Abstract
Objectives Several lines of emerging pharmacological and epidemiological evidence imply that overall risks related to COVID-19 may be reduced by green tea catechins. Therefore, it may be expected that: (i) higher green tea consumption is associated with lower COVID-19 morbidity and/or mortality; (ii) countries with higher per/capita green tea consumption would be less affected by COVID-19. The aim of this study was to assess the second possibility.
Study design This was a cross-sectional study.
Methods Among countries with at least 3 million population (n=134), countries with relatively high (above 150 g) per/capita green tea consumption have been tentatively identified (n=21); (ii) normalized to population values of COVID-19 cases (morbidity) and deaths (mortality) for groups of countries with high and low per/capita green tea consumption were compared.
Results Striking differences in COVID-19 morbidity (and mortality) between groups of countries with ‘high’ and ‘low’ green tea consumption. The differences were still observed after the adjustment for the onset of the disease. Besides, preliminary analysis using multiple linear regression approach suggests that the associations are present at the level of individual countries.
Conclusion Evidence supporting the idea that green tea constituents could reduce overall risks related to COVID-19 has been obtained. The results are promising and are in line with emerging evidence from other studies including pharmacological ones. Nevertheless, because of limitations of this study the idea still should be considered as a hypothesis requiring further assessment.
Introduction
There is emerging evidence suggesting that green tea constituents, in particular catechins such as (-)-epigallocatechin (EGC), and (-)-epigallocatechin-3-gallate (EGCG) may be protective in regard of COVID-19. As briefly outlined below this could be due to (i) direct anti-viral activity and (ii) the influence on factors associated with COVID-19 severity and mortality.
Green tea catechins, are known to have anti-viral activity against several viruses [1-4] including viruses causing respiratory diseases [5, 6]. Recent target-based virtual ligand screening study suggests that at least one catechin (EGCG), is likely to target papain-like proteinase (PLpro), an indispensable enzyme in the process of coronavirus replication and infection of the host [7]. Inhibition of SARS-CoV-2 3CL-Protease by EGCG has been recently demonstrated in vitro [8].
Additionally, several factors associated with COVID-19 mortality are likely to be affected by green tea constituents. Indeed, there is evidence that green tea catechins: lower cholesterol levels [9, 10]; have anti-diabetic [11] and anti-obesity effects [12]; are beneficial in cardiovascular disease [13].
Besides, EGCG is protective in diseases with uncontrolled immune activation, thus, may ameliorate severity of COVID-19 [14] since the latter is related to overreaction of the immune system [15]. Finally, green tea catechins can act as ionophores for zinc ions, while the latter are considered as potentially beneficial in relation to COVID-19 [16, 17].
Altogether these results suggest that green tea catechins could reduce overall risks related to COVID-19. If so, it may be expected that countries with higher level of green tea consumption are less affected by COVID-19.
To address this hypothesis a preliminary ecological analysis was performed in this study. For this purpose: (i) among countries with at least 3 million population (n=134), countries with relatively high (above 150 g) per/capita green tea consumption have been identified (n=21); (ii) normalized to population values of COVID-19 cases (morbidity) and deaths (mortality) for groups of countries with high and low per/capita green tea consumption were compared.
Methods
Ethics Statement
Ethical approval was not deemed necessary as this was an analysis of publicly available data.
Identification of countries with relatively high per/capita green tea consumption
It has been previously reported that higher per/capita green tea consumption is typical for some countries in Asia, North Africa and Middle East [18]. Values for several countries in Asia (China, Vietnam, Japan and Indonesia) have been reported [19]. There is also information regarding Taiwan, arbitrary considered as a green tea consuming in the framework of this work because of predominant consumption of the Paochong tea (oxidized by only (8-12%) and Oolong tea (semioxidised) [20]. (The value of green tea consumption in Taiwan, indicated in the Table 1 (750 g) was arbitrary estimated as equal to one half of that reported [21] for total tea consumption in this country.) The relevant and reliable information regarding green tea consumption in the other countries is scarce [22], it is not as trivial as it may seem to find it. Therefore, to estimate the values of per/capita green tea consumption in the other countries the following approach was used. It was assumed that the difference between annual green tea import and export in a particular country (except countries with considerable green tea production) is roughly equivalent to annual green tea consumption. Thus, the difference between import and export divided by population should give a reasonable estimate of per/capita consumption. Indeed, such an approach is used for estimation for countries that do not produce tea [22].
