Psychoneuroendocrinology 143 (2022) 105840
Available online 16 June 2022
0306-4530/© 2022 Elsevier Ltd. All rights reserved.
The association between Mediterranean diet adherence and allostatic load in older adults
Anik Obomsawin, Danielle D ’ Amico, Alexandra J. Fiocco
*Toronto Metropolitan University, Department of Psychology, Institute for Stress and Wellbeing Research, Canada
A R T I C L E I N F O Keywords:
Allostatic load Dietary pattern Older adults Mediterranean diet Biological sex
A B S T R A C T
Allostatic load (AL) is a multisystemic index of biological wear and tear which is associated with poor health outcomes. In recent years, researchers have examined the association between dietary pattern intake and AL;
however, no studies to date have examined the relationship between AL and consumption of a Mediterranean diet. Blood and urine samples were collected from 201 community-dwelling older adults who completed a Food Frequency Questionnaire (FFQ). A Mediterranean Diet Score (MDS) was calculated based on previous recom- mendations and a sex-based AL index was calculated using a count-based approach for 16 biomarkers associated with neuroendocrine, immune, cardiovascular, or metabolic function. It was hypothesized that a higher MDS would be associated with lower AL, and that this association would be particularly robust for the immune and metabolic subcomponents of the AL index. In support of the primary study hypotheses, generalized linear models revealed a significant inverse relationship between MDS and AL (ß = −0.03, P =0.037). However, sub- components of the AL model were not significantly associated with MDS. Exploratory sub-group analyses by sex suggested that the association between AL and MDS was more robust in male than in female participants. The current findings are interpreted with caution given the study design and sample characteristics. Nonetheless, these findings contribute to the literature supporting the Mediterranean diet as an important lifestyle behavior that may minimize AL, and therefore support healthy aging.
1. Introduction
Allostatic load (AL) is a construct that was first introduced by McE- wen and Stellar in 1993, which represents the biological wear and tear of an organism through repeated allostatic responses that are required to adapt to a changing, or stressful, environment (McEwen and Stellar, 1993). The AL model recognizes the interdependence among biological systems, including the neuroendocrine, cardiovascular, metabolic, and immune systems, which work together to foster the organism’s health and longevity. According to the model, extended activation of primary mediators of the neuroendocrine system exerts their effects on subsidiary systems through the process of allostasis which, over time, result in a shift in cardiovascular, metabolic, and immune parameters. These sub- clinical shifts in secondary outcomes eventually progress to the final stage, referred to as allostatic overload, whereby biological dysregulation across multiple systems result in tertiary, or clinical, outcomes (Juster
et al., 2011).
Allostatic load was first operationalized in the context of aging in the MacArthur Studies of Successful Aging, which provided a count-based index of allostatic load (Seeman et al., 1997). Specifically, the AL index was calculated by summing biological parameters that fell within the “highest risk” quartile of the sample distribution across inter- connected systems (Seeman et al., 1997). Since its conceptualization, a growing body of research has shown that AL increases with age and is associated with poorer cognitive and physical functioning in later adulthood (Crimmins et al., 2003). Furthermore, higher AL in later life has been found to predict greater declines in cognitive and physical functioning and is associated with increased risk of mortality (Crook et al., 2018; Karlamangla et al., 2002; Seeman et al., 1997, 2001, 2004;
Booth et al., 2015; Costa de Robert, 2018; Goldman et al., 2006).
In recent years, researchers have examined the association between dietary pattern intake and AL due to the intricate relationship between
Abbreviations: AL, Allostatic load; CRP, C-reactive protein; DHEA-S, Dehydroepiandrosterone-sulfate; FFQ, Food Frequency Questionnaire; GLM, Generalized linear model; HbA1c, Glycosylated hemoglobin; HDL, High-density lipoprotein; HPA, Hypothalamic-pituitary-adrenal; IGF-1, Insulin-like growrth factor-1; MDS, Meditteranean diet score; MetS, Metabolic syndrome.
* Correspondence to: Department of Psychology, Toronto Metropolitan University, 350 Victoria St., Toronto, ON M5B 2K3, Canada.
E-mail address: [email protected] (A.J. Fiocco).
Contents lists available at ScienceDirect
Psychoneuroendocrinology
journal homepage: www.elsevier.com/locate/psyneuen
https://doi.org/10.1016/j.psyneuen.2022.105840
Received 26 January 2022; Received in revised form 10 April 2022; Accepted 13 June 2022
stress and food intake (e.g., Epel et al., 2004) and the extant literature supporting the role of dietary intake as a lifestyle behavior that de- termines health and wellbeing across the lifespan. Previous research in the Boston Puerto Rican Health Study showed that greater perceived stress among older adults is associated with lower intake of protein, fruits and vegetables; and higher intake of salty snacks and sweats (Laugero et al., 2011). Furthermore, higher cortisol levels, a core component of the AL-neuroendocrine system, was associated with higher intake of saturated fats. A prominent body of literature shows that consuming an unhealthy dietary pattern that is high in red meat, refined grains and sugars; and low in fruits and vegetables, is associated with increased risk for metabolic disorders, cardiovascular disease, cancer, and cognitive impairment in later life (Newby and Tucker, 2004;
Cordain et al., 2005). Conversely, healthy dietary patterns that are rich in fruits and vegetables, fish, whole grains, and legumes are found to associate with more optimal health outcomes (Newby and Tucker, 2004;
Martinez-Lacoba et al., 2018). In particular, the Mediterranean diet has received increasing attention as an important lifestyle behavior in the prevention of poor health outcomes.
