activity status, we used the International Physical Activity Questionnaire (www.institutferran.org/documentos/scoring_short_ipaq_april04.pdf).

At the scheduled visits, staff members who were unaware of the study group assignments queried patients about all medical events and hospitalizations and measured their weight and blood pressure, along with assessing medication use and obtaining blood for analysis at the hospital laboratory. All study participants completed a three-day food record with a picture booklet of portion sizes of typical foods. The average intakes of total energy, total, saturated, monounsaturated and polyunsaturated fat (proportion of the total daily energy intake), carbohydrates and protein were calculated. The Mediterranean diet score was adapted from the Mediterranean diet scale of Trichopoulou et al.,¹¹ as described previously.⁸ The median intakes of food groups associated with a Mediterranean diet were calculated and the participants received a point on the scale if they measured above the median consumption for fish, fruit, legumes, nuts, ratio of monounsaturated to saturated fat, vegetables and whole grains. Red and processed meat consumption below the median received 1 point, otherwise 0 points. For alcohol, a value of 1 was assigned to men who consumed 10–50 g/day and to women who consumed 5–25 g/day. For fat intake, we used the ratio of monounsaturated lipids to saturated lipids, rather than the ratio of polyunsaturated to saturated lipids, because in Campanian County monounsaturated lipids are used in much higher quantities than polyunsaturated lipids. Thus the Mediterranean diet scores ranged from 0 to 9; higher scores indicated closer adherence to a Mediterranean diet.

Quantification of circulating EPCs

Peripheral blood cells were analysed for the expression of surface antigens by direct flow cytometry as described previously.¹² Quantitative analysis was performed on a BD FACSCalibur cytometer (Becton-Dickinson) and 1,000,000 cells were acquired in each sample. A morphological gate was used to exclude granulocytes. We gated CD34⁺ or CD133⁺ peripheral blood cells in the mononuclear cell fraction and examined the resulting population for the dual expression of KDR. Triple-positive cells were identified by the dual expression of KDR and CD133 in the CD34⁺ gate. Data were processed with the use of the Macintosh CELLQuest software program (Becton-Dickinson). Measures were repeated twice in two separate blood samples. The same trained operator, who was blinded to the characteristics of the patients, performed all the tests throughout the study. The results from flow cytometry were expressed as the number of cells per 10⁶ events.

Measurement of CIMT

Ultrasonographic scans of the carotid artery were performed by expert sonographers who were specifically trained and blinded to the treatment group, as reported previously.¹³ Each participant was examined with a single ultrasound machine (high-resolution B-mode ultrasound scanner) equipped with a high-frequency (>7.5 MHz) linear transducer with a limit of detection < 0.1 mm. All images were stored electronically. The measurement was made in the 1 cm segment proximal to the dilation of the carotid bulb and always in plaque-free segments. For each patient, three measurements on both sides were performed on the anterior, lateral and posterior projections of the far wall; the readings were then averaged. An independent observer who was blinded to the treatment group and trained in the interpretation of CIMT images performed offline analysis. Paired CIMT measurements in the same arteries showed a high degree of reproducibility with a mean difference in CIMT of 0.020 mm and an intra-class correlation coefficient of 0.97 (p < 0.001). CIMT regression was defined as a decrease of 0.020 mm in the mean CIMT.

Laboratory determinations

The parameters measured included HbA1c (high-pressure liquid chromatography traceable to the Diabetes Control and Complications Trial reference method), anthropometrics (weight), waist circumference, lipids [total and high-density lipoprotein (HDL) cholesterol], high-sensitive C reactive protein (CRP) and insulin. An end of trial (EOT) sample was also collected when each participant left the study. All measurements were made in the hospital’s chemistry laboratory.

