Association Between Dietary Theobromine and Cognitive Function in a Representative Elderly American Population: A Cross-Sectional Study

Background: Despite reports on neuroprotective effects of dietary theobromine intake, whether dietary theobromine can exert benecial effects on cognitive function is unclear. We aimed to investigate the association between dietary theobromine and cognitive function in old American population. Methods: We collected data from the 2011-2012 and 2013-2014 cycles of the National Health and Nutrition Examination Survey, a cross-sectional survey. Daily dietary theobromine was treated as a continuous variable and a log transform. Cognitive function was measured by four tests: Consortium to Establish a Registry for Alzheimer's Disease (CERAD) Word Learning tests, CERAD delayed recall test, animal uency test, and digit symbol substitution test (DSST). We conducted linear regression analyses and subgroup analyses to study the association between theobromine intake and cognitive performance. Basic characteristics, lifestyle factors, disease history, and nutritional intake were adjusted in these models. Results: A total of 2,845 participants were included in this study. Daily theobromine intake was not signicantly different between the 2011-2012 and 2013-2014 cycles. The CERAD-immediate and delayed recall scores were signicantly different between these two cycles, but not the animal uency score or digital symbol score. The daily dietary theobromine intake in log form was positively associated with immediate recall score (β, 95% CI, P: 0.661, 0.222-1.101, <0.01), delayed recall score (β, 95% CI, P: 0.232, 0.016-0.449, 0.04), and DSST score (β, 95% CI, P: 1.395, 0.140-2.649, 0.03) in the fully adjusted model, but not with the animal uency score (β, 95% CI: 0.001, -0.122-0.907, 0.13). Sensitive analyses showed that L-theobromine intake was linearly associated with cognitive performance. Conclusions: Daily theobromine intake was associated with cognitive performance in a large population. However, further interviews. Trained interviewers used an automated data collection system under the guidance of an examination protocol (11). Based on previous studies, we also collected data on nutrients reported to be associated with cognitive function, including total energy intake (12), protein (13, 14), lutein (15), zeaxanthin (15), folic acid (16–18), vitamin B12 (17, 19), added vitamin B12 (17, 19), vitamin D (20, 21), magnesium (22), iron (23), zinc (23), copper (23), selenium (23), alcohol, and caffeine (24, 25). Dietary theobromine and other potential confounding dietary nutrients were treated as continuous variables. DSST an of execution a representative old American signicantly with the CERAD-immediate/delayed recall score uency


Background
Age-related cognitive decline, characterized by impairment of episodic memory, working memory, and attention, can affect the quality of life (1).
Nutritional conditions are reported to be involved in this degenerative cognitive impairment (2,3). Since previous studies have reported a protective effect of chocolate (4,5), and theobromine is one of its main active components (6), it could also be associated with cognitive function.
In animal experiments, dietary theobromine is suggested to exert cognitive protection. Theobromine intake is reported to be capable of crossing the blood-brain barrier in mice (7), which provide theobromine chance to interact directly with brain tissue. Further, theobromine is estimated to exert protective effects through neurotransmitter regulation. For example, Mendiola-Precoma et al. (8) found that theobromine intake could improve A1 receptor expression. Theobromine might play a role in phosphodiesterase inhibitors to enhance motor learning skills (9).
Despite this evidence, few clinical trials have investigated the association between dietary theobromine and cognitive function. Therefore, we included a large, representative sample of ≥ 60-year-old American participants from the cross-sectional National Health and Nutrition Examination Survey (NHANES) dataset to study the association between daily theobromine intake and cognitive performance. We hypothesized that daily theobromine intake would be positively associated with cognitive performance.

Study population
The NHANES survey was a complex strati ed, multistage sampling designed cross-section survey conducted by the Centers for Disease Control and Prevention of USA to assess the American health and nutritional status (10). A survey cycle has been running every 2 years since 1999. In this study, we collected data from the 2011-2012 and 2013-2014 cycles. Participants > 60 years were eligible for cognition test. A total of 3,632 participants were eligible for the cognitive function questionnaire. A total of 508 participants did not answer the cognitive function questionnaire; 269 participants were not available for dietary theobromine intake data. Finally, 2,854 participants were analyzed (Fig. 1). Ethics review board of National Center for Health Statistics approved survey protocol of NHANES, including ethic protocol.

