Meta-analysiscited by the Tayebat archivefull text checked A-SOD-08 in-depth review
Sugary drinks, diet drinks, juice and diabetes
Consumption of sugar sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: systematic review, meta-analysis, and estimation of population attributable fraction.
Overview
Imamura F et al.. BMJ 2015 · PMID 26199070 · DOI 10.1136/bmj.h3576
Each daily sugary drink was linked with 13-18% more diabetes; diet drinks and juice also showed (possibly biased) links.
Does the study support the specific claim? Yes · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Are SSB, ASB and fruit juice associated with T2D independently of adiposity?
- How was the study done?
Systematic review and random-effects meta-analysis of prospective cohorts; PAF estimation (US, UK).
- How many people, animals or samples were studied?
17 cohorts; 38,253 cases; 10,126,754 person-years; Adults without diabetes
- What was actually found?
Per serving/day: SSB +18% (9-28%; I2 89%) and +13% (6-21%) after adiposity adjustment; ASB +25% and +8% (publication bias and residual confounding indicated); fruit juice +5% (-1 to 11%) and +7% (1-14%). PAF: 8.7% USA, 3.6% UK.
- How large was the observed effect?
+18% per serving (+13% adiposity-adjusted)
- What do the results mean in everyday life?
Neither diet drinks nor juice are proven healthier alternatives to sugary drinks.
- What are the study's limitations?
High heterogeneity; observational studies; protocol not registered; ASB and juice results likely biased
- What does it say about Elawady's specific claim?
Supports C-SOD-01; mixed for C-SOD-10; limits C-SUG-06 (juice association).
- Do the results apply to Germany and Europe?
High (UK PAF estimated).
- What can explicitly not be concluded from this study?
Causation; that diet drinks or juice cause diabetes; diet-drink effects on mortality
Scientific in-depth review
1. Full reference
- Authors
- Imamura F, O'Connor L, Ye Z, Mursu J, Hayashino Y, Bhupathiraju SN, Forouhi NG
- Journal
- BMJ
- Year
- 2015
- Identifier
- PMID 26199070 DOI 10.1136/bmj.h3576 · PMC4510779
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Medical Research Council Epidemiology Unit Core Support (MC_UU_12015/5); one author AHA postdoctoral fellowship. Authors declare no competing interests; sponsor had no role. No included study was industry funded.
2. Research question
Are SSB, ASB and fruit juice associated with T2D independently of adiposity?
3. Study design and population
- Study design
- Meta-analysis – Systematic review (PRISMA; protocol not registered, available on request) and random-effects dose-response meta-analysis of prospective cohorts, with Cochrane-based risk-of-bias assessment, GRADE, calibration for measurement error and publication bias, and population attributable fraction estimation (USA, UK)
- Population
Adults without diabetes
- Country
Multinational; PAF for USA and UK
- Period
Searches 31 May 2013, updated 10 February 2014 (PubMed, Embase, Ovid, Web of Knowledge; no language or date limits); PAF for 2010-2020
- Sample size
17 cohorts (21 articles of 16 cohorts plus 1 unpublished); 38,253 cases over 10,126,754 person-years; PAF based on 4729 US (NHANES 2009-10) and 1932 UK (NDNS 2008-12) adults
- Inclusion and exclusion criteria
Prospective studies of adults >=18 years free of diabetes, assessing beverage consumption and incident T2D, mean follow-up >=2 years
4. Methods and endpoints
- Exposure / intervention
Per one serving/day (also per 250 mL/day) of SSB, ASB and 100% fruit juice
- Comparator
Lower intake (dose-response per serving/day)
- Endpoints
Incident T2D
- Statistical methods
Random-effects meta-analysis (prespecified); I2; estimates before/after adiposity adjustment and crude; calibration for within-person dietary variation and self-reported diabetes; meta-regression; Egger test, contour-enhanced funnel plots, trim and fill; influence analysis; simulation of residual confounding by imperfectly measured adiposity; GRADE; PAF with validated risk algorithms assuming causality
5. Actual results
- Results
Per serving/day, before/after adiposity adjustment: SSB RR 1.18 (95% CI 1.09-1.28; I2 89%) and 1.13 (1.06-1.21; I2 79%); ASB 1.25 (1.18-1.33; I2 70%) and 1.08 (1.02-1.15; I2 64%); fruit juice 1.05 (0.99-1.11) and 1.07 (1.01-1.14). Adjustment for confounders attenuated SSB from 1.25 to 1.18 and ASB from 1.48 to 1.25. Calibration for within-person variation strengthened SSB to 1.28 (1.12-1.46). ASB after publication-bias calibration 1.22 (0.98-1.52; n.s.). Residual-confounding simulation (r=0.80 for adiposity): SSB 1.20 (1.04-1.38), ASB 1.01 (0.81-1.25), fruit juice 1.12 (1.03-1.22). Fruit juice null in studies with objective T2D ascertainment (1.08, 0.97-1.20) and repeated diet measures (0.98, 0.86-1.11). GRADE: SSB moderate; ASB and fruit juice low. PAF (adiposity-adjusted): USA 8.7% (3.9-12.9%), UK 3.6% (1.7-5.6%); adiposity-unadjusted 11.9% and 4.9%. No study was industry funded.
