Short-term benefits of an unrestricted-calorie traditional Mediterranean diet, modified with a reduced consumption of carbohydrates at evening, in overweight-obese patients.

Abstract The Mediterranean diet (MeD) is believed to promote health; nevertheless, changes in the nutritional patterns in the Mediterranean area (increased intake of refined carbohydrates/saturated fats; reduced fibers intake; main calorie load shifted to dinner) led to reduced MeD benefits in recent decades. We retrospectively investigated the effects of a MeD with a low intake of refined carbohydrates in the evening (“MeDLowC”) on body weight (BW) and metabolic profile of overweight/obese subjects. According to their adherence to MeDLowC, subjects were classified into 44 (41%) individuals with “excellent” adherence and 63 (59%) with “poor” adherence. Nutritional counseling induced an improvement in BW, glucose metabolism and liver transaminases in both groups, with an increased magnitude of these effects in the “Excellent” adherence group. “Excellent” adherence to MeDLowC improved insulin sensitivity and lipid metabolism. In conclusion, MeD with a restriction of carbohydrates in the evening significantly ameliorates obesity and associated metabolic complications.


Introduction
The Mediterranean diet (MeD), described more than 60 years ago by Ancel Keys as the "elixir of life" in the Cilento region in Italy, is considered one of the healthiest dietary patterns available (Keys et al. 1986). The traditional MeD is a pattern of nutrition characterized by a high intake of vegetables, fruit, whole grains, legumes, nuts and virgin olive oil [being enriched in fibers, antioxidants, monounsaturated and polyunsaturated fatty acids (PUFAs)], and a low consumption of red meat, processed meat, and foods rich in refined and highglycemic-index carbohydrates and unhealthy fats (Trichopoulou et al. 2003;Chrysohoou et al. 2004;Panagiotakos et al. 2006;Saura-Calixto & Goni 2009;Trichopoulou et al. 2009).
The MeD has been shown to be effective in reducing the risk of metabolic syndrome (MS) and cardiovascular disease (CVD), and in increasing life span (Esposito & Giugliano 2002). Although the MeD has long been celebrated for its beneficial impact on health, this concept is debated since some "Mediterranean" patterns have been shown to be not completely devoid of side effects, being associated to obesity, hypertension and high risk of stroke due to the increased consumption of white flour, wine, salt, and by increased total energy intake (Stamler 2013). In the last decades, Mediterranean regions are also progressively losing protection from CVD (Tunstall-Pedoe et al. 1994) since they are rapidly moving from the healthy MeD toward a new dietary pattern particularly enriched in refined grains, saturated fats, sugars, pre-packed food and processed meat (Bonaccio et al. 2014). This pattern of nutrition resembles some features of the "Western diet", which is a well-known risk factor for developing MS and its complications (di Giuseppe et al. 2008;Buijsse et al. 2009;Boffetta et al. 2010;Crowe et al. 2011;Cooper et al. 2012).
These dietary modifications, associated with changes in daily habits such as less physical exercise and a shift of the calorie loads to dinner, could explain the loss of the protective role of the traditional MeD against CVD, and why Italy is suffering from an obesity epidemic and an increase in type 2 diabetes (T2D) prevalence (Monesi et al. 2012). We investigated the hypothesis that the most detrimental tendency of these new dietary habits in Italy may be represented by a high intake of food enriched in refined carbohydrates especially during the evening, when "fast" sources of energy are not needed since the resting phase does not require an excessive burning of nutrients for producing energy (Shimomura et al. 1999;Repa et al. 2000;Denechaud et al. 2008).
A chronic intake of food at high glycemic index [GI; namely the ability of a specific nutrient to raise postprandial glycaemia (Scazzina et al. 2016)], rapidly digested and absorbed, is associated also to peaks in post-prandial insulinemia. In the liver, insulin is essential not only in the control of "fast-to-fed" metabolic shift in carbohydrate metabolism (inducing hepatic glucose uptake, glycolysis and glycogen synthesis, and inhibiting gluconeogenesis and glycogen utilization), but also in promoting de novo lipogenesis (DNL) and very low-density lipoprotein (VLDL) secretion; these processes are all involved in hepatic steatosis and dyslipidemia (Choi & Ginsberg 2011;Kawano & Cohen 2013). We hypothesized that "bringing back" the MeD to a lower content of refined and high glycemic index carbohydrates, and reducing the quantity of carbohydrates and fructose during the evening (when carbohydrates have less chance of being utilized, they are converted into lipids for storage), could have beneficial effects on the body weight (BW) and on the metabolic profile of overweight/ obese patients without the need to apply calorie restriction, or to drastically reduce the total amount of carbohydrates ingested. To address our theory, we retrospectively assessed the impact of an unrestrictedcalorie MeD with a low intake of refined carbohydrates in the evening (MeDLowC protocol) on weight loss and cardiovascular risk (CVR) factors in overweight/obese patients.

