中国武汉高校学生从奶茶摄入咖啡因的风险评估
Probabilistic health risk assessment of caffeine intake from milk tea among college students in Wuhan, China by Monte Carlo simulation
该研究首次对武汉高校学生每日从奶茶摄入咖啡因进行了概率风险评估。通过HPLC测量了154份奶茶样本中的咖啡因浓度,并结合横断面调查数据,采用蒙特卡洛模拟估算每日咖啡因摄入量。
Abstract
Milk tea has become increasingly popular among young Chinese consumers, yet caffeine exposure through this beverage has not been systematically quantified. This study conducted the first probabilistic risk assessment of daily caffeine intake from milk tea among college students. Caffeine concentrations were measured in 154 freshly prepared milk tea samples across five tea base categories using HPLC. Consumption data were obtained from a cross-sectional survey of 1,495 college students in Wuhan, China. A Monte Carlo simulation was applied to characterize variability in caffeine concentration and milk tea consumption and to estimate daily caffeine intake. Caffeine exposure was additionally expressed on a body-weight basis for habitual intake (mg/kg bw/day) and single-occasion intake (mg/kg bw), and compared with the corresponding EFSA reference levels. Uncertainty in the summary exposure estimates was quantified using bootstrap resampling. The overall median caffeine concentration in milk tea was 178.2 mg/kg (range: 30.7–476.9 mg/kg), with borderline differences across tea bases (P-value = 0.057). Monte Carlo simulation yielded a median daily caffeine intake of 13.1 mg/day (95% CI 12.9–13.3) and a P95 of 29.0 mg/day (95% CI 28.2–29.7), with no simulated intake exceeding 400 mg/day. The median body-weight-normalized habitual exposure was 0.21 mg/kg bw/day (95% CI 0.21–0.21), with a P95 of 0.50 mg/kg bw/day (95% CI 0.49–0.52), and no simulated exposure exceeded the EFSA adult reference level of 5.7 mg/kg bw/day. For single-occasion exposure, the median was 1.45 mg/kg bw (95% CI 1.42–1.48), while the P95 was 3.35 mg/kg bw (95% CI 3.25–3.45), with 7.7% of simulated exposures exceeding 3.0 mg/kg bw. Overall, caffeine exposure from milk tea was low on a habitual basis, although higher exposure levels were observed for some simulated single-consumption occasions. These findings provide a more comprehensive characterization of milk–tea-derived caffeine exposure among college students.
Introduction
Caffeine (1,3,7-trimethylxanthine) is a naturally occurring xanthine alkaloid found in the leaves, seeds, and fruits of various plants, most notably tea, coffee beans, and cocoa beans1,2. Its central nervous system stimulatory effects are well documented and include enhanced alertness, improved cognitive performance, and reduced fatigue3,4. Epidemiological evidence also suggested that habitual moderate caffeine intake may be associated with a lower risk of several chronic diseases, including type 2 diabetes, cardiovascular disease, Parkinson’s disease, and certain cancers5,6.
Despite these possible benefits, excessive caffeine intake can produce a range of adverse effects. Acute high doses have been linked to anxiety, insomnia, restlessness, tachycardia, and gastrointestinal disturbances7, and it has been suggested that chronic excessive consumption may exacerbate underlying psychiatric disorders, though the evidence base for this association is still evolving8. The European Food Safety Authority (EFSA) conducted a comprehensive risk assessment and concluded that habitual caffeine consumption of up to 400 mg/day (approximately 5.7 mg/kg bw/day) does not raise safety concerns for healthy non-pregnant adults; for single doses, caffeine intake up to 200 mg (approximately 3 mg/kg bw) does not raise safety concerns for healthy adults9. Consistent with this assessment, a subsequent systematic review of over 380 studies found that caffeine intakes below 400 mg/day were not consistently associated with adverse cardiovascular, behavioral, reproductive, or developmental effects10.
The sources and levels of dietary caffeine intake vary considerably across populations. In Western countries, coffee is the predominant source. Mahoney et al. reported that U.S. college students consumed a mean of 159 mg/day of caffeine, with coffee and energy drinks as the primary contributors11. In China, by contrast, tea has historically been the dominant caffeine contributor. A national assessment estimated average daily caffeine intakes of 123 mg/day for Chinese adult males and 116 mg/day for females, with traditional tea leaves contributing the largest proportion12. Yong et al. further applied a probabilistic Monte Carlo framework specifically to tea consumption and reported a median caffeine intake of 104 mg/day among Chinese adult tea consumers, with approximately 10% of this population exceeding the 400 mg/day threshold13. However, these estimates were derived from data spanning all adult age groups, and the caffeine intake profile of Chinese college students remains largely uncharacterized. College students differ from the general adult population in several respects that may influence caffeine consumption, including higher levels of academic stress, more irregular sleep patterns, and distinct beverage consumption habits shaped by campus food environments and peer influences.
