Associations of fats and carbohydrate intake with cardiovascular disease and mortality in 18 countries from five continents (PURE): a prospective cohort study
Mahshid Dehghan, Andrew Mente, Xiaohe Zhang, Sumathi Swaminathan, Wei Li, Viswanathan Mohan, Romaina Iqbal, Rajesh Kumar, Edelweiss Wentzel-Viljoen, Annika Rosengren, Leela Itty Amma, Alvaro Avezum, Jephat Chifamba, Rafael Diaz, Rasha Khatib, Scott Lear, Patricio Lopez-Jaramillo, Xiaoyun Liu, Rajeev Gupta, Noushin Mohammadifard, Nan Gao, Aytekin Oguz, Anis Safura Ramli, Pamela Seron, Yi Sun, Andrzej Szuba, Lungiswa Tsolekile, Andreas Wielgosz, Rita Yusuf, Afzal Hussein Yusufali, Koon K Teo, Sumathy Rangarajan, Gilles Dagenais, Shrikant I Bangdiwala, Shofiqul Islam, Sonia S Anand, Salim Yusuf - Lancet, 2017
In this large international prospective cohort, higher carbohydrate intake was associated with higher total mortality, while higher total fat and each type of fat (including saturated fat) were associated with lower total mortality. Total fat, saturated fat, and unsaturated fats showed no significant association with myocardial infarction or cardiovascular mortality, and saturated fat was inversely associated with stroke. The authors concluded that global dietary guidelines should be reconsidered.
Key findings
- Higher carbohydrate intake was associated with higher total mortality (Q5 vs Q1: HR 1.28; 95% CI 1.12-1.46; p-trend = 0.0001), but not with cardiovascular disease or cardiovascular mortality.
- Higher total fat intake was associated with lower total mortality (Q5 vs Q1: HR 0.77; 95% CI 0.67-0.87).
- Higher saturated fat intake was associated with lower total mortality (Q5 vs Q1: HR 0.86; 95% CI 0.76-0.99) and lower risk of stroke (HR 0.79; 95% CI 0.64-0.98).
- Total fat and saturated and unsaturated fats were not significantly associated with risk of myocardial infarction or cardiovascular mortality.
Effect measures
- Hazard Ratio: 1.2895% CI 1.12-1.46
- Hazard Ratio: 0.7795% CI 0.67-0.87
- Hazard Ratio: 0.8695% CI 0.76-0.99
- Hazard Ratio: 0.7995% CI 0.64-0.98
Why this evidence tier (Moderate)
- Risk of bias:
- Large, well-conducted prospective cohort, but observational: subject to confounding by overall diet quality, socioeconomic factors, and reverse causation, especially given the very wide range of countries and food environments.
- Precision:
- Very large sample (135,335) yields precise estimates; several intervals are narrow and exclude the null.
- Directness:
- Directly measures macronutrient intake against mortality and cardiovascular events, but the lowest-carbohydrate populations differ from low-carbohydrate diets studied in trials, limiting transfer to dietary advice.
- Consistency:
- Carbohydrate-mortality and fat-mortality directions are consistent across analyses, though they sit in tension with some Western-population cohorts and with replacement RCTs.
- Funding / COI:
- The verified primary source is the result, not the full funding statement: PURE is coordinated by the Population Health Research Institute (Hamilton, Canada) with Canadian public and institutional support. The complete funding statement is in the paywalled full text and was not verified in this session, so no specific industry-sponsor claim is asserted here; funding details are marked pending verification.
Moderate certainty. A large, precise cohort whose findings challenge low-fat guidance, tempered by observational confounding; the full funding statement was not verified this session.
- Population:
- 135,335 individuals aged 35-70 years enrolled 2003-2013 across 18 countries on five continents (PURE cohort); median follow-up 7.4 years; dietary intake by validated food frequency questionnaires.
- Conflicts of interest:
- Coordinated by the Population Health Research Institute with Canadian public and institutional funding. The complete funding and disclosure statement is in the paywalled full text and was not verified in this session; no specific industry-sponsor claim is asserted (pending verification).
- Funding:
- Coordinated by the Population Health Research Institute (Hamilton, Canada) with Canadian public and institutional support; the complete funding statement is in the paywalled full text and is marked pending verification.
Limitations
- Observational cohort; associations are subject to confounding and reverse causation and cannot establish causation.
- Diet measured by food frequency questionnaire at baseline, with measurement error and no capture of dietary change over follow-up.
- Very wide range of countries and food environments; the highest-carbohydrate populations may differ systematically in ways residual adjustment cannot capture.
- The complete funding and disclosure statement is in the paywalled full text and was not verified in this session (marked pending).