Is Milk Good or Bad for You? Testing the “Phosphorus Cancels Calcium, So It’s a Net Negative” Claim Against the Evidence [2026]

[Level 2–3 (Moderate)] [Food & Nutrition] [Series: Health for Serious Middle-Agers]

⚠️ Before You Read

This article is educational information based on peer-reviewed studies (cohorts, RCTs, Mendelian randomization) on milk and dairy. It is not a substitute for medical advice, diagnosis, or treatment. If you have a family history of breast or prostate cancer, reduced kidney function, or lactose intolerance, consult your physician about how to consume dairy. No specific food is guaranteed to “cure” or “prevent” disease.

Key Points (3-line summary)

• The claim that “milk’s phosphorus cancels calcium’s benefit, making it a net negative” is not supported by current evidence (leaning toward rejected). Milk’s Ca/P ratio is actually favorable; the phosphorus to worry about is the inorganic phosphorus in processed foods.
• Milk is neither a superfood nor a poison. It helps bone density and colorectal cancer, but raises some concern for prostate and breast cancer — the direction differs by outcome.
• Bottom line: centered on fermented dairy (yogurt, cheese), in moderate amounts, dairy is likely beneficial for Japanese people. Heavy consumption of non-fermented milk warrants caution. Conditioning on type, amount, and individual factors is closest to the truth.

1. Is the “Phosphorus Cancels Calcium” Claim True?

Milk’s calcium-phosphorus balance

Let’s address head-on the concern many readers share: “Milk is said to be good for bone because of calcium, but it’s also high in phosphorus — so isn’t it a net negative?”

Per 100 g of milk: calcium ~110–118 mg, phosphorus ~90–93 mg — a Ca:P ratio of roughly 1.2–1.3:1. What is considered “unfavorable” for bone is a diet with a Ca:P ratio below ~0.5 (phosphorus excess, calcium deficiency). Milk’s ratio is calcium-dominant and within the recommended range, so the notion that drinking milk is itself bad for bone does not hold up.

Indeed, even when calcium is adequate, a low Ca:P ratio raises blood PTH (parathyroid hormone) and promotes bone resorption (Kemi et al., Br J Nutr 2010). Conversely, keeping the Ca:P ratio high — the direction milk contributes — favors bone. The average US adult Ca:P ratio is ~0.6 (phosphorus-heavy), but that is driven by excess phosphorus from processed foods plus insufficient calcium, not by milk.

What you should really watch: inorganic phosphorus in processed foods

This is the crucial point. Phosphorus comes in two forms with very different absorption:

• Organic phosphorus (natural foods — milk, meat, fish): absorbed at ~40–60%. Plant sources even less (phytic acid).
• Inorganic phosphorus (food additives — processed cheese, ham/sausage, soft drinks, instant-food phosphates): absorbed at 80% to nearly 100%.

High blood phosphorus and FGF23 (a phosphorus-regulating hormone) are linked to vascular calcification, cardiovascular risk, and death even in people with normal kidney function. But the main culprit is inorganic phosphorus from processed foods. Milk’s organic phosphorus is absorbed moderately, and with its calcium-dominant ratio there is no data showing it cancels calcium’s benefit.

Verdict: the “milk’s phosphorus cancels calcium, making it a net negative” claim is not scientifically valid — it leans toward rejected. If you want to limit phosphorus, cut ham, processed cheese, and soft drinks (inorganic phosphate additives), not milk.

2. Milk/Dairy and Bone and Fractures

Bone is where two opposing lines of evidence exist — and the epicenter of the “milk is bad for bone” idea.

The cohort where heavy milk raised fractures and death (Michaëlsson 2014)

A large Swedish cohort (Michaëlsson et al., BMJ 2014) followed 61,000 women and 45,000 men and reported that women drinking ≥3 glasses/day had all-cause death HR 1.93 and hip fracture HR 1.60 — worse, not better. The authors hypothesized that D-galactose in milk promotes oxidative stress and inflammation (an animal-derived hypothesis).

However — the authors themselves stated it is observational and cannot show causation, with possible residual confounding and reverse causation (frail people choosing milk). Crucially, the result has not been replicated in other cohorts. Critical appraisals cite confounding by alcohol and overall diet, and reverse causation, and the D-galactose hypothesis is now judged “poorly supported.” Michaëlsson alone does not establish milk as harmful.

The RCT where dairy reduced fractures (Iuliano 2021) — the highest-quality evidence

On the other side is a randomized controlled trial closest to causation.

Iuliano et al., BMJ 2021: across 60 aged-care facilities and 7,195 residents (mean age 86), increasing milk, yogurt, and cheese to boost both calcium and protein produced all fractures −33%, hip fractures −46%, and falls −11%, with no adverse effect on blood lipids.

This is direct causal evidence that in older people short on calcium and protein, dairy supplementation actually reduces fractures. Note the population was calcium/protein-deficient aged-care residents.