Cognitive function
In the 2011-2014 NHANES cycles, survey participants aged ≥ 60 years and quali ed to understand English, Spanish, Korean, Vietnamese, traditional or simpli ed Mandarin, or Cantonese were eligible for cognitive function examination. Three cognition tests were employed: the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) word learning and recall test, the animal uency test, and digit symbol substitution test (DSST). These tests have validated reliability in evaluating cognitive function in Americans (26)(27)(28). Participants completed a cognitive test using online tests. The CERAD word learning and recall tests assess immediate and delayed learning ability for new verbal information, respectively (28). The animal uency test evaluates a component of executive function, categorical verbal uency (29). DSST is a performance module from the Wechsler Adult Intelligence Scale that evaluates processing speed, sustained attention, and working memory (30).
Bad cognitive performance was de ned as the lowest quartile of these four scores (31).

Covariates
Demographic variables, including age, sex, race, education, marital status, home status, employment status, smoking status, body mass index (BMI), history of disease, hypertension, diabetes, sleep disorder, and depression were also noted. Age was treated as a continuous variable and split into three subgroups (> 60, ≤ 70; <70, ≥ 80; <80) in strati ed analyses. Race was categorized as Mexican American, other Hispanic, non-Hispanic white, non-Hispanic black, and other races including multi-racial. Education was classi ed as less than 9th grade, 9-11th grade (including 12th grade with no diploma), high school graduate or equivalent, some college degree, and college graduate or above. Marriage was categorized as married/living with partner, widowed/divorced/separated, and never married. Home status refers to whether participants owned their living house or other arrangements. Smoking status was de ned as never ("never smoked or smoked < 100 cigarettes in their life"), previous smoker ("smoked ≥ 100 cigarettes in their life and currently no longer smoking"), and current smoker ("smoked ≥ 100 cigarettes in their life and currently smoking"). BMI was classi ed as underweight (< 18.9 kg/m 2 ), healthy or normal weight (≥ 18.9, < 25 kg/m 2 ), overweight (≥ 25, < 30 kg/m 2 ), or obese (≥ 30 kg/m 2 ). If participants were diagnosed with asthma, anemia, psoriasis, celiac, arthritis, emphysema, liver condition, chronic bronchitis, cancer, or malignancy, the history of disease was binary recorded as yes, or no. Heart disease was positively recorded with previous congestive heart failure, coronary heart disease, angina pectoris, and heart attack medically con rmed. Hypertension, diabetes, and sleep disorders were de ned by a previous medical diagnosis. Depression was de ned as a patient health questionnaire-9 (PHQ-9) score > 5(32).  Table 1, the distribution of daily theobromine was not normal. Thus, we applied a log transformation in later association analysis. We applied linear regression analyses to study the association between theobromine intake and cognitive scores and bad cognitive performance, adjusting for age, sex, race, education, marital status, home status, employment, smoking status, BMI, history of disease, heart disease, history of stroke, hypertension, diabetes, depression, sleep disorder, and dietary nutrients (including energy, protein, lutein, zeaxanthin, folic acid, vitamin B12, vitamin B12, vitamin D, magnesium, iron, zinc, copper, selenium, and caffeine). Daily theobromine intake was analyzed as a continuous variable and a log transform was employed in these analyses. We also conducted subgroup analysis to study the association between theobromine intake and cognition function scores by basic characteristics using multiple linear regression analysis with adjusted variables in the full model, except for the subgroup variable. P < 0.05 was considered statistically signi cant. All analyses were performed using R 4.0.0.