Outcomes

The primary outcome measures were changes in circulating levels of EPCs and the CIMT of the common carotid artery after the treatment period (EOT). Investigations were carried out at the start of the study, after 2 and 4 years, and at the EOT. We followed patients for 8.1 years to assess trial outcomes. Participants were censored to the date of the last follow-up, i.e. at the time they had to leave the study with an HbA1c level > 7% for at least 3 months.

Statistical analysis

We analysed the data by intention-to-treat. To evaluate the repeated blood measurements over time, we estimated generalized estimating equations to account for the non-independence of the same variable in the same patient over time. Changes from baseline to treatment visits were assessed with the one-sample t test and Wilcoxon signed rank test within the group. Baseline and follow-up group comparisons were assessed with Student’s t test or the Wilcoxon rank sum test for continuous variables and Fisher’s exact test for categorical variables. Fisher’s exact test was also used for comparing the proportions of patients in the two groups that demonstrated progression or regression of the mean CIMT. The correlation between changes in levels of EPCs and corresponding changes in the CIMT was assessed using Spearman’s coefficients of correlation. Participants were divided into groups of low (scores of 0–3), middle (4–5) and high (6–9) adherence to the Mediterranean diet based on their diet scores for analysis. Comparisons were conducted using linear regression to test for trends across the Mediterranean dietary pattern score groups for EPCs and CIMT. We adjusted for potential confounders in multivariable analyses and these included age, sex, HbA1c, total energy intake, weight, physical activity, smoking, blood pressure and lipid concentrations. All statistical tests were two-sided with a 5% significance level. We conducted all analyses using SPSS, version 10.05 (SPSS, Chicago, IL, USA).

Results

Between 1 January and 30 December 2004, we assessed 260 candidates for eligibility. Of these, 25 declined to participate and 20 did not meet the inclusion criteria. Thus 215 participants were randomly allocated to either a Mediterranean diet (n = 108; 50% women) or a low-fat diet group (n = 107; 51.5% women). Equal numbers of patients withdrew from the groups during the trial (Figure 1). Participants in the Mediterranean diet group were followed for 8.1 years as they all reached the original primary endpoint (need of diabetes drug) 2 years later than those of the low-fat group (6.1 years). The mean follow-up time times were 5.2 years in the Mediterranean diet group and 3.5 years in the low-fat diet group. At the EOT (30 September 2012), 102 participants had completed the trial in the Mediterranean diet group and 99 in the low-fat diet group.

Figure 1.

Trial profile.

The two treatment groups were well matched for demographic and clinical characteristics (Table 1), including the use of concomitant drugs for hypertension and dyslipidaemia. The mean age of participants was 52 years and all were free from apparent CVD. At baseline, the EPC cell counts did not differ between the two groups (Table 1). There was a significant increase in some EPC phenotypes at the EOT in the Mediterranean diet group compared with the low-fat diet group. Both the CD34⁺KDR⁺ and CD34⁺KDR⁺CD133⁺ counts showed a significant increase in the Mediterranean diet group compared with their respective basal concentrations (p = 0.010 for both) or with the low-fat diet group (p = 0.011 for CD34⁺KDR⁺ and p = 0.034 for CD34⁺KDR⁺CD133⁺). These differences in CD34⁺KDR⁺ and CD34⁺KDR⁺CD133⁺ counts between the two diet groups were significant starting from the second year of follow-up and persisted in the subsequent analyses. No other significant change was recorded for the other EPC phenotypes at different times of evaluation (Table 2). There was no sex-based difference.

Table 1.

Characteristics of the study participants.

