Apolipoprotein E gene variants shape the association between dietary fibre intake and cognitive decline risk in community-dwelling older adults

Abstract

Background

healthy dietary patterns have been associated with lower risk for age-related cognitive decline. However, little is known about the specific role of dietary fibre on cognitive decline in older adults.

Objective

this study aimed to examine the association between dietary fibre and cognitive decline in older adults and to assess the influence of genetic, lifestyle and clinical characteristics in this association.

Design and participants

the Invecchiare in Chianti, aging in the Chianti area study is a cohort study of community-dwelling older adults from Italy. Cognitive function, dietary and clinical data were collected at baseline and years 3, 6, 9 and 15. Our study comprised 848 participants aged ≥ 65 years (56% female) with 2,038 observations.

Main outcome and measures

cognitive decline was defined as a decrease ≥3 units in the Mini-Mental State Examination score during consecutive visits. Hazard ratios for cognitive decline were estimated using time-dependent Cox regression models.

Results

energy-adjusted fibre intake was not associated with cognitive decline during the 15-years follow-up (P > 0.05). However, fibre intake showed a significant interaction with Apolipoprotein E (APOE) haplotype for cognitive decline (P = 0.02). In participants with APOE-ɛ4 haplotype, an increase in 5 g/d of fibre intake was significantly associated with a 30% lower risk for cognitive decline. No association was observed in participants with APOE-ɛ2 and APOE-ɛ3 haplotypes.

Conclusions and relevance

dietary fibre intake was not associated with cognitive decline amongst older adults for 15 years of follow-up. Nonetheless, older subjects with APOE-ɛ4 haplotype may benefit from higher fibre intakes based on the reduced risk for cognitive decline in this high-risk group.

Key Points

• Dietary fibre intake was not associated with cognitive decline during 15 years in this cohort of 848 older adults from Italy.

• Higher intakes of dietary fibre showed a protective effect against cognitive decline only in subjects with Apolipoprotein E (APOE) ɛ4 haplotype.

• These results may contribute to personalised dietary fibre recommendations to prevent cognitive decline in older adults.

Introduction

Age-related cognitive decline, considered a precursor of dementia [1], is an important public health concern lacking effective treatments. Thus, identifying modifiable risk factors for cognitive decline is important to develop effective strategies of prevention. Amongst possible prevention strategies, diet quality has been associated with age-related cognitive decline [2]. Dietary patterns characterised by high intake of plant-based foods, such as the Mediterranean diet, have been associated with lower risk for cognitive decline and dementia [3–5]. Despite the positive impact on cognition of healthy dietary patterns characterised by high consumption of food groups rich in fibre, whether there is a specific effect of dietary fibre is unclear.

Dietary fibre refers to plant-derived non-starch polysaccharides, resistant oligosaccharides, lignin and resistant starch that are resistant to human digestive enzymes [6]. National dietary guidelines currently include an optimal dietary fibre intake, which for adults in western countries are in the order of 30–35 g/d for men and 25–32 g/d for women [7]. However, average intakes do not reach this level in any country [8]. Dietary fibre is known to enhance gastrointestinal [9], immune and metabolic health [10, 11], but its influence on cognitive function has not been fully elucidated. Evidence from animal studies using synthetic, extracted or single foods high in fibre shows beneficial effects on cognition [12, 13], but data from human cohorts are still scarce, especially in older adults. In human studies, diets rich in fibre have been associated with better cognitive scores [14–17], whereas the opposite was observed for simple carbohydrates [14, 18, 19]. Vercambre et al. [17] showed an association between higher fibre intake and lower rate of cognitive decline in a French cohort study of older women. Furthermore, a recent study showed an inverse association between dietary fibre and disabling dementia in a long follow-up study in Japanese middle-aged adults [20]. Indeed, dietary fibre intake, together with other nutrients, was associated with measures of improved brain integrity, such as larger total brain volume and lower white matter hyperintensity volume in dementia-free older adults [21]. The impact of dietary fibre on cognitive function might be mediated by its influence on gut microbiota. Dietary fibre intake shapes gut microbiota composition and/or functionality, increasing the output of fermentative end-products, such as short-chain fatty acids (SCFAs) [22]. These metabolites have shown the potential to modulate brain function via the ‘microbiota-gut-brain’ axis [23].