The Mediterranean diet is a widely investigated food-based eating pattern that is characterized by high consumption of fruits, vegetables, legumes, nuts, whole grains, olive oil, fish, and seafood; as well as low consumption of high-fat dairy and red meat, and moderate alcohol consumption (F´eart et al., 2013). Research shows that adherence to a Mediterranean diet is associated with reduced risk of depression (S´anchez-Villegas et al., 2009), neurodegenerative disease including Alzheimer’s disease (Sofi et al., 2008), cancer, metabolic disorders, cardiovascular disease, and overall mortality (Sofi et al., 2008, 2013;
Martinez-Lacoba et al., 2018). It is postulated that neuroprotective nu- trients embedded within the Mediterranean diet may influence health outcomes by modulating various systemic biomarkers, including markers of inflammation and metabolic functioning (Kastorini et al., 2011; Stachowicz and Lebiedzinska, 2016), which are notable secondary ´ mediators of the AL model.
To date, a handful of studies have examined the association between dietary pattern intake and AL (Dimitratos et al., 2021; Mattei et al., 2011, 2013; Beydoun et al., 2019; Kusano et al., 2016). Although pre- vious research suggests that dietary pattern intakes may differentially associate with AL, these studies are limited by the use of data-derived dietary patterns, which minimize the translational significance of study findings into real-world public health guidelines. Indeed, there is a paucity of research that has examined the association between dietary intake and AL using translational food-based eating patterns, such as the Mediterranean diet. To this end, the overall objective of the current study was to examine the association between the Mediterranean di- etary pattern and AL among older adults, a growing segment of the population who are at increased risk of poor physical and cognitive health outcomes. More specifically, this study examined: 1) the associ- ation between the Mediterranean diet and AL index, with the hypothesis that a lower Mediterranean diet intake would associate with higher AL;
and 2) the association between Mediterranean diet intake and sub- systems of the AL index, with the hypothesis that the Mediterranean diet would most robustly associate with the immune and metabolic sub- systems of the AL index. This study also explored the association be- tween components of the Mediterranean diet and AL, and whether the association between dietary intake and AL differs by sex.
2. Material and methods 2.1. Participants
As part of a larger study that examined biopsychosocial correlates of cognitive function among older adults, a total of 201 community- dwelling older adults aged 60 years and older were recruited from the Greater Toronto Area in Ontario, Canada using community and online advertisements and the university’s older adult participant pool.
Exclusion criteria entailed having learned English after the age of 12, presenting with uncorrected vision or hearing impairment, and having received a diagnosis of a neurological disorder (e.g., dementia, stroke, Parkinson’s disease), substance abuse/dependency, schizophrenia spectrum disorder, bipolar disorder, or a learning disability. Further, older adults were excluded if they reported a history of chemotherapy or radiation treatment, head injury, or had undergone general anesthesia in the last year. These exclusion criteria were implemented due to their influence on cognitive task performance, which was the primary outcome of interest in the larger study.
2.2. Measures
2.2.1. Demographic and health
Participants completed a sociodemographic and health question- naire which indexed age, biological sex, ethnicity, years of education, perceived socioeconomic status (low, middle, high), retirement status, English as a primary language, smoking status (yes/no) and the presence (yes/no) of medical conditions including diabetes, hypertension, and thyroid disorder. Participants also completed the 10-item Perceived Stress Scale (Cohen et al., 1983), which measures the degree to which situations over the previous month have been appraised as stressful; the Insomnia Severity Index, a 7-item questionnaire that measures sleep difficulty over the previous two weeks, with higher scores indicated greater sleep difficulty (Morin, 1993); and the Beck Depression In- ventory, a 21-item questionnaire that measures characteristic attitudes and symptoms of depression over the previous two weeks (Beck et al., 1961).
2.2.2. Food Frequency Questionnaire
Participants completed the EPIC-Norfolk Food Frequency Question- naire (FFQ), a 130-item semiquantitative questionnaire that records the average frequency in which various food items are consumed during the previous year (Bingham et al., 1997). Frequency of consumption for each food item was ranked on a 9-point scale which ranged from “never or less than once per month” to “more than 6 times per day”. The FFQ has been validated against a 16-day weighted food diary for use in older adults (Bingham et al., 1997).
Mediterranean diet scores (MDS) were calculated based on a method established by Sofi et al. (2013). Using cut-off values based on pre- determined nutritional guidelines, a value of 0, 1, or 2 was assigned for each of the nine components of the Mediterranean diet: 1) vegetables, 2) fruits, 3) legumes, 4) fish and seafood, 5) wholegrains, 6) meat, 7) dairy, 8) alcohol, 9) olive oil. Healthy components (i.e., vegetables, fruits, le- gumes, fish and seafood, wholegrains) were assigned a value of 0 for low intake, 1 for moderate intake, and 2 for high intake. Components that are unhealthy if consumed frequently (e.g., red meat and high fat dairy) were assigned a value of 0 for high intake, 1 for moderate intake, and 2 for low intake. Alcohol consumption was assigned a value of 0 for high intake, 1 for low intake and 2 for moderate intake. Olive oil consumption was assigned a value of 0 for nonconsumption, 1 for low to moderate intake, and 2 for high intake. The total MDS was calculated by summing the scores for each component, with higher scores reflecting greater adherence to the Mediterranean diet. To examine subcomponents of the Mediterranean diet in exploratory analyses, legumes and wholegrains were amalgamated into one component, as were fruits and vegetables, as these foods are commonly grouped by federal dietary recommenda- tions. The remaining components included meats, dairy, fish and sea- food, alcohol, and olive oil. Cut-off values and associated MDS scores are detailed in Supplementary Table 1.
