White rice and diabetes in 21 countries

Overview

Bhavadharini B et al.. Diabetes Care 2020 · PMID 32873587 · DOI 10.2337/dc19-2335

Very high white rice intake (>=3 cups/day) linked to 20% higher diabetes risk.

Does the study support the specific claim? No · Methodological quality: high

Original publication

Plain-language summary

Scientific in-depth review

1. Full reference

Authors

Bhavadharini B et al.

Journal

Diabetes Care

Year

2020

Identifier

PMID 32873587 DOI 10.2337/dc19-2335

Verification

verified – PubMed efetch 2026-10-25

Retracted

no

Corrections

none

Extent checked

full text

Conflicts of interest and funding

Investigator-initiated PURE funded by PHRI, CIHR, Heart and Stroke Foundation of Ontario, Ontario Ministry of Health and unrestricted grants from pharmaceutical companies (AstraZeneca, Sanofi, Boehringer Ingelheim, Servier, GSK, Novartis, King Pharma) plus national bodies incl. Dairy Farmers of Canada/National Dairy Council, Unilever Health Institute (Brazil) and the South Africa Sugar Association. Co-author V.M. is involved in the promotion of healthier varieties of rice. Funders had no role in design, analysis or reporting.

2. Research question

White rice and incident diabetes

3. Study design and population

Study design

Cohort study – Multinational prospective cohort (PURE)

Population

PURE participants aged 35-70 years without baseline diabetes, with complete diet data and >=1 follow-up visit (148,858 with diet data minus 16,485 with baseline diabetes); mean age 50 years; South Asia n=26,419, China n=41,727, rest of world n=64,227

Country

21 countries: Argentina, Bangladesh, Brazil, Canada, Chile, China, Colombia, India, Iran, Malaysia, occupied Palestinian territory, Pakistan, Philippines, Poland, South Africa, Saudi Arabia, Sweden, Tanzania, Turkey, United Arab Emirates, Zimbabwe (Egypt not included)

Period

Enrolment from 2005 (PURE baseline); outcome events to 3 July 2019; mean follow-up 9.5 years

Sample size

132,373; 6129 cases

Inclusion and exclusion criteria

Included: complete FFQ data, at least one follow-up visit; excluded baseline diabetes (n=16,485).

4. Methods and endpoints

Exposure / intervention

Cooked white rice from country-specific validated FFQs at baseline only (mixed dishes disaggregated); categories <150, 150-<300, 300-<450, >=450 g/day (1 cup = 150 g). Rice type (e.g. parboiled, basmati) not recorded; brown rice not on the FFQ.

Comparator

<150 g/day

Endpoints

Incident diabetes: physician diagnosis, use of oral agents or insulin, or fasting plasma glucose >=7.0 mmol/L; 90.7% of 6,129 cases documented, 9.3% self-report only

Statistical methods

Multivariable Cox frailty model with random intercept for centre; adjusted for age, sex, BMI, waist-hip ratio, family history, smoking, urban/rural, education, wealth index, physical activity, energy intake, whole and refined grains, fruit and vegetables; regional stratification; pooled regional heterogeneity (I2).

5. Actual results

Results

Fully adjusted HRs vs <150 g/day, overall: 150-<300 g 1.12 (1.01-1.24); 300-<450 g 1.25 (1.10-1.43); >=450 g 1.20 (1.02-1.40); P trend 0.003. South Asia >=450 g: 1.61 (1.13-2.30), P trend 0.02. Rest of world >=450 g: 1.41 (1.08-1.86), P trend 0.01 (only 108 events in top category). China >=450 g: 1.04 (0.77-1.40), P trend 0.38 (300-<450 g: 1.34, 1.05-1.70). Middle East, South East Asia and South America separately: same direction, not significant. Regional heterogeneity I2 7.7% (P=0.369).

Absolute risks

Crude cumulative incidence over mean 9.5 years, overall: 4.1% (<150 g/day; 2,960 events) vs 5.5% (>=450 g/day; 1,619 events); South Asia 4.8% vs 8.0%; China 2.91% vs 2.3%. No adjusted absolute risk differences reported.

Authors' conclusion

Higher white rice consumption is associated with increased risk of incident diabetes, most marked in South Asia; in other regions the association was modest (China null). Replacing highly polished white rice with healthier rice varieties or other cereals, or adding legumes, may help in rice-staple populations.

6. Strengths

• Largest prospective study on rice and diabetes; 21 countries on 5 continents with wide intake range
• Mostly documented diabetes diagnoses (90.7%); fasting glucose tested in 97.3%
• Extensive adjustment incl. wealth, education, other grains; centre random effects
• Low between-region heterogeneity

7. Weaknesses

• Diet measured once at baseline by FFQ; no rice type data (parboiled, basmati, etc.)
• Residual confounding possible (authors list unmeasured alcohol, air pollution, pesticides; South Asian high consumers differ markedly: lower BMI, fibre 10.8 vs 24.2 g/day, 71% energy from carbohydrate)
• No OGTT/HbA1c: undiagnosed diabetes and outcome misclassification possible
• Non-monotonic dose-response overall and in China; top category in 'rest of world' only 108 events
• Abstract conclusion calls other regions 'modest, nonsignificant' although pooled rest-of-world HR 1.41 (1.08-1.86) was significant (internal inconsistency)
• Table typo: South Asia top-category median given as '379 (694-1,099)'
• Egypt not included; Middle East subgroup underpowered
• Funding includes pharmaceutical, dairy and sugar organisations; one author promotes healthier rice varieties

8. Risk of bias

Orienting tool

ROBINS-E

Type of assessment

structured assessment based on the full text by a single reviewer (no formal double review)

Full formal assessment

no

Overall risk of bias

some concerns

Rationale

Overall unchanged. Simplified single-reviewer judgement from the full text; a formal ROBINS-E assessment would need two independent reviewers and the protocol.

Full-text review (25 October 2026): Added full category HRs (non-monotonic), country list (Egypt absent), outcome definition, crude cumulative incidence by category (4.1% vs 5.5%), funding (pharma, dairy, sugar association; author promoting healthier rice), lack of rice-type data, internal inconsistency in abstract. Bias overall unchanged with full ROBINS-E domains.

9. Assessment of the statistical interpretation

HR 1.20 for >=450 vs <150 g/day is a modest relative association; the dose-response is not monotonic (300-<450 g/day HR 1.25). Regional estimates are imprecise; Middle East-specific results not significant. Crude 9.5-year incidence 4.1% vs 5.5% overall.

10. Relevance to the specific claim

Contradicts unlimited rice (C-GRN-05).

11. Independent contrary findings

12. Transferability to Germany and Europe

Moderate.

13. Overall methodological judgement

Strong, large multinational cohort; supports a modest association of very high white rice intake (>3 cups cooked/day) with incident diabetes, mainly in South Asia; weak or null in China; no information on rice type.

14. Open questions

Effect at typical Egyptian intakes and with Egyptian rice varieties; parboiled vs raw vs basmati rice; whether replacing rice with other staples changes risk.

15. Linked fact checks

• Rice in all forms (including white rice) is allowed without restriction; 'rice is sugar' and two spoons of rice a day sustain life. – Studies contradicting or limiting the claim
• Basmati rice is 'wrong' (in another video: 'not good ... rubbish'); ordinary Egyptian white rice is the rice to eat. – Studies with mixed or unclear results