The intermediate conclusion from meta-analyses

Bone mineral density (BMD) consistently rises with milk. For fractures, milk alone shows no clear preventive effect, whereas cheese and yogurt are more consistently protective. In sum: “drinking milk strengthens bones” is an overstatement, but “people short on calcium/protein who supplement with dairy have fewer fractures” holds. For calcium-deficient Japanese people, this is not a minor point.

3. Milk/Dairy and Mortality, Cardiovascular Disease, Diabetes

Large international cohorts favor dairy overall

The PURE study (Dehghan et al., Lancet 2018) followed 136,000 people across 21 countries and reported that consuming >2 servings/day of dairy was linked to lower total mortality, major cardiovascular events, and stroke (major CVD −22%). Even full-fat dairy showed no harmful signal, with benefits larger in low- and middle-income countries.

The keys: fermented vs. not, and fat content

• Fermented vs. non-fermented: In a dose-response meta-analysis (Front Nutr 2026), yogurt 200 g/day and cheese 15 g/day were linked to lower mortality and CVD death (robust, zero heterogeneity). Non-fermented milk was less consistent, with some reports of neutral-to-slight increase at high intake.
• Full-fat vs. low-fat: Cohorts predominantly report lower total and cardiac mortality with low-fat/skim milk than full-fat (e.g., Norway, AJCN 2025). Some reviews argue “food type matters more than fat,” so the debate continues.
• Type 2 diabetes: Total dairy shows no significant association (RR 0.98), but yogurt at 1 serving/day is linked to ~17% lower risk (Chen et al., BMC Med 2014). Fermented dairy’s edge reappears.

Summary: for cardiovascular disease, mortality, and diabetes, fermented dairy (yogurt, cheese) is clearly favorable. Non-fermented milk is neutral-to-mildly-unfavorable depending on conditions, with low-fat predominantly reported as better.

4. Milk/Dairy and Cancer (Direction Reverses by Site)

Cancer is where the direction flips by site — the most important point. You cannot label milk “good” or “bad” for cancer overall.

• Colorectal cancer (lower risk, most consistent): Milk, dairy, and calcium “probably” lower colorectal cancer (WCRF/AICR, meta-analyses). Mendelian randomization is also protective (milk, colorectal cancer OR 0.89). Mechanism: calcium binds carcinogenic secondary bile acids and free fatty acids.
• Prostate cancer (mild increase, weak-to-moderate): Total dairy, milk, and cheese show mild risk increase (RR ~1.01–1.03), but with substantial PSA-screening bias confounding.
• Breast cancer (increase signal, moderate): In the Adventist Health Study-2 (Fraser et al., 2020), breast-cancer risk rose dose-dependently with milk intake (~+50% at 1 cup/day). Mendelian randomization also suggests causation (genetically predicted milk intake, breast cancer OR 1.12). Still, this is special-population, low-intake data awaiting confirmation.

Summary: colorectal cancer ↓ (strong) / prostate cancer ↑ (weak-moderate) / breast cancer ↑ (moderate). Because directions oppose, evaluate by individual risk (sex, family history) — you cannot call it “good” or “bad” overall.

5. A Japan-Specific View

High lactose intolerance, chronically low calcium

East Asians, including Japanese people, have 70–100% lactose intolerance (lactase non-persistence) (5–15% in Europeans). Large amounts of non-fermented milk often cause abdominal symptoms, while fermented dairy (yogurt, cheese) has reduced lactose and suits Japanese people better. This aligns neatly with the health-outcome edge of fermented dairy.

At the same time, Japanese adults consume ~480–560 mg/day of calcium, below the recommended 600–700 mg/day — a chronic shortfall. Against this backdrop, using dairy to supplement calcium and protein carries relatively greater value.

Japanese cohorts favor milk

• JACC study: Men who drink milk daily have a high probability of lower stroke (especially cerebral infarction) death. One analysis found total stroke RR 0.53 in the highest dairy-calcium group.
• Large Japanese follow-ups repeatedly show milk/dairy linked to lower stroke death.

Japanese epidemiology is more favorable to milk/dairy than some Western cohorts (e.g., Michaëlsson). For Japanese people facing calcium shortfall and stroke (especially cerebral infarction), dairy supplementation carries real rationale.

6. Verdict: Is Milk a Net Positive or Negative?

Final judgment on the “phosphorus cancels calcium” claim

Not valid — leaning rejected. Four reasons: (1) milk’s Ca:P ratio ≈1.2–1.3 is calcium-dominant and bone-favorable; (2) milk’s phosphorus is organic with moderate absorption; (3) the real harm comes from inorganic phosphate additives in processed foods (80–100% absorbed); (4) low Ca:P ratio / phosphorus excess is a problem of the whole diet (“too much phosphorus + too little calcium”), and adding milk actually improves the ratio. “Milk = net negative because of phosphorus” does not hold.

