**Dietary Habits and Risk of Early-Onset Dementia in an Italian Case-Control Study**

**Tommaso Filippini 1, Giorgia Adani 1, Marcella Malavolti 1, Caterina Garuti 1, Silvia Cilloni 1, Giulia Vinceti 2,3, Giovanna Zamboni 2,3, Manuela Tondelli 3,4, Chiara Galli 3,4,5, Manuela Costa 6, Annalisa Chiari <sup>3</sup> and Marco Vinceti 1,7,\***


Received: 29 October 2020; Accepted: 27 November 2020; Published: 29 November 2020

**Abstract:** Risk of early-onset dementia (EOD) might be modified by environmental factors and lifestyles, including diet. The aim of this study is to evaluate the association between dietary habits and EOD risk. We recruited 54 newly-diagnosed EOD patients in Modena (Northern Italy) and 54 caregivers as controls. We investigated dietary habits through a food frequency questionnaire, assessing both food intake and adherence to dietary patterns, namely the Greek-Mediterranean, the Dietary Approaches to Stop Hypertension (DASH), and the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diets. We modeled the relation between dietary factors and risk using the restricted cubic spline regression analysis. Cereal intake showed a U-shaped relation with EOD, with risk increasing above 350 g/day. A high intake (>400 g/day) of dairy products was also associated with excess risk. Although overall fish and seafood consumption showed no association with EOD risk, we found a U-shaped relation with preserved/tinned fish, and an inverse relation with other fish. Similarly, vegetables (especially leafy) showed a strong inverse association above 100 g/day, as did citrus and dry fruits. Overall, sweet consumption was not associated with EOD risk, while dry cake and ice-cream showed a positive relation and chocolate products an inverse one. For beverages, we found no relation with EOD risk apart from a U-shaped relation for coffee consumption. Concerning dietary patterns, EOD risk linearly decreased with the increasing adherence to the MIND pattern. On the other hand, an inverse association for the Greek-Mediterranean and DASH diets emerged only at very high adherence levels. To the best of our knowledge, this is the first study that explores the association between dietary factors and EOD risk, and suggests that adherence to the MIND dietary pattern may decrease such risk.

**Keywords:** early-onset dementia; dietary habits; MIND diet; DASH diet; Mediterranean diet; risk; prevention

#### **1. Introduction**

Dementia is a syndrome, usually of a chronic or progressive nature, characterized by impairment of cognitive functions beyond what might be expected from normal ageing [1,2]. Early-onset dementia (EOD) is a heterogeneous group of cognitive disorders characterized by an onset of dementia symptoms before the age of 65 [3]. Such an age cut-point has not been established based on biological differences between younger and older subjects, but mainly on the socio-economic implications of dementia diagnosis at a younger age [3,4]. Indeed, the main feature of EOD compared to late-onset forms is a higher impact at two levels: First, on affected people, particularly in terms of their social functioning and working life [5]; second, on family members, especially when young children are still present [6].

EOD prevalence has been estimated to range between 38 and 420 cases per 100,000 inhabitants, with an annual incidence between 2.4 and 22.6 new cases per 100,000 inhabitants [4], the most common forms being Alzheimer's dementia (AD), frontotemporal dementia (FTD), and vascular dementia [7].

Little is known about EOD etiology, not least in comparison with the determinants of late-onset dementia. Genetic mutations apparently only account for a small percentage of EOD, around 10% [8,9]. Therefore, the role of environmental factors and modifiable lifestyles including diet seems particularly relevant [10–14]. Among the dietary habits involved, a high consumption of vegetables, fruit, and fish [15–17], and adherence to the Mediterranean diet or other dietary patterns (e.g., the Dietary Approaches to Stop Hypertension (DASH) and the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet) have been associated with a slower cognitive decline and decreased risk of all-age dementia [12,18–21]. Interestingly, recent studies have evaluated the correlation between AD brain biomarkers and dietary patterns characterized by a higher intake of fresh fruit and vegetables, whole grains, fish and low-fat dairies, along with a lower intake of sweets, fried potatoes, high-fat dairies, and meat. Data provided evidence of protective effects against the risk of developing AD, suggesting that dietary interventions may play a role in the prevention of cognitive decline [22,23].

In this study, we investigated EOD risk in an Italian population in relation to dietary habits, including food consumption and adherence to dietary patterns.

#### **2. Methods**

Following approval by the Modena Ethics Committee (*n*. 186/2016), we performed a case-control study on environmental and lifestyle risk factors for EOD in the province of Modena, Northern Italy. We recruited EOD cases from newly-diagnosed patients referred to the Cognitive Neurology Network of Modena province, including the Modena Policlinico-University Hospital Memory Center (Modena, Italy) and the Carpi Hospital Neurology Department (Carpi, Italy), in the period October 2016–2019 [7,24]. Cases are referred to this Network by either Neurology Units, primary care services, or general practitioners through specific pathways activated by the Modena Local Health Authority to identify and care for subjects with dementia. Inclusion criteria encompassed EOD diagnosis, residence in the Modena province, and presence of a reliable caregiver. The diagnosis of EOD subtypes has been established according to the most recent clinical criteria, as previously described [7]. The gene mutation status was not available for all participants since it is not routinely part of the clinical workflow. Among those tested, one subject was a carrier of *SP1* gene mutation. We recruited controls from caregivers of subjects with a diagnosis of early or late-onset dementia referred to the same Cognitive Neurology Network. All subjects provided a written informed consent.

