Dark chocolate or milk chocolate: what the head-to-head trials show

A meta-analysis of randomized trials lasting at least 4 weeks found that cocoa extract or dark chocolate with 70% cocoa or more lowered LDL cholesterol by 9.47 mg/dL (95% CI 5.20 to 13.75) and systolic blood pressure by 2.52 mmHg. Milk chocolate was not tested, and the pooled trials mix extracts with actual chocolate. Trials that put dark and milk chocolate side by side are all short. In 30 young adults, small-letter contrast sensitivity was 0.15 logCS higher (95% CI 0.08 to 0.22) about two hours after a dark bar than after a milk bar, and nobody knows how long that lasts. In 58 people who form calcium kidney stones, three days of dark chocolate raised urinary oxalate from 30 to 36 mg a day, and the same oxalate load from milk chocolate did not. Stone formation itself was not measured.

How much is actually in it

• A placebo-controlled trial in 64 adults gave capsules with 19 mg caffeine and 250 mg theobromine, a dose combination the authors call ecologically relevant. — 64 adults, double-blind placebo-controlled, six mornings (n = 64) Evidence B source
  19 mg caffeine plus 250 mg theobromine vs placebo, six mornings
• Per 100 g, UK data give milk chocolate 519 calories, 226 mg calcium and 57 mg magnesium, against 510 calories, 33 mg calcium and 89 mg magnesium for plain (dark) chocolate. — UK composition data (n = -) Evidence E source
  Milk vs plain: energy 519 vs 510 calories; sugars 56.0 vs 62.6 g; protein 7.3 vs 5.0 g; calcium 226 vs 33 mg; magnesium 57 vs 89 mg; iron 2.05 vs 2.30 mg

What a real portion looks like

• US survey data put a regular dark chocolate bar at 45 g, a snack size at 15 g and one piece at 8 g. — FNDDS portion weights, dark chocolate candy (n = -) Evidence E source
  Piece 8 g; snack size 15 g; regular 45 g; king size 90 g

What your body absorbs

• In calcium stone formers, 3 days of dark chocolate raised urinary oxalate by about 20%, while milk chocolate with the same oxalate but 430 mg calcium did not. — 58 calcium kidney stone formers, 94 mg oxalate a day from chocolate for 3 days (n = 58) Evidence B source
  Urinary oxalate 36 vs 30 mg/24 h after dark chocolate; no rise after milk chocolate
• In a crossover study in 6 adults, epicatechin absorption was similar from dark chocolate and from chocolate made with extra milk protein. — 6 adults, chocolate confections and cocoa beverages, serum followed 6 h (n = 6) Evidence B source
  Similar AUC among chocolate matrices; Cmax 32 nM dark vs 25 nM high milk protein chocolate
• In a crossover trial in 21 adults, 1.4 oz (40 g) of cocoa taken in 8.5 fl oz (250 ml) whole milk gave the same total urinary epicatechin metabolites as cocoa in water, though the metabolite profile shifted. — 21 healthy volunteers, 40 g cocoa powder in 8.5 fl oz (250 ml) whole milk or water (n = 21) Evidence B source
  No significant difference in total metabolites excreted over 24 h

What cooking and storage change

• Direct dark versus milk chocolate trials in the corpus last hours or days, USDA retention tables do not cover chocolate, and no US health claim names chocolate or cocoa. — Corpus, retention and claim datasets (n = 0) Evidence E source
  No long-term head-to-head trial; no retention factors; no US claim