Study population
From 2007 to 2013, 153 asymptomatic overweight and obese subjects (body mass index, BMI >25.0 kg/m 2 ) were treated at the Ambulatory of Clinical Nutrition (Head: Prof. Antonio Moschetta) of the Clinica "Augusto Murri" (Director: Prof. Giuseppe Palasciano) of the "Aldo Moro" University of Bari (Italy). We included in this retrospective analysis patients with at least one criterion of MS according to the Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults classification (Adult Treatment Panel III, ATP III) (Grundy et al. 2005) without history of complications (i.e. coronary heart disease, stroke and cancer), endocrine disorders or acute/chronic disease (e.g. liver and renal failure, autoimmune diseases, etc.). Data on lifestyle, educational achievement, history of illness and smoking, and physical activity were recorded during the initial (T0) visit at the time of the first nutritional counselling. The diagnosis of MS was made in the presence of three or more criteria for MS according to ATP III (Grundy et al. 2005). None of the subjects was supplementing their diet with vitamins and/or other nutraceuticals. However, 46 patients were excluded from the retrospective analysis because of missing information on dietary, anthropometric or lifestyle variables, or for dropping out the control visits. Thus, a final number of 107 subjects (mean age: 53.0 ± 1.0 years; 51 M: 56 F) were included in this retrospective analysis. Baseline (T0) characteristics of the study population are summarized in Table 1. The retrospective analysis was approved by the Ethical Committee of the Azienda Ospedaliero-Universitaria Policlinico di Bari (Italy); according to law and in absence of written informed consent, patient records were anonymized and de-identified prior to analysis.

Basal nutritional patterns
The Mediterranean dietary pattern consists of: (a) daily consumption of non-refined cereals (i.e. "wholegrain" bread, pasta, rice, etc.), vegetables, fruits, olive oil and non-fat or low-fat dairy products (i.e. cheese, yoghurt and milk), (b) weekly consumption of potatoes, fish, olives, pulses and nuts, and more rare poultry, eggs and sweets, and monthly consumption of red meat and meat products. This pattern is also characterized by moderate consumption of wine (1/2 glasses of wine a day), which usually accompanies meals. To assess the adherence to the MeD of our patients at baseline, we used the diet score developed by Panagiotakos et al. (2006), as summarized in Supplemental Table 1; the greater is the adherence to the MeD, the higher is the score (up to 55).

