The Mediterranean Diet (MedDiet) is associated with a reduced low density lipoprotein cholesterol (LDL-C), even in the absence of weight loss [1,2]. However, dietary interventions generally have a limited long-term efficacy in reducing LDL-C and there is a decreasing adherence to this dietary pattern in Mediterranean countries in Europe [3,4].

The National Cholesterol Education Program (NCEP) Adult Treatment Panel (ATP) III report [5] recommends incorporating plant stanols/sterols into a diet aimed at lowering LDL-C by 10% to 15%. This dietary strategy may reduce LDL-C in individuals with hypercholesterolemia with a low cardiovascular (CV) risk [6] and avoid lipid-lowering medications in high risk patients who cannot take them due to adverse effects [7-11].

It is well accepted that lowering LDL-C using pharmacological and non-pharmacological therapies can reduce the rates of CV events [12]. Thus, although the evidence regarding their efficacy is conflicting [13-15], a growing number of functional foods and nutraceuticals have been proposed to reduce serum lipid levels.

LDL-C can be reduced further through a combination of ingredients with cholesterol-lowering properties, such as plant sterols/stanols, viscous plant fibres from oats, barley, and nuts, soymilk, and soy meat analogues. This leads to an LDL-C reduction from 20% to 30% and a triglyceride reduction of 16% in individuals with hyperlipidemia compared with diet alone [16]. This dietary intervention, namely the Portfolio Diet (P-Diet) is now a recognised approach to lowering cholesterol [17]. However, one of the limitations of this dietary pattern is that it includes several core food groups which may be hard to consume on a continual basis.

The aim of the current pilot study was to verify the feasibility of a simple dietary strategy (called the P-MedDiet) in reducing undesirable plasma LDL-C levels by a MedDiet containing a combination of foods with cholesterol-lowering properties but fewer core food groups than the classical P-Diet, in individuals with hypercholesterolemia. We compared its effect with other dietary interventions such as a MedDiet incorporating a sterol-enriched yogurt (SEY) and MedDiet alone.

Between January and December 2018 we retrospectively reviewed existing medical records of a population composed of adult patients aged over 20 years, attending the outpatient lipid clinic of the “Mater Domini” Azienda University Hospital in Catanzaro, Italy. All of the patients had hypercholesterolemia which was initially managed by a dietary intervention (such as the MedDiet with or without functional foods) or were intolerant to lipid-lowering medications or had refused them [18]. We restricted our analysis to new outpatients who had initiated a dietary treatment with or without functional foods in that period. We included only subjects who had continuously followed the treatment, and had a follow-up LDL-C measurement in the 4 to 8 weeks after the start of the dietary treatment, before a lipid-lowering drug was prescribed.

In accordance with the definition of desirable serum cholesterol levels used by the ATPIII of the NCEP [19], we enrolled individuals with a fasting total cholesterol concentration equal or over 200 mg/dL which may or may not be associated with an elevated triglyceride concentration.

We excluded individuals who took lipid-lowering medications (except in the case of intolerant individuals in whom the interruption duration was more than 3 months), those with secondary causes of hyperlipidemia—as cholestasis, hypothyroidism, pregnancy, sepsis, acute intermittent porphyria, oral contraceptive use, nephrotic syndrome, chronic renal failure, corticosteroid therapy, immunosuppression, past and current alcohol abuse (>20 g of alcohol per day; 350 mL (12 oz) of beer, 120 mL (4 oz) of wine, and 45 mL (1.5 oz) of hard liquor each contain 10 g of alcohol—who were affected by debilitating diseases and those consuming only one type of functional food (for example only soy or only oat or only green tea), as ascertained from their clinical records. We also excluded subjects who were following a different diet from the Mediterranean one (e.g., vegetarian, very low-calorie, intermittent fasting, low-carbohydrate, low-fat diet) and those with any energy restrictions.

All of the 72 enrolled individuals had completed data. The mean age was 52±12 years. In this population, 14% were intolerant to lipid-lowering medications, 31% refused to take any type of lipid-lowering agents, and 55% had no indications for statins. The low-adherence rate was 8%, 4%, and 4% in MedDiet, MedDiet+SEY, and P-MedDiet, respectively.