What it does to your health

• In a crossover trial in 30 young adults, small-letter contrast sensitivity was 0.15 logCS higher 1.75 hours after a dark chocolate bar than after a milk chocolate bar. — 30 adults without eye disease, mean age 26, single bar each session (n = 30) Evidence B source
  +0.15 logCS (95% CI 0.08-0.22), P < .001; composite +0.20 log U (0.10-0.30)
• A meta-analysis of 31 trials in 1986 adults found that at least 4 weeks of cocoa or 70%-plus dark chocolate lowered LDL cholesterol by 9.5 mg/dL and systolic pressure by 2.5 mmHg, with no change in weight. — Adults with and without comorbidities, cocoa extract or >=70% cocoa dark chocolate for >=4 weeks (n = 1986) Evidence A source
  LDL -9.47 mg/dL (95% CI -13.75 to -5.20); SBP -2.52 mmHg (-4.17 to -0.88); DBP -1.58; weight and HbA1c unchanged
• A meta-analysis of cohort studies with 908,390 people links each extra 5 g of chocolate a day with 2% lower cardiovascular risk, without separating dark from milk chocolate. — 21 risk estimates from prospective cohorts (n = 908390) Evidence C source
  Pooled RR 0.98 per 5 g/day (P < 0.001)
• In a 27-day crossover trial in 42 young men, a daily 46 g milk chocolate bar in place of a high-carbohydrate snack raised HDL cholesterol by 3.1 mg/dL (0.08 mmol/L) without worsening LDL. — 42 normocholesterolemic young men on a Step 1 diet, 27 days per period (n = 42) Evidence B source
  HDL +0.08 (0.03) mmol/L; triglycerides -0.06 (0.03) mmol/L; LDL response not adversely affected

Safety, limits and interactions

• After a severe reaction to a dark chocolate with undeclared milk, testing found milk protein levels in unlabelled products that could trigger reactions in up to 68% of milk-allergic adults. — Complaint sample plus dark chocolate batches and brands, probabilistic risk model for milk-allergic adults (n = -) Evidence D source
  Up to 68% of adult milk-allergic consumers could react
• True cocoa allergy is rare: a case report of a 2-year-old confirmed by oral challenge with dark chocolate notes most reactions to chocolate come from milk, peanut or tree nuts in it. — One 2-year-old girl with tree nut anaphylaxis (n = 1) Evidence D source
  Anaphylaxis on dark chocolate challenge; IgE to cocoa confirmed
• Lab testing of 72 US dark chocolate and cocoa products from 2014 to 2022 found 43% above California Prop 65 daily limits for lead and 35% for cadmium, though 70 of 72 were under the FDA lead reference level. — 72 consumer cocoa-containing products bought 2014-2022 (n = 72) Evidence D source
  Exceeding Prop 65 MADL: lead 43%, cadmium 35%, arsenic 0%; 97.2% below FDA IRL for lead

What people believe that the data does not support

• The EU refused a general claim that cocoa flavanols maintain and promote normal blood pressure. — EU health claim register (n = -) Evidence E source
  Non-authorized

Who this works differently for

• In 14 postmenopausal women given equal calories of each, ad libitum energy intake afterward was about 340 kJ lower after 80% cocoa dark chocolate than after 35% cocoa milk chocolate. — 14 healthy postmenopausal women, 2099 kJ of chocolate per session, one week apart (n = 14) Evidence C source
  1355 vs 1693 kJ (P = 0.008) dark vs milk; glucose and insulin transiently higher after milk chocolate
• A meta-analysis of 145 trials found flavan-3-ol foods such as cocoa lowered office blood pressure by 2.8/2.0 mmHg, and by 5.9/2.7 mmHg in people with raised or high blood pressure. — 5205 participants in 145 RCTs, about 586 mg flavan-3-ols a day from cocoa, tea, grape, apple or pure compounds (n = 5205) Evidence A source
  Office BP -2.8 (95% CI -3.9 to -1.7)/-2.0 mmHg; hypertensive -5.9 (-10.0 to -1.8)/-2.7 mmHg

