Cohort studycited by the Tayebat archivefull text checked A-SOD-16 in-depth review
Added sugar and heart deaths in the US
Added sugar intake and cardiovascular diseases mortality among US adults.
Overview
Yang Q et al.. JAMA Intern Med 2014 · PMID 24493081 · DOI 10.1001/jamainternmed.2013.13563
People getting 25% or more of calories from added sugar had nearly three times the CVD death risk of those under 10%.
Does the study support the specific claim? Yes · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Is % energy from added sugar associated with CVD mortality?
- How was the study done?
NHANES trend analysis and NHANES III linked mortality cohort.
- How many people, animals or samples were studied?
31,147 (trends); 11,733 (cohort); 831 CVD deaths; US adults (nationally representative)
- What was actually found?
Fully adjusted HRs across quintiles 1.00, 1.07, 1.18, 1.38, 2.03 (1.26-3.27); 10-24.9% vs <10%: HR 1.30 (1.09-1.55); >=25%: 2.75 (1.40-5.42). Most adults (71.4%) consumed >=10% from added sugar.
- How large was the observed effect?
Highest vs lowest fifth of added-sugar intake: HR 2.03 (1.26-3.27); >=25% vs <10% of calories: HR 2.75 (1.40-5.42); about 1 extra CVD death per 22 people over 15 years in the highest fifth.
- What do the results mean in everyday life?
Total added sugar matters, from any source.
- What are the study's limitations?
Observational; diet measured once by 24-hour recall at baseline; possible residual confounding; no association seen in non-Hispanic black participants; total mortality not significantly raised.
- What does it say about Elawady's specific claim?
- Do the results apply to Germany and Europe?
Moderate.
- What can explicitly not be concluded from this study?
Causation.
Scientific in-depth review
1. Full reference
- Authors
- Yang Q, Zhang Z, Gregg EW, Flanders WD, Merritt R, Hu FB
- Journal
- JAMA Intern Med
- Year
- 2014
- Identifier
- PMID 24493081 DOI 10.1001/jamainternmed.2013.13563 · PMC10910551
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Conflict of interest: none reported. Authors at CDC (disclaimer: views not official CDC position); no separate funding statement in the full text.
2. Research question
Is % energy from added sugar associated with CVD mortality?
3. Study design and population
- Study design
- Cohort study – Two components: serial cross-sectional NHANES trend analysis (1988-1994, 1999-2004, 2005-2010) and a prospective cohort analysis of NHANES III (1988-1994) linked to the National Death Index through 2006.
- Population
US adults (nationally representative)
- Country
USA
- Period
NHANES 1988-2010 (trends); cohort baseline 1988-1994, mortality follow-up to 2006 (median 14.6 years)
- Sample size
Trends: 31,147 (11,733 + 8,786 + 10,628); cohort: 11,733 adults, 831 CVD deaths in 163,039 person-years
- Inclusion and exclusion criteria
Non-pregnant adults >=20 years; excluded incomplete first-day 24-h recall, history of myocardial infarction, stroke or heart failure, diagnosed diabetes, cancer at baseline, BMI <18.5, missing covariates.
4. Methods and endpoints
- Exposure / intervention
Usual % of daily calories from added sugar (24-h recalls; MyPyramid Equivalents Database; NCI measurement-error method with regression calibration); secondary: sugar-sweetened beverage servings/week
- Comparator
Lowest quintile (midvalue 7.4% of calories) or <10% of calories
- Endpoints
CVD mortality
- Statistical methods
Survey-weighted Cox models with Box-Cox-transformed usual intake (NCI method); HRs at quintile midvalues; adjusted for age, sex, race/ethnicity, education, smoking, alcohol, physical activity, antihypertensive medication, family history of CVD, Healthy Eating Index, BMI, systolic BP, total cholesterol, total calories; number needed to harm at 15 years with bootstrap CIs; stratified and sensitivity analyses.
5. Actual results
- Results
Mean % calories from added sugar 15.7% (1988-1994), 16.8% (1999-2004), 14.9% (2005-2010); 71.4% consumed >=10% in 2005-2010. Fully adjusted HRs for CVD mortality by quintile (midvalues 7.4%, 11.4%, 14.8%, 18.7%, 25.2%): 1.00, 1.07 (1.02-1.12), 1.18 (1.06-1.31), 1.38 (1.11-1.70), 2.03 (1.26-3.27); P=.004. 10.0-24.9% vs <10%: HR 1.30 (1.09-1.55); >=25%: HR 2.75 (1.40-5.42). SSB >=7 servings/week vs <=1/week: HR 1.29 (1.04-1.60). Total mortality, highest quintile: HR 1.28 (0.94-1.75), not significant. No association among non-Hispanic black participants (interaction P=.09).