Our hypothesis was that higher intake of dietary fibre would have a protective effect on cognitive decline in older adults. The main objective was to evaluate the relationship between dietary fibre intake and the incidence of cognitive decline in a prospective study of community-dwelling older adults, the Invecchiare in Chianti, aging in the Chianti area (InCHIANTI) study. In addition, a secondary aim was to assess whether other genetic, lifestyle and clinical factors modulate the effect of dietary fibre intake and the risk of cognitive decline.

Methods

Study design

The InCHIANTI is a prospective study including community-dwelling older adults living in the Chianti geographic area (Tuscany, Italy) [24]. The baseline data were collected in 1998–2000 and three follow-up assessments were conducted every 3 years up to year 9 and then one last assessment by year 15. The Italian National Institute of Research and Care of Aging Institutional Review and Medstar Research Institute (Baltimore, MD, USA) approved the study protocol, and all participants signed an informed consent.

The current report followed the Strengthening the Reporting of Observational Studies in Epidemiology-Nutritional Epidemiology guidelines [25].

Study population

One thousand four hundred and fifty-three community-dwelling adults were randomly selected from the population registries of two Italian communities: Greve in Chianti and Bagno a Ripoli, using a multistage, stratified sampling method. At baseline, 1,155 participants aged ≥ 65 years were evaluated. Sixteen participants were excluded because they had missing data on dietary questionnaires. In addition, we excluded 123 participants who died within the first 3 years of follow-up, 42 participants with dementia at baseline and 126 participants without two consecutive Mini-Mental State Examination (MMSE) evaluations during the follow-up (Figure 1). In total, the final data set comprised 848 participants and 2,038 observations.

Figure 1.

Flowchart of participants included in the analysis.

Dietary intake

Habitual dietary intake was assessed by trained interviewers using the Italian version of the food frequency questionnaire developed and validated in the European Prospective Investigation into Cancer and Nutrition-Italy study [26]. This questionnaire asked how often (daily, week and monthly) the consumption of 198 food and beverages items occurred in the past year, considering their respective portion sizes. Nutrient data for specific foods were obtained from the Food Composition Database for Epidemiological Studies in Italy [27]. Fibre content of foods in the Database was measured using a modified version of the AOAC method [27]. For this analysis, dietary data at baseline and during follow-up (at 3, 6 and 9 years) were considered.

Mediterranean diet adherence score

Adherence to Mediterranean diet was computed using a 9-point linear scale described by Trichopoulou et al. [28], that ranged from 0 (no adherence) to 9 (high adherence).

Covariates

Covariates were selected based on existent literature and associations with the outcome of cognitive decline in the InCHIANTI study [17, 29, 30]. Age, sex, years of education and physical activity were assessed through standardised questionnaires. Smoking habits were self-reported, and participants were classified into never, former or current smokers. Height and weight were measured, and body mass index (BMI) was calculated in kg/m². Depression was assessed with the Center for Epidemiologic Studies Depression scale. Comorbidities and its definitions are detailed in Supplementary Methods.

Apolipoprotein E haplotypes

Overnight fasting blood samples were used for genomic DNA extraction as previously described [31]. Apolipoprotein E (APOE) variant genotypes (e2, e3, e4) were defined by two single-nucleotide polymorphisms (SNPs), rs429358 and rs7412. These two SNPs were genotyped using TaqMan assays (Applied Biosystems, Inc. [ABI], Foster City, CA, USA) [32].