The bottom line

Per 100 g, UK composition data give milk chocolate 519 calories, 226 mg calcium and 57 mg magnesium, and plain chocolate 510 calories, 33 mg calcium and 89 mg magnesium. UK plain chocolate is a low-cocoa dark, so a 70% bar will differ. Sugar is similar, 56.0 g in milk and 62.6 g in plain. A regular dark chocolate bar weighs 1.6 oz (45 g) in US survey data and one piece 8 g. Milk does not seem to block cocoa. In 6 adults, epicatechin absorption was similar from dark chocolate and from chocolate with extra milk protein. In 21 volunteers, 1.4 oz (40 g) of cocoa stirred into 8.5 fl oz (250 ml) of whole milk gave the same total urinary metabolites as cocoa in water. Milk chocolate has its own small trial: in 42 young men on a lower-fat diet, a daily milk chocolate bar in place of a high-carbohydrate snack for 27 days raised HDL cholesterol by 3.1 mg/dL (0.08 mmol/L), in an older study run on men only. In 14 postmenopausal women given equal calories of chocolate, later free eating was 1355 kJ after dark and 1693 kJ after milk chocolate, from single meals in an order that was counterbalanced rather than randomized. Across 145 trials of flavan-3-ols from cocoa, tea, grape, apple or pure compounds, office blood pressure fell by 2.8/2.0 mmHg, with heterogeneity high and evidence moderate. Cohort data on 908,390 people link each extra 5 g of chocolate a day with 2% lower cardiovascular risk (RR 0.98), without separating dark from milk. That is an association. The EU refused the general claim that cocoa flavanols maintain normal blood pressure, which means the evidence fell short at assessment. Milk chocolate is made with milk, so it is out for anyone with a milk allergy. The less obvious cautions sit with dark chocolate. After a severe reaction to a dark chocolate with undeclared milk, testing found milk protein in unlabelled products at levels a model predicts could trigger reactions in up to 68% of milk-allergic adults, based on one index case. If you have a milk allergy, a dark bar without a milk warning is not proof of no milk. True cocoa allergy is rare, and a case report of a 2-year-old notes that most reactions to chocolate come from the milk, peanut or tree nuts in it. Of 72 US dark chocolate and cocoa products bought from 2014 to 2022, 43% exceeded California Prop 65 daily limits for lead and 35% for cadmium, while 70 of 72 stayed below the FDA lead reference level. Milk chocolate was not tested there. People who form calcium kidney stones have the oxalate result above to weigh.