We administered a questionnaire tailored to collect personal characteristics and clinical, occupational, and environmental factors potentially affecting the central nervous system [25], and a detailed food frequency questionnaire (FFQ). The latter is a validated semi-quantitative FFQ developed within the European Prospective Investigation in Cancer (EPIC) project, in a version specifically validated for the population of Northern Italy [26,27]. The EPIC-FFQ was designed to estimate the intake of 188 food items over the previous year in terms of frequency and amount. Photos of serving sizes were also used to assist with proper completion by participants.

Foods and beverages were categorized into major food groups and subgroups based on the common EPIC-SOFT classification, as previously reported in detail [28,29]. The final list of food categories included the following items and subcategories: Cereals and cereal products, meat and meat products, milk and dairy products, eggs, fish and seafood, vegetables, mushrooms, legumes, potatoes, fresh and dry fruits, sweet products, oils and fats, and beverages. We also computed alcohol (ethanol) intake by conversion of all quantities of alcoholic beverages into grams of ethanol per day using a methodology previously described [30]. We also computed scores for three diet quality patterns defined *a priori*: The Greek Mediterranean (GM) diet [31], the Dietary Approaches to Stop Hypertension (DASH) diet [32,33], based on a methodology described elsewhere [34], and the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet [19,35]. In more detail, the GM diet is based on Mediterranean diet scales [31], and scoring is calculated on median intake levels of nine items: Vegetables, legumes, fruit and nuts, dairy products, cereals, meat and meat products, fish, alcohol, and the monounsaturated/saturated fatty acid ratio [34,36]. The range of possible scores was 0–9, with higher scores indicating higher adherence. The DASH diet was originally designed to reduce blood pressure [32,33], and it has been suggested to be neuroprotective [37]. DASH diet adherence scores were calculated according to previous studies [34,38] based on eight components: Fruits, vegetables, nuts and legumes, low-fat dairy products, whole grains, sodium, sweetened beverages, red and processed meats. Overall, possible scores ranged from 8 to 40, with higher scores indicating higher adherence. Eventually, the MIND diet was developed as a hybrid of the Mediterranean and DASH diets associated with slower cognitive decline and decreased incidence of Alzheimer's dementia [19,35]. MIND diet scores were based on the intake of 15 items, namely whole grains, green leafy and other vegetables, berries, red meat, poultry, fish, legumes, nuts, fast/fried food, olive oil and other fats, cheese, sweets and alcohol/wine. Scores ranged from 0 to 15, with higher scores meaning higher adherence.

In data analysis, we used a multivariable unconditional logistic regression model to estimate the EOD risk associated with dietary factors and patterns. We performed an analysis on the overall population for risk of EOD, also performing a stratified analysis according to the type of diagnosis, namely early-onset Alzheimer's dementia (EO-AD) and early-onset frontotemporal dementia spectrum (EO-FTD). Regarding dietary factors, we calculated the odds ratio (ORs) and 95% confidence intervals (CIs) according to the increasing tertiles based on the distribution in the control group using the lowest tertile as a reference category. We also modeled the relation between dietary factors and EOD risk using the restricted cubic spline model with three knots (10, 50, and 90 percentiles). We included in the multivariable model as potential confounders and effect-modifiers sex, age, educational attainment (as years of education), and total energy intake (kcal/day). We used "logit", "mkspline", and "xblc" routines of the Stata-16.1 statistical package (Stata Corp., College Station, TX, USA, 2020) for statistical analyses.

#### **3. Results**

#### *3.1. Characteristics of the Study Population*

Of the 150 eligible participants, only 144 could be contacted. We recruited 112 subjects, 58 EOD cases, and 54 controls with an average response rate of 78%. Reasons for non-participation were unwillingness to contribute to the research and lack of time to fill out the questionnaire. In addition, four cases were excluded as they did not return a reliable and complete FFQ, leaving 54 cases and 54 controls for the final analysis. Table 1 reports the characteristics of the study participants. The average age at the questionnaire filling date was 65 years (66 for EOD cases and 64 years for referents), with a higher proportion of women (57%). The mean age of EOD onset was 59.8 years (59.7 years for EO-AD and 59.8 for EO-FTD), ranging from 45 to 65 years. Alzheimer's dementia (EO-AD) was the most frequent diagnosis (*n* = 30), followed by frontotemporal dementia spectrum disorders (EO-FTD, *n* = 18), vascular dementia (*n* = 4), or other rarer diseases (Supplemental Table S1).


**Table 1.** Characteristics of the study participants. Early-onset dementia (EOD), early-onset Alzheimer's dementia (EO-AD), and early-onset frontotemporal dementia spectrum (EO-FTD).