2.2.3. Allostatic load
Total AL was calculated using 16 different biomarkers associated with either metabolic, immune, cardiovascular, or neuroendocrine function. Metabolic biomarkers included 1) serum levels of high-density lipoprotein (HDL) cholesterol and total cholesterol, to assess lipid levels,
2) glycosylated hemoglobin (HbA1c), an integrated measure of glucose metabolism, 3) waist-to-hip ratio, a measure of adipose tissue deposi- tion, 4) insulin, a measure of metabolic regulation, and 5) plasma creatinine, a measure of kidney function, muscle and protein meta- bolism. Immune biomarkers included 1) C-reactive protein (CRP), a systemic marker of inflammation, 2) insulin-like growth factor (IGF-1), a regulator of oxidative stress, 3) fibrinogen, a pro-inflammatory protein involved in blood coagulation, and 4) albumin, a protein with antioxi- dant properties involved in the regulation of intravascular osmotic pressure. Cardiovascular biomarkers included systolic and diastolic blood pressure. Neuroendocrine biomarkers included 1) 24-hour urinary cortisol level, an integrated measure of hypothalamic-pituitary-adrenal (HPA) axis activity, 2) 24-hour urinary epinephrine and norepineph- rine levels, integrated measures of sympathetic nervous system activity, and 3) dehydroepiandrosterone-sulfate (DHEA-S), a functional HPA axis antagonist (See Juster et al., 2010 for review of AL biomarkers). All biomarkers were assayed and analyzed at the CORE laboratory at the St.
Michael’s Hospital.
The AL index was calculated in males and females separately, ac- cording to their respective biomarker distributions (see Supplementary Table 2). Using the count-based calculation method, biomarker values that fell above the 75th percentile of the sample distribution were categorized as 1 and those below the 75th percentile were categorized as 0; except for DHEA-S, HDL cholesterol, IGF-1 and albumin, which were categorized as 1 for values that fell below the 25th percentile and 0 for values that fell above the 25th percentile (Seeman et al., 1997). Sub- system scores (metabolic, immune, cardiovascular, neuroendocrine) were created by summing the relative biomarker scores and a total AL index score was calculated by summing all biomarker scores.
2.2.4. Procedure
Older adults who met eligibility criteria during an initial telephone screening were instructed to fast for 12-h before their morning visit to the CORE laboratory at St. Michael’s Hospital. Following provision of consent, participants underwent blood draw and were then provided with two urinary collection bottles to sample 24-h cortisol and 24-h catecholamine levels. Participants were instructed on how to collect urine samples and were asked to return the samples the following week.
Following blood collection, participants had a light low-glycemic breakfast, provided anthropometric measurements, three blood pres- sure measurements, and completed a testing session at the Stress and Healthy Aging Research Laboratory, which included the completion of a questionnaire battery. The questionnaire battery comprised psychoso- cial questionnaires, a demographic and health survey, and the FFQ. This study was approved by the Ryerson University Research Ethics Board (REB 2014–164).
2.3. Statistical analyses
Analyses were conducted using the R Statistical Package (RStudio, 2016). Analyses were considered statistically significant at P-value <
.05. Goodness-of-fit statistics were used to compare different distribu- tions of the response variable and to select the optimal distribution.
Generalized linear models (GLM, Poisson distribution, log link function) using the “stats” package in R were created to determine the relationship between MDS (independent variable) and AL (dependent variable). To determine the association between MDS and specific subsystems of the AL index, 4 additional models were created for each subsystem (i.e., metabolic, immune, cardiovascular, and neuroendocrine). Following the creation of unadjusted models, sex, age, and years of education were entered as a priori covariates. Bivariate correlations were conducted to identify additional covariates for a fully adjusted model. Factors that displayed a significant Kendall’s tau correlation coefficient (p <.05) were included in the final fully adjusted models. Multicollinearity of predictor variables was assessed using the ‘VIF’ function in R.
To explore whether specific components of the MDS were associated
with total AL scores, a model with scores for all nine dietary components (i.e., vegetables, fruits, legumes, fish and seafood, wholegrains, meat, dairy, alcohol, olive oil) was created. To explore potential sex-specific associations, each of the aforementioned models were stratified by sex. Influential observations were identified by potting Cook’s distance.
Assumptions of linearity and independence of errors were assessed using residual plots and the chi-square test for residual deviance. Please see Supplementary material for R syntax.
3. Results
3.1. Participant demographics
As shown in Table 1, among the 201 participants recruited, 63.2% (n
=127) were female with a mean age of 68.7 ±7.0 years. The majority of the sample was Caucasian (86.0%, n =172) and a majority reported being of middle socio-economic status (69.5%, n =139), with a mean education of 16.7 ±3.3 years. Few participants were diagnosed with diabetes (5.5%, n =11), 20.9% (n =42) reported a diagnosis of hy- pertension, 13.9% (n =28) reported a diagnosis of thyroid disease, and the mean Beck Depression Inventory score was 6.4 ±6.5, with 12.4% (n
=28) meeting the cut-off for depression (score >14). A total of 85.5% (n
=171) reported engaging in regular physical activity, 69.6% (n =119) were non-smokers, and the mean sleep difficulty score was 6.77 ±5.36, indicating low sleep difficulty within the sample.
