Impact of legumes and plant proteins consumption on cognitive performances in the elderly

The neuropsychological assessment was conducted by a medical assessment alongside the use of the MMSE and ADAS-cog.

MMSE is a global test of cognitive function with components of orientation, attention, calculation, language and recall [10]. A score of 20 or below is indicative of cognitive impairment [11, 12]. A validated Italian version was used [14].

The ADAS-cog is a psychometric scale, measuring memory disturbances, language, praxis, attention and other cognitive abilities [15]. The range of scores is from 0 to 70 and the scale of the ADAS-cog is reversed, where 0 represents no errors and 70 represents errors on all items [15, 16].

To reduce the potential for practice effects during subsequent visits, different word lists in the neuropsychological tests were used. In addition, the investigators performing the cognitive tests were blinded to the patients’ clinical data and nutrient intake.

In all patients we assessed the presence of the known classical cardiovascular (CV) risk factors and anthropometric characteristics. The following criteria were used to define the distinct CV risk factors: diabetes: fasting blood glucose ≥126 mg/dL or antidiabetic treatment; hyperlipidemia: total cholesterol >200 mg/dL and/or triglycerides >200 mg/dL or lipid lowering drugs use; hypertension: systolic blood pressure ≥130 mmHg and/or diastolic blood pressure ≥85 mmHg or antihypertensive treatment; overweight: 25 kg/m² ≤ BMI < 30 kg/m²; obesity: body mass index (BMI) ≥30 kg/m²; smoking: current smoker: who has smoked more than 100 cigarettes in their lifetime and smoke cigarettes every day or some days [19, 20]. Written informed consent was obtained from participants. The investigation conforms to the principles outlined in the Declaration of Helsinki [21].

Dietary intake data were assessed by a 24-h recall and a seven-day diet record and calculated using nutritional software MetaDieta 3.0.1 (Metedasrl, San Benedetto del Tronto, Italy). The 24-h recall was collected via an interview by a dietitian who used images associated with a comprehensive food list in the program. All participants were also given a food diary, measuring sheet with life-size images of a spoon, cup and bottle sizes for food diaries. The INRAN (National Institute of Food Research) 2000 and IEO (European Institute of Oncology) 2008 database serves as the source of food composition information in the program [22]. We focused on major food groups, like consumption of fruits fresh and dry, virgin olive oil, legumes, fish, cereals, red meat, animal fats/margarines and cake/pies. The resulting database was exported into SPSS for analysis [23, 24].

The measurement of the systemic blood pressure (systolic blood pressure—SBP and diastolic blood pressure—DBP) of both arms was obtained by auscultatory blood pressure technique with aneroid sphygmomanometer. Clinic BP was obtained in supine patients, after 5 min of quiet rest as previously described [25].

Venous blood was collected after fasting overnight into vacutainer tubes (Becton & Dickinson, Plymouth, England) and centrifuged within 4 h. Serum glucose, total cholesterol, high density lipoprotein (HDL)-cholesterol, triglycerides and creatinine were measured with Enzymatic colorimetric test. Low-density lipoprotein (LDL) cholesterol level was calculated by the Friedewald formula [26]. Quality control was assessed daily for all determinations.

All tests were performed after a 12 h overnight fast. Body weight was measured before breakfast with the subjects lightly dressed, subtracting the weight of clothes. Body weight was measured with a calibrated scale and height measured with a wall-mounted stadiometer. BMI was calculated with the following equation: weight (kg)/height (m)². Waist circumferences and hip circumferences (WC and HC) were measured with a nonstretchable tape over the unclothed abdomen at the narrowest point between costal margin and iliac crest and over light clothing at the level of the widest diameter around the buttocks, respectively, as described in the past [27].

Dietary guidance was provided to all enrolled patients to promote a “healthy diet” by an experienced dietitian, not including a total energy reduction or weight loss goals, nor was physical activity promoted for the entire duration of the study. The enrolled patients were contacted again after 6 months by way of telephone, reminding them to maintain their dietary pattern. In those with nutrient deficiencies or severe obesity, additional assistance was provided so as to achieve their correct nutrient intake. However, these were excluded from the follow-up study.

Data are reported as mean ± SD. To find a correlation between ADAS-cog (basal and follow-up) and a specific food or dietary pattern with an r equal to 0.15–0.20 and 80% power on a two-sided level of significance of 0.05, 190 subjects are required. Changes in MMSE and ADAS cog score from baseline to follow-up were compared using paired Student’s t test (two tailed).Significant differences were assumed to be present at p < 0.05 (two-tailed). All comparisons were performed using SPSS 17.0 for Windows (IBM Corporation, New York, NY, USA).

