Association between tea consumption and cognitive function in cognitively healthy older adults and older adults with mild cognitive impairment

Data source and the participants

The analytical sample was derived from Shanghai Brain Aging Study, a prospective cohort study established in July 2016. Data for the current study are restricted to baseline measures, collected between July 2016 and June 2018 across nine community sites in Shanghai. To be included in the study, participants were required to be 55 years of age or older, and to be living in Shanghai. As illustrated in figure 1, a total of 2026 participants were recruited into the study. Participants were deemed ineligible if they presented with factors that would prevent completion of the cognitive testing protocol, including impairments in vision, hearing and illiteracy. For the current analytical sample, 177 participants were removed due to missing cognitive assessments or diagnosis, gender or age, and/or tea consumption data, which resulted in a total sample of 1849 participants.

Request permission

Ascertainment of cognitive status

Cognitive status was determined by a comprehensive assessment which included a medical history review, a physical examination, cognitive assessment (normed reference scores by sex and education level), self-reported daily activities of living and neuroimaging biomarkers. MCI was diagnosed according to the core diagnostic criteria of the National Institute on Aging and the Alzheimer’s Association14: (1) concern regarding a change in cognition provided by the participant, the participant’s informant or a clinician; (2) objective evidence of impairment in one or more cognitive domains, including memory; (3) preservation of independence in functional abilities; (4) not demented. A total of 816 participants met the criteria for MCI.

Cognitive performance

Participants completed the Montreal Cognitive Assessment (MoCA, a measure of global cognitive function15), the Hopkins Vocabulary Learning Test (HVLT, a measure of verbal episodic memory16 17) and the Digit Span Test (DST, a measure of short-term rote memory (DST forward) and working memory (DST backward)).18 Cognitive assessments were administered by trained psychologists, who received standard training before the study and achieved the same evaluation level.

Tea consumption and potential study covariates

Being a tea drinker (vs non-tea drinker) was operationalised as drinking tea (green, oolong or black) more than three times a week, for at least 6 months. Participants also responded to questions pertaining to demographic characteristics (age, sex, years of education), medical history (diagnosed hypertension, heart disease, hyperlipidemia), smoking history (defined as currently smoking or cumulative smoking over 6 months at any point during the lifespan) and history of alcohol intake (defined as alcohol consumption of more than 20 g/day on average over 6 months).

Statistical analyses

Data distribution met the assumptions of normality. Between-group differences between tea and non-tea consumers were assessed for all demographic and health-related variables for the entire sample (the CH group and the MCI group). Age, sex and education were considered a priori covariates. Other variables with significant differences between tea consumers and non-tea consumers were treated as additional covariates in subsequent analyses (ie, fully adjusted models).

To examine the association between tea consumption and cognitive function, independent analyses of covariance were conducted for each of the cognitive performance scores. Analyses were first conducted for the entire sample and then for each of the cognitive status groups.

All statistical analyses were conducted using IBM SPSS Statistics for Windows, V.23.0. The significance level was set as p<0.05 for two-tailed tests.

Participant characteristics by cognitive status

Within a sample of 1849 participants, 1033 were categorised as CH and 816 were categorised as MCI. Of the study sample, 65.6% was female, with a mean age of 69.50 (8.02) years. As expected, compared with the CH group, the MCI group was significantly older (70.25 (8.48) years vs 68.90 (7.60) years, t=3.567, p<0.001), reported fewer years of education (9.65 (4.04) years vs 11.12 (3.37) years, t=−8.350, p<0.001) and were less likely to be tea consumers (41.3% vs 47.1%, χ²=6.099, p=0.014). Groups did not differ with respect to sex, smoking or alcohol history (p value ranged from 0.081 to 0.794). Compared with CH participants, MCI participants were more likely to report a history of heart disease (22.9% vs 17.2%, χ²=7.967, p=0.005) and were less likely to report a history of hyperlipemia (37.6% vs 45.0%, χ²=8.572, p=0.003). Groups did not differ for self-reported hypertension or diabetes (p value ranged from 0.114 to 0.136). As expected, the MCI group performed more poorly on all cognitive tests compared with the CH group (table 1).

Participant characteristics by tea consumption

Of the 823 participants who were labelled as tea consumers, as displayed in table 2, tea drinkers were more likely to be male, to report more years of education, and to report a history of smoking and alcohol intake. These between-group differences were found in the entire sample and in each of the cognitive status groups. Self-reported heart disease was more common among non-tea drinkers, relative to tea drinkers; however, this difference was restricted to the CH group. Scores of cognitive tests were generally higher among tea drinkers relative to non-tea drinkers (table 2).

Association between tea consumption and cognitive function within the entire sample

Controlling for age, sex and education, analyses of covariance (ANCOVAs) revealed a statistically significant association between tea consumption and HVLT total (F(1, 1844)=10.65, p=0.001, η_p ²=0.006) and delayed recall score (F(1, 1841)=4.36, p=0.037, η_p ²=0.002) within the entire sample. The results did not change in the fully adjusted model for HVLT total or delayed recall score. The association between tea consumption and DST forward was trending in the initial adjusted model (F(1, 1841)=3.26, p=0.071, η_p ²=0.002) and reached statistical significance in the fully adjusted model (F(1, 1540)=4.65, p=0.031, η_p ²=0.003). No significant associations were found for performance on the MoCA or DST backward (table 3).