The information about green tea import and export for particular countries was obtained from Commodity Trade Statistics Database | United Nations Statistics Division database [23]. Since for some countries recent records are unavailable, to minimize time bias, it was arbitrary chosen to use values for year 2016 as a standard (still for a few countries older records in the database were used in this work).
Using reported values for China, Vietnam, Japan and Indonesia [19], an estimate for Taiwan and the approach mentioned above it has been found that per/capita green tea consumption is above 150 grams in twenty one countries. However, it is essential to note that the list (Table 1) is probably incomplete (see considerations below), because the approach used for the estimate has several limitations. Namely, it cannot be excluded that some countries with considerable green tea per/capita consumption have not been identified among those producing green tea (for instance, Myanmar, South Korea, Papua New Guinea). Additionally, some countries may have not been identified because of absence of the relevant records in the UN database (for instance, Tajikistan). Finally, the estimates may be biased by import/export operations not reflected in the UN database. This is especially likely for ‘smaller’ import/export operations and thus for less populated countries. For this reason detailed screening of the database was limited to countries with at least 3 million of population.
Data regarding COVID-19 morbidity and mortality
All data were obtained from open resources. Information on total number of cases and total number of deaths was obtained from ‘Worldometers info. Coronavirus’ [24]. (The information on Worldometer is based on official daily reports and is considered as a reliable (for instance, [25, 26]).
For some comparisons information about COVID-19 morbidity (defined as number of cases per million population) and mortality (defined as number of deaths per million population) for a specific date was directly obtained from [24]. For others, considering that the onset of the disease was not simultaneous in different countries an adjustment for this factor was made as described below. Date when the total number of confirmed cases in a particular country became 50 (or more) was determined based on information provided in [24] and considered as a reference time point (referred further as the onset date). Then information about the total number of cases/deaths at a date six months after the onset date was obtained from [24] and normalized to population. These adjusted values were (also) used for comparisons. Six-month period after the onset date has been chosen as a compromise between longer period for analysis and a number of countries available for analysis (performed in November 2020). Since in some countries the onset occurred in June and even July (2020) some countries were excluded from a comparison of the adjusted values. Calculated values reported in the tables are (typically) rounded to integer, while values used for statistical comparisons were not rounded.
Statistical analysis
Since the variables of COVID-19 morbidity and COVID-19 mortality do not have a normal distribution [25], non–parametric statistic was primarily employed for the analysis as suggested in [25, 27]. Namely, Wilcoxon (Mann-Whitney U Test) for Unpaired Data and Spearman Rank Correlation were used for the analysis. On the hand, for multiple linear regression analysis, an approach similar to that previously reported [25] was used. In this analysis, morbidity and mortality per million of population were transformed to the common logarithm (log10) to adjust for normality of the distribution as suggested previously [25].
Factors, included in multiple linear regression analysis (beside green tea consumption) were: population density [28], percentage of population aged above 65 [29], percentage of urban population [30]. In a complementary analysis additional variable, namely Human Developmental Index (HDI) based on access to health and education services and income [31] was added to the model.
‘KyPlot’ software was employed for statistical assessments.
Results
A tentative list of countries with relatively high (above 150 g) per/capita green tea consumption was created using the approach described in Methods by screening relevant information for countries and territories listed in [24] and population with at least three million (134 countries and territories). Among those 21 were identified as countries with ‘high’ (above 150 g) per/capita green tea consumption (Table 1).