The MeDLowC and adherence to the nutritional approach
Over a period of 6 months, all subjects were asked during our 1 h nutritional counselling (performed by our medically qualified team of Internists with nutritional interest) to: (1) avoid high-glycemic-index carbohydrates, and food rich in fructose and alcohol in the evening (fruit was strongly recommended as a substitution for snacks and allowed until afternoon); (2) reduce the intake of red and processed meat; (3) prefer fibers (vegetables) and fish in substitution; (4) consume moderate amount of extra virgin olive oil (D'Amore et al. 2016) and nuts; (5) reduce white sugar and sweets; (6) substitute white (refined) with unrefined flour (also in pasta, pizza and bread) or whole grains (especially if minimally processed) (Buijsse et al. 2009;Mitjavila et al. 2013;McGuire 2016;Mozaffarian 2016). We also suggested to have dinner early in the evening (before 8 pm), when possible (Garaulet et al. 2013). Subjects consumed three main meals and two snacks per day with no reduction in portion size. We also suggested to the patients to drink at least 2 l of water per day, to reduce the intake of salt, to quit smoking and, to keep physical activity (walking) at an average of 7000-10,000 steps per day, when possible (Pate et al. 1995;Sisson et al. 2010;Mozaffarian 2016). On top of these widely accepted lifestyle recommendations, the core issue of our nutritional approach was a progressive reduction in carbohydrate intake from breakfast to dinner, minimizing them as much as possible during dinner. We clearly explained to the patients the rationale of this request to maximize the compliance. A checklist was provided to all participants to easily identify the foods that were forbidden in the evening. (Supplemental Table 2). After six months, adherence to the dietary regimen and physical activity was recorded and all patients were subjected to clinical assessment (study design summarized in Supplemental Figure 1). For the purpose of this retrospective analysis, participants have been classified according to their adherence to the MeDLowC as "excellent" (patients who declared an intake of "forbidden nutrients" at dinner not more than one time a week according to the checklist), or "poor" (patients who declared an intake of the "forbidden nutrients" at dinner more than one time a week according to the checklist).

Blood pressure, anthropometric and laboratory measurements
Body weight and waist circumference (WC) were taken according to standardized procedures (Lohman & Ramr 1988;D'Amore et al. 2013). Systolic and diastolic blood pressures (SBP and DBP) were averaged from three measurements taken after a 10 min rest in the sitting position using a manual sphygmomanometer. After overnight fasting, serum was collected for the assessment of standard biochemical markers of glucose and lipid metabolism, liver and renal function, and inflammation by standard biochemical methods. The homeostatic model assessment for insulin resistance (HOMA-IR) was calculated by the following formula (Fasting plasma glucose Â fasting plasma insulin/405) (Matthews et al. 1985). Cardiovascular risk prediction was assessed using the scoring system of the Progetto Cuore (Palmieri et al. 2004).

Statistical analysis
Data were reported as mean ± standard error (SEM) and frequencies and percentages from the baseline variables. Comparisons of clinical and biochemical parameters among groups ("excellent" versus "poor") were performed for baseline values with the Pearson v 2 test and the Mann-Whitney U test. All clinical parameters on a continuous scale were checked for normal distribution using the Kolmogorov-Smirnov goodness of fit test. Time dependent changes in continuous variables were assessed (separately in the two groups of adherence) with the non-parametric Wilcoxon Signed Rank Test. In order to compare the time dependent changes of each continuous variable between the two groups of adherence to the MeDLowC protocol, we used a multiple regression analysis. This method provided a way of accounting for potentially confounding factors (i.e. physical exercise, statins and therapies for diabetes and hypertriglyceridemia). The strength of the linear relationship between continuous variables was calculated with Pearson's correlation coefficient. Statistical tests were conducted as two tailed hypothesis with an alpha value of 0.05 using NCSS software (NCSS, Kaysville, UT, www.ncss.com).

Baseline characteristics of the study population
One hundred and seven individuals were included in this study. At baseline, subjects had an average BMI of 34.1 ± 0.6 kg/m 2 (25.0 BMI 29.9: 28 patients; BMI !30.0: 79 patients). Seventy-five participants (69%) were positive to 3 or more NCEP-ATP III criteria (Grundy et al. 2005) for MS; 90% with increased WC, 77% with blood hypertension, 50% with increased glycaemia, 50% with hypertriglyceridemia and 48% with low HDL cholesterol. Twenty-nine patients were taking antidiabetic medications (sulfonylureas and/or biguanides), 58 antihypertensive drugs (ACE inhibitors, angiotensin II receptor antagonists and others), 23 statins and five fibrates/PUFAs X3 (details in Table  1). When recommended by good medical practice guidelines, we added antihypertensive drugs (n: 9) and statins (n: 11). Thirty-two percent of patients had an alcohol intake !20 g/day, while 18% were smokers. As expected in our region, the adherence to the MeD (although not complete) was quite high (score was 32.9 ± 1.1 where the maximum adherence is equal to 55.0), mainly driven by the frequent intake of olive oil, moderate amount of alcohol, fruits and vegetables, dairy products, poultry, legumes and fish (Supplemental Table 1).