Within the entire sample, 21 had incomplete biomarker data and 9 had incomplete FFQ responses. Missing FFQ items were imputing using multiple imputation for participants with no more than 10% of items missing. Participants had a mean total AL score of 4.06 ±2.06, which did not differ by sex (female’s mean total AL score ±SD =4.04 ±2.09;
male’s mean total AL score ±SD =4.09 ±2.01; t =0.14, p =0.89).
Participants had a mean total MDS of 10.89 ±2.50, which did not differ by sex (women’s mean total MDS ±SD =11.07 ±2.18; men’s mean total MDS ±SD =10.60 ±2.98; t = − 1.18, p =0.24). Greater MDS was associated with higher energy intake (τ =0.15, p =0.004).
3.2. Assessment of covariates
Bivariate correlations revealed that total AL score was associated with hypertension status (τ =0.17, p =0.008). Bivariate correlations for Table 1
Socioeconomic and health demographics for older adult participants (n =201).
Participant Characteristic Mean±SD (min-max), or n (%)
Age; Mean ±SD (range) years 68.65 ±6.95 (60–95)
Sex; n (%) female 127 (63.2)
Ethnicity; n (%) Caucasian 172 (86.0)
Perceived socio-economic status; n (%)
Low 39 (19.5)
Middle 139 (69.5)
High 22 (11.0)
Education; Mean ±SD (range) years 16.70 ±3.34 (10–29) Physical activity; n (%) regular 171 (85.5) Smoking status, non-smoker; n (%) 119 (69.6) Sleep difficulty score; Mean ±SD (range) 6.77 ±5.36 (0–21) Diabetic status (controlled); n (%) 11 (5.5) Hypertension status (controlled); n (%) 42 (20.9) Thyroid disease status (controlled); n (%) 28 (13.9) Beck Depression Inventory score; Mean ±SD
(range) 6.41 ±6.54 (0–38)
Perceived Stress Scale score; Mean ±SD (range) 12.05 ±6.46 (0–32) Total Mediterranean diet score; Mean ±SD (range) 10.89 ±2.50 (5–17) Total allostatic load score; Mean ±SD (range) 4.06 ±2.06 (0–9)
Note: Sample size for differs for the following variables: Ethnicity = 200;
Physical activity =200; Perceived socio-economic status =200; Smoking status
=171; Sleep quality score =195; Total Mediterranean diet score =192; Total allostatic load score =180; SD =standard deviation
each AL subsystem revealed that the metabolic subsystem AL score was correlated with diagnoses of diabetes (τ =0.16, p =0.012) and hyper- tension (τ =0.17, p =0.009). There were no significant associations between a priori covariates (i.e., age, sex, and years of education) and total AL score or subsystem scores (p-value range= 0.128– 0.997).
Furthermore, no other associations were observed between participant characteristics and AL. See Supplementary Table 3 for the full correla- tion table.
3.3. Association between Mediterranean diet and allostatic load 3.3.1. Total MDS and total AL score
In the unadjusted model, higher MDS was significantly associated with lower total AL (ß = − 0.03, P =0.024). This relationship remained significant after adjusting for sex, age, and years of education (ß =
− 0.04, P = 0.019). After including hypertension status in the fully adjusted model, results did not change significantly (ß = − 0.03, P = 0.037). See Table 2 for the fully adjusted model.
3.3.2. Total MDS and AL subsystem scores
In unadjusted models for each AL subsystem, higher MDS was significantly associated with lower metabolic subsystem score (ß =
− 0.06, P =0.018), but was not associated with immune (ß = − 0.04, P = 0.154), cardiovascular (ß = − 0.01, P =0.768) or neuroendocrine (ß = 0.02, P = 0.623) subsystem scores. Upon adjusting for sex, age, and years of education, results did not change for metabolic (ß = − 0.06, P = 0.013), immune (ß = − 0.04, P =0.140), cardiovascular (ß =0.01, P = 0.778) or neuroendocrine (ß =0.02, P =0.563) subsystem scores. Based on bivariate correlations, a final fully adjusted model was created for the metabolic subsystem analysis, controlling for diagnoses of diabetes and hypertension. The association between MDS and metabolic subsystem score became marginally significant in the fully adjusted model (ß =
− 0.05, P =0.051). See Supplementary Table 4 for the fully adjusted models.
3.4. Association between MDS components and total AL score
In the unadjusted model using scores for all components of the MDS, no specific components were significantly associated with total AL. After adjusting for sex, age, years of education, and hypertension status, there were still no significant relationships between total AL and components of the MDS. See Supplementary Table 5 for fully adjusted models.
3.5. Association between Mediterranean diet and allostatic load stratified by sex
3.5.1. Total MDS and AL total score
In unadjusted models, higher MDS was significantly associated with lower AL in males (ß = − 0.04, P =0.046) but not in females (ß = − 0.03, P =0.253; see Fig. 1). After adjusting for age and years of education, higher MDS was still significantly associated with lower AL in males (ß =
− 0.04, P =0.038) but not in females (ß = − 0.03, P =0.280). After including hypertension status as a covariate, results became marginally significant in males (ß = − 0.04, P =0.054) and did not change in fe- males (ß = − 0.025, P =0.284; see Table 2).