Characteristic	Mediterranean diet (n = 108)	Low-fat diet (n = 107)
Sex (M/F)	54/54	52/55
Age (years)	52.4 ± 11.2	51.9 ± 10.7
Body weight (kg)	86.0 ± 10.4	85.7 ± 9.9
Waist circumference (cm)	98 ± 10	98 ± 10
Body mass index (kg/m2)	29.7 ± 3.4	29.5 ± 3.6
HbA1c (%)	7.75 ± 0.9	7.71 ± 0.9
Plasma glucose (mg/dL)	162 ± 34	159 ± 33
Serum insulin (pmol/L)	108 ± 43	115 ± 50
HOMA of insulin sensitivity	5.2 ± 1.7	5.3 ± 1.8
CRP (mg/dL)	2.65 (1.81–4.92)	2.79 (1.71–5.21)
Total cholesterol (mg/dL)	221 ± 35	216 ± 33
HDL cholesterol (mg/dL)	43 ± 10	43 ± 10
Non-HDL cholesterol (mg/dL)	178 ± 71	168 ± 69
Blood pressure (mm Hg)
Systolic	139 ± 12	140 ± 12
Diastolic	87 ± 8	86 ± 8
CIMT (mm)	0.84 ± 0.17	0.84 ± 0.16
Smoking	21 (21.3)	22 (21.5)
Drug treatment
Antihypertensive drugs	26 (24)	25 (23.3)
Lipid-lowering drug	16 (15)	15 (16)
Physical activity (min/week)	45 ± 12	43 ± 13
EPC phenotypes (n/106)
CD34+	357 (291–421)	373 (301–442)
KDR+	79 (52–94)	73 (47–89)
CD133+	242 (193–300)	253 (199–308)
CD34+CD133+	197 (154–235)	188 (143–222)
CD34+KDR+	44 (39–50)	45 (40–52)
CD133+KDR+	27 (18–35)	32 (24–41)
CD34+KDR+CD133+	17 (13–22)	19 (13–23)

Table 2.

Changes in EPC phenotypes from baseline.

Variable	Mediterranean diet	Low-fat diet	Difference (95% confidence interval)	p
CD34+ (n/106)
Year 2	20 (19)	9 (9)	11 (−6 to 28)	0.112
Year 4	31 (23)	12 (13)	19 (−4 to 42)	0.223
EOT	26 (21)	10 (11)	16 (−6 to 38)	0.253
KDR+ (n/106)
Year 2	10 (8)	4 (5)	6 (−3 to 15)	0.196
Year 4	6 (7)	10 (9)	−4 (−13 to 5)	0.267
EOT	14 (13)	6 (9)	8 (−8 to 24)	0.348
CD133+ (n/106)
Year 2	12 (11)	14 (15)	−2 (−18 to 14)	0.570
Year 4	9 (10)	5 (5)	4 (−4 to 12)	0.285
EOT	15 (14)	9 (10)	6 (−5 to 17)	0.325
CD34+CD133+ (n/106)
Year 2	3 (4)	−2 (3)	5 (−3 to 13)	0.366
Year 4	4 (3)	0 (2)	4 (−3 to 11)	0.417
EOT	16 (13)	2 (4)	14 (−2 to 30)	0.129
CD34+KDR+ (n/106)
Year 2	25 (17)	3 (4)	22 (5 to 39)	0.028
Year 4	21 (19)	7 (6)	14 (1 to 27)	0.050
EOT	22 (17)	1 (2)	21 (6 to 36)	0.011
CD133+KDR+ (n/106)
Year 2	4 (4)	2 (3)	2 (−2 to 6)	0.185
Year 4	2 (2)	1 (3)	1 (−2 to 4)	0.342
EOT	3 (3)	1 (2)	2 (−2 to 6)	0.158
CD34+KDR+CD133+ (n/106)
Year 2	10 (4)	0 (2)	10 (3 to 17)	0.041
Year 4	8 (3)	−1 (2)	9 (2 to 16)	0.045
EOT	9 (3)	1 (2)	8 (2 to 14)	0.034

The actual number of survivors were: year 2, n = 85 Med diet/64 low-fat diet; year 4, n = 50/29; EOT, n = 102 Mediterranean diet/99 low-fat diet.

EOT: end of trial.