This knowledge gap is rendered more salient by the rapid growth of milk tea consumption among young Chinese consumers in recent years. Milk tea, a beverage prepared by combining tea infusion with milk, sweeteners, and other flavoring ingredients, has become a prominent feature of contemporary youth culture in China. According to an industry report published by the China Chain Store and Franchise Association, the market value of new-style tea beverages in China reached approximately 104 billion yuan in 2022, with nearly 486,000 milk tea stores operating nationwide14. College students constitute a core consumer segment for this market. Despite this considerable market penetration, research on the health implications of milk tea consumption remains limited in scope. The existing literature has focused primarily on psychological outcomes, with studies reporting associations between milk tea consumption and depressive symptoms, anxiety, and suicidal ideation among Chinese youths15,16. While these findings warrant attention, they leave largely unaddressed the more fundamental question of dietary exposure: milk tea derives its caffeine content from the tea base used in its preparation, yet no study to date has systematically quantified the levels of caffeine that consumers ingest through this increasingly popular beverage.
The present study was designed to address this gap. The primary objective was to quantitatively assess caffeine exposure from milk tea among college students in Wuhan, the city with one of the largest student populations in China, where the number of full-time college students exceeded 1.16 million in 202317. A probabilistic Monte Carlo simulation framework was adopted to account for variability in both consumption and concentration, thereby enabling estimation of habitual and single-occasion caffeine exposure and comparison with the corresponding EFSA reference levels. Secondary objectives were to characterize caffeine concentrations across the different tea bases used in commercially available milk tea, to examine whether caffeine intake differed meaningfully across sociodemographic and behavioral subgroups, and to estimate the milk tea consumption level corresponding to a 5% probability of exceeding 400 mg/day of caffeine.
Materials and methods
In the present study, we determined the caffeine contents in five types of milk tea commonly sold and consumed in Wuhan, China. Determinations were completed in the laboratory of the Wuhan Center for Disease Control and Prevention. We performed a risk assessment for caffeine intake from milk tea in Wuhan college students by combining caffeine content data with milk tea consumption data from a cross-sectional survey.
The study was conducted in accordance with the Declaration of Helsinki. The survey was approved by the ethics committee of the Wuhan Center for Disease Control and Prevention (WHCDCIRB-K-2024014), and all participants provided written informed consent. All methods were performed in accordance with the relevant guidelines and regulations.
Milk tea sampling
A total of 154 freshly prepared milk tea samples were collected between March and April 2024 in Wuhan, China. Samples were obtained from 36 chain milk tea brands (defined as brands operating at least three stores in Wuhan) and 11 non-chain milk tea brands, all located within a 1 km radius of five participating universities. For each store, at least three types of milk tea with different tea bases were obtained. The tea bases were classified into five categories: oolong tea (n = 46), green tea (n = 45), black tea (n = 33), jasmine tea (n = 12), and others (n = 18). Others included mixed tea bases and unspecified types. Milk tea with a milk cap or fruit-based tea was excluded.
Determination of caffeine content in milk tea
Samples were analyzed by a high-performance liquid chromatography (HPLC) method as described in the Chinese national standard GB 5009.139–2014 “National Food Safety Standard-Determination of caffeine in beverage”18. In short, 1 g (accurate to 0.001 g) of homogenized milk tea sample was weighed, and potassium ferrocyanide and zinc acetate were added for protein sedimentation. The solution was adjusted to 10 mL, shaken, ultrasonicated, centrifuged, and filtered. The caffeine content in the filtrate was determined by HPLC using a Waters HPLC system (Waters Corporation, USA). The HPLC method was validated for accuracy, precision, the limit of detection (LOD), and the limit of quantitation (LOQ). The LOD of caffeine was 0.07 mg/kg, and the LOQ was 0.21 mg/kg.
Milk tea consumption survey
Milk tea consumption data were obtained from a cross-sectional survey conducted among college students in Wuhan from March to April 2024. A total of 2,384 participants were recruited from five universities using convenience sampling. Food consumption data were collected using a self-administered questionnaire. Participants reported their milk tea consumption frequency (times in the past 30 days), average volume per occasion (mL), and body weight (kg). The daily milk tea consumption (g/day) was calculated as (frequency / 30) × volume per occasion, assuming a density of 1 g/mL. Only individuals who reported consuming milk tea at least once within the 30 days were included as milk tea consumers in subsequent analyses.
Quantitative probabilistic dietary intake assessment
The following equation was used to calculate daily caffeine intake from milk tea:
$$EDI=F\times C$$
where EDI (mg/day) is the estimated daily intake of caffeine, F is the daily consumption of milk tea (kg/day), and C is the caffeine concentration in milk tea (mg/kg).