- Absolute risks
Predicted 10-year T2D event rate 11.0% (USA, 20.9 million events) and 5.8% (UK, 2.6 million); attributable to SSB under causal assumption: 1.8 million (USA) and 79,000 (UK) adiposity-adjusted, 2.6 million and 126,000 unadjusted.
- Authors' conclusion
Habitual SSB consumption associated with T2D independently of adiposity; ASB and fruit juice associations likely biased but neither are healthy alternatives.
6. Strengths
- Four databases, no language limits, duplicate review and extraction
- Formal risk-of-bias assessment (Cochrane NRSI-based, 7 domains) and GRADE
- Adiposity-adjusted and unadjusted estimates; measurement-error calibration
- Extensive sensitivity, publication-bias and residual-confounding analyses
- PAF estimation with national data
7. Weaknesses
- Protocol not registered
- High heterogeneity for SSB (I2 79-89%)
- Observational cohorts; residual confounding and reverse causation, especially for ASB (authors)
- Dose-response partly derived from categorical estimates and ORs
- PAF assumes causality and no change in lifestyle
- Eligible cohorts excluded for lack of information
- Bias assessment summarised by one author then discussed (subjectivity acknowledged)
8. Risk of bias
- Orienting tool
- AMSTAR 2
- 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
- low
- Rationale
High-quality review; only non-critical weakness in protocol registration. Confidence in SSB estimate moderate (GRADE), ASB/fruit juice low. Overlaps with A-SOD-01 primary studies (not independent confirmation). Structured single-reviewer judgement from full text; a formal assessment would require two independent reviewers and the protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| Item 2 protocol established before review (critical) | some concerns | Protocol exists (available on request) but not registered. |
| Item 4 comprehensive search (critical) | low | Four databases plus hand search, no language/date limits, author contact for unpublished data. |
| Item 5/6 duplicate selection and extraction | low | Full texts reviewed in duplicate; standardised duplicate extraction. |
| Item 7 excluded studies justified (critical) | low | Supplementary table S1 lists cohorts not meeting criteria. |
| Item 9 risk of bias of included studies (critical) | low | Cochrane NRSI-based tool with seven domains; studies with potential bias identified. |
| Item 11 appropriate meta-analytic methods (critical) | low | Random effects, dose-response, calibration, meta-regression. |
| Item 13 RoB considered in interpretation (critical) | low | GRADE ratings and explicit downgrading of ASB and fruit-juice evidence. |
| Item 15 publication bias (critical) | low | Egger, contour-enhanced funnel and trim-and-fill with calibrated estimates. |
| Item 16 COI | low | Funding and COI reported; industry funding of primary studies checked. |
Full-text review (25 October 2026): Added search/selection and RoB methods, confounder attenuation, measurement-error calibration (SSB 1.28), ASB after publication-bias calibration (1.22, n.s.) and residual-confounding simulation (1.01), fruit-juice subgroup null findings, GRADE ratings, PAF details, funding; AMSTAR 2 items added, overall unchanged (low).
9. Assessment of the statistical interpretation
SSB: robust, moderate-certainty association of about 13-18% per daily serving (up to 28% after measurement-error calibration); PAF of 4-13% (USA) under causal assumptions. ASB: association largely explained by residual confounding/publication bias (1.01 in simulation; 1.22, n.s., after publication-bias calibration). Fruit juice: unstable, null in objectively ascertained studies.
10. Relevance to the specific claim
Supports C-SOD-01; mixed for C-SOD-10; limits C-SUG-06 (juice association).
11. Independent contrary findings
None.
12. Transferability to Germany and Europe
High (UK PAF estimated).
13. Overall methodological judgement
High-quality meta-analysis; strong support for the SSB-diabetes association (moderate certainty) and weak, probably biased evidence for ASB and fruit juice.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | low |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | taken out of context |
14. Open questions
Causality for ASB and juice.
15. Linked fact checks
- Regular consumption of sugar-sweetened soft drinks increases the risk of type 2 diabetes (and metabolic syndrome); one to two cans per day raise type 2 diabetes risk by about 26%, and each daily serving by about 18%. – Studies supporting the claim
- Diet soda is not a safe alternative to sugary soda: artificially sweetened drinks are linked to higher mortality, especially circulatory deaths. – Studies with mixed or unclear results
- Implicit in the system's lists: sugar-cane juice, canned juices, honey, jam and sweets are acceptable while sugar-sweetened soft drinks are harmful. – 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
(Narrative) A newer BMJ meta-analysis 'raised the number' to an 18% jump in diabetes risk for every daily serving.
- A. What was actually studied?
SSB, ASB, juice and T2D.
- B. What did the authors find?
SSB +18% per serving (13% adjusted for adiposity); ASB and juice associations likely biased.
- C. What conclusions do the authors themselves draw?
SSB associated with T2D; ASB and juice not healthy alternatives.
- D. What does the archive derive from it?
Presents 18% per serving as an increase over the earlier 26% estimate.
- E. Is this derivation scientifically justified?
Not justified as framed: 18% per serving and 26% highest vs lowest (A-SOD-01) are different metrics, not an increase; the archive omits the adiposity-adjusted 13%, the low-certainty and probably biased ASB result, and the fruit-juice findings.
- Adequacy of the interpretation (by the archive)
- taken out of context