Over the entire follow-up period, the Mediterranean diet, but not the low-fat diet, significantly reduced the mean CIMT relative to the baseline (Table 3). At the EOT evaluation, there was a significant difference of −0.025 mm (p = 0.024) favouring the Mediterranean diet. Compared with the low-fat diet group, the rate of regression in CIMT was higher in the Mediterranean diet group (51 vs 26%), whereas the rate of progression was lower (25 vs 50%) (p = 0.032 for both). At the EOT, there was an inverse association between the changes in CIMT and the increase in circulating levels of both CD34⁺KDR⁺ (r = −0.24; p = 0.020) and CD34⁺KDR⁺CD133⁺ (r = −0.28; p = 0.014) counts. These associations were reduced, but still significant (r = −0.19; p = 0.039 and r = −0.20; p = 0.042, respectively) when adjusted for potential confounders in multivariable analyses, including age, sex, HbA1c, plasma glucose, HOMA, weight, blood pressure and lipid concentrations.

Table 3.

Changes in assessed variables from baseline.

Variable	Mediterranean diet	Low-fat diet	Difference (95% CI)	p
Weight (kg)
Year 2	−4.9 (2.5)	−3.7 (2.1)	−1.2 (−2.1 to −0.3)	0.001
Year 4	−3.8 (2.5)	−3.2 (2.8)	−0.6 (−1.6 to 0.4)	0.112
EOT	−3.7 (2.1)	−2.9 (1.9)	−0.8 (−1.4 to 0.1)	0.045
Waist circumference (cm)
Year 2	−4.4 (2.8)	−3.3 (2.5)	−1.1 (−1.8 to −0.3)	0.011
Year 4	−3.0 (1.7)	−2.6 (2.0)	−0.4 (−0.9 to 0.1)	0.061
EOT	−3.2 (1.8)	−2.6 (1.8)	−0.6 (−1.1 to −0.1)	0.042
HbA1c (%)
Year 2	−1.1 (0.9)	−0.5 (0.4)	−0.6 (−1.0 to −0.2)	0.001
Year 4	−0.9 (0.6)	−0.5 (0.4)	−0.4 (−0.9 to −0.1)	0.011
EOT	−0.7 (0.5)	−0.4 (0.3)	−0.3 (−0.6 to 0)	0.050
Plasma glucose (mg/dL)
Year 2	−38 (20)	−20 (19)	−18 (−31 to −5)	0.001
Year 4	−30 (19)	−14 (14)	−16 (−27 to −4)	0.042
EOT	−10 (10)	−4 (5)	−6 (−12 to 0)	0.050
HOMA of insulin sensitivity
Year 2	−2.1 (0.9)	−1.1 (0.7)	−1 (−1.8 to −0.3)	0.014
Year 4	−1.5 (1.0)	−0.9 (0.6)	−0.6 (−1.1 to −0.1)	0.050
EOT	−1.5 (1.1)	−0.8 (0.7)	−0.7 (−1.2 to −0.2)	0.043
Total cholesterol (mg/dL)
Year 2	−16 (13)	−8 (8)	−8 (−14 to −2)	0.015
Year 4	−10 (9)	−4 (4)	−6 (−15 to 3)	0.158
EOT	−10 (9)	−5 (5)	−5 (−10 to 0)	0.050
HDL cholesterol (mg/dL)
Year 2	5 (5)	0 (1)	5 (1.9 to 6.9)	0.013
Year 4	3 (4)	1 (1)	2.8 (0.1 to 5.5)	0.042
EOT	3 (3)	0 (1)	3 (0.5 to 6.5)	0.017
Non-HDL cholesterol (mg/dL)
Year 2	−11 (6)	−8 (6)	−3 (−14 to −1.9)	0.015
Year 4	−7 (4)	−3 (3)	−4 (−13 to −2.7)	0.041
EOT	−7 (4)	−5 (4)	−2 (−10 to 0.2)	0.053
Systolic blood pressure (mmHg)
Year 2	−4.5 (3.7)	−1.4 (1.7)	−3.1 (−5.4 to −0.6)	0.014
Year 4	−2.5 (2.8)	−1.0 (1.1)	−1.5 (−4.5 to 1.2)	0.127
EOT	−2.4 (2.6)	−0.6 (0.7)	−2.1 (−4.0 to −0.2)	0.043
Diastolic blood pressure (mmHg)
Year 2	−3.2 (2.8)	−2.5 (4.0)	−0.7 (−2.4 to 0.7)	0.228
Year4	−2.9 (2.6)	−1.5 (1.4)	−1.4 (−4.0 to 1.2)	0.359
EOT	−2.5 (1.7)	−1.0 (1.1)	−1.5 (−4.0 to 2.0)	0.174
CIMT (mm)*
Year 2	−0.032 (0.07)	−0.010 (0.06)	−0.022 (−0.045 to −0.0)	0.050
Year 4	−0.022 (0.08)	−0.003 (0.07)	−0.019 (−0.035 to −0.003)	0.031
EOT	−0.026 (0.07)	−0.001 (0.06)	−0.025 (−0.040 to −0.005)	0.024
CRP (mg/dL)
Year 2	−0.93 (0.01)	−0.38 (002)	−0.55 (−0.32 to −0.94)	0.029
Year 4	−0.76 (0.02)	−0.25 (0.03)	−0.51 (−0.24 to −0.88)	0.038
EOT	−0.41 (0.01)	−0.12 (0.01)	−0.29 (−0.26 to 1.25)	0.079

