Moderators of the Impact of (Poly)Phenols Interventions on Psychomotor Functions and BDNF: Insights from Subgroup Analysis and Meta-Regression

d: Recent anti-aging interventions have shown contradictory impacts of (poly)phenols regarding the prevention of cognitive decline and maintenance of brain function. These discrepancies have been linked to between-study di erences in supplementation protocols. This subgroup analysis and meta-regression aimed to (i) examine di erential e ects of moderator variables related to participant characteristics and supplementation protocols and (ii) identify practical recommendations to

anti-aging interventions have shown contradictory impacts of (poly)phenols regarding the prevention of cognitive decline and maintenance of brain function. These discrepancies have been linked to between-study di erences in supplementation protocols. This subgroup analysis and meta-regression aimed to (i) examine di erential e ects of moderator variables related to participant characteristics and supplementation protocols and (ii) identify practical recommendations to design e ective (poly)phenol supplementation protocols for future anti-aging interventions. Methods: Multiple electronic databases (Web of Science; PubMed) searched for relevant intervention published from inception to July 2019. Using the PICOS criteria, a total of 4303 records were screened. Only high-quality studies (n=15) were included in the nal analyses. Random-e ects meta-analysis was used, and we calculated standard di erences in means (SDM), e ect size (ES), and 95% con dence intervals (CI) for two su ciently comparable items (i.e., psychomotor function and brain-derived neurotrophic factor (BDNF)). When signi cant heterogeneity was computed (I 2 >50%), a subgroup and meta-regression analysis were performed to examine the moderation e ects of Nutrients2020,12, 2872; doi:10.3390 /nu12092872 /journal/nutrients

Nutrients2020,12, 2872 2 of 23 participant characteristics and supplementation protocols. Results: The reviewed studies support the bene cial e ect of (poly)phenols-rich supplementation on psychomotor functions (ES= 0.677, p=0.001) and brain plasticity (ES=1.168,p=0.028). Subgroup analysis revealed higher bene cial impacts of (poly)phenols (i) in younger populations compared to older (SDM= 0.89 vs. 0.47 for psychomotor performance, and 2.41 vs. 0.07 for BDNF, respectively), (ii) following an acute compared to chronic supplementation (SDM= 1.02 vs. 0.43 for psychomotor performance), and (iii) using a phenolic compound with medium compared to low bioavailability rates (SDM= 0.76 vs. 0.68 for psychomotor performance and 3.57 vs. 0.07 for DBNF, respectively). Meta-regressions revealed greater improvement in BDNF levels with lower percentages of female participants (Q=40.15, df=6, p<0.001) and a skewed scatter plot toward a greater impact using higher (poly)phenols doses. Conclusion: This review suggests that age group, gender, the used phenolic compounds, their human bioavailability rate, and the supplementation dose as the primary moderator variables relating to the bene cial e ects of (poly)phenol consumption on cognitive and brain function in humans. Therefore, it seems more advantageous to start anti-aging (poly)phenol interventions in adults earlier in life using medium ( 500 mg) to high doses ( 1000 mg) of phenolic compounds, with at least medium bioavailability rate ( 9%). Keywords:antioxidant; aging; psychomotor functions; brain plasticity; brain functions; cognition 1. Introduction Aging, a complex biological process, is inescapably connected with age-related health decline, a ecting several aspects of cognitive functioning [1]. A primary factor associated with age-related cognitive decline is the development of neuroin ammation [2]. This nervous tissue in ammation may be initiated by a variety of cues, such as traumatic brain injury [3] infection, autoimmunity [4] toxic metabolites, or a disequilibrium redox state in favor of prooxidants (e.g., reactive oxygen species (ROS)) [5,6]. Regarding this, increased lipid peroxidation and protein oxidation in aged hippocampus and the cerebral cortex have been shown to increase the susceptibility of neurons to apoptosis via the generated reactive protein oxidation and lipid peroxidation products (e.g., 4-hydroxy-2-nonenal) [6,7]. In a state of chronic oxidative stress, ROS-induced intracellular signaling pathways are altered,