Outcome assessment

Global cognitive performance was assessed using the MMSE, and cognitive decline was defined as ≥3 points decline in MMSE between two consecutive follow-up evaluations, according to previous publications [33, 34]. Number of participants, time periods and cognitive decline cases during the follow-up are shown in Supplementary Figure 1.

Statistical analysis

In 54 participants (6.4%), BMI and renal function at baseline were missing and values were imputed using the baseline median. During the follow-up, missing values in dietary data (k = 79 observations, 4%) were imputed by the Last Observation Carried Forward. Dummy variables were created to identify participants with imputed values at baseline and during the follow-up to be excluded in sensitivity analyses.

Fibre intake was adjusted by energy using the residual method [35]. Energy-adjusted fibre intake (g/d) was categorised into tertiles for the descriptive analysis at baseline. Continuous variables are presented as mean (SD) or median (interquartile range). Categorical variables are expressed as percentages. Differences in baseline characteristics, comorbidities and dietary intake data across tertiles of energy-adjusted fibre intake and between participants according to their APOE haplotype were assessed using generalised linear models (GLM) adjusted for age and sex. We evaluated the differences in energy-adjusted fibre intake during the follow-up using linear mixed models with participants as a random effect. Intra-class correlation coefficients were calculated to assess changes in fibre intake during the follow-up.

Cognitive decline was considered an event because of the skewed distribution of MMSE scores and the non-linear trajectories of MMSE scores over time in older adults [36]. Therefore, Cox regression was used instead of linear mixed models for analyses. Time-to event was defined as the time elapsed from baseline to the one at which cognitive decline was first detected (see Supplementary Figure 1 for details). Three time-dependent Cox models were used for main analyses controlling for potential confounders selected a priori. Model 1 was adjusted for age at baseline (years), sex and total energy intake (kcal/d). Model 2 was further adjusted for BMI, years of education, smoking, alcohol intake (g/d), physical activity and APOE haplotype. Model 3 was additionally adjusted for comorbidities: impaired renal function (IRF), diabetes, depression, chronic obstructive pulmonary disease (COPD), hypertension, cancer and cardiovascular diseases (CVD). Restricted cubic splines were used to model non-linear associations between fibre intake and cognitive decline.

Interaction analyses were conducted according to the risk factors for dementia described by the 2020 report of the Lancet Commission on Dementia prevention, intervention and care [37]. Interaction terms between energy-adjusted dietary fibre (as g/d) and sex, smoking, physical activity, alcohol intake, diabetes, hypertension and APOE haplotype in relation to cognitive decline were added to the Cox regression models. Sensitivity analyses are detailed in Supplementary Methods. For the statistical analyses, SPSS version 25.0 (IBM, USA) and R 4.0.5 (R foundation, Vienna, Austria) were used. P-values < 0.05 were considered statistically significant.

Results

General characteristics

The studied population consisted in 848 participants (56% women) with a mean age of 74 ± 7 years at baseline (Table 1). Amongst comorbidities, the most frequent were hypertension (59% of participants), IRF (33%), depression (33%) and CVD (21%). Mean MMSE was 26 ± 3, and daily total fibre intake was 20 ± 6 g/d.

Table 1.

Baseline characteristics of the population according to tertiles of energy-adjusted dietary fibre intake in the InCHIANTI study.