Table 1 shows the clinical characteristics of participants according to the dietary treatment. At baseline, the groups were comparable for age, BMI, and basal LDL-C. A total of 56 individuals had hypercholesterolemia alone while 16 individuals had a diagnosis of mixed hyperlipidemia (Table 1). Table 2 shows the baseline and follow-up clinical characteristics of participants according to the treatment group. At baseline, the groups were comparable for all of the characteristics (see ANOVA test between dietary treatments). Creatinine and TSH were all in the normal range (not shown). LDL-C decreased at the second time point in the participants taking the P-MedDiet as well as in participants in the MedDiet group with or without a SEY (paired t test) (Table 2). However, the LDL-C reduction was significantly different between groups (see absolute change in Fig. 1). Since age, gender, weight change, follow-up duration, lipid disorders, and basal LDL-C were correlated with LDL-C reduction (univariate correlation; table not shown), we adjusted the LDL-C reduction for all these factors, which were −21±4, −23±4, and −44±4 mg/dL (Fig. 1; −13%, −14%, and −26% in the Fig. 2) in the MedDiet, MedDiet+SEY, and P-MedDiet, respectively, P<0.001 (MedDiet+SEY vs. P-MedDiet, P=0.027). No variables were significantly different between groups at the follow-up visit.

In addition, there was a trend towards a lower triglyceride concentration with P-MedDiet than the control groups (−13, −6, −18 in the MedDiet, MedDiet+SEY, and P-MedDiet, respectively; P=0.70 [due to a large SD]). Supplemental Fig. S1 shows individual LDL-C changes for the participant in each groups. When we excluded participants with hypertriglyceridemia the LDL-C reduction was −22±17, −27±23, and −44±20 in the MedDiet alone, MedDiet+SEY and P-MedDiet, respectively (P<0.007). In the population with only mixed hyperlipidemia, plasma triglycerides improved only with P-MedDiet (−5, 0, −23 mg/dL in the MedDiet alone, MedDiet+SEY, and P-MedDiet, respectively) (data not shown in the tables).

Dietary intake characteristics of participants at baseline and follow-up are showed in Table 3. At baseline, energy and nutrients intake were not significant different between groups. No nutrients significant differed at follow-up between groups except for saturated fats (absolute values 9, 9, and 6 g in the MedDiet, MedDiet+SEY, and P-MedDiet, respectively; P<0.001) (Table 3), polyunsaturated fats (6, 7, and 9 g in the MedDiet, MedDiet+SEY, and P-MedDiet, respectively; P<0.001), and plant sterols (not quantifiable vs. 1.6 and 1.4 g in the MedDiet, MedDiet+SEY, and P-MedDiet, respectively; P<0.001) and chitosan. Supplemental Fig. S2 shows the change in the nutrients intake for the participants in each group. There was a significant reduction in energy intake only in the MedDiet+SEY group (P=0.011, paired t test) (Supplemental Fig. S2A). Fats and monounsaturated fatty acid intake was reduced significantly in the MedDiet and P-MedDiet groups (paired t test) (Supplemental Fig. S2B, S2E). Polyunsaturated fats intake was lower in the MedDiet+SEY group (P=0.033) (Supplemental Fig. S2F) but significantly higher in the P-MedDiet (P=0.012) (Supplemental Fig. S2F) with a greater increase in the P-MedDiet group than in the controls (P=0.002 between group with ANOVA test). In all the dietary treatments the cholesterol and saturated fats intake decreased significantly during the study.

The P-Diet, which includes a combination of ingredients with cholesterol-lowering properties (such as plant sterols/stanols, fibres from oats and barley, nuts and soy products), is a recognised alternative approach lowering cholesterol by 20% to 30% in individuals with hyperlipidemia [16]. In this study we assessed the feasibility of a new dietary approach to lowering cholesterol, combining a MedDiet with a P-Diet (resulting in a P-MedDiet). In Italy, where our study was conducted, the population already follows a Mediterranean style diet. Since the various core food groups of the P-Diet may be hard to consume on a consistent basis, we assessed the efficacy of a simplified P-MedDiet pattern, without nuts and soy and adding chitosan.

This decision was also due to the fact that nuts and soy (or soy products) are rarely reported in our medical reports. In Italy, soy is not used to reduce LDL-C because the European food safety authority (EFSA) has not confirmed its cholesterol-lowering effect [12]. Due to the risk of allergies, we frequently observe that some patients refuse to consume nuts [25].