The actual number of survivors were: year 2, n = 85 Med diet/64 low−fat diet; year 4, n = 50/29; EOT, n = 102/99. * Cumulative CIMT changes are reported.

CIMT: carotid intima-media thickness; CRP: C reactive protein; EOT: end of trial; EPC: endothelial progenitor cells; HDL: high–density lipoprotein.

At the EOT, changes in weight (−0.8 kg; p = 0.045), waist circumference (−0.6 cm; p = 0.042), HbA1c (−0.3%; p < 0.0050), plasma glucose (−6 mg/dL; p = 0.050), HOMA (−0.7; p = 0.043), total cholesterol (−5 mg/dL; p = 0.050), HDL cholesterol (3 mg/dL; p = 0.017) and systolic blood pressure (−1.5 mmHg; p = 0.043) were significantly greater in the Mediterranean diet group than in the low-fat group (Table 3). There was no sex-based difference.

Figure 2 shows the circulating EPCs at EOT grouped by level of adherence to the Mediterranean diet scores. Diabetic patients with the highest scores (6–9) had the higher circulating level of CD34⁺KDR⁺ phenotype than the diabetic patients who scored < 3 points on the scale (P for trend = 0.010). The association persisted significant when adjusted for potential confounders, including age, sex, weight, energy intake, physical activity level and smoking status. A similar association between adherence to Med diet and the level of CD34⁺KDR⁺CD133⁺ phenotype was also recorded (P for trend = 0.011). Similarly, there was an inverse association between adherence to Mediterranean diet and CMIT at EOT, which was 0.79 (SD 0.16) mm in the high adherence group, 0.83 (0.17) mm in the middle adherence group, and 0.86 (0.18) mm in the low adherence group (P for trend = 0.010).

Figure 2.

Association between adherence to a Mediterranean diet and two phenotypes of circulating EPC levels. Adherence to a Mediterranean diet is quantified as low (L, n = 21), moderate (M, n = 40) or high (H, n = 41) with the relative scores.