Among the other 113 countries: in 15 the overall annual balance between import and export of green tea was negative; for 12 countries no records were found in the UN database. Since information about production of green tea in countries with negative import/balance was not available, the group of these 113 countries was considered as a group with (i) low and (ii) undetermined consumption. Finally, a group of countries with the estimated consumption below 150 g was considered as a ‘low-consuming’ (n=86). Information about total number of COVID-19 cases (per one million population) and total number of COVID-19-death (per one million population) was obtained from ‘Worldometer’ [24] as provided on November 18, 2020 for the countries from the above mentioned three groups and summarized in Table 2.
COVID-19 morbidity (per one million population) and mortality (per one million population) in relation to per/capita green tea consumption.
Both, COVID-19 morbidity and mortality were strikingly (several times) lower in the Group 1 as compared to Group 2 (and Group3). Besides, the differences were highly significant (Table 2).
These results support the hypothesis that green tea constituents could reduce overall risks related to COVID-19. On the other hand, this preliminary analysis has several limitations. In particular, the onset of the disease was not simultaneous in different countries. To adjust for this the approach described in Method section was used.
A summary for the adjusted values is shown in Table 3. Notice a reduction of ‘N’ in the groups as compared to Table 2. This is due to late onset of the disease in some of the countries (Values for four countries in Group 2 and two countries in Group 3 were not available at a time of analysis (November 23, 2020). Nevertheless there were still profound and statistically significant differences in COVID-19 morbidity and mortality between Group 1 and Group 2/Group 3 (Table 3).
Adjusted for the onset COVID-19 morbidity (per one million population) and mortality (per one million population) in relation to per/capita green tea consumption.
To check for possible association between per capita green tea consumption and COVID-19 morbidity/mortality at the level of individual countries corresponding correlation coefficients were calculated. Namely, Spearman Rank Correlation was used to calculate correlation coefficients between: (i) the adjusted for the onset value of morbidity (or mortality) and (ii) per capita green tea consumption in a particular country. Thus, the total number of countries considered in this analysis was 105.
No statistically significant correlation between green tea consumption and the morbidity was observed (Rho=-0.13; P=0.17), while there was a weak but significant correlation between estimated green tea consumption and the mortality (Rho=-0.2; P=0.036).
To further study potential associations linear regression analysis accounting for several potential confounders was used. In this analysis: (i) morbidity and mortality per million population were transformed to the common logarithm (log10) to adjust for normality of the distribution (as suggested earlier [25]); (ii) in addition to green tea consumption, factors, reported previously as important confounders were included. In regard of the latter: population density [32]; percentage of population aged above 65 [32] and percentage of urban population [25] were included. Mauritania was excluded from the linear regression analysis as the apparent outlier in terms of estimated green tea consumption. (Indeed, value of 3400 g/capita is much greater than Q3+1.5*IQR for the group with high green tea consumption (where Q3 is 3-rd quartile, IQR interquartile range), thus was considered as outlier according to Tukey’s rule [33].) Additionally, since number of deaths (per million) in Mongolia and Cambodia were 0 (as of November 23, 2020) and could not be transformed to the logarithm, these countries were excluded from the analysis in regard of the mortality.
It has been found that there are: weak but statistically significant correlations between morbidity (and mortality) and green tea consumption; stronger correlations between morbidity (and mortality) and percentage of urban population. Plots for COVID-19 morbidity and mortality in relation to these two factors are shown in Fig. 1; additional details about results of the analysis are provided in Supplementary Table S1 (A,B).
COVID-19 morbidity (A, C) and mortality (B, D) in relation to % of urban population (upper panel) and green tea consumption (lower panel). Morbidity and mortality values are adjusted for the onset of the disease (6-month after the onset). Countries with values available as of November 23, 2020 were included. Values of regression coefficient R and significance levels denoted were calculated using the multiple linear regression model, N –number of countries included in the analysis. See text for more details. On plots C and D empty circles represent values for the countries with ‘high’ green tea consumption; for some of those standard two letter country codes are provided (AF-Afghanistan; CN –China; JP –Japan; LY-Libya; MA - Morocco; ML –Mali; MN-Mongolia; TW-Taiwan; UZ - Uzbekistan).
