Carbonated drinks: the calcium loss belongs to the caffeine

The belief worth checking first is that carbonation pulls calcium out of the body. A crossover trial in women aged 20 to 40 who habitually drank 680 mL of carbonated beverages a day compared 4 of them, 2 with caffeine and 2 without, 2 acidified with phosphoric acid and 2 with citric acid. Extra calcium in the urine appeared only after the caffeinated drinks, at about 0.25 mmol, which is the amount already reported for caffeine on its own. Phosphoric acid without caffeine produced none, and it added nothing to what the caffeine did. So the acid and the gas leave urinary calcium alone, and caffeine moves it by an amount the trial authors read as compensated later in the day. That urine was collected over 5 hours after one breakfast, so it measures a short-term loss and not bone, and the abstract never says how many women took part. Where cola does move bone markers is at a dose nobody reaches by accident. Eleven men aged 22 to 29 on a low-calcium diet drank 2.5 L of cola a day for 10 days and their bone resorption markers rose. The other period of that trial replaced the cola with 2.5 L of milk, so the displacement of milk and the cola itself cannot be told apart.

The numbers, per 100 g

What cooking and storage change

• In 18 young adults, how the drink reached the mouth changed dental plaque acidity: a plastic glass gave the biggest pH swing, and a straw differed significantly from it at 5 and 10 minutes. — 18 subjects aged 18 to 25 years randomly divided into 3 groups of 6, drinking a carbonated beverage of pH 2.98 from a plastic glass, through a straw, or directly from the bottle, plaque sampled at 5, 10, 20 and 30 minutes (n = 18) Evidence B source
  Largest mean plaque pH change with the plastic glass, reported as 5.29, at all time intervals; significant difference between glass and straw at 5 and 10 minutes (P less than 0.05); no difference between straw and bottle

What it does to your health

• When 11 young men drank 2.5 L of cola a day for 10 days on a low-calcium diet instead of 2.5 L of milk, bone resorption markers rose. — 11 healthy men aged 22 to 29 years on a low-calcium basic diet, two 10-day periods with a 10-day washout, 2.5 L of cola daily in one period and 2.5 L of semi-skimmed milk in the other (n = 11) Evidence B source
  Serum C-telopeptide and urinary N-telopeptide both rose in the cola period (P less than 0.001), with phosphate, 1,25-dihydroxyvitamin D, parathyroid hormone and osteocalcin also higher; serum calcium and bone-specific alkaline phosphatase unchanged
• In 45 Japanese adults who finished a 12-week trial, a daily 500 g of plain carbonated water cut snacking by about 1.5 times a week and body weight by 0.72 kg against still water. — 46 healthy Japanese adults aged 20 to 64 with BMI 23 to under 30, randomised open-label placebo-controlled parallel groups, 500 g of carbonated or non-carbonated water once daily for 12 weeks, 45 completed (22 carbonated, 23 placebo) (n = 45) Evidence B source
  Snacking frequency -1.49 times per week at week 12 (95% CI -2.93 to -0.05); alcohol intake -24.44 g per week (95% CI -44.02 to -4.86); body weight -0.72 kg (95% CI -1.37 to -0.07); BMI -0.26; AST -4.05 U/L and ALT -9.47 U/L; no beverage-related adverse events
• In 13 young women, carbonated water in the mouth alone raised fullness and lowered hunger ratings even when it was spat out rather than swallowed. — 13 young women fasting overnight, randomised crossover of 4 conditions at 250 mL each: carbonated water swallowed, water swallowed, carbonated water sham-fed, water sham-fed, with electrogastrography from 20 minutes before to 45 minutes after (n = 13) Evidence B source
  Fullness up and hunger down after swallowing carbonated water, and transiently after sham-feeding it; gastric electrical power rose only when a drink was swallowed, and the rhythm shifted in opposite directions for oral stimulation and for ingestion
• In 8 healthy volunteers given a radiolabelled meal, carbonated water left gastric emptying unchanged and instead held more of the meal in the upper stomach. — 8 healthy volunteers, crossover comparison of carbonated against still water with a radiolabelled mixed meal, emptying and intragastric distribution measured by scintigraphy (n = 8) Evidence B source
  Emptying of solid and liquid and the lag phase identical for both drinks; proximal stomach held 74 per cent of solids against 56 per cent (P less than 0.05) and 43 per cent of liquids against 27 per cent (P less than 0.05) with carbonated water