A Monte Carlo simulation was conducted to estimate the distribution of caffeine intake from milk tea among college students in Wuhan. Unlike deterministic approaches that rely on fixed point estimates, probabilistic Monte Carlo methods account for variability in both consumption and concentration by defining input parameters as distributions rather than single values, thereby providing a more complete characterization of exposure and enabling estimation of the probability that intake exceeds a health-based guidance value19,20. Caffeine concentration data for each of the five tea base categories were fitted to lognormal distributions. Daily milk tea consumption was treated as an empirical distribution based on the survey data, with participants randomly sampled with replacement in each of 200,000 iterations for the overall assessment and 50,000 iterations for each subgroup. The tea base for each iteration was determined according to the proportion of each tea base among the 154 sampled products: oolong tea (29.9%), green tea (29.2%), black tea (21.4%), jasmine tea (7.8%), and others (11.7%). In each iteration, a caffeine concentration was randomly generated from the fitted distribution corresponding to the selected tea base, and the daily caffeine intake was calculated as the product of consumption (kg/day) and concentration (mg/kg). For the body-weight-normalized habitual exposure assessment, daily caffeine intake (mg/day) was divided by the corresponding participant's body weight (kg) and expressed as mg/kg bw/day. For the single-occasion exposure assessment, single-occasion caffeine intake was calculated using the reported average volume consumed per occasion and the simulated caffeine concentration, assuming a milk tea density of 1 g/mL, and was then divided by the corresponding participant's body weight (kg) to obtain exposure in mg/kg bw. The resulting habitual and single-occasion exposure estimates were compared with the corresponding EFSA reference levels of 5.7 mg/kg bw/day for habitual intake and 3.0 mg/kg bw for single doses, respectively.
The distribution of simulated daily caffeine intake was summarized using the mean, standard deviation (SD), median, 75th percentile (P75), and 95th percentile (P95). The P95 represents the 95th percentile of the simulated intake distribution. The proportion of iterations in which daily caffeine intake exceeded the EFSA adult reference level of 400 mg/day was also calculated. For the body-weight-normalized analyses, the proportions of simulated habitual and single-occasion exposures exceeding the corresponding EFSA reference levels of 5.7 mg/kg bw/day and 3.0 mg/kg bw, respectively, were calculated. Bootstrap resampling of the simulated exposure distributions was performed 1,000 times to obtain 95% confidence intervals for the summary exposure estimates. Each bootstrap replicate drew 5,000 values with replacement from the simulated distribution, and the 2.5th and 97.5th percentiles of the resulting bootstrap distribution were used as the 95% confidence interval.
Statistical analysis
Caffeine concentrations in milk tea were described as means, standard deviations (SD), medians, minimum, and maximum. Daily milk tea consumption and caffeine intake were described as means, SD, medians, and interquartile ranges (IQR). For the deterministic estimation of caffeine intake, each individual's daily intake was calculated as daily consumption (kg/day) multiplied by the median caffeine concentration of all 154 samples (178.2 mg/kg). Differences in caffeine concentrations across tea bases were compared using the Kruskal–Wallis rank-sum test. Differences in daily milk tea consumption and caffeine intake across population subgroups were compared using the Kruskal–Wallis test. All tests were two-sided, and P < 0.05 was considered statistically significant. Data cleaning, Monte Carlo simulation, bootstrap resampling, and all statistical analyses were performed in R software (version 4.1.2; R Development Core Team).
Results
Caffeine concentrations in milk tea with different tea bases
A total of 154 milk tea samples were analyzed across five tea base types: oolong (n = 46), black tea (n = 33), green tea (n = 45), jasmine tea (n = 12), and others (n = 18). The overall median caffeine concentration was 178.2 mg/kg, ranging from 30.7 to 476.9 mg/kg (Table 1). Caffeine concentrations tended to differ across tea bases, although the difference did not reach the threshold of statistical significance (P-value = 0.057). Green tea and black tea bases had relatively high median concentrations (187.7 and 181.0 mg/kg, respectively), while oolong tea had the lowest (150.6 mg/kg). Jasmine tea showed the highest median of 208.3 mg/kg, though this estimate is limited by the small sample size (n = 12). Pairwise comparisons are displayed in Fig. 1.
Caffeine concentrations in milk tea with different tea bases. Boxplots display medians, interquartile ranges, and individual data points for each tea base category. Asterisks indicate statistically significant differences between pairs (*P < 0.05).
Daily milk tea consumption
After excluding non-consumers and records with implausible values, 1,495 college students (67.2% female; 85.2% aged ≤ 20 years) were included. The median daily milk tea consumption was 83.3 g/day (IQR: 66.7–83.3 g/day), with a mean of 76.8 ± 22.4 g/day. Consumption was homogeneous across subgroups (Table 2): the median was identical across gender, age, and BMI categories, and varied only modestly by monthly allowance and price per cup. Kruskal–Wallis tests indicated statistically significant but quantitatively small differences across monthly allowance groups (P-value = 0.0003), price per cup groups (P-value = 0.0140), sleep duration groups (P-value = 0.0293), and age groups (P-value = 0.0426), while gender and BMI showed no significant differences (P-values > 0.05).