Results
The recruited participants completed a cognitive questionnaire together with the rst questionnaire from the 2011-2012 and 2013-2014 NHANES surveys. As shown in Fig. 1, this study included 2,854 elderly American participants.
The mean age of study participants was 69.4 years. Study participants had a mean score of 19.0 for the CERAD word list learning test, 5.9 for the CERAD recall test, 16.7 for the animal uency test, and 46.2 for the digital symbol test. The mean CERAD word list learning test score was 1.  Table 1. Abbreviations: Consortium to Establish a Registry for Alzheimer's disease, CERAD. Figure 2 presents a weighted distribution of daily theobromine intake in log form with respect to quartiles of immediate recall score, delayed recall score, animal uency score, and DSST score. Daily theobromine intake was not signi cantly different among these quartile subgroups. As shown in Supplementary Fig. 1, daily theobromine intake did not follow a normal distribution. However, the log transformation distribution did. Thus, we used a continuous form and a log transform of daily theobromine intake to analyze the association between dietary theobromine and cognitive performance, as shown in Table 3 In the initial model, log transformation of daily theobromine intake was signi cantly associated with immediate recall test score (β, 95%CI, P value: 0.588, 0.151-1.025, 0.01), animal uency test score (β, 95%CI, P value: 0.729, 0.194-1.264, < 0.01), and DSST score (β, 95%CI, P value: 1.525, -0.002-3.053, 0.05), but not with the delayed recall test (β, 95%CI, P value: 0.172, -0.041-0.386, 0.11). In the fully adjusted model, dietary theobromine intake was positively associated with immediate recall score (β, 95% CI, P: 0.661, 0.222-1.101, < 0.01), delayed recall score (β, 95% CI, P = 0.232, 0.016-0.449, 0.04), and DSST score (β, 95%CI, P value: 0.009, 0.004-0.015, < 0.01). Daily theobromine intake was not associated with the poor cognitive function measured by the four indexes, independent of whether confounding variables were adjusted or not.
Theobromine intake was positively associated with animal uency score in never-smokers (β, 95%CI To further analyze any non-linear associations, we conducted a curve tting analysis between theobromine intake and cognitive performance. A non-linear association was found, as presented in Supplementary Fig. 3.

Discussion
In this study, we found a linear association between daily theobromine intake and DSST score, an indicator of execution ability, in a representative old American population. However, dietary theobromine intake was not signi cantly associated with the CERADimmediate/delayed recall score and animal uency score.
Previous studies have explored different diet patterns affecting cognitive function. Fernández-Fernández et al. (34) reported that an LMN diet, mainly composed of theobromine, could enhance cognitive reserve function in mice. One of the active components of chocolate is methylxanthine, of which theobromine is the main constituent. A randomized controlled trial (RCT) focusing on the psycho-pharmacologically of methylxanthines showed a protective effect on cognitive function (35). In a Portuguese prospective cohort study, 531 participants aged ≥ 65 years with normal cognitive function were followed for a median of 48-month to detect the association of chocolate intake and cognition impairment as measured by the Mini-Mental State Examination (4). In this study, researchers reported that long-term chocolate intake was inversely associated with cognitive decline. Another RCT in Japan also investigated the effect of chocolate intake on cognitive function (5). The intervention group was received dark chocolate daily for 30 days. The modi ed Stroop color word test and digital cancellation test were conducted to test the association between dark chocolate intake and cognitive function, nding that dark chocolate has a bene cial role in cognitive function.
We selected a large, representative elderly American population to further study the association between dietary theobromine and cognitive performance, nding a signi cant association between daily theobromine intake and DSST score. However, Mitchell et al. reported non-effect of theobromine (700 mg) alone or combination of theobromine (700 mg) and caffeine (120 mg) on DSST scores in 29 female healthy participants (36). In our study, 1000 mg/d theobromine intake was signi cantly associated with improvement in DSST score, from 4 to 15. Thus, a higher theobromine intake and a larger study sample size might be necessary to observe a positive association between dietary theobromine and DSST scores.
Islam et al. (7) reported that theobromine could improve the working memory of rats through the CaMKII/CREB/BDNF pathway. DSST is a working memory task that re ects execution ability. Thus, more studies will be required to study whether dietary theobromine intake could exert cognitive protection through similar mechanisms.
There are several limitations to our study. First, owing to the intrinsic limitations of the cross-sectional design, our study cannot conclude a causal association between theobromine intake and cognitive performance. Rigorous, prospective cohort studies or RCTs will be required to validate our results. Second, we excluded participants with unavailable daily dietary data or incomplete cognitive test scores, which could have biased our ndings. In Supplementary Table 1, we compare the basic characteristics of included and excluded participants, nding that the majority were balanced between these two groups. Third, because of the questionnaire design, we could not analyze side effects related to theobromine intake, such as heart rate and blood pressure. Thus, we cannot estimate the negative impact of theobromine intake. Further research on the use of theobromine in older adults is warranted.
Jie Liu, Li Jie Gao, designed the study, Li Jie Gao, Chao Hua Cui, Zheng Zhou Yuan, Wen Jing Ge, Qian Liu, and Jie Li, collected and cleaned the data, Li Jie Gao analyzed the data, Li Jie Gao and Jie Liu analyzed and interpreted the result, Li Jie Gao Na Liu, and Chun Yan Lei written the article. All authors reviewed this manuscript.