What people believe that the data does not support

• In a trial that separated the ingredients, extra calcium in urine after carbonated drinks came only from the caffeinated ones, and phosphoric acid on its own produced none. — women aged 20 to 40 years who habitually drank 680 mL or more of carbonated beverages a day, incomplete random block design, 4 carbonated beverages (2 with caffeine, 2 without, 2 acidulated with phosphoric acid and 2 with citric acid) against water and milk (n = not stated in the abstract) Evidence B source
  Excess calciuria about 0.25 mmol, the same as reported for caffeine alone, and confined to caffeinated beverages; phosphoric acid without caffeine produced none and did not add to the caffeine effect; acidulant type had no acute effect
• A panel of 11 doctors reached 100 per cent agreement that people with reflux-like symptoms should limit carbonated beverages, and said in the same paper that the evidence behind such advice is low quality. — 11 gastroenterologists and primary care doctors, modified Delphi process with 2 online meetings and 2 voting rounds, 21 consensus statements graded with GRADE (n = 11 experts) Evidence E source
  Limiting carbonated beverages is one of 5 dietary statements that reached the strongest 100 per cent agreement, alongside avoiding personal triggers, limiting alcohol and coffee, and not overeating

Who this works differently for

• Among 1000 women aged 44 to 98, bone mineral density at 4 sites showed no association with any type of carbonated beverage once the usual confounders were taken out. — 1000 community-dwelling older white women aged 44 to 98 years, Rancho Bernardo cohort, bone density at 4 sites, adjusted for age, obesity, calcium intake, exercise, tobacco, alcohol, thiazides, estrogen and thyroid hormone (n = 1000) Evidence C source
  No association between bone mineral density and intake of any type of carbonated beverage after adjustment
• A systematic review that took 52 studies to full text found carbonated beverages and drinking acidic drinks at bedtime to be the most consistent dietary risks for dental erosion in adolescents. — adolescents aged 10 to 19 years with permanent teeth, 449 articles screened and 52 taken to full-text review, cohort and cross-sectional studies (n = 52 studies) Evidence C source
  Most consistent findings implicate the erosive potential of carbonated beverages and consumption of acidic drinks at bedtime; results were not consistent between cohort and cross-sectional designs

The bottom line

Plain carbonated water is a different product from cola, and it has one trial of its own. Forty-five healthy Japanese adults with a BMI from 23 to under 30 drank 500 g of carbonated or still water once a day for 12 weeks. The carbonated group snacked about 1.5 times less a week, drank 24 g less alcohol a week and weighed 0.72 kg less, and every one of those confidence intervals only just clears zero. The trial was open-label, so everyone knew which water they had, and its authors call the analyses exploratory with no adjustment for testing many outcomes at once. The plausible route is the change in behaviour, and nothing here shows the gas doing anything by itself. Two small mechanism trials say where the feeling of fullness comes from. In 13 young women, 250 mL of carbonated water raised fullness and lowered hunger even when it was sham-fed and spat out, which puts the signal in the mouth. In 8 volunteers given a radiolabelled meal, carbonated water left the speed of gastric emptying unchanged and held 74 per cent of the solid meal in the upper stomach against 56 per cent with still water. Appetite ratings and scintigraphy are not food eaten or weight lost. Teeth are where the acid matters. A systematic review that read 52 studies in full found carbonated beverages and acidic drinks at bedtime to be the most consistent dietary risks for dental erosion in adolescents, with cohort and cross-sectional studies disagreeing and the authors asking for better cohorts. Erosion follows acid rather than gas, which is why a cola at pH 2.98 and plain sparkling water are not the same exposure. In 18 young adults, 6 to an arm, such a cola dropped plaque pH most when it came from a plastic glass, and a straw differed from the glass at 5 and 10 minutes. Bone at ordinary intake looks untouched. Among 1000 women aged 44 to 98 in one community cohort, bone mineral density at 4 sites showed no association with any type of carbonated beverage after adjustment for age, obesity, calcium intake, exercise, tobacco, alcohol and several medicines. The authors limit that conclusion to modest intake. Reflux is advice rather than evidence. Eleven gastroenterologists and primary care doctors reached 100 per cent agreement that people with reflux-like symptoms should limit carbonated drinks, and the same paper says the evidence behind dietary advice of this kind is low quality, with few interventional studies. This page carries no separate list of who should be careful, because every caution on it is tied to a particular drink rather than to carbonation: acid on adolescent teeth, the reflux advice, and 2.5 L of cola a day on a low-calcium diet. Three things this page cannot tell you. We found no study of whether carbonation changes what a person absorbs from a drink, none linking carbonated drinks to kidney stones under that name. And nothing here measures the sugar in sweetened carbonated drinks, so no sentence on this page clears a sugary soda of anything.