Weight loss
At the 6 months follow up (T1), 44 subjects (41%) showed an "excellent" adherence to the MeDLowC protocol, while 63 individuals (59%) were classified into the "poor" adherence group. No differences were detected in the two groups according to gender, medical treatment, basal anthropometric measures and serum biochemistry at baseline; the only difference that resulted statistically different (p ¼ 0.05) between the two groups was represented by fibrate/PUFAs X3 treatment (five patients undergoing fibrate/PUFAs X3 exclusively in the "poor" adherence group; see Supplemental Table 3). The groups were also comparable for medical treatment and exercise at the end of the observation period (data not shown). Six months after the nutritional counseling, BW and BMI ( Figure  1(A,B)) were significantly decreased in both groups after the MeDLowC, but the reduction was more evident in the "excellent" adherence group (BW D%: À1. 7% versus À9.4% in the "poor" and "excellent" adherence groups, respectively). Only six patients with "poor" adherence (10%) reached the target of 8% BMI reduction (1998), while the target was achieved in 26 patients (60%) with "excellent" adherence to the MeDLowC (v 2 test, p < 0.001). In the "excellent" adherence group, we observed a significant reduction in WC that was not confirmed in the "poor" adherence group (D%: À1.3% versus À6.8% in the "poor" and "excellent" adherence groups, respectively; Figure  1(C)). Finally, MeDLowC did not impact SBP and DBP (Supplemental Figure 2). All the changes in anthropometric measures were significantly greater in the MeDLowC "excellent" than in the "poor" adherence group. Anthropometric values after the MeDLowC protocol in the "poor" and "excellent" groups are presented in Supplemental Table 4.

Glucose homeostasis
It is well established that MeD and low-carbohydrate diets ameliorate glucose homeostasis (Brehm et al. 2003;Foster et al. 2003;Stern et al. 2004;Gardner et al. 2007;Shai et al. 2008;Frisch et al. 2009;Brinkworth et al. 2009;Elhayany et al. 2010;Hu & Bazzano 2014;Saslow et al. 2014). In line with these evidences, after six months the MeD nutritional intervention significantly reduced fasting glucose (GLC) in both patient groups, and the entity of this reduction was more evident in patients with low carbohydrate intake in the evening (p < 0.05 at multivariate analysis; D%: À2.7 ± 1.9 versus À9.6 ± 1.2% in the "poor" and "excellent" adherence groups, respectively; Figure 2(A)); interestingly, therapy for diabetes mellitus (kept in consideration as a confounding factor in the multivariate analysis), only tended to be beneficial but did not reach statistical significance (p¼0.07; Means are represented as bars: (grey) "poor" adherence group; (white) "excellent" adherence group. Lines represent patients (before/after). Error bars represent SEM. Time dependent changes in continuous variables were assessed with the nonparametric Wilcoxon Signed Rank Test (p value in sections 1 and 2 of each variable plot). In order to compare D% of continuous variables between the two groups we used a multiple regression analysis (adjusted p value in section 3 of each variable plot; p value of the confounding variables in Supplemental Table 5).  Supplemental Table 5). Changes in GLC levels were significantly and positively correlated with BMI (R ¼ 0.3; p < 0.01; Figure 3(A)) and WC (R ¼ 0.2; p < 0.05; Supplemental Table 6). HbA1c was reduced exclusively in the "excellent" adherence group (p < 0. 001) while no differences were detected according to adherence (p ¼ NS; D%: À2.7 ± 1.9%, versus À6.5 ± 3. 9%, % in the "poor" and "excellent" adherence groups, respectively; Figure 2(B)) at multivariate analysis (HbA1c appeared to be dependent mainly from antidiabetic therapy; p < 0.05, Supplemental Table 5). Both insulinemia and HOMA-IR were significantly reduced in the "excellent" adherence group (D%: À41. 6 ± 4.4 and À47.1 ± 5.1% respectively; Figure 2(C,D)) while HOMA-IR was slightly but significantly reduced also in the "poor" adherence group (D%: À11.6 ± 8.1, p < 0.05). Changes in insulinemia and HOMA-IR were strongly associated with adherence to the MeDLowC with no significances associated to medical treatments kept in consideration as confounding factors (multivariate analysis). Changes in fasting insulinemia and HOMA-IR were positively and statistically correlated with those in BMI (R ¼ 0.5; p < 0.001; Figure 3(B,C)), and in WC (R ¼ 0.4; p < 0.01; Supplemental Table 6).
Our results clearly confirm that the MeD strongly ameliorates glucose homeostasis; nutritional counseling, even if incompletely undertaken (i.e. "poor adherence" group), is still weakly effective in ameliorating IR.