	Total	Dietary fibre
(n = 848)		T1  (282)	T2  (283)	T3  (283)
Age (years)	74 ± 7	74 ± 7	73 ± 6	73 ± 6
Women (%)	56	53	57	57
BMI (kg/m2)a	27 ± 4	26 ± 3	27 ± 4	27 ± 3
Education (years)	5 ± 3	5 ± 3	5 ± 3	5 ± 3
Current smoking (%)	27	29	25	27
Physical activity (%)
Sedentary	18	22	17	14
Light	45	40	48	46
Mod-High	37	38	35	40
Hypertension (%)	59	57	61	58
Diabetes (%)	12	7	14	13*
IRF (%)	33	36	31	31
CVD (%)	21	21	20	20
COPD (%)	7	9	6	6
Cancer (%)	6	5	5	7
Depression (%)	33	37	30	31
MMSE (0–30)	26 ± 3	25 ± 3	25 ± 2	25 ± 2
APOE haplotypea (%) ε2 ε3 ε4	14 71 15	15 16 69	12 73 15	14 72 14
MDS (0–9)	4 ± 2	3 ± 1	4 ± 1	5 ± 1**
Dietary characteristics
Energy (103 kcal/d)	1.9 ± 0.6	1.9 ± 0.6	1.8 ± 0.5	1.9 ± 0.5
Dietary fibre (g/d)	20 ± 6	15 ± 4	18 ± 3	24 ± 4**
Protein (g/d)	76 ± 20	77 ± 20	72 ± 19	77 ± 19
Total lipids (g/d)	62 ± 19	62 ± 20	57 ± 17	64 ± 19
SFA (g/d)	22 ± 8	23 ± 3	20 ± 6	21 ± 7**
MUFA (g/d)	33 ± 11	31 ± 10	30 ± 9	35 ± 11**
PUFA (g/d)	7 ± 2	7 ± 2	6 ± 2	7 ± 2*
Carbohydrates (g/d)	242 ± 76	241 ± 79	231 ± 71	253 ± 75
Alcohol (g/d)	8 (0–27)	13 (1–32)	7 (0–19)	2 (0–13)**

MDS, Mediterranean diet score. Data for continuous variables are shown as mean ± SD or median (interquartile range). Cut-off points for energy-adjusted dietary fibre intake (g/d) were: tertile 1, 7.6–18.5 g/d; tertile 2, 18.5–22.1 g/d; tertile 3, 22.1–45.7 g/d.

^aData available for 810 participants.

^* P for trend < 0.05, **P for trend < 0.001 using age- and sex-adjusted GLM.

Overall, baseline characteristics across the participants classified according to energy-adjusted fibre intake were similar. Participants with the highest consumption of energy-adjusted dietary fibre had a higher prevalence of diabetes, a higher adherence to Mediterranean diet, a lower alcohol and saturated fatty acid (SFA) intakes, and a higher intake of monounsaturated fatty acid (MUFA) and polyunsaturated fatty acid (PUFA) (Table 1).

Regarding dietary sources of total fibre intake, major contributors were fruit and cereals (43 and 35%, respectively), followed by vegetables (16%), legumes (5%) and nuts (0.4%). During follow-up, energy-adjusted fibre intake remained stable with an intra-class correlation coefficient of 0.44. Nonetheless, fibre intake decreased during follow-up and differences were only significant between the first and the last dietary evaluation (9-year follow-up) [mean difference (95% CI): −2.7 (−3.1 to −2.2) g/d].

Association between fibre intake and cognitive decline

During follow-up, 549 participants developed cognitive decline. Time-dependent Cox regression models showed that energy-adjusted fibre intake was not associated with cognitive decline for the 15-years follow-up (P > 0.05, Supplementary Table 1). The results remained the same when fibre intake was included in a non-linear model (P > 0.05), as well as in sensitivity analysis (Supplementary Table 2).

A statistically significant interaction was only observed between APOE haplotype and fibre intake in relation to cognitive decline risk (P for interaction = 0.02, Figure 2). In participants with APOE-ɛ4 haplotype, a higher fibre intake was significantly associated with a lower risk of cognitive decline [HR (95% CI) per 5 g/d of fibre intake: 0.70 (0.52–0.95)]. Conversely, in participants with the APOE-ɛ2 and APOE-ɛ3 haplotypes, this association was not statistically significant [0.77 (0.55–1.08) and 1.04 (0.91–1.19), respectively, for a 5 g/d difference of fibre intake].

Figure 2.

