In 1976, Dongsuh Foods sold Koreans coffee powder, creamer and sugar measured into one packet under the name Maxwell House Coffee Mix. That stick taught a country how to drink coffee. Its successor brand, Maxim, still moves about 5.3 billion sticks a year, roughly 170 every second, according to the Korea Times. For most of half a century, a Korean cup of coffee was sweet and milky, and it came stirred out of a paper sleeve.
Then the cafés came, and they came fast. Statistics Korea counted 51,551 coffee shops in 2016 and 100,729 by the end of 2022, about twice the number of convenience stores, Korea JoongAng Daily reported. Koreans now drink an average of 405 cups a year, against a world average of 152.
A national diet survey covering almost exactly that boom has now been used to ask what the move from stick to café cup did to the caffeine Koreans swallow. The answer depends on which part of the distribution you look at. Average daily caffeine from beverages rose by 1.66 milligrams a year. Over 2016 to 2024 that slope works out to roughly 13 milligrams, about one 25-milliliter mouthful of brewed coffee. Over the same years, the odds that an adult crossed 400 milligrams on a given day rose about 5 percent a year (odds ratio 1.05, 95% CI 1.02–1.07). The middle of the distribution barely moved while the heavy end grew.
One day of drinks, one number per liter
The analysis appeared October 19 in Frontiers in Nutrition. It comes from a team at Pusan National University Hospital, with Jinmi Kim as first author and Yu Hyeon Yi as corresponding author. The team pooled nine cycles of the Korea National Health and Nutrition Examination Survey, 2016 through 2024, and ended up with 48,991 adults. Each of them described a single day of drinking in a 24-hour recall. The researchers turned every caffeinated beverage into milligrams using one fixed average concentration per category from the Ministry of Food and Drug Safety, and applied the same numbers in every year so the cycles stayed comparable. The work was funded by the hospital’s Biomedical Research Institute grant, and the authors declared no commercial or financial conflicts.
Coffee dominated the results. Of the 31,056 adults who drank something caffeinated on their recall day, 90.8 percent drank coffee. Across the whole period, mixes still reached more caffeine consumers than brewed coffee did, 44.6 percent against 34.2 percent, but the two trend lines ran in opposite directions. Brewed coffee climbed year over year while the mix steadily lost ground, as News-Medical summarized.
The strongest coffee is the one people are leaving
The authors suggest that the shift “from coffee mixes toward brewed and specialty coffee” may partly explain the rise, because these products “typically contain higher caffeine concentrations and are often consumed in larger portion sizes.” The paper’s own conversion table cuts against the first half of that sentence. It assigns prepared coffee mixes 571.9 milligrams per liter, roasted bean or capsule coffee 522.8, and liquid coffee 352.9. By the study’s own accounting, the only coffee stronger per liter than brewed is the mix Koreans are supposedly giving up.
Those concentrations never change between 2016 and 2024, so the model cannot see a café adding a third espresso shot or a roaster selling stronger beans. The entire trend has to come from how much people said they drank and which category each drink fell into. If brewed coffee is pushing people past 400 milligrams, it has to be doing so through volume: bigger cups, more cups, more coffee occasions in a day, or a change in who is drinking. The paper reports none of those over time. The brewed category also lumps a café Americano together with a capsule made at home, so the survey can’t say how much of the growth happened at a café counter. The explanation the authors name is concentration, and their own table contradicts it. A follow-up that reported volume per drinking occasion by beverage type and year would test the portion-size half of their claim.
The 400-milligram line needs some context. It comes from regulators, including the European Food Safety Authority in its 2015 opinion, and it describes a habitual daily intake that raises no safety concern in healthy adults who are not pregnant. It is a reference level, not a cliff. A single recall can’t distinguish a habitual 400-milligram drinker from someone who pulled an all-nighter on Tuesday, a limit the authors acknowledge when they write that their method “may not fully represent habitual intake.” Compounded over eight years, a 5 percent annual rise implies odds about 48 percent higher in 2024 than in 2016. That figure describes the odds on any given day. It is not a count of new heavy drinkers, and the paper’s text gives no year-by-year prevalence numbers.
Average intake rose faster in men, 2.22 milligrams a year, than in women, 1.10, and fastest in adults aged 19 to 44, at 2.79. Adults 65 and older are an odd case. Their average intake showed no significant change (0.44 milligrams a year, with an interval that crosses zero), yet their odds of a 400-milligram day rose 12 percent a year (odds ratio 1.12, 95% CI 1.01–1.23). That estimate rests on a rare event. Across the whole sample, the paper’s tables count 85 men and 15 women aged 65 or older over the line on their recall day, or 1.2 and 0.2 percent after weighting.