Lipid homeostasis
Intriguingly, an "excellent" adherence to the MeDLowC led also to a significant amelioration of the lipid profile which was absent in the "poor" adherence group; we found a significant reduction of TGs (Figure 4(A)), which was positively correlated with changes in insulinemia and HOMA-IR (R ¼ 0.4; p < 0.01; Figure 5(A,B)). Changes in TGs were strongly associated with the adherence to the MeDLowC (Figure 4(A)), with no confounding factors at multivariate analysis [i.e. exercise and medical treatment (including fibrate/PUFAs X3)] (see Supplemental Table 5)].
Total and LDL (low-density lipoprotein) cholesterol were also significantly reduced only in the "excellent" adherence group (D%: À11.5 ± 2.5% and À11.7 ± 3.4% respectively; Figure 4(B,C)); multivariate analysis indicated significance according to the adherence to the MeDLowC protocol (p < 0.001) and statin therapy (with regard to total cholesterol: p < 0.01; Supplemental Table 5). Total and LDL cholesterol variations (%) were also positively correlated with D% BMI in the overall population (R ¼ 0.3; p < 0.01; Figure 3(D,E)). On the other hand, in both groups HDL-c tended to increase but did not reach statistical significance, appearing to be driven mainly by physical activity (p < 0.05 according to multivariate analysis; Supplemental Table 5) and, weakly, by the reduction of BMI (R ¼ À0.2; p < 0.05; Figure 3(F)).
Our results confirm the beneficial effects of MeD on lipid metabolism, with a particular regard to TGs and cholesterol levels, therefore leading to a reduction of CVR prediction according to the "Progetto Cuore" CVD risk score (Supplemental Table 4).

Effects of MeDLowC in patients with MS
We then focused our analysis in the subgroup of our patients with an established diagnosis of MS (three or more ATPIII criteria). After 6 months follow up in this subgroup, 26 subjects (42%) showed an "excellent" adherence to the MeDLowC protocol, while 36 individuals (58%) were "poorly" compliant. The two groups at baseline were comparable for gender, medical treatment, anthropometric measures and serum biochemistry also in this subset of our cohort (Table 2). Six months after the nutritional counseling, all the changes previously described in the main cohort regarding anthropometric measures (BW, BMI, WC), glucose (GLC, HOMA-IR) and lipid (TGs, Total and LDL cholesterol) homeostasis, and liver damage (ALT) were also confirmed in the subset of MS patients, thus confirming that MeDLowC is a powerful tool to ameliorate MS (Table 2).