Daily caffeine intake from milk tea
Deterministic estimate
Using the median caffeine concentration (178.2 mg/kg), the median daily caffeine intake was 14.9 mg/day (IQR: 11.9–14.9 mg/day), approximately 3.7% of the EFSA adult reference level of 400 mg/day. Caffeine intake varied minimally across subgroups (Table 3), consistent with the homogeneous consumption pattern. The lowest median intake (11.9 mg/day) was observed in the > 2000 CNY monthly allowance group, followed by the < 1000 CNY group (12.5 mg/day), while most other subgroups had a median of 14.9 mg/day.
Probabilistic Monte Carlo assessment
To account for variability in both consumption and caffeine concentration, a Monte Carlo simulation with 200,000 iterations was performed. The simulation yielded a mean daily caffeine intake of 14.4 mg/day (95% CI 14.2–14.6), a median of 13.1 mg/day (95% CI 12.9–13.3), and a P95 of 29.0 mg/day (95% CI 28.2–29.7) (Table 4). None of the 200,000 iterations exceeded 400 mg/day, yielding an exceedance probability of 0.0%.
Subgroup-specific estimates were similarly low. The P95 intake ranged narrowly from 27.7 to 31.7 mg/day across all strata, with the highest value observed among consumers purchasing milk tea priced above 20 CNY per cup (31.7 mg/day; 95% CI: 30.9–32.5). This value remains less than 8% of the EFSA adult reference level of 400 mg/day. The full results are presented in Table 4. The simulated intake distribution (Fig. 2) was right-skewed and concentrated at low values, with approximately 95% of iterations falling below 30 mg/day. The 400 mg/day threshold lies far outside the observed range.
Distribution of simulated daily caffeine intake from milk tea. The histogram displays the results of 200,000 Monte Carlo iterations in which daily caffeine intake was estimated by combining the empirical distribution of milk tea consumption with lognormal-fitted distributions of caffeine concentration for each tea base. A small number of simulated values beyond the 99.9th percentile are not displayed for visual clarity but were retained in all statistical calculations.
Body-weight-normalized habitual and single-occasion caffeine exposure
When normalized to body weight, the mean habitual caffeine exposure from milk tea was 0.24 mg/kg bw/day (95% CI 0.23–0.24), with a median of 0.21 mg/kg bw/day (95% CI 0.21–0.21) and a P95 of 0.50 mg/kg bw/day (95% CI 0.49–0.52). No simulated habitual exposure exceeded the EFSA adult reference level of 5.7 mg/kg bw/day. For single-occasion exposure, the mean and median were 1.63 mg/kg bw (95% CI 1.61–1.66) and 1.45 mg/kg bw (95% CI 1.42–1.48), respectively. The P95 was 3.35 mg/kg bw (95% CI 3.25–3.45), and 7.7% of simulated single-occasion exposures exceeded the EFSA reference level of 3.0 mg/kg bw (Table 5).
Contribution of different tea bases to caffeine intake
The contribution of each tea base to total caffeine intake was estimated by Monte Carlo simulation (Table 6). Green tea and black tea contributed the largest shares (21.8% and 21.7%), followed by jasmine tea (20.1%), oolong tea (18.4%), and others (18.1%). The five tea bases contributed relatively evenly to the estimated caffeine intake, with the difference between the highest and lowest contributions being only 3.7 percentage points.
Consumption level corresponding to a 5% exceedance probability
The daily milk tea consumption corresponding to a 5% probability of exceeding 400 mg/day of caffeine was approximately 1,150 g/day, equivalent to about 2.3 cups of 500 mL (Fig. 3). This estimate considers caffeine intake from milk tea only and does not account for additional caffeine exposure from other dietary sources.
Probability of exceeding 400 mg/day of caffeine at different daily milk tea consumption levels. The curve was generated by combining fixed consumption volumes (200–4,000 g/day, in 50 g increments) with the empirical distribution of caffeine concentrations from all 154 milk tea samples. For each volume, 50,000 Monte Carlo iterations were performed. The dashed red lines indicate the consumption volume (approximately 1,150 g/day, equivalent to 2.3 cups of 500 mL) at which the exceedance probability reaches 5%.
Discussion
In this study, we conducted a probabilistic risk assessment of caffeine intake from freshly prepared milk tea among 1,495 college students in Wuhan, integrating primary data on caffeine concentrations in 154 milk tea samples with a consumption survey and Monte Carlo simulation. The median daily caffeine intake was estimated at 13.1 mg/day (95% CI 12.9–13.3), with a P95 of 29.0 mg/day (95% CI 28.2–29.7). No simulated daily intake exceeded the EFSA adult reference level of 400 mg/day. These findings were consistent across gender, age, monthly allowance, price per cup, sleep duration, and BMI strata. When normalized to body weight, habitual caffeine exposure from milk tea remained below the corresponding EFSA adult reference level of 5.7 mg/kg bw/day, with a P95 of 0.50 mg/kg bw/day. For single-occasion exposure, the P95 was 3.35 mg/kg bw, with 7.7% of simulated exposures exceeding the EFSA reference level of 3.0 mg/kg bw.