- Absolute risks
831 CVD deaths in 163,039 person-years. Adjusted number needed to harm at 15 years versus Q1: 265 (166-715) for Q2, 109 (67-297) for Q3, 53 (33-152) for Q4, 22 (13-66) for Q5.
- Authors' conclusion
Most US adults consume more added sugar than recommended; a higher percentage of calories from added sugar is associated with significantly increased CVD mortality, and regular SSB consumption is associated with elevated CVD mortality; results support recommendations to limit added sugars. Authors state observational results should not be directly interpreted as causal.
6. Strengths
- Nationally representative sample
- Measurement-error correction (NCI usual-intake method)
- Mortality via National Death Index linkage
- Broad confounder adjustment including diet quality (HEI); results robust to HEI components and sodium
- Absolute effect reported as number needed to harm
7. Weaknesses
- Single baseline exposure; not updated over up to 18 years (authors)
- Only ~8% of NHANES III participants had a second 24-h recall for the usual-intake model
- Residual and unmeasured confounding acknowledged by authors
- Adjustment for BMI, blood pressure and cholesterol (possible mediators) may bias estimates
- Highest-quintile estimate imprecise (CI 1.26-3.27) and depends on modelling assumptions (Box-Cox, midvalue contrasts)
- No association among non-Hispanic black participants
- Total mortality association not significant
- Self-reported dietary intake
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
Structured single-reviewer judgement from the full text; a complete formal ROBINS-E assessment requires two independent reviewers and a protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| Confounding | some concerns | Extensive adjustment incl. HEI, smoking, activity; residual confounding possible; high-sugar consumers were younger, less physically active and had lower diet quality. |
| Exposure measurement | some concerns | 24-h recalls with NCI correction, but few NHANES III participants had repeat recalls; added-sugar database matching; single baseline measure. |
| Selection of participants | low | Representative sample; exclusions of prevalent disease reduce reverse causation; complete-case exclusions modest (905 for missing covariates); sensitivity analysis with n=14,338 showed moderate attenuation. |
| Post-exposure interventions | some concerns | Diet changes over follow-up not captured. |
| Missing data | low | Mortality linkage nearly complete; probabilistic matching. |
| Outcome measurement | low | Death-certificate-based CVD mortality (ICD-10 I00-I78); some misclassification of cause possible. |
| Selection of reported result | some concerns | No pre-registered protocol; non-linear modelling choices; but multiple cut-points and sensitivity analyses reported. |
Full-text review (25 October 2026): Numbers confirmed. Added inclusion/exclusion criteria, NCI measurement-error method and full adjustment set, number needed to harm (265 to 22 at 15 years), supplementary SSB result HR 1.29 (1.04-1.60), non-significant total mortality HR 1.28 (0.94-1.75), null result in non-Hispanic black participants, and authors' caution on causality. Full domain-level ROBINS-E (overall some_concerns, unchanged).
9. Assessment of the statistical interpretation
Clear, monotonic dose-response with modest HRs at middle quintiles and a large but imprecise HR at the top. Absolute differences are meaningful only at high intake (NNH 22 over 15 years for ~25% vs ~7.4% of calories). Possible over-adjustment for mediators would bias toward the null rather than inflate the association.
10. Relevance to the specific claim
Supports C-SOD-03; contradicts C-SUG-01.
11. Independent contrary findings
None.
12. Transferability to Germany and Europe
Moderate.
13. Overall methodological judgement
Relevant to total added sugar.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Causality; role of SSBs versus other sources of added sugar; why no association among non-Hispanic black adults.
15. Linked fact checks
- Sugar-sweetened soft drinks raise the risk of coronary heart disease, cardiovascular death and early death (e.g., two or more daily servings: +35% coronary heart disease in women; +31% cardiovascular and +21% total mortality). – Studies supporting the claim
- White sugar is the greatest product humans ever made; any amount of sugar per day is beneficial, and one can eat a kilogram of white sugar or honey. – Studies contradicting or limiting the claim
Use by the Tayebat archive
Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.
- What the archive says about the study
Listed in the archive bibliography only (no archive interpretation card).
- A. What was actually studied?
Added sugar and CVD mortality.
- B. What did the authors find?
Dose-response association.
- C. What conclusions do the authors themselves draw?
As above.
- D. What does the archive derive from it?
Supports soda harm.
- E. Is this derivation scientifically justified?
Consistent in direction. It concerns total added sugar (all sources, including sweets and desserts), not only soda; the SSB-specific result (HR 1.29 for >=7 servings/week) is a supplementary analysis. Association, not proof of causation (authors' own caveat).
- Adequacy of the interpretation (by the archive)
- not applicable