Consumption of ultra-processed foods associated with weight gain and obesity in adults: A multi-national cohort study

Summary

Background

There is a worldwide shift towards increased consumption of ultra-processed foods (UPF) with concurrent rising prevalence of obesity. We examined the relationship between the consumption of UPF and weight gain and risk of obesity.

Methods

This prospective cohort included 348 748 men and women aged 25–70 years. Participants were recruited between 1992 and 2000 from 9 European countries in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Two body weight measures were available, at baseline and after a median follow-up time of 5 years. Foods and drinks were assessed at baseline by dietary questionnaires and classified according to their degree of processing using NOVA classification. Multilevel mixed linear regression was used to estimate the association between UPF consumption and body weight change (kg/5 years). To estimate the relative risk of becoming overweight or obese after 5 years we used Poisson regression stratified according to baseline body mass index (BMI).

Results

After multivariable adjustment, higher UPF consumption (per 1 SD increment) was positively associated with weight gain (0·12 kg/5 years, 95% CI 0·09 to 0·15). Comparing highest vs. lowest quintile of UPF consumption was associated with a 15% greater risk (95% CI 1·11, 1·19) of becoming overweight or obese in normal weight participants, and with a 16% greater risk (95% CI 1·09, 1·23) of becoming obese in participants who were overweight at baseline.

Conclusions

These results are supportive of public health campaigns to substitute UPF for less processed alternatives for obesity prevention and weight management.

Introduction

In 2016, more than 39% of the world population was affected by overweight or obesity (body mass index, BMI ≥25 kg/m²) and it is projected that the prevalence of obesity will increase further in the years to come [1]. Obesity is defined as a state of excess body fatness and is the consequence of a sustained positive energy balance [2]. Dietary factors are among the many factors that can contribute towards an energy imbalance. Several characteristics of foods and drinks are known to influence the amount consumed, including energy density and portion size [2]. Continued efforts are needed to identify additional modifiable factors for the prevention of weight gain and obesity [3].

Globally, the consumption of industrially processed foods, so-called ultra-processed foods (UPFs), increased in the last few decades [4] representing nowadays 50%–60% of daily energy intake in some high-income countries [[5], [6], [7]]. In contrast to fresh or minimally processed foods, UPFs tend to have a higher energy density [8], and they can trigger a higher eating rate/energy intake rate [9]. These properties may result in energy overconsumption and weight gain when consuming a diet with a large proportion of UPFs [8,9].

UPFs are defined by the NOVA food system classification as products formulated mostly or entirely from food constituents, not found in home cooking, and culinary ingredients such as fat, sugar, and salt [10]. During manufacturing they undergo physical and chemical processes, such as extruding, prefrying, or hydrogenation [11]. Typically, UPFs are mass-produced packaged breads, sugared breakfast cereals, buns, biscuits, sweet or savoury packaged snacks, instant soups and noodles, processed meat as well as certain industrially pre-prepared meals [12]. These foods provide for many people, particularly in urban areas or for those with extensive or unusual working hours, easily accessible and affordable sources of energy [13].

Several cross-sectional and four prospective cohort studies, using NOVA, suggested positive associations between a higher consumption of UPFs and excess body weight [[14], [15], [16], [17], [18], [19]]. In a meta-analysis of 13 cross-sectional and one prospective observational study, average positive associations between UPF consumption and overweight or obesity were reported [14]. However, there was evidence for publication bias and heterogeneity across studies was substantial (I [2]≥85%), which hampers the possibility to draw robust conclusions. Different study designs and different level of adjustment for confounders may at least partly explain the large heterogeneity in previous studies. Further evaluation whether UPFs promote energy overconsumption and weight gain in diverse populations is therefore warranted.

We investigated relationships between UPF consumption and weight change among adults in a multi-national setting, which allowed assessment of potential heterogeneity across study populations with different underlying dietary habits, while applying a uniform adjustment for confounders. We also investigated associations with risk of developing overweight or obesity.