Sources

1. carbw-r4-05 · randomised clinical study
   Highest mean pH drop (5.29) was recorded when consumed with plastic glass at all time intervals...The use of a straw and direct consumption of beverage from the bottle could limit harmful effects on dentition.
   Oral Health Prev Dent, 2005 · checked 2026-09-24
2. carbw-r4-02 · randomised controlled trial
   This controlled crossover intervention study included 11 healthy men (22-29 years) who were given a low-calcium basic diet in two 10-day intervention periods with an intervening 10-day washout. During one period, they drank 2.5 l of Coca Cola per day and during the other period 2.5 l of semi-skimmed milk.
   Osteoporos Int, 2005 · checked 2026-09-24
3. carbw-r4-06 · randomised controlled trial
   No study beverage-related adverse events were observed...Given the exploratory nature of the analyses and multiple outcomes without multiplicity adjustment, findings should be interpreted cautiously.
   Sci Rep, 2026 · checked 2026-09-24
4. carbw-r4-07 · randomised crossover trial
   After ingestion, significantly increased fullness and decreased hunger ratings were observed in the CW group...Our results suggest that CO2-induced oral stimulation is solely responsible for the feeling of satiety.
   J Nutr Sci Vitaminol (Tokyo), 2017 · checked 2026-09-24
5. carbw-r4-08 · randomised controlled trial
   However, the proximal stomach contained a greater proportion of solids (74 +/- 7% vs 56 +/- 8%, P < 0.05) and liquids (43 +/- 5% vs 27 +/- 4%, P < 0.05) with carbonated water as opposed to still water...In conclusion, carbonated water did not alter overall gastric emptying but profoundly modified intragastric distribution of the meal.
   Dig Dis Sci, 1997 · checked 2026-09-24
6. carbw-r4-01 · randomised controlled trial
   Relative to water, urinary calcium rose significantly only with the milks and the 2 caffeine-containing beverages...Phosphoric acid without caffeine produced no excess calciuria; nor did it augment the calciuria of caffeine.
   Am J Clin Nutr, 2001 · checked 2026-09-24
7. carbw-r4-09 · consensus statement
   One reason for this may be the lack of emphasis in management guidelines owing to 'low-quality' evidence and a paucity of interventional studies...Eleven statements achieved the strongest (100%) agreement: five are related to diet and include identification and avoidance of dietary triggers, limiting alcohol, coffee and carbonated beverages, and advising patients troubled by postprandial symptoms not to overeat
   Eur J Gastroenterol Hepatol, 2024 · checked 2026-09-24
8. carbw-r4-03 · cohort study
   One thousand women 44 to 98 years of age had bone mineral density measured at four sites and provided medical and behavioral histories, including type and quantity of carbonated beverages consumed...Bone mineral density levels were not associated with intake of any type of carbonated beverage after adjustment for age, obesity, calcium intake, exercise, and current use of tobacco and alcohol, thiazides, estrogen, or thyroid hormone.
   Am J Public Health, 1997 · checked 2026-09-24
9. carbw-r4-04 · systematic review
   The most consistent findings implicated the erosive potential of carbonated beverages and the consumption of acidic drinks at bedtime...Although results were not consistent between cohort and cross-sectional studies, this review suggests certain dietary risk factors may contribute to dental erosion in adolescents.
   Int J Paediatr Dent, 2020 · checked 2026-09-24