of prooxidants (e.g., reactive oxygen species (ROS)) [5,6]. Regarding this, increased lipid peroxidation and protein oxidation in aged hippocampus and the cerebral cortex have been shown to increase the susceptibility of neurons to apoptosis via the generated reactive protein oxidation and lipid peroxidation products (e.g., 4-hydroxy-2-nonenal) [6,7]. In a state of chronic oxidative stress, ROS-induced intracellular signaling pathways are altered, leading to dysregulation of the in ammatory response [8]. This loss in the regulation of signal transduction by the cells is accompanied by an increased production of damage-associated molecular patterns and an increased secretion of proin ammatory molecules, which act together to promote neuroin ammation and may play an important role in neuron dysfunction and ultimately cognitive decline [2,8,9]. Given the decreased activity of many endogenous antioxidant enzymes, such as superoxide dismutase, glutathione peroxidase, and catalase in aged hippocampus and the cerebral cortex [10,11], a huge array of existing literature has highlighted the importance of exogenous antioxidants as potential anti-aging agents [12–16]. The most widely known exogenous antioxidants are carotenoids (lycopene, lutein, zeaxanthin, - and -carotene, -cryptoxanthin), vitamin E ( - and-tocopherol), vitamin C, and (poly)phenols [17], with the latter exhibiting the highest antioxidant capabilities [18]. Considering the increased e ectiveness of (poly)phenols in counteracting age-related oxidative stress, recent human studies have examined the role of natural (poly)phenols-rich products in the prevention of cognitive decline and maintenance of brain function [15,16]. Several studies have shown that the consumption of (poly)phenols-rich supplementation can bene t cognitive decline in older adults [19–21], as well as in young- and middle-aged populations [22–24]. Additionally, polyphenols bind to nuclear estrogen receptor (ER ) and (ER ), thus inducing neuroprotective e ects. These e ects are particularly noticeable in human cells, where they mimic or inhibit actions of endogenous estrogens [25]. Contrarywise, other large-scale clinical trials investigating similar populations have failed to produce similar results [26–29].

Nutrients2020,12, 2872 3 of 23 Discrepancies between these ndings have been suggested to be linked to the between-studies di erences in supplementation protocols, amongst others [15]. Speci cally, a recent systematic review and meta-analysis by our research group underlined the speci city of phenolic compounds, (poly)phenol dose, and bioavailability as the primary determinants of the e cacy of (poly)phenols-rich supplementation in anti-aging interventions [15,16]. The results of this publication suggested that an intermediate dose of (poly)phenols with intermediate to high human rates of bioavailability are necessary in order for them to successfully cross the blood–brain barrier, exerting signi cant e ects on cognitive function and brain health [15]. Furthermore, the application of anti-aging interventions to young, healthy individuals has theoretical advantages to reduce the onset of brain-related aging processes [30,31]. Our research group additionally underlined age as an important moderator in (poly)phenol e cacy in anti-aging interventions [16]. Another recent systematic review focusing on young- and middle-aged populations suggested that only a low- to medium-dose of phenolic components is necessary at younger ages to elicit promising e ects on brain health [16]. Although the aforementioned suggestions in our previous reports [15,16] are of importance to aid in designing future anti-aging (poly)phenol interventions, the suggested phenolic doses, bioavailability rate, and target age group eventually need to be con rmed by a subgroup meta-analysis of comparable data from age-group studies. Moreover, given the large range of phenolic compounds (i.e., avanols, anthocyanidins, avones, iso avones, avonols, and avanones/ avanonols) and their natural fruits and vegetables sources (e.g., celery, onions, oregano herbs, broccoli, green tea, dark chocolate, red wine, soy, citrus fruit, leeks, berry fruits, and parsley) [13,32], it is also important to identify the most e ective phenolic compounds during anti-aging preventive interventions. Finally, there is extensive literature documenting the moderating e ects of gender on cognition and brain aging process (i.e., rate of blood ow, pattern of glucose metabolism, and receptors activity), with some evidence suggesting that women show less age-associated cognitive decline, while men undergo more progressive decreases in frontotemporal brain volume [33]. Therefore, performing a gender subgroup analysis of data

Finally, there is extensive literature documenting the moderating e ects of gender on cognition and brain aging process (i.e., rate of blood ow, pattern of glucose metabolism, and receptors activity), with some evidence suggesting that women show less age-associated cognitive decline, while men undergo more progressive decreases in frontotemporal brain volume [33]. Therefore, performing a gender subgroup analysis of data from anti-aging (poly)phenol intervention is warranted. To overcome the gaps in recently published meta-analyses, this review was designed to (i) examine di erential e ects of moderator variables related to participants' characteristics (i.e., gender/age) and the supplementation protocol (i.e., nature, dose, phenolic compound's bioavailability, and the duration of the intervention period) and (ii) identify practical recommendations to design e ective (poly)phenol supplementation protocols for future anti-aging interventions based on optimal phenolics' compounds, dose, bioavailability rate, optimal intervention period, and optimal target population. 2. Materials and Methods The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines was followed in the present systematic review [34]. 2.1. Search Stratgies and Sources of Data Up to July 2019, an electronic comprehensive systematic search was performed using electronic databases (PubMed and Web of Science), without applying any time limits. The search was limited to only English publications. The search strategy as well as the search terms were similar to the ones utilized by Ammar et al. [15,16]. Additionally, in order to minimize the risk of missing relevant publications, the reference lists of included manuscripts and similar journal citations identi ed from Google Scholar were reviewed. Two independent researchers considered each of the articles for their inclusion appropriateness. A discussion with a third researcher was performed in cases of uncertainty to determine the nal inclusion or exclusion of the paper. Further information on the search process and inclusion and exclusion criteria are presented in Table.