Association between tea consumption and cognitive function by cognitive status

Controlling for age, sex and education, ANCOVAs revealed significant associations between tea consumption and HVLT total (F(1, 1028)=10.51, p=0.001, η_p ²=0.01) and delayed recall score (F(1, 1026)=6.39, p=0.012, η_p ²=0.006) within the CH group, but not the MCI group (p=0.330 and p=0.928, respectively). The findings did not significantly change in the fully adjusted models. The association between tea consumption and DST forward was trending in the initial adjusted model for the CH group (F(1, 1026)=3.10, p=0.078, η_p ²=0.003), and reached statistical significance in the fully adjusted model (F(1, 1026)=4.47, p=0.035, η_p ²=0.005). No additional associations were found in the CH group or the MCI group (table 3)

Main findings

A body of research suggests that tea consumption is a modifiable lifestyle behaviour that may decrease the risk of cognitive impairment and dementia in later adulthood. Few studies have examined the benefit of tea consumption on specific domains of cognitive function and there is a paucity of research that has examined the relative benefit of tea consumption on cognitive performance among cognitively healthy and cognitively impaired older adults. The current study addressed this gap in knowledge by examining the relative benefit of tea consumption for cognitive function in healthy non-demented older adults and older adults with MCI.

The current study suggests that the benefit of tea consumption is relatively small, and that this benefit is restricted to cognitively healthy older adults, and does not extend to older adults with MCI. This finding is novel because it is the first study to examine the association between tea consumption and cognitive performance in cognitively healthy and cognitively impaired older adults separately. These results further call to question the benefits of tea consumption once cognitive impairment is expressed through standardised neurocognitive testing. However, it is also important to recognise that MCI participants were less likely to be tea consumers relative to the cognitively healthy group.

Similar to findings reported by Feng et al,13 tea consumption among cognitively healthy older adults was associated with better performance on tasks of episodic memory (HVLT) and short-term memory (DST forward). However, in contrast to Feng et al,13 no significant relationship was observed for executive functioning (DST backward) or global cognitive function (MoCA). Conflicting findings may be attributed to the age of the analytical sample, with the current sample being older (on average 69 years of age relative to 64 years of age). Conflicting results may also be attributed to the discrepancy in categorisation of tea consumption. Rather than dichotomising non-tea consumers and tea consumers (ie, consumption of three or more cups of tea per day), Feng et al 13 created an ordinal variable according to frequency of tea drinking, with three or more cups per day as the maximum consumption. Despite differences in statistically significant associations, both studies reported a small effect of tea consumption on tests of memory. Although Feng et al 13 speculated that the small effect was due to sample characteristics, namely that the sample was high functioning, the current study does not support this conclusion.

The current results suggest that tests of short and delayed verbal memory, which are hippocampal dependent, are more sensitive to the benefits of tea consumption. Within the current study, it is plausible that the association between tea consumption and tests of memory were not significant as the hippocampus is a primary structure that is compromised in persons with MCI and dementia.19 While a research has shown that tea consumption can delay the onset of MCI and dementia,11 it may be surmised that once neural pathology is present, the benefit of tea consumption is void. Indeed, in a placebo-controlled study by Ide et al,20 12 months of green tea consumption (relative to placebo powder) failed to improve cognitive function among older adults with cognitive impairment, despite improvements in biomarkers of oxidative stress. While in healthy older adults, a placebo-controlled study by Lianghui et al 21 showed that treatment with Shentai Tea Polyphenol (Jendari) capsules over a 90-day period improved memory, language and executive function, relative to controls.

The current study contributes to the growing body of literature that examines the benefits of tea consumption on cognitive health in late life by examining this association in cognitively healthy and cognitively impaired older adults.

Limitations

Results from this study must be considered in light of several study limitations. The first limitation pertains to quantifying tea consumption and the type of tea consumed. A dichotomous measurement was used, with tea consumption defined as persons who consumed three or more cups of tea per day over the last 6 months. Unfortunately, this study did not quantify intake in millilitres per day of tea or strength of the brew (ie, dilution). A research suggests that tea consumption may have a dose-response effect, with the greatest benefit observed for 500 mL/day of tea consumption.11 Further, while a majority of participants reported the consumption of green tea (65.81%), approximately 27% consumed a combination of green, black and/or oolong tea, which differed in level of fermentation, and their neuroprotective properties. As such, the neuroprotective mechanisms remain unclear. The second study limitation pertains to the cross-sectional evaluation, which precludes causal statements in the association between tea consumption and cognitive functioning.

Implications

This is the first study to examine the association between tea consumption and cognitive performance on tests of memory and executive function in cognitively healthy and older adults with MCI. Although previous researches suggest that tea consumption may delay the onset of MCI and dementia, the current study findings suggest that the benefit of tea consumption may be negated once pathological impairment exists. These findings further suggest that the relative benefit of tea consumption on cognitive function is small and thus must be considered in combination with other lifestyle behaviours, including dietary pattern intake and physical activity, to ensure optimal cognitive function in late life.