The caffeine intake levels observed in this study are markedly lower than those reported for Chinese adult tea consumers. Yong et al. estimated a median caffeine intake of 104 mg/day from tea among Chinese adults, with a P95 of 675 mg/day and 9.9% of consumers exceeding 400 mg/day13. This substantial discrepancy can be attributed to two factors. First, the caffeine content of dry tea leaves (27–43 mg/g) is approximately two orders of magnitude higher than that of milk tea (median: 178.2 mg/kg, i.e., ≈ 0.18 mg/g), reflecting the dilution of tea by milk, sweeteners, and other ingredients during milk tea preparation. Second, the daily consumption volume of milk tea in our study (median: 83.3 g/day) was considerably lower than the typical consumption of brewed tea among Chinese adults. The substantially lower caffeine concentration and modest consumption volume jointly explain the low caffeine intake from milk tea observed in our study.
The levels reported here are also lower than or comparable to caffeine intakes from other beverages in young adult populations. Mahoney et al. reported that U.S. college students consumed a mean of 159 mg/day of caffeine from all sources, with coffee and energy drinks as the primary contributors11. In a study of New Zealand tertiary students, Stachyshy et al. found that coffee contributed 68% of total caffeine intake21. Data on caffeine intake specifically among Chinese college students remain scarce. Chen et al., using data from the China Health and Nutrition Survey and the National Food and Beverage Consumption Survey, estimated that among Chinese adult consumers (≥ 18 years), the average caffeine intake was 123 mg/day for males and 116 mg/day for females, with tea leaves, coffee, and sodas as the primary contributors12. However, that study focused on the general adult population rather than college students specifically. Li et al. investigated caffeine-related knowledge and behaviors among Chinese university students from an ecopharmacovigilance perspective and noted that most students demonstrated positive attitudes toward controlling caffeine consumption, but quantitative intake data were not reported22. A study of medical students in Chongqing found that caffeinated beverage consumption was associated with late chronotype and elevated BMI, though caffeine intake was not quantified in absolute terms23. In our population, milk tea contributed approximately 13–15 mg/day, suggesting that for Wuhan college students, milk tea plays a relatively minor role in total dietary caffeine exposure compared with the dominant contribution of coffee in Western populations. However, in the absence of comprehensive caffeine intake data for Chinese college students that includes all dietary sources, the relative contribution of milk tea to total caffeine exposure in this population cannot be determined.
To our knowledge, this is the first study to quantitatively assess caffeine exposure from milk tea, a beverage that has experienced remarkable growth among young Chinese consumers, using a probabilistic Monte Carlo framework. The China Chain Store and Franchise Association reported nearly 486,000 milk tea shops nationwide by the end of 2022, and the market value exceeded 104 billion yuan in 2022, with a compound annual growth rate surpassing 20%14. The global bubble tea market was valued at USD 2.4 billion in 2019 and is projected to reach USD 4.3 billion by 202724. Despite this immense popularity, prior research on milk tea has focused predominantly on its psychological correlates15,16 or consumer purchasing behavior25, with virtually no quantitative exposure assessment for caffeine. Our study filled this gap by providing the first population-level caffeine intake estimates from milk tea, along with a characterization of the contributing factors.
A second novel finding of this study is the striking homogeneity of caffeine intake across diverse population subgroups. While Kruskal–Wallis tests identified statistically significant differences across monthly allowance, price per cup, and sleep duration groups, the absolute differences were trivial: median estimates differed by less than 2 mg/day, and P95 estimates by no more than 4 mg/day across the most divergent subgroups. This pattern is notable because prior research has documented significant socioeconomic and behavioral gradients in the consumption of sugar-sweetened beverages (SSBs) among young adults26,27. Pei et al. reported that limiting tapioca milk tea intake significantly reduced added sugar consumption among Taiwanese young adults, suggesting that milk tea can be a substantial contributor to dietary sugar28. However, our findings indicate that when it comes to caffeine specifically, milk tea consumption behavior is remarkably uniform across economic and demographic lines. This suggests that among college students, milk tea functions primarily as a social and recreational beverage rather than a functional source of caffeine, an interpretation supported by our finding that 86.2% of respondents cited personal enjoyment as their primary consumption motivation, while only 20.5% reported consuming milk tea to increase energy.