Outcome-by-outcome summary

Evidence-based practical guidance

Milk/dairy cannot be flatly labeled “good” or “bad” — the peer-reviewed literature shows it depends on how you consume it. As a guide:

1. Type: Center on fermented dairy (yogurt, cheese) — most consistently favorable for cardiovascular, diabetes, and bone outcomes, and suited to lactose-intolerant Japanese people.
2. Amount: Moderate (about 1.5–2 servings/day of dairy total, or 1–2 glasses of milk) is sensible. Habitual >3 glasses/day of non-fermented milk is hard to actively recommend given mortality and prostate/breast-cancer considerations.
3. Fat: For cardiovascular and mortality outcomes, low-fat is slightly favored in most reports (debate ongoing).
4. Individual factors: With lactose intolerance, use fermented dairy or small split servings. Those with a family history of prostate or breast cancer should avoid heavy consumption. Conversely, elderly and women at high risk of calcium deficiency or osteoporosis gain a large fracture-prevention benefit from dairy’s calcium and protein.
5. Where phosphorus really matters: not milk, but phosphate additives in processed foods and soft drinks.

In one line: the “milk is a net negative because of phosphorus” claim lacks support and is rejected. For Japanese people, dairy — centered on fermented types and in moderate amounts — likely benefits bone and stroke by supplementing scarce calcium and protein. Heavy consumption of non-fermented milk warrants caution regarding mortality and some cancers. Conditioning on type, amount, and individual factors is closest to the truth.

7. FAQ

Q1. So should I just not drink milk? A. No. Especially for calcium-deficient Japanese people, moderate dairy likely benefits bone and stroke. But center on fermented dairy and keep non-fermented milk to 1–2 glasses/day.

Q2. I worry about phosphorus. Should I cut milk? A. If phosphorus concerns you, cut phosphate additives in processed foods and soft drinks, not milk. Milk’s organic phosphorus is moderately absorbed and, given its calcium-dominant ratio, is unlikely to harm bone.

Q3. Yogurt or milk — which is better? A. For cardiovascular, diabetes, and bone outcomes, yogurt (fermented dairy) has more consistently favorable data and suits lactose-intolerant Japanese people. Center on yogurt and cheese first.

Q4. Full-fat or low-fat? A. For cardiovascular and mortality, low-fat is slightly favored in most reports, though debate continues. In moderation, no need to overthink it.

Q5. I’m worried about breast/prostate cancer. A. If your risk (e.g., family history) is high, avoiding heavy non-fermented milk is the cautious choice. Cheese, yogurt, or moderate amounts need not be feared excessively — discuss individual decisions with your physician.

8. Summary

• The “phosphorus cancels calcium, so milk is a net negative” claim leans rejected. The Ca:P ratio is favorable; watch inorganic phosphorus in processed foods.
• Milk is neither superfood nor poison. Colorectal cancer ↓ and BMD ↑ are pluses; prostate/breast cancer are mild concerns — direction differs by outcome.
• Centered on fermented dairy (yogurt, cheese), in moderation, dairy is likely beneficial for Japanese people. Heavy non-fermented milk warrants caution.
• Conditioning on type, amount, and individual factors (lactose tolerance, family history) is the answer most faithful to the evidence.

References

• Michaëlsson K, et al. (2014). Milk intake and risk of mortality and fractures in women and men: cohort studies. BMJ 349:g6015. PMID: 25352269
• Iuliano S, et al. (2021). Effect of dietary sources of calcium and protein on fractures and falls in older adults in residential care (RCT). BMJ 375:n2364. PMID: 34670754
• Dehghan M, et al. (2018). Association of dairy intake with cardiovascular disease and mortality in 21 countries (PURE). Lancet 392:2288-2297. DOI: 10.1016/S0140-6736(18)31812-9
• Kemi VE, et al. (2010). Low calcium:phosphorus ratio and serum PTH / bone resorption. Br J Nutr. PMID: 19781123
• Chen M, et al. (2014). Dairy consumption and risk of type 2 diabetes: meta-analysis. BMC Med. PMID: 25420418
• Systematic review & meta-analysis (2022). Dairy products and prostate cancer risk. Br J Nutr. PMID: 35945656
• Meta-analysis (2020). Dairy products, calcium and colorectal cancer risk. Cancer Epidemiol Biomarkers Prev. PMID: 32855265
• Mendelian randomization (2023). Milk consumption and risk of twelve cancers (UK Biobank + FinnGen). Int J Epidemiol. PMID: 36473423
• Fraser GE, et al. (2020). Dairy, soy, and risk of breast cancer (Adventist Health Study-2). Int J Epidemiol.
• Dose-response meta-analysis (2026). Fermented dairy and mortality. Front Nutr. PMC12873575
• Harvard T.H. Chan School of Public Health, The Nutrition Source: Phosphorus (organic vs. inorganic absorption)
• JACC study (Japanese cohort: dairy calcium and stroke death) PMC7550994 / PMC4275440

Evidence as of September 2026. Most are observational and cannot exclude residual confounding; higher-quality causal evidence is limited to the Iuliano RCT (fractures) and Mendelian randomization (cancer). Some effect sizes/PMIDs should be verified against the original papers.