ESC 2018 - Munich

Low-carbohydrate diets and all-cause and cause-specific mortality: A population-based cohort study and pooling prospective studies

Presented at the ESC congress 2018 by: Mohsen Mazidi (Sweden)

Introduction and methods

Obesity is a major health problem worldwide, raising the risk for several chronic diseases, including cardiovascular disease (CVD), hypertension, type 2 diabetes mellitus, and cancer. Different diets have been suggested to reduce weight, such as low-carbohydrates diets (LCD), which have become popular in the last 20 years. However, the long-term safety of LCDs is controversial, and some available studies suggest that LCD may increase the risk of CVD, cancer and death. Therefore, this cohort study evaluated the link between LCD and mortality in a large US cohort, after which the results were validated by performing a meta-analysis of available prospective studies that examined this association.

Data were used of the US National Health and Nutrition Examination Surveys (NHANES) study (1999-2010), which included 24,825 participants (mean age of 47.6 years) who were followed for a mean of 6.4 years to examine the energy intakes of carbohydrates for each individual. Subgroups of participants were made by subdividing the communities into quartiles of carbohydrate consumption and association between mortality and LCD was evaluated for obese (BMI ≥30kg/m2) and non-obese patients (BMI <30kg/m2). Levels of carbohydrate intake of the respective quartiles were not provided.

The meta-analysis included seven prospective cohort studies with 447,506 participants (mean follow-up 15.6 years) published up to February 1, 2018.

The primary outcome was all-cause mortality and cause-specific mortality, including CVD, stroke and cancer.

Main results

NHANES data

• After full adjustment, participants in Q4 of LCD had highest risk of total (HR: 1.32, 95%CI: 1.14-2.01, P<0.001), cardiovascular (HR: 1.51, 95%CI: 1.19-1.91, P<001), cerebrovascular (HR: 1.50, 95%CI: 1.12-2.31, P<001) and cancer mortality (HR: 1.35, 95%CI: 1.06-1.69, P<0.001).
• After full adjustment, the association between LCD and total mortality was stronger in obese (Q4 HR: 1.19, 95%CI: 1.11-2.25) than non-obese (Q4 HR: 1.48, 95%CI: 1.37-2.01) participants and in participants ≥55 years (Q4 HR: 1.52, 95%CI: 1.41-1.79, P<0.001), compared to younger participants (HR: 1.17, 95%CI: 1.10-2.01).

Meta-analysis

• A positive association was observed between LCD and total (RR: 1.15, 95%CI: 1.07-1.22, p<0.001), CVD (RR: 1.13, 95%CI: 1.02-1.24, p<0.001) and cancer mortality (RR: 1.08, 95%CI: 1.01-1.14, P=0.02).

Conclusion

Both individual data and pooled prospective cohort studies on LCD showed increased risk of total and cause-specific mortality in adults, suggesting that LCD should not be recommended. Further studies are needed to assess the effect of diet duration and to clarify underlying mechanisms involved in these associations, which can help policy makers to increase public awareness about the role of LCDs, and their potential harmful effects.

Discussion

The dominant questions during the discussion were about the exact amount of carbohydrates in the low-carb-diet, as well as the total amount of calories in each diet. These data were not available. It was postulated that the increased mortality in the LCD group might be explained by the fact that the missing carbohydrates were replaced by more dietary fat .

- Our reporting is based on the information provided at the ESC congress -