The even contribution of different tea bases to total caffeine intake (range: 18.1%-21.8%) represents a third novel observation. Despite statistically significant differences in caffeine concentration across tea bases, oolong tea had a median of 150.6 mg/kg, while green and black tea exceeded 180 mg/kg; the market share and consumer choice distribution resulted in remarkably balanced contributions. This has practical implications for risk communication: unlike coffee, where the caffeine content difference between Arabica and Robusta can be a factor of two29, consumers of milk tea do not need to be advised to choose specific tea bases to reduce caffeine intake. Oolong-based milk tea, despite its lower caffeine concentration, does not substantially alter total caffeine exposure at the population level because of its high market share (29.9% of samples). This finding also validates our probabilistic approach, which weighted the caffeine concentration of each tea base by its market proportion, thereby providing a more realistic exposure estimate than would be obtained from a simple deterministic calculation using the overall mean concentration.
From a public health perspective, simulated daily caffeine intake from milk tea among Wuhan college students remained well below the EFSA adult reference level of 400 mg/day. Even among the highest-exposure subgroup, consumers purchasing milk tea priced above 20 CNY per cup, the P95 intake reached only 31.7 mg/day, representing less than 8% of this reference level. Our simulation further demonstrated that to maintain a ≤ 5% exceedance probability, daily milk tea consumption should not surpass approximately 1,150 g, equivalent to about 2.3 cups of 500 mL. This threshold exceeds the consumption level of the vast majority of college students in our sample, whose median daily consumption was 83.3 g.
However, the low caffeine risk does not imply that milk tea consumption is without potential health consequences. Several studies have documented that milk tea products often contain substantial amounts of added sugars and trans fatty acids. Shi et al. analyzed 122 retail handcrafted milk tea products in Shanghai and found that the median sugar content was 8.5 g/100 mL, with some products exceeding 15 g/100 mL30. Min et al. reported that a typical 473 mL serving of boba milk tea can contain approximately 38 g of sugar31. Trans fatty acids, introduced through non-dairy creamers commonly used in milk tea, have been identified as a risk factor for cardiovascular disease, insulin resistance, and systemic inflammation32,33. Furthermore, Qu et al. recently documented addictive-like consumption patterns of milk tea among Chinese youths, with associations with depression, anxiety, and suicidal ideation15. Future research should therefore extend the risk assessment framework applied in this study to other milk tea constituents, particularly added sugars and trans fatty acids, and evaluate the potential metabolic and psychological health effects of frequent milk tea consumption.
Several limitations of this study should be acknowledged. First, the cross-sectional design precludes causal inference. Second, consumption data were self-reported over a 30-day recall period and may be subject to recall bias; the use of a 30-day reference period, while capturing habitual consumption patterns, may introduce greater recall error compared with shorter recall periods such as 24-h dietary recalls34. Third, individual tea base preferences were not recorded; the Monte Carlo simulation assumed that tea base selection was random and proportional to the market share observed in our 154 samples. If certain consumers systematically prefer specific tea bases, their actual caffeine intake may differ modestly from our estimates. However, given that even the P95 intake in the highest-exposure subgroup (31.7 mg/day) remained far below the safety threshold, any such misclassification is unlikely to alter the risk conclusion. Fourth, the density of milk tea was assumed to be 1 g/mL; actual density may vary depending on the specific formulation. Fifth, total dietary caffeine intake from coffee, brewed tea, soft drinks, energy drinks, and chocolate was not assessed, precluding an evaluation of the relative contribution of milk tea to overall caffeine exposure. The EFSA reference levels apply to caffeine from all dietary sources, whereas the present assessment considered caffeine exposure from milk tea only. Finally, the convenience sampling strategy and single-city design may limit generalizability to college students in other regions of China, where milk tea consumption patterns may differ.
Conclusions
In summary, this probabilistic risk assessment found that habitual caffeine intake from milk tea among Wuhan college students was low. The median daily intake was 13.1 mg/day, the P95 was 29.0 mg/day, and no simulated daily intake exceeded the EFSA adult reference level of 400 mg/day. Body-weight-normalized habitual exposure also remained below the corresponding EFSA adult reference level of 5.7 mg/kg bw/day. For single-occasion exposure, 7.7% of simulated values exceeded the reference level of 3.0 mg/kg bw. Daily intake varied only modestly across the socioeconomic and behavioral subgroups examined; the contributions of different tea bases to total intake were well balanced, and the estimated consumption level corresponding to a ≤ 5% probability of exceeding 400 mg/day (1,150 g/day) far exceeded the typical daily consumption observed in this population. Given the presence of added sugars and trans fatty acids documented in commercially available milk tea, future studies should extend this risk assessment framework to these constituents and evaluate total dietary caffeine intake to more fully elucidate the health implications of milk tea consumption among young adults.
Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author (email: hosan9174@163.com) upon reasonable request.
Abbreviations
- BMI:
-
Body mass index
- CI:
-
Confidence interval
- CNY:
-
Chinese yuan
- EDI:
-
Estimated daily intake
- EFSA:
-
European Food Safety Authority
- HPLC:
-
High-performance liquid chromatography
- IQR:
-
Interquartile range
- LOD:
-
Limit of detection
- LOQ:
-
Limit of quantitation
- P75:
-
75th percentile
- P95:
-
95th percentile
- SD:
-
Standard deviation
References
Heckman, M. A., Weil, J. & Gonzalez de Mejia, E. Caffeine (1, 3, 7-trimethylxanthine) in foods: A comprehensive review on consumption, functionality, safety, and regulatory matters. J. Food Sci. 75(3), R77-87 (2010).