3.5.2. Total MDS and AL subsystem scores
In unadjusted models, higher MDS was associated with lower metabolic subsystem scores in males (ß = − 0.08, P=0.012) but not in females (ß = − 0.03, P=0.432). MDS was not associated with immune, cardiovascular, or neuroendocrine subsystem scores when examined in either sex. Upon adjusting for age and years of education, higher MDS remained significantly associated with lower metabolic subsystem scores in males (ß = − 0.09, P=0.007) but not in females (ß = − 0.02, P=0.449). Upon including diabetes and hypertension diagnoses as covariates in metabolic subsystem analyses, the association between MDS and metabolic subsystem scores remained significant in males (ß =
− 0.08, P=0.016) and did not change substantially in females (ß =
− 0.02, P=0.609). Total MDS was not associated with immune, car- diovascular, or neuroendocrine load scores in either sex in isolation, after adjusting for age and years of education (see Supplementary Table 4).
3.5.3. MDS components and AL scores
In the unadjusted model using scores for all components of the MDS, no components of the MDS were significantly associated with total AL for either males or females. Results did not change significantly after adjusting for age, years of education, and hypertension status (see Supplementary Table 5).
4. Discussion
The AL model is a multisystemic predictor of physical and cognitive- related health outcomes in later adulthood, which may be modulated by dietary intake. The current study presents novel findings suggesting that adherence to a Mediterranean dietary pattern may help minimize the biological wear and tear that stem from allostatic responses accrued over the lifespan. As hypothesized, greater adherence to the Mediterranean diet was significantly associated with lower AL scores, comprised of biomarkers across four interconnected biological systems. Further, this association was not statistically significant for individual subsystem scores of the AL index, emphasizing the utility of taking a multisystemic approach.
The current findings align with the growing body of literature that supports the importance of dietary pattern intake in minimizing AL (Mattei et al., 2011, 2013; Beydoun al, 2019; Dimitratos et al., 2021). In a sample of participants aged 45–75 years from the Boston Puerto Rican Health Study (Mattei et al., 2011), three dietary patterns emerged using factor analysis which comprised of meat, processed meats and French fries (i.e., meats pattern); traditional rice, beans and oils (i.e., traditional pattern); and sweats, sugary beverages and dairy deserts (i.e., sugars Table 2
Association between Mediterranean diet score (MDS) and allostatic load (AL) Total and subsystem scores, within the entire sample and stratified by sex.
Total Allostatic Load
β-estimate Std.
Error z- value p-
value Pseudo- R2
Total Sample 0.256
Intercept 1.518 0.436 3.481 <
0.001 Mediterranean diet
score -0.033 0.016 -2.084 0.037
Age 0.005 0.005 0.915 0.360
Sex -0.028 0.080 -0.351 0.726
Years of education -0.007 0.012 -0.576 0.564 Hypertension status 0.204 0.090 2.268 0.023
Females 0.250
Intercept 1.619 0.593 2.730 0.006
Mediterranean diet
score -0.025 0.024 -1.072 0.284
Age 0.004 0.007 0.578 0.563
Years of education -0.016 0.015 -1.068 0.285 Hypertension status 0.287 0.116 2.473 0.013
Males 0.297
Intercept 1.360 0.655 2.077 0.038
Mediterranean diet
score -0.042 0.022 -1.928 0.054
Age 0.005 0.008 0.587 0.557
Years of education 0.009 0.020 0.434 0.665 Hypertension status 0.082 0.142 0.579 0.562
Note: Fully adjusted generalized linear models. Nagelkerke pseudo-R2 values are specified for each model.
pattern). Dividing each dietary pattern into quintiles, it was found that increasing intake of the meats pattern from the lowest to the highest quintile associated with higher AL, composed of 10 biomarkers across neuroendocrine, cardiovascular and metabolic systems. More recently, in a sample of young and middle-aged adults, Dimitratos et al. (2021) examined sex differences in the association between six dietary pattern clusters derived from the Healthy Eating Index and AL. Sex differences were found for the second dietary pattern cluster (Cluster 2=low adherence to vegetables, greens and beans, fruits, whole grains, fatty acids, added sugars and saturated fat), suggesting that this dietary cluster associated with lower AL in women, but not in men. Notably, red meat and processed foods were not included in the derived clusters, which are important food components to consider in Western pop- ulations that tend to overconsume these food products (Cordain et al., 2005).
Although the aforementioned studies are interesting from an analytical point of view, it may be difficult to derive public messaging based on data-driven clusters which prevent comparisons across studies and minimizes effective knowledge dissemination for the development of guidelines and policy. Conversely, the Mediterranean dietary pattern has been extensively studied and research-informed adherence scores have been developed to facilitate collaborative research efforts and to inform clinical practice at an individual level (Sofi et al., 2013). Indeed, a strength of the current study is the use of well-established cut-off scores in estimating adherence to a Mediterranean dietary pattern. It is also important to note that individual components of the Mediterranean dietary pattern did not significantly associate with AL, suggesting that these foods do not act in isolation and may interact in complex ways to support interconnected biological systems (Tapsell et al., 2016). Future research is encouraged to use these established guidelines in order to foster replication and to draw comparisons across culturally distinct cohorts.
Due to the antioxidant properties of the Mediterranean diet, which reduces postprandial inflammation and cholesterol, and further con- tributes to glucose regulation (Szabo et al., 2021), it was hypothesized that the Mediterranean dietary pattern would most robustly associate with the immune and metabolic subsystems of the AL index. However, this hypothesis was not supported, highlighting the importance of taking a multisystemic approach involving the insidious shift of biological parameters across primary and secondary mediators (Seeman et al., 1997). This finding also supports previous research suggesting that the AL model is distinct from metabolic syndrome (MetS; McCaffery et al.,
2012) and may display greater predictive value in health outcomes compared to MetS (Mattei et al., 2010). Unlike MetS, the AL model addresses whether dietary pattern intake associates with interconnected biological markers that may be within the subclinical range, which is of value in the context of prevention research.