Discussion
The obesity pandemic is a social health problem associated with increased morbidity (MS and its complications), and mortality (Vacca et al. 2011(Vacca et al. , 2015. Epidemiological trials show that the genetic background is not sufficient per se to justify the onset of obesity, since environmental factors and lifestyle (mainly nutrition and physical exercise) are associated . Lifestyle counseling ameliorates hypertriglyceridemia; excellent adherence to MeDLowC protocol further reduces TGs levels. Excellent adherence to the MeDLowC also leads to a reduction of total and LDL cholesterol, while HDL cholesterol tends to increase. Means are represented as bars: (grey) "poor" adherence group; (white) "excellent" adherence group. Lines represent patients (before/after). Error bars represent SEM. Time dependent changes in continuous variables were assessed with the nonparametric Wilcoxon Signed Rank Test (p value in sections 1 and 2 of each variable plot). In order to compare D% of continuous variables between the two groups we used a multiple regression analysis (adjusted p value in section 3 of each variable plot; p value of the confounding variables in Supplemental Table 5).
with BW gain and the associated cardio-metabolic complications (Vacca et al. 2011(Vacca et al. , 2015. The Mediterranean dietary pattern is believed to be one of the most effective nutritional approaches for preventing and treating MS and CVD (Trichopoulou et al. 2003;Chrysohoou et al. 2004;Panagiotakos et al. 2006;Tzima et al. 2007;Shai et al. 2008;Trichopoulou et al. 2009;Kastorini et al. 2011;Jones et al. 2012). Nevertheless, Southern Italian MeD is not devoid of side effects since has been associated to obesity, hypertension and high risk of stroke due to the high consumption of white flour, wine, salt and by increased total energy intake (Stamler 2013). Recent studies have shown that adherence to the MeD in Italy is also becoming progressively incomplete (Bonaccio et al. 2014), therefore potentially reducing the traditional CVD protection that has been shown in the past. Surprisingly, Italy is nowadays characterized by prevalence and incidence of CVD and cancer really close to those observed other Northern European and American countries (Tunstall-Pedoe et al. 1994;Cameron et al. 2004;Jemal et al. 2011) where "western diet" is the main nutritional pattern. In the last decades, the Italian diet has been characterized by an increase in the consumption refined carbohydrates and meat, with a reduction of the consumption of fruit and vegetables (Pala et al. 2003). Refinement of cereals has been introduced quite recently (during the industrial revolution) to allow flour to last longer; whether refined carbohydrates should be considered a part of "MeD" pattern is quite controversial, and a "traditional" MeD should be characterized by a low proportion of refined carbohydrates (Trichopoulou et al. 2003;Chrysohoou et al. 2004;Panagiotakos et al. 2006;Trichopoulou et al. 2009;Stamler 2013). Changes in nutritional habits could hence have led to an increased prevalence of obesity and MS and their complications (Mendez et al. 2006;Tortosa et al. 2007;van Dam & Seidell 2007;Boffetta et al. 2010;Gonzalez & Riboli 2010;Burger et al. 2011;Crowe et al. 2011;Cooper et al. 2012). While high carbohydrate intake has been associated with increased weight gain (van Dam & Seidell 2007) and development of glucose and lipid metabolism unbalances (Coulston et al. 1983), low carbohydrate diets induce a favorable effect on BW, glucose and lipid metabolism (Brehm et al. 2003;Foster et al. 2003;Stern et al. 2004;Gardner et al. 2007;Shai et al. 2008;Brinkworth et al. 2009;Frisch et al. 2009;Elhayany et al. 2010;Hu & Bazzano 2014;Saslow et al. 2014). Nevertheless, both the glycemic index and the choice of food that contains carbohydrates play a major role in the protective/worsening function of carbohydrates in obesity and MS (e.g. pasta consumption has been recently associated to reduced BMI) (Scazzina et al. 2013;Pounis et al. 2016;Scazzina et al. 2016), as well as the amount ingested. In fact, when carbohydrate intake exceeds the needs (due to an excessive intake or a high glycemic index leading to an increase of circulating glucose and insulin), both glucose [through the activation of carbohydrate-responsive element-binding protein (ChREBP), a key regulator of glucose metabolism and fat storage] and insulin [through the promotion of sterol regulatory element binding protein 1c (SREBP 1c), a major lipid synthesis regulator] stimulate DNL in the liver leading to steatosis, hypertriglyceridemia and lipid accumulation in adipose tissue (Shimomura et al. 1999;Repa et al. 2000;Denechaud et al. 2008;Vacca Figure 5. Correlations of insulin D% and homeostatic model assessment for insulin resistance (HOMA-IR) D% versus triglycerides (TGs) D%. In the overall population, D% of insulin (A) and HOMA-IR (B) were significantly and positively correlated with percent changes in TGs after six months. "Poor" adherence patients in grey; "excellent" adherence patients in white.   . 2015). Additionally, while in the past the lunch was traditionally the main meal (Garaulet et al. 2013), nowadays a focus on dinner is becoming the prominent behavior due to the demands of modern lifestyles. This fact may have led to an additional worsening of the protective features attributed to the MeD. In fact, timing of food intake, and in particular eating late in the night, can negatively predict weight loss effectiveness in dietary intervention (Garaulet et al. 2013). Different theories link circadian rhythms to the control of the metabolic balance and the development of obesity and MS; glucose intake and physical inactivity have been shown to be able to desynchronize the circadian clock leading to metabolic imbalances, and increased risk of developing metabolic diseases (Froy 2007(Froy , 2011. In line with these evidences, the present retrospective study was aimed to investigate the short-term impact of an unrestricted-calorie MeD with a low proportion of refined carbohydrates in the evening combined with moderate daily physical exercise on weight loss and metabolic profile. In particular, we focused our nutritional approach on a progressive (from breakfast to dinner) reduction of the intake of refined carbohydrates, alcohol and fructose. Although fruit is proven to be beneficial for health, we suggested to our patients to avoid fruit at dinner since, in the hypercaloric glycogen-replete state (fruit is traditionally used in Italy as a dessert), intermediary metabolites from fructose metabolism are used as preferential substrates for DNL, and excessive fructose consumption has been associated to hepatic insulin resistance, steatosis and dyslipidemia (Hebbard & George 2011;Lustig 2013).
In this retrospective analysis, we tested the hypothesis that a traditional MeD modified with a reduced intake of refined carbohydrates in the evening (when excessive energy intake is not supported by adequate physical exercise), could significantly promote weight loss reduction and the amelioration of metabolic homeostasis. Also, we did not restrict calorie intake and the total amount of carbohydrates, but we promoted the choice of healthy macronutrients (e.g. carbohydrates at low glycemic index). After six months from baseline, all patients declared a better adherence to the Mediterranean lifestyle, but only 44 patients (41%) accurately followed our nutritional suggestions reducing carbohydrate intake in the evening; 37 patients (35%) also increased physical exercise [18 patients (41%) in the "excellent" group and 19 (30%) in the "poor" group]. In line with scientific evidence (Tzima et al. 2007;Shai et al. 2008), being counseled on healthy lifestyles was associated with a significant amelioration in BW and BMI in both groups. The most intriguing results were achieved regarding metabolic homeostasis (insulinemia and IR improved by $40% in the patients with "excellent" adherence). As a consequence of reduced insulinemia, and in line with previous reports showing a beneficial effect of the reduction of carbohydrate intake on glucose metabolism (Brehm et al. 2003;Foster et al. 2003;Stern et al. 2004;Gardner et al. 2007;Shai et al. 2008;Brinkworth et al. 2009;Frisch et al. 2009;Elhayany et al. 2010;Hu & Bazzano 2014;Saslow et al. 2014), MeDLowC ameliorated GLC, HbA1c and TGs. We also found intriguing beneficial effects in total and LDL cholesterol, and in liver transaminases, while HDL, which tended to increase, appeared to depend on physical exercise rather than on dietary changes.
Although the present study had several limitations such as the retrospective design, the short duration of the observation period, the absence of a positive/negative controls (healthy controls were not seen in our clinic, and would have been not ethical to exclude patients with increased CVR from dietary recommendations), and the lack of an accurate daily nutritional intake record (that would have allowed correlation between macronutrients and clinical benefits), our retrospective analysis points to clear benefits of an unrestricted-calorie traditional MeD in the management of obesity, and its complications in clinical practice. If it is already widely accepted that a beneficial dietary pattern should emphasize a reduction in refined carbohydrates and high intake of minimally processed whole grains (Mozaffarian 2016), our results introduce the concept that carbohydrates could be detrimental when their intake is prevalent at dinner rather than per se; indeed removing unhealthy carbohydrates at dinner "turbo-boosted" the efficacy of MeD to promote BW loss and metabolic homeostasis. In this respect, we propose here a novel unrestrictedcalorie MeD with low intake of refined carbohydrates in the evening as a possible alternative intervention in overweight/obese patients. Further prospective studies are needed to prove the long term effectiveness of this nutritional regimen in obesity, MS, diabetes, CVD and cancer. of the retrospective analysis. All authors read and approved the final manuscript.