Sources

1. dcm-r5-10 · RCT
   Treatment capsules contained either an ecologically relevant dose combination of 19-mg caffeine and 250-mg theobromine or a placebo.
   Psychopharmacology (Berl), 2005 · checked 2026-09-29
2. dcm-r5-14 · dataset
   CoFID 17-648 Chocolate, milk: Energy 519 kcal; Total sugars 56.0 g; Protein 7.3 g; Calcium 226 mg; Magnesium 57 mg; Iron 2.05 mg. CoFID 17-491 Chocolate, plain: Energy 510 kcal; Total sugars 62.6 g; Protein 5.0 g; Calcium 33 mg; Magnesium 89 mg; Iron 2.30 mg
   CoFID 2021 (Public Health England), codes 17-648 and 17-491 · checked 2026-09-29
3. dcm-r5-15 · dataset
   FNDDS 91705300 Dark chocolate candy (fdc_id 2710336): 1 piece 8 g; 1 fun/snack size 15 g; 1 pouch/regular size 45 g; 1 large/king size 90 g; Quantity not specified 45 g
   USDA FNDDS 2021-2023, food code 91705300 (fdc_id 2710336) · checked 2026-09-29
4. dcm-r5-03 · RCT
   The consumption of dark chocolate induced a significant increase in mean urinary Ox (36 +/- 14 versus 30 +/- 10 mg/24 hr) not observed in the milk chocolate group.
   J Ren Nutr, 2003 · checked 2026-09-29
5. dcm-r5-04 · RCT (crossover)
   Six subjects consumed each product in a randomized crossover design, with serum EC concentrations monitored over 6 h post consumption. Areas under the serum concentration-time curve (AUC) were similar among chocolate matrices.
   J Agric Food Chem, 2009 · checked 2026-09-29
6. dcm-r5-05 · RCT (crossover)
   The results show that milk does not significantly affect the total amount of metabolites excreted in urine.
   Br J Nutr, 2008 · checked 2026-09-29
7. dcm-r5-17 · search result
   corpus_search.py "milk chocolate": head-to-head trials acute (29710322, 28314639, 28572069) or 3-day (12563622); retention_search.py dark chocolate: no method; claims_search.py dark chocolate: nothing in 21 CFR 101; euclaims_search.py cocoa: 0 authorised, 7 non-authorised
   Bioma corpus and dataset search (PubMed corpus, USDA R6, 21 CFR 101, EU register snapshot 2026-09-21), 2026-09-29 · checked 2026-09-29
8. dcm-r5-01 · RCT (crossover)
   Among the 30 participants (9 men and 21 women; mean [SD] age, 26 [5] years), small-letter contrast sensitivity was significantly higher after consumption of dark chocolate (mean [SE], 1.45 [0.04] logCS) vs milk chocolate (mean [SE], 1.30 [0.05] logCS; mean improvement, 0.15 logCS [95% CI, 0.08-0.22 logCS]; P < .001).
   JAMA Ophthalmol, 2018 · checked 2026-09-29
9. dcm-r5-06 · meta-analysis
   Yet, there was a reduction in total cholesterol (-8.35 mg/dL, 95% CI -14.01; -2.69 mg/dL), LDL-c (-9.47 mg/dL, 95% CI -13.75; -5.20 mg/dL), fasting blood glucose (-4.91 mg/dL, 95% CI -8.29; -1.52 mg/dL), SBP (-2.52 mmHg, 95% CI -4.17; -0.88 mmHg), and DBP (-1.58 mmHg, 95% CI -2.54; -0.62 mmHg).
   Nutrients, 2024 · checked 2026-09-29
10. dcm-r5-08 · meta-analysis of cohorts
    Within the updated meta-analysis of cohort studies (21 risk estimates, 908,390 participants and 65,407 events), greater consumption of chocolate (per 5 g/day) was associated with a lower risk of CVD incidence and mortality (pooled relative risk = 0.98, P value < 0.001; P for nonlinearity < 0.001).
    Eur J Epidemiol, 2022 · checked 2026-09-29
11. dcm-r5-09 · RCT (crossover)
    When subjects consumed a milk chocolate bar instead of the high-carbohydrate snack, high-density-lipoprotein (HDL) cholesterol was 0.08 +/- 0.03 mmol/L higher (P < 0.01) and plasma triglycerides were 0.06 +/- 0.03 mmol/L lower (P < 0.05).
    Am J Clin Nutr, 1994 · checked 2026-09-29
12. dcm-r5-11 · case report with risk assessment
    The results show that milk protein concentrations in unlabelled products reach levels that may elicit allergic reactions in up to 68% of the adult allergic consumers.
    Food Addit Contam Part A Chem Anal Control Expo Risk Assess, 2010 · checked 2026-09-29
13. dcm-r5-12 · case report
    Suspected allergic reactions to cocoa are frequently reported by patients, yet most of these cases are caused by sensitization to other more common allergens contained in chocolate products, such as milk, peanuts or tree nuts.
    Front Pediatr, 2026 · checked 2026-09-29
14. dcm-r5-13 · food analysis
    Our analysis reports that 43, 35, and 0% of the products tested exceed Prop 65 MADLs for heavy metal concentrations, respectively, of Pb, Cd, and As, while 97.2% (70 of 72) fall below US FDA IRL limits established for Pb.
    Front Nutr 2024 · checked 2026-09-29
15. dcm-r5-16 · health claim register
    POL-HC-6949 Cocoa flavanols Maintenance and promotion of a normal blood pressure Non-authorised
    EU Register on nutrition and health claims, claim POL-HC-6949, snapshot 2026-09-21 · checked 2026-09-29
16. dcm-r5-02 · crossover study
    Ad libitum energy intake was significantly lower following dark (1355 ± 750 kJ) compared with both milk (1693 ± 969 kJ; P = 0.008) and white (1842 ± 756 kJ; P = 0.001) chocolate consumption.
    Appetite, 2017 · checked 2026-09-29
17. dcm-r5-07 · meta-analysis
    Interventions decreased office (-2.8 [95% CI -3.9, -1.7]/-2.0 [-2.6, -1.3] mmHg) and 24 h-ambulatory BP (-3.7 [-5.8, -1.6]/-2.6 [-4.5, -0.8] mmHg) after chronic repetitive consumption.
    Eur J Prev Cardiol, 2025 · checked 2026-09-29