Thus, it appears that COVID-19 morbidity and mortality are indeed associated with green tea consumption at the level of individual countries. Of note, these associations were still statistically significant when additional important [32] variable, namely Human Developmental Index (HDI) based on access to health and education services and income [31] was added to the model (see Supplementary Table S2 (A,B))
Discussion
The main finding of this study is that higher green tea consumption is associated with lower COVID-19 morbidity/mortality at the level of ecological study. Indeed, there are striking differences in COVID-19 morbidity/mortality between groups of countries with ‘high’ and ‘low’ green tea consumption. The differences were still observed after the adjustment for the onset of the disease. Besides, preliminary analysis using multiple linear regression approach suggests that the associations are present at the level of individual countries. At this level, this is presumably the first evidence supporting the idea that green tea constituents may reduce the risks related to COVID-19. These results are in agreement with several lines of emerging evidence implying that green tea catechins may be effective in prevention/treatment of COVID-19 or amelioration of its severity [7-11, 14] (as outlined in the Introduction). Nevertheless, in the context of this study, it would be fair to mention that green catechins are not the only constituents of green tea. Other constituents may be also of importance for potential beneficial effects.
Limitations of the study and potential future directions
For most of the countries green tea consumption was estimated based on the annual balance between its import and export. Although this approach is used for coarse estimates [22], apparently, it has several limitations as outlined in Methods. Thus, the list of countries with ‘high’ green tea consumption should be considered as a tentative. It cannot be excluded that some countries with considerable green tea per/capita consumption have not been identified among those producing green tea (for instance, Myanmar, South Korea, Papua New Guinea). It would be of interest to obtain relevant information in regard of these countries, include it in the analysis and check how this would affect associations considered in this work.
Other limitations include multiple factors that may differentially affect COVID-19 morbidity and mortality in distinct countries. For instance, the administrative strategies to prevent transmission; population density, percentage of urban population, percentage of older population; condition-specific mortality risks; TB infection and BCG vaccination [25, 27, 34]. Potential role of only some of these factors was assessed in this study. On the other hand, potential confounding factors included in the linear regression model in this study are, actually, the majority of those identified as consistent (and at the same time relatively independent) in relation to COVID-19 in previous studies [25, 32]. Additionally, it appears that there are causal relations between green tea consumption, and several condition-specific risks (such as risks of cardiovascular diseases [13], diabetes [11], etc). Therefore, inclusion of these risk factors in the regression model as confounders would be rather controversial because a confounding variable should be an independent factor for outcome of interest [35]. It is important to notice that results of any ecological study, taken alone, could only provide the information about a correlation rather than a causal relation. Thus, experimental studies are required to either confirm or reject the potential causal relation. Importantly, green tea catechins are considered as safe by FDA and already have been investigated in experimental studies in relation to some infections. Moreover, results of at least one randomized placebo-controlled study [36] strongly suggest that green tea catechins can protect against acute upper respiratory tract infections (thus, quite likely, infections induced by members of coronavirus family). Therefore, it may be expected that green tea catechins can also protect against the infection induced by SARS-CoV-2. If this is the case, it would have very important implications. Possible efficacy of green tea catechins in treatment of COVID-19 and amelioration of its severity should be also assessed in observational and experimental epidemiological studies as potentially very important.
Conclusion
Evidence supporting the idea that green tea constituents could reduce overall risks related to COVID-19 has been obtained. The results are promising and are in line with emerging evidence from other studies including pharmacological ones. Nevertheless, because of limitations of this study the idea still should be considered as a hypothesis requiring further assessment.
Data Availability
Data derived from public domain resources
Author statements
Ethics Statement
Ethical approval was not deemed necessary as this was an analysis of publicly available data.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Competing interests
None declared.
Supplementary materials
Acknowledgments
I would like express my gratitude to personnel of the open resources for creating excellent opportunities for research.
Footnotes
More recent data have been analyzed. Therefore, the Tables and the Figure are revised. Supplementary Materials section has been added.