Interaction between APOE haplotype and energy-adjusted fibre intake (g/d) for the association with cognitive decline. Time-dependent Cox regression model, including ^*energy-adjusted dietary fibre (residual method) adjusted for baseline age, sex, BMI, years of education, smoking, alcohol intake (g/d), physical activity, comorbidities (IRF, diabetes, depression, COPD, hypertension, cancer and CVD) and APOE haplotype. No. of subjects at baseline = 810. No of events = 522. No. of observations = 1,955. Shaded areas are 95% confidence intervals.

We compared the participants according their APOE haplotype (n ɛ2 = 110, n ɛ3 = 578, n ɛ4 = 122) to better characterise them. Participants with APOE-ɛ4 haplotype had lower BMI and higher alcohol intake. However, they showed no statistically significant differences in other sociodemographic, clinical or dietary characteristics (Supplementary Table 3).

Discussion

The current study explored the association between dietary fibre intake and the development of cognitive decline in a cohort of community-dwelling older adults during 15 years of follow-up. Overall, we found that a higher intake of dietary fibre was not significantly associated with a lower risk for cognitive decline. However, a higher intake of dietary fibre showed a protective effect against cognitive decline in subjects with APOE-ɛ4 haplotype, a genetic risk factor extensively associated with an increased risk for impaired cognitive function [38].

Up to date, few studies have assessed the specific association of dietary fibre intake with cognitive function in older adults. A 13-year follow-up study of older French women found higher risk of cognitive decline in participants with lower intake of total and soluble dietary fibre [17]. Similarly, a cross-sectional analysis of the US National Health and Nutrition Examination Survey showed that higher dietary fibre intake was associated with improved scores in the Digit Symbol Substitution Test, a surrogate marker for some domains of cognitive function, in older adults aged 60 years and older [39]. Randomised controlled trials reviewed by La Torre et al. [40] showed that dose and type of dietary fibre may have differential effects on cognitive function. This could be a possible explanation for our null association in the whole study. Indeed, in an Australian cohort of adults aged 55–65 years habitual consumption of high-fibre breads (but not from fruits and vegetables) was associated with better cognitive function [2]. Further research assessing the potential differential effects on cognitive outcomes depending on the dietary source of fibre is needed in large cohorts. Our results describe for the first time an interaction between fibre intake and APOE haplotype. In participants with APOE-ε4, dietary fibre intake showed a significant protective association with cognitive decline. Up to date, APOE-ε4 haplotype is considered the strongest genetic risk factor for late onset Alzheimer’s disease and has been recently associated also with poorer cognition and greater risk of dementia at older ages [38, 41]. Increasing evidence points that APOE-ε4 carriers could respond differently to dietary and metabolic treatments [42–44]. Indeed, it has been shown that a high fat diet improved cognition in APOE-ε4 carriers, whereas it worsened Alzheimer’s disease biomarkers for other APOE variants [45]. However, the interaction between APOE-ɛ4 and fibre intake has not been reported before. In the light of the present findings, it is possible that some of this protection could come from different responses elicited by dietary fibre, revealing a nutrigenetic interaction. APOE is involved in lipid transport, cholesterol homeostasis, synaptic plasticity, mitochondrial function and insulin signalling [46]. Thus, APOE gene alleles have been described to change the response to diet at many levels. APOE4 allele has been described to alter the endosomal trafficking of cell surface receptors mediating lipid and glucose metabolism [47], leading to a modulation of the metabolic response to diet in APOE-ε4 carriers [48]. Indeed, previous metabolomic studies have shown that energy metabolism pathways regarding fatty acid oxidation, ketone bodies and glucose utilisation are different in APOE-ɛ4 carriers, in comparison with APOE-ɛ3 homozygotes [49, 50]. Two recent analyses conducted in older adults from the French Three-City Study showed that amongst APOE-ε4 carriers, a diet high in refined carbohydrates and low in fibre was associated with higher long-term risk of dementia and Alzheimer’s disease [51]; and that a high glycemic load diet was associated with cognitive decline [52]. All together, these reports underscore the importance of personalised nutrition based on the APOE-ɛ4 gene variant to reduce the risk for cognitive decline and Alzheimer’s disease, as recently suggested [43].