The composition of the diets consumed by participants in the Mediterranean diet and low-fat diet groups did not differ at baseline (Supplementary Table 1, available online). The daily energy intake decreased in both groups during the study without statistically significant between-group differences at the EOT. The percentage of carbohydrate intake decreased in the Mediterranean diet group compared with the low-fat diet group and the percentage of monounsaturated and polyunsaturated fatty acid intake increased. At the EOT, the increase in circulating levels of CD34⁺KDR⁺CD133⁺ cells was positively associated with changes in monounsaturated fat (r = 0.28; p = 0.010) and negatively associated with changes in carbohydrates (r = −0.19; p = 0.021). There was no other significant association between any other EPC phenotype and nutrient intake.

Discussion and conclusions

We assessed circulating EPCs and CIMT at baseline and at the EOT, including some points (two and four years) of follow-up. Compared with the participants randomly allocated to a low-fat diet, the participants randomized to Mediterranean diet showed significant increases in the circulating numbers of some EPC phenotypes (namely CD34⁺KDR⁺ and CD34⁺KDR⁺CD133⁺) associated with a significant regression of CIMT at the end of treatment. This association was in part mediated by improvements of some CVD risk factors during treatment with a Mediterranean diet. These findings are novel because no prior long-term trial has assessed changes in EPCs and CIMT in response to dietary interventions in type 2 diabetes and they suggest a role for a Mediterranean diet in the prevention of progression of atherosclerosis in patients with newly diagnosed type 2 diabetes and free of apparent CVD.

EPCs may contribute to the maintenance of the endothelium by replacing injured mature endothelial cells. The number of circulating EPCs is reduced in patients with CVD,¹⁴ diabetes mellitus^6,7 and multiple coronary risk factors,¹⁵ which seems to be consistent with the hypothesis that atherosclerosis is also linked to consumptive loss of the endothelial repair capacity.

Both circulating CD34⁺KDR⁺ and CD34⁺KDR⁺CD133⁺ cell counts were significantly increased in the Mediterranean diet group, whereas they remained stable in the low-fat diet group. The analysis at the EOT showed that the patients with diabetes with the highest scores (6–9) of adherence to the Mediterranean diet had the highest levels of circulating CD34⁺KDR⁺ and CD34⁺KDR⁺CD133⁺ cells, whereas no other association could be found with the other five EPCs phenotypes. The effect of a Mediterranean diet seems to be limited to these particular EPC phenotypes that are best linked to CVD. In particular, the CD34⁺KDR⁺ level independently predicts cardiovascular events and atherosclerosis progression in patients with coronary heart disease¹⁴ and negatively correlates with measures of CIMT.¹⁶ CIMT is a commonly used and direct assessment of early atherosclerosis and is generally considered to be a reliable surrogate endpoint of vascular outcomes.¹⁷ The average annual increase of CIMT in untreated patients is 0.015 mm/year and thus requires large sample sizes or a long duration of follow-up to show effects of treatment. In intervention trials^18,19 with follow-ups ranging from 1 to 2.4 years, there was no significant effect of a Mediterranean diet enriched with olive oil on the CIMT.

To date, the most compelling evidence for cardiovascular protection by a Mediterranean diet has come from the results of the PREDIMED study,² which reported a 30% reduced risk of CVD events with two Mediterranean diets (supplemented with extra virgin olive oil or nuts) in people at high cardiovascular risk, although the total mortality remained unaffected. The pre-specified diabetes mellitus subgroup (about 50% of the entire population) demonstrated similar results. Against this background, the results of the largest controlled trial in overweight or obese adults with type 2 diabetes showed that an intensive lifestyle intervention focusing on weight loss did not reduce the rate of cardiovascular events,²⁰ although it may have multiple metabolic benefits.