Ashihara, H. & Suzuki, T. Distribution and biosynthesis of caffeine in plants. Front Biosci 9(2), 1864–1876 (2004).
Smith, A. Effects of caffeine on human behavior. Food Chem. Toxicol. 40(9), 1243–1255 (2002).
McLellan, T. M., Caldwell, J. A. & Lieberman, H. R. A review of caffeine’s effects on cognitive, physical and occupational performance. Neurosci. Biobehav. Rev. 71, 294–312 (2016).
van Dam, R. M., Hu, F. B. & Willett, W. C. Coffee, caffeine, and health. N. Engl. J. Med. 383(4), 369–378 (2020).
Poole, R. et al. Coffee consumption and health: Umbrella review of meta-analyses of multiple health outcomes. BMJ (Clin. Res. ed) 359, j5024 (2017).
Wikoff, D. et al. Systematic review of the potential adverse effects of caffeine consumption in healthy adults, pregnant women, adolescents, and children. Food Chem. Toxicol. 109, 585–648 (2017).
Temple, J. L. et al. The safety of ingested caffeine: A comprehensive review. Front. Psychiatry 8, 80 (2017).
EFSA Panel on Dietetic Products N, Allergies: Scientific Opinion on the safety of caffeine. 13(5):4102. (2015).
Doepker, C. et al. Key findings and implications of a recent systematic review of the potential adverse effects of caffeine consumption in healthy adults, pregnant women, adolescents, and children. Nutrients 10(10), 1536 (2018).
Mahoney, C. R. et al. Intake of caffeine from all sources and reasons for use by college students. Clin. Nutr. 38(2), 668–675 (2019).
Ye, C. et al. Trends of caffeine intake from food and beverage among Chinese adults: 2004–2018. Food Chem. Toxicol. 173, 113629 (2023).
Yong, L. et al. Quantitative probabilistic assessment of caffeine intake from tea in Chinese adult consumers based on nationwide caffeine content determination and tea consumption survey. Food Chem. Toxicol. 165, 113102 (2022).
Association CCSF: 2022 New Tea Beverage Research Report. In. Beijing: China Chain Store & Franchise Association; (2023).
Qu, D. et al. New form of addiction: An emerging hazardous addiction problem of milk tea among youths. J. Affect. Disord. 341, 26–34 (2023).
Wu, Y., Lu, Y. & Xie, G. Bubble tea consumption and its association with mental health symptoms: An observational cross-sectional study on Chinese young adults. J. Affect. Disord. 299, 620–627 (2022).
Wuhan Municipal Bureau of Statistics, Wuhan Survey Team of National Bureau of Statistics: Wuhan Municipal Statistical Communiqué on National Economic and Social Development 2023. In. Wuhan; (2024).
China NHaFPCotPsRo. National Food Safety Standard—Determination of Caffeine in Beverages: GB 5009.139-2014 (Standards Press of China, 2014).
Products EPoPP, Residues t: Guidance on the Use of Probabilistic Methodology for Modelling Dietary Exposure to Pesticide Residues. 2012, 10(10):2839.
Pekmezci, H., Sipahi, S. & Başaran, B. Health risk assessment of dietary chemical exposures: A comprehensive review. Foods https://doi.org/10.3390/foods14234133 (2025).
Stachyshyn, S., Ali, A., Wham, C., Knightbridge-Eager, T. & Rutherfurd-Markwick, K. Caffeine consumption habits of New Zealand tertiary students. Nutrients 13(5), 1493 (2021).
Li, S., Guo, J., He, B., Zhu, Y. & Wang, J. Environmental knowledge, behaviors, and attitudes regarding caffeine consumption among Chinese university students from the perspective of ecopharmacovigilance. Environ. Sci. Pollut. Res. 28(5), 5347–5358 (2021).
Zhang, Y. et al. Caffeinated drinks intake, late chronotype, and increased body mass index among medical students in Chongqing, China: A multiple mediation model. Int. J. Environ. Res. Public Health https://doi.org/10.3390/ijerph15081721 (2018).
Bubble Tea Market by Base Ingredient, Flavor, and Component: Global Opportunity Analysis and Industry Forecast, 2020–2027 [https://www.alliedmarketresearch.com/bubble-tea-market]
Istijanto, I. & Handoko, I. What approach and avoidance factors drive Gen-Z consumers to buy bubble tea? An exploratory study. Young Consum. 23(3), 382–396 (2021).
Han, E. & Powell, L. M. Consumption patterns of sugar-sweetened beverages in the United States. J. Acad. Nutr. Diet. 113(1), 43–53 (2013).
