Moderation Is Key for Carbs in Long-term Health

Diana Phillips

August 20, 2018

New research suggests that the diet pendulum related to "optimal" carbohydrate intake, which has swung back and forth between diets based on very high and very low carbohydrate intake, belongs somewhere in the middle to improve long-term health for most people.

Data from an observational study including 15,428 people links both low- and high-carbohydrate diets to an increased risk for all-cause mortality. Moderate carbohydrate consumption is associated with the lowest mortality risk for the three diet types, Sara B. Seidelmann, MD, clinical and research fellow in cardiovascular medicine at Brigham and Women's Hospital in Boston, Massachusetts, and colleagues report in an article published online August 16 in the Lancet Public Health.

Further, a meta-analysis of carbohydrate intake studies, which the investigators updated with this and other recent research, confirms the U-shaped relationship, they report.

However, authors of an accompanying editorial say the results should be interpreted with caution, and that a single, sweeping conclusion may be too simplistic.

In fact, the authors themselves note that a slightly different picture emerges when looking at different types of low-carbohydrate diets. Whereas low-carb diets that rely heavily on animal-based proteins and fats are linked to a higher mortality risk than diets with moderate carbohydrate intake, low-carb diets that consist of primarily plant-based proteins and fats are associated with a longer life, the authors explain.

"Our findings suggest a negative long-term association between life expectancy and both low carbohydrate and high carbohydrate diets when food sources are not taken into account," the authors write. Whereas the data also provide evidence to warrant discouraging animal-based, low-carbohydrate diets, "replacement of carbohydrates with predominantly plant-based fats and proteins could be considered as a long-term approach to promote healthy ageing," they note.

Prospective Cohort Study

For the new study, the researchers analyzed data from the ongoing, prospective, observational Atherosclerosis Risk in Communities (ARIC) study of cardiovascular risk factors in four US communities. The study population included 15,428 adults, aged 45 to 64 years, who completed a modified food frequency questionnaire at enrolment between 1987 and 1989 and at the third of five subsequent follow-up visits. Participants without complete dietary information and those who reported extreme caloric intake, defined as less than 600 or more than 4200 kcal/day for men and less than 500 or more than 3600 kcal/day for women, were not included in the analysis. The researchers used a standardized nutrient database to derive nutrient intakes from the participants' food frequency questionnaires.

With a median follow-up of 25 years, there were 6283 deaths. The adjusted multivariate model showed a U-shaped association between the percentage of energy consumed from carbohydrate (mean, 48.9%; standard deviation [SD], 9.4) and mortality. The highest risk for mortality was observed in participants with the lowest carbohydrate consumption, the authors report, whereas the lowest was seen in individuals in 50% to 55% carbohydrate intake.

Table. Carbohydrate Intake and All-Cause Mortality, ARIC Cohort

Carbohydrate Intake, % energy <30% 30%-40% 40%-50% 50%-55% 55%-65% >65%
N 315 2242 6097 3026 3034 714
Mortality, N (%) 163 (51.7%) 986 (44.0%) 2533 (41.5%) 1162 (38.4%) 1150 (37.9%) 289 (40.5%)
Hazard ratio, fully adjusted (95% confidence interval)* 1.37 (1.16-1.63) 1.37 (1.16-1.63) 1.11 (1.03-1.19) Reference 1.01 (0.93-1.10) 1.16 (1.02-1.33)
Model adjusted for age, race, sex, ARIC test center, total energy consumption, diabetes, cigarette smoking, physical activity, income level and education. Table adapted from Seidelmann et al.

Meta-analysis Reveals Similar U-shaped Curve

To contextualize the findings, the researchers expanded a meta-analysis previously published in 2012 to include the current study, along with two additional studies. In total, the meta-analysis included 432,179 participants from eight cohort studies investigating carbohydrate intake, with 40,181 deaths.

As in the ARIC study, compared with moderate carbohydrate consumption, the risk for all-cause mortality in the pooled population was significantly greater among participants with low carbohydrate consumption, defined as less than 40% of energy intake (hazard ratio, 1.20; 95% confidence interval, 1.09 - 1.32; P < .0001), and those with high carbohydrate consumption, defined as more than 70% of energy intake (hazard ratio, 1.23; 95% confidence interval, 1.11 - 1.36; P < .0001).

In addition, the meta-analysis showed that mortality increased when animal-derived fat and protein were substituted for carbohydrate, and decreased when these substitutions were plant-based. "These data suggest that the source of the protein and fat substituted for carbohydrates in the diet might notably modify the relationship between carbohydrate intake and mortality," the authors write.

Unlike previous analyses of carbohydrate intake, which searched for trends within quantiles of carbohydrate consumption, the current investigation relies on continuous carbohydrate intake data. This approach, the authors note, "provides more granular information and allowed us to identify a more U-shaped relationship between carbohydrate consumption and risk, which might otherwise not have been evident."

Reminder to Avoid Extremes

In an accompanying commentary, Andrew Mente, MD, and Salim Yusuf, MD, from the Population Health Research Institute, Hamilton Health Sciences and McMaster University, Ontario, Canada, suggest the study design limits the generalizability of the findings.

"Such differences in risk associated with extreme differences in intake of a nutrient are plausible, but observational studies cannot completely exclude residual confounders when the apparent differences are so modest," they write.

The editorialists also stress that the meta-analyses results should be interpreted with caution in light of the potential "group think" bias that can occur when researchers choose analytical approaches to produce findings that align with a preconceived outcome.

"These analyses are based on models that the investigators created, meaning they are subject to their assumptions," Yusuf said in an interview with Medscape Medical News. "In order for the findings to be clinically useful or meaningful, there has to be better data and more transparent analyses, and the findings have to be replicated by other groups."

In this regard, Mente and Yusuf advocate for a collaboration among the investigators involved with the large studies included in the meta-analysis to transparently pool the individual data. In addition, future observational studies should use new methods, such as triangulation, "to assess whether there is a coherent pattern of information about the links between consumption of a nutrient such as carbohydrates with a panel of physiological or nutritional biomarkers and clinical outcomes," they write.

Although the study findings do indicate that there may be a sweet spot of carb consumption around 50% of energy, "that conclusion is oversimplistic," Yusuf said. "Different foods that are high in carbs, such as legumes, appear protective, whereas others, such as polished rice, in high amounts may not be."

Although Yusuf says the findings do not have any specific clinical implications "at the moment," the report offers a reminder about dietary extremes: "People should not rush to adopt extremes in diet of any kind. Neither extremely low carbs, low fats, or low salt," he said. "The nutrition field is full of extremes, often without the evidence to back it up."

Some of the study authors disclose financial relationships with the California Walnut Commission and Dairy Management Inc, Novartis, and Zogenix. The editorialists have disclosed no relevant financial relationships.

Lancet Public Health. Published online August 16, 2018. Article full text, Commentary full text

For more news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.