Nutrients2020,12, 2872 4 of 23 Table 1. A summary of the search strategy and the inclusion and exclusion criteria adopted in the present systematic review and meta-analysis. Search Strategy Item Search Strategy Details String of keywords ((polyphenol) OR ( avonoids) OR (polyphenolic compounds) OR (iso avone) OR ( avanol) OR (phytoestrogen) OR (resveratrol)) AND ((cognitive performance) OR (cognitive) OR (cognitive function) OR (cognition) OR (brain function) OR (executive function) OR (attention) OR (working memory) OR (brain imaging) OR (neuroimaging) OR (neural) OR (magnetic resonance imaging) OR (MRI) OR (fmri) OR (grey matter) OR (gray matter) OR (brain volume) OR (brain structure) OR (electrophysiology) OR (EEG) OR (event related potential) OR (neuroblast) OR (neuroblast) OR (cerebral blood ow) OR (CBF) OR (regional perfusion) OR (brain-derived neurotrophic factor) OR (BDNF) OR (cerebrovascular responsiveness) OR (CVR) OR (pulsatility index) OR (transcranial doppler) OR (TCD) OR (near-infrared spectroscopy) OR (NIRS) OR (cerebral hemodynamics) OR (total hemoglobin) OR (total-Hb) OR (oxygenated hemoglobin) OR (oxy-Hb) OR (deoxygenated 2 hemoglobin) OR (deoxy-Hb)) NOT ((mice) OR (animals) OR (Parkinson's) OR (stroke) OR (Alzheimer's) OR (dementia) OR (cancer) OR (lesions) OR (diabetes) OR (injury) OR (patients) OR (rats) OR (disease) OR (impairment)] Searched databases Web of Science and PubMed; up to July 2019 Inclusion criteria (i) English language published primary research (up to July 2019), (ii) research in healthy adult humans, (iii) original investigations researching e ects of (poly)phenol-rich supplementation on brain health, (iv) no major methodological issues (i.e., lack of a comparative control, not blinded, or inappropriate/ incorrect statistical analyses) Exclusion criteria (i) studies written in any non-English language, (ii) congress, meeting, conference, or workshop publications, (iii) studies conducted in diseased individuals or a individuals greater than 55 years of age and (iv) studies that did not include supplementation. Findings from sources such as encyclopedias, reviews, case studies, or book chapters were not included. Time lter None applied (search from inception) Language lter English PICOS Participants: healthy adults (>18 years of age) Intervention: chronic and/or acute (poly)phenols-rich supplementation Comparative: Any Outcome: cognitive function (e.g., neuroplasticity, overall cognition, executive function, processing speed, verbal memory, language psychomotor performance, visual memory,

supplementation. Findings from sources such as encyclopedias, reviews, case studies, or book chapters were not included. Time lter None applied (search from inception) Language lter English PICOS Participants: healthy adults (>18 years of age) Intervention: chronic and/or acute (poly)phenols-rich supplementation Comparative: Any Outcome: cognitive function (e.g., neuroplasticity, overall cognition, executive function, processing speed, verbal memory, language psychomotor performance, visual memory, attention) and brain activity, neuroprotective measures (e.g., brain perfusion, cerebral blood ow (CBF), cerebral hemodynamics, and neuroin ammation) Study design: controlled clinical trial

Nutrients2020,12, 2872 5 of 23 2.2. Study Selection The process used for selecting articles is outlined in Figure. EndNote X8 (produced by Clarivate Analytics, Philadelphia, USA) was used to remove duplicate articles in the initial search results. Following the removal of duplicates, the titles and abstracts of all unique hits were screened by two authors for eligibility, and disagreements were resolved by consensus. Then, inclusion and exclusion of the article list was conducted via full text screening in accordance to the Participants, Intervention, Comparative, Outcome and Study design (PICOS); inclusion and exclusion criteria are detailed in Table. Reasons for article exclusion were recorded during this process.Nutrients 2020, 12, x FOR PEER REVIEW 6 of 24 to examine publication bias. Sensitivity analyses and cumulative meta-analysis were also conducted to assess the stability and the reliability of the findings. Statistical significance was set at p < 0.05. 3. Results 3.1. Study Selection The predefined search strategies yielded a preliminary pool of 4303 possible papers. A totalof 1688 duplicates and 2385 non-clinical trials were removed. Then, 230 published papers were screened by titles and abstracts for eligibility, and 38 published studies met the inclusion criteria. After a careful review of the 38 full texts, 15 articles included enough comparable data (based on the used cognitive test and the tested brain parameters) to be used in the present subgroup meta-analysis (Figure 1). Figure 1. Flow diagram of the literature selection process. 3.2. Study Characteristics Figure 1.Flow diagram of the literature selection process.