The age pattern in women held up through every adjustment the authors tried, and the pattern in men did not. The authors adjusted for income, education, occupation, body mass, smoking, alcohol, breakfast habits and sleep. After all of that, women aged 30 to 44 had an adjusted odds ratio of 1.62 (95% CI 1.21–2.16) for a 400-milligram day compared with women aged 19 to 29, and 3.7 percent of them crossed the line on their recall day, against 2.5 percent of the younger group. Women aged 45 to 64 had nearly the same adjusted odds ratio, 1.61 (95% CI 1.19–2.17), although their weighted share, 2.6 percent, was close to the youngest group’s. Men aged 30 to 44 started with a crude odds ratio of 1.53, which fell to 1.21 (95% CI 0.93–1.56) once the same factors were included.
| Result | Value (aOR) | Lower bound | Upper bound |
|---|---|---|---|
| Women 30 to 44 | 1.62 | 1.21 | 2.16 |
| Women 45 to 64 | 1.61 | 1.19 | 2.17 |
| Men 30 to 44 | 1.21 | 0.93 | 1.56 |
Decades of research on coffee are reassuring for most adults, and the exceptions show up in specific places. A 2017 BMJ umbrella review of meta-analyses by Robin Poole and colleagues found that three to four cups a day were associated with about 17 percent lower all-cause mortality than none, and it concluded that coffee “seems generally safe within usual levels of intake.” The harms it did flag were concentrated in pregnancy, where high intake was associated with low birth weight (odds ratio 1.31) and pregnancy loss (1.46), and in fracture risk among women (1.14). These are observational associations with the usual confounding, and none of them were measured in the Korean respondents. Still, the group of women with both the highest adjusted odds and the highest weighted share, ages 30 to 44, covers much of the childbearing years, and the Korean analysis never breaks out pregnancy status.
What the stick was hiding
The swap has another effect, and this one could work in drinkers’ favor. Each mix stick replaced by a black Americano removes a serving of sugar and powdered creamer. A Chung-Ang University analysis of 14,132 Korean adults estimated about 5.5 grams of sugar per mix serving. That benefit only holds if people switch to black coffee, though, since the brewed category also includes lattes and sweetened café drinks, and snapshots of different people in different years can’t show anyone actually switching. Earlier Korean studies point in opposite directions. A 2014 KNHANES analysis of 17,953 adults in Diabetes Research and Clinical Practice found that instant-coffee drinkers had higher odds of metabolic syndrome (odds ratio 1.37) and blamed the sugar and creamer. The Chung-Ang study found that women who drank mixes had lower prevalence than women who drank no coffee at all (0.82).
That makes Korea’s café boom a natural experiment that nobody designed. A whole country is trading a small, sweet, creamy dose for a larger and often unsweetened one, and the published surveillance so far tracks only the caffeine. KNHANES already measures fasting glucose and waist circumference in the same people it asks about their drinks. Linking those measurements to the beverage data would show whether the café cup adds more risk through caffeine than it removes through sugar.
Two things will shape what happens next. The coming KNHANES cycles will show whether the odds of a 400-milligram day keep compounding at 5 percent a year. The other decision belongs to the Ministry of Food and Drug Safety. Its high-caffeine warning rule, which took effect after December 31, 2021, covers liquid foods with at least 0.15 milligrams of caffeine per milliliter. By the study’s figure, brewed coffee carries about three and a half times that. The rule requires the warning on a package’s main display panel, along with a caution for children, pregnant women and people sensitive to caffeine. The ministry has yet to decide whether that warning should follow coffee off the convenience-store shelf and onto the cups served at Korea’s 100,729 café counters.
Sources
- Frontiers in Nutrition: Kim et al., beverage-based caffeine intake trends among Korean adults, 2016 to 2024 (2026)
- BMJ: Poole et al., coffee consumption and health, umbrella review of meta-analyses (2017)
- Nutrients: coffee consumption patterns and metabolic syndrome in Korean adults (2019)
- Diabetes Research and Clinical Practice: Kim et al., instant coffee consumption and metabolic syndrome in Korean adults (2014)
- FoodNavigator: South Korea enforces mandatory high-caffeine warning labels (2020)
- Korea JoongAng Daily: Cafes nationwide top 100,000 to double convenience stores (2024)
- Korea Times: Maxim Mocha Gold marks 37 years of sweetening coffee culture (2026)
- News-Medical: Caffeine intake rises among Korean adults as brewed coffee gains ground (2026)