Another potential and complimentary explanation could be related to gut microbiota composition and activity [53–56]. Human and murine evidence have shown a negative association between Ruminococacceae, Prevotellaceae and APOE-ɛ4 status [54, 57, 58]. Interestingly, Ruminococcaceae and Prevotellaceae are bacterial families that metabolise complex polysaccharides into SCFAs [59, 60]. Specifically, a reduced amount of these bacteria has been causally linked to inflammation [61, 62], suggesting that they might contribute to the protective effects of APOE-ɛ2 and -ɛ3 alleles against cognitive decline. This hypothesis is further supported by an increase in Prevotella and other SCFA producing bacteria in asymptomatic, APOE-ɛ4 transgenic mice supplemented with inulin (a type of soluble dietary fibre) [56].

The main strengths of this study are the prospective design and the long follow-up in a well-established cohort of older adults, allowing us to test the influence of clinical, genetic and dietary data. Additionally, we included repeated dietary assessments to reduce bias from measurement errors in dietary questionnaires. Last, we report the association of fibre intake with incident cases of cognitive decline. In the present study, we considered cognitive decline as a longitudinal event, and not as a cognitive outcome. This choice represents an advantage because: (i) cognitive decline trajectories might not be linear in older adults [63], (ii) MMSE scores suffer from ceiling effect leaving clinically significant cases of cognitive decline undetected if using MMSE cut-off scores and (iii) there is a lack of consensus on cut-off values of MMSE scores [64].

Important limitations to this working definition of cognitive decline include the use of a three points difference in consecutive MMSE evaluations to define cognitive decline, and the relatively small sample size. Moreover, the selected population had a mean age of 74 years old; thus, there is a potential for survival bias. Although Cox models were adjusted for age, education and other comorbidities, residual confounding may still remain. Another limitation is that the food database used in the InCHIANTI study did not allow us to differentiate between soluble and insoluble fibre intake. In a recent study, different associations were observed between dietary fibre subtypes and dementia [20]. Last, high levels of dietary fibre may be associated with higher adherence to healthy dietary patterns, such as the Mediterranean diet. In the present study, the additional adjustment for Mediterranean diet did not change the main results (Supplementary Table 2).

In conclusion, our results showed a null association between dietary fibre intake and cognitive decline amongst community-dwelling older adults from a Mediterranean country. Nonetheless, a nutrigenetic interaction between dietary fibre intake and APOE gene variants on the risk for cognitive decline was observed. Higher levels of dietary fibre may reduce the risk for cognitive decline in participants with APOE-ɛ4 haplotype, which genetically determines a higher risk for dementia and impaired cognitive function. Further studies are needed to confirm this distinctive effect of dietary fibre and tailor nutritional recommendations for preventing cognitive decline in older adults.

Supplementary Data

Supplementary data mentioned in the text are available to subscribers in Age and Aging online.

Data Availability

Access to InCHIANTI study data was granted upon request to coordinators of the study (S.B., L.F. and A.C.). Data are available upon reasonable request through the study website.

Declaration of Conflicts of Interest

None.

Declaration of Sources of Funding

Joint Programming Initiative ‘A Healthy Diet for Healthy Life’ (INTIMIC J PI HDHL) Project ‘AC19/00111’, and CIBERFES funded by Instituto de Salud Carlos III and co-funded by European Regional Development Fund ‘A way to make Europe’, Ministry of Economy and Competitiveness (MINECO) (PID2020-114921RB-C21). C.A-L. awarded by grant 2017SGR1546 from the Generalitat de Catalunya’s Agency AGAUR and ICREA Academia 2018. T.M. would like to thank the Ayuda IJCI-2017-32534 financed by the MCIN/AEI/10.13039/501100011033.