Zagorsky, J. L. & Smith, P. K. Who drinks soda pop? Economic status and adult consumption of sugar-sweetened beverages. Econ. Hum. Biol. 38, 100888 (2020).
Pei, Y. L. et al. The effect of limiting tapioca milk tea on added sugar consumption in Taiwanese young male and female subjects. J. Med. Invest. 65(1.2), 43–49 (2018).
Dam RMv, Hu FB, Willett WC: Coffee, Caffeine, and Health. 383(4):369–378. (2020).
Zehuan S, Zhuo S, Qi S, Mengying Q, Zhengyuan W, Jiajie Z: Nutrient content of 122 kinds of retail handcrafted milk tea products in Shanghai. J. Environ. Occupat. Med., 40(7):756–760. (2023).
Min, J. E., Green, D. B. & Kim, L. Calories and sugars in boba milk tea: Implications for obesity risk in Asian Pacific Islanders. Food Sci. Nutr. 5(1), 38–45 (2017).
Islam, M. A. et al. Trans fatty acids and lipid profile: A serious risk factor to cardiovascular disease, cancer and diabetes. Diabetes Metab. Syndr. 13(2), 1643–1647 (2019).
Mozaffarian, D., Katan, M. B., Ascherio, A., Stampfer, M. J. & Willett, W. C. Trans fatty acids and cardiovascular disease. New England J. Med. 354(15), 1601–1613 (2006).
Thompson, F. E. & Subar, A. F. Chapter 1 - Dietary assessment methodology. In Nutrition in the Prevention and Treatment of Disease (Fourth Edition) (eds Coulston, A. M. et al.) 5–48 (Academic Press, 2017).
Acknowledgements
All authors gratefully acknowledge the Wuchang District Center for Disease Control and Prevention and the Wuhan Field Epidemiology Training Program for their valuable support and assistance throughout this study.
Funding
This research was supported by the National Key Research and Development Program of China (grant number FC1601503) funded by the Ministry of Science and Technology of the People’s Republic of China and the Major Prevention Projects of Hubei Province Health Committee (grant number WJ2019H303) funded by the Health Commission of Hubei Province, China.
Author information
Author notes
Yating Wu, Haifeng Zhang, Yue Zhou, and Liya Ge have contributed equally to this work.
Authors and Affiliations
Wuhan Center for Disease Control and Prevention, Wuhan, 430021, People’s Republic of China
Yating Wu, Liya Ge, Xiaoqi Zhou, Xiaomin Wu, Junling Liu & Zhenyu He
Department of Nutrition and Food Hygiene, Hubei Key Laboratory of Food Nutrition and Safety, MOE Key Laboratory of Environment and Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, People’s Republic of China
Haifeng Zhang
Wuchang District Center for Disease Control and Prevention, Wuhan, 430061, People’s Republic of China
Yue Zhou & Fangzhao Ming
School of Public Health, Wuhan University of Science and Technology, Wuhan, 430065, People’s Republic of China
Yue Zhou
Jianghan Center for Disease Prevention and Control, Wuhan, 430015, People’s Republic of China
Zhi Zhao
Dongxihu Center for Disease Prevention and Control, Wuhan, 430040, People’s Republic of China
Ji Xin
Jiangxia Center for Disease Prevention and Control, Wuhan, 430200, People’s Republic of China
Shibo Hu
Wuhan Children’s Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430015, People’s Republic of China
Yiyang Guo
Authors
- Yating Wu
- Haifeng Zhang
- Yue Zhou
- Liya Ge
- Xiaoqi Zhou
- Xiaomin Wu
- Fangzhao Ming
- Zhi Zhao
- Ji Xin
- Shibo Hu
- Yiyang Guo
- Junling Liu
- Zhenyu He
Contributions
YTW, HFZ, YZ, and ZZ contributed equally to this work. ZYH and YYG conceived the study design. YTW, HFZ, YZ, ZZ, JNL, XQZ, XMW, FZM, JX, and SBH contributed to the acquisition of data. YTW, HFZ, YZ, and ZZ carried out data analysis. YTW, HFZ, YZ, and ZZ drafted the first version of the manuscript. ZYH had primary responsibility for the final content. All authors read and approved the final manuscript.
Corresponding authors
Correspondence to Yiyang Guo, Junling Liu or Zhenyu He.
Ethics declarations
Competing interests
The authors declare no potential conflicts of interest.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Wu, Y., Zhang, H., Zhou, Y. et al. Probabilistic health risk assessment of caffeine intake from milk tea among college students in Wuhan, China by Monte Carlo simulation. Sci Rep 16, 37385 (2026). https://doi.org/10.1038/s41598-026-75010-4
Received: 15 June 2026
Accepted: 05 October 2026
Published: 09 November 2026
Version of record: 09 November 2026
DOI: https://doi.org/10.1038/s41598-026-75010-4
Keywords
来源:Nature 健康护理主题 · nature.com