Nutrients2020,12, 2872 6 of 23 2.3. Data Collection Using a pilot-tested extraction form, two authors independently collected data, and disagreements were resolved via mutual consensus. Data that were extracted included (i) characteristics of participants (i.e., sex, age, participant numbers), (ii) supplementation protocol (i.e., nature, dose and bioavailability of the used phenolic compounds, and the duration of the intervention period), and (iii) key outcomes from anti-aging based-(poly)phenols intervention on psychomotor performance (i.e., evaluated during Trail Making Test (TMTa) and Reaction Time Test (RTT)) and brain plasticity (i.e., evaluated using the brain-derived neurotrophic factor (BDNF) measurements). TMT: Traditionally, trail making tests re ect a multiple of cognitive processes such as shifting, attention, sequencing, visual search and scanning, psychomotor speed, exibility, abstraction, ability to execute and adjust an action plan, and the capability to sustain two simultaneous trains of thought [30]. TMTa includes 25 numbered circles (1–25) distributed throughout a sheet of paper., Then, the patient draws lines connecting each number in ascending order, as quickly as possible, without ever lifting the writing instrument [35]. RTT Example: A white square appears 30 times, at random intervals, in a computer screen's center. A “yes” button is pressed as soon as the stimulus is visible [23]. TMTa and RTT scores are calculated as the time required to complete the task in seconds; higher scores indicate increased levels of impairment [20,35]. 2.4. Quality Assessment Methodological study quality was assessed via the Physiotherapy Evidence Database (PEDro) scale [15,16,36]. PEDro is a reliable and objective tool based on the Delphi list developed by Verhagen et al. [37]. To identify which randomized controlled trials are externally (criteria 1) and internally (criteria 2–9) valid with interpretable results, each paper was independently assessed twice by two independent authors using the 11-item checklist to yield a maximum score of 10 (the sum of awarded points for criteria 2–11). Points are only awarded when a criterion is clearly satis ed. A score of 9–10 on the PEDro scale were considered to be of “high quality”, scores of 5–8 were considered to be of “moderate quality”, and studies that scored below 5 were considered

checklist to yield a maximum score of 10 (the sum of awarded points for criteria 2–11). Points are only awarded when a criterion is clearly satis ed. A score of 9–10 on the PEDro scale were considered to be of “high quality”, scores of 5–8 were considered to be of “moderate quality”, and studies that scored below 5 were considered to be of “low quality” [15,16,38]. 2.5. Statistical Analysis The software “Comprehensive Meta-Analysis” (CMA for Windows, version 3, Biostat, Englewood, NJ, USA) was used for the meta-analysis (MA). Given the level of cognitive task variability and brain measurement techniques between included studies, only psychomotor performance (TMTa or RTT) and BDNF were su ciently comparable and thus included in this MA. To calculate e ect size, performance in psychomotor tests was recorded in seconds (s), and BDNF concentrations were collated using ng/mL. For those studies where net changes in the groups were not directly reported, e ect size (ES) was computed through subtraction of the values at the intervention endpoint from baseline. Calculations of standard deviations of mean di erences were completed via (SD=square root ((SD pre-treatment) 2 +(SD post-treatment) 2 (2R SD pre-treatment SD post-treatment))); the correlation coe cient (R) was assumed to be 0.5 [39,40]. ESs as well as their 95% con dence intervals (CI) were calculated via Cohen's method, re ecting the standardized di erence in means (SDM) between measured parameters (i.e., psychomotor performance and BDNF), in response to (poly)phenols-rich supplementation and placebo. ES were determined to be trivial (ES<0.2), small (ES of 0.2 to<0.6), moderate (ES of 0.6 to <1.2), large (ES of 1.2 to<2.0), very large (ES 2.0), and extremely large (ES>4.0) [41]. Forest plots were utilized to illustrate point estimates of the e ect size and 95% con dence intervals. A positive ES value in BDNF and a negative ES value in the psychomotor test (i.e., time to completion in seconds) indicated that (poly)phenols-rich supplementation enhanced outcomes. For the forest plots, each individual study is represented through a black square, and the overall e ect is represented by a red diamond. Statistical heterogeneity was calculated

and 95% con dence intervals. A positive ES value in BDNF and a negative ES value in the psychomotor test (i.e., time to completion in seconds) indicated that (poly)phenols-rich supplementation enhanced outcomes. For the forest plots, each individual study is represented through a black square, and the overall e ect is represented by a red diamond. Statistical heterogeneity was calculated by computing Q [42] andI 2 [43]. When substantial

The study analyzes the impact of (poly)phenols on psychomotor functions and BDNF through subgroup analysis and meta-regression.