Although not a primary objective of the current study, differential associations by sex were explored and suggest that AL may be more sensitive to dietary intake among males relative to females. Potential sex-specific associations for dietary pattern intake have previously been reported in older adults. In a sample of 1268 community-dwelling older adults, it was found that consumption of a Western dietary pattern, characterized as high in red and processed meats, was associated with poorer global cognition among men, but not women (D’Amico et al., 2020). Differential associations between dietary pattern intake and AL were also reported by Dimitratos et al. (2021). Among females, Cluster 6 (defined as relatively poor adherence to fruits, whole grains, and sodium recommendations) associated with lower AL relative to Cluster 2 (defined as relatively poor adherence to vegetables, fruits, whole grains, fatty acids, added sugars and saturated fat recommendations); however, among males, Cluster 2 associated with lower AL relative to Cluster 6.
Given the relatively small sample size, sex-based outcomes should be interpreted with caution. Further, additional research that examines sex-specific associations is necessary in order to develop appropriate dietary guidelines that align with an individual’s biological milieu.
For over three decades, the Mediterranean dietary pattern has been widely investigated through observational and clinical trial designs, advocating a wide range of health benefits including decreased risk for depression, cardiovascular disease, neurodegenerative disease, and overall mortality, all of which have been found to associate with greater AL (Juster et al., 2010; Marin et al., 2011). It may be surmised that regulation of AL serves as an underlying mechanism in the observed health benefits. Future research is needed to determine whether changes in clinical endpoints following dietary recommendations may be explained by changes across multisystemic parameters of the AL index.
Limitations of the current study include the cross-sectional design, which precludes causal statements in the relationship between dietary intake and AL. Furthermore, retrospective reports of dietary intake may be prone to memory bias and measurement error (Ravelli and Schoeller, 2020). However, the FFQ used in the current study has been validated against a 16-day weighted food diary for use in older adults (Bingham et al., 1997). Sample characteristics of the analytical sample must also be considered. On average, the sample was healthy overall, evidenced Fig. 1.Association between total allostatic load score and Mediterranean diet score by sex.
by a relatively low AL index scores which ranged from zero to nine, out of a possible score of 16. Furthermore, the sample characteristics impede generalizability to non-European, racially and culturally diverse pop- ulations, as well as populations experiencing food insecurity. Indeed, poorer dietary intake is often associated with financial constraints that determine food expenditure following other household costs, such as housing. In a systematic review of 151 studies that examined dietary intake quality and cost of food across socioeconomic status groups from multiple countries, it was found that lower quality diets not only cost less but were more frequently consumed by lower socioeconomic status groups (Darmon and Drewnowski, 2015). As previous research suggests that AL is elevated among persons with a low socioeconomic status relative to a high socioeconomic status (Szanton et al., 2005), additional research is needed to examine the association between dietary pattern intake and AL among culturally diverse older adults who reside in low socioeconomic neighborhoods. Finally, the current findings must be interpreted with caution due to the relatively small sample size.
Accordingly, future research is needed to investigate the association between dietary intake and AL in a larger, more diverse sample in order to better elucidate the potential moderating effects of socioeconomic status, ethnicity, sex, and gender.
5. Conclusion
The Mediterranean diet, comprised of foods that are high in anti- oxidants and anti-inflammatory properties, may minimize dysregula- tion of interconnected biological systems that may develop with aging.
These findings are novel and important; however, additional research is needed to confirm whether the association between the Mediterranean diet and AL is sex-dependent, and whether the relationship between diet and health outcomes is mediated through AL mechanisms.
Conflict of Interest
The authors have no conflicts of interest to disclose.
Acknowledgment
Alexandra J. Fiocco is a member of Team 5 (Diet and Prevention) of the Canadian Consortium on Neurodegeneration in Aging (CCNA). Anik Obomsawin and Danielle D’Amico are trainee members of Team 5 of CCNA. The CCNA is supported by a grant from the Canadian Institutes of Health Research (CIHR) with funding from several partners. This study is partially supported by a Toronto Metropolitan University Harry Rosen Research Grant. Anik Obomsawin and Danielle D’Amico are supported by the Social Sciences and Humanities Research Council of Canada (SSHRC). We would like to thank the participants who contributed their time to this study. We would also like to acknowledge Dr. Vivian Huang, Laura Krieger and the research assistants who contributed to the collection of data.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the online version at doi:10.1016/j.psyneuen.2022.105840.
References
Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for measuring depression. Arch. Gen. Psychiatry 4, 561–571. https://doi.org/10.1001/
archpsyc.1961.01710120031004.
Beydoun, M.A., Nkodo, A., Fanelli-Kuczmarski, M.T., Maldonado, A.I., Beydoun, H.A., Popkin, B.M., Evans, M.K., Zonderman, A.B., 2019. Longitudinal associations between monetary value of the diet, DASH diet score and the allostatic load among Middle-aged urban adults. Nutrients 11 (10), 2360. https://doi.org/10.3390/
nu11102360.