We obtained food patterns by using the Principal Components Analysis (PCA) [28]. We also obtained, for the first time, “nutrient patterns”, using macronutrients and micronutrients as variables in place of foods groups. The food/nutrient variables were correlated together (correlation coefficients r > 0.4). The orthogonal rotation (obtained with varimax option) was used to derive food patterns (non correlated components). To fix the number of components to retain, the two screen plots of the eigenvalues, that derived from the correlation matrix of the standardized variables, were examined. The eigenvalue is a value that indicates the proportion of the variance in consumption explained by each component. According to the Kaiser criterion, the number of components that should be retained from principal components analysis is equal to the number of eigenvalues that are greater than one. The food/nutrients patterns were named according to scores of the foods/nutrients that correlated most with the factor (>0.4). A higher absolute values in the correlation coefficients indicated that the food/nutrient variable contributes most to the construction of the component. Participants were globally grouped into four different classes according to food patterns and four different according to nutrient patterns (components). The Pearson’s correlation was performed to identify the variables which correlated with the patterns above cited. We analyzed the correlation with the following variables: MMSE (basal and at follow-up), ADAS-cog (basal and at follow-up), age, education level, BMI, WC, glucose, LDL, HDL, triglycerides, SBP, DBP. Furthermore, a stepwise multivariable linear regression analysis was used to test the association between MMSE and ADAS-cog (after 1 year) and each dietary pattern, if this last was in correlation with these neuropsychological tests in the univariate analysis. Finally, we performed a logistic regression analysis to assess which pattern was correlated with an improvement on ADAS-cog test after 1 year.

Table 1 shows the characteristics of the study population. The mean age was 70 ± 4 years, the MMSE score was 24 ± 1 and ADAS-cog was 16 ± 7. The participants’ nutritional intake and food groups used in the PCA are showed in Table 2. Globally, plant-based protein sources represented 37.5% of the total protein consumed.

MMSE and ADAS-cog scores improved at the second time point (MMSE: from 24.4 ± 1 to 25.6 ± 1, p < 0.001; ADAS-cog: from 15.4 ± 5 to 13.1 ± 5, p < 0.001). Total loss at follow-up were 70 individuals, of whom 59 failed to attend to visit, 1 died, and 20 were untraceable.

Figures 1 and 2 shows the screen plot of the eigenvalues, indicating the number of components (food and nutrients patterns) obtained. In our investigation, a four-(component) food patterns, namely: (1) cereals/meat/fish/olive oil pattern, (2) cakes/fruit pattern, (3) animal fats/margarines pattern and (4) legumes pattern; solution was selected. In addition, four-nutrient patterns, namely: (1) animal protein pattern, (2) vegetal oils pattern; (3) fats pattern and (4) plant proteins/polyunsaturated fats pattern, were identified. The correlation coefficients indicated that the food or nutrient variable contributes most to the construction of the component. To simplify this presentation, only the significant data are here shown, i.e. the only patterns associated with MMSE and/or ADAS-cog score and/or improvement on ADAS-cog, that were the following:

As food patterns: (1) legumes pattern (explained variation 66%; 26 g/day = 3 serving/week); (2) animal fats/margarines pattern;

As nutrients patterns: (1) plant protein/polyunsaturated fats (variation 79%); (2) fats pattern. Table 3 shown the factors which were correlated with these patterns. In particular, legumes pattern was correlated with MMSE and ADAS-cog at basal and after 1 years (with 12-months ADAS-cog: r = −0.12 and p = 0.06; with 12-months MMSE: r = 0.21 and p = 0.01) and plant protein/polyunsaturated fats with an improved ADAS-cog test after 1 years(r = 0.18 and p = 0.03). In the stepwise multivariable linear regression analysis (Table 4), legumes pattern was found to be associated with the both ADAS-cog and MMSE after 1 year (B = −0.10, p = 0.006; 95% CI −1.79/−0.30 and B = 0.23, p = 0.01; 95% CI 0.04/0.42, respectively). Finally, vegetal proteins/polyunsatured fats pattern was found to be associated with the improved ADAS-cog (categorical variable) after 1 years (B = 0.584, p = 0.04; OR 1.79, CI 0.04–0.42, Table 5) (improved individuals were all those with a reduction of the ADAS-cog score after 1 years; not improved were those with unchanged or increased score).