We found that the LDL-C reduction was significantly different between groups (Fig. 1). However, in line with several previous studies [16,17], the LDL-C reduction was higher in the P-MedDiet group than in the control groups (−13%, −14%, and −25% in the MedDiet alone, MedDiet+SEY, and P-MedDiet, respectively) (Fig. 1). It has been demonstrated that a nutraceutical combination comprising red yeast rice, berberine, policosanol, astaxanthin, coenzyme Q10, and folic acid was effective in reducing LDL-C by 23 mg/dL and triglycerides by 13 mg/dL [26] while there was an LDL-C reduction of 44 mg/dL and a triglycerides reduction of 18 mg/dL with our P-MedDiet. To date, no studies exist on the effects of a similar diet, except for one carried out on patients with human immunodeficiency virus (HIV) dyslipidaemia [27].

Our study was not designed to determine which of the foods were responsible for the effects observed, but to assess the combined impact of a MedDiet incorporating a combination of functional foods. However, we observed a significant change in the intake of nutrients during the study (Table 3, Supplemental Fig. S2). In particular, we found a greater intake of polyunsaturated fats in the P-MedDiet than in the control groups (Supplemental Fig. S2F). This finding is consistent with the concept that, in term of lowering LDL-C, a diet rich in monounsaturated fat is as effective as a diet rich in polyunsaturated fat [28].

The cholesterol lowering effects induced by this diet may take place through, for example, reduced intestinal cholesterol absorption due to β-glucan, plant sterols, and chitosan [29,30], interference with transport-mediated processes of cholesterol uptake in liver by plant sterols [31], decreased hepatic cholesterol synthesis by green tea [32] and increased bile acid loss by the action of dietary fibre [33]. Furthermore, fish protein has been shown to decrease serum cholesterol by increasing fecal cholesterol and bile acid excretion [34].

Another important finding of this study regards the plasma triglyceride level which improved only in the P-MedDiet group, as shown in a previous study on another type of P-Diet [17]. It is well known that there is a linear relationship between serum cholesterol levels and the risk of CV events [35]. The incremental decrease in LDL-C by non-pharmacological therapies can translate into reductions in CV event rates [12]; thus, our results are of important clinical significance.

In this study, we tested several functional foods whose health claims relating to the cholesterol-lowering effect have already been approved by EFSA, including a SEY in the control group. We selected a SEY due to its wide commercialization and acceptability. Furthermore, the lipid-lowering effects of this SEY are well accepted and, most importantly, yogurt is consumed daily. The P-MedDiet includes key components of the MedDiet as well as the consumption of a SEY. Our P-MedDiet derives from the MedDiet as well as the P-Diet. We, thus, decided to compare our P-MedDiet with the other two dietary patterns to better understand the net effect of certain products (oats, barley, chitosan, green tea) with a lipid-lowering effect.

In our study all groups lost weight to the same extent. This finding reinforces our hypothesis regarding the greater lipid-lowering effect obtained by combining functional foods in the MedDiet. While the results of this study cannot be generalized due to the type of sampling used, these results could still be useful in designing future intervention trials aimed at testing the efficacy of new combinations of functional foods. In addition, there have been several randomized, controlled trials combining the use of plant sterols and statin medications which show that in combination plant sterols have a significant additive effect [36]. Thus, we hypothesize that a P-MedDiet, used as an adjuvant to drug therapy, may be effective in reducing serum total and LDL-C.

In this study, some weaknesses need to be addressed. Firstly, the nature of the study (pilot) and our sample make it difficult to generalize these results. Experimental designs or larger longitudinal studies would be necessary to confirm these findings. However, this type of study is helpful in detecting relationships among different phenomena, in order to generate hypothesis. Larger and more inclusive studies are needed on this important issue. Secondly, due to the study design, this investigation did not address P-MedDiet sustainability in the longer term. Despite these limitations, our study has strengths. Since the P-MedDiet differs from all other P-Diets tested to date, our study is original. For the first time, we demonstrated the feasibility of a new and simple dietary approach, namely the P-MedDiet, which involves consuming high amounts of legumes, fish, fruit, and vegetables, incorporating a new combination of functional foods with cholesterol-lowering properties, such as oats or barley, plant sterols, chitosan, and green tea, excluding soy and nuts, in reducing LDL-C by 25% in the short term in individuals with hypercholesterolemia.