Bingham, S.A., Gill, C., Welch, A., Cassidy, A., Runswick, S.A., Oakes, S., Lubin, R., Thurnham, D.I., Key, T.J., Roe, L., Khaw, K.T., 1997. Validation of dietary
assessment methods in the UK arm of EPIC using weighed records, and 24-hour urinary nitrogen and potassium and serum vitamin C and carotenoids as biomarkers.
Int. J. Epidemiol. 26, S137–S151.
Booth, T., Royle, N.A., Corley, J., Gow, A.J., Hern´andez, M.D.C.V., Maniega, S.M., Ritchie, S.J., Bastin, M.E., Starr, J.M., Wardlaw, J.M., Deary, I.J., 2015. Association of allostatic load with brain structure and cognitive ability in later life. Neurobiol.
Aging 36 (3), 1390–1399.
Cohen, S., Kamarck, T., Mermelstein, R., 1983. A global measure of perceived stress.
J. Health Soc. Behav. 24, 385–396.
Cordain, L., Eaton, S.B., Sebastian, A., Mann, N., Lindeberg, S., Watkins, B.A., O’Keefe, J.
H., Brand-Miller, J., 2005. Origins and evolution of the Western diet: health implications for the 21st century. Am. J. Clin. Nutr. 81 (2), 341–354.
Costa de Robert, S., 2018. Impact of allostatic load on cognitive level, memory and left ventricular mass. Argent. J. Cardiol. 87 (1) https://doi.org/10.7775/rac.v87.
i1.13508.
Crimmins, E.M., Johnston, M., Hayward, M., Seeman, T., 2003. Age differences in allostatic load: an index of physiological dysregulation. Exp. Gerontol. 38 (7), 731–734.
Crook, Z., Booth, T., Cox, S.R., Corley, J., Dykiert, D., Redmond, P., Pattie, A., Taylor, A.
M., Harris, S.E., Starr, J.M., Deary, I.J., 2018. Apolipoprotein E genotype does not moderate the associations of depressive symptoms, neuroticism and allostatic load with cognitive ability and cognitive aging in the Lothian Birth Cohort 1936. PLoS One 13 (2), e0192604.
D’Amico, D., Parrott, M., Greenwood, C., Ferland, G., Gaudreau, P., Belleville, S., Laurin, D., Anderson, N.D., Kergoat, M.-J., Morais, J.A., Presse, N., Fiocco, A.J. 1, 2020. Sex differences in the relationship between diet pattern adherence and cognition among older adults: the NuAge study. Nutr. J. 19, 58. https://doi.org/
10.1186/s12937-020-00575-3.
Darmon, N., Drewnowski, A., 2015. Contribution of food prices and diet cost to socioeconomic disparities in diet quality and health: a systematic review and analysis. Nutr. Rev. 73 (10), 643–660. https://doi.org/10.1093/nutrit/nuv027.
Dimitratos, S.M., Hercules, M., Stephensen, C.B., Cervantes, E., Laugero, K.D., 2021.
Association between physiological stress load and diet quality patterns differs between male and female adults. Physiol. Behav. 240 (5), 113538 https://doi.org/
10.1016/j.physbeh.2021.113538.
Epel, E., Jimenez, S., Brownell, K., Stroud, L., Stoney, C., Niaura, R.A.Y., 2004. Are stress eaters at risk for the metabolic syndrome? Ann. N. Y. Acad. Sci. 1032 (1), 208–210.
F´eart, C., Samieri, C., All`es, B., Barberger-Gateau, P., 2013. Potential benefits of adherence to the Mediterranean diet on cognitive health. Proc. Nutr. Soc. 72 (1), 140–152. https://doi.org/10.1017/S0029665112002959.
Goldman, N., Turra, C.M., Glei, D.A., Lin, Y.H., Weinstein, M., 2006. Physiological dysregulation and changes in health in an older population. Exp. Gerontol. 41 (9), 862–870.
Juster, R.-P., McEwen, B.S., Lupien, S.J., 2010. Allostatic load biomarkers of chronic stress and impact on health and cognition. Neurosci. Biobehav. Rev. 35, 2–16.
https://doi.org/10.1016/j.neubiorev.2009.10.002.
Juster, R.-P., Bizik, G., Picard, M., Arsenault-Lapierre, G., Sindi, S., Trepanier, L., Marin, M.-F., Wan, N., Sekerovic, Z., Lord, C., Fiocco, A.J., Plusquellec, P., McEwen, B.S., Lupien, S.J., 2011. A transdisciplinary perspective of chronic stress in relation to psychopathology throughout life span development. Dev. Psychopathol.
23, 725–776. https://doi.org/10.1017/S0954579411000289.
Karlamangla, A.S., Singer, B.H., McEwen, B.S., Rowe, J.W., Seeman, T.E., 2002.
Allostatic load as a predictor of functional decline: MacArthur studies of successful aging. J. Clin. Epidemiol. 55 (7), 696–710.
Kastorini, C.M., Milionis, H.J., Esposito, K., Giugliano, D., Goudevenos, J.A., Panagiotakos, D.B., 2011. The effect of Mediterranean diet on metabolic syndrome and its components: a meta-analysis of 50 studies and 534,906 individuals. J. Am.
Coll. Cardiol. 57 (11), 1299–1313.
Kusano, Y., Crews, D.E., Iwamoto, A., Sone, Y., Aoyagi, K., Maeda, T., Leahy, R., 2016.
Allostatic load differs by sex and diet, but not age in older Japanese from the Goto Islands. Ann. Hum. Biol. 43, 34–41. https://doi.org/10.3109/
03014460.2015.1013985.