The effect of the Mediterranean diet on cardiovascular health has been explained by a beneficial effect on classical risk factors. Several meta-analyses of randomized controlled trials showed a favourable effect of a Mediterranean diet, compared with other diets, on body weight, total cholesterol and HDL cholesterol.^21,22 A secondary outcome analysis of the PREDIMED trial²³ reported benefits of a Mediterranean diet enriched with olive oil compared with a low-fat diet on body weight (0.43 kg) and waist circumference (0.55 cm), which were similar to our results at the EOT (0.8 kg and 0.6 cm, respectively). Alternative mechanisms for the cardiovascular benefits of a Mediterranean diet include an inhibitory effect on circulating inflammatory markers.^3,4 In this perspective, the beneficial effect of a Mediterranean diet on subclinical atherosclerosis may be related, at least in part, to the down-regulation of the expression of several proatherogenic genes involved in vascular inflammation,²⁴ coupled with reduced oxidative stress.²⁵

No prior dietary trial has assessed the long-term effect of the Mediterranean diet on EPCs. Marin et al.²⁶ studied 20 healthy elderly patients who were randomly allocated to three different diets, each for four weeks in a cross-over design. Compared with participants assigned to a low-fat or a high-carbohydrate diet, the participants consuming a Mediterranean diet enriched with monounsaturated fatty acids presented significant increases in circulating CD34⁺KDR⁺CD133⁺ cells. In our study, we found both a quantitative (EPC increased in the Mediterranean diet group) and qualitative (the greater the adherence to Mediterranean diet, the greater the increase in EPCs) association between a Mediterranean diet and the number of both CD34⁺KDR⁺ and CD34⁺KDR⁺CD133⁺ cells. The Mediterranean diet we used in the trial⁸ contained a relatively low carbohydrate energy content (43.7% of the total energy intake at the EOT) and a relatively high monounsaturated fatty acid content (18.3%), which may explain, at least in part, the positive association between increases in EPC levels and monounsaturated fat intake and the negative association with decreased carbohydrate intake.

We acknowledge some limitations of this study. First, the assessment of circulating EPCs and CIMT was not the primary endpoint of the MÈDITA trial, which, in theory, could have resulted in group imbalance; however, the groups were well comparable for both EPC count and CIMT. Second, our sample consisted of patients with type 2 diabetes at diagnosis and is therefore not representative of the overall diabetic population. This potential limitation is counterbalanced by having evaluated a more homogeneous diabetic group, i.e. those patients who for the first time address dietetic treatment for their hyperglycaemia and away from confounding linked to drug treatment for diabetes. Third, the evaluation of diet was based on self-reports, whereas more objective measurements may have resulted in more pronounced associations with EPC counts. Food diaries are considered to be a reliable method for dietary studies, but misreporting may occur. However, this trial succeeded in achieving a difference in macronutrient composition, namely carbohydrate and total fat, large enough to evaluate the association between single macronutrients and EPC counts. Fourth, given its nature of censored study, the number of participants available for evaluation of EPC levels decrease as a function of time. On the other hand, the analysis at the EOT included all participants who had their assessment of EPCs and CIMT at the time they were leaving the study. Progression or regression of carotid atherosclerosis remains a controversial surrogate for cardiovascular effects in type 2 diabetes.⁹ The strengths of this study were its randomized design, the long-term follow-up and the changes achieved in the participants’ dietary habits according to the interventions. This last point seems particularly important as larger interventional trials often fail to maintain a difference in macronutrient intake large enough to accomplish the aim of the study.

In conclusion, this study is the first dietary intervention trial demonstrating a sustained and long-term effect of a Mediterranean diet in increasing circulating EPC levels and preventing the progression of subclinical atherosclerosis in patients with newly diagnosed type 2 diabetes and free from apparent CVD.

Author contribution

DG and KE contributed to the conception and design of the work. MIM, GB, MP, MG, and MC contributed to the acquisition, analysis, or interpretation of data for the work. MIM, DG, and KE drafted the manuscript. GB, MP, MG and MC critically revised the manuscript. All authors gave final approval and agreed to be accountable for all aspects of work ensuring integrity and accuracy.

Acknowledgements

ClinicalTrials.gov identification number NCT00725257.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: this work was supported in part by the Second University of Naples, and by the Associazione Salute con Stile. The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing the report.