Laugero, K.D., Falcon, L.M., Tucker, K.L., 2011. Relationship between perceived stress and dietary and activity patterns in older adults participating in the Boston Puerto Rican Health Study. Appetite 56 (1), 194–204. https://doi.org/10.1016/j.
appet.2010.11.001.
Marin, M., Lord, C., Andrews, J., Juster, R., Sindi, S., Arsenault-Lapierre, G., Fiocco, A.J., Lupien, S.J., 2011. Chronic stress, cognitive functioning and mental health.
Neurobiol. Learn. Mem. 96, 583–595. https://doi.org/10.1016/j.nlm.2011.02.016.
Martinez-Lacoba, R., Pardo-Garcia, I., Amo-Saus, E., Escribano-Sotos, F., 2018.
Mediterranean diet and health outcomes: a systematic meta-review. Eur. J. Public Health 28 (5), 955–961.
Mattei, J., Noel, S.E., Tucker, K.L., 2011. A meat, processed meat, and french fries dietary pattern is associated with high allostatic load in Puerto Rican older adults. J. Am.
Diet Assoc. 111 (10), 1498–1506. https://doi.org/10.1016/j.jada.2011.07.006.
Mattei, J., Bhupathiraju, S., Tucker, K.L., 2013. Higher adherence to a diet score based on American Heart Association recommendations is associated with lower odds of allostatic load And metabolic syndrome in Puerto Rican adults. J. Nutr. 143, 1753–1759.
Mattei, J., Demissie, S., Falcon, L.M., Ordovas, J.M., Tucker, K., 2010. Allostatic load is associated with chronic conditions in the Boston Puerto Rican Health Study. Soc. Sci.
Med. 70 (12), 1988–1996. https://doi.org/10.1016/j.socscimed.2010.02.024.
McCaffery, J.M., Marsland, A.L., Strohacker, K., Muldoon, M.F., Manuck, S.B., 2012.
Factor structure underlying components of allostatic load. PLoS One 7 (10), e47246.
https://doi.org/10.1371/journal.pone.0047246.
Morin, C.M., 1993. Insomnia: Psychological Assessment and Management. Guilford Press, New York.
Newby, P.K., Tucker, K.L., 2004. Empirically derived eating patterns using factor or cluster analysis: a review. Nutr. Rev. 62 (5), 177–203.
Ravelli, M.N., Schoeller, D.A., 2020. Traditional self-reported dietary instrucments are prone to inaccuracies and new approaches are needed. Front. Nutr. 7 https://doi.
org/10.1016/j.neubiorev.2009.10.002, 90-16.
RStudio Team (2016). RStudio: Integrated Development for R. RStudio, Inc., Boston, MA URL http://www.rstudio.com/.
S´anchez-Villegas, A., Delgado-Rodríguez, M., Alonso, A., Schlatter, J., Lahortiga, F., Majem, L.S., Martínez-Gonz´alez, M.A., 2009. Association of the Mediterranean dietary pattern with the incidence of depression: the Seguimiento Universidad de Navarra/University of Navarra follow-up (SUN) cohort. Arch. Gen. Psychiatry 66 (10), 1090–1098.
Seeman, T.E., Singer, B.H., Rowe, J.W., Horwitz, R.I., McEwen, B.S., 1997. Price of adaptation—allostatic load and its health consequences: MacArthur studies of successful aging. Arch. Intern. Med. 157 (19), 2259–2268.
Seeman, T.E., Crimmins, E., Huang, M.H., Singer, B., Bucur, A., Gruenewald, T., Reuben, D.B., 2004. Cumulative biological risk and socio-economic differences in mortality: MacArthur studies of successful aging. Soc. Sci. Med. 58 (10), 1985–1997.
Seeman, T.E., McEwen, B.S., Rowe, J.W., & Singer, B.H. , 2001. Allostatic load as a marker of cumulative biological risk: MacArthur studies of successful aging. Proc.
Natl. Acad. Sci. USA, 98(8), 4770–4775.
Sofi, F., Cesari, F., Abbate, R., Gensini, G.F., Casini, A., 2008. Adherence to Mediterranean diet and health status: meta-analysis. Br. Med. J. 337, 337.
Sofi, F., Macchi, C., Abbate, R., Gensini, G.F., Casini, A., 2013. Mediterranean diet and health status: an updated meta-analysis and a proposal for a literature-based adherence score. Public Health Nutr. 17 (12), 2769–2782.
Stachowicz, M., Lebiedzi´nska, A., 2016. The effects of diet components on the level of cortisol. Eur. Food Res. Technol. 242, 2001–2009. https://doi.org/10.1007/s00217- 016-2772-3.
Szabo, Z., Koczka, V., Marosvolgyi, T., Szabo, E., Frank, E., Polyak, E., Fekete, K., Erdelyi, A., Verzar, Z., Figler, M., 2021. Possible biochemical processes underlying the positive health effects of plant-based diets – a narrative review. Nutrients 13, 2593. https://doi.org/10.3390/nu13082593.
Szanton, S.L, Gill, J.M., Allen, J.K., 2005. Allostatic Load: A Mechanism of Socioeconomic Health Disparities? Biological Research For Nursing 7 (1), 7–15.
https://doi.org/10.1177/1099800405278216.
Tapsell, L.C., Neale, E.P., Satija, A., Hu, F.B., 2016. Foods, nutrients, and dietary patterns: interconnections and implications for dietary guidelines. Adv. Nutr. 7 (3), 445–454. https://